Conservation and Restoration of Glass

Conservation and Restoration of Glass

Butterworth-Heinemann Series in Conservation and Museology

Series Editors: Arts and ArchaeologyAndrew OddyBritish Museum, London

ArchitectureDerek LinstrumFormerly Institute of Advanced Architectural Studies, University of York

US Executive Editor: Norbert S. BaerNew York University, Conservation Center of the Institute of Fine Arts

Consultants: Sir Bernard FeildenDavid BomfordNational Gallery, LondonC.V. HorieManchester Museum, University of ManchesterSarah StaniforthNational Trust, LondonJohn WarrenInstitute of Advanced Architectural Studies, University of York

Published titles: Artists’ Pigments c.1600–1835, 2nd Edition (Harley)Care and Conservation of Geological Material (Howie)Care and Conservation of Palaeontological Material (Collins)Chemical Principles of Textile Conservation (Tímár-Balázsy, Eastop)Conservation and Exhibitions (Stolow)Conservation and Restoration of Ceramics (Buys, Oakley)Conservation and Restoration of Works of Art and Antiquities (Kühn)Conservation of Brick (Warren)Conservation of Building and Decorative Stone (Ashurst, Dimes)Conservation of Earth Structures (Warren)Conservation of Glass (Newton, Davison)Conservation of Historic Buildings (Feilden)Conservation of Historic Timber Structures: An Ecological Approach to

Preservation (Larsen, Marstein)Conservation of Library and Archive Materials and the Graphic Arts

(Petherbridge)Conservation of Manuscripts and Painting of South-east Asia (Agrawal)Conservation of Marine Archaeological Objects (Pearson)Conservation of Wall Paintings (Mora, Mora, Philippot)Historic Floors: Their History and Conservation (Fawcett)A History of Architectural Conservation (Jokilehto)Lacquer: Technology and Conservation (Webb)The Museum Environment, 2nd Edition (Thomson)The Organic Chemistry of Museum Objects, 2nd Edition (Mills, White)Radiography of Cultural Material (Lang, Middleton)The Textile Conservator’s Manual, 2nd Edition (Landi)Upholstery Conservation: Principles and Practice (Gill, Eastop)

Related titles: Concerning Buildings (Marks)Laser Cleaning in Conservation (Cooper)Lighting Historic Buildings (Phillips)Manual of Curatorship, 2nd edition (Thompson)Manual of Heritage Management (Harrison)Materials for Conservation (Horie)Metal Plating and Patination (Niece, Craddock)Museum Documentation Systems (Light)Risk Assessment for Object Conservation (Ashley-Smith)Touring Exhibitions (Sixsmith)

Conservation and Restoration ofGlass

Sandra Davison FIIC, ACR

Glass ConservatorThe Conservation Studio, Thame, Oxfordshire

OXFORD AMSTERDAM BOSTON LONDON NEW YORK PARISSAN DIEGO SAN FRANCISCO SINGAPORE SYDNEY TOKYO

Butterworth-Heinemann is an imprint of ElsevierLinacre House, Jordan Hill, Oxford OX2 8DP, UK30 Corporate Drive, Suite 400, Burlington, MA 01803, USA

First edition 1989Paperback edition 1996Second edition 2003Reprinted 2006

Copyright © 1993, 2003, Elsevier Ltd. All rights reserved

No part of this publication may be reproduced, stored in a retrieval systemor transmitted in any form or by any means electronic, mechanical, photocopying,recording or otherwise without the prior written permission of the publisher

Permissions may be sought directly from Elsevier's Science & Technology RightsDepartment in Oxford, UK: phone: (�44) (0) 1865 843830; fax: (�44) (0) 1865 853333;email: [email protected]. Alternatively you can submit your request online byvisiting the Elsevier web site at http://elsevier.com/locate/permissions, and selectingObtaining permission to use Elsevier material

NoticeNo responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made

British Library Cataloguing in Publication DataA catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication DataA catalog record for this book is available from the Library of Congress

ISBN–13: 978-0-7506-4341-2ISBN–10: 0-7506-4341-2

Printed and bound in Great Britain

06 07 08 09 10 10 9 8 7 6 5 4 3 2

For information on all Butterworth-Heinemann publicationsvisit our website at books.elsevier.com

About the author vi

Preface vii

Acknowledgements ix

Introduction xi

1 The nature of glass 1

2 Historical development of glass 16

3 Technology of glass production 73Part 1: Methods and materials 73Part 2: Furnaces and melting techniques 135

4 Deterioration of glass 169

5 Materials used for glass restoration 199

6 Examination of glass, recording and documentation 227

7 Conservation and restoration of glass 242Part 1: Excavated glass 243Part 2: Historic and decorative glass 271

Appendix 1 Materials and equipment for glass conservation and restoration 345

Appendix 2 Sources of information 347

Bibliography 349

Index 367

Contents

Sandra Davison FICC ACR trained in archaeo-logical conservation at the Institute ofArchaeology (London University), and hasworked as a practising conservator for thirty-five years. Fourteen years were spent as aconservator at The British Museum, and aftera brief spell abroad, she has continued in herown private practice since 1984. Sandra haslectured and published widely, including adefinitive work, Conservation of Glass (withProfessor Roy Newton, OBE), of which thisvolume is a revised and enlarged edition.

In addition to working for museums in theUnited Kingdom, France, the Czech Republic,Malaysia and Saudi Arabia, she has taughtglass restoration in the UK, Denmark, Norway,the Netherlands, the USA, Egypt, Mexico andYugoslavia.

In 1979 she was made a Fellow of theInternational Association for the Conservationof Historic and Artistic Works (IIC), and in2000 became one of the first conservators tobecome an accredited member of the UnitedKingdom Institute for Conservation (UKIC).

About the author

Conservation of Glass, first published in 1989,was intended to serve as a textbook for conser-vation students, conservators and restorersworking on glass artefacts within museums,and those restoring painted (stained) glasswindows in situ. It was written by two authorswith very different, but complementarybackgrounds and experience in the conserva-tion of glass. Roy Newton, a glass scientist(now retired), has worked in glass manufac-turing, on the archaeology of glass and on theproblems concerned with the conservation ofmedieval ecclesiastical painted windows.Sandra Davison, a practising conservator forover thirty years, has conserved a great varietyof glass artefacts, published and lecturedwidely, and teaches the principles and practiceof glass conservation in many countries.

In this edition, written by Sandra Davison,the section concerning painted glass windowrestoration has been removed, with the inten-tion of producing a separate volume at a laterdate. However, information concerning thehistory and technology of glass window-making has been retained as backgroundknowledge for conservators preserving panelsof glass held in collections. The revised title,Conservation and Restoration of Glass, reflectsthe closer involvement of conservators indeveloping conservation strategies for dealingwith glass in historic houses and elsewhere inthe public arena. The volume includes sectionson the historical development and treatmentof mirrors, chandeliers, reverse paintings onglass and enamels.

Conservation and Restoration of Glassprovides an introduction to the considerablebackground knowledge required by conserva-tors and restorers concerning the objects intheir care. Chapter 1 defines the nature of

glass in terms of its chemical structure andphysical properties. Chapter 2 contains a briefhistory of glassmaking, illustrating the chang-ing styles of glass decoration, and the histori-cal development of light fittings (in particularchandeliers), flat glass, mirrors, reverse glasspaintings and micromosaics and enamels.Chapter 3 consists of two parts. The firstdescribes the use of the raw materials fromwhich glass is made and the historical devel-opment of methods of glass manufacture; thesecond is concerned with the development offurnaces and melting techniques. The mecha-nisms by which glass deteriorates, in differentenvironments, are described in Chapter 4,together with an outline of experiments under-taken for commercial/industrial concerns, todetermine the durability of glass. The materi-als used in the processes of conservation andrestoration of glass are discussed in Chapter 5.The examination of glass, described in Chapter6, outlines both simple methods for use byconservators, and those more elaboratetechniques which can be of use for analysis,research and the detection of fakes. Finally, inChapter 7, the details of conservation andrestoration techniques, based on currentpractice in several countries, are described andillustrated. Conservators/restorers should notnormally undertake complicated proceduresfor which they have not had training orexperience; but specialized areas of glassconservation are outlined in Chapter 7 in orderto identify the problems that will requireexpert attention. Information concerningdevelopments in glass conservation, whichmay also include details of treatments thathave proved to be unsuccessful, can be foundin conservation literature and glass conferenceproceedings.

Preface

This Page Intentionally Left Blank

There have been significant developments andgrowth in glass conservation. The author hasattempted to reflect this by inviting commentsfrom a number of conservators and restorers(in private practice or museum employment),conservation scientists and experts in relatedfields, working in Britain, Europe and NorthAmerica.

In particular, the author is greatly indebtedto Professor Roy Newton for undertaking theenormous amount of research for Conserva-tion of Glass, of which this book is a devel-opment; and to the following colleagues fortheir valuable assistance (and who, unlessstated otherwise, are in private practice):

Chapter 1: Angela Seddon (Professor ofMaterials Science, University of Nottingham).Chapter 2: Phil Barnes (enamels); Simone Bretz(reverse paintings on glass; Germany); JudyRudoe (micromosaics; Assistant Keeper,Department of Medieval and Modern Europe,British Museum); Mark Bamborough (paintedglass windows); Tom Kupper (plain glazing;Lincoln Cathedral); Eva Rydlova (Brychta glassfigurines; Czech Republic). Chapter 3 part 1:Paul Nicholson (Egyptologist, University ofBristol); part 2: David Crossley (industrialarchaeologist, The University of Sheffield) andthe late Robert Charleston (glass historian andformer Curator of the Department of Ceramicsand Glass, Victoria and Albert Museum).Chapter 4: Ian Freestone (Deputy Keeper,Department of Scientific Research, BritishMuseum). Chapter 5: Velson Horie (conserva-

tion scientist, Manchester Museum, University ofManchester). Chapter 6: Angela Seddon(University of Nottingham) and Ian Freestone(British Museum). Chapter 7: Victoria Oakley(Head of Ceramics and Glass Conservation,Victoria and Albert Museum) and PatriciaJackson (UK), Rolf Wihr (Germany), CarolaBohm (Sweden), Raymond Errett (retired) andSharon Smith-Abbott (USA) (glass object conser-vators); Alison Rae and Jenny Potter (conserva-tors of ethnographic material – beads; OrganicArtefacts Section, Department of Conservation,British Museum); Annie Lord (textile conserva-tor – beads; The Conservation Centre, NationalMuseums and Galleries Merseyside, Liverpool).

Thanks are also due to Vantico (formerlyCiba Speciality Polymers), Duxford, Cambridgefor technical advice and for a generous granttowards research. Finally to my family, T.K.and E. Lord, without whose gift of a computerthis book would not have been written, toWBJH for patience with computer queries andendless photocopying, and Steve Bell fortechnical support.

The sources of illustrations (other than thoseby Roy Newton and the author) are statedbriefly in the captions. Every effort has beenmade to trace copyright holders. The authorand publishers gratefully acknowledge thekind permission, granted by individuals,museum authorities, publishers and others, toreproduce copyright material.

S.D.2002

Acknowledgements


The conservation of glass, as of all artefacts,falls into two main categories: passive conser-vation, the control of the surrounding environ-ment to prevent further deterioration; andactive conservation, the treatment of artefactsto stabilize them. A storage or display environ-ment will consist of one of the following: (i)natural climatic conditions (especially paintedglass windows and glass mosaics in situ); (ii)modified (buffered) climatic conditions inbuildings and cases with no air conditioning;(iii) controlled climatic conditions, where airconditioning has been installed in museumgalleries or individual showcases, to holdtemperature and relative humidity withincarefully defined parameters. Environmentalcontrol is a discipline in its own right(Thomson, 1998) and outside the scope of thisbook. However, conservators need to beaware of the basic facts in order to be able toengage in discussions regarding display andstorage conditions, and the choice of materi-als for display, and packaging for storage andtransport. The prevention of further damageand decay by passive conservation, representsthe minimum type of treatment, and normallyfollows examination and recording. Reasonsfor not undertaking further conservation mightbe lack of finance, facilities, lack of an appro-priate treatment or the sheer volume of glass,e.g. from excavation.

Active conservation, as the term implies,involves various levels of interference.Minimal conservation would include ‘first aid’,photography, X-radiography (where appropri-ate), a minimal amount of investigative conser-vation such as surface cleaning, and suitablepackaging or repackaging for safe storage.Partial conservation entails the work abovebut with a higher degree of cleaning, with or

without consolidation. Full conservation workwould additionally involve consolidation andrepair (reconstruction of existing fragments),supplemented by additional analytical infor-mation where appropriate. Display standardconservation might include cosmetic treatmentsuch as restoration (partial or full replacementof missing parts) or interpretative mounting fordisplay. Restoration of glass objects may alsobe necessary to enable them to be handledsafely. It should only be carried out accordingto sound archaeological or historical evidence.The level of conservation has to be agreedbetween a conservator/restorer and the owner,custodian or curator, before work begins.

Historically, glass conservation was not aseasily developed as it was for ceramics, forexample. The fragile nature of glass made itdifficult to retrieve from excavations, and thetransparent quality of much glass posed thedifficulty of finding suitable adhesives andgap-filling materials with which to work. Theuse of synthetic materials and improvementsin terrestrial and underwater archaeologicalexcavation techniques have resulted in thepreservation of glass which it was not formerlypossible to retrieve; and continues to extendthe knowledge of ancient glass history,technology and trade routes. Early treatmentsusing shellac, waxes and plaster of Paris wereopaque or coloured and not aestheticallypleasing (Davison, 1984). Later, rigid transpar-ent acrylic materials such as Perspex (US:Plexiglas) were heat-formed and cut to replacemissing areas of glass. Advantages were theirtransparency and only slight discoloration andembrittlement with age. However, theprocesses were time-consuming, and thereplacements did not necessarily fit wellagainst the original glass. Unweathered glass

Introduction

surfaces are smooth, essentially non-porousand are covered with a microscopic layer ofwater, so that few materials will adhere satis-factorily to them. It was only with thecommercial formulation of clear, cold-settingsynthetic materials, with greater adhesiveproperties, that significant developments inglass conservation were achieved. Epoxy,polyester and acrylic resins could be polymer-ized in moulds in situ, at ambient tempera-

tures with little or no shrinkage. However,restoration involves interference with the glassin terms of the moulding and casting processes(Newton and Davison, 1989). Recentapproaches to glass conservation and restora-tion have been the construction of detachablegap-fills (Hogan, 1993; Koob, 2000), and themounting of glass fragments or incompleteobjects on modern blown glass formers, or onacrylic mounts.

xii Introduction

The term glass is commonly applied to thetransparent, brittle material used to formwindows, vessels and many other objects.More correctly, glass refers to a state of matterwith a disordered chemical structure, i.e. non-crystalline. A wide variety of such glasses isknown, both inorganic (for instancecompound glasses and enamels, and even thesomewhat rare metallic glasses) and organic(such as barley sugar); this book is concernedonly with inorganic glasses, and then onlywith certain silicate glasses, which areinorganic products of fusion, cooled to a rigidcondition without crystallizing. The termancient glasses is that used by Turner(1956a,b) to define silicate glasses which weremade before there was a reasonable under-standing of glass compositions, that is beforethe middle of the seventeenth century (seealso Brill, 1962). In this book, for convenience,the term glass will be used to mean bothancient and historic silicate glasses.Understanding the special chemical structureand unique physical properties of silicateglasses is essential in order to appreciate boththe processes of manufacture of glass objectsand the deterioration of glass, which maymake conservation a necessity.

Natural glasses

Before the discovery of how glass could bemanufactured from its raw ingredients, manhad used naturally occurring glass for manythousands of years. Natural silica (the basicingredient of glass) is found in three crystallineforms, quartz, tridymite and cristobalite, andeach of these can also occur in at least two

forms. Quartz is the most common, in theform of rock crystal, sand, or as a constituentof clay. Rock crystal was fashioned into beadsand other decorative objects, including, inseventeenth century France, chandelier drops.If quartz is free from inclusions, it can bevisually mistaken for glass.

Sudden volcanic eruptions, followed byrapid cooling, can cause highly siliceous lavato form natural glasses (amorphous silica), ofwhich obsidian is the most common. Inancient times, obsidian was chipped andflaked to form sharp-edged tools, in the samemanner as flint (Figure 1.1). Other forms ofnaturally occurring glass are volcanic pumice,lechatelierite or fulgurites and tektites. Pumiceis a natural foamed glass produced by gasesbeing liberated from solution in molten lava,before and after rapid cooling. Lechatelierite isa fused silica glass formed in desert areas by

1

1

The nature of glass

Figure 1.1 Since prehistoric times, obsidian has beenused to fashion tools. The spearhead shown here is amodern example, made in Mexico.

lightning striking a mass of sand. The irregu-lar tubes of fused silica (fulgurites) may be ofconsiderable length. Lechatelierite has alsobeen discovered in association with meteoritecraters, for example at Winslow, Arizona.Tektites are small rounded pieces of glass, ofmeteoric origin, found just below the surfaceof the ground in many parts of the world, andwhich appear to have come through theatmosphere and been heated by fallingthrough the air while rotating. Their composi-tion is similar to that of obsidian, but theycontain more iron and manganese.

Man-made glasses

In order to understand the nature of man-made glass, it is first necessary to defineseveral terms for vitreous materials, some ofwhich have previously been used ambiguouslyor incorrectly (Tite and Bimson, 1987). Thereare four vitreous products: glass, glaze, enameland (so-called, Egyptian) faience, whichconsist of silica, alkali metal oxides and lime.Glass, glaze and enamel always contain largequantities of soda (Na2O) or another alkalimetal oxide, such as potash (K2O), andsometimes both, whereas Egyptian faiencecontains only quite small amounts of alkalimetal oxide. It has formerly been supposed,that because of the difficulty of reaching andmaintaining the high temperatures required tomelt glass from its raw ingredients, in ancienttimes, the raw ingredients were first formedinto an intermediate product known as frit.However, there is limited evidence for thispractice. In the fritting process, raw materialswould be heated at temperatures just highenough to fuse them, and in doing so torelease carbon dioxide from the alkali carbon-ates. The resulting mass was then pounded topowder form (the frit). This was reheated athigher temperatures to form a semi-moltenpaste which could be formed into objects, orwas heated at higher temperatures at which itcould melt to form true glass.

A silicate glass is a material normally formedfrom silica, alkali metal oxides (commonlyreferred to as alkalis) and lime, when thesehave been heated to a temperature highenough to form them into a homogeneousstructure (formerly and ambiguously termed

glass metal). Chemically, glass, glaze andenamel can all be identical in composition, thefundamental difference being their method ofuse in antiquity. The coefficient of thermalexpansion of a glass was not important whenit was used alone (unless it was applied on adifferent glass, as in the manufacture of cameoglass), whereas in a glaze or an enamel anydifference in thermal expansion between themand the base on which they were fused couldcause the glaze or enamel to crack or becomedetached from the base material. In practice,glasses and enamels needed to have a lowmelting point, remain plastic as long as possi-ble while cooling and, apart from the veryearliest glasses, be translucent or transparent(in contrast to the early glazing of earthenwarewhere coloured decoration had been impor-tant).

A glaze is a thin vitreous coating applied toanother material to make it impermeable, orto produce a shiny decorative appearance.Glaze was sometimes applied with the bodymaterial before firing, but more often it wasapplied to the object after it had received afirst firing, following which the object wasrefired to form the glazed surface (Figure 1.2).

Faience is composed of fritted silica withabout 2 wt per cent of lime (CaO) and about0.25 wt per cent soda, lightly held togetherwith a bonding agent such as water. Theresulting paste was shaped by hand or in anopen mould and then heated until the limeand soda had reacted enough (fused suffi-ciently) to hold the silica particles together.During the formation process, faience objects

2 Conservation and Restoration of Glass

Figure 1.2 A thick layer of glaze covering a stonewarebowl.

formed a glazed surface with a similar compo-sition to the body, usually coloured blue orgreen with copper compounds. (Strictly speak-ing the term faience, derived from the nameof the Italian town of Faenza, should refer tothe tin-glazed earthenware made there.) Toreduce confusion the material discussed hereshould be referred to as Egyptian faience, orpreferably, glazed siliceous ware (see Plate 2and Figure 3.2), (Nicholson, 1993; Smith,1996).

The pigment known as Egyptian Blue, firstused in Egypt during the third millennium BC,and during the next 3000 years, in wall paint-ings, and as beads, scarabs, inlays andstatuettes, is the mineral (CaO.CuO.4SiO2) =(CaCuSi4O10). X-ray diffraction analysis hasshown that, in addition to this compound, theonly crystalline materials were quartz andtridymite (another of the crystalline forms ofsilica) (Chase, 1971; Tite et al., 1981).

A enamel resembles a glaze in that it is alsofused to a body of a different material, in thiscase, metal (see Figures 3.33–3.38, 7 57 and7.58); however, the term enamel is also usedto describe vitreous pigments used to decorateceramics and glass (see Chapter 3).

Chemical structure and composition

Zachariasen (1932) established that the atomsand ions in silicate glasses are linked togetherby strong forces, essentially the same as incrystals, but lacking the long range orderwhich is characteristic of a crystal. Crystallinesilica (quartz) melts sharply at 1720°C from itssolid state, to a liquid, just as ice melts to formwater at 0°C. This melting point is scientificallyreferred to as the liquidus. When the silicaliquid (molten glass) is cooled from above theliquidus, the randomly distributed moleculeswill endeavour to adopt a less random config-uration, more like those of crystals. However,an alternative three-dimensional structureforms because the crystallization process ishindered by the high viscosity of the glass,and the presence of the network modifiers.The melt becomes more and more viscous asthe temperature is lowered until, at about1050°C it sets to form a solid glass (a stateformerly but no longer referred to as a super-cooled liquid). Moreover, the density of thatglass is less than that of the original quartz

because there are now many spaces betweenthe ill-fitting molecules.

However, in order to form a usable glass itis necessary to add certain oxides to the silica,which act as network modifiers, stabilizers andcolourants, and which also have a markedeffect on the structure of the resulting product.When network modifiers are added, they havethe effect of considerably lowering the viscos-ity of the melt (see Figure 1.8). Thus there isthe potential for a different type of crystalcontaining atoms from the modifiers, to form inthe sub liquidus melt, provided the melt hasbeen held at the liquidus temperature for longenough. Thus a glass with the molar composi-tion 16Na2O, 10CaO, 74SiO2 can form crystalsof devitrite (Na2O.3CaO.6SiO2); which grow ata rate of 17 μm per minute at a temperature of995°C, the optimum temperature for growth ofdevitrite in that composition of glass. The totalchemical composition of the glass remainsunaltered (i.e. no atoms are added orsubtracted from those already in the glass),although the composition will change locally ascrystals of devitrite separate from the bulk glass.

Ancient glasses have such complex compo-sitions that devitrification occurs much lesseasily than in modern glasses, so that ifcrystals of devitrite are present in a sampleundergoing examination, there may be doubtsconcerning the antiquity of the glass.However, the enormous block of glass madein a tank furnace in a cave at Bet She’arim, inIsrael, was found to be heavily devitrified(with the material wollastonite, CaSiO3) as aconsequence of containing 15.9 wt per cent oflime (Brill and Wosinski, 1965). The opalizingagent in some glasses may be a devitrificationproduct itself, which forms only when suitableheat treatment is given to the glass. Devitritedoes not occur as a mineral in nature.

Early historians and archaeologists haveoccasionally used the term devitrification inquite a different sense, meaning loss of vitre-ous structure to describe glass that has weath-ered with loss of alkali metal ions, of otherconstituents of the glass and probably a gain inwater content. This ambiguous use of the termshould be avoided (Newton and Werner, 1974).

Network formersThe principal network former in ancientglasses is silica (SiO2). Silicon and oxygen in

The nature of glass 3

crystalline silica (quartz) are arranged in adefinite pattern, the units of which arerepeated at regular intervals forming a three-dimensional network consisting of tetrahedrawith a silicon atom at the centre and anoxygen atom at each corner; all four of theseoxygen atoms form bridges to silicon atoms ofthe four neighbouring silicon tetrahedra. Othernetwork formers are the oxides of boron(B2O3), lead (PbO) (Charleston, 1960) andphosphorus (P2O5). The presence of boron isimportant for clarifying glass compositions.However, it is difficult to analyse and so mighteasily be missed, especially since ancientglasses typically contained only 0.01 to 0.02per cent (whereas some Byzantine glassescontained 0.25 per cent boron). Boron enteredthe glass by way of the ash obtained byburning plants containing boric oxide. Themineral colemanite (hydrated calcium borate)(Ca3B6O11.5H2O) is found in western Turkey,and may have been used in glassmaking.

The concept of network-forming oxides isillustrated in Figures 1.3 and 1.4. Figure 1.3shows the regular structure of an imaginarytwo-dimensional crystalline material. Withinthe broken line there are 16 black dots (repre-senting atoms of type A) and 24 open circles(representing atoms of type O); hence theimaginary material has the composition A2O3

and its regular structure shows that it iscrystalline. If the imaginary crystalline materialA2O3, shown in Figure 1.3, has been melted,and is cooled quickly from the molten state,the resultant solid might have the structureshown in Figure 1.4. Here the broken lineencloses 24 black dots and 36 open circles andhence the composition is again A2O3 but thestructure is irregular and non-crystalline, repre-senting the amorphous, glassy or vitreous stateof the same compound. Note that theamorphous structure contains spaces and thusoccupies a greater volume than the crystallineone, and hence the crystal has a higher densitythan the glass, even though the chemicalcomposition is the same.

Network modifiersFigure 1.5 shows a structure which is nearerto that of silicate glass. It is again a simplifiedtwo-dimensional diagram, and the key to itnow mentions the word ion. Ions are atomsthat have been given an electrical charge, by


Figure 1.3 Schematic two-dimensional representationof the structure of an imaginary crystalline compoundA2O3.

Figure 1.4 Structure of the glassy form of thecompound in Figure 1.3.

adding or subtracting one or more electrons;cations having lost electrons, have a positivecharge, and anions having gained electrons,have a negative charge. The network-formingatoms are represented by black dots withinshaded triangles (atoms of silicon), and thenetwork modifying ions (positively chargedcations) are cross-hatched circles lying in thespaces of the network. Each network-formingtriangle (silicon atom) is accompanied by threeoxygen atoms (shown by small circles), whichcan be of two kinds. There are bridgingoxygen atoms (shown by plain open circles)which are shared between two triangles, thusjoining them together and forming part of thenetwork. There are also non-bridging oxygenions (shown by circles with a central dot)which belong to only one triangle; each ofthese thus bears a negative charge which isneutralized by a positive charge on one of thecross-hatched circles (cations). (Strictly, the Si-O-Si bonds are ‘iono-covalent’. They are notionic enough to refer to the oxygen as ions,and the Si as a cation. In the case of the Si-O non-bridging bonds, the Si-O bond is stilliono-covalent, but the negative charge on theoxygen gives it the ability to form an ionicbond to a cation in a nearby space.) It shouldbe noted that there is a very small amount ofcrystalline material in the diagram, near ‘A’ inFigure 1.5, where four triangles are joinedtogether to form a regular (hence crystalline)area. (This can occur also in ancient glasses,

where micro-crystallites can be detected.) Atall other points the triangles form irregularchains, which enclose relatively large spaces(and hence the density of the glass is less thanthat of a corresponding crystalline form).These spaces in the network have beencreated by the network-modifying cationswhich bear one or more positive electricalcharges, and which can be considered to beheld, by those electrical charges, to be more(or perhaps rather less) loosely bound in thoseenlarged spaces.

The monovalent cations (which bear onlyone positive charge, having lost an electron toan adjacent non-bridging oxygen ion) areusually the alkali metal ions, either sodium(Na+) or potassium (K+), which bring withthem one extra oxygen ion when they areadded to the glass as soda or as potash.Because these cations bear only a singlepositive charge, they can move easily fromone space in the network to another (looselybound). Thus, when the glass is placed inwater, it becomes less durable because thecations (the smaller of the cross-hatchedcircles in Figure 1.5) can move right out ofthe glass into the water, thus making the waterslightly alkaline. In order to maintain theelectrical neutrality of the glass, these cationsmust be replaced by another cation such asthe oxonium ion (H3O).

In the case of the divalent alkaline earthcations (the larger cross-hatched circles), eachbears a double positive charge (being associ-ated with two non-bridging oxygen ions, thecircles with dots inside). These are usuallyCa++ or Mg++, added to the glass as lime(CaO) or as magnesia (MgO), but otherdivalent alkaline earth ions may also bepresent. The double electrical charge on themholds them nearer (more tightly bound) totheir accompanying non-bridging oxygens,making it much harder for them to move fromone space to another. Thus divalent alkalineearth cations play little or no part in carryingan electric current through the glass. Becausethey are associated with two non-bridgingoxygen ions, they strengthen the network,thus explaining why they help to offset thereduction in durability produced by the alkalimetal cations. However it should be noted thatin Figure 1.5 the double ionic linkages (tocircles with dots) are not immediately obvious.


Figure 1.5 Schematic two-dimensional representationof glass, according to Zachariasen’s theory.

It is these linkages which determine the verydifferent effects that the monovalent anddivalent cations have on the durability of glass.

Notable advances have been made in theunderstanding of the structure of glasses. Forexample, it is now realized that the networkis actually loosened in the vicinity of themonovalent cations, channels (rather thanmerely larger spaces) being formed in whichthe cations can move even more easily thanwas formerly realized.

Phase separationDespite the essentially homogeneous nature ofbulk glasses, there may be minute areas,perhaps only 100 nm (0.1 m) in diameter,where the glass is not homogeneous becausephase separation has occurred. These regions(rather like that near ‘A’ in Figure 1.5) canhave a different chemical composition fromthe rest of the glass, i.e. the continuous phase(Goodman, 1987). Phase separation can occurin ancient glasses, and can have an effect ontheir durability, because the separated phasemay have either a greater or a lower resistanceto deterioration. The amount of phase separa-tion can be seen through an electron micro-scope.

ColourantsThe coloured effects observed in ancient andhistoric glasses were produced in three ways:(i) by the presence of relatively small amounts(about one per cent) of the oxides of certaintransition metals, especially cobalt (Co),copper (Cu), iron (Fe), nickel (Ni), manganese(Mn), etc., which go into solution in thenetwork; (ii) by the development of colloidalsuspensions of metallic, or other insolubleparticles, such as those in silver stains (yellow)or in copper or gold ruby glasses (red ororange); (iii) by the inclusion of opalizingagents which produce opal and translucenteffects. The production of coloured glasses notonly depends on the metallic oxides presentin the batch, but also on the temperature andstate of oxidation or reduction in the furnace.Of course the exact compositions of ancientglasses were complex and unknown, beinggoverned by the raw materials and furnaceconditions, so that the results could not beacccurately determined.

Dissolved metal oxides/state ofoxidationColoured glasses can be produced by metaloxides dissolving in the glass (similar to thecolours produced when the salts of thosemetals are dissolved in water), although theresultant colours will also be affected by theoxidizing or reducing (redox) conditions inthe furnace. In the traditional sense, a metalwas oxidized when it combined with oxygento form an oxide, and the oxide was reducedwhen the metal was reformed. The positioncan be more complicated when there is morethan one state of oxidation. For example, iron(Fe) becomes oxidized when ferrous oxide(FeO) is formed, and a blue colour isproduced in the glass (because Fe2+ ions arepresent), but it becomes further oxidized whenmore oxygen is added to form ferric oxide(Fe2O3), which imparts a pale brown or yellowcolour to the glass (due to the Fe3+ ionspresent). However, the situation is rarely sosimple and usually mixtures of the two oxidesof iron are present, producing glasses ofvarious shades of green. When a chemicalanalysis of glass is undertaken, it is customaryto quote the amount of iron oxide as Fe2O3,but that does not necessarily imply that all ofthe iron is in that state.

The oxidation process occurs when an atomloses an electron, and conversely, reductiontakes place when an atom gains an electron.Consider the two reversible reactions set outin equations (1.1) and (1.2), where e– repre-sents an electron, with its negative charge. Inequation (1.1) the forward arrow shows thatan electron is lost when Fe2+ is converted toFe3+.

Fe2+ Fe3+ + e– 1.1

Mn3+ + e– Mn2+ 1.2

The combined effects of equations (1.1) and(1.2) is equation (1.3), which shows that thereis an equilibrium between the two states ofoxidation of the manganese and of the iron(Newton in Newton and Davison, 1989).

Fe2+ + Mn3+ Fe3+ + Mn2+ 1.3

But the Fe3+ and Mn2+ are the more stablestates, and hence the equilibrium tends to


move to the right. Thus, when the conditionsduring melting of the glass are fully reducing(the equilibrium has been forced to the left,for example by producing smoky conditionsin the furnace atmosphere) the ironcontributes a bright blue colour due to the Fe2

ions (corresponding to FeO) and themanganese is in the colourless form so that ablue glass is obtained. When the conditionsare fully oxidizing (the equilibrium has beenmoved to the right by the addition of oxidiz-ing agents; by changing the furnace conditionsto have short, bright flames; or by prolongingthe melting time), the iron contributes abrownish yellow colour and the manganesecontributes a purple colour, so the glassappears brownish violet. When the conditionsare intermediate, a variety of colours areobtainable, such as green, yellow, pink, etc.including a colourless glass when the purplefrom the manganese just balances the yellowfrom the iron. This is the reason why, if thereis not too much manganese, it will act as adecolourizer for the glass which would other-wise be greenish in colour.

These conditions have been experimentallystudied by Sellner (1977) and Sellner et al.,(1979), who produced a forest-type glass inwhich the colouring agents were onlymanganese (1.7 wt per cent MnO) and iron(0.7 wt per cent Fe2O3). A variety of colourswas obtained, from pale blue, when thefurnace atmosphere was fully reducing (withunburned fuel present and a very low partialpressure of oxygen in the waste gases)through green and yellow to dark violet whenthe furnace atmosphere was fully oxidizing(plenty of excess oxygen in the waste gases).

Sellner et al. (1979) also examined samplesof glass excavated from two seventeenth-century glassworks sites, one at Glassborn/Spessart and the other at Hilsborn/Grünenplan, both in Germany. The composi-tions of the glasses at both sites were similarto each other, but the former factory hadproduced green glass and the latter hadproduced yellowish to purple glass. Measure-ments by electron spin resonance showed thatthe green glass had been melted under reduc-ing conditions and the Hilsborn glass hadbeen melted under oxidizing conditions. Thus,the colour of the glass had been determinedby its having been made using beechwood ash

(which contains both iron and manganese),and the furnace atmosphere, and not by theaddition of manganese. The origin of colourin these glasses has also been investigated bySchofield et al. (1995), using synchrotronradiation.

Greenish colours can be obtained fromcopper. For archaeological reasons it may benecessary to discover whether tin or zinc isalso present, because the presence of tinwould suggest that bronze filings might havebeen added to the batch, whereas the presenceof zinc would suggest the use of brass waste.

However, the presence of appreciableamounts of a particular oxide need not neces-sarily indicate a deliberate addition of thatmaterial. For example, Figure 1.6 showsremarkable differences in the potash andmagnesia contents of Egyptian Islamic glassweights, manufactured either before, or after,845 AD. Brill (1971a) suggested that the earlierexamples were made with soda from thenatron lakes, whereas the later ones couldhave contained potash derived from burntplant ash. There are still many problems andambiguities to be solved regarding the compo-sitions of ancient glasses, by analyses ofsamples from known provenances. However,there are many cases where the colouringagent is so strong that there is no problem.Figure 1.7 shows the contents of metal ions infive kinds of ancient glass; sometimes only


Figure 1.6 Chronological division of Egyptian Islamicglass weights into high- and low-magnesium types.(From Sayre, 1965).

0.02 per cent of cobalt is sufficient to producea good blue colour. The deliberate productionof an amber colour in ancient glass was in theform of iron-manganese amber describedabove, or carbon-sulphur amber. (They can bedistinguished from each other because theFe/Mn colour has optical absorption bands at380 and 500 nm, whereas the C/S colour hasits absorption bands at 430 and 1050 nm.)

The metals strontium (Sr), lithium (Li) andtitanium (Ti) enter glasses as trace elements inthe raw materials, in calcium carbonate forexample; beach sand containing shells is highin strontium in comparison with limestonewhich is low in its content, and therefore theamount present in glass is an indicator as towhether shell was a deliberate addition.Strontium is a reactive metal resemblingcalcium, lithium is an alkali metal resemblingsodium, but is less active and titanium resem-bles iron.

Colloidal suspensions of metalsQuite different colouring effects are obtainedwhen the metals do not dissolve in the glass,but are dispersed (as a colloid) in the glass;the colour is then produced by light diffrac-tion, and is therefore related to the size of thedispersed metal particles. For example, coppercan produce red, orange or yellow colours.

The dichroic colour of the Lycurgus Cup (Plate4), made in the fourth century AD, is a strik-ing example, appearing transparent wine redin transmitted light, and translucent green byreflected light. This dichroic effect is producedby colloidal gold and silver.

The rich red colour in medieval cathedralwindow glass was produced by the presenceof dispersed copper, but another red, with adistinct tint of purple, was produced bydispersed gold. The production of gold andcopper ruby glasses is complicated because thestrong colour does not develop (strike) untilthe glass is reheated (Weyl, 1951).

Copper ruby glasses have certainly been inuse since the twelfth century. One problem intheir use was the very intense colour produced:a piece of red glass only 3 mm thick (about thethinnest which could be used as window glass),would have appeared black instead of red. Twodifferent techniques have been used at differ-ent times to overcome the problem. In thetwelfth and thirteenth centuries, a transparentred glass was produced by distributing the redcolour in a series of very many extremely thinlayers. It is not known exactly how the layeredeffect was obtained, because the copper-containing glass had to be reheated before thecolour appeared (i.e. before it strikes), whichwould have melted the glass layers together. Itmay have been that the multi-layered effectmay have been obtained accidentally whilsttrying to produce an extremely diluted copperred glass. A poor distribution of the copper inthe melt perhaps influenced the strike of thecolour in that some layers became red whilstothers did not. From the fourteenth centuryonwards, the technique of flashing, in which athin layer of red glass was laid on a base ofcolourless glass, was used to produce transpar-ent red glass. Flashed glass appears bright redwhen viewed from the front, but when viewedthrough the edge, the layers of clear andcoloured glasses can be seen.

Gold ruby glasses were probably in usefrom the sixteenth century, but its extensiveuse in the seventeenth century follows fromthe use of Purple of Cassius (a purple pigmentconsisting of a mixture of colloidal gold andstannic acid) by Johann Kunckel (1679).Kunckel evidently did not completely masterthe art of developing the full colour becauseonly a small proportion of the melts seem to


Figure 1.7 Colour element patterns in cobalt-blueglasses dating from the second millennium BC. (FromSayre, 1965).

have been satisfactory. After Kunckel’s deathin 1705, the production of gold ruby glasscontinued in Bohemia, and certainly until theeighteenth century. The excavation ofKunckel’s glassworks, on Pfauen Island, nearPotsdam, caused a resurgence of interest inthe work (Schulze, 1977). Neutron activationstudies on the excavated samples of glassshowed that the depth of colour was relatedto the concentration of gold, faintly colouredsamples contained about 0.03 per cent gold,and the more strongly coloured samplescontained 0.07 per cent, confirming datapublished by Kunckel (1679). In thenineteenth century the owners of glassworkshad a custom of tossing a gold sovereign intothe gold ruby batch. Gold dissolved in aquaregia would have already been added to thebatch to produce the colour (Frank, 1984), andso it would seem that the custom of adding acoin was either to impress the workmen, orto confuse industrial spies (Newton, 1970).

DecolourizersIf iron is the only colouring oxide present itwill produce a blue colour in its reduced form,but a much paler yellow is produced whenthe iron is oxidized. As seen in equation 1.3above, manganese oxide can oxidize the ironto the yellow ferric state, and a slight excessof manganese will produce a pale purplewhich is complementary in colour to theyellow and thus effectively neutralizes itproducing a virtually colourless glass. Thus,for at least the last few centuries, manganesehas been deliberately used as the decolourizerfor iron. There are also other oxidizing agents(such as the oxides of arsenic and ofantimony) which can turn the blue from theiron to a very pale yellow, but it does notneutralize it in the same way that the purplecolour of the manganese neutralizes theyellow of the ferric iron. Since no other colouris neutralized by this process, it is fortunatethat iron is the predominant impurity in sandwhich produces undesired colour.

Lead glassesLead-rich glasses are relatively uncommon. Inthe West, they were used to produce red andyellow opaque glasses in antiquity, and certaintransparent glasses in the medieval period. Inthe Far East, lead-rich glasses were produced

in China. The amount of lead found in ancientglasses was probably not enough to alter theirworking properties or appearance, and there-fore it is unlikely to have been a deliberateaddition, but derived from the sand. In factlead oxide seems to have been an uninten-tional ingredient of glass until Roman times.Lead-containing glasses probably existed asearly as the second millennium BC, since leadwas one of the ingredients mentioned inMesopotamian cuneiform texts of that date.Analysis of a cake of red glass dating from thesixth century BC showed that it contained 22.8per cent PbO by weight, giving the impressionthat 0.25 per cent of the glass composition.However, since lead is a very heavy element,the true position is seen to be quite differentwhen the glass composition is calculated on amolar percentage, the lead oxide then beingonly 9.3 per cent. Thus 9.3 per cent of themolecules in the glass are lead oxide, andtherefore lead glasses can be regarded assilicate glasses containing some 10 per cent ofdivalent network-modifying lead oxide.

Before the use of lead oxide in the makingof lead glass in the seventeenth century, leadwas used in the form of litharge, produced byblowing air over the surface of molten lead.When litharge is further oxidized, it becomesred lead. Its use required special furnaceconditions, as its conversion back to metalliclead would discolour the glass and damagethe crucibles or pots.

In the seventeenth century George Ravens-croft, working in England, produced a clear,brilliant glass by adding as much as 30 percent lead oxide to the glass batch. Lead isso heavy that it can represent 50 weight percent of a glass. Figure 1.8 shows how thedensity of a glass is closely related to its leadcontent.

Opacifying agentsThe most ancient glasses were opaque due tothe presence of masses of tiny bubbles, orother dispersed materials within the viscousbatch. Deliberate incorporation of air bubblescan be a way of producing opaque, somewhatopalescent glasses. However, the majority ofopal glasses were produced by the use ofrelatively small number of opalizing agents,which form microcrystalline areas within theglass. Different opalizing agents were used in


three distinct eras of glassmaking (Turner,1957a,b, 1959; Rooksby, 1959, 1962, 1964;Turner and Rooksby, 1959, 1961).

Table 1.1 shows that Roman, and pre-Romanwhite opal glasses (or blue if cobalt was

present) contained calcium antimonate,whereas by the fifth century AD the opacifierin common use was tin oxide or, occasionally,calcium fluorophosphates. The use of tin oxidecontinued until the eighteenth century, when itwas replaced by calcium fluoride or lead arsen-ate. Similarly, yellow opaque glasses containedlead antimonate in the early period, and a lead-tin oxide later on. It should, however, be notedthat Bimson and Werner (1967) found cubiclead-tin oxide as the yellow opacifier in therare first century AD gaming pieces found atWelwyn Garden City (Hertfordshire, UK). Thusthe date for the use of this material should beregarded as being much earlier than formerlysupposed. The opaque red glasses (haemat-inum or aventurine) contain copper andcuprous oxide (Cu2O, which is always red) andthey also contain tin or lead (Weyl, 1951). Theorigins of Roman opaque glasses, especiallythose containing antimony, have beendiscussed by Mass et al. (1998).

Physical properties of glass

As explained at the beginning of the chapter,crystalline materials have a definite structure,whereas amorphous ones do not, and there-fore only rather general statements can be


Table 1.1 Opacifying agents in glass, 1450 BC to AD 1961 (from Bimson and Werner, 1967)

Period Type of glass Opacifying agent Number of specimens

1450 BC tofourth century AD

Fifth century AD

toseventeenth century AD

Eighteenth century AD

topresent day

Opaque white and blue

Opaque yellowOpaque red

Opaque white and blue

Opaque yellow and greenOpaque red

Opaque white

Ca2Sb2O7

(occasionally CaSb2O6)Cubic Pb2Sb2O7

Cu2OCu2O+Cuor Cu

�SnO2 usually�3Ca3(PO4)2.CaF3

occasionallyCubic Pb5SnO4

CuCu+Cu2O rarelyCu+SnO2 sometimes

�3Pb2(AsO4)2.PbO(apatite-type structure)CaF or CaF3+NaF(Na2Ca)2Sb2O6F

15

10

8

10

417

7

4

Many1

Figure 1.8 Graph relating the density of lead glass toits lead content.

made about a material which, when hot, isductile but when cold is brittle, and fracturesif there is a sudden change of temperature.The thermal history of glass is of particularimportance, because glass that has beencooled quickly retains an imprint or ‘memory’of its state at the moment before it was cooled.In the example of a viscous glass melt whichis cooled very slowly from a temperature T1 toa lower temperature T2 energy available formolecular movements is gradually reduced,but (because the rate of cooling is very slow)the network has enough time to readjust itselfand become more compact. (In some casesdevitrification crystals can form when theglass is cooled too slowly at the liquidustemperature.) The spaces in the silicatenetwork will close somewhat, and the glass atT2 will be denser than it was at T1 (this is quitea different process from that of thermalcontraction, which also brings about a slightincrease in density). If the same glass is cooledsuddenly from T1 to T2, the viscous glass doesnot have time for the viscous network tocompact, and the glass at T2 has the lowerdensity which would be characteristic of T1.For this reason, T1 is known as its fictivetemperature, and this demonstrates the slightuncertainty about defining the properties of aglass at any particular temperature. Thisconcept appears again, later in the chapter,under transition point (Tg).

Viscosity of molten glassGlass is generally regarded as being a rigidmaterial, and is recognized as such in every-day use, but depending on the composition ofthe glass, it becomes plastic at temperaturesabove circa 900°C, when it can be worked invery many ways, and into a variety of forms.The viscosity of a liquid is a measure of itsresistance to flow, but compared with otherliquids, molten glass has two special proper-ties: (i) it is very much more viscous than anyother liquids, and (ii) it has an enormousviscosity range depending on the temperature.Figure 1.9 shows a plot of the logarithm ofthe viscosity against temperature for a widerange of glasses. Each division on the lefthand scale represents a 100-fold change inviscosity, and the full extent of the scale repre-sents a change of 1020, or one hundred million,million, million times. Water is shown right at

the bottom. Treacle (molasses) in a warmroom is one thousand times more viscous, butthe most fluid glass shown in the diagram (atpoint F) is ten times even more viscous; whenglass articles are manufactured the viscosity isabout ten times even greater.

The viscosity changes with temperature sorapidly that special terms are used to describeits viscosity at various stages in the manufac-turing process. Figure 1.9 shows that theworking point (103 Nsm–2) of a glass is at aviscosity of 1000 Nsm–2, but at the softeningpoint (6 �106 Nsm–2), the glass is 6000 timesmore viscous than that (when ‘soft’, it is muchtoo viscous to be worked). At the annealingpoint (5 �1012) of the glass it is about a milliontimes even more viscous and the strain pointis about 10 times more viscous still. There isalso a transition point which can have aviscosity as much as 1000 times higher thaneven strain point (5 �1013), and is discussedlater in the chapter. The working range is thedifference in temperature between the workingpoint and the softening point, and thus it canbe seen why neither fused silica (A), nor 96per cent silica (B), can have a working rangewithin ordinary furnace temperatures. (In fact


Figure 1.9 Viscosity-temperature curves for varioustypes of glasses. (After Brill, 1962).

special kinds of electric furnace are required toprocess those very hard glasses on a commer-cial basis, for example, in making fused silicacrucibles, or other highly special chemicalapparatus.) There are also marked differencesin behaviour between different types of glass.Glass C (a laboratory type borosilicate glass)has a working range of 370°C, whereas glass F (high lead optical glass) has a workingrange of only 220°, but that is in the tempera-ture range 580–800° and glass F will cool moreslowly than glass C, which has a temperaturerange of 830–1220°. Glass C has a wider rangein which it can be manipulated, but it alsoloses heat more rapidly and may thereforehave to be re-heated in the furnace glory holemore frequently. Thus both the working range,and the actual temperature, have to be consid-ered when fashioning glass articles. Glass C isreferred to as a hard glass because it requiresa higher temperature for working. It has beensuggested that the viscosity of glass might beexplained by theories of thermodynamicsbased on the interaction of thermally excitedsound waves within fluids.

Because the viscosity increases continuouslywith decreasing temperature, without thediscontinuity of melting which is so character-istic of crystals, it has been suggested that coldglass should show plastic flow if measuredover very long periods of time. Cold glassunder tension does not flow at room temper-ature, because irreversible flow of glass atroom temperature requires a stress of at leastone-tenth of the theoretical breaking strengthof the glass, whereas commercial glasses haveso many surface defects that they fractureunder tensile stresses of only one-hundredthof the theoretical breaking strength. There isactually no evidence for the supposed coldflow of glass under its own weight, becausemany of the alleged examples are actuallystatistical (Newton, 1996).

The process of annealing glass (controlledcooling to relieve the internal stresses whichare formed because the thermal conductivityof hot glass is low) is actually an example ofslow plastic flow of glass when the viscosityis in the range 1011 to 1013 Nsm–2, corre-sponding to temperatures of the order of500°C. When a glass object is formed, theoutside surfaces cool very rapidly, become stiffand contracts thermally, long before the inside

cools. The thicker the glass, the greater thedifference in cooling rate between the surfaceand the interior. The subsequent internalcontraction puts the surfaces into a great stateof compression, resulting in a mechanicallyunstable condition. Thus, unless glass iscooled slowly (annealed), it will contain inter-nal (frozen) strains which may cause it toshatter spontaneously (Lillie, 1936).

An extreme case of frozen strains in glass isthat of Prince Rupert’s Drops (LacymaeBatavicae; Larmes de Verre; or Tears Glass).The tadpole-shaped pieces of glass werenamed after Prince Rupert, a nephew ofCharles I of England, who produced the glassdrops in 1661 (Moody, 1988). They are madeby dropping a gather of molten glass (notmerely hot glass), into cold water. The suddenchilling of the glass by the water freezes theoutside, while the fluid inside contracts sostrongly that a space, containing a vacuum,not an air bubble), forms in the centre. Thecompressed outside will resist blows with ahammer, but the breaking of the tail, or evenscratching of the surface, will cause the wholeobject to shatter.

AnelasticityGlass is also described as anelastic, because itpossesses internal friction, and absorbs energywhen vibrated. Thus, when a glass vessel islightly struck the walls can vibrate and mayemit a musical note. The vibrations die awaybecause the alkali metal ions in the spaces ofthe silicate network absorb energy when theyjump from one vacancy in the network toanother, producing internal friction. There aregenerally two absorption peaks, the one at thelower temperature being due to the motion ofthe alkali ions in the network whereas thesecond one, at a higher temperature, is associ-ated with the diffusion of oxygen ions(Mohyuddin and Douglas, 1960). Differentalkalis have different temperatures at whichthe first peak occurs; thus lithium ions havethis peak at about –50° C, sodium ions absorbenergy at about –20°C, and potassium ions atabout +30°C. However, at room temperatures,i.e. below 30°C, potassium ions move easilyand less energy is absorbed, so that themusical note can be heard for longer;potash–lead–crystal wine glasses can ring fora second or so, when lightly struck. In the


case of sodium ions, the energy is absorbedat room temperature and below, so that theglass does not ring when struck.

Thermal expansionThe vast majority of materials expand whenthey are heated. Glasses have a somewhatsmall coefficient of linear thermal expansion inthe range 0.5–1.0 � 10–7 per degree C, whichcan actually be calculated from their chemicalcomposition. Silica itself has the lowest expan-sion (with a value of only 0.05 in terms of thevalues given above) whereas the majority ofthe other constituents have values in theregion of 1.7, except for the alkali oxides,which have by far the largest contribution,being 4.32 for soda and 3.90 for potash. Thusthe thermal expansion of a glass dependsgreatly on the amount of alkali oxide in it. Intheory therefore, ancient glasses will havehigher rates of expansion than modern glass.In particular, the low silica, high lime, highpotash medieval glasses will have about twicethe expansion of modern soda–lime glasses.

Transition point (Tg)Figure 1.10 shows a representative thermalexpansion curve (curve A) for a glass whichhad been chilled suddenly after forming,before it had had time to adopt the somewhatmore ordered structure of the glassy state.When heated, it has the large expansion value(0.8 � 10–7) typical of a liquid (having disor-dered molecules), At about 500°C themolecules have achieved enough freedom to

become more ordered, and the expansionfalls, until the random (liquid) state has beenfully reached. Curve B represents a well-annealed glass, well below the glass transitiontemperature (see also the discussion of fictivetemperature, earlier in this chapter). It can beseen that the curve has a lower starting value(0.2) and a fairly constant slope (both charac-teristic of a solid) up to a temperature of about580°C. There is then a relatively suddenincrease in expansion to values that corre-spond to those of a liquid, as the structurebecomes more random.

Optical propertiesApart from certain single crystals, such as rockcrystal, naturally occurring solids are not trans-parent, transparency being more a characteris-tic of a liquid, than that of the solid state.Glass being amorphous is more akin to aliquid, which is structurally the same as anindefinite molecule. Ordinary glasses transmitvisible light and also some ultra-violet andinfra-red light (to which they are transparent).If the wavelengths (i.e., the frequencies) of theincoming radiation are in resonance with thefrequencies of the molecular vibrations withinthe glass, the radiation is absorbed and theglass is said to be opaque.

Glass also has unique optical properties. Forexample it can transmit images in an enlargedor diminished form, or invert them. A brokenor cut glass surface can reflect light in thecolours of the spectrum, (when the glasscauses the light to rebound from its surface).Glass actually reduces the velocity of lightwhich travels through itself, and hence aconvex piece of glass can cause the emerginglight to appear as if it had come from a differ-ent direction (i.e. the light is refracted). Thisrefractive effect is measured by the refractiveindex (RI) of the glass, and characteristicrefractive indices are listed in Table 1.2.(Technically, the RI is calculated from the ratioof the sine of the angle of the incident ray tothe sine of the angle of the refracted ray, whenthe light is refracted from a vacuum.)

The index of dispersion of a transparentmaterial is a measure of the extent to whichthe RI changes with the wavelength (colour)of the light; for example, it determines thewidth of the spectrum produced by a prism ofthe material in question. Also, the image


Figure 1.10 Thermal expansion of (A) a chilledsample and (B) an annealed sample of the same glass.

produced by a simple lens can be colouredbecause it also acts slightly as a prism, but theeffect can be eliminated by making acompound lens from two pieces of glass,having different dispersions. If the composi-tion of the glass is known, the index of disper-sion can be calculated. Thus it tends to becorrelated with the RI, and a cut lead crystaldrinking glass is attractive because it has bothhigh refraction and high dispersion.

A knowledge of the RI may be relevant inthe conservation of transparent glasses. Whenjoining two pieces of glass the RI of theadhesive should ideally match that of theglass, and the join would then disappearcompletely from view (see Tables 1.2 and 5.2;Figure 5.1). In the case of some ancientglasses the RI would have to be speciallydetermined, and the cost of doing that mighttherefore have to be considered.

DensityThe density (mass per unit volume) of glassescan fall within a very wide range, from 2400to 5900 kg m–3, depending on their composi-tion (Figure 1.11), being related to the RI.Certain glasses containing lead have a veryhigh density. SI units tend to be cumbersome,and hence it is useful to refer to the specificgravity (i.e., the relative density, compared towater, where the density is 1.000). Scholes(1929) lists density factors for soda–lime–silicaglasses. Huggins and Sun (1946) showed howthe density can be calculated from the chemi-cal composition of the glass.

HardnessThe property of hardness cannot be definedeasily, because it depends on several otherproperties of the material (whether it is also

brittle, elastic, plastic, etc.). A useful referenceis the Mohs scale of hardness, which is basedon the fact that each material is softer (andtherefore scratched by) all others harder thanit (i.e. having a higher number in the scale):1, talc; 2, gypsum; 3, calcite; 4, fluorite; 5,apatite; 6, orthoclase; 7, quartz; 8, topaz; 9,corundum; 10, diamond. Depending on theircomposition, glasses occupy positionsbetween 4.5 and 6.5 on the scale. The termshard and soft can, however, be used in otherways in connection with glass. High-leadglasses are sometimes called soft because theyare easier to cut and engrave. Hard glass canalso refer to that which does not stain easilywith silver.

BrittlenessGlass is brittle and fractures easily but, whenit is newly formed, and has a perfect surface,it is extremely strong due to the nature of itsinter-atomic bonding. In practice, however,defects arise very easily on the surface merelyby the action of atmospheric moisture, or fromextremely slight abrasion, or even by slightpressure (Figure 1.12). These defects concen-trate any applied stress at the apex of thedefect (Figure 1.13) in a way that is extremely


Figure 1.11 Graph showing the relationships betweendensity and refractive index for various types of glass.Point H is the poorly durable glass H in Figure 4.18.

Table 1.2 Comparative refractive indices ofsome transparent materials

Material Refractive index

Diamond 2.4173Glass 1.5–1.7Quartz (fused) 1.458Ethanol (at 25°C) 1.359Water (at 25°C) 1.332Air (at 0°C and 760 mm) 1.000 293

damaging. Under such stress, the strong bondsbreak and fracture occurs, so that the effectivestrength of glass in tension is only about one-

hundredth of the theoretical strength. Thisability of glass to fracture easily has been putto use since ancient times, by chipping andflaking obsidian and lumps of cold, solid glassto form artefacts.

Fractures on glass can be visually analysedto determine their origins and the directions inwhich they were propagated. Fractures thatoccurred rapidly, at about 2 km/s are easier tostudy than those that propagated at a rate ofonly a few millimetres per century. At theactual origin of a recent crack the broken edgebears a characteristic mirror area which issurrounded by grey areas, hackle marks, and,finally, rib marks, which indicate the directionin which the fracture travelled. Murgatroyd(1942) observed that rib marks are alwayscurved, and that their convex faces show thedirection in which the crack grew. If the glasshas broken due to excessive heating, the ribmarks are well spaced on the cold side, butare crowded together on the heated side. Ifthe outside of a vessel has been given a sharpblow, the area which received the blow maybe crushed, with a surrounding ring of cracksforming an impact cone.


Figure 1.12 Section through impact cones on damagedglass.

Figure 1.13 Diagnostic markings on the edges offractured glass.

The natural glass obsidian occurs in allvolcanic regions of the world, and sincePalaeolithic times was fashioned into tools,weapons and objects of trade, by primitivepeoples (see Figure 1.1). On the basis ofchemical analysis of obsidian artefacts, and ofmaterial from volcanic flows, it has been possi-ble to assign a provenance to many artefacts;and to determine the trade routes along whichobsidian artefacts were disseminated. Obsidianis highly durable glass, and consequently doesnot at present pose any conservationproblems.

The date and place of origin of man-madeglass may never be known precisely; but it isgenerally agreed to have originated in north-ern Mesopotamia (Iraq) prior to circa 2500 BC.However, in ancient times, the mouth of theRiver Belus in Phoenicia (now the RiverNaaman in Israel) was associated with glass-making for many centuries. The association ofthe River Belus with glassmaking, wasmentioned by the Roman historian Pliny (AD

23–79), who drew much information fromGreek sources (themselves a mixture of first-hand information and legend; Greek merce-naries, travellers and writers were visitors tothe eastern Mediterranean from the seventhcentury BC). The account by Pliny (AD 77)concerning glassmaking has been somisquoted, that it is given here in full:

That part of Syria which is known asPhoenicia and borders on Judea contains aswamp called Candebia on the lower slopesof Mount Carmel. This is believed to be thesource of the River Belus, which, after travers-ing a distance of five miles, flows into the sea

near the colony of Ptolemais (Akko). Itscurrent is sluggish and its waters unwhole-some to drink, although they are regarded asholy for ritual purposes. The river is muddyand flows in a deep channel, revealing itssands only when the tide ebbs. For it is notuntil they have been tossed by the waves andcleansed of impurities that they glisten.Moreover, it is only at that moment, whenthey are thought to be affected by the sharp,astringent properties of the brine, that theybecome fit for use. The beach stretches fornot more than half a mile, and yet for manycenturies the production of glass dependedon this area alone. There is a story that oncea ship belonging to some traders in naturalsoda put in here and that they scattered alongthe shore to prepare a meal. Since, however,no stones suitable for supporting theircauldrons were forthcoming, they rested themon lumps of soda from their cargo. Whenthese became heated and were completelymingled with the sand on the beach a strangetranslucent liquid flowed forth in streams; andthis, it is said, was the origin of glass. (Engle,1973a; Newton, 1985b)

According to the Roman historian Josephus,‘numbers of ships are continually coming totake away cargoes of this sand, but it nevergrows less’. Similar statements were madeabout the sand at the mouth of the RiverVolturnus (north of Naples in Italy).

Analysis of the sand from the River Belushave confirmed its substantial lime content(8.7 per cent CaO), which would enable stableglass to be made in the absence of any instruc-tion to add lime, (which was not actually

16

2

Historical development of glassmaking

specified as an ingredient for glassmaking untilcirca AD 1780).

Throughout historical time man-made glasshas been regarded as a special material, andit is not difficult to see why this should havebeen so, since glass must have seemed to havehad magical origins. To take sand and plantashes and, by submitting them to the trans-muting agencies of fire, produce colouredliquids which, whilst cooling, could be shapedinto an infinite variety of forms and textures,which would solidify into a transparent mater-ial with the appearance of ‘solid water’, andwhich was smooth and cool to the touch, was,and still is, the magic of the glassworker’s art.Glass can be fashioned into many shapes inways that are not possible with any othermaterial. It has unique optical properties: forexample, glass can transmit images in anenlarged or diminished form, or invert them;a broken or cut glass surface can reflect lightin the colours of the spectrum. Certain typesof glass are especially appealing, particularlylead crystal glass by virtue of its weight, itsgreat clarity, its ring when lightly struck, andwhen cut, the sparkle and colours which ariseas a result of its high refractive index anddispersion.

In consequence of the supposed magicalproperties of glass and the technologicalsecrets associated with its production, glass-makers were often granted a higher socialstatus than was given to other craftsmen; andfrom time to time throughout history, speciallegislation was passed for their benefit. Inancient Egypt for example, glass was regardedas being more precious than gemstones.During the first phase of the Roman Empire,when the best glass was being made in Syria,the Syrian glassmakers were regarded as CivesRomani (Roman citizen). Once glassmakinghad been established throughout the Near Eastand the West, measures were being taken tosafeguard the technological secrets of thetrade. For instance, in medieval France glass-making methods could only be passed onthrough the male line, and then only betweenmembers of a few specific families such asHennezal, Thietry, Thisac and Bisseval. In1369 Duke John I of Lorraine granted lettersof privilege to glassmakers to encourage themto settle in Lorraine; and in 1448 Jean deCalabre granted a charter to the makers of

glass in the Forest of Darney in the Vosges.The Italian city of Venice became an impor-tant glass centre in the middle of the eleventhcentury when glassmakers from Constanti-nople settled there to make the mosaics forSan Marco. The glassmakers of Venice eventu-ally became so powerful that they were ableto form a guild in 1220; emigration of guildmembers was forbidden on pain of death(Forbes, 1966).

Another privilege, this time for glassvendors, existed in England in 1579, wherelaws were in force against rogues andvagabonds, but ‘glass men of good behaviour’were exempt from prosecution if theypossessed a licence from three justices of thepeace (Charleston, 1967). The restriction of thesecrets of glassmaking to certain specifiedfamilies, or craft communities, has led to theperpetuation of glass terminology which hasbeen handed down not merely over genera-tions, but over centuries. Glassmakers alongthe Phoenician coast in the first to sixthcenturies AD were using terms similar to thoseused in Babylonia in the seventh century BC

and, following a study of sixteenth-centuryItalian glassmaking texts, Engle (1973b)suggested that some of the early glassmakingfamilies of Europe may have originated inareas where Aramaic was spoken. In addition,family names in Hebrew, Flemish, French andEnglish have been studied with a view totracing the relationships between glassmakingfamilies as they emigrated from Asia Minorthrough Sicily, Lombardy, the Rhineland andLorraine to Britain (Engle, 1974). Freestone(1991) gives an account of glassmaking fromMesopotamian to medieval times. Histories ofglassmaking have been produced by Tait(1991) and by Liefkes (1997).

The first glassmakers, 2500–1200 BC

Most authorities claim northern Mesopotamia(Iraq) as the birthplace of glass at a time priorto 2500 BC. Comparatively few glass objectshave been excavated there, but this may wellbe due to the relative humidity of the soil, andto the rise of the water-table in historic times,causing the destruction of much of the earlyglass which was inherently unstable in itschemical composition (and therefore relatively

Historical development of glassmaking 17

water-soluble). However, it is known fromobjects which have survived burial, thatcoloured vitreous glazes were extensively usedin the Jemdet Nasr phase of Mesopotamia, theBadarian civilisation of Egypt, and in the earlyAegean in the fourth millennium BC, for cover-ing steatite and sintered quartz beads in imita-tion of semi-precious stones such as turquoise,lapis lazuli and red jasper; later glass beadswere developed for the same purpose. Smallobjects could be hand-formed or cast usingsimple tools and finished by abrading. Fewglass items are known until the first core-formed vessels were made in western Asiasometime before 1500 BC. The Mesopotamianevidence was summarized by Moorey (1994).

The development of core-forming was thetechnological breakthrough which producedthe first glass vessels, and which therebyallowed glassmaking to become an industry inits own right. This may have developed fromthe technique of winding glass around a coreto form glass beads, but the connection isunproven. Not long after the core formingtechnique was discovered, polychrome vesselsbegan to be made of mosaic glass (datingmainly from circa 1350–1250 BC). These wereformed of pieces of monochrome opaqueglass, fused together and subsequently shapedaround a core or possibly slumped over orinto a form. Fragments of mosaic glass recov-ered from a palace site to the west of Baghdadwere made of sections of multi-colouredmosaic canes. Inlaid panels from the same sitewere formed by pressing turquoise blue andwhite glass into a red glass base whilst theglass was still in a pasty state, to form patternsand birds. Occasionally marbled glass wasproduced in imitation of veined stone.

Contemporary with the core-formed andmosaic glass vessels are a wide variety ofmonochrome or polychrome objects, includingbeads of many different types, jewellery inserts,plain and decorated pendants, furniture inlaysand figurines of deities, demons and animals.Many of these were made in moulds, but thereis no contemporary evidence to show whetherthe glass was poured into open moulds orwhether moulds were pressed down ontolumps of soft glass on a flat surface.

During the later sixteenth and fifteenthcenturies BC, glassmaking evolved rapidly innorthern Mesopotamia. Mesopotamian glass

vessels have been excavated over a wide areaof the Middle and Near East: Persia (Iran),Elam and Babylonia in the east to Syria andPalestine on the Mediterranean coast; and atother centres of Late Bronze Age civilizationin Cyprus and Mycenaean Greece.

During this period, the Levant played animportant part in the trade in raw glass andin finished products. The Levant was the areastretching from ancient Antioch (Antakya inmodern Turkey), down the coast of Syria,Phoenicia (modern Lebanon), Palestine/Israel,and included the island of Cyprus. Very fewglass vessels have been found on Late BronzeAge (Mycenaean) sites in Greece. Exclusive tothat area however, are ornaments of translu-cent glass, mainly bright blue, and normallywith flat backs and suspension holes, anddating from 1400–1200 BC (Nightingale, 1998).The almost exclusive use of bright blue glasssuggests that it was imported, probably fromEgypt, as analysis has shown that the compo-sitions of the Mycenaean glass is the same asthe blue glass being used in Egypt at that time(Shortland, 1999). The blue glass wassometimes used in combination with gold foil.Steatite moulds in which the ornaments weremade by pressing the glass into them, havebeen found on many sites. Glass, ivory andgold were used as inlays for luxury items ofpersonal ornamentation, palace furnishingsand weapons.

The Egyptian glassmaking industry began inthe fifteenth century BC, about the same timeglass starts to be mentioned in Mesopotamiancuneiform tablets. From circa 1450 BC theEgyptian pharaoh Tuthmosis III made militaryconquests in Syria and up to the Meso-potamian borders, and it is possible that as aresult of this contact, Asiatic glassworkers weresent to Egypt to found the glassmaking indus-try there. Glassworking complexes were estab-lished at Malkat in the early fourteenthcentury, at Tell el-Amarna, the new capital cityof Akhenaton (Amenhotep IV, c. 1352–1336BC) and at el-Lisht, an early twelfth-centurynecropolis. Glass was not produced in anyquantity until the reign of Amenophis IIIonwards (c. 1390 BC). This is far later than inneighbouring countries, which is surprising inview of the Egyptians’ mastery of manufactur-ing techniques. Individual glass beads, proba-bly manufacturing aberrations of glazed


composition, were made a thousand yearsearlier, and a few scarabs are known fromcirca 1900 BC. Moreover, the basic material –an alkaline calcium silicate– is the same as thatof the glaze produced in pre-DynasticBadarian period (c. 4000 BC) to coat stonebeads, and later in the manufacture of glazedcomposition. The only difference is that glasswas not used to produce objects in its ownright. The Egyptian term for glass was iner enwedeh or aat wedhet, both meaning ‘stone ofthe kind that flows’.

One of the earliest glass vessels known is asmall turquoise blue jug from the tomb ofTuthmosis III, with an elaborate yellow andwhite patterning of stylized tamarisk trees,threads, dots and scales incorporating a hiero-glyphic text with the prenomen Menkheperre(British Museum, London) (Figure 2.1).

Egyptian glass is the most common typeknown from this period, many exampleshaving been found in the tombs of theEighteenth (1570–1293 BC) and Nineteenth(1293–1185 BC) Dynasties. The vessels aresmall and served mostly for holding perfumesand ointments or as tomb gifts and cultobjects, and copy the shapes of contemporaryvessels of pottery, stone and faience. Theserichly coloured vessels are almost opaque, dueas much to the desire to imitate semi-preciousstones in glass as to the technological limita-tions. Core-forming persisted as an importantglassmaking technique for many centuries.

The glassmaking industry reached its peakin the mid-fourteenth century, both in westernAsia and Egypt, but continued to flourish andspread until circa 1200 BC. For all practicalpurposes glassmaking then came to an endwhen Egypt and Syria were invaded by thePhilistines. With the downfall of the variouskingdoms under the impact of the invaders,there was no longer a market for the fine andexpensively produced glass articles. There isan almost total absence of glass finds from theend of the second and the beginning of thefirst millennia BC.

This phase in ancient history is marked bythe eclipse of the great empires and theemergence and migratory movements of newpeoples and tribes, in the Aegean and NearEast. Not until the resurgence of the greatempires in the eighth and seventh centuries BC

was there again the necessary stability andconcentration of wealth and resources for therenewed production of glass. Yet glassmakingexpertise must have continued somewhere,because the re-emergence of the demand inthe eighth century BC, brought about themanufacture of articles by all four of theearlier techniques with increasing degrees ofsophistication in Egypt, Mesopotamia andelsewhere

Western Asia and the Mediterraneancirca 900–300 BC

The resurgence of the glass industry in theninth century BC took place against abackground of cultural revival that affected thewhole of Western Asia, the Levant and theMediterranean world. The earliest use of glass


Figure 2.1 One-handled jug bearing the name ofTuthmosis III (c.1504–1450 BC). Opaque light blue, withyellow, white and dark blue opaque trails, and whiteand yellow powdered glass fired on. Core-formed, withground and polished surface, on rim and underneaththe base. Intact and unweathered; some bubbles andsandy impurities in the glass. H 88 mm, GD 38 mm.Second quarter of the fifteenth century BC. Egypt. (©Copyright The British Museum).

on a large scale in the Iron Age was as inlays,often in ivory plaques and panels used todecorate furniture. The glass inlays were eithermonochrome, different shades of blue as wellas red, green or yellow, sometimes with coldpainted or possible enamelled designs, or ofpolychrome glass forming rosettes, circles andsquare patterns. Glass inlays in ivory plaqueswere all of monochrome glass, and most havebeen assigned to craftsmen in Phoenicia(Lebanon) on stylistic grounds. In the tenth oreleventh century BC, glass beads were beingmade in the delta of the River Po, showingthat glass technology had reached Italy.

Vessels, also of monochrome glass, beganto be made around the middle of the eighthcentury BC, and were made by the lost waxmethod, or the technique of slumpingsoftened glass into moulds. Polychrome glassvessels were made by the core-formingtechnique, but although mosaic inlays weremade, mosaic glass vessels were very rare until

the late third century BC. A class of luxuryvessels in greenish or yellowish or naturalgreen monochrome glass was produced at thistime, possibly in Phoenicia. Drinking vessels,mostly in the form of hemispherical bowls,were made in the eighth and seventh centuriesBC. These were probably made by the slump-ing process and undecorated or with simpledecoration of horizontal cut grooves or ridges,or rarely, with geometric patterns or glassinlays. A group of tall perfume flasks (alabas-tra) were probably made by the lost waxprocess and shaped by grinding and abrading;a squat example bearing the name of theAssyrian king Sargon II (721–705 BC) is one ofa series which were produced during theseventh, sixth and possibly the fifth centuries.

In the mid-eighth century BC, core formingwas revived in Mesopotamia, most notably inthe form of alabastra, but the products weredull compared with those produced in theBronze Age. Mesopotamian core-formed


Figure 2.2 Core-formed vessels for cosmetics and scented oils from Mediterranean workshops operating between550 and 50 BC, together with an earlier Mesopotamian example (front). (© V&A Picture Library).

vessels reached other countries, notably Persia(Iran) where they seem to have led to theestablishment of a local industry at Susa; theisland of Rhodes (Greece); and Italy (Etruria)in the seventh and sixth centuries BC.

The core-formed products of Mediterraneanworkshops in production circa 550–50 BC

were the most numerous and widespread.Shapes were copied from Greek vases inpottery and metal, the most common formsbeing alabastra, amphoriskoi, araballoi andoinochoai (jugs) made of dark blue glassdecorated with white, yellow and turquoiseglass trailed, combed into patterns of zigzags,festoons or feather patterns (Figure 2.2). Thevessels were used as containers for perfumes,scented oils and cosmetics; and were widelytraded, as far as the Black Sea, the Balkansand Gaul (France). The final flowering of theMediterranean core-forming industry tookplace in the late Hellenistic period, betweenthe second half of the second century and themid first century BC. Only the alabasta andamphoriskoi were made and these weresmaller than those produced earlier. Themajority have been excavated in Syria,Palestine and Cyprus, where they were proba-bly made. Others were imported into Egypt,where a renaissance in all branches of the artstook place during the Saite Twenty-SixthDynasty (c. 664–525 BC). The technique ofinlaying glass into another material re-emergedduring the reign of Amasis (c. 570–526 BC).During the fifth and fourth centuries BC, cleargreenish or colourless glass bowls with cutdecoration copying metal vessels, were madein the Persian Empire. Some may have beenproduced in the western provinces in AsiaMinor (Turkey). In the fifth and fourthcenturies BC yellowish and greenish clear glasswas also being made in Greece. Excavationsin Olympia in the workshop of the Greeksculptor Phidias show that glass was being castinto clay moulds.

Hellenistic glass, circa 325 BC to AD400

During the Hellenistic Period (late fourth tosecond century BC), new shapes and decora-tions were introduced into core manufacture,although there was a decline in aesthetic

quality, and in the production of glass inMesopotamia. Contemporaneously, there weremajor developments in glassmaking, both fromtechnical and artistic points of view, notablyof engraved gold leaf enclosed between twolayers of glass, mosaic and cameo glasstechniques. In this period there appear thehemispherical mould-cast bowls made oftransparent, almost colourless glass, in theAssyrian tradition. These bowls were lathe-finished and mostly decorated with mouldedand/or cut ribs and lines in imitation of theirmetal prototypes (Figure 2.3). Outstandingamong this type of bowl are the sandwich-gold glass vessels dating from the late thirdcentury AD (Figure 2.4). These were formed oftwo glass bowls enclosing gold leaf decora-tion, which were ground and polished withsuch precision that the outer glass fitted soperfectly over the inner that no adhesive orfusion was required to hold them together);others may have been fused at the rim. Onlya few sandwich bowls have been found, andalthough distributed over a wide area, fromthe north Caucasus, central Anatolia and Italy,


Figure 2.3 Deep bowl with band of bosses. Greenishcolourless glass, now with an iridescent and flakingsurface. Cast in a two-piece mould and finished bycutting and grinding; the bosses are in relief and theremainder of the design is in antaglio. H 92 mm, D205 mm. Late third century BC. Canosa, Apulia, Italy. (©Copyright The British Museum).

it is generally accepted that they were madein Alexandria.

The technique of producing mosaic glasswas difficult and complex, the required designbeing built up from canes of variouslycoloured glass into a slab of material (seeFigure 3.10). When heated and pulled fromboth ends the slab could be drawn out into along rod, which retained the original sectionaldesign in miniaturized form along its wholelength. The rod was then cut into smallsections in which the design recurred eachtime. The discs were used as inlays for wallsand furniture, or fashioned into beads andvarious kinds of jewellery. Also, as mentionedabove, patterned sections were arranged inmoulds and fused together. The resultingvessels, mostly small cups and bowls, transmitlight with a polychrome brilliancy (Figure 2.5).Sometimes sections of coloured rods werefused together to create variegated patches inthe body of the glass; or thin threads of glasswere twisted into rods which were then fusedtogether in moulds to form elaborate vesselsof lace glass.

Vessels and plaques made by the cameotechnique were composed of two or morelayers of glass. The upper layers were then cutaway to reveal the base colour, which thenformed a background to the relief design ofmythological figures, vine leaves and othermotifs of Hellenistic art. However, much of thecelebrated cameo glass, such as the PortlandVase and Auldjo Jug (British Museum,

London), is of the early Imperial period, datingfrom the late first century BC/early first centuryAD) (see Figures 3.19 and 7.20).

All the techniques mentioned above arethought to have been either invented orperfected in Alexandria, the cultural andindustrial centre of Hellenistic civilization,founded by Alexander the Great in 332 BC.Despite the considerable information aboutAlexandria as a glass centre, only a smallamount of glass has been found there. Thescarcity of glass in places where glass musthave been abundant seems to be due to thefact that broken glass was often collected andre-melted as cullet to form new glass batches.

There were other important glass-manufac-turing centres in this period, some with longtraditions of glassmaking. Those mentioned bythe Roman historian Pliny the Elder includeSidon in Lebanon, Acre and the area aroundthe mouth of the River Belus north of theMount Carmel range in Israel, Campania inItaly, Gaul and Spain. In addition toAlexandria, the Roman historian Strabomentions glassmaking in Rome; and the firstcentury poet Martial refers to a hawker from


Figure 2.4 A bowl of sandwich gold glass. Canosa,Apulia, Italy. Found in a tomb with seven other vessels.Circa 275–200 BC.

Figure 2.5 Segment of a large plate of mosaic glass,formed from sections of multicoloured canesinterspersed with segments of yellow, opaque white oroccasionally gold foil sandwiched between twocolourless layers. H 50 mm, D 308 mm. Late thirdcentury BC. Canosa, Apulia, Italy.

across the River Tiber (in Rome), who barteredsulphur matches for broken glass.

Glass of the Roman Empire, AD100–400

The invention of glass-blowing

Around the turn of the millennium, glass-blowing was invented, probably in the Syrio-Palestinian area long associated withglassmaking. Despite the fact that glass-blowing revolutionized the production of glassvessels, no mention was made of it bycontemporary writers. Glass-blowing turnedglass into a cheap commodity, which could bemass produced; and no doubt provided thestimulus for the proliferation of glasshousesthroughout the Roman Empire.

At its height, the Roman Empire includedthe countries which are now the UnitedKingdom (except Northern Ireland), France,Spain, Portugal, parts of the Netherlands,Germany, Belgium, Switzerland, EasternEurope, Greece, Turkey, the Middle East andNorth Africa. Thus all the major glassmakingcentres came under the domination of Rome.In addition the art of glassmaking was spreadand important centres established throughoutthe Empire. However, the glass productionremained essentially Roman, with only minorregional variations until the collapse of theRoman Empire in the West soon after AD 400(Lemke, 1998). Thus glass dating from the firstto the fourth centuries AD may more accuratelybe described as Roman than, for instance,Spanish or Gallic (Harden et al., 1968; VonSaldern, 1974, Tait, 1991). Glass ceased to beexclusively a luxury product, the stylesbecame largely simple and functional, and infact glass became more widely used fordomestic purposes during the Roman periodthan at any subsequent time or place until thenineteenth century. Glass containers wereparticularly valued as shipping and storagecontainers because they were light, transpar-ent, reusable and did not impart a taste totheir contents.

Glasses were packed in straw to survivelong journeys by land and sea. Some contain-ers were square-shaped for easier packing(Figure 2.6). Besides the utilitarian glassware,

mould-blown bottles were widely made, infanciful shapes such as animals, human heads,fruit, sea-shells and as souvenirs of gladiator-ial contests. Some glassmakers incorporatedtheir names in their moulds, the best knownbeing that of Ennion, a Sidonian whoemigrated to Italy (Harden, 1969a). At thesame time that utilitarian glassware wasbecoming commonplace, some of the mostlavish glass ever made was being produced,for example, the gold-sandwich glasses. ManyRoman glassworkers sought to imitate rockcrystal with clear glass, and other semi-precious materials. Layered stones, such asthose used for producing cameos, wereimitated in glass and carved in high relief.Techniques of cold painting, enamelling andgilding on glass were also highly developed.Other vessels were decorated with scratchedor wheel-abraded designs. Other products ofthe Roman glasshouses were jewellery,window-panes, lamps, mirrors, mosaictesserae, cast glass panels imitating jasper,porphyry and marble, and opus sectile (panelsmade up of flat glass pieces and set in mortar).A survey of glasses taken to be lenses hasshown that their focal length was too short tohave improved sight; their most probable usewas as magnifying aids for engravers.

In the third century, glassmaking reached apeak, both in quantity and quality of products.During the third and fourth centuries Egypt


Figure 2.6 A mould-blown square bottle of the typecommonly used to transport liquids, later first or secondcentury AD; a blown triple-bodied flask, probablythird–fourth century AD; and a mould-blown, barrel-shaped jug, third century AD. All made in Westernworkshops: the bottle and jug were found atFaversham, Kent, South Eastern England. H of bottle20 cm.

also had a considerable blown glass industry,which had not existed there previously. Alarge number of blown glass vessels with localstylistic features such as the fashioning of thebases, was found in the excavations at Karanis(Harden, 1936). It is interesting to note thatthe Emperor Aurelian (AD 270–275) hadimposed a duty on Egyptian glass imported toRome, presumably to offset its cheapness. Thesuccess of the industry meant that it becamesubject to heavy taxation at various times. TheEmperor Alexander Severus (AD 222–235)imposed taxes on all artisans. In the followingcentury, the Emperor Constantine (AD 337)eased the burden of taxation in order that thevitrearii could perfect their skills and bring uptheir sons in the family crafts.

Until the turn of the third century AD thereis evidence of strong continuous links betweenglassmakers in the Middle East and the West,largely formed as a result of the migration ofworkers, mainly in the east to west direction.Contemporary literary sources mention Syrianglass manufacturers working in the Romanprovinces; and glassmakers’ quarters wereestablished in every large city. During the firstcentury AD glass-blowing was introduced tothe glassmaking district of Campania (theprovince around Naples); and many blownvessels have been found at Pompeii andHerculaneum, both of which were destroyed

by the eruption of Vesuvius in AD 79. Theaccurate dating of ancient vessels is oftenmade difficult by the fact that much of it doesnot have a recorded provenance and was notrecovered from excavations. However, thewealth of glass objects found in use inPompeii at the time of the eruption, shows therepertory of glass vessels current in the thirdquarter of the first century AD. (Much of theglass from these sites has been recovered fromthe cemeteries, and is therefore much olderthan that buried during the eruption ofVesuvius in AD 79.) New glassmaking centresarose in the north of Italy in the valley of theRiver Po, and at Aquileia on the Adriatic coast.From northern Italy, glass was exported as faras Britain.

A group of distinctive vessels appeared inthe fourth century AD. These were the polyg-onal bottles (mainly hexagonal and octagonal),either without handles or with a single handle,and bearing moulded symbols on the sides.The most familiar and prominent of thesesymbols was the seven-branched candlestick(menorah) of the Jewish faith, while otherswere an arch supported by two columns(apparently symbolizing the Temple portals),palm trees and branches, and other designs ofuncertain significance. Although the exactprovenance of the polygonal bottles isunknown, it is generally supposed that they


Figure 2.7 Fondi d’oro bowlfragment with emerald greenblobs, and with golddecoration within the innerfaces of the blobs and theouter surface of thecolourless glass bowl.Greatest dimensions: 10 mm(smaller portion), 168 mm(larger portion). Second halfof the fourth century AD.From St Severin’s parish,Cologne. (© Copyright TheBritish Museum).

were first produced in Palestine. Bottles almostidentical in shape but with Christian symbolsare also known. Apparently both types ofvessel were made in one workshop butprovided with different symbols according tothe religion of the customer.

Other glass objects with religious symbolswere the gold-glass bases (Ital. fondi d’oro) inwhich a gold leaf etched or painted with adesign was enclosed between two layers ofglass (Figure 2.7). The technique was popularin Romano-Byzantine times, and was usedalmost exclusively for religious iconography,both Jewish and Christian. (Many gold-glassvessels were embedded in the walls of thecatacombs outside Rome, where they acted asgrave markers.) Religious symbols also appearon a category of objects of a personal charac-ter, such as bracelets and amulets, stampedwith representations of menorah, lions, frogs,human masks, and also elaborate scenes andinscriptions.

Fashion and innovations spread with thecontinuous traffic of glassmakers with theresult that types of glass originally made in the East began to be produced in the West.

Especially noteworthy are the two groupsmade from the second century onwards in theRhenish centre of Cologne. One of theseincludes vessels with cut and engraved decora-tion. The other group bears the type ofdecoration known as snake thread trailing(Figure 2.8), which began to be made in Syriain the late second century and then, about onehundred years later, appeared in a somewhataltered form in the Rhineland and in Britain(Figure 2.9), the Western examples oftenbearing trailed decoration of a different colourfrom that of the body of the vessel (Harden,1969b).

By the middle of the fourth century, nodoubt as a result of the division of the RomanEmpire, East–West contact effectively ceased,and the different glassmaking centres devel-oped their own glass styles. Glassmaking thusbecame less international, and more provin-cial, so that regional types of mould-blown,cut- and thread-decorated glasses are foundwithin a limited range of distribution. Forexample, the Syrian double unguent bottle islater than fourth century, and is not found inthe West. In due course the regional styles


Figure 2.8 Flask of greenish glass, with blue enamel-like weathering and flaking. On the body, three windingapplied ‘snake’ coils, flattened and bearing a criss-crossdesign, ending in a triangular head. H 155 mm, D (rim)30 mm, D (body) 81 mm. Late second century AD.Idalium, Cyprus.

Figure 2.9 Flask of greenish colourless glass, appliedcoloured threads on the body. Similar to that shown inFigure 2.8, but found in the Rhineland. H 213 mm.Third century AD. Cologne.

developed into the glass types of the Teutonicnorth on the one hand, and the Syrian, Iranianand Egyptian styles of the Islamic period onthe other.

Roman Gaul had a flourishing glass indus-try; some glass was already being made inGaul before the influx of Sidonian andAlexandrian immigrants. One of the Gallicfactories made cylindrical bottles, which werestamped on the base with the name Frontiniusor its abbreviated form FRON. Although glasswas imported to Britain during the Romanoccupation, there is archaeological evidencethat it was also manufactured locally atLondon, Colchester, Wroxeter and Mancetter,on a modest scale. Production would mainlyhave been of simple vessels and bottles, andsome window glass. The industry may nothave survived long after the Roman departure,or it may have continued in isolated areas.

Islamic countries

Gradually the prominent Mesopotamian andSyrian glassmakers established themselvesthroughout the Roman Empire, and were againimportant in the development of MiddleEastern glass, which culminated in the distinc-tive and sophisticated wares of Islam. With thedecline of Rome, the seat of power transferredto Constantinople (Istanbul) in AD 305. Despiteits magnificence and importance, Con-stantinople appears never to have had a tradi-tion of glassmaking. This may be explained bythe fact that since it was so close to the estab-lished Syrian glasshouses of Tyre and Sidon,there was never any great necessity to set upan independent manufacture when the bestglass was so close at hand. It may also be thecase that whatever glass was made inConstantinople, closely followed in the Syriantradition and is not easily identifiable. Glass ofthe period is similar to that found throughoutthe Roman Empire, but during the Sassanianperiod (c. 100 BC to AD 600) leading up to theadvent of Islam, a tradition of cut glass devel-oped. For this purpose the glass needed to bethicker than for the earlier blown andmoulded styles. Cutting generally took theform of facets or geometric patterns and wasdeveloped to a very high standard (Figure2.10).

Glass vessels of the Byzantine period (fourthto seventh centuries AD) demonstrate imagina-tion and great technical skill, but the forms arerather heavy. There is an absence of clear glass,and the coloured glass was not as vivid as hadpreviously been the case, and was generallyimpure. The vessels were irregular in shape andbadly proportioned; the decoration is intricateand over-profuse. Cosmetic vessels in the formof two, sometimes three or even four tubeswere widespread in the Near East. The major-ity of these vessels were found in tombs,usually with the metal spatulas for applying thecosmetics still inside one of the tubes.

Extremely common during this period arethe conical cups, which were used as lamps;these were filled with water and oil on whicha wick was floated. The lamps were placed inholders or suspended by a chain from theceiling. Other lamps were in the form ofstemmed bowls or cups with a hollow projec-tion in the centre to hold the wick. Similartypes, placed in metal holders, were used forlighting in the Middle Ages. Glass was animportant element in mosaics, a major art ofthe Byzantine period. Itinerant mosaicistsdecorated Byzantine churches in RomanRavenna, and in mosques in Damascus andCordoba (Spain) with splendid wall mosaics.The synagogue mosaics in Israel includedmany glass tesserae, especially of colours notfound in natural stone.


Figure 2.10 Bowl with cut decoration, of thickgreenish glass with heavy iridescence. Hemispherical,with a rounded rim and base. Exterior decorated withlarge circular facets in quincunx; four horizontal bandson the side with one large central facet on the base,making it stable. H 75 mm, D 103 mm. Fifth to sixthcentury AD. Persian; said to have been found at Amlash.

After the Arab conquest of the Middle Eastin AD 635, and the establishment of a capitalat Damascus, there was a rapid move awayfrom the Roman traditions of glassmaking. Thechange in the balance of power affected glassproduction, which stagnated until the rise ofthe Abbasid dynasty, and the transfer of thecapital to Baghdad in Mesopotamia (Iraq), inAD 750, which was outside the mainstream ofan area which had been unsettled for manyyears. By this time the whole of the MiddleEast had become settled under the rule ofIslam and new styles in glass slowly began toemerge to suit the tastes of a new society. Inthe early stages of their conquest, the Arabsadopted the art of the countries over whichthey ruled, and had their palaces built anddecorated by local craftsmen. Only at the endof the first millennium AD did Islamic art beginto assume an individual character. As withRoman Imperial art, though to a lesser degree,the development of Islamic art was remarkablyuniform, whether in Persia, Mesopotamia,Syria or Egypt, centres of influence movingfrom country to country in the wake of shift-ing centres of government.

Mesopotamia, important in the ancient glass-making world, again came to the fore; glasskilns were probably more common thanpottery kilns in medieval Mesopotamia andsouthern Persia. Islamic glassmaking centresdeveloped on the Euphrates river east ofAleppo (Syria); at Samara on the River Tigris(Mesopotamia); at Siraf, an early Islamic porton the Persian Gulf; at Nishapur (Neyshabur),an important trading centre in northern Persia;and at Fustat south of Cairo (Egypt) which hadtaken over from the Roman glassmakingcentres such as Alexandria. There was muchemphasis on mould-blown patterns and thecutting, engraving and polishing of glass,followed by pincering with tongs, lustre paint-ing and gilding and enamelling. The moststriking was the cut glass, surviving examplesof which are either linear or facet cut.

A characteristic vessel of the Islamic periodis the mould-blown flask with a globular bodyand long narrow neck. A fine group of suchflasks, dating from the eleventh and twelfth-centuries AD, and typical of the Gurgan districtin north-eastern Iran, is displayed in theHaaretz Collection, Tel Aviv (Israel), besidethe clay moulds in which they were blown.

The relationship between Islamic cut glass andsimilar glass of an earlier period is not clear.The technique of glass-cutting was alreadyknown in the Late Bronze Age and muchpractised in Roman times, but did not reachits peak until the Islamic period. In Iran (andpossibly also in Iraq) a tradition of cutting –from powerful relief work in the form ofbosses, to delicate intaglio figural engraving –developed into a brilliant Persian-Meso-potamian school of relief cutting on glassduring the ninth and tenth centuries. The glasswas mainly colourless, the designs beingoutlined by deep, notched lines. This engrav-ing was occasionally executed on glass casedwith an overlay of emerald green or blueglass. Parallel with this luxurious relief engrav-ing went a simpler or rougher style of intaglioengraving.

Lustre painting was a characteristic form ofdecoration from the eighth century, especiallyin Egypt where it may have originated. Theearliest surviving example of lustre painting ison a glass bowl dated AD 773. This technique,which involved applying pigments, and firingthem under reducing conditions in the kiln, toproduce golden or silver iridescence, probablydeveloped simultaneously in Egypt and Meso-potamia. The surviving examples includefragments on which different hues wereobtained by repeated firings in the kiln, andvessels on which lustre spots have beenapplied to the interior and exterior of theglass.

The art of gilding glass may also have origi-nated in Egypt. Gilding formed the basicelement in the technique of gilding andenamelling glass, which developed in Syria,centred on Damascus, during the late twelfthand thirteenth centuries. The gilt and enamelglasses, largely beakers, bowls, flasks andmosque lamps, made mostly during thethirteenth and fourteenth centuries AD, areconsidered to be the highpoint of Islamic glassart (Figure 2.11). The so-called mosque lampsare in fact lamp-holders in which small glasslamps were placed. The usual shape of amosque lamp (holder) was a large vase witha splayed neck. On the body were small glasslugs to which chains for suspending the lampfrom the ceiling were fastened. Often thedonor’s name was included in the enameldecoration. Two main styles of glass


enamelling are discernible: one using enamelheavily laid on, usually in horizontal bands ofintricate abstract patterns interspersed withquotations from the Koran, the other a finelinear style in red, both on a gold ground.

Glass vessels were exported from Damascusto every part of the Islamic world, and evenas far as China. The beauty of Islamic glasswas already appreciated in medieval Europe.In circa 1025 AD a ship of unknown origin anddestination, carrying a cargo of approximately1 tonne of raw and scrap glass, foundered atSerçe Liman off the south west coast ofTurkey. Excavated by Bass (1980), the shipalso yielded more than eighty intact engravedbeakers and bottles of Islamic manufacture.These were found in the ship’s living quarters,and were perhaps used by the crew, or wereintended for use of items of trade. Vesselsbrought from Syria and Palestine by pilgrimsand crusaders are now in many churches,

monasteries, museums and private collections.Under the influence of Far Eastern Art,

coming in the wake of the Seljuk and Mongolconquerors in 1258, the earlier heavier formsand styles of decoration became freer andmore naturalistic, consisting of arabesque andfloral designs. Enamel work declined at theend of the fourteenth century, and themanufacture of gilt and enamelled ware ceasedalmost entirely after the sack of Damascus bythe Mongol chief Timur (Tamerlaine) in 1402during the conquest of the Middle East. Thegeneral decline in the production of Islamicglass gave Venice the opportunity to expandand to take over markets that had previouslybeen supplied by the East.

Chinese glass

Glassmaking never achieved any great signifi-cance in Far Eastern decorative arts, possiblydue to the development of other materialswith glossy translucent surfaces such as jade,lacquer and porcelain. Until the end of theeleventh century the manufacture of glass inChina was repeatedly stimulated by the importof glass items from Central Asia and furtherwest. Chinese glass beads made during thefifth to the third centuries BC, whilst having ahigher lead and barium content, are similar tothose produced in Western Asia. During theHan Dynasty (206 BC to 220 AD), Roman glassentered China along the silk route. HanChinese glass was treated rather as a semi-precious stone, and was often carved usingjade-working techniques. The shapes of theobjects were typical of those found in jade andlacquer. Glassblowing was introduced fromWestern Asia around the fifth century AD. Theglass is found associated with Buddhism andburied in tombs of members of the Imperialfamily, such as that of Li Jingxub (AD 608), LiTai (AD 631) and Li Shuang (AD 668). In theFamesi at Fufeng (Shaanxi province), glass isincluded amongst other highly prized objectsin a ninth-century inventory. During the SongDynasty (960–1279), glass was used to makeegg-shaped objects of unknown use, gourd-shaped bottles and occasionally, foliate bowls(Brill and Martin, 1991).

From the Yuan period (1279–1368) onwards,the main area of production for Chinese glass


Figure 2.11 Mosque lamp holder of colourless glasswith a yellowish tinge and containing many bubblesand impurities. Six suspension rings trailed on body.Heavily decorated in naskh script in red, blue, blackwhite, green and yellow enamels (the last two badlyfired). H 350 mm. Middle fourteenth century AD. Syria.(© Copyright The British Museum).

was Boshan (Shadong province) in northeastChina. Although excavations at Boshan haverevealed furnace sites and glass rods in consid-erable numbers, existing glass vessels of thisperiod, which are turquoise in colour, are rare.In the Qing Dynasty, under the EmperorKangxi (1662–1722), a glasshouse was estab-lished in 1696 within the Imperial Workshopin the Forbidden City of Beijing. Theglasshouse was supervised by a German Jesuitcalled Stumpf and resulted in a flourishingproduction which greatly increased the statusof glass. However glass produced during thefirst half of the eighteenth century sufferedfrom the defect of crizzling, like Western glassof the same period. The vessels were oftypical Chinese shapes, often reflecting thoseof ceramic, jade and bronze objects. Thevessels were decorated by various techniquessuch as moulding, incising, carving, diamond-point engraving, overlay (cameo) andenamelling. Glass was made as a cheap alter-native to jade, and to imitate stones such asaventurine, jasper, lapis lazuli, pink quartz,realgar and turquoise. Enamelled decorationon glass was usually applied to opaque whiteglass which resembled porcelain or enamel.The art of enamelling glass reached its apexin the Qianlong period (1736–1797) when anumber of high quality pieces were produced,undoubtedly for imperial use. They aremarked with the characters Guyuexuan (OldMoon Pavilion). By the seventeenth centuryglass, in addition to its decorative use in theform of vessels, was also used for utilitarianpurposes in the form of lamps and lanterns,window-panes, blinds, scientific instruments,lenses and spectacles. Occasionally glass wasused for Buddhist figures and writing materi-als. It was used in imitation of precious stonesin jewellery, hairpins, toggles and plaques.Reverse paintings on glass, produced mainlyin Canton, were executed in large numbers forexport to the West. The technique was paral-leled by the highly skilled decorativetechnique of painting the interior of smallsnuff boxes.

Early Medieval Europe AD 400–1066

There were marked cultural changes after thefifth century, when barbarian incursions

replaced central Roman imperial power. Thechanges were reflected in glassmaking by ageneral technical decline. The glass wasinferior in quality and colour to Roman glass,and the vessel shapes were generally simpler.However, they were decorated with additionalglass applied with considerable manipulativeskill (Harden, 1956, 1969a, 1971; Henderson,1993b).

In the northern European countries glass-making tended to move away from the centresof population into the forests, which suppliedfuel for the furnaces. It is possible that natroncontinued to be transported (perhaps as‘chunk glass’, see Chapter 3, Part 2, and Seibel,2000) to these countries even after the collapseof the Roman Empire in the West, by overlandroutes through the Brenner Pass in Switzer-land, by sea around the Iberian Peninsula, orup the Rivers Vistula and Danube (Besborodovand Zadneprovsky, 1963; Besborodov andAbdurazakov, 1964). However, at some timebefore the tenth century, the ash produced inglass furnaces was substituted for the ashes ofmarine plants, which had been the almostuniversal fluxing agent used in Roman glass-making. This change to potash derived fromthe ashes of burnt trees, especially beech,resulted in a change of both alkali and limecontents of the glass, which is known as forestglass (Ger. Waldglas; Fr. verre de fougère –fern glass). The comparisons of the locationsof the northern glasshouses to the distributionof beech pollen in AD 1000 is discussed byNewton (1985b). The northern forest glass-makers, conditioned by their raw materials,produced mainly green and brown glass, anddecorated it with furnace-wrought embellish-ments of simple rib moulding, applied trailsand blobs, mostly in the same colour glass asthe body of the vessel itself. The vessels fallinto several categories: simple palm cupswithout handles, bag beakers, cone cups upto 265 mm in height and tapering to a pointedbase, a variety of squat pots and bottles, andclaw beakers (Ger. Rüsselbecher) (see Figure7.31).

In the course of the later Middle Ages glasswas improved to produce a substantial mater-ial of beautiful quality and a variety of greentones, used in a characteristic range of shapesof great originality and charm. Little is knownof glassmaking in the Rhineland from the


eighth to the fifteenth centuries. A few speci-mens have been found which provide enoughinformation to show that glass was madeduring this period, but it appears to have beenconfined to small, crudely made utilitarianvessels. Two important pieces are reliquariescontaining parchments dated 1282 and 1519.The former was discovered in a church atMichelfield near Hall, and is a small jardecorated with trailed threads reminiscent ofthe trailed-thread snake vases made inCologne in the third century. The latter is ashort parallel-sided beaker called an igel, withapplied decoration in the form of spikesresembling those of a hedgehog (Ger. Igel).

The general term for applied blobs of glassis prunts (Ger. Nuppen); and prunts are one ofthe most characteristic features of northernEuropean glass from the late fourteenth centuryonwards. It is possible to draw a parallelbetween these prunts and the projections onlate Roman glass vessels, and on Seine–Rhineclaw beakers (see Figure 7.31). Whereas in theearlier glasses the prunts were hollow blownand drawn out to form the distinctive longclaws, the later prunts were restricted to solidlumps of glass, which owed their appearanceand decorative effect to the manner in whichthe surface was finished. They were drawn outin several styles, for example to produce thinspikes resembling thorns, drawn out and foldedover to form loops from which small rings ofglass were suspended, and drawn into curlsand pressed back onto the surface of the vesselto resemble pigs’ tails. They were also flattenedand moulded to produce a beaded surface, ontypes commonly known as raspberry or straw-berry prunts.

Gradually the squat igel became taller whilstretaining its parallel sides until glasses weremade which were in excess of 300 mm high.This type of vessel became very popular andacquired different names according to theirintended use and style of decoration. Oneversion, decorated with a row of prunts, whichresembles broken-off leaf stalks, was termedkrautstrunk (Ger. cabbage stalk). Anotherversion was a plain glass divided into zonesby horizontal trailed rings, the passglas (Ger.),which each drinker in turn was expected todrain to the next division in one breath.

The known glassmaking centres at thistime were Cologne, Liège, Namur, Amiens

and Beauvais. It seems likely that glass wasexported to Britain from northern Gaul, andfrom the Rhineland, during the first sevencenturies AD, but there was certainly somelocal production, at least from the seventhcentury. Glassmakers seem to have beenworking in the Kentish kingdom in theseventh century because bag beakers andsquat jars are more prolific there than on thecontinent; moreover, a glass furnace wasfound in the cloisters of Glastonbury Abbey,near Bristol, beneath the medieval levels.With the spread of Christianity, the practiceof burying grave goods with the dead slowlydeclined in Britain, northern France and theRhineland. However, the custom continuedin Scandinavia until the beginning of theeighth century so that the major source ofglass from this period is from Scandinavianexcavations.

There may have been no real break in thesouthern glassmaking tradition between lateRoman times and the emergence of theVenetian glass industry in the thirteenthcentury, although there is as yet not muchevidence of glassmaking activity in south-eastEurope between the fourth and the eighthcenturies. In contrast with the northern forestglassmaking sites, the manufacture of glass insouthern Europe and the Mediterraneancountries largely remained at sites near thecoast such as Alexandria, Sidon, Damascus,Aleppo, Corinth, Aquilea, Murano, Florenceand Barcelona. The towns offered manyadvantages: there were customers immediatelyat hand, especially the wealthy ones; therewere churches and cathedrals with a demandfor window glass; the towns provided commu-nication and banking facilities; there was animpetus for innovation, such as the develop-ment of Venetian cristallo; and glassworkers’guilds could be formed to protect the interestsof the industry. As a result of increasingdemand, window glass was made morefrequently, and many examples are nowknown of the early type of small crown glasswindow panes. The crown glass windowpanes in the church of San Vitale in Ravennaalmost certainly came from the windows of theapse when the basilica was dedicated in AD

547. One bears an outline drawing of Christnimbed and enthroned. Glass tesserae wereused in wall mosaics; the mosaics at Ravenna


were no doubt made locally as were those atTorcello.

Further evidence of the produce of thesouthern glasshouses was found in the cargoof the Gnaliç wreck. This vessel, probably theGagiana, was on its way from Venice to theLevant, when it sank off the Yugoslavian coastin 1585. The cargo included 648 roundwindow panes, 170 mm, 185 mm and 205 mmin diameter, and 86 types of glass objects.These included goblets, flasks, vases, pitchers,large square plates and two types of mirror,squared and round, which were obviously ofVenetian origin. Some objects were of verythin transparent glass with greyish, greenish orpurplish tints. Many were decorated with verti-cal filaments or threads of white opaque glass.There were also several small bottles of cobalt-blue glass and a group of glasses with delicatediamond-point engraving (Brill, 1973;Petricioli, 1973).

The glassmaking of southeastern Europe inthe medieval period is as yet only imperfectlyunderstood, but it is evident that good qualitycrystal glass had evolved by the thirteenthcentury at the latest. Many glasses, made ofclear almost colourless glass, have now beenexcavated in Italy and other parts of Europe,including England, and can be dated to thetwelfth to fourteenth centuries. The formsmade were long-stemmed wine glasses andprunted beakers, often of considerable finesseand delicacy. Considerable use was made ofopaque red glass, not only for whole vessels,but also for decoration. However, none of theglass can be attributed to the Venetianglasshouses with any confidence. Some of theclear glasses with prunts and trailed bluethreads seem to have had their antecedents inthe Byzantine glasshouses situated in Greece,one of which has been excavated in Corinth.It is, however, possible that a number ofItalian glasshouses were involved in theirproduction, but by the fifteenth century Venicehad become the most important of these; andthe eastern predominance in glassmaking wasusurped by Venice. The technique of glassmanufacture and its decoration had by thistime largely developed, and therefore thehistorical development of glassmaking tends tobecome a list of changes in style or decora-tion which occurred at various times in differ-ent countries.

The pre-eminence of Venice

As previously mentioned, glassmaking inVenice may have had a continuous existencesince Roman times, unless there was a migra-tion of glassmakers from Aquilea, which isknown to have had an established glassmak-ing tradition. Certainly a glassmaking industrywas already established in Venice in the tenthcentury; and in records dating from 1090,mention is made of a phiolarius, which showsthat vessel glass was being made there(Charleston, 1958). However, before the mid-fifteenth century, Venetian glassware wasundistinguished and it is therefore difficult, ifnot impossible, to differentiate it from anyother glass of the period. Whether or not therewas some glassmaking on a small scale sinceRoman times, the pre-eminence of Venice inthe glassmaking industry came about througha chain of circumstances: the accidents ofgeography, time, political power and therebirth of the arts (the Renaissance) producedin Venice a standard of glassmaking both ofquality and imagination which had not beenachieved previously. A lagoon of low-lyingswampy islands was an unlikely place tocreate a city. The problems of construction,communication and health would make anysite on firm ground seem more favourable, butthe earliest settlers were probably refugeesfrom the effects of the barbarian invasionswhich swept the area after the fall of theRoman Empire in the West, for whom thelagoons may have seemed to offer security. Asat various times in other countries, turbulentand troubled times had an inhibiting effect onthe practice of the domestic arts and crafts. IfVenice was able to establish herself while thesurrounding region was in a state of turmoil,she would have offered a refuge whereartisans could practise their arts with theminimum of interference.

In 1204 Constantinople fell to the Crusadersand the immediate beneficiary was Venice,whose fleets had carried them there. Theglassmakers surely then obtained all the glass-making information and assistance theyrequired. At first, the Venetians may only havebeen glassmelters (i.e. not glassmakers), formasses of cullet are known to have beenimported in the form of ships’ ballast; andthere was a law of the Marine Code, as late


as 1255, which permitted vitrum in massa etrudus (crude lump glass) to be put on boardships as ballast.

That the glassmakers flourished, however, iscertain, for by 1271 the first records of a glass-makers’ guild appear; and shortly afterwardsin 1291 the glass furnaces were moved to theisland of Murano in the Venetian lagoon toavoid the risk of fire in a city comprised ofsmall and densely populated islands. It mayalso have been that having become importantand powerful enough to form a guild, it wasfelt that a tighter control could be maintainedover the industry if the glassworkers wereassembled in one locality. Glassmakers werehighly regarded in Venice and the city recordsshow that some became powerful and impor-tant men ranking with nobility in a city, whichhad a highly developed class system.

As communication and trade developedthroughout Europe, Venice and Genoabecame the natural focal points for the traderoutes of the world. There was intense rivalrybetween the two cities but Venice eventuallyemerged the victor, and became the cross-roads for land and sea traffic from east to westand from north to south. With this positionsecured, Venetian coffers swelled with thetaxes and duties she levied in commercialtraffic; Venice became the most powerful cityin the Mediterranean. Venetian merchantstravelled the trade routes of the world, and inthe course of their dealings with the MiddleEast they would have encountered Syrian andMesopotamian glassmakers so that the best ofMiddle Eastern glass would have found its wayback to Venice.

During this period the trade in glassware,which of course had been taking place forsome three millennia, started to bedocumented. Trade routes were established,for example, between Genoa and Syria, theMediterranean countries and Asia as far east asChina, and between Murano and Dubrovnik.

With the overthrow of the Middle East byTamerlaine in 1402, the decline of the glasstrade in the Middle East left a vacuum in thesupply of high-quality glass. Since Venice wasin control of all the main commercial routesand had established a considerable glass tradeof her own, it was natural that she should stepin to fill the breach. In the face of theonslaught of the barbarian hordes the local

inhabitants would flee wherever they could tofind refuge, and what more natural than thatthe glassmakers would make for a placewhere, from contact with visiting merchants,they knew that there would be an opportunityto continue the practice of their skills? Thus,in addition to having new markets opened toit, Venice also had a second infusion of skillsfrom the Middle East.

As the world once again became a moresettled place, prosperity increased and therecame about a rebirth of artistic endeavour andachievement – the Italian Renaissance, duringwhich talented artists, able to satisfy theaesthetic demands of a new society, weresponsored by wealthy patrons of the arts.Glass painters are recorded in Murano as earlyas 1280; and colourless glass was almostcertainly made there in the thirteenth century.From the middle of the fifteenth century thesetwo branches of the arts experienced anunparalleled expansion in what was by thena highly specialized industry. Trade betweenVenice and the East declined afterConstantinople was captured by the Turks in1453; and the Venetians then built up anactive trade with the West based on the façonde Venise style which enabled them tocontinue to dominate the glass industry.

The earliest Venetian glass vessels of anyquality were goblets, usually flat-bottomed andwith straight tapering sides mounted on apedestal foot. They were made of colouredglass, usually blue or green, with enamelleddecoration around the bowl (Gasparetto,1973). Previously, stemmed glasses had beenrare. At first the glass bowls were set on aplain ribbed pedestal in which stem and footwere one. Gradually the stems were madetaller and were decorated with hollow blownbulbs (knops), until for the first time glasseswith separate stems and feet were produced.Once the stem and foot were able to be madeseparately there was no limit to the ingenuitywhich could be used in fashioning the stem:hollow knops with moulded lion masks, stemswith a central feature of glass threads drawnout in the shape of serpents or figures-of-eight,winged stems with pincered fringes, and manyothers (Tait, 1968, 1979). In a further devel-opment the Venetians improved their clearglass by the addition of lime to the soda–silicamixture, and purified the soda to produce a


fine clear glass (cristallo), which captured thepopular taste of the fifteenth century.

One of the great abilities of the Venetianglassmakers was their skill in the manipulationof the material. They acquired dexterity incontrolling the molten material, which enabledglasses to be produced of a delicacy, and witha degree of elaborate decoration, which noone else could equal (Figure 2.12). Such glass-ware was made for a wealthy and sophisti-cated market, and outside Venice that wouldmean principally the European nobility. ThusVenetian glass was for a long time the prerog-ative of the rich and educated. Later cameplates, tazzas, flasks, chalices and a widevariety of other domestic vessels in blue, greenor purple glass and later in cristallo. The glass-ware was enamelled and gilded and was the

product with which Venice first entered theworld market. The decorative themes were oftypical Renaissance inspiration: triumphs,allegories of love etc.

Inspired by classical Rome, the Venetianglassmakers also produced mosaic, millefiori(Figure 2.13), aventurine (with copper parti-cles) and calcedonio glasses, the latter in imita-tion of Roman agate glass produced 1500 yearsearlier, and which was itself an imitation of thenatural stone. Enamelled glass gradually wentout of favour, except for customers in north-ern Europe, and mould-blown or exquisiteplain forms such as aquamaniles or nefs(decanters, see Figure 2.12) succeeded. Thesewere sometimes decorated with bands orcables of opaque white (lattimo) threads, thatis, filigrana, and occasionally by gilding,diamond-point engraving (although the thinsoda glass was generally unsuited to engrav-ing), cold painting behind glass (Ger.Hinterglasmalerei) (see Figure 7.71), or asurface craquelure (ice-glass, Figure 2.14). Asthe fifteenth century progressed, furnace-wrought decorations of applied threads orbosses, often of fantastic forms, became morepopular, and in the seventeenth century thistendency was sometimes carried to extravagantand not always attractive lengths.


Figure 2.12 Nef (or ship) ewer of colourless glass, thebowl in the form of a boat with a spout forming aprow. H 343 mm. Sixteenth century. Venice.

Figure 2.13 Millefiori miniature ewer of opaque blueglass, with canes of purple, green, brown and blue. H127 mm. Early sixteenth century. Venice.

Venice was jealous of her position in theworld’s markets, and to ensure that she keptit the Venetian Senate placed highly restrictiveconditions on its craftsmen, especially theglassmakers. They were forbidden on pain ofdeath to practise their skills anywhere outsideVenice, or to impart their knowledge toanyone other than Venetians. However, thedemand for Venetian glassware from theCourts of Europe was such that considerableinducements were offered to persuadeVenetian glassmakers to risk the consequencesand to set up their glass pots outside Venice.There are records of Venetian ambassadorsusing bribery and blackmail to persuade theunfortunate truants to return home. The tidecould not be entirely stemmed, and slowly but

surely at other glassmaking centres throughoutEurope the stylistic Venetian elements – thefaçon de Venise – were introduced by migrantworkers from Venice itself, and from Altarenear Genoa which was the traditional rival ofVenice and where the glassmakers’ corpora-tion had a deliberate policy of disseminatingworkers and techniques. The Italian knowl-edge thus spread throughout Europe, reachingVienna in 1428, Sweden in the 1550s andEngland by 1570 at the latest, so that Venicelost its supremacy in the glassmaking industry.Each country showed just a little individualityof style, which marked its product from thetrue Venetian glass.

Inevitably Venice made enemies whoconsidered her power too great, or whoseinterests conflicted with her own. As a resultthere were many attempts to break her influ-ence. When the Portuguese found a new routeto the Far East via the Cape of Good Hope atthe end of the sixteenth century, goods couldbe shipped to and from the Far East withoutthe necessity of paying high duties to shipthem through Venice. Her importance on thetrade routes thus undermined, Venice startedto decline, and the Republic finally collapsedin 1797. Long before this the glass trade innorthern Europe had overtaken the Venetianmanufacture, and from the end of the seven-teenth century she had lost her importantposition to numerous rivals. Taste was begin-ning to change, looking more towards moresolid, colourless crystal glass truly resemblingnatural rock crystal. Glass of this nature wassuccessfully produced in Britain and Bohemia.

Europe from the Middle Ages to theIndustrial Revolution

Northern European glasshouses

BohemiaRecords of glassmaking in Bohemia date fromthe fourteenth century, and because of theabundant supplies of wood for fuel, and ofraw materials for the glass, the craft soonestablished itself there. Glassmaking becamesuch an important part of the industrial life ofBohemia that, by the end of the nineteenthcentury, there were 56 glasshouses in opera-


Figure 2.14 Large beaker of ‘ice-glass’, created byplunging the hot glass into water for a moment andimmediately re-heating. The roughened surfaceresembling ice became popular in Venice in thesixteenth century and spread to Northern Europe, whereit remained in vogue into the seventeenth century. H209 mm. Late sixteenth century. Southern Netherlands.

tion. Several of the glass styles, the roemer, forexample, were common to the whole of north-ern Europe, regional characteristics not devel-oping until the eighteenth century. The roemer(Figure 2.15) originated as a large, cup-shapedbowl with a hollow stem, which was blownin one piece, and mounted on a long narrowpedestal foot made from a thread of glasswound round a conical pattern. The stemconnection was ornamented with prunts andmilled collars at the top or bottom. As timepassed, the foot assumed more importanceand became taller, while the bowl tended todecrease in size. In later years, the pedestalwas formed in one piece with a corrugatedouter surface to represent the original threadof glass. (The name rummer, applied to thelarge beer glasses made in Britain in the 1780s,may be a corruption of the word roemer).Contemporary with the development of theroemer, a number of other forms appeared:kuttrolf, stanenglas, igel (hedgehog) glass,passglas, daumenglas (or daumenhumpen).The daumenglas was a more robust, barrel-shaped vessel with a series of indentations inthe sides to provide finger-grips.

During the sixteenth and seventeenthcenturies, German glasses (humpen) becamegradually taller and wider. They were usually

decorated with coloured enamels, and arenow classified according to the method ofdecoration. Reichsadlerhumpen bore thedouble-headed Imperial eagle, and on theoutspread wings were painted the coats-of-arms of the 56 members of the GermanicConfederation. The thinly blown kurfursten-humpen depicted the emperor and the sevenelectors.

Although the Venetians had suppliedenamelled glass for the German market, this isusually easy to distinguish from the Germandomestic products since these were morerobust and the brighter colours more heavilylaid on. Whatever it lacked in finesse, Germanenamelling succeeded by its sheer exuberance.Schwarzlot was another style of painting onglass, which gained some popularity duringthe seventeenth century. This consisted ofoutlined pictures in black with the clear areasoccasionally infilled with a brown wash. Fineexamples of schwartzlot technique wereproduced by Johan Schaper (1621–1670).Bohemian glass had begun to acquire itsindividual character in the late sixteenthcentury with the introduction of cutting andengraving. Glass made in the façon de Venisehad spread northwards through Europe, andwith it such techniques as the art of cutting


Figure 2.15 (a) Green glass Roemer with a globular bowl and a stem decorated with raspberry prunts. H 156 mm.Mid seventeenth century. (b) Green glass Daumenglas with six fingergrips (fingernapfen). H 200 mm. Seventeenthcentury. Germany or Netherlands. (c) Greenish glass Stanenglas with prunted and trailed decoration. H 426 mm.First half of the fifteenth century. Germany. (d) Reichsadlerhumpen, pale green glass with enamelled decoration. H290 mm. 1654. Bohemia.

(a) (b) (c) (d)

glass as if it were rock crystal. This caught theimagination of the Bohemian glassmakers, andit was from this beginning that the greatBohemian tradition of cutting and engravingglass arose.

In the early seventeenth century anotherinnovation by the Bohemians was combinedwith their skill in engraving to produce someof the most remarkable examples of thistechnique ever seen. This was the develop-ment of a new glass, which required theaddition of lime to the potash–silica batch thenin use, to produce a perfectly clear, solidcrystal glass. The glass proved to be ideal fordecorating by wheel-cutting and engraving;techniques which had been developed inPrague before 1600, and then transplanted toNuremberg, where a school of engravingflourished throughout the second half of theseventeenth century. The elaborate gobletswith multi-knopped stems which were madein Nuremberg copied contemporarygoldsmiths’ work.

A hardstone engraver, Caspar Lehman ofPrague, was the greatest exponent of therevival of glass engraving in Europe. Lehmanpresumably used the same skills and equip-ment for both crafts, a tradition that was tocontinue until relatively modern times. WhenLehman died in 1622, the patent to engraveglass was taken over by George Schwanhardtthe Elder, who founded a school of brilliantglass engraving in Nuremburg, which flour-ished until the eighteenth century. Wheel-engraved glass was produced in Holland,shortly after the middle of the seventeenthcentury. The most prominent Dutch engraverwas Jacob Sang of Amsterdam. English leadglass was also suitable for engraving, andblank vessels were exported in large quanti-ties to be engraved in Holland.

Political troubles eventually led to thedispersal of the engravers throughout Ger-many, where engraving became a less popularmethod of decorating glass. Towards the endof the seventeenth century the taste for façonde Venise had declined and been replaced bythe new style of wheel-engraving which madeuse of the qualities of the new glass, andwhich became highly developed in Bohemiaand Silesia, a speciality being imposing gobletsengraved in high relief (Ger. hochschnitt), theground of the design being cut back by means

of a water-driven wheel (see Figure 3.18).From Silesia the art of engraving was trans-planted to Potsdam, one of the most accom-plished workers being Gottfried Spiller.

Goblets with round funnel bowls hadknopped stems cut with facets, well propor-tioned but less ornate than the Venetian equiv-alent. The glass was thicker and the bowlswere often cut with vertical panels and haddomed covers to match. The bowl, the coverand sometimes even the foot would beprofusely covered with highly ornate baroquedecoration interrupted by coats of arms orformal scenes in enamel. For special courtcommissions gold was often applied to thesurface of the glass in relief. Constant handlingof such glass would quickly spoil its glory andtherefore the highly burnished fired gilding ofthe Potsdam glasshouses, patronized by theElectors of Brandenburg, was far more practi-cal, and rather more spectacular. Frequentlythe wheel-engraved decoration on Potsdamglasses is heavily gilded and the overall effectis little short of ostentatious. In the course ofthe eighteenth century glass engraving spreadthroughout Germany and Central Europe,centres of special importance being Nurem-berg, Kassel, Gotha, Weimar, Dresden andBrunswick, with Warmbrunn and others inSilesia.

Later in the eighteenth and nineteenthcenturies more naturalistic themes were intro-duced, which included woodland scenes withdeer. So highly was this combination of glassand engraving regarded that Bohemian artistswere persuaded to go to Venice and Spain tointroduce the technique there. Some colouredglass was made in Bohemia, and the mostimportant researcher connected with its devel-opment was Johann Kunckel, a chemistworking in Potsdam. Kunckel (1679) produceda variety of colours varying from rose pink topurple using Purple of Cassius.

Although cutting and engraving were by farthe most important of the decorative processesto be developed in Bohemia, several othertechniques, well-known in antiquity, wererevived and given a distinctive Bohemianquality: for example, about 1725, the art ofenclosing a pattern of gold leaf and transpar-ent enamel between two layers of glass (Ger.Zwischengoldglas). The inner layer of glasshad a lip on the outer surface onto which the


outer layer joined, thus making a close-fittingjoint which completely protected the delicatefilm of gold leaf inside. At about the same timeIgnatius Preissler was establishing a reputationfor decorating glass objects with painted glasssubjects of chinoiserie scenes enclosed inbaroque scrolls painted in black enamel,frequently highlighted with gold.

In the early nineteenth century the blackbasalt wares of the Wedgwood factory inBritain were very popular, and this taste wasreflected in black glass vessels, the Hyalithglass made in Bohemia from 1822 onwards.One of the manufacturers of Hyalith wasFrederick Egermann, who was responsible fora number of new varieties of glassware, whichimitated natural stones such as agate, jasperand marble, as well as transparent glasscoloured with chemicals. In the earlynineteenth century, transparent and opaquewhite glass, were combined to produce theoverlay glass whose popularity lasted through-out the century. It usually consisted of a layerof white over a base of transparent green, redor blue glass. The white glass was then cutthrough to form windows through which thecoloured glass underneath could be seen. Thewhole was then decorated with gilding orcoloured enamel. The idea was taken one stepfurther later in the century when the overlaywas reduced to one or two medallions. Thesewere decorated in coloured enamel, withportraits or with bunches of flowers, while thebody of the vessel was covered with a finemeandering pattern in gilt. This style of warewas usually made as ornamental ewers orvases.

A cheaper but effective method of produc-ing coloured glass was to cover the outside ofa clear vessel with a film of coloured glass.This was applied either as a stain or bydipping the gather in a pot of coloured glass.The commonest colours were fluorescentyellow produced with uranium and red, butgreen and amethyst examples are known.Some glasses cut with vertical panels had eachpanel coated with a different colour. If thevessel was embellished with engraved decora-tion the thin layer of colour was easilyremoved to show the clear glass underneath.Popular subjects were views of spa towns onbeakers and tumblers, which were sold assouvenirs. Although they were probably not

very expensive when they were made, theengraving was always of a high standard.

In the late nineteenth century when the ArtNouveau movement began, a Viennese, LouisLobmeyr, established a factory to produceglass of a high artistic and technical qualityemploying Bohemian craftsmen. This was areaction against the vast quantities of cheaplymade mass-produced glassware which wasmade everywhere during the second half ofthe nineteenth century. This effort to reviveglassmaking as an art was continued by theLoetz glassworks, where beautifully colourediridescent glass was made.

The Low CountriesThe early history of glassmaking in the LowCountries is as vague as that of the rest ofnorthern Europe. Geographically the areacomprises the country between the Rhine andthe Mosel rivers (now forming part ofGermany), the seven provinces of theNetherlands, and the general area of Belgium.It was rarely all under one rule at any onetime, and was variously controlled by theSpanish and Austrians. Glassmaking in theLow Countries first came to prominence withthe introduction of Venetian glass in thefifteenth century. The Venetian cristallo whichwas such an improvement over its predeces-sors was displayed and advertised in all thecourts of Europe, and seems to have attractedparticular attention in the Low Countries.There are records of glasshouses being estab-lished to make the façon de Venise in Antwerp(1537), Liège (1569) and Amsterdam (1597).The trade flourished, and the area became themost important in Europe outside Venice toproduce the Venetian style of glassware. Itsdistance from Venice, however, led to slightdifferences in the character of the glass but itis still difficult to distinguish the glass made inlocal glassworks from that imported fromVenice.

While the façon de Venise flourished, otherglassworks in the principal towns as well asmany others in minor centres continued toturn out masses of utilitarian wares in thetraditional styles of the region. The local tradi-tion still made use of waldglas, but by thesixteenth and seventeenth centuries the green-ish glass had been developed further, and awhole range of blue-greens produced. The


roemer was one of the native styles whichacquired great popularity, and several types,identified by the prunts applied to the stems,have been assigned a provenance in the LowCountries. One of these employed prunts thatwere flattened and smoothed to the pointwhere they blended into the wall of the vessel.This produced different depths of colouraccording to the varying thickness of the glass.Two other prunt variations were thosemoulded on the surface with a face ofNeptune, and those with beads of blue glassapplied to the centres. During the seventeenthcentury roemer stands were made. They weretall pedestals in precious metals designed tohave a roemer clamped into a mounting onthe top. Due to the proportions of the roemersthen in use this had the effect of producing atall metal goblet with a glass bowl. The totaleffect was reminiscent of the elaborate tallgoblets made by the Nuremberg goldsmiths.Another style peculiar to the Low Countrieswas the tall, narrow flute. This was a verythinly blown glass up to 450 mm high and50 mm wide. The funnel bowl was usuallymounted on a single hollow knop and aspreading foot. Very few of these fragileglasses have survived.

From the end of the seventeenth century thelead glass being produced in Bohemia andBritain became increasingly popular withDutch glass engravers, by virtue of its clarityand density. The Bohemian glassmakers hadan efficient marketing organization, and set upwarehouses in the Netherlands to stock theirproducts. The Newcastle light baluster stylewas particularly favoured and, during theeighteenth century, British lead glass was muchimitated by local glasshouses. The traffic wasnot all one way however; elaborate basketsand dessert services made in Liège during theseventeenth and early eighteenth centurieswere popular in Britain. These vessels weremade of threads of glass built up in layers toproduce an open-mesh design. They stood onplain glass plates decorated with a borderexecuted in the same mesh style.

Engraving was the only applied method ofdecorating glass which achieved any degree ofimportance in the Low Countries, and through-out the seventeenth and eighteenth centuriesthere were a number of artists whose workwas of the highest order. The technique as

developed in the Netherlands was quite differ-ent from that practised in Bohemia. Dutchengraving was carried out on the surface ofthe vessel with a delicacy equal to that ofpainting, while Bohemian engraving dependedon a good thickness of glass and was moreclosely allied to sculpture. One of the earliestrecorded Dutch engravers, Anna RoemersVisscher (1583–1651) was one of a very fewwomen known to have been engaged in thiswork. She specialized in a free-flowing styleof calligraphic engraving, a method alsofavoured by Willem von Heemskerk(1615–1692) in the seventeenth century, andby Hendrik Scolting in the eighteenth. Thefinest seventeenth-century work, was pro-duced by artists such as Frans Greenwood,David Wolff and Jacob Sang. The first twospecialized in stipple engraving, a methodwhereby the surface of the glass was markedby repeated light blows with a diamond-pointed tool. The density of the marks thusproduced determined the variations in toneand shading of the completed picture. Theresulting effect was rather similar to a thinphotographic negative. It required a completesureness of touch and considerable artisticability for its success (see Figure 7.27e).Because of these demands on the artist, stippleengraving was never widely practised, butthere are a few engravers working in contem-porary Britain who can produce stippleddecoration comparable in quality to that of theeighteenth-century Dutch masters. Jacob Sangspecialized in wheel-engraving, and his workhas a delicacy, precision and sureness of touchwhich few engravers have equalled. Foremostamong his work are the glasses which heengraved with coats of arms.

Britain AD 1500–1850During the first half of the sixteenth centurythere was an influx of French glassworkers toBritain from Lorraine. It appears that they werenot popular locally and eventually they left theWeald and settled in other parts of thecountry. The principal cause of their unpopu-larity was the rate at which the forests weredestroyed by use of wood fuel in their glassfurnaces. Glassworkers were in competitionwith ironworkers for wood fuel; the latterhowever were static, while the glassworkerswith their much lighter glass pots could move


on whenever the local supplies of fuel wereexhausted. Crossley (1967, 1972) calculatedthat the furnace at Bagot’s Park in Staffordshire(AD 1535) would use circa 130 tonnes ofwood per month (i.e. 1.6 hectares. of 15-year-old coppice) and that as a consequence themodern area of Bagot’s Park would bedenuded of trees in 15–20 years. This fuelconsumption would correspond to about60 Whg-1 of energy, compared with only4 Whg-1 from coal at the beginning of thetwentieth century, and less than 1 Whg-1 inthe best oil-fired present-day practices. Withthe consumption of forests at such a rate theBritish government became alarmed about theloss of trees for building ships for the Navy,and in 1615 James I issued a ProclamationTouching Glass, banning the use of wood formaking glass. Nevertheless the landownerswelcomed the use for glassmaking of other-wise unsaleable timber.

By the early seventeenth century coal wasbeginning to be used as an alternative sourceof fuel. This encouraged the dispersal of theFrench glassmakers from the south of Britainand led them to settle in districts where coalwas readily available, such as Stourbridge andNewcastle, where their influence was to be feltfor two hundred years or more. Scottish coalhad been used at first but Lady Mansell sentWilliam (Roaring) Robson to Newcastle and hethen found that the Newcastle coal wascheaper than the Scottish coal (Watts, 1999).

One of the Lorraine glassmakers was JeanCarré who, in 1567, obtained a licence for 21years to set up a glasshouse in London, toproduce glass in the façon de Venise since theVenetian product was still highly regarded inBritain. To this end Carré imported severalVenetian glassmakers and was the firstmanufacturer in Britain to use soda instead ofpotash as a source of alkali, possibly as aresult of his Venetian contacts.

When Carré died in 1572, one of hisVenetian craftsmen, Jacob Verzelini, took overthe licence and in 1574 was himself granted a21 year licence for the making of drinkingglasses, ‘... suche as be accustomablie made inthe towne of Morano’ (Douglas and Frank,1972), but in fact the Verzelini glass was muchplainer and more functional than the conti-nental examples of façon de Venise. Only ahandful of glasses exist which can be attrib-

uted to the Verzelini glasshouse. They are allgoblets in façon de Venise, engraved bydiamond-point and bearing dates between1577 and 1586. Verzelini ran the glasshousefor 17 years before retiring, and he finally dieda rich and respected citizen of London in 1606.With Verzelini’s retirement the monopoly tomake glass in Britain passed through severalhands until Sir Robert Mansell gained controlin 1618, having bought up existing monopo-lies (a history of the London glasshouses isgiven by Watts, 1999).

Mansell was successful as a result of the lawof 1615 prohibiting the use of wood fuel forfiring glass pots, since he had acquired thepatents covering the use of coal as a fuel forthat purpose. He established a number ofglasshouses in England, which made glassusing Spanish barilla (containing soda andlime) as a source of alkali, coal fuel andemploying workmen from Altare in Italy.Mansell maintained control of the glass indus-try for about thirty years, but does not seemto have survived the Civil War or theCommonwealth (1640–1660). It was only onthe restoration of King Charles II to the thronethat glassmaking began to flourish. In 1663George Villiers, Duke of Buckingham,petitioned Charles II for what was probably arenewal of the Mansell monopoly. With theaid of a Frenchman, Jean le Cam, Villiersopened a glasshouse at Vauxhall in London tomake looking glasses and imitations of rockcrystal. However, Villiers never secured thetotally monopolistic power of his predeces-sors.

In 1664 the Glass Sellers’ Company wasformed, and from that date onwards the glasstrade was largely dictated by that company.Two factors illustrate this clearly. First, insurviving correspondence between JohnGreene, a glass seller in London and AllesioMorelli in Venice, from whom he bought glass,Greene states exactly how the orders are tobe executed, often complaining of the qualityand enclosing sketches of the shapes andstyles to be supplied. The second factorconcerns the backing of George Ravenscroftby the Glass Sellers’ Company to pursueresearch into the development of a new typeof glass.

Thus the trade had changed in 100 years from an industry based on individual


semi-itinerant glassmakers to a properly organ-ized and commercial enterprise controlled bya regular trade association. This was to provethe springboard for the ascendancy of Britishglass over the next 150 years.

George Ravenscroft and lead glassVenetian glass was thinly blown and delicate,and fine pieces executed for wealthy familieswere undoubtedly treasured and carefullypreserved, so that many still survive. However,little glass made for everyday use remains.Indirectly, the glassmaker George Ravenscroftwas to change this situation by producing amore robust type of glass; and thus relativelylarge quantities of eighteenth-century domes-tic glass still exist. Ravenscroft was not a glass-maker by trade, and had reached the age of55 years before setting up a glasshouse at theSavoy in London in 1673, in order to carry outexperiments to create a new type of glass.

The late seventeenth century was a periodof intense research and experiment by men ofculture and education, to advance scientificknowledge for the benefit of their fellow men,the Royal Society being a product of this era.Ravenscroft’s early experiments were unsuc-cessful, the glass soon crizzling as a result ofthe lack of balance of constituents, but iteventually showed such promise that in 1674the Glass Sellers’ Company established him ina glasshouse in Henley-on-Thames, Berkshire,to pursue his researches in seclusion. This wason the condition that when a successfulformula was achieved, the vessels would bemanufactured to the Company’s own specifi-cations; however, Ravenscroft took the precau-tion of obtaining a seven year patent on hisnew ideas (Macleod, 1987; Moody, 1988).

Most early lead glass showed distinct tingesof colour, usually grey, green or yellow, whichwere the result of impurities in the raw materi-als. Undoubtedly the glassmakers endeavouredto produce a glass that was colourless, andadded various chemicals to counteract theimpurities. However this was very much a caseof trial and error, and success was achievedmore by chance than by chemical control. Amore predictable result could be achieved byusing purer materials, for example, sand freefrom contaminants could be obtained fromareas near King’s Lynn in Norfolk, and nearNewcastle-upon-Tyne, sand from these areas

being shipped to all the British glassmakingcentres.

By 1676 such progress had been made thatthe Glass Sellers’ Company issued a certificateexpressing its satisfaction with the new glassof flint, as it was then called, and givingpermission for the glassware to be identifiedby means of a seal bearing a raven’s head.From this badge a small number of glassvessels can be attributed to the Ravenscroftglasshouses. The secret of the glass produc-tion was the addition of lead oxide to the rawmaterials, which produced a soft and brilliantglass of great refractive quality. The Venetiancristallo had looked clear and transparentpartly on account of its thinness and also onthe purification of the soda which had beenused, but the new lead glass remained clearand transparent when much thicker glass wasblown. Lead glass did not lend itself to theextravagances of the Venetian style, butbecause of its lustre, plainer shapes made adisplay which was just as effective (Figure2.16) (Charleston, 1960).

Ravenscroft terminated his agreement withthe Glass Sellers’ Company in 1679; and diedin 1681, about the same time as his patentexpired. The glasshouse was continued byHawley Bishopp, who had worked withRavenscroft at Henley on behalf of the GlassSellers’ Company, and who presumably knewthe lead glass formula. However, before manyyears had passed lead glass was in commonuse in all the glassmaking centres of Britain.

For the first few years after the introductionof lead glass, vessels continued to be made inthe Venetian tradition. Indeed, Ravenscroft hadimported two Venetian glassmakers to work inthe Savoy glasshouse. As early as 1670, thedesigns which John Greene had sent to Venicehad shown a tendency to simplify theVenetian styles, and as time went by severalfactors combined to create a particularlyBritish style. As the demand for glass grew,and new glasshouses opened, there couldnever have been enough skilled Venetianworkers to staff them all. Therefore increasingnumbers of British glassworkers would havehad to be trained and it is unlikely that theywould have acquired, and exactly imitated, theabilities of their Venetian instructors.

Lead glass became so popular that, as previ-ously mentioned, various attempts were made


on the Continent to produce lead glass in thefaçon de l’Angleterre. This was not achieveduntil the late eighteenth century, largely dueto the fact that in order to protect lead glassfrom the effects of smoke it was necessary touse covered pots, and this was not the practicein continental glasshouses. The British glass-making industry, however, was already usingcovered pots before the introduction of leadglass, as a direct result of the need to use coalas an alternative fuel to wood (Newton, 1988).

During the period under review the variouschanges of style and decoration followed eachother in a regular sequence, which has beencatalogued by Barrington Haynes (1959). Thetransitional period from about 1680 to 1690 isusually referred to as the Anglo-Venetian,when such Venetian characteristics as spikedgadroons, trailed decoration and pinceredstems still appeared on glassware. Thesegradually disappeared to give way to the firstperiod which can be described as beingpeculiarly British: the period of heavy balusterstems, 1690–1725. The knop formations on thestems of the glasses derive from the hollow

blown knopped stems of Venetian glasses, andwere the last signs of Venetian influence onBritish glass. The stems of British glasses of theperiod were either solid or contained airbubbles (tears). The early baluster stems werenotable for their size and weight. The bowlsare usually solid at the base, and the knopsappear in a variety of forms with such descrip-tive names as cylinder, egg, acorn andmushroom knop. As time passed the propor-tions of the glasses became more refined; theheight increased, the weight decreased, andwhere the stem had previously been madewith one knop, combinations of knops becamefashionable. This style reached its peak withwine glasses made in the Newcastle-upon-Tyneglasshouses in the north of England. TheNewcastle light baluster was tall with a multi-knopped stem often containing rows of tiny airbeads, and of clear colourless glass.

One of the most famous glassmakingfamilies in Newcastle at this period was thatof Dagnia. The family was of Italian origin,and seems to have arrived in Newcastle fromBristol in about 1684. Until that time only


Figure 2.16 A basin and ewer made by George Ravenscroft. The basin is cloudy, having become crizzled. (©V&A Picture Library).

sheet glass and bottles had been made inNewcastle, and it is quite possible that theDagnias brought with them the recipe for thenew lead glass and began the quality glasstrade which culminated in the fine wineglasses referred to above. Contemporary withthe baluster period was another group ofglasses, which owed their introduction to theaccession of George I to the throne of Englandin 1714. To mark this event glasses with theso-called Silesian stems were produced. Theseglasses had tapering, four-sided, mouldedstems, which the English glassmakers soonmodified to six or eight sides.

The glasses retained their popularity alongwith the baluster glasses until the introductionof the Glass Excise Act of 1745. From theseventeenth century onwards the BritishGovernment had considered glassmakers to bea fruitful source of revenue. In 1695 WilliamIII had introduced a window tax, which greatlyreduced the output of glass and caused muchunemployment in the trade. The duty wassubsequently repealed, but was levied again in1745 as a duty on raw materials. As a resultglassware became plainer in shape and lighterin weight to minimize the effect of the tax; andplain-stemmed glasses containing less leadwere introduced (Charleston, 1959). From themid-eighteenth century public taste changed,the more ornate rococo and chinoiserie stylesbecoming popular, and the glassmakerschanged their styles to meet the demand. From1750 onwards air-twist and opaque-twist stemsmade their appearance. Both these styles weresuited to an art which put a premium on theamount of raw material used since they permit-ted decorative glass to be made without thelarge quantities of glass necessary for produc-ing baluster stems. Air-twist and opaque-twiststems remained popular until 1777 when afurther duty was imposed on the enamel rodsused in the production of the opaque twiststems; these were then made in less quantity,and cutting began to re-emerge as the mainform of decoration.

In the 1770s cutting again became popular(Bickerton, 1971), but this time as a methodof decorating the glass stems with facets. Theplain stems were cut with a series of scallopsso that the intersecting edges formeddiamonds or hexagons. The art of glass-cuttingspread rapidly and from 1780 onwards a wide

range of glassware was being made with moreand more elaborately cut designs. Cutting wasable to exploit the lustrous and refractivequalities of lead glass (see Chapter 1 for defin-itions of refractive index and dispersion), andfor the next 50 years British glassmakingreached a peak of quality in both glass anddecoration. After 1830 styles tended to becometoo intricate and the art of cutting generallywent into decline. Throughout the periodmany other domestic articles, besides vessels,were made in glass, for example chandeliers,candlesticks, jugs, bowls, decanters, bottlesand sweetmeat dishes. Besides cutting, otherforms of decoration were employed such asengraving, gilding and enamelling, but theyrarely achieved either the degree of quality orpopularity which they had attained on theContinent. Engraving on glass is rare in theearly eighteenth century, and when it beganto make an appearance after 1730, wheel-engraving was the method of productiongenerally preferred over hand-engraving.

In 1820 an annual licence was introducedwhich greatly disrupted glass manufacture.The operation of the new law was verycomplicated; furnaces were locked by inspec-tors, and 12 hours’ notice in writing had to begiven before a glass pot was filled. Moreover,the regulations made it virtually impossible tointroduce new types of glass, such as the typerequired for making lenses or for bottles thatwould be acid-resistant (Douglas and Frank,1972). These restrictions were lifted in 1845,and the glass industry immediately entered aperiod of rapid growth.

Diamond-point engraving by British artistsin the eighteenth century is virtually unknown.On the other hand, large quantities of the fineNewcastle light balusters were exported andmost of those which survive bear Dutchengraving, both by wheel-engraving and bydiamond-point work, which is of an exceed-ingly high standard (see Figure 3.16). Gildingoccurs either as oil gilding which was notpermanent and became barely detectable as aresult of constant handling, or as gilding whichwas fired onto the glass to produce a perma-nent decoration. James Giles, a Londonengraver and decorator, executed a number ofglasses decorated with fired gilding, and hisstyle consisting of sprays of flowers andinsects, is easily recognizable. In the last


quarter of the eighteenth century the Jacobsfamily in Bristol also produced a great deal ofgilded glassware; unusually, many of thesepieces were signed.

Enamelling was only carried out by a fewartists. Most examples are attributed either toMichael Edkins or to William and Mary Beilby,although there are a few enamelled glasses,which do not conform to the style of either ofthese. The enamelled designs attributed toEdkins are nearly all executed on dense,opaque, white glass, which was made inStourbridge or Bristol to resemble porcelain.Edkins is recorded as having worked in bothtowns, and the decorations consist of birds,flowers and chinoiserie designs. The Beilbyfamily, brother and sister, worked in Newcastlefrom 1762 to 1778. They painted small scenesof rural pursuits or classical ruins on sets ofwine glasses, usually in white enamel. Thefamily must have been widely known andhighly regarded since a small number ofsurviving goblets carry coats of arms inpolychrome decoration, which had beenexecuted to special orders for titled families(see Figure 3.29). There was also a demand forcoloured glass. While early eighteenth-centurypieces do exist, they are rare, and the major-ity of surviving coloured glass dates from thelate eighteenth and early nineteenth centuries.The principal colours used were green, blueand amethyst, while amber and red are rare.

A method of dating glass of this period isby reference to the many pieces which wereengraved to commemorate a person or eventwith a known date. Probably the best knownBritish commemorative glasses are those relat-ing to the Jacobite movement, from the failureof the Rebellion of 1745 to the death of PrinceCharles Edward in 1788. During this periodseveral societies sympathetic to the Jacobitecause flourished, and it was fashionable,though treasonable, to drink a toast to BonniePrince Charlie from glasses bearing his portraitor emblems representative of the movement.Occasionally Latin mottoes expressing hopefor his return from exile were added, such asFiat, Audentior Ibo and Revirescat.

Britain after 1850 (industrialglassmaking)Towards the middle of the nineteenth centurythe orderly progression of one style to

another, which had occurred during the previ-ous 150 years, collapsed. Influences fromabroad, notably Bohemia, had a marked effecton design in Britain, and the introduction ofnew production techniques, such as pressmoulding from America, led to cheaper glass-ware which catered for a much wider marketthan ever before. National and internationalexhibitions, culminating in the GreatExhibition of 1851, introduced many newstyles to the British public. The purpose-builtbuilding in which the Great Exhibition washeld, the Crystal Palace, was an enormousglass and cast iron construction. One of itsprincipal features was a glass fountain weigh-ing about 4 tonnes built by the firm of F & COsler of Birmingham.

At the exhibition the finest designs and thenewest manufacturing methods from all partsof the world were displayed. When it wasover, glassmaking in Britain changed radically:as the Industrial Revolution gathered momen-tum and manufacturing units became larger;small firms using traditional hand methodsbecame uneconomic. This led to the concen-tration of the glass industry into fewer compa-nies. Glassmaking declined in areas such asBristol and London, and became establishedprincipally in the Midlands, aroundStourbridge and Birmingham, and along therivers Tyne and Wear around Newcastle,where there was an abundance of fuel andraw materials. Whereas in earlier years everyglassmaker had been content to produce whatwas in popular demand, after 1851 eachmanufacturer endeavoured to produce designswhich were quite different from those of hiscompetitors. To protect their interests designswere registered, and during the second half ofthe century literally thousands of patterns forglassware were registered.

Changes in style and taste were so rapid thatit becomes difficult to date glass made after1850; but attribution to particular factories isoften easier, since it was the practice of somefactories to mark their products with the nameof the company.

The taste for continental glass was encour-aged by the Great Exhibition, and Bohemianstyles of coloured and overlay glass as well asthe French style of opal and frosted glassdecorated with coloured glass buttons becamevery popular.


Glass had to be manufactured to suit allpockets, and there was a demand for mass-produced wares for the working classes, aswell as for high-quality glass products for thewealthy. A good example of this dual produc-tion is the difference in quality that occurredin Bohemian-style ruby glass. Expensive glasswas made, by covering a clear glass vesselwith a layer of red glass. Expert craftsmenwould then cut through the outer layer toexpose panels of clear glass. Vessels producedin this way were known as Biedermeierglasses. On the other hand, cheap glasseswere produced by first cutting facets in thevessel, after which the surrounding areas werefilled in with a ruby stain. To produce theformer, foreign craftsmen were often broughtto Britain to work in their own tradition, whilethe cheaper versions could be made withalmost any class of labour.

Among the immigrant workers who settledin the Stourbridge area were several decora-tors who established high reputations for theirskill: Frederick Kny, William Fritsche and PaulOppitz all executed the most beautiful engrav-ing in the Bohemian tradition, and their workwas shown at many exhibitions. Jules Barbewas a gilder in the French style whose workhas never been bettered in Britain. All thesecraftsmen worked at one time or another forthe firm of Thomas Webb & Son. Webb’s,which was founded in 1856, and which is stilloperating, is one of several firms that estab-lished the high reputation of Stourbridge glassin the nineteenth century. Others wereRichardsons, Stevens & Williams, and Boulton& Mills. In Birmingham, George Bacchus &Sons, Rice Harris & Sons, and F & C Oslerwere well-known glassmaking firms.

These British glassmakers did not take theirinspiration solely from European sources. Forexample, in 1885 Queen Victoria received agift of ‘Burmese’ glass (which had recentlybeen invented at the Mount Washington GlassFactory in America); and in a short whileThomas Webb & Son had agreed to makeBurmese glass under licence in Britain. Thisunusual glass was shaded from pink to yellow,an effect which was achieved by adding goldand uranium to the raw materials.

Cameo glass was redeveloped in thenineteenth century by John Northwood, afamous Stourbridge glassmaker, who was

associated with the firm of Stevens & Williams.Cameo glass was always extremely expensivesince it required great skill and patience onthe part of the carver. One of the best-knowndecorators in this field was George Woodall,who worked for Thomas Webb & Son. Hespecialized in carving classical figures inflowing robes; signed examples of Woodall’swork command very high prices. Variousmethods were invented to produce glasswhich had the appearance of cameo glass butwhich could be mass-produced.

Although glassmaking largely died out inLondon, the oldest glassmaking firm in Britainwas a London company, J Powell & Sons,which was established about 1700 under thename Whitefriars Glass (Evans et al., 1995;Jackson, 1996, Watts, 1999). The firm madeglass for William Morris and has been respon-sible for fine handmade glassware in thetwentieth century. The other main centre ofglassmaking in the nineteenth century was inthe north of England. The area aroundNewcastle-upon-Tyne produced much of thecheaper pressed glass for the mass market.The technique of press moulding was devel-oped to a very high standard, and a widevariety of goods was made. The most soughtafter glassware was made from slag glass. Thiswas an opaque glass in a variety of colours,the most typical of which had a purple andwhite marbled effect. The three most impor-tant companies which made slag glass, andwhich often added their trade mark to themoulds so that their products could be identi-fied, were Sowerbys Ellison Glass Works, G.Davidson & Company, and H. Greener &Company. In addition to moulded glass muchcoloured decorative glass was made in thenorth of Britain. It derived from the Venetiantradition with its liberal use of applied glassdecoration. Coloured glass baskets, spill vasesand candlesticks were also popular.

In the twentieth century several attemptshave been made to establish artist-craftsmenstudios in order to break away from the massproduction methods of the large factories.Among the better known of these are thenames Greystan and Monart.

IrelandThere is documentary evidence for glassmak-ing in Ireland from 1258 onwards, when


French glassmakers began to operate there.They probably made their way to Ireland fromthe glassmaking centres in Sussex. In 1586 aCaptain Thomas Woodhouse acquired a patentgiving him the sole right to make glass inIreland for eight years. However, the venturedoes not appear to have been very successful,for in 1589, George Longe, a trained Britishglassmaker with considerable interest in theBritish glass trade, bought the patent. Longe isthe first recorded experienced British glass-maker; until this time the glassmakers had allbeen French or Italian. The Act of 1615banning the use of wood for firing glassfurnaces was not extended to Ireland until1641, so it is quite possible that there was afurther influx of French glassworkers from thesouth of Britain after the Act was passed. After1641 glassmaking in Ireland seems to havedeclined.

The first glasshouse producing lead glasswas set up in Dublin in 1690 by a CaptainRoche, barely 10 years after Ravenscroft’spatent had expired. Roche went into partner-ship with Richard Fitzsimmons, and thecompany continued in business until about1760. At the same time as the Excise Act of1745 imposed a duty on glass made in Britain,Irish glass was exempted, but its export toBritain was prohibited. This badly affected theIrish glassmaking trade since the number ofthe island’s inhabitants able to afford goodquality glassware was too small to support anygreat manufacturing capacity. At the same timethe effect on the Irish home trade was aggra-vated by the fact that imports from Britainwere not banned. However, in 1780 the grant-ing of free trade and the lifting of all restric-tions on the industry put new life into the Irishglass trade. The art of glass-cutting wasbecoming fashionable in Britain at this timebut, as mentioned previously, heavy dutieshad an inhibiting effect upon this type ofdecoration. The British manufacturers there-fore saw the creation of free trade for Irelandas a golden opportunity to cater for the publicdemand without having to pay the heavyduties they would incur at home.

The reputation of Irish glass was madebetween 1780 to about 1830; and glasshouseswere established in Cork, Waterford, Dublinand Belfast. Those amongst them which werefinanced by Irishmen, such as the Penrose

brothers at Waterford, relied entirely, in thefirst instance at least, on imported Britishcraftsmen to produce the glass. In 1825 theIrish free trade was ended, and the advantageof manufacturing in Ireland was lost. This,combined with an increasing floridity of style,caused a decline in the trade and the impor-tance of Irish glass.

Before the use of cut decoration, theproduct of the Irish glasshouses was indistin-guishable from that of the British factories.After 1780, when cutting became fashionable,the same patterns were produced in bothcountries, but there were certain characteris-tics, which help to distinguish the Irishproduct. Most important among these was theoccasional practice of impressing the factoryname on the base of tableware, particularlydecanters. Examples are the names of B.Edwards, Belfast, Cork Glass Company, andPenrose, Waterford. Certain styles of engrav-ing are associated with particular factories, andsome styles of cutting are peculiarly Irish(Warren, 1970).

Irish glass is now synonymous with thename Waterford. There is no doubt that theWaterford glass factory did acquire a very highreputation for its products, but there wereother important factories also operatingbetween 1780 and 1850: Cork Glass Company;Waterloo Glass House, Cork; Terrace GlassWorks, Cork; B. Edwards, Belfast; Belfast GlassWorks; Richard Williams & Company, Dublin;and Charles Mulvaney, Dublin.

The history of the Waterford Glass House iswell documented by advertisements, factoryrecords and correspondence betweenmembers of the Penrose and Gatchell families.The factory was founded in 1783, by thebrothers George and William Penrose; wealthymen who knew nothing of glassmaking, butcould see the opportunities which free tradeoffered. Initially the factory was staffed byBritish craftsmen under the leadership of JohnHill of Stourbridge, who was an experiencedglassmaker. Hill stayed for three years andthen left, apparently in disgrace. Beforeleaving, Hill handed on his glassmakingsecrets, for mixing the raw materials, to aclerk, Jonathan Gatchell. As the holder of suchimportant information, Gatchell grew to be animportant figure in the business, and in 1799,with two partners, he bought out the then


remaining Penrose brother. Thereafter thecompany remained in the hands of theGatchell family until it closed in 1851.

North America 1608–1940

Glassmaking in North America began in 1608,little more than a year after the first settlershad arrived. The glasshouse owned by theVirginia Company was situated at Jamestown(near Richmond, Virginia) and covered an area15 m by 11 m. It was considerably larger thanany of the contemporary medieval glasshousesin the Chiddingfold region of southern Britain.The furnace itself was rectangular and, like somany of the British examples, it was found tobe devoid of useful fragments of the articleswhich had been made in it because the sitewould have been cleared of glass for cullet(Harrington, 1952). It seems that the James-town glasshouse had been set up to make useof the extensive source of fuel from the forestswhich were now available in the New World,and thus to provide cheap glass for use inBritain. This at a time when British glassmak-ers were discouraged, and eventually prohib-ited, from burning wood in their furnaces, andcoal-firing was still in its infancy. There washowever not a large local demand for glass,either as tableware or as windows, in thepioneer Colonial houses, and the project wasabandoned in 1624.

During the next 150 years there wereattempts to set up glasshouses, one of thelonger-lived being the glassworks built in 1738by Caspar Wistar at Alloway, in Salem Co.,New Jersey. This began the German domina-tion of the American glass industry that wasto continue until the nineteenth century.Wistar died in 1752 and his son, Richard,managed the modest business in an effectivemanner until 1767. Some of the glass madethere has been found to contain about 17 percent lead oxide, which probably got into theglass from the addition of English 30 per centlead oxide cullet in an attempt to improve theglass. A more spectacular venture was startedin 1763 by a ‘Baron’ Stiegel, first atElizabethtown (near Lancaster, Pennsylvania)and then at Manheim nearby. Stiegel ran theglassworks in a flamboyant manner, eventuallybecoming bankrupt and ceasing operations in1774. The products of the Stiegel and Wistar

glasshouses possess considerable artistic merit,but care needs to be exercised when usingcontemporary sources of information asevidence for the quality of the glass because,at that time, it was politically expedient toconceal from the British Colonial authoritiesthe success which had been achieved byAmerican glassmakers. Thus Benjamin Franklininstructed his son William, Governor of NewJersey, to report in 1768 that Wistar made only‘coarse window glass and bottles’. Stiegelimported glassworkers from Europe; andeighteenth-century American glassware wasmade in the German and British styles, thusmaking it difficult to distinguish from importedwares. Exceptions to this were the mould-blown flasks and bottles, on which the decora-tion was often of a commemorative nature, forexample, those dating from after 1830, bearingthe legend ‘Union and Liberty’.

The Amelung New Bremen Glassmanu-factory was set up in 1784, by JohannFriedrich Amelung, after the War of Inde-pendence (1775–1783). It was sited at NewBremen on the Monocasey River in FrederickCounty, Maryland. Amelung had previouslyworked at the famous Grünenplan glassworksin Germany (south of Hanover) where glasshad been made since the fourteenth century.The British tried very hard to prevent Amelungfrom going to America to set up the glass-house, first by arranging for the HanoverianGovernment to forbid emigration and secondlyby trying to capture the vessel during itspassage through the English Channel, butAmelung avoided both hazards.

The Amelung New Bremen factory wasburnt down in 1790 but some manufacturewent on until 1795 (Schwartz, 1974; Hume,1976; Lanmon and Palmer, 1976). The site wasexcavated in 1963 (see Figure 3.70) and analy-ses were carried out on many samples of theglass. Their ‘fine glass’ (colourless, purple orblue) was found to be a high-potash glass witha moderate lime content (16 per cent K2O and9 per cent CaO), but the green, aqua andamber glasses had much less potash and muchmore lime (5.7 per cent K2O and 19.6 per centCaO). Brill and Hanson (1976) suggest that aconstant batch composition had beenemployed for both types, but purified woodash (pearl ash) had been used in the formerand the unpurified ash had been used in the


latter. Other interesting features were thepresence of antimony and certain traceelements; and those glasses made before 1790contain some lead oxide (0.11–0.64 per centPbO) while those made after 1790 containmuch less (0.0–0.04 per cent PbO).

Apart from the successful AmelungGlassmanufactory, the intense demand forforeign commodities (following the War ofIndependence) operated as a strong deterrentto the construction of glassworks in the UnitedStates. For this, and for other reasons, such asforeign competition and local taxes, glassmanufacture in the United States of Americadid not make much progress, but between1786 and 1800 glassmaking was carried out inat least six New England states, and monopo-lies were soon granted; for example, in 1787the Boston Glass Manufactory had a 15-yearmonopoly, its capital stock was exemptedfrom all taxes, and its workmen were relievedof military duties.

The British blockade of the NapoleonicEmpire brought prosperity to America, and thiswas reflected in an increase in the number ofglasshouses, so much so that the first nationalsurvey of manufactures, in 1810, showed that22 glassworks existed, again all on the easternseaboard, with a total output of more than $1million; window glass accounted for four-fifthsof the total, the amount of lead glass beinginsignificant.

The end of the war with Canada(1812–1814), however, not only put an end torapid expansion of the glass industry but alsobrought distress to many of the glasshousesthat had recently begun operations. In 1817the New England Glass Company was formed,under the ownership of E.D. Libbey. It wasthreatened in 1885 by a paralysing strike ofworkmen. Libbey, however, broke the strikeand transferred his whole company across theAlleghenies to Toledo, Ohio, where he madea fresh start in 1888. The new companyultimately became the Owens-Illinois GlassCompany, the largest and most influentialmaker of glass containers in the world (Meigh,1972).

The Census of 1820 showed that deepindustrial depression had returned to America,the total annual value of the glass outputbeing only about $0.75 million, much less thanin 1810, but 5.4 million ft2 (5 �105 m2) of

window glass was made, high-quality crownglass being produced in Boston, and cylinderglass elsewhere.

Manufacturing was again concentrated inthe east coast states, and an excellent sourceof sand was found at Milville (New Jersey),now the homes of the Wheaton GlassCompany and the Kimble Glass Company.Wood was generally used as the fuel, exceptat Pittsburgh where coal was used, and thefurnace designs were based on those ofGermany, whence the migrant glassmakershad come. Some details are available aboutthe consumption of wood in Canada, wherethe Ontario Glass Manufacturing Companyleased 1500 acres of land in 1810. The areawas richly covered in beech, hickory, oak andmaple, but fourteen years later the activities ofthe glassmakers had cleared 200 acres;however, the owners of the land sold thiscleared area as valuable farmland.

There was a resurgence in the building ofglasshouses between 1825 and 1831, and ahigher proportion of lead glass was made.When the fourth Census was carried out in1850 the annual value of the glass producedhad increased to $4.6 million, and then to $8.5million in 1860. The Civil War (1861–1865)brought about a considerable increase inproduction. In 1860 there were 1416 workersin 13 factories but by 1870 these figures hadrisen to 2859 and 35, respectively.

In the early nineteenth century the evenolder technique of moulding glass in a closedmould was revived and improved by theprocess of press moulding. By this method flator tapered vessels were formed in a two-piecemould; one half being used to compress theslug of molten glass into the shape of theother half so that at the point at which thetwo halves met the mould was full (and therewas usually a mould-line to confirm this). Withthis method far more intricate patterns couldbe produced with much sharper definition,and perhaps even more importantly, articlescould be produced at a faster rate. The earli-est recorded patent for press moulding wastaken out in 1829 when presses were manuallyoperated (see also Cable, 1998).

During the next 50 years a whole succes-sion of patents was issued covering bothimprovements in power presses and in mould-ing processes whereby several articles could


be moulded at one time. Press mouldingenabled a wide variety of glassware to bemade at prices that put cheap copies of cutand decorative glass into every home. Similardevelopments took place in France, Bohemiaand Canada.

In the context of the exchange of ideasbetween Britain and America, it is interestingto note that Harry Northwood, son of one ofthe greatest of the Stourbridge glassmakers,John Northwood, emigrated to America in1885, and three years later was running hisown glassworks. As well as the production ofutilitarian wares, the continuing search fornovelty culminated in a wide variety of flasksand bottles made to represent people, animals,birds, fruit, buildings and railway engines toname but a few. Towards the end of thenineteenth century American glassmakersbecame involved in the Art Nouveaumovement, and much elegant and decorativeglass was made in the Art Nouveau style. LouisComfort Tiffany (1848–1933) and FrederickCarder (1863–1963) were prolific glassworkersin the style. Tiffany had a studio in New Yorkwhere a whole variety of imaginative colouredglass was produced whose chief attraction wasits iridescence. One of Tiffany’s largest under-takings was a remarkable glass curtain for thestage of the National Theatre of Mexico, theBellas Artes, in Mexico City. Carder learnt hisskills at Stourbridge, UK, later emigrating toAmerica and helping to found the SteubenGlassworks, Corning, New York State. Hereglass artists have continued to create individ-ual works which explore to the full the possi-bilities of shape and decoration.

The first use of natural gas in the glassindustry (1881) gave another great impetus toglass manufacture, and there was an appre-ciable movement of glassworks to the gasfields of Pennsylvania, Indiana and Ohio. Thetotal number of glassworks increased by 50per cent between 1880 and 1890 and the firstcontinuous-melting tank was installed atJeannette (Pennsylvania) in 1888. Glass-workers’ Unions and Manufacturers’ Associa-tions started to be formed around 1880. Afterthat time the story is one of growth andremarkable improvements in manufacturingmachinery, the Arbogast press-and-blowmachine being invented in 1882, the OwensSuction machine in 1904 etc. These develop-

ments and the many others in making windowglass are recorded by Douglas and Frank(1972) and, as a historical record, by Cable(1998).

Summary of nineteenth- and twentieth-century glassmaking in Europe

The second half of the nineteenth century wasa period of enormous technical achievement,with Britain and Bohemia leading the field inthe middle of the century, but France comingvery much to the fore in artistic glassmakingtowards the end of it, and the great industrialpower of America gradually making itself felt,albeit mainly in styles imported from Europe.Apart from the traditional techniques ofenamelling and gilding, wheel-engraving andwheel-cutting, this period saw the introductionof transfer printing and acid-etching; surfacetreatments such as the revived ice-glass andacid-etched satin finishes; silver and gold leafdecoration, aventurine, and an electrolyticsilver deposit process; bubbles of glasstrapped within the glass; glass shading fromone colour to another owing to heat treatment(Burmese, Peach Blow etc.); wrought decora-tion of every kind, with twisted and ribbedelements, opaque and coloured twists andthreads; and applied drops, and threading (bymechanical process).

From the nineteenth century, iridescentglassware was produced in Europe, andgained great popularity, e.g. Loetz; carnivalglass. Opalescent glasses treated with metallicoxide were heated in a controlled atmosphereto develop the iridescent effect. It wasproduced commercially in 1863 by J & J Lob-meyer, and thereafter by many glasshousesunder various patents.

All these methods were used in an infinitevariety of combinations, reflecting the eclectictaste of the period. Of perhaps the greatestsignificance, however, was the perfection ofpress-moulding in America about 1825, for thisbrought decorated glass within the means ofthe poorer classes of society. In France in the1870s the multiplicity of techniques washarnessed to the making of individual worksof art, by Eugène Rousseau and Emile Gallé.Rousseau was greatly inspired by Japanese art;early glass pieces reflect this in their themes,forms and their decorative composition. Later,


however, Rousseau developed original shapesand decorative techniques of enclosed red,green and black markings, of craquelure andof deep wheel-cutting. Rousseau abandonedglassmaking in 1885. Gallé had begunmanufacturing glass in 1867, and after a periodof experimentation with every availabledecorative technique in a variety of historicstyles, Gallé also came under the influence ofJapanese art. About 1885 Galle developed alyrical style in which vases with mainly floralpolychrome designs were made by compli-cated casing techniques, with much use ofwheel-cutting and acid-etching. Such vasesepitomize the Art Nouveau style in glass andwere highly influential until the First WorldWar (Arwas, 1977). Of equal complexity andsophistication were the glasses of theAmerican Louis Comfort Tiffany, who usedembedded drawn threads and other forms,combined with an iridescent surface treatmentto produce glass objects equally characteristicof the period.

From the 1830s to the 1940s, radioactiveuranium (first discovered in 1789) was widelyused to produce colours ranging throughyellow, green, orange-red and black, in glass,glazes and enamels on decorative objects. Themost well-known colours are those in vaselineglass and fiesta red glass (Strahan, 2001).

After the disruption of the 1939–1945 war,a neutral country – Sweden – took the lead inglass fashion. Two artists, Simon Gate andEdvard Hald, employed by the OrreforsFactory, devised a variant of Gallé cased glass,Graal glass, covering the cut and colouredlayers with a colourless coating, thus embed-ding them and imparting a smooth surface tothe glass. They also produced designs forwheel-engraving on crystal glass. The generalSwedish ideal was to produce well-designedgoods for everyday use, apart from the luxurywares; and in this their lead was followed byother countries in northern Europe.

A more personal style of glassmaking,however, was developed in France by MauriceMarmot, a Fauve painter who not onlydesigned, but made glass, using techniques oftrapped bubbles, powdered oxides, and deepacid-etching in glass objects which recoveredthe forms and styles of the 1930s. This sponta-neous approach to glassmaking was reflectedin the work of other French contemporaries:

Henri Navarre, André Thuret and Jean Sala; andeven the commercial firm of Daum abandonedthe traditions of Gallé for glass made underMainot’s influence; while René Lalique applieda highly sophisticated taste to mechanicalproduction by designing glasses with modelleddecoration in high relief which could beenlivened by acid-etching and enamelling.

The northern European striving afterfunctionalism was echoed in Murano in 1920,notably in the work of Paulo Venini, ErcoleBarovier and Flavio Poli. After the decline ofVenetian glassmaking in the eighteenthcentury, the skills and traditions were keptalive through the nineteenth century mainly bythe enterprise of Antonio Salviati, and were atthe service of the new style when it came. Thesimple shapes of Venini were often decoratedby the thread and mosaic techniques in theopaque white (lattimo) and coloured glassestraditional to Venice, while both Venini andBarovier invented new surface treatments togive their various glass products mottled,granulated or dewy effects. The work of thesemasters continued after the Second WorldWar, but this period has been chiefly charac-terized by the emergence of the studio glass-maker, first in America – Harvey K. Littleton,Dominick Labino and others – and then inEurope. Working single-handed, at small one-man furnaces, studio glassmakers haveproduced glassware in an infinite number ofshapes, often non-functional and sculptural,decorated with bubbles, embedded colouredmetallic powders, and other furnace-madeornaments, latterly supplemented by wheel-engraving (Charleston, 1977).

Glass micromosaics

The taste for miniatures is seen in many artforms during the eighteenth century, at theend of which, the technique of creating small-scale glass mosaics was developed in theVatican workshops in Rome. These came tobe known as micromosaics (Ital. mosaico inpiccolo) to distinguish them from the largerarchitectural works. The technique was usedto create portraits and pictures to embellishdecorative objects such as jewellery, snuffboxes and caskets and the tops of tables(Figure 2.17).


Two important events heralded its begin-ning. One was the excavation in Hadrian’svilla in Tivoli, in 1737, of the finest ancientRoman mosaic wall panel ever discovered(now in the Capitoline Museum, Rome). Thepanel (known as the Doves of Pliny), measur-ing 98 � 85 cm (381⁄2 � 331⁄2 in) and wascomposed of tiny stone tesserae arranged torepresent doves drinking from a bowl. Thesecond was the development of a wide rangeof opaque matt colours by Alessio Mattioliwho was in charge of the production ofmosaic tesserae produced for the Vaticanstudio from circa 1730 to 1750. These addedconsiderably to the range of transparent, shinycoloured glass tesserae produced in Venice,and enabled mosaicists to imitate paintingsmore successfully.

During the nineteenth century, micromo-saics gradually began to look less like theancient Roman mosaics, owing to an increas-ing emphasis on naturalistic effect. In the lateeighteenth micromosaics a recurring featurewas the use of two-dimensional neoclassicalmotifs silhouetted against a contrastingbackground.. The next development was anattempt at greater naturalism in modelling,achieved by enhancing the effect of the gradu-ated colours by varying the angle at which thetesserae were set; and by adopting a three-quarter, as opposed to frontal or profile viewof the image, which was still set against acontrasting monochrome background. The

ultimate exercise in sculptural modelling wasthe execution of tesserae en grisaille in imita-tion of marble statues. Other highlights of themicromosaic technique were landscapes,portraits, notably papal portraits, and multi-coloured flower subjects.

As the tourist industry developed towardsthe second half of the nineteenth century, theproduction of micromosaics became acommercial enterprise, and standardsinevitable dropped as easily portable souvenirswere produced. However, expensive, high-quality work continued to be produced ondemand up to the end of the nineteenthcentury.

The glass itself was produced in the form ofcakes (smalti) in a vast range of colours,sometimes two or more colours were blendedtogether. The cakes were then broken, and thechips softened in the furnace-mouth whilstbeing held by long tongs. The glass was thendrawn out to form long thin threads (smaltifilati), which were cooled in order to hardenthem, and finally sliced into thin sections. Thetesserae are so minute that some of the micro-mosaics are composed of more than 5000pieces per square inch. The portrait or picturewas created by placing a support covered witha slow-drying adhesive on an easel. Thesections of coloured canes held in tweezers,were arranged by colour and shade one at atime. The support was usually made from athin sheet of copper metal with a turned up


Figure 2.17 Fragment of a broken micromosaic plaque, which depicts a bridge viewed from below, showing themany individual slivers of coloured glass that form the design. W of rectangular glass surround 270 mm. (©Copyright The British Museum).

rim, but slabs of marble, hardstone or glasswere hollowed out to create supports formicromosaics. The adhesive was composed oflime burnt from marble and finely powderedTravertine stone, mixed to a thick consistencywith linseed oil. Once the adhesive holdingthe glass canes in place had hardened (whichmight take weeks or months depending on thesize of the work), any gaps in the surface werefilled with wax, the entire uneven surfaceabraded level. This was done first with a hardstone to flatten the surface, then with emery,and finally with lead to impart a polish. Finallythe surface was waxed and polished. A fulldescription of the development of micromo-saics, and an explanation of the glassmakingtechniques associated with it are given byRudoe in Gabriel (2000).

Enamels (glass fused to metal)

Objects composed of vitreous coatings fusedto a metal backing are known as enamels. Thetechnique of using enamel colours was knownin Egypt from circa 1400 BC; it was used to alimited extent in Greece, but extensivelypractised in Byzantium in the sixth century, inVenice from the fifteenth century, andelsewhere in Europe from the sixteenthcentury. Since earliest times, enamels havebeen made in the form of jewellery andcontinue to be produced as jewellery, candle-sticks, boxes, panels, dressing table sets etc.,and since the nineteenth century, for commer-cial and industrial purposes such as signs andadvertisement plaques (Maryon, 1971;Cosgrove, 1974; Speel, 1984; Wicks, 1985).There are several types, classified mainly withFrench terms, by the way in which the enamelfrit is attached to the metal, i.e. cloisonné,champlevé, basse taille, plique à jour andpainted (see Figures 3.33-3.38).

Enamel resembles a glaze in that both vitre-ous materials are fused to a substrate –ceramic in the case of glaze and metal in thecase of enamels. However, the term enamel isalso used to describe a colour of similarcomposition used to decorate ceramic andglass objects. An enamelled object is formedby applying a vitreous substance in the formof a dried frit, to a metallic surface such ascopper, silver or gold. They are then fused at

a relatively low temperature in an enamellingoven. Some enamelling techniques relied on aframework of metal bands or wires to containdifferent coloured enamels, whilst the successof others depended upon a more sophisticatedknowledge of enamel composition and firingtemperatures. With each colour having to befired at a successively lower temperature, thetechnique was, and continues to be, painstak-ingly slow, and is prone to irreparable flaws.However the result are permanent brilliantcolours, which do not fade.

Most early enamels failed because the highlyfusible frit never actually fused to the metallicsubstrates. Enamels must be so formulated asto have a co-efficient of contraction roughlyequivalent to that of the metallic substrate; andits melting point must be approximate to butlower than that of its backing to ensure fusion.For these reasons most enamels are alead–soda or lead–potash glass with or withoutthe addition of colourants and opacifiers. Onvery thin or extensive areas of metalwork, thecontraction of the enamel on cooling might besufficient to cause the metal to warp. Tocounteract this, the reverse of the object mightalso be enamelled, as process known asenamel backing or counter enamelling. Wherenecessary, the surface of the finished enamel,when cold, was made even and polished withfine abrasives. Opaque enamel usually requireda lower firing temperature (petit feu) thantranslucent enamel, about 300°C. Highertemperature firing was known as grand feu.Translucent enamelling involves the firing oftransparent layers of enamel, onto a metalguilloche surface, engraved by hand or engine-turned. There may be as many as five or sixlayers of enamel, each of which had to be firedseparately at successively lower temperatures.

Sometimes gold leaf patterns or paillons orpainted decoration or scenes were incorpo-rated in the design. This effect was achieved,by applying and firing the gold leaf or enamelonto an already fired enamel surface, afterwhich it was sealed with a top layer ofenamel, and refired. The completed enamelthen required careful polishing with a woodenwheel and fine abrasives, to smooth down anyirregularities in the surface, and then finishingwith a buff.

The characteristic milky quality of some ofFabergé’s translucent enamels was obtained by


mixing four to six parts of translucent enamelto one part of opaque enamel, producing asemi-opaque or opalescent enamel. Enamelswere usually applied to the metal in such a wayas to form a level surface with the surroundingmetal, by providing a sunken area into whichthe frit could be fused. These areas might beprovided for in the original metal casting, or cutout with scorpers, a technique known aschamplevé (or en taille d’épargne). Used by theRomans, Celts and Medieval enamellers, thechamplevé technique was perfected in Mosanenamels of the twelfth century, and used exten-sively in Limoges, France in the twelfth to thefourteenth centuries.

In cloisonné work, the enamel was placed incompartments (cloisons) formed by a networkof thin metal bands soldered onto the surfaceof the metal object to be decorated. The tops ofthe metal bands remained exposed, definingthe pattern of the cells filled with differentcoloured enamels. Cloisonné is one of the mostancient enamelling processes and is probablyNear Eastern in origin. It was used by the Celtsand by Byzantine enamellers. The Pala d’Oroin St Mark’s, Venice, is a masterpiece ofcloisonné work. From Byzantium the techniquewas transmitted to China in the fourteenthcentury, and eventually throughout the Orient(Cosgrove, 1974).

In filigree enamelling, thin wires enclose theenamel in the same manner as strips of metalin the cloisonné technique. It is thought tohave originated in or near Venice in thesecond half of the fourteenth century, and isbest known from fifteenth century Hungarianexamples.

The technique of plique à jour involvessupporting the enamel at the edges in a metalframework, without the support of a metalbacking. The framework was produced byfretting (piercing through and removing mostof) a small sheet of metal, which was thentemporarily backed with a material such assheet mica, to which enamel would notadhere. Frits formulated to produce translucentenamels, were placed in the compartments,fused and cooled. On removal of the tempo-rary backing, light could shine through thecoloured enamels, producing an effect similarto that of a stained glass window.

Apart from the use of plain, or colouredenamels, a number of techniques for enamel

decoration were employed which were verysimilar to those on glassware, such as mosaicand millefiore. The metalwork itself could bedecorated before application of enamel, as forexample, in the technique of basse taille,which entailed chasing or engraving the metalsurface with a delicate design in low relief,before the application of translucent enamel.The engraving could be seen through theenamel, which itself appeared lighter over theshallow cutting and darker over the deeperareas. A refinement of the champlevétechnique, basse taille, was first perfected inthe Paris enamels of the early fourteenthcentury. The Royal Gold Cup in the BritishMuseum is a magnificent example of thetechnique. The cup was made circa 1530 inBurgundy or Paris, and depicts scenes fromthe life of St Agnes. An almost identicalprocess, lavoro di basso rilievo, was developedindependently in Italy.

The process of en resille sur verre entailedpacking the enamel frit into gold-linedincisions engraved on a medallion of blue orgreen glass, to which it fused on firing. Thisdifficult technique was only adopted duringthe second quarter of the seventeenth centuryin France, where it was mainly used todecorate miniature cases. A rich technique ofcovering figures or decorative devices formedin the round, with opaque enamel, en rondebosse, was used in Paris at the beginning ofthe fifteenth century on large reliquaries, andin England, where it was known as encrustedenamelling. The Dunstable Swan Brooch inthe British Museum is a fine example of thisprocess.

Enamel colours could be applied to a metalfoundation to produce a picture or paintedenamel. (Glass painters are recorded inMurano, Italy as early as 1280.) The colourswere formed of metallic oxides mixed with aglassy frit of finely powdered glass andsuspended in an oily medium for ease ofapplication with a brush. The object, such asa plaque or box, was first covered with a layerof white opaque enamel and fired. The designand/or inscriptions were applied in differentcolours, and fired in a low temperature mufflekiln (approx. 500–700°C). The medium burnedout during firing. Sometimes different firings atsuccessively lower temperatures were requiredto fuse different colours, in order to prevent


them from running into one another. Theprocess was invented in Limoges, France inthe late fifteenth century. Led by father andson Jean and Henri Toutin, French goldsmithsdeveloped a sophisticated technique for paint-ing polychrome miniatures in enamel on gold.A locket bearing Henri Toutin’s signature infull and the date 1636 is the earliest date toappear on one of the pieces which havesurvived.

In England and the Netherlands, the style ofpainting was copied in monochrome on thebacks of many items of jewellery during themiddle decades of the sixteenth century. Fromthe mid-sixteenth century enamels were oftenpainted on black grounds instead of white,which thus formed the basis for grisaille paint-ing. If the colours were used on a blackground, thin layers of gold- or silver foil couldbe inserted between coats of translucentenamel to provide a warm or cold reflectivetonality. Large painted enamels were producedin Limoges, including portrait plaques, platters,plates, dishes and ewers. From the sixteenthcentury onward, goldsmiths enriched theirwork with touches of coloured enamel.

In order to protect soldered joints in themetalwork, these had to be coated with rouge,whiting or plaster-of-Paris before the enamelwas fired. A cheap form of enamelware madein various parts of Europe in the seventeenthcentury came to be known as SurreyEnamelling. The enamel was applied to brassobjects, cast with recessions to hold the frite.g. candlesticks, andirons, sword hilts andhorse harness.

During the sixteenth century, paintedportrait miniatures had developed in England,from a technique of illuminating manuscripts,known as limning. The technique of produc-ing enamel portrait miniatures was introducedto England from Sweden and France in the1680s by foreign enamellers such as Petitot,Boit and Zincke. As the technique developed,larger and more ambitious works were under-taken; for example, Henry Bone (1755–1834),made large copies of Old Master paintings.

Between 1842 and 1918, the Russian firm ofFabergé (Peter Carl Fabergé 1846–1920),produced enamelled jewellery, cases, boxesand ornaments decorated with precious andsemi-precious stones, and enamels of thehighest quality in terms of their evenness and

smoothness of texture. The company special-ized in the covering of comparatively largesurfaces or fields, termed as en pleinenamelling, since the enamel was applieddirectly to an object and not to plaquesattached to the surface. Enamelled itemscontinue to be produced both by individualsand in small workshops, and industrially forcommercial purposes.

Pictures created on or of glass

There are many types of historical picturescreated on or of glass. In the case of picturescomposed of glass beads or of minute glasstesserae (micromosaics), conservators may berequired to deal with problems of glass deteri-oration or of the materials used to bond theglass or tesserae to a substrate. In the case ofpictures painted either on the front or reversesurface of a panel of flat glass, the usualdamage is the fact that the glass panel hasbroken. However, problems associated withthe paint itself will necessitate the involvementof a specialist painting conservator.

• Paintings in oil or water colours on thefront of a panel of glass, which wereextensively produced by amateur artists inCentral Europe from the seventeenthcentury onwards.

• Reverse paintings on glass (commonlyknown by the German term, hinterglas-malerei), in which the design was executedin reverse order (i.e. details first,background last) on the back of a sheet ofglass (see below). This group also includesreverse foil engraving (Figure 2.18),amelierung, eglomisé (sometimes referredto as verre eglomisé), painting combinedwith mirroring (mirror-painting), andmezzotints (or English glass pictures(formed by adhering prints to the under-side of flat or convex glass, removing thepaper whilst leaving the ink and thenpainting). (Silhouettes were produced bythe techniques of reverse foil engraving orbackpainting) (Figure 2.19).

• Luminaries – coloured etchings on glass.• Photographic images on glass which are

composed of different photographicmaterials and processes. They include


lantern slides – a composite made up of aphotographic image, binder and support,which are encased by a paper matt andcover glasses and bound around the edgeswith a gummed black paper tape; imageson opal plate glass (see Figure 7.69) chrys-toleum – hand-tinted or paintedphotographs adhered to the underside ofslightly convex glass panels; and colouredphotographic images (see Figure 7.70),frequently mistaken for very fine miniaturepaintings on glass.

• British Victorian crazes or manias, decalco-manie, potichimanie, and vitre mania,dating from the mid-nineteenth century.Decalcomanie, was the art of decoratingany plain smooth surface with brightly

coloured paper scraps. Potichimanie, wasthe art of decorating a plain glass or porce-lain vase or jar (potiche) by pasting thecoloured paper scraps to the interiorsurface. In both techniques, the entiredesign was then given a coat of protectivepaint or varnish. In vitre mania, decorativepatterns produced as transfers wereapplied to window glass by a simpleprocess to resemble stained glass windows.Around 1850, the term Pearl painting orOriental painting was used to describe amontage of coloured tinsel and silver paperpressed on to the glass and then framed.Flower studies were painted on glass intransparent oil stains, the background filledin with opaque light tints or with the solid-ity of lamp-black. Tin-foil was crushed,then smoothed out and placed behind thefloral forms so that it glittered whenviewed through the glass.

• Architectural glass, which includes mosaicsand opus sectile, stained glass windows andcloisonné glass windows and panels.

• Micromosaics, composed of minute glasstesserae adhered to a sheet of copper orother substrate, used as jewellery and tomake small souvenirs and copies of easelpaintings (see Figure 2.17).

• Enamels, vitreous material fused to a metalsubstrate and used for decorating smallobjects such as jewellery and boxes (seeFigures 3.4, 3.29, 7.57 and 7.58).

• Pictures composed of glass beads adheredto a piece of flat glass or other substrate.

• Pictures composed of coloured wax on aflat piece of glass.

• Chinese nineteenth-century glass snuffboxes painted on the interior.

History and technology of reverse glassdecoration with cold painting and foilengraving

As an art form, the beginnings of paintingsexecuted in reverse on glass (hinterglas-malerei) can be traced back to ancient times.Painting on the reverse of colourless or tintedbut transparent materials, depended on itseffect on the particular material used. Thematerials include coloured or lightly tintedglass, rock crystal, mica, amber and tortoiseshell. The earliest surviving example is a rock


Figure 2.18 Detail of a reverse foil engraving on glass.

Figure 2.19 Silhouette on the reverse side of glass,paint and gilding, with a pencil drawing in thebackground.

crystal plate dated to 1500 BC, now in theHeraklion Museum, Crete.

The technique of decorating glass objectswith cold painted designs and engraved goldfoil is ancient. Cold painting was known invarious parts of the Roman Empire; numerousexamples of polychrome images on thereverse surfaces of free-blown objects survive,but complete objects in good condition arerare (see Figure 3.30). Most specimens seemto have originated in the easternMediterranean, between the first and thirdcenturies AD. The technique of fondi d’oro wasbeing developed in the third century AD, inRome, in Syria and other Islamic countries,mostly in the form of decorative bowls andbeakers. The bases of some of these itemshave been found embedded in the plasteredwalls of Roman catacombs, where they hadacted as grave markers. In late Hellenistictimes, the technique of enclosing engravedgold leaf between two layers of glass(sandwich gold) and the eggshell techniquewere practised. The techniques of cold paint-ing and foil engraving seem to have beenforgotten after the collapse of the RomanEmpire: a vessel with cold painted decorationis not known again until the sixteenth century.

The production and high quality of reversepaintings on glass was much greater than isgenerally recognized. In the eighteenth andnineteenth centuries it was practised at thelevel of folk art, which was of course muchmore common than the expensive commis-sioned pictures. Despite the enormous output,the art form has not made an impression onthe history of art. Glass paintings in museumcollections are not prominently displayed;most being kept in storage. Very few museumspublish catalogues of their collections(Pettenati, 1978). There are some regionalpublications (Aigner, 1992; Schuster, 1980,1983, 1984). Research into the art of hinter-glasmalerei was carried out by Keiser (1937),Staffelbach (1951) and Ritz (1972). Research byFrieder Ryser in Berne (Switzerland),highlighted the range of hinterglasmalerei andcorrected mistaken ideas concerning theirtechnology (see Ryser in Lanz and Seelig,1999). There have been five important exhibi-tions accompanied by catalogues (CorningMuseum of Glass, 1992; Salmen (ed.), 1995,1997 (Murnau); Seelig, 1999 (Munich/Zurich);

Lanz and Seelig, 2000 (Romont/Zug) and since1998 research into Swiss hinterglasmalerei ofthe seventeenth century has been conductedby the Swiss Centre for Research andInformation on Stained Glass in Romont.

Paintings produced by painting the designin reverse on glass, are generally known bythe German term hinterglasmalerei, howeverthe term is recognized in several otherlanguages: reverse painting on glass (Eng);peinture sous verre (Fr.); vetri dispinti (Ital.);pintura en vidrio (Sp.).

Nowadays, paintings for wall-hanging arethe most commonly encountered glass objectspainted in reverse. However, glass vesselssuch as tankards and double walled bowlswere also decorated in this way. Their qualityand quantity was always dependent upon theavailability of the glass. In times of the highestproduction of clear glass, it was possible tocommission expensive paintings; whereasmass produced paintings were generallyexecuted on uneven crown, cylinder or plateglass. The refractive properties of the irregularglass surface created an effect, which couldnot be achieved with any other material ortechnique. In the production of hinterglas-malerei, the normal process of painting wasreversed. The artist began by applying thehighlights and the final detail of the image,progressing by applying successive layers ofpaint creating the larger areas of the design,ending with the background. All the detailshad to be correct, as it was not usually possi-ble to make corrections without destroying theunderlying work. The painting, viewedthrough the clear or tinted glass acquireddepth. When the painting was turned forviewing through the glass, the elementspainted on the left and right of the image werereversed, an important factor to be consideredwhen lettering was incorporated.

The technique was demanding for an artistas the work had to be visualized in its entiretybefore work began. Very often a drawing,engraving or print was translated into anoutline pattern, (the riss) to be laid under the glass sheet. The outline was redrawn onthe glass surface following the pattern. Themechanics of this process allowed the devel-opment of simple routines and repetitive massproduction of patterns in the nineteenthcentury. Each colour was applied in turn by a


different painter until the painting wascompleted. It has been calculated that in Sandl(Austria) between 1852 and 1864 the outputof a single family workshop was 386,000hinterglasmalerei (Eswarin, 1982; Knaipp,1998; Ryser, in Lanz and Seelig, 1999; Bretzand Ryser, 2000).

Judging from the lack of early documentsconcerning instruction or recording actualexperience, there appears to have been nostrict rules for the choice and use of materialsfor the production of hinterglasmalerei. Manywere created, by painting directly onto a cleansheet of glass. More often, the side of the glassto be painted was coated with a transparentground. Some of the priming substances thathave so far been identified include egg white,raw linseed oil, gelatin and spar varnish. Inaddition to oil, egg-, gum-, and casein temperawere used to produce reverse paintings onglass.

Documents dating from the sixteenth toeighteenth centuries indicate that the glass wascoated with drying oils. Linseed oil and nut oilmixed with turpentine, or spike oil used as adiluent, were preferred, as these also acted asa binding medium for the paint. Fatty acidscould be released from linseed oil, whichwould have a stabilizing effect on the paint.Certain metals (lead, cobalt and manganese)accelerated the drying of the oil. All kinds ofpaint available on the market were used forproducing the paintings, however for commis-sioned paintings, only the best and mostexpensive were used. The poisonous nature ofsome pigments containing arsenic or mercurywas disregarded (Wieck, 1981; and see contri-butions by Bretz and by Haff in Lanz andSeelig, 2000). (From about 1900, aniline dyeswere used to provide colour.) Besidespigments, many other materials used toproduce pictures on glass have been identi-fied. These were used singly or in combina-tion: bronze, silver or gold powders, motherof pearl, wax, paper, parchment, linen, silk,coloured copper engravings, painted gelatineleaves or transfers (decalcomania). After thepainting was complete, it was sometimes givena protective coating of whatever material theartist had to hand, such as linseed oil and drypigments. Other forms of backing were cloth,poor quality wood pulp paper or a woodveneer.

Reverse foil engravingReverse foil engraving is the decorativetechnique of applying gold leaf (occasionallysilver leaf or both silver and gold) to a pieceof glass, after which the design is engravedthrough the leaf with a fine needle point, andviewed through the glass (Figure 2.18). Greattechnical ability is required, for as with paint-ing in reverse on glass, no corrections can bemade. The opaque property of the foil doesnot permit a pattern to be copied from anoutline placed beneath the glass. On comple-tion, the back of the engraving was coveredwith a layer of black or coloured paint toprovide contrast and protection for the gold.Most of the engravings from the Renaissancehave been attributed to centres in NorthernItaly, notably Padua, where an account ofworking on gold foil appears at the end of thefourteenth century (Cennini, c. 1390).

The foil engraving technique found its wayto Saxony and Austria in the form of zwischen-gold beakers, goblets and decanters; to theLow Countries in landscapes and silhouettesby the Dutch engraver Jonas Zeuner(1727–1814); and into folk art in the earlynineteenth century in Silesia and Bohemia. Ablack inner drawing and translucent paint forthe shadows, etched in grisaille manner, wereadded to colourful opaque paint, or as aspecial feature, this kind of hinterglasmalereiwas mounted in lead, and was often signedand dated.

By the fourteenth century, gold foil engrav-ings were being used in Italy and France, todecorate small glass panels for religiousartefacts: crosses, reliquaries and house altars.From the fifteenth century onwards, to thebeginning of the Thirty Years’ War in Europe,hinterglasmalerei were produced by highlyskilled painters in Burgundy, Flanders andLower Rhine. The same artists were producingstained glass paintings and hinterglasmalerei.From the fifteenth century onwards, most ofthe subjects were derived from contemporarywoodcuts and engravings, after paintings bywell-known artists, and original blocks madefor producing prints. During the sixteenthcentury almost every reverse glass paintingproduced made extensive use of gold leafbeneath clear glass or transparent pigment.The refractive properties of the irregular glasssurface and protection from oxidation and


discoloration of the metal created an effectthat could not be achieved with any othermaterial or technique. The size of the glasspanels available governed the size of thepaintings which could be produced. Thedesigns or the patterned borders were oftencopied from contemporary engravings. InVenice, large circular dishes of clear glass andbowls with elaborate lattimo cane twists werefurther embellished with polychrome paintingon the plain base, often of a female head froma contemporary woodcut.

In the High Renaissance and in Mannerismthere were two distinguishable schools ofpainting, to the north and south of the Alps.In Lombardy, little reverse paintings in mutedcolours, usually on rock crystal, wereproduced in the form of amulets, pictures forhanging on walls, reliquaries and small altars;and in Nuremberg paintings with etched silver-or gold leaf combined with colourful lacquerswere produced for cabinets, backgammonsets, beakers, caskets, decorative mirrors, ringsand thimbles. Extra sparkle was given toreverse metal foil engraving with translucentlacquers by inserting a layer of wrinkled tin orsilver foil beneath them. This technique,known as amelierung, can be traced back to1532 in Nuremberg, and should not beconfused with emalieren (fired enamels). Theamelierung achieved its high point in theSwiss art of hinterglasmalerei in the first halfof the seventeenth century; the Swiss painterHans Jakob Sprüngli (1559–1637) was one ofthe outstanding practitioners. In a deliberatedeceptive technique, the naked figures in hisimages were painted on parchment and gluedonto the reverse painting on glass.

In the Austrian Tirol and in Venice, duringthe sixteenth and seventeenth centuries,3–4 mm thick glass plates, mostly with curvededges, were painted in reverse in addition tolarge circular dishes of clear glass and bowlswith elaborate lattimo cane twists on the plainbase. Themes were taken from mythology andfrom the Old and New Testaments. Tastes andfashion were beginning to change by the earlyseventeenth century. The stylistic and concep-tual attitudes of high art were adopted, theforms became more supple, colour shades tocreate an illusion of depth, and the paintingsexecuted with meticulous attention to realisticdetail. From the end of the seventeenth

century the gradual development of new ideasin the arts and sciences with their chiefcreative sources in France, Germany andEngland, changed popular designs from thebaroque to the rococo.

There was a considerable output of reverseglass paintings in Central Europe between themiddle of the eighteenth and the last quarterof the nineteenth centuries. In Augsburg, amajor centre of the graphic arts and printing,accomplished individuals such as JohannWolfgang Baumgartner (1709–1761), producedexceptional reverse glass paintings. In additioncraftsmen glass artists produced multiplecopies of paintings. Paintings from Augsburg(middle of the eighteenth century) influencedthe development of production in the‘Augsburg style’, in other localities. Because ofthe high demand, hinterglasmalerei waswidely exported (to Italy, Spain, Portugal,North America as far as the West Indies) anda warehouse was established in Cadiz (fromabout 1750) to export paintings from theproduction centres in the Bavaria andBohemia region, also the Black Forest andAlsace. In some of these areas a few paintersbecame known by name, but the vast major-ity worked anonymously, turning outenormous numbers of pictures on a widevariety of subjects. Religious subjects (e.g. thenames of saints who protected the household)were a favourite theme for ardent Catholiccustomers, as well as allegories of The Seasonsor The Continents.

It is generally agreed that the Bohemianexport of reverse glass paintings introducedthe concept to Romania, creating a virtuallyindividual style in the so-called Romanianicons. In southern Italy, particularly Sicily,production of reverse glass paintings flour-ished with other forms of folk art providingthe themes. Spanish folk production seems tohave been centred around the southern regionof Andalusia, other areas being Barcelona andToledo. Outside Europe, export led to theindigenous production of reverse glass paint-ings as far away as India, Persia (Iran),Indonesia and the Orient and North America.Production continued into the nineteenthcentury but was usually of inferior quality.

Around 1910, Wassily Kandinsky(1866–1944), Gabriele Münter (1877–1962) andother artists of the group Der Blaue Reiter in


Murnau (Bavaria) produced hinterglasmalerei(Eswarin, 1982; Knaipp, 1988; Ritz, 1972;catalogues: Corning Museum of Glass, 1992;Ryser and Salmen,1995; Salmen, 1997).

EglomiséA technique known as eglomisé, had devel-oped in the Netherlands during the sixteenthand seventeenth centuries, and was alsopractised in Saxony, Bohemia and Austria untilabout 1725, to decorate small objects such asjewellery, snuff boxes and caskets. Moreaccomplished artists painted pictures for walldecoration, using themes copied fromcontemporary engraved copper-plate prints ofcity views and landscapes.

The term eglomisé or verre eglomisé (Ital.agglomizzato) is derived from the name of aneighteenth-century French art dealer andpicture framer, Jean Baptiste Glomy. Glomysurrounded his drawings and prints with aborder of gilding and colour painted behindthe glass, a technique which then becamefashionable. Unfortunately, the term eglomiséhas been and continues to be used to describeall types of painting or gilding behind glass ofany age. Technically it is a layer of black painton a glass panel etched and then gilded ordecorated with translucent paint and thengilded (Steinbrucker, 1958; catalogues: Lanzand Seelig, 1999; Lanz and Seelig, 2000).

Paintings on mirrored glassAlthough the technique of producing paintingson mirrored glass probably originated inEngland during the late seventeenth century,the best surviving examples are Chinese (seeFigure 7.72). Trade and missionary contacts ofthe eighteenth century introduced the conceptto the Far East. Glass panels, notably bevelledplate glass made at the Vauxhall glassworks insouth London, were exported to China(especially Canton) to be decorated in thismanner for the European market (until thenineteenth century), using contemporary printsas sources for subjects. An amalgam of tin andmercury was first applied to one side of theglass in order to form the mirror. The amalgamwas then scraped away from the areas to bepainted in oils. The best quality Chinese paint-ings on mirrored glass were executed in thefinest transparent colours. They showed a highartistic ability. Poorer quality paintings were

produced for the mass export market. (Thetwentieth century saw the revival of commer-cially produced mirror paintings in the form ofadvertisements.)

MezzotintsThe technique dates back to the middle orsecond half of the seventeenth century.Mezzotints were introduced to England byPrince Rupert who bought the working detailsfrom the inventor Ludwig von Giegen. Thediarist Samuel Pepys mentions mezzotints asbeing new to England in 1665. The discoveryof mezzotint engraving, and its popularity inEngland, would seem to be the key to theorigin of reverse glass prints, since the twoprocesses are inextricably linked. Clarke(1928), ‘venture(s) to assert that the art (ofapplying paper prints to glass) came intobeing through the “call for colour” toMezzotint Engravings’. After mezzotint engrav-ings became firmly established in England, andafter publication in 1687 of The Art of Paintingin Oyl by the clockmaker, John Smith, printstransferred to glass began to be produced.

Tremain (1988) gives a detailed account ofthe history technique and conservation ofreverse glass prints. There were severalvariants of the method for producing mezzo-tints, but the general procedure was to soaka print in water for about four hours toremove the size from the paper, and thenallow it to dry. A sheet of fine Bristol glasswas then covered with an adhesive (normallyVenice turpentine) and the dry print laid facedown upon it. When the adhesive had set, theback of the print was dampened, and thepaper rubbed away with a sponge or thefingertips leaving only a thin tissue with theink adhering to it, attached to the glass. (Aftercirca 1770 the dry print was removed by theuse of glass paper.) After this had again driedout, the design was coloured in, details first,background last.

The earlier, more beautiful mezzotints owetheir brilliance to the fact that only the small-est amount of paper remained on the glass.Many display the date, a title and the name ofthe engraver and the painter below the paint-ing. Later examples made by amateurs or lessskilled practitioners have a much more opaqueand heavier appearance, owing to the greateramount of paper remaining, the required


transparency being produced by the applica-tion of a white varnish (Ryser, 1991). Backpaintings were generally produced anony-mously, although as suggested above, in itsearlier days, the engravers of the prints proba-bly produced the mezzotints themselves.During the eighteenth century portraits inmezzotint transfers were favoured, as well asthe Four Seasons, the Months, the Four Timesof Day, the Five Senses and the Elements,followed in the nineteenth century by shippingscenes.

Chinese snuff bottlesIn China during the nineteenth century, clearrock crystal or glass snuff bottles began to bedecorated by painting on the interior surface.The basis of the technique was similar to thatused on sheet glass, where the layers of painthad to be applied in reverse order, i.e. thedetails first, background last, the result beingviewed through the glass. However, the artistrequired considerable skill to create delicatedesigns such as portraits and landscapes, usingan angled pen or brush, inserted through thetiny mouth of the bottle.

Architectural glass

Historic literary sources and archaeologicalevidence show that at least from Roman times,glass was used in the form of small windowpanes and for the decoration of walls in theform of glazed bricks, opus sectile, and asmosaic tesserae on floors, walls and ceilings.The conservation treatment of mosaic floors –lifting techniques in particular – is wellcovered in conservation literature (Novis, 1975;ICCROM, 1983; see also Proceedings of theConferences of the ICOM Committee forConservation of Mosaics).

The use of glass as an architectural medium,particularly since the nineteenth century, isoutside the scope of this volume, since it isnormally an architectural or building concern.Even so, a glass conservator may be consultedwith regard to glass technology, deteriorationmechanisms and methods of recording.Decorative glass continues to be used to formwindows, wall coverings, skylights and roofs,and during the twentieth century, toughenedglass has been developed for use as load-

bearing walls, floors and staircases, e.g. theGlass Gallery, Victoria and Albert Museum,London, (McGrath et al., 1961).

Mosaics

The architectural process of embedding smallpieces of roughly squared glass fragments(tesserae) in floor cement to form pictures wasdeveloped from early Greek mosaicscomposed of coloured pebbles. The techniquewas extensively used in the Roman andByzantine periods. The Roman historian Plinyrecords (Forbes, 1966, ref. 232) that,

When pavimenta lost their place of honouron the ground, they took refuge in thevaulted ceilings, and glass was used for theirconstruction. Such an application of glass isquite recent, for when Agrippa, in the bathswhich he constructed at Rome, adorned theterra-cotta walls of the hot rooms withencaustic paintings, and finished the otherrooms in white, he would unquestionablyhave put glass mosaics in the ceilings if suchfashion had existed, or if the glass which wehave described as figuring in the walls ofScaurus’ theatre had found a place in theceilings.

There are also historical references to glasspillars (Forbes, 1966 quoting ancient sources,refs 230 and 231), which stood for instance inthe temple of Aradus.

Mosaics were composed of tesserae, madefrom coloured glass, mostly about onecentimetre square in cross section, which wereembedded in mortar, scored with a roughoutline of the design, while it remained soft(two to four days) (see Figure 3.39). The firstmosaics were formed of brightly coloured leadglass with a limited colour range of red andyellow or intermediate hues. Later, the glasstesserae were formed from glass pastes, smalti(cakes) and metal leaf. Glass pastes appear tohave been made in a limited colour range. Thebrightest of these were of dark, transparenthomogenous glass, whilst the lighter shadeswere produced either by reducing the amountof colourant in the glass, and/or by theaddition of a white opacifying agent. Specialhues could be obtained by adding colouredcrystalline materials to the molten glass, such


as brownish-orange fired clay, or a vitreousopaque yellow intermediate product(galliano).

In Roman times the technique of adheringleaves of silver and gold and their alloys tohot glass was discovered. The metal leaf wasapplied to cast glass slabs 5–10 mm thick, orto a thin sheet of blown glass (the cartellina),less than 1 mm thick, which would protect themetal against oxidation, and add to its brilliantappearance. The two glasses were then heatedin the furnace until they began to soften andwere then pressed together to ensure goodadhesion, and thus embedding the leaf. Thefinal appearance of the metal-leaf tesseraedepended upon the purity of the metal, thenature of the support and whether colour wasdeliberately added to the cartellina.

The use of the gold background in mosaicwork was particularly associated withByzantine architecture. The earlier mosaics atRome and Ravenna generally have back-grounds of a dark blue colour, which isparticularly fine at SS Cosma and Damiano inRome, and in the tomb of Galla Placidia inRavenna. Fiorentino-Roncuzzi has published astudy of the mosaics at Ravenna in the churchof San Vitale (dedicated AD 547) and theBasilica of Sant’ Apollinare in Classe (c. AD

500). The mosaics at San Sophia inConstantinople (sixth century) had goldbackgrounds, as did all the later examples inItaly from the ninth and tenth centuries only.The finest examples are at San Sophia, SanMarco in Venice, and the Cappella Palantinaat Palermo.

According to a twelfth-century treatise byTheophilus (Hawthorne and Smith, 1963) theByzantines,

made sheets of glass a finger thick (20 mm)and split them with a hot iron into tinypieces, and cover them on one side with goldleaf, coating them with clear ground glassbefore firing them in a kiln to fuse the goldleaf into place. Glass of this kind interspersedin mosaic work embellishes it very beauti-fully.

Smalti were produced in the Murano glass-works from the fifteenth century, allowing theproduction of a larger range of intenselycoloured glass with a greater surface brilliance,

and ease of cutting due to the high percent-ages of lead oxide present (producing a ‘soft’glass). Vasari (1511–1571) describes the prepa-ration of mosaic cubes. First the glass is madeopaque by the addition of tin oxide, andcoloured with metallic oxides. The moltenglass is ladled out in small quantities onto ametal plate and pressed into cakes 20 mm indiameter, and from 10 to 12 mm thick. Thecold glass was annealed and cracked intotesserae, the fractured surface being used toform the upper surface of the mosaic becauseit has a brighter surface. The thickness of theglass cake therefore determined the texture.

The production of leaf tesserae wasextremely skilled, and in times of less techni-cal ability, e.g. during the eighteenth century,it was replaced by cold gilding on glass. Theearly Christian mosaics in Ravenna (notably inthe mausoleum of Galla Placidia; the RavennaCathedral baptistery, the basilica of Sant’Apollinare Nuovo and the Church of SanVitale), cover immense wall surfaces andvaults. Within an architectural frameworkhighlighted by decorative design, the mainareas are covered by representations offigures. Some vaults are damasked with floralmotifs. The mosaics in the church of San Vitale(dedicated AD 547), and in the basilica of Sant’Apollinare in Classe (c. AD 500) have beenpublished by Fiorentino-Roncuzzi (1967). Inthe past ten years or so, several scientificreports concerning the study and analysis ofancient mosaic tesserae have been published(Freestone et al., 1990; Veritá, 2001). Inmodern times glass mosaics have been usedto cover the exterior surfaces of buildings, e.g.Liverpool Cathedral, but without the samesuccess.

Deterioration mechanisms in mosaicscomposed of glass tesserae are the leaching ofalkali (sodium) by water, followed by theformation of micro-cracks. These are propa-gated by the crystallization of sodium(calcium) carbonate in them. In pollutedenvironments, sodium, calcium carbonate istransformed into sulphate, which becomesacidified and hastens decay. In addition, watermay penetrate metal leaf tesserae if the bondbetween the two layers of glass enclosing theleaf has not been complete. Water alsopenetrates between the individual tessera andif there are structural faults in the building,


which they adorn, may penetrate behind themosaic causing their disruption.

Opus sectile

The site of Kenchreai, one of the two ports ofancient Corinth, was excavated between 1965and 1968, during which fragments andsections of one hundred panels of glass opussectile were recovered from the sea (seeFigures 7.7a–d). The panels are thought tohave been made in the Egyptian city ofAlexandria (or in Italy), and to have beenstored in a warehouse when Kenchreai wasdestroyed by an earthquake, circa 365 AD, andthe shoreline of the harbour was submerged.Each panel was composed of colouredfragments and moulded shapes of glass set ina rosin/marble plaster matrix to form architec-tural and pictorial compositions depictingbuildings, human figures and Nilotic plantsand birds. This was supported by a backingof thick ceramic tiles, which in addition to thecured resin, would have resulted in a heavy,sturdy construction. Because of the uneven-ness of this backing, the panels were packedin pairs, face to face for shipping. Immersionin salt water for just over 1600 years causedthe glass surfaces to become stuck to eachother; however, the natural disaster allowedthe recovery of more opus sectile than hasbeen found anywhere else in the ancientworld.

In 1995, as a result of renewed archaeolog-ical interest in the site of Kenchreai, several ofthe original excavators and conservationspecialists re-examined the panels andfragments. Koob, Brill and Thimme (1996) givea comparison and assessment of previoustreatments applied to the opus sectile panels.

As was the case with so many ancienttechniques, that of opus sectile was resurrectedby Victorian craftsmen in nineteenth century-Britain. Coloured glass blown for windowswas occasionally used for mosaics, and led tothe ‘invention’ at Powell’s WhitefriarsGlasshouse in London of a new opaque mater-ial known as opus sectile and described at thetime as ‘standing halfway between tile paint-ing and stained glass’. The technique wasemployed as a means of reducing waste: ithad been discovered that fragments of flintglass contaminated with clay, and therefore

normally scrapped, could be ground to a finepowder (fritted), and baked to produce asolid, durable material with an eggshell surfaceand an almost unlimited range of colours.Large panels of opus sectile could be madefrom the frit, which also formed the basis ofPowell’s production of glass tiles. These weremade by packing decorative iron moulds withpowdered glass which was then fused byheating. The technique was first experimentedwith in the 1860s, and by the 1880s it beganto be used widely to create church memorialsand decorative wall effects. Many decorativeschemes of opus sectile can be seen inchurches throughout Britain. Both opus sectileand glass tiles continued to be made until theSecond World War.

Figure 3.41 shows an example of late-nineteenth-century opus sectile work in a highVictorian public house, The Crown Bar(Belfast, Northern Ireland). Glass opus sectilepanels decorate the spandrels above the bararcade. The panels are in the design of bowlsof fruit with stylized foliage formed in blue,green, deep red, purple and clear glass,attached to rough wooden backboards withputty. The putty was coloured amber, yellowor green depending upon which colour glasswas to be laid on it. Some of the glass wasbacked with gold or silver coloured foil(which may actually be aluminium). The glasspieces were roughly cut and the edges notsmoothed off. Putty either oozed up betweenthe pieces or was deliberately applied,coloured black to resemble lead cames. Someof the blackening undoubtedly results from thesmoky bar-atmosphere. It is assumed that aBelfast firm was brought in to provide all theglass ornamentation for the Crown Bar; but thesource of their manufacture is unknown.

Historical development of flat glass

The methods by which glass was produced inflat, sheet form have changed considerablyover the ages. Small pieces of flat glass couldbe produced by casting glass into openmoulds or by cutting squares of glass formedby the cylinder or crown methods (see Figure3.43). Syrian glassmakers were the firstproducers of blown glass crowns (specula) inthe first century. Small windows of greenishglass, cast into wooden moulds, and bearing


their imprint, are known from Roman times.Flat sections of glass cut from medieval crownand cylinder glass were termed panes. Squareor diamond-shaped panes used (particularly ingrisaille windows are called quarries (from theFrench, carré, square).

Up until the seventeenth century, broad andplate glass were produced by the cylindermethod. The technique of producing plateglass by casting glass onto a flat surface wasinvented in France. Bernard Perrot of Orleansand Louis Lucas de Nehou, an official at theTourlaville glassworks at Cherbourg whereplate glass was already being made, bothclaim its invention. In December 1688, LettersPatent granted the monopoly of glassmanufacture by the casting process for theFrench home market, and, later also forexport, to several Frenchmen acting throughAbraham Thevart. Following a chequeredcareer, the factory was reconstructed in 1702,after which manufacture was concentrated atworks in Tourlaville and at St-Gobain inPicardy. By 1725, the annual output hadprobably reached about 700 tons, by 1750about 850 tons, and after 1760 more than 1150tons.

MirrorsThe first mirrors were probably reflectivesurfaces made of obsidian, polished stone orrock crystal, and in the classical world andChina, of pieces of dark glass or metal with ahigh tin content (speculum). The first mentionof a genuine glass mirror with a reflectivebacking of tin is by Aphrodisias (c. 220 AD,Forbes, 1966). Looking glasses or mirrorsformed of flat glass coated on the reverse withan amalgam (tin–mercury) came into generaluse in the late Middle Ages, once suitable flatglass could be produced. This type of mirrorwas prevalent from the sixteenth until thetwentieth century, and is therefore thecommonest type found in museum collectionsand historic buildings.

The first centre for the production ofamalgam mirrors was Murano in Venice,where in 1507 brothers Andrea and Domenicodel Gallo obtained a twenty-year privilege(patent) from the Ten Men’s Court in Veniceto make mirror glass according to a newmethod. Since the glass was made by thecylinder process, it was difficult to make

mirrors larger then a cylinder of one metrelength. In 1687, a process of producing largesheets of plate glass by casting it onto an irontable and then rolling it flat with a metal cylin-der was invented in France. The centre of themirror industry moved from Italy to Franceand from there gradually spread to othercountries.

Plain glass windowsThe use of glass in an architectural contextwas relatively slow to develop. It was one ofseveral more or less translucent materials,including alabaster, mica and shell, used bythe Romans to form small windows. Thematerials were set into the masonry, or intodecorative frames of plaster, wood or bronze.There is archaeological and literary evidencefor the use of window glass throughout theRoman and Muslim worlds. Glazed windowswere in use by the first century AD, particu-larly in the northern part of the RomanEmpire. In the United Kingdom, fragments ofRoman window glass have been excavated atHartfield, East Sussex (Money, 1976), Stonea,Cambridgeshire and Caerleon in Wales (Boon,1966). The gradual replacement of wood orplaster lattices by malleable lead resulted in aflexible and more versatile construction. Whenand where the use of lead began is notknown, but it is possible that its use forwindow framing may have been suggested bythe use of thin metal strips to separatecoloured areas of enamel in the cloisonnétechnique.

Decorative windows appeared in Christianchurches at a very early date. The poetPrudentius (348 to c. 410), in describing thewidespread use of glass in Constantinople,wrote, ‘In the round arches of the windows inthe basilica shone glass in colours withoutnumber’. The windows may have been anabstract coloured mosaic as no subject matteris mentioned. Although there is much literaryevidence for them, no complete windowssurvive from the fifth and sixth centuries.

During the Middle Ages glass was made inthe forest glasshouses of France, especially inNormandy and Lorraine, the Low Countries,parts of Germany and other parts of Europe.The output would have included windowglass, with a greenish caste, made by thecylinder and crown glass processes. Although


the cutting of crown glass was arduous andcomplicated by the bull’s eye, it yielded asmooth and highly polished surface. Althoughcostly to produce, crown glass was highlyfavoured. Glass discs spun to a diameter of122–183 cm (4–6 ft) and after cooling andannealing, was cut into small panes of glass,rarely more than 30�20 cm (12�8 in), and setin parallel rows. The central bull’s eye formedwhen the pontil rod was cracked off wasincorporated, the gaps between the roundelswere filled with triangular pieces of cut glass.The small panes necessitated the use ofglazing bars in the larger window spaces ofthe eighteenth and early nineteenth centuries.

The glassmaking families of Normandy andLorraine were the finest manufacturers ofwindow glass and their products were widelyexported. In the late fourteenth century thecylinder process was refined in Lorraine byincreasing the size of the cylinders to obtainlarge and even sheets (sometimes referred toas the Lorraine Method). Numerous paneswere cut from the sheets and arrangedgeometrically (often into lozenge patterns) andheld in place by lead strips. The windowswere set into wooden frames ready to beinstalled. They were highly valued and movedwith the owner when a house was sold.

With the aid of glassworkers fromNormandy and Lorraine, Jean Carré establishedtwo furnaces in the Weald area of southernEngland, to manufacture window glass. As aresult, cheaper home-produced windowscould be afforded by others than the wealthy,and began to be used in carriages and shipsin addition to buildings. Initially the glass wasundoubtedly of inferior quality to thatproduced in France, however by circa1580–90 its quality had substantially improved.

The French religious wars of 1560, and inEngland, the technical changes in glassmakingbrought about by James I’s proclamation of1615, banning the use of wood as fuel in glassfurnaces, had major repercussions for the wayin which glass was to be developed and used.During the seventeenth to the nineteenthcenturies, window tax and duty in England,through the way duty was collected, encour-aged one method of production over another,and stifled creative development, giving Europethe lead in glass production techniques.

During the late seventeenth century wood

largely replaced the use of lead as supportsfor tall windows, as it was both lighter inweight and less expensive. Timber mullionsand transoms supported leaded lightcasements. Sash windows evolved during theneo-classical movement, with the need to letmore light into buildings through the installa-tion of larger panes of glass. In England woodwas used in the manufacture of sash windows,introduced at Windsor Castle in 1686. Afterthis date, wood framed windows becameincreasingly common in Europe, and aftercirca 1740, in North America. In GeorgianEngland, the box sash window was developedwith astragal divisions; and it became fashion-able to insert glass fanlights above the mainentrance doors. The fanlight above the mainentrance door of Marble Hill House,Twickenham (built 1724–7) is amongst the firstto be documented. The majority of fanlightsoriginate between 1780 and 1830. Togetherwith iron railings and balconies, they form themost conspicuous forms of house decorationon the late Georgian house.

The availability of sheet glass enabled largespectacular buildings to be erected, built ofglass supported in a metal framework, such asthe Palm House, Kew Gardens (usingimproved German sheet glass produced byChance), and the Crystal Palace, both inLondon. The Palm House was built between1844 and 1848, to house new plant discover-ies coming into the country from expeditionsaround the world. Between 1985 and 1989, thebuilding underwent major restoration work,which reduced it to its bare bones beforebeing rebuilt. During the course of work,16,000 panes of glass and 10 miles of glazingbars were used. The Crystal Palace was builtto house the Great (Trade) Exhibition of 1851,and was later destroyed by fire.

Painted and stained glass windowsIt is not known where the technique of stain-ing on window glass originated. Staining onvessel glass was practised in Egypt in the sixthand seventh centuries. Archaeologicalevidence is pushing back the date of the earli-est known enamelled glass. Fragments ofcrown glass bearing the painted outline ofChrist in Benediction from San Vitale inRavenna (Italy) have been dated to circa 540AD. Silver stained (and enamel painted) glass


windows flourished as a major Christian artform in Gothic Europe (mid twelfth to thesixteenth centuries), being originally made forecclesiastical buildings (see Figure 3.45). Laterthey were also made for public buildings andprivate dwellings. The technique originated inmedieval France at the time of the building ofthe great churches and cathedrals, and mayhave been inspired by other art forms employ-ing similar materials and techniques: thesetting of coloured glass and gemstones injewellery, where they were held in place bymetal claws over holes cut in the metalwork,the manufacture of cloisonné enamels and theuse of glittering glass mosaics to cover wallsand domes in early Christian churches such asthose in Ravenna. In the Islamic world,coloured glass was set into a pierced marbleframework to form decorative windows, e.g.Topkapi Palace, Istanbul.

The oldest surviving stained glass windowsare thought to be the Prophet Windows inAugsburg Cathedral, circa 1065, and theAscension window at Le Mans, circa 1145.These examples are so developed that theremust have been precursors which have eitherperished or been deliberately destroyed.

Windows were produced in great quantityin France in the fourteenth century, during theperiod of the building of the great churchesand cathedrals. The names of the medievalcraftsmen remain largely unknown, but mayoccasionally be found in church records, e.g.John Thornton of Coventry, under whosedirection the great east window of YorkMinster was glazed between 1405 and 1408.This window measures circa 23.4 � 9.8 m, andwas badly damaged by fire in the twentiethcentury. The most renowned survivingwindows are in the French cathedrals atChartres and Le Mans, and in the Sainte-Chapelle, Paris.

Glassmaking in the medieval period wasclearly dominated by the great demand forwindow glass for the cathedrals and churchesbeing erected at the beginning of the Gothicera. Barrelet states that for the cathedral ofChartres alone, at least 2000 m of glass wasrequired over a period of 30 years; this wouldcorrespond to about 8 m of glass (c. 20tonnes). Large quantities of fuel might havebeen difficult to obtain at the best of times; itwould not have been surprising therefore, if

the need for so much fuel caused a shortageat coastal glassmaking sites. Migration of theglassworkers to well-wooded sites may havebeen a natural consequence, especially northof the Alps where most of the new cathedralswere being constructed.

A succession of French glassmakersemigrated to Britain to carry on the glass trade;there are no records of native Britons manufac-turing glass at the time. Thus from the earliesttimes French glassworkers acquired a wide-spread reputation, in particular for the produc-tion of window glass. Documents dating fromthe eighth century include requests for theirassistance in supplying glazing for churchesand monasteries in Britain. The trade wouldhave further increased after the Normanconquest of Britain in 1066, when churchbuilding began to be carried out on a largescale. The first recorded French glassmakerworking in Britain was Lawrence Vitrearius ofNormandy, who obtained a grant of land atDyers Cross in Sussex in AD 1220. In 1240 hewas commissioned by Henry III to make plainand coloured glass for Westminster Abbey inLondon. Succeeding generations of theVitrearius family were recorded until 1301.Then, in circa 1343, a John Schurterre settledat Chiddingfold in Sussex and became animportant glassmaking figure. The Chidding-fold area of Sussex (in the Weald), remainedthe centre of British glassmaking for manyyears, having as it did an abundant supply ofwood-fuel, and beechwood and bracken whichcould be burned to produce ash (alkali). Asixteenth century Venetian map of Britainshows two glassmaking sites, that atChiddingfold, and another at Guildford inSurrey.

Medieval painted and stained windows weremade of small pieces of glass, but the size ofthe individual pieces has increased over thecenturies. The finest glass windows are thoseof the thirteenth and fourteenth centuriesmade with coloured glass; those of thesixteenth century and later were made bypainting with coloured enamels, mostly oncolourless glass panels.

Following the increase in realism found infifteenth-century painted glass, sixteenth-century glass painters increasingly tried toachieve the effects of easel painting. By theseventeenth and eighteenth centuries, the


leading, which had played such an importantpart of the design of medieval windows, wasreduced to holding the large squares or rectan-gles of clear glass together. The entire glasssurface was then painted like a canvas. By thelate sixteenth century, translucent enamelswere available in a wide range of colours.Their total adoption was perhaps acceleratedby the increasing difficulty of obtainingcoloured glass, as wars and political unrest inEurope gradually ruined or made glassmakingareas inaccessible.

The decline of the art during the second halfof the sixteenth century can be attributed to anumber of factors. Internal causes were techni-cal innovation, changes in fashion and the useof new materials; external causes were thereformation, religious conflict and neglect. Therise of classicism hastened a dislike for the de-corative narrative of the Gothic stained glasswindow in favour of smaller, simplerwindows. During the sixteenth centuryBiercheiben (celebration windows) broughtthe ecclesiastical art of stained glass toordinary homes. To celebrate the building ofa new home or family occasion, a local artistwould be commissioned to paint a suitablepicture or coat of arms in transparent colours,on a small pane of glass which would then beincorporated into a window. The panelsbecame enormously popular and wereproduced by tradesmen and glaziers with aflair for art, rather than by artists. Popularthemes were biblical scenes and rural activi-ties. Bierscheiben were exported abroad. Smallleaded stained glass roundels, and panels,bearing designs executed in fired enamelpigments and – occasionally – yellow silverstain from which they derive their name, werealso made for secular use.

In Venice, during the seventeenth andeighteenth centuries, small glass plaquesdepicting picturesque city views derived fromcontemporary prints and paintings wereexecuted in cold painting and fired enamellingtechniques. In Bohemia, Germany and Spain,engraved clear glass plaques depictedcommemorative scenes: e.g., a panel engravedby Felix Ramos in the late eighteenth century,showing the facade of the Royal Palace of SanIldefonso (Museo Arqueologico Nacional,Madrid). In the nineteenth century hochschnittplaques were produced by the German glass-

maker Wilhelm von Eiff (1890–1943) depictingfigural subjects in a contemporary manner.

Stained glass window production wasrevived in the nineteenth century whenmedieval methods of manufacture were inves-tigated and practised. One of the greatestBritish designers was Sir Edward Burne-Jones(1833–98), whose designs were executed byWilliam Morris (1834–96), e.g. those inBirmingham Cathedral; by the American JohnLa Farge (1835–1910) and others. In thetwentieth century windows were designed by,among others, the artist Marc Chagall(1889–1985).

Traditional windows and panels continue tobe made and restored in numerous studios. Inaddition new art forms have been developed,such as glass-in-concrete, appliqué glass andfused glass windows, usually but not exclu-sively in modern churches and cathedrals suchas those at Coventry, Warwickshire (UK) andRouen (France). Lee et al. (1982) published anillustrated comprehensive guide to the world’sbest painted glass windows.

The term stained glass used to describethese decorative, coloured glass windows, isnot strictly correct, since the majority containlittle silver staining, the technique only havingbeen discovered in the fourteenth century. Atechnically more accurate term would bepainted glass windows, since the overallcolour derives from colourants introduced atthe glass-melting stage, i.e. to the batch, orfrom the use of flashed glass, i.e. clear glasswhich has been covered with a thin layer ofcolour, e.g. ruby red; and from which thepieces of glass to form the design are subse-quently cut. Onto this, details were applied inenamel colours, which were fired for perma-nency. Silver staining became popular duringthe fourteenth and fifteenth centuries, but laterwindows were often enamel painted withsmall amounts of staining on features such ascrowns, rings and canopies.

In the process of creating a traditionalstained glass window, the design is conceived,sketched and then a full-size drawing(cartoon) is produced. From this, the cut-line,a tracing of the lead-lines is taken. The centreof each lead-line is carefully drawn in orderthat the coloured glass pieces placed over thetracing can be accurately cut to shape. Minoradjustments may have to be made to the glass


edges. In medieval times this process, termedgrozing, was carried out using a metal toolwith a hooked end (the grozing iron), whichleft the glass with a characteristic nibblededge. From the eighteenth century, a sharpmetal tool or diamond wheel was used, whichcreated a smooth straight edge. Decorativedetail is applied in the form of enamel coloursand/or yellow (or silver) stain, and renderedpermanent by fusing them to the glass atrelatively low temperatures in a muffle kiln.

The technique of yellow staining developedin the early fourteenth century. It colours clearglass yellow (or more rarely blue glass green)by the application of a solution of a silvercompound to the exterior surface and firing it.The earliest firmly dated example of theyellow stain technique is a window from theparish church of Le Mesnil-Villeman, Manche(France). Yellow stain ranges from pale lemonto orange and is produced by applying asolution of silver compound to the surface ofthe glass, which when fired, turns yellow.Details such as facial features can be definedwith a brownish-coloured enamel made withiron oxide termed grisaille (Fr. grissailler, topaint grey). Grisaille was also used to creategeometric patterns of regular design or offoliate design, leaded into or painted on tintedglass. Smear shading was a method of shadingdetails such as drapery and facial features byapplying pigment sparingly with parallel brushstrokes. The technique was prevalent until thefourteenth century. Stippling creates the effectof minute points of light all over the glass.

The glass pieces are then leaded-up in aframework of lead cames to form panels. Aseries of panels is combined to form largewindows. Individual small panels or largerwindows are cemented into the stonework ofa window aperture and supported at intervalsby being tied-in to horizontal metal glazingbars (ferramenta).

Another type of window glass was made bythe Norman slab (or bottle glass) method. Agather of molten glass was blown into asquare mould. The resulting square block wasthen cut so that each side became a small flatpiece of glass. (A modern development is theuse of slab glass blocks (Fr. dalle de verre),cast glass pieces usually about 2.5 cm thickand 30 cm long, which are set in concrete orin epoxy resin, as opposed to being leaded-

up. Its use creates windows with a monumen-tal appearance dominated by the dark frame-work; which can be load-bearing.) Traditionalstained glass windows continue to be made byindividuals and by small workshops. Theirmanufacture is not a process that can beindustrialized.

The historical care of painted and stainedglass windows has, as is the case with allwindows, been linked with the routine mainte-nance of the buildings in which they wereinstalled. During the windows’ lifetime,damaged glass and lead cames have beenremoved and replaced by successive genera-tions. Only a fraction of the stained glassproduced during the Middle Ages has survivedto the present day. Few survive intact andwithout considerable change in their originalappearance, resulting from decay, paint lossand restoration. In some countries such as TheNetherlands and Norway, the losses have beencatastrophic. In Britain, during the seventeenthand eighteenth centuries, the Calvinists, funda-mentalists and Cromwell’s army broke as manychurch windows as they could, since pictorialart was then regarded as idolatrous. In Francethe iconoclasm associated with the FrenchRevolution was also responsible for the loss ofmuch painted window glass. A considerableamount of French glass was bought by patronsof the arts in Britain, and much of it wasinstalled in British churches.

In Britain, France and Germany much moremedieval glass survives, but windows haveoften been removed from their originalsettings, and some reduced to fragments.Examples of glass from the poorest to thefinest buildings have found their way intoauction rooms, and been sold into public andprivate collections. It is these smaller pieces ofglass which an objects conservator might berequired to treat.

Cloisonné glassDuring the last years of the nineteenthcentury, a London company attributed to itselfthe invention of a decorative form of glasspanel, after which it was named: TheCloisonné Glass Co. (at 40 Berners Street,Oxford Street, London W.) However, it isknown that the company of Barthels & Pfisterwere also making cloisonné artistic windows


in London at that time. The main feature ofcloisonné glass panels was the combination ofboth transmitted and reflected light effects (seeFigure 7.68).

Decorative forms do not develop in isola-tion, and as its name suggests, cloisonnéstained glass derived from the techniques ofstained glass and cloisonné enamel manufac-ture (and perfectly described the technique ofcontaining coloured, tiny glass chips andglobules within a framework of metal strips,between two sheets of glass). Initially, theglass was only produced in London, but laterunder English patent, it was produced in otherEuropean countries, notably Spain, and in theUnited States of America.

There is little information regarding thetechnique, but the following information wastaken from publicity material issued by TheCloisonné Glass Company. The panels weremade up, by first outlining the design withthin gilt or silvered metal wire secured withtranslucent cement (adhesive) to a sheet ofclear glass (back-plate). The cells thus formedwere filled with pot metal coloured glass ingranular form (< 1 mm), either globules orsquares, and secured in place with adhesive.The Cloisonné Glass Co. boasted a stock ofraw glass comprising over eight hundreddifferent shades and tints. A second sheet ofclear glass was then placed over the designand sealed by binding the edges of the glass‘sandwich’ with tin foil. According to theCompany, the top sheet of glass was notabsolutely necessary but served to give thecloisonné layer a smooth surface, so that itcould easily be washed and kept clean. It also,of course, afforded protection to the work.Panels thus formed could be fixed in placewith putty by any glazier, or by the use ofwooden beading for internal decorations. Theywere used as windows, fanlights, door panels,lanterns and ceiling lights, window blinds,screens (including fire screens), and partitions.They were used for internal and externaldecorations in houses, public buildings andalso in piers, yachts, liners and railwaycarriages, for decorative fascias, signs – illumi-nated or opaque, for the decoration of furni-ture such as cabinets, cupboards, book-cases,screens, writing desks, over-mantels andpianos etc. The panels could in fact be usedas an alternative in any situation where stained

or leaded glass, mosaic, fresco or temperawould be appropriate.

It was claimed that, as in the case of stainedglass panels, cloisonné stained glass wassuitable and preferable for decorations whichwere seen at night, when no light is transmit-ted through it, or where a good opaque effectwas desired in addition to the transparency. Itwas further claimed that the double sheet ofglass excluded any draughts and was absolutelywater-tight. For indoor work, cloisonné stainedglass could be supplied without the top sheet.The work was laid on a slate base and wassaid to be perfectly durable. The Company’sdescription of this type of work states that thebase could be gilt, thus obtaining a greaterbrilliancy and richness of colour as a result ofthe gilt layer reflecting light through thecloisonné layer. The panels were said to be lessliable to break than ordinary window panes.The best quality window glass was used in theirmanufacture, in 21, 26, 32 oz weights, or plateglass, according to the required size or situa-tion. The process of manufacture, being a coldone, the glass sheets were not subject to anystrain. Panels could be made on glass varyingin thickness between 1⁄4 in and 3⁄8 in and also ofcurved glass or of plate glass between 1⁄2 in and5⁄8 in. It was recommended that the panels wereset in a 1⁄4 in rebate to hide the tin foil edging,or that the tin-foil be painted a ‘suitable’ colour.As is the case with stained glass windows,glazing bars could be introduced to supportvery large panels. In the case of the front glassplate being broken, this could ‘easily bereplaced by any glazier’. Should the back sheetor both sheets become broken, the glass couldeasily be repaired at the works in a few daysat small cost. If the back plate, which wasalways made stronger than the front, was onlyslightly damaged, another sheet of glass couldbe put behind the original to protect the work.This second sheet had to be well embedded inputty to prevent any dust entering the spacebetween the two glass sheets.

Historical development of light fittings

Lanterns and lampshadesThe first reference to glass lamps dates fromthe fourth century AD. Standing and hanginglamps were used for illuminating churches andother buildings. Contemporary mosaics depict


various types of lamps in use, which seem tohave been made of glass. The vessels are ofvarious shapes; some were used as true lampsburning a wick, and some as candlesticks,some as single lights and some as elements ina polycandelabra. Simple glass bowls or dishessuspended from metal chains and containingoil and a wick to provide light, of a typeknown since antiquity, are still in use inmosques. On some much later examples ofglass dish lamps, an urn-shaped reservoir wassuspended centred over the dish, while onothers the fixtures were equipped with threeor four reservoir burners attached to andstanding up from the metal rim.

So many of these devices have beenconverted to electricity that original metalworkdoes not often survive. Glass dish lightssuspended from the ceiling by chains becamepopular in nineteenth century Europe, as theyoffered a practical alternative to candlelight. Awick soaked in colza oil was usually held ina central reservoir, and was gravity fed tosmall sumps beneath the burners which werecovered with decorated or opaque glassshades.

In the earliest hand-held lanterns, a candlewas protected by panels of thin animal horn(Eng. lanthorn). Metal-framed hanging lanternsbecame popular in the eighteenth century,when they were glazed, and when decorationof the frame and glass panels became increas-ingly ornate. The glass panes may be held inplace with linseed putty and/or by clips incor-porated in the metal framework. Checks onthe metal framework, the security of any doorhinges and on the fixing arrangements willneed to be made. In the case of sphericallanterns, each glass panel will be curved. Ifone or more of these requires replacing, acomplete panel can be removed and a plaster-of-Paris mould taken from the convex sideinto which a glassworker can heat and slumpa piece of glass of the correct thickness, toform a new panel. The edges of the new panelcan be chipped (grozed) in order to adjust itto the correct dimensions.

Following the introduction of electric light-ing (the first light bulb was produced in 1879),lampshades of numerous varieties began to bemade. Particularly well-known are the highlydecorative lampshades produced by LouisComfort Tiffany in the United States; and those

produced by Gallé and Daum in France.Tiffany lampshades are composed of colouredfragments of glass assembled to form designs,and held in a thin copper framework, theentire effect being similar to that of a stainedglass window.

Candlesticks and candelabraGlass holders to support one or more candles(candelabra) were produced in England,France, Spain, Germany and Italy (Venice)from the 1600s onwards, then in Ireland, andby 1760 in America. Many originally followedthe designs of their metal prototypes. From themid-seventeenth century onwards it wasfashionable to decorate the metal frameworkof light fittings with glass or rock crystalornaments, such as pendants and facettedbeads, which caught and reflected the light. InEngland, branched candle-holders for tableuse, with two or more arms for candle sockets,were commonly made of glass circa1760–1880. Originally advertised as girandoles,they were known as candelabra from circa1792. Early examples were plain, and againmany originally following the designs of theirmetal prototypes. After circa 1714, cutting onthe shaft reflected fashionable styles ofcontemporary wine glass stems; after circa1765 cutting became increasingly more lavish,and the candelabra might be decorated with astar or crescent finial at the top. After circa1775, many were decorated with pendantdrops (lustres), which continued to feature onnineteenth-century examples.

Wall light and sconcesWall lights and sconces were made in Venicefrom the early 1600s and shortly afterwards inEngland. Made of glass or glass combined withmetal giltwood or ceramic, they had one ormore branches for holding candles. One typewas designed with a light-reflecting back-plate,usually of polished metal or mirror glass (thelatter by the late seventeenth century), calleda sconce by circa 1712. Their design followedthe style of table candelabra of the period, andthey offered a practical solution to lightinglarge rooms. In Ireland, cut glass wasproduced in Belfast, Dublin, Waterford andCork, the Waterford factory opening in 1783.Distinctive wall lights unique to the Irish glass-making repertory combined an oval mirror-


glass back-plate framed by clear and blue glasssegments (and rarely, green), fitted with ametal hook on top, from which a miniaturechandelier was suspended.

ChandeliersThere are few publications concerned with the historical development of chandeliers(Mortimer, 2000). Determining age can becomplicated by the fact that from the end ofthe eighteenth century it was already fashion-able for chandeliers to be dismantled and re-formed to the latest fashions, includingchanges in the ornamentation; and that theymay have begun life as gasoliers and havebeen converted to electricity. Thus one cannotalways be certain how much alteration hastaken place. The huge nineteenth-centurymarket for lighting fixtures has furnished therestoration trade with a readily availablesupply of glass components, which can beused at random. Later pressed glass items mayhave been substituted for cut glass. It may bepossible to obtain information on style fromeighteenth-century trade cards of London glassdealers; or to obtain particulars of orders andinvoices for purchases and cleaning, if recordsexist with the owner, or in a local library,museum or County Record Office.

Another difficulty arises from the use ofterms to describe chandeliers (see below).Similarly, there appears to be no universallyrecognized terminology for describing thecomponents of chandeliers; Davison (1988)and Reilly and Mortimer (1998) haveattempted to rectify this. In the event of therebeing insufficient evidence to reconstruct achandelier, a particular time period in itshistory will have to be chosen, in consultationwith an historian of historic lighting. At sometime in its life, a period chandelier should bedismantled, examined and correctlydocumented. Any broken, original parts thathave to be replaced should be recorded,labelled and kept in a location where they willnot be discarded, lost or confused with eachother. This includes worn textile or passe-menterie suspension chain covers.

True glass chandeliers evolved in Englandand also on the Continent in Bohemia, France,Spain and Italy (Venice) early in theeighteenth century. In the twentieth century,chandeliers were produced in the USA, follow-

ing the Art Nouveau style. The general use of the term chandelier to describe a free-standing light is relatively modern, being de-rived from the French for tallow candles(chandelles). Early eighteenth-century adver-tisements list them as lustres or branches,names also interchangeable with candelabra,girandole and wall-light, so that the few liter-ature references to any of them can bemisleading.

The first chandeliers, used to light churchesand public buildings, were made of iron orbrass. Some wealthy houses were lit withsilver chandeliers, but many of these weremelted down when they became unfashion-able. The basic chandelier consists of a centralmetal support, disguised by glass globes orurns. The central support of each chandelieris normally formed over a metal pipe or rod,threaded at both ends. The metal suspensionrings, receiving plates, tubes and washers areusually of silvered brass or Sheffield plate.(Sheffield or plate-silver fused onto a copperbase-was a cheap substitute for solid silver. Itis thought to have been discovered by aSheffield cutler, Thomas Boulsover, in 1742.The process was used extensively from circa1760.) A glass bowl at the base conceals apierced metal disc from which numerous load-bearing glass arms (branches) radiate, eachwith a socket or grease-pan for a single candle(sometimes adapted for gas or electricityfittings during the nineteenth and twentiethcenturies).

The general component assembly remainedfairly constant, whilst incorporating structuralimprovements and changes in decoration. Thisstandard framework inspired by metal proto-types was adapted widely throughout Europeand resulted in the appearance of severaldistinctive and readily identifiable types. Glasschandeliers may be small and simple, or largeand complex in construction. In either casethey are by their nature fragile.

The design of the first true glass Englishchandeliers, made of lead crystal, was basedon early Flemish or Dutch brass chandeliers,and appeared early in the eighteenth century(the Georgian Period). They are firstmentioned in an advertisement dated 1727, inwhich John Gumley, a London glassmakerbased in Lambeth, offered ‘Looking Glasses,Coach Glasses and Glass Schandeliers’ for sale.


Georgian chandeliers are characterized bylarge baluster stem-pieces and scroll-shapedarms of glass. One of the earliest survivingexamples is a chandelier made for ThornhamHall in Suffolk in 1732, and now in Winterthur,Delaware (USA). The arms of Georgianchandeliers were solid and had flattened sideswith thumb cuts above and below. One endof each glass arm was formed into acombined, narrow-lipped greasepan and atube (socket or nozzle), which held a candle.There were no canopies or dressings. Theother end of the arms, were set in metal potswith plaster-of-Paris. The earliest chandeliershad square pots; from the mid eighteenthcentury round pots were made of cast brass.The earliest pots were engraved with thenumbers of the arms, cast brass pots werestamped with numbers and/or letters.Corresponding numbers appeared on thecircular cast brass plate (the receiving plate)cut with corresponding square or round holesinto which the arms were slotted. The platewas hidden from view by a glass bowl. Themetalwork was usually gilt finished, sometimeswith silver-plated tubes inside the glass stem-pieces to disguise the iron shaft. Very fewchandeliers survive with the combinedtube/drip-pan/arms, but short examples of thistype were often fitted to the aprons of pierglasses, with frames of gilt gesso, the combi-nation being known as a sconce. These all-in-one constructions were difficult to clean, sothat the drip-pans were soon made to beremovable.

Famous examples of chandeliers are thosemade in 1771 by William Parker of Fleet Street,London, and now hanging in the AssemblyRooms, Bath (UK). In the late 1730s and1740s, moulded chandeliers appear to havebeen popular. They frequently included somecutting, but were principally of reticulatedglass with rope twist arms. Soon makers cutthe arms to provide sparkle, and more elabo-rate surface treatments were devised, e.g. largeflat diamonds with cross cuts, the commonestdecoration between 1750 and 1770. As armsbecame more elaborate, curved canopies (orshades as they were then called) were addedand the whole chandelier was hung withapplied ornaments. By circa 1745, chandelierstyles became even more elaborate andshallow cutting extended to cover all the glass

elements – sockets, grease-pans and branches– in diamond and star-shaped patterns. Shortlyafterwards, cut glass pendant drops wereintroduced and applied to the framework –initially in small numbers, but by circa 1770in increasingly greater numbers.

Chandeliers followed rococo, then neo-classical styles, with extensive and sophisti-cated cutting. Those designed by the architectdesigners Robert and James Adam, featuredtwo tiers of arms with the upper tier support-ing triangular spikes or spires in place oflights. The solid arms were usually groundwith flat sides (hexagonal in section) with fourrows of thumb cuts. Other features were topand bottom canopies, elongated and eleganturn-shaped shafts, large pear-shaped pendantdrops, and smaller and pear-shaped dropswired into chains and hung between thebranches in swags, each pendant facetted inmeticulous detail. The Van Dyke edged drip-pans and candle nozzles were removable, thenozzles being attached to the arms with metalferrules. By 1800, decorative ornaments hadbecome so profuse that the underlying struc-ture of the chandelier – the shaft and branches– was practically obscured from view.

During the Regency Period in England (c.1810–1820), the country was at war withFrance, and harsh taxes were introduced toraise money. One of these, the Glass ExciseAct, exacted a heavy tax on all glass madeeach day at the glasshouses. To avoid the tax,chandelier manufacturers bought crystal dropscut by garret workers from broken pieces ofglass to resemble icicles. The drops werestrung together and hung in tiers from the topto bottom of decorative metal frames to formtent and bag, or waterfall designs. The shaftand the branches, being completely hidden,were usually undecorated. These styles ofchandelier became popular throughoutEurope.

By the mid 1800s Britain was prospering. In1832 the Glass Excise Act was repealed andchandelier manufacturers reverted to the glassarm designs. However, the chandeliersthemselves became bolder and weightier withdeep cutting and heavy swags of dressings.One of the most prolific manufacturers wasMessrs Perry & Co., of London, many ofwhose products survive. The chandeliers oftenfeature rope- or barley twist-shaped arms,


those on the lower tier being S-shaped. Theywere heavily dressed with swags nearly alwayscomposed of button-shaped lustres pointed oneither side, and with pear-shaped pendantdrops. Two of the most popular cuttingdesigns for the stem pieces were hollow andsplit and step. The metalwork was nearlyalways silver plated, and if a makers’ markwas applied, the wording Perry & Co. wouldbe stamped on the receiver plate or the topshackle.

The Birmingham Company of F & C Osler,established in 1807 by Thomas Osler and histwo sons Follett and Clarkson, came to promi-nence in the nineteenth century, creatingspectacular crystal chandeliers, fountains andfurniture, much of it for export to the Indiansub-continent. The firm was commissioned tomake a giant crystal fountain for the GreatTrade Exhibition of 1851, held in the CrystalPalace erected in Hyde Park, London.Containing 4000 kg (4 tonnes) of lead crystalglass, the fountain stood 8.25 m (27 ft) high.The Crystal Palace was later dismantled andre-erected in Sydenham, South London, thefountain taking pride of place in the centralnave. The Palace and fountain were destroyedby fire on the night of 30 November 1936.

As the popularity of chandeliers and decora-tive lighting began to decline, the factorycontinued, by cutting glass blanks supplied bylocal factories. The range of lighting becamesmaller and less ornate. In 1924 F & C Oslermerged with Faraday & Son Ltd (founded byRobert Faraday in 1814) and continued totrade until 1965. In 1985 the firm of Wilkinsonplc found and purchased the remnants of theOsler Company, including brass and bronzecasting patterns and old working drawings,which have made it possible to restore origi-nal Osler products with confidence.

Many early chandelier styles were repro-duced, for either candle or gas, during theVictorian period (1837–1901) with only slightvariations in the cut decoration. Thoseintended to be lit by gas had hollow armsthrough which gas was transported to theburners. The advent of electricity broughtabout several changes in the style of chande-liers, which now incorporated large areas ofmetal, electrical wiring and glass light bulbs,which were often aesthetically unpleasing.Mehlman (1982), Reilly and Mortimer (1998)

and Mortimer (2000) give accounts of thehistorical changes in chandelier design, andthe last two, a full glossary of terms associatedwith chandeliers.

Chandeliers have been and continue to bemade throughout the world; the greatestproducers being Germany and Bohemia, witha large export trade from circa 1720 to thepresent day. The Czechoslovakian glass indus-try began making Bohemian crystal glasschandeliers soon after the English manufactur-ers, whose designs were often copied. Atpresent the Czech Republic is the greatestproducer of glass chandeliers.

The Venetian style of chandelier haschanged little over the centuries. On the earli-est versions, glass arms, leaves and flowerswere simply pushed into holes bored orcarved into a wooden base. In later examples,glass flowers and leaves were added todecorate wooden or iron frames. Ironworkwas usually painted silver. The finest Venetianchandeliers were composed of both clear andcolourless glass, and were richly adorned withnaturalistic motifs such as flowers, leaves andfruit.

In the sixteenth century the more elaboratetypes of French chandelier were decoratedwith rock-crystal drops, and in the seven-teenth, with many-facetted pieces of crystalwhich reflected the light of the candles; theywere called chandeliers de crystal. Expensivecrystal was often replaced by cut glass, fromthe late seventeenth century, when the termlustre began to be applied (as it still is) to allchandeliers whether or not they incorporatepieces of crystal or glass. Cage chandeliersformed around a brass or bronze frame aretypical of the eighteenth- and nineteenth-century production of France. Early exampleswere dressed with rock crystal (quartz)pendants, and later examples with less expen-sive glass drops.

Scandinavian and Russian glass chandeliersare difficult to differentiate from one anotheras they all took a similar form. They generallyhave light, more decorative metalwork usuallywith a brass or gilt finish, and small slenderglass drops. Sometimes a blue glass dish orother coloured element was incorporated inthe design. Sommer-Larson (1999) describesthe conservation of three chandeliers, made atNostetangen for the church in Kongsberg,


Norway. The chandeliers have iron frame-works, and glass arms decorated with glassdrops and other ornamentation.

A number of flamboyant, painted glasschandeliers in the chinoiserie style hang in thebanqueting room and the music room atBrighton Pavilion in Sussex (UK). They aremore properly described as lanterns, beingcomposed of large panes of glass held inmetal ribs and similar in appearance to largeinverted umbrellas. The banqueting roomchandelier was created by the Regencydesigner Robert Jones and made by theLondon firm of Bailey and Saunders in 1817.The lamp itself, consisting of cut glass lustres,was created by Messrs Perry & Co. It issuspended by chains of gold radiating from

the centre of a large and brilliant star, a beltof rubies, pearls and garnets encircling itsbase. The body represents a fountain in fullplay, the cut crystals forming the water beingextremely realistic. Six winged dragons, eachholding a beautifully shaped water lily in itsupturned mouth, crouch around the base.Crystals in the form of festoons, shields andtassels are suspended from the dragons’ claws.The entire chandelier is 30 ft high and weighsnearly a ton. It has undergone many changesduring its lifetime. In 1999 it was surveyed andgiven a light cleaning by conservators (Oliver,1999). The chandeliers in the music roomwere badly damaged by fire, and water fromhoses, during an arson attack in 1975, andwere subsequently restored (Rogers, 1980).


Raw materials

The materials for making ancient glasses werenaturally occurring rocks and minerals: amixture of silica, alkali and lime, which alsocontained trace elements. The silica wasgenerally obtained from sea sand in whichthere were ground seashells, the source oflime. The alkali was natron, obtained fromdried lake deposits or from ash produced byburning saltmarsh plants (containing a highamount of soda). From the ninth century AD

ash produced by burning forest plants (havinga high potash content) was also used as asource of alkali for glassmaking. In order toproduce a durable glass, i.e. glass that wouldnot be easily soluble by water, the raw ingre-dients had to be chemically balanced. Inancient times this would have been achievedby trial and error, and by the empirical knowl-edge thus gained. The addition of too muchalkali to the glass batch had the effect ofincreasing the solubility of the resulting glassin water. However, other oxides such ascalcium oxide (lime) acted as stabilizers,thereby to a considerable extent counterbal-ancing the deleterious effect, and enablingreasonably durable glasses to be made.

To produce glass, the raw materials had tobe heated to about 1000°C, at which temper-ature they slowly melted to form a liquid andreacted together. Since it was impossible forthe first glassmakers to reach and maintainhigh temperatures long enough to melt glassin any quantity, the melting temperatureneeded to be considerably lowered to around900°C. This was achieved by the addition ofalkali, which acted as a network modifier and

flux. The amount of network modifier requiredto lower the temperature depended on thestate of development of furnace technology inany particular era. Ancient glass was uninten-tionally coloured by the presence of impuri-ties such as iron in the raw materials, whichresulted in the glass being yellow or green. Asglassmaking technology evolved, a secondgroup of metallic oxides, such as those ofcopper, iron, cobalt and manganese, weredeliberately used to colour glass. The deliber-ate production of colours in ancient glasseswas complicated since the metal oxidesproduced different colours depending uponthe firing conditions, i.e. the oxidizing orreducing atmospheres in the furnace. Mostearly glasses were translucent or opaque as aresult of the low fusion temperatures, whichallowed microscopic air bubbles to remain inthe viscous glass as it cooled.

In the production of glass one other ingre-dient was commonly added to the basic ingre-dients, that is, scrap glass (cullet), which wasnormally of about the same composition as thebatch to be melted. The function of the culletwas purely physical, it acted as a nucleusaround which the new glass formed, andhelped to eliminate unevenness such as cordsand striae in the new batch. The use of scrapglass in this way probably accounts for thelack of waste glass products on many glass-making sites.

Apart from the raw materials, the process ofmaking glass required a source of fuel (origi-nally wood, from the seventeenth century,coal); a furnace; and crucibles (or pots) madefrom refractory (heat-resisting) materials; inpractice, clay free from fluxes. The glass was

73

3

Technology of glass productionPart 1: Methods and materials

fashioned into artefacts and then allowed tocool, whereupon it congealed into a solid. Thecooling had to proceed slowly in an oven(later called an annealing oven, a lehr or leer),to relieve stresses set up during the manufac-ture, otherwise the products would shatter.

It is known that in ancient times, until atleast the ninth century AD, centres of glass-making evolved, which distributed chunk glassto other glassworking centres, which did nothave the technology to produce glass from itsraw materials.

Silica (the network former)

Several inorganic oxides have the ability toform vitreous materials. Of these silicondioxide (silica) is by far the most important(although boric oxide, sodium borate, was alsoinfrequently (accidentally) used in antiquity).

The main source of silica for ancient glassproduction was sand. In the Near and MiddleEast sands relatively pure and free from ironwere widely obtainable and therefore thecommon source (for example, Egyptian sandscontain only 1–3 per cent iron). Two othersources of sand exploited for ancient glass-making were those at the mouth of the RiverBelus in Phoenicia (now the River Na’aman inIsrael), which contained 87 wt per cent of lime(Turner, 1956c; Engle, 1973a, 1974), and thosefrom the River Volturnus (the modern RiverVolturno), north of Naples in Italy. In theMiddle Ages, when glassmaking spread toEurope, it was found that the sands in manyareas were too impure to be of use in glass-making without prior treatment, consisting ofwashing and then burning or being brought tored heat to remove organic matter. In Italy, forexample, suitable sand could be found at themouth of the River Volturnus; but that at themouth of the River Tiber, and on the Italiancoast above Ostia, contained large amounts ofvolcanic debris. Another source of silica wascrushed flint or quartz pebbles obtained fromriver-beds. Pliny (77 AD) mentions the follow-ing ingredients of the glassmaker: soda(nitrum), limestone (magnes lapis), shinypebbles (calculi), shells (conchae), andexcavated sands and sandstones (fossilesharenae). In 1645 the English diarist JohnEvelyn was in Venice, and referred to whitesand, ground flints from Pavia, and the ashes

of seaweed brought from Syria as the rawmaterials used for glassmaking (Forbes, 1966).In 1979, Brill filmed glassmaking in the ancienttradition in Herat, Afghanistan. The rawmaterials were quartz pebbles, plant ash andscrap copper. White quartz pebbles werecollected from a dry river-bed, one donkeyload providing enough silica for one week’sglass production. Plant ash (ishgar) wasprepared by nomads, plant twigs beingheaped together in a shallow pit and burned,after which the ashes were left to coolovernight. To test its quality the glassmakerstasted the ishgar, choosing that which had asweet taste. Before being placed in the furnaceto melt, the quartz pebbles were broken,ground and mixed with the plant ash. Tocolour the glass bright turquoise, a lump ofscrap copper was heated and the surfaceoxide scraped off and added to the batch. Thefurnace was fuelled with hardwood broughtfrom the hills by donkey every two weeks; thiswas the most expensive commodity. A sawn-off rifle barrel was used as a blowpipe(Johnston, 1975).

Alkalis (the network modifiers)

The principal network modifiers of whichancient glasses were composed were theoxides of sodium and potassium. Until aboutAD 1000 the oxide of sodium (natrium –natrum in ancient times) was used universallyand introduced to the glass batch either assodium carbonate in the form of soda crystals(Na2CO3) or as sodium nitrate (NaNO3).Sodium oxide could also be produced byburning saltmarsh plants, usually Salicorniakali, or seaweed. Commoner sources of alkaliincluded natural deposits resulting from dryingand evaporation of land-locked seas and lakes;arid salts obtained by deliberate evaporationof sea or river water in pans or pits.Undoubtedly natron, a natural sodiumsesquicarbonate (Na2CO3.NaHCO3.2H2O) fromthe Wadi Natrun, north-west of Cairo, wouldhave been used in the glassmaking activitiesthere. The composition of natron is complexand variable: the sodium carbonate contentvaries from 22.4 to 75.0 per cent, sodium bi-carbonate from 5.0 to 32.4 per cent, sodiumchloride from 2.2 to 26.8 per cent, sodium sul-


phate from 2.3 to 29.9 per cent plus water andinsoluble material. The use of sodium oxideas a modifier resulted in the production ofsoda glass, which remained plastic over a widetemperature range, thus lending itself to elabo-rate manipulative techniques (Toninato, 1984).

During the height of the Roman Empire,glassmaking had, at least by AD 50, spread fromSyria and Egypt to western areas, and thennorthwards so that, by about AD 100, glass-makers were operating in the Rhineland. Theglasses made in Europe at the time containedlow amounts of magnesia and potash, andsince there do not seem to have been anysources of soda with the same characteristicsavailable in Europe, it seems reasonable toassume that the glasses were made usingnatron rather than the ashes of maritime plants(which contain significant amounts of magne-sia and potash). The deposits of sodiumsulphate in Germany, even if they had beenknown at the time, would probably have beentoo difficult to exploit, given the primitive stateof glassmaking technology (Turner, 1956c). Itis possible that natron continued to be trans-ported to the Rhineland even after the collapseof the Roman Empire in the west. Althoughsuch a practice might seem to have beenuneconomical, it must be remembered that theearly glassmakers were extremely conservativeabout the use of tried and tested raw materi-als.

A dramatic change in glassmaking practicebegan to occur around AD 1000, in that potashbegan to replace soda as the regular source ofalkali. The exact reason for this change isunknown, but it may have been the result ofa greater demand for glass (and thereforefuel). Only relatively small quantities of sodawould have been required for the manufactureof glass vessels and ornaments, but, with theonset of the Middle Ages, and especially afterthe Gothic Revolution, there would have beena large demand for window glass, first forchurches and cathedrals and later for palaces.Newton (1985b) has suggested that beech-wood ash was used to make coloured glass,and this would have introduced a high levelof potash into the glass batch. Once it wasrealized that the use of beechwood ash wouldenable a variety of colours to be produced,the glassmakers would have moved into areaswhere beech forests existed, thus ensuring

both a plentiful supply of fuel and alkali.Newton (1985b) has shown that there was ascarcity of beechwood south of the Alps, andhas related the glassmaking centres in north-ern Europe to the distribution of beechwoodpollen in AD 1000. Nevertheless, care must beexercised before trying to draw a picture thatis too simple. Geilmann and Bruckbauer(1954) showed that the manganese content ofbeech wood (as with any component of anyplant) depended on the place where the treehad grown, the maturity of the wood, etc.

Freestone et al. (1999) examined Anglo-Saxon vessel glass from south-eastern Britain,and window glass from the monastery atJarrow (Northumbria) and identified severalcompositional groups. The fifth- to sixth-century claw beakers and cone beakers, andthe window glass were of the low magnesia,low potash (natron) type. The sixth- toseventh-century globular beakers and palmcups were of the high magnesia, high potash(‘plant ash’) type. Three possibilities for theseresults are (i) the recycling of old Roman glass(cullet), (ii) the continued supply of chunkglass from the Mediterranean, and (iii) the useof specific raw materials. A survey of glassfragments from Hamwic (Saxon Southampton,UK) (Hunter and Heyworth, 1999) combinedtypological and compositional data, to suggestan emerging glass industry in middle Saxontimes. The study covers the time in earlyChristian England, when knowledge of glassproduction was only slowly developing, butglass from pagan graves, the usual source ofsuch early glass, was less commonly available.The known fragments are dominated byvessels of the palm cup/funnel beaker variety,some of them being decorated with appliedcoloured reticella rods.

Ancient soda, obtained by burning maritimeor desert plants, was usually contaminatedwith enough lime to produce durable glass.However if the soda was reasonably pure, aswas the case with some natron from somesources, it was necessary to mix it withcalcareous sand in order to impart stability tothe glass. The addition of too much lime, forexample in ash prepared from beechwood inmedieval times, resulted in the glass being lessdurable. Lime was not intentionally added tothe glass batch in ancient times, or in fact untilthe seventeenth century.

Technology of glass production 75

Lead oxide was also an unintentional ingre-dient of ancient glass until Roman times. Itwas introduced directly to the batch, as oneof the two oxides, litharge (PbO) or red lead(Pb3O4), or indirectly as white lead or basiclead carbonate (2PbCO3.Pb(OH)2, or as galena,the basic sulphide ore (PbS). Litharge wasproduced by blowing air over the surface ofmolten lead, an extremely dangerous healthhazard. When the resulting litharge was furtheroxidized, it became red lead. Its use in glass-making required special furnace conditions, asits conversion back to metallic lead woulddiscolour the glass and damage the crucible(or pot).

The high refractive index of lead glass givesa brilliance to the glass, and when the glassis facet cut, spectral colours are reflected fromthe facets. Lead was used with great successin the seventeenth century to produce lead(flint) glass vessels suitable for facet cutting.The usual source of barium oxide, anotherunintentional ingredient, was barium sulphideore (BaSO4, barytes or heavy spar). The inclu-sion of aluminium oxide (Al2O3, alumina),which occurs naturally in clays or, in a fairlypure form, as emery or corundum, and in thehydrated form as bauxite (Al2O3.xH2O), wouldraise the melting point of the batch, andimprove the durability of the glass. Aluminiumoxide was in fact an inevitable contaminant,being introduced from the fabric of thecrucibles themselves.

Colourants

Metallic oxidesThe common colourants for glass wereprecisely the same as those used in makingglazes. Unlike the potters, however, who onlyhad to alter the atmosphere of a kiln toproduce oxidized or reduced colourings, theglassworkers needed the ability to achievethese effects within the batch. Normally thecolourants would be in a reduced condition,but the glass modifiers such as the oxides ofantimony and arsenic could produce oxidizingconditions within the molten glass, and thusoxidized colourings were produced. Doubtlessancient glassworkers learnt this fact empiri-cally.

Coloured glass was generally pot coloured(i.e. in the batch), but in the Middle Ages,

certain colours, especially red, were flashed,that is, clear glass was covered with a thinlayer of coloured glass. To produce pot-coloured glass, metallic oxide colouring agentswere added to the molten batch in thecrucible. Thus, when formed, the glass artefactwould be uniformly coloured throughout itsthickness. In the manufacture of flash-colouredglass, a gather of clear molten glass wascovered by a second gather from a pot ofmolten coloured glass, so that when thecoated gather was blown into a cylinder, cutand flattened out, it formed a sheet of glasswhich was coloured on one side only, thegreater thickness of the glass being clear.Flashed ruby is by far the commonest colourafter the fourteenth century, but examples ofothers, notably blue, are known. The processof flashing glass was described by Theophilusin the twelfth century; and imperfectly flashedglasses are found in the earliest examples ofglass windows. The flashed or coloured sideof the sheet is generally found on the innerface of windows where it would not bedamaged by weathering.

Colloidal suspensions of metalsMetals do not dissolve in glass, but exist ascolloids in suspension. Gold and copper rubyglasses contain those metals dispersed on asubmicroscopic scale. Other highly decorativeglasses could be made to resemble the semi-precious stones agate, aventurine, chalcedony,onyx etc. by dispersing flakes of metal ormetallic oxides in the glass before the artefactwas formed from it. An alternative method ofmaking onyx glass was to place two differentlycoloured opal glasses in the same crucible,and to make the gather before the mixing wascomplete. The cold glass could then be cutand polished so that the new surface displayedalternating layers of the colours.

On the whole, copper was used by theEgyptian and Mesopotamian glassmakers toproduce bluish-green glass, and cobalt toproduce blue or violet glass, although mixturesof copper, cobalt or manganese to produceblue glass were also common. Copper oxidesmay be introduced to the glass melt as any ofthe common copper ores, although the oxidesand carbonates were the most usual. Underoxidizing conditions, copper oxide coloursglass blue or green, depending partly upon


which other glass modifiers are present. Thus,with lead oxide, copper oxide will producegreens; with sodium or potassium oxide thecolour will be turquoise blue. Under reducingconditions copper oxide will produce a dullred colour. Copper oxide is usually present inthe proportions of 2–5 per cent, above whichit causes the colour to darken even to black.The Romans used copper and cobalt toproduce blue glass, but were equally awarethat this colour could be achieved by the useof ferrous iron, and that a reducing atmos-phere in the furnace would deepen the bluecolour by increasing the number of ferrousions present. The making of blue glass contin-ued from Roman through Islamic times andreceived a boost during the Renaissance whenthe Venetians began producing a rich blueglass. Blue glass was also made on a smallscale in the northern Rhine area. It becamevery popular in Britain during the eighteenthand nineteenth centuries, for example thedeep blue glass being made at Nailsea, nearBristol.

In ancient times, red glass was generallymade, by using copper in a reducing furnaceatmosphere. This resulted in a brilliant redopaque glass. A fine red glass from copperwas made in Egypt from the time of theEighteenth Dynasty. The Roman writer Pliny(AD 77) mentions an opaque red glass calledhaematinum, implying it was of local manufac-ture. In medieval times red glass was stillmade with copper, but manganese was usedto make a pale rose red or pink glass.Manganese oxide may be introduced to thebatch either as the dioxide or carbonate ofmanganese, and will produce purple-brown orshades of violet depending upon which othermodifiers are present. For example, with ironit produces black. Manganese is generally usedas about 2.6 per cent of the glass metal; andits colour fades if fired above 1200°C. It wasnot until the end of the seventeenth centurythat a clear ruby or pink glass could be madeconsistently. Iron in its ferric state, (i.e. firedunder oxidizing conditions), can colour glassyellow or amber. Any of the ores from whichiron may be smelted may be ground andadded to the glass metal. When the ore ispoorly ground, or when a red clay containssmall nodules of iron ores, these will producelocal patches of excess iron oxide in the glaze

appearing as dark brown or black spots in thesurface.

The Venetians made a particularly finegreen glass in the Renaissance period. Cop-per or iron could be used to make the greencolour under reducing conditions. Thebrownish-green colour of some English jugsmade at the end of the eighteenth century isthe result of iron impurities, since to avoidexcess duty, the glassmaker had made thevessels from bottle glass.

The Venetians continually experimentedwith coloured glasses, and produced a richpurple glass, probably by the use ofmanganese. Venetian opaline glass of theseventeenth century was made by usingarsenic and calcined bones in the batch. Whenheated, these materials struck an opalescentwhite colour. The Venetians also continuedthe ancient tradition of imitating semi-preciousstones such as chalcedony, jasper, onyx andagate. From the late 1820s, Friedrich Eger-mann, working in Bohemia, made lithyalinglass in imitation of natural stones. This wasa polished opaque glass marbled with red andother strong colours.

From at least the second century BC, blackglass was made, by using an excess of ironoxide, that is, 10 per cent or more of the glassbatch. However, an excess of any colouringoxide will colour glass so deeply that it takeson a black appearance. The only instance ofblack glass being produced in Britain in thefirst half of the seventeenth century is afragment found adhering to part of a crucibleat Denton near Manchester. A combination ofiron, manganese and sulphur in the glass,coupled with a smoky atmosphere in thefurnace, produced a black opacity. The nextinstance of black glass being produced isfound in southern Bohemia, where hyalith (adense, opaque jet-black glass) was made from1820.

OpacifiersAncient glassmakers used antimony toproduce a white opacity in glass. There wasno great interest in opaque white glass,however, until its potential as an imitation ofChinese porcelain was seen by Italian glass-makers. Opaque white glass produced byadding tin oxide or arsenic to the batch wasmade at Venice before 1500 and continued


thereafter, being especially popular in theeighteenth century. The products certainly hadthe appearance of porcelain, but feel different(cool) to the touch, and are generally consid-ered to have none of the aesthetic qualities ofeither material. Tin oxide was normally addedas cassiterite (SnO2), although it could alsohave been added as an impurity in the rawmaterial, for example, when calcined pewterwas added to a batch to produce lead oxide.Tin oxide is not strictly speaking a colourant,but produced a white opacity in glass by exist-ing as a colloid, that is, as a mass of minutecrystals.

Modern chemistry has added a new range ofcolours to glassmaking, such as yellow-brownsobtained from titanium, red from selenium,purple and blue from nickel and yellow-greenfrom chromium. Chromium oxide is generallyadded directly as the oxide chromite, whichalso contains iron, or as a chromate or bichro-mate. It is a difficult colour to use since theresult so often depends not only on the othermodifiers present but also on the temperatureto which it is fired. Thus with lead, chromiumoxide will produce reds at low temperaturesand brown or green at higher temperatures.With tin it will produce reds or pinks. Nickeloxide may be introduced from a number ofsources, chiefly the mineral millerite (NiS).Alone it produces drab greens, but with ironit produces browns, and if used in greatconcentrations, black fluorspar and zirconia areused to produce opaque white glass.

DecolourizersAncient so-called colourless glasses had in factalways displayed a greenish tint due to thepresence of iron from the sand, which existedin the iron (II) state in the glass. Initially, carein the selection and washing of proven materi-als, was probably the most common steptaken to avoid the inclusion of contaminants.It was then discovered that glass containingiron oxide could be decolourized by convert-ing the blue colour of reduced iron to theyellow colour of the oxidized state, forexample, by altering the melting conditions orby adding oxidizing agents such as the oxidesof manganese or antimony to the glass batch.Traces of antimony have been found in glassesdating from at least 2000 BC, probably presentas an impurity in the raw materials. Colourless

glasses could well have been accidentallymade, by using a raw material which,unknown to the glassmaker, containedantimony or manganese. Having obtained adesirable result, the use of the material wasencouraged (Newton, 1985b). For example,the ash from a certain type of plant (one notcontaining antimony or manganese) mighthave always produced a green-tinted glass (theiron was not decolourized) whereas the ashfrom a different type of plant (one that didcontain enough antimony or manganese)might have produced a less green or evencolourless glass. For example, the brownmarine alga Fucus vesiculosus can concentratemanganese to the extent of 90 000 ppmcompared with 4800 ppm in vascular plants ingeneral. The effect of this on the colour of theglass could well have been noticed by thosewho burnt the plants, or by the glassmakers.However, there is an additional technologicalcomplication because the amount of decolour-ization, and even the amount of opalization(in the case of excess antimony), dependedupon the oxidation–reduction situation in thefurnace. Geilmann and Bruckbauer (1954)studied samples of beechwood ash fromdifferent localities and from various parts ofthe trees, and found great differences in themanganese content of the samples.

Manganese was used increasingly from thefirst century AD as a decolourizer. Added tothe batch as manganese dioxide, it oxidizedthe greenish colour of the glass, and the resul-tant yellow tinge was compensated for by thepurple manganese. Nevertheless, the firstauthor to mention manganese explicitly wasBiringuccio (1540, translated by Smith andGnudi, 1942).

Depending upon the firing conditions andamount added, antimony can react as adecolourizer or, as mentioned above, anopacifier. During glass-melting with antimonypresent in the batch, a rise in temperature willconvert opaque glass to crystal clear glass. Thedevelopment of colourless glass by theRomans was a gradual process; the use ofantimony as a decolourizer reached a peak bythe second century AD. Colourless glass formedthe medium favoured by glass-cutters in thesecond and third centuries (Harden, 1969a).Utilitarian ware continued to be made ingreenish glass.


Ancient white opaque glasses containapproximately the same amount of antimonyas the colourless high antimony glasses, butthose which are opaque become colourlesseither when melted under reducing conditionsor at an elevated temperature (Sayre, 1963)Thus another possible explanation of the high-antimony situation may be that the glassmak-ers obtained either clear or opaque glassesdepending on the type of flame in the furnaceand the amount of organic matter in the batch.

Sayre (1963) also commented on twoinstances where parts of a glass article did notcontain the decolourizer, and claimed that the examples were another indication that theantimony was added for decolouration, in thefirst case a colourless high-antimony glass hada blue thread decoration with a low antimonycontent, and in the second the body of a beadwas low in manganese whereas an outer layerwas high in manganese. However, the outerparts of the article might have been made fromimported glass (Newton, 1971b,c).

Bearing in mind the empirical and precari-ous procedures of glassmaking used in ancienttimes, it seems unlikely that the glassmakerswere knowledgeable enough to be able to addabout 1 per cent of an ingredient with anyconfidence and also disperse such a smallamount uniformly throughout the batch(which is not an easy operation even withmodern techniques). Moreover, there seems tobe no reason to believe that antimony wasknown in Mesopotamian times (Brill, 1970c),or that manganese was deliberately added tothe glass batch earlier than AD 1540 (Turner,1956a).

However, whether or not Sayre (1963) wascorrect about the deliberate use of antimonyand manganese, the analyses have resulted inan invaluable wealth of analytical informationabout glasses from many different eras andgeographical locations.

A clear glass luxury-ware was in productionas early as the eighth century BC, the glass-makers probably being based in Gordion inTurkey and Nimrud in Mesopotamia. Theseglasses were mostly bowl-shaped and wereground, cut and polished with a high degreeof technical skill. By the third century BC thesame type of bowl was being produced inAlexandria in Egypt. However, it was theRomans who achieved the regular production

of a transparent and almost colourless glass.The ancient world made a careful distinctionbetween ordinary glass (vitrum) and crystalglass (crystallum or crystallina). The accountsby Strabo (first century BC) (Thorpe, 1938)suggest that crystal glass may have beeninvented between 20 and 7 BC. It had becomewell enough known for Pliny (77 AD) todescribe the manufacture of colourless glassmade from the fusion of ground sand andsoda, followed by a second fusing with shellsand stone referred to as ‘magnum lapis’,perhaps limestone or dolomite. Pliny referredto the product as ‘vitrum purum’ (pure glass)and remarked that the glass was colourlessand transparent and as nearly like rock crystalas possible.

It was only after circa 1500 that the realvogue for colourless glass began, and it thenbecame so popular that it displaced stronglycoloured glass from the market.

Post-Roman glass compositions

In the Dark Ages, glass of the soda–lime typehad continued to be made as the glassmakersin Western Europe were still able to obtaineither natron from Alexandria or marine plantash from the Mediterranean countries. Aroundthe tenth century, however, European glass-makers began using ash from bracken andother woodland plants as a source of alkalithus producing green-tinted potash glass(forest glass). A different forest glass compo-sition (high lime/low alkali) became promi-nent in the sixteenth century in theEichensfeld region of Germany. Wedepohl(1993) suggested that an increase in zinccontent might indicate the use of ash preparedfrom bracken and reeds, as an alternative fuelto beechwood. Jackson and Smedley (1999)have investigated glassmaking in Staffordshire(UK) during the fourteenth to sixteenthcenturies and concluded that documentaryevidence and vegetation history suggest thatbracken not beechwood was the main sourceof alkali there.

In the fifteenth century, Venice was pre-eminent in its glassmaking expertise and there-fore went to great lengths to keep glassmakingknowledge secret. Glassmakers were forbid-den on pain of death to practise their skills


outside Venetian territory. However, briberyand the competition of opposing glassmakingpractices in Altare (near Genoa) encouragedglassmaking technology to be disseminated.Information regarding seventeenth-centurybatch formulations and their use in Murano(Italy) Zecchin (1987) contains a facsimile ofthe original notebook used by GiovanniDarduin (1584–1654), along with a transcrip-tion and modern commentary. Elsewhere inEurope glassmakers producing soda–limeglasses produced vessels in the style of thosefrom Venice (façon de Venise) but without thesame degree of brilliance in the glass.

Towards the end of the seventeenth centuryboth Germany and Britain competed withVenice in producing glasses which weresuperior in quality to the original cristallo. Thefourth book of L’Arte Vetraria (Neri, 1612) wasdevoted to the discussion of various leadglasses and their use for counterfeitingprecious jewellery. It was translated intoEnglish by Merrett in 1662, and the workwould undoubtedly have been known toGeorge Ravenscroft, credited with the discov-ery of lead glass in England. In 1673, the GlassSellers’ Company had commissionedRavenscroft to produce a substitute for Vene-tian cristallo. The first attempts using indige-nous flint and potash to replace pebbles, andsoda produced from the Spanish sea plantbarilla used by Venetian glassmakers, wasunsuccessful since the glass was inclined tocrizzle. This crizzling took the form of anetwork of tiny cracks caused by the break-down in the chemical structure within theglass due to an excess of alkali. To remedythis, oxide of lead was substituted for aproportion of the potash. Eventually thisincreased to as much as 30 per cent of themix. A glass of high refractive index produc-ing a brilliance was the result. Sand soonreplaced the flints but the term flint glasscontinued to be used to designateRavenscroft’s glass. Lead glass was heavier andless fluid than Venetian cristallo but itslustrous appearance established it as the leaderin the production of clear glass from the endof the seventeenth century.

The lead glass was free from seed (minuteair bubbles), because molten lead glass issome 100 times more fluid than a soda–limeglass at the same temperature (see line F in

Figure 1.9); appeared very bright because ithad a high refractive index (Figure 1.11); andsparkled when decoratively cut because of itshigh optical dispersivity; and had a clear ringat room temperature when struck, because itsalkali (potash) ions were bound more closelyto the lead–silica network than are the sodaions in soda–lime glass, and thus absorb lessenergy when in vibration. Lead can be intro-duced into the glass as litharge (PbO), as redlead (PbO3), or as white lead or basic leadcarbonate, (PbCO3.Pb(OH)2).

By 1680 Germany had established a solidclear glass, apparently first made in northernBohemia, by adding lime to stabilize thepurified potash. Potash continued to be usedas a source of alkali even after the change ofwood to coal as a source of fuel. The richpotash deposits at Stassfurt (in Germany) wereexploited in 1861 (Douglas and Frank, 1972).

Use of cullet

A method of reducing the viscosity of moltenglass was to add cullet (broken or scrap glass,either from the glassmaking sites themselves,in which case their approximate compositionwould be known; or obtained as an article ofcommerce brought from another location, andof less certain composition). Cullet is differentfrom chunk glass, which was manufacturedfrom raw ingredients and sold to glassmakers(Freestone and Gorin-Rosen, 1999).

Fuel, crucibles and furnaces

Apart from the raw materials for making theglass, a source of fuel was required – origi-nally wood, or dung; and from the seven-teenth century coal (in modern times, gas, oiland electricity). The glass was melted incontainers made of refractory (heat-resisting)materials such as clays, which were free fromfluxes. Initially these were small crucibles, andas glass technology developed, pots were usedthat increased in size as time progressed andhigher furnace temperatures could bemaintained. Furnaces are discussed in Part 2of this chapter.

The molten glass was traditionally called‘metal’, by analogy with molten metal,however the term is misleading in connectionwith glass. All glass artefacts fashioned from


molten glass had to be annealed (cooledslowly) immediately, in an annealing oven,usually attached to the main furnace.

Crossley (1999) stressed the importance ofinvestigating residues from glassmaking sites.Glass fragments found on site may representglass made during the final use of the furnace,and may therefore not be representative of theearlier output. Material resembling solidifiedfoam (spilled or ladled from the crucibles)may reveal information regarding the sourcesof raw materials.

Manufacture of glass

As previously mentioned, coloured vitreousglazes were used extensively in Egypt andMesopotamia in the fourth millennium BC forcovering objects in imitation of semi-preciousstones. The first essential in making a goodglaze was to reduce all the ingredients to afine particle size by crushing. They were thenmixed with water with or without the additionof an emulsifier to form a suspension. Theglaze was applied, by brush or by dipping theobject into it. The object was then heated ina furnace to distribute the glaze evenly, andto fuse the glaze particles. In antiquity theformulation of any glaze had to be arrived atby trial and error; and when one considers thecost of such experimentation to the potter itis hardly surprising that once a satisfactoryglaze had been devised it remained in use,unaltered, over a long period of time. Intheory it would have been possible to haveproduced a wide range of glazes with differ-ent melting points by altering the proportionsof silica, lime and alkali (soda or potash), butin practice too much alkali resulted in a glazeprone to crazing, while too much silicaresulted in a glaze with a melting point eithertoo high for the primitive furnace or for thebody to which it was to be applied; also, thecoefficient of thermal expansion could becometoo high or too low. It was for this reason thatsoda–lime glazes were only used on objects ofquartz, steatite or Egyptian faience (itself asoda–lime–quartz composite material) but noton pottery.

The transition from using glass for glazingto its being manufactured as a material in itsown right was probably very slow. Since

sustained high temperatures are required inorder to melt the raw materials which produceglass, it seems likely that the art of glassmak-ing originated where there were closedfurnaces able to achieve such temperatures,such as those used for firing pottery or smelt-ing metal. The potter’s craft was already longestablished by the time that glass made its firstappearance, and the similarities between glazeand glass production are apparent.

Tait (1991) has published sequences ofphotographs of Bill Gudenrath, a modernmaster glassworker, demonstrating the tech-niques currently thought to have been used byancient glassmakers.

In any critical discussion of the productionof ancient glasses it may be necessary todistinguish between three or four differentactivities: glassmaking, including fritting, glass-melting and glass-forming.

Glassmaking (fritting)

Glassmaking is the preparation of molten glassfrom its basic raw ingredients, which in thecase of most ancient glasses was calcareoussand and soda. The glass could then befashioned into objects, or could be allowed tosolidify, and then be transported as ingots orbroken chunk glass, for use by glassworkersat other sites, who did not have the technicalability to melt glass.

In ancient times when furnaces could onlyreach the required high melting temperaturesof the order of about 1000°C for short periodsof time, it was extremely difficult to melt theraw materials to produce glass. The resultingproduct was a viscous, semi-glassy materialfull of air bubbles (seed) and unmelted batchmaterials (stones). (At higher, modern furnacetemperatures, the viscosity of the glassbecomes low enough for the seed to escape,and all the batch material will react to producea good quality glass in a single operation.)Therefore, although there is limited evidencefor this, it is thought that ancient glasses hadto be made in at least two operations: frittingand melting. During the fritting process, theraw materials were heated to a point whichallowed solid state reactions to occur betweenthe sodium carbonate in the plant ash ornatron and the silica in the sand, after whichthe mass was cooled and ground to a powder


(frit). This process might be repeated severaltimes to refine the frit, after which the frit wasmelted at a higher temperature to form ahomogeneous molten vitreous material (glassmelt) which could be used to manufactureartefacts (Turner, 1956c).

The fritting process was investigated byHowarth et al. (1934). The process was foundto occur slowly at 700°C, and to increaserapidly with temperature until, at 850°C, therewas a marked change in appearance of the frit,to that of partly melted sugar. This is no doubtassociated with the melting of sodium carbon-ate (melting point 851°C; potassium carbonatemelts at 891°C) (Turner, 1930, 1956c), and itindicated that the furnace temperature was toohigh. If melting occurs, the viscous mass causesgreat difficulties in handling; all contemporarysources (for example, cuneiform tablets, Pliny,Theophilus and Neri) state that only moderatetemperatures should be used, and that the fritshould be kept well stirred so that it should notmelt and lumps should not form. Nevertheless,nearly complete glassmaking reactions canoccur if heating is continued for long enough.At 750°C Howarth et al. (1934) obtained 98.5per cent reaction in 15 hours, and 98.1 per centreaction at 800°C in 10 hours, but muchdepends on the actual proportions of the rawmaterials, the reactions being more rapid whenhigher proportions of silica are present.

When completely reacted frit is ground andwell mixed by hand, it can be melted to forma good quality, homogeneous glass at temper-atures of the order of 1000–1100°C. Much ofthe first-quality Roman glass was no doubtmade in this way, perhaps using more thanone stage of quenching the melt in water,grinding and remixing by hand. This processof multiple quenching was also used to purifythe glass by removing undesirable inclusions.

The subsequent melting was carried out inspecial kilns or crucibles, sometimes in severalstages to permit unfused elements to be drawnoff from the molten mixture. After the mixturehad turned to glass, it was allowed to cool offin the crucibles, which later had to besmashed to release the glass ingots formedinside. The ingots were sold to glassworkshops, where they were remelted andturned into finished objects.

Large quantities of a wide variety of ancientglass have been recovered by the Institute of

Nautical Archaeology (Turkey) from a lateBronze Age shipwreck site (c. 1350 BC), andmedieval glass from a wreck at Serçe Liman(c. 1025 AD). (The shipwrecks were namedafter their locations.) It is thought that bothships were carrying glass for trade.

In 1962, the late Bronze Age shipwreck wasdiscovered less than 1 kilometre northeast ofthe tip of Ulu Burun, a cape near Kaș inTurkey. It lay only 50 m from the shore on asteep slope at depths ranging between 44 and51 m. Excavated in 1984, the ship was foundto have been carrying a large cargo of rawgoods, amongst which were at least twelveround glass ingots, and manufactured goodsincluding glass and faience beads. Oneamphora (KW8 in area J/K 12 was filled withtiny glass beads, solidly concreted together(see Figure 7.8). Not far from Pithos 250, lyingloose under a thin covering of sand, were twoglass ingots (KW 3: max. diam. 0.154; min.diam. 0.125; thickness 0.055 to 0.069; wt2343 g) and (KW 4: max. diam. 0.156; min.diam. 0.141; thickness 0.056 to 0.068; wt2607 g) (Figure 3.1). Roughly discoid inappearance, they are truncated cones with theedges of their smaller faces rounded, and theirlarger, flat faces of rougher texture; the facesare not parallel. They were described as cobaltblue at the time of their discovery, and chemi-cal analysis has subsequently confirmed cobaltas the colouring matter (Brill, 1988). They arethe earliest glass ingots known, and are ofgreat interest as evidence of ancient trading inraw glass. It has been suggested that they are


Figure 3.1 Blue glass ingot, from the Bronze Ageshipwreck at Ulu Burun, Turkey. (Courtesy of theInstitute of Nautical Archaeology, Texas).

probably of Egyptian origin, and that they mayhave been made in cylindrical vessels verysimilar to those found at Tell el-Amarna.Further analysis will be required to prove thistheory (Nicholson et al., 1997). The ingotswere unusually clean, with few traces ofconcretion or weathering layers although theywere scarcely buried. Presumably they hadbeen abraded by sea-bed action and thereforeit is surprising that they are the largest andheaviest raised from the site.

About 1025 AD a merchant ship of unknownorigin and nationality sank in 36 m of waterwithin the rocky confines of Serçe Liman, anatural harbour on the southern Turkish coastopposite the Greek island of Rhodes. Thewreck was excavated between 1977 and 1979and yielded the largest amount of medievalglass yet known, having been loaded with3000 kg of cullet in baskets, both in the formof chunks of raw glass and broken vessels,stored in the aft hold area of the ship. Inaddition, over eighty intact vessels, includingengraved bowls and beakers, presumablyhaving belonged to the merchants themselves,were found in the living quarters at the bowand stern (Bass, 1978, 1979, 1980, 1984).

Egyptian faience

The glazed quartz material known as Egyptianfaience was, as the term suggests, first madein ancient Egypt (Peltenburg, 1987). Glazedsteatite and faience beads are known from anumber of Predynastic sites (5500–3050 BC). Itcontinued to be manufactured throughoutancient Egypt into the Later Periods (1070 BC

to 395 AD). The material was also produced inMesopotamia (4300–3900 BC). It would seemthat Egyptian faience was initially produced inimitation of the highly valued, semi-precioustones, turquoise and lapis lazuli. During thenext five thousand years, manufacturingcentres were established throughout NorthAfrica, the Aegean and Asia, often alongsideglassmaking operations (Shortland and Tite,1998). Developments in glass technologyproduced an increasing range of colours.

By 2000 BC. Egyptian faience objects weretransported to Europe and even to Britain byestablished trade routes. By the first millen-nium AD, the material had largely fallen out ofuse except in the Near East. During the twelfth

century its manufacture was revived in Syriaand Persia, where white faience-type bodieswere produced in imitation of Chinese porce-lain and stoneware.

The vitreous material Egyptian faience is nota true glass, since it is not homogenous on amicro scale. The body of faience artefacts arecomposed of crushed quartz or sand withsmall amounts of lime, metallic salts and eithernatron or plant ash, made into a paste withwater and organic resin, e.g. gum of traga-canth. The basic mixture forming the quartzcore varied in colour from white, grey to pinkor orange depending on the impuritiespresent. (Small amounts of clay weresometimes added to improve the plasticquality of the faience. However, the brownbody colour was reflected through the glaze,producing a dull, muddy appearance. Thiscould be avoided by applying a surface layerof quartz to provide a white ground for thecoloured glaze.) The core was then coatedwith a soda–lime–silica glaze, which wasgenerally a bright blue–green colour due tothe inclusion of copper-based salts either frommetal parings or crushed minerals containingcopper such as malachite (see Plate 2). Thetypical faience mixture was thick at first andbecame soft and flowing as it began to bedeformed, though it cracked if deformed toorapidly.

The earliest objects manufactured werebeads, amulets and inlays, which could behand-modelled and refined by surfaceabrasion. From about 2040 BC open mouldswere used to make complete objects, or partsof objects, which were joined together withglaze slurry whilst still damp. Wheel-throwingwas possible to a certain extent, but the result-ing vessels tend to be very thick-walled. Oncethe required shapes had been formed, theobjects were finished by grinding and smooth-ing, and decorated by incising the surface, andafter circa 1782 BC, painting on details such asfacial features and hieroglyphics, with a blackor brown slurry, or pigment wash colouredwith manganese and iron oxides. The decora-tion and glazed object were then glazed byone of three methods (efflorescence, cemen-tation, application) and fired at temperaturesin the range of 800–1000°C (Figure 3.2).

The faience paste shrank by 4–12 per centon drying, and if clay had been added to the


mixture, differential shrinking could causecracking and loss of glaze during firing.Cracking sometimes also occurred during thefiring of objects composed of several piecesluted together.

The efflorescence technique was probablythe one most commonly used by the ancientEgyptians. It was essentially a self-glazingmethod, in which water soluble alkali salts suchas the carbonates, chlorides, sulphates ofsodium (and less commonly potassium), in theform of plant ashes or natron, were mixed withthe quartz material forming the core of theobject. During the process of drying out, thesalts migrated to the surface to form an efflo-resced crust or bloom. When fired, the crustmelted and fused with the fine quartz, copperoxide and lime. The technique produced varia-tions in glaze thickness, it being thickest inareas of greater evaporation and drying such asthe exterior surface. Where the object had beenin contact with a surface or where there hadbeen secondary working of the object’s shapeby cutting away the surface, efflorescence wasreduced or prevented. This glazing technique

produced the greatest amount of interstitialglass (the glassy phase in the gaps between thequartz particles), and the interface between theglaze and core is usually clearly defined.

Cementation was also a self-glazingtechnique, in which the unglazed but dryfaience core was buried in a glazing powdercomposed of lime (calcium oxide), ash, silica,charcoal and colourant. On heating, thepowder partially melted and that which wasin contact with the quartz core reacted withand glazed it. Unreacted powder wascrumbled away from the glazed object oncompletion of the firing. Objects produced bythe cementation process can be recognized bya fairly uniform, but often thin, all-over glaze,the absence of drying marks. In the case ofsmall objects there may also be a lack ofdrying marks, whereas larger examples mayexhibit rough areas where they rested duringfiring. Occasionally the glaze may be thickeron the underside of pieces produced by thecementation technique. There is very littleinterstitial glass, and the interface between theglaze and core is generally well defined.


Figure 3.2 Glazing techniques for Egyptian faience: (a) application of a glaze slurry; (b) efflorescence of a glaze;(c) cementation of glaze.

Application, as the term suggests, entails theapplication of a glazing powder, or slurry tothe faience core, either by immersion of theobject or painting it with the glaze. Thepowder held in the slurry may comprisequarts, lime and natron, which have beenground up together, or raw materials whichhave been fritted together and then crushed.The porous quartz body absorbed some of themixture so that the mixture adhered to it ondrying. On firing, the coating melted to forma glaze, which was often fairly thick. Objectsproduced in this way are recognized by thepresence of drips, brush marks or flow linesand the greater thickness of the glaze layer.They may also have a clear edge where theapplication was ceased in order to enable thepiece to be handled or to prevent it fromadhering to the kiln supports. The amount ofinterstitial glass is small and the interfacebetween glaze and core not well defined.

The forming of the body and glaze variedconsiderably over time. In comparison withsmall local workshops, those enjoying royalpatronage are likely to have had access to thebest materials and time to experiment. Thethree basic techniques of producing Egyptianfaience objects are not always readily identifi-able, especially since they were sometimescombined. Nicholson (1993) reproduces atable of methods of Egyptian faience manufac-turing techniques by period and illustrates thetechniques with diagrams and photographstaken through a scanning electron microscope.Sode and Schnell (1998) have documentedtechniques of faience manufacture in modernEgypt.

Faience beads

The vast majority of surviving faience objectsare cylindrical beads (Figure 3.3). Faiencebeads are characteristic of the Bronze Agewhen bronze-melting furnace temperatureswere not high enough to fuse glass satisfacto-rily, but from which slags were readily avail-able as a source of colourant. The earliestexamples seem to have been made inBadarian contexts (fifth millennium BC). In theNear East, faience beads may have been madein millions; hundreds of thousands have beenfound in one site alone, the Grey Eye Templeat Brak in the Lebanon; and a site at Nineveh

in Iraq yielded a layer 2 metres thick. Thebeads generally measure some 6–30 mm long,are 2–10 mm in diameter, have an axial holeand most are segmented.

Relatively few faience beads (some thirty inall) are quite different in shape, being eitherdisc-shaped (quoit beads) or star-shaped, andmuch larger, about 30 mm in diameter. Theseare found almost solely in the British Isles(Stone and Thomas, 1956) and some authors(Newton and Renfrew, 1970; Aspinall et al.,1972) have argued that these two types weremade locally, possibly at Glenuce and Culbinin Scotland (UK). Tite et al. (1983) carried outextensive research, first in Iran where moderndonkey beads were being made at Q’om, andsubsequently in the laboratory. Briefly, thecore of the bead was composed of finelyground quartzite, made into a paste with gumtragacanth, and the glaze was made from theash of Salsola kali and Salsola soda, mixedwith slaked lime, powdered quartz, charcoaland copper oxide. After drying, the shapedbeads were embedded in the glazing powderand the whole mass was fired at 1000°C. Thebeads were then removed from the glazingpowder (which apparently had not melted at1000°C) and were found to have a bright,glossy blue glaze. If the firing temperature was


Figure 3.3 Egyptian faience beads, typical of thesegmented type found in Egypt and other Mediterraneansites. Scale of millimetres below.

only 900°C the glaze was coarse and pale, butat a temperature of 1110°C the glazing powderfused into a useless block. Thus the heatingregime was fairly critical and it seems surpris-ing that the plant ash did not fuse at a lowertemperature.

Wulff et al. (1968) undertook chemical andpetrographic analyses of the plant ash, thebead body and the glazing powder from beadsfound in Iran. It was concluded that the grainsin the core had sintered together through theformation of cristobalite (a crystalline modifi-cation of silica) at their points of contact; andthat the glazing action had been brought aboutin the vapour phase through the formation ofsoda (Na2O) vapour, resulting from the inter-action of the sodium carbonate (Na2 CO3) withthe slaked lime (Ca(OH)2), which then reactedwith copper chloride (CuCl) vapour formed bythe interaction between sodium chloride(NaCl) in the plant ash and the copper oxide(CuO).

Noble (1969) adopted quite a differentapproach by mixing glazing material with corematerial before the beads were shaped. Whenthe bead then dried out the sodium carbonateand bicarbonate (NaHCO3), together with thecopper oxide, had migrated to the surface asan efflorescence. Thus, when the bead wasfired at 950°C, the alkali-rich efflorescencemelted to form a glassy blue layer. If thecopper was omitted, the resultant glaze waswhite but if iron oxide had been used theglaze was yellow. Manganese would give apurple glaze and Noble, in commenting onWulff et al. (1968), claimed that the aboveprocess for making faience articles was correctbecause ancient multicoloured articles existand these could not have been made by thevapour process. However, the natron used inNoble’s experiments may not have been repre-sentative, since the material seems to havemuch less Na2CO3 and NaHCO3 than the 14analyses quoted by Douglas and Frank (1972).

Glass-melting

Glass-melting is the preparation of moltenglass by the melting down of prepared frit,chunks or ingots of raw glass or of cullet(broken, waste glass). Glass-melting does notrequire a knowledge of or access to rawmaterials, and is technologically an easier

process to carry out than glassmaking. In somecases glassmaking and glass-melting werecombined by adding cullet to the raw materi-als during the melting process, in which case,the chemical characteristics of glass melt maynot be as well defined as material made bythe glassmaking process, i.e. if the cullet hadbeen imported from another geographicalarea. On the other hand, a long-continued re-use of cullet in one area could perpetuate achemical characteristic of the glass long afterone of the raw materials had ceased to beavailable. For example, such a practice mightbe one reason for the continual melting ofsoda glass in the Rhineland long after Romanglassworkers had left the area.

Glass-forming

Glass-forming is the heating and reshaping ofa piece of glass without actually melting it.Many glass articles are said to have been madeat certain sites whereas they were really onlyformed there. For example, a Romano-Britishpurple and white glass bracelet from TraprainLaw in Scotland was formed by reheating apiece of a Roman pillar moulded bowl(Newton, 1971b). Since the glass was plasticrather than molten, the purple and whitelayers were deformed only, and did not mergeinto one another.

Annealing

Annealing is a process which is applied to allglass artefacts in order to relieve stresses setup as a result of rapid cooling, or differentrates of cooling of various components, forexample, body, handle, foot. Failure to annealglass in a suitable manner would leave it weakand brittle. Annealing is usually carried outimmediately after the shaping process. Theglass is reheated in a special annealing furnace(lehr) to the required annealing temperature(which depends upon the composition of theglass), and is then allowed to cool at acontrolled rate, whilst still in the furnace. Thesmall amount of viscous flow, which takesplace during the annealing process, leads tostress relaxation within the glass which is thusrendered relatively stress-free. It is thusunlikely that conservators will encounter badlyannealed glasses since these would have


fractured during their manufacture, or shortlyafterwards during their functional life.

Shaping glass artefacts

The development of glass bead-making

The earliest vitreous material had been usedin Egypt before circa 4000 BC, as a glaze tocover beads of stone and clay in imitation ofcoloured semi-precious stones. Later, circa2500 BC, when furnaces were able to bemaintained at temperatures high enough tosoften glass, the same material was used inEgypt to make beads, which were the firstobjects to be made entirely of glass. Glassbeads are known from Mycenae from thesixteenth to the thirteenth centuries BC. Glasspaste beads made at Mycenae circa 1300 BC

are in the form of small thin tablets of which

the ends are ribbed and perforated for thread-ing. They were circular, rectangular or trian-gular, and usually of blue or pale yellow glass.The beads are decorated with reliefMycenaean motifs, such as rosettes, ivy andspirals. They were formerly believed to havebeen used for necklaces, and as decoration ongarments, but a recent view is that the beadswere also used to adorn diadems and theskulls of skeletons (Yalouris, 1968).

Glass beads are found in most periods andcultures (Guido, 1977; Dubin, 1987; Kock andSode, 1995; Sode, 1996). They can be madeby at least seven methods (Figure 3.4): (i)winding threads of glass round a rod; (ii)drawing from a gob of glass which has beenworked into a hollow; (iii) folding glassaround a core, the join being visible on oneside; (iv) pressing glass into a mould; (v)perforating soft glass with a rod; and (vi)blowing (though cylindrical blown glass beads


Figure 3.4 Techniques ofbead making: (a) threadingsoft glass around a wire;(b) folding soft glassaround a wire; (c)perforating a lump of softglass; (d) perforating anddrawing out a lump ofglass; (e) fusing powderedglass in an open mouldand drilling holes throughthe corners; (f) drillingthrough a cold lump ofglass; (g) forming a mosaicglass bead.

are rather exceptional); (vi) drilling through asolid block of glass. The same methods ofdecoration were applied to glass beads as toglass vessels, that is, tooling the soft glass,thread inlay, mosaic inlay, millefiore etc. Anagate effect was produced by casing glasstubing with layers of differently coloured glassand then chamfering down the ends of cutlengths to expose the various layers. Examplesof this type are the aggry (aggri) beadsexcavated in Africa.

The early decorative patterns on beads fromcirca 1500 BC are stripes and spots; later devel-opments are eye beads and beads with zig-zagsand chevrons. Egyptian beads were exported tomany countries. These early beads werenormally made of opaque glass, frequently bluewith decoration in yellow and white. Since themanufacture of glass beads could only becarried out in furnaces hot enough to melt iron,their production in Britain is mainly associatedwith the Romano-British period; although onthe Continent large annular beads and armletswere made in the La Tene I period (fifthcentury BC). A comprehensive study of prehis-toric and Romano-British beads found in Britainand Ireland (Guido, 1977) has shown that somebeads were indigenous, but that many wereimported. In some cases the beads or braceletswere made by reshaping fragments of Romanglass articles, and there is a strong suggestionthat the highly coloured glass used for theapplied decoration (spirals, eye spots etc.) wasan article of trade imported from a glassworkswhich specialized in making coloured glass,perhaps in Gaul.

Most of the methods used in making anddecorating glass beads, can be applied to themanufacture of bangles. If not cut from a solidpiece of glass, however, the commonestmeans of producing bangles are either bybending a glass rod round and fusing the twoends together; or by first blowing a hollowglass cylinder and then cutting it into shortlengths. In the latter instance it is not uncom-mon to find that the glass gathering has beencased with several layers of differentlycoloured glass.

Manufacturing glass vessels

Fourteen hundred years before the inventionof glass-blowing, four very different tech-

niques were already in use for the productionof glass artefacts: core-forming; moulding;cutting or abrading (cold glass); and mosaic.

Cold, solid glass can be shaped anddecorated by cold working, i.e. by glyptictechniques such as cutting and engraving.However, having the property of becomingmolten (plastic) when heated to sufficientlyhigh temperatures, means that practically allancient (and modern) techniques of manufac-ture use glass in that condition and thereforecan be described as hyaloplastic. Since glasscan possess a remarkably wide range ofviscosities (and thereby working ranges)depending on its composition and temperature(see Figure 1.9), glass artefacts can be madein an almost infinite variety of shapes. Noother material has these properties or permitssuch a varied manufacture to be undertaken.In ancient times knowledge of the workingranges would have been part of the long-heldsecrets of the glassmaking process.

With the invention of glassblowing, newtechniques of producing glass artefacts weredeveloped, and along with them the methodsof decolouration. These techniques have thusbeen in use for about two millennia, usuallywith increasing sophistication, despite theintroduction of industrial mass manufacturingmethods in the nineteenth and twentiethcenturies. Consequently, glassmaking tech-niques will be described without attempting todeal with them in strict chronological order,although where possible an indication will begiven of the date of their introduction. Toolsfor hand-working glass have remained essen-tially unaltered; but increased sophistication ofglass designs led to changes in composition,that is choice of raw materials, and in furnacedesign. These will be discussed in the light ofsurviving ancient glass and historic glassobjects, of contemporary literary sources, andmodern melting trials and hypotheses onancient glass production. Modern methods ofglassworking will be mentioned only briefly,since they are outside the scope of this book.

Core-formed vesselsThe majority of pre-Roman glass vessels weremade by the core-forming method (Figure3.5). Small vessels formed in this manner wereformerly termed sand-core vessels after Petrie(1894) introduced the term, without supplying


evidence that sand had been used to form thecore.

Labino (1966) carried out a series of exper-iments, which showed that core-formedvessels could be made by trailing glass on toan entirely organic core that had a coefficientof expansion greater than that of the glass sothat the vessel would not crack on cooling.However, Labino did not disclose the exactcomposition of the core material in order toprevent its use in the production of fakes.Bimson and Werner (1967) examined twocore-formed vessels – a Cypriot Bronze Agescent bottle and an Eighteenth Dynasty modelcoffin, inside which, remains of the cores werefound. Both samples suggested that the origi-

nal core had been made in two layers from afriable porous mass consisting of fragments ofplants (probably as dung), a highly ferrugi-nous clay and ground limestone, the outerlayer being largely ground limestone. Bimsonand Werner (1969) examined a further 62samples of core material and found that early(Eighteenth Dynasty) material did not differsubstantially from the two samples mentionedabove, although later material (after 750 BC)consisted of sand grains cemented togetherwith iron oxide. Wosinski and Brill (1969)examined ten examples of core material fromthree basic types of core-formed vessel andfound that the core materials were mixtures ofsand, clay and plant material (again possiblydung), but did not confirm the presence of asurface layer of ground limestone (Figure 3.6).

It was formerly supposed that core formedvessels were made by dipping the core intothe molten glass, but this has been shown notto be the case (Stern, 1998). A core shaped tothe interior form of a vessel would be madein a soft but firm material such as a mixtureof dung and sand, on the end of a woodenrod. A trail of molten glass was wound aroundthe core and when a sufficient thickness ofglass had been built up, the glass and corewere repeatedly reheated, rotated and rolledsmooth (marvered) on a flat stone surface. Theglass could be reheated and decoration addedin the form of trails of molten coloured glass.The trails could be combed into patterns suchas swags and zig-zags, and embedded in theglass surface by reheating and marvering.


Figure 3.5 Stages of core-forming glass vessels: (a)winding a soft glass thread around the core andmarvering it to shape; (b) winding soft coloured glassthreads around the vessel and embedding them bymarvering; (c) the mouth, foot and handles are formed;(d) when cold, the core is scraped out; (e) the finishedvessel.

Figure 3.6 Interior of a broken core-formed vessel,showing remains of the burnt core.

Additional parts such as rudimentary handles,stems and footrings might be added. Oncooling, the rod was removed and as much aspossible of the core picked out, followingwhich a rim could be trailed on and shaped.The carrot and lentil shapes of the earliervases and bottles were natural marver shapesand these vessels were usually about80–120 mm in height, although a few largerexamples exist. Core-formed vessels weremainly produced in Egypt but examples havebeen excavated in Mesopotamia, Cyprus, Syria,Crete, Rhodes and mainland Greece. Stern(1998) has shown that it is possible to applyfinely ground glass to a wetted core, bypacking it against the core with a wet brush,heating the powder to a temperature of 593°Cto glaze it over, followed by a firing at 816°C.

Four distinctive types of core-formed vesselsare known from the later Egyptian period (seeFigure 2.2): alabastron (cylindrical or cigar-shaped); amphorisk (pear-shaped); aryballos(globular); and oinochoe (jug with one handleand a flat base). These containers (unguen-taria) seem to have been luxury articles andwere used for ointments, perfumes andcosmetics. Core-formed vessels continued tobe made throughout the Hellenic period,slowly retiring in the face of blown glass andmillefiore-type vessels.

Casting in open or closed mouldsCasting into open moulds was a techniquealready in use in the pottery and faienceindustry. It was used for the production ofopen glass vessels such as bowls, dishes andwide-necked bottles, and plaques. Threetechniques were probably used: (i) fusing ofpowdered or chipped glass, poured or hand-pressed into a mould; (ii) direct pouring andmanipulation of the glass into the mould; and(iii) the lost wax (Fr. cire perdue) process. Thefigurine of Astarte from Atchana, Turkey(Figure 3.7), dating from the early fifteenthcentury BC, was cast in a one-piece mould, theback being flattened by pressure (Harden etal., 1968). Numerous clay moulds, whichcould have been used for moulding glazedquartz fritware, have been found in Egypt.Phidias (c. 450 BC) cast small pieces of glassin clay moulds in situ at Olympia for thefamous statue of Zeus (Bimson and Werner,1964b).

Fusing powdered glass in situA closed bowl-shaped mould formed of twoparts could be made of refractory clay; or bycarving pieces of wood to the desired shape,and when fitted together, lining the woodenmould with clay to act as a barrier. The mouldpieces were then fitted together leaving aspace between them which represented the


Figure 3.7 Figurine of Astarte, originally translucentgreenish-colourless, now appearing opaque white.Plaque cast in a one-piece mould, the back flattened bypressure, and moulded in the front in the form of astanding figure of the goddess. L 85 mm, W 23 mm, T16–18 mm. Fourteenth or thirteenth century BC. Atchana,Turkey. (© Copyright The British Museum).

thickness of the glass artefact to be cast. Themould would then be heated to 700°C and asthe temperature increased powdered orfragmented glass were dropped through anopening in the top of the mould. The glasseventually filled the mould (including theorifice which left a knob on what was tobecome the base of the bowl, and which hadto be ground flat after the glass hadhardened). When the melting temperature of1000°C had been reached, the furnace wasallowed to cool over a long period of time,and the mould taken apart to release the glassvessel (Figure 3.8).

Alternatively, coloured glass pastes could beplaced in the mould (Fr. pâte de verre or pâtede riz). Glass ground to a powder was mixedwith a fluxing medium so that it would meltreadily, and coloured. It was pressed into theclay by hand and fired. Coloured glass pastescould be freely modelled like clay. The variedcolouring suggested by semi-precious stoneswas obtained by the positioning of differentpowdered ingredients in the mould. Someexamples were built up into polychrome highrelief or figures by successive layers added tothe mould, and sometimes after refiring theywere refined by being carved. This processwas known in ancient Egypt. It was revivedin France in the nineteenth century to formartificial jewels and other decorative articles,especially by the sculptors Henri and JeanCross. Other exponents of the technique wereAlbert Dammouse (c. 1898), FrancoisDecorchement (c. 1900–1930), Emile Gallé andGabriel Argy-Rousseau (Newman, 1977).

The moulding operation using a two-piecemould could be hastened by pressing on the

upper (male) mould if the mould material wasstrong enough. Such pressure will also producea sharper impression of any pattern on the insideof the external (female) mould. During thenineteenth century and subsequently, a mechan-ical mould-pressing technique using metalmoulds was used to provide cheap imitations ofdeep-cut glass. Glass made by this process bearsthe exact design and the contours of the twoparts of the mould. Nineteenth-century glasswarewhich has external decoration (such as raisedribbing) and corresponding internal concavities,will have been made by the blown three-mouldprocess, or by pattern moulding also knownfrom the nineteenth century onwards as opticmoulding (Newman, 1977).

The lost wax processGlass artefacts, such as figurines with acomplicated shape with many undercuts,could be made by the lost wax process (Fr.cire perdue; Ger. Wachsausschmelzfabren).The desired object was modelled or carved inwax. The model was then encased in a mouldof refractory material such as clay, incorporat-ing airholes and pour-holes at the base. Themould was heated and inverted, so that thewax ran out and powdered glass was thenintroduced to the hot mould, which wasreheated, thus melting the glass to form a solidartefact. Alternatively, if a hollow glass objectwas required, such as a pillar-moulded bowl,an internal (female) mould was made of woodor refractory clay and covered with hot waxapplied with a brush (Schuler, 1959a). Thewax was shaped with a template or by build-ing up and carving the wax to form the exter-nal shape of the object. Small metal barsincorporated into the female mould andprotruding through the wax into the malemould kept the mould pieces apart during thecasting process. A hole was left in the base ofthe male mould in order for the molten waxto run out when heated. Thus the space insiderepresented the form of the glass object to becast. Powdered glass was then introduced intothe mould and heated in order to melt it. Oncooling, the moulds were removed and theresult was a hollow glass object (Figure 3.9).

Mosaic glassMosaic glass is composed of thin sections, cutfrom plain or coloured glass rods. The


Figure 3.8 Forming a glass bowl by placing crushedglass into a mould and heating the mould until theglass fuses.

Refractorymaterial

simplest way of producing a fine, short glassrod was to pour the molten metal very slowlyfrom the crucible or ladle so that as it fell itcooled and solidified. In the same way a glassrod can be formed by dipping a bait into thesurface of molten glass and then slowly raisingit. The molten glass adheres to the bait andbehaves rather like treacle, necking in rapidlyto form a thin stream. Treacle will continue torun back from the bait but glass will soonfreeze and stiffen (depending upon the rate atwhich the bait is raised, the temperature abovethe glass and the composition of the glass). Arod of surprisingly constant thickness can bemade by this method, thicker rods beingobtained with a slower rate of draw.

At a much later date, longer and more evenlengths of rod could be made, by fixing agathering of molten glass metal to an iron postor a plate on a wall and pulling out a lengthof rod with the aid of a punty by walkingaway from the gather. Much later again glasstubing was produced in the same way except


Figure 3.9 Casting a glass bowl by the lost wax (Fr.cire perdue) process. (a) The material used for themould was refractory clay. After the core was formed itwas dried in air, (b) hot wax was applied and shapedwith a template. Ribs of wax were shaped and applied,and a lump of wax was applied at the top to form acavity for holding glass. (In a two-part mould clay canbe used instead.) (c) The wax was covered with morerefractory clay and left to dry. (If the mould was two-part, the join between the pieces was coated with athin wash of clay.) (d) The wax was melted out withthe mould inverted. (If the mould is two-part, the partsare separated, cleaned of clay and reassembled, usingrefractory clay to seal the joint.) The mould was placedin a furnace and heated slowly to 700°C. At this pointsmall pieces of glass were placed in the hollow part ofthe top of the mould. The temperature was graduallyincreased, and as the glass melted down, more wasadded. (The temperature might have been held at 800°Cif the process were to take longer.) The glass floweddown and filled the space in the mould; eventuallythere was excess glass in the hollow space at the top.When the temperature reached 1000°C and glass wasobserved at the top, the mould was assumed to havecompletely filled, and the furnace was cooled. Whencold enough to handle, the mould was taken out andbroken apart.

(a)

(b)

(c)

(d)

(e)

that the glassworker drew out the gather witha blow iron, walking backwards and blowinggently down the iron at the same time in orderto maintain the cavity in the tube. Needless tosay, this operation required considerable skill.

Monochrome rods of different colouredglass were then sliced into sections whichcould be placed adjacent to one another andfused to form a pattern by heating (Figure3.10). The technique was used to formplaques in which the mosaic slices wereadhered to a backing with bitumen. In thecase of a face, the image was formed as onehalf only, two slices from the same rod beingplaced side by side, one being turned over so

that a completely symmetrical face wasproduced (Schuler, 1963).

The mosaic technique was also used to formglass vessels. A group of vessels dating fromthe late fifteenth to the early thirteenthcenturies BC from Tell al Rimah, ‘Agar Quf andMarlik in north-west Iraq were made by themethod, the different coloured sections beingarranged in patterns or in zig-zag bands(Figure 3.11), probably on the outer surface


Figure 3.10 Stages in forming a mosaic plaque with adesign of a mask. (a) Pre-formed coloured glass canesassembled to form half of the mask, in large section(b). (c) The assemblage is drawn out to the desired size(d), from which thin slices are cut (e). (f) Two slicesare placed face-to-face to form (g) a complete mask.

(a)

(b)

(c)

(d) (e)

(f) (g)

Figure 3.11 Fragments of mosaic beaker, composed ofcircular sections of white, red, yellow and dark brownopaque rods, set in horizontal zig-zags to form thewhole vessel in a pattern which has the same aspectinside and outside. Built on a core, covered with anouter mould and fused, exterior ground and polished.The colours are often obscured by enamel-likedeterioration. L of largest fragment 77 mm, W of largestfragment 60 mm, average T 4 mm, average D of sections2 mm. Fifteenth century BC. Tell al Rimah, Iraq. (©Copyright The British Museum).

of a circular mould having the interior shapeof the vessel to be produced. The sections ofrod may have been fixed to the mould withan adhesive, which burnt out during firing or,in some cases, an outer mould may have beenused to keep the sections in place. The mouldwas then heated sufficiently to soften thesections with the minimum amount of distor-tion required to fuse them together to form amosaic vessel. In some cases the glass rodswere multi-coloured in concentric rings. Whenthe mosaic was properly fused, and the article

had cooled, it was removed from the mould,ground smooth and polished internally andexternally (Figure 3.12). Mosaic bowls such asthose mentioned above were the prototypes ofthe mosaic glassware of the late Hellenic andlater periods.

The distinction is not always made betweentrue mosaic and mosaic inlay, although thetwo are quite different. In true mosaic thecoloured glass pieces formed the entire thick-ness of the article, whereas in mosaic inlaythin sections of glass rod were backed by aglass of uniform colour. In Venetian millefioriglass, which sought to imitate true Romanmosaic, the sections of coloured glass wereembedded in clear glass and then blown tothe final shape of the vessel. Millefiore is oftenincorrectly used to describe any type ofmosaic inlay.

The making of millefiori artefacts requiredgreat skill and patience, especially whencomplicated rod patterns were involved. Therods were produced by first casing a glasscane or tube with several layers of colouredglass on an appreciably larger scale than thatwhich was finally required. Each layer couldbe reheated and shaped by marvering (round),pressing on the marver (square) or rolling ona corrugated surface (flower shape). (Lateriron moulds were used to produce theseshapes.) When the design was complete, theglass rod was reheated to its softening pointand drawn out so that its diameter decreased,the design decreasing proportionately until itwas about 20–30 m in diameter. The rod wasthen sectioned, the slices placed adjacent toone another, sometimes embedded in clearglass and fused by reheating (Figure 3.13). Asa variation, sections of more than one mille-fiori rod or slices cut at an angle instead oftransversely across the rod could be fusedtogether, the latter producing elongated stripesof the design.

In AD 1495, Marc Antonio Sabellico, thelibrarian of San Marco in Venice (Italy), wroteof the inclusion of, ‘all sorts of flowers, suchas clothe the meadows in Spring, in a little ball’(of glass). The objects described, soundremarkably similar to the paperweights whichwere made in Murano in Italy and Baccarat, St-Louis and Clichy in France, and which were tobecome fashionable in the nineteenth century.Paperweights often had thin designs such as


Figure 3.12 Stages in the formation of a reticelli bowl(multi-cane), and a mosaic (cane-sectioned) bowl with areticelli rim. (a) Rings of reticelli canes are placed overa former. (b) When the former has been completelycovered, the whole is heated until the canes fuse. (c) Areticelli rod is formed to a similar diameter to that of amould. (d) This cane is used as an accurate perimeterwithin which to fuse sawn cane sections to create a flatdisc for sagging. (e) The disc is softened over, or into amould to create a bowl, which is then finished bycutting and polishing.

flower petals (millefiori) incorporated in thebody of coloured glass by picking up a thinsection of glass on a gather and then takinganother gather over the top (Newman, 1977).Figure 3.14 shows a piece of jewellery with adecorative centre made in the same manner.

Cutting, grinding and abradingCutting or grinding glass is the only techniqueinvolving removal of glass from a cold mass,which was used to produce a complete vessel.Techniques for stone- and gem-cutting werewell established before the advent of glass-making, and it was natural that they should beextended to cutting glass once this materialcould be produced in massive form. On theMohs scale of hardness, cold glass has ascratch hardness between 4.5 and 6.5 depend-ing on its composition, and it can be cut orground by harder materials such as quartz or flint (scratch hardness 7.0). Cold-cutting orgrinding were used at an early date for finish-ing glass objects or vessels which had beencast (Harden, 1968).

The head-rest of the Pharaoh Tutankhamun(1352 BC) (Cairo Museum) was cut from twoblocks of glass, the join between them beingcovered by a gold band. A much later exampleof glass abrasion, an alabastron bearing thename of King Sargon of Assyria (772–705 BC)(British Museum, London), bears internal spiralgrooves and an internal knob at the bottom,which are the result of its having been cold-worked (Figure 3.15).


Figure 3.13 Stages in the formation of millefioreobjects. (a) A cane is formed of concentric layers ofcoloured or patterned glass. (b) The soft glass cane isrolled over a corrugated surface to shape it into aflower. (c) The flower shape is embedded in clearglass. (d) The rod is drawn out and cut into sections,which (e) are fused together and encased in clear glass.(f) The rod is again drawn out, and (g) cut into slices.

Figure 3.14 Silver cross with a central design inmillefiore, and a group of drawn millefiore glass canes.

Figure 3.15 Glass vase engraved in cuneiform scriptwith the name of King Sargon II of Assyria. The interiorgrooves show that the vessel was ground from a solidblock of glass. H c. 150 mm. Nimrud, 722–705 BC. (©Copyright The British Museum).

In general, the techniques of casting glassinto moulds, followed by the discovery of glassblowing replaced the production of glassvessels by abrasion of solid glass blocks.However, the technique continued as a meansof further embellishment as, for example, in theproduction of diatreta, as a means of surfacedecoration (cutting and engraving) or simplyfor finishing off cast glassware (Harden, 1968).

Good glassware was often (and still is)given a final polish using fine, hard abrasive,and where this was well carried out, much orall of the evidence that would suggest howthe vessel had been made was removed – forexample, the casting flashes which remainedon the glass at the junctions of the mould-pieces.

The process of cold-cutting glass was broughtto an extremely fine art in making diatreta (cagecups). As the name suggests, the cups bear apattern which stands out from the (moulded)cup as an open network or cage. The majorityof the pattern was completely undercut so thatit stood free from the main vessel (the cup),being held to it by a small number of glassbridges. For many years there was speculationas to how cage cups were made, and it wasnot until 1880 that the glasscutter Zweiselsucceeded in reproducing an example.Fremersdorf (1930) correctly described how thediatretarius had produced cage cups; and finallyin 1964, a German glasscutter, Schäfer (1968,1969), made a perfect copy of the cage cupfrom Daruvar (Zagreb, Yugoslavia). The illus-tration of a cage cup by Brill (1968) also makesit quite clear that a cutting technique had beenused in the example shown.

Two outstanding examples of cage cups area virtually complete diatreta in the RomischGermanisches Museum, Cologne; and theLycurgus Cup in the British Museum, London(Plate 4). The latter was skilfully carved in highrelief, the backs of the major figures beinghollowed out in order to maintain an eventranslucency. At one point (behind a panther),the vessel wall was accidentally pierced(Harden and Toynbee, 1959; Harden, 1963).

Surface decoration by cutting glass

Abrasive techniques for the decoration ofglassware have as long a history as glass-

making itself; there is evidence for glassengraving in Egypt as early as the sixteenthcentury BC. Engraving probably developedfrom the stone- and gem-cutting trade, beingexecuted with the same pointed instruments.Harden (1968) drew attention to simple formsof incised decoration in three different eras.Cut or engraved patterns were rare in the pre-Roman period; during the Roman period goodengraving declined in Egypt but excellentwork was produced in Syria (Harden, 1969a);and in the post-Roman era, engraved Christianmotifs began to appear on glassware (Harden,1971).

Diamond-point engraving is, as its namesuggests, the technique of using a diamond-point to scratch the glass surface. Earlyexamples are said to date from Roman times,although it has been argued that these wereexecuted with pointed flint tools. Diamond-point engraving was used to decorate Islamicglass. In the sixteenth century the technique wasused by glass decorators in Venice (Figure3.16), and in Hall-in-Tyrol (Austria), where anglasshouse under the patronage of the ArchdukeFerdinand produced a great deal of diamond-engraved Venetian glass; and by German, Dutchand British engravers, firstly in the façon deVenise and subsequently in local styles.

Diamond-point engraving was first used inBritain on glassware made in the glasshouseof Jacopo Verzelini. In Holland, during theseventeenth century, diamond point engravingwas mainly used by amateur decorators,especially for calligraphy (Tait, 1968). Thetechnique was superseded in the eighteenthcentury by wheel-engraving and by


Figure 3.16 Detail of a diamond-point engraved dish,with the arms of Pius IV. (c) V&A Picture Library.

enamelling. However, it continued to be usedto produce stipple engravings from the 1720sto the late eighteenth century, and in Britainfor producing Jacobite and other commemo-rative glassware. In the technique of stippling,grouped and graded dots were engraved witha diamond-point on the surface of the vessel.These represented the highlights of the design.The diamond-point was set in a handle, whichmay have been gently struck with a smallhammer to produce a single small dot on theglass. In the best examples of stippling thedecoration can be compared to a delicate filmbreathed upon the glass (see Figure 7.27e for a nineteenth-century example). FransGreenwood, a native of Rotterdam, broughtthe art of stippling to its greatest height in thefirst half of the eighteenth century. In the lastforty years of the eighteenth century, stippledengraving was practised by numerous artists,the most famous being David Wolff inHolland.

Formative and decorative wheel-cuttingand wheel-engraving

Wheel-cutting and wheel-engraving of glassare basically the same technique in which arotating abrasive wheel is used to cut into thesurface. In early times the same equipmentwould have been used for both methods:rotating wheels of various sizes propelled bya bow lathe. Gradually cutting came to becarried out with large wheels and was charac-terized by large-scale geometric designs,usually relatively deeply incised (see Figure7.26); whereas engraving was executed withsmall wheels and was usually the method usedto produce fine or pictorial work since it waspossible to achieve great detail (Figure 3.17;Charleston 1964, 1965). Rotary abrasion (usingemery powder or perhaps powdered quartzsand) had been carried out as early as theThird Dynasty (c. 2100–l800 BC), and engrav-ing tools of that date have been excavated, butfirm evidence for the rotary wheel-cutting ofnarrow lines seems to be scarce. It is notknown what type of abrading equipment wasused to decorate glass in Roman times. Thereis a suggestion that an all-purpose tool couldhave been used, which would be adapted asa lathe, drill or cutting or engraving wheel asrequired. The Romans used abrasives such

as emery for cutting and pumice stone forpolishing glass.

With the decline of the Roman Empire inthe West, glass engraving died out, principallybecause there was no fine quality glass madeafter this time which was suitable for engrav-ing. In the East the technique never ceased,and glasses with cut decoration can be tracedin continuity through Sassanian to Islamictimes. In the ninth and tenth centuries aschool of relief cutting flourished in Persia andprobably also in Mesopotamia, which was notrivalled until the end of the seventeenthcentury in Europe.

A description of wheel-engraving datingfrom AD 1464 states that a diamond-point wasused, ‘and with wheels of lead and emery; andsome do it with little bow (archetto)’. By thefifteenth century gem-cutting wheels of lead,pewter, copper, steel and limewood were


Figure 3.17 Detail of a wheel-engraved design on aJacobite wineglass bowl.

being used to engrave and polish glass. By thesixteenth century the principle of continuousrotary movement had been established. Adrawing dating from 1568 depicts a foot-operated treadle with a large flywheel. Wheel-engraving was to reach its greatest heightsonce Bohemia and Germany produced apotash-lime glass suitable for this technique inthe early seventeenth century.

Intaglio decoration (Ital. cave relievo; Ger.tiefschnitt) was created by engraving or cuttingbelow the surface of the glass so that theapparent elevations of the design werehollowed out, and an impression taken fromthe design produced an image in relief. Thebackground was not cut away, but was left onthe plane of the highest areas of the design(see Figure 3.17). Intaglio work was carriedout on rock crystal and other semi-preciousstones and on glass by wheel-engraving inmedieval Rome, but more particularly in glassin Germany and Silesia from the seventeenthcentury. For greater effectiveness the designwas left matt or only partially polished.

The opposite decorative technique wascameo work (Ger. hochschnitt; Eng. highengraving) which involved the formation of adesign in relief by cutting away the ground(Figure 3.18). The cameo technique, a combi-nation of wheel-cutting and engraving, wasused in Roman times. Cameo carving oflayered stones seems to have originated inPtolemaic Alexandria; and since Alexandriawas also one of the most important centres ofancient glassmaking, it is reasonable tosuppose that cameo glass was an Alexandrianinvention in imitation of gem-stones. Cameoeffects simulating agate and other similarlymulti-coloured stones were commonlyproduced by casing, for example, a colouredglass with a white one, followed by cuttingaway the white glass in low relief.

The earliest known example of cameo glassdesign points to the same conclusion. This isa fragment of a plaque in the Department ofEgyptian Antiquities of the British Museum,which shows a man’s leg and part of a bullin a purely Egyptian style, which cannot belater than the third century BC. The white-on-blue blank for the plaque must have beenmoulded, since glassblowing had not beeninvented; but the relief was unquestionablycarved. However, cameo glass that can be

certainly ascribed to pre-Roman times is rare,and the bulk of that which has survived proba-bly dates from the early Roman Imperialperiod. Two of the most notable specimens,the Blue Vase (National Museum, Naples) andthe Auldjo Jug (British Museum, London – see


Figure 3.18 An early eighteenth-century Silesian gobletand cover with an acanthus design cut and carved inhigh relief (hochschnitt). (© V&A Picture Library).

Figure 7.40), were found in Pompeii andtherefore pre-date the volcanic eruption ofMount Vesuvius, which destroyed the city inAD 79.

The first step in the manufacture of a cameoglass vessel such as the Portland Vase, was toblow the blue glass body and to coat it witha layer of white glass reaching up to a leveljust above the shoulder of the vase (Figure3.19). To achieve this the glass-blower mayhave dipped a partially inflated paraison ofblue glass into a crucible of molten white thusgathering a white layer over the blue; or a cupof white glass may have been formed and theblue glass blown into it. Having cased the blueglass with the white, the body was formed byfurther blowing and marvering. The averagethickness of the blue glass is 3 mm.

Considerable manipulative skill wasrequired to blow a two-layered vessel, but theglassworker’s chief difficulty was to preparetwo differently coloured glasses with the samecoefficient of thermal expansion (and thereforealso contraction) as an essential condition if

they were not to crack or split apart oncooling. The annealing process itself wouldhave had to be carefully controlled. When theglass body was complete, the handles wereformed from glass rods and attached, theirlower ends to the white glass covering theshoulder, and the upper ends to the blue glassof the neck. The vessel was then handed to aglass engraver to carve the frieze, cut theornament of the handles, and bring the wholeto a finish. No doubt the frieze was copiedfrom a model in wax or plaster, and theengraver would have begun by incising theoutlines of the design on white glass, then allthe white glass would have been removedfrom the background in order to expose theblue glass, and lastly the figures and featuresthereby left in block relief would have beenmodelled in detail.

The grooved treatment of the drapery androcks reveals the use of engraving wheels,which were probably used for grinding awaythe background and other relatively coarsework; but for the more intricate and delicatedetails, small chisels, files and gravers wouldhave been required (Haynes, 1975).

It was not until the nineteenth century thata glass cameo vessel such as the Portland Vasewas produced again. (The Portland Vase issuch an exceptional work of art that consid-erable effort has been devoted to making acopy.) In 1786 Josiah Wedgwood made aceramic copy in Jasper ware; and in the early1830s a prize of £1000 was offered for a copyin glass. This was achieved in 1876 by JohnNorthwood Senior, but, even using nineteenth-century techniques of glassmaking, there wasan expansion mismatch between the blue andwhite glasses, and the vase cracked after threeyears’ work had been spent on it (Northwood,1924). Other cased articles made byNorthwood have also shown evidence ofstrain between the layers of glass. However,cameo glass was popularized to such anextent that its production became commer-cialized by the end of the century. In order toincrease production cameo workers worked asteams rather than as individuals; and the largerareas of unwanted overlay were removed bydipping the vessel into hydrofluoric acid,having first covered the areas to be retainedwith a protective acid-resistant material suchas wax.


Figure 3.19 Stages in forming a cased (cameo) glassvessel. (a,b) A gather of transparent blue glass is blownand dipped into molten opaque white glass. (c) Thevessel is shaped. (d) The mouth, foot and handles areformed. (e) The design is engraved through the whiteglass. (f) The Portland Vase. Late first century BC orearly first century AD. H 245–248 mm.

(a) (b)(c)

(d) (e) (f)

Allied to cameo-cutting was the work of thediatretarius previously discussed, in which aglass blank was almost entirely undercut bylapidary means to produce the delicate decora-tive cage.

Cutting in high relief was used by theChinese on scent and snuff bottles, and in thenineteenth century for the decoration of glassin the Art Nouveau style which arose in the1930s, the name being derived from that of aParis gallery devoted to interior decoration. ArtNouveau was adopted as a movement inBritain by William Morris (1834–1896) and hiscontemporaries. The style was adapted forglassware by Emile Gallé (1846–1904) inFrance and Louis Comfort Tiffany (1848–1933)in the United States, much of the decorationresulting from cane cutting.

Facet cutting

It is evident that ancient glass-cutters knewhow to exploit the refractive effects of glassby cutting facets in addition to engraved lines.Harden (1969a) illustrates facet-cut bowls fromthe second century AD but states that by thefourth century the art of facet-cutting haddegenerated to a mere abrasion of the surfaceboth in the East and in the West. Some twohundred years later, in the Sassanian area ofeastern Iraq and western Iran, bowls wereproduced with regularly spaced deep circularor hexagonal facets. Another 600 years later inthe twelfth century, the Hedwig glasses(Pinder-Wilson, 1968; Harden, 1971), possiblyproduced in Russia, portray animals mosteffectively with large sweeps of the cuttingwheel. In contrast with the cutting of fine lineswith a narrow wheel, the use of a large andwide wheel seems, not surprisingly, to havebeen established much earlier.

It is only towards the end of the seventeenthcentury that a genuine distinction betweenglass cutting and glass engraving can be made.For the first time it is obvious that differenttypes of equipment for cutting and for engrav-ing were being developed. The glassengravers’ equipment was light enough at thisperiod to be carried, and resulted in a numberof travelling glass engravers who wouldengrave any design on the spot for customers.The most famous of these was Georg FranzKreybich who travelled Europe at the end of

the seventeenth century engraving glass. Onthe other hand, the glasscutters’ equipmentused for facetting, intaglio, deep cutting, orroughing out for finer engraving was hardlyportable. The large interchangeable wheelswere rotated on a heavy, hand-turned cuttingmachine, a form of equipment, which surviveduntil the modern period (Figure 3.20).

By the end of the seventeenth century water-power was in use for turning the wheels to cutglass. Water-power was probably used toenable all-over facetting to be carried out as anobligatory prelude to the engraving of potashglasses in Bohemia and Silesia in the eighteenthcentury. The highest development of facet-cutting, to produce designs in colourless glass,occurred with the cutting of full lead crystalglass. Cut glass has become synonymous withthe deep wheel-cutting used on Irish glass fromthe late eighteenth century onwards, and alsoon modern cut wine glasses and decanters(Warren, 1981). Cut glass products enjoyedgreat popularity in Britain in the laternineteenth century. In this style of glassware


Figure 3.20 Glass cutting and engraving. An engraverholds a goblet beneath a copper wheel powered by atreadle (omitted from the print). On the table areseveral other wheels, and bowls of abrasive paste. (©V&A Picture Library).

angular cuts were made into the vessel which,when polished, act as prisms with adjacent cuts,producing a brilliant effect. The glass blankwould first be marked with the pattern using amixture such as white lead and gum water.Following the design, deep cuts would beroughed in against an iron wheel fed withabrasive such as sand. For coarse work,overhand-cutting where the vessel was presseddown onto the wheel from above was usual,whilst underhand-cutting where the vessel waspressed up onto the wheel from below wasused for more delicate work. After the initialdesign had been cut, fine-grained, water-cooledstones, which required no abrasive, were usedto smooth the first cuts and to add finer lines.(Besides water-power, steam-powered cuttingmills were in existence by the beginning of thenineteenth century.) Finally the cuts could bepolished with lead or wooden wheels or withrotary brushes charged with tripoli or puttypowder. After the second half of the nineteenthcentury a method of plunging the glass vesselsinto a mixture of hydrofluoric and sulphuricacid was used to polish cut glassware.

Two techniques related to engraving aresand-blasting and acid-etching.

Sand-blastingThe technique of decorating glass by sand-blasting was invented in 1870, by the AmericanBenjamin C Tilghman. Some early sand-blasting was steam-powered, but this was soonsuperseded by compressed air power. Parts ofthe glass to be left plain are covered with astencil plate of steel or an elastic varnish orrubber solution painted on to form a protec-tive shield. A stream of sand, crushed flint orpowdered iron is then directed onto thesurface of the glass in a jet of compressed air.The type of finish is varied by altering the sizeof the nozzle directing the abrasive, the size ofthe abrasive, or the pressure of the air. Sand-blasting is normally carried out in a closedcircuit in a cabinet which can be sealed. Thetechnique has rarely been applied to vesselglass except for lettering on mass produceditems. The main use of sand-blasting has beenon glass panels for decorative architectural use.

Acid-etchingThe effects of abrasion can be duplicated, byetching glass with hydrofluoric acid (HF). A 90

per cent solution of HF possesses the uniqueproperty of dissolving silica:

SiO2 + 4HF – SiF4 = 2H2O.

The effect depends greatly on the strength ofthe acid, and whether sulphuric acid is alsopresent since some combinations of the acidmixtures polish the glass instead of producinga rough-etched surface (Schweig, 1973).Hydrofluoric acid is both difficult to prepareand extremely dangerous to use from thehealth point of view. It seems that the earliestexample of acid-etching was made bySchwanhardt at Nuremberg in 1670. Thisexample has been the centre of much specu-lation in the past since it was not until 1771that Scheele discovered hydrofluoric acid.However, it appears that the Nuremberg articlewas treated with a mixture of calcium fluorideand sulphuric acid which produced hydroflu-oric acid. The method of acid-etching gener-ally employed was to coat the glass with waxafter which the design was scratched throughthe protective layer onto the glass. The glasswas then placed in a bath of hydrofluoric acidand was only etched in the areas from whichthe wax had been removed. In France, Gallé,Daum and Marmot used the technique toproduce deep bold patterns on glass. Acid-etching is now used mainly to produce inter-esting surface textures and patterns onarchitectural glass. Glass can also be polishedwith a mixture of demineralized water,sulphuric acid and hydrofluoric acid.

Glass-blowing

It is remarkable that none of the contempo-rary Roman sources give any indication of afundamentally important occurrence in thefield of glassmaking, namely, the invention ofglass-blowing; nor of the great increase inoutput of glass vessels which was broughtabout by its use. Later authors (Augustinianand Tiberian) refer to the use of glass-blowingwithout saying when it was invented (Harden,1969a). Literary sources have traditionallygiven the birthplace of glassblowing as thePhoenician coast. However, on archaeologicalevidence it seems more likely that the inven-tion took place in the Aleppo-Hama-Palmyra


area of Syria, or in Israel. Grose (1977)reviewed the evidence for the introduction ofglass-blowing, and concludes that discoveriesmade in Israel in 1961 and 1970 point to glass-blowing having been invented there sometimein the period 50–40 BC.

Despite the fact that there are no examplesof blown glass earlier than this date, anancient Egyptian wallpainting depictingworkers blowing into tubes, was cited asevidence of glass-blowing. The depiction infact shows Egyptian metallurgists blowing acharcoal fire with reed tubes ending in claynozzles. Harden (1969a) suggested that solidmetal rods had long been used for gatheringmolten glass from crucibles to introduce it intomoulds, and that it was but a short step tousing a hollow tube to inflate the glass into amould by blowing. Schuler (1959b), however,believes that free (off-hand) blowing precededmould-blowing since more skill was requiredto handle a blowpipe in the vertical positionrequired for mould-blowing.

Remains of glasshouses have been uncoveredat Rishpon near Herzlia, Tiberias, Sussita(Hyppos) and in several places along the sea-shore in Israel. During the excavation of a firstcentury AD cemetery at Acre a group of glasses(mainly piriform unguentaria) were found.These may have been the product of a localfactory using the glass sand of the Belus district,which is known to have been shipped south,even to Alexandria. At Bet She’arim (SheikhAbreiq) near Nazareth a glasshouse was foundwhich was working as late as the second andthird centuries AD, and which produced veryaccurately shaped glass vessels. Many claymoulds were also found on the site. A contem-porary glasshouse at Sussita on Lake Tiberiasyielded a large quantity of glass vessels andwasters (Forbes, 1966). The thinnest and mostbeautiful glass was produced at Tiberias.

The blown glass of this period is referred toas Sidonian (a term comparable to the muchlater façon de Venise), since it was a traditionof blown glass shapes and techniques used bySyrians and Jews, which were handed fromfather to son, a training which was encouragedin the fourth century by exemption from taxes.The ancestors of the Jewish glassmakers mayhave learned something of glass manufactureand glass painting during the Babylonian Exile(seventh to fourth centuries BC).

Sidonian (Syrian and Jewish) glassmakersexploited the glass-blowing technique, migrat-ing west and northwards and establishingglasshouses abroad. Mould-blown glass wasbeing made in Italy by the first century AD,and in the Alpine provinces and north-westernEurope by AD 40–50. Glasshouses werefounded in the first century AD in Cologne,while the north Gallic and Belgic glasshouseswere probably in full production by thesecond century. Their free- and mould-blownproducts were of a style directly related to theproducts of the central and eastern RomanEmpire. The Egyptian workers based atAlexandria continued to specialize in cuttingglass and in producing fine-coloured waressuch as mosaic bowls, and did not adopt theglass-blowing technique until some time afterthe second century AD.

Glass-blowing techniques

The invention of glass-blowing brought abouta complete revolution in the manufacture ofglass artefacts. Glass ceased to be a luxurymaterial since scores of vessels could be blownin a day in contrast with the use of the labori-ous core-forming process. Another great advan-tage was the possibility of making articlesthinner and lighter and thus more acceptablefor domestic purposes than was generallypossible using the earlier glassmaking pro-cesses. There was also a great increase in thevariety, shape and number of small hollowarticles, and for the first time larger containersand window-panes could be produced.

From the very beginning, many of thevessels have the appearance of having beenripped off the blow iron as fast as they wereproduced. Handles and ornamentation alsohave the appearance of having been trailed onquickly. Speed of production gave early blownglass a pace and spontaneity which had previ-ously been unknown. The shapes quicklygrew more composite and within a fewcenturies glassware stood on the tables ofordinary citizens. At Karanis in Egypt a housedating from this period was excavated andyielded no less than eight oval glass dishes,sixteen bowls, five conical lamps, two drink-ing cups, two jars, two flasks and two jugs. Atfirst blown-glass flasks were, like their fore-runners, used for perfumes and other such


luxury items and, like the earlier glass vessels,were transported to southern Russia, Gaul orGermany packed in plaited straw covers likethe modern chianti bottles. Bottles in this typeof packing are shown in a mosaic at El Djem(Tunisia).

The modern process of glass-blowing isdescribed below but it is likely that the ancientprocess differed little from this, except that theblowpipe may have been simpler, and thefurnace would have been far less sophisti-cated.

The glassworker’s blowpipe is a steel tubeabout 1.5 m long, tapered to a mouthpiece atone end. The gathering head is made bywelding on a 100 mm length of thick-walledtube, made of wrought iron (or other alloyresistant to oxidation) having a diameter from25 to 80 mm, the wider tubes being used formaking the heavier ware. Several pipes areused by each chair or shop (a team of workersworking from one pot, producing hand-blownglassware) so that a clean pipe at the righttemperature will always be ready for thegatherer who starts the cycle of operations(Figure 3.21).

Molten glass will wet an iron rod which ishot enough to prevent the formation of a skin

when it touches the glass, and hence thegatherer can collect from the pot a quantity ofglass (the gather) whose size depends on thediameter of the pipe head, the temperature ofthe glass and the number of turns given to thepipe. Before blowing starts, it is necessary toform a skin of chilled glass (or the gather willperforate at the hottest point) and to makesure the glass is distributed evenly around thepipe. This is carried out by marvering, orrolling the glass on a hard, smooth surface,probably of stone (and in later times iron).

Large gathers, especially those made bydouble gathering, are shaped by blocking orturning the gather in a shaped block of wetwood (usually pear wood); the steamproduces an air-cushion which prevents thewood from burning. The gatherer will nowpuff or blow lightly down the pipe to form asmall bubble of air and produce an internalviscous skin; the still molten glass between thetwo skins can now be manipulated to producethe desired distribution of glass needed in thefinal article. After puffing down the pipe, athumb is immediately placed over the orificeand the air is allowed to expand until thecorrect size bubble (Fr. paraison or parison)is produced.


Figure 3.21 The interior ofthe cone at the Aston FlintGlass Works, Birmingham.Cones were spacious andprovided good workingconditions for the glassblowers, working around thecentral furnace. Such furnacesnormally held between eightand twelve pots. The furnacefor pre-heating pots (the ‘potarch’) can be seen in thebackground on the right, andthe annealing tunnel on theleft.

Many glass articles are produced off-hand,that is, without the use of moulds. The blower(gaffer) sits at a glassworker’s chair, a specialseat invented in the mid-sixteenth century,which has long arms on which the blowpipeis rolled whilst the glass is being shaped (seeFigure 3.21). The term chair has beenextended to describe the team for which thechair is the centre of operations (Charleston,1962).

Manipulation of the glass was a naturaldevelopment of glassmaking once a bubblehad been blown. By reheating and blowingalternately, and by holding the blowpipeabove or below his head, the glassworkercould control the shape and thickness of theblown vessel. If the glass-blower was trying tomake a vessel to a preconceived design itrequired a much greater degree of skill toachieve it by free blowing than by the use ofa mould carved to the desired shape of theartefact. However, free blowing allowedspontaneity of design and it was perhaps forthat reason that its use continued throughoutthe history of glassmaking up to and includ-ing the present day.

In the blowing of exceptionally large vesselsa subterfuge could be used. Having blown abulb, the worker took a mouthful of water andejected it down the blowpipe quickly puttinghis thumb over the mouth-piece. The steamgenerated inside the glass bulb then continuedthe blowing process until it was released.

By successively reheating the article in aglory hole (a small furnace or an openingwhich leads to the hot interior of the furnace)a gaffer could manipulate the body of thevessel into a variety of shapes before the metalcooled; adding extra small gathers (brought tohim by a servitor) to form stems, handles,bases etc., using basic tools (see Figure 3.22).

The marvering surface itself could incorpo-rate shaped hollows in order to producedefinite patterns, or the glass could be shapedwith the aid of a battledore (Fr. palette), a flatwooden board or even with a wad of wetnewspaper. Tongs could be used to produceraised ridges and knobs by pinching outportions of the vessel’s walls, or the pointedend of a reamer (a flat-bladed tool with apointed end) could be used to produce ribsand furrows in the glass. It seems probablethat the earliest decorative marks were madeon glass at this stage, that is, before the moltenglass had completely hardened.

When the main body of the article had beencompleted a solid iron rod, the punty (origi-nally pontil), which had been heated andlightly coated with glass, was attached to thecentre of its base. The blower then wetted theglass near the end of the blowpipe or touchedit with cold metal so that the glass vesselbroke away from the blowpipe and wassupported only on the punty iron. The opentop was then softened in the glory hole, cutoff at the desired height with shears and, if


Figure 3.22 The tools ofthe glassmaker, taken fromDe Arte Vitraria, Neri–Merrett.First Latin edition,Amsterdam, 1662.

required, further opened or reduced withtongs (Ital. pucella). The vessel was thencracked off the punty iron and carried awayfor annealing. Subsequently, when cold, thescar from the punty (the punty or pontil mark)may be largely or even entirely ground awayand polished. (Glass which has been madeoff-hand has a fire-finished surface which isinitially more durable than the surface formedagainst a mould because the surface becomesalkali-deficient when exposed to the flames ofthe glory hole; Bruce, 1979.)

Wine bottles were most commonly made byblowing a bulb, marvering it into shape, andthen pushing in the closed end to produce thecharacteristic dimple foot, so carefully repro-duced in the modern machine-made bottles.The bases of the bottles were substantiallythicker than their necks. In blown glass gener-ally there is a tendency for the material to bethicker at the furthest point from the blowiron.

Mould-blowingThe technique of blowing glass into mouldsprobably developed simultaneously with thediscovery of glass-blowing. Once a gather ofglass had been taken from the furnace on theend of a blow-iron, it was marvered, blowninto a small, elongated bulb and introducedinto a two-piece mould, then blown to fill it.Until modern times the moulds would havebeen made of clay or wood (the latter waskept wet during use to prevent the wood fromburning) (Figure 3.23).

Pattern mouldingWhen glass was blown into a mould with apattern carved on its interior surface, thepattern was impressed in reverse on the glass(see Figure 3.23). Designs on a pattern-moulded vessel could be varied once thevessel had been removed from the mould byblowing or by twisting or swinging the viscousglass on the end of the blow-iron. Forexample, designs such as ribbing on the sidesof a blown vessel were produced by blowingthe glass into a ribbed mould. If the mould-blown vessel was removed from the mouldand then gently reblown after reheating, thethicker glass between the ribs, being hotterthan that forming the ribs, expanded so thatthe ribs were pushed to the interior surface of

the vessel instead of remaining on the outersurface. Glass which has been mould-blownand then additionally blown to increase thesize of the object or to soften or modify thelines of the pattern by blowing into a plainmould is sometimes now referred to as havingbeen optic blown.

The process of moulding glass lends itselfto considerable variation, including theproduction of highly decorative effects such asthe inclusion of rows of internal air bubblesin various patterns such as spirals (see Figure3.24), and the incorporation of white andcoloured threads of glass.


Figure 3.23 Mould-blowing techniques. (a) A gather ofglass. (b) Free-blown and shaped by marvering. (c) Thebase indented. (d) The spout and handle formed. (e)Blowing into a two-part mould. (f) Blowing into apattern mould. (g) Introducing air twist into a stem bypressing spikes into the base of a blown gather, or (h)dipping it into a mould containing spikes. (i) Drawingthe gather off the spikes. (j) Twisting the stem.

(a) (b)

(c)(d)

(e) (f)

(g) (h)

(i)(j)

Air bubbles and air-twistsA highly decorative effect was produced in thestems of British drinking glasses by twisting arod of glass in which were embedded threadsor tapes of opaque white or coloured glass (asdescribed above), or air bubbles and air-twists.The technique of deliberately incorporating airin glass stems dates from circa 1735, and waspopular from the 1740s until the 1760s. It hasbeen stated that there are over 150 varieties indifferent forms and combinations. Numerousexamples are illustrated by Bickerton (1971).

Air-twists were produced by two basicmethods: by making slots in a rod of glass,then drawing the rod until it became thin andtwisting it to make the columns of air spiral;or by moulding a pattern of circular holes orflat slits in the top of a glass rod, coveringthem with molten glass, and then drawing andtwisting the rod to make a spiral pattern of air(used to make multiple series twists). Aquarter twist in a four-column stem wouldappear to produce a complete twist of 360°.In the early examples the twist was irregularlyformed and spaced, but later the threadingwas uniform; it was sometimes carried downfrom the bowl of a two-piece glass. Anotherprocess involved placing several rods contain-ing elongated tears into a cylindrical mouldwith grooves on its interior surface, thencovering them with molten glass and, afterwithdrawing the mass on the punty, attachinganother punty and twisting until the desiredpattern was produced. In some examples onetwist is concentrically within another (made byrepeating the process with the mould, thustwice twisting the inner rod into a tightertwist); these are termed double series air-twistor even triple series air-twist. Occasionally anair-twist stem includes one or more knops,and the twist continues unbroken (butsometimes slightly distorted) through theknops (Figure 3.24).

Paperweights or other solid glass objectsdecorated with a pattern of regularly spacedair bubbles, were made in a spot-mould. Themould was sectional with small spikesprotruding in the interior. When the paraison(sometimes already having an interior decora-tive motif) was introduced, the sectional partsof the mould were tightened around it forcingthe spikes into the glass. In this way smallcavities were formed so that when the piece

was cased, air bubbles were trapped thusforming a pattern (Newman, 1977).

Casing and cuppingCased glassware is made of two or more layersof differently coloured glass. To achieve thisthe glassblower either dipped a partiallyinflated mass of glass (paraison) into acrucible containing glass of a different colour,thus gathering it over the top (casing), orformed a cup of glass into which the secondglass was blown (cupping). Having cased (orcupped), for example, a blue glass with awhite glass, the body of the vessel would bebrought to the required size and shape byfurther blowing and by rolling on a metalmarver. (In making flashed window glass, thecoloured glass is usually gathered first, termedthe post and then the colourless glass isgathered over the top of it.) When the glass


Figure 3.24 Detail of an air twist in a wineglass stem.

had been annealed, the outer layer of glasswould be carefully abraded away to form adesign in which the underlying glass wasexposed or seen in varying degrees throughthe outer layer. An early example of casedglass dating from the first century AD is thePortland Vase (see Figure 3.19).

Technological features of the Portland Vasehave attracted much attention since it was firstdiscovered (probably in 1582). For instance,there is a complication in its supposedmanufacture by blowing since the opal whiteglass casing only extends part way up thetranslucent blue glass vessel; and there are twosubstantial blue glass handles added over thewhite layer. The coefficients of expansion ofthe blue and white glasses apparently matchedeach other so perfectly that there is noevidence of strain in the vase (Bimson andFreestone, 1983).

FiligranaFiligrana (Ital. vetro filigranato) is the Italianterm which has been applied to glass artefacts,originally made in Murano circa 1527–1529, ofclear glass with various styles of decorationproduced by embedding threads of solidopaque white-glass (lattimo) forming latticino(or latticinio); or of coloured glass (or evenoccasionally a single white thread). It is nowused to refer to all styles of decoration onclear glass made by a pattern formed byembedding threads of glass, including vetro afili (threaded glass), vetro a reticello (glasswith a small network formerly termed a redex-elo and a redexin in Murano) (Ger. Netzglas),and vetro a retorti (or retortoli), originallymade in Murano but now termed zanfirico orsanfirico in Venice and Murano. Vetro de trina(lace-glass) is a term which has been usedloosely to describe various types of filigranaglass, and is now considered to be superflu-ous and of no historical significance. ModernVenetian glassworkers find it hard to achievethe accuracy and lightness of design infiligrana of their predecessors.

In vetro a fili the opaque white and/orcoloured glass threads are embedded in clearglass in continuous lines without any crossingof the threads, the lines being in a spiral orhelix pattern (e.g. on plates) or in a spiral orvolute pattern (e.g. on vases). In vetro areticello, the threads are embedded in clear

glass in the form of criss-cross diagonalthreads forming an overall diamond latticenetwork. There are three separate varieties,depending on whether the pattern is madewith fine threads, coarse threads, or fine andcoarse threads running in opposite directions.As the threads protrude slightly, tiny airbubbles (sometimes microscopic) are en-trapped within each criss-cross diamond, thesize depending on the process of production.On a few rare examples there are, instead ofthe air bubbles, thin wavy lines running inonly one direction between the rows of whitethreads. The style has been used on vases,bowls, jugs etc. and also on plates (where thenetwork often becomes distorted towards thecentre or the edge of the plate). Severalprocesses have been documented, and inMurano, according to local glassmakers,several methods have been used (Figure 3.25);the first two methods detailed below are themost authoritatively stated:

(i) A bulb (or cylinder) of blown glass in amould, on which have been picked up ona gather parallel threads of glass (almostalways lattimo) running diagonally in onedirection (resulting from twisting the parai-son after gathering the threads) and whichis then blown into another similar bulb (orcylinder) with threads running in theopposite direction; the two bulbs (or cylin-ders) become fused together. The difficultyarises from the necessity of having thethreads on each bulb exactly equidistant sothat they will form equal and similardiamonds (except for distortions towardthe extremities of the glass).

(ii) A bulb of glass is made, as describedabove, with diagonal threads, and then halfof the bulb is bent into the other half andfused, thus making the criss-cross threads.This method assures equal spacing, but isa difficult process.

(iii)A bulb of glass is made, as describedabove, with diagonal threads, and then theglassblower sucks on the tube andcollapses the further half of the bulb intothe nearer half making a double wall withcriss-cross threads.

(iv)A bulb of glass is made, as describedabove, with diagonal threads, then it is cutand twirled to make, by centrifugal force,


a flat open plate; another such plate ismade with threads running in the oppositedirection, and then the two plates are fusedtogether by heating. This method alsopresents the difficulty of obtaining equalspacing of the threads on both plates.

In all these methods there develop, withinthe interstitial spaces made by the crossedthreads, small (sometimes microscopic)bubbles, but occasionally they are elongated

into thin wavy lines, and the enclosed areassometimes have unequal sides. After the pieceof glass with the crossed threads is made it isreheated, then blown and manipulated intovarious forms such as vases, plates and jugs.

Vetro a retorti is the Italian term applied toa style of decoration with parallel adjacentcanes (vertical or spiral) of glass havingembedded threads in various intricate patterns(as in the stems of some British wine glasses)and made by flattening the canes and fusing


Figure 3.25 Stages informing a filigrana vessel,and (o–p) details of thefiligrana patterns. (a) Row ofcoloured glass canes, (b)picked up on a gather ofglass or arranged around thesides of a mould and (c)picked up on a gather ofglass. (d) The free endsgathered in onto an iron, and(e) drawn out and twisted.(f) Rows of twisted, colouredglass canes picked up on (g)a flattened gather of glass.(h) The free ends attached toan iron and drawn out. (i)The bottom cut off withshears. (j) After the base ofthe bowl has been shaped,the glass forming the knop isattached and (k) shaped. (l)A clear glass foot is attached,and the top of the vessel issheared off. (m) The mouthof the vessel is widened androunded off. (n) Thecompleted vessel is detachedfrom the iron and annealed.(o) Detail of Vetro a retortishowing parallel canes withtwisted designs. (p) Detail ofVetro a reticello showingcrossed opaque white glassthreads and entrapped airbubbles.

(a)(b)

(c)

(d)(e)

(f)(g)

(h)

(i) (j) (k)

(l)

(m)(n)

(o)(p)

them together. It includes decoration madewith opaque white or coloured glass threadsor both embedded in the fused canes. Thestyle has been used for plates, bowls, vasesand, occasionally, for paperweights (Newman,1977).

Sand-mouldingSand-moulding invented in 1870 is a specialform of moulding used to make glass linersfor silver vessels. A wooden block is carvedto fit exactly into the metal vessel, thenremoved and forced into a bed of damp sand.It is carefully withdrawn so that its exactimpression remains in the sand. A gather ofglass is then blown into the impression toform the glass liner. When cool it is abradedand polished.

Lamp-working (at the lamp; at theflame)

Lamp-working is the technique of manipulat-ing glass at the lamp by heating it with a smallflame (Figure 3.26); and was probably discov-ered in the Roman era. Examples of piecesproduced in this way range from small figuresand objects to large composite three-dimen-sional scenic groups.

The technique of Verre de Nevers is closelyallied to lamp-work. Small figures made of

opaque fusible glass, were produced at Neversand elsewhere in France, in the late sixteenthand early seventeenth centuries. Examplesvary from 25 to 150 mm in height and are verydetailed. Single figures often have stands oftrailed glass threads; a figure without such astand has usually been broken from a grotto(tableau). Some animals are mould-blown,with applied thin glass threads (Fr. verrefrisée).

Occasionally large groups were made, suchas a crucifixion scene or several figures withanimals. They were made with portions ofglass rods softened at the lamp and thenmanipulated with pincers or other instruments,and often fastened on an armature of copperwire. The names of some artists are known,including Jean Prestereau (1595) and his sonLéon.

Such ware, being made of white opaquematerial, might be mistaken for porcelain orfaience. Most examples display coloureddetails. Similar figures were made in Venice,Germany, England and Spain in the seven-teenth century and later; they are not readilydistinguishable or dateable.

Decoration of glass artefacts

Since glass production began, glassmakers anddecoratori have sought to improve simpleglass shapes. The numerous techniques usedcan thus be divided broadly into two groups:those produced by the glassmaker whilst theglass was still hot; and those produced by thedecorator on cold glass.

Glassmaking techniques of decorationinclude colouring (and decolourizing), freeblowing and shaping, including special effectsproduced by tooling and marvering, and bypattern moulding. Embellishments producedby the glass decorator include cutting, engrav-ing, cold painting, gilding and lustre andenamel painting. Lustre and enamel paintingand some forms of gilding, however, requirethe use of a kiln to complete the process.

Glassmaking techniques

Colouring, free blowing and blowing intopatterned moulds have already been discussedunder methods of glass vessel production.


Figure 3.26 Lamp-worked figures in the tradition ofJaroslav Brychta. Made by Jaroslav Janus, Zeleny Brod,Czech Republic.

Tooling the glassIt seems probable that tooling hot glass wasone of the earliest forms of decoration. Someearly Islamic glass bowls and saucers bearpatterns such as bulls eye circlets, rosettes etc.which were impressed on their sides by meansof patterned tongs whilst the glass was stillhot. Cylindrical bowls of the same period bearother designs such as lozenges and small birdsimpressed in the hot glass by shaped pincers(Harden, 1971). Pincering was a fairly generaltechnique (Figure 3.27)

Addition of glass blobs and trailsThe addition of glass blobs and trails has beenused since the late second millennium BC fordecorating core-formed vessels with lines,dots, rings and hieroglyphic inscriptions(Harden, 1968). After trailing on the design ofhot glass, probably from a metal rod, the trailswere marvered flush with the surface,although in some areas which would be diffi-cult to marver such as the foot or neck, thetrails remained in relief. The Romans wereparticularly fond of trailing in relief. Threadsmarvered into the glass body could becombed by dragging a pointed tool acrossthem before the trails cooled.

Newman (1977) refers to a rare type ofdecoration (underglaze) where the colouredtrails have themselves been covered withtransparent glass, as in the case of the hollowfish-shaped container dating from theEighteenth Dynasty (1567–1320 BC) and nowin the Brooklyn Museum, New York. Trailing

was especially popular in the sixteenth andseventeenth centuries AD and is still carried outin modern glassmaking. Allied to trailing is theintricate stem work so popular with Venetianand façon de Venise glassmakers in the seven-teenth century, though in these cases the trailsare mostly free-standing.

Another relatively simple form of applieddecoration is the use of blobs and prunts onglass vessels. Sometimes while still hot theblobs were stamped into designs such asraspberry or strawberry prunts with a metaldie. The addition of prunts to a vessel besidesbeing decorative also helped it to be heldmore easily in the absence of a handle. Afurther development was the application ofhot blobs of glass to a vessel while it was stillon the blowpipe. The vessel was then reblownso that the hot blobs blew out further than thecooler walls of the body. These were thendrawn out with pincers, and blown at thesame time to keep them hollow, thenreattached lower down the vessel (Harden etal., 1968).

The earliest prunted beaker seems to be thedolphin beaker of the fourth century, but theybecame common in Saxon times (clawbeakers) (Harden, 1971), and particularly so inthe fifteenth century (Harden, 1971; Tait,1968). In the daumenglaser from Germany thistechnique has been put into reverse. After theblobs were applied to the heated body of the vessel, the glassmaker sucked through theblowing iron so that the blobs were made toextend into the interior of the vessel, forminghollow finger-grips on the outside (Newman,1977).

Other forms of applied decoration owemore to the skill of the artist trained in othermedia than to the skill of the glassmaker.

Special effects produced by marveringA variety of special effects can be producedby spreading various materials (such aschopped coloured glass rod, glass powder,chalk etc.) on the marver. The first gather isformed, marvered so that the powdered mater-ial is incorporated in the surface, and then asecond gather is made. The chopped colouredrod spreads out like coloured worms, and thepowdered chalk decomposes to produce parti-cles of lime and copious bubbles of carbondioxide. Shapes resembling petals can be


Figure 3.27 Examples of hot-worked (tooled) glass:leaves and canes.

produced by making several gathers, spread-ing chalk on the top of each, and pressing ina spike so that all the layers are depressed andtrumpet-like shapes are produced (Newman,1977).

Ice-glassIce-glass (cracked glass, Italian, vetro a ghiac-cio), which has a rough irregular outer surfaceresembling cracked ice, was first produced inVenice in the sixteenth century, by twoprocesses: plunging the partially blown gather-ing of hot glass momentarily into cold waterand immediately reheating lightly so as not toclose the cracks caused by the sudden cooling,and then fully blowing to enlarge the spacesin the labyrinth of small fissures (Figure 3.28);or rolling the hot glass on an iron marvercovered with small glass splinters (sometimescovered) which adhered to the surface andbecame fused when lightly heated, a processwhich also removed the sharp edges.

Ice-glass was produced in Liège, Belgium,and in Spain in the seventeenth century. It wasrevived in England in about 1850 by ApsleyPellat, who called it ‘Anglo Venetian glass’. Innineteenth-century France the first methoddescribed above produced a glass called verrecraquelé, and the second method an effectcalled broc à glaces. A third method developedin modern times by Venini involves the use ofhydrofluoric acid and produces vetro corroso.Ice-glass has been produced in the United

States since the nineteenth century, where it istermed overshot glass.

Davenport’s Patent glassThis was decorated with the intention ofimitating engraving or etching although itresembled neither. The process was patentedin 1806, by John Davenport of Longport,Stoke-on-Trent, an English potter and glass-maker. The process involved covering theouter surface of the glassware with a pastecontaining powdered glass, then removing thesurface paste so as to leave the intendeddesign and quickly firing at a low temperatureto fuse the glass powder onto the surfacewithout melting it. The designs were oftenheraldic insignia and sporting scenes. Suchglassware was usually inscribed Patent on alabel made and affixed by the same method.

Decorating the surface of cold glass

EnamellingEnamelling is the process of decorating thesurface of glassware by the application of avitreous material, coloured with metallicoxides, to the surface, after which the enamelis fixed by low temperature firing in a mufflekiln. Enamel decoration is in the form ofscenes, figures, inscriptions and heraldry.

Enamel colours are metallic oxides mixedwith a glass frit of finely powdered glasssuspended in an oily medium (formerly honey)


Figure 3.28 Detail ofVenetian frosted or ice glass,created by plunging the hotglass into water for amoment and immediately re-heating. The roughenedsurface, resembling ice,became popular in Venicein the sixteenth century andspread to Northern Europe,where it remained in vogueinto the seventeenth century.

Fine shallow cracks, where thesurface has not been verystretched

Deep cracks in a rough surfacewhere the glass has been greatlystretched by blowing

for ease of application with a brush. Thecolours are applied to the surface of the glassartefact and fixed by low temperature firing, inthe range 700–900°C, in a muffle kiln, duringwhich the medium burns out. A flux mixedwith the enamel colours lowers the firing pointto below that of the glass to which they havebeen applied. Enamel colours differ from coldcolours in that they are fixed to the glasssurface and are therefore not so readily affectedby wear; the firing results in a smooth surfaceonly slightly palpable to the fingers; and thefinal colours of the enamels are not oftenapparent upon application but only after firing.

In gilt enamelling the decoration was firstoutlined with pen and brush and fixed in thekiln. Then the colours were spread on theoutlying areas and the vessel fired a secondtime at a lower temperature. The vitreouscolours are opaque and cover the surface ina thick coating. The numerous gilt designswere enclosed in red enamel lines for addedemphasis. The colours used for enamellingmodern glass tableware are usually based onglasses that contain substantial amounts oflead oxide, in order to reduce the fluxingtemperature (e.g. 590°C) to below that of thesoftening point of the glass, and to produce aglossy finish on the fired enamel.

Enamelling was used extensively in Venicefrom the fifteenth century and elsewhere inEurope from the sixteenth century onwards.The technique was used in China in theeighteenth century.

The process of enamelling glass is atechnique that can be dated back to thefifteenth century BC: a small Egyptian jugbearing the name of Tuthmosis III (c.1504–1450 BC) was decorated with powderedglass fired to the body of the vessel (seeFigure 2.1). Some Roman glassware wasdecorated with enamel painting, which wascarried out both in the East and in Italy in thefirst century AD (Harden, 1969a). Vesselsdecorated in Egypt tended to feature human,animal and plant motifs; whereas those depict-ing wild beasts in the arena or in gladiatorialcombat seem to have been made in theRhineland. Italian glassware from the sixth toseventh centuries returned to trailed decora-tion (Harden, 1971) or had marvered splashesand blobs of white, yellow, or red glass allover their surfaces.

Islamic enamelling, consisting of opaquevitreous enamelling (with gilding) as decora-tion on glassware was made between 1170and 1402, when Tamerlane (Timur) sackedDamascus and the local glass industry wasmoved to Samarkand, and some glasswareassociated with Fustat (Egypt) said to datefrom about 1270 to 1340. It is of five tenta-tively identifiable groups based on the criteriaof style and date rather than any provenconnection with the places of productionidentified with the groups, that is Raqqa, Syro-Frankish, Aleppo, Damascus, Chinese-Islamicand Syrian glassware (see Figure 2.11).

In the third quarter of the fifteenth centurythere was a phase in Venice when gorgeouslydecorated ware was made (Harden, 1971).Enamelling was executed at Murano on clearcoloured glass and on opaque white glass.From the late fifteenth century enamelling wascarried out in a manner similar to that ofIslamic glassware of the thirteenth andfourteenth centuries, the technique being saidto have been re-invented by Anzolo Barovier.It was first used on sapphire blue glasswareand other coloured transparent glassware, butin Venice, towards the end of the fifteenthcentury, coloured glass tended to be super-seded for this purpose by colourless glass. Bythe end of the sixteenth century enamellingwas little used there except on opaque whiteglass, which again became popular in theeighteenth century.

The earliest German enamelling on glasswas in the second half of the sixteenthcentury, copying the Venetian enamelledarmorial glasses that had previously beenimported. The German enamelling is found onthe typical German humpen and stangenglas,where the style evolved to cover almost theentire surface with painting subordinating theglass itself. Bright colours created the decora-tive effects, but the painting was not ofsuperior quality. The early motifs were coatsof arms, followed by a great variety ofsubjects, for example, painting of a religious,allegorical, historical or scenic nature, orshowing artisan or guild activities, or satiricalor family scenes, usually with dates or longdatable inscriptions added. The place ofproduction is usually unidentifiable, except insome special types such as the Ochsenkopfglas(Franconia), Hallorenglas (Halle) and


Hofkellereiglas (Saxony or Thuringia). In theseventeenth century enamelling became morerestrained and skilfully executed, with theartists usually signing the pieces. A new styleof enamelled decoration, Schwarzlot, wasintroduced by Johann Schaper and his follow-ers, Johann Ludwig Faber, Abraham Helmbackand Hermann Benckertt. The use of opaqueenamel was superseded by transparentenamel, introduced in about 1810 by SamuelMohn and also used by his son GottlobSamuel Mohn and Franz Anton Siebel. In the1750s some enamelling was carried out onopaque white glass.

The earliest mention of enamelling on glass-ware in England is a notice by a Mr Grillet in1696, but no examples are extant. Enamellingwas introduced by artists from Germany andthe Low Countries, and was of the type knownas thin or wash enamel, more suitable for thesofter English lead glass, and was executedwithin outlines previously etched on the glass.Later dense enamel became the usual medium,first with designs of festoons and flowers, butlater landscapes, figure subjects and coats of

arms. It was in this period that the members ofthe Beilby family, William and his youngersister Mary, produced their remarkableenamelled ware. They always worked as a pair(in the period 1762–1778) and their associationhas been claimed as one of the greatest in thehistory of artistic glass, their fame bringingcommissions from many noble families inEngland. William and Mary Beilby, and possi-bly Michael Edkins, were the leadingenamellers of the eighteenth century; theydecorated on clear glass and opaque whiteglass, respectively (Figure 3.29). Some warewas enamelled by glasshouses in Stourbridge inthe nineteenth century, and possibly by Londonjewellers and silversmiths doing work on glassboxes and scent bottles, as well as by artistswho decorated porcelain in imitation of Sèvresporcelain and who did similar work on glass-ware in the mid-nineteenth century (Charleston,1972). (See Painted and stained glass windows.)

Cold paintingCold painting (Ger. Kaltmalerei; Ital. di pintofreddo) of glass dates back to the Roman


Figure 3.29 TheBeilby family,working inNewcastle uponTyne, wereresponsible formuch of theenamelled glassmade in Englandfrom the 1760s.William Beilby hadtrained as anenameller inBirmingham. (©V&A PictureLibrary).

period but, lacking the durability of firedenamels and therefore being easily rubbed off,it is a poorer form of coloured decoration(Figure 3.30). In this process lacquer coloursor oil paint were applied to the surface ofglassware. Cold painting is particularly effec-tive when applied to the back of the surfacethrough which it is to be viewed, andprotected by a layer of varnish, metal foil, orby another sheet of glass (Ger. Hinter-glasmalerei). The process was sometimes usedon humpen and other large glasses ofwaldglas. This was possibly because theglasses were too large to be enamelled andfired in a muffle kiln; because the glass wasnot sufficiently durable to withstand firing; orbecause the painting was carried out bypeasants, without access to a furnace.

In the nineteenth century there was a greatinterest in the leisure-art of painting scenes inreverse on the backs of glass plates, sheets of

glass, mirrors etc. so that the design wasprotected by the glass when viewed from thefront (Figure 3.31). There was a brightluminosity in such paintings since there wasno air gap between the paint and the glass(Stahl, 1915; Newman, 1977; Bretz in Lanz andSeelig, 2000). The development of reversepainting on glass is given in Chapter 2.

Lustre paintingLustre painting is the term given to the processof applying metallic oxide pigments to theglass surface and firing under reducing condi-tions to produce a metallic iridescent effect (alustre) (Lamm, 1941; Newman, 1977). Theprocess could be used to produce a groundcompletely covering the surface, or simply toproduce a design. Oxides of gold, copper orsilver (and, in modern times, platinum andbismuth) were dissolved in acid and, afterbeing mixed with an oily medium, were


Figure 3.30 The Daphne Ewer, cold-painted andgilded. H 222 mm. Circa late second or early thirdcentury. Roman Empire, possibly Syria. (CorningMuseum of Glass).

Figure 3.31 Artist copying a European print on glass.Water-colour on paper. China (Canton). Circa 1770. 34.1� 41.5 cm. (© V&A Picture Library).

painted on the glass. Firing in a reducedatmosphere, smoky and rich in carbon monox-ide, caused the metal to fuse into a thin film,producing a non-palpable, evenly distributedmetallic flashing. Gold and copper yielded aruby colour and silver a straw yellow colour.This style of decoration appears on Islamicglass from the ninth to the eleventh centuries;most specimens having been found in andattributed to Egypt, where the technique isthought to have originated. Examples of lustre-painted glass in Egypt are dateable to the sixthif not the fifth century (Harden, 1971). Thetechnique continued to be practised until thelate twelfth century (Pinder-Wilson, 1968).

IridescenceAncient glasses may become iridescent as aresult of the weathering process havingproduced layers of silica at their surfaces,which are interspersed with air spaces withinwhich light can cause optical interferencephenomena; and as a result of deposition ofmetallic oxide ions on the surface. However,glass can be deliberately made iridescent byspraying on a solution of stannic chloridefollowed by firing in a reduced atmosphere toproduce a transparent layer of stannic oxide,which is thick enough to produce opticalinterference colours. Decorative yellow,orange or red effects can be produced onglass by firing on compounds of silver andcopper, and then refiring.

GildingThe technique of gilding is the process ofdecorating glassware on the surface, or on theback of the glass, by the use of gold leaf, goldpaint, or gold dust. The history and techniqueof gilding have been comprehensivelydescribed by Charleston (1972). Earlyexamples of the application of gold sheet toglass are Tutankhamun’s head-rest (CairoMuseum), and the plain gold bands on alidded eye-paint container (1457–1425 BC) inthe British Museum.

In the process of fired gilding, gold wasapplied to the outer surface of a glass by usinggold leaf pulverized in honey, or powderedgold, and affixed by low temperature firing toassure reasonable permanency. The result-ing appearance was dull with a rich andsumptuous effect but the gold could be

burnished to brightness. Gold could be morepermanently fixed to a glass surface by brush-ing on an amalgam of gold and mercury. Low-temperature firing caused the mercury tovaporize, leaving a gold deposit forming thedesign, which could then be burnished tobrightness. Mercuric gold has a thin, metallic,brassy appearance quite unlike the dull richcolour of honey gilding; it is much cheaperand easier to fix.

Some German diamond-point engraving issupplemented with gilding. By the seventeenthcentury, decoration on the surface of glass wascarried out by firing gold leaf, using one ofseveral methods (see above), or, especially inVenice, Hall-in-Tyrol, Austria, and theNetherlands by unfired gold painting. Suchmethods of decoration were for gold borderson engraved glassware, and in Germany onsome Court glassware with gilt in relief.Gilding on the surface of English glassware inthe eighteenth and nineteenth centuries wascarried out by Lazarus Jacobs, William andMary Beilby, Michael Edkins and James Giles.In Spain gilding was applied to wheel-engraved glassware at La Granja de SanIldefonso in the eighteenth century, andsimilar work was produced in Germany.

The process of unfired gilding involvedapplying gold, in a medium, to the outersurface of glass, resulting in a lack of perma-nency. It was a primitive method by which apreparation of linseed oil was applied to theglass with a brush after which the gold leafwas laid on and the oil allowed to dry. As itwas unfired the gold readily rubbed off and,of course, it could not be burnished. Thismethod was used in Britain on some opaquewhite glass and on countrymarket glass of thenineteenth century. It was also used on someGerman, Bohemian and Spanish glass, whichhas lasted better than that done in England,perhaps due to the fixative employed. Wheresuch gilding has been rubbed off, the designmay still be observed by the pattern on theglass left by the fixative. This method ofgilding is also called cold-, oil-, size-, lacquer-,or varnish-gilding and is similar to applyingcold colours.

Acid gilding has been used more recently,usually to produce decorative borders. Adesign, or ground, is etched on the glass withhydrofluoric acid, gold leaf is then applied


overall. After being burnished, the polishedraised areas contrast with the matt etchedareas.

Sandwiched gold leafGold leaf, with engraved designs, could besandwiched between layers of glass. Some ofthe earliest sandwich gold glasses are theHellenistic bowls found at Canosa in Apulia,Italy, dating to the third century BC, and nowin the British Museum London (Figure 3.32).

Heraclius described the manufacture ofgold-decorated glass bowls of the Canosa type(Forbes, 1966):

I obtained several bowls of high sparklingglassThese I painted with a brush dipped in aresin called gumThen onto the golden bowls IBegan to put leaves of gold, and when Ifound them driedI engraved little birds, and men, flowers andlions According to my desire. Then I coatedthe bowlsWith thin layers of glass for protection, blownat the fireAnd when this glass had enjoyed the evenheatIt enclosed the bowls perfectly in a thin layer.

The outer bowl, which reaches just below therim of the inner one, has not been fused oradhered to the inner bowl but the two layershold together simply because they match eachother perfectly. Sandwiching gold between twolayers of glass protected the gold leaf, but therewas always the disadvantage that air bubblesmight get between the two layers and disfig-ure the design. Later examples, where the goldleaf lay under only parts of the surface, towhich it was fused with a layer of glass (fondid’oro, see Figure 2.7), were found in thecatacombs in Rome and in the Rhineland inthe fourth century AD (Painter, in Harden et al.,1968). Examples of gold leaf protectedbetween two layers of glass are known fromthe Parthian or Sassanian periods in Persia,perhaps the second to fourth centuries AD, andfrom Egypt and Syria, when a piece of glasswas gilded and engraved on its reverse andthen fused to a layer of clear glass. A leaf ofgold, or silver, fixed to the back of a sheet of

colourless glass and formed into a pattern orscene, is known as verre eglomisé (see Chapter7). Highlights are produced by scraping awaythe foil to reveal the glass beneath it.

Some Roman glass gilding was done byapplying gold leaf to a hot bubble of glasswhich, when blown, would break the leaf intospeckles. Another method used on Romanglass and later on Venetian glass, was sprink-ling granular gold dust on to molten glass.Islamic glass in the twelfth to fourteenthcenturies was very rarely decorated only withgilding, applied by the use of colloidal goldand then fired; but mainly such glasscombined gilding with enamelling, forexample mosque lamps. Some Venetian glass-ware of the sixteenth century has gilding as


Figure 3.32 Detail of a gold-glass bowl from Canosa,Italy, showing its fine decoration of a floral design ingold leaf, sandwiched between two layers of colourlessglass. In this early period, the layers were not fusedtogether, as in late Roman and Renaissance sandwichgold glass. (© Copyright The British Museum).

the sole decoration, but normally enamellingwas combined with it. The history andtechniques of gilding glass have been compre-hensively described by Charleston (1972); andthe production of gold tesserae is discussedbelow under Architectural Glass.

In Bohemia a technique of sandwiching goldbetween two layers of glass was developed inthe 1730s to decorate a type of drinking glass(Ger. Zwischengoldglas). The vessels weredecorated with gold leaf by a process wherebythe outer surface of one glass was coated withgold leaf, the design engraved through it, andthen a bottomless glass was sealed over thetop (using a colourless resin) to protect thedesign. The inner glass had been ground downwith great exactness for almost its whole heightso as to permit the outer glass to be fittedprecisely over it. On early rare Bohemianexamples the joint showed at the top of therim, but on later ones the rim of the inner glasswas of double thickness for a distance of about10 mm, so that a projecting flange fitted overthe outer glass. The outer glass projectedslightly at the bottom, and the space below, inthe case of a beaker, was filled with a glassdisc with similar gold engraving and sealed bytransparent colourless resin. The outer surfaceof the double-walled glass was sometimesfurther decorated by cutting 12–18 narrowvertical facets or flutes. Silver leaf wassometimes used instead of gold. A frequentform of this type of vessel was a small straight-sided beaker, and popular decorations werehunting scenes, views of monasteries,Bohemian saints and armorial bearings, all verydelicately engraved. The best examples datefrom the 1730s, but others were made untilabout 1755. Such decoration was also used onbeakers, goblets and other double-walledvessels, sometimes combined with cold paint-ing on coloured glass or with enamelling. Suchware is sometimes termed double glass. JohannJosef Mildner revived and elaborated thetechnique in 1787 (Newman, 1977).

Nineteenth- and twentieth-centuryindustrial glassmaking: a summary

The nineteenth century was a period of greatchange (yet the rate of change was certainlyeclipsed in the twentieth century), and only a

few highlights can be recorded here. Greatimprovements in melting efficiency becamepossible when the Siemens regenerativefurnace was introduced in 1861 (Douglas andFrank, 1972), to be followed by the tankfurnace. There were 126 bottle-houses inBritain in 1833, but the number rose to itspeak of 240 in 1874 (Meigh, 1972), thereafterdeclining as the tank furnaces became largerand more efficient. The first semi-automaticmachinery for making jars (Arbogast) wasintroduced in America in 1882, and the firstfor making narrow-neck bottles was inventedby Ashley in Britain in 1886.

The construction of the Crystal Palace in1851 was an extraordinary achievement forwhich nearly one million square feet of glasswas required (Hollister, 1974). The first safetyglass was invented in 1874 and shown at theMotor Show of 1906. The first electric lampbulbs were made at Lemington (nearNewcastle-upon-Tyne) for Sir Joseph Swan in1860; to be followed by the CorningGlassworks in America who supplied Edisonin 1881 (Douglas and Frank, 1972). Opticalglasses were first studied seriously by Dollondin 1758, but the materials were of poor quality(Douglas and Frank, 1972). In 1798, Guinanddiscovered how to make glass homogeneousby stirring it during melting (the only effectivemeans of obtaining homogeneity), butcommercially satisfactory optical glass was notmade until 1848, when Bontemps (1868)joined Chance Brothers in Birmingham. Thefirm’s leadership in this field was notmaintained, however, but passed to Germanywhen, in 1846, Carl Zeiss opened a workshop.Zeiss was later (1875) joined in the venture bya physicist, Abbé. Between them Zeiss andAbbé made a spectacular range of glassesduring the 1880s. These had new opticalproperties which enabled great advances to bemade in the design of lenses for cameras,microscopes and telescopes.

Even more remarkable advances were madein glassmaking in the twentieth century,Douglas and Frank (1972). The first half of thecentury could be said to have been dominatedby engineering-type developments whereasthe post-1950 period has seen quite extraordi-nary changes in the compositions of glasses,especially in the fields of non-silicate glasses.

In making flat glass, the first Lubbers cylin-


der machine was introduced in 1903, but theFourcault machine for sheet glass was devel-oped in 1913 and the mammoth machine forthe continuous grinding and polishing of plateglass (it was about 400 m long) was inventedby Pilkington Brothers in the 1930s (Douglasand Frank, 1972). This process was replacingall previous methods of making plate glasswhen, in 1959, the announcement of the floatglass process, in which a perfectly smooth andbrilliant fire-polished glass is floated from abath of molten tin, secured for PilkingtonBrothers the world lead in flat glass technol-ogy. The automatic machinery for makingbottles and jars also developed greatly duringthe first half of the century.

The Owens suction machine was firstsuccessful in 1903 and it rapidly dominated theglass container-making industry all over theworld, partly because of its technical efficiencybut partly because a cartel was set up regard-ing its use which excluded all others.Gradually, however, between the First andSecond World Wars, various types of gob-fedmachine overtook the cumbersome Owensmachines and displaced them. The numbers ofbottles and jars increased from 2.9 thousandmillion per annum in the United States in 1918(and 0.5 in the United Kingdom) to 15.3 (3.1in the United Kingdom) in 1950 to 44.3 (6.9United Kingdom) in 1977; the 1977 total forthe United Kingdom and the United Statesrepresented 14.3 million metric tons of glass.

The understanding of the chemical constitu-tion of glasses, and the relationship betweenit and the physical properties has increasedgreatly: Pyrex glass was developed in 1915(Society of Glass Technology, 1951); thedelicate colouring produced by rare-earthoxides was discovered in 1927; top-of-the-stove ware was introduced in 1935; and glassfibre was first produced on a commercial basisin 1938. Remarkable improvements were madein optical glasses in which entirely new typesof glass were made: the fluoroborates,phosphates, germanates, and all-fluorideglasses possess different combinations ofrefractive index and dispersion, which hadnever been anticipated in the previous century(Douglas and Frank, 1972). Special ultra-purehigh-transmission glasses have been devel-oped for lasers and optical communicationsystems and a range of glass-ceramics has

been produced which have a zero coefficientof expansion. The photosensitive silver-containing glasses, which were first introducedin 1950, have been developed for specialpurposes, such as sun-glasses which adjusttheir absorption coefficient according to thelight intensity, and polychromatic glasseswhich can develop any colour in the spectrumaccording to the extent to which they areexposed to ultra-violet light and a subsequentheat treatment.

Enamels

Enamels are coloured vitreous powders,applied and fused at relatively low tempera-tures between 500 and 700°C in a muffle kiln,to glass, ceramics or metal substrates to formdecorative designs. Enamel colours whenapplied to glass and ceramic objects and towindow glass, are referred to as enamelling orenamelled decoration. Enamel objects areformed when dry vitreous powders areapplied to a metal substrate, normally of gold,silver, copper, bronze or iron (Maryon, 1971).

Most early enamels failed because the highlyfusible frit never actually fused to the metallicsubstrates. Enamels must be formulated so asto have a co-efficient of contraction roughlyequivalent to that of the metal base; and itsmelting point must be approximate to, butlower than that of the metal, to ensure fusionoccurs. For these reasons, enamels arecomposed of soda or potash glass with orwithout the addition of colourants and opaci-fiers. On thin or extensive areas of metalwork,the contraction of the enamel on coolingmight be sufficient to cause the metal to warp.To counteract this, the reverse of the objectmight also be enamelled, a process known asenamel backing or counter enamelling.

Enamel is generally a comparatively softglass; a compound of flint or sand, red leadand soda or potash, melted together toproduce an almost clear glass with a bluish orgreenish tinge known as flux or frit. It is madein different degrees of ‘hardness’, that is, themore lead and potash it contains the morebrilliant but softer it is. The clear flux or fritis the base from which coloured enamels aremade by the addition of metallic oxides. Theinclusion of 2–3 per cent of an oxide to the


molten flux is generally sufficient to producecolour. The enamel, after being thoroughlystirred, is poured out onto a slab in cakesabout 110 mm in diameter. For use it is brokenup, ground in a mortar to a fine powder,thoroughly washed, dried and applied to themetal. The work is placed in a furnace untilthe powdered enamel fuses and adheres to itsmetal base. Soft enamels require less heat tofire them and are therefore more convenientto use, but not so durable. In creating adesign, the hardest colours are applied andfired first, and successively softer ones fusedduring subsequent firings, at lower tempera-tures. Thus an object such as a plaque, maybe fired a dozen or more times, dependingupon the number of colours to be applied.

Opaque enamel usually required a lowerfiring temperature (Fr. petit feu) than translu-cent enamel, about 300°C. Higher temperaturefiring was known as grand feu. Translucentenamelling involves the firing of transparentlayers of enamel, onto a metal guillochesurface, engraved by hand or engine turned.There may be as many as five or six layers ofenamel each of which had to be firedseparately at successively lower temperatures.Sometimes gold leaf patterns or paillons orpainted decoration or scenes were incorpo-rated in the design. This effect was achievedby applying and firing the gold leaf or enamelonto an already fired enamel surface, beforebeing sealed with a top layer of enamel, whichwas then fired. The completed enamel thenrequired careful polishing with a woodenwheel and fine abrasives, to smooth down anyirregularities in the surface, and then finishingwith a buff.

Enamel objects are classified according tothe relationship of the frit to, and the struc-ture of, the metal base (Figure 3.33). Dippedenamelling, by which a heated metal core wasdipped into and coated with molten glass andshaped to the desired form, was introducedinto Hellenistic jewellery in the third centuryBC. The technique was used mainly to producependants for ear-rings. However, the enamelwas usually applied in such a way as to forma level surface with the surrounding metal, byforming compartments into which the frit wasfired. This was achieved by soldering thinmetal wires or bands (Fr. cloisons) onto themetal to form cells, a technique known as

cloisonné; or by providing for sunken areas inthe original metal casting, a technique knownas champlevé (Fr. for ‘raised field’), or en tailled’epergne (Fr. for ‘economic cut’). Anothergroup of enamels did not rely upon havingmetal boundaries, but on knowledge ofcolours and firing temperatures. These include


Figure 3.33 Types of enamelling on metal: (a)cloisonné; (b) champlevé; (c) plique à jour; (d) en rondebosse.

painted and encrusted (Fr. en ronde bosse or‘in the round’) enamels.

Cloisonné is the most primitive form ofenamel. Each mass of powdered glass wasplaced in a separate compartment formed fromstrips of metal wire to which, and to thebackground if it had one, the enamel is fusedby heating in a kiln (Figure 3.34). On verythin metalwork the contraction of the enamelon cooling might be sufficient to cause themetal to warp, and to counteract this thereverse face of the object might also beenamelled, a process known as counter-enamelling or enamel backing. In 1453cloisonné enamelling virtually ended with thefall of the Byzantine Empire, but it continuedin the form of filigree enamelling. In thisprocess, thin twisted wires of copper or silverwere enclosed the enamel in the cloisonné

manner. The technique is thought to haveoriginated near Venice in the second half ofthe fourteenth century, and is best known forfifteenth century Hungarian examples.

Enamels could also be made without apermanent backing, being held to the metalonly at the edges – plique à jour (Fr. for‘against the light’ (Figure 3.35). The areas tocontain the enamel were fretted and given atemporary backing of sheet mica or somesimilar material to which the enamel wouldnot adhere. Once the enamel had been fusedinto the framework by heating, the backingwas removed leaving translucent enamel likea painted glass window, the lead lines of awindow being replaced by the metal cloisonsof the enamel. This type of work is fragile andnot suitable for objects that are subject torough handling.

In the second group, known as champlevé,or en taille d’epargne, the powdered enamelwas fused into cells, cut with chasing tools,carved, stamped or cast into the metal base-plate. To ensure a good grip for the enamels,it was customary to leave the floor of therecesses rough. The metal base was normallythick, and so did not require counter-enamelling to prevent its distortion. In theearliest examples only lines of design wereincised into the base plate, but over time,


Figure 3.34 Detail of cloisonné enamel on a vase, inwhich the strips of metal defining the design can beclearly seen. (Courtesy of S. Dove).

Figure 3.35 Plique à jour enamel cross. The emptyring on the right shows that the enamel (now missing)had no backing.

more of the base was cut away, leaving onlythin partitions, resembling the earlier cloisonnéwork in appearance. Beautiful examples ofthis work are known from Celtic and Anglo-Saxon Britain. Another centre from whichgreat quantities of champlevé work came wasthe town of Limoges in central France, fromwhere in the twelfth, thirteenth and latercenturies thousands of reliquaries, crosses,altar vessels and other works were sent to allparts of Europe. In most of the early workfrom Limoges each decorated space was filledwith broken opaque colours, often blues andgreens touched with creamy white. The

ground was generally gilded, and the head offigures often made by repoussé work in fairlyhigh relief from a separate piece of copper orbronze, rivetted on and gilded prior to beingenamelled.

A sub division of this group is known asbassetaille (Fr. for ‘shallow cut’) enamel. Inthis process, a layer of translucent enamel wasfused over a design cut in low relief, chasedor engraved on a metal base-plate. Sometimesa panel produced primarily by chasing wassharpened by a certain amount of engraving.Considerable traces of such work can be seenon the Royal Gold Cup (British Museum,London) (Figure 3.37), although the majorportion of the work was executed by chasing.The cup was made circa AD 1530, probably inBurgundy or Paris, and decorated with scenesfrom the life of St Agnes. In many medievalworks executed in the bassetaille technique,the enamel extended right across the panel,with no metal surface left exposed. Thecolours were prevented from running into oneanother by mixing a little gum of tragacanthwith them, and by allowing each colour to drybefore the next was added alongside it. Inpanels with figures, even the flesh wascovered with a layer of clear enamel, whichallowed the modelling of the face and handsto show through. In the case of the RoyalGold Cup, the whole of each figure, tree,scroll or piece of furniture represented iscovered from side to side with enamels, which


Figure 3.36 Tie-pin decorated in champlevé enamelwith symbols of the four playing card suites.

Figure 3.37 The Royal GoldCup, detail of the bassetailleenamel, showing a scenefrom the story of St Agnes. H236 mm. 1380–1. France. (©Copyright The BritishMuseum).

extend as a level surface right across them.Each is thus shown in silhouette against thegold background. The detailed modelling ofthe faces, hands, the folds of the robes, andother objects are seen through the enamel, thedeeper depressions appear richer in tone thantheir shallower neighbours.

The process of en resille sur verre entailedpacking the enamel frit into gold-linedincisions engraved on a medallion of blue orgreen glass, to which it fused on firing. Thisdifficult technique was only adopted duringthe second quarter of the seventeenth centuryin France, where it was mainly used todecorate miniature cases.

A rich technique of covering figures ordecorative devices formed in the round, withopaque enamel, en ronde bosse (Fr. for ‘inrounded relief’), was used in Paris at thebeginning of the fifteenth century on largereliquaries; and in England, where it wasknown as encrusted enamelling, and was usedto decorate irregular surfaces, for example, theshoulders of finger rings, to ornament themounts of cups or a figure in high reliefformed by repoussé work and chasing. TheDunstable Swan in the British Museum,London is a fine example of this process. Fromthe sixteenth century, goldsmiths enrichedtheir work with touches of coloured enamel.One of the principal problems arising from theuse of enamel in this way concerns themanner in which the object can be supportedduring firing. An elaborately constructed fingerring may have a number of soldered joints,which would melt if exposed to the heat ofthe furnace. Such joints must be protected, bypainting them with rouge or whiting beforethe enamel is fired. A pendant jewel built upfrom a number of separately formed piecesheld together by solder, may have enamelleddecorations on surfaces inclined at manyangles. During the firing process, the soldercan be protected with plaster-of-Paris.

Enamels of the third group come under thegeneral classification of painted enamels,although as with colours on a canvas, thematerial was sometimes applied with a paletteknife. The process was invented in Limoges,France in the fifteenth century. Paintedenamels have a plain foundation of a sheet ofmetal, generally slightly domed, and as a rulethe whole surface on both sides of the metal

is covered with enamel (Figure 3.38). Thetechnique relied upon a more sophisticatedknowledge of enamel compositions and firingtemperatures. With each colour having to befired at successively lower temperatures, inorder not to destroy the previous work, thetechnique is painstakingly slow and prone toirreparable flaws.

Cloisonné and champlevé enamels had beenmade for many centuries before it was discov-ered that the metal outlines between the differ-ent colours were not essential to thepermanence of the work, valuable though theywere from the decorative point of view.Towards the end of the fifteenth century thecraftsmen at Limoges in France began, in theenamelled pictures which they fitted into thework, to leave out the metal divisionsaltogether. The way in which they worked hasbeen followed, with little variation, byenamellers ever since.

Wetted finely ground enamel frit mixed withoil for ease of application was painted over adesign scratched in a metal base-plate andeach colour allowed to dry before the nextwas applied. The object, such as a plaque orbox, was first covered with a layer of whiteopaque enamel and fired. The design and/orinstructions were applied in different colours,and fired in a low temperature muffle kiln(approx. 500–700°C). The medium burned outduring firing. Different firings at successivelylower temperatures were required to fusedifferent colours of the same hardness, inorder to prevent them running into oneanother.

If a copper plaque were given a coat of flux


Figure 3.38 Detail of a Chinese painted enamel dish.(Courtesy of J. McConnell).

before enamelling, the enamels would showmuch more brilliantly. Copper, with a coatingof clear flux, appeared a bright golden or apink coppery colour according to the compo-sition of the flux used and to the temperatureat which it was fired. Silver seen through clearflux resembled white satin; whereas the colourof gold hardly changed at all. Any design onthe metal, whether scratched in with a steelpoint, or drawn with a lead pencil, showedclearly through the flux. The surface of theflux could, however, be roughened, thenwashed over with hydrofluoric acid to avoidmilkiness in the colour, and the design trans-ferred to it. The coloured enamels were thenapplied where required. The colours could bemodified, shading if necessary, gold or silverfoil added to make a more brilliant patch,parts of the design outlined, or touches ofgold added.

Gold leaf or foil may be employed invarious ways in enamel work: it may be usedto cover part of the background, such as figureas in some primitive Italian and other paint-ings. For such work a foundation of translu-cent yellow enamel was provided over whichthe gold could be laid. Before laying the gold,the surface was moistened and kept moist bybreathing gently on the film of gold while itwas applied. A temperature just high enoughto fuse the enamel was sufficient to fix thegold. If any defects appeared, the gold wasbrushed with a glass brush, another layer ofgold placed over the first to cover the gaps,and refired. It was not necessary to cover thegold with enamel.

In the technique of Swiss enamelling, aneutral-coloured enamel was fused over a goldpanel to form a matt surface. A picture wasthen painted, fired and given a translucentcovering. Examples are found on nineteenth-century plaques decorated with pictures ofSwiss girls wearing typical cantonal peasantdress.

Stars, fleurs-de-lys, rosettes or other devicescould be stamped out in low relief in gold foiland laid on a foundation layer of enamel toform a diaper or other patterns. They wereoverlaid with a coating of translucent enamel,and the work fired. Translucent colours lookedbrighter when fired on a background of goldor silver than on copper or on the black ordark blue background employed by many

medieval enamellers (Maryon, 1971).Apart from the use of plain coloured

enamels, a number of techniques for enameldecoration were employed which were verysimilar to those used on glass, such as mosaicand millefiore.

Architectural glass

By far the most widespread use of glass in thearchitectural context was in the form ofpainted window glass, from medieval timesonwards. However, in Roman times, glass inthe form of mosaics and opus sectile had beenused to decorate the interiors of buildings (seeFigure 3.39) and glass with a greenish tint hadbeen used for small window-panes (see Figure7.28f).

Mosaics

The architectural process of embedding smallpieces of roughly squared glass (tesserae) incement on walls or floors to form a picturewas first developed by the Greeks, and thenused extensively in the Roman and Byzantineperiods. Roman mosaics were generallycomposed of stones, but there are earlyexamples made of glass in Rome; and seventhcentury AD examples occur in Ravenna andTorcello, Italy. According to Theophilus (trans.Hawthorne and Stanley Smith, 1979), theGreeks made sheets of glass a finger (20 mm)thick and split them with a hot iron into tinysquare pieces and covered them on one sidewith gold leaf, coating them with clear groundglass ‘as a solder glass before firing in a kilnto fuse the gold leaf in place; Glass of thiskind, interspersed in mosaic work, embellishesit very beautifully’. Vasari (1511–1571) gives adescription of the preparation of mosaiccubes. First, the glass is made opaque by theaddition of tin oxide and/or coloured by theaddition of metallic oxides. When the glasswas sufficiently melted and fused it was ladledout in small quantities onto a metal table andpressed into circular cakes about 20 cm indiameter and from 1.0 to 1.25 cm thick. Thesolidified glass was then annealed, cooled andcracked into tesserae. The fractured surfacewas generally used to form the upper surfaceof the mosaic since it had a more pleasing


surface and richness of colour. The thicknessof the glass cake therefore regulated thetexture.

Gilded tesserae were produced by hermeti-cally sealing gold leaf between two sheets ofglass (Figure 3.39). The technique of produc-ing the fondi d’oro or glass vessels adornedwith designs in gold and found in the Romancatacombs, was of the same nature (see Figure2.7). According to Vasari a glass disc wasdamped with gum-water and gold leaf appliedto it. The gold-covered disc was then placedon an iron shovel in the mouth of the furnace.The glass covering was either made of a glassbubble or from broken glass bottles (cullet) sothat one piece covered the entire disc. Thedisc was then held in the furnace until italmost reached red heat when it was quicklydrawn out and cooled. In order to set thetesserae on walls or ceilings, a cartoon(drawing) was first pounced (pricked through)portion by portion on soft cement. Sometimesthe resulting outline was coloured as a guidebefore the application of a thick cement andsetting of the tesserae. The stucco cementremained soft for two to four days dependingupon the weather. It was made of lime cementmixed with water, that is travertine, lime,pounded brick, gum-tragacanth and white ofegg, and once made was kept moist withdamp cloths while the tesserae were set(Maclehose, 1907).

Opus sectile

During the excavations at Kenchreai, theeastern port of Corinth (Greece), during the1960s, over 100 panels of opus sectile worked

in glass were found, stacked on the floor of abuilding in the harbour area. The building andothers associated with it were in the processof being remodelled or redecorated, whenthey were submerged during a seismic distur-bance in AD 375. Other examples of opussectile are known (Figure 3.40), but theKenchreai panels are the most extensive.

The panels can properly be regarded as astage in the historic evolution of the use ofrichly coloured glasses fitted or fused togetherto form ornamental designs. The origins of thistradition lie in the early glass of Egypt, anddevelop through the Nimrud ivory inlays andthe fused mosaic glass vessels like those fromHasanlu and Marlik. But the most immediatetechnological precursors of the Kenchreaipanels are the whole array of objects from theworld of ribbon glass and millefiori, includingminiature fused mosaic plaques. TheKenchreai panels might be considered to bejust a slightly different version of other opussectile work, such as those at Ostia (Italy),except that they were executed in glassinstead of stone. But the separate units inmany of the details and highlights of thepanels are themselves polychrome pieces ofglass. These details were not formed by simplyarranging separate, tiny units of differentlycoloured monochrome glasses. They weremade by a series of intricate processes inwhich carefully shaped bits of monochromeglasses were fitted, fused together by heating,and while still softened, manipulated invarious ways into complicated polychromedesign components like the details of fishscales and bird feathers. This characteristicputs the work squarely into the millefiori tradi-


Figure 3.39 Mosaic panel of Roman glass. 160 mm.100 BC to AD 100. (© V&A Picture Library).

Figure 3.40 Fragment of opus sectile wall revetment.Cast and mosaic glass set in resin. Roman Empire. Latefirst or early second century AD. L 7.5 cm. (CorningMuseum of Glass).

tion. It is a technique that could only havebeen executed in glass and should not beconstrued as purely an inlay or mosaictechnique, which could have been executed,for example, with shaped, polished, colouredstones. The Kenchreai panels are a product ofthe pyrotechnological arts, and not simply thelapidary art (Scranton, 1967; Ibrahim et al.,1976).

The panels may not only be a stage in thistradition, but could conceivably be regarded asits culmination; for it does not appear tocontinue. In the West a few inlays in casketsand jewellery occur later and eventually thecloisonné technique evolves. In the Byzantineworld mosaics flourished on a grander scalethan ever before, but still as a part of their ownmosaic tradition: one colour for each cube, andonly cube after cube, with few other geometri-cal shapes being utilized. With the decline ofthe Roman world, the centre of glassmakingmoved eastwards, and along with the otherchemical arts, glassmaking flowered in theIslamic world. But by then glass artists hadbecome totally preoccupied with glass-blowing,three-dimensional forms and transparency.

The panels were of various sizes and shapes– squares and oblongs – among which thesmallest dimension would be around 0.3 cm,and the largest almost 2.0 cm. According toone hypothesis, a surface of appropriate sizeand shape was prepared first on a table, or atray with raised edges. On this was then laida kind of pavement of large potsherds cutfrom coarse amphorae, approximately rectan-gular in shape. Over these potsherds, andimpressed into the interstices, was laid a thickcoat, perhaps 20 mm over the potsherds, of akind of plaster of which the chief ingredientswere pine resin and finely pulverized marble.This was brought to a smooth surface. On thissurface were laid pieces of glass of variouscolours, cut to various shapes (about 2 mmthick), in order to create the desired pattern,and these were glued down with more resin.According to a second hypothesis, the piecesof glass would be laid face down on the tray,and covered with the coat of resin plaster intowhich the potsherds would be impressed.

When found, some of the glass was firmand hard, some had disintegrated to a powder,and some was preserved in one of a varietyof states in between. Although the panels were

colourful when found, the colours presentedwere apparently not the same as the originalcolours, in most cases at least. Thus thecharacteristics of the various kinds of glass atthe time when they were applied to the panelscan only be inferred. However, it would seemthat some kinds, notably those used for humanflesh and for the masonry of buildings, werecut from large thin sheets. Of other kinds,notably the neutral backgrounds, it has beensuspected that they were trowelled into placewhile in a more or less pasty state, though thisis not certain. Some thin dividing lines, andespecially the curved stems of flowers, mayhave been set in place as still-soft rods, orconceivably even as an extruded paste. Morecomplex forms such as the seed-pods of lotusplants or the bodies of certain fish or birdsgive the appearance of having been made ina kind of millefiori and fused-glass mosaictechniques, externally moulded so that incross-section the block would have theintended inner markings. Some particularshapes, such as the heads of birds or animals,may have been built up plastically in separatemoulds, with certain details engraved orimpressed on the face intended to be visible.

Mosaic and opus sectile techniques wererevived in the eighteenth century in the UnitedKingdom (Figures 3.41 and 3.42).

Manufacture of window glass

Mould-cast glass panes

In Roman times, small panes of glass wereproduced by casting molten glass into shallowtrays and spreading it out manually as itcooled. There is no contemporary reference tothis process, but a number of mould-cast glasspane fragments have been excavated, forexample, at Pompeii in Italy, and on severalRomano-British villa sites. There is a fragmentof Roman window glass from Caerleon inWales (Boon, 1966), which seems to havebeen cast on wet wood judging from theimpressions of wood grain on one side of it,although this has been disputed. Fragments ofan almost complete pane of window glassdating from the second century AD wereexcavated in 1974 in a bath-house complex atthe Romano-British iron-working settlement atGarden Hill, Hartfield in East Sussex (Money,


1976). The pane (see Figure 7.28f) measured235 mm � 255 mm and is now in the BritishMuseum. It has one flat side on which thereare marks, which suggest that it was cast ona bed of sand. Two edges are rounded whilethe others are grozed. The one remainingoriginal rounded corner bears a mark made bypushing the molten glass into the mould witha pointed instrument. There may have beenother panes in the bath complex; and theywere certainly used elsewhere on the sitesince fragments representing several paneswere unearthed in several locations. Fragmentsof window glass excavated at Stonea,Cambridgeshire in 1981 bear marks of woodgrain on one side, and one bears evidence ofhaving been manipulated to fill the mould.

Cylinder glass

The most prevalent and long-lived methods ofproducing early Roman window glass,however, were the spinning process, whichproduced crown glass; and the cylinderprocess, which produced cylinder glass (alsoknown as hroad, sheet, spread or muff glass,later Lorraine glass or German sheet) (Figure3.43). Cylinder glass tended to be irregular inthickness between the ninth and thirteenthcenturies, and it lacked flatness, although itwas glossy on both sides, unlike the Romancylinder glass with one matt and one glossyside, probably because the kilns were not ashot as the Roman type. Chambon (1963) givesa very detailed account of the developmentsin making cylinder glass.

Cylinder glassmaking seems to have beenthe earlier invention, but cylinder and crownglassmaking repeatedly replaced each otherduring the fourth to the nineteenth centuriesdepending on the fashion; for example, Tudorleaded windows had crown glass, after whichthe methods were superseded by mechanicalmethods of sheet glass production. Cylinderglass was certainly used in Roman buildingsin Britain, Italy and Greece (at Corinth) beforecrown glass became the fashion.

The cylinder process enabled larger andmore even sheets of glass to be made thanthose produced by the mould-cast method. Agathering of glass was first blown into a broadbulb up to 1.5 m in length, and then given acylindrical shape by swinging it back andforth (to lengthen it) and marvering (to keepit cylindrical). The ends were then cut off andthe cylinder split longitudinally by a hot iron,and then reheated on the flat bed of a kiln.When it was hot enough, the glass could beflattened into a sheet with iron tongs and asmooth piece of iron or wood called a rake,croppie or flattener (see Figure 3.44). Thisdescription is simplified in its essentialsbecause the exact method of productionvaried from era to era. For example,Theophilus states (Dodwell, 1961; Hawthorneand Smith, 1963) that, after the bulb wasblown, the end was pierced (by blowing itout after heating the end in a glory hole), thehole was widened to equal the bulb’s widestdiameter, and then pinched together so thatthe pontil could be attached. At other times


Figure 3.41 Detail of nineteenth century glass opussectile in the Crown Liquor Saloon, Belfast, NorthernIreland. (Courtesy of The National Trust).

Figure 3.42 Detail of a nineteenth-century panel madeof glass mosaic and glass frit opus sectile. St George’sChurch, Crowhurst, Kent (UK).

the cylinder was cut with shears while stillhot; there is evidence that this was themethod used in Roman times (Harden, 1959).Cylinder-made glass can be identified by theshape of the air bubbles (seed) which itcontains because they become elongated instraight parallel lines consistent with theelongated shape of the glass bulb. The edgesof a sheet may also identify it as being madeby the cylinder process since the top andbottom of the cylinder (the long sides of the resultant sheet) were rounded in the flame and hence become slightly thickened to produce a thumb edge (Harden, 1961). The longitudinal split made in the cylinder(the edges of which are the short sides of thesheet) may show shear marks, or may bequite sharp. Much Roman cylinder glass is ofthe one matt, one glossy sided type becausethe cylinder was opened, and then presseddown by the croppie onto the sanded floorof the kiln. Green (1979) considered that theearliest type, which had a very flat and matt

lower surface, and an undulating uppersurface, was opened in a too-hot kiln. Laterthe glassmakers’ increased skill resulted in thelower edge not being roughened, and eventu-ally panes were produced which remainedglossy on both sides. In 1973 Hardenexamined Romano-British glass excavated atShakenoke Farm near Oxford and concludedthat equal quantities of matt/glossy sided anddouble glossy sided window glass were usedconcurrently during the third and fourthcenturies.

Harden (1961) describes some large panesof Roman glass, one measuring 395mm �260 mm, and another, an almost completepane of cylinder glass measuring 535mm �305 mm. Two pieces, excavated in 1972 atHartfield, Sussex, were 255mm � 235mm and275 mm by more than 215 mm, and theyvaried in thickness from 2 to 5 mm (Harden,1974). Other large pieces from Italy, alsolisted, were 330mm � 270 mm and 267 mm �267 mm (Harden, 1961).


Figure 3.43 Techniques ofmaking (a) cylinder glass, (b)crown glass and (c) Normanslab for use as window-panes.

Moulding the cylinder

Blowing and rolling Shaping for crown glass

Transferring from pipe torod Widening the hole

The final disc marked forcutting

Moulding Dividing the sidesRemoving the end

It seems likely that northern Gaul and theRhineland exported glass to Britain during thefirst seven centuries of the millennium, butthere was certainly some local production, atleast from the late seventh century. In AD 675Benedict Biscop imported workmen from Gaulto help in constructing the monastery atMonkwearmouth in Sunderland (UK), ‘in theRoman manner’, and glaziers to glaze thewindows of the church, porticus and refectory.

That these glaziers melted glass on the site hasbeen confirmed by Cramp’s excavations; andthe glass all seems to have been made by thecylinder process. The glassmaking techniqueseems not to have become firmly established,for in AD 758, another appeal was made byCuthbert for glassmakers to be sent to Britain.Glass with similar characteristics to that foundat Monkwearmouth was excavated by Hunter(1977) at Escomb near Durham, Repton nearDerby, Brixworth near Northampton andHamwic (the Saxon settlement in South-ampton). The fragments of Saxon glass wereall quite small, the largest being only about65 mm long.

In the fourteenth century there was a notice-able improvement in the quality of cylinderglass which probably first took place inBohemia. The knowledge of its productionwas taken to Lorraine by four known glass-making families, and thus cylinder glassbecame known as Lorraine glass (Fr. verrefaçon de Lorraine or verre en table). Also inthe fourteenth century, the Venetians used thecylinder process to make glass for mirrors inlieu of the wasteful crown glass method. Inthe eighteenth century a device was intro-duced to help maintain the shape of the cylin-der during working. This was a wet woodenmould ‘shaped like half a cannon’, and animproved surface finish was achieved byopening the cylinder out and flattening it outon a sheet of clean glass. The hand cylindertechnique continued in use until the nine-teenth century when it was replaced bymachine drawn cylinder glass (Lubbersprocess).

Crown glass

The first examples of crown glass date fromthe fourth century AD, both in the East (fromJerash, Samaria) (Harden, 1959, 1961) and in the West (from Chichester, Sussex)(Charlesworth, 1977). These early crowns werequite small, being some 150–200 mm in diame-ter, and (at least in the East) they weremounted in pairs in plaster frames. It was notuntil a much later date that larger crowns werecut into panes (quarries).

The flat panes of glass (crowns) wereproduced by blowing a gather of glass into aglobular shape, transferring it from the


Figure 3.44 Production of early flat glass. Uneven andscarred glass was evened and polished in a grindingshop. Large sheets of glass were fixed to the spokes ofa wheel (B). Emery and wet sand (A) were spread onthe surfaces, which then ground each other smooth.Finally the glass had to be polished with felt buffersaffixed to jointed ribs which maintained a constantpressure (C). Plate from Diderot’s Encyclopédie.

blowing iron to the pontil, cutting it open andthen after reheating in the glory hole, thepontil would be spun rapidly so that bycentrifugal force the glass assumed the shapeof a flat disc up to 1.2 cm in diameter. Theglass was then annealed, either used whole orcut into rectangular- or diamond-shapedpanes.

Pieces of crown glass can be identified bythe curved lines in which the seed lie (withthe pontil at the centre) and by the curvedripples in the surface (see Figure 3.43). Thecomplete crown has a much thicker centre,called a boss, bullion or bull’s eye with therough circular mark in the centre where it wasbroken from the pontil when cold; the piecescut from near the boss are thicker than thosecut from near the edge.

Harden (1971) drew attention to the sixthcentury crown glass window panes found atthe church of San Vitale in Ravenna in such aposition that they almost certainly came fromthe windows of the apse. They are mostlymonochrome panes, 170 to 260 mm in diame-ter, although one fragmentary piece bears anoutline drawing of Christ nimbed andenthroned and this may be a distant precursorof painted glass windows. Bovini (1964) andHarden (1971) suggest that these small crownsform a link between the fourth-century onesmade in the East and the Normandy modecrowns of the fourteenth century. In Britain,however, cylinder blown glass continued to beused in the tenth century, as shown by asample from Thetford, Essex (Harden, 1961).

The surfaces of crown glass were alwaysbright and shiny because they had beenheated in a glory hole and had not touched asanded surface such as that which impairedthe lower surface of cylinder glass. Thus abrighter and more transparent window glasscould be made by the crown process, thanwas possible by the cylinder process (at leastin antiquity), and the relative competitionbetween the two processes depended on thefashion at the time. Thus a demand for largewindows could only be met by the cylinderprocess whereas leaded light windows couldbe made better with crown glass. The crownglass process was later termed the Normandymethod, its invention having been uncertainlyattributed to Philippe de Caqueray ofNormandy who established a glassworks in

Normandy in 1330 where crowns of500–600 mm diameters were produced; despitethe fact that crown glass had been in use formany centuries.

In the sixteenth century Dutch crowns hadan even better reputation than those fromNormandy and the discs became larger so that,by 1700, the diameter was 800–850 mm. In1724 an Arrêt du Conseil d’Etat obliged eventhe Normands to make discs which were 38 pouces (about 1010 mm) in diameter(Chambon, 1963), but this was evidently quitedifficult to do because few were made thatsize, being generally only 30 to 36 pouces(800-960 mm) in diameter. The manufacture oflarge crowns was more complicated thanshown in Figure 3.44; Chambon (1963, PlateII) illustrates 16 different operations.

Chance Brothers continued to make crownglass at least until 1832. Chance (1919)commented that ‘Crown glass excels in brillianceand transparency, but yields in other respects toplate and sheet ... [the centre lump] and thecircular form of the tables, prevent the cuttingof large rectangular panes from them, and thereis much waste... And, as the tables have a slightconvexity the panes, unless flattened, show adistinct curvature and distortion of vision. ...Owing to ... the British public [being] habituatedto the small bright panes, it survived there formany years longer than on the Continent.

Norman slab

A simple technique for making window glassinvolves blowing a square-sided glass bottleinto a mould, and then cutting it into the fourside panels and the base panel (see Figure3.43). It was not possible to make large areasof glass in this manner; the largest squareRoman bottles would have provided panelsthat were only some 200 mm � 100 mm.

The technique is now called Norman slabalthough its origin is unknown (Harden, 1961).When whole panels survive they can be recog-nized by their thin edges, because it is noteasy to blow glass into the corners of a squaremould and maintain the same thickness as inthe rest of the bottle. The edges may alsoshow some curvature if the cut does not runexactly down the edge of the bottle; butpieces cut from the panel may not be recog-nizable as such. Nevertheless, the bubbles will


not be elongated in parallel rows as is oftenthe case with cylinder glass, and observationsthat some of the Jarrow glass had roundedbubbles could support a suggestion that theNorman slab technique might have been used(Cramp, 1975).

Flat glass sheets

Plate or cast glass, run out onto a flat metaltable, had to be quite thick, to allow for thegrinding and polishing processes which werenecessary to render the glass clear enough tobe transparent. The hardened glass had arough, dull surface, which had to be groundand polished with abrasives in order toachieve the brilliance required for its use aswindows and mirrors (Figure 3.44). Around1845 Frederick Masson and J. Conquerorinvented a machine for squeezing molten glassbetween two rollers, opening the way for thedevelopment of many different processes forthe making of sheet glass. However there wasstill little advancement in the size or flatnessof the panes produced. From this time,polished sheet glass became widely available,and resulted in the production of expansivewindows, reaching from floor to cornice, andusually reserved for grand buildings. The mainproducer of rolled glass in Britain during thenineteenth century was the company of JamesHartley of Sunderland, whose products canstill be seen in many Victorian railway stations.

During the eighteenth and nineteenthcenturies, glass sheet introduced to Britain byChance with the aid of Bontemps was termedGerman sheet. (The sheets were in fact largecylinders, also known as patent plate andimproved cylinder glass.) Following the devel-opment of German sheet, drawn cylinder glasswas produced about 1923. The continuousflow process was introduced in 1921 on thecontinent (1923 in Britain), but the glass stillhad to be polished. In 1933 the first verticaldrawn float glass was produced. In 1923,Pilkingtons (UK) invented the first continuouson-line grinding and polishing process,followed by the development of the twingrinding and polishing process in 1937. Theresulting high quality, thick glass is used forlarge display windows, mirrors, table-tops etc.In 1959, Pilkingtons (UK) introduced the floatglass method of producing sheet glass, in

which a continuous ribbon of molten glass, upto 3.30 metres wide, moved out of a meltingfurnace and floated along the surface of amolten bath of tin. The controlled atmosphereand temperature prevent irregularities fromforming in the glass, and the resulting flat glasscan be produced to a thickness rangingbetween 2.5 and 25 mm. The glass is fire-polished and annealed without the need forgrinding and polishing. The electro-floatprocess, developed in 1967, enables the glassto be tinted as it passes over the float bath.

Painted and stained window glass

A number of decorative techniques were usedto embellish glass windows, especially thosein ecclesiastical buildings. During manufacturethe glass could be coloured, or clear glasscould be flashed with a thin coating of


Figure 3.45 Detail of a medieval stained glass windowshowing the design formed of leaded stained andpainted pieces of glass.

coloured glass on one side. The panesthemselves could be painted, enamelled,stained, etched or engraved (Figure 3.45).

Throughout the Middle Ages decorativewindows were composed of pieces of whiteand coloured glass, cut to fit the basic design(cartoon), and held together in a frame-workof leads. With the exception of black enamelpainting and silver–yellow stain, the colour inmedieval window glass came from the glassitself, the pot-metal having been coloured sothat the colour ran through the entire thick-ness of the pane and was as permanent as theglass itself. Certain colours, however, particu-larly ruby, were so dense in tone that theywould not transmit light sufficiently. In thetwelfth and thirteenth centuries, therefore, thetranslucent red colour was produced bymaking multi-layered glasses. It is not knownexactly how such glasses came to be made butthere are so many very thin layers, perhaps asmany as fifty, which bend back on each otherin the manner of hairpins, that it seems likelythat they were not made by a process ofrepeated gathering alternatively from colour-less and red glass pots. One suggestion regard-ing the method of manufacture ofmulti-layered glasses was that red glass wasstirred into a pot of colourless glass, and thegather for the cylinder was made beforeproper mixing of the glass had taken place.However, the conditions necessary for produc-ing the ruby colour exist in a suitably non-homogeneous manner when the colour isprepared in a very dilute form, and thus thelayering may at first have been an accidentalconsequence which was later deliberatelyexploited (Cable, personal communications,1979). Apparently it is much easier to producea homogeneous red colour when enoughcopper has been incorporated to produce ared glass, which would be too dense to usein the thickness of 5 mm required for awindow.

There are, in addition, some twelfth-centuryglasses which are multilayered in about halftheir thickness, the other half being of green-ish glass. It is possible that the greenish partis a copper-containing glass, which failed tostrike during the essential reheating process,or it may be an early attempt at producing aflashed glass. The multi-layered part is,however, so thick that glasses of this type

should not, strictly speaking, be described asflashed glass.

The difficulty of producing transparent redglass was finally overcome in the fourteenthcentury by the production of flashed glass.The manufacturing process is relatively simple:a first small gather (the post) was taken fromthe pot containing the glass batch which,although colourless in appearance, had beenso formulated as to strike a red colour duringan essential reheating process. A second,much heavier gather of colourless glass wasthen taken over the top of the first. The glasswas blown to form a cylinder, which was thencut, opened out and flattened. (Crown glasscould also be flashed.) Finally, the flashedglass pane was subjected to a controlledreheating process to strike the colour. Theoriginal necessity of flashing dense colourswas frequently turned to advantage by themedieval glazier. By abrading the thin layer ofcoloured flashing, both white and colour couldbe obtained on one sheet of glass. Forexample, if a figure had a ruby tunic trimmedwith white cuffs, the glazier abraded the areaof the cuff from a piece of flashed rubyinstead of having to cut and lead on extrapieces of white glass. In heraldry, with intri-cate charges on coloured grounds, thistechnique was particularly useful, especially asall the gold charges could also be obtained bypainting the white abraded areas with yellowstain.

Applying enamels and yellow stainIn the initial stages of painting, the glass is laidover the drawing, the line work then beingtraced in glass paint on the surface of theglass. Using a sable-haired ‘rigger’ or tracingbrush the painter gathers sufficient paint toproduce a firm black line. Too little paint onthe brush will produce a streaky or washyline, too much produces a blob, which is diffi-cult to control. Alternatively, a piece of glassmay be coated overall with an opaque layerof paint, the design, be it lettering or diaper,then being scratched out of the blackbackground by means of a sharp stick, needleor dart. Lettering on most twelfth- andthirteenth-century glass was carried out in thisway.

With one or two exceptions the methods ofglass painting have changed little since the


eleventh century. Since that time two colour-ing agents have been used to decoratewindows in association with clear andcoloured glass panes. The first of these wasthe opaque black or dark brown enamelwhich was used for painting all the mainoutlines or trace lines, the washed tones andthe shading on the glass; the second colour-ing agent was a silver–yellow stain. Theenamel was originally made from a highlyfusible green lead glass mixed with copper oriron oxide, a binding medium such as gumarabic (sugar, treacle or vegetable oil) and aflux. The use of copper and iron oxides variedat different periods, which explains thechanges in colour from a full black to asomewhat reddish-brown. The ingredientswere mixed together cold, as opposed tobeing mixed under heat, as were the latertranslucent enamels. The binding medium hadthe effect of enabling the enamel to flow andto adhere to the glass, and also hardened thepaint, thus making it more resistant to acciden-tal damage before firing. The addition of a fluxwas necessary to lower the softening point ofthe enamel below that of the substrate glass.After the painting was completed, the piecesof glass were fired, the flux of lead glassmelting and fusing the enamel onto thesubstrate (Lee et al., 1982). The effect ofcoloured jewels in a painted glass window, forexample, on the hem of a robe, can easily beproduced by employing spots of differentenamel colours; but at earlier periods fourother processes were used.

Theophilus (trans. Dodwell, 1961; Haw-thorne and Smith, 1963), described a techniquein which the jewels were cut from a piece ofcoloured glass, and fixed to the substrate glasswith a thick ring of (black) enamel paint whichwas then fired. If the jewel subsequently felloff, the thick ring of paint remained to showwhere it had been. A Romanesque example ofthis type of jewel survives on the Jesse Treewindow (1225–1230) at Regensburg Cathedral,west Germany; and fifteenth-century examplesare known from the St Cuthbert window inYork Minster, and St Michael’s Church,Spurriergate, York (UK). A second techniquefor producing jewels is somewhat similar tothat described above. Because the glass isreheated, the technique is misleadingly referredto by art historians as ‘annealing’. In this

method, the two glasses (i.e. the jewel and thesubstrate glass) were sealed together by apply-ing a layer of ground green glass (perhaps alead glass) between them before refiring. Thepowdered glass melts and acts as a solder.Examples of such jewels do not seem to beknown. The third technique was to insert circu-lar jewels into the substrate glass. Holes weredrilled, somewhat larger than the jewels to beinserted, and the jewels were then leaded intothe holes. There are examples of this techniquein the UK in St John’s Church, Stamford, inBrown’s Hospital, Stamford, in CanterburyCathedral and elsewhere. The fourth technique,occasionally confused with the third approach,was to lead-in small pieces of coloured glassusing ordinary glazing techniques, as can beseen in York Minster, for example the north-ernmost window of the east wall of the NorthTransept.

During the sixteenth century a type oftranslucent enamel was developed. Theenamels were made by mixing a flux of highlyfusible lead glass with various metallic oxidecolourants. With this type of enamel the oxidewas dissolved in the flux by strong heatproducing the required colour. The enamelwas therefore merely a highly fusible colouredglass ground to a fine powder, which, mixedwith a suitable medium, could be applied tothe window pane as a paint. On being firedonto the pane of clear glass it regained thetransparency it had partially lost through beingpowdered, and became a thin coat of trans-parent coloured glass upon the surface towhich it had been applied. It is an anomalythat some of the earliest surviving windowglass has perfectly preserved paint while, fromthe fifteenth century onwards, an irregular butsteady decline in condition of the paintpersisted until some of the worst is found inthe first half of the nineteenth century.Although some of this is attributable to theaddition of borax to the paint, some iscertainly due to inconsistency of firing, as testi-fied by variations within the bounds of asingle window. The firing was carried out ina muffle kiln (Fr. petit feu) at temperatures thatdid not cause the substrate to melt. While thetemperature for fusing pigment into glass isbetween 600 and 620°C, much depends on thetime that the glass is held at a given temper-ature. In the case of medieval firing, the


temperature was gradually built up in the clayfurnace, and the fire extinguished, orwithdrawn, when the temperature was judgedto be correct. The heat was thus retained bythe clay walls, and the resultant slow coolinggave a good period of annealing.

The application of silver nitrate (or silversulphide) to the back of the glass to producea yellow colour when fired was introducedearly in the fourteenth century and revolu-tionized painted glass design by providing theability to incorporate more than one colour ona single piece of glass, for instance, goldenhair or crown on the head of the Virgin.

The medieval practice of preparing silverstain appears to have been to cut some silvermetal into small pieces or to use it in thinsheets, and to burn it with sulphur in acrucible, which converted it into a sulphide.This was finely ground and mixed with anearthy vehicle such as pipe-clay which madethe stain easier to apply. Yellow stain inmedieval glass was usually painted on clearglass and was chiefly used to heighten detailsof figure or canopy work or grisaille groundpatterns, but in later times it was painted oncoloured glass in order to change the tint ofthe piece, or part of a piece. For instance,using blue glass and yellow stain, a green hilland the sky could be painted on one piece ofglass. In figure painting, both hair and facecould be painted on one piece of glass; anddecorated borders to white robes werefrequently stained yellow. Broadly speaking,fourteenth-century stain was golden in colourand some deepening came about until the redstain of the sixteenth century, and its associ-ated lighter tints, brought about the high pointof the stainer’s art, at the time when avail-ability and variety of pot metal colours wasseverely limited.

While at first confined to the yellow-on-white combination, it was extended within thenext century to produce a green tint by apply-ing the stain to blue. It was also realized that,by abrading away the coloured surface offlashed glass and staining the resulting whiteareas yellow, the interpretation of heraldryinto glass became much simpler and enabledsuch charges as golden lions on a ruby groundto be carried out without complex leading.

The effect of silver stain depends on theamount of stain applied and the number of

applications, the temperature at which it isfired and the chemical composition of theglass. The finest staining on glass is known askelp, its alkali content being said to be derivedfrom seaweed, although it probably containedadditional metallic oxides such as tin. It wasextensively used from the sixteenth to thetwentieth centuries and could produce a stainwhich was almost a true red and whichmaintained a clarity equal to that of pot metalyellow or red glass. It was used by Peckitt inYork Minster (UK).

The first use of enamels had been on asmall-scale for transferring complex heraldiccharges on to glass, and for colouring the fruitand flowers of garlands, but subsequently itwas used for the flesh tones of figures andultimately it superseded most of the pot metalcolours. However, at no time were the coloursas transparent as those of silver-stained glass,or of coloured pot metal, nor did the enamelshave a durability comparable to that of thecoloured glass.

Mirrors

In the manufacture of seventeenth-centurymirrors on plate glass a piece of tin foil,approximately a tenth of a millimetre thick,was laid on a perfectly flat piece of marble.The protective layer of tin oxide was removedfrom its surface by rubbing mercury over thefoil until the surface was shiny. Next a layerof mercury was poured over the foil to adepth of 3–5 mm (this was the quantity ofmercury which could lie on a horizontal tinplate, and corresponded to circa 50 kg ofmercury per square metre of glass). A scum oftin oxide accumulated on the surface of themercury and was removed with a glass edge.Meanwhile, the glass sheet had beenthoroughly cleaned by rubbing it with washedash and polishing it with a linen cloth. In asingle movement, the glass sheet was slid overthe mercury with the leading edge dipped intothe mercury and without scraping the tin. Inthe case of small mirrors, a thin sheet of paperwas laid on the mercury, the glass then laidon top and gently pressed down whilst thepaper was drawn out. The glass was thenpressed down to make good contact with thetin foil. A great deal of mercury ran out, butexcess was removed, by weighting the glass,


and then tilting the marble support. Theamalgam process began the moment themercury came into contact with the metal foil.Complete conversion of the tin to tin–mercurycompound occurred within 24 hours. Theweights were then removed and the mirrorlifted off the marble support. However theamalgam was still very soft and mobile so thatthe mirror had to immediately be laid flat,glass side down, and then tilted very slowly,first toward one side and then to the corner,so that enough of the liquid phase could flowand evaporate away to leave a solid, stableamalgam layer. The hardening process tookup to a month, depending on the size of themirror (Figure 3.46).

The tin–mercury amalgam mirror is a two-phase reflective coating, the mercury havingreacted with the tin to form a layer of crystalscontaining about 19 wt per cent of mercuryalloyed with the tin. The voids between thecrystals are filled with a fluid containing about0.5 per cent tin in mercury. The production ofamalgam mirrors was technically difficult,dangerous and time-consuming and thereforecostly. The glass itself had to be colourless,

free of bubbles and of uniform thickness,before being ground and polished on bothsides. Application of the amalgam involved theuse of a large quantity of mercury, whichresulted in the workers suffering from mercurypoisoning.

The technique of producing amalgammirrors remained largely the same, with onlyminor developments. An eighteenth-centuryEnglish account of mirror-making mentions themixing of lead, tin and marcasite (probablyantimony or bismuth) with the mercury inorder that ‘the glass may dry sooner’. Severallater sources mention that the tin foil cancontain up to 2 per cent copper, which resultsin a better adhesion to the glass. Howeversuch additions were probably uncommon.

There were two methods of making curvedmirrors. A free-flowing liquid amalgam wasformed by melting bismuth, lead and tintogether and then mixing them with mercury.Whilst still warm, the mixture could be pouredinto concave glass or into hollow glass balls,and then poured out again, leaving a thin filmon the glass which hardened on cooling toproduce a convex mirror. Concave mirrors were


Figure 3.46 Glass mirror manufacture in the mid eighteenth century. A flat piece of glass which will form thebase of a mirror. The uneven surface is ground and polished. The glass is slid onto a sheet of tin foil coated withmercury, and weighted with rocks. A rim around the table prevents the liquid from running off. When the excessmercury has been squeezed out, the finished products are stacked upright to drain further and to harden.

formed by pouring the same amalgam into agap between curved glass and a varnishedplaster of Paris mould. Pure tin amalgam couldalso be used to make curved mirrors by firstpressing tin foil into a plaster mould taken fromthe curved glass. Mercury was then poured intothe foil to produce a concave mirror or into thecurved glass, after which the foil was pressedinto place, to form a convex mirror. The glassand the foil-covered mould were then pressedtogether. During each process, the glass had tolie uppermost during the amalgam maturingprocess to allow the tin alloy crystals to floatupwards in the heavy mercury to form themirrored surface against the glass.

In the middle of the nineteenth century thetechnique of depositing a thin layer of silveron glass by adding an aldehyde to a silvernitrate solution was discovered. This methodwas quick and relatively safe; however the firstsilver mirrors were not durable, and so did notcompete effectively with tin amalgam mirrorsuntil around 1900. Tin amalgam mirrorscontinued to be made in the first decades ofthe twentieth century (Schweig, 1973; Child,1990).

Tin amalgam and silver mirrors differ fromone another in colour and reflectivity; tinamalgam has a bluish tint and reflects signif-icantly much less light than silver mirrors,which appear slightly yellow. The glass itselfalso contributes to the difference in appear-ance since the older glass is generally darker(more grey). However, it must be remem-bered that both types of mirror becomedarker and foggier with the onset of corro-sion. Hadsund (1993) suggests that it issometimes possible to differentiate betweenthe two types of mirror by viewing themthrough a piece of thin (tracing) paper, whenthe silver mirror will appear brighter.Amalgam mirrors were very seldom painted atthe time of manufacture, normally only whenthey were for use in damp rooms or at sea.Sometimes mirrors are found which havebeen painted on the reverse in an attempt atpreservation of the mirrored surface. Thereverse side of silvered mirrors was alwaysprotected by several layers of paint, based onred lead. Modern silvered mirrors are copperplated before being painted in order toprovide greater protection.


Mesopotamian glassmaking

The earliest records of glass technology, in theform of cuneiform tablets, were found inMesopotamia. The earliest text giving a recipefor glaze was found near Tell ‘Umar (Seleucia)on the River Tigris. The tablet dates from theseventeenth century BC but presumes a well-established glassmaking tradition. Amongst themany thousands of Assyrian cuneiform tabletsexcavated from the site of the library ofAshurbanipal (668–627 BC), a small numberrecord information concerning the manufac-ture of coloured glass and glazes. Thesetablets were evidently copies of originalsproduced in the last centuries of the secondmillennium BC, showing that the essentialprinciples of glassmaking were understood inabout 1700 BC. An important philologicalaspect of the cuneiform glassmaking texts isthat the seventh-century copies contain the

earlier Sumerian words and their laterAkkadian translations; moreover, the formerare related to words used in Ur III texts andthe latter are related to words used by glass-makers on the Phoenician coast in Romantimes. Thus the glassmaking tradition seems tohave been an extremely conservative one, thesame terms (and no doubt the same tried andtested formulae) being handed down for morethan two millennia.

Comparatively few glass artefacts have beenexcavated in Mesopotamia, but this may wellbe due to the moisture content of the soil, andto the rise of the water-table in historic times,causing the destruction of much of the earlyglass which was inherently unstable in itschemical composition (and therefore slowlysoluble in water) (Oppenheim, 1973). There isalso the fact that glassmaking would have onlybeen carried out on a small scale, and thatwasters would have been reused as cullet.

Part 2: Furnaces and melting techniques

The cuneiform texts with specific referencesto glass have been translated at various times,notably by Thompson (1925), Oppenheim etal. (1970) and Brill (1970b,c). However, trans-lation is notably difficult. The earlier transla-tions contain many errors, based on a lack ofunderstanding of glassmaking techniques. Thephilologist needs the help of a glass technol-ogist to interpret the methods and materialsdescribed; the glass technologist who advisedThompson was not aware of the compositionof the plant ash used, and its profound effecton the composition of the glass. Thompsonerroneously referred to the use of humanembryos ‘born before their time’ as aconstituent of Babylonian glass. In Oppen-heim’s translation, however, the relevant wordis kubu (images); images were in fact set upas part of the ritual for building a furnace.Thompson also referred to the use of arsenicand antimony in glassmaking, but Turner(1956a) pointed out that the arsenic is only atrace element in Babylonian or Assyrianglasses, and Brill (1970b,c) concluded thatsince there was no description of ingredientscontaining antimony in cuneiform texts, therewas no reason to believe that the Assyriansknew of antimony as such. Similarly, it seemsthat Douglas and Frank (1972) were misledinto believing that lime was specified as aningredient of Assyrian glass. Despite theirtechnical content, the texts were not technicalinstructions, but, ‘They have to be considered,strange as it may seem, as literary creationswithin a complex literary tradition ... [were]subject to certain stylistic requirements; theirwording and their literary forms were histori-cally conditioned ...’ (Oppenheim et al., 1970).

In Mesopotamia there were many wordsused for glass, glass intermediates and glass-like substances (Oppenheim et al., 1970). Acoherent interpretation of the cuneiform textswas made by Brill (1970b,c). As a result ofinvestigation, it seems that the naga plant maywell have been Salsola kali or a similar plant;and that ahussu, the plant ash prepared fromit, probably contained about 55 per centNa2CO3, 8 per cent KCl, 8 per cent MgO, 4per cent NaCl, 5 per cent CaSO4 etc.Immanakku was ground quartzite pebblesfrom a river bed, probably containing 95 percent SiO2, 2.5 per cent A12O3 etc.; and zukuwas an intermediate reaction product (a frit)

obtained by heating the silica with plant ashat ‘a heat which has the colour of the red ofred grapes’, that is, probably less than 850°C.After fritting was complete, the cold zuku wasground up and the colouring agents wereadded (see below); the mixture was thenheated to ‘a heat which is yellow’ (probablyabout 1000°C) to produce the final glass.

Chief among the materials added at thissecond stage of glass melting were urudu.hi.a(slow copper) and sipparuarhu (fast bronze).These materials have not yet been identifiedwith any certainty but it is possible that slowcopper was copper oxide (CuO) and that fastbronze was a (metallic) alloy of copper, tinand lead which would melt, form a layerbelow the glass, and confer a blue colour toit (Brill, 1970b).

Having established the probable nature ofthe ingredients of zuku, Brill (1970c) thenheated them together for 10 hours at 920°C toobtain a frit, which was cooled, pulverizedand then heated for 16 hours at 1100°C toproduce a well-melted homogeneous piece ofpale bluish glass of very good quality. Theresultant zuku glass contains 56.0 per centSiO2, 23.8 per cent Na2O, 6.6 per cent CaO,5.6 per cent MgO, 3.8 per cent K2O, 2.2 percent Al2O3 and various minor ingredients. Thusthe glass was not unlike typical Mesopotamianglasses; and from its composition it would beexpected to have a reasonable durability.

Mesopotamian furnaces and meltingprocedures

The religio-magic preparations associated withthe setting up of a furnace, described in trans-lation by Oppenheim (1973), reflect the limitsof the Mesopotamian glassmakers’ technicalknowledge, which was essentially empirical innature.

According to Thorpe (1938) it would seemas though there were three types of furnacedescribed in the cuneiform texts. Theglasshouse (the bît kûri or house of thefurnaces) contained the kûri sa abni orfurnace for the pot metal in which the glassbatch was formed. The kûri ša siknat ênâit-pel-ša (furnace with a floor of eyes) was thefounding furnace (‘that where the workmenwork’). A much later usage of the term ‘eye’in this context is that of the Italian occhio or


lumella, that is the circular opening betweenthe siege (middle or founding) storey of thethree-storey furnace and the upper (annealingor tower) storey. The Mesopotamian furnacecould probably achieve temperatures as highas 1000–1100°C. Finally the kuri sa takkanni(furnace of the arch) appears to have been adoor (bab kûri) through which the finishedarticles were introduced. This corresponds toa lehr or annealing furnace.

The fire was made of logs from poplar trees,found growing along the River Euphrates, andthe duration of the found was, in certain cases,as long as seven days, during which it musthave been difficult to maintain the hightemperatures required. However, there is nodoubt that by the seventh century BC,Mesopotamian glass furnaces had developedenough for this to be possible. Although nodescription exists of furnaces constructedspecifically for glassmaking, and no furnaceshave been excavated, Forbes (1966) suggestedthat the reverbatory furnace (a preconditionfor glass-blowing), may have originated inMesopotamia.

The texts make no distinction between thebatch of unmelted materials and the frit, bothbillu and abnu (stone) being used to describethe mix. The glass pot or crucible (taptuzakatu) had to be clean and it was stilted(nimedu, stilt or support) so that it did nottouch the furnace ceiling. Several types ofmould appear to have been used (open andclosed), the moulding processes possiblyhaving been derived from bronze-working.The cuneiform texts also mention the hook orrake (mutirru, Syr. mattara, Lat. rutabulum),and a ladle (su’lu) for moving molten glassfrom large glass pots to smaller ones and forskimming the batch. When producing the frit,the instructions were:

When the glass assumes the colour of ripe(red) grapes, you keep it boiling (for a time)[this is probably the evolution of carbondioxide from the reaction between thesodium carbonate and the silica]. [Then] youpour it [the glass] on a kiln-fired brick ... Youput [it] into a kiln which has four fireopenings and place it on a stand ... You keepa good and smokeless fire burning [so thatthe flames come out of the openings] ... Notuntil the glass glows red do you close the

door of the kiln and stir it once ‘towards you’[with a rake] until it becomes yellow [hot].After it has become yellow [hot], you observesome drops [forming at the tip of the rake].If the glass is homogeneous [without bubbles]you pour it [inside the kiln] into a new dabtu-pan ... (Oppenheim et al., 1970)

The constructional differences between thefritting furnace and the fusion furnace havebeen noted by Oppenheim et al. (1970). Thereare a number of technologically importantobservations to be made here: the use of ared heat (less than 850°C) for the first frittingand a yellow heat (1100°C) for the secondfiring, and also the use of a smokeless fire, sothat reducing conditions were avoided. Byexperiment, these conditions can now be seento be important for the success of the opera-tion. Nevertheless, the variability in composi-tion of the plant ash must have been afrequent source of failures. Analyses of plantash from bushes, bracken and seaweedshowed that the soda content could varywidely, from 14.2 to 42.5 per cent. Eachcomplete glassmaking operation might haveproduced about 3500 cm3 of zuku frit and800 cm3 (about 2 kg) of the copper-containingred glass (Brill, 1970b,c).

Mesopotamian glass

Transparent glassesIn view of such limited knowledge, it isunlikely that the glassmakers could make theslight deliberate additions of antimony ormanganese which, it has been suggested, wererequired to decolourize the glass (Sayre, 1963).Such additions were probably accidental asimpurities in the raw materials forming theglass batch (Newton, 1985b).

Analytical results for Mesopotamian trans-parent glasses are quoted by Turner (1954b)and by Brill (1970c). The glasses would beexpected to be reasonably durable. Twosamples quoted by Turner (1954b) showedweathering, equivalent to losses of substanceof 0.18 mm per century (sample A) and0.04 mm per century (sample B). However, allthe non-red glasses from this era have muchthe same composition, and Turner (1954b)commented that it is a remarkable fact thatglass from Nimrud, from Eighteenth Dynasty


Egypt, and from Knossos in Crete have somuch resemblance in composition despite thefact that they are spread over 700 years and adistance of 1700 km. This could be explainedby the possibility that all the glassmaking wascarried out in Syria, and only glass-meltingwas performed in the other centres. A darkblue ingot found at Eridu (about 175 km WNWof Basra) and dated to about 2000 BC (Brill,1970b,c), seems to have been an ingot forremelting and this aspect of glassworkingoccurs repeatedly in the history of glass – thatis, that the secret of making good quality glasswas by no means known to all glassworkersand many of them, perhaps even most, had toobtain their best glass from other sources(Forbes, 1966, 1961; Brill, 1970c), and this stillapplied almost four thousand years later(Charleston, 1963, 1967; Oppenheim et al.,1970; Newton, 1971b,c).

The possibility that the raw glass (cullet)was all made in Upper Syria, and that it wasbased on a traditional recipe, with manufac-turing processes that were well-guardedsecrets, may be a more feasible explanation ofthis uniformity in composition than the alter-native hypothesis that different glassmakersexperimented with different local raw materi-als, even though they might be following atraditional recipe. (However, it has to be bornein mind that any glass having a distinctlydifferent composition from those survivingmay have had such a poor durability that itmay have entirely perished during the inter-vening millennia.)

Opal glassesThere are two kinds of Mesopotamian glasswhich deserve special mention: the green and yellow opals which contain yellow lead antimonate (Pb2Sb2O7) and the copper-containing sealing wax red glasses which hadto be used in a special manner (in this earlyperiod they do not contain much lead, incontrast to the high-lead sealing wax red ofthe seventh century BC). The use of leadantimonate is particularly interesting because itseems that the Assyrians had no direct knowl-edge of antimony as such.

Brill (1970d) points out that the yellowPb2Sb2O7 was added to the blue transparentglass (tersı�tu) to form a greenish-coloured

artificial lapis lazuli, and hence the leadantimonate is likely to be either anzahhu orbu�su. It also seems that anzahhu wasprepared by craftsmen other than the glass-makers and that it had been known over aperiod of fifteen centuries; it could thereforehave been an article of trade and its use bythe Assyrians need not imply a knowledge ofthe properties of antimony, either as an opaci-fier or as a decolourizer. The sealing wax redglasses are of even greater interest; theircolour is due to the presence of colloidalcuprous oxide (Cu2O) and its manufacture hasits difficulties even today. It must be preparedunder suitable reducing conditions, or theCu2O will be oxidized to give a transparentblue glass (Brill, 1970d), and the cuneiformtexts specify that closed containers and longfiring times should be used, and that the glassshould be cooled within the kiln.

Cable (personal communication, 1979)melted some of these glasses and points outthat the cast slabs all have a black tarnishedmetallic lustre. If the surface film of metallicoxide was ground away, and the piecesreheated, the surface blackened again within2 mm at 550°C, at which temperature the glassis still quite solid, and thus such a glass couldnot have been hot-worked, as was the casewith the other glasses. Turner (1954b) statedthat, in 1952, the archaeologist Mallowandiscovered traces of furnaces at Nimrud, andnearby cakes of sealing wax red glass hadfragments of charcoal on them (which wouldhave helped to maintain the necessary reduc-ing conditions until the glass was cold).Secondly, that from such cakes of opaqueglass Egyptian craftsmen cut thin plates, whichwere ground and polished to use as inlays onfunerary furniture. It thus seems likely that thesealing wax glasses were used only forlapidary purposes and not for hot-working.Brill (1970c) also remarked on the dark surfaceof the glass.

Egyptian furnaces and meltingprocedures

Due to lack of archaeological evidence beforethe late twentieth century, it had been gener-ally held that the Egyptians had failed to learn


the secret of glassmaking, despite the fact thatthey were skilled craftsmen in general, andincluded glassmakers, i.e. makers of artefactsfrom glass cullet, imported from Upper Syria(Turner, 1954a). However, work carried out inEgypt in 1993 and 1994, which included theconstruction and firing of a glassmaking kiln(see below), has demonstrated that the ancientEgyptians did in fact have the technology toproduce glass from its raw materials(Nicholson, 1997). Reuren et al. (1998)published details of a glass colouring workswithin a copper-centred Bronze Age industrialcomplex at Qantir in the Nile Delta.

Oppenheim (1973) gives evidence forstating that ‘The craftsmen who produced themagnificent glass objects for the EgyptianCourt depended for their basic raw materials,or for the essential ingredient thereof, onimports from Asia’. It is evident, from theurgency of the appeals from the Kings ofEgypt for ehlipakku and mekku (both wordsfor raw glass, the former of Hurrian origin andthe latter West Semitic) that the Egyptians hadinitially failed to learn the secret of theAssyrian fritting technique. Communicationbetween Egypt and Mesopotamia would havebeen infrequent, and of course the Meso-potamians would have guarded their glassproduction recipes, thus maintaining the statusquo.

In the light of the foregoing, it is perhapssuprising that the earliest Egyptian glasswork-ing complex known, at Tell el-Amarna(excavated by Flinders-Petrie in 1891), is datedover 100 years later than the discovery ofglassworking in Egypt. Tell-el-Amarna, thenew capital of Akhenaton, required a largeamount of decorative work, with the resultthat factories sprang up to supply the materi-als. Glazes and glass were the two principalmanufactures, in which a variety and brilliancywas achieved that was never reached in earlieror later times in Egypt. So far as the use ofglazes is possible, this period shows thehighest degree of success, and the greatestvariety of application.

The sites of three or four glass factories, andtwo large glazing works were discovered; andalthough the actual workrooms had almostvanished, the waste heaps were full offragments which showed the types of productand their method of manufacture. Frits made

in Egypt from the Twelfth Dynasty onwardwere composed of silica, lime, alkalinecarbonates and copper carbonate varying from3 per cent in delicate greenish blue, up to 20per cent in rich purple blue. The green tintswere always produced when iron was present,which was usually the case when sand wasthe source of silica.

One of the first requisites therefore was toobtain raw materials free from iron. Thequestion of how this was achieved wasanswered with the discovery of a piece of apan of frit, which had been broken in thefurnace and rejected before the frit hadcompletely combined. This contained chips ofwhite silica throughout the mass, which wereclearly the result of using crushed quartzpebbles as a source of silica instead of sand.The lime, alkali and copper had alreadycombined, and the silica was in the course ofsolution and combination with the alkali andlime. The carbonic acid in the alkali and limehad been partly liberated by the dissolvedsilica, and had raised the mass into a spongypaste. With longer continued heating the silicain other samples had entirely disappeared, andformed a mixture of more or less fusiblesilicates. These made a pasty mass when keptat the temperature required to produce thefine colours. The mass was then moulded intopats, and heated in the furnace until thedesired tint was reached. On being cooled, asoft, crystalline, porous friable cake of colourwas produced.

Amongst the furnace waste were manywhite quartz pebbles. These had been laid asa cobble floor in the furnace, and served as aclean space on which to roast the pats ofcolour, scraps of which were found adheringto the cobbles. The floor also served to layobjects on for glazing; superfluous glaze hadspread over the pebbles in a thin green wash.Doubtless this use of the pebbles was two-fold; they provided a clean furnace floor, andthey became disintegrated by the repeatedheating so that they were the more readilycrushed for mixture in the frits afterwards.

The half-pan of uncombined frit shows thesize and form of the fritting pans: about254 mm across and 76 mm deep. Among thefurnace waste were also many pieces of cylin-drical jars, about 178 mm across and 127 mmhigh. These jars almost always bore runs of


glaze on the outside from the base to the rim;the glaze, being of various colours, blue,green, white, black etc., evidently leaked fromthe pans, hence the jars must have stoodmouth downward in the furnace to supportthe fritting pans and glass crucibles above thefire.

Of the furnaces used for glass-melting, therewas no certain example; but a furnace discov-ered near the great mould and glaze factorycontained a great quantity of charcoal, but notrace of pans, jars or glass. The furnace wasan irregular square varying from 1092 mm to1448 mm at the sides. It was originally about889 mm high, but the roof had beendestroyed. The northern door was 737 mmhigh and 381 mm wide, to admit the northwind and to serve for tending the furnace onthe windward side. The south or exit door was406 mm high and 330 mm wide to allow gasesto escape. Probably the glazing furnaces werebased on the same principle, and perhapseven the same furnace would be used forvarying purposes.

Of the stages of production of the glassthere is ample evidence. The crucibles inwhich it was melted were deeper than thefritting pans, being about 58 and 76 mm indepth and diameter. Their form was knownfrom the shape of unused pieces of glassshowing the section of the vessels in whichthey cooled. Many such pieces of glass werefound retaining the rough surface, and evenchips of crucible adhering to them, while theancient top surface shows the smooth meltedface, with edges drawn up by capillary action.

The upper part of the glass was often frothyand useless. The presence of the froth (ofcarbonic acid expelled by the meltingreaction) suggests that the glass was fused inthe pans. The manner in which the cruciblehad been chipped off the lump of solidifiedglass in every case shows that the glass wasleft to cool in the crucible so as to allow thescum gradually to rise and the sediment tosink. If the glass had been poured out thesefeatures would not have been found, on thecontrary masses of cast glass should have beenin evidence. While the glass was being madesamples were taken out by means of pincers,to test the colour and quality, and many ofthese samplings were found showing theimpression of the round-tipped pincers.

Analysis of the crucibles showed that theywere not made of clay in the true sense butseemed to consist of mud and sand mixedtogether. Heating trials showed that thecrucibles would have begun to vitrify at atemperature of 1100°C and that they fused toa black mass after 1 hour at 1150°C. It seemslikely that the highest temperature forprolonged heating would be lower than1100°C, and since no really high temperaturecould have been attained therefore, the glasswas probably worked in a pasty state, mainlyby drawing out threads and constantly marver-ing them on a flat slab. This supposition seemsto be borne out by the excavated evidence(Petrie, 1894).

Petrie’s excavations have since been supple-mented by those of the Egypt ExplorationSociety (Nicholson, 1995; Nicholson and


Figure 3.47 Kilns excavatedat Tell el-Amarna, in Egypt,site 045.1. Kiln 3 (left) andkiln 2 (right) are believed tohave been used for meltingglass. An extensive layer ofvitrified clay can be seen onthe right side of kiln 3. Thescale rods are 2 metres long.(Courtesy of P.T. Nicholson,Egypt Exploration Society).

Jackson, 1997). Jackson et al. (1998) believedthat the long-held contention that all raw glasshad to be imported from Syria might be incor-rect; and remarked that neither Petrie nor hiscolleague Howard Carter had given the exactlocations for the glazing factories, which hadbeen discovered in the 1890s. A protonmagnotometer survey was undertaken in anarea believed to be near one of Petrie’sexcavations. This identified an area thoughtworthy of excavation; preliminary work wasundertaken in 1993 and was followed up thefollowing year (Nicholson, 1997).

During the 1994 excavations, two largefurnaces were revealed (Figure 3.47), roughlycircular, with a 1.5 m internal diameter, andformed of three layers of brickwork. Thefurnaces contained vitrified material (knownlocally as khorfush), which proved to be theremains of vitrified brick and of a ‘sacrificialrender’ (lining material), making it clear thatthe furnaces had not been used for smeltingmetal. In order to determine whether suchlarge furnaces could have reached andmaintained the high temperatures required forglassmaking, a replica was constructed, usinglocal materials as representative as possible ofthose used in ancient times, and glass made init. The raw materials used were plant ashprepared from seaweed from Penarth (SouthWales, UK), local sand and modern cullet inthe form of crushed green bottle glass, and thefuel was wood. After five hours’ firing, temper-atures of 1100–1150°C were achieved, and asatisfactory glass obtained. The experimentthus demonstrated that glassmaking could havebeen in existence in Egypt by circa 1350 BC.

Glassmaking during the RomanEmpire

In contrast with the considerable amount ofglass surviving from the Roman period, thereis a dearth of evidence regarding the actualglassworking operations (Strong and Brown,1976). Remains of a Roman furnace site atTrier excavated in 1922 helped to identify thetype of crucible used. Iron tubes excavated atBadajoz (Spain), may have been glassblowingpipes (Lang and Price, 1975).

Unlike Roman pottery kilns, which weresunk in the ground, glass furnaces were built

at ground level or even raised above it. Theglass furnaces mentioned by Pliny (AD 77)were probably small beehive-shaped hearthfurnaces with one or two compartments forannealing the glass. The earliest contemporaryevidence of Roman glassworking would seemto be a representation on a clay lamp, attrib-utable to the first century AD, of what mayreasonably be taken to be two glassworkers ata furnace. The details are far from clear, butit seems that the furnace was in two tiers atleast; presumably a stoke-hole below (repre-sented by the filling hole in the lamp itself)and a chamber above (Figure 3.48).

During the latter half of the twentiethcentury, excavations in many countriesformerly under Roman occupation, haverevealed the remains of glassworking opera-tions (Jackson et al., 1991; Seibel, 2000;Nenna, in Kordas, 2002; Skordara et al., inKordas, 2002). However, the degree to whichglassworking of chunk glass imported fromMediterranean glassmaking sites was carriedout, and actual manufacture of glass from itsraw materials in the north is still under consid-eration.

Glassmaking had spread to Gaul (France)and the Rhineland (Germany) by at least 50AD. The structural remains of Roman glass


Figure 3.48 A scene on the discus of a lamp showingtwo glassworkers at a glass furnace. First century AD.Asseria, Dalmatia.

furnaces had been excavated at Eigelstein nearCologne before the Second World War (Figure3.49). Although the excavation plans of thesite were destroyed during the war, it isknown that the remains represented the lowercourses of both circular and rectangular struc-tures, somewhat separate from each other, butwithout any clear indication of their relation-ship. The excavations showed that thefurnaces had been rebuilt, each rebuildingbeing carried out over the previous furnaces,which had been razed to their foundations.Between the layers fragments of completelycolourless glass were found (Doppelfeld,1965).

Five glassworking sites excavated in theHambacher forest area of the Rhinelandrevealed two types of furnace, the one circu-lar and the other semi-circular, along withrefractories and glass articles. The raw mater-ial may have been chunk glass, possiblyimported from the Mediterranean (Hadera,Israel) (Seibel, 2000).

In Britain, the excavation of putative Romanglass sites found at Wilderspool andMiddlewich in Cheshire, Mancetter inWarwickshire, Wroxeter in Shropshire andCaistor St Edmund in Norfolk have revealedlittle structural detail. Five workshops werefound at Wilderspool; the furnaces weredescribed as small oval ovens with outlets andflues. Work in 1964–5 and 1969–71 on aRoman glass site at Mancetter revealed a small,almost circular (880 mm � 770 mm) furnace(Hurst Vose, 1980). It is presumed that theglass made in this furnace was melted fromcullet.

Brill (1963) carried out melting experimentsusing a gradient furnace, and concluded thata typical silica–soda–lime Roman glass wouldrequire a final melting temperature (i.e. whenground up and remixed after the fritting opera-tion) of 1100°C; and that the glass would havehad to be held at a temperature of at least1080°C for satisfactory glass-blowing opera-tions to be performed.

Near Eastern glassmaking

Surviving in situ in a cave in the ancientJewish necropolis of Bet She’arim, Israel, is amassive glass slab, measuring approximately3.40 � 1.95 � 0.45 metres, and which isestimated to weigh about nine tonnes. It wasextensively investigated by Brill and Wosinski(1965), and was originally dated to the fourthcentury AD. The date now assigned to the slabis early ninth century AD. The glass wasobviously melted in a tank furnace in site,however the process failed due to the compo-sition of the glass, which contained far toomuch lime so that it either did not melt fullyor devitrified extensively upon cooling.

Although the slab is the only one intact andin situ, it is not unique in the ancient glass-making world. A group of seventeen tankfurnaces, each of which produced large slabsof glass has been excavated at Bet Eli’ezernear Hadera (Israel) (Figure 3.50). On thebasis of associated finds, the group is dated tothe sixth to seventh centuries AD (Perrot, 1971;Freestone and Gorin-Rosen, 1999). Thefurnaces seem to be connected to the produc-


Figure 3.49 Plan of thefurnaces excavated atEigelstein near Cologne,Germany, showing successiverebuildings of the furnaces onthe same site. Overallmeasurements: c. 4.9 m.

tion of primary unformed glass from localsources of sand and alkali, and their situationto be a consequence of the nearby source offine sand, with a long proven used for glass-making, at the mouth of the River Belus.

Evidence for primary glassmaking is alsofound on a large scale in the region ofAlexandria and the Wadi el-Natrun in Egypt.It seems likely that in the late Byzantineperiod (sixth to seventh centuries AD, andperhaps earlier), raw glass was manufacturednear the source of raw materials and thenexported for fabrication into glass objects atother sites.

The analysis of chunk glass from Bet Eli’ezer,Dor and Apollonia, all dated to the sixth andseventh centuries, and later Islamic glass fromBarias (eleventh to thirteenth centuries), and agroup of glass vessels from Ramla (eighth toeleventh centuries) showed them all to besilica–soda–lime glasses; but that the levels ofpotash and manganese were much higher inthe later Islamic glass from Barias than in theByzantine glass. The transition from the lowermanganese/high potash composition to thehigh manganese/low potash composition hadbeen observed by Sayre and Smith in 1961,and is generally considered to reflect a changeof raw materials.

Throughout the Roman and Byzantineperiods, glasses produced in the Mediterra-nean area were based on the use of natron, arelatively pure mineral soda, as a flux. How-ever, in the Islamic period, a plant ashcontaining potash, magnesia and lime inaddition to soda, was introduced (Henderson,1985). Plant ash had been used to make glassin Mesopotamia in the Parthian and Sassanianperiods (Smith, 1963a,b; Sayre, 1965), so thischange in the source of alkali does not repre-sent an entirely new glass technology. InEgypt, the changeover from natron based, lowmagnesium glass to a high magnesium plantash has been dated to the middle of the ninthcentury AD, based on the analysis of glassweights inscribed with the names of officials,and therefore datable (Sayre and Smith, 1974;Gratuze and Barrandon, 1990). The two typesof glass have been found together in eighth-to ninth-century dumps at Raqqa in Syria(Henderson, 1995); and glass samples fromRamla indicate that the new high magnesiumcomposition was introduced in the eighth andninth centuries with some overlap in the useof natron-based glass (Freestone, 2001).

The major change in glass composition,which occurred in the Near East during theeighth and ninth centuries, had an impact asfar away as northwestern Europe, where achange from natron-based silica-soda–limeglass to a potash-rich forest glass compositionwas taking place about the same time (Hunter,1981; Henderson, 1991, 1993a).

Paradoxically, it is only with the Dark Agesthat evidence becomes available to indicatewhat early glassmaking furnaces were like.Under the unifying influence of the RomanEmpire, glassmaking practices in differentparts of the Empire remained much the same.Around the time of the collapse of the Empire,however, a marked cleavage in glass technol-ogy occurred, the origins of which are still notcompletely certain. Different styles of glass-making furnace developed north and south ofthe Alps and it has been suggested that thiswas the result of the supply of maritime plantshaving become scarce at a time when therewas a great demand for coloured glass innorthern Europe. This caused the glassmakersto migrate to forested areas where a supply ofalkali could be obtained in the form of plantash. The distribution of glassmaking sites north


Figure 3.50 Schematic plan and reconstructed sectionof a sixth- to seventh-century AD furnace at Bet Eli’ezer,Hadera, Israel (Drawing: Michael Miles). (CorningMuseum of Glass).

of the Alps closely corresponds to the distrib-ution of beech woods (Barrelet, 1953; Newton,1985b). Newton (1985b) has suggested that thenorthern type of furnace (which had a gooddraught) was adapted for using beechwood,whereas the southern type of furnace (in areasdeficient of beech forests) were able tocontinue using traditional sources of alkali,and continued to be of the traditional furnacedesign.

The southern European type ofglassmaking furnace

The first reasonably full description of a south-ern type of furnace is given in a Syrianmanuscript (in the British Museum), whichapparently cannot be dated earlier than theninth century AD:

The furnace of the glassmakers should havesix compartments, of which three aredisposed in storeys one above the other ...the lower compartment should be deep, in itis the fire; that of the middle storey has anopening in front of the central chambers,these last should be equal, disposed on thesides and not in the center[?], so that the firefrom below may rise towards the centralregion where the glass is and heat and meltthe materials. The upper compartment, whichis vaulted, is arranged so as uniformly to roofover the middle storey; it is used to cool thevessels after their manufacture.

Not all the details of this account are clear,but the essentials are that the furnace was inthree storeys, with a fire chamber at thebottom, a central chamber into which the heatrises to melt the glass, and a vaulted uppercompartment in which the glass may cool.

This arrangement may be seen in the earli-est certain representation of a glass furnace(Figure 3.51) in a manuscript dating from 1023in the library at Monte Cassino, a text of thework De Universo by Hrabanus Maurus(Archbishop of Mainz c. 776–856). Again thedetails are not unequivocally clear, but theartist appears to have attempted to representa cylindrical structure, although the roof isshown as a simple tent shape. The manuscriptmay have been a copy of an earlier one dating

from the fourth or fifth century AD. The mainbody of the furnace has been interpreted asbeing rectangular in plan, but this seemsunlikely from the elliptical rendering of theglory holes, which would certainly not in anycase be made at the corners of a rectangularstructure; furthermore, the tent-like upperstorey is carefully depicted as being adaptedto the curved structure below. The essentialfeatures, however, are these: a single stoke-hole at the bottom, a middle chamber withmultiple glory holes giving access to glasspots, and an upper compartment in which aglass may be seen annealing. If indeed thefurnace represented was cylindrical or roundin section, the illumination provides a mostuseful link, for on the one hand the laterfurnaces of this type were almost always circu-lar, while on the other hand the remains of aseventh/eighth-century glassmaking complexon the Venetian island of Torcello, excavatedin 1961/2 (Gasparetto, 1965, 1967; Taba-czynska, 1968), include the foundations of acircular structure which may have been afritting furnace. However, it seems far morelikely to have been the main (founding)furnace since it stood in the middle of thecomplex, as would be natural for the workingfurnace to which all surrounding structures are


Figure 3.51 The earliest illustration of a glass furnacethree tiers high. The craftsman on the three-legged stoolis blowing a vessel, another vessel can be seenundergoing the annealing process in the topcompartment of the kiln. Illustrated miniature from themanuscript of Hrabanus Maurus, De Universo, datable to1023 AD. Abbey of Monte Cassino. (Codex 132).

ancillary. A study of the site shows thattemperatures of 1270°C had been reached inthe melting furnace.

The glass mosaics used for decorating theBasilica of Santa Maria Assunta on Torcellostarted in AD 639 were probably made in thefurnaces described above, but the glassworkswas so close to the church (only about 35 mto the west of its walls) that it seems likelythat it was dismantled when the basilica wascompleted. The spread of broken pots, glasswaste etc. on the site was probably the resultof clearance and disturbance at the time ofdemolition.

On a glass site at Corinth in Greece(Weinberg, 1975) (Agora, South Centre site)however, dating from the eleventh/twelfthcentury, the foundations of a square furnacewere unearthed which, from the absence ofany ancillary furnace, were reasonably takenby the excavators to have been the loweststorey of a three-tiered arrangement (Wein-berg, 1975). The original conclusions havebeen modified by later finds, notably somelimestone blocks covered with glass drippingswhich seem to suggest that either small rectan-gular tanks (320 mm � 270 mm) or squarecontainers were used in the glassmakingprocess. Finds of refractory pots with glassadhering, however, found in association withthe Agora north-east glassmaking complex (ifsuch it is) suggest the more normal method offounding the metal. Fragments of glass foundin the Agora South Centre have been reinter-preted as pontil wads and suggest that (unlessthey were cullet, which seems somewhatunlikely) glass was actually worked in thevicinity of this furnace (i.e. that it was notmerely a fritting or annealing furnace). Thefact that there was so little space surroundingthe furnace structure (barely 500 mm on threesides) is the strongest counter-indication.

A furnace excavated at Monte Lecco in theApennines about 30 km from Genoa, andprobably dating from the late fourteenth orearly fifteenth century clearly revealed a circu-lar structure, with a central fire trench runningbetween two roughly segmented solid sieges(banks on which the glass pots stood). Thefire trench was dug down into an ash pit inthe ground at the front, and had an outlet atthe bank, presumably for the clearance ofashes. The furnace has been tentatively recon-

structed as a three-tiered structure on the basisof a picture, perhaps dating from circa 1590,in the Oratorio of St Rocco at Altare. Theabsence of any structures in the immediatevicinity of the furnace suggests that all theprocesses were carried out in the one furnace,and that this was therefore likely to have beena three-tiered structure.

As the finds at Torcello suggest, the circu-lar furnace (perhaps in three storeys) wasprobably of the first type used by the Venetianglassmakers who, from the mid-fifteenthcentury at the latest, began to dominate theworld markets with their superior crystal glass.Two illuminations in manuscripts in theVatican library portray glassmakers at work(Charleston, 1978). The cruder of the twoshows two glassblowers sitting on three-legged stools, while another man attends tothe stoke-hole. On the furnace is writtenunequivocally fornax vitr (glass furnace). It isrepresented clearly as three storeys, the uppertwo set slightly back, and this detail isrepeated in the second miniature. Here,however, four out of the six ribs are visible,rising from the broader ledge just below thelevel of the bocca. This ledge no doubtprovided space for a marver. The bocca isshown as a round-topped arched opening, andthe glassblower to the right holds his iron withhis right hand in the circular boccarella (orlittle mouth). In front there is a circularaperture through which perhaps the tiseurcould check the condition of the fire: alterna-tively, it may indicate in false perspective thecentral opening of the furnace. From thesixteenth century there survives an eyewitnessaccount of an Italian furnace. In 1508, PederMånsson, a Swedish priest living in Rome,interested himself closely in the glass industry,then unknown in Scandinavia, and compiledan account of it (Månsson, 1520). Månssondescribed the furnace as follows:

The second furnace, in which the glass is tobe founded, is more difficult to build. It mustbe entirely built and walled up with dampclay capable of resisting the fiercest heat, andmust be in the middle of a wide, roomyhouse. You lay the foundation wall round ina circle, with a diameter of 3.8 m (12 1⁄2 ft). Atthe point where the furnace mouth is to bebeneath the ground, you lay no wall


foundation. This furnace must have threearches, one above the other. The first andlowest arch must occupy the whole interiorright up to the walls, and not be higher than760 mm (2 1⁄2 ft) from the floor and 508 mm(1 ft 8 in) in diameter. [This must mean ‘inwidth’. Vault would perhaps give a cleareridea of the construction than arch.] On thisarch must be set the pots in which the glassfrit is put. Then you arrange the wall outsideso that it has six thin ribs, and between eachset of ribs you reserve an opening, throughwhich the glass mass is drawn out, worked,inserted and handled. Directly in front ofeach opening a pot must be set in the furnacewith the batch in such a way that the heightexactly suits the opening. The second arch ismade 1.27 m (4 ft 2 in) high above the first,extending over the whole furnace, except fora round opening, 250–380 mm (10–15 in) indiameter in the middle of the arch. Round thetop of the opening there must be a ledge, sothat the glasses, when put there to cool, maynot fall down into the furnace. The third andtop arch extends over the whole furnace, andthere must be 1.0 m (3 ft 4 in) between it andthe second arch, and three openings 250 mm(10 in) broad, through which the smokedischarges and the glasses are put in to cool.The furnace mouth below in the earth shouldbe 510 mm (1 ft 8 in) broad. You stoke it withdry wood, the length of which correspondsto the inner breadth of the furnace; and forthis purpose one digs out the earth in frontof the furnace mouth ...

This suggests that there may have been a firetrench in the furnace, and this detail togetherwith the mention of digging out the earth infront of the furnace, strongly recalls thefurnace at Monte Lecco. A woodcut illustrat-ing a glass furnace of this type appears in Dela Pirotechnia (Biringuccio, 1540) (Figure3.52). This shows a hive-shaped furnace withsix external ribs, the stoke-hole to the front,the tiseur carrying an armful of faggots withwhich to replenish it. The glassblowers sit oneither side on three-legged stools. In front ofeach (seen better on the right-hand side of thewoodcut) is a screen to protect the blowerfrom the glare of the glory hole, throughwhich the glass pots are reached. Projectingfrom the side of the furnace below is a marble

shelf supported on an arch, the rudimentaryform of the modern marver, on which theglass drawn from the pot is rolled to smoothit as a preliminary to blowing.

Biringuccio describes the furnace as follows:

Now in order to complete the purification, around furnace is built of rough bricks from aclay that does not melt or calcine from thefire. Its vault has a diameter of about fourbraccia (550 mm) and a height of sixbracchia (3.32 m). It is arranged in this way.First a passage for the fire is made whichleads the flames into the middle of thefurnace; around the circle at the bottom ashelf 3⁄4 braccio wide (100 mm) is made onwhich are to be placed the pots that hold theglass, and this must be about one bracchio(140 mm) above the ground. Around this fiveor six well-made little arches are built assupports for the vault, and under these aremade the little openings which allow one tolook inside and to take out the glass forworking at will. Then the vault is continuedto cover the glass, and only in the middle isa little opening of a palmo (280 mm) or lessleft. Above this vault another vault is madewhich seals up and covers the whole; this istwo braccia (280 mm) high above the first sothat it completes the reverbatory furnace. Thisis the cooling chamber for the works whenthey have been made, for if they did notreceive a certain tempering of air in this, allthe vessels would break as soon as they werefinished when they felt the cold.


Figure 3.52 Woodcut depicting glassworkers at afurnace, illustrating Vannoccio Biringuccio, De LaPirotechnia, Venice, 1540.

By a curious convention, the illustrator ofBiringuccio’s work, in representing the topstorey, or annealing chamber of the furnace,has shown the holes through which the glasseswere inserted to anneal as seen from theoutside, but has cut away the central panel ofthe furnace wall to reveal on the inside thehole through which the heat ascends from thefounding chamber to the annealing chamberabove. This cut is defective in many ways. Itdoes not for example illustrate the arches madeabove the siege to accommodate the workingopenings. These arches were structurallynecessary to support the vault above themwhen the glass pots were changed, as wasnecessary when an old pot cracked or other-wise became defective, and a new, preheatedpot had to be substituted quickly with theminimum disturbance to the general workingof the furnace. Biringuccio (1540) gave someindication of how this was achieved:

After six or eight months from the time theywere made, when you wish to put them [theglass pots] in the furnace in order to beginwork, that place which you left open underthe arches is a quarter closed with a wall andonly enough space is left to allow one of thesaid vessels to enter ...

The aperture was then closed with clay,making two small holes from one large one sothat the worker can take out the glass with histube from whichever vessel he wishes in orderto work it. In the other opening he keepsanother iron tube so that it will be hot. Outsidein front of these openings there is a supportmade of a marble shelf placed on an arch.Above this shelf and in front of the opening forthe glass a screen is made to serve as a protec-tion for the eyes of the workers and to carryan iron support which holds up the tube ...

These hooks (halsinelle) or supports for theblowing iron and pontil are not shown inBiringuccio’s woodcut. Probably at a quiteearly date it was found convenient to incor-porate the working opening in a slab of fireclay, which could be removed and replacedwhen the pots were changed. Fragments of suchslabs were found on the early seventeenth-century glassmaking site at Jamestown inVirginia, USA (Harrington, 1952) and havebeen reported from glassmaking sites in

Denmark. The structure of these archedopenings, not shown in Biringuccio’s woodcut,is illustrated in the woodcuts to what isperhaps the most systematic account of glass-making surviving from the sixteenth century,the twelfth chapter of Georgius Agricola’s DeRe Metallica (1556) (Figures 3.53 and 3.54).

In many particulars Agricola was obviouslydependent on Biringuccio as a source ofinformation, but the situation on glassmakingwhich he described was far more complex,probably because in Agricola’s time, in Saxonyand elsewhere in Germany, the indigenous


Figure 3.53 Woodcut illustrating G. Agricola, Die ReMetallica, Basle, 1556, showing the three-tier ‘secondglass furnace’, i.e. annealing furnace.

‘northern’ tradition was being penetrated bythe Venetian tradition. The detailed woodcutsin De Re Metallica show a three-tiered furnace.In Figure 3.54 the arched openings to thesiege storey of the furnace can be seen (butwith a square bocca and no boccarella), and

in the sectional rendering there is an indica-tion of the vault supporting the siege itself, thecentral holes allowing the passage of the heat(the upper one square instead of circular), andthe arched opening allowing access to theannealing chamber. In his text Agricola refersto ‘six arched openings’ but the illustrationwould suggest that there were at least eight.A similar discrepancy appears in his descrip-tion of the two-storey ‘second furnace’ whichis said to be strengthened on the outside withfive ribs whereas ‘in the wall of the upperchamber between the ribs there should beeight windows ...’. The illustration suggeststhat there would have been at least eight ribs.

These illustrations were interpreted in asixteenth-century wall-painting by G.M. Butteriwhich decorates the studiolo of Francesco I de’Medici in the Palazzo Vecchio in Florence. Itshows the Grand Duke’s glasshouse, which isknown to have been of the Venetian type, andwhich seems to have begun operating about1568/9. In the background of the wall-painting can be seen the great glowingfurnace, the gaffers seated on their three-legged stools before the glory holes, from theglare of which they are protected by the fire-clay screen described by Biringuccio, whilst toone side and slightly lower down are theboccarelle accommodating three or moreirons. The master to the left warms-in his glassat the glory hole, resting the iron on thelowest of the halsinelle. Above the masters’heads glow the apertures leading to theannealing chamber, through one of which aservitor (to the left of the picture) is placing afinished glass to cool. To the right of theseated gaffer, with his back to the onlooker,can be seen the stoke-hole, toward which thetiseuer, apparently stripped to the waist, bringsa fresh bundle of faggots. Above the furnaceis a framework of beams on which the woodfor fuel was set to dry, with a resultant verygrave risk of fire. (When the Crutched Friarsglasshouse in London burned down in 1575 ithad ‘within it neere fortie thousand billits ofwood’; Thorpe, 1961.)

The furnace in Butteri’s painting appears tobe a developed form of that illustrated byBiringuccio or Agricola. The ribs of the furnacedelineated as rather thin in Biringuccio andAgricola here have become of considerabledepth, at least at the base, thus providing a


Figure 3.54 Woodcut illustrating G. Agricola, De ReMetallica, Basle, 1556, depicting a southern glassfurnace of the sixteenth century. The furnace has threesections, the lower one for the fire, the middle one forthe glass pots, and the upper one for annealing theglass. The glass blowers are working around the furnaceand the vessels are packed in the large box seen in thebottom right-hand corner. In the background the sale ofthe glass is being discussed, and a pedlar carries awaythe vessels. In the sixteenth century the customer eitherordered his glass direct from the maker or bought itfrom the travelling hawker.

much wider working surface for the gaffer.This detail is seen again in seventeenth-century representations of Italian-typefurnaces.

One curious feature of the Florence furnaceis its asymmetrical form at the right-hand side,where the vault appears to project in anoverhang. It seems possible that this may bethe beginning of what later became the tunnellehr. This development may be found inBiringuccio. The text is not easy to follow, butthe French translation by Jacques Vincentoffers an easier interpretation. Vincent (1556)writes, ‘this cooling-off is effected by a certainopening made on the left-hand side’,Biringuccio says ‘at the back, and this channelis shaped like a trumpet; from it all the cooledvessels are skillfully drawn by means of a longiron, one after the other, in three or four goes,until they reach the mouth and are takenoutside’. This ‘trumpet-shaped opening’ isperhaps the beginning of the lehr, and theman standing on the right of Biringuccio’swoodcut is no doubt performing the office ofmoving the glasses along from hotter to coolerpositions.

Merrett (1662), in a translation of Neri (1612)Dell’Arte Vetraria (The Art of Glass) (seeTurner, 1962, 1963), wrote:

‘The Leer (made by Agricola, the thirdfurnace, to anneal and cool the vessels, madeas the second was to melt the Metall, and tokeep it in fusion) comprehends two parts, thetower and leer. The tower is that part whichlies directly above the melting furnace with apartition betwixt them, a foot [300 mm] thick,in the midst whereof, and in the sameperpendicular with that of the secondfurnace, there’s a round hole [Imperat. andAgricola make it square and small] throughwhich the flame and heat passeth into thetower; this hole is called Occhio or Lumella,having an Iron ring encircling it called theCavalet or Crown; on the floor or bottom ofthis tower the vessels fashioned by the Mrs[masters] are set to anneal; it hath 2. Boccasor mouths, one opposite to the other, to putthe Glasses in as soon as made, taken witha Fork by the Servitors, and set on the floorof the tower, & after some time these Glassesare put into Iron pans, Agricola makes themof clay call’d Fraches, which by degrees are

drawn by the Sarole man all along the Leer,which is five or six yards [4.5–5.5 mm] long,that all the Glasses may cool radatim, forwhen they are drawn to the end of the Leerthey become cold. This leer is continued tothe tower, and arched all along about fourfoot [1.2 m] wide and high within. The mouththereof enters into a room, where the Glassesare taken out and set. This room they call theSarosel.

This structure may be seen on the frontispieceof the 1669 Latin edition of Neri, published inAmsterdam (Figure 3.55), and a similar


Figure 3.55 Frontispiece to the 1669 Latin translationof A. Neri, L’Arte Vetraria.

engraving illustrates the 1752 French edition ofthe same work (Figure 3.56). In thesefurnaces, however, the glasses appear to havebeen put directly into the lehr through smalldoors (N in Figure 3.56). The same featureappears in Diderot and D’Alembert’sEncyclopédie (1772) (Figure 3.57). The samevolume also gives a vivid view of the interiorof the lehr seen from the Sarosel room, withthe fraches moving in two lines down thetunnel (Figure 3.58). An improved version oftunnel lehr was invented by George Ensell, atCoalbrook in the British Stourbridge glass-making district, about 1780, but it is notabsolutely certain what the improvement was.It may well have been the provision of aseparately heated lehr, perhaps one with twotunnels, for large and small objects respec-tively, such as became standard in Britishglasshouses in the nineteenth century. (By themid-nineteenth century the pans were movedalong the lehr mechanically and the doublelehr had become a quadruple one.)

An illustration of an Italian furnace of themid-eighteenth century appeared as thefrontispiece of Due Lettere di Fisica al SignorMarchese Scipione Maffei by Gian-LodovicoBianconi, published in Venice in 1746 (Figure3.59). Although it is difficult to envisage justhow this furnace worked (it seems to havederived some of its features from Kunckelalthough its basic principles are different, and it

seems to have one storey too many), the natureof its long lehr, with central archway, is unmis-takable. A second piece of evidence concerningmid-eighteenth century furnaces is a tin-glazedpottery (maiolica) model of a furnace in theScience Museum, London. In this model, thelong tunnel lehr is clearly in evidence, with the glass visible in the tower, apparently heatedby the updraft from the founding furnaceconveyed through a chimney, which does notconnect directly with the floor of the lehr.


Figure 3.56 Plate IV from the French translation of A.Neri’s L’Arte Vetraria (Art de La Verrerie, Paris, 1752)showing the ‘Amsterdam’ furnace.

Figure 3.57 Plate from Diderot, Encyclopédie showing‘Verrerie en Bois, Coupe et Plans d’une petite Verrerie apivette: et Coupe de La Cave it braise’ (pl. vol. X. pl. 3)showing the grille composed of very short bars to spanthe firing channel.

Figure 3.58 Plate from Diderot, Encyclopédie showing‘Verrerie en Bois, l’operation de retirer les Feraces et lestransporter au Magasin’ (pl. vol. X, pl. 22).

The Venetian-style glasshouse had onefurther subsidiary furnace. This was used forthe preliminary roasting (fritting) of the silicaand ash. Biringuccio described this process asfollows:

Then put all these things [that is, silica, soda,and manganese for decolourizing] mixedtogether into the reverbatory furnace madefor this purpose, three braccia long, twowide, and one high, and apply enough of thestrong flames of a wood fire by means of thereverberator, so that the composition ismelted well and is converted all into onemass.

Månsson confirms this general picture, addingthe detail: ‘The mixture is often stirred andturned around with an iron hook’. No reallyexplicit representation, however, is availableuntil the publication of the Encyclopédie(Diderot and D’Alembert, 1772) whichcontains an engraving showing the furnace inplan and section, and an illustration of thefurnace-man at work raking the frit from themouth of the furnace (Figure 3.60). Agricola’sillustration of a fritting furnace also shows around construction, although the text seems toindicate an oblong structure (‘Their firstfurnace should be arched over and resemblean oven. In its upper compartment, six feetlong, four broad, and two high, the frit iscooked ...’). The Encyclopédie version issquare on plan although the internal shape iscircular. It has the advantage that it is fired at

the side, for the greater convenience of theworker stirring the frit (Charleston, 1978).

The northern European type ofglassmaking furnace

There is no evidence to throw light on thenorthern type of furnace as early as that avail-able for the southern type. The earliest sourceis a chapter in a twelfth century manuscript,Schedula Diversarum Artium (Treatise onDiverse Arts), by Theophilus Presbyter (trans-lated by Dodwell, 1961 and by Hawthorne andSmith, 1963). It is thought that Theophilus wasthe Benedictine monk Roger of Helmars-hausen, and that the manuscript was compiled


Figure 3.59 Frontispiece to G.L. Bianconi, Due Letteredi Fisica, Venice, 1746. (Corning Museum of Glass).

Figure 3.60 Plate from Diderot, Encyclopédie showing‘Verrerie en Bois, Plan et coupe de la Calcaisse, etl’Operation de retirer la fritte cuite’ (pl. vol. X. pl. 15).

between AD 1110 and 1140. If the identificationis correct, the author was a practising metal-worker who was personally able to carry outmost of the techniques he described and whowould certainly have described the glassmak-ing process with the insight of a craftsman:

If you have the intention of making glass, firstcut many beech wood logs and dry them out.Then burn them all together in a clean placeand carefully collect the ashes, taking carethat you do not mix any earth or stones withthem. When the ashes have been well mixedfor a long time, take them up with the ironshovel and put them in the smaller part ofthe kiln over the top of the hearth to roast[i.e. to frit]. When they begin to get hot,immediately stir them with the same shovelso they do not melt with the heat of the fireand run together. Continue this throughout anight and a day. (Dodwell, 1961; Hawthorneand Smith, 1963)

Theophilus also gives precise instructions formaking crucibles out of clay, and for makingglass articles, but the important point to notehere is that the frit must not be allowed tomelt (and, by implication, the small part of thefurnace must not be allowed to get too hot)so that the solid-state reactions can continuefor a long time (a night and a day) withoutany molten glass being formed that would trapthe released carbon dioxide as bubbles in themelt.

After this build a furnace of stones and clayfifteen feet [4.5 m] long and ten feet [3 m]wide in this way. First, lay down foundationson each long side one foot [300 mm] thick,and in between them make a firm, smooth,flat hearth with stones and clay. Mark offthree equal parts and build a cross-wallseparating one-third from the other two. Thenmake a hole in each of the short sidesthrough which fire and wood can be put in,and building the encircling wall up to aheight of almost four feet [1.2 m], again makea firm, smooth, flat hearth over the wholearea, and let the dividing wall rise a littleabove it. After this, in the larger section, makefour holes through the hearth along one ofthe long sides, and four along the other. Thework pots are to be placed in these. Then

make two openings in the centre throughwhich the flame can rise. Now, as you buildthe encircling wall, make two separatewindows on each side, a span long and wide,one opposite [each] of the flame openings,through which the work pots and whateveris placed in them can be put in and takenout. In the smaller section also make anopening [for the flame] through the hearthclose to the cross-wall, and a window, a spanin size, near the short wall, through whichwhatever is necessary for the work can beput in and taken out.

When you have arranged everything likethis, enclose the interior with an outer wall,so that the inside is the shape of an archedvault, rising a little more than half a foot[150 mm], and the top is made into a smooth,flat hearth, with a threefinger-high lip allaround it, so that whatever work or instru-ments are laid on top cannot fall off. Thisfurnace is called the work furnace ... Nowbuild another furnace ten feet [3 m] long,eight feet [2.5 m] wide, and four feet [1.2 m]high. Then make a hole in one of the facesfor putting in and taking out whatever isnecessary. Inside, make firm, smooth, flathearth. This furnace is called the annealingfurnace. ... Now build a third furnace, six feet[1.8 m] long, four feet [1.2 m] wide, and threefeet [900 mm] high, with a [fire] hole, awindow and a hearth as above. This furnaceis called the furnace for spreading out andflattening. The implements needed for thiswork are an iron [blow] pipe, two cubits longand as thick as your thumb, two pairs oftongs each hammered out of a single pieceof iron, two long-handled iron ladles, andsuch other wooden and iron tools as youwant. (Smedley et al., 1998)

The illustration of the model of Theophilus’furnace made at the Science Museum, Londonis probably incorrect in showing four workingholes per side instead of two. The feature ofthe holes made in the siege to take pots isunique, and one is tempted to wonderwhether Theophilus was not misled, by seeingin a furnace the ring of glass left on the siegewhen a broken pot was removed. The recon-structions (Theobald, 1933) have always beenmade very trim and square; the furnaces wereprobably always somewhat more rough and


ready in practice. It should be noted that inthis instance the small furnace was used forfritting, and a separate furnace for annealing.These procedures were often reversed, andsometimes the subsidiary furnace was used forboth fritting and annealing. For the units ofmeasurements used see Hawthorne and Smith(1963).

The treatise entitled De Coloribus et ArtibusRomanorum (On the Colours and Arts of theRomans), attributed to a certain Eraclius (preAD 1000), contains chapters on glassmakingwhich have been added, probably in thetwelfth or thirteenth century, to an existingmanuscript. These describe a tripartite furnaceof which the largest section is in the centreand is the founding and working furnace. Thishas one glory hole on either side, apparentlywith two pots to each, perhaps in the samemanner as the Theophilus furnace. To the leftof this should be a smaller furnace used forfritting and pot arching, and to the right a stillsmaller compartment presumably for anneal-ing. The actual ground plan of these furnacesis not prescribed, and they may not necessar-ily have been rectangular. (For the text, seeMerrifield, 1949.) The fire trench down themiddle of the whole furnace is clearlyindicated, and it may be assumed that therewere solid sieges to either side of it, ratherthan the improbably flimsy structure repro-duced by Maurach.

That rectangularity was by no means the rulein practice seems to be shown by the evidenceof a glass furnace at Glastonbury (UK). Thisappears to date from late Saxon times(between the eighth and the eleventh centuriesAD). Although not enough of the structure hassurvived to permit an exact reconstruction, theground plan appears to have been oval.Similarly, four furnaces attributed to the ninthcentury AD excavated at Nitra in Slovakia hadoval ground plans. This has been interpretedto suggest that the Bohemian furnace was ofa type ‘entirely different from the well-knowndescription’ in Theophilus (Hejdová, 1965). Itmay well be, however, that the essentialelement in the northern tradition was not somuch the rectangularity of the ground plan asthe fact that the main and subsidiary furnaces(or at least one of them) were on the samelevel and shared their heat either by having thesame fire trench running the length of the

composite furnace (as in Theophilus) or bytransmitting the heat laterally from the mainfurnace to the subsidiary one, or by both. Anarrangement of this kind can be seen in thefamous illustration to Mandeville’s Travels in amanuscript in the British Museum, London(Plate 1). This manuscript is thought to havebeen compiled in Bohemia in about 1420.There is no description accompanying theoriginal manuscript, but that given by Kenyon(1967) seems to be the most interesting andthe following is adapted from it.

All round there is forest, with a man carry-ing fuel in a basket, two others carrying ashin sacks, one man digging sand from a hill ina clearing in the forest with a stream at itsfoot, and another carrying the sand to theglasshouse in a shoulder hod. The glasshousehas a rough shingle-roofed open shed with itsstoke-hole entrance sheltered by a roof madeof wood billets drying on a heavy timberframe. The furnace appears to be rectangular,having rounded and domed corners with acircular flue on top. Part of the roof is missing,perhaps to allow the furnace gases to escape.The furnace may have had four crucibles, twoeach side, and the annealing furnace is builton at the end. Vessel glass is being made andthe master, in a hat, inspects a jug with ahandle; a workman is taking a vessel out tofinish its annealing in a large storage jar. Twoglassmakers, wearing sweat rags on theirbrows, are shown; one is gathering glass froma crucible with his blowing iron. The boystoker is attending to the fire. The stream inthe background suggests a need for water, forwashing the sand, mixing the ash, andperhaps preparing the crucible clay.

The whole lively scene, with its emphasison the temporary woodland shack is repre-sentative of a northern glasshouse. The artist’srendering of the crown of the furnace wouldsuggest an oval ground plan.

An actual furnace of about this periodexcavated at Skenarice (in the Semily districtof the present day Czech Republic) revealedan oval ground plan extended by two paral-lel lines of masonry, perhaps the original firetrench. Another structure close by and appar-ently of rectangular form has been interpretedas being the fritting furnace. Yet anotherfurnace of this general type has beenexcavated in the Czech Republic at Ververi


Bityska. That the Bohemian furnace was notalways oval, however, is suggested by theexcavation of a late fifteenth/early sixteenthcentury at Pocatky (in the district ofPelhrimov, Czech Republic). Here three wallsof a furnace were preserved to a height of400 mm and revealed a rectangular groundplan. A further complex of three rectangularfurnaces, dating from some time later in thesixteenth century was excavated at Rejdice.The same pattern was confirmed by finds atthe nearby contemporary glassworks at Syriste(founded 1558). Whether the variationbetween oval and rectangular furnaces inBohemia is a question of date or of function,it is difficult to say in the light of current avail-able evidence.

In Britain, the rectangular furnace appearsto have been the rule from, at the latest, thefourteenth century onward. At Blunden’sWood, Hambledon (Surrey), a roughly rectan-gular furnace dating roughly from about 1330(Figure 3.61) had a central fire trench roughly3.2 m long by probably originally 610 mmwide, with a hearth at either end; to each sideof this was a siege for two pots, 2.6 m longby 690 mm wide and 610 mm high (Figure3.62). This showed the unusual feature of(apparently) a cavity between the siege andthe outside wall, either for insulation purposesor possibly as a means of constructing thevaulted roof. This main furnace was accom-panied by two subsidiary structures, presum-ably for the operations of the fritting,

pot-arching and annealing (Wood, 1965)(Figure 3.63).

A similar furnace, brick-built and almost6.4 m long and 770 mm wide, with clearlymarked firing chambers at either end, wasfound well-preserved at Fernfold, Sussex, on asite connected with Jean Carré, founder of the‘modern’ glass industry in Britain in 1567. Acomparable furnace of early sixteenth-centurydate, 3.7 m long, built of brick and stone andwith sieges for three pots each side, wasdiscovered at Bagot’s Park, near AbbotsBromley, Staffordshire, and one ancillaryfurnace was excavated in the vicinity. Almost300 kg of cullet was found on the site. Arectangular furnace site 3.7 m long by 1.2 mwide was found at St Weonards, Herefordshire


Figure 3.61 Plan of themedieval glasshouse atBlunden’s Wood, Hambledon,Surrey. (After Wood, 1965).

Figure 3.62 Reconstructed section of the glasshouse atBlunden’s Wood, Hambledon, Surrey (see Figure 3.58).(After Wood, 1965).

in 1961. It was probably built of brick andstone, a square of large stones in the centreof a burned area probably representing thefoundations of the founding chamber. Furtherrectangular furnaces, dating from the mid-sixteenth century, with three pots per siege,were excavated at Knightons, Alfold, Surrey,in 1973 (Figure 3.64). Associated with thesefurnaces was a pair of smaller rectangularfurnaces, perhaps for spreading window glass(Kenyon, 1967).

The picture at Blunden’s Wood is repeatedin essence at a late sixteenth- to early seventeenth-century glasshouse site at BlorePark, in Eccleshall, northwest Staffordshire.Here a fairly well preserved stone furnacefoundation was found and excavated, revealinga more or less square furnace with a long fire

trench running east and west, and a siege oneither side 300 mm high, 420 mm wide and860 mm in length, accommodating two pots oneach side. This furnace appeared to becomplete in itself, and no other structure wasexcavated, although traces were found inneighbouring mounds. There is therefore nomeans of knowing how the ancillary processeswere carried out at this site (Pape, 1933). At themore or less contemporary site of Vann, nearChiddingfold, Sussex, however, the foundationsof a larger, brick-built structure were excavated,the main furnace being an oblong (3.7 m �1.7 m) at the corners of which were four diago-nally projecting, fan-shaped wings (Figure3.65). These were no doubt originally used forannealing, fritting and pot-arching. There was agreat concentration of medieval glasshouses in


Figure 3.63 Reconstruction by James Gardner of a medieval glasshouse, based largely on Blunden’s Wood. A =main working furnace with hearth at each end; B = apertures for pots; C = hearth; D = small furnaces forannealing and pre-heating pots; E = fuel (billets of beechwood); F = cullet (broken glass); G = raw material (sandetc.); H = new pots ready for furnace (pots not made at this site); I = marver on which glass was rolled andsmoothed; J = water trough for cooling; L = bed of charcoal and sand on which to rest finished crown of glass; M= finished crowns of window glass; N = blowpipe and other tools.

the Chiddingfold area; at least 36 sites havebeen positively identified. (There are howeverby comparison 186 known medieval glassfurnaces in France.) An interesting feature ofthe medieval forest glasshouses is that uponexcavation they were found to contain verylittle broken glass; and it would seem that allthe cullet was carefully collected when theglassmakers abandoned each site (perhapsbecause the furnace collapsed or because thesupply of wood was exhausted in the vicinity)for use at the next.

The only excavated British glass furnacewhich does not fall into the general pattern ofa rectangular ground plan, with or with-out wings, is that at Woodchester inGloucestershire, excavated around 1904. Thecircular structure was apparently 4.9 m indiameter with an internal diameter of 3 m,making the waIls 900 mm thick, a feature

which must arouse some doubt. The firinghole was 1 m wide on the outside and 910 mmon the inside, exceptionally wide. If anyfeature still remains, a new excavation wouldbe highly desirable. An adjacent rectangularstructure 2.7 m by 2.1 m was interpreted as anannealing furnace. It has been suggested thatthe round furnace at Woodchester may beexplained by the presence there of Flemishglassmakers. All these glasshouses made greenforest glass, and their structure was describedby Merrett in 1662:

The Green Glass furnaces are made square,having at each angle an arch to anneal theirGlasses ... For green Glass on two oppositesides they work their Metall, and on the othersides they have their Calcars, into whichlinnet holes are made for the fire to comefrom the furnace, to bake and to prepare their


Figure 3.64 Plan of the glass furnace at Knightons,Alfold, Surrey. Circa 1550. Note the three six-potfurnaces, furnace 2 overlying furnace 1, and the two-chamber annealing furnace 4. (After Wood, 1965).

Figure 3.65 Plan of the Vann Copse furnace atHambledon, Surrey, showing four ‘wing’ furnacesattached to the main furnace. A = fire chamber; B =sieges; C = hearth; D = (?) annealing chambers; E =hearth lip (tease hole); W = taper wing wall. (Based ona rough sketch by S.E. Winbolt, modified by A.D.R.Caroe).

Frit, and also for the discharge of the smoke.But they make fires in the arches, to annealtheir vessels, so that they make all theirprocesses in one furnace onely.

This general layout of furnace survived inBritain and France until at least the 1770s,although the adapted Venetian-type furnacewas by now well established in both countries.Its characteristic of heating subsidiary furnacesthrough linnet holes recalls the two-chamberfurnaces of Theophilus and the Mandevilleillumination (Plate 1). This plan appears tohave survived in Germany until at least theend of the seventeenth century.

A four-pot bottle furnace with two (circular)wings was in operation from at least 1777 inGravel Lane, Southwark, London (Figures 3.66and 3.67). Here one wing was used forfritting, the other for heating bottle culletbefore its transference to the pots. Annealingwas carried out in seven independentsubsidiary furnaces (00 and ww on the plan),and pot arching in another (t on plan). Theorigins of this type of furnace are at presentuncertain. A furnace with a ground plan of thistype excavated at Heindert (Canton d’Arlon,

Luxembourg) is supposed to date from at leastCarolingian times. In Britain, apart from Vann(see above), three-winged furnaces have beenexcavated at Hutton and Rosedale in Yorkshire(Crossley and Aberg, 1972) and at Kimmeridgein Dorset in 1980/81.

At Hutton, an earlier furnace of plain fire-trench type, was overlaid by a second furnacewith two fan-shaped wings to north-east andsouth-west, these being incorporated with athird overlying two-pot furnace. A secondfurnace, standing apart, may have been anannealing furnace for the first phase. The lastphase of the furnace was dated to the latesixteenth century by thermo-remnant magnet-ism. At Rosedale (Figure 3.68) a two-potfurnace of similar date had four fan-shapedwings.

Both furnaces were built of stone and clay(Frank, 1982). Excavations at Kimmeridge in1980/81 revealed the foundations of an earlyseventeenth-century glasshouse having fourfan-shaped wings (Figure 3.69). The siegeswere badly eroded so that it was not possibleto be certain how many crucibles had beenset in the furnace, but there appeared to beample space for four pots. Underground air


Figure 3.66 View of a bottle-house in Gravel Lane, Southwark, London. Circa 1777. c represents the grill.(Drawing by C.W. Carlberg).

passages were necessary to provide draughtfor the local oil-shale, which was used as asource of fuel. The floors of the air passageswere slabbed with stone and beneath them adrain channelled water towards the sea. Stonesteps at the ends of the air passages farthestfrom the furnace gave access for raking outash. The furnace was contained within arectangular cover building with a shale-tileroof supported on timber on a stone founda-tion wall. This building protected the furnaceand the working areas located on either sideof it between the wings, which were presum-ably used for ancillary processes.

From a previous era of archaeology there isa record of a furnace excavated at Buckholtnear Salisbury in Wiltshire, which appears tohave had a winged construction. The Buckholt

glassmakers seem to have been, with one possi-ble exception, French by origin; and the Vannglasshouse appears to have produced greenglasses of the normal Wealden ‘Lorraine’ type.

The two Yorkshire glasshouses also pro-duced characteristic green vessels of the sametype, but it seems that higher temperatureshad been employed than in the glasshouses ofthe Chiddingfold (Weald) area of southernBritain (Cable and Smedley, 1987). On thewhole, it seems likely that the winged furnacewas of French origin. That it enjoyed a widerdiffusion at a later date is shown by the Britishexamples quoted above and by the following:a stone-built furnace in a glasshouse atKarlova Hut, in the Czech Republic, foundedin 1758; and a furnace in the Amelung factoryat New Bremen, Maryland, USA (Figure 3.70),


Figure 3.67 Plan of the glasshouse in Gravel Lane, Southwark (see Figure 3.66). qq are explained as ‘openingsinto the vaults of these two furnaces, through which the heat is communicated from the founding-furnace’; uu areannealing furnaces, the largest, accommodating 800 bottles, is a pot-arching furnace; v is a furnace for chemicalapparatus and other large pieces, the annealing furnace for which is at ww.

built after 1774 (and probably after 1784)(Hume, 1976). Both these furnaces appear tohave had a plain firing trench, without a grill,no doubt reflecting the use of wood ratherthan coal as a fuel. Amelung had begun hiscareer at Grünenplan south of Hanover inGermany, and in Maryland took over a factorythat had previously been run by Germans.Therefore it may reasonably be concluded thatthe winged furnace was also established in theGerman-speaking countries by, at the latest,the middle of the eighteenth century. It seemslikely that the six-pot furnace with twoannexed annealing furnaces was used at theNotsjö factory in Finland in 1799 (Seela, 1974).

Kunckel illustrated a German furnace (ArsExperimentalis, or Experiments in the Art ofGlass, 1679) (Figures 3.71 and 3.72). Here theworking furnace and the fritting or annealingfurnace are simply two chambers of the samestructure, with the fire chamber runningthrough their combined length (see Figure3.72). It is not clear from the engravingswhether there was also a linnet-hole connect-ing the upper chamber (that is, the workingstorey of the main furnace and the actualannealing chamber of the subsidiary one), andKunckel provided no explanatory text.However, the illustrations were reproducedwith explanatory text in the 1752 Frenchedition of Neri (1612):


Figure 3.68 Plan of agreen-glass furnace excavatedat Rosedale, Yorkshire. (AfterD.W. Crossley and F. Aberg).

It is important to note that the subsidiaryfurnace was used for annealing as well as forfritting; the opening marked � on the plan isdescribed as an opening for placing the glassto anneal. This type of furnace was clearlydomesticated in France about as early as thewinged furnace was established inGermany/Bohemia, for among the supple-mentary plates of Diderot (1751–1771) is aseries of engravings illustrating this doubletype of furnace, entitled ‘Round furnace orFrench furnace’. There are changes of detail,notably a square plan for the subsidiaryfurnace and four ribs to the beehive-shapedfounding furnace.

That the type of furnace represented in theFrench edition of Neri was continuously in useup to the date of Kunckel’s book is indicatedby the results of an excavation carried out onthe site of a glasshouse at Trestenhult, inSweden, dating from about 1630. That it wasessentially a German-type furnace is confirmed

by the fact that this factory was under theleadership of a certain Påvel Gaukunkel,probably one of a West German family ofglassmakers, and that it made green glass ofthe German waldglas type. This furnace hadan octagonal founding furnace with four pots,and a stoke-hole running throughout thelength of the entire structure, which includedan annealing furnace of circular plan, setslightly askew to the axis of the main furnace,with the result that the square hole permittingthe passage of heat from the fire chamber tothe annealing chamber is off-centre (Figure3.73). The two furnaces are interconnected atfirst floor level by a linnet-hole (K on thedrawing, Figure 3.73) (Roosma, 1969).

A furnace of the same basic type wasalready known to Agricola (Figure 3.74) in thesixteenth century, although in his illustrationthe annealing furnace is rectangular in plan,perhaps a throwback to the earlier traditionsof Eraclius and Theophilus. Agricola describedthe furnace in the following terms:


Figure 3.69 Plan of the early seventeenth-centuryglasshouse at Kimmeridge, Dorset, based on the1980/1981 excavations. Features: 1–4, wings; 5, airpassages; 6, steps giving access to air passages; 7–8,working areas; 9, fire-box; 10, robbing pit; 11, wallfooting. (Courtesy of D.W. Crossley).

Figure 3.70 Aerial view of part of the Amelungglasshouse, USA, excavated in 1963, showing thewestern melting furnace with its four wings, which arebelieved to have housed small fritting ovens. Beyondthese foundations, to the right, can be seen the remainsof a pair of annealing ovens, while in the top leftcorner stands the east melting furnace. The modernmetal roof in the background covers the pair of frittingovens excavated in 1962. (From Schwartz, 1974.)(Corning Museum of Glass).

The second furnace is rounded, ten feet [3 m]broad and eight feet [2.5 m] high, and strength-ened on the outside with five ribs [460 mmthick]. This again consists of two chambers, theroof of the lower being [460 mm thick]. Thislower chamber has in front a narrow openingfor stoking the logs on the ground-levelhearth; and in the middle of its roof is a biground aperture opening into the uppercompartment so that the flames may penetrateinto it. But in the wall of the upper chamberbetween the ribs there should be eightwindows so large that through them thebellied pots may be put on the floor roundthe big aperture. ... At the back of the furnaceis a square opening, in height and breadth 1palm, through which the heat may penetrateinto the third furnace, adjoining. This is oblongeight feet [2.5 m] by six feet [1.8 m] broad,similarly consisting of two chambers, of whichthe lower has an opening in front for stokingthe hearth ...’ (Winbolt, 1933).

This is in all essentials the Trestenhult furnace,except for the ground plan and for the fact thatapparently each furnace here has its ownfirebox, one set at right-angles to the other. But


Figure 3.71 View of Kunckel’s furnace (from the 1752French translation of A. Neri, L’Arte Vetraria).

Figure 3.72 Plan of Kunckel’s furnace, shown inFigure 3.71.

Figure 3.73 Reconstructed section of a green glassfurnace found at Trestenhult, Sweden, circa 1630.

the express mention of interconnection at thefirst floor level is a vital point and such a featuremay have also existed in the furnace ofMandeville’s illustration (see Plate 1). Evidencefor further furnaces of this type is to be foundin the German-speaking areas of Central Europe:

(i) A humpen attributed to ChristianPreussler, with a view of the Zeilberg

glasshouse, Bohemia 1680 (Museum ofApplied Arts, Prague).

(ii) A second Preussler family humpen of1727, showing a virtually identical furnace(Figure 3.75). To the right may be seenthe subsidiary furnace with somethingstretched across the access door, as in thecase of the 1680 humpen, which seems toshow pot arching in progress. To the leftis a separate secondary furnace (forfritting?).

(iii) A glasshouse at Reichenau (probably thaton the Buquoy estates at Gratzen on theAustrian side of the South Bohemianborder), shown in Figure 3.76. The houseclearly made both window and vesselglass (the latter by the muff process), withthe circular furnace for vessel, the struc-ture to its left being designated by the keyas the annealing furnace, and that to theextreme left as the Taffel Offen, presum-ably the furnace for window glass. Thestructural connection, if any, betweenthese two last is unclear from the engrav-ing, but the lehr is clearly connected withthe circular vessel-glass furnace, althoughit seems to be in use for annealing (andperhaps spreading?) muffs. Probably thisshould be regarded as a bastardizedversion of the German-type furnace.

(iv) The illustration from the Augsburg Bible of1730, showing all the features of theKunckel engraving, and perhaps deriveddirectly from it.


Figure 3.74 Woodcut illustrating G. Agricola, De ReMetallica, Basle, 1556, showing a ‘second furnace’ withconjoined annealing furnace. H shows the clay ‘tunnels’in which the glasses were annealed.

Figure 3.75 Humpen of the Preussler family, depicting a glasshouse. Bohemia, 1727. (After Partsch, DerPreusslerhumpen, 1928).

(v) Excavations of the Junkernfelde furnace.

Three sites excavated in Denmark seem tosuggest that the German-type furnace was alsoin use there:

(i) Hyttekaer in Tem Glarbo, in the generalarea of Arhus, Jutland. Here six furnaceswere discovered, not all necessarily of thesame period, of which two appeared tohave a common fire trench, one beingcircular, the other an oblong structure.Two further furnaces were apparentlycircular on plan, but although close toeach other, their fire trenches werealigned roughly at right angles. Of theremaining furnaces apparently only thefire trenches survived, without any indica-tion of their external shape.

(ii) Tinsholt, south of Aalborg in Jutland. Atleast one furnace (or two contiguousfurnaces) here apparently correspond tothe main Hyttekaer furnace, although thepublished details are somewhat unclear.

(iii) Stenhule in Tem Glarbo (see above).Three furnaces were found (discountingan apparently earlier circular foundation),of which the founding furnace, probablyfor six pots, may have been almost square,3.4 m � 2.7 m, with chamfered corners. Itsvery long firebox, projecting on the westside, may have heated a long, narrowsubsidiary (annealing?) furnace, whichwould have resembled that at Reichenau.

The two remaining furnaces, apparently circu-lar and oval on plan respectively, were possi-bly for fritting and spreading of window glass.However, the site at Nejsum in Vendsysselnorth of Aalborg, the plan of which seemsclearer than some of those mentioned, revealsa central fire trench projecting a long way outat either side of an almost circular furnace plan– very much as at Fernfold in the Weald, exceptthat the furnace there was rectangular. TheStenhule furnace could well be reconstructed inthe same sense. It should be noted that all theglassmakers associated with these Jutland siteswere Germans, some specifically from Hesse.

A glasshouse at Henrikstorp in Skane,Sweden, had a ground plan suggesting anexceptionally long fire trench, and may havebeen a structure resembling that at Reichenau.Its reconstruction with four pots on one sideof an offset fire channel, and with a circularannealing furnace, seems very improbable.Finally, a glasshouse dating from about 1530excavated at Junkersfelde, in eastern West-phalia, revealed a four-pot furnace rectangularin plan, with chamfered angles (as atTrestenhult and Stenhule), apparently leadingto an annexed annealing furnace.

Development of coal firing and thechimney furnace

With the exception of the glasshouse atKimmeridge in Dorset, all the furnaces hitherto


Figure 3.76 Engraving showing the glasshouse at Reichenau, from a drawing by Clemens Benttler. Middleseventeenth century.

described were fired by wood. It fell to Britainin the sixteenth century, with the increasingcompetition among wood-using industries forthe output of coppiced woodlands, to developthe use of coal for this purpose. The latesixteenth and early seventeenth centuries werefilled with the clamour of rival patentees claim-ing monopolies for their particular systems ofusing coal for this and that industrial process.In glassmaking, sufficient progress had beenmade by 1615 to enable the government of theday to issue a ‘Proclamation touching Glass’,forbidding the burning of wood in the glassindustry (Crossley, 1998).

The essential feature of the coal-burningfurnace was the use of a grill of iron bars onwhich the coals could be laid and rakedperiodically, the ashes falling into a pit below.In general, the furnace at Kimmeridge, opera-tional between 1615 and 1623, resembled thewood-burning furnaces that have beenexcavated on both sides of the EnglishChannel (see Figure 3.69). There was a centralpassage between sieges on which thecrucibles had been set. At each corner was atriangular wing. The wings had formed thebases of four subsidiary structures in whichthe preparation of raw materials, and theworking and annealing of the glass, could be

carried out. They may also have been used forthe prefiring of crucibles. Outside the sieges,between each pair of wings, there wereplatforms on which lay fragments of stonefloors. The main difference between Kim-meridge and the woodburning furnaces of theLorraine tradition were the size of the fireboxand the long vaulted passages used to supplyair. When burning wood, it was possible tolay fires between the wings and to allow theflames to travel into the arched structure inwhich the pots were set. However, a coal firegave a shorter flame, and therefore had to beplaced in the centre of a furnace. AtKimmeridge a central block of brickwork laybetween the sieges in a stone-lined channelapproximately 1.4 m deep and 1 cm wide. The‘coal’ (bituminous shale) was placed upon thecentral block, or above it on fire-bars, ofwhich no trace survived at the time of excava-tion (1981). Another example of an earlyseventeenth-century four-pot furnace with anash-pit below the sieges, and an apparentlyvaulted firing chamber, was excavated byHurst-Vose (1980) and Burke at HaughtonGreen, Denton near Manchester (Figure 3.77).

The development of air passages is to beexpected at this time. The account by Merrett(1662) of furnace procedure described air


Figure 3.77 Plan of the seventeenth-century glasshouse at Haughton Green, Denton (near Manchester). The mainfurnace lies to the bottom right, with (?) annealing chambers to the left. (From Hurst-Vose, 1980).

passages, and the use of fire-bars, ‘Sleepers arethe great Iron bars crossing smaller ones whichhinder the passing of the coals, but give passageto the descent of the ashes’. This appears to bethe earliest reference to this device, but it isclearly shown in the drawing of a Londonglasshouse made by a Swedish architect visitingEngland in 1777–8 (see Figure 3.67). The grillsystem seems to have been adopted on theContinent. It is shown in an illustration byKunckel in 1679 (see Figures 3.71 and 3.72)and is clearly seen in the illustrations by Diderotand D’Alembert in their Encyclopédie of 1772.The provision of an ash chamber added, as itwere, an extra storey to the furnace.

Clearly the provision of an adequate draftwas of critical importance in the firing of acoal furnace, and it was inevitably in this fieldthat the British made a further importantcontribution to glass technology. The princi-ple was to have a cone-shaped building tohouse the furnace, capable of having all itsoutlets closed. Underground flues, no doubtarranged with regard to the prevalent wind,led to the furnace itself. By closing the conedoors and opening the flues, a tremendousthrough-draft was created, the great coneacting as a chimney. Glassworks cones werebuilt only in Britain and the Hanover regionof Germany (which at that time had strong


Figure 3.78 Plate from Diderot, Encyclopédie showing‘Verrerie Angloise, Vue et Coupe sur la Longueur de laVerrerie’ (pl. vol. N. pl. 2 of article ‘Verrerie Angloise’).

Figure 3.79 Plan of the glasshouse shown in Figure3.78 (pl. 1); b in Fig. 2 is the grill. The ‘linnet-holes’ tothe ‘wing’ furnaces are visible in the engraving.

royal connections with Britain), but not inFrance or Belgium. Bontemps (1868) in aGuide de Verrier (Guide to Glass) states, ‘It isgenerally known that the English glasshousesare huge cones which surround the furnace.The English furnaces are thus situated underbig chimneys which encourage energeticcombustion in a way that is impossible in ourFrench furnaces, which have a louveredopening above the furnace for the combus-tion products’ (Figures 3.78 and 3.79).

The advantages of this system weredescribed in Diderot and D’Alembert’sEncyclopédie (1772). The writer of the articleVerrerie stressed the importance of the singleupward movement of air facilitated by thecone superstructure, all the outward exits ofwhich were shut during founding, allowingthe air to enter the building only through threebroad ducts laid in the foundations under thefurnace. This increased the efficiency of theBritish furnace enormously in comparison withthat of the French type, the former being able,all other things being equal, to make in 12days the same number of bottles as a Frenchfactory made in 15 days.

The same points are chosen for commentby a Danish visitor to London in 1727:

It is known that the English glass is of highquality, and this is ascribed to the intenseheat of the coal which is used for thispurpose; I therefore also visited two of theirglass furnaces, which stand on an eminence,and in the ground there is an open passagefrom both sides to a very large iron grill,which is in the middle, and the coals on andabout it; by means of this passage the coalsreceive the air they need and one can stirthem with appropriate instruments.

Gunther (1961b) pointed out that the conesserved a dual purpose, being both a glassfactory (i.e. a manufacturing area) and a wastegas extraction system. Gunther was particularlyinterested in their technical performance andquotes the general height as being 15–25 m;the volume of masonry as being 1100 m3 (thecone at Obernkirchen weighing 2800 tonne);the (diluted) waste gas temperature as 400°C;the exit velocity as 1 ms–1, concluding that, atany time, the cones enable coal to be burntmore efficiently than in any other manner for

achieving melting temperatures, and gave amore equable environment for the workmen.However, the cones no longer had this advan-tage in fuel efficiency once the Siemens regen-erative furnace had been introduced (Gunther,1961a).

It is not certain exactly when the glass conewas first devised. Godfrey (1975) has an inter-esting discussion of the development of thecone furnace, including some eyewitnessdescriptions of the Winchester House furnace,probably from as early as 1610. There is,however, some misunderstanding of thefunctions of the glass cone, which is quitedifferent from a chimney leading from thefurnace. The point is that the heat should bedrawn over the glass pots and not out of the crown of the furnace. The cone and theunderground passages simply augment thedraft, which follows this course, emerging atthe working holes of the furnace.

Captain Philip Roche was building a conein Dublin as early as 1696 but this was proba-bly preceded by a period of experiment. In1702 The London Gazette reports the existenceof a glass cone, ‘94 Foot high and 60 Footbroad’ (28.7 m � 18.3 m). The 10.7 m (35 ft)high building constructed in 1621 atBallnegery was evidently a normal frame andshingle glasshouse with presumably no effecton the draft. The Belfast News Letter for 19August 1785 referred to a new glasshouse31.1 m (120 ft) high ‘being the largest of anyin Great Britain or Ireland’. In 1823 a cone32.0 m (150 ft) high was recorded in the samecity. In 1784, a prohibition was promulgatedin Dublin against any glasshouse chimney lessthan 15.2 m (50 ft) in height.

Possibly the great heat reported by Merrett(above) was due to this device as early as 1662.More probably the cone was invented betweenthis date and the end of the century. A fewcones still survive although they are no longerin use (Ashurst, 1970; Lewis, 1973). In Britainthey are now preserved as Ancient Monumentsand can be seen at Alloa (Scotland), Lemington(near Newcastle-upon-Tyne), Stourbridge (nearBirmingham) and at Catcliffe and Gawber inYorkshire. The cone glasshouse at Gawber wasprobably constructed in the eighteenth centuryand is known to have been in ruins by 1823(Figure 3.80). It illustrates the use of under-ground flues to induce the draft under the


furnace. On the site of the main furnace werefound lengths of fire bar, 2–3 ft (600–900 mm)long by 11⁄4 in (30 mm) square section, and asandstone block carved with a slot into whicha fire bar exactly fitted, to form the detachablefire grate. This seems to be the only archaeo-logical record of the dimensions of fire bars. InGermany there are examples of glassworkscones at Obernkirchen and Steinkrug (nearHanover) and at Gernheim (near Minden).

The difference between the heat needed formaking drinking glasses and that for windowglass is made explicit in an entry in the diaryof Sir James Hope in 1647 concerning theWemyss (Scotland) glasshouse:

That window glasses and drinking glassescannot be made in one fornace because thoserequyre a great deall stronger heatt thanthese: That the fornace for those is yrfore

long vaulted; and for these round bothowever yt could not make window glass,nather possiblie could find workmen whohave skill of both ...

These observations were based on the experi-ence of an Italian glassmaker, ChristopherVisitella. The great heat reached in the green-glass furnaces is emphasized by Merrett: ‘Theheat of those furnaces, is the greatest that everI felt ... The workmen say tis twice as strongas that in the other Glass-furnaces ...’

One further British invention resulted fromthe use of coal for firing. This was the coveredglass pot. That open pots continued to beused in circumstances where the fumes of thecoal made little difference (e.g. in bottlemaking) is proved by the details of the interi-ors of British bottle houses given by Carlbergin 1777–78 (see Figure 3.68 and 3.69) and by


Figure 3.80 Plan of thecone glass-house atGawber, South Yorkshire,showing three flues or airintakes, feeding air fromthe exterior to the centralfurnace. Sl-Sj2 refer tosections drawn duringexcavations. Site of theearlier (Phase I) furnace isshown in the top left-handarea. T2 (Trench 2)confirmed that Flue 2continued into Flue 3. Flue1 was added at a laterdate. Sector ‘a’ seemed tobe a smith’s hearth formaintenance ofglassworkers’ equipment.Sector ‘b’ contained amine-shaft, which antedatesthe main cone structure.Sector ‘c’ was probably the‘lehr’ or annealing area,with a sand store adjoiningthe exterior wall. (DenisAshurst).

the illustrator of the Encyelopédie in 1772.Here the old type of pot is clearly visible. Itseems more likely, therefore, that the devel-opment of the covered pot was associatedwith the evolution of lead crystal, for this isirreparably damaged by the sulphurcompounds which result from the combustionof coal. Curiously enough the earliest mentionof it appears to come from Norway. In 1756at the Nostetangen factory there wereprepared ‘16 English covered pots’. The earli-est pictorial representation would appear todate from as late as 1802 (Newton, 1988).

The theory that there was a southern tradi-tion of furnace building (in which the found-ing furnace was normally circular andincorporated a third storey for annealing); anda northern tradition in which the foundingfurnace was normally rectangular or oblong onplan and the annealing furnace often on a levelwith it and interconnecting by means of acommon fore channel, and usually also by

linnet holes, seems in general to have beenborne out by archaeological discoveries. Thedifferences are discussed by Newton (1985b)in terms of the availability of beech trees. Thenorthern tradition clearly bifurcated at somepoint, probably in the sixteenth century, intothose furnaces that were built with wings forsubsidiary firing processes on the one handand, on the other, two-chamber furnaces of thetype represented in the Mandeville manuscriptand Kunckel’s engravings. These general devel-opments must be viewed in the context of allthe variables imposed in particular cases bysite, available building materials etc.

Twentieth- and twenty-first-century glass-making furnaces, which are outside the scopeof this discussion of ancient and historictechnologies of glassmaking, are of sophisti-cated design: the Siemens regenerativefurnace, cold top electric furnace and floatglass furnace make use of gas, oil or electric-ity as sources of fuel.


Thousands of ancient glass artefacts havesurvived exposure to burial environments.However it is reasonable to suppose that theserepresent only the more durable (stable) of theglasses manufactured in the past, and thosethat were buried in conditions which werefavourable for their survival (and which sincetheir recovery have been stored in conditionsideal for their preservation). Historic glassesthat have never been buried, and Europeanmedieval glass windows (buried and in situ)have also survived in great number.

Much of the investigation into the mecha-nisms of glass deterioration has centredaround the decay of European medievalwindow glass (Newton, 1982b; 1985a). Verylittle has been written concerning the deterio-ration of archaeological glass (Geilmann, 1956,1960; Shaw, 1965; Knight, 1999; Freestone,2001); and the understanding of the processesof its deterioration is limited. There is,however, a substantial body of literature con-cerning investigation and experimental re-search relating to the chemical principlescontrolling the weathering (or durability) ofcontainer glasses (e.g. Hench, 1975b; Paul,1977). Whilst the same scientific principlescontrol the interaction of archaeological andhistoric glasses to their environments, thepurely scientific approach to glass deteriora-tion does not translate easily to archaeology/conservation; and the results contained thereincannot be used to predict the nature of actualdeterioration products which will be formedon complex glasses over centuries or millen-nia, except in the most general terms.

In 1936, Harden stated that

The term weathering is applied to any changefor the worse on the surface ... of glass that

is caused, during the passage of time, bycontact with outside influences ... the termcovers, therefore, a wide range of phenom-ena ... it is quite impossible to foresee whattype or degree of weathering will beproduced on a piece of glass after preserva-tion for a fixed time under seemingly fixedcircumstances ... even a very slight change inenvironment may produce a markedly differ-ent kind or degree of weathering on twoparts of the same vessel ... No strict rules cantherefore be formulated. The causes andeffects of weathering are as manifold as theyare elusive.

This statement is still true, despite the fact thatit has been scientifically proven that deterio-ration is related to the chemical compositionof glass and its environment (Cox et al., 1979;Newton and Fuchs, 1988, Freestone, 2001).

In order to be clear about the processes,deterioration will be used to describe the decayof buried or submerged glass, weathering, todescribe decay to due atmospheric influences,and corrosion avoided, since this term refersessentially to deterioration of metals.

There are three main aspects to the deteri-oration of glass artefacts. Subject to theirenvironment, they may deteriorate (i) as aresult of physical damage, (ii) from superficialdisfigurement, and (iii) from chemical deterio-ration.

Physical damage

The fragility of glass renders it susceptible tophysical damage by mechanical shock. Thecauses of physical damage to glass can result

169

4

Deterioration of glass

from (i) manufacturing defects; (ii) impactdamage; (iii) thermal shock damage; (iv)abrasion; and (v) previous treatment.

(i) Glasses may break if they were poorlycrafted, or were of an impractical design;there were weaknesses in the glass (stria-tions, seed, stone, see Figure 4.1) or thebatch recipe was poorly formulated(leading to chemical instability, discussedbelow). Thin, blown objects may springout of shape due to release of tension inthe glass. Some objects are particularlysusceptible to impact damage: those madeof more than one piece, figures withprojecting limbs, tall, top heavy objects;and those with rounded bases.

(ii) Impact damage is the most common formof physical deterioration to objects held incollections and which are otherwise ingood condition, resulting from accident,careless handling or packing, or vandalism(Figure 4.2). The glass breaks when itcomes into contact with a harder materialor by falling, cracking and breaking.Objects suffer some disruption in the formof scratches, cracks, chips and completebreakage, which may lead to loss offragments. Vibration from pedestrianmovement and all forms of traffic cancause objects to fall over, or move alongand off shelves.

(iii) Damage by thermal shock occurs whensudden warming or cooling of the glass

causes uneven rates of expansion orcontraction. The resulting stress causes theglass to crack or break. This type ofdamage most often occurs when glass iswashed in hot water, or when candles areallowed to burn too close to the glass(Figure 4.3). Glass should never be driedin direct sunlight or by artificial sources ofheat (Bimson and Werner, 1964a). It is notalways possible to tell how glass willbehave under heat stress. Even if heatedto 90°C over 8 hours, glass may crizzle tothe extent of becoming opaque.Enamelled decoration may alter in appear-ance or spring off and gilding becomedulled. Glass subjected to fire will crizzleor shatter. In extreme heat where temper-atures may rise above the original firingtemperature of the glass, glass will melt.Lead glass, for example, may be reducedto black molten lumps. Tar and carbondeposits may be difficult or even impossi-ble to remove, especially from deterio-rated surfaces and within cracks.

(iv) Abrasion of glass occurs as (a) a result ofuse during its lifetime (e.g. scratches fromeating utensils); (b) the use of harsh clean-ing methods; (c) previous conservation/restoration (removal of materials used forrepair and filling), and scratches made byscalpels, abrasives (sandpaper) and metaltools (Figure 4.4). (It has to be borne inmind that conservation techniques havethe potential to harm glass at the time of


Figure 4.1 A stone in glass, which has subsequentlycaused the glass to crack.

Figure 4.2 Large wine cooler broke into three pieces(see Figures 7.25a–b).

application or in the future). Generally anysurface decoration, which was unfired(cold painting and gilding) or was fired atlow temperatures, e.g. gilding, lustre andenamel decoration, tends to be suscepti-ble to abrasion (Figure 4.5). (In additionto physical damage, irreversible chemicaldamage is caused by the improper use ofcommercial cleaning agents, acids andalkalis etc., discussed below.)

Superficial disfigurement

Foreign material, from a number of sources,may accumulate on the surface of glass (andinside hollow vessels). These derive from (i)use; (ii) encrustations during burial; (iii) stains

from metal corrosion products or from metalties or rivets; (iv) excess use of conservationmaterials; and (v) atmospheric pollution.

(i) Stains and residues resulting from use maybe required to be kept as archaeological/historical evidence, e.g. remains of crema-tion, food, drink, medicines, ink etc. Thesemay have become trapped inside closedforms such as bottles, or held within flawsin the glass, such as cracks, chips, pinholes and crizzled surfaces (in the lattercase alkali salts can become trapped).Limescale can form on the interior ofcontainers such as vases (Figure 4.6).

(ii) Glass is normally recovered from excava-tion covered by soil (see Figure 6.1).Thick calcium deposits of organic originmay form during burial (especially in amarine environment), or black sulphideduring burial in anaerobic deposits (Figure4.7). Encrustations of carbonates,sulphates and silicates may form to oblit-erate the surface of buried glass inclimates where there is sufficient precipi-tation to dissolve these compounds (whichare nevertheless regarded as insoluble,since they are poorly soluble in water),but where there is also sufficient evapo-ration to permit them to be depositedagain (Figure 4.8). Being the least solublein water, silicates are the least common,but where they exist in the soil in greatquantity, as in the Near East, a silicatecrust may form on the surface of glass, or

Deterioration of glass 171

Figure 4.3 Glass cracked by the heat of a candle,which had been allowed to burn too close to it.

Figure 4.4 Fragments of glass deliberately scratched byfiling during a damaging attempt at repair.

Figure 4.5 A gilded design, which has almostdisappeared as a result of abrasion.

from mortar (if the glass has been embed-ded in a wall for instance). Glass incor-porated into a building structure may becovered with mortar.

(iii) Metal deposits can result from burial ofglass in close proximity to deterioratingmetal objects (Figure 4.9), or from metalrivets, ties and dowels used to repairhistoric glass objects (Figure 4.10).

(iv) Excess use of materials for repair/restora-tion such as pressure-sensitive tapes(Figure 4.11), adhesives and filling materi-als can leave deposits, or become trapped


Figure 4.6 Limescale deposits inside a glass vessel.

Figure 4.7 A Roman silica–soda–lime glass afterconservation, which has a thin opaque blackenedsurface, thought to have resulted from its burial inanaerobic conditions, where hydrogen sulphide becamedeposited with the leached surface. (Courtesy of S.M.Smith).

Figure 4.8 A deposit of insoluble salts, blending withthe weathering crusts and obscuring the surface of theglass.

Figure 4.9Iron corrosionproductsdeposited on aglass flask,duringprolonged burialnext to an ironobject.

in a deteriorated glass surface. Adhesivelabels can accentuate the deterioration ofglass, presumably by holding moisture incontact with the glass (Figure 4.12).

(v) Atmospheric pollution in the form ofwater- or wind-borne chemicals, andaccumulating particulate matter, causedeterioration of glass (Figure 4.13). Thetype and concentration of pollutingagencies will depend on location. Thesemay include incomplete fuel combustionassociated with traffic and industry, build-ing materials, and skin and clothing parti-cles introduced by visitors.

Chemical deterioration

Factors affecting chemical deterioration

The chemical processes associated with thedeterioration of ancient and historic glassesarise as the result of the internal composition


Figure 4.10 Rivets used to repair the foot of awineglass, which have subsequently corroded.

Figure 4.13 Medieval glass which has formed anencrustation due to prolonged exposure to theatmosphere. Part of the crust has fallen away, exposingthe blue glass beneath. (Courtesy of S. Strobl).

Figure 4.12 Damage to the base of a wineglasscaused by prolonged contact with a gummed paperlabel. (Victoria and Albert Museum, London).

Figure 4.11 Gilding removed from the surface of aglass vessel, on Sellotape used to tape fragmentstogether.

of the glass being attacked on a molecularlevel by external forces, principally involvingwater (Figure 4.14). In simple terms, when aglass reacts with water, or with an aqueoussolution, chemical changes occur at the glasssurface, which then progress into the body.Two mechanisms are involved: de-alkalization(commonly referred to as leaching), andnetwork dissolution.

Besides the presence of water, other inter-dependent factors affecting the decompositionof glass, fall into two main categories: (i) thoserelating to the glass itself, and (ii) those relat-ing to the environment. Those relating to theglass are the nature of its surface, compositionof the bulk glass, phase separation or otherinhomogeneities in the glass, and the firingprocess. Those relating to the burial, aqueousor atmospheric environment, are the amountof water or water vapour present, the natureof the attacking solution, i.e. the pH, theparticular ions present and their concentration,the presence of complexing agents and salts,the time of exposure to attack, temperature,organisms, and to some extent pressure, in thecase of burial at depth on land or beneath thesea (see Figure 4.14). Glass in collections, andto a much greater extent, medieval windowglass in situ, is affected by atmospheric pollu-tion, condensation, humidity and solarization.

The presence of water and, above all,accumulated moisture, is the most important

factor initiating and sustaining forms of glassdecomposition. Without water, glass canremain in excellent condition for centuries,little, if any, attack occurring in dry conditions.Water is essential for the replacement byprotons of the diffusing alkali ions (leaching)and the subsequent hydration and dissolutionof the silica network. It also removes thesoluble salts formed from such reactions. Evendamage caused by dehydration is the conse-quence of previous attack by water. Literaryreferences to the leached glass layer refer toit in different terms, such as being alkali-deficient, silica-rich, or as hydrogen glass, orthe gel-layer.

The damaging effect of water on glass wasappreciated as early as the sixteenth century.In a letter dated 1595, concerning glassimported from Venice, the instructions fromthe importer stated that the glass should be’carefullij packt up and with thorou drij weeds,for if the weeds be not well drijed or doe takeanij wett after theij be packt theij staijne andspoijle the glasses’ (Charleston, 1967).

The fact that water is the primary agent ofthe environment for causing the deteriorationof glass, was established by Lavoisier as earlyas 1770. The earliest scientific examination ofweathered glass was that undertaken byBrewster (1863), on iridescent glasses fromNineveh; which demonstrated that the play ofcolours on the glass surface resulted from the


Figure 4.14 Processes of glass deterioration.

Bulk (mass) glass

Composition:SiO2, Na2O, K2O

CaOAl2O3, P2O5

MgO, MnO, FeO,CuO

trace elements

Leaching andnetwork

dissolutionSlO2 K2

+

Al2O3 Fe2+

P2O3 Mn2+

H+H+

H2OH2O

H2O H2O

interference between rays of light reflectedfrom thin alternating layers of air and weath-ered glass crusts. If the gaps were injectedwith water the iridescence disappeared, onlyto return when the water evaporated. Brewster(1855, 1863) and Loubengager (1931) foundthat deterioration of the glass surface by waterproceeded downwards faster than it didsideways; a fact subsequently verified inexperiments undertaken during the twenty-firstcentury.

De-alkalization (ion exchange)In Chapter 1 it was shown that the negativelycharged alkali and metal cations are free tomove around within the glass network, fromone space to another. In a damp or wetenvironment, they are extracted from the glassby water, to form a sodium or potassiumhydroxide solution. However, since the electri-cal neutrality of the glass must be maintained,hydronium ions (H3O

+) from the waterexchange positively charged hydrogen ions(protons) for the alkali ions leaching from theglass network. This process results in ahydrated silica-rich surface layer, sometimesreferred to as a gel layer. Water moleculesreact with the non-bridging ions in the glassto produce hydroxyl ions, which migrate outof the glass with the alkali cations. The waterthus contains an excess of hydrogen ions,which increases its alkalinity and potential forattacking glass.

H3O+ + =Si–O––Na + → =Si–OH + H2O + Na

(4.1)

Since the hydrogen protons are much smallerthan the sodium or potassium ions theyreplace, the alkali-depleted surface layer willhave a smaller volume than the underlyingglass. Furthermore, the hydrogen ions have astrong covalent bond with oxygen, whichresults in a contraction and weakening of theglass network (Ernsberger, 1980).

Potash glasses have about half the durabil-ity of soda glasses because the potassium ions,are larger than those of sodium ions, and takeup more space in the glass network. Thus,when potassium ions are leached out of theglass, they allow a greater number of watermolecules in. The resulting reduction involume of the leached layer causes shrinkage

to occur, and further shrinkage takes place ifa hydrated alkali-deficient (hence silica-rich)layer then loses water, e.g. upon excavation.Thus dehydration of glass, often thought of asa cause of deterioration, merely highlights thedeterioration which has already occurred. Thedecrease in volume can lead to microporosityof the surface layer and this may be the causeof the many-layered effects found in thesurface crusts of some medieval glasses(Scholze et al., 1975).

If, however, the alkali-deficient surface layerremains undamaged, at least for a time, it maycause the alkali diffusion rate to decrease,since the ions must diffuse through it in orderto reach the unaltered glass. Therefore the rateof extraction, which initially decreased withthe square root of time, becomes slower, andlinear with time. (Thus a false impression canbe given that the glass has become moredurable as time progresses; Hench, 1975a,b;Hench and Sanders, 1974; Clark et al., 1979.)

The nature of the leached surface layer onglass is of fundamental importance in under-standing glass deterioration/durability. Unlessthe leach solution is frequently renewed bywater, the accumulation of alkali ions in thewater leads to an increase in the pH value.The speed at which the critical value of pH 9is reached (at which point more rapid de-alkalization occurs), depends on manyparameters, two of these being the surfacearea of the glass (SA) and the volume (V) ofwater involved.

Network dissolutionThe water of the attacking solution will usuallybe either slightly acidic or alkaline dependingon the soil or the atmospheric conditions. Asmentioned above, under alkaline conditions,the oxygen bridges of the glass network itselfmay be broken down and the silica dissolve:

=Si–O–Si + =OH– → =Si–OH + =Si–O–

(4.2)

This reaction is strongly pH-dependent (e.g.Adams, 1984). In general, the rate of alkaliextraction from glass in a buffered system ata pH of less than 9 is constant and indepen-dent of pH. There are exceptions to this, forexample in a soda–lime–silica glass where thelime content is greater than 10 mol per cent


there is a rapid increase in the rate of extrac-tion when the pH drops below 3. Theincreased extraction is due to the increasedsolubility of lime, whose extraction was other-wise negligible. In contrast to alkali extraction,silica extraction is almost non-existent belowpH 9, but increases noticeably above this. Atthe critical point of pH9, both the silica andthe divalent network modifiers (calcium,magnesium, lead etc.) can be leached out ofglass, and eventually total dissolution of thesurface will occur, opening up the glassbeneath to further attack.

In effect, the processes of glass deteriora-tion, i.e. ion exchange and network dissolu-tion, are in competition. If leaching of alkaliproceeds more rapidly than dissolution of thenetwork, the glass surface will develop aleached layer composed of hydrated silica. If,however, network dissolution is the predomi-nant mechanism, a hydrated silica layer willnot form and the glass will be graduallydissolved away. Since most ground water is ofintermediate pH, ion exchange appears to bethe dominant mechanism, so that a hydratedsilica-rich layer is formed on many buriedarchaeological glasses. Network dissolutionappears to be a more progressive phenomenain glasses above ground, for examplewindows. Whereas in the ground alkalisleached from the glass are washed away, inthe atmosphere they tend to remain as saltson the glass surface, creating alkaline condi-tions, which favour breakdown of the network(Freestone, 2001).

The glass

The nature of the glass surface

The surface of a glass is highly influential inthe effects of the deterioration processes thatoccur on it since the amount of constituentsreleased during decomposition is proportionalto the exposed surface area (SA/V). Thesurface is a defect in itself, because all surfaceions are in a state of incomplete coordination.This asymmetry of the surface producesabnormal interatomic distances and hence thespace occupied by the surface ions is greaterthan usual, enabling replacement by ions of alarger or smaller radius to occur. Thus, the

subsurface glass layer is porous on a micro-scopic scale, and this enables surface reactionwith molecules of water, sulphur dioxide,oxygen and hydrochloric acid etc. to takeplace to a considerable depth.

As the ratio of surface area to volume(SA/V) of the leaching solution increases, thereis an increase in the extraction of silica.However, this can be attributed to the accom-panying increase in pH of the solution onrelease of the alkali. In contrast, the quantityof alkali extracted does not vary with chang-ing SA/V. It would be expected that theincrease of pH would suppress further releaseof alkali from the glass, but this possibility iscounteracted by the release of silica from thenetwork as the pH increases (Newton andSeddon, 1992). This dissolution of the silicanetwork causes alkali to pass into solution,and reduces the thickness of the surface gel(leached layer).

The original glass surface would rarely havebeen homogeneous, even though it may haveoriginally had a slightly protective surface oncompletion of manufacture. The surface wouldhave been physically altered by internal orexternal stress, for example by cold working(Pilosi in Kordas, 2002), or accidental damage.On the other hand, a slightly protective surfacecould be formed, at least in the initial stagesof surface alteration (see above). In eithercase, the surface would have begun to bechemically altered (leached by moisture) fromthe moment it was made. As a result of theseinterferences, the glass surface would bemicroscopically uneven, or macroscopicallyrough, thus presenting a greater surface areato the attacking solution (SA/V). In otherwords, the glass would have sites of damage,which would be more prone to deteriorationthan others.

The composition of the bulk (mass) glass

The composition of glasses, i.e. the ratio ofsilica, alkali (soda, potash or less commonlylead) and calcium oxide (lime), additives suchas metallic oxides and opacifiers and the inclu-sion of trace elements, all have a part in deter-mining their durability. The majority of ancientand historic glasses are of soda–lime–silicatecomposition: a network of silica and othermetallic oxides in the roles of glass formers,


modifiers and intermediates (see Figures 1.4and 1.5). Network formers are limited to thoseoxides which have high bond strengths, themost common being silicon oxide (silica).Theoretically, fused silica would form a highlydurable glass, but such high temperatureswould be required to melt it that, for practi-cal reasons, modifiers always had to be addedto act as fluxes, which lower the meltingtemperature. The concentration of silica is acrucial factor in determining decomposition ofglass. Experimentally, it has been shown thatbelow 62–66 mol per cent, the silicon atomsin the glass become associated with amodifier, sodium, calcium etc., as their secondneighbours. Consequently, there is always aninterconnecting path of neighbouring siliconoxide (Si–O) groups, which provide suitablesites for the movement of interchanging ionsbetween the solution and the glass. Thedecaying glass has a tendency to form asurface crust. Above 66 mol per cent of silica,the SiO2 groups are isolated by Si–O–Sigroups, which suppress the movement of ions involved in the leaching process(Bettembourg, 1976; Collongues, 1977; Cox etal., 1979; Perez-y-Jorba et al., 1975; Schreiner,1987; Newton and Fuchs, 1988).

As mentioned above, modifiers break up thesilica network, bonding ionically with theglassy network and altering properties such asviscosity, thermal expansion and durability.They have relatively low bond strengths andfall into two main groups: fluxes and stabiliz-ers. Fluxes, materials with the chemicalcomposition R2O, the oxides of sodium(Na2O), potassium (K2O), present in about 15per cent, which reduce the viscosity of thebatch. It is largely the presence, type andquantity of the modifiers which impair thehighly durable nature of a pure silica network.

Alkaline earth oxides with the chemicalcomposition RO, usually lime or calcium(CaO) and magnesium (MgO), are referred toas stabilizers. Present in about 10 per cent ofthe batch, they prevent crystallization fromoccurring as the batch cools, and improves thechemical stability of the glass, making it lesssoluble in water. If lime (CaO) is added to theglass in increasing amounts up to 10 mol percent, there is a rapid decline in soda extrac-tion, owing to the increasing stability of thesurface (gel) layer. The stability is due to the

presence of CaO increasing the coupling ofthe vibrational modes of the silica non-bridging oxygen modifier bonds to thebridging of the Si–O–Si network. It would beexpected that the replacement of one Ca2+ ionby two protons (H+) would have the sameeffect as replacing two ions from the network,but in the latter case a much more porouslayer is formed. If the lime content is increasedabove 15 mol per cent, the resistance to deteri-oration starts being drastically reduced. Limewas not specified as a constituent of pre-seventeenth-century glasses, and hence stable(durable) glasses were prepared more or lessaccidentally either from calcareous sands (inancient times) or from high-lime wood ash (inmedieval times); thus it is not surprising thatthere was a period when the addition of extralime was regarded as being positively harmful(Turner, 1956a).

Some minor components in glass such asalumina (Al2O3), phosphorous (P2O) and iron(Fe2O3), are intermediates present in glass ina forming or modifying position. They derivefrom (i) trace elements in the raw materials,(ii) extractions in the clay crucibles or pots inwhich the glass was melted, or (iii) haveentered the glass network from the surround-ing environment. The intermediate oxides allhave a low solubility (intermediate bondstrengths), and therefore their presence withina glass surface may increase its resistance todissolution, by immobilizing the alkali ions, sothat they are no longer free to move throughthe silicate network.

As mentioned above, the greater thepercentage of silica, the less tendency there isfor it to be extracted from the glass, whereasin the case of alkali, the greater the percent-age, the greater its potential to be extracted.In general the greater proportion of glassmodifier, such as calcium, to the alkali sodiumor potassium, the more stable the glass.However, if a glass containing large amountsof alkali does deteriorate, the calcium will beleached out along with it. The remaining silicamay be so small a proportion as to be unableto maintain the internal bonds, thus the glasswill disintegrate.

It was formerly believed that, becausedifferently coloured glasses decay in differentmanners, the colour per se influenced theweathering of medieval window glass. This


arose from the mistaken impression that allglass made for the same window (and hencemade at the same time by the same glass-maker) would have essentially the samecomposition (Newton, 1978). In fact, there waslittle control over raw materials in medievaltimes, and the sources of colouring oxidesmight be highly contaminated by, for example,clay or lime, both of which would profoundlyaffect the weathering of glass made with it. Inaddition, the concentration of colourant ionsis low, usually less than 1 per cent, and istherefore not significant. It is the bulk compo-sition of the coloured glass which is different,and which explains the difference in durabil-ity. For example, in 1972, attention was drawnto a late twelfth-century border panel from thenave clerestory of York Minster in which allthe pink glass had resisted weathering, andhad shiny surfaces, whereas the green glasshad rough crusted surfaces. Partial analyses ofthe glasses revealed that the green piecescontained relatively more lime, compared withpotash, than did the pink pieces and it wasthis difference in composition, rather than thecolour per se, which was responsible for thedifferences in weathering (Hedges andNewton, 1974). In some cases colouredmedieval glass is found to be better preservedthan colourless glass.

In view of the above discussion concerningthe effects of composition, it can be appreci-ated that the rate of deterioration ofancient/historical glasses is greatly influencedby production technology, governed in turn bycultural and geological factors (see Chapters 2and 3). The difference in composition of glass,for instance, is the reason for the difference indeterioration between relatively stable Medi-terranean soda–lime–silica glass and poorlydurable medieval glass, made from plant andbeechwood ash, and containing too muchpotash and sand lime to ensure its stability.The hydrated layer on much Roman glass witha typical composition of 15–20 per cent Na2O,5–10 per cent CaO, 2–3 per cent Al2O3 andcirca 70 per cent SiO2, is negligible, only a fewtens of micrometres thick depending upon itsburial environment. On the other hand,medieval European forest glass, which is lowin silica, high in alkali (potash) and alkalineearths, is notoriously susceptible to deteriora-tion, and typically develops substantial

hydrated layers, even after a few centuries. Forexample, such a glass might contain 15 percent potassium monoxide (K2O), 15 per centcalcium oxide (CaO3) and 55 per cent silicondioxide (SiO2), plus small quantities (circa 1per cent) of other oxides, e.g. iron andmanganese, derived from the use of impuresand, but mainly from the wood ash.

Soda–lime–silica glasses produced fromplant ash in the Near East also differ from theglasses described above, in that they have areduced silica content, lower aluminium andhigher alkaline earths (calcium and magne-sium). This results in their being less stablethan Roman soda–lime–silica glass. Forexample, glass found in Mesopotamia, datingfrom the Sassanian period (fourth century AD),can be excavated with opaque surface crustsseveral millimetres thick, alongside Romanglass having a thin iridescent skin. Glass ofthis type may be recovered in an extremelygood state of preservation from the Egyptiandesert because of the very low ambienthumidity. There is little data concerning thedeterioration of lead–silica glasses. However,the lead-rich opaque red enamels on medievalCeltic metalwork are often extremely heavilydeteriorated in comparison with other (lowlead) colours. Chinese glasses producedduring the Han dynasty (206 BC to 220 AD) ofbarium–lead–silica composition (see Chapter2), frequently developed a thick deterioratedsurface layer during burial.

Colourless cristallo glass was made inVenice from the fifteenth century, by purifyingplant ash by the solution and precipitation ofthe alkalis. Since the stabilizing alkaline earthswere also removed during the process, theresulting low-lime glass was inherently unsta-ble and susceptible to attack by atmosphericmoisture. Such glass is said to be sufferingfrom glass disease, or to be sick or weeping(see Figure 4.15). As mentioned above, theoptimum lime content is about 10 mol percent of CaO in the final glass, but some glasseswere made in the middle of the seventeenthcentury by Ravenscroft (particularly in theperiod 1674–76), and in Venice at the begin-ning of the fifteenth century, which now havecrizzled or weeping surfaces (Freestone, 2001).Since the early stages of glass deteriorationoccur at a molecular level, they may gounnoticed. Later the glass object may appear


dull or misty as deterioration productsaccumulate on the surface. In humid condi-tions, these attract moisture to the surface,which may then exhibit drops of moisture ora general slipperiness (Figure 4.15). If theatmosphere becomes drier, alkali salts aredeposited on the surface of the glass (Figure4.16). As the process continues, a network ofminute fractures (crizzles) develops in thesurface layers of the glass, a stage known asincipient crizzling. Under magnification, andlater with the naked eye, a network of veryfine cracks can be seen (Figure 4.17). If theprocess continues unchecked, the network ofcracks extends deeper into the glass surface,causing small flakes to spall away. The glassbecomes entirely crizzled and exhibits anoverall cloudiness and lack of transparency. Inextreme cases, the glass may lose its mechan-ical strength and collapse, disintegrating

entirely. The term crizzling has been used todescribe such deteriorated glass since at leastthe sixteenth century. Although these are theclassic examples, many other examples arefound in glass from other countries andcenturies. These glasses are characterized bylime contents which are usually less than 5mol per cent CaO (and sometimes less than 1per cent) and their situation is frequently madeworse because they contain potash as thealkali instead of soda. It was formerly believedthat pink glasses had a greater tendency tocrizzle than clear glass. However Brill (1975)showed that the pink colour of crizzled glassis confined to the surface of otherwise colour-less glass. The phenomena of weeping andcrizzling may also occur on enamelled metalobjects. Chemically unstable glass has beeninvestigated by Bakardjiev (1977a,b), Scholze(1978), Ryan et al. (1993, 1996), Hogg et al.


Figure 4.15 Close-up view of a weeping glass vessel,showing the drops of alkaline moisture, which form onthe surface. (© Copyright The British Museum, London).

Figure 4.16 Deposits of alkali salts on the interior ofa weeping glass ewer. (© Copyright The BritishMuseum).

(1998, 1999) (other articles are published inthe Proc. of the 18th Int. Congr. on Glass, SanFrancisco, 1998, CD ROM).

The triangular diagramIn 1975 a triangular diagram (Figure 4.18) wasformulated, on which the compositions ofglasses were plotted, as an aid to understand-ing the relative weathering behaviour of differ-ent glasses, (Newton, 1975; Iliffe and Newton,1976). The diagram remains a good generalguide to which chemical compositions of glassare more durable than others.

The compositions of the glasses are obtainedby analysis, and converted to molecularpercentages of the constituent oxides. Theseare then grouped into three categories, thenetwork formers, the alkaline earth networkmodifiers and the effective alkali content.However, the fact that most of the glasses lieon the line Z to J in Figure 4.18, means that,in the majority of cases, it is necessary only toknow the molar proportion of network formersin order to establish approximately where theglass will lie in the diagram. Thus, as a firstapproximation, the problems of analysis andcalculation can be reduced by determiningonly the silica content.

The constituents of the glass exercise theireffect on the weathering of that glass by virtue

of the numbers of their molecules, in associa-tion with each other, and not by the weightof those molecules (which is the conventionusually adopted, in reporting the results ofanalyses of glasses). The difference betweenthe numbers of molecules and the weight of the molecules would not be significant ifall the molecules had similar weights, but thisis by no means the case. The two main alkalisdiffer considerably, the molecule of potash(K2O) being 52 per cent heavier than themolecule of soda (Na2O). A molecule of lime(CaO) is 39 per cent heavier than the moleculeof magnesia (MgO), and lead oxide is 550 percent heavier than MgO. To obtain the relativenumbers of molecules, the weight percentagesare divided by the factors in Table 4.1 (which


Figure 4.17 A crizzled posset pot, on which thenetwork of fine surface cracks can be seen. (©Copyright The British Museum).

Figure 4.18 A triangular representation of glasscompositions in which highly durable glasses (A, M, R)are placed near the centre, and the least durable glasses(H, J, K) fall at the bottom and to the right; glasses ofthe composition W and Z may even decay in somemuseum atmospheres. A = modern float glass; B =window glass (1710) from Gawber, Yorkshire; C = sheetglass (1855) from St Helens, Lancashire; D = uncrustedglass (1535) from Bagot’s Park, Staffordshire; E = glassfrom Rosedale; F = crusted glass from Bagot’s Park; G =crusted glass from Hutton, Yorkshire, all sixteenthcentury; H = badly decayed, heavily crusted medievalAustrian window glass; J = badly decayed, heavilycrusted glass medieval glass from the Lady Chapel, ElyCathedral; K = medieval glass from Weobley Castle; M= durable Saxon glass from Monkwearmouth; R =Roman glass; P, Q = pink and green twelfth-centuryglass from York Minster; W = ‘weeping glass’; Z =‘crizzled glass’. (Iliffe and Newton, 1976).

are each one-hundredth of the correspondingmolecular weight), and then the results areadjusted so that their total is 100 per cent. Themethod of calculation is shown in the workedexample given in Table 4.2. The first columnis the composition of the glass in terms of itsconstituent oxides; the second column is theusual weight percentage distribution of theseoxides; and the third column gives the factors,taken from Table 4.1, by which the weightpercentages are divided. The fourth columngives the relative molar proportions, and theirtotal is 162.4, so that each has to be dividedby 1.624 in order to obtain the molar percent-ages given in the fifth column. It should benoted that the weight percentages of MgO,CaO and K2O are quite different, but the molarpercentages are almost the same.

These molar proportions of oxides (Table4.2) are combined according to the followingrules: the alkaline earth component, RO, isobtained by adding together all the oxideswith this formula, i.e. CaO + MgO + MnO +CuO + PbO etc. In this example it is CaO +MgO, or 11.0 + 10.8 = 21.8 per cent.

The network formers are almost entirelysilica (SiO2) but forest-type glasses oftencontain as much as 5 per cent P2O5 and thismust be added to the SiO2, as well as anyTiO2, ZrO2, SnO2 etc. In addition, it was shownin Chapter 1 that Al2O3 occupies a specialplace, each molecule being able to immobilizean alkali ion, and it can also be incorporatedinto the network. Thus the total of thenetwork oxides has to be increased by twicethe A12O3 (and similar oxides, such as Fe2O2,Cr2 O3 etc.). In the example, the total ofnetwork-forming oxides (called ‘SiO2’ becauseit is mainly SiO2) is given by SiO2 + 2(A12O3)= 65.9 + 1.4 = 67.3 per cent.

The remaining component in the triangulardiagram is the alkali oxide, called ‘R2O’ becauseit represents only the available alkali, and notthe total alkali, some of the alkali being immobi-lized by the alumina. Thus ‘R2O’ = K2O +Na2O–A12O3 = 10.9 + 0.7 – 0.7 = 10.9 per cent.

The total of ‘RO’ + ‘SiO2’ + ‘R2O’ = 21.8 +67.3 + 10.9 = 100 per cent. When the threevalues are plotted in the triangular diagram(Figure 4.18), it should be noted that in each


Table 4.1 Factors for converting weight percentages into relative molar percentages, forfrequently occurring glass-making oxides

Oxide Factor Oxide Factor Oxide Factor

SiO2 0.601 PbO 2.232 Al2O3 1.020K2O 0.942 CuO 0.796 Fe2O3 1.597Na2O 0.620 MnO 0.710 Sb2O3 2.915Li2O 0.299 ZnO 0.814 As2O3 1.979CaO 0.561 CoO 0.749 B203 0.696MgO 0.403 TiO2 0.799 P2O5 1.420BaO 1.534 Sn02 0507 SO3 0.801

Table 4.2 Worked example, converting weight percentages to molar percentages

Oxide Weight percentage Factor Relative molar proportions Molar percentage

SiO2 64.3 0.601 107.0 65.9K2O 16.7 0.942 17.7 10.9Na2O 0.7 0.620 1.1 0.7CaO 10.0 0.561 17.8 11.0MgO 7.1 0.403 17.6 10.8Al2O3 1.2 1.019 1.2 0.7

100.0 162.4 100.0

case, the coordinate lines run parallel to theside which becomes the base when the trian-gle is turned so that the numbers can be readin a horizontal manner.

In Figure 4.18, the upper, central areaindicates the range of durable glasses. Thefour points D to G confirm the fact that glassdurability decreases when the silica content isless than 66.7 mol per cent, the line separat-ing D and F being at 65 mol per cent ofnetwork formers. The glasses in the bottomright-hand corner are poorly durable. The tworemaining points, the stars at W and Z,indicate, respectively, weeping and crizzlingglasses. Thus it can be seen that the durabil-ity of glass decreases very markedly as soonas the composition of the glass moves a littleto the left of the group A, M, R. The impor-tance of the RO (lime) content can now beseen by considering the glasses Z, M, A, P, D,F, J; they all have about 15 mol per cent ofavailable alkali but the RO varies from lessthan 5 mol per cent in Z to nearly 40 mol percent in J and the durable glasses lie in anarrow band between 10 and 20 per cent RO.

Whilst the triangular diagram is a convenientmethod of displaying the effects of differencesin glass composition, it does have limitations.For example, it does not discriminate betweensoda and potash, yet the soda glasses haveabout twice the durability of the potashglasses, or between the effects of lime andmagnesia. Neither does it discriminatebetween different types of glass surface. It iseven probable that there is a central plateauof very highly durable glass, and that thedurability will fall off, suddenly and precipi-tously, especially as the composition moves tothe left but also as one moves towards thebottom right-hand corner; if so, it could havea bearing on the existence of highly durablepre-first century AD glasses and completeabsence (through deterioration) of glasses withless durable compositions.

The environment

Environmental factors affecting decompositionare: (i) water in the context of damp to water-logged wetland sites, or underwater burialsites; or in the atmosphere both inside andoutside buildings (precipitation, humidity,

condensation); (ii) pollution of the atmosphereby gases which combine with water toproduce acid solutions and/or form depositson the glass. The effects of these agencies arelinked to time and temperature. In the case ofdeep land or underwater burials the effects ofpressure may also have to be considered (seeFigure 4.17). Other environmental factors tobe considered in connection with the deterio-ration of glass are those of temperature andtemperature fluctuations, solarization, micro-organisms, and, more recently, vibrationscaused by road, rail and air traffic.

Burial

Physical aspectsPhysical damage can occur as the result ofanimal or root action, ploughing, soil pressure,or land movement. Buried glass may also havebeen broken accidentally or deliberatelybefore burial (e.g. glassmaking chunk glass,cullet, wasters, ritual etc.). Physical disintegra-tion of glass may also occur as the result ofrepeated dissolution and crystallization ofsoluble salts within it, particularly on exposureto the air (Vandiver, 1992; Freestone, 2001).

Chemical aspectsSince water is the primary agent of corrosion,glass buried in a dry environment might beexpected to survive indefinitely. However,even in desert areas, water tables can rise dueto annual river flooding, irrigation, habitation,dam construction and earth movement. In adamp or wetland environment, where glass iscontinually exposed to water, its constituentswill be leached out, particularly if the attack-ing solution is alkaline in nature. Glass ishardly affected by weak acid pollutants.

Newton and Seddon (1999) discuss variousconditions where water could becomesaturated with ions and thus have a reducedleaching effect on glass (e.g. within stagnantwaterlogged sites such as the bottom ofdisused wells). In soils, especially in peat orother humic horizons, there are numerouschelating or complexing agents, such asamines, citrates and acetates, many of whichcan have a detrimental effect on the durabil-ity of buried glass.

In Western Europe, the burial environmentbelow circa 30 cm is much more stable than


above ground. The RH is rarely less than 98per cent, and the temperature is low and fairlyconstant. Above ground the RH varies widelyand the temperature can change by severaldegrees within minutes. However, despite thehigh RH, the soil is not waterlogged; rather,each particle of soil is surrounded by a thinlayer of water, so that the soil atmosphere (thespace between the particles) is in close contactwith water, and almost saturated.Consequently glass that on excavation appearsdry, may in fact be hydrated and suffer a rapidand irreversible change in appearance, as thewater evaporates.

Visible symptoms of deterioration

In general, glasses of a soda–lime–silicacomposition are more stable, and thereforesurvive burial better than those of asilica–potash–lime composition. The mainvisual varieties of glass surface alteration aredulling, iridescence and the formation oflamellar or enamel-like crusts, which may havebecome blackened. Deteriorated surfaces mayalso exhibit cracking, flaking and the forma-tion of pits. Glass may also be recovered fromburial in an apparently undeteriorated condi-tion, appearing shiny, transparent and reflec-tive. Rapid, irreversible changes in theappearance of heavily deteriorated glass canoccur on exposure to the atmosphere (i.e.upon excavation). It may lose up to 23 percent water by weight, becoming opaque andfragmentary.

DullingDulling describes the condition of glass whichhas lost its original clarity and transparency,and become translucent (Figure 4.19). It iseasily distinguished from dulling that is due toscratches, abrasion or stains.

IridescenceIridescence describes a variegated coloration ofthe surface of glass, sometimes occurring alone(Plate 5), and sometimes associated with othertypes of weathering. The thickness of thesurface layers containing a concentration ofmetal oxides (in the order of hundreds ofnanometres) causes light interference, resultingin the vivid iridescent colours, such as gold,purple and pink. When found alone, it is first

visible in filmy patches, and it may thenbecome abraded in powdery form. If undis-turbed, the iridescence may develop into athick layer, which may flake away. Thick layersof glass flaking away may eventually weakenthe glass so severely that it collapses entirely.

Spontaneous crackingSpontaneous cracking of glass occurs as theresult of deterioration processes.

(i) The formation of a thick hydrated surfacelayer, which then undergoes dehydrationwith extensive loss of silica and thus adecrease in volume. The resulting strainon the surface is relieved by the formationof fractures. The process is likely to be thecause of a surface being covered with anetwork of tiny cracks similar in appear-ance to frost on windows (termed frost-ing). The cracks may be confined to thesurface layers, or penetrate deeper (Figure4.20), allowing ingress of attackingsolutions, and may eventually lead to thetotal collapse of the glass (Wiederhorn,1967). In such cases, moisture within thecracks and layers of decomposed glass,may be all that holds the object together,so that it disintegrates on drying (see alsocrizzling, below).


Figure 4.19 Dulling, the simplest type of weathering,in which glass loses its original clarity and transparency.(Courtesy of J.M. Cronyn).

(ii) The outer layers of hydrated silica are nodifferent from silica gel (and are referredto by glass scientists as the gel layer).Being hygroscopic, they behave in thesame manner in response to changes inatmospheric water (relative humidity) orliquid water. The layers absorb and desorbwater (vapour) at the same time expand-ing and contracting, which leads to theformation of microscopic surface cracks.These allow agents of destruction topenetrate deeper into the glass.

(iii) Spontaneous cracking following the physi-cal scratching of glass (on a fragment ofRoman window glass from GreatCasterton, Leicestershire) was reported byHarden (1959) (see Figure 4.21): ‘Theoriginal scratches (of unknown depth anddate) have subsequently developed

subsidiary branches. After some onethousand eight hundred years’ burial thecentral cracks are some 1.5–2.5 mm deepand the side branches are 0.3–0.7 mmdeep.’ Cramp (1975) also reported thistype of spontaneous cracking on glassfrom the Monastery at Jarrow (UK), butincorrectly described the phenomenon asa feather pattern incised on the surface.

(iv) A relatively slight scratch in the surface ofthe glass may subsequently lead to theformation of shell fractures on one side ofthe crack. These characteristic fracturesgrow in a curved manner and reach thesurface of the glass again so that a smalllenticular fragment can leave the surface.A pit remains, which is usually elliptical inshape, except where it abuts the scratchthat initiated it, and is shallow, the sidesbearing conchoidal fracture marks.Scratches in the surface may penetrate asoluble surface layer and thus lead togreatly accelerated corrosion beneath thescratch.

Milky or enamel-like surfacesAs the name suggests, milky or enamel-likedeterioration products are usually opaquewhite, but they may also be light brown ormottled brownish-black in colour. In its incip-ient stage, when visible merely as small spotsor streaks of white, it has been termed milkyweathering, and may sometimes be confusedwith stone inclusions. The spots, which repre-sent sites of deterioration, gradually progressinto the body of the glass. It may flake awayin small crystals, leaving pits in the glass.


Figure 4.20 Spontaneous cracking of glass duringburial, termed frosting, since the overall effect,resembles frost on windows.

Figure 4.21 Spontaneous cracking, in the form of‘feather-type’ cracks.

On the other hand, the patches of milkyweathering may remain and develop over theglass surface, forming layers ranging in thick-ness from a few micrometres to millimetres.The most developed form, enamel-like weath-ering, appears as a thick coating varying incolour from white to brownish-black, over alarge part of the entire surface of an object(Figure 4.22). This too has a tendency to chipoff, exposing highly iridescent pits and thinlamellae. When highly deteriorated, with littleor no recognizable glassy core remaining, anobject may collapse (Figures 4.23 and 4.24). Adeteriorated glass surface may actually consistof many layers, usually sub-parallel to theoriginal surface. (It has formerly beensuggested that these layers may have offereda method of dating glass, since it wasobserved that the number of bands on someglasses was close to their calendar ages inyears. However, this is not the case, seeChapter 6.) The layers may represent varia-tions in the hydrated silica concentration, since

in many decay products these are virtually theonly components present, with only smallconcentrations of sodium or potassium and ofcalcium. In addition to leaching of the alkali,


Figure 4.22 The surface of a glass torso of Aphrodite,covered by a thick milky-white weathering surface,which in some areas has flaked away to reveal aniridescent layer beneath. H 95 mm. Circa first to secondcentury AD. Near East or Italy. (Corning Museum ofGlass).

Figure 4.23 Section through a fragment of deterioratedglass, showing the invasion of the interior. (Courtesy ofJ.M. Cronyn).

Figure 4.24 A fragment of glass with a thickblackened enamel-like surface. The glass has become soweakened through deterioration that pieces are fallingaway. (Courtesy of J.M. Cronyn).

less soluble chemical components of the glassmay be concentrated in the surface layers. Forexample, where lead is present, even in lowconcentrations, lead salts (carbonates,phosphates, sulphates) may be precipitated asthe result of lead leaching out and anioniccomponents leaching in. Similarly iron,aluminium, manganese, tin and antimonyoxides, all of which commonly occurred inancient glasses, may be enriched in the deteri-oration layers.

Black discolorationExcavated glass may have an opaque black-ened layer on the surface (see Figures 4.8 and4.24). The darkening or blackening may bedue to (i) oxidation of iron and manganeseions present in the glass (see below) (Shaw,1965, Knight, 1999); (ii) the action of sulphur-reducing bacteria actively producing hydrogensulphide in anaerobic conditions; or (iii),rarely, due to the formation of lead sulphide.This only occurs in glasses having a high leadcontent (not usual in medieval glass), buriedin anaerobic conditions where sulphate-redu-cing bacteria are actively producing hydrogensulphide (HS) (Smithsonian Institution, 1969).

Potash glass used to make much medievalglass in northern Europe has a tendency todevelop dark spots (pits) and an opaquelamellar crust during burial, which is oftendark brown or black in colour in contrast withthe featureless fissured and pale colour of theatmospheric corrosion product found onmedieval glass windows. The darkening effectis due to the oxidation of iron (II) andmanganese (II) ions present in the originalglass, derived from sand and beechwood ashused in its manufacture. In their reduced statesthe Fe(II) and Mn(II) cause the pale greencolour of most glass, but during the leachingprocess, the ions are hydrated and oxidized.Pale pink hydrated Mn(II) ions and pale greenFe(II) ions are converted into dark brownMnOOH and FeOOH. The result is a darkamorphous precipitate of hydrated iron (III)and manganese (III) oxides held in the poresof the leached hydrated silica layer. The oxida-tion process is most rapid in the highlyalkaline conditions at the leached layer/bulkglass interface. In section, under magnification,a brownish-black dendritic invasion of theinterior of the glass may also be observed,

apparently following the lines of cracks intothe glass (Figures 4.24 and 4.25). The stain-ing appears to have diffused into the glass oneither side of the cracks. (X-ray diffractiongives only a very weak pattern for quartz.) Ithas been stated formerly that iron andmanganese diffuses into the glass from thesurrounding soil. This is not the case,however, since iron and manganese oxides areonly soluble at very low pH and arecompletely insoluble at pH 8 or 9, which isthe level present at the glass surface as aconsequence of the leaching of the alkalications.

Freshwater

Ancient and historic glass artefacts may berecovered from freshwater environments suchas the bottoms of rivers and lakes, where theywill have been subject to physical damage,insoluble salt deposition and chemical deteri-oration. Depending on the purity or pollutionof the water, the glass may or may not beaffected by salts (Singley, 1988).

Seawater

A study of the effects of the marine environ-ment on inorganic artefacts has beenpublished by Weier (1973). Ancient andhistoric glass artefacts may be recovered fromshipwrecks, where the presence of the glassranges from cullet used as ballast, cargo and


Figure 4.25 Magnified section through a fragment ofglass, showing deterioration proceeding from the surfaceto the interior, presumably following the lines of cracks.(Courtesy of J.M. Cronyn).

glass artefacts for the use of crew and passen-gers. In 1969, 120 panels of glass opus sectilewere recovered from an ancient warehouse inKenchreai, the ancient port of Corinth inGreece (see Figures 7.7a–e). Apart from a fewknown fragments of opus sectile, these panelsare unique, and continue to be the subject ofmuch investigation (Koob et al., 1996,Moraitou in Kordas, 2002).

The wrecks at Ulu Burun and Kaș off thesouthern coast of Turkey (Bass, 1979, 1980,1984) and that at Sadana Island (Ward, 1998)have been excavated by the American Instituteof Nautical Archaeology, in Turkey and Egyptrespectively. Conservation of the glass fromthese sites has been reported by Pannell(1990). In 1969 a wreck dating from circa 80BC was discovered off the Greek island ofKythera south of the Peleponnese. The shipwas found to have been carrying a cargo ofglass amongst which were polychrome andgold glass vessels. Roman glass has beenrecovered from Baia near Naples (Branda etal., 1999). Cox and Ford (1989) investigatedglass excavated from the wrecks of TheAmsterdam (Hastings, UK) and theDrottengina af Sverige (Lerwick Harbour,Shetland Islands, UK). Over many years, aseries of shipwreck sites has been investigatedand partly excavated by teams of maritimearchaeologists and conservators from the

Western Australian Museum in Freemantle(Pearson, 1975; MacLeod and Davies, 1987;Corvaia et al., 1996; MacLeod and Beng, 1998).The glass recovered from the wrecks hasundergone investigation and analysis.

Physical aspectsThe condition in which glass artefacts arerecovered from a marine environment willhave been determined by the physical andchemical aspects of the seawater and thebottom sediments in which it has lain (Weier,1973) (Figure 4.26). Dramatic physical damagecan occur on objects recovered from a marineenvironment (or salt laden land site) by theaction of soluble salts upon drying. Solublesalts dissolved in water within the object will,upon evaporation, crystallize within or uponthe surface, causing disruption. If the relativehumidity of the air increases, the salts willdissolve, only to reform when the relativehumidity falls. If this action is repeated contin-uously, the object will be severely damaged ordestroyed (Figure 4.26).

Seawater is a physiologically balanced,dilute salt solution containing the knownelements, some dissolved gases and traces ofmany organic compounds. Apart from a fewconstituents produced or consumed by biolog-ical activity, the composition of unpollutedseawater is relatively constant. Seawater with


Figure 4.26 Deterioration of glass in a marine environment.

Artefact/water/atmospheric interface(temporary, as artefacts would berapidly broken or moved by watercurrents, tides and wave action)

Artefacts/water interfaceArtefacts subject to erosion by waterand water-borne particlesEncrustations/depositions/concretions

Artefact/sediment/interstitial waterinterfaceEncrustations/depositions/concretions

Artefacts/sediment interface below 50cmArtefacts may be subject to anaerobicbacterial action

a salinity of 34.325 per ml is consideredstandard. The dissolved oxygen content ofseawater is determined by temperature,pressure, salinity and biological activity. Thesolubility of oxygen increases with decreasingtemperature and decreasing salinity.

The physical aspects of light, temperatureand water movement determine the kind ofmarine life that will inhabit the sea, and inturn will affect the number of organisms, theoxygen concentration and the pH of the water(seawater is only slightly alkaline, pH 7.8 to8.2). Light and temperature will effect diurnalor seasonal cycles. Pollution by domesticsewage or organic wastes greatly increases theconcentration of organic derivatives such asnitrogen compounds and phosphates, and mayresult in a marked deficiency in the oxygenconcentration. Archaeological objects found onthe seabed may have been in an environmentwith little or no oxygen depending on pollu-tion, organic decay and microbiological activ-ity. Sulphur-reducing bacteria have been

known to blacken glass objects, both in thesediments and under concretions.

The sea-bed is predominantly rock, sand, siltor clay. Objects deriving from a ship wreckedin a rocky area may be physically destroyed, orbe washed into cracks and crevices in the rock,where they may lie protected and becomecovered over with sand. Objects lying on sandoverlying bedrock or on a sandy bottom sweptby currents can be rapidly destroyed. Silt, beingof smaller grain size, does not settle wherethere are strong currents and therefore artefactsdeposited in it will probably not be subject towater movement and may be protected by thesilt. Clay binds tightly to any object buried init as a consequence of its colloidal propertiesand small grain size. However it is more likelythat artefacts will be deposited on the claysurface having an overlay of fine silt, and there-fore be subject to physical damage. Hollow orbroken glass objects will become filled withdebris, which may harden on exposure to theair (Figure 4.27).


Figure 4.27 Marine concretion inside a glass flask,which has hardened on drying. Soluble salt crystalshave formed on the surface of the concretion. (Courtesyof the Institute of Nautical Archaeology, Texas).

Figure 4.28 Marine encrustation on a bottle. (Courtesyof A. Alonso Olvera).

Because of their fragility, glass artefacts areunlikely to survive if they are exposed on theseabed. However, if they are initially protectedin a wreck or other shelter, other than byburial, they may become covered by marinefouling organisms, which leave behind theirmud homes, or calcareous shells or skeletons(see Figure 4.28). These take a number offorms:

• Semi-motile fouling organisms: (i) seaanemones and allied forms; (ii) wormswhich build more or less temporary,loosely adherent tubes of mud and sandfor protection. These organisms frequentlyabandon their tubes – often 20 cm (8 in)or more in length – and move to anotherlocation; (iii) certain crustacea, e.g.corophium, build small temporary sandand mud tubes which they cement tomaterial submerged in salt water; (iv)various molluscs, e.g. mussels, attachthemselves to objects by means of a matof very strong chitinous hairs (a complexorganic material), and this remains firmlyattached to the artefact after the musselshave died.

• Sessile organisms: organisms that buildhard calcareous or chitinous shells, andwhich cannot survive without becomingfirmly attached to a suitable base: (i)annelids, which form coiled or twistedtubes; (ii) barnacles, which constructcone-shaped shells built up of laminatedplates; (iii) encrusting bryozoa, colonialanimals which form flat, spreading, multi-cellular, coral like patches; (iv) molluscs ofseveral species, e.g. oysters and mussels;(v) corals, of which there are many forms.Organisms without hard shells: (i) marinealgae: green, brown or red filamentgrowths; (ii) filamentous byozoa – fern- ortree-like growths; (iii) coelenterates(hydroids), such as tubularia, with stalk-like or branching growths, each branchterminating in an expanded tip; (iv)tunicates (sea squids) – soft spongymasses; (v) calcareous and siliceoussponges.

Fouling usually begins with the arrival ofmarine bacteria or other unicellular marineorganisms. They form a thin film on the

artefact, which then provides a favourablefoothold for macro-organisms. These them-selves are commonly in the minute larval orundeveloped juvenile form and secure afoothold in a manner characteristic to theirparticular group. A hard, smooth surface suchas glass provides a firmer footing than a softmaterial. They help to protect the glass byslowing down the diffusion of ions, and there-fore it is possible that glass artefacts may sufferlittle damage despite the fact that it is in aslightly alkaline medium of seawater. Mostconcretions on objects recovered from the seaare porous (those on iron being the mostporous because of the rapidity with whichthey were formed).

Chemical aspectsSeawater is a buffered, very slightly alkalinesolution, with a pH ranging from 7.5 to 8.5.When in equilibrium with the carbon dioxideof the atmosphere, the pH of seawater rangesfrom 8.1 to 8.3. The removal of carbon dioxideby photosynthetic processes of marine plantsincreases the pH; the decomposition oforganic matter removed from the influence ofthe atmosphere decreases the pH. Certaingeneralizations can be made concerning thepH of marine sediments: the pore-size of thesediment can sometimes be an indication ofthe oxidation state, sandy bottoms allowing forthe movement of water containing air (result-ing in a higher pH); while a finer grained siltgenerally promotes reducing anaerobic condi-tions by the prevention of water circulation(resulting in a lower pH). The colour of asediment is also an indication: highly oxidizedsediments are usually brown, whilst thereduced acidic sediments vary from grey togreen to black. The colour is affected by theamount of marine humus and the amount ofhydrogen sulphide produced by bacterialactivity. The greatest bacterial activity might beexpected to occur in coastal areas receivingland drainage. However, as a result ofsedimentation, by which bacteria are carried tothe sea bottom by sediment particles, greaterbacterial concentrations are found within thesediment.

Carbonate precipitationMost of the dissolved carbonate in the sea(derived from atmospheric carbon dioxide) is


in the form of hydrogen carbonate (HCO3).The CO2 content decreases as the pHincreases, there being no free CO2 in solutionin seawater more alkaline than 7.5, most of itoccurring in combination as bicarbonates andcarbonates. If the temperature, salinity or pHof the seawater increase, carbonates will beprecipitated, usually as calcium, magnesiumand/or strontium carbonate. Carbonate concre-tions covering inorganic materials recoveredfrom the sea can be formed in four ways, butonly the first two are relevant to glass: (i)physically (by the growth of organisms, seeabove); (ii) biochemically (by various micro-biological activities which ultimately lower thepH); (iii) physio-chemically (by the solutionand re-precipitation of carbonates, especiallystone); (iv) electrochemically (by precipitationon the cathodic and anodic areas, especiallymetals). The sulphate reducing bacteria playan important role in carbonate precipitationbecause they release H2S into the environ-ment, which rapidly lowers the pH.Concretions formed in the sea are oftenheavily stained by iron salts, and it has beenestablished that the hydrogen sulphidereleased by the sulphate-reducing bacteriareacts with the ferrous irons in the water orsediment to form iron sulphide which precip-itates with the calcium carbonate.

The atmosphere

Chemical effects of air pollutionExamination of European medieval windowglass has provided evidence of glass decayoccurring as the result of atmospheric environ-ments (Newton, 1982b; Römich, 1999a).Glasses held in collections for many decadesmay also have been affected by atmosphericpollution, humidity, incorrect storage condi-tions (Ryan et al., 1993, 1996) or previousattempts at conservation and restoration.

It had long been supposed that air pollu-tion, and in particular sulphur dioxide (SO2),had been responsible for the deterioration ofmedieval window glass (Fitz et al., 1984).Nevertheless, there seems to be no directevidence that sulphur dioxide actually attacksthe glass, although it accounts for the presenceof the sulphates always found on the surface.Similarly, carbon dioxide (CO2) does not attackglass directly, but combines with water, and

converts the hydroxides produced by theattack of water on glass, to carbonates; itaccounts for the calcite frequently found aspart of the weathering crust. The damagingeffect of so-called acid rain is the result ofatmospheric CO2 and SO2 being dissolved byrainwater, thus forming weak acids whichattack the glass. (Carbonic acid may also helpin the formation of alkali carbonates, which,being hygroscopic, will attract water to theglass.)

Three stages are involved: first, an attack ofthe glass by water to produce hydroxides; thenthe subsequent conversion of hydroxides tocarbonates by the carbon dioxide in theatmosphere; and finally conversion of carbon-ates to sulphates (Newton, 1979, 1987).Sulphites may play a role in the chain ofevents. It is also likely that other industrialpollutants such as hydrogen fluoride mayaccount for damage to medieval window glass.The durability of medieval glass windows hasbeen reviewed by Newton (1982b).

The weathering crusts found on medievalwindow glass in situ are a form of enamel-like weathering. The crust may be more than1 mm thick, and the window rendered quiteopaque, thus no longer fulfilling its originalfunction of allowing light into a building (seeFigure 4.13). The crust may be very soft andpowdery, or hard and flinty. It may appearwhite, brownish or even blackish, and a greatdeal of gypsum (calcium sulphate) is usuallypresent. Scott (1932) seems to have been thefirst to demonstrate that the weatheringproducts (on fourteenth-century paintedwindow glass from Wells Cathedral, Somerset,UK), consist largely of sulphates. This isaccounted for by SO2 in the air, arising fromnatural sources (Newton, 1982b). The mostinformative studies of weathering crusts onmedieval window glass have been carried outby Collongues (1974, 1977; Collongues andPerez-y-Jorba, 1973; Collongues et al., 1976;Perez-y-Jorba et al., 1975, 1978, 1980) and Coxet al. (1979).

Visible symptoms of deterioration

Formation of surface pits and crustsPits and crusts form on the surface of glassduring burial, immersion or in the atmosphere,but have been studied mainly in the context


of medieval window glass. The deteriorationof glass resulting from the formation of surfacepits and crusts was originally regarded asbeing essentially two different phenomena.Bimson and Werner (1971) suggested that pitsonly formed in potash glasses, and that crustsonly formed on soda glasses, but in fact thisis not the case. It seems likely that the forma-tion of surface crusts is actually preceded bythe formation of pits in the glass surface. Coxet al. (1979) demonstrated that undeterioratedglass from York Minster (UK) generally had aSiO2 content greater than 60 mol per cent; thatcrusted glasses usually contained less than 60mol per cent SiO2; and that pitted glassesgenerally lay in the range 57 to 63 mol percent SiO2.

It has been noted on fragments of medievalwindow glass from more than one source,that, at least in the early stages, (i) pits havea distinctive cylindrical shape; (ii) they seemto be uniform in size on any given piece ofglass; (iii) the size of pits varies from onepiece of glass to another; and (iv) pits form insome glasses but not in others. The reasonsfor these phenomena have not been fullyexplained. However, it is thought that the size,depth and distribution of pits are related to thesize and distribution of the water droplets onthe glass surface (see solution marks andcondensation); and that defects in the glasssurface may also play a part in promoting theirformation, by holding water in specific sites bysurface tension. The formation of crusts onglass, or in the pits which contain whitedeposits, may sometimes have a protectiveaction, by preventing aggressive agents fromreaching the glass; or conversely, by retainingwater and alkali in close contact with theglass, may have an accelerating action on itsdeterioration. Fissures in deteriorated glasswould enable water to enter pits where itwould be held by surface tension and accel-erate the attack. This may also account for thefact that pits develop downwards, i.e. into theglass, and for their distinctive cylindrical shape(Cox and Khooli, 1992).

HumidityAtmospheric moisture in the form of humidity(water vapour) and condensation can causethe deterioration of glass objects, Egyptianfaience and enamels held in collections, glass

stored in damp conditions, and to medievalglass in situ. It is a particular problem incountries having a climate of high tempera-tures and humidity (Walters and Adams, 1975).

In the case of opaque glass objects havingwhat appears to be a stable surface, any hetero-geneity in the interior, which may be affectedby humidity, will be concealed as long as theouter shell remains intact. Such a conditionexisted in the case of a large Egyptian scarabof the Eighteenth Dynasty, which had been onexhibition at the British Museum (London) formany decades. There had been no reason toregard it as being abnormal in any respect,when it was reported as having suddenly devel-oped cracks. On examination it was found tobe so hygroscopic that it would not remain dryfor 2 minutes at a time. The glass scarab fell topieces; from an examination of its interior, itbecame obvious that the condition was differ-ent from that of the exterior.

A white salt efflorescence formed on partsof the fractured surface but no such crystal-lization occurred on the exterior. It had onlyrequired a micro-crack to form in the surfaceof the object for moisture to have access tothe hygroscopic material inside. The mobilizedsalts then crystallized and caused the disinte-gration of the scarab. A similar example isshown in Figure 7.54.

Chemically unstable glass, which isdescribed as sick, weeping, sweating orcrizzling, repeatedly generates droplets ofmoisture on its surface, if exposed to condi-tions of high humidity. This is the result ofalkali being leached out of the glass by theatmospheric moisture (see discussion ofweeping and crizzled glass in the earliersection on chemical deterioration, and Figures4.15–4.17).

During the Second World War (1939–45), agreat deal of European medieval window glasswas placed in protective storage. Unfortunatelythis was often in damp cellars. Exposure to100 per cent RH, and probably also surfacemoisture, for 5 years or more seems to havedeveloped a hydrated surface layer on theglass. This was subsequently prone to rapiddeterioration, probably due to the opening upof fissures in the hydrated layer as it dried outwhen the windows were reinstalled. (Thus theglass could have seemed to be in good condi-tion when it was removed from storage, and


to have decayed more rapidly during thefollowing time period.)

CondensationStudies of the effect of condensation havesuggested both that glass decays more rapidlywhen exposed to cycles of condensation ordrying than it does when exposed to a contin-uous level of high humidity (Simpson, 1953),and that the reverse is also true (Newton andBettembourg, 1976). This apparent contradic-tion may be explained by the extent to whichcondensed water has run off the glass surface.If the condensation was so heavy that thedroplets merged into a film which was thickenough to run off the glass, leached alkaliwould also be removed. If, however, thecondensation occurred in minute discretedroplets, which then dried out, leached alkaliwould remain on the surface, to form nucleiaround which droplets formed during the nextphase of condensation would accumulate(Adlerborn, 1971). Solutions of high pH wouldthen form at each site, furthering the attack onthe glass. It seems also that deterioration ismore rapid at a liquid/air/glass boundary(Weyl and Marboe, 1967). It seems reasonableto conclude that the formation of pits in aglass surface may be the result of water andleached alkali accumulating in droplets.

The effect of enamel colours, yellowstaining and leadingEnamel paint, and/or the yellow stain, fired onto stained glass can either play a protectiverole, or promote localized deterioration of theunderlying glass (Newton, 1976b). Their effectshave been investigated in connection withmedieval window glass, but are also known onvessel glass (Figure 4.29). The paint is gener-ally more durable than the glass so that, afterweathering has occurred, the painted line-workmay be raised above the rest of the surface ofthe glass. Less frequently it can have thereverse effect, and marked deterioration hasoccurred where the glass had been painted. Inthe case of the yellow stain, Knowles (1959)stated, incorrectly, that ‘The yellow stain whichis applied always to the back of the glass (i.e.facing the outside of the building) invariablyprotects it from corrosion.’

Various hypotheses have been forwarded inan attempt to explain the phenomena; for

example, the paint could encourage corrosionif it had too high an alkali content, and itmight protect the glass if it lost much of itsalkali content to the atmosphere during firing.Similarly, the protective effect of silver staincould readily be explained on the grounds thatthe sodium ions in the surface have beenreplaced by silver ions, and leaching of thesilver would not make the solution alkaline inthe way that sodium ions would. Also, thesilver compounds were applied in a mediumor slurry of various binders. However, none ofthese plausible hypotheses has yet been putto the test.

It was formerly believed that the enamel onthe interior surface of medieval window glassseemed to extract material from the glass, andrender it more liable to deterioration on theexterior, a phenomenon termed back-matchingcorrosion (Knowles, 1959). It has now beenshown (Newton, 1976b) that the corrosion onthe outside is restricted to areas where theoriginal artist, in a wish to strengthen thedesign on the inside, had applied matting(smeared or stippled paint) on the outside tomatch (or supplement) the paintwork on theinside. The matting, being porous, retainedwater, thus encouraging deterioration to occur.

Another phenomenon associated with thedeterioration of glass in connection withpainted decoration is the existence of ghost


Figure 4.29 Detail of a medieval glass beaker,showing the effect of the loss of enamelled decoration.

images of the design. These have been shownto be caused by the volatilization of alkalifrom the paint on another piece of glass fromthe same panel when both were fired, one ontop of the other, in the same kiln (Newton,1976b). The volatilized alkali had left the glassmore prone to deterioration, and over eightcenturies it had lost 0.2 mm in thickness.

The leading on window glass has also beenreported as having an effect on the durabilityof the glass. In some cases the corrosionappears to be enhanced along the sides of theleading, perhaps by virtue of moisture contain-ing dissolved atmospheric pollutants beingretained under the lip of the leads if thecement had fallen out. Since lead is a betterconductor of heat than glass, there is greatertendency for dew to form along the leads.What are more difficult to understand are themany cases where the middle part of a pieceof glass has corroded but the corrosion isabsent in a band within about 20 mm of theleading. If the leads have been cemented well,the original thickness of weathered glass canbe estimated by measuring the part which hadbeen protected; the original thickness couldnot have been greater than the width of the‘heart’ of the leading.

Physical effects

SolarizationIn 1825 Faraday produced a scientific reporton the effect on glass of solarization. Glass,which had originally been colourless, has beenknown to develop a marked purple tint whenexposed to sunlight for a long time. Thisphenomenon, known as solarization, was firstshown to have been produced by the interac-tion of Fe2O3 and MnO, leading to the forma-tion of Mn2O3 and Fe2O3 (Pelouze, 1867)(compare Equation 1.3 in Chapter 1). Theeffect is particularly noticeable in examples ofnineteenth-century house windows, whichhave assumed a marked purple hue in thecourse of time. This effect can be seen, forexample, in the windows of old houses alongBeacon Street in Boston, Massachusetts, USA(Brill, 1963). Examples can also be seen inAntwerp in Belgium, and most restorers ofnineteenth-century window glass will haveencountered it.

Micro-organismsOrganic growths may be found on glass storedin damp conditions, and on neglectedmedieval church windows, and may be associ-ated with glass deterioration. Obviously themost difficult problems associated with organicgrowths on glass are found in hot, humidcountries. Micro-organisms such as mosses,lichens, liverworts and algae, do not attackclean glass. While they do not require anynutrients since they obtain food by photosyn-thesis, they require dirt, grease or pitting as asubstrate to provide a foothold on the glass.Lichenous growths probably do not attackglass directly but promote corrosion bytrapping moisture next to the glass and thushelp to accelerate decomposition. The mostcommon types of lichen found on churchwindows are Diploica lanescens Ach.,Pertusaria leucosora Nyl. and Lepraria flavaAch., all frequently being found on smoothglass. Most have a crust-like growth and arebest adapted to growth on an exposedsubstratum because the whole of their under-surface is attached to the support. Most lichensare found on the outer side of the glass, partic-ularly on north-, west- and east-facingwindows, where there is good air circulation,a certain intensity of light and some humidity;but wind and hot sun-rays are inhibiting. Itseems that water is retained between the glassand the lichen by capillary action. The watercontains CO2 from respiration of the lichen. Asthe hyphae become more or less turgid due togrowth and water availability, so the glass issubjected to pressure and chemical attack.Fungi may also be found on glass. Theyrequire an organic source of nutrient, whichseems to be available from CO2 in the atmos-phere (Tennent, 1981).

It has been suggested that ‘silicophage’bacteria may attack buried glass, but no detailsappear to have been published (Winter, 1965,Newton, 1982b). Examination of a batch ofblackened water-logged glass from the marineunderwater archaeological site HMS Sapphire(1696) at Bay of Bulls, Newfoundland, byFlorian (1979), showed that the blackeningwas due to the deposition of ferrous sulphide,and that much of the deterioration of suchglass is associated with microbiological organ-isms – bacteria being present on all surfacesexamined (the samples had been stored in


fresh water for nearly a year prior to exami-nation).

Previous treatmentsThe use of chemicals to ‘clean’ glass or toremove encrustations, causes irreversibledamage. Chemicals may be difficult to removeentirely, even by washing, and should only beused when absolutely necessary, for exampleuse of acid to reduce thick marine concretionto a thickness that can be removed mechani-cally. (If acid reaches exposed areas of glassit will damage it.) The results of treatment withchemicals, adhesives, consolidants, mouldingand gap-filling materials should be known,and as far as possible, be reversible withoutdamage to the artefact under treatment. Theuse of commercial cleaning agents ofunknown or uncertain composition (andwhich may contain chelating or complexingagents), the over-use of acids, alkalis etc. ortheir non-removal after treatment doesirreversible damage to glass.

Ancient glass artefacts are unlikely to beplaced in a dishwasher, but incredibly casesare known of historic glass collections being

cleaned in this way. Chemicals in the deter-gent or finishing powders can do irreversibledamage, particularly to lead glasses, on whicha white bloom forms (Figure 4.30).

EDTA (ethylenediaminetetra-acetic acid) hasbeen recommended for removing weatheringcrusts and other stubborn surface encrustation,particularly in relation to the cleaning ofmedieval window glass (Bettembourg, 1973;Bettembourg and Burck, 1974). However,since EDTA also attacks glass (Paul andYoussefi, 1978), cleaning has to be verycarefully controlled, and the EDTA washed offas soon as glass becomes exposed.

EDTA favours rapid lead extraction inpotash lead silicate glasses, and as a conse-quence increases the extraction of potash. Thisis not due to the formation of complex K+

salts, as potassium does not form a stablecomplex, but to the removal of the blockingeffect of the lead from the hydrated silicasurface. As the concentration of EDTAincreases, the K2O/PbO extracted ratio de-creases. Barham (1999) used selected chelat-ing agents to remove obscuring deposits, thusrevealing the painted decoration on medievalwindow glass.

Some complexing agents can decrease therate of corrosion, as is the case with ethanol,which is adsorbed onto the surface of thehydrated silica and forms insoluble ethylsilicate, which acts as a protective layer,retarding leaching.

Glass science

Investigations related to the chemical durabil-ity of commercial and industrially producedglasses were included in the discussion of thedeterioration of ancient and historic glasses byNewton in Newton and Davison (1989).However, the distinction between the resultsobtained and the natural deteriorationproducts found on ancient and historic glasswas not made. Most of the studies of chemi-cal durability have been carried out on glassesof simple (binary, i.e. silica and alkali) compo-sition, whereas ancient glasses are complex,by virtue of impurities in the raw materialsused, and the fact that their manufacture wasbased on empirical knowledge.

As previously mentioned, the parametersgoverning the deterioration of glass are inter-


Figure 4.30 Permanent bloom (cloudiness) on awineglass, the result of being cleaned in a householddishwashing machine.

dependent and complex. Experiments havebeen helpful in understanding the fundamen-tal principles of glass deterioration, but cannotbe used as an indicator to the actual deterio-ration products which will have occurred oncomplex archaeological glasses that will havebeen subjected to uncontrolled, naturalenvironments over long periods of time. Theresearch may prove helpful in predicting theresponse of archaeological, and more likely,historical glasses to various storage and displayconditions. This would be of particular signif-icance in countries having a moderate to highrainfall and/or a high variable humidity.

Understanding the altered surfaces ofdeteriorating glassAs mentioned above, the surface of a glass ishighly influential on the effects of the deteri-oration processes that occur on it. However,there are a limited number of possible permu-tations of fresh glass-leached layer-precipitatedphase formation. As the results of experiment,Hench (1977, 1982) and Hench and Clark(1978) described these in terms of six generictypes of glass surface formed as a result ofexposure to an aqueous solution. These areshown in Figure 4.31. The six types of glasssurface do not of course describe the manydifferent compound surfaces which form onthe surfaces of archaeological glass. Only threeof them (Types II, IV and V) would seem tobe relevant to the deterioration ofancient/historical glass.

The Type I surface has an extremely thin(less than 5 nm) hydrated layer and no signif-icant change in surface composition, either byloss of alkali or of the silica network. It repre-sents an extremely durable glass (such as vitre-ous silica) exposed to a solution with a neutralpH. It is therefore unlikely that there are anyancient glasses with a Type I surface.

The Type II surface possesses a silica-richprotective film, where the alkali has been lostbut the silica network has not been damaged.The glass is very durable; it has a low alkalicontent and the pH of the leaching solution(for one reason or another) does not riseabove 9.0. However, if flat glass having a TypeII surface is stacked closely together in moistconditions, a large surface area (SA) is exposedto a small volume of water (V) trappedbetween the layers of glass, where it gradually


Figure 4.31 Six types of glass surface defined by Hench(1982), interpreted in archaeological/conservation terms, bythe author. (a) Type I, an extremely thin hydrated surface(5 nm) with no significant change to surface composition;an extremely durable glass surface, which would not occurin antiquity. (b) Type II, a silica-rich protective film withloss of alkali, but no further changes to the silica network.Ancient and historic glasses may develop this type ofsurface. (c) Type IIIA, a double protective film consistingof aluminium silicate or of calcium phosphate, deliberatelyformed on modern glasses in laboratory experiments. (d)Type IIIB, multiple protective layers deliberately formed onglasses used for the storage of radioactive waste. (e) TypeIV, a silica-rich film, which however is not thick enoughto prevent loss of alkali or destruction of the silicanetwork. Ancient and historic glasses may develop thistype of surface. (f) Type V, a glass surface which is slowlysoluble in water, and may therefore develop solutionmarks and corrosion pits. Ancient and historic glasses maydevelop this type of surface.

–[SiO2]→

Dis

tan

ceD

ista

nce

Dis

tan

ceD

ista

nce

Dis

tan

ceD

ista

nce

Bu

lk

Bu

lk

Bu

lkB

ulk

Bu

lk

Solution

Solution

Solution

Solution

Solution

SolutionTotal dissolution

Protectivefilmonglass

Dualprotectivefilmsonglass

Nowprotectivefilm onglass

Selectiveleaching,film breakdowntotal dissolution

Glass surface

Inert glass

Type I

Type II

Type IIIA

Type IIIB

Type IV

Type V

Inertglass

Originalglass/solutioninterface

Selectiveleaching

Water WaterAlkali Surface

The silica layer thickens

Silica rich layer

Complex glasses (for storageof nuclear waste)

Glass surface

Glass surface

Water

Water Pts

Solubleglass

Alkali

Alkali

Slowly soluble glass

Solution marks

Dualprotectivelayers

Multipleprotectivelayers

Silica rich layer

–[SiO2]→

–[SiO2]→

SiO2

SiO2

Al2O3–SiO2CaO–P2O5CaO–SiO2

CaCO3, SrCO3

Zn(OH)2

Na3+

Fe(OH)3

Al2O3–SiO2

CaO–P2O5

CaO–SiO2

–[Oxides]→

–[SiO2]→

–[SiO2]→

becomes highly alkaline. As a result of this, thesurface loses its durability. This effect can beobserved, for instance, on poorly stored lanternslides, or panels of window glass.

Glasses with Type II or Type IIIA surfacesappear quite glossy and unmarked, andundeteriorated, with any original surfaceblemishes easily visible. Nevertheless, thesurfaces would be largely leached of alkali.

The Type III layer is a double protectivesurface film of aluminium silicate or calciumphosphate, and can exist in two versions, IIIAand IIIB. The addition of aluminium oxide orphosphorus oxide to the glass results in theformation of an aluminium silicate – or acalcium phosphate-rich glass on top of a silica-rich layer. These protective films can beproduced by adding the oxides mentionedabove to the glass batch, by alkali extractionfrom a glass containing the oxides, or byprecipitating the alumina or phosphate fromsolution. The addition of sufficient Al3+ ions tothe water, for example, as aluminium chloride(AlCl3) at a concentration of more than 25 ppmof Al3+, can greatly reduce the leaching ofalkali and destruction of the glass network inboth acid and alkaline conditions, due to theformation of a Type IIIA surface. At first (forexample, in the first few hours, depending onthe Al2O3 content of the glass), the attack onsuch a glass can be relatively rapid, a Type Vsurface being formed, but within a day or twothe situation becomes reversed due to theprecipitation of an alumina-silicate complex onthe surface (Hench and Clark, 1978).

Archaeological glass may contain traces ofalumina and phosphorus, but it is unlikely thatthe amounts would be sufficient to have a sig-nificant effect on its deterioration/preservation,especially since there are more dominantparameters at play. Many medieval windowglasses contain substantial amounts of bothAl2O3 and P2O and the formation of Type IIIAsurfaces must be considered as part of thegeneral situation regarding their complexweathering behaviour. (The Type IIIB surface,added by Hench (1982), is characterized bythe formation of multiple protective layerscomposed of oxides, hydroxides and hydratedsilicates. It is found on complex glasses for thestorage of nuclear waste.)

The Type IV surface is also a silica-rich film,but the silica concentration is insufficient to

protect the glass from migration of alkali ordestruction of the network. Many medievalwindow glasses may be considered to fall intothis category because of their low silicacontent.

The Type V surface is slowly soluble in theleaching solution. There is extensive loss ofalkali accompanied by loss of silica, butbecause the attack on the glass is uniform, thesurface composition remains the same as thatof the bulk glass, and thus it superficiallyresembles that of the durable surface Type I.However, as the glass is attacked, solutionmarks develop on the surface. These areparticularly noticeable on transparent samplesviewed in oblique illumination (Newton,1982a). The Type V surface also has a markedability to form corrosion pits, because the localconcentration of alkali in surface defects(solution spots) can exceed pH 9.0, the criti-cal point at which attack on the glass becomesprogressive. This mechanism, and the mannerin which it was investigated, is described indetail by Sanders and Hench (1973). Type Vsurfaces are extremely susceptible to scratch-ing and abrasion when the scratches aredeeper than 0.2 μm. The scratch produces asmall trough in which the extracted alkali cancollect and progressive degradation of thenetwork then occurs at that point. Scanningelectron micrographs in Sanders and Hench(1973) show very fine scratches produced by600-grit silicon carbide, which have developedinto relatively deep (perhaps 0.4 μm) troughsby exposure to a small amount of water for216 hours (this glass had a very low durabil-ity). When 120-grit silicon carbide was used asthe abrasive, the resultant troughs are perhaps1 μm deep.

Experimental research into chemicaldurability of glass

Accelerated ageing testsIn order to obtain an idea of the deteriorationprocesses associated with glass, within a shortperiod of time, accelerated ageing tests areundertaken in laboratory conditions. Theseinclude increasing the surface area of the glassto the volume of the attacking solution, raisingthe temperature, cyclical increase and decreaseof the relative humidity and exposure of glassto acid or alkaline environments. As a result


of such experiments, there is a large body ofliterature concerned with the durability ofglass (e.g. Hench, 1975b; Paul, 1977). Whilstthe results of accelerated ageing tests can givean indication of the natural weatheringprocesses, which might be expected to occurover time, it cannot be assumed that the twoprocesses correlate exactly.

Increasing the surface area of glass to theattacking solutionMany accelerated ageing tests involve increas-ing the ratio of glass surface area to the attack-ing solution volume (SA/V), often by using theglass in crushed form. This procedure mayyield valid information on the properties of thebulk of the glass, but these are likely to bevery different from the properties of theancient glass surface, where other parameterscome into play (Ethridge et al., 1979).

Raising the temperatureRaising the temperature of its environment,increases the rate of attack on glass by aggres-sive agents.

ThermodynamicsAttempts have been made to predict thedeterioration of glasses from their compositionby thermodynamic approaches, first proposedby Paul (1977) and developed by Jantzen andPlodinec (1984). Thermodynamics relate theloss from glass as a result of deterioration witha calculated free energy of hydration for theglass composition concerned.

Paul (1977) states that, for most silicateglasses, the quantity leached in a given timeis doubled for every 8–15°C rise in tempera-ture, depending upon the composition of theglass and the type of alkali ion in question.Bacon (1968) gives the following formula:

log t1 = log t2 + (T2 – T1)/23.4 (4.3)

where t1 and t2 are the times for leaching attemperature T1 and T2 in °C. This producesa doubling in the rate of leaching for a rise of7°C. There is the possibility that reactions willoccur at higher temperatures, which hardlyoccur at all at ambient temperatures (Henchand Clark, 1978). In addition, the thickness ofthe protective high-silica surface layer formedon heating is far less than that formed atambient temperatures.

In terms of thermodynamics, there is a goodcorrelation where compositional differencesbetween glasses is high, for example betweenvolcanic obsidian glass and soda–lime–silicaglass. However, in terms of differentiatingbetween compositionally similar archaeologi-cal glasses, the method is less useful. As anexample, Cox and Ford (1993) calculate similarfree energies of hydration for Roman glassesexcavated from Wroxeter, (Shropshire, UK),which show apparent deterioration rates(derived from the thickness of the hydratedlayers) differing by factors approaching anorder of magnitude. Freestone (2001)expressed the opinion that a difference incorrelation rate of this order is crucial, in thatit can determine whether a thin glass vessel,of 1–2 mm wall thickness, has a thin layer ofdeterioration or is totally destroyed. Thusminor differences in glass composition andlocal environmental factors are likely to be ofgreater significance than free energies ofhydration, calculated using present models, indetermining the relative rates of decay ofarchaeological glasses (Freestone, 2001).

Cyclical humidityA test that may yield results similar to thoseobtained, particularly on historic glass and onmedieval window glass in situ, is the cyclicincrease and decrease of humidity (Simpson,1953; Bacon and Calcamuggio, 1967; Römichin Kordas, 2002).

The effect of acids on glassExperimentally, El-Shamy et al. (1972) studiedthe effect of acids on glass, down to pH 1.0,and showed that the pH was not significantunless there was some 20 mol per cent of CaOin the glass. In that case, the extraction of NaOand CaO increased greatly as the pH wasreduced below 4. Such high-lime glasses arereadily soluble in mineral acids (El-Shamy etal., 1975). El-Shamy (1973) conducted experi-ments involving glasses made to medievalcompositions, and showed that the extractionof CaO by 0.5 mol HCl (hydrochloric acid)was greatly increased when the CaO or MgOcontent of the glass was as much as 15 molper cent. When the CaO and the MgO wereboth present at 15 mol per cent, the extrac-tion of both was more than doubled. Neitherof these papers discusses what happens to


medieval-type glasses exposed to weak acids,such as carbonic acid (a solution of CO2 inwater) or sulphurous acid (a solution of SO2

in water), when the pH value is between 6and 7, corresponding to rainwater or evenpolluted rainwater. Ferrazzini (1977) applied 1mol HCl on a medieval glass sample, andproduced an ion-exchanged layer 1 mm thickin one hour.

Burial experimentsShort-term burial experiments have been carriedout by Römich (Römich et al., 1998; Römich,1999a,b; Römich in Kordas, 2002) on a numberof potassium-rich model glasses. These wereburied in garden soil saturated with water atvarying ranges of pH 7.8, 3.5 and 8.9, at roomtemperature and for periods of between sixweeks and seven months. After nine months,the following results were obtained: (i) at pH3.5, a single deterioration layer, cracks and localdeterioration had occurred; (ii) at pH 7.8approximately 50 per cent of the glass haddeteriorated; (iii) at pH 7.8, a layered structurehad formed over the entire glass surface anddeep cracks had formed in the glass. Thesamples were analysed by light microscopy,scanning electron microscopy on cross sectionsand infrared spectroscopy.

In 1970, a long-term glass burial experimentwas begun at Ballidon (Derbyshire, UK), withthe intention of recovering samples at differ-

ent times until the year 2482. Nine types ofglass were buried in a mound, at ground leveland mid-mound level. In 1986, 28 sampleswere removed for examination and 46commercial samples of glass were buried. Notall the original samples were found. Of thoseexhumed, six of the nine types showed nosignificantly visible signs of deterioration andthus would seem to have been of a toodurable nature to reveal the intended infor-mation. The six types were: no. 1, a simulatedRoman glass; no. 4, polished plate glass; no.5, ‘as produced’ plate glass; no. 7 Pyrexovenware; no. 8 high-quality soda–lime–silicaglass; and no. 9, a high lead optical glass. Theworst signs of deterioration were slight irides-cence on sample no. 8 and the emphasis ona few old scratches on samples nos. 5, 7 and8. The three types of glass which did showappreciable signs of deterioration were no. 6E,glass marbles, which had developed wingfractures at impact points where the marbleshad rolled against one another after manufac-ture. No. 2, a simulated medieval glass,displayed a surface iridescence which tendedto run in lines, perhaps following otherwiseimperceptible differences in homogeneity. Theonly type which showed obvious deteriorationwas no. 13, a simulated Saxon type of glass,which had developed an extensively crazedsurface layer about 260 μm thick which couldeasily be removed (Newton, 1992).


The materials used in the processes of conser-vation and restoration fall into severalcategories: solvents and reagents, adhesivesand consolidants, surface coatings, modelling,moulding and casting materials, pigments anddyes. In addition, there are those used toconstruct supports to facilitate the safe displayof fragile pieces, and those used for safestorage and packaging.

It is important to bear in mind that allchemicals are potentially dangerous, and careis required wherever they are stored, used ordiscarded. Manufacturers’ data sheets shouldbe obtained, and safety instructions containedwithin them heeded. Whenever appropriate,safety goggles, fume or dust extraction masksand protective gloves and clothing should beworn, and fume extraction facilities used.Solvent hazards fall into three categories:flammability, toxicity and corrosiveness. Allchemicals should be clearly marked with theappropriate hazard labels. Detailed informa-tion on materials used for conservation isgiven by Horie (1987).

Cleaning agents

The materials chosen to clean a particularpiece of glass will depend on the condition ofthe glass and any applied surface decoration,and upon the substances to be removed.Extraneous matter can range from mud,calcareous or iron deposits or other accretionfound on glass recovered from land, water ormarine sites, to previous repair or restorationmaterials. In addition, products formed duringthe decomposition of glass may have to beremoved. A wide range of materials may there-fore be required for cleaning glass: water,

detergents, chelating (sequestering) agents,acids, organic solvents and biocides. It mustbe remembered that these materials, especiallythose which are formulated commercially, mayhave deleterious effects on glass if used injudi-ciously. Their uses are described in Chapter 7,but their important properties in relation touse on glass, and dangers to the conservator,will be discussed in this chapter. Mechanicalmethods of cleaning and the use of cleaningagents both have advantages and disadvan-tages, which should be evaluated before workbegins.

Water

Despite the fact that water is the primary agentfor causing glass to deteriorate, it is also themost commonly used solvent for cleaningglass. The reason for this apparent contradic-tion is that if glass and any decorative surfaceare sound, short exposure to water does notharm it. Water will, however, remove poorlyadhering paint or gilding and loose glassflakes from the surface of deteriorated glass(during burial or conservation cleaning).Prolonged exposure, e.g. during burial orsubmersion, will cause dissolution of alkalifrom the glass. Water may be used in severalforms, e.g. sea water (gradually replaced withtap water), tap-water, or de-ionized or distilledwater which have been treated to removedissolved salts. Tap water usually containscalcium and magnesium hydrogen carbonates,chlorides and sulphates. The amounts varywith the district; in hard-water areas theamount of salts can be more than 170 mg perlitre. The hydrogen carbonates which causetemporary hardness are stable in solution onlywhen the water is acidified with dissolved

199

5

Materials used for glass conservation

carbon dioxide. These salts can therefore beprecipitated from ground or atmospheric waterby boiling. The sulphates and chlorides cannotbe removed by boiling and cause permanenthardness.

When droplets of water evaporate on a glasssurface, tiny deposits of salts may be visible,e.g. calcium, sulphates and magnesium (alsopotassium or sodium hydroxides (or carbon-ates) in the case of unstable glass). Suchdeposits are commonly erroneously referred toas water stains. Very visible deposits ofcalcium carbonate can sometimes be formedin vessels such as flower vases that haveconstantly contained water. In order to preventsuch deposition therefore, purified water(which has had the salts removed) should beused when treating glass objects. Impurities inwater can be removed by two methods: distil-lation (to produce distilled water); and treat-ment by ion-exchange resins (producingde-ionized water). Washing of glass with watershould be kept to a minimum; glass shouldnot be left to soak, if possible (obviously thiscannot apply to glass from wet environments,stored in water).

Detergents

The term detergent is applied both to thechemical compounds (surfactants) and thecommercial products (formulae) that act in asolvent to aid the removal of soiling andcontaminants. Surface active materials for usein water can be divided into three groups:anionic, cationic and non-ionic. The termsrefer to the nature of the polar, hydrophilic,group in the molecular structure of the surfac-tant. Surface-active compounds have a two-fold nature, since each molecule has a polarand a non-polar end. One end is compatiblewith the solvent used and the other will inter-act with the soiling substances. The mostcommonly used non-ionic detergents used inconservation were the nonylphenol ethoxy-lates (NPEs), examples being Lissapol,Synperonic and Triton brands. Their use isbeing phased out as a European CommunityDirective (EC PARCOM Directive 92/8),(although still available in Canada and theUSA), because their degradation productsmimic oestrogen, and when discharged intothe environment, affect marine life. A number

of surfactants is being tested to ascertain theirsuitability for use in conservation (Fields,2000). The manufacturer of the Synperonicrange of surfactants has introduced SynperonicA7, an alcohol ethoxylate which is readilybiodegradable, and conforms to EEC directive82/242. Glass is most commonly washed withan inert detergent dissolved in distilled water,although detergents have been developed withthe increased use of non-aqueous (i.e. dry-cleaning) solvents.

A commercial detergent formula is a mixtureof chemicals each of which plays a role in thecleaning operation. A typical formulation forwashing powder contains alkali, surface-activechemicals, chelating agents, suspension andthickening agents. Many of these ingredientscan damage glass or other surfaces. The useof commercial products cannot be recom-mended because their ingredients areunknown in detail and thus the effects of theiringredients on ancient glasses cannot beforeseen.

It is unlikely that the commonly used deter-gents have any serious toxicity if used sensi-bly. However, long immersion of the hands indetergent solutions will extract the oily protec-tive chemicals from the skin and leave it opento invasion by infection; since the use ofbarrier creams may contaminate an object, itis preferable that conservators with a sensitiveskin should wear protective gloves (Davidsohnand Milwidsky, 1978).

Chelating (sequestering) agents

Many metal ions are stable in water solutionsonly when the pH is in the correct range, orwhen appropriate anions are present.Chelating agents can stabilize metal ions overa wide range of solvent conditions. A chelat-ing agent has a strong reaction with the metalion, enclosing it in a protective complex(Richey, 1975), for example, ethylenedi-aminetetra acetic acid (EDTA):

(5.1)


The characteristic of chelating agents is theformation of stable, multiple bonds betweenthe chelating agent and the metal ion. Thematerials that usually have to be removedfrom glass with chelating agents are weather-ing crusts that consist of calcium carbonatesand sulphates in combination with silica. Themetal ions that have to be removed are thesame as those that make up the glass, and itis therefore difficult to apply any solution tothe weathering crust that will not penetrateand react with the glass beneath, or react withthe glass as soon as the crust has beenremoved. For this reason the use of chelatingagents must be assumed to affect the under-lying glass and it is the conservator’s job tominimize the contact of these reagents withthe glass surface (Ferrazzini, 1977). The use ofpastes made up with a thickening agent suchas Sepiolite (magnesium trisilicate) orcarboxymethylcellulose sodium salt mayconfine the action of the solution to thesurface to which it is applied. However,chelating agents will extract the more accessi-ble atoms first and it is therefore possible thatthe weathering crust may be weakened beforeany significant attack on the underlying glasstakes place. The weakened crust may then beremoved by gentle mechanical means (Bauer,1976), but the removal of crusts from glass isa drastic treatment that needs careful consid-eration. Chelating agents can conveniently bedivided into three groups: polyphosphates,aminocarboxylic acids and hydroxycarboxylicacids. Their use has been largely in the fieldof medieval stained glass conservation,however, a few references to this work aregiven here, as there may be a need to considertheir use, as a last resort, to clean other typesof glass.

Polyphosphates

Polyphosphates were introduced as watersofteners in the 1930s to chelate and keep insolution calcium and magnesium ions. Theyare still widely used in commercial detergents;Calgon, for instance, is a polyphosphate wherethe value of n (below) is about 12 (Albrightand Wilson, 1978). They can therefore be usedto dissolve calcium and magnesium salts fromhard crusts.

Bettembourg (1972, 1973) used a solution(solution A) consisting of an aqueous solutionof 10 per cent sodium thiosulphate (NaSO.5HO) and 5 per cent sodium pyrophosphate(NaPO.10HO) to clean medieval windowglass. The reported effects of these polyphos-phates on glass seem contradictory, and it ispossible that much may depend on the natureof the actual piece of medieval glass beingcleaned. Frenzel (1970) recommended the useof Calgon, but Frodl-Kraft (1967) used it onlywith caution, provided the solution processwas halted before the actual surface of theglass was reached. Four years later, however,Frodl-Kraft discontinued the use of Calgonbecause it was found to creep along the glasssurface, under the painted decoration, andloosen it (Frodl-Kraft, 1970, 1971). Ferrazzini(1977, Figures 2 and 3) has published illustra-tions showing how Calgon produced cracksand other types of damage on glass surfaces,which become apparent on drying or aging.

Polyphosphates tend to hydrolyse insolution to the orthophosphate anion, whichforms insoluble salts with alkaline earths, andshould therefore not be used on glass.

Aminocarboxylic acids

The most commonly used chelating agent ofthis class is EDTA:

The tetra acid can be partly or totallyreacted with sodium hydroxide, and various

Materials used for glass conservation 201

(5.2)

(5.3)

EDTA products with different proportions ofsodium are available (Richey, 1975, Table 1).EDTA can be used to hold calcium andmagnesium ions over a wide range of pHvalues between pH 7 and pH 11, and it willchelate lead from a lead-containing glass (Pauland Youssefi, 1978; Olsen et al., 1969). Thereis a large number of other aminocarboxylicacids which are similar to EDTA but providea range of potentially useful properties. Anexample is N, N-bis (2-hydroxyethyl) glycine,HOOC-CH2N(CH2 CH2 OH)2. This does notchelate calcium and magnesium ions but doeshold other metal ions in solution (Davidsohnand Milwidsky, 1978). Solution B used byBettembourg (1972a, 1973) is an aqueoussolution of the sodium salt of EDTA, 30 gl–1

buffered with 30 gl–1 ammonium hydrogencarbonate (NH4.HCO3). At first the solutionwould seem to have been used without anyother caution, soaking entire glass panels (stillin their leading) in baths of EDTA solution for2–3 hours, or until the required degree ofcleaning had been achieved. However, in1975, Bettembourg and Perrot (1976) changedthe procedure to three successive applicationsof the solution on cotton wool swabs. Bauer(1976) claimed that EDTA does not harm thesurface of glass, whereas Ferrazzini (1977) hasillustrated damage caused by EDTA to poorlydurable glass surfaces. No doubt, like theexperience with Calgon, much depends uponthe type of glass which is being treated, butthere is sufficient evidence of the deleteriouseffects of EDTA on glass to show that greatcaution must be exercised in its use.

Hydroxycarboxylic acids

Hydroxycarboxylic acids act in two differentways at different pH values. Below pH 11these acids form only weak complexes withalkaline earths, and a large excess of thechelating agent is necessary to ensure thedissolution of calcium ions. Above pH 11 theyare more effective than EDTA or polyphos-phates as chelating agents for calcium.Hydroxycarboxylic acids chelate iron and othermultivalent ions over the whole pH range.

At pH 7, therefore, gluconic acid might beused to extract copper or iron staining fromglass with relatively little effect on the alkalinemodifiers of the glass. Citric acid is well

known for the damaging effect that it can haveon glass surfaces (Bacon and Raggon, 1959).As previously discussed, chelating agents arenon-specific in their action and will attackglass, especially at high pH values, causingdamage by complete removal of the surface,or by selective leaching of the alkaline earthmodifiers (Paul, 1978). Glass surfaces that havebeen treated with chelating agents are there-fore made more liable to deterioration(Ferrazzini, 1977). The solutions must there-fore be applied locally. As yet there is nochemical treatment for safely removing theiron staining from glass surfaces; improve-ments to chelating agents may, however,produce reagents that are more ion-specific. Areview of chelating agents by Richey (1975)outlines the toxic hazards and main points ofinterest in their use.

Hydrogen peroxide

Aqueous hydrogen peroxide solutions can beused to bleach out organic stains on glasssurfaces (Moncrieff, 1975). Hydrogen peroxidebreaks down in solution to evoke an activeform of oxygen H2O2 → H2O + [O] that canreact with and decolourize organic material. Apreservative, typically parts per million ofphosphates, is added to solutions to slowdown the spontaneous production of oxygen,and alkali can be added to increase the rateof oxidation. Hydrogen peroxide is purchasedas a solution in water. The concentration ofthe solution is normally indicated by thevolume of oxygen gas which one unit volumeof solution will produce. For example, 100 vol.hydrogen peroxide is a 30 per cent solution.


(5.4)

The solution should be diluted to a 10 vol.solution for use on glass.

As oxygen can be evolved in large amounts,which rise as bubbles to the surface, hydro-gen peroxide should not be used or storednear highly inflammable substances such assolvents. Hydrogen peroxide can oxidizeorganic material such as skin or clothing andit should be used with care, especially whenin concentrated solution.

Mineral acids

Before the deleterious effects of applying acidsto glass were fully understood, the mineralacids – hydrochloric, nitric and hydrofluoric –were used to remove surface deposits. All ofthese can attack poorly durable ancientglasses, especially when the lime content ishigh; and should not normally be used onglass artefacts (despite the customary use ofnitric and chromic acids for cleaning labora-tory glassware). Alternative, less invasive treat-ments are now available.

Mineral acids are supplied in the form ofconcentrated solutions: hydrochloric acid(HCl) 30 per cent; nitric acid (HNO3) 75 or 90per cent; sulphuric acid (H2SO4) 98 per cent;hydrofluoric acid (HF) 40 per cent. All ofthese, but especially sulphuric acid, becomevery hot when added to water. For this reasonacid should always be added to water, neverthe other way around. This will reduce thedanger of concentrated acid boiling andspitting when it comes into contact with water.

Mineral acids can cause burns on the skin,and hence protective clothing (gloves, eye-shields and aprons) is needed, and scrupulouscleanliness is important. Dilute solutionssplashed on clothing, or on the bench, willevaporate, become more concentrated and belikely to cause burns hours after the acidbottles have been put away in a safe place.

Hydrofluoric acid is particularly dangerousbecause it differs from other acids in the wayit attacks living tissue. The concentrated acidcan attack the surface of the skin, as do otheracids, but the fluoride ion diffuses throughhealthy skin and fingernails to precipitatecalcium in the tissues beneath, thus causingintense pain which can occur some hours aftercontact with dilute fluoride solutions hasceased. Although still used commercially for

etching glass, hydrofluoric acid should notnormally be used for conservation work.However, should its use ever become neces-sary, it is of the utmost importance that advicebe sought upon the stringent safety precau-tions to be taken and the specific First Aidtreatment required in the event of an accidentoccurring. Antidotes for poisoning with hydro-fluoric acid are calcium gluconate gel in theUnited Kingdom and zephiran chloride (abenzalkonium chloride) in the United States.

Organic solvents

Organic solvents are used for three purposesin conservation: for removing greasy dirt, forapplying or diluting polymers in solution, andfor removing polymers. The term solvent isusually assumed, as here, to mean a mobileorganic liquid, but it can be extended to coversolids, such as those polymers that can dis-solve dyes etc. by absorption. Polymer-solventinteractions are important, and it is thereforenecessary to understand those properties(Horie, 1987).

As previously mentioned, all solvents arepotentially dangerous to those using them, andto the environment, in terms of their toxicityand flammability, and care is requiredwhenever they are used. Fume extraction facil-ities should be used wherever possible. It issometimes possible to exchange a dangeroussolvent for a less hazardous one, yet still retainsuitable solvent properties, and this shouldalways be done.

All organic solvents, except the highlyhalogenated ones, are flammable. Any firemust start in an air/vapour mixture, and hencethe more volatile the solvent, the more readilyit is ignited. A good indication is the closedcup flash point, which is the lowest tempera-ture at which a spark above the liquid willcause the vapour to ignite. The lower the flashpoint, the greater the hazard, and the vapourscan travel some distance from an open vesselcontaining the solvent. Hence naked flames,cigarettes, electric switches and other sourcesof sparks must not be used when solvents areemployed. Even the non-flammable chlori-nated solvents such as trichloroethylene canbe dangerous when exposed to naked flamesor cigarettes because they can be converted tothe poisonous gas phosgene.


All solvents are toxic to some extent.Perhaps the most common and insidiousdamage is caused by breathing solvent vapourover extended periods of time; for this reasonthe effects and relative dangers of solventvapour have been extensively studied (seebelow). However, liquid solvents can beabsorbed through the skin and they dissolvethe protective chemicals from the skin, therebyallowing ease of entry for infection.

There are two different kinds of vapourtoxicity, both of which may occur with thesame solvent:

• narcotic effects, causing drunkenness orpoisoning, which wear off as the solvent iseliminated from the body (e.g. ethanol);and chronic effects, which persist long afterthe solvent was initially absorbed (e.g.methanol). The cancers induced by somesolvents fall into the category of chronicdamage.

The type of toxicity caused by two chemicallysimilar solvents, such as ethanol and methanol,is not necessarily the same. Ethanol in moder-ate doses causes drunkenness and evenunconsciousness, but the effects wear off asthe ethanol is eliminated from the body.Methanol, on the other hand, is metabolizedinto products that cause permanent damage tothe nervous system and may lead to blindness.

The danger of breathing solvent vapours isassessed in each country by its relevantauthority, which should be consulted beforeusing a material. In the United Kingdom therisk is summarized in the OccupationalExposure Limit (OEL), and in the United Statesof America in the Threshold Limit Value (TLV).In both cases, these specify the maximumconcentration of the solvent vapour in the airthat can be tolerated by a worker, withoutsignificant health risk. The measurement of theconcentration of a solvent vapour has beenmade much easier in recent years by thedevelopment of simple and relatively inexpen-sive instruments. For example, the Draegersystem works by drawing a known amount ofair through a tube of chemicals, which indicatethe concentration of vapour in the air bychanging colour.

The OEL (HSE 2000) and TLV (ACGIH 2000)listings are updated annually and are part of

a wider assessment of risk, also published bythese and other similar authorities.

Biocides

Micro-organisms can cause difficulties duringthe conservation of glass. Archaeological glassexcavated from waterlogged conditions isfrequently dirty and hence can support fungaland bacterial growths; stained glass can act asa support for the growth of mosses andlichens.

When archaeological glass from wet sites isstored, biocides should be added to preventspoilage of the excavated material. Alkalinesodium salts should not be used because it islikely that they will cause corrosion duringstorage, and the quaternary ammo-compoundscan form a hydrophobic layer (see above). 2-Hydroxybiphenyl is the only appropriatematerial that seems to have been investigated(for conservation use on wood). A saturatedstock solution in water (0.07 per cent) can beadded to packaged material awaiting stabiliza-tion. The weight of solution should be approx-imately equal to the weight of preservedmaterial, thus producing a 0.035 per centsolution, which is sufficient to kill fungi. Toolow a concentration of 2-hydroxybiphenylshould not be used because it may notimmediately kill the organisms and they maybecome resistant to the biocide. Algae may bemore troublesome to control, but they rely onlight for photosynthesis to occur and hencethe treated packages of glass should be storedin cool dark locations. Any package that hasbeen treated with a biocide should be markedwith the name of the material used, theamount added and the date.

Adhesives, consolidants and lacquers

Before the advent of synthetic polymers, thematerials used for the consolidation, repairand restoration of ancient glasses were animalglues, natural waxes and resins, notablybeeswax and shellac (Davison, 1984; Horie,1987), and plaster of Paris. Their only advan-tage was that they held the fragments together,at least for a limited period. All had seriousdisadvantages compared with modern materi-als: glue shrinks, wax flows and attracts dirt,


natural resins are brittle and plaster of Paris isopaque. These traditional materials are stillused in some countries, and may be found onancient glasses from collections.

Animal glues are composed of polyamidesderived from the protein connective tissue(collagen) derived from bones and skin. Thevarious types of glue (hide, bone or fish) areprepared by boiling the raw material in waterwith hydrolysing catalysts, which break downthe collagen structure. The glues range frompale to dark brown in colour. They areapplied in a warm liquid state and gel oncooling. As the gel dries, it shrinks throughloss of water and this may cause sufficienttension to detach the surface of fragile glassobjects. Animal glues and consolidants can beremoved by the use of warm water appliedeither on cotton wool swabs, or, if the condi-tion of the glass will allow it, by immersion.

Shellac is the alcohol-soluble portion of theexudate of the lac insect Tacchardia lacca. Itis a polyester in which 20–40 per cent of theresin is aleuric acid; other components of theresin include aldehydes, hydroxyls, othercarboxylic acids and unsaturated groups. It isdark brown in colour and is thereforeunsightly. Shellac is initially soluble in ethanoland other solvents, but on ageing, polymerizes(cross-links) into a virtually insoluble mass.This mass is often difficult if not impossible toremove from fragile glass.

In contrast to animal glues and shellac,Canada balsam, the resinous component of theexudate from the pine Abies balsama, has arefractive index of 1.52, which is close to thatof colourless soda glass. It has therefore beenwidely used for the cementing of opticalelements in lenses, but apparently not for therepair of glass antiquities.

Adhesives based on dextrin have been usedboth to repair glass and as gum on labels tomark glass. Dextrins are polysaccharidesproduced by breaking down starches to formmore soluble products; and they suffer thesame form of degradation as cellulose.Moreover, the low molecular weight of dextrinmakes the product hygroscopic and thereforeunsuitable for the conservation of glass.

During and after the Second World Warthere were considerable advances in commer-cial and industrial polymer chemistry, oneresult of which was to make available many

new materials, and further developments havetended to create a particular product for theintended end-use. But it is also necessary toevaluate the materials for permanence of theirproperties. In practice, however, it is usuallynecessary to make a compromise by selectingthe material that has the least number of disad-vantages with regard to the particular conser-vation task in hand. It has often been the casethat, once a polymer has been found that satis-fies enough of the requirements, it tends to bewidely recommended until it is superseded bya new material. However, the formulation maybe changed by the manufacturer withoutnotice and this may affect the result when itis used in conservation.

At present the materials used in conserva-tion, and discussed in this chapter, are: cellu-lose derivatives, epoxy, polyester and acrylicresins (including cyanoacrylates and photoac-tivated or ultraviolet curing adhesives), vinylpolymers, polyurethanes and silicones. Wherepossible, the chemical composition of materi-als should be known, also their toxicity beforeand during curing, and/or their potential as acause of skin ailments such as dermatitis. Theuse of a fume-cupboard, extraction fan,goggles or protective gloves may be necessary.In general, products should be easily availableat a reasonable cost, and preferably in smallamounts, to reduce wastage due to expiry ofshelf-life. The shelf-life of unmixed compo-nents should be known. Horie (1987) reviewsof the use of polymers in conservation.

Theory of adhesion

In order for materials to fulfil their roles asadhesives, consolidants or gap-fillers, it isobvious that they must have a reasonableadhesion to glass. In order for this to be so,materials used to conserve and restore glassmust have a strong attraction for glasssurfaces; thus, when applied, they will flowand cover the glass so wetting it. They mustthen set (cure) to prevent movement of thefragments or vessels being treated. They mustbe able to adjust to strains set up during andafter setting, should not put undue strain onthe glass, and be as unobtrusive as possible.It is also desirable that adhesives, consolidantsand gap-fillers should remain soluble overconsiderable periods of time.


Glass has a high-energy surface and, whenclean and dry, will be readily wetted by anyadhesive or contaminant. It is therefore essen-tial that the surfaces to be adhered should bethoroughly cleaned to remove grease and dirt.However, the surface of glass is very hygro-scopic and in normal humidities has manymolecular thicknesses of water lying on it. Thenon-hydrogen bonding organic liquids do notspread spontaneously on the water-bearingglass surfaces. The area of contact betweenorganic liquids and glass can be increased bylowering the level of relative humidity. It istherefore likely that poor initial wetting of theglass surface by the adhesive will result fromthe use of adhesives that cannot displace thewater from the glass surface. Having onceformed the adhesive bond, an adhesive (orgap-filler) may gradually be displaced from thesurface by the attraction of glass for water.This displacement may be further induced bythe deterioration reactions that occur at glasssurfaces exposed to water; alkali ions in theglass migrate to the surface to produce theirsoluble hydroxides.

All polymers are permeable to water to agreater or lesser extent. All adhesives thereforeallow the penetration of water through theirfilms, and on to the glass surface. The watermolecules can react in the pores of the glasssurface that have not been coated withpolymer and also at the interface of the glassunder the coating. The alkaline solutionformed in the pores and at the interface willabsorb more water by osmosis, which willgradually generate a hydrostatic pressureconfined by the adhesive. This pressure willput a stress on the polymer–glass adhesionand promote further loss of adhesion. Thehydrolysis reactions take place faster on highlyalkaline glasses than on the more stablemodern glasses.

Silane treatment

It has been suggested that the problem ofadhesion might be solved by pre-treating theglass surface with a coupling agent beforeapplication of an adhesive (Errett et al., 1984).Such pre-treatment may be particularly usefulin the field of medieval stained glass, wherethe glass is to remain in situ, but its use ofglass in museums is not normally necessary

and would render repairs irreversible by meansthat would not damage the glass. However, itis worth discussing the use of silanes in orderto elucidate on the above statement. There arevarious families of coupling agents, includingphosphate and chromate, but the silane familyhas the most widespread use, and has alreadyshown promising results in the field of stone-restoration.

Silane coupling agents are a class ofmonomers in which both the silicon groupand the organic radicals contain reactivegroups. They are of the general form R-Si(Y)3,Y being commonly –OCH3 methoxy or –OC2H5

ethoxy, and R is a reactive organic group. TheY groups hydrolyse with water to formreactive silanol R-Si(OH)3 groups which canthen react with themselves and with thesurfaces to which they are applied. The Rgroup can be chosen to react with a polymerwhich is subsequently applied to the surface;for example, unsaturated R groups are chosento improve adhesion of vinyl-reactingpolymers such as polyesters, and amine-containing R groups can be chosen forepoxies.

A typical silane coupling agent is amino-propyltrimethoxysilane: H2N-CH2–CH2–CH2-Si (OCH3)3. This is used to promote glass–epoxy adhesion, but many other silanes, with different organo-functional groups, areavailable.

The silane, applied in a dilute solution,reacts with water, on the glass surface and inthe air, to form silantriols by eliminating thehydrolysable groups:

The silantriols react with hydroxyl groupson the surface, and with themselves, to forma tightly cross-linked three-dimensional layerattached to the surface with strong covalentbonds.

A polymer system that will react with theorgano-functional groups can then be applied,the amino groups on the silane reacting withepoxy groups and bonding chemically to theadhesive. Coupling agents considerably


(5.5)

improve the adhesion of polymers to glass,both initially and on prolonged exposure towater. The silane groups immobilize thesilicon atoms to which they are attached andreduce the adsorption and reaction of waterthus delaying failure of an adhesive bond.However, coupling agents will not be able toprevent the diffusion of water molecules intothe pores and onto the surface completely.The use of coupling agents can thereforeincrease the initial strength and life expectancyof an adhesive join considerably, but the bondwill eventually fail due to water penetrationand glass decay. However, their use on glassartefacts in the relatively dry atmosphere inmuseums may ensure that adhesive bonds willbe effectively non-reversible.

Strain

Stress and strain on glass can result during andafter curing of polymers, which invariablyshrink on gelation due to loss of solvent, orto a polymerization reaction. This shrinkagesets up stresses at the interface between thepolymer and the glass surface that makes thebond susceptible to both mechanical andchemical attack, for example, by the absorp-tion of water, and the glass may deteriorate inspecial ways (Newton et al., 1981). The effectsof such stress will depend upon the force(elastic modulus) necessary to maintain theadhesive stretched over the non-yieldingsurface of the glass. A large shrinkage of theadhesive will cause only slight stress when theelastic modulus of the polymer is low. If,however, the elastic modulus is high, such asin the case of epoxy resins, then large stressesare created even though there is only a smallshrinkage of the resin during curing.

Strains may also arise from the mismatch ofthermal expansion coefficient between the

glass and the polymer (see Table 5.1). Theamount of strain produced at the interface willdepend on the temperature change and theforce necessary to compress the polymerwhen it tends to expand more than the glass.Strain may remain locked in the polymersystem and thus remain a potential source ofweakness in the bond; or cause cracking ofthe glass, in the polymer, or along theadhesive interface. Alternatively the polymermay slowly flow to relieve the strain.

Refractive index

Reflections result from light striking an inter-face between two materials with differentrefractive indices, for example air and glass, orpolymer and glass. The larger the differencein refractive indices, the greater the degree ofreflectance. Theoretically, the refractive indexof an adhesive should match that of the glassexactly in order to avoid reflections from thesurfaces of breaks, etc. (Figure 5.1).

For purposes such as joining the glasselements in an optical instrument, the refractiveindex of the adhesive need only match that ofthe glass to ±0.02. In optical elements, however,the light strikes the glass/adhesive interfacenearly at right angles, but as the angle becomesmore glancing the reflection increases. Lightstrikes a repaired glass object from all anglesand hence cracks and joins will be renderedvisible by glancing reflections from the cracksurfaces, unless the refractive indices match by±0.01 (Tennent and Townsend, 1984a).

The refractive index of a material varies withthe wavelength, that is, the colour of the light.The value usually stated for a material is the


(5.6)

Table 5.1 Coefficients of expansion ofpolymers and glasses

Material Coefficient of Tensile linear expansion modulus(cm °C–1 � 10–6) (108 Pa)

Epoxy resins 45–90 24Polystyrene 60–80 28–41Silicones 200–400 0.6Medieval window 8–15 –

glassModern window 8 –

glass

refractive index found using sodium (yellow)light (this is the nD value). A slight mismatchof refractive index may lead to the formationof colours within cracks in the glass (Tennentand Townsend, 1984a,b). An added difficultyin matching the indices is the change of refrac-tive index that occurs at a glass surface whenweathering takes place.

However, there are other sources of reflec-tion; when glass vessels break, they tend tospring out of shape, releasing the stresses inthe glass, and a reconstruction is likely toleave slight misalignment between fragmentswhich create thin lines of reflective surfacesacross the vessel. Excess adhesive is frequentlyremoved from a join before it hardens, byusing solvents, or mechanically after setting. Iftoo much adhesive is removed, reflections willarise from the dry edges left uncovered.

Deterioration of polymers

The polymers used for glass conservation areto a large extent derived from organicmolecules and, therefore, like other organicmaterials, they are susceptible to oxidation anddeterioration. The main environmental influ-ences acting on polymers used in conservationare light (Tennent and Townsend, 1984a,b;Horie, 1987), oxygen and water; heat onlyrarely causes problems.

The mechanisms of deterioration are charac-teristic for each polymer and they arefrequently difficult to unravel, but somegeneral remarks can be made. Any or all ofthe following reactions can be caused by light,oxygen, water, or other impurities in thepolymer film: cross-linking of chains; breakingof the chains (chain scission); formation ofchromophores (groups which absorb light)leading to yellowing of the polymer. Thecross-linking reaction in thermoplastics willconvert a soluble film into an insoluble one.

Chain-breaking reactions cut the polymerchain, thus reducing the molecular weight. (Itis this reaction that causes shortening of thecellulose chains in paper, thus weakening itrapidly when exposed to light and/or air.)Those polymers containing ester groups alongthe chain, such as polyester resins andpolyurethanes, can be hydrolysed. They can,therefore, be slowly weakened by chainscission when exposed to water.

Yellowing of resins is caused by the absorp-tion of blue or ultraviolet light by the chromo-phores. In most of the cases studied, thechromophores arise from loss or alteration ofside groups to form conjugated double bondsand carbonyl groups along the chain. Thesechromophores are reactive and will absorbmore energy from the light, causing furtherdeterioration. The energy contained in light,especially ultraviolet radiation, is sufficientlypowerful to break chemical bonds inpolymers; it frequently initiates deterioration,and is a common cause of failure in polymers.All the forms of deterioration described aboveare caused by chemical changes in the molec-ular structure, but the physical state of thepolymer can change in response to theenvironment. If the temperature rises above itsglass transition temperature (Tg) a polymer hassufficient mobility in its molecules to be ableto flow. For this reason even very toughthermoplastics, such as Nylon, cannot be usedwhere a constant force is applied. Creep understress (cold flow) results in the stretching ofjoins made with polyvinyl acetate. Movementof the polymer chains can also occur on asmaller scale and a particle of dust or dirt lyingon the surface of a polymer above its Tg canbe slowly incorporated and become fixed inthe polymer film; this effect can be seen inemulsion paints where the Tg is less than 4°C,and in polyethylene bags (Tg = –20°C), bothof which attract and hold dirt, even afterwashing. In addition, water will dissolve in apolymer film, causing slight swelling whichmay create stresses between the glass and thepolymer.

The majority of polymer properties can bealtered by adding another material. Forexample, polyvinyl chloride has mechanicalproperties which are similar to polymethylmethacrylate but the addition of softeningagents (plasticizers) makes the PVC polymerflexible. The plasticizers are liquids that act asnon-volatile solvents, but they can slowlymigrate or evaporate, causing the polymer tobecome brittle.

ReversibilityThe deterioration of polymers may or may notmake their removal convenient. In any event,products used for conservation should haveclearly established methods of removal (US: re-


dissolubility), which avoids damage to the glassat any time in the future. The concept ofreversibility, that is, that of taking the objectback to its state before treatment, is basic tothe whole of conservation, nevertheless it mustbe honestly stated that many conservationprocesses are not reversible. Everything, fromdrying the article to removal of weatheringcrusts, transforms the object in some way, andhence some information about the original stateof the object may be lost during its treatment.However, the condition of some objects is sopoor that action must be urgently taken to stopthem disintegrating (Oddy and Carroll, 1999).

Reversibility must be seen on two levels:whether the gross treatment of an object,coating, adhesive, etc., can be removedwithout harm to the object; and whether thetreatment, even after it has been reversed,distorted the information that could have beenobtained from the original state of the object,for example by chemical analysis. Studies ofthe traces of adhesive left after self-adhesivetapes were stripped from glass plates haveshown that minute quantities of adhesiveremain, and can be detected, on an apparentlyclean, smooth surface. The same may well betrue of the polymers alleged to have beencompletely extracted from the pitted andporous surfaces encountered in conservation.Therefore the treatment of an object with anypolymer cannot be considered completelyreversible because of the likelihood of tracesremaining, but the treatment used must berecorded for future reference.

In practice, however, reversibility must beconsidered on a less rigorous basis, andpolymers have been placed in two categories:permanently soluble polymers and polymersthat form cross-links. It must always be bornein mind that it is the particular use of thepolymer (the process) that determines whetherthe treatment is reversible, not the potentialsolubility of the polymer. For example, themajority of cross-linking resins will swellconsiderably when soaked in solvents, andthis swelling will usually disrupt any edge-to-edge join, or surface coating, sufficiently topermit mechanical removal of the bulk of theresin. This use of cross-linking resins is there-fore reversible when the surfaces of the objectare smooth and non-porous, even though thepolymer is not soluble.

Swelling of a polymer, whether cross-linkedor not, will put stress on the object to whichit is attached. Therefore, it is better to removeas much adhesive or coating as possible bymechanical means before swelling, in order toreduce the final stress on the object. The stressis likely to be low when the polymer is freeto expand in at least one direction, forexample, when used as a coating. However,when the polymer has penetrated into poresor between fragments of an object, theswelling of the polymer will tend to push thefragments of the object apart.

The process of consolidation uses thepolymer to penetrate, harden and thereby bindfragments of an object together. It is in sucha situation that removal of the consolidant willdo most harm because the swelling will tendto disrupt the consolidated part of the object.Unfortunately the swelling of the polymer willtend to put the object into tension, the kindof force that most friable structures are leastable to withstand. Thus any process involvingthe partial or complete consolidation of aporous object with any polymer should beconsidered to be irreversible.

As conservation methods improve, treat-ments that were considered to be irreversiblemay be able to be reversed. For example, thepossibility of removing organic materials bythe use of oxidizing plasmas has been demon-strated. This technique may permit theremoval of so-called insoluble coatings.

The polymer selected to fulfil a particularrole will depend upon the influences that thebond is expected to resist. A glass vesselstanding in a museum case has less need fora water-stable adhesive than glass in a windowwhich is frequently wetted, and which isinaccessible to further conservation work.Although it would be undesirable, the vesselcould be restored every few years, but paintedglass in a window must be expected to remainin place for more than one hundred years.

It is unlikely that an adhesive joint cancombine reversibility with durability. Theachievement of durability in the glass–polymerbond seems to require chemical reactionbetween the polymer and the glass surface.The durable bond will degrade in time, withthe loss of some part of the glass surface. Thereversible glass–polymer bond will operate byphysical attraction and so it can be displaced



Table 5.2 Comparative table of optically clear epoxy resins used for glass conservation(manufacturers’ data)

Trade/chemical Mixing Pot life Viscosity of Curing RI Commentsname proportions @ 25°C mixed resin time

Wt. Vol. @ 25°C

Ablebond 342-1(Ablestick Lab.California USA)

Bisphenol Adiglycidyl etherPolyoxypropylenediamine

AralditeAY103/HY956

(Vantico)Bisphenol Adiglycidyl etherModified aliphaticpolyamine

Araldite 2020(formerlyXW396/XW 397)(Vantico)

Bisphenol Adiglycidyl etherAliphatic polyamine

Epotek 301-2(EpoxyTechnology Inc.USA)

Bisphenol Adiglycidyl etherAliphatic amine

Fynebond(FyneConservationServices, UK)

Bisphenol A,epichlorohydrinPolyoxypropylenediamine

HXTAL-NYL-1(Conservators’Emporium, USA)

HydrogenatedBisphenol A,diglycidal etherPolyoxypropylenetriamine

10

3

10

1.6–1.8

10

3

10

3.5

10

3.2

3

1

10

4.6

10

1.8–2.0

10

3.5

10

4.1

8 hours

1.5–2 hours

8 hours

1 hour

8 hours

15 hours

200 cps

140 cps

130 cps

100 cps

No data

No data

48 hours

24 hours

16 hours

24 hours

36–48hours

7 days

1.565

No data

1.553

1.564

1.565

1.549

Resin crystallizedat roomtemperature. Nolongermanufactured

Especiallyformulated foruse on glass

Resin crystallizesat roomtemperature.Designed foroptical filters

Resin crystallizesat roomtemperature

Formulated forglass by Dr N.TennentExpensive in UKBest ageingperformance

Formulated forglass by HerbertHillary

physically by water in the environment; thusit is less durable.

To create the most durable chemical bondsbetween most polymers and glass, a couplingagent should be used. The improvement inadhesion will be gained only when both thepolymer and the glass react with the couplingagent. This implies the use of one of thereactive, cross-linking, polymers as adhesives,consolidants or lacquers.

Types of synthetic polymer

Cellulose derivativesCellulose derivatives have been used for theadhesion and consolidation of antiquities formany years, and their properties are wellunderstood. Cellulose nitrate (sometimesincorrectly referred to as nitrocellulose) wasthe first major plastic in commercial use,having been derived from natural cellulose in1838, and produced industrially as early as1845. Plasticized with camphor, cellulosenitrate was patented in Britain in 1864, andwas discovered independently in America in1869, where it was marketed as Celluloid.Cellulose nitrate is formed by the reaction ofnitric acid with cellulose under carefullycontrolled conditions, and externally plasti-cized, commonly with dibutyl phthalate.Plasticized cellulose nitrate has a refractiveindex of 1.45–1.50. References to the early useof cellulose nitrate on decayed glass relate toa product called Zapon (Pazaurek,1903), andanother marketed as Durofix (Plenderleith andWerner, 1976) In the form of HMG, Durofixand Duco Cement (US), cellulose nitrate is stillin common use (Koob, 1982). In the form ofthe lacquer Frigiline, it was suggested for useas a consolidant for decayed enamels. A moresuitable polymer would be Paraloid B-72 (US:Acryloid B-72, an ethyl methacrylate/methylacrylate copolymer).

However, cellulose nitrate is known to havepoor ageing qualities (Hedvall et al., 1951;Koob, 1982). Despite the fact that attention hasbeen drawn to the long-term instability ofcellulose nitrate, it still remains a most usefuladhesive for the restoration of glass, becauseof its comparative lack of colour, ease ofapplication from a tube, and reversibility inacetone. As the product HMG it is widely usedfor the repair of archaeological glass in Britain.

Soluble nylonBefore the use of the chemically stableproduct Paraloid B-72 (methyl methacrylatecopolymer) as a consolidant in many areas ofconservation, a product known as, Solublenylon (N-methoxymethyl nylon) was widelyused in the 1950s and 1960s. Marketed asCalaton CA and CB and Maranyl C109/P, itwas usually made into a 3 per cent w/vsolution in industrial methylated spirits. Its useon flaking glass surfaces has been mentionedby Dowman (1970), and Melucco (1971), whoadded that soluble nylon acted as a holdingtreatment, not prejudicing future treatments ofthe glass. The problem of cross-linkingtogether with the fact that soluble nylonattracts dust onto the surface of objects onwhich it has been used, has resulted in thestrong discouragement of its use (Sease, 1981).

Undeteriorated glass is almost impermeableto solvent vapour, and hence resins thatrequire solvents to evaporate in order to effecttheir cure are inconvenient for use in repair,consolidation or restoration. For this reason,resins that polymerize in situ, and do notrequire access to the air for evaporation ofsolvent, are more suitable. In addition, theshrinkage of a polymerizing polymer is consid-erably less than a solvent-deposited polymerand hence strains are less likely to be exertedupon the glass.

Epoxy resinsEpoxy resins used in conservation are typicallycomposed of two parts, a di-epoxy componentand a polyamine cross-linking agent, both ofwhich are compounded with diluents andcatalysts (Table 5.2).

The hydroxyl groups formed in this reactioncan take part in and contribute to furtherreaction with epoxy groups.

The epoxy resin shown in (5.8) iscommonly used, and is a bisphenol A andepichlorohydrin condensate. The value of n


(5.7)

usually should be less than two in order tomaintain the epoxy as an easily worked liquid.There is a wide range of amines available foruse; many of the low viscosity amines are skinsensitizers and have a dangerously highvolatility. For domestic use less unpleasantliquid polyamides are used with a catalyst. Thepolyfunctional epoxies and amines ensure across-linked polymer of high strength. Theadvantages of epoxy adhesives and gap-fillersare their relatively high adhesion and lowshrinkage on curing. However, there is alwayssome shrinkage in the polymerization ofmonomers. The shrinkage of an epoxy resinoccurs at two stages, the first during the liquidstate, and the second (greater amount) aftergelation. A disadvantage of the bisphenolA/amine adhesives is their tendency to yellowwith time and light exposure, owing to theformation of chromophores arising from thebenzine rings that react with amine break-down products. Bisphenol A epoxies, whencured with anhydride cross-linking agents at80°C, form stable, non-yellowing polymers.Slight discoloration of the resin may be notice-able in time, when it is used on optically clearglass.

A disadvantage of aliphatic epoxies is theirgreater sensitivity to moisture. The use ofepoxy resin has been suggested for consoli-dation of archaeological glass on site (Wihr,1977), but the resin yellows, is difficult toremove from the glass surface, and is irremov-able from porous materials such as decayedglass. In general, epoxy resins should not beconsidered for use as glass consolidants,particularly as there are more suitable productsavailable, such as Paraloid B-72.

An American epoxy resin, Ablebond 342-1,was formerly widely used in the UK and theUnited States in the 1970s and 1980s for glassbonding, but is no longer available. A disad-vantage of this product was that the hardenertended to crystallize at room temperature sothat, on being measured for use, the hardenerhad to be warmed to liquidize it, before

mixing it with the resin. Other resins used areAraldite AY103/HY956 and Araldite 2020(formerly known as XW396/XW397 in itsexperimental stage), the latter specificallydesigned for bonding glass.

At present excellent results in the repair andrestoration of glass artefacts are being obtainedby the use of a number of optically clearepoxy resins: Araldite 2020, HXTAL NYL-1 andEPO-TEK 301-2 (both US products) andFynebond (UK). Unlike other optically clearepoxy resins, the chemicals in HXTAL NYL-1and Fynebond, whilst not being speciallysynthesized, are optimized for conservationuse. HXTAL NYL 1 is very expensive to importto the UK, and takes a week to cure fully. Theresin component of Fynebond tends to crystal-lize at room temperature and so has to bemelted before being mixed with its hardener.

When testing resins for bonding glass, itshould be noted that their specification can bewithin the manufacturer’s tolerance but not upto the standard required for conservation. Thisaccounts for the colour differences that occurfrom batch to batch and for the fact that testscarried out on the products by conservationscientists may be meaningless if the batchtested is at the high end of the manufacturer’stolerance, but a batch used by a conservatoris at the lower end. Other factors such as theage of the resin when tested will rendercomparative tests on epoxy resins (and othermaterials) meaningless. Many factors affect thediscoloration of epoxy resins, not all of whichare fully understood. However it is known that(i) the hardeners are not exactly the same foreach type of epoxy resin, and may cause longterm instability, (ii) some resins containadditives (e.g. dibutyl phthalate plasticizers)which can increase their tendency to yellow,and (iii) the purity of both components isimportant as even minor impurities maydecrease the stability of a resin. Down (1986)has reported on the yellowing of epoxy resins.Messenger and Lansbury (1989) and Augersonand Messenger (1993) have reported on the


(5.8)


Figure 5.1 The results of RI measurements on a variety of adhesives, and historic glasses, compared with glassstandard.

Adhesives Historic and artistic glass Glassstandards

PVARESINS

ACRYLIC

RESINS R

OMAN

&

EGYPTIAN

GLASS

POST

MEDIEVAL

VESSEL

GLASS

LEAD

GLASS

MEDIEVAL

STAINED

GLASS

DGEBA

EPOXY

RESINS

SILICONE

RESINS

CELLULOSE

POLYESTER

NITRATE

RESINS

RESINS


ADHESIVES1 Arbosil Aquarium Sealant (silicone). Adshead Rateliffe & Co.

Ltd, Belper, Derbyshire DE5 1WJ, UK2 Eccosil 2 CN (silicone). Emerson & Cuming (UK) Ltd, 1

South Park Road, Scunthorpe, South Humberside DN172BY, UK

3 Silastic 732 RTV (silicone) Dow Corning Ltd,4 Dow Corning 3140 RTV (silicone) Reading Bridge5 Dow Corning 3144 RTV (silicone) House, Reading, Berks7 Sylgard 184 (silicone) RG1 8PW, UK6 Loctite Clear Silicone Sealant

(silicone) Loctite UK, Loctite30 Loctite 357 (UV curing, urethane Holdings Ltd,

acrylic) Watchmead, Welwyn33 Loctite 350 (UV curing acrylic) Garden City, Herts38 Loctite ‘Glass Bond’ (UV curing AL7 1JB, UK

system)8 Rhodopas M (PVA): after casting Rhodia (UK) Ltd,

10 from solution (8), as supplied (10) 14 Essex St, LondonWC2R 3AA, UK

9 Vinalac 5254 (PVA). Vinyl Products Ltd, Mill Lane,Carshalton, Surrey, UK

11 UHU ‘All-purpose’ (PVA). Beecham UHU, Brentford,Middlesex, UK

12 Evostik ‘Resin W’ (PVA) Evode Ltd, Common35 Evostik ‘Clear’ (cellulose nitrate) Road, Stafford, UK13 Berol ‘Merlin’ (PVA). Berol Ltd, Oldmedow Road, King’s

Lynn, Norfolk, UK14 Paraloid F10 (acrylic)15 Paraloid B72 (acrylic): after19 casting from solution (15), Rohm & Haas (UK)

as supplied (19) Ltd, Lennig House,16 Paraloid B82 (acrylic) 2 Mason’s Avenue,18 Paraloid B67 (acrylic) Croydon CR9 3NB,20 Paraloid B66 (acrylic) UK23 Paraloid A30 (acrylic)25 Paraloid B44 (acrylic)17 Butvar B76, B79 (polyvinyl

butyral, PVB) Monsanto Ltd,22 Butvar B72, B74, B90. B98 Monsanto House,

(PVB) 10–18 Victoria Street,26 Formar 12/85 (polyvinyl formal, London SW1 0NQ,

PVF) UK. (Manufacturer’s31 Formar 5/95E, 6/95E, 15/95E data)

(PVF)21 Plastogen G (acrylic)28 Durafix (cellulose nitrate) Frank W. Joel Ltd,34 HMG (cellulose nitrate) Oldmeadow Road,36 Frigilene (cellulose nitrate) Hardwick Industrial37 Ercalene (cellulose nitrate) Estate, King’s Lynn,40 Rutapox 1200 (epoxy) Norfolk PE30 4HH,44 Plastogen EP (epoxy) UK57 Rutapox 1210 (epoxy)24 Lascaux Acrylglasur 40X (acrylic). Alois K. Diethelm

Farbenfabrik, CH-8306 Brüttisellen, Switzerland27 Technovit 4004A (acrylic). Rubert & Co. Ltd, Dennings

Road, Cheadle, Cheshire, UK29 Bostik 5293 adhesive (cellulose nitrate). Bostik Ltd,

Ulverscroft Road, Leicester LE4 6BW, UK32 L.R. White Resin (acrylic). London Resin Co., P0 Box 24,

Basingstoke, Hants RG22 5AS, UK39 HXTAL NYL-1 (epoxy) Conservation Materials50 Ablebond 342-1 (epoxy) Ltd, 340 Freeport51 Devcon ‘2-ton’ (epoxy) Blvd, Sparks, NV52 Devcon ‘5-minute’ (epoxy) 89431, USA

41 Opticon UV-57 (UV-curing system). Opticon Chemical, POBox 2445, Palos Verdes Peninsula, CA 90274, USA [6]

42 Norland Optical Adhesive 63(UV-curing system) Leader House, 117–120

47 Norland Optical Adhesive 61 Snargate Street, Dover,(UV-curing system) Kent CT17 9DB, UK

59 Vitralit DAC(UV-curing system)

43 Tiranti Embedding Resin (polyester). A. Tiranti Ltd, 20Goodge Place, London W1, UK

45 Epotek 301 (epoxy). Epoxy Technology Inc., 65 GroveStreet, Watertown, MA 02172, USA

46 Metset Resin SW (polyester): first batch (46), second batch48 (48). Metallurgical Services Laboratories Ltd, Reliant Works,

Brockham, Betchworth, Surrey RH3 7HW, UK49 Araldite MY790/X83-19/DY040 Ciba-Geigy Plastics

(epoxy) and Additives Co.,55 Araldite AY103/HY951 (epoxy): Plastics Division,58 first batch (55), second batch (58) Duxford, Cambridge60 Araldite AY1OS/HY951 (epoxy) CB2 4QA, UK53 Epofix (epoxy). Struers, Copenhagen, Denmark54 CXL (epoxy). Resin Coatings Ltd, Colebrand House. 20

Warwick Street, Regent Street, London W1R 6BE, UK56 Permabond E27 (epoxy). Permabond Adhesives Ltd,

Woodside Road, Eastleigh. Hants S05 4EX, UK

HISTORIC AND ARTISTIC GLASS*

1 GMAG/DA/96.38a. Cristallo (greenish) Venetian Tazza, 16thcentury

2 NHT. Clear (2) and violet (3) glass from Alexandrine3 Zwischengoldglas fragment, 1st century AD

4 GMAG/BC/17.64. Dark blue transparent glass from Romanbottle, early centuries AD

5 R/NM10754.248. Bohemian glass, wheel-engraved, mid-18thcentury

6 GMAG/DA/97S0bv. Spanish (Catalonia) jug, clear glass,28 18th century: vessel glass (6), ornament glass (28)7 NHT. Clear glass from Islamic lamp fragment, 13th/14th

century AD

8 Egyptian vessel stamp, 8th century AD [19, Table 1]9 Eleven Egyptian vessel stamps, coin weights, etc., 8th

century AD [19, Table 1]10 NHT. Clear glass from Islamic lamp fragment, 13th/14th

century AD

11 R/NM10754.270 Russian (?) covered beaker, wheel-engraved,mid-18th century

12 GMAG/A/03.185gp. Cypriot glass flask, early centuries AD

13 Egyptian beads: one pale green from Tell Asmar, 2700–2500BC [15, p.10, specimen no. 7]; two blue from Qau, 7525[16]

14 NHT. Transparent amber (14) and blue (16) glass from16 Alexandrine lozenge-shaped decorative glass, c. 1st century

AD

15 R/NM8326. Bohemian glass, wheel-engraved, 1725–1750 AD

17 GMAG/A/03.185iz. Cypriot glass dish, early centuries AD

18 Egyptian blue bead from Tell el-Amarna, XVIII Dynasty [16]19 GMAG/DA/97.50cq. Southern Spanish bull-shaped ornament,

green glass, 19th century?

* Explanation of registration codes: GMAG/DA, GMAG/BC,GMAG/A = Glasgow Museums and Art Galleries, Departmentof Decorative Art, Burrell Collection, Department ofArchaeology, respectively; NHT = N.H. Tennent’s collection; R= Rijksmuseum, Amsterdam.

�

�

�

�

�

�

�

�

�

�

��

�

�


20 GMAG/DA/97.50bu. Spanish (Catalonia) jug, clear glass,16th century?

21 Three Egyptian red beads from Armant, 1200A [16], XIIDynasty

22 Egyptian vessel stamp, early-middle 8th century AD [19,Table 1]

23 Egyptian ring weight, 8th century AD [19, Table 1]24 Egyptian green glass bead from Qau, 612 [15, p. 14, speci-

men no. 1], XI Dynasty [16]25 Egyptian coin weight, early 9th century AD [19, Table 1].26 GMAG/DA/96.38c. Venetian wine glass, late 16th century27 Egyptian coin weight, late 8th/early 9th century AD [19

Table 1]29 GMAG/DA/93.93w. Blue (29) and opal (54) glass from54 Venetian wine glass, 16th/17th century30 NHT. Transparent dark blue vessel glass fragment, Tell el-

Amarna, XVIII Dynasty31 GMAG/BC/17.49. Pale green glass from Roman jar, early

centuries AD

32 Arabic bead from Assint, probably 14th–18th century AD [16]33 NHT. Transparent turquoise vessel glass fragment, Tell el-

Amarna, XVIII Dynasty34 Transparent pink bead from Armant, 1200A [16]3536 Excavated medieval stained glass from St Andrews37 Cathedral (35; StA1) and Elgin Cathedral (36; EC4. 37;55 EC5, 55; EC1). See [17] for analytical data.38 GMAG/BC/45.470. Clear glass from Flemish roundel, early

16th century39 GMAG/DA/97.50bo. Spanish (Andalusia) vase, green glass,

17th century40 Clear dark blue bead from Armant, 1213D [16]41 GMAG/DA/27.90e.3. Water-glass, Scottish (John Baird

Glassworks), c. 188142 GMAG/BC/45.484. Clear, blue and green glass from

German (Nuremburg) stained glass panel, c. 1470–148043 GMAG/BC/45.462. Clear glass from Flemish roundel, 16th

century44 GMAG/DA/88.65a. Clear (44) and yellow (49) glass from49 ruby glass goblet, Rhine Glass Works Co., c. 188845 GMAG/BC/45.17. Clear glass border dated 1577 (enclosing

a 14th century roundel)

46 GMAG/BC/45551. Clear glass from Dutch panel, dated 168647 GMAG/BC/45.499. Clear glass from Swiss panel, late 16th

century48 GMAG/BC/Inv 151. Wine glass. Dutch/British, mid-18th

century50 R/NM10754.78. Wine glass, stipple engraved. Dutch, late-

18th century51 R/NM4015. Wine glass, diamond engraved, English,

1725–175052 R/RBK 1966.61. Wine glass, wheel-engraved in Holland,

English, c. 175053 GMAG/45.32. Clear glass from English panel, 14th century

GLASS STANDARDS

1 Glass FK3: Schott Glaswerke, Hattenbergstrasse 10, D-6500Mainz, W Germany

2,2' Glass Standard No. 4 Society of Glass Technology,3 Glass Standard No. 5 20 Hallam Gate Road,11 Glass Standard No. 3 Sheffield S10 5BT, UK4 Zinc Crown, ZC 5086125 Hard Crown, HC 5245927 Extra Light Flint, Pilkington Brothers plc,

ELF 541472 Research & Development8 Extra Light Flint, Laboratories, Ormskirk,

ELF 548456 Lancashire L40 5UF, UK15 Light Flint, LF 56742819 Light Flint, LF 579411

Standards prepared for the6 Glass Standard B Corning Museum to9 Glass Standard D simulate the composition of

ancient glasses [21]10 Standard 76-C-15112 Standard 76-C-158 Series of synthetic13 Standard 76-C-149 medieval stained glass14 Standard 76-C-144 compositions prepared by16 Standard 76-C-150 Pilkington Brothers for17 Standard 77-C-33 Professor R.G. Newton18 Standard 76-C-145 [22]20 Standard 76-C-14721 Standard 76-C-148

control of the refractive index of the epoxyresins Ablebond 342-1 and HXTAL NYL-1. InFigure 5.1 a comparison is made betweenvarious optically clear epoxy resins that havebeen used for glass repair.

Polyester resins

Polyester casting resins are widely used forjoining glass elements in optical instruments,but there are few recommendations for their

use as adhesives in the conservation of glassantiquities. However, those incorporating anultraviolet light absorbant are extremely usefulas gap-fillers, and will be discussed in moredetail in the section on casting materials.

Casting and laminating polyesters aresolutions of unsaturated polyesters in an unsat-urated reactive monomer. Typical polyestersare made from propylene glycol, maleicanhydride and phthalic anhydride, and have ashort molecule (molecular weight 2000), of thegeneral formula shown in (5.9).

�

�

�

�

�

�

�

This represents a highly unsaturatedpolymer. Styrene is usually chosen as thereactive monomer because of its low cost andease of use. Methyl methacrylate can replacesome of the styrene when resistance to lightis required. The cross-linking between thechains is achieved by copolymerizing the sty-rene with the unsaturated bonds in the chainby using a peroxide initiator. This results in across-linked polymer that is both hard andrigid. The properties of the cross-linked, curedresin can be varied, by altering any of theseveral components that make up the formu-lation. All polymerization reactions result inshrinkage, and polyester formulation is noexception. A typical value for the shrinkage is8 per cent by volume, most of which occursafter gelation. Heat generated during polymer-ization can lead to charring and cracking whenlarge masses of polyester resins are being cast.(Large amounts of resin are not commonlyused in the conservation of glass.) The refrac-tive index of stabilized polyester resins istypically 1.54–1.56.

Acrylic resins

The monomers from which acrylic polymersare made fall into two groups, acrylates andmethacrylates. The methacrylates were one ofthe first synthetic resins used to coat glass(Hedvall et al., 1951), and as consolidants theyare still amongst the most popular. Acrylates

and methacrylates are nominally derived fromacrylic and methacrylic acids, respectively.These acids can be esterified with alcohol toproduce a wide range of monomers:

The polymers made from the acrylates tendto have lower Tg points than the equivalentmethacrylates. For example, a polymethylmethacrylate such as Perspex has a Tg of105°C, whereas polymethyl acrylate has a Tgof 3°C. Polymethyl methacrylate sheet doesnot yellow on ageing and typifies the majoradvantage of acrylics; that is, their lack ofcolour change and resistance to oxidation.However, both the polyacrylates and thehigher polymethacrylates will cross-link underthe influence of ultraviolet light and willeventually become insoluble. Polymethylmethacrylate is the most stable of the acrylicpolymers, reacting only very slowly to ultravi-olet light. The presence of methyl methacry-


(5.9)

(5.10)

late in copolymers increases the resistance todeterioration by light disproportionately to itsconcentration in the polymer. For this reasonmethyl methacrylate is much used in bothacrylic and other polymers.

As polymethyl methacrylate has too high aTg for many adhesive and coating uses, softeracrylic polymers, which still retain colourstability, are prepared. Poly(butyl methacry-late), which has frequently been used in thepast, has been shown to cross-link in ultravi-olet light (Feller, 1972), and hence suggestionsfor its use, often under the trade-name ofBedacryl 122X (now Synocryl 9122X) shouldbe reconsidered. Wihr (1977) used a solutionof Plexigum 24, a polybutyl methacrylate, andstated that it did not yellow and adhered wellto the glass. However, this material could beexpected to cross-link in time. Paraloid B-72is now being widely used in conservation asa consolidant both on vessels and paintedwindow glass. Paraloid B-72 is a very stableresin with a Tg of 40°C and a refractive indexof 1.49. Poly(alkyl methacrylates) with longerside chains have lower Tg values and may beuseful for conservation purposes. Copolymersof the methacrylates with acrylates such asethyl- or 2-ethylbexylacrylate are used toachieve softer products with good colourstability. The relatively good stability of acrylicpolymers over other polymers has resulted ina large number of speciality adhesives andcoatings designed for use where degradationis a problem.

Cyanoacrylate resins

Cyanoacrylate resins have been recommendedfor use in repairing glass (Moncrieff, 1975;André, 1976) because of their ease of appli-cation. Cyanoacrylates polymerize in situ in afew seconds at room temperature without theaddition of a catalyst, the reaction beingpromoted by the presence of moisture orweak bases present on the glass surface.

The cured products are high molecularweight polymers that can be dissolved inorganic solvents. The dilute alkali solutions onthe glass surface cause deterioration of thepolymer; thus cyanoacrylate adhesives aregenerally unsuitable for glass restorationexcept for effecting temporary repairs, forexample when pressure-sensitive tape cannot

be used on a delicate surface. Care should betaken in their use since cyanoacrylate resinsbond very readily with skin tissue. In order toprolong their shelf-life the resins should bestored at a temperature of –20°C.

Ultraviolet curing (photosensitive) acrylicresins

Acrylic resin formulations that cure onexposure to ultraviolet (UV) light have beenspecially developed for joining glass to glassand other substrates. The curing reactionresults in a highly cross-linked polymer:

The only brand of UV curing acrylate whichhas been tested for use in conservation andpublished (Madsen, 1972) is the Americanproduct Opticon UV57 which was found byMoncrieff (1975) to have poor physical proper-ties both in the uncured and cured states. Asecond American UV curing adhesive, NorlandOptical Adhesive 61, has been used for tackingsmall fragments of glass, which are difficult totape. The use of UV curing adhesives shouldnormally be considered to be non-reversibleespecially on porous glass surfaces where it islikely that mechanical keying will hold thepolymer in the pores of the glass therebymaking the bond virtually impossible toseparate. The bond may only be consideredto be reversible if it has been made between


(5.11)

(5.12)

two smooth surfaces, which can be cleanedmechanically after the join has been disruptedby soaking in solvents such as acetone ordichloromethane. Alternatively, prolongedsoaking in water can destroy the adhesivebond (but not dissolve the adhesive) if nocoupling agent has been used.

An unexpected danger has been noticedwhen exposing glass to UV light. A test-piece,which was joined with a UV curing adhesive,crizzled during the exposure to UV radiation.

Polyurethane resins

Polyurethane resins adhere well to polarsurfaces because the isocyanate groups reactwith adsorbed water and hydroxyl groups; thetightly cross-linked network is fairly resistantto swelling by solvents. They are not recom-mended for direct contact with glass, butpolyurethane foams are useful for block liftingfragile objects and structures (see LiftingMaterials, below).

Silicones and silanes

The term silicone is used as a loose descrip-tion for the many compounds formed fromsilicon and organic radicals, and it thereforecovers a series of compounds analogous tocarbon compounds, but containing silicon inthe main chain.

Silicon materials used in conservation(Horie, 1987) may be divided between thesilanes and the silicones. The silanes are fairlysimple molecules that are formulated to reactstrongly and irreversibly. These alkoxysilanespenetrate well and are used primarily (ascoupling agents) to increase adhesion betweena surface and a polymer, though they may alsoadd to consolidation and reinforcement (Errettet al., 1984). Silicone rubbers (US: silastomers)are used where a flexible, stable rubber isrequired for joining fragments, as mastic, or asa moulding material, and are of the roomtemperature vulcanizing (RTV) type. Theyhave the general formula:

The rubbers start as viscous liquids(composed of large silicone molecules plusfillers etc.) that then react to form a cross-linked rubbery solid, more or less rigiddepending on the formulation. The largemolecules have reactive end groups that cross-link through small reactive molecules (silanes).Traces of water and metal catalysts arenormally required. The liquid pre-polymers areavailable as two-part formulations where thecatalyst is added, or as one-part formulationswhere water is excluded until they are used.All the formulations contain a proportion ofun-reacted silicone oil that, because it has alow surface tension, can migrate away fromthe rubber to cause staining of porous materi-als, and may act as a release agent.

The rubber products made from silanes areinherently water white and have exceptionalweathering stability. They change hardly at allwith exposure to light and water, but theyhave little strength compared with otherelastomers. The two-part (RTV) silicones donot have much adhesion to surfaces such asmetal, but have sufficient adhesion to glass toneed a separating agent when silicone rubbermoulds are made. The one-part curingsilicones can have appreciable adhesion,depending on their cross-linking functionalgroups, although they may require couplingagents for maximum adhesion. Silicones areinsoluble in solvents but they will swellconsiderably in aliphatic, aromatic and chlori-nated hydrocarbons.

Clear silicone rubbers with a refractive indexof 1.43 are used commercially to repairdamaged plate glass windows. However, sincethey have a very low Tg of –123°C they attractdirt and within a few weeks the lines ofsilicone rubber become grey and then black-ened as a result of dirt held on, and eventu-ally in, the polymer. The products are notsuitable for repairing archaeological andhistoric glasses.

Silane coupling agents were discussed in thesection on adhesion, and silicone mouldingmaterials will be discussed under mouldingmaterials.

Vinyl resins

Polyvinyl acetate (PVAC) has been used inconservation for many years, including the


(5.13)

consolidation of glass, but has largely beensuperseded by the use of Paraloid B-72(Dowman, 1970; Brill, 1971a; Majewski, 1973;Hutchinson, 1981). PVAC is stable, has a refrac-tive index of 1.46 (similar to that of glass), agood resistance to yellowing and it is availablein a wide range of molecular weights. Themolecular weight is related to the viscosity ofthe solution and is important for achievingadequate penetration, for example, a PVACmolecule with a molecular weight (MW) of51 000, equivalent to 600 monomer units perchain (the degree of polymerization, DP),would have a minimum diameter of 13 nm,and thus it could not penetrate pores as smallas that. However, it has a low Tg (28°C) andis therefore prone to cold flow, and to attract-ing and absorbing dirt, when the MW is low(<57 000). In order to inhibit this tendency toattract dirt, a top coat of a compatible polymerwith a higher T should be applied. PVACappears to be a useful consolidant fordehydrated glass, as a 25 per cent solution intoluene, or as more dilute solutions in alcohol.

Polyvinyl chloride (PVC) solutions (Derm-O-Plast SC, ArcheoDerm) have been recom-mended for consolidating degraded glass(Wihr, 1977). Unfortunately, PVC is one of themost unstable polymers in commercial produc-tion and should therefore not be used in directcontact with antiquities; it will degrade,discolour, release hydrochloric acid andprobably become insoluble with time.

Polyvinyl butyral (PVB) has a refractiveindex of 1.49 (and is used as the interlayer oflaminated windscreens for motor vehicles); ithas good adhesion to glass and a fairly satis-factory Tg of 62–68°C, but it cross-links onprolonged exposure to light and should there-fore be considered to be insoluble in the longterm. Of the grades of PVB that have beenrecommended for glass conservation are thosewith lower molecular weights, for example,Butvar B-98 (Monsanto) and Mowital B20H(Hoechst). PVB has been used by Vos-Davidse(1969).

Vinyl polymers that are polymerized inemulsion can be significantly different fromthose polymerized by other techniquesbecause surfactants and other emulsifiersbecome inextricably incorporated in thepolymer. The advantage of emulsions is thatwater (inexpensive and non-toxic) is used as

the diluent, and the absence of hazardoussolvents has encouraged the use of emulsionswhere large open areas of liquid are required,such as in the conservation of textiles andpaper.

A polymer emulsion has a very high solidscontent with a low viscosity. The viscosity ofthe emulsion is often deliberately increased bythe manufacturer, to make handling easier.The high content of solids permits thickerfilms to be deposited, with less shrinkage thanis the case with films applied from solvents.Up to 10 per cent of the solids content of anemulsion consists of the emulsifiers and stabi-lizers. When the emulsion dries to form a film,the emulsifiers in the polymer film are knownto cause sensitivity to water and a lowering ofclarity. Many of the emulsifiers cause yellow-ing of the film, for example, films of PVACcast from emulsions turn yellow far morerapidly than those cast from solution. Furtherproblems arise when polymeric emulsifierssuch as polyvinyl alcohol (PVAL) orpolyacrylic acid have been incorporated intothe film. Both can undergo cross-linkingreactions and thus reduce the solubility of thepolymer to which they are attached.

Emulsifiers usually work best with a properbalance of ionic materials in the solution.There are three categories of emulsifiers:anionic, non-ionic and cationic. The cationicemulsifiers are rarely used. Anionic emulsifiersrequire acid conditions, while non-ionicemulsifiers can withstand a wider pH range.Before using any emulsion on glass, conser-vators should ensure that it has a pH at whichthat glass is stable. The films formed fromemulsions are frequently soft, with Tg belowroom temperature. As the water evaporatesfrom an emulsion, the polymer particles areforced closer together. These particles mustcoalesce if a coherent film is to form andtherefore the particles must be soft enough toflow into one another. In general the polymerparticles must be above their Tg for coales-cence to occur and thus the polymer film isusually above its Tg when set. Films formedat room temperature will therefore have a Tgbelow room temperature and will absorb dirt.Volatile solvents may be added to theemulsion in order to soften the polymertemporarily. Only small quantities can beadded before the solvents destabilize the


emulsion. Emulsions are usually formulated bymanufacturers to meet a particular need. Fromthe conservation point of view, many of thedetrimental effects of commercially availableemulsions could be reduced by changes inproduct formulation.

Modelling materials

Missing areas of glass may be modelled up insitu prior to moulding, using one of severalcommercially produced modelling materials.Those most commonly used are potters’ clayand Plasticine, a putty composed of petroleumjelly, fatty acids and whiting. The oilysubstances enable the Plasticine to be workedto a smooth surface; however, it does notadhere very well to glass, and will contami-nate the surface. Plasticine residues should beremoved with cotton wool swabs moistenedwith a degreasing solvent such as acetone ortoluene before adhesives are applied to theglass. Aloplast is a modelling material similarto Plasticine in texture, but formulatedespecially for use against polyester resins. Theuse of the buff-coloured version is to bepreferred, since colour from the dark blueversion may discolour the resin.

Damp potters’ clay adheres very well toglass and is easily worked to a smooth surfacewith a spatula dipped in water. In addition,any fragments of glass that can be positionedaccurately but do not actually join to the bodyof a vessel (i.e. floating fragments) may beheld in position by placing them in situ on aclay former. The disadvantage of using clay isthat moisture contained within it separates anyadhesive used in the repair from the glass.Backing joins with tape, rubber latex or thinsheets of wax does not prevent this fromoccurring. In fact, adhesive on Sellotape andmasking tape breaks down with moisture toform a messy substance, removal of the waxsometimes causes the joins to fail, and latexflows into tiny cracks and chips and is diffi-cult to remove without dismantling the glass.

Moulding materials

Materials used for taking moulds from glassartefacts should have the following properties:

They should not harm the object physically byadhering too strongly to the surface, by pullingoff glass projections, or by heat generation.They should not harm the glass chemically bycontaminating or reacting with it. Mouldingmaterials should reproduce all the fine detailsof the original without distortion. The viscos-ity and thixotropic properties should be suffi-ciently variable by the manufacturer orconservator to allow the materials to beadapted to meet different requirements. Theyshould preferably be available at reasonablecost and have an adequate shelf-life. Mouldsmust be able to withstand heat of polymer-ization of the proposed casting material andmust not react with that material. For mould-ing glass the most suitable materials are tough-ened dental wax and silicone rubber (US:silastomer), which is available in severalgrades of thixotropicity. Large silicone rubbermoulds are sometimes given added strengthby the addition of an outer case or mother-mould, constructed of plaster of Paris, or ofpolyester resin incorporating glass-fibrematting.

Dental waxes

Dental waxes are composed of a number ofdifferent waxes and are often supplied assheets measuring 180 mm � 82 mm � 1.5 mm.Dental waxes are available in various gradesof hardness, in sheet sizes up to 305 mm �203 mm, and in thicknesses of 0.4 to 3.0 mm.For making moulds, the sheets or parts ofthem can be softened by gentle heating inwater or warm air before shaping them overthe glass object. Before casting, the waxmould must be coated with polyvinyl alcohol(PVAL) (which is difficult because it tends torun into pools) as a release agent between thewax and the resin to be cast against it. ThePVAL tends to crawl back from the waxsurface into pools. This can be alleviated, byadding a drop of non-ionic detergent to thePVAL to break the surface tension, and ifnecessary by continuously gently brushing therelease agent until it has almost set. It may benecessary to apply a second coat when thefirst has dried; this must be done carefully inorder not to disrupt the first layer, which willhave dried to form an easily disruptable skinover the wax surface.


Silicone rubber (US: elastomer)

The majority of silicone rubber products usedfor moulding cure by catalytic elimination ofalcohol to form cross-links between thechains. When the alcohol evaporates, therubber shrinks but the amount of shrinkage issmall (less than 1 per cent) and occurs over aperiod of a few days. However, shrinkage of2.2 per cent has been observed to occur overa number of years. Silicones are insoluble insolvents but can be swelled considerably bythe use of aliphatic, aromatic and chlorinatedhydrocarbons. Wihr (1977) has suggestedswelling silicone rubber back to size by expos-ing it to organic solvents, but this would seemto be an unreliable method. In the majority ofcases silicone rubber requires no release agentbetween it and the glass surface, althoughinstances of silicone rubber adhering to glassand porcelain have been known (Morgos etal., 1984), and therefore preliminary tests mustbe undertaken. A release agent such as petro-leum jelly or an organic lacquer must be usedif silicone rubber is to be cast against a curedsection of silicone rubber, or the two willadhere. Thin layers of silicone rubber may tearwhen peeled off an object, but thick layers arehard-wearing and the moulds are reusable. Ifnecessary, it may be over-catalysed to shortenits setting time, for instance, when siliconerubber is being used to reattach a siliconerubber mould to glass.

Inert fillers such as kaolin, talc or aerosolsilica may be added to thicken mobile gradesof silicone rubber. Unfortunately, many curedsilicone rubbers are prone to tearing and mustbe applied in thick section and often with arigid case mould for support. Rubber latexshrinks too much to be of use in mouldingsuch a precise material as glass; a small shrink-age will mean that details such as trailedthreads on the glass will not match up withthose on the cast. A mould must remain stablefor several days or weeks whilst restoration isin progress.

Hot-melt preparations

Hot-melt preparations such as gelatine,Formalose and Vinamold (PVC) should beavoided for direct use on glass since the heatmay cause damage. However, Formalose, a

gelatine material containing glycerine to keepit flexible (Wihr, 1977), is useful for repro-ducing the interior shape of an object with anarrow neck where plaster or rubber cannotbe introduced. The Formalose can be pouredhot into a plaster mould and, on cooling, itsets and begins to shrink uniformly. A watchmust be kept and when there is a gapbetween the Formalose and the plaster, repre-senting the thickness of the glass vessel to bereproduced, resin can be cast into the gap.The Formalose core must be supported awayfrom the plaster walls. The method isdescribed by Petermann (1969).

Plaster of Paris

Plaster of Paris (calcium sulphate) is preparedby heating gypsum (CaSO4.2H2O) to drive off some of the combined water, forming2CaSO4.H2O. On adding more water, thecalcium sulphate rehydrates, forming inter-locking crystals, which set to a rock-like masswith very slight expansion, typically 0.5 percent. Various grades of plaster are availablewith different setting times, expansions andparticle fineness. Dental plasters are availablewhich set to form very hard solids withminimal expansion. The material is cheap andis a useful product for the construction ofcase-moulds over silicone rubber. However, itrequires release agents between plaster-to-plaster surfaces and plaster-to-resin ones. It isrigid and hence any undercuts on the glassmust be moulded separately, preferably insilicone rubber. If incorrectly placed, it is diffi-cult to remove without causing damage.

Release agents

Release agents must prevent adhesionbetween objects, moulds and casts; the agentchosen will depend upon the materials beingused. As previously mentioned, silicone rubberonly requires the use of a release agent suchas petroleum jelly or organic lacquer when itis being cast against a section of cured siliconerubber. The surface of plaster of Paris mouldpieces, however, must be sealed as each ismade to prevent it adhering to the adjacentpieces; and release agents must be applied tofacilitate removal of resin if it is cast directlyinto a plaster mould.


Shellac in a solution of industrial methylatedspirit (IMS; US: grain alcohol), or a solution ofPVAL, can be used to seal the surface of dryplaster of Paris mould-pieces, after which thesurface is coated with soft soap, Vaseline,detergent, petroleum jelly or a wax emulsion.If the resin is to be cast directly into plastermoulds, a specially formulated release agentsuch as Scopas (PVAL) (supplied by Tiranti)must be applied to the plaster surface. Releaseagents may be supplied as colourless orcoloured, in order that it can be seen whenall the surfaces have been covered adequately.However, traces of colour may remain on thecast. It is preferable therefore to use a colour-less release agent and to take care in its appli-cation.

Silicone release agents are available asliquids or in spray cans, but are not recom-mended for use in glass conservation sincetraces of silicone oils will remain on the cast,and if they are not removed completely theycan prevent paint or adhesives bondingproperly to the surface. The same is true forpolytetrafluoroethane (PTFE) dispersions inaerosol cans.

Casting materials

The requirements for casting materials for usewith glass are very severe. It should be possi-ble to pour the material into moulds; it shouldset with minimal shrinkage to form a hardsolid; and it should be crystal clear and remaincolourless indefinitely. No materials meet thisspecification completely, although some comeclose to it.

Polyester resins

Many clear polyester embedding resins areavailable, but it is obviously outside the scopeof this book to discuss them all. Suffice it tosay that before restoration work begins, theresin to be used should be tested by mixingand casting it into a mould of the same mater-ial to be used in the final reconstruction. Thiswill ensure that the resin’s shelf-life has notexpired and its performance is as expected(i.e. the formula of the product has not beenaltered by the manufacturer since it was lastordered). Because of their high mobility,

polyester resins are normally cast into closedmoulds. Small quantities of resins may beaccurately weighed on a digital balance, ordispensed from a graduated disposable syringe(minus the needle), provided that care is takennot to introduce air bubbles with the resin.

Where polyester resin is applied as a flatsurface against a one-sided mould, it must beapplied in thin layers, each layer beingallowed to gel in turn. However, subsequentlayers of resin should be applied before eachhas fully cured, in order to prevent the forma-tion of visible interfaces in the cast. If the gapto be cast is very thin, the cast may bestrengthened, by incorporating a layer of fineglass-fibre surfacing tissue in the resin as it iscast. However, the fibre filaments will remainslightly visible within the resin (see Figure7.31). Clear polyester resins may be colouredwithout noticeably altering their transparencyby the addition of minute quantities of translu-cent polyester pigments.

Disadvantages in the use of polyester resinsare the shrinkage of 8 per cent during curing(though this can partly be compensated for bytopping up the mould as the resin polymer-izes), the emission of styrene for some consid-erable time after curing, and the fact that theresin surface often remains tacky for sometime. Reasons for this latter phenomenon areinterference from atmospheric moisturecausing the cessation of chain-building mecha-nisms during polymerization, ageing of thehardener, or, if the resin and hardener arestored under refrigeration, their use beforehaving reached room temperature, thusinhibiting a complete chemical reaction fromoccurring. Hardening of the surface may beaided by polymerizing the cast in a dry atmos-phere, for example, in a sealed cabinetcontaining trays of silica gel. Warming the castin an oven is not recommended since it maycause premature discoloration of the resin.Polyester embedding resins abrade and polisheasily.

Polymethyl methacrylate resins

Plastogen G with Lumopal hardener, used byWihr (1963) and Errett (1972) is transparentand mobile and is therefore normally cast intoclosed moulds (see above). The liquid resin ismixed with 0.25–0.5 per cent hardener


(powder) that is difficult to assess in smallquantities, and the addition of too muchhardener may cause premature discolorationof the resin. Plastogen G has a 15-minute pot-life but cannot be worked during that timebecause a skin quickly forms over the surface;it also has an extremely powerful, unpleasantsmell. After mixing, and if necessary colour-ing, the resin should be covered and left tostand whilst air bubbles escape.

The methacrylate Technovit 4004A istranslucent and therefore can only be used asa gap-filler on opaque glass. The polymer(powder) is mixed with the liquid monomerin the ratio of five parts to three, but theproportions are not critical and the setting-timemay be varied, by changing the amount ofpowder. When mixed in the recommendedproportions, Technovit 4004A sets at roomtemperature in 10–15 mm, but can be workedwith a spatula during this time. This productis guaranteed by the manufacturer not toshrink or expand on curing. It emits heat ifcast in large amounts. Technovit adheres wellto glass, is relatively hard and can be abradedand polished. This particular grade ofTechnovit is no longer manufactured (Koob,2000), but other grades are available.

Epoxy resins

Epoxy resins can be used to gap-fill losses inglass. They can be pigmented with pre-mixedcolours in epoxy paste (although these areusually opaque), or with dyes. In general,epoxy resins do not respond well to polishingafter abrasion. For large areas of restoration,polyester resins are a cheaper alternative.

Colouring materials

Materials used for colouring fall into twocategories: those used for mixing in with resinand those applied to the resin after it hascured. Colours for mixing in with the resinmay be transparent or opaque dependingupon the desired effect; enough coloured resinmust be produced to complete the restorationand allowance must be made for areas whichmay have to be cast more than once, or mayhave to be made good. This is important sincea colour can rarely be exactly matched a

second or third time. Hardener is then addedto small amounts of the coloured resin asrequired for use, and the resin is allowed tostand before use to enable air bubbles toescape before being introduced into a mould.Casts can be made colourless and thencoloured by hand or by air-brushing withpigments and media. This may have theadvantage of being able to remove and re-apply colour, provided that the retouchingmaterials can be removed from the castwithout spoiling its surface.

Most glass on display in museum cases isexposed to high levels of illumination andhence light-resistant pigments are needed inany restored portions. Improvements inpigment technology have provided the conser-vator with a fairly wide palette of colours. Alist of light-resistant pigments is given byThomson (1998). Other pigments are used byWihr (1963), Errett (1972) and Staude (1972),and Davison (1998). Occasionally, pigmentscan produce adverse effects with somereactive polymers and hence it is then moresatisfactory to purchase ready-mixed colours.There are colours for polyesters, silicones andfor epoxies; their light stability must bechecked before use. Transparent epoxy andpolyester resins used as adhesive and castingmaterials often require the use of dyes ratherthan pigments to colour them, in order toretain their transparency. The range of light-stable dyes available for use in polymers islimited (Horie, 1987).

Retouching lacquers

The lacquers usually employed for retouchingthe restored portions of glass objects are fre-quently those used in the restoration of cera-mics, porcelain in particular. Transparency andretention of colour are important. In the UK,Rustin’s Clearglaze (a urea-formaldehyde/melamine-formaldehyde mixture catalysedwith butanol normal/sulphuric acid) is usedfor this purpose. It forms a hard clear coatingwhen catalysed, and has reasonable colourretention. However, this type of polymer isalmost unaffected by solvents (exceptdichloromethane), and it would be most diffi-cult to remove from a glass surface, or evenfrom a resin cast without spoiling it.


Enamelled decoration may be copied onresin casts using any pigments and media usedfor ceramic restoration, provided that theyadhere to the cast and the solvent does notdamage it. Gold decoration may be copied inleaf gold applied on a size, or as liquid metal-lic paints, though the latter will probablydiscolour on aging.

Lifting materials

Materials for the removal of artefacts fromarchaeological sites are discussed byWatkinson and Neal (1998). See alsopolyurethane foam below.

Polyethylene glycol

Polyethylene glycols (PEG) are widely recom-mended for providing reinforcement andbulking during removal of objects such aswood from the ground. Various grades areavailable, but it is the harder, higher molecu-lar weight polymers (4000 and 6000) that areused for this purpose. PEG will shrink slightlyon freezing, and also on cooling to roomtemperature. PEG can be removed by re-melting it, or cleaning the artefact in solventssuch as water, alcohols, trichloroethane ortoluene. There is only one reference to theimpregnation of archaeological glass withPEG, where it was used to provide mechani-cal strength to the remains of severely deteri-orated medieval potash glass alembics(Bimson and Werner, 1971; see Chapter 7).

Polyvinyl chloride

Vinyl polymers have already been discussed,but the use of polyvinyl chloride in the fieldshould be mentioned here. A polyvinylchloride solution, Derm-O-Plast SG, has beenused during excavation, for consolidating boththe soil and fragile objects (Wihr, 1977). Suchtreatment should be considered irreversible,and may have a long-term deleterious effecton the consolidated object. The product hasnow been discontinued but a new product(Archaeo-Derm), presumably of the same type,has been substituted (Filoform). PVC is one ofthe most unstable polymers in commercialproduction and should not be used in contact

with artefacts. Small traces remaining aftercleaning will degrade, discolour, releasehydrochloric acid and probably become insol-uble (Watkinson and Neal, 1998).

Flexible polyurethane foam

Flexible polyurethane foams are relativelyexpensive, but may be convenient materials touse, for lifting and for packaging excavatedartefacts and fragile structural remains (e.g.glass kilns, Price, 1992), owing to their light-ness. Polyurethane foams are formed bymixing an isocyanate (generally toluene di-isocyanate), with a polyhydroxyl component(a polyester or a polyether), each of which isavailable in a range of formulations. This inturn creates a range of products of varyingproperties, for example, flexibility and stabil-ity. The greatest variability is the polyolcomponent, which may be polyester orpolyether. Various catalysts such as aminesand tin compounds are incorporated in thecomponents; these may increase the degrada-tion of the foam and its effects on objects. Thefoaming action itself is frequently achieved by adding small amounts of water, which react with the isocyanate to produce carbondioxide. The adhesion of foamed poly-urethanes is similar to that of the poly-urethanes used for coatings; their resistance tosolvents is greater, but they are less stable onageing (Moncrieff, 1971; Watkinson and Leigh,1978).

On mixing, the components expand to formfoam within 3 minutes, the exact time beingdependent upon the prevailing temperature.All the isocyanates are very toxic but thesmaller and more volatile compounds are themost dangerous because of the ease of breath-ing in the vapour. For example, toxic fumes(TLV, 0.02 ppm) are given off by theisocyanate component in polyurethane foam


(5.14)

and this can survive unchanged in thecompleted foam. Thus extreme care must betaken when using products based onpolyurethane. The flexible foams are resilientopen cell structures. In general, the polyether-derived foam is preferred for packaging,having better cushioning properties and beingless likely to degrade. However, bothpolyester and polyether foams will graduallydeteriorate on ageing, especially whenexposed to light; polyester foams in particularcan disintegrate, forming a sticky powder.

Packaging materials

Flexible food-wrapping films are oftensuggested for keeping objects wet duringstorage. Such products as Clingfilm (US:Saranwrap) are polyvinyl chloride films heavilyplasticized with a material such as dioctyladi-pate, and are prone to degradation especiallyif exposed to light. Polyvinyl chloride filmswill retain water to a certain extent, but therate of water diffusion through them is greaterthan through polyethylene sheeting of thesame thickness. For storage periods longerthan a few days therefore, heavy gaugepolyethylene sheeting, heat-sealed around theedges, would provide better protection forobjects.

Self-adhesive (pressure-sensitive) tapes arewidely used in conservation for packagingpurposes and do not present any problems.Where greater strength is required, plastic tapeor string may be used. However, certain diffi-culties may be encountered in the use of self-adhesive tapes for temporarily holdingfragments of glass in place whilst adhesivescure. The tapes are coated with a tackysubstance, which is in reality an extremelyviscous liquid. In warm conditions this canflow into the pores of an object, and mayincrease the adhesion of the tape to thesurface of objects if the tapes are left inposition for longer than necessary. Thus onremoval the tapes may lift fragments of glasswith it and traces of the adhesive may remainin the object. Tape alone should never beused for supporting fragments of glass duringstorage, since it will degrade to a brittle orsticky mass with some resulting damage toobjects. If there is any risk of endangering the

object, solvents should be used to soften theadhesive before attempting to remove thetape. The tapes use a natural rubber, orsimilar, contact adhesive, which can degrademore or less rapidly to form an intractablebrown mass. As it may well be impossible toremove the last traces of the adhesive fromthis type of self-adhesive tape, it would bewise to use only those tapes which use a morestable adhesive, such as an acrylic, which havebeen introduced in recent years.

Nevertheless, efforts should always bemade, by using solvents, to remove the lasttraces of any adhesive which remain afterstripping the tape from the object. A usefulmixture for removing self-adhesive tape fromantiquities is made up as follows: 5 ml toluene,5 ml. 1,1,1-trichloroethane, 1 ml concentratedammonia and 10 ml industrial methylatedspirits. Fragile objects will need to be packedin acid-free tissue (never cotton wool) and/orinert foam inside strong cardboard boxes.Paper used for storage must be long-lasting,and as such will normally consist of highlypurified cellulose, perhaps buffered to ensurethat the paper is acid-free. The term acid-freemeans only that the paper is currently non-acidic, and is no guarantee that deteriorationwill not produce acids or other harmfulemissions. However, for the majority of chemi-cally insensitive materials higher specificationthan acid-free is unnecessary.

Plastazote and Ethazote are chemically inertpolyethylene foams composed of polyethylene(sometimes copolymerized with vinyl acetate),which has been blown into foamed sheets.The foam is made of closed cells, which resultin increased stiffness and resistance tocompression. It is available in a range of thick-nesses, split or laminated from the standardsheet size; and is easily cut with a sharp knife.

Cardboard boxes are made of compressedpaper pulp (usually recycled), which is madeinto board of various thicknesses by buildingup layers of thinner plies, and frequently facedwith brown paper. The boards commonlyused range from 1.0 to 2.2 mm thick. Theboard is bent to shape and fixed by staples(stitches), brass being preferable to stainlesssteel. On ageing, fibre-board oxidizes andweakens, a process which is accelerated indamp conditions since it is permeable andabsorbs water vapour. The lifespan of a good


fibre-board box in good storage conditionsseems to be in excess of twenty years. Inpoorer conditions and over longer periodsthere is increased danger of the box collaps-ing. Correx board, a white, ethylene propylene

copolymer extrusion, can be used to formmounts and boxes to store and display glassobjects. It is similar in appearance to corru-gated cardboard, but unlike card, is chemicallyinert (Navarro, 1999; Lindsey, 2000).


One of the most crucial stages of the conser-vation process is the initial examination of anobject. Only after examination will it be possi-ble to determine the condition of an object, tosuggest whether conservation is actuallynecessary, and if so, the most appropriateapproach. The processes of observing anddocumenting observations and actions, should,in theory, be commenced at the same time.

Recording and documentation

The primary purpose of making records is toensure that the information gained fromexamination and during the processes ofconservation is not lost. The information willbe valuable to the interpretation of the object,will provide data against which the conditionof the object can be monitored, and will beavailable for reference if the object has to bere-treated at a future date.

Condition/conservation reports shouldcontain the following information:

(i) The condition of the object (appearanceand strength of the surface and body;number and type of breaks and conditionof the break edges; presence of solubleor insoluble salts; stains and concretions;and the effects of any previous treat-ment).

(ii) Archaeological evidence (type of glass,provenance and approximate date ofmanufacture; ancient repair; function/evidence of use, and possibly contents).

(iii) Technical evidence (method of manufac-ture and decoration; evidence of the use

of tools, moulds etc.; inclusions such asair bubbles).

(iv) Diagnosis of condition and suggestedconservation treatment.

(v) Factors other than condition which mayaffect the choice of treatment. (What theobject is needed for; level of conserva-tion, e.g. for stabilization to display;future reversibility of the treatment ifpossible; different approaches to theconservation/restoration.)

(vi) Further evidence which may becomeapparent during the course of treatment,and which may cause the treatment to bemodified.

(vii) Measured drawings and sketches; photo-graphs; X-radiographs; analysis resultswhere appropriate.

For larger institutions, the future of documen-tation lies with information technology andcomputerization, but such an approach maybe inappropriate for small conservation labora-tories or some countries. Notebooks(sometimes referred to as Day Books) can beused for noting brief observations, remindersand drawings, which can then be transferredto a standard condition or conservation reportform which allows a more systematicapproach to recording to take place. Theforms could be enclosed in an envelope,which would bear the basic details on theoutside. This approach would enable reports,photographs, X-radiographs and small samplesto be kept together.

In order to extend the lifespan of recordforms or cards, they should be made ofarchival quality paper and inks, particularly in

227

6

Examination of glass, recording anddocumentation

hot and humid climates, and filled in by theconservator using permanent ink, also ofarchival quality. The format should include asection for standard information whichdescribes, identifies and locates the object.

The majority of the record document willcontain a description of the conservationprocesses, and should include informationabout the techniques and materials used,giving both the proprietary names (tradenames) and chemical names of materials, e.g.Araldite 2020 (epoxy) resin. Whilst drawingsmay help to give more information, photogra-phy has the potential to be more accurate.Ideally, several photographs should be taken,before, during and after conservation. Thechoice of whether to use black and white orcolour photography or transparencies is deter-mined by personal preference, cost or organi-zation policy. A scale of measurement (inmillimetres/centimetres) and an identificationshould normally be included in every photo-graph. Some glass objects will need to bephotographed by transmitted light in additionto reflected light, or in a raking light whichwill highlight surface details.

During work on the site of an archaeologi-cal excavation, any first aid or emergencytreatment given to preserve objects should beaccurately documented. The records willprovide clear information for conservatorscarrying out future treatments in a laboratory.

Condition surveys of collections will requirespecially formulated record forms, which inaddition to recording the condition of objects,can be used to accompany loans of objects forexhibitions.

Basic examination techniques

The majority of objects a glass conservator willencounter routinely will not require morespecialized techniques or equipment thanthose described in the following section.

Visual examination

Visual examination is the first stage in examin-ing an object when it arrives in a conservationlaboratory; and in order to be reliable it shouldbe carried out in good lighting conditions. Araking light (light directed at the surface at an

angle almost parallel to the surface plane) willhighlight surface irregularities and textures.Placing transparent glass on a light box willreveal irregularities within the glass itself.Besides observing the nature of any damageor deterioration, visual examination mayidentify the type and colour of the glass andany surface decoration. These physical charac-teristics relate to the nature of the glass, andwill affect the choice of treatment.

Simple magnificationA simple hand-held magnifier can extend thepowers of the naked eye up to ten times,depending on the magnification of the individ-ual lens. Other useful aids are a magnifier,which fits against the eye, or a jeweller’s loop,which attaches to spectacles.

Optical microscopyIt is possible to obtain more detailed informa-tion by viewing an object through a micro-scope. Microscopes range from the simplebinocular types, which produce a stereoscopic(3-dimensional image), which can be used atthe bench, to much more sophisticated instru-ments with zoom facilities, fibre-optic lightingand attachments for cameras.

The examination process

The glass surface and the nature ofobscuring deposits

As is the case with much excavated material,glass will normally be found covered in a layerof obscuring deposits (Figure 6.1). These arenormally removed mechanically or withdistilled water. If the deposits are resistant, itwill be necessary to identify them in order todetermine the appropriate treatment for theirremoval (see, for example, the tests for deter-mining insoluble salts below). Where possible,a small sample is removed for testing. Beforeits removal, material adhering to the glassshould be examined as it may reveal thepresence of other materials in association withthe glass. For example, in the case of glassbeads, there may be traces of fibres from thecords on which they were threaded (see Plate8), or of metal rods on which they werewound during manufacture. The presence ofmetal corrosion products on glass (or enamel)


will indicate that it had been had been set intoa metal base, or that it had been buriedadjacent to metal artefacts.

As the deposits are cleared away, a searchmust be made for any applied decoration,such as gilding or painting, to avoid itsremoval. As work progresses, more will berevealed of the decoration (if present) andstate of deterioration of the glass. Finally, theshape and colour of the object will berevealed, and details of its method of manufac-ture sought. If, at any stage, an item of inter-est is uncovered, it should be photographed.The dimensions of sherds and thicknesses ofvessel walls should be measured, immediatelyin the case of heavily decayed glass, in caseit flakes or crumbles without warning. Bearingin mind that it may be necessary to analysesoil found within a hollow glass object, todetermine whether there are traces of originalcontents present, soil samples should be takenfor further examination or analysis.

The glass bodyBy viewing glass by transmitted light (e.g. ona photographic light box), it will be seen to beeither translucent or opaque, clear or coloured.If the glass is opaque, it should be determinedwhether this was the original condition, orwhether it is the result of deterioration. If theglass is transparent, it may be clear andhomogeneous, or it may contain inhomo-geneities such as air bubbles (seed) (Figure6.2), striae and cords (streaks in the glasswhich have a different refractive index fromthe body of the glass), or opaque inclusions(stones) (see Figure 4.1). Viewing the objectby means of transmitted light may be helpfulin distinguishing between glass and hardgemstones, e.g. if air bubbles are seen to bepresent). The glass may be coloured through-out (colour having been added to the batchduring manufacture) or flashed (a thin appli-cation of coloured glass to clear glass,sometimes found in window glass) or cased (athick application of one or more coloured glassgathered over another, which when cold, werecarved to varying depths to produce a design).

If the furnace temperature had been insuf-ficient to melt all the quartz grains (sand) intoglass, microscopic examination of the materialwould reveal anisotropic quartz grains embed-ded in an isotropic base (i.e. crystals inamorphous glass). When viewed in a polar-iscope, quartz grains are pleiochroic, i.e.display different colours according to theangles in which they are viewed.

Examination of glass, recording and documentation 229

Figure 6.1 Obscuring mass of soil adhering toexcavated glass.

Figure 6.2 Inhomogeneities in glass seen bytransmitted light from a photographic light box: airbubbles which cross the breaks in glass fragments, andwhich can be an aid to confirming the accuracy of joinsduring conservation.

Indications of fabrication techniquesClose examination of a glass artefact will oftenreveal details of its manufacture. There aremany types of information which may berevealed during examination and initial clean-ing; archaeologists and conservators must becontinually aware of the possibilities.Examples of the evidence to look for are:order of manufacture, for example, handlesetc. added after blowing, flash lines (seammarks) resulting from mould blowing or press-ing; jagged broken marks on the bases ofvessels, left by a punty iron (punty mark); andcircular grinding marks on vessels groundfrom a casting or from a solid block of glass,especially on the interior, where less troublewould be taken to polish the marks away. Ifthe article was cast, or pressed into a mould,there may be flow lines where the glassflowed into projections of the mould (such asa rimmed base) or into a corner. These flowlines can be identified by the lines of seed, orthe elongated shape of the air bubbles, orlines of striae, which follow them (Bimson andWerner, 1964b). If seed are present, they maybe spherical (showing that the glass wasundisturbed as it cooled), or elongated(showing that they were stretched as the glassflowed while it was being manipulated whenhot to form the artefact). If the air bubbles arelarge and distort the glass surface, they arecalled blisters; and if broken, weathering ofthe glass may have taken place inside them.Flat bubbles can also occur between the layersof successive gathers. Striae and cords can beobserved by holding the glass against abrightly lit area with a dark edge (such as alight box), and moving it back and forth acrossthe edge, or by using a cord-detector.

Differentiating glass from othertransparent materialsBeads made of crystalline quartz, calcite orfluorite are often mistaken for glass beadseven though they have quite a different chemi-cal composition and can be distinguished froma true glass (for other materials which can bemistaken for glass beads, see Chapter 7).Quartz is harder than glass and twice as refrac-tive; calcite, although also twice as refractive,is much softer and has a hardness of onlythree, compared with glass at six; fluorite hasa hardness of four but, unlike quartz and

calcite, is isotropic. There are also differencesin specific gravity but these tests should beused only as a rough guide, and the exactcomposition of a particular glass is not easyto determine.

Evidence of previous repair andrestorationArchaeological glass that has not come directlyfrom an excavation may have undergone previ-ous repair or restoration. There may have beenattempts to disguise the work by the applica-tion of adhesive, mud and glass flakes; or anoutright attempt at forgery (see Chapter 7).

Historical glass is even more likely to haveundergone repair and restoration. If the workhas not been carried out by a conservator,there are not usually any records of the work.One of the advantages of glass being trans-parent is that the application of restorationmaterials is normally visible. However, as thiscannot be taken for granted, the glass must becarefully examined.

Use of a hand-held metal detector (such asthose used by electricians to detect hiddenwiring), can assist in revealing metal compo-nents such as wire, rivets and dowels, whichmay have been incorporated in previousrestorations. This may be especially usefulwhen examining previously restored archaeo-logical glass, which has been heavily coveredwith mud and glass flakes in an attempt todisguise the restoration, and heavily over-restored historical glass.

Ultraviolet (UV) lightUltraviolet (UV) light provided by a hand-heldlamp in a dark room can reveal syntheticmaterials used to restore glass and enamels,and forgeries, since different materialsfluoresce slightly differently (Goldstein, 1977;Bly, 1986). A trained eye may be required todetermine the differences in fluorescence,particularly if it is being used to determine orcompare glass compositions (Brain, 1999).

X-radiography (X-rays)X-rays can be useful for detecting metalcomponents in previous restorations, themarrying together of pieces from differentobjects or other types of fakes and forger-ies. Air bubbles, striations, inclusions, stresspoints may become visible and assist in deter-


mining manufacturing techniques. Although X-radiography is normally outside the scope of small laboratories, conservators may makearrangements with extra mural sources havingX-ray facilities, such as a museum, universityor hospital.

Infra-red (IR) photographyInfra-red photography can be useful for reveal-ing indistinct surface decoration such asgilding or unfired paint, obscured by overly-ing material which is difficult to removewithout causing damage; or gold decorationwhich has virtually disappeared (leaving aghost of a pattern on the glass surface).

Universal indicator papersUniversal indicator papers can be used todetermine the pH (alkalinity or acidity) of thesurface of a glass, its immediate environment(e.g. soil) or of a liquid, which it is proposedto use in the conservation process. Forexample, some fungicides are highly alkaline.Electrically operated pH meters are available,but the use of universal indicator papers issimple and adequate for conservationpurposes.

Strongly alkaline solutions (those with a pHvalue of 9.0 or greater) will eventually causethe breakdown of most glasses and certainlyall ancient glasses. Similarly, acid solutions(those with a pH value of 5.0 or less) willattack the high-lime medieval glasses such asH and J in Figure 4.18. It is therefore usefulto have a simple test with which to measurethe pH. If an historic glass is found to have adamp, greasy surface, it is in a sweating orweeping condition, and the pH (alkalinity) ofthe surface liquid can be tested.

Universal indicator papers are available inthree types: the full-range paper, where pH 1shows as a red colour, through shades oforange to pH 7 (a greenish yellow) to a deepblue colour at pH 14; the narrow-range paperfor acid solutions, where pH 4.0 is yellow andpH 5 is blue; and the narrow-range paper foralkaline solutions where pH 7.0 is green andpH 9.0 is dark blue. A small piece of indica-tor paper is torn off, moistened with the liquidto be tested, and laid on a white tile; after 30seconds the colour is compared with thecolour chart supplied with the indicatorpaper.

Methods of examination for research

Measurement of densityA simple test for identifying lead-containingglasses is to measure the density. Lead oxidehas a pronounced effect on the density of glass(see Figure 1.8, which shows how density isrelated to lead content). Lead glasses made byRavenscroft, have a density of about 3150 kg m3,compared with 2460 kg m3 for soda–lime–silicaglass, 3580 kg m3 for opaque Bristol blue glass;but some lead glass beads can have a densityof 6000 kg m3. The density of an object can bedetermined, by weighing it first in air and thenin water. The loss of weight of the glass inwater (weight in air – weight in water) isapproximately equal to the volume of the glassin ml and hence the density = weight in air (g)divided by the volume (ml). Further accuracycan be maintained by making corrections forthe density of the air and the temperature ofthe water, but this is rarely worthwhile forancient glasses. The presence of an air-twiststem, a tear stem, a hollow stem, or even manyseed in the glass, will invalidate the result (theapparent density will be less than the true oneowing to the presence of the air inside theglass). If the glass is in small pieces, or if afragment can be obtained, the most accuratedeterminations are made by the sink–floatmethod (Scholze, 1977).

Some Chinese glasses contain substantialamounts of barium, which has a marked effecton their density. Care must be taken to distin-guish between lead- and barium-containingglass.

Measurement of refractive index (RI)The refractive index (RI) and the density ofglass are closely correlated, but differentrelationships apply to different types of glass.The RI. and density of the glass can be calcu-lated by various empirical formulae, from theanalysed composition of the glass. However,the simplest procedure is to refer to the graphin Figure 1.11, in which it can be seen thatthe lead crystal glasses (with potash as thealkali), lie on a line at the right-hand side ofthe graph, the average value of C being 0.19.The weight percentage of PbO is indicated onthe line.

Soda–lime–silica glasses lie on a muchsteeper line to the left of the diagram, with an


average value of C = 0.21, and it connects thepoint for fused silica (at the bottom) with aglass containing 60 per cent SiO2, 20 per centCaO and 10 per cent Na2O (at the top), inter-mediate values of CaO being spaced, more orless equidistantly, on the line. (C is a constantand differs for each type (composition) ofglass; Huggins and Sun, 1946).

Ancient glasses lie between the two lines;Roman glass lies at the point R (70 per centSiO2, 10 per cent CaO, 20 per cent Na2O), butmedieval glasses fall within the shaded area,because the MgO, A12O3 and P2O3 which theycontain have more effect on the density thanon the RI. It is interesting to note that quartz,which has the same chemical composition asfused silica, falls in quite a different positionon the graph (at Q) because the crystallinematerial is more dense, and has a much higherRI than the amorphous (fused) silica. The RIof a glass object is of importance when choos-ing materials for conserving glass (Tennentand Townsend, 1984a,b) (see Figure 5.1).

Examination by immersion in liquidsIt is often difficult to see details inside a pieceof glass, even when the glass is transparent,because (i) the surface may be scratched,abraded, etched or sand-blasted, (ii) may becurved so that it acts as a lens and distorts thedetail, or (iii) there may be reflections fromsurface features such as cut glass designs. Thevisual effects of such surface features can beremoved, by immersing the glass object in aliquid that has the same or close RI. If the RImatch is exact, the surface of the glass seemsto disappear, and its internal features, such asseed, striae and stones become easily visible.

The RI of soda–lime–silica glasses lies in therange 1.51–1.52, whereas full-lead crystal glasshas an RI of 1.565 (see Figure 1.11). The RIof a glass of known composition can be calcu-lated (Gilard and Dubrul, 1937) and it can beseen that high-lime glasses will also have ahigh RI. The poorly durable medieval glassesat the bottom of the triangular diagram inFigure 4.18 are glasses of that type, and thosehaving a composition of 50 per cent SiO2, 15per cent K20 and 35 per cent CaO, would havean RI of 1.59 and a density of 3.1, that is, atpoint H in Figure 1.11.

Convenient immersion liquids, and their RIvalues, are: xylene (1.490), chlorbenzene

(1.525), nitrobenzene (1.553), aniline (NB acarcinogen) (1.586) and quinoline (1.624).Bimson and Werner (1964b) used the immer-sion technique to prove that one of the minuteheads on the Tara Brooch (found in CountyWexford, Ireland) was in fact glass rather thana carved gemstone. Immersion in toluene (RI= 1.496) revealed the presence of air bubbles,and characteristic markings showed that theheads had been made by casting the glass intomoulds. The detection of striae is greatlyaffected by the orientation of the sample. Forexample, none may be seen when lookingthrough a piece of window glass, yet they arequite obvious when looking through the edgeof the glass.

Detection of strainIf glass contains numerous cords and striae, orhas been badly annealed, it will exhibit signsof strain, which can be detected by means ofa strain-viewer (Werner et al., 1975). Badlyannealed ancient glass will not usually beencountered since it is unlikely to havesurvived, but there may be occasion to lookfor the presence of strain-producing inhomo-geneities. Glass subjected to strain (bending orstretching) becomes bi-refringent, that is, itacquires two different refractive indices. Thusa single ray of light entering the glass willemerge as two separate rays. This phenome-non can be used to detect, and to measure,the frozen strains resulting from unsatisfactoryannealing. Briefly, when plane polarized lightpasses through a material of a bi-refringentnature, it becomes elliptically polarized, andhence strain-viewers make use of polarizedlight produced either by reflection from apolished metal surface, or by the use ofPolaroid filters. There are various commercialinstruments available; or a strain-viewer can bemade, in which the polarization of light isproduced by two pieces of black plate glass,placed at Brewster’s angle (56.50).

Fracture analysisFracture analysis is a technique that enablesthe origin of a fracture in glass to be identi-fied by showing the directions in which thefracture propagated itself. The technique maybe useful in the study of well-preserved glass.Fractures which occur at speed (experi-mentally) are easy to study, but those which


occur perhaps as slowly as a few millimetresper century are more difficult to understand.In the great majority of cases the break startsat a single point and the fractures spread outfrom this point of origin. If the glass is brokeninto a number of fragments the smallest onesare closest to the point of origin. A fracturearises from a tensile stress in the glass andhence it starts at right angles to the stresswhich produced it. However, if the stress isgreat the crack, once started, may fork so thatthere is an acute angle between the two newbranches. Cracks rarely join up and hencethese acute-angled forks point towards theorigin of the cracks.

At the actual origin of the crack the brokenedge bears a characteristic mirror area which(see Figure 1.13) is surrounded by grey areas,hackle marks and, finally, rib marks whichmay extend a long distance from the originand indicate the direction in which the fracturetravelled. The rib marks are places where thefracture has hesitated briefly before continuingits start–stop advance, and they represent theleading edge of the fracture at successivemoments. The rib marks are always curvedand present their convex face to the directionin which they were formed. They are thusimportant for determining the direction inwhich the crack was growing. The investiga-tor simply follows the reverse direction to thepoint at which the rib marks point the otherway. When glass breaks due to excessive localheating, the rib marks are spaced well aparton the cold side and they crowd together onthe hot side. The origin of the fracture isalways at the surface of the glass and the ribmarks generally face the same side as theorigin of the fracture. If the outside of a vesselhas been given a sharp blow, the area whichreceives the blow may be crushed (causing abruise consisting of powdered glass), with asurrounding stressed ring from which a familyof cracks start. This ring of cracks forms animpact cone, which may separate from theglass in the form of a plug (see Figures1.12 and 1.13).

Another feature of the fractured surface isthe size of the mirror area. The energy of afracture is directly related to the diameter ofthe mirror area; for example, a mirror diame-ter of 0.1 mm corresponds to a stress of about200 MNm–2 (30000 psi), whereas one of

2.5 mm corresponds to only some 40 MNm–2

(6000 psi). Discussions of fracture analysis canbe found in Ernsberger (1977) and Scholze(1977).

Instrumental methods of analysing glassfor scientific research

In addition to physical and chemical methodsof examination, a large and ever increasingnumber of instrumental methods used in otherareas of scientific analysis is being applied tothe examination of the surface and bulkcompositions of ancient and historic glassobjects. The composition of a weathered glasssurface and the body of a glass differ substan-tially as a result of the leaching processes,which have occurred over time, and thereforeit will be necessary to remove surface mater-ial in order to obtain a true sample of the glassbody. Through the application of instrumentalmethods of analysis, information can beobtained concerning condition, provenanceand fabrication technology. Analysis deter-mines (i) the reduced composition of theglasses, i.e. the main elements which charac-terize them: the oxides of silicon, sodium,potassium, calcium, iron, magnesium andaluminium; (ii) the additives, i.e. opacifiersand colourants: the metallic oxides; and (iii)the trace elements which indicate differencesin raw materials from different sources: oxidesof boron, titanium, manganese, antimony andsulphur.

A brief introduction to the instrumentalmethods of analysis is given below, stating thepurposes which they serve. They can beconveniently grouped as (i) electron beammethods, (ii) X-ray methods, optical spectro-metric methods, (iii) mass spectrometricmethods, and (iv) oxidation state methods.Full details can be found in Mass (1999).Before scientific methods of analysis can beused with confidence, it must be certain thatthe results of analyses in one laboratory willbe essentially the same as those from the sameglass from another laboratory. Otherwise,observed differences in composition etc. couldbe due either to true compositional differ-ences, or to those inherent in the particularanalytical procedures of the laboratoriesconcerned, the accuracy of which will dependvery much upon the expertise of the scientist


undertaking the work. It is for this reason thatanalyses of glasses have not been included inthis volume.

The establishment of standard referenceglasses is therefore important in order toachieve any meaningful results (particularly asglasses have such a complex nature).

A conservator will normally be party to theidentification of a need for analysis, and maysupply the samples for analysis. This is partic-ularly the case if glass is fragmentary, other-wise a conservation scientist may prepare thesample. (Note that many of the instrumentalmethods of analysis are destructive, in thatthey require a sample, albeit minute, to betaken from the glass.) Very few conservationlaboratories will contain the specialized equip-ment for scientific examination and analysis,but it is often possible to work in collabora-tion with a research laboratory in a museumor university.

Electron beam methodsElectron beam methods of analysis involvemeasuring the signals generated by a beam ofhigh energy electrons impinging on a (glass)sample. The SEM (scanning electron micro-scope) can be used to image the surface ofancient glasses; SEM-EDS or energy dispersiveX-ray analysis, AES or auger emissionspectrometry, and EPMA or electron probe X-ray microanalysis, are all used for the compo-sitional analysis of ancient glass surfaces. TheSEM can be used in conjunction with electronmicroprobe analysis to provide a micro chemi-cal analysis of the area in the field of view ofthe SEM.

Scanning electron microscopy (SEM)Scanning electron microscopy is a techniqueused to image the surface of a (glass) sampleand reveal features that are less than 0.4 μmor 400 nm in diameter). The SEM is usuallyoperated at magnifications between 5000� to100 000�, and it has an extraordinary depthof field. Three-dimensional views, which canbe photographed, can be obtained by tiltingthe sample under magnification. The only(slight) disadvantage of the use of an SEM isthe need to apply a thin and invisible electri-cally conductive coating such as carbon orgold to the glass sample, in order to preventthe build-up and release of electrical charges

which would appear as bright streaks acrossthe image (Adlerborn, 1971).

SEM has been extensively used to study thesurfaces and microstructures of ancient andhistoric glasses: SEM has been used to detectfaked weathering surfaces (Gairola, 1960;Werner et al., 1975); opacifiers and colourantsin glasses (Brundle et al., 1992; Freestone, 1993;Henderson, 1993a; Verita, 1995; Freestone andBimson, 1995; Freestone and Stapleton, 1998;Mass, 1999). Krawczyk-Barsch et al., (1998)used SEM and ion-beam slope-cutting to studythe thickness of weathering on glass. Hogg etal., (1999) drew attention to the possiblechanges in the surface of a specimen, due tothe inevitable dehydration of the surface whichoccurs when a specimen is placed in the highvacuum. The use of an environmental SEM (e-SEM) may overcome these problems.

Energy dispersive X-ray spectrometry (EDS)SEM-EDS has been widely used to study thecomposition of ancient and historic glass:second millennium BC Mesopotamian glass(Vandiver, 1982); pre-Malkata Egyptian glass(Lilyquist and Brill, 1993); thirteenth- tosixteenth-century glass and glass waste inGuildford Museum, Surrey (UK) (Mortimer,1993); early Venetian enamels (Freestone andBimson, 1995); glass recovered from an


Figure 6.3 Scanning electron micrograph ofdeterioration layers on glass (�700).

eighteenth-century shipwreck (Corvaia et al.,1996).

Electron probe microanalysis (EPMA), alsoknown as wavelength dispersive X-rayanalysis (SEM-WDS) or electron microprobeanalysisThe electron microprobe can either be usedwith the SEM (as above) or separately. Theprinciple is similar to that of X-ray fluores-cence (XRF) except that electrons are used asthe exciting radiation instead of X-rays. EMPAis widely used for the chemical analysis ofglass samples as it requires very small samples,and can be applied to broken surfaces, or cutsections of glass (Brill, 1968; Verita, 1985;Henderson, 1988, 1991; Barrera and Velde,1989; Mortimer, 1995; Hartmann et al., 1997;Wedepohl et al., 1997).

It is of great value in interpreting complexchemical situations such as the contents of pitsin a glass surface. It can be seen that air pollu-tion does not attack painted glass becausesulphur-containing compounds are found onlyat the surface of a pit, and not at the bottomwhere corrosion reactions are occurring. Thedisadvantages are that it analyses the surfacelayer only; lighter elements are harder tomeasure than the heavier ones, and alkalielements can be forced deeper into the glassby the charge on the electron beam.

Verita et al. (1994) compared two related tests,using wavelength dispersive systems (WDS) orEMPA and energy dispersive systems (EDS). Itwas concluded that both methods were poorlysensitive to elements lighter than sodium; andthat EDS could be inaccurately interpreted dueto peak overlap. The disadvantage of WDS isthat unless precautions are taken, the higherenergy used with it can cause sodium ions tomigrate away from the test area, resulting in alower result for the element, unless the micro-probe beam is defocused. EDS with its limitedaccount rate capability is regarded as being lessaccurate, but is more readily available than WDS(see also Mass, 1999).

Auger electron spectrometry (AES); augermicroprobe analysisIn analysis by AES, the surface of the glasssample is bombarded in such a way as tocause the emission of Auger electrons from itssurface, by means of which elements within

the glass can be identified (Mass, 1999). Theresults are limited to the first few atomic layersof the surface, so that AES was used to inves-tigate the durability of medieval window glass(Dawson et al., 1978) and by Pollard (1979).

X-ray methodsX-ray methods of analysis are all non-destruc-tive, i.e. do not require a sample to be takenfrom the (glass) objects. The techniques ofanalysis produce signals which result from theinteraction of the atoms in a (glass) samplewith incident X-rays: diffracted X-ray beams(XRD or X-ray diffraction); outer shell electronejection energies (XPS or X-ray photoelec-tron spectrometry, also known as ESCA, elec-tron spectrometry for chemical analysis): andfluorescent X-rays (XRF or X-ray fluorescence).XRD is used to identify crystalline phases inglass; XRF and XPS determine the surfacecompositions of glass. XPS will also identifythe oxidation states of surface atoms in glass.

X-ray diffraction (XRD)X-ray diffraction is frequently used in the studyof ancient glass to identify undissolved rawmaterials, opacifying agents, devitrificationproducts and undissolved raw materials (New-ton and Davison, 1989). When a beam ofdiffracted X-rays is passed through powderedcrystalline material, diffraction patterns areformed which can be detected photographi-cally as a series of curved lines that can beinterpreted as having come from the particu-lar material.

Since it determines crystalline phases, thetechnique can also be used to differentiatebetween glass and semi-precious stones,Egyptian faience or metallurgical slags. XRDhas been used by Dandrige and Wypyski(1992), to identify crystalline opacifying agentspresent in medieval enamels; by Hoffman(1994) in the study of crystalline colourant/opacifiers in Merovingian beads from Germangrave deposits; and McRay et al. (1995) in theidentification of the lead arsenate crystalsresponsible for the colouring of Venetiangirasole glass.

X-ray diffraction can be useful in determin-ing whether a glass object has been repaired,especially if the repair has been obliterated bythe application of weathering products fromelsewhere (Plate 7).


Goldstein (1977) illustrated two examples ofrepairs detected by X-rays; a large amphorawith a tip reconstructed from a moderndropper tube, and an ewer to which a plasticshandle had been added (see Figure 7.49). Thepossibility of darkening of lead glass by theuse of X-rays must be considered.

X-ray fluorescence (XRF)XRF is a non-destructive technique which iswidely used for the chemical analysis ofmaterials because it is rapid and accurate, andcan be carried out on equipment that is avail-able commercially (Brill and Moll, 1961). Asuitably prepared test piece (hence not trulynon-destructive) is irradiated with suitableprimary X-rays, and the sample then emitssecondary X-rays that are characteristic of theelements that make up the sample. Thesecondary X-rays can be detected in two ways:the earlier technique measured theirwavelengths (Hall et al., 1964; Hall andSchweizer, 1973); energy-dispersive XRF testsmeasure their energies, which appear as peakson a video monitor; the positions of the peaksindicate the chemical elements, and the heightsof the peaks measure their concentrations. Acomputer program assists in the interpretationsand makes allowance for the mutual interfer-ences of certain elements. The sample can beaffected by the test; Brill (1968) mentionsdamage by radiation burns, and lead-containingglasses can be darkened unless suitable pre-cautions are taken. Cox and Pollard (1977)showed that older weathered glasses possessedan ion-exchange layer from which alkali ionshad been removed so that the surface layer(even though it looked quite unaltered) had acomposition different from that of the interior;they found it necessary to grind part of thesurface away to a depth of 0.5 mm and thenpolish it smooth. Pollard (1979) used thetechnique extensively in his thesis, and manyother workers have found it invaluable.

X-ray photo-electron spectrometry (XPS),also known as electron spectrometry forchemical analysis (ESCA)This method analyses the surface of glass bymeans of photoelectrons emitted from thesurface when it is bombarded by suitable X-rays. XPS has not been widely used in thestudy of ancient glass. Hench (1975a) used it

to follow the early stages of weatheringprocesses which occur in ancient glass;Lambert et al. (1978) used XPS to identify thecopper oxidation states responsible for theblue colour of ancient Egyptian glass.

Secondary ion mass spectroscopy (SIMS)This is highly specialized, but it has the advan-tage that any element or isotope (includinghydrogen) can be analysed, and the alkali ionsare not driven further into the glass, which isthe case with EMPA (Hench, 1975a).

Infrared reflection spectroscopy (IRRS)This again is restricted to the study of glasssurfaces, and it has been used by Hench et al.(1979) to characterize the weathering ofmedieval glasses, and for predicting the weath-ering behaviour of others.

Particle methodsIn these techniques of analysis, a (glass)sample is placed into a beam or flux of parti-cles and the characteristic radiation resultingfrom the sample–particle interaction ismeasured. NAA (neutron activation analysis)and PIXE (particle induced X-ray emission,using protons instead of neutrons) are used tomeasure the major, minor and trace elementcompositions of bulk glass samples and theirsurfaces respectively.

Particle induced X-ray emission (PIXE)Conservation scientists have frequently usedPIXE to study ancient glasses: Fleming andSwann (1994) studied the production formulaeand colourants used in the manufacture ofRoman onyx glass; McGovern et al. (1991)studied glass beads from mid-second millen-nium BC sites in Iran; Swann et al. (1990) andVandiver et al. (1991) studied glass workshopdebris from Tell el-Amarna in Egypt; Germain-Bonne et al. (1996) investigated the composi-tions and deterioration of fifteenth- andsixteenth-century painted enamels; andBorbely-Kiss et al. (1994) classified late Romanglass seals.

Neutron activation analysis (NAA)NAA is a valuable technique for determiningthe bulk chemical composition of a glass, andit is particularly suitable for detecting minor


and trace elements, but it has two disadvan-tages; it is restricted to the availability of anuclear reactor and, although it is described asbeing non-destructive, some samples, forexample those that contain much antimony orcobalt, are rendered so radioactive that it maybe many years before they are safe to returnto the owner.

NAA has been frequently applied to theelemental analysis of ancient glass to study itsprovenance, colourants and opacifying agents.For example, the technique has been used toidentify the origins of the dichroic colour of theRoman Lycurgus cup (British Museum,London); and in the study of medieval windowglass (Brill, 1965; Olin et al., 1972). Hancock etal. (1994) and Kenyon et al. (1995) applied nondestructive NAA to the study of the colourantsand opacifying agents in European glass tradebeads, by developing a system of short irradi-ations which allowed the beads to be returnedto their collections after two weeks.

Optical methodsOptical methods of analysis include ICP–OES(inductively coupled plasma-optical emissionspectrometry), and AAS (atomic absorptionspectrometry), which have been routinely usedto determine major, minor and trace elementsin the compositions of ancient glass. They arebased on the measurement of bands of UV-VIS radiation resulting from electronic transi-tions, which have been excited in inorganicsamples (in this instance glass) at high temper-atures.

Inductively coupled plasma-optical emissionspectrometry (ICP-OES)ICP–OES is a destructive form of analysisrequiring a sample of 0.1–0.5 g. It has beenused in the study of many ancient and historicglasses: medieval Hungarian glass (El-Nady etal., 1985); glass bead fragments from Hun-garian graves of the great migration periodand glass fragments from the king’s palace atBuda Castle (Zimmer, 1988); Saxon glass fromSouthampton and Winchester (UK) (Heyworthet al., 1989); and Roman glass from Augusta,Praetoria (Mirti et al., 1993): compositions ofVenetian girasole glass (McCray et al., 1995);compositional variations among Romano-British glass from Colchester, Essex (UK)(Baxter et al., 1995).

Atomic absorption spectrometry (AAS)Atomic absorption spectometry is a destructivemethod of analysis requiring a sample of0.05–0.50 g., and is used in the study ofelements used to determine the concentrationof metal ions in solution. The greatest limita-tion to AAS is that the measurement for all theelements must be made serially because thelight source has to be changed for eachelement. Consequently this technique is moreappropriate for the study of groups ofelements already known to be present in aglass sample, than for the study of entirelyunknown glass compositions. However, AAShas been extensively used in the study ofancient glass compositions: ancient Egyptianglass (Brill, 1973); early American glass fromthe New Bremen glass manufactory (Brill andHanson, 1976): Renaissance Venetian glass(Brill and Barnes, 1988); and early Islamicglass (Brill, 1995); characterization of medievalScottish cathedral glass (Tennent et al., 1984b);and the identification of a modern forgery ofearly Roman or late Hellenistic glass from theAshmolean Museum, Oxford (UK) (Newtonand Brill, 1983). Salem et al. (1994) used AASin conjunction with atomic emission spectrom-etry to measure the concentration of alkali andalkali earth ions in acidic solutions, which hadbeen applied to medieval-type glass.

Mass spectrometry methodsThese methods of analysis include ICPMS(inductively coupled plasma mass spectrome-try), SIMS (secondary ion mass spectrometry)and lead isotope ratio determination, and areused to determine the major, minor and traceelement compositions of ancient glasses aswell as their isotopic compositions. They allrequire destructive sampling of the object tobe studied.

Inductively coupled plasma massspectrometry (ICPMS)The advantage of IPCMS is that many elementscan be analysed simultaneously; the disadvan-tages are that the technique is costly andrequires destructive sampling to obtain100–250 g of glass.

Secondary ion mass spectrometry (SIMS)SIMS is a bulk analytical technique used todetermine the major, minor and trace elements


(from hydrogen to uranium) in a (glass)sample (which should be less than 2.5 cm indiameter and 1.0 cm thick), by bombarding itwith the heavy primary ions (such as oxygenor caesium). It will detect elements in thesurface of the glass in the parts-per-millionrange. Since its range is limited to a depth of50–100 A, SIMS is a valuable aid to the studyof weathering phenomena on ancient glasssurfaces, such as medieval window glass(Schreiner et al., 1984). It has been used tostudy the deterioration of ancient glass inmuseum environments (Ryan et al., 1996), thedeterioration of weeping Venetian glass(Rogers et al., 1993) and Hogg et al., 1999).

Isotope ratio analysisLead and oxygen isotope ratio analysisrequires destructive sampling of the object tobe studied. There are four stable isotopes oflead, created by radioactivity. Three of these,(206Pb, 207Pb and 208Pb) are produced bythe decay of 238U, 235U and 232Th respec-tively. Thus the ratios of the four lead isotopesin a lead ore deposit will depend upon thegeological age of the deposition of the ore.Theoretically then, the lead isotope ratios in alead-containing glass will have the potential toprovide information about the particular leaddeposit exploited for its manufacture (Brill,1967a,b, 1968, 1970a,b; Brill et al., 1970;Barnes et al., 1978). For example, lead fromDerbyshire (UK) has 206 Pb/204 Pb = 18.6 and208 Pb/207 Pb = 2.46, whereas lead of Italianorigin has values of 18.8 and 2.48, respec-tively. Although these pairs differ only byabout 1 per cent they are nevertheless statis-tically significant. However, since several leadsources can have the same geological age,lead isotope ratios alone can only be used toeliminate potential lead sources, not to assignthe provenance of an object. Ore depositscontaining lead isotope ratios different fromthose of an object can be disregarded aspotential sources of lead for the object;conversely, ore deposits with lead isotoperatios similar to those of an object can beconsidered as potential sources of lead for theobject.

Unlike the concentrations of trace and minorelements, lead isotope ratios will not beaffected by the chemical and pyrotechnologi-cal transformations that raw materials undergo

during their conversion to glass. However theinterpretation of lead isotope ratios is madedifficult by the long-range trade of materialscontaining lead, which can result in the use oflead from several sources in one object; andby the frequent practice in antiquity ofmelting-down and re-using glass (and metal)objects.

This technique of analysis has beenfrequently applied to the study of colouredopaque ancient glasses with high leadcontents, such as red and yellow opaqueglasses, primarily to identify the sources oflead used, or for identification of groups ofglass which were prepared from the samesource of lead ore: Egyptian (Brill et al., 1970,1974), Japanese and Chinese glasses (Brill etal., 1991), third millennium BC glass bead fromNippur (Vandiver et al., 1995).

In a similar manner, the proportion of theoxygen isotope 180 can be characteristic ofglasses made from different sources of rawmaterial (Brill, 1968, 1970a, 1988).

Radiation monitoring of potash glassRadiation monitoring of potash glass can be ofuse in determining whether painted windowglass is medieval or a later replacement. Theformer will contain potash, derived from thebeechwood ash from which it was made, andtherefore contain the natural radioactiveisotope of potash (40K), the B radiation fromwhich can be detected with a standard radia-tion monitoring badge shown as a darkeningof the badge over a period of approximatelytwo months. Glass made later (certainly afterthe sixteenth century) will have been madefrom an alkali, which was predominantly soda.Potassium contains a naturally occurring,weakly radioactive isotope (40K), and hencemedieval glass will slowly cause darkening ofa radiation-monitoring badge over a period ofabout two months (Hudson and Newton,1976).

Oxidation state methodsOxidation state methods of analysis includeESR (electron spin resonance) and UV–VIS(ultraviolet and visible spectroscopy), which inthe study of ancient glasses are used primar-ily for the determination of the oxidation statesof the glass colourants and clarifying agents.


Electron spin resonance (ESR), also knownas electron paramagnetic resonance (EPR)ESR requires destructive sampling to obtain a15–20 mg sample of glass. Sellner (1977) andSellner et al. (1979) used ESR to measure thestates of oxidation of iron and manganese inmedieval glasses from two sites, and thusshow that a wide range of colours could beobtained by using beechwood ash as thesource of alkali and varying the state of oxida-tion of the glass.

Ultraviolet and visible spectroscopy(UV–VIS)UV–VIS can be used to identify transitionmetal ions present in glass, and their state ofoxidation. UV–VIS has been used to study thecolours produced by iron and sulphur in lateRoman glass (Schreurs and Brill, 1984); ironoxidation states in medieval stained glass(Longworth et al., 1982); in the study ofRoman and of post medieval glass (Green andHart, 1987); and identification of the colour-ants used to prepare the red and yellowstained glass windows of Toledo Cathedral(Spain) (Fernandez Navarro and La Iglesia,1994).

Rarely used and out-dated analysistechniques

Chemical analysisThe chemical analyses of glasses by traditionalmethods have largely fallen into disuse. Thereasons for this are that chemical analyses areextremely time consuming, require highlytrained analysts, need substantial amounts ofglass as samples, and require preliminarydissolution of the glass with hydrofluoric acid,or by fusion with sodium carbonate. (SeeChapter 5 for the dangers of using hydrofluo-ric acid.) Nevertheless, in competent hands,excellent results can be obtained by chemicalanalysis of very small glass samples, andwould still be used as a reference before theintroduction of a new instrumental method ofanalysis.

Emission spectroscopy (spectrography)The analysis of glass by emission spectrogra-phy, which required the use of a massiveamount of equipment, has now been super-seded by other methods. This destructive

technique of analysis required a tiny fragmentof glass to be totally destroyed (volatilized inan electric arc discharge), so that individualelements, seen as a visible spectrum, could beidentified. It will analyse for almost anyelement in a semi-quantitative manner, can beused over a wide range of concentrations, andwas the classical technique of analysis used inthe early part of the twentieth century. Sayreand Smith (1974) used it for their study ofglass from the New Kingdom to early Islam.Newton and Renfrew (1970) made use ofresults which had been obtained a generationbefore to study the origins of British faiencebeads. Emission spectroscopy has beenreplaced by XRF.

Counting the weathering layersGlasses, which have weathered in conditionswhere the alkali has leached out in an unusualmanner, may develop a thick crust on thesurface. The crust may be 4 mm thick and,when a section is examined under a micro-scope, it can be seen to consist of a multitudeof very thin layers (see Figure 6.1). Brewster(1855) found twenty or thirty layers in one-fiftieth of an inch; the layering phenomenawas recorded by Fowler (1881), Raw (1955)and Geilmann (1960). Brill and Hood (1961)noted that, in eleven out of about twohundred samples examined, the number oflayers in the crust was approximately the sameas the number of years that they had beenburied, or immersed in fresh- or seawater.Despite the fact that eleven samples out oftwo hundred cannot be considered to bestatistically significant, it was hypothesized thatthe layers represented an annual phenomenonwhich could be used for dating the glasses (inthe same way that annual growth rings areused to date trees). Newton (1966, 1969,1971a) concluded that there were too manyinconsistencies in the technique to enable it tobe reliable. These include the effect of temper-ature (ten layers were produced in four hoursin an autoclave); and evidence that layerscould merge into each other (Newton, 1972).Douglas (in Newton, 1971a, p. 7) suggestedthat the layers were in fact the result of thealternation of two weathering processes,which accidentally took about a year tocomplete a cycle. Shaw (1965) used EMPA toshow that the silicon and calcium contents of


the crust alternately rose and fell with aperiodicity of 6–8 μm, the same as that of thevisible layers in the crust. Newton and Shaw(1988) illustrate a case in which the weather-ing layers are at an angle to one another other.It would seem that the formation of surfacelamellae does not correspond to yearly cycles,but form as a result of minute changes in thecomposition of the glass.

Hydration rind datingLanford (1977, 1986) found that the alkali-deficient layer on the surface of weatheredglass were thickest on the oldest glasssamples, and suggested that measurement ofthe weathered layer could be used for datingsamples. However, it proved difficult tosubstantiate this hypothesis, and the researchwas abandoned.

Fission track datingFission track dating depends on the assess-ment of damage done to glass over a consid-erable length of time, by the tracks fromnuclear fission, caused by the spontaneousdisintegration of uranium atoms that itcontains (Nishimura, 1971). However, man-made glasses rarely contain uranium, and arenot old enough for sufficient fission tracks tohave formed (Brill et al., 1964

Radiocarbon (14C) datingRadiocarbon dating is used to date carbon-containing materials such as wood. Howeverthere have been two instances where thetechnique has been indirectly useful in datingglass. Glass from the great slab at Bet She’arimwas found to contain 3 per cent of dissolvedcarbon dioxide; Brill (1968) stated that radio-carbon dating had been carried out on asample of the glass, but the results seem notto have been published. Fiorentini-Roncuzzi(1970) dated Byzantine mosaics at Ravenna toAD 345–695, by radiocarbon dating the strawbinder in the original mortar in which the glasstesserae had been embedded.

Thermoluminescence (TL)TL is used to date ancient pottery by releas-ing energy (as thermoluminescence) storedfrom the time the pottery was fired. (Anysubsequent application of heat will havenegated the result, i.e. in effect having become

the last firing; Brain, 1999.) A modification ofthe technique, radioactively induced TL, hasbeen used for characterizing obsidian (Huntleyand Bailey, 1978).

Beta-ray backscatteringBeta-ray backscattering has been used todetermine the lead content of glasses by irradi-ating the surface with electrons, and thenmeasuring those which are scattered backagain by the atoms of lead. Emeleus (1960)used it to study numerous samples of the firstglasses made by George Ravenscroft in theseventeenth century (see Chapter 2). Asahinaet al. (1973) used it to determine the leadcontent of a Japanese blue glass bowl.

Use of a profilometerAt one time it was thought that the chemicalcomposition of a glass could be deduced bymaking sufficient well-chosen measurementsof physical properties but that is now consid-ered to be unlikely.

A profilometer can be used to measure thesmoothness of the surface of a glass; this hasbeen of value in considering how any rough-ness affects the strength of glass. The Britishversion of the profilometer (the Talysurf), canmeasure the roughness in 25 nm. Whenenamel paint is fired on to glass, it createsdepressions in the glass (i.e. an image), lessthan one wavelength of visible light, butwhich can be measured by means of aprofilometer. Newton (1974b) used thetechnique to detect the image left by painteddecoration on medieval window glass, whichhad been removed by inappropriate cleaning.There are also optical methods of studyingsurfaces, such as the Schmaltz light cut(Vickers Projection Microscope Handbook).

Infra-red reflection spectroscopy (IRRS)IRRS has been used to characterize the weath-ered surface of medieval glasses (Hench et al.,1979; Schreiner et al., 1999). The techniquedepends upon observing the changes thatoccur in the vibrations of the ‘silicon-bridgingoxygen stretching’, and ‘silicon non-bridgingoxygen bond’ in a molecule.

Ion beam spectrochemical analysis (IBSCA)IBSCA is an extremely sensitive technique ofanalysis, which, for example, permits studies


to be made of the relative durabilities of differ-ent areas of any piece of glass (Hench et al.,1979; Rauch, 1985; Lanford, 1986).

Techniques of the futureTechniques of analysis are continually evolv-ing. Those which may become applicable to

the study of glass are laser-induced break-down spectroscopy (LIBS) (Mansfield et al.,1998), atomic force microscopy (Techmer andRädlein, 1998), Brillouin light scattering(Cavaillé et al., 1998), voltammetry of immobi-lized microparticles (Perez-Arantegui et al.,2000) and X-ray tomography (mCT) (Römichand Lopez, in Kordas, 2002).


Conservation has two aspects: first, the controlof the environment to minimize the decay ofartefacts and materials (passive conservation);and secondly, treatment to arrest decay and tostabilize them where possible, against furtherdeterioration (active conservation). It hasbecome increasingly common for environ-mental control to be allowed for in the finan-cial budget to refurbish or construct museumgalleries. However, environmental control isby no means universal, either within museumsor storage areas. In relation to glass, specialstorage conditions are required for damp orwet archaeological glass, and for storing/displaying historical weeping glasses, Egyptianfaience and enamels. Restoration is the exten-sion of conservation, by which part or all ofincomplete objects are reinstated without falsi-fication, using synthetic materials. This may bedone to render an object safe for storage ordisplay, or to aid in its interpretation.

In considering conservation and restorationof objects, the ethics of conservation need tobe continually borne in mind. Briefly stated,these are: the assessment of risks to the objectassociated with conservation (Ashley-Smith,1999); that all treatment should be adequatelydocumented, and that there should be nostructural and decorative falsification ofobjects. In addition, it is generally agreed that:all conservation/restoration processes should,as far as is practicable, be fully reversible,even after a number of years; that wherepossible decayed parts of objects should beconserved in situ, and where this is not possi-ble, loose fragments should be carefullylabelled, packaged and placed in storage; andthat the natural consequences of ageing of

the original material (iridescence and flakes in the case of glass) should not normally bedisguised or removed.

Glass is a difficult material to conserve,especially when it retains its transparency ortranslucency. It can be perfectly clear,coloured in an almost infinite range, or beopaque. It is always fragile and in some cases,its composition results in it being chemicallyunstable.

Different approaches are taken in the way inwhich conservators deal with archaeological,i.e. excavated, ancient glass, and historicalglass. In the case of archaeological glass, apolicy of minimum intervention is generallyfollowed. Removal of weathering layers solelyfor the purpose of enhancing the glass isconsidered unethical. However, in rareinstances, it may be necessary to remove someor all of mud-encrusted deterioration productsin order to identify the object or its true colour.Missing areas of a glass artefact should only befilled where necessary to add support or aidinterpretation, and where possible, the gap-fillsmade away from the object. Glass fragmentscan be mounted on a Perspex (US Plexiglas)frame or stand, or adhered to a former madeof resin or glass. It is difficult to blow glass tothe shape of an ancient glass of inexactmeasurements, and therefore this can be a time-consuming and consequently expensive option.

Before the advent of synthetic resins, theonly was of achieving a transparent fill inglass, was to cut and insert a piece of glassfrom a similar object. Although this practice isstill carried out in some countries, it is gener-ally considered to be unethical. In the case ofdecayed archaeological glass and enamel, it is

242

7

Conservation and restoration of glass

not necessary to match the refractive indicesof conservation materials to the glass, whereasin the case of uncoloured, transparent histori-cal glasses, matching the refractive index canresult in repair and restoration being lessvisible. Repairs to ancient glasses are notexpected to be invisible. However, those tohistorical glass, which may, for example, behigh art objects representative of the culturein which they were produced, presentationpieces, or simply of aesthetic or sentimentalvalue, need to be as unobtrusive as possible,since the emphasis is on viewing the glass as

an art object. In order to restore missing areasof glass, moulds are often taken directly off acorresponding area of the object, sealed inplace over the missing area, and a syntheticresin poured into the mould in situ. For thisto be done, the glass and any surface decora-tion must be sound and firmly attached to oneanother.

The case histories used in this chapter toillustrate conservation procedures are intendedto be taken as treatments that have beendevised to meet specific problems, and whichcan be adapted and modified.

Conservation and restoration of glass 243

The first stage of conservation takes placebefore an excavation begins: in theory thereshould be no excavation without prior consid-eration being given to conservation of thefinds, although in practice, the budget forpost-excavation work is often very limited.Liaison between the excavator, conservatorand the curator of the museum or store inwhich the finds will ultimately be deposited,should determine the amount of time, money,space and administration available to deal withany glass that may be retrieved from excava-tion. On-site conservation consists of correctlifting, labelling, packaging, storage and trans-port, processes which are well documented(Dowman, 1970; Sease, 1988; Cronyn, 1992).However some preliminary preservation work,such as de-salting and cleaning, may becarried out if time permits. In the laboratory,more detailed and therefore time-consumingconservation can be undertaken.

During archaeological excavations, glassmay be found in the form of whole, brokenor completely shattered artefacts, pieces beingeither more or less in their original positions,or disturbed and scattered over a wide area.The most common glass artefacts encounteredare vessels, window panes, manufacturingwaste, wall or floor mosaics, or small objectssuch as bracelets and beads, the lattersometimes attached to textiles and ethno-graphical materials. Glass may also be foundas a minor constituent of objects (Plate 3),such as inlay in furniture and jewellery

(Bimson, 1975; Cronyn et al., in Bacon andKnight, 1987). The greatest quantity ofexcavated glass, however, consists of individ-ual fragments from any of these sources.Nevertheless, shapes, profiles and other infor-mation such as details of the technology canbe retrieved from fragmentary evidence. Asarchaeology and its associated techniquescontinue to develop, new questions will beasked of the evidence, and therefore allexcavated glass should be retained for futureexamination (Wihr, 1968, 1977; Eshøj, 1988;Newton and Davison, 1989).

Water content

One of the major uses of glass vessels is forthe storage of liquids, which would tend tosuggest that glass is impervious to water.However, this is not strictly true, as alkali isleached out of the glass during prolongedcontact with water. The mechanisms involvedin the deterioration of glass are discussed inChapter 4. The water content of deterioratedburied glass can vary from saturation ofmaterial excavated from marine or water-logged land environments, though damp inglass from the majority of sites in cool,temperate climates, to dry from sites inwarmer seasons or climates. In hot seasons orsemi-arid climates, the glass may be partiallydesiccated (free from surface water). In aridenvironments, such as desert tombs or caves,

Part 1: Excavated glass

the glass can be totally desiccated although, ina physical sense, there will always be somefree water. Surviving archaeological glassrepresents only the best formulated glass of amuch larger output.

During its burial, glass will have reactedwith its particular environment and may evenhave achieved a state of equilibrium with it.Assessing and dealing with the water contentof excavated glass is crucial to its survival.Upon excavation, glass is suddenly exposed toa new set of environmental conditions,perhaps after hundreds of years; thus it cannotbe considered surprising if it immediatelyreacts to the change in environment. Glasswhich may seem to be in good conditionupon excavation, may form iridescent layerswithin a few minutes or hours. The change inappearance is not a result of a rapid increasein the rate of deterioration of the glass, but issimply a revelation, as a result of dehydration,of the deterioration that has already takenplace. The free water, which is present innewly excavated glass, maintains the trans-parency and also holds weathered layerstogether by virtue of surface tension. Hencedrying may cause the glass to crumble orflake. Once the free water has evaporated,allowing air to enter the weathering crust, it isextremely difficult (if not impossible) to renderthe glass translucent again. The free water willalso contain dissolved salts, derived eitherfrom the environment or from the decayingglass itself. On drying, these salts will crystal-lize out, disrupting the weathering crust and(if allowed to remain) ultimately destroying itby the oscillation of their volumes as theambient humidity changes. Consequently, theglass must be photographed immediately onexposure, being kept damp between shots, bycovering it over with damp tissue andPolythene sheeting. (If a large amount of glassis involved, it may be advisable to lift the glassand place it immediately into a controlledenvironment pending photography.) Animportant aspect of on-site conservation,therefore, is the provision of first-aid treat-ment, with adequate packaging incorporatingbasic environmental control, in order topreserve glass in the condition in which it wasfound. Correct treatment and regular monitor-ing can delay the further deterioration of glassafter excavation for a considerable period of

time. Ideally, packaging should be devised tocover the short period between excavationand laboratory conservation. In practice,however, as mentioned above, this period ofnon-treatment often becomes permanent dueto lack of funds, and this fact must be takeninto consideration when planning the work.

Removal of glass from the ground

Conservation time (and therefore money) canbe saved if the glass can be correctly lifted bythe archaeologists. If, however, conservatorsare on site, they will be required to work asquickly as practicable without unduly disrupt-ing the progress of the excavation. Con-servators should check that the area in whichthey are working has been recorded,especially if the earth has to be removed toany depth in order to recover artefacts. Liftingmethods will vary with the state of the glassand the nature of the surrounding soil.

The total extent of glass deterioration mayor may not be immediately visible, dependingon its state of preservation, and its degree ofsaturation with water. Glass from dampdeposits must therefore initially be kept dampafter exposure. After recording and photogra-phy, it may be convenient to allow a small,representative sherd to dry out slowly awayfrom direct heat. If no appreciable deteriora-tion occurs, the glass may be allowed to dryout. If deterioration of the sample occurs, thebulk of the glass must be packed in dampacid-free tissue or Plastazote in self-sealPolythene bags (Figure 7.1) or polypropylene


Figure 7.1 Wet glass fragments stored in a self-sealPolythene bag.

boxes and stored or sent for laboratory treat-ment.

Depending upon their structural strength,glass objects or fragments may be simply liftedout of the ground. However, where there aresigns of flaking or of gilding or painting, alayer of soil must be allowed to remainattached to the glass to prevent the surfacelayer becoming detached. Complete vesselsmust be totally uncovered before removal toprevent damage from occurring as they arelifted. In the case of fragmented glass, therecording of the spatial relationship of thepieces may be essential to aid later recon-struction, or in the case of window glass, inter-pretation of the iconography. Thus, beforethese are removed individually, they must beplotted, drawn and photographed.

For economic and other reasons, full conser-vation of all excavated material may not bepossible or desirable. A selection process mayhave to be adopted, following discussionsbetween the archaeologist, conservator and sitedirector or museum curator, to ensure that allthe relevant criteria are considered for gainingthe maximum information. The most importantis the archaeological significance of the glass,not only in relation to the discipline as awhole, but also to its own particular context.Selected glass may be simply recorded in termsof amount; or cleaned and drawn for study orpublication. Other glass may be conserved andrepaired. Subsequent full reconstruction (andrestoration or mounting) may be restricted tothose objects required for display.

Certain categories of glass may best beremoved from site immobilized in a block ofsoil: glass fragments or objects which are sothin or degraded that they are in danger ofdisintegrating; an extensive spread offragments in dry or damp deposits; or objectsthat will not support their own weight(especially if filled with debris). The aim ofthe lifting technique is (i) to preserve the glassin the condition in which it was excavated;and (ii) to render the glass itself immobile; or(iii) to render the soil around the glassimmobile so that it will support the glassduring its removal, storage and/or transport toa laboratory. Methods that may prejudicefuture conservation treatments should not beused, and all methods should follow theconservation ethic of minimal intervention.

Immobilization of the glass aloneIt is difficult to immobilize glass by consol-idation when it is still in the ground, and itmay be premature to do so for severalreasons. First, because the glass cannotusually be cleaned thoroughly, dirtinevitably becomes consolidated on thesurface. It is not always possible to removeeither the dirt or consolidant at a later stagewithout damage to a flaking glass surface.Secondly, the consolidant is unlikely topenetrate throughout the weathering crust,so that spalling may occur at a later date.Thirdly, damp or wet glass would have tobe dried before consolidation (which maycause damage), or a water-miscible consoli-dation system must be used. In order toovercome the problems posed by working indamp conditions, Bimson and Werner (1971)suggested adapting a technique usingCarbowax 6000 (polyethylene glycol) for usein the field. An extremely dilute solutionwould have to be used in order to ensuregood penetration and hygroscopicity of thewax may cause difficulties in moist storageconditions. However the consolidant wouldbe resoluble in water. Wihr (1977) and Ypey(1960–61, 1965) advocated the use ofDermoplast SC Normaal (polyvinyl chloride)in a ketone hydrocarbon solution for consol-idating glass in situ. However, polyvinylchloride is unstable and will cause futureconservation problems. Emulsions such asthose of polyvinyl acetate may be useful indamp conditions, but they may not penetratefully, and are difficult to remove at a laterdate. Another approach used in the past wasto dehydrate the glass first and then to applya resin. Wihr (1977) described such amethod in which Araldite AYIO3/HY956(epoxy) was applied to a glass vessel in theground after the vessel had been dehydratedby filling it with two changes of acetone.Apart from causing damage to the glass bydehydration, potentially dangerous, grindingand abrading techniques had to be used toremove excess resin. Impregnation of glassin this way, still carried out in somecountries, has the advantage of restoringtranslucency in some cases, but the long-term effects are likely to be deleterious.More suitable methods of lifting glassuntreated, are described below.


Immobilization of a soil blockThe simplest and quickest method of provid-ing support is to isolate the glass on a platformof soil, and then to push a thin metal sheetor spade into the soil well below the glass.The soil platform may be strengthened, bycovering it with aluminium foil, held in placeby and wrapping the block with gauze

bandages dipped in plaster of Paris (Figure 7.2a–e). Such a system should only be consid-ered as a temporary measure in order toremove the glass from its burial place; it is notstrong enough to support the glass duringtransport to a museum. Another temporarysupport of this nature was used by Garlake(1969) in the dry climate of South Africa:


(a) (b)

(c) (d)

(e)

Figure 7.2 The removal of archaeological glass fromthe site by isolating the soil in which it lies, with aplaster of Paris case. (a) Glass within a block of earth,protected by aluminium foil, over which the plaster caseis applied. (b) Sides wrapped with gauze bandagedipped in plaster of Paris, and a metal sheet slidbeneath to facilitate its removal. (c) Block inverted andthe base protected by foil prior to applying plaster ofParis. (d) Completed case secured with plaster bandageand removed from site. (e) Removal of the case at alater date.

polyvinyl acetate was painted around theedges of the soil platform. Consolidating thesoil in this way, with a moisture-curing resinsuch as Quentglaze (polyurethane), has alsobeen suggested (Dowman, 1970), but as sucha resin is not re-soluble it must not be allowedto come into contact with the glass itself.

Another method for immobilizing the soilblock is to freeze it with dry ice (frozencarbon dioxide). This method has been usedin Sweden (Arrhenius, 1973) for lifting archae-ological materials on site, and it is conceivablethat this technique might occasionally be ofuse in temperate climates, if there wasimmediate access to freezer storage.

A more sophisticated method for immobi-lizing a soil block is the construction of aprotective casing with plaster of Paris or

polyurethane foam (see Figure 7.3). In thismethod, both the glass and the surroundingsoil are rendered immobile by enclosing themin a rigid casing. Great care must be taken toprevent the materials of which the casing isconstructed from coming into contact with theglass itself. The block is isolated with damppaper tissue, followed by aluminium foil orthin vinyl plastic film (e.g. Clingfilm; US:Saranwrap).

Plaster of Paris is cheap but also heavy, sothat for the construction of large casings it maybe preferable to use polyurethane foam, whichalthough expensive, is light, rigid and easilycut with a knife for removal at a later date.The foam is formed when two components aremixed together, setting hard in about 5–10minutes. The foaming reaction is exothermic.


Figure 7.3 Removingarchaeological glass from siteby isolating the surroundingsoil, in a plaster of Paris orfoamed polyurethane casing.(1a) Glass on a pedestal ofearth and protected by foil orvinyl wrapping. (2a)Assembly strengthened bywrapping it with cottonbandage dipped in liquidplaster of Paris, forming acocoon. (2b) Surrounded bya collar of corrugatedcardboard, andcommencement of applicationof foaming polyurethane. (2c)Flat sheet of wood placed onthe foam to produce a flatsurface. (3a) Inverted blockafter inversion andapplication of more foam tocomplete the casing. (3b) Thecasing secured with straps.Alternatively a case could bemade of six plaster sides,applying a release agentbetween the pieces as theywere made.

If there is any risk of the heat affecting thedecayed glass, an insulating layer of closelypacked sand or earth or of Plastazote couldbe placed over the isolating film before thefoam is applied; or the foam could be appliedin several thin layers. It may not cure in coldor damp conditions, and there is a potentialhealth hazard in its use since one of thecomponents is isocyanate-based and can causesevere irritation of mucous membranes and ofthe skin. Great care must therefore be takennot to allow it to come into contact with theskin or be inhaled when it is being mixed, orwhen the foam is cut at a later date (Moncrieff,1971; Escritt and Greenacre, 1972; ICI, 1977;Watkinson and Leigh, 1978). Protective glovesand a face-mask must be worn (and fume ordust extraction utilized if the foam is mixed orcut indoors).

The procedure for making a casing arounda small glass object is shown in Figure 7.3,and described below. The measurementsquoted can be adjusted according to the sizeof the object to be lifted. The object is isolatedon a platform of soil some 25 mm larger thanitself. This platform is then undercut as far aspossible leaving the object on a pedestal about50 mm high. Any undercuts in the object arefilled with soft soil or tissue. A collar of corru-gated cardboard, rigid plastic, wood or metal,is placed around the pedestal, allowing a gapof 20–30 mm between it and the collar andallowing the collar to stand 20 mm proud ofthe uppermost surface of the object. Soil isheaped around the outside of the collar,blocking any gaps and holding it down. Toprevent the foam coming into contact with theglass, the object and pedestal are then coveredwith a piece of clear vinyl plastic film, thinPolythene sheeting or aluminium foil. A smallquantity of foam is applied from an aerosolcan, or is mixed according to the manufac-turer’s instructions, and poured into the collar,ensuring that it runs beneath the pedestal. Thisprocess is repeated until the foam nears thetop of the collar, when a piece of wood isplaced over it and light pressure exerted toproduce a flat surface. The surface should bemarked with the site-name, year, orientationand details of its position on site, also withthe word ‘TOP’. The pedestal and collar arethen undercut by inserting a thin metal sheet,and the whole case is inverted. The surface

thus exposed is isolated as described aboveand a layer of foam applied. The completecase is then re-inverted so that it stands on itsoriginal base. The case should be wrapped inPolythene sheeting and stored in cool condi-tions until it is transported to a conservationlaboratory.

Glass that can only be removed by usingthe lifting methods described above mayalready show considerable signs of deteriora-tion in the ground. If it has been lifted simplyin a block of soil, the moisture content of theburial environment should be maintained asfar as possible in packaging. Thus, blocks ofsoil and glass should be placed in well-sealedPolythene bags or boxes, stored in coolsurroundings, and treated in a laboratory assoon as possible. If soil blocks are allowed todry out, the glass will become dehydrated, andthe soil block may crumble within the casing,causing the destruction of the glass within.Therefore the casing should only be left inposition temporarily, and should be openedand its contents fully excavated and treated assoon as possible.

Occasionally entire glass kilns are removedfrom site, either in sections (Hurst-Vose, 1980),or as a block lift (Price, 1992).

Glass from waterlogged land sites

Glass found in waterlogged land sites shouldbe kept wet before lifting by covering it withwet paper, saturated plastic sponge and sheetsof Polythene. After lifting, the glass must bekept wet, but not necessarily immersed inwater, unless the presence of large quantitiesof salts is suspected, or long-term storageenvisaged. For a short period it can be placedin self-seal or heat-sealed, water-tightPolythene bags together with a few millilitresof distilled water. Since such bags are nevertotally water-tight, they should be placed inplastic boxes with tightly fitting lids, which canbe carefully stacked on top of one another,separated by padding, such as shreddedPolythene sheeting or plastic sponge. Toprevent the growth of fungus or bacteria, aneutral biocide can be added, no tissue orother organic packing or labelling materialsshould be included, and the glass should bekept cool or refrigerated. In such waterloggedconditions, hydrolytic breakdown of the glass


can continue, especially when the pH withinthe bags rises. Thus this type of storage shouldnot be needlessly prolonged. However, long-term storage is sometimes inevitable.

The possibility of deep-freezing glass at atemperature of –20°C has been suggested(Arrhenius, 1973; Nylén, 1975). However,further work on freeze-drying excavated glassis essential before such a method can berecommended unreservedly. Another possibil-ity during prolonged periods of storage is tobegin the process of consolidation, by replac-ing the water with an inert solvent, whichallows the glass to be de-watered, but whichprevents air entering the weathering crust. Thewet glass should be immersed in a 50 per centmixture of ethanol and water and, afterapproximately an hour, moved into increas-ingly solvent-rich mixtures until the glass canbe stored in a pure solvent. The experiencesof rescuing damaged glass objects, following adisastrous flood in the Corning Museum ofGlass (New York State), have beendocumented by Martin (1977).

Removal of glass from marine orfreshwater sites

Glass fragments or objects are recovered fromrivers, lakes (Singley, 1998) and the sea, mostnotably from submerged buildings, or

shipwrecks, when glass was being carried ortraded in the form of complete objects or ascullet, which also acted as ship’s ballast(Pearson, 1975, 1987). Depending upon theircondition, individual objects or fragments ofglass may be lifted from their burial sites, withor without the aid of supports, and kept wetuntil treated (Figure 7.4).

The support and lifting of a fragile glassbottle from the sea is described by Turner (inPiercy, 1978). A case bottle retrieved from ashipwreck off Mombasa was found virtuallycomplete, but the glass itself had begun toexfoliate, was badly cracked and only heldtogether by the compacted clay inside. Sinceremoval of the clay would cause disintegrationof the bottle it was decided to trim down themud to leave a 5 mm lining, by excavating itthrough a hole in one side of the bottle fromwhich glass had been lost during burial. It wasthen proposed to reinforce the mud liningwith netting and resin. To support the glasswhile the mud was being removed, the bottlewas faced with netting held in place withPVAC (polyvinyl acetate), but because of thefragile condition of the glass the facing couldonly be applied while the bottle was stillsupported in water. The level of the water wasreduced to expose the upper face alone,which was dried thoroughly by swabbing withIMS and acetone. The surface was consoli-dated with PVAC and a synthetic net facing


Figure 7.4 Storage of a large quantity of objects from a marine site, in a tank of water. The tank is constructed ofconcrete, and has metal roll-top covers (right), which are operated by ropes tied to their handles.

applied. The facing was secure enough towithstand soaking in water while the othersides were faced. The mud was then removedto leave a thin layer on the inside of the bottle.At this stage the bottle could be taken out ofthe water and dried, after which a nettingsupport was applied to the inside layer of mudwith 5 per cent Paraloid B-72 (in acetone), anacrylic resin which is not soluble in IMS. ThePVAC and net outer facing could be safelyremoved with this solvent with no risk ofweakening the lining. Finally, the surfaceswere cleaned and given a final consolidatingcoat of PVAC. If the conservator had access toa museum laboratory (and time), an attemptmay have been made to remove the lastvestiges of mud from the case bottle.

In some instances, however, solidifiedmarine concretion, inside glass vessels, whichhave subsequently become partially or totallydestroyed, may represent the original shapeand form of the artefacts. It is advisable todetermine the state of preservation of theartefacts, and to group the objects accordingly.This will aid in packaging, and in alerting thelaboratory conservator to the conditions of theobjects before they are unpacked. Ideally,there should be no attempt to clean the glasson site since evidence of manufacture, origi-nal design and shape preserved in the weath-ering crust may be removed. If glass from amarine site is allowed to dry immediatelyupon recovery, salts will crystallize in theweathering crust and disrupt it (Pearson,1975), and concretions will harden to acement-like state (see Figure 4.27). Salts musttherefore be removed as quickly as is practi-cable (Macleod and Davies, 1987).

DesalinationDesalination is accomplished by slowly reduc-ing the salt content of the water in which theglass was found, followed by immersion inchanges of fresh water. The salt level of thewater used for desalination should be known.The glass should not be placed directly intofresh (tap or distilled) water in case anosmotic pressure develops between the salt-laden glass and the wash water, causing thewater to force entry into the decayed weath-ering crust and disrupt it. Desalination inwater can be carried out by one of threemethods (Figure 7.5):

(i) A static immersion process in which theobjects are placed in a sealed container ofwater, which is then changed at regularintervals. For the first day the water shouldbe changed twice, e.g. morning andevening, then once a day at the same timeof day. In cases where the amount ofobjects being desalinated is too large orfragile to be moved, samples of the washwater can be syphoned off for testing. Theprocess is very slow, as pockets of salts(registering a relatively high conductivity)will accumulate in the solution, reducingthe efficiency of salt removal, by reducingthe osmotic pressure difference betweenthe salt in the object(s) and that in thesolution. It is, however, probably the bestmethod for use on extremely fragile glass,although the lack of water movement hasto be balanced against the amount of timethe glass needs to remain immersed.

(ii) Gentle agitation of the wash solution willprevent pockets of high salt concentrationfrom occurring, thus an optimum osmoticpressure differential is maintained therebyincreasing the efficiency of the desaltingprocess. Stirrers or pumps can be used toagitate the water gently. Olive andPearson (1975) suggest the conversion ofold washing machines with a cyclic onehour on and one hour off washingprocess, by which the desalting processwas increased by a factor of four timesthat of the static immersion process.

(iii) A flow through immersion process is moreefficient but requires a constant supplyfirst of tap- and then of distilled water.The pressure or movement of the watershould not be allowed to disrupt theweathering crust from the glass surfaceand, if possible, the object should not belifted in and out of the bath, as sponta-neous drying will affect the deterioratedglass.

Where there is a high salt content, the initialdesalination may be carried out using tapwater provided it, itself, does not have a highsalt content (or even diluted sea water forobjects recovered from marine excavations).The tap water may be allowed to run througha bath containing the objects, or changedseveral times a day. When conductivity


readings on the wash water indicate that thesalt level of tap water has been reached,distilled or de-ionized water is used tocomplete the washing. When tap water isused, the tap is turned off at night and thewash water left to stand overnight before asample is drawn off and tested.

Monitoring the salt content of the washwaterThe treatment progress can be monitored byusing a conductivity meter to measure theelectrical conductivity of the wash water. Thereadings are a measure of the amount ofdissolved salts in the water, not of which saltsare present. (A specific chloride conductivitybridge is manufactured, but it is important toremember that chlorides are not the onlydestructive soluble salt. The specific test for

the presence of chloride ions, with silvernitrate, is given below.) The meter measuresthe concentration of salts in the wash waterby measuring the current passing through, thetwo being in proportion to each other. Theconductivity cell is dipped into the washwater, and stirred gently to ensure that it isnot in a salt-free or salt-laden pocket, and thereading taken (Figure 7.6). After each readingthe wash water is changed. The volume of thewash water in every bath should be the same.

Conductivity readings on the wash water aretaken at regular intervals, e.g. morning, after-noon and evening. The regularity of thetesting will to some extent depend upon thequantity of material under treatment, and uponthe conservator’s general workload. Providedthat the glass will withstand continuouswashing, treatment should continue until


Figure 7.5 Options fordesalinating porous objects.

Day/damp objectcontaining soluble

salts

Wet objectcontaining

soluble salts

Crumbling Fragile Sound

Sound Fragile

Bath of pure IMS

Allow to drynaturally Consolidate

Bath of paraloid 8-72(dissolved in acetone)

in IMS

ConsolidateStabilizeenviron-

ment

Desalinatein stillwater

(distilled)

Desalinatein running

watertap → distilled

Desalinatein agitated

watertap → distilled

Desalinatein still water

(distilled)+ monitor

Monitor saltremoval

(conductivitymeter)

conductivity readings taken at regular intervalsremain constant over a period of a few days.It is not necessary to remove every trace ofsoluble salts from objects (and they will bereintroduced to some extent throughhandling).

Once the desalination is complete, the glassshould not be allowed to dry out since theapparent state of preservation in the wet stateis no real indication of what the object maylook like once it has dehydrated. The glassshould be kept wet temporarily (bearing inmind that prolonged contact with waterpromotes its deterioration). Following desali-nation, wet glass may need to be consolidated(see section on consolidation, later in thechapter).

Chemical tests for identifying soluble andinsoluble salts

Soluble saltsThe wash water used for desalinating objectscontaining soluble salts can be electronicallymonitored to detect the presence of the salts:chlorides and nitrates (see above). Alterna-tively, it may be possible to remove a sampleof salt crystals from an object prior to desali-nation, dissolve the sample in distilled water,and test the water.

Chlorides1 The sample is dissolved in distilled water.2 Ten millilitres of the solution is placed in

a clean test-tube which has been rinsedwith distilled water.

3 Three drops of dilute nitric acid are addedto remove any carbonate ions, whichwould confuse the test result.

4 Three drops of molar silver nitrate solutionare added. The formation of a cloudy whiteprecipitate indicates that chlorides arepresent. This is an extremely sensitive testwhich will detect a few parts per millionof chloride.

Nitrates1 The sample is dissolved in distilled water.2 A few drops of sulphuric acid are added

to acidify the solution.3 Ten millilitres of the solution is placed in

a clean test tube, which has been rinsedwith distilled water.

4 A few crystals of fresh ferrous sulphate areadded and dissolved.

5 The test tube is tilted, and a few millilitresof concentrated sulphuric acid are trickleddown the side of the test tube into thesolution. The formation of a brown ring offerrous nitrate indicates that nitrates arepresent.

Insoluble saltsDeposits of insoluble salts are usually carbon-ates, sulphates, or silicates, either combined oron their own. When a small drop of diluteacid is placed on a sample of the deposit, astrong effervescence will suggest that it iscomposed of carbonates, whilst a less vigor-ous reaction would indicate the presence ofsulphates. Silicates do not react with mostacids, and if necessary would have to beremoved from an object mechanically.

Sulphates1 A sample is placed in 10 ml of distilled

water in a clean test tube, which has beenrinsed with distilled water.

2 A few drops of hydrochloric acid areadded to remove any carbonate ions.

3 A few drops of barium chloride solutionare added. The formation of a whiteprecipitate indicates that sulphates arepresent.

If samples of the objects are to be taken forbiological or analytical study, either in the fieldor at some later date, the glass should be storedin 70 per cent ethanol. This will prevent further


Figure 7.6 Use of a conductivity meter to monitor thedesalination process.

biological attack from solutions not experiencedby the object during burial, and will preservethe micro-organisms present on the glass. Ifpossible, soil and water samples from theregion of retrieval should be taken for studypurposes. The glass, stored in ethanol or freshwater and fungicide, should be packed in heat-sealed Polythene bags and the fragmentsseparated from each other by wads of polyesterfoam or bubble packing. The objects should becoded so that the more fragile glass is placednear the top of the container and thus moreeasily retrieved for conservation.

Investigation of glass deterioration resultingfrom storage systems for waterlogged archaeo-logical glass, have been undertaken by Earl(1999). These included the commonly recom-mended immersion and high humidity environ-ments; low temperature and ethanol/watermixtures; and novel approaches, based onresearch for the glass industry, with the poten-tial for improved stabilization, notably solutionscontaining phosphate, aluminium, calcium andsilicate ions and pH buffers. At intervalsthroughout the study, Fourier transform infra-red microscopy was used to examine the glasssurfaces. The position and relative intensity ofthe bands, due to the stretching mode ofSi–O–Si (bridging oxygen and Si–X (non-bridging oxygen), change according to the na-ture and extent of deterioration. Some of thestorage environments were also evaluatedusing scanning electron microscope and energydispersive X-ray analysis to examine the glasssurfaces. The development of surface effectssuch as crystal growth and highly localizeddepletion of alkali and alkaline earth metalions was noted. As a result of the findings, therelative efficacy of the environments wasassessed. It was concluded that the highhumidity systems and those containing ionsthat affect the surface chemistry of the glassshould be avoided in the conservation ofarchaeological glass, and that ethanol/watermixtures at low temperatures are worthy offurther investigation. The best of these wasstorage at temperatures below 0°C in a 50/50vol/vol ethanol water mixture.

Glass opus sectile panels fromKenchreai, GreeceDuring excavations between 1965 and 1968 ofthe submerged site of Kenchreai, one of the

two ancient ports of Corinth in Greece(Ibrahim et al., 1976), a cache of over onehundred opus sectile glass panels was recov-ered from the sea (Koob, Brill and Thimme,1996) (Figures 7.7a–c). The panels had beenin temporary storage still in their shippingcrates, leaning against the walls at an anglefairly close to the vertical with one long sideresting on the floor. There were four to tencrates in each of nine stacks (Figure 7.7a). Asa result of an earthquake, which destroyed theport in antiquity, the entire building wassubmerged, and tilted, so that the floorssloped downward towards the south. It mustbe assumed that the panels themselves weredisturbed by the shock. Most important, theywere waterlogged, so that the plaster slab onwhich the glass was affixed lost its cohesionand the panels were no longer solid or rigid;they could not be moved without crumblingto pieces. Presumably other damage wascaused by the fact that the edges of the crateresting on the floor would have beendislodged as they slid away from the walls,allowing the panels to sag and warp.

The building and its contents were subse-quently abandoned, filled in with debris andbuilt over. During one thousand six hundredyears of submerged burial, continuous reactionof the glass with chemicals in the sea waterand with decaying animal and vegetable life,resulted in extensive decomposition of muchof the glass. The parts of most crates highestabove the floor were eroded by wave actionor by subsequent despoiling of the walls toobtain building material.

Clearance of the site began in 1965, andexcavation began in 1968, when largestorage/desalination tanks were constructed onthe shore. During excavation, work had to becarried out in shallow water, in order toprevent the building from drying out, althoughpumps constantly lowered the water leveltemporarily. Various lifting techniques wereused to remove the stacks of opus sectile; in1968, two entire stacks were block-lifted andtransferred to the freshwater storage tanks onshore. The water was replaced daily over aperiod of one month, and readings of thesalinity of the water taken at the end of eachday. The results (unpublished) seemed toshow that by the end of the period of immer-sion, the salts had been removed. The panels



Figure 7.7 (a) Stacks containing panels of opus sectilewith preserved end-pieces of the crates in situ, afterpreliminary cleaning. Kenchreai, Greece. (b) Cleanedand consolidated fragment of opus sectile (panel number2 layer B), fixed to a sheet of fibre-glass screening. (c)Part of a panel with potters’ clay isolating the glassfragments and filling the voids. (d) The voids filled(over the clay) with Vel-Mix. (e) Restored section of apanel, similar to that shown in (a–d), on display inNauplion, Greece. (Courtesy of T.K. Lord).

(a)

(b)

(e)

(c)

(d)

in these stacks were the only panels to be fullydesalinated. The crates were then taken toNauplion (Greece) with the intention ofdissecting and conserving them, as timepermitted. The panels had been packed face-to-face, two to a crate, and had through theprocesses of deterioration become fused toone another. Keeping part of the stack in itsaccustomed state of saturation with water,whilst spending an indefinite amount of timeconserving the exposed panel, was impracti-cable, because the moisture from the wet partof the stack would move continuously towardthe dry part. Therefore, the glass panels couldnot, excepting in two or three instances, befully dissected, cleaned and consolidated.Although there has been extensive loss, whathas survived is of such a nature that it haspreserved a disproportionately large amount ofevidence.

Although very few records were kept, itseems certain that the majority of panelsreceived only minimal treatment. A thickcoating of polyvinyl acetate (PVA) resin‘dissolved in a mixture of different solvents’(Ibrahim et al., 1976) was applied to the backof some of the panels. The majority of thepanels were backed with bandage/cottongauze impregnated with PVA, and they werethen wrapped in yellow flannel cloth forprotection and supported on sheets ofMasonite.

In 1971, seven fragments of the opus sectilepanels were sent to the Corning Museum ofGlass (USA) to undergo conservation researchand restoration. It was hoped to be able todevise a method of conservation which wouldstabilize the fragments, and which could beused at a later date on the remaining panelsin Greece. The treated panels were eventuallyreturned mounted for display in the IsthmiaMuseum in 1976.

Upon arrival at Corning, the panel fragmentswere examined visually and by radiography,before being stored in their original packinguntil conservation experiments could begin.Unfortunately, Corning Museum itself wasinundated by a flood during 1972, and it wasnearly three weeks before the panels wereretrieved. The glass was covered with mud, alayer of mould-growth developed on theflannel, and the Masonite supports hadwarped. The flood waters had softened the

original background material of the panels,and the consolidant, so that the gauze and theflannel cloth adhered to and in some areashad become partially embedded in the panelsurfaces. The panels were cleaned superficiallyand set aside, worked on when time permit-ted.

Conservation experiments began in 1974. Anew set of X-radiographs was made in orderto achieve a better understanding of exactlywhich areas of each panel were of glass, andwhich were of matrix material. With a fewexceptions the glasses were heavily weatheredand had little resemblance to their originalcolours. The shapes which stood out mostclearly in the radiographs were the areas ofred, yellow and green glass, owing to the lead,tin and copper in their compositions (knownfrom chemical analysis of similar glass in otherpanels). The materials used in making thepanels were the glasses themselves, the plaster(an adhesive mix of resin and crushedmarble), the pottery backing tiles whichprovided support and the wooden slats of thecrates in which the panels had been shippedin antiquity.

Preliminary experiments were carried out onthe fragments, as a result of which a treatmentfor the cleaning, consolidation and mountingof the panels was devised and applied to one of the smaller panels. The treatment con-sisted of removing the gauze backing from thesurfaces of the panel with ethanol, and thencovering the surfaces with polyester gauze.The fragment and its polyester support wasthen placed on a wooden stretcher and furthercleaned by soaking it in ethanol baths, whichremoved any foreign matter and any resinousdeposits on the surface of the glass. Afterthese baths, cotton wool swabs moistenedwith ethanol were used to remove the finaltraces of dirt. The fragment was then placedon a fibre-glass support and immersed indilute baths of ethanol and 15 per cent AYAF,polyvinyl acetate, in order to consolidate it.On removal from the final bath, excess consol-idant was removed with ethanol, and a secondframe was placed over the panel to completethe consolidating and mounting procedures.From this and a number of other experiments,it was clear that care would have to be takennot to saturate the panels with water, alcohol,acetone or any other solvent, since after


prolonged contact the fragments became weakand sticky.

The specific aims of the conservation proce-dures were defined as follows. To clean andconsolidate the panels as thoroughly as possi-ble, and to provide sufficient rigidity to allowthem to be exhibited; to protect the glass fromfurther deterioration; and where possible, toclean individual glass pieces so as to exposethe original colour and shape of the undeteri-orated glass.

Initially it was necessary to decide if anattempt should be made to separate each pairof panels (they had been crated two panels toa crate, placed face to face). Over thecenturies, the complex of materials supportingthe panels had coalesced so that each pair hadin effect become one. Thus when they hadbeen excavated only the reverse side of eachpanel was visible. It soon became evident thatit would be virtually impossible to separate thepairs of panels since such an operation wouldcompletely destroy the integrity of both faces.Thus it was decided to expose and treat one(reverse) face only, and as far as possible, toleave the (reverse) face of the second panelvisible for examination.

Previous attempts at removing weatheringcrusts from the glasses had consisted ofmechanically scraping the surfaces with smalltools. However, this was not only tedious andtime-consuming, but damaged the glass. It wastherefore decided to clean the glass fragmentswith the aid of an Airbrasive unit, using micro-scopic glass beads and crushed glass as theabrasives.

Once the panels had been superficiallycleaned, they were fumigated in order to killmicro-organisms which might have beenpresent as a consequence of the flooding andwet storage. This was carried out by placingthe panels in a sealed fumigation chambercontaining two dishes of water to cultivate themicro-organisms under a vacuum of 28 psi.Oxyfume 12, a fumigant–sterilant gas, was thenintroduced for a period of 15 hours. Followingthis treatment, the panels were removed fromthe chamber, the Masonite supports removed,and the panels placed on sheets of cardboard.They were then ready to undergo the cleaningtechniques which had been formulated.

First, the flannel wrapping was lightly moist-ened with water and cut away from the top

surface of the panel. Alcohol and acetonewere applied on cotton wool swabs to loosenthe layers of gauze, which had become fixedto the upper side of the glass fragments. Thegauze was then cut away with scissors andscalpels. The sticky resinous substance, whichcoated many surface areas, was removed withacetone. The panel was turned over and thecleaning procedures repeated. A sheet of fibre-glass screen was laid on the glass and adheredwith a 12 per cent solution of AYAF, polyvinylacetate, in alcohol; the solution also acted asa consolidant to some extent (Figure 7.7b).

The panel was left overnight so that thealcohol could evaporate. The panel was thenturned over in order that the other side couldbe cleaned mechanically and with acetone toremove gauze and other foreign matter fromthe surface. The surface was then carefullycleaned in the Airbrasive unit using crushedglass as the abrasive, using a pressure ofapproximately 60 psi with a powder flow of2.5–3.0. For stubborn areas the pressure wasincreased to 80 psi with a powder flow of3.0–4.0. Residual powder was blown off theglass by attaching a quick release 3 mm highpressure air nozzle to the air compressor,followed by further cleaning with alcohol onswabs where necessary.

In general, the Airbrasive unit proved tobe an excellent method for cleaning the opussectile fragments. The principal disadvantagewas the possibility of working right throughthe glass if the abrasive was concentrated onone spot for too long. However, within a fewseconds of beginning the cleaning of acorner or edge, it became apparent as towhether or not there was any glass remain-ing beneath the deteriorated layer. If therewas no glass present the treatment wasstopped immediately. Another problem wasthe tendency of the abrasive powder toadhere to the background material, fromwhich it was difficult to remove. Some of thepowder was removed with compressed airafter which the remainder was removed withcotton wool swabs moistened with alcohol,taking care that a minimum amount ofsolvent came into contact with the glass.When the cleaning had been completed, amore thorough consolidation process wascarried out, after which the panels weremounted for exhibition.


A wall of potters’ clay was laid around theedge of each fragment a few centimetres fromit, and also around the edges of any voids inthe interior. The wall was of the same heightas the fragments, so that the material used togap-fill missing areas would not have to befiled down with the risk of damage to origi-nal material. The panel, still on its layer offibre-glass screen, was then placed on a sheetof aluminium foil. Kerr Vel-Mix stone colouredwith Liquitex acrylic pigments and texturedwith sand was used to fill the missing areasand to form a surrounding support. Vel-Mix issimilar to dental plaster but is harder andmuch more durable. The mixing proportionsare 22–25 ml water to 100 g of powder(approximately 1 part water to 4 parts powderby volume). The powder was added to thewater and mixed for about 1 minute. It wasthen poured into the areas to be filled whereit settled out forming a smooth surface, andset in approximately 10 minutes. On setting,the texture was improved by working the Vel-Mix with wooden tools (Figure 7.7c).

Where necessary, the filler was reducedwith sanding discs attached to a flexible drill.The fibre-glass screening which extendedbeyond the newly formed border, was thentrimmed. The panel was turned over and alayer of filling material poured around theedges and into the voids to raise the level tothat of the fragments. The majority of this sidewas left exposed so that future study of theglass is still possible. The Vel-Mix was thencoloured with Liquitex pigments, and a thinlayer of Liquitex Matte Varnish applied to therestored areas in order to reduce the glossyappearance of the paint.

To protect the edges of the panel, whichwere vulnerable to chipping, it was edgedwith lead came secured with a layer of DowCorning Clear Seal (mastic). Where necessarythe lead cames were cut and soldered to fitthe panel, soldered areas being smoothed withsteel wool. To restore a glassy appearance tothe surfaces of the cleaned glass, each piecewas coated with Paraloid B-72 in acetone;Paraloid B-72 was not, however, applied tothe opaque surface crusts.

Upon their return to Greece in 1976, theopus sectile panels from Kenchreai were in astable condition. The panels had been cleanedas thoroughly as possible, so that either the

original colours were visible, or a clear under-standing of the form and design of thefragments were apparent. Any extraneousforeign material had been removed, eithermechanically or chemically, and the panelshad been treated so that any further deterio-ration would be retarded (Corning, 1976).Mounting the panels on screens of fibre-glassenabled them to be framed for display (Figure7.7e), or more readily handled for studypurposes (Rothaus, Brill and Moraitou et al.,in Kordas, 2002).

Cleaning and repair on site

Ideally, first-aid treatment should beginimmediately glass is exposed, not after it hasbeen lying in a ‘finds tray’ for several hoursor days. Cleaning of glass finds on site shouldnormally be kept to the minimum required todefine them, since they should be thoroughlyexamined both before and during cleaningand this is difficult and too time-consuming toachieve during an excavation. Exceptions tothis rule may be the cleaning of large quanti-ties of glass waste (cullet) from glass-makingsites or shipwrecks; and cleaning of importantglass finds prior to marking. Where cleaninghas to be carried out on site, laboratorymethods should be used, adapting them to suitlocal conditions. Deposits are usually mucheasier to clean off the glass when they aredamp as they will be softer; if allowed to dryout, the soil or other deposit may contract anddamage the delicate glass surface.

For example, large quantities of a variety ofancient glass have been recovered by theInstitute of Nautical Archaeology in Turkey,from a Late Bronze Age shipwreck (c.1350 BC)lying in 44–51 metres of water off the cape ofUlu Burun near Kaș in Turkey; and ofmedieval glass from a wreck lying in 36 metresof water, within the rocky confines of SerçeLiman, a natural harbour on the southernTurkish coast opposite the Greek island ofRhodes (Bass, 1978, 1979, 1980, 1984). Theglass from the shipwrecks was mechanicallycleaned by hand, using small brushes andscalpels (Figure 7.8). Occasionally the carefuluse of a hammer and small chisel was neces-sary to separate large masses of glass from thesurrounding concretion (Figure 7.9). During


this process the glass was kept wet anddesalted from sea to tap water. The glass wasfound to be in good condition, and afterdesalting could be dried naturally and repaired(Pannell, 1990).

In considering the repair and/or reconstruc-tion of broken glass artefacts, i.e. the physicaljoining together of fragments, there are threeimportant factors to consider. The condition ofthe glass itself, the type of adhesive to beused, and the method of supporting the vesselwhilst the adhesive sets. On-site repair of glassmay have to be undertaken if this is the onlytreatment it will have, and/or the glass isrequired for the purposes of study, drawingfor publication, display, or to maintain thecontinuity of several sherds, which jointogether. A point to remember is that the glasswill take up more storage space as partially orfully repaired objects, than it will as fragments,and may be more likely to become furtherdamaged, than a properly stored group ofsherds.

In attempting to join decayed glass, penetra-tion of the deteriorated layer along the edgesto be joined, and strong bonding with theremaining glass core by the adhesive, areessential, otherwise the weathered layer willsimply pull away from the glass at a later date.Thus the chosen resin must act both as anadhesive and as a consolidant. Filling gaps iniridescent or consolidated glass is extremelydifficult and should be attempted only wherethis is absolutely necessary for the safety of

the object. In choosing treatments for individ-ual conservation problems, the conservatormust be aware of the condition of the glass,and understand the effects of the processesand materials used.

Marking glassMarking glass on site is difficult, and time-consuming, if glass is present in large quanti-ties. Dry glass which is neither crumbling norflaking, but which has a porous surface, canbe marked by first creating a writing surfaceon a small area of the artefact. This is achievedby applying a coat of lacquer or consolidantto the area to receive the marking. When dry,the data is applied using black waterproof,fibre-tipped pen, or ink applied using a fine-pointed wooden stick as an applicator in ordernot to scratch the glass (Figure 7.10). A


Figure 7.8 Retrieving glass beads from an amphorafound on the Bronze Age shipwreck at Ulu Burun,Turkey. (Courtesy of Institute of Nautical Archaeology,Texas).

Figure 7.9 Glass objects found in the shipwreck atSerçe Liman, Turkey, being removed from a mass ofconcretion by the careful use of a small hammer andchisel. (Courtesy of the Institute of NauticalArchaeology, Texas).

second layer of lacquer is applied over theink. It is probable that most excavated glasswill be unsuitable for direct marking withoutsuch surface treatment. (Glass with a soundsurface, particularly historical vessels etc.which have not been buried, may be markedas described above, but without the applica-tion of a lacquer to the surface.) If the glassis damp or wet, the data should be recordedon a waterproof label (using a waterproofpen) placed in the Polythene bag containingthe glass, and repeated on the exterior of thebag.

Storage of dry and treated glass

Consideration must be given to the questionof long-term storage of glass fragments orobjects, which should be such that evenuntreated glass will be protected from furtherdamage. A small proportion of glass from each

site should, where possible, be conservedsimply by correct packing and storage, withoutany intervention in terms of cleaning or intro-duction of synthetic resins, in order to providea source for future analysis of uncontaminatedmaterial, should this be required.

Glass from dry deposits should bemaintained in dry conditions. The fragmentsshould be placed in self-seal Polythene bagsand laid horizontally in strong cardboardboxes padded with Plastazote inert polyesterfoam or pads of acid-free tissue paper. Cottonwool (US: surgical cotton or cotton batting)must be avoided in direct contact with theglass as its threads can be difficult to removefrom delicate artefacts. Entire vessels shouldbe well padded with acid-free tissue andpolyethylene foam and placed in strong boxeswith well fitting lids (Figure 7.11). To preventfurther dehydration, excavated glass shouldnever be stored in less than 42 per centrelative humidity (RH), even if the glassappears to be well preserved.

It is conceivable that glass may be recov-ered from conditions of low humidity, whichhave caused it to become permeated with anetwork of fine cracks (see Chapter 4), thoughnone has been reported at the time of excava-tion. Such cracks may also result from stressesset up by the weight of soil. In some cases itmay well be that all that holds the glasstogether is a small amount of chemicallycombined water or surrounding soil, so thatthe glass may disintegrate on lifting.


Figure 7.10 Marking glass with its site identification.

Figure 7.11 Stem of an excavated wine glass packedin polyethylene foam and a strong cardboard box.(Courtesy of O. Theofanopolou).

Treated excavated glass, although robust,should never be stored in either excessively dryor wet conditions because further dehydration orhydrolysis can occur through a layer of consoli-dant. The glass should be kept cool becauseconsolidants and adhesives may begin to creepas the temperature rises, or become tacky andpick up dirt, or adhere to packing materials. Toavoid this last problem, consolidated glass couldbe covered with acid-free tissue or silicone paper.

Sherds should be stored horizontally inperforated Polythene bags. Vessels must bestored in dust-free cases or boxes withadequate padding of inert foam or acid-freetissue paper (never newspaper or cotton wool).Where glass is patinated or iridescent, handlingmust be kept to a minimum to prevent furtherdamage to the surface. The humidity should beregularly monitored by the installation of athermometer and a hygrometer. It is normallysufficient to note the readings of these instru-ments at regular intervals, perhaps twice a day.However a continuous record can be providedby installing a thermo-hygrograph, whichrecords both temperature and relative humidityon a paper chart. Such instruments require re-calibrating every month. The environment forstorage and display of artefacts is discussed indetail by Thomson (1998).

Problems of storage may arise when glass isfound in association with some other materi-als, such as metals, (e.g. enamels) or ethno-graphic material (Lougheed and Shaw, 1986).Recommendations for packaging and storagecan be found in Watkinson and Neal (1998),Sease (1988) and Cronyn (1990).

Laboratory treatments

The conservation treatments carried out in alaboratory may include any or all of the follow-ing processes: excavation, examination andreport writing (including photography), clean-ing, consolidation, repair, and possibly alsorestoration with synthetic resins or glass replace-ments (Fisher and Norman, 1987; Newton andDavison, 1989; Hogan, 1993; Koob, 2000).

Examination

Visual examination is one of the fundamentalprinciples of archaeological conservation.

Before and during cleaning, the conservatorshould look for clues as to the technology,decoration, use, type of burial deposit andassociated material, as well as the actual shapeand state of deterioration of the artefact.Records must be kept at the time of anyfeatures noted. A good light source and theaid of a �4 magnifying lens or a binocularreflecting-light microscope are essential (seeChapter 6).

Cleaning

As previously mentioned, cleaning ofexcavated glass involves the removal ofobscuring soluble and insoluble deposits toreveal the shape, decoration and originalsurface of the glass. It may be virtually impos-sible to remove disfiguring deposits fromfragile glass even if the glass is consolidatedprior to or during the cleaning process (seeFigure 4.7). Routine cleaning of glass in collec-tions may also be necessary, or the removalof old restoration materials prior to treatment.The parameters governing these cleaningmethods are relatively clear compared withthose concerning the removal or non-removalof a decayed weathering crust.

When glass deteriorates in the ground, itdoes not increase in volume and thus theoriginal surface and dimensions of the glasswill be represented by the deterioratedsurface. It follows therefore that if the wholecrust is removed, or if part of a thick crust isremoved in order to create a smooth patina,the dimensions of the artefact are irretrievablyaltered. Even though it may not be immedi-ately apparent, the deterioration may haveproceeded right through the glass and removalof the crust in such a case could lead tocomplete destruction of the object (see Figure4.23). In the case of cuprous red glass, whichhas become superficially green, there may bean argument for removing some or all of thesurface, in order to reveal the original colour.Such treatment was carried out on escutcheonson Anglo-Saxon hanging bowls from SuttonHoo in Suffolk (UK) (Bruce-Mitford, 1975). Inthe case of a hanging bowl from Lincolnwhere a similar problem of discoloration wasencountered, the original surface was retained(Foley and Hunter, in Bacon and Knight,1987). It is possible that a green crust need


not be removed completely in order to revealits true colour, since the last traces could berendered transparent by consolidation with aresin, a test first being made with toluene toascertain if enough crust has been removed.However, removal of any or all of the crustwill destroy the flush finish of the glass in itssetting, and this discrepancy in level may haveto be made up with a suitable resin tomaintain the coherency of the design. Thusremoval of weathering crusts on enamels isnot to be recommended in general. Hughes(in Bacon and Knight, 1987) has shown thatimportant archaeometric data can be retrievedfrom the weathering crusts of enamels.

Nevertheless, since the weathering crusts onglass are often iridescent, opaque or black,and hence the original colour and trans-parency cannot be seen, removal of crusts hasoften been advocated on aesthetic grounds.Although this cannot necessarily be condoned,it may be that the removal of light opaquepowdery deposits, which do not form asubstantial surface of the original glass, maybe justified. Should removal of deteriorationcrusts become necessary it is probably bestcarried out mechanically, as this treatment canbe confined to small areas, by using scalpels,wooden picks and soft paint brushes.

Removal of soluble or removabledepositsThe removal of soluble salts from glass from wetland sites or marine excavations was describedearlier as a treatment that could be undertakenduring a period of temporary storage. However,if salt removal has not already been achieved inthe field, it must be carried out in the labora-tory as soon as possible.

If glass has been lifted from an excavationinside a protective casing, its micro-excavationis carried out in the laboratory. If the glass isparticularly badly deteriorated, then only smallareas at a time are exposed. These are cleanedand consolidated before the next is uncovered.Where it has been necessary to consolidateglass prior to cleaning, the overlying depositsmay be softened by placing the artefact in avapour of the solvent in which the consolidantwas dissolved, for a short period of time. Thedeposits can then be removed mechanicallywith a soft sable brush, a scalpel and solventon cotton wool swabs.

Glass bearing unfired pigments, or flakingpaint or gilding must be treated with the great-est of care, and must not be washed. It maybe desirable to consolidate the decoration, andto re-affix flaking paint or gilding, with aconsolidant, bearing in mind that such treat-ment would be irreversible.

By careful cleaning of a small area it shouldbe ascertained whether or not decayed glassis actually being held together by dirt orcalcium concretions. These may fall offunaided, but may be encouraged to do so byusing small wooden toothpicks and softbrushes. Barely damp cotton wool swabs ofdistilled water, propan-2-ol, or industrialmethylated spirits (IMS) can be used toremove the last of the dirt but care must thenbe taken not to use excess amounts of IMS orblanching of the glass may occur (Dowman,1970). Lal (1962–63) suggests binding flakingglass with cotton thread before it is cleaned inorder to keep any loose fragments in place.However, this is likely to cause physicaldamage to the glass. In general, dirt should beleft on the glass rather than risk the removalof a flaking surface (see Figure 4.7). Soundarchaeological glass may be washed in water,preferably distilled or de-ionized. Artefactsfrom muddy deposits, especially urban excava-tion, are often covered in a film of grease,which should be removed from the glass usingwater or other solvents before it is dried,consolidated or repaired. In rare cases, a 25per cent aqueous solution of hydrogen perox-ide may be used to loosen stains and dirt heldin cracks on this type of glass. Metal depositsmay be removed by mechanical methods, orby the controlled use of chemicals. Howeverthey may be so bound up in the decayed glasssurface, that they are best left untreated. TheRoman glass flask shown in Figure 7.12 wascleaned using a wooden pick. It had beencovered with a deposit of iron (see Figure4.9), and the ease with which the deposit fellaway, exposing an undeteriorated glasssurface, raises a question of its authenticity.

It is sometimes possible to remove mudtrapped inside hollow parts of an undecoratedbroken vessel, by flushing it out with jets ofwater from a pipette. Alternatively if the glasssurface appears to be sound, the object maybe immersed in water in an ultrasonic tank fora few seconds at a time, in order to flush out


the cavities. Ultrasonic transducers generatehigh-frequency (non-audible) sound waves,typically 20–50 kHz, which interact in a liquidto produce minute bubbles. These bubblesexpand and contract violently and it is theshock waves produced by this action whichspeed up the loosening of dirt. This effect,combined with the vibration induced by theultrasonic waves in friable materials, can causerapid disintegration of the structure of thematerial. For this reason it is most useful forremoving mud and porous encrustation, butcan also cause severe damage to friablesubstrates. It is essential that the progress ofcleaning should be constantly monitored.

Ultrasonic cleaning has been successfully andsafely used on medieval window glass in goodcondition (Gibson and Newton, 1974), butdamage has been shown to be severe onmedieval painted window glass, and there isalways the potential danger of promotingcracks. In particular it should be noted that theaqueous cleaning solutions recommended bysuppliers of ultrasonic equipment are frequentlyharmful to glass. Only mild cleaning agentsshould be used, especially when such a power-ful cleaning method is employed. Occasionallyencrustations have been removed from glassobjects using the Airbrasive technique.

Dilute concentrations of acids have beenused in attempts to remove carbonate, and

even iron and manganese, from within theweathering crusts, in order to produce bright-ened silica surfaces. However, the resultingeffervescence tended to disrupt the fragileweathering crusts, and the blackening seemedlittle improved.

The only method attested for the removal of lead sulphide (PbS) was carried out at the Conservation Analytical Laboratory of the Smithsonian Institution (SmithsonianInstitution, 1969). A blackened, nineteenth-century goblet was lightened by immersion indilute hydrochloric acid for one minute; thiswas said to convert the PbS to yellow sulphur,much of which could be removed with a softbrush. Any sulphur caught within the weath-ered layer was dissolved by immersion incarbon disulphide (CS2) for 24 hours, leavingthe glass colourless but hazy because theweathering had been made visible (note thatCS2 is very toxic, with a TLV of 20 ppm).

Dilute mineral acids have been used toremove weathering crusts entirely; however,even dilute acids may affect the remainingglass and will penetrate deep into the glassthrough cracks and flaws endangering thewhole object.

Lîbiete (1998) reported two methods ofcleaning glass beads dating from between thethird and seventeenth centuries, which hadbeen excavated from the Daugmale Castlemound in Latvia. The first method involvedsoaking the beads in a warm solution(40–50°C) of acetic acid, and then rinsing themin distilled water. In the second, the beadswere boiled for 5 minutes in alternatesolutions of acetic and alkaline solutions (3per cent acetic acid and 3 per cent potassiumhydroxide, after which they were boiled inchanges of distilled water to remove residualchemicals. The beads were dried by immers-ing them in ethyl alcohol for approximatelyone hour, removed and left to dry in air. Beadfragments were repaired with acrylic adhesive.The beads were coated with a 7 per centsolution of polyvinyl butyral (C2H5OH) toprevent further decay.

Subsequent examination confirmed the effec-tiveness of the methods, and that the deterio-ration process was halted without changing theglass structure. However, this may only be shortterm, and such drastic intervention is not to berecommended. It is possible that sequestering


Figure 7.12 The Roman glass flask shown in Figure4.9, after removal of iron deposits.

agents in the form of pastes could be usedeither to remove weathering crusts or selectiveions within them. For example Catechol, 1,2-dihydroxtbenzene, C6H4(OH)2, sequesters silica,and tetrasodium ethylenediaminetetra-aceticacid (EDTA), (HOOC.CH2)2N.CH2, sequesterscalcium and magnesium, but these have bothbeen shown to attack glass at high pH values,even though the damage may not be visible atfirst. It should be remembered that sequester-ing agents will also affect enamelling andgilding, and hence if employed, they must beused with the greatest of caution.

Repair

Repairs to archaeological glass may follow thesame general procedures as those describedfor decorative and utilitarian glass (see below).However, it is unlikely that epoxy resin canbe used to repair the majority of fragilearchaeological glasses since it is too strong,and difficult to remove at a later date, withoutcausing damage. Repairs are generally madeusing a cellulose nitrate based adhesive, e.g.HMG (UK), Duco cement (US), or polymethylmethacrylate, e.g. Paraloid B-72 (Acryloid B-72- US) (Figure 7.13a,b).

Consolidation

Consolidation of deteriorated glass or fragilesurface decoration with a resin may be under-

taken when the glass is liable to crumble orflake upon drying or handling. It has alsobeen used to re-introduce a measure of trans-parency to thin surface crusts. The treatmentshould be considered to be irreversible, evenwhen resins that are normally re-soluble, areused. This is because resins swell duringsolvation and the glass is unlikely to surviveboth this swelling and the prolonged immer-sion required to remove a consolidant. It isalso unlikely that all traces of a resin could beremoved if required at a later date.Consolidation should therefore only be carriedout where it is absolutely necessary for thesurvival of the artefact. The question ofwhether to consolidate before or after clean-ing is problematic. Pre-consolidation maymake subsequent removal of surface dirt froma fragile artefact more difficult. On the otherhand, it may not be possible to clean fragileglass before it has been consolidated.Therefore each case must be judged on itsown merit.

It is not only necessary to consider the glasscrust as the substrate; there may also be asolid glass core to which the consolidant mustadhere. The surface of the core is normallyassociated with water molecules, but (theamorphous silica of the weathering crust beinghydrophilic) it is usually highly hydrated.Thus, to obtain enhanced penetration andadhesion, a consolidation system whichdisplaces or incorporates this water must be


Figure 7.13 (a) A fragile and flaking glass bowl, (b) repaired piece by piece, with cellulose nitrate adhesive.

(a) (b)

employed. Preliminary drying out of a crust,especially by heat or vacuum, must be avoidedsince this would lead to shrinkage and disrup-tion. In general, the approach to the consoli-dation of extensively weathered glass has beeneither that of preliminary de-watering withsolvents, or that of using a water-miscibleconsolidation system, and of choosing aconsolidation system which combines lowviscosity with a high deposition of solids onsetting.

The term consolidant is used to define a resinused for the purposes of consolidation, andconsolidation system to define the liquid phase inwhich the consolidant is applied, whether it bemolten, in solution, or as a polymer pre-treatment.When choosing a consolidation system, theconservator must bear in mind the nature andproperties of the resins and solvents to be used(see Chapter 5), the condition of the decayedglass to be treated, and the manner in which theconsolidation system is applied (Figure 7.14a–h).


Figure 7.14 Options forconsolidation of porous glassobjects. (a) Brushing; (b)spraying; (c) dipping; (d)adding a consolidation systembeside an object; (e) partialimmersion; (f) totalimmersion; (g) vacuumimpregnation; (h) drying anobject after consolidation.

Systems that are too viscous fail to penetratethe glass, resulting in the formation of skinswhich spall off, whilst those with too low asolids content fail to fill the interstices.Viscosity varies with consolidant, and withconcentration of the consolidation system.Thus, for example, Paraloid B-72 in a 20 percent solution of toluene at 21°C has a viscos-ity of only 29 � 10–3 Nsm–2. Polyvinyl acetate(of approximately 520 degrees of polymeriza-tion) in a 20 per cent solution has a viscosityof 40 � 10–3 Nsm–2. Reduction of viscosity canbe achieved by a number of means, mostcommonly by altering the concentrations ofthe consolidation system. Another methodwould be the application of heat, but temper-atures in excess of 60°C must be stringentlyavoided in order to safeguard the glass; andconsideration must be given to the vapourpressure and flash point of the solvent in use.

If the glass in question is not completelywaterlogged, that is, the decayed surface layerscontain air, then the consolidation system maybe best applied under conditions of reducedpressure. The glass can either be immersed ina consolidant and the pressure reduced, or thepressure can be reduced and the solutiondripped beside the glass, until it is covered.The former method may give better support toa whole, fragile vessel but greater disruption islikely when the pressure is reduced. If thedripping method is used, for example, in caseswhere the glass is liable to lose flakes insolution, sufficient consolidant to cover thefragment should be dripped into the consoli-dation chamber before the vacuum is released.

Reduced pressure is used primarily toremove air from the pores and capillaries ofdecayed glass, which would otherwise impedethe penetration of the consolidation system.Also, to introduce a low pressure into thepores so that, when the assembly is returnedto atmospheric pressure, more of the consoli-dation system is forced into them due to thepressure differential. This lowest practicalpressure will depend in part on the apparatusand pump available, and in part on the vapourpressure of the liquid at the given temperature(because a liquid boils when the ambientpressure reaches its vapour pressure); solventswith high vapour pressures will evaporateeven when the pressure is only slightlyreduced. However, it should be remembered

that the vapour pressure of a solution is lowerthan that of a pure solvent. Reduced pressurewill therefore tend to concentrate the solutionand, if taken to excess, it could cause theconsolidant to solidify. An excessively lowpressure should, however, be avoided becausethere is a danger that any sealed bubbles(seed) which are very close to the surface ofthe solid glass might burst.

The atmospheric pressure should bereduced only slowly in order to prevent airrushing from the glass and causing it to disin-tegrate and to allow time for the pressure inpartially sealed pores to come to equilibrium.This is especially true when the glass isimmersed in a liquid before the pressure isreduced, because the viscosity of the liquidretards the process. Again, the reducedpressure should only be slowly increased inorder to avoid air rushing into cavities wherethe solution has not fully penetrated. Theequipment for consolidating glass underreduced pressure is shown in Figure 7.14).

After consolidation by immersion, the glass isremoved from the consolidation system with apair of plastic tweezers, and dipped briefly insolvent to dilute surplus consolidant of thesurface of the object. The glass is placed onsilicone (US: glassine) paper and excess consol-idant allowed to drain off the surface. Thisprocess may be aided by the use of cottonswabs soaked in solvent or simply by absorbingsolvent with tissue. It is then advisable to replacethe object in a solvent atmosphere whilst it driesout, either in a glass receptacle or, if in theground, by covering the glass with a Polythenebag (Garlake, 1969). Drying the glass in this waywill slow the rate of evaporation so that the resinremaining on the surface will be more thinlydispersed and therefore appear matt. Coveringthe glass also prevents dust from adhering to theconsolidant (see Figure 7.14h).

Generally speaking, a matt surface willrender surface details, such as engraving orpaint, more readily visible on an unevenweathered surface. Hedvall et al. (1951)achieved such a finish by gently heating theconsolidant bath, but this is an unnecessaryprocedure, adding the potential risk of damageto the glass, or of causing a fire.

Spraying, dripping or painting the consoli-dant solution on the glass are far less effica-cious than the methods described above but


they might have to be used on individualfragments or objects on site; or in the case oflarge or very delicate objects (Figures 7.14a–cand 7.15). In these cases, the first applicationsmust have a low viscosity in order to achievedeep penetration, and hence several dilutecoats are much more helpful than one thickone. A pipette or syringe may be useful forimpregnating areas of an object that are diffi-cult to reach with a brush.

Silane as a pre-treatment has been success-fully used as a coupling agent for the bondbetween hydrated glass, weathering crusts andthe acrylic consolidant Paraloid B-72 on exten-sively weathered glass by Errett et al. (1984).The long-term stability of the technique isunknown, and it is unlikely to have a wideapplication on archaeological glass, beingmore suited to the conservation of windowglass in situ.

A method of consolidating fully decayedglass was developed by Bimson and Werner(1971) for the treatment of a fragile potashglass alembic head dating from the fourteenthor fifteenth century AD. The alembic head hadbeen successfully excavated, along withseveral other fragments of medieval distillingapparatus, but had been subsequently severelydamaged by poor storage. More than half theobject had been crushed to tiny flakes and theremainder was in an extremely fragile condi-tion and could not be safely handled. Theproblem was not only to consolidate thosefragments so that they could be safely

handled, but also to use a consolidant thatwould serve as a gap-filler, because most ofthe fragments were joined only at one or twopoints. Glass and amorphous silica arehydrophilic, and therefore Carbowax 6000, apolyethylene glycol wax which is alsohydrophilic, was chosen as the consolidant.

The procedure used was as follows. Eachfragment of glass was supported on a raft ofaluminium foil. This was laid on the surfaceof the molten Carbowax 6000 in a dish, placedin an oven at a temperature of about 80°C. Itwas feared that the Carbowax 6000, having arelatively high viscosity, might disintegrate theglass further, but this did not occur. After 2hours the raft with its fragment of glass hadsunk to the bottom of the molten wax and allthe bubbles of displaced air had separatedfrom the glass surface. The fragments werethen gently removed from the wax bath bylifting the corners of the raft, and replaced inthe oven on a pile of absorbent tissues. Whenexcess molten wax had run off, the aluminiumraft was carefully removed, and any remainingwax blotted away with paper tissues. Thealembic fragments were allowed to cool toroom temperature, when it was found thatthey had sufficient mechanical strength to behandled safely, and their appearance showedremarkably little change.

In order to join treated fragments, the waxon their edges was melted using a Microweldtorch, which restricted the heating area to afew square millimetres. This enabled wax inany given area, to be melted without soften-ing joins already made, and badly fitting joinsto be strengthened by melting small lumps ofthe wax to fill gaps. Whilst the treated glasswas not in a strong physical condition, severalyears later it was possible to handle it in orderto fit mounts for display.

The following method for temporary storageand consolidation of wet excavated medievalwindow glass was developed by Hutchinson(abstract No. 416 in Newton, 1982b) in theEnglish Heritage conservation laboratory inLondon (formerly the Historic Buildings andMonuments Commission for England). Onexcavation, the glass was kept wet and packedin single layers between sheets of polyesterfoam, of a grade that retained water well. Theglass was then placed in two self-sealPolythene bags, one inside the other, and


Figure 7.15 Application of a consolidant (10 per centParaloid B-72 in toluene) to flaking glass by brush.

labelled both on the outside and on a water-proof label inside the outer bag. The glass waskept in a cool place. (If a refrigerator is usedto store the wet glass there must be no possi-bility of the water freezing.) During temporarywet storage the water surrounding the glassshould not be changed; it was found that thepH of the solution does not rise above 8.0.

Once in the laboratory, the glass wasunpacked and whilst being kept wet as muchmud and loose encrustation as possible wasremoved mechanically. The fragments werethen consolidated with a mixture of PEG 4000(polyethylene glycol) and Vinamul 6815, apolyvinyl acetate emulsion. This acted as abulk filler and consolidant. The mixture wasprepared as follows: the required amount ofVinamul 6815 was diluted with an equalamount of distilled water and the requiredamount of PEG 4000 was dissolved in waterin the proportion 500 g to 1 litre of water. Theconsolidation system itself was composed offour parts diluted polyvinyl acetate emulsion,two parts distilled water, and two partspolyethylene glycol solution, all parts beingmeasured by volume.

The glass fragments were laid flat on coarseNylon netting in a shallow container. Thenetting ensured that the consolidantcompletely surrounded the glass. It wasallowed to overhang the sides of the containerfor ease of lifting the delicate fragments aftertreatment. Several layers of net and glass couldbe placed in one container. The consolidantwas then slowly added to the dish, ensuringthat the fragments did not float (they did notif waterlogged), until there was at least 10 mmof liquid above the glass. The top of thecontainer, clearly labelled with its contents,was covered with plastic film to preventevaporation, and left for approximately 2months. After this time, the fragments wereremoved from the consolidation system, andexcess wax removed with paper tissues. Therewere some loose flakes of glass, which couldnot be re-attached. The glass was then placedon silicone paper and allowed to dry. Repairswere carried out at this stage, using a cellu-lose nitrate adhesive (HMG). Finally all thesurfaces of the glass (including the edges)were painted with a 30 per cent solution ofpolyvinyl-acetate in toluene. Since toluene isless volatile than, for example, acetone, the

solution did not form into brush lines on thesurface while drying. Lacquering the glass inthis way altered the optical appearance of thesurface, thus enabling the painted design to beseen through the thin layers of dirt anddecayed glass remaining. Glass treated by thismethod has been found to be in good condi-tion several years later; favourable storageconditions will be a major contributing factorto this. Experiments with the method arecontinuing; however, it may be that the use ofParaloid B-72 and aqueous solutions of acrylicemulsions described below will prove to bemore widely accepted.

Koob (1981) began experimenting withacrylic colloidal dispersions as a consolidantfor newly excavated archaeological bone.Since the smaller particle size and low viscos-ity of the colloidal dispersion system permit-ted better penetration at higher concentrations,it was concluded that they were the mostsuitable consolidant in a water-based systemfor the consolidation of fragile materials,especially bone and other organic materialsbecause of their near neutral pH. Furthermore,resistance to high temperatures, and charac-teristically good working properties, areassured by the glass transition temperature,and the minimum film temperature respec-tively. In addition, the relatively high moisturebarrier, due to an acrylic film’s low waterpermeability and absorption, protected objectsagainst climatic fluctuation. After drying, theacrylic resin film was hard and durable.

In 1983 an aqueous solution of Primal WS-24 acrylic emulsion was used to consolidatefragments of excavated glass beads andpainted window glass in Denmark (Roberts,1984). The glass beads, dated to between AD

400 and 700, arrived for treatment packed ina self-seal Polythene bag. Since condensationhad formed on the inside of the bag, it wasat first difficult to determine their exactquantity and condition (Figure 7.16a).However, it could be seen that the fragmentswere in an extremely friable condition (due inpart to improper lifting and packaging), andthat they would require consolidation beforebeing cleaned. To this end, a number of testswere carried using 20, 50 and 100 per centsolutions of Primal WS-12, WS-24 and WS-50acrylic emulsions on silica gel lying on dampsoil to represent the beads. A 50 per cent


aqueous solution of Primal WS-24 was chosenas the solution with which to consolidate theglass under moist conditions.

The bag containing the bead fragments wasthen placed on a glass plate for support andcut open to reveal the contents to be one glassbead and one bead fragment. The exposedsurfaces of the beads were consolidated witha few drops of 100 per cent Primal WS-24 (i.e.as supplied). While the earth was still moistfrom condensation and the Primal solution, itwas removed from the intact glass bead, usingsoft brushes and distilled water. Cleaning inthis manner was terminated if the glass beganto split, and consolidation was renewed witha 50 per cent aqueous solution of Primal WS-24. The consolidation medium was applieddrop by drop, until the glass was saturated,and excess was removed with cotton woolswabs. In order to slow down the evaporationrate of the water, the bead was placed in anopen plastic bag under a slightly raised glassbeaker. After an hour the bead was transferredto silicon release paper (US: glassine paper)and was further consolidated with a 50 percent aqueous solution of Primal WS-24. Whenexcess was apparent, consolidation was termi-nated. The glass bead was allowed to dryovernight (Figure 7.16b).

Final cleaning was accomplished by dissolv-ing the consolidant on the surface, withtoluene, and cleaning the bead with a softbrush. The cleaned surface was then re-consolidated with an 8 per cent solution ofParaloid B-72 in toluene glass (Figure 7.16c).After treatment, the glass bead could be seento be turquoise in colour with a marvereddesign of bands of white and orange-red.

A method of consolidating excavated glasswith Primal WS-24 and Paraloid B-72 was alsoused to conserve a fragment of a painted glassroundel (kabinettscheibe) from Slagelse inDenmark (Roberts, 1984). The roundel had adiameter of 72 mm and was decorated with afired enamel design and script on the front,and with evidence of a silver stain on thereverse. The fragment was received for treat-ment in a waterlogged condition, but hadsubsequently been allowed to dry out. Upondrying, the glass surface and consequently thepaint layer began flaking and cupping on bothsides of the fragment (Figure 7.17a).Treatment was begun using a 3 per cent


Figure 7.16 (a) Excavated wet glass bead fragments ina self-seal Polythene bag, obscured by gravel. Lightcleaning being undertaken with a sable brush. (b) Glassbead after consolidation with 50 per cent Primal WS-24.(c) Glass bead after removal of excess solvent.

(a)

(c)

(b)

solution of Paraloid B-72 in toluene applied toselected areas and allowed to penetrate thesurface layers by capillary action. Since it wasobserved that moisture relaxed the flakingsurfaces of both the glass and the paint layer,a water-soluble acrylic consolidant, Primal WS-24 (as a 20 per cent solution in water), wasintroduced with a brush. Capillary actiondistributed the consolidant. After drying, thesurface of the glass was brushed with asolution of 8 per cent Paraloid B-72 in toluene.In areas of heavy flaking, such as the paintedscript, the surfaces of both the glass and paintwere gently warmed with a heated spatula(temperature controlled at 25°C), in order toaid the flattening and relaying of the paint.

During this process, some of the flakes broke.The results can be seen in Figure 7.17b.

The use of an electrically heated spatula,applied to the glass and/or paint usually overa small piece of Melinex sheet, in conjunctionwith small amounts of solvent and/or consol-idant to relax the surfaces, is often successfulin relaxing and relaying flaking designspainted on glass (see also hinterglasmalerei atthe end of this chapter).

Surface treatmentsOther methods have been used to revealsurface detail. A thick layer of polyvinylacetate was applied to the surface of theblackened sherds of medieval window glass toreveal paint (Hutchinson, 1981; Newton, 1982bentry No. 416); and the application of a glossyresin (Plexigum P24) over the primary treat-ment, a fluoro-silicate, Durol-Polier-Fluat SD(Karl, 1970). However, none of these methodscan be recommended for reasons previouslydiscussed.

In order to adhere flaking glass surfaces andcracked enamels, solutions of various resins,such as polyvinyl acetate or the methacrylatecopolymer Paraloid B-72, have been appliedas a surface treatment (Wihr, 1977; Dove,1981). Improved transparency of decayed glasscrusts on small glass objects has beenvariously achieved by surface treatment withParaloid B-72 after solvent de-watering, orwith epoxy resin after air-drying. The successof such an operation depends upon the condi-tion of the glass undergoing treatment. Whilstsuch treatment may also reduce the visualeffect of iridescence, neither it nor the consol-idation systems previously discussed canrestore transparency to a thin weathering crustcompletely, and are not recommended forgeneral use.

The use of organic lacquers has beensuggested as a protective film for weepingglass; and Hedvall et al. (1951) have describeda technique in which such a glass was impreg-nated with a polymethyl methacrylate undervacuum. Lacquering may appear to delayfurther deterioration, but all organic lacquersare permeable to water vapour, and thechemical reaction of disintegration willcontinue. Thus the reaction products will betrapped at the interface, and their build-upmay eventually cause the total collapse of the


Figure 7.17 (a) Flaking and cupping of a fragment ofpainted glass roundel (kabinettscheibe). (b) The samefragment after consolidation with a 20 per cent aqueoussolution of Primal WS-24, and surface treatment with 5per cent Paraloid B-72 in toluene.

(a)

(b)

glass object. The latest approach to conserv-ing unstable (weeping and crizzling) glass byalkali ion replacement is discussed later in thechapter.

Corvaia et al. (1996) describe the treatmentof an exfoliant surface layer on several greenglass bottles recovered from the wreck of theZuytdorp (1712) in 1988. The bottles wereextremely weathered and stained. Scanningelectron microscope/energy dispersive X-rayanalysis (SEM/EDXRA) and thermal analysisshowed that the hydrated outer layer of glasshad suffered a calcium loss of approximately21 per cent and that water represented 41 percent of the total mass of the glass. Treatmentof the surface layer with calcium acetatesolution was successful in re-depositing cal-cium ions into the weathered layer. Exfoliationof the degraded layer was prevented by treat-ment combining the calcium acetate solutionand an emulsion of Primal AC-235 (copolymerpolybutylacrylate/polymethyl methacrylate). Theaddition of Primal AC-21 slowed down the rateof calcium diffusion into the glass, but resultedin allowing the glass to be air-dried withoutexfoliation of the surface glass. (See also relay-ing of paint on hinterglasmalerei, at the endof the chapter.)

Consolidation of wet glassExperiments are continuing into the preserva-tion of glass from waterlogged environments;it seems that consolidation with acrylicemulsion, or with Paraloid B-72 after de-watering through solvents, will produce themost satisfactory results.

Allowing damp or waterlogged glass todehydrate normally results in loss of trans-parency, surface flaking, or even total collapseof the glass. Therefore, water in the surfacelayers and core of the glass must be displacedby a solvent. This in its turn will be replacedby or incorporated in the consolidationsystem, e.g. Paraloid B-72 in acetone.Hydrophilic solvents, such as ethanol, propan-2-ol, acetone or butanone, are suitable.Without allowing the glass to dry, it is firstplaced in a 50:50 mixture of water and solvent.Every few hours, it is passed through a succes-sion of baths of increasing concentration ofthe solvent in which the consolidant isdissolved. Finally the glass can be transferredto the consolidation system itself.

Reisman and Lucas (preliminary report,undated) carried out experimental researchinto retrieval and consolidation of glass fromunderwater environments in 1978. The methodof consolidation was as follows. The objectwas weighed on removal from its last waterbath, and then immersed in a bath of ethanol(anhydrous ethanol). The ethanol bath waschanged every few days to ensure that thewater was being replaced. After a fewchanges, the specific gravity of the ethanolbath was tested to ensure that it was water-free. When the ethanol exchange wascomplete, the glass object was re-weighed. Itwas then submerged in the consolidatingmaterial, PVA-AYAA 5 per cent in 95 per centethanol. The object was left in the consolidantfor 15 minutes, after which a vacuum wasdrawn to 5 in Hg and held for 15–20 minutes.The percentage concentration of the solutionwas increased to 7 per cent by adding morePVA–AYAA, the vacuum drawn to 5 in Hg, andafter 15 minutes, to 7 in. After a further 15minutes, the percentage of PVA-AYAA wasincreased to 10 per cent, the vacuum appliedand if possible the percentage of thePVA–AYAA further increased. The highestconcentration of consolidant possible consis-tent with impregnation should be achieved,especially if the glass surface is insecure.

The object was then removed from thevacuum and placed in a chamber over ethanolvapour for several hours to force the PVA-AYAAinto the surface of the crust even further and toprevent the surface from drying with a glossyappearance. During this time the object wasobserved to note any shrinkage, delamination orspalling, of which there was none. The objectwas then removed from the desiccator andallowed to dry at ambient temperature.

Reinforcing

It is possible that heavily decayed glass maynot be strong enough to handle even afterconsolidation, and that a support in the formof a backing material will be required.Working on very decayed glass from Sardis inTurkey, Majewski (1973) used Japanesemulberry tissue with polyvinyl acetate (PVAC),the resin which had already been used forconsolidating the glass. It would have beenpreferable to have chosen a resin that would


not dissolve in the same solvent as the consol-idant in case the backing material needed tobe removed at a later date. Artal-Isbrand(1998) filled losses in a fragile, opaque Romanglass with Japanese tissue paper coloured withwater-colour paints, and secured to the glasswith Paraloid B-72 (see also Figure 7.31).

Restoration

Gap-filling missing areas in archaeologicalglass is best undertaken by reinforcing theglass as described above, if the glass is partic-ularly fragile. Where a glass vessel has largeareas missing but does not warrant total recon-struction because, for example, it will remainin storage, it may be partially restored for safe

handling, e.g. during study, photography ordrawing for publication. Strips of fine glass-fibre tissue cut to size and impregnated withcellulose nitrate adhesive or epoxy orpolyester resins are used to bridge gaps in theglass and to hold floating fragments in theircorrect positions. Total reconstruction of smallvessels is also possible by this method, withthe results shown in Figure 7.31. It is notaesthetically pleasing but may be useful as atemporary measure (Newton and Davison,1989). Artal-Isbrand (1998) used Japanesetissue paper as a gap-filling material (seeabove). Where the condition of the glassallows, gap-filling may be undertaken by themoulding and casting techniques described forhistorical glass (Wihr, 1963, 1968, 1977;Davison, 1998; Newton and Davison, 1989).


Cleaning

Glass is an immensely versatile material,which, as discussed in Chapters 2 and 3, canbe formed and decorated with infinite variety.For at least a thousand years, glass has beenused to form items for storing and servingfood and drink. Other liquids stored in glasshave ranged from perfume, paint, ink andchemicals. In use for lighting, glass wasformed into candlesticks, candelabra, taper-sticks, wall lights, sconces, lanterns, lamps andchandeliers. It was used to form decorativeinlays and panels on furniture (Davison, 1992).Other decorative uses were for armorial,masonic and commemorative wares, animalsand figures, friggers, centre-pieces, such asepergnes, nefs and domes, paperweights andsculptures. For personal use, glass was usedto form looking-glasses, and jewellery in theform of beads, bracelets and rings, ear-rings,inlays and mosaic, and paste in imitation ofprecious and semi-precious stones. Thorensen(1998) discusses the use of glass in theproduction of classical intaglios. Glass paneswere used as substrates in the production ofplaques and panels, mirrors, paintingsexecuted in reverse, mirror pictures, verre

eglomisé and photographic images. Glass alsohas many varied uses scientifically, commer-cially and industrially.

As in every other field of conservation,certain generalities concerning repair andrestoration can be laid down. There are,however, certain glass objects, which, becauseof their shape, require specialist treatment. Yetothers, such as paintings in reverse on glass,mirrors and chandeliers, require the assistanceof other specialist conservators or companies.In these instances only general guidance willbe given, indicating the range of treatmentsavailable.

Removal of previous repair andrestoration materials

In the case of glass artefacts which have beenin a collection for a number of years, theremay be evidence of previous repair andrestoration. This often consists of the use ofadhesives ranging from animal glues to epoxy-and rubber-based compositions, with wax orplaster of Paris as the replacement for missingareas of glass, coloured with a variety of paints(Wihr, 1963), as shown in (Figures 7.18 and7.43).

Part 2: Historic and decorative glass

It may not always be possible for theconservator to know which materials havebeen used. Provided they can be removedwithout damage to the glass, a sample of thematerial can be removed and tested in orderthat the correct solvent may be chosen.Otherwise, the usual practice is to beginremoval of previously applied materials, withthe least toxic solvent, i.e. cold water, and ifthis has no effect to try warm water, followedby acetone, industrial methylated spirits (IMS),white spirit, xylene and dichloromethane. Theobject may either be exposed to solventvapour in a closed container, or the solventmay be applied directly by means of a smallsable paint brush, pipette, or cotton woolswabs laid along the joins (Figure 7.19a–f).Once softened, the material may be


Figure 7.18 Old, unsightly restoration on a wing-handled cup, gap-filled with plaster of Paris and paintedsilver. From Canosa, Italy. (© Copyright The BritishMuseum).

Figure 7.19 Options forremoval of materials used forrepair and restoration. (a)Mechanical cleaning with orwithout solvents, using ascalpel, cotton wool brush orcotton wool swab. (b) Cottonwool rolls soaked in solvent,or a paste of water andenzyme detergent, placedalong previous joins. (c)Vapour solvent inside a self-seal Polythene bag, or alidded container. (d)Immersion in solvent. (e)Removal of metal rivets. (f)Removal of lead-covered wireties.

Mechanical cleaningwith or without

solvents, using ascalpel, cotton wool

bud or brush(a)

Cotton wool stripssoaked in water orother solvent, or apaste of water andenzyme detergent

(b)

Solvent vapourinside a self-sealpolythene bag.

Object supportedon foam or tissue

Solvent vapourinside a glass

container(c)

Immersion in solvent.Object standing on

weighted foam(d)

Lid

(e)

(f)

WeightFoam

OR

Metal rivetsembedded in

plaster

GlassGlass

Glass Glass Glass Glass Glass Glass

Rivet cut with plierswhere possible torelease it

Plaster excavated

Lead cut with pliers andcut away with a scalpel,

wires cut and easedout of the glass

Lead coveringWire

Removalof wires

completely removed using a sable paint brushor small cotton wool swabs and solvent, or bycareful use of a scalpel. Force should neverbe applied since fragile glass may break beforea strong adhesive is dislodged.

Williams (1989) described the dismantling ofthe Portland Vase prior to its subsequentrepair. The interior of the vase was lined withthree layers of damp blotting paper, held inplace by a thin layer of plaster of Paris. Threelayers of damp blotting paper were applied tothe exterior, followed by a number of flexibleplastic tourniquets to give support. The vasewas then placed inside a glass dessicator overa bowl of warm water. After three days, theold adhesive (animal glue) had softened, andas the blotting paper was rolled away, theglass fragments, 189 in all, could be removedin a controlled manner (Figure 7.20).

Washing sound glass

Prior to washing, glass objects (excavated orotherwise) that have been in collections forsome years should be carefully examined forsigns of previous restoration, as this may notbe immediately apparent. If the glass has notbeen repaired or restored and is otherwisesound, it can be washed in tepid water (neverhot) in a plastic bowl (to prevent breakageshould the vessel slip). A few drops of a non-ionic detergent may be added to the water (iftoo much is used the glass will not be visiblethrough the suds) and only one object shouldbe placed in the bowl at a time. Care must betaken not to exert pressure on the glass duringcleaning, especially on fragile areas such as arim or stem of a drinking vessel. When clean,the glass is laid on paper towelling to drain.The towelling will absorb water and alsoprevent the glass from sliding on a wetsurface. Breaks may occur on the slightestimpact and therefore the working surfacesshould not be overcrowded, especially as it issometimes difficult to judge how much spacethere is between transparent objects. Drinkingglasses should not be held by the stem whilstbeing dried, in case they snap under pressure,but should be supported by cupping the handunder the bowl. A soft, lint-free cloth, or papertowelling, is used to dry the glass, taking carenot to exert any pressure. Drying should becarried out over a bench spread with a thicklayer of towelling, in case the glass shouldslip. The glass should be free from surfacemoisture before being displayed or stored.

DecantersThe most commonly encountered problemswith decanters are:

(i) part of the bowl having split off due to thevessel having been filled with hot water;

(ii) so-called water-marks around the bowl;(iii) a white ‘bloom’, especially on lead glass;(iv) small holes at the corners of the base of

square-shaped decanters;(v) the decanter stopper having become

jammed in the neck, after which attemptsto remove it may have caused the stopperto have broken, leaving the lower partfixed in position, or the decanter neck tohave broken.


Figure 7.20 Damp blotting paper applied to thesurface of the Portland Vase in order to facilitate itsdismantling. The cameo-glass amphora known as thePortland Vase, after its former owners, of cobalt-blueglass cased with opaque white, in which a mythologicalscene is cut cameo-fashion in relief. H 245–248 mm, GD177 mm, T (at bottom of broken edge of blue glass)3 mm. Perhaps from Rome where it was probably madein the early first century AD. (© Copyright The BritishMuseum).

A broken bowl may be repaired in the usualway; however, the owner would be welladvised not to put the decanter to everydayuse since if it contains liquid all the time thebond may eventually fail without warning dueto the penetration of liquid between the glassand adhesive.

There are many home remedies recordedfor the removal of ‘water-marks’ around thebowls of decanters. These include filling thebowl with water and newspaper (therebyforming a weak acid solution); the use oftablets marketed for cleaning dentures, forsterilizing babies’ or wine bottles or for de-scaling kettles; use of weak or concentratedorganic acids; and scouring the inside withmaterials ranging from sand to lead shot. Someof these remedies work some of the time, butare not recommended for use, because of therisk of damaging the object. Sometimes themark may seem to disappear whilst the glassis wet, only to re-appear when the surfacedries out. It is of course important to realizethat the mark may be a deposit of lime-scale(calcium) or other insoluble salt, or maysimply be a mark formed when wine has beenallowed to evaporate. However the surface ofthe glass itself may have deteriorated,especially in the case of old glass which hasbeen in long term use.

If the use of a weak concentration ofhydrochloric acid does not remove a deposit,it is probably best left alone. The interior ofthe decanter can be commercially abraded/polished with mild abrasives, or a thin layerof glass removed by swilling the interior withhydrofluoric acid (HF). This is an extremelydangerous acid and must only be used byspecialist companies competent to deal withit (see Chapter 5). Pinkish stains, the remainsof wine, port etc., can be removed by fillingthe decanter with a weak, tepid aqueoussolution of an enzyme detergent for a fewhours. If necessary a Nylon bottle-brush(wrapping the end of the metal holder withplastic tape, so that it does not scratch theglass) can be used.

A white ‘bloom’ may form on the surface ofglass, especially lead glass, which has beenput through a dishwashing machine. This isthe result of a reaction between the glass andchemicals in the finishing agent, and isirreversible (see Figure 4.30).

There are also home remedies for therelease of stoppers that have become wedgedin the neck of their decanters. Some of thesemay also work some of the time, and to someextent whether they do or not will dependupon how long the stopper has been in placeand what attempts have previously been madeto free it. It is sometimes the habit of peoplereplacing a stopper after pouring wine etc. togive the stopper a final pat, and this maycause it to become wedged in place. The mostobvious solution is to place the decanter inwarm to hand hot water (gradually increasingthe warmth) and, whilst the stopper isimmersed gently trying to pull it free. If thisdoes not work, the introduction of liquid soapor thin lubricating oil may suffice. However, ifthese are not thoroughly removed, they maydry out and act as adhesives. Another remedymay be to warm the decanter itself in water(thereby slightly expanding the glass), whilstapplying ice to the stopper (preventing it fromexpanding), so that just enough discrepancy isachieved to free the stopper. Tapping thestopper gently all round may loosen it, but ifextreme care is not taken this approach maybreak the neck, as will forceful twisting of thestopper. A stopper may remain firmly stuckand there may be nothing to be done.

In the case of stoppers that have broken offin the neck, these may be dislodged by adher-ing a wooden handle to the stump, afterwhich any of the methods described above,may be applied. If the stump is removed, theadhesive securing the ‘handle’ is dissolved andthe stump adhered to the rest of the stopperif this is present. As a final resort, it may bepossible for the remains of a stopper to becommercially ground out. It is not advisableto effect a repair to a stopper, part of whichhas become wedged in situ, as there is littleguarantee that adhesive will not run inside theneck and thus compound the problem.

Laser cleaning

Laser beams have been successfully used toremove (ablate) defined areas of dirt depositsin other fields of conservation such as stone(Cooper, 1998). Ablation is the term used todescribe the removal of surface dirt layer bylayer. As early as 1975, Asmus advocated theuse of laser beams for cleaning glass.


However, the equipment is extremely expen-sive, and much research on the technique isstill needed. The Fraunhofer Institut fürSilicatforshung in Germany is undertaking aresearch project aimed at examining theadvantages and limitations of laser cleaningheavily corroded medieval window glass bythe application of KrF excimer laser beams(wavelength 248 nm) to model glasses(Römich in Kordas, 2002). It has been shownthat it is impossible to remove corrosion crustson model glasses without significant damageto the sensitive gel layer (the softened layerbeneath the corrosion). Further experimentsare necessary to determine the appropriatelaser parameters for cleaning heavily corrodedglass with an uneven surface before originalglass can be laser cleaned, and the effect oflaser irradiation on the corrosion crust itself.The influence of the colour and compositionon the laser treatment has to be investigated.Laser treatment may not be applicable toobject glass conservation except perhaps forcleaning large quantities of marine encrustedor fire-blackened deteriorated glass. Glass‘jewels’ on the Albert Memorial in Londonwere successfully cleaned by the use of Nd-Yag lasers.

Repair

Decorative glass, more specifically glassvessels, form one of the most difficult groupsof antiquities to repair for a number ofreasons. The edges of un-decayed glass arenormally smooth and therefore do not providea key for good adhesion; the surfaces of glassobjects are covered with many molecularlayers of adsorbed water, thus reducing thebond strengths of adhesives; and since mostglasses are transparent, the repair or restora-tion is more visible than it would be on othermaterials. In comparison with other materialstherefore, relatively little early work has beenpublished in the field of glass object conser-vation, the majority having been concernedwith painted European medieval glasswindows in situ. However, this situation hasgreatly improved in the past four decades, asthe discipline of glass conservation hasexpanded. The choice of resins suitable forrepairing and restoring glass is continually

changing, allowing ever more complicatedprocedures to be undertaken (Petermann,1969; Errett, 1972; Wihr, 1963, 1968, 1977;Staude, 1972; Fiorentino and Borelli, 1975;Martin, 1977; Davison, 1988, 1998; Jackson,1982a, 1983; Fisher, 1998; Newton andDavison, 1989; Brain, 1992; Depassiot, 1997).

Glass restoration can entail complicated andtime-consuming operations, which althoughresulting in results acceptable for museum orhistoric house display, may not be aestheticallypleasing enough for a private collector. Evenskilled restorers are limited by the materialsavailable.

Choice of materials

The choice of adhesives for repairing glass isdiscussed in Chapter 5, with easily reversibleadhesives such as those based on cellulosenitrate and methyl methacrylate copolymersbeing the most suitable for the repair ofdecayed archaeological glass. Glass which isbadly deteriorated, can lose up to 75 per centof its weight, and glass which has lost itssurface through time or misadventure can bevery thin and delicate. In this condition glassmay be joined by adhesives that set by loss ofsolvent because the solvent can to someextent evaporate through decayed glasswhereas it does not do so through un-deteriorated glass. Furthermore, adhesion ofdecayed glass will be facilitated by the keyingof adhesive into the surface of micro-cracks. Ifthe glass has been consolidated, the charac-teristics of the adhesive must match those ofthe consolidant, remembering that the solventused to apply the adhesive will probably causethe consolidant to swell. A similar problem isencountered when attempts are made toremove the adhesive.

Cellulose nitrate, marketed in Britain asHMG and Durofix (and in the US as Ducocement), has been used successfully forjoining decayed glass and it is convenient touse. Its long-term instability has beenmentioned (Koob, 1982) and this should betaken into account. No long-term tests havebeen carried out on the stability of the jointwhen cellulose nitrate has been used onconsolidated glass. Polyvinyl acetate has alsobeen used as an adhesive but it has atendency to plastic flow if kept in a warm


environment. Paraloid B-72 formulated as anadhesive can also be used to repair fragileglass (Taylor, 1984). However it tends tobecome stringy very quickly and to contain airbubbles, due to rapid evaporation of theacetone solvent, and is therefore not as conve-nient for use as cellulose nitrate. The majordisadvantage of weaker adhesives is that arestored object may eventually fall apartcausing further damage, not only to the objectitself, but possibly to others nearby. However,epoxy resins have continued to be used forrepairing and even consolidating decayedglass in some countries.

Once such resins penetrate and consolidatethe weathering layers they must be consideredas irreversible. Fiorentino and Borrelli (1975)suggest that this irreversibility can be circum-vented by the use of a stable soluble resinsuch as Paraloid B-72 as a primer underneaththe epoxy, but several hours of immersion ina solvent are still required to break down suchjoins. However, in the absence of satisfactoryalternatives to provide strong joints for vesselsor display pieces such a system might beuseful. For much archaeological glass, strongadhesives are neither necessary nor advisable.

In the publications cited above the materi-als used for glass vessel restoration are epoxy,polyester and polymethacrylate resins whosetrade names, compositions and properties willto some extent vary with the country of origin.

Literature references to the use of epoxyresins for the repair of glass include AralditeAY103/HY956 and Araldite 2020 (formerlyknown as Araldite XW956/XW957 in its exper-imental stage), Fynebond (UK), and EPOTEK-1, Ablebond 342-1 and HXTAL-NYL-1 (Fisher,1992) (the last three manufactured in the US).Fynebond and HXTAL-NYL-1 are speciallyformulated for conservation use. Details canbe found in Chapter 5. Williams (1989) usedHXTAL-NYL-1 for the reconstruction of thePortland Vase, in combination with ultra-violetcuring adhesive by which the fragments wereheld in position during application and curingof the epoxy resin.

It has been reported that on ageing, epoxyresins may exert enough force on the glass topull flakes from the surface (Bimson andWerner, 1971; Moncrieff, 1975), however norecent cases of this occurring are known.Epoxy- and polyester resins cannot be

dissolved, only softened and swelled by theuse of solvents.

Supporting glass fragments during repair

While adhesive is setting, glass fragments canbe supported in a number of ways dependingupon their size and weight (Figures 7.21 and7.22a–e). Glass having a sound surface maybe supported by strips of pressure-sensitivetape, ensuring that the tape is of a type thatcan easily be peeled from the glass withoutdamaging it, e.g. Scotch Magic Tape 810,and/or partially immersing it in a tray filledwith glass beads, rice, dried pulses or sand,placing a paper barrier between the glass andthe sand (Figure 7.22a). Small plastic clampscan be used if necessary.

In many cases the objects are self-supporting,provided that the fragments to be joined arecorrectly balanced, and the adhesive in onecrack is allowed to set before the next frag-ment is added. However, it may be necessaryto use drops of faster-setting adhesives suchas cyanoacrylate on surfaces that are delicateor decorated and therefore difficult to tapewithout causing damage to the glass. A dentalproduct useful for the temporary support ofglass fragments is ‘Sticky Wax’ (beeswax,cerecine wax and gum dammar mix). Suppliedin pencil-thick sticks, small drops of the waxcan be melted with an electrically heatedspatula and dropped across cracks and joinsbetween fragments. The wax becomes stickywhen heated, but it hardens quickly oncooling (see Figure 7.22b). Once the glassfragments have been secured with adhesive,


Figure 7.21 Supporting glass fragments during repairwith adhesive, with strips of pressure-sensitive tape,spots of cyanoacrylate adhesive or ‘Sticky Wax’, ormetal ‘bridges’ secured with adhesive.


Figure 7.22 Supporting objects during repair: (a) asand-tray with a paper barrier; (b) with ‘Sticky Wax’and thin strips of pressure-sensitive tape; (c) a large,heavy object supported by wide elastic bands; (d) aclamp stand; (e) stumps of Plasticine.

(a) (b)

(c)

(d)

(e)

the wax is easily removed from the glass byslight pressure from a fingernail or bluntscalpel blade.

A method of supporting sound glassfragments by the use of metal bridges (termedstaples) originated in Germany (Augustin-Jeutter, 1997; Depassiot, 1997) (see Figure7.21). The bridges were easily made frombrass wires of different diameters by bend-ing them with small pliers. Although time-consuming to make in any number, they couldbe re-used. The size of the bridges dependedon the size, weight and number of glassfragments to be supported. The bridges weresecured to the glass surface at angles acrossthe joins with cyanoacrylate adhesive.However as this adhesive does not make astrong bond with the glass, drops of 5-minutecuring epoxy resin were applied to each endof the bridges with a cocktail stick. Epoxyresin was then applied to the breaks in orderto repair the glass. Once this had cured thesmall metal bridges were removed using ascalpel. The larger ones were easier to removeusing a rotating rubber disc on a flexible drill.The rotating disc was applied to the top of themetal, and the heat produced by frictionsoftened the adhesive in a few seconds. Whilstthe metal is still hot it could be removed witha pair of pliers, or with a scalpel. Adhesiveremains were removed with a scalpel and/oracetone on cotton wool swabs. This methodof supporting ancient or historic glassfragments cannot be recommended for generaluse, because of the amount of interferencewith the glass surface, the possibility that notall the adhesive used to affix the bridges willbe removed, and the potential risk of damagewhen removing the bridges. It could, however,prove useful for supporting fragments ofheavy glass objects such as sculptures.

Increasingly, heavy glass figures, especiallythose made in Venice, are requiring repair andrestoration. Some are extremely large andconsequently heavy, and pose particularconservation problems. Apart from the diffi-culties of supporting heavy, awkward shapes,there is also the question of how to bondbroken fragments. In general, the glass shattersand chips easily, and it is therefore inadvis-able to attempt to insert dowels. The brokenedges may be keyed by scratching themdeeply with a diamond point. Epoxy resin is

the adhesive most likely to hold the glass. Ifthe glass is opaque, missing chips of glasssurrounding the breaks can be filled using athick paste composed of colourless epoxyresin, fumed silica and dry artists’ pigments(see Figure 7.29). Over a long period of time,epoxy resin can separate as a sheet, from thesmooth surfaces of heavy glass items.

Procedure

First, the glass fragments should be sorted, forexample rim fragments from pieces of thebody and base, and laid out as nearly as possi-ble in their correct positions (Figure 7.23).Secondly, the broken surfaces are cleanedwith acetone on tightly wound cotton woolswabs (US: cotton buds) to remove greasefrom handling. (Uniform wetting by theacetone often indicates that the adhesive willdisperse uniformly.) If the glass is to berepaired with a cellulose nitrate adhesive suchas HMG, repair begins by working from theheaviest section, normally the base or rim, andallowing the adhesive on each piece to setbefore applying the next, or supporting thefragments as described above.

Adhesives which are not available in tubeform may be conveniently applied withsyringe, cocktail stick or fine paint brush. Tinyfragments of glass can be manipulated withhand-held or vacuum tweezers. André (1976)suggests manipulating fragments by attachingthem to a small ball of clay on the end of astick, but this cannot be recommended since


Figure 7.23 Fragments of glass laid out in order, priorto repair.

the clay may adhere to loosely adheringdecoration or flakes of glass, and detach them.

Fragments can be manipulated with hand-held- or vacuum tweezers. Those which fitinto a hole can be manoeuvred by attachinga cocktail stick to act as a handle, using water-soluble adhesive (Figure 7.24). Fragments ofglass which cannot be located should be putin labelled containers and kept with the object.They may be of use for analysis at a later date.Care must be taken not to apply an excessiveamount of adhesive to weathered or flakingglass since it will be difficult to remove fromthe surface. Excess adhesive may be removedwith cotton swabs moistened with acetone.

If the glass is sound, and the vessel doesnot bear vulnerable decoration such as gildingor unfired painting, epoxy resins may be usedfor its reconstruction. Glass that can berepaired in this way will also need to be ableto withstand future removal of the resins bysoftening and swelling them with commercialpaint strippers. These have been traditionallybased on dichloromethane, e.g. Nitromorswater or solvent miscible systems, or Desolv292, but are now being phased out of themarket in the UK for health and safetyreasons, and being replaced with NitromorsSuperstrip (based on dimethoxysulphide).

If the glass is to be reconstructed with anepoxy resin, work again proceeds from theheaviest section, using narrow strips of self-adhesive tape placed on alternate sides of theglass to support the fragments, as shown inFigure 7.21 (assuming that the interior of the

vessel is accessible). Spots of cyanoacrylateresin or of Sticky Wax placed at each end ofevery break (Figure 7.21) will provide tempo-rary support, especially where the interior ofthe vessel cannot be reached after repair, orwhen the humidity of the surrounding air issuch that it causes tape to peel away from thesurface of the glass.

No epoxy resin is applied at this stage andit can thus be ensured that all the fragmentsfit together accurately, there being no build-up of adhesive to distort the shape. Thecomponents of the adhesive to be used aremixed together using a wooden stick (a metalspatula may affect the hardening or the colourof the adhesive). To ensure that the resinhardening has begun before the adhesive isapplied to the glass, the mixture is allowed tostand for a few minutes, covering thecontainer to prevent dust from settling in it.The adhesive is then applied along the joinsin the assembly with the wooden stick (Figure7.25a–c). Being highly mobile, it seeps intothe cracks and by capillary action is drawnalong, even behind the thin strips of tape,completely filling the voids.

After about an hour, any excess resin thatmay have run down the surface of the glassmay be removed with cotton wool swabsmoistened with acetone. However, over-cleaning with solvent at this stage will weakenthe adhesive or even prevent it from setting.It is therefore better to apply the resinsparingly, and to leave the excess to setovernight, before removing small excesseswith a scalpel and cotton wool swabs moist-ened with acetone. In some cases, such ascone-shaped or narrow-necked vessels, whereboth sides of the vessel are not accessible, itmay be necessary to reconstruct the vessel intwo halves with tape on both sides. The twohalves can then be married together to ensurethat they will fit, then separated and epoxyresin applied to the cracks in each half. Whenthis has set, the tape can be removed from theinside and the two halves joined with tape onthe outside and secured with adhesive (seeFigures 7.45a,b).

The effect of heat on glass and glazes gener-ated from, for example, too rapid drying indirect sunlight, or conservation techniques, isunpredictable. For several reasons, notablydehydration of the glass, which may lead to


Figure 7.24 The last fragment inserted into a closedglass vessel, using a cocktail stick adhered with water-soluble adhesive, as a temporary handle.

visible crizzling (Brill, 1975, 1978), or evensurface damage (Bimson and Werner, 1964b),any repair using direct heat from a desk lampor infra-red lamp is not to be recommended.Despite the danger, several techniques for thetreatment of glass which involve heating atsome stage have been described in the past(Schroder and Kaufman, 1959; Staude, 1972;Wihr, 1968, 1977). It would, however, beacceptable to warm an adhesive before apply-ing it to the glass (but this might cause prema-ture discoloration of the adhesive).

SpringingWhen thin blown glass breaks, the object may‘spring’, i.e. move slightly out of its originalshape (see Glass domes below). This is causedby the release of tension created either whenthe glass was made (particularly in the area ofa bowl where the glass has been stretched byblowing, and the tension not removed byannealing) or by hydration of the surface. Theresult is that the glass fragments do not fittogether correctly when reconstructed.Because glass is fragile it may not be possibleto make a successful repair using strips oftape, or a tourniquet, to close the gap afterthe adhesive has been applied, as issometimes the case with ceramics.

CracksBroken glass may be found in a damagedcondition in which cracks do not run rightthrough the glass (so-called running or travel-ling cracks). Although these may have noeffect upon the joining of other fragmentsunless the glass has sprung out of shape, theirpresence in a glass artefact is potentiallydangerous. As breaks in the glass surface,cracks are points of weakness from whichdamage may propagate if the artefact issubjected to mechanical or thermal stress. Ascracks constitute glass/air/glass interfaces, theypresent a plane for reflection of light, whichis a distraction to the eye and thus detractsfrom the aesthetic qualities of the object.

Historically, the lengthening of cracks wassometimes arrested by inserting rivets, or bydrilling through the glass at a point justbeyond their closed end. This action is notrecommended for ethical reasons, and becauseit may result in further damage and in anycase results in an unsightly repair similar in


Figure 7.25 (a) Introducing epoxy resin into breaks ina glass vessel (see Figure 4.2), by capillary action. (b)The two parallel lines across the centre represent eachglass surface, the area in between having been filledwith epoxy resin, whose refractive index closelymatches that of the glass itself. (c) The cracks are onlyvisible as hair-lines.

(c)

(a)

(b)

appearance to the repairs shown in Figures4.10 and 7.26.

In order to prevent cracks from lengthening,it is usually possible to introduce a highlymobile epoxy resin into them by capillaryaction, by applying resin along the cracks witha cocktail stick (see Figure 7.25a–c). Bearingin mind the dangers of heating glassmentioned above, it may help the resin topenetrate cracks if the glass is warm. Onmodern glass, this action can be made moreeffective if the object is alternately gentlyheated with a warm air blower and thenallowed to cool several times. Of course ifcracks can be gently eased open it will beeasier to introduce the resin, but this actionmay cause cracks to lengthen dramatically!

For the repair of cracked, optically clearglass, a close match of refractive indices (RI)is desirable (Tennent and Townsend, 1984a)(see Figure 5.1 and Table 5.2). This might bedifficult to obtain for the following reasons:the RI of glass changes with composition; thesurface RI of glass changes on weathering andthe RI value of a recent crack could be differ-ent from that of an old one in the same object.It may be possible to formulate a range ofadhesives of different RI values, so that theadhesives could be matched to the glass.However, since adhesives are expected toyellow with age, thus disclosing the area ofrepair, exact matching may not be important,except in special cases, although Tennent andTownsend (1984a) point out that yellowing of

a crack may not be important when the crackis to be viewed endwise. There is also theethical consideration of whether cracks shouldbe concealed.

DowellingIn the case of broken wine glasses or glassfigures, it may be necessary to insert an acrylicor glass dowel to strengthen a repair (Figure7.27a–e). This is a potentially dangerousoperation since it necessitates the drilling ofsmall holes in the glass, which may cause theglass to fracture. Dowelling cannot thereforebe recommended as a general procedure.

Larney (1975) refers to a dowellingtechnique which involves drilling a hole free-hand into each half of the wine glass stem,and inserting a piece of glass-rod grounddown to fit. A problem encountered in usingthis technique is the difficulty of locating theexact centre of the glass stem, particularly onan uneven surface. Unless the centre point ofeach half of the broken stem is accuratelydetermined, the dowel cannot be inserted.This can be achieved by placing a small dotof easily removable ink or paint on one sideof the stem, and by touching it accuratelyagainst the other half, marking it. It is alsodifficult to ensure that the holes are drilledvertically into the stem. Alternatively, it maybe possible to drill oversized holes in the glassstems, to allow some leeway in inserting thedowel. However this risks causing furtherdamage. The minimum thickness of glass rodavailable is 3.0 mm and in practice this wouldhave to be considerably ground down to beused as a dowel, unless the glass can beheated and drawn out by a glassworker.

The problems associated with this techniquewere overcome by producing a metal collaron a lathe, which fitted over the stem of thewine glass and through which a central dowelhole could be drilled into the glass (Jackson,1982a) (Figure 7.27b). The size of the holewas reduced by using a 1.0 mm Perspex rod(methyl methacrylate) (US: Plexiglas) as thedowel. Having measured the diameter of thewine glass (the greatest diameter if the stemis not truly round) with a micrometer, (a) ashort length of mild steel rod, with a diame-ter sufficient to enable it to fit over the glassstem to form a sleeve with an adequate radialthickness (Figure 7.27c) was placed in a lathe


Figure 7.26 An unsightly repair on a heavy cut glassbowl, in which lead-covered wire ties have been usedto secure a fragment.


Figure 7.27 (a) A broken stem of a wine glass, whichrequires the insertion of a dowel to strengthen therepair. (b) A metal collar used to centre the drill on thebreak edges of the wine glass stem. (c) Section throughthe metal collar. (d) The use of a diamond-tipped drillto form a dowel-hole in each section of the wine glassstem. (e) The repaired wine glass stem incorporating aPerspex dowel secured with epoxy resin. (Courtesy ofP.R. Jackson).

(a) (b)

(c)

(e)

(d)


Figure 7.28 Mounting glass fragments for display orstorage. (a) A blown glass form supporting an Islamicbottle. (b) A slumped glass form, shown with theplaster of Paris mould taken from the original glassdish. (c) A Perspex form supporting an Islamic glasslamp. (d) A three-part Perspex mount for enclosing apane of moulded green window glass from a Romanvilla at Garden Hill, Hartfield, East Sussex (afterrestoration). Upper and right-hand sides grozed. H 255 mm, D 235 mm. (e) A resin cast (mount) onwhich fragile glass is displayed. (© Copyright TheBritish Museum).

(a)

(b)

(e)

(d)

(c)

and faced. The centre point was marked, witha centre bit, and a hole, of a diameter accom-modating a 1.0 mm glass drill bit (b), drilledto a depth in excess of 15 mm. This wasfollowed by a drill with the same diameter asthe stem (a) and drilled to a depth of 10 mm.The rod was cut to a length of 15 mm,reversed in the lathe and faced. The metalcollar was placed over the section of brokenstem attached to the foot of the glass. Usinga 1.0 mm MM Glazemaster diamond-tippedglass-drill, a hole was drilled into the stem(Figure 7.27d). The broken stem attached tothe bowl was drilled in the same way. APerspex rod was cut to the correct length toform a dowel and secured in the holes withepoxy resin, thus bringing the base and bowlof the wine glass together (Figure 7.27e). Theuse of a metal collar as a jig to hold the drillbit in place produced very accurate results.The dowel holes were only 1.0 mm in diame-ter and were correctly centred, aligned andvertical (Jackson, 1982a).

Mounting fragmentary objectsObjects that are too fragmentary to enablethem to be interpreted can be mounted forstorage or display. The most aestheticallypleasing mounts are those made from blownglass, acrylic sheets and rod or resin formers(Figures 7.28a–f).

Restoration

Restoration (replacement of missing areas witha synthetic resin) may be undertaken toimprove the stability of an object such as aglass vessel, to aid in its interpretation, or toimprove its appearance for display purposes.If a great deal of the vessel is missing it maynot be worth restoring in terms of the timetaken to carry out the work, unless the glassis historically important or is of particularvalue to a collection.

There are a number of factors to be takeninto account when considering the restorationof incomplete glass artefacts:

(i) whether the condition of the glass willpermit the work to be undertaken;

(ii) whether such a procedure is desirable onethical grounds; and

(iii) whether restoration is feasible.

Feasibility is usually determined by thepercentage of the object remaining, the shapeand thickness of the glass, the condition andtype of any applied decoration present(Davison, 1998), and the materials available.

Restoration, as opposed to reconstructionfrom existing fragments, is essentially a mould-ing and casting operation. Moulds may betaken from the glass itself or from a modelledclay or Plasticine (US: Plastilina) former repre-senting the missing glass. The moulds are thensecured over the area to be replaced and aclear resin is cast into them. The choice ofmaterials is governed by the effect desiredfrom the restoration, cost, availability andproperties in relation to the job in hand(Jackson, 1982b, 1984; Davison, 1988, 1998).

Restoration techniques can be convenientlyclassified as shown below. The techniques aredescribed and illustrated with diagrams andactual case histories: however, it is part of aconservator’s task to adapt restoration tech-niques as required, since each restoration is anindividual undertaking.

• Filling losses in opaque glass• Filling tiny losses• Partial restoration• Detachable gap-fills• Gap-filling with casts from moulds taken

from the original glass• Gap-filling with casts from a mould taken

from a clay model or a previous restora-tion

• Gap-filling where the interior of a vessel isinaccessible for working

Filling losses in opaque glassLosses in opaque glass can be filled with pastemade by mixing optically clear epoxy resinwith fumed silica and dry artists’ pigments(Figure 7.29), or with epoxy putty such asMilliput (UK) coloured with epoxy- or acryliccolours (Figures 7.30a,b).

Filling tiny lossesPressure-sensitive tape can be used as abacking behind tiny areas of loss, such asholes or chips, in order to fill them. A strip ofpressure-sensitive tape is fixed across the losson one side, and a drop of resin applied.When the resin has cured, the tape isremoved. Small losses on flat or gently curved


glass can be filled by sealing a piece of waxcoated with polyvinyl alcohol to the exteriorof the glass in such a way as to cover the loss.The object is then supported in a sand tray,

so that the area to be filled lies in a horizon-tal plane, in order that resin can be pouredinto the gap and find its own level. If the glassis thick-walled, the resin can be poured in oneor two thin layers, provided that each layer isonly allowed to set, not fully cure, or thelayers may not bond to each other.

Partial restorationWhere a glass vessel has large areas missingbut does not warrant total reconstruction, forexample because of the time involved, orbecause the object is destined for storage oroccasional study, it may be partially restoredfor safe handling. Strips of glass fibre surfac-ing tissue cut to size and impregnated withcellulose nitrate adhesive, Paraloid B-72,epoxy- or polyester resins, are used to bridgegaps in the glass and to hold floatingfragments in their correct positions. Totalreconstruction of small vessels is also possibleby this method, with the results shown inFigure 7.31. It is not aesthetically pleasing but


Figure 7.29 An opaque glass sculpture restored withcoloured epoxy paste.

(a)

(b)

Figure 7.30 An opaque mosaic glass bowl restoredwith coloured epoxy putty.

Figure 7.31 An Anglo-Saxon claw beaker partiallyrestored with polyester resin and glass-fibre surfacingtissue. (© Copyright The British Museum).

may be useful as a temporary measure forstudy or drawing purposes.

Gap-filling with casts from moulds takenfrom the original glass

Silicone rubber dental impression compoundSmall three-dimensional features on glass, suchas knobs, handles and applied decoration suchas prunts, can be reproduced by taking amould with a rapid-setting dental siliconeimpression compound such as Amsil (UK)from another identical object. The mould ispeeled off, or slit, to enable it to be removed,supported in a sand tray and filled with resin.Once the resin has cured, the cast is removedand the ends trimmed to fit the broken glass,to which it is attached with epoxy resin(Figure 7.32). The cast usually has a mattfinish and can be given a coat of epoxy resinto restore its clarity.

Casting resin into wax or silicone rubbermouldsFor replacing all but the smallest missing areasof glass, and glass with surface detail, resinmust be cast into a closed mould, made fromsheets of dental wax (which is cheap, but doesnot reproduce very fine detail), or siliconerubber (which is expensive but reproducesfine detail). The moulds are then filled with asynthetic resin to form casts, which replace themissing areas of glass. Once the resin hascured, it can be sanded and polished if neces-sary, although it can rarely be brought toabsolute clarity. Casts can be made of clearresin and be left unpainted, although this

approach may result in their being visuallydistracting. Alternatively, the casts can bepainted by hand, or by airbrush if the glasscan be safely masked off, or the resin can bepigmented before being cast.

The manufacturers of epoxy- and polyesterresins produce pigmented pastes for colouringthe resin. Only a few of these are transparent,and are polyester-based. The colours are inter-mixable and their brightness can be toneddown by the minute addition of black opaquepolyester pigment. The coloured pastes are sostrong that very little is needed to colour theresin. Consequently, whilst it is not generallyrecommended to mix resins, it is possible tomix tiny amounts of epoxy pigmented pastewith polyester resins, without impairing itscuring.

Where coloured resin gap-fills are to bemade, it is recommended that a batch ofcoloured resin be prepared in advance (Plate8). As each cast is made, a small amount ofcoloured resin is measured out and catalysed.The batch will ensure that exactly the samecolour is used for each cast, and that extraresin is available should a cast have to bemade more than once. In general it is betterto tint the resin to match the original glass asseen by transmitted light, the reasons for thisbeing that the cast will absorb colour from thesurrounding glass (especially if original glassis also behind the cast glass, e.g. in the caseof a bottle). Also if the cast is made to matchthe colour of glass seen by reflected light, itwill appear too dense (Plate 9).

Wax mouldsWhere the gap to be filled is the edge of arim or foot of an object, a two-part waxmould, open at the edge, is made from smallpieces of dental wax (Figures 7.33 and7.34a–j). Without touching the flat surfaces ofthe wax sheets, so as not to mark them, twopieces of wax, slightly larger than the area tobe gap-filled, are cut, and slightly softenedwith a hot air blower (or in warm water). Itis then gently pressed over an existing area ofglass – one on the inside and one on theoutside of the vessel – which conforms to thesame shape as the missing area (i.e. is on thesame horizontal plane) (Figures 7.34 and7.35a). If the wax is over-heated so that itssurface becomes tacky, it will adhere to the


Figure 7.32 A mould made of silicone dentalimpression compound, and an epoxy resin cast used toreplace a missing glass knob on a lid.

glass and tear when removed. The sides of thewax, which will be in contact with the resin,are brushed with a solution of polyvinylalcohol, to which has been added one or twodrops of a non-ionic detergent to break itssurface tension. (Otherwise the separatingagent has a tendency to collect in pools onthe wax – Figure 7.34c.) The application ofthe separating agent facilitates easier removalof the wax after the resin has cured, but also

prevents colour from the pink or red waxbeing absorbed by the resin.

When this has dried, the wax piece formingthe inside of the mould is fixed in position byrunning the point of a small electrically heatedspatula around its edges (Figure 7.34d). Thecloseness of the fit of the wax against the glasscan be checked at this point and, if necessary,improved by slightly warming the wax andpressing it against the glass. Care must be


Figure 7.33 Stages inconstructing a wax mould.(a) A section through a two-part wax mould. (b) Shapingwax over matching areas ofglass. (c) Coating the waxwith release agent. (d) Theedges of the wax melted witha heated spatula, to seal it tothe glass surface. (e) Resinintroduced to the mould. (f)Excess resin above the rim ofthe glass to allow forshrinkage. (g) Plan view of awax mould made on a thickglass object, which requiresends, to form a dam tocontain resin.


Figure 7.34 The neck of a whisky dispenser restoredby casting resin into a wax mould. (a) A missingsection of rim. (b) Two pieces of wax cut to shape,warmed and shaped by pressing it onto the glass, onepiece on either side. (c) The sides of the wax whichwill be next to the resin, coated with release agent. (d)Wax sheet secured to the interior of the glass. (e) Waxsheet being sealed to the exterior of the glass, bymelting the edge with a heated spatula. (f) Mould beingfilled with resin. (g) Removing the exterior sheet ofwax. (h) Removing the interior sheet of wax. (i) Theresin cast before final cleaning with white spirit. (j) Thecompleted restoration.

(a) (b)

(f)

(c)

(e)

(d)

taken not to distort the wax by pressing itwhere it is unsupported by the glass. The waxforming the outside of the mould is then fixedin position with the heated spatula (Figure7.34e). It is difficult but important to ensurethat the wax fits tightly against the glass, orthe resin, when introduced, will seep over theglass leaving a large amount to be cleaned offafter the mould is removed. The resin ispoured into the mould, through which it canbe seen (Figure 7.34f). Once the resin has set,the wax is carefully removed (Figure 7.34g,h),by releasing the edges with the tip of a metalspatula, and then peeling it off.

Wax remaining on the glass and cast shouldbe removed with the spatula or fingernailtaking care not to scratch the surface. Thesurface can then be wiped with swabs ofcotton wool and white spirit (US: Stoddartsolvent) to remove the last vestiges of wax(Figure 7.34i). Any polyvinyl alcohol whichhas remained on the cast can be removed withwater. The top of the resin cast is then

trimmed and if necessary abraded andpolished (Figure 7.34j) (Staude, 1972; Davison,1998).

Generally speaking, if the cast is found tobe too thick, resulting in a step around theedges, it is better to remove the cast and startagain. If the step is on the inside of a vesselmade of dark glass, or with a narrow neck,where it will not be seen, there is no need toremove the cast. Although it is possible toabrade a resin cast as described below, itrequires a great deal of time and the cast cannever be brought to the clarity of unworkedresin. Excesses of resin above the rim or footcan be removed by the careful use of a dentaldrill and burs, and if necessary, lightly filedand sanded, though care must be taken not tocrack the cast or separate it from the glass byapplying too great a pressure. If the castshould become separated from the glass, it canbe reattached with adhesive, provided that theadhesive is not allowed to spread onto andmark the surface of the cast.


(g) (h)

(i) (j)

After sanding with progressively finer gradesof abrasive such as Micromesh, the surfacemay be polished using a fine abrasive pastesuch as Solvol Autosol, on a small buffingwheel held in the drill-head. (However, caremust be taken in the use of abrasive pastes,since they may become irremovably trappedin small deficiencies in the cast such as airbubbles, or at the junction between the castand original glass.) During any operationinvolving the use of a dental drill, it is wiseto step down the current with a rheostat inorder to prevent friction from melting thesurface of the resin, and to prevent therestored object from being accidentally spunout of the conservator’s hands.

In cases where the area to be filled lieswithin the main body of the glass, a differentmethod of filling the mould has to be adopted.The wax moulds are shaped as describedabove and the piece of wax which will formthe inner side of the mould, coated withpolyvinyl alcohol and secured to the glass bymelting its edges. Before applying releaseagent to the piece of wax which will form theouter side of the mould, small airhole(s) anda slightly larger pour-hole are made throughthe wax, at the highest point of its curve. Theholes are made by pushing a small metalprong through the wax, from the concaveside, which ensures that the displaced waxfrom the holes is on the outside of the mould(see Figure 7.35). The pour-hole is thencarefully enlarged with a scalpel, and the wax

forming the outside of the mould is fixed inposition with a heated spatula.

The mould is filled with resin dripped intothe pour-hole from a wooden cocktail stick ormetal spatula. If the area to be filled iscomplex, with several undercuts, the mouldmay take time to fill. It may be helpful toattach a small funnel made of aluminium foilto the edge of the pour-hole with a heatedspatula. The glass is supported with the funneluppermost in the sand-tray, and the resinpoured into the funnel (or directly into themould), until it begins to seep out of theairholes. At this point the funnel may bebroken off so that excess resin does notcontinue to flow out over the surface of theglass. (It may sometimes be possible to use amedical syringe without the needle to intro-duce resin into the mould, but it is not easyto control the flow of the resin, or to preventthe introduction of air bubbles.) Once theresin has fully cured, the wax mould iscarefully removed, taking care not to mark thecast. The cast is then cleaned and if necessarytrimmed and polished (see Figure 7.34j; seealso Plate 5).

DomesGlass domes were made from the eighteenthcentury onwards in France to encase glassfigurines (such as those made at Nevers andelsewhere), and in other parts of Europe toprotect clocks from dust and grime, therebyhelping to ensure accurate timekeeping. In


Figure 7.35 A hole in a globular glass lampshade,restored by dripping resin through a small hole in theouter side of a two-piece wax mould.

Figure 7.36 Missing section of a glass dome restoredby casting resin into a mould made of a doublethickness of dental wax.

nineteenth-century England, numerous largedomes (or shades) were made to protectornamental displays of wax fruits, artificialflowers, flowers made of shells, stuffed birdsand animals, glass ships and elaborate ‘birdfountains’.

Domes were made by the broad (cylinder)glass process. A bubble of glass was workedinto a rounded cylindrical shape, one end ofwhich was cut and the other left dome-shaped, the whole being given a smooth finishby re-heating it before the final annealingprocess. Domes continue to be made, butusually on a smaller scale, and tend to becircular in shape.

Antique glass domes are difficult to repairand restore. Some are very large and theirgeneral shape, flat on two sides, with curvedends and a domed top, make them difficult tohandle and to support whilst undergoingrepair. In addition the glass tends to be thinand brittle, and to spring out of shape whenbroken. Glass domes can be repaired withepoxy resin, and missing areas restored withpolyester embedding resin cast between sheetsof wax, coated on one side with a separatingagent. If the restoration is being carried out ona curved part of the dome, it may be neces-sary to construct the mould using two sheetsof glass on either side. This has the effect ofstrengthening the mould, and thus preventingthe wax sheets from being sucked together asthe resin cures (Figure 7.36). It is sometimespossible to obtain replacement domes from aspecialist company, which keeps a stock ofantique domes or the original moulds.

Silicone rubber mouldsWax is generally unsuitable for mouldingsurfaces with intricate detail, those which aretwo-dimensional, e.g. deeply cut glass, orwhich curve sharply both from top to bottomand from side to side. In these instances, andwhere the area of loss is larger than dentalwax sheets, silicone rubber can be used tomake the moulds (Figure 7.37a–h). Where theloss is at the glass rim or the edge of a foot,a thin coating of catalysed silicone rubber ispoured over the rim and both sides of anintact section of the glass vessel, covering anarea slightly larger than that which is to bereplaced (Figure 7.38a–e). The rubber will runvery thin at the glass edge so that the glass

can be seen through it, nevertheless it willhave coated the surface.

When this has cured, a second layer ofsilicone rubber, mixed to a stiff consistencywith an inert filler such as silicone mattingagent (e.g. Santocel), is applied over the firstlayer to strengthen it and to prevent therubber from tearing when it is eventuallyremoved. However, the silicone rubber layershould not be so thick as to prevent it frombeing slightly flexible. When the siliconerubber has cured, a slit is made in the mouldalong the rim with a scalpel. The slit can beas wide and as long as possible provided thatthe mould will remain intact at either end. Thesealed ends will locate the mould on the glassrim on either side of the area of loss. Openingup the mould in this way will prevent airbubbles from being trapped in the cast. Theslit is made at this stage, for if it were madeafter the mould was in position over the gap,the mould might be inadvertently cut belowthe level of the glass rim, or loose pieces ofrubber might fall into the cavity.

The mould is then removed from the glassby loosening it on both sides and slowlypeeling it off, and is secured over the missingarea as follows, using silicone rubber as an‘adhesive’. A small amount of catalysedsilicone rubber is applied around both sidesof the missing area of glass and over the edge,taking care that the amount of rubber and itsproximity to the gap does not result in rubberrunning onto the break edges when themoulds are replaced (Figure 7.38b). When therubber ‘adhesive’ has cured, more siliconerubber, mixed to a stiff consistency withfumed silica, can be applied around the edgesof the mould, as a secondary line of defenceagainst resin escaping from the mould. Resinis then dripped or poured through the slit intothe mould, which is occasionally slowly andgently squeezed to expel the air (taking carenot to introduce air). The process is repeateduntil the mould is full. It may be necessary toadd a little more resin, just before it gels, tocompensate for a small amount of shrinkage.The resin is then allowed to set, after whichthe mould is removed and any necessaryfinishing work carried out to the cast (Figure7.38c,d).

Where a missing area is a hole in the mainbody of the glass vessel, a comparable area of


the vessel will have to be moulded both insideand out with silicone rubber. The moulds arethen peeled off and attached over the missingarea, using uncured silicone rubber. A hole ismade with a scalpel through the siliconerubber on one side of the vessel (normally thaton the outside for ease of working), and theresin is introduced as described above.Alternatively, two thin clay stumps, orchimneys formed from paper or plastic drink-

ing straws, can be attached to the glass beforethe mould is made. These when eventuallyremoved will have formed holes through thesilicone rubber mould. A similar approach canbe taken in the case of an uneven rim, inorder to form pour-and airholes in a mould(Figures 7.39a,b). Rigid materials such asPerspex should not be used in this way, asthey will prove difficult to remove from thesilicone rubber.


Figure 7.37 Stages in theconstruction of a siliconerubber mould. (a) An area ofglass outlined with aPlasticine wall. (b)Application of a thin layer ofsilicone rubber, followed by athick second layer. (c)Plasticine removed, and a slitmade in the top of themould. (d) Mould peeled offthe glass and secured overthe area to be filled. (e)Resin poured into the mould.(f) Similar process where themissing area is in the bodyof the vessel. (g) Stumps ofPlasticine, which will form(h) pour- and airholesthrough the mould.


Figure 7.38 Restoration of the rim of a decanter. (a)Silicone rubber used to form a mould over an area ofglass similar to that to be replaced. (b) A thin strip ofsilicone rubber trailed close to the broken edge, onboth sides of the glass, in order to secure the siliconerubber mould. (c) The silicone rubber mould securedover the area to be filled. Note the slit in the rim,through which the mould will be filled with resin, andthrough which air will escape. (d) The polyester resincast after finishing and polishing. (e) Close-up view ofthe restored rim.

(a) (b)

(c) (d)

(e)

A more complicated version of the siliconerubber moulding technique had to be devisedin order to restore the Auldjo Jug, a cameoglass on display in the British Museum(Figures 7.40a–l). This vessel has a narrowneck and therefore access to the interior formoulding purposes was severely restricted(Jackson, 1985). The jug had been broken atsome time in its past and a considerableamount of the body was missing. It had previ-ously been repaired, probably with animalglue, and gap-filled with plaster of Pariscoloured dark blue to match the remainingglass. The plaster restoration had becomedamaged and unsightly and it was thereforedecided to remove the plaster in order toeffect a more accurate, light-weight andaesthetically pleasing restoration. In order toachieve this it was necessary to produce twosilicone rubber moulds, which conformed tothe inner and outer profiles of the jug. Sincethe plaster was actually dimensionally correct,it was repaired, and a silicone rubber mouldmade over it. The positions of funnels throughwhich to introduce resin into the area at

present occupied by the plaster was decidedupon. To form them, plastic drinking strawswere attached to the plaster with small lumpsof Aloplast (a modelling compound similar toPlasticine, but especially formulated for usewith polyester resins).

The globular shape of the Auldjo Jug deter-mined that the outer silicone rubber mouldwould have to be made in two sections inorder to facilitate its removal from the vessel.To contain the uncured rubber, a wall ofAloplast was laid over the glass surface so thatit divided the jug into two sections. A thin stripof lead wire was embedded in one side of thewall to form a key in order to reposition themould accurately prior to casting the resin(Figure 7.40a). The jug was then laid on itsside and a thin layer of silicone rubber wasbrushed over the uppermost side thus cover-ing the previous restoration and remainingglass fragments. When the silicone rubber hadcured it was reinforced with glass fibre wovenmatting and a second layer of silicone rubber.The jug was then turned over and supportedwith the opposite side uppermost. The


Figure 7.39 (a) Clay chimneys placed on the edge of the existing glass form pour- and airholes through thesilicone wax mould. (b) The completed restoration using polyester resin to replace the missing area of glass.

(a) (b)

Aloplast wall was removed and polyvinylalcohol painted along the exposed edge ofsilicone rubber to act as a separating agentbetween it and the second part of the mouldwhich was made as previously described(Figure 7.40b). A two-piece mother mouldwas then made over the silicone rubber usingrigid polyester laminating resin strengthenedwith heavy grade glass fibre matting. Smallholes were made through the flanges of themould pieces where they joined each other inorder to incorporate nuts and bolts, which

would secure the mould during the castingprocess (Figure 7.40c). This completed theconstruction of the outer mould, which wasthen carefully dismantled and removed fromthe jug. The adhesive and plaster used in theprevious restoration was easily removed fromthe glass fragments by soaking the jug inwarm water (Figure 7.40d).

Since access to the interior was restricted itwas not possible to make or to remove amould of the interior profile in the samemanner as the outer moulds. Therefore a


Figure 7.40 (a) A wall of Aloplast dividing the previously restored Auldjo Jug into two sections; and incorporatinga thin lead strip to form a key. The thin wooden sticks will form holes through the mould into which small nutsand bolts will be inserted to secure the mould pieces during the casting process. (b) The completed silicone rubbermould. (c) Completed glass fibre mother-mould secured by nuts and bolts; and incorporating plastic straws to formpour- and airholes. (d) Fragments of the Auldjo Jug after removal of the old restoration. (e) Comparison of differenttypes of silicone rubber, from which an inner mould is to be made: Silicoset 105 (left) and Rhodorsil 11504A(right). (f) Fragments of the Auldjo Jug repositioned in one half of the outer mould, showing the extent of themissing areas of glass. (g) Missing areas of glass filled with Aloplast. (h) Forming the inner mould with siliconerubber. (i) Coloured polyester rigid laminating resin coating the inner silicone rubber mould to prevent the nextlayer of resin from seeping between the mould and the glass. (j) Introducing the coloured polyester embeddingresin from a syringe. (k) Removal of the silicone rubber mould. (l) Completed restoration of the Auldjo Jug afterpolishing the resin cast.

different method had to be devised. Thecompleted mould would have to be thin andflexible enough to be pulled out through thenarrow neck after the restoration wascomplete; firm enough to support the weightof the glass neck and handle of the jug duringrestoration; and capable of adhering to theinner surface of the glass fragments in orderto prevent resin from flowing behind them. Inorder to choose a moulding material thatwould fulfil these criteria, a number of exper-iments were carried out using a narrow-necked laboratory jug to represent the AuldjoJug. After several attempts at filling a toyballoon with air, sand or water, it was foundto be impossible to fit the balloon closelyagainst the sides of the bottle.

A method was then devised to produce arubber skin similar to the balloon, but whichconformed closely to the inner surface of thebottle. For this purpose several differentgrades of silicone rubber and rubber latexwere tested. Each rubber moulding materialwas poured into a bottle, which was thenslowly rotated so that a thin skin formed overthe interior, before the excess rubber waspoured out through the neck. The rubber wasthen left to cure and the resulting mouldpulled out of the bottle, thus testing the tearstrength of the rubber. Other observationsmade were the extent to which the materialhad coated the inside of the bottle, and theease with which the rubber could beremoved. As a result of these tests, twobrands of silicone rubber were selected forfurther testing: Silcoset 105 and Rhodorsil11504A, although neither product possessedall the required qualities (Figure 7.40e).Silcoset 105 was extremely fluid, coated theglass surface well and was self-supporting.However, when cured it was rather rigid andthus proved difficult to remove from thebottle. The addition of silicone oil as a thinnerproduced a less rigid mould but considerablyreduced the tear strength of the Silcoset 105.Rhodorsil 11504A produced a very thin, flexi-ble mould which was easily removed from thebottle but which was not self-supporting.Various attempts were made to support themould in situ by filling the bottle with sand,water or vermiculite granules, of which thelatter proved to be the most successful.Vermiculite had the advantages of being dry,

extremely light, easy to use, and, although thegranules tended to compress slightly, easy toremove. Rhodorsil 11504A supported byvermiculite granules were therefore chosen asthe materials from which to make the interiormould.

The fragments of the Auldjo Jug werecoated with polyvinyl alcohol to preventexcess resin from adhering to them duringthe casting, and were reconstructed usingAblebond 342-1 (epoxy resin, no longer avail-able), and laid in one half of the outer mould(Figure 7.40f). Areas of the jug from whichthe glass was missing were then filled withAloplast which was modelled and smoothedto represent the thickness of the original glass(Figure 7.40g). Fragments of glass whichoverlapped the edge of the mould, (i.e.which overlapped into the other half of themould) were carefully removed withoutdisturbing the Aloplast model. These andother remaining fragments were laid in thesecond half of the outer mould and the areasof missing glass filled with Aloplast as previ-ously described. The two sections of thesilicone rubber mould and the glass fibreouter mould were then brought together andlocated in their correct positions by means ofthe keys incorporated in their adjoiningedges, and secured with small nuts and bolts.The interior of the jug was then inspectedwith the aid of a small dental mirror toensure that the glass fragments and Aloplasthad remained undisturbed. The seam-linebetween the modelling compound in the twohalves of the mould was smoothed down onthe inside using a Perspex tool madeespecially for the purpose.

Rhodorsil 11504A was then poured into thejug through the neck opening and the vesselslowly rotated so that the rubber flowed overthe interior completely covering the glass andAloplast (Figure 7.40h). Excess rubber wasthen allowed to drain out by inverting themould. When the Rhodorsil 11504A had curedthe interior of the vessel was again inspectedto ensure that the rubber had completelycovered it, after which the jug was packedwith vermiculite granules to support the mouldduring the casting process. At this stage thejug with its missing sections temporarilyreplaced with Aloplast was totally enclosedbetween two moulds. The next step was to


remove the Aloplast and to replace it withpolyester resin. In order for this to beachieved, one half of the outer mould wascarefully dismantled so that the Aloplast couldbe removed carefully with a spatula thusexposing the inner rubber mould surface. Thiswas cleaned with acetone, and a thin layer ofpolyester gel coat (a thixotropic paste), whichhad been coloured with blue polyesterpigment paste was brushed over the mouldand over the broken edges of the glass (Figure7.40i). The outer mould was then replacedand the process repeated on the other half ofthe jug. The two sections of the mould werethen again bolted together. This procedureensured that the gel coat resin was in closecontact with the inner mould and the edgesof the glass fragments, and that the castingresin would be less able to flow behind theremaining glass.

A batch of clear polyester resin wascoloured with opaque blue polyesterpigment to match the glass. The bulk of theresin was catalysed and introduced to themould from a syringe through the plasticstraws incorporated in the outer mould toform funnels for this purpose (Figure 7.40j).A small amount of the coloured resin waskept un-catalysed in order to effect anynecessary repairs to the cast. Air bubbles inthe resin were encouraged to rise and toescape through other straws forming airholesby gently tapping the mould. The resin wasthen allowed to cure for three days beforeremoving the exterior mould. The vermiculitegranules were poured out of the jug and theinner rubber mould removed in one piece(Figure 7.40k).

Excess resin on the outer surface of the castin the form of seam-lines and stumps whichhad formed in the pour- and airholes wasremoved with a small metal grinding wheelattached to a flexible drive drill. Tiny faults inthe cast in the form of holes caused bytrapped air bubbles were filled using colouredresin applied with a syringe. When the resinhad fully cured the restored areas werepolished to a glossy finish using a felt polish-ing buff attached to the dental drill, and SolvolAutosol, a mild abrasive paste (Figure 7.40l).In the case of simple globular shapes, it maybe possible to use a toy balloon to form aninner mould.

Replacement by casts from a mouldtaken from a clay model or a previousrestoration

Clay modelsIn the case of glass vessels having extensive,or more complex, missing areas, or where araised design must be copied, a former isconstructed of modelling clay on to which theremaining fragments are placed. The missingpart of the design is modelled in the clay, andthe entire assembly is then moulded, afterwhich the clay is removed in order that resinmay be cast into the areas of missing glass(Figures 7.41 and 7.42a–d). Figure 7.41 showsthe restoration of a missing section of applieddecoration on a Saxon cone beaker. Figure7.42a shows a group of Roman glass vessels,which were restored by modelling the missingareas in clay and then moulding and castingthem.

The first step is to make any joins in theoriginal glass. For this a cellulose nitrateadhesive such as HMG will suffice, becausethe various processes that follow will causethe fragments to part, and the adhesive willbe easy to remove. The fragments are thenplaced over a former of clay, which may beturned or modelled up by hand, on a sheet ofplate glass on a turn-table (Figure 7.42b). Thearea of plate glass around the model is coated


Figure 7.41 The missing section of an Anglo-Saxoncone vessel, modelled in clay, prior to being moulded.

with Vaseline or liquid soap to ease theeventual removal of the mould. It is usuallybetter to invert the glass vessel so that theeven rim rests on the plate glass and the vesseldoes not bear the weight of the clay. Wherefragments are missing, the clay is built up tothe outer surface of the glass and any model-ling of the design is carried out, loosefragments being floated into the design. Ifnecessary, two clay stumps are attached at thehighest point of the clay model, which willeventually form pour- and airholes through themould.

A mould of the entire structure is thenmade. Silicone rubber is poured over thewhole assembly, to form a thin layer, ensur-ing that it entirely covers the glass and clay.

A wall of clay can be laid around the glass,about 20 mm from it, to contain the flow ofrubber. When the rubber has cured, a secondlayer, mixed to a stiff consistency with an inertfiller, is applied over the first layer tostrengthen it and to prevent the rubber fromtearing when it is eventually removed, butensuring that the mould remains slightly flexi-ble (Figure 7.42f). (If the object being restoredis relatively small, with no great curvature, thesilicone rubber layer may suffice, so that asupporting plaster of Paris or resin mother-mould will not be needed.) When the siliconerubber has cured, the mould is carefullypeeled off the plate glass and inverted. Takingcare not to disturb the glass fragmentsattached to the interior of the mould, the clay


Figure 7.42 (a) A group of Roman glass vessels restored by modelling their missing sections in clay, makingsilicone rubber moulds and casting in the missing areas with resin. (b) Remaining fragments mounted on a clayformer. (c) After moulding the exterior, the assembly is inverted, and the clay removed. (d) A thin layer of clayreplacing the missing glass.

(a) (b)

(c) (d)

is removed from the interior, except for a thinlayer, which fills the missing areas of thevessel (Figure 7.42d,e). The top edges of therubber mould are coated with a release agentsuch as Vaseline, and a silicone rubber mouldmade of the interior of the vessel, as describedabove (Figure 7.42f).

When the rubber has cured, the mould isremoved, followed by the remaining clay,again not disturbing the glass fragments. Wateron cotton wool swabs is used to remove finaltraces of clay from the glass edges. Fragmentsthat do become dislodged can be replacedusing a small amount of silicone rubber as an‘adhesive’. Faults in the mould can be madegood by filling them with clay and sealingthem with polyvinyl alcohol release agent

(Figure 7.42g). The surface of the glass iscoated with polyvinyl release agent, followingwhich the interior mould is replaced andsealed to the outer mould with silicone rubber.This will prevent the mould from floatingwhen the resin is introduced (Figure 7.42h).Once the resin has cured, the moulds areremoved and any finishing of the cast carriedout (Figure 7.42i).

Moulds made around large or heavy claymodels and glass objects (Figure 7.43a) mayrequire the addition of a plaster of Parismother-mould (Figure 7.43b–d), about 40 mmthick, constructed with a flat top, so that it willstand safely when inverted. (Note that a spher-ical shape cannot be moulded in plaster ofParis in fewer than three pieces since the


(e) (f)

(g) (h)

(e) Moulding the interior. (f) Removal of the interior mould. (g) Interior mould secured with silicone rubber, andresin introduced through a gap at the front edge. (h) The completed restoration.

curvature would prevent its removal – Figure7.43b.) No release agents are required eitherbetween the glass and the silicone rubber orbetween the silicone rubber and the plaster,but a release agent such as Vaseline or liquidsoap is required between the plaster pieces,and between them and the plate-glass. Theplaster pieces should be numbered in order ofmanufacture since they will be removed inreverse order, and, on completion, the plasterpieces are secured to one another with cottonbandage or scrim dipped in plaster of acreamy consistency. When the plaster has set,the whole assemblage is inverted on to its flattop (which thus becomes the base). Thesilicone rubber flange is trimmed back to theinner edge of the plaster, and the bulk of the clay former removed with a spatula, takingcare not to dislodge any of the glass fragmentsnow supported in the silicone rubber mould.If any fragments are loosened, they should bereattached with silicone rubber (see below).Clay is left in the areas of missing glass andsmoothed with a spatula until it is level withthe inner surface of the vessel, i.e. represent-ing the missing glass. Vaseline is applied tothe top edge of the silicone rubber mould anda thin layer of silicone rubber poured over theinterior. When this has set, a second thickenedlayer is applied as before, and a 10 mm wideflange modelled up at the rim.

The interior of the vessel is filled with aplaster piece-mould in order to support thesilicone rubber, and on which to key a plasterlid on which the mould will stand duringcasting. Figure 7.43c demonstrates the way inwhich the plaster pieces should beconstructed, chamfering the edges of eachpiece so that they do not overlap one another.To ensure that these plaster pieces can eventu-ally be removed easily, each one should beremoved after it has set, sealed with shellac,


Figure 7.43 Diagrams showing (a) a silicone rubbermould over a missing area of glass which has beenmodelled in clay; (b) the construction of a plaster ofParis mother-mould over the silicone rubber mould; (c)the construction of a plaster of Paris piece-mould overthe silicone rubber inside the vessel; and (d) thecompleted plaster of Paris mother-mould.

and Vaseline (white petroleum jelly) appliedas a release agent before casting the nextpiece against it. The tops of the pieces mustbe flush with the top edge of the rubber andthe outer mother-mould, except for the centrepiece which should project for ease ofremoval. (This being the last piece made, willbe the first to be removed.) The top edges ofthe interior plaster piece-mould, and themother-mould, are given a coat of shellac andVaseline. Plaster of a creamy consistency isthen applied over the top area in order toform the lid about 40 mm thick.

When the plaster has set, the lid is removed,followed by the inner piece-mould, the innerrubber mould (which may have to be slit witha scalpel and later resealed with siliconerubber), and the clay. On no account shouldthe mother-mould be disturbed. Every trace ofclay must be removed with moist swabs ofcotton wool, and any loose fragments of glassrejoined with HMG.

It is very likely that floating fragments ofglass may become loose at this stage. The onlysatisfactory method of replacing them is to usesilicone rubber as an adhesive (for thispurpose it may be overcatalysed to save time)and to wait for this to set before proceedingwith the restoration. When all the fragmentsare secure in the mould, a 5 mm wide, normalmixture of silicone rubber is painted aroundthe areas of the missing glass, leaving a 25 mmgap between the silicone rubber and theedges. The inner silicone rubber mould is thenreplaced before the silicone rubber sets andthe strip of silicone rubber seals it to the glasssurface (the 25 mm gap from the edge of thebreak is to allow for the liquid rubber tospread out during replacement of the mould).This prevents resin seeping into the gapcreated by removing the mould and breakingthe suction between it and the glass. This isextremely important since the resin finds itsway into the smallest gap.

Next, the upper edges of the silicone rubbermoulds are degreased with a tissue and sealedtogether with a normal mixture of siliconerubber; this again is to prevent resin seepage.When the silicone rubber has set the plasterpieces are replaced, followed by the lid, whichis sealed to the mother-mould with plaster andbandage as previously described. The wholeassemblage is then inverted and catalysed

resin is poured into the mould through onehole until it appears in the other, indicatingthat the mould is full (Figure 7.43d). Whenthis occurs, the resin is left to cure, after whichthe mould can be dismantled and the glassand cast removed together. Any joins thatcome apart are remade with HMG or, if thecondition of the glass allows, it may be recon-structed using epoxy resin.

Previous restorationIf a previous restoration needs replacement foraesthetic reasons, but is dimensionallyaccurate, a mould may be taken from it beforeit is removed (see restoration of the AuldjoJug, described above). This will dispense withthe need to model-up the missing area in clayor Plasticine. Figure 7.44 illustrates such acase, in which only the base, and half thecircumference, had remained of a blue glasspyx. The earlier restoration had consisted offilling the remains with plaster of Paris,shaping it to the original dimensions andcolouring it blue to match the glass. Therestoration was heavy and obscured theinterior of the jar, and therefore it was decidedto remove the plaster and to replace it with amore aesthetically pleasing polyester resin.However, since the restoration was dimen-sionally correct, a mould was taken from theplaster prior to its removal.


Figure 7.44 Small blue glass pyx restored by filling itwith plaster of Paris, painted to match the colour of theremaining glass.

The pyx was inverted on a piece of plateglass, and thickened silicone rubber pouredover it to a thickness of 15 mm. Once thesilicone rubber had cured, it was slit downone side, using a sharp scalpel, to allow theglass to be released. The slit was carefully cutagainst the original glass so that when the slitwas later joined using silicone rubber as‘adhesive’, a seam mark would not be left inthe cast. Meanwhile, the plaster of Paris wasremoved from the glass by soaking in water,followed by the use of a scalpel. Once theglass fragments were clean and dry, joins weremade with a cellulose nitrate adhesive. Theglass was then replaced in the mould in itscorrect position and the missing area of glassfilled with pieces of toughened dental wax, cutto shape and melted together where necessaryusing a heated spatula. (The wax happened tobe exactly the same thickness as the walls ofthe vessel.) A short length of acrylic rod(Perspex; US: Plexiglas) was fixed at each endof the wax (by softening the wax) to formholes through to the interior mould which wasto be constructed. (Plastic drinking straws orwax stumps would have been a better choice,as the rigid Perspex rods proved extremelydifficult to release from the silicone rubber.)The top edge of the silicone rubber wasgreased with Vaseline to act as a separatingagent, and thickened silicone rubber appliedto the interior to a thickness of 15 mm. Therelative positions of the two mould pieceswere marked, by drawing lines across the jointwith a felt-tip pen.

Once the silicone rubber had set, it wasremoved, followed by the acrylic rods anddental wax. At this stage several of the joinsmade in the original glass broke down andhad to be remade with HMG. The glass wasagain inserted in the outer mould and theinner mould placed in position. The twomoulds were sealed using catalysed siliconerubber around their junction to prevent resinescaping. Tiranti’s clear polyester embeddingresin, coloured with blue translucent pigment,was then poured into the mould through onehole until it began to appear at the other, thusindicating that the mould was full. After a fewminutes the resin had settled in the mould andit was topped up. When the resin had set, thejoint between the two moulds was slit with ascalpel, and the inner mould was released.

The two resin stumps, left in the pour- andairholes, were abraded away with a drill andthe circles left in the resin were polished untilthey were hardly visible.

By leaving the glass and cast in the outermould during these operations it was hopedto prevent vibration causing separation of thecast from the glass fragments, but once theouter mould was slit open, the glass and castseparated from one another as they wereremoved. This was turned to advantagebecause a thin layer of resin had seepedbehind the glass in the mould. This was easily


Figure 7.45 Repair of glass objects with narrow necks,where the interior is inaccessible. (a) A section of glasspieced together in situ, taped and bonded, thenremoved and gap-filled, replaced and bonded. (b) Arepaired section of glass removed from the side of avessel, in order to gain access to the opposite side,which requires gap-filling. (c) A repaired objectseparated above an area of loss, to enable gap-filling tobe carried out. (d) A repaired object separated alongthe edge of a missing area, to enable gap-filling to becarried out.

released from the glass using a scalpel, andcould be detached from the now separate cast,which could be polished before joining it tothe original glass, with epoxy resin.

Gap-filling where the interior of thevessel is inaccessible for workingIf the vessel to be restored has a narrow neck,so that the interior cannot be reached, it willbe necessary to adopt special methods ofrestoration (see Figure 7.45a–d and 7.46a–c).If the vessel is damaged only in the body, thatis, the neck has not become detached, thefragments must be taped in position andepoxy resin introduced to all the cracks exceptthose around the perimeter of the damagedarea. Care has to taken to ensure that the resindoes not flow to the perimeter, which can be

given a coat of PVA release agent, to ensurethat the two parts of the object can beseparated after the resin has cured.

Once the resin has set, the damaged areacan be removed as one piece, missing areaswithin it backed with tape, wax or siliconerubber, depending upon their size, and filledwith resin. When the resin has hardened andany necessary cleaning has been carried out,the repaired section may be taped in positionon the vessel, and the perimeter cracks sealed(Figure 7.45a and 7.46a–c).

If, however, the vessel is much damaged,and has a large area of glass missing from thebody, it may still be possible to reconstruct itwith epoxy resin as described above, leavinga large area unglued around the edges so thatit can be removed (Figure 7.45b). This allows


Figure 7.46 (a) Broken Roman glass flask with anarrow neck, which makes the interior inaccessible forworking purposes. (b) Completed restoration of thevessel. (c) Closer view of the restored area of thevessel.

(a)

(b)

(c)

access to the interior of the vessel for ease ofworking. When gap-filling of the missing glassis complete, the glass can be replaced and theperimeter cracks sealed. In the case of a tallglass amphora, a sheet of dental wax wasshaped over the outside of the glass, coatedwith polyvinyl alcohol, and then transferred tocover the missing area and fixed there byrunning a warm spatula around its edges. Alarge area of glass was removed from theopposite side of the vessel to enable work tocontinue in the interior. A piece of glass-fibresurfacing tissue was cut to the shape of thegap to be filled, laid in it and covered with athin layer of polyester resin. When this had seta second layer of resin was introduced and,when this had hardened, the wax mould wascut away, the polyvinyl alcohol removed withwater, and the resin cleaned and polished.(See also restoration of the Auldjo Jug,described above.)

Another approach is to reconstruct a brokenobject in two parts, and separate it along joinswhich do not run through an area that needsto be filled (Figure 7.45c). Restoration ofmissing areas is carried out, after which thetwo parts are joined together. If the separationhas to be made along a line adjoining amissing area (Figure 7.45d), the procedure isas for Figure 7.45c, except that the resin castwill have to be carefully abraded along its freeedge in order that the two parts of the glasswill fit together.

Small holes, especially those in the cornersof decanters, can only be filled with somedifficulty, and the owners should be advisednot to put the vessel to everyday use. First, awax mould is taken from a complete cornerand its inner surface coated with separatingagent. The mould is then sealed over the holeon the outside, by melting its edges onto theglass using an electrically heated spatula. Thedecanter is then positioned on an angle so thatthe corner is at its lowest point. Epoxy resinis then carefully introduced to the mould, adrop at a time, through the neck off the tipof a long bamboo satay stick. Great care mustbe taken not to introduce more resin than isrequired to fill the hole, since of course it willnot be possible to smooth its surface once theresin has cured. Similarly, any resin, which isdropped on the side of the glass must beremoved before curing, using swabs of cotton

wool lightly moistened with acetone. Once theresin has cured, the mould can be removedand any excess wax cleaned away with swabsand white spirit.

Detachable resin fillsDetachable resin gap-fills for the restoration ofmissing glass rim, feet and handles are easierto make than those for replacing losses withinthe body of a glass (see restoration of theoinochoe described below). This technique hasthe advantages over casting the resin in situof not creating undue stress on weak joins orglass during moulding and casting of the resin,and of minimizing contact with the surface ofthe glass during the finishing processes(sanding and polishing) (Koob, 2000).

Thermosoftening resinsAn alternative method of replacing missingareas of glass was to cut the shapes from pre-formed acrylic sheets (Perspex; US: Plexiglas),bend them to the required curve after soften-ing with a hot air blower, and attach them tothe glass with adhesive. However, this is alengthy process requiring accuracy in cuttingand filing, and the finished result is in no wayas aesthetically pleasing as casting in themissing fragments with a clear resin; it is there-fore not recommended for general use.

An alternative approach makes use of thefact that methacrylates are thermoplastic,softening at temperatures around 600°C. Oncecured, they can be softened by heat andreshaped by hand. The thermosofteningproperty of Technovit 4004a (a grade ofpolymethyl methacrylate, no longer available)was exploited by Errett (1972) to reshapesmall casts and thus dispense with the needto make complicated moulds. The techniquewas also used by Jackson (1983) to restore ablue glass, core-formed Italic oinochoe (jug)with prunts, dating from the sixth or fifthcentury BC (Figures 7.47a–f ).

The oinochoe was about 80 mm in heightand the glass itself was in fairly good condi-tion (Figure 7.47a). There was some surfaceweathering but no actual flaking occurring.The neck, spout and handle were missing butthe foot and prunts had previously beenrestored with wax. It was decided to leavethe previous restorations intact since theywere dimensionally and aesthetically stable,



Figure 7.47 Restoration of the neck, spout and handleof a glass oinochoe (jug). (a) The remaining glass; (b)the neck and spout modelled in clay, prior to moulding;(c) neck and spout cast in Technovit 4004a. The castwas produced in a silicone rubber mould made over themodel shown in (b). (d) The handle was made bycasting resin into a drinking straw. (e) Spout and handleshaped and attached to the glass. (f) The completedrestoration coloured to tone with the original glass.

and to reconstruct the neck, spout andhandle, which could be copied from a photo-graph of a similar oinochoe. From the illus-tration it can be seen that in order to producea one-piece cast, the mould required wouldbe both difficult and time-consuming toproduce. In particular, the thin, curved flangeforming the spout would present problems inthe moulding process. Experiments wereundertaken in heating and reshaping simpleresin casts to form complicated shapes. As apreliminary to the project, several discs of asize and thickness similar to those of thespout were cast in Vosschemie (polyester)resin and Technovit 4004a. When cured, thediscs were heated in various ways (i.e. warmair, warm water and boiling water) until theybecame pliable enough to fold into shape.The folded discs were then held in cold wateruntil they hardened. It was found that thepolyester resin disc softened at a much lowertemperature than the methacrylate disc,which had in fact to be placed in boilingwater for a few minutes until it could beshaped easily. The polyester resin discsoftened easily in warm air, but in warmwater it became opaque. The discs of resinwere left overnight under a desk lamp(approximately 27°C) to determine whetherthey would distort under display conditions.The polyester discs flattened considerablywhereas the shape of the methacrylate discdid not alter at all.

The neck and spout of the oinochoe weremodelled in situ in fine-quality potters’ clay(Figure 7.47b). The curved flange of the spoutwas not actually modelled to shape but wasleft as a flat disc and the neck was modelledwithout a hole through the middle. Once theclay was leather hard (after 4 hours) the modelwas removed from the oinochoe and mouldedwith Dow Corning Silastic E RTV (siliconerubber). After curing, the rubber mould wasremoved from the clay and the mould cleanedwith water. The mould was filled from the topwith Technovit 4004a, which was allowed tocure. The rubber was then cut away revealingthe cast (Figure 7.47c). Using the appropriatesized drill bit, a vertical hole was drilledthrough the centre of the resin cast. Holdingthe piece by the neck with tongs, the flat resincast was placed in boiling water until it hadsoftened, removed and the edges of the disc

bent up by hand to form the flange shown inFigure 7.47e. When the required shape hadbeen achieved the resin was immediately heldin cold water until it hardened. The surfacewas then finished with fine glass paper. Themoulded neck and spout were then attachedto the broken neck of the oinochoe withAraldite AY103/HY956 (epoxy resin) and gapsbetween the glass and the cast filled withTechnovit 4004a.

The handle was produced by a similar butmore direct method (Figure 7.47d). A rod ofTechnovit 4004a was cast using a glass tubeas a mould. The tube was filled with resin;when this was hard (approximately 30 mm)the glass was broken away. As shown in thephotograph, plastic straws could have beenused as moulds, but in this particular case theglass tube happened to be of exactly the rightdiameter. The resin rod was subsequently cutto the length of the original handle, heating inboiling water, bent to shape and hardened incold water. The handle was attached to theoinochoe with Araldite AY103/HY956 (Figure7.47d). Deka glass colours (fabric dyes) wereused to colour the reconstruction to match theglass (Figure 7.47f).

For this particular object the method usedproved to be very successful. The workingtime was substantially reduced, first because itwas not necessary to make a complicatedpiece-mould from a clay model; and secondlybecause the methacrylate resin used cured in30 mm whereas polyester resin would havetaken 24 hours. Furthermore, there were noproblems with mould relocation on theoinochoe or removal of flash-lines left on thecast by use of a piece-mould. As the Technovit4004a could be softened and reshaped after ithad been cast into a basic form it was veryeasy to make minor adjustments to the curvesof the spout and handle. Thus the methodproved to be both much more flexible and lesstime-consuming than casting alone and shouldbe considered whenever a complicated shapeis to be copied and water-white clarity of therestoration is not a priority.

Thermohardening resinsA method of filling gaps in glass with pre-formed casts of epoxy resin was devised byHogan (1993), in order to enable a badlydamaged green glass bottle to be exhibited.


The vessel, approximately 25 cm high, had anunusually thin base (c. 1 mm.) in comparisonto its thicker, heavier, neck and rim (c. 3 mm.).About a third of the original vessel wasmissing. The bottle was reconstructed in twoparts, the main body and the neck and shoul-ders, using a UV curing acrylic (Loctite 350engineering adhesive).

The joins between the top and bottom ofthe vessel were not substantial and somefilling was necessary to give support. Amethod of support with minimum use ofstrategically placed resin infills was devised.Detachable resin fills were constructed awayfrom the object and adhered to the glass aftercuring. Sheets of epoxy resin AY103/HY956were cast, to correspond to the varying thick-ness of the glass vessel in simple hexagonalmoulds, made from sheets of dental wax.When the cured resin was removed from themould a thin film of wax remained on itssurface, and was removed with a spatula,followed by cleaning with white spirit. A sheetof dental wax was placed across the gaps inthe glass that were to be filled, and the outlineof the gap scribed onto the wax. The shapeswere then cut out of the wax and placed insitu in the vessel, to ensure that they fittedwell, and secured there with strips of pressure-sensitive tape (Sellotape). Each wax shape wasremoved in turn, using acetone to release theSellotape, and laid on a resin sheet of therequired thickness. Its outline was scribedonto the resin. The resin was then gentlyheated with a hot air blower to make it pliableenough for the shape to be cut out withscissors. The edges of the resin sherds werefiled where necessary with fine metal needlefiles to ensure close contact with the glass.Where appropriate the resin sherds werereheated and the curvature was modified tocomply with the contour of the glass vessel.The finished pieces were placed in positionand secured with HMG (cellulose nitrate)adhesive. The restored areas were thenpainted with one coat of Rustin’s Clear PlasticCoating (urea formaldehyde) coloured withMaimeri Restoration colours to match thecolour of the glass.

As this method of gap-filling provedsuccessful, variations upon the techniqueusing different resins were used to restoreother glass vessels. A clear Anglo Saxon glass

was gap-filled using the same method butusing clear HXTAL NYL-1 epoxy resin to formthe cast. The resin was heated gently byimmersing the pieces in warm water, beforebeing shaped and bonded in position withHMG. Another small, delicate flask neededsupport to its neck in order to connect it tothe body and base. A mould was made ofthe interior of the neck, by forming a cylin-der of dental wax. A sheet of Araldite 2020(epoxy resin) was cast in a wax mould aspreviously described. Before it hadcompletely cured, the resin was removedfrom the mould, and while still flexible, itwas formed around the plaster core usingCling Film plastic food wrapping as a barrier,and left to cure fully. Once the resin hadcured, the plaster core was removed by split-ting the resin with a scalpel blade. The resinbacking, now in two sections, was insertedin the neck of the flask, and adjustmentsmade to secure a good fit. It was thenbonded with HMG, giving full support to theneck. It is essential that the HXTAL NYL-1and Araldite 2020 resin casts have fully curedbefore being bonded in position with HMGto prevent discoloration of the epoxy resin.

The method proved to be extremely effec-tive, both in terms of the final appearance ofthe glass vessels, and in the support provided.Being able to work on the resin fills awayfrom the glass surface, avoiding the need forindividual moulding and casting of sherds,makes this a quick and safe method of fillingfragile glass. The fact that the resin pieces arebonded in position with cellulose nitratemakes future removal easy and safe forarchaeological glass. Disadvantages may be inthe heating of the resin, which could acceler-ate discoloration. However, no immediateyellowing of the HXTAL NYL-1 or Araldite2020 was observed. In the case of the AralditeAY103 used on the green glass, this aspectmay not present a problem. A previous caseof heating HXTAL NYL-1 epoxy with a hot airblower caused yellowing of the resin. In viewof this the use of warm water to soften theresin is preferred. This method could, intheory, be used to complete the gap-filling ofan entire vessel, depending on the intricacy ofthe missing shapes. Detachable fills for restor-ing ancient glass are also discussed by Koob(2000).


Commercial ‘restoration’

There are companies who quite legitimately useglassmaking techniques to reproduce copies ofancient and historic glass objects (see below),and replacement parts for chandeliers, andobjects such as silver salts and sugar bowls,requiring blue glass liners. Others grind andpolish chips on the edge of glass items such aswine glasses, and ‘polish’ the interior of vasesand decanters (by removing the surface withfine abrasive or hydrofluoric acid). However,there are others who specialize in cutting andmodifying glass, and even marrying parts ofsimilar objects together for the glass antiquetrade. Glasses can be expertly cut in such a waythat joins made with adhesive between two ormore pieces can be concealed. Sometimessilver mounts are transferred from one glassobject to another, or added to conceal joins.

Replicas and fakes

From the foregoing section it will be realizedthat the conservator is not only able to carry outrepair and restoration to great effect, but that itis also possible for replicas of glass objects tobe made (von Saldern, 1970). Wihr (1963, 1968)made copies of a whole series of Roman glassvessels, and illustrates three of them: a dishengraved with a design of Abraham and Isaac,an extremely delicate diatreta glass, and aPortland vase (Wihr, 1963). Methods of glassreproduction are also given by Petermann(1969).

Replicas of glass vessels may be producedfor loan to institutions and/or for studypurposes where the original is too valuable orfragile to be handled (Goldstein, 1977). Withthe synthetic materials now available, thetransparency and colouring of the replica canbe so good that, visually, it can be very diffi-cult to distinguish between the original andthe copy (Mehlman, 1982). However, differ-ences both in the weight and feel of thesynthetic material, compared with the glass,will usually enable one to distinguish betweenthe two. The ability to produce copies of glassvessels raises the possibility of replicas beingpassed off as original, but this would beunusual for the reasons mentioned above. Itwould, however, be easier for a broken vessel

with a heavy weathering crust to be heavilyrestored and presented as being complete.

It is also possible to copy glass vessels inmodern glass and, if necessary, create a patinaby chemical treatment. In the last quarter ofthe nineteenth century archaeological excava-tions, the development of museum collections,exhibitions of famous glass collections andpublications such as The Stones of Venice(Ruskin, 1852) and pattern books full ofdesigns, created a fashion for historical repro-duction of glass in glass, throughout Europe.Bly (1986) records the types of glasswarewhich have been reproduced or faked, andgives indications as how to identify them.

A more serious possibility of faking wouldbe that of modern decoration of an ancientglass vessel, for instance by engraving a designon the surface; or by artificial ageing (perhapsby etching with hydrofluoric acid) to increaseits monetary value. Pilosi and Wypyski (1998)describe two ancient glass vessels with moderndecoration in the Metropolitan Museum of Art.No doubt further such examples will continueto be exposed in other collections. Glass is stillbeing made, altered or ‘restored’ with theintention of deceiving the buyer. Ancient glassis particularly easily faked, as the modern glasscan be artificially aged with chemicals, ordisguised by applying flakes of iridescence andmud, sometimes bound in an adhesive, to itssurface. Parts of similar objects may be cut andmarried together beneath a muddy or fakeiridescent surface (Plate 7). Modern sources offorgeries are Israeli, Turkish and Egyptianglassblowers, working in primitive conditions.Knowledge of glass style and technologyusually enables a conservator to identify fakes:for example, the visual appearance of thesurface, mould flash lines where there shouldnot be any, drilled holes in what should behollow core-formed vessels (Figure 7.48a),incorrect patterns and colours (Figure 7.48b).By the use of a binocular microscope or ascanning electron microscope, it may be seenthat detail such as engraving continues unbro-ken over a damaged surface, or that the‘engraving’ is in reality etched, so that theedges of the lines are polished, not cut(Werner et al., 1975). Examination under ultra-violet light may enable areas of restoration tobe identified, if they fluoresce differently fromthe original glass.


In the 1970s a number of glass figures ofbulls, purporting to be ancient, appeared forsale at much the same time in severalcountries, a fact which in itself raised suspi-cion. The figures were exposed as forgeries byWerner et al. (1975). It was discovered that thelegs had been added by lamp-workingtechniques, rather than having been drawn outof the softened body of glass with pincers.Their flaky, iridescent covering containedapatite, showing that the finished articles hadbeen exposed to the vapour of hydrofluoricacid (discovered in 1770). Newton and Brill(1983) identified a ‘Roman’ weeping glassbowl as a forgery or hoax. Chemical analysisof the glass showed it to be free from traceelements and to contain only 0.33 per centCaO and 0.01 per cent MgO. The lack ofmodifying oxides seemed to indicate that thebowl had been made either by a moderninexperienced glassworker or had been madefrom laboratory glass. X-radiography can alsoprove useful in detecting fakes, especially ifmetal components have been incorporated, orparts of more than one object have been fittedtogether beneath a convincing deteriorationlayer or restoration (Figure 7.49).

Storage and display

Materials used for display and storage of glass– wood, adhesives, varnishes, paints andtextiles – can become carbonyl pollutants by

emitting acetic, formic, formaldehyde andacetaldehyde acids. Unstable glass is particu-larly vulnerable to attack from these acids,


Figure 7.48 A fake Egyptian core-formed vessel. (a) The pattern is incorrect, the object is heavy, and whenviewed from the top (b) it can be seen that the vessel is not hollow, but has been drilled.

(a) (b)

Figure 7.49 Radiograph of a glass ewer in which canbe seen the addition of a ribbed ewer spout and neckto a pincer-decorated bottle. (The addition of a snake-like handle of plastic, seen as a faint shadow on theright of the radiograph, is not visible in the illustration.)Note that there are several non-joining pincer-decoratedrosettes and concentric circles ‘floating’ in the body ofthe vessel. H 146 mm. (Corning Museum of Glass).

resulting in the formation of white crystallinesalts on the surface and within hollow partsof vessels. Ideally display cabinets should beconstructed from metal and glass; this has theadded advantage of glass objects being highlyvisible. Glass objects should not be allowed totouch each other, and unless by deliberatedesign, shelves should not be overcrowded.Vessels that have more than one componentneed special care; lids and stoppers shouldeither be stored separately (and identified), ortied in place with lightweight Nylon fishingline, knotted at the ends, so that they cannotfall when the vessel is lifted. Smaller items ofglass should be placed at the fronts of theshelves so that they are easily seen. If mouldgrowth is noticed, for instance on animal gluerestoration, it must be removed with a dilutesolution of a disinfectant such as Panacide onswabs. It may be necessary to dismantle andrestore such a vessel, using a more suitableadhesive.

Display cases and the shelves within themshould be stable and level so that glassexhibits cannot move due to vibrations causedby visitor-movement, passing vehicles, trains,underground railways or aeroplanes etc. Ifnecessary, glass objects can be supported byacrylic mounts; and mirrors may be used toreflect light upward on to details of decorationso that the objects need not be placed immedi-ately over light bulbs. Ideally, lighting shouldbe external and properly situated so as not tocause a heat build-up, and infra-red absorbingfilters should be used. The effects of heatbuild-up can be very serious on certain typesof glass, for example glass with incipientcrizzling, painted surfaces or weatheringproducts (Brill, 1975, 1978), because it lowersthe relative humidity of the air. The deleteri-ous effects of spotlighting glass vessels fordramatic exhibition are appreciated by conser-vators but not by all curators or collectors.

Enamels should be kept in a constanttemperature so that the differential expansionbetween the glass and the metal does not leadto disruption of the object. Those enamels inwhich it has not been possible to stabilize themetal, must be kept in an atmosphere that isneither too damp to allow the metal tocorrode, nor too dry to desiccate the glass. Inthe case of copper alloys this would be about35 per cent RH. Brill (1978) has discussed a

method of controlling humidity in museumcases. Special storage conditions are requiredfor unstable glass and enamels (see below); itmay be necessary to install a de-humidifier instorage cases containing large numbers ofunstable glasses/enamels (Figure 7.50).

Navarro (1999) describes and illustrates themounting of some 1,500 small glass items fordisplay in the glass gallery of the Victoria andAlbert Museum, which opened in 1994(Oakley, 1999). The majority of the pieces arefragments, but also include complete items,such as beads, cameos, medallions, cutlery,bowls and bottles and glass tiles all between1 cm and 20 cm in size. The methods bywhich the fragments were secured, were


Figure 7.50 Storage of weeping and crizzling glassand enamels in a relatively air-tight case incorporating adehumidifier which maintains the atmosphere inside at aconstant relative humidity. (© Copyright The BritishMuseum).

embedding, pinning or sewing onto PlastazoteLD 45 (a cross-linked, closed-cell, chemicallyinert polyethylene foam), used singly or incombination depending upon the require-ments of each piece of glass (Figure 7.51a–c).

They were mounted in two chests of eightdrawers each, and constructed in metal, glass

and plastic. The drawers incorporate acomplex braking system designed to reduceimpact when the drawers are closed, and theirdesign is such that no two drawers can beopened simultaneously. Each drawer containsthree white plastic trays and is sealed with asheet of clear laminated glass. The trays werelined with two 4 mm layers of white Correxboards (ethylene propylene copolymer extru-sion) and one 9 mm layer of white Plastazote.The glass fragments were mounted as follows:each tray was lined with paper, and theobjects arranged for display. An accuratepencil line was then drawn around each pieceof glass, and then as each was transferred toa lined drawer, its accession number writtenbeside the outline in order to identify it. Theplan served as a permanent reference, andwas filed with other related documentation.

The glass fragments were then secured tothe Plastazote by the methods mentionedabove. The most common of these was to cutthe shape of the object out of the Plastazoteusing a scalpel blade, thus forming a hollowin which to house the object. It was thenpinned into the Plastazote with headless orNylon-headed stainless steel pins varying inthickness between 0.3 mm and 0.6 mm, theheads of which had been coated with ParaloidB-48N (methyl methacrylate copolymer) toprevent them from scratching the objects, andto provide a good hold. The pins, were cut,bent and shaped to match the requirements ofeach object. They were long enough topenetrate through the Correx board. Wherethe pins were in contact with break edges ofwith large areas of glass, they were coveredwith heat-shrinkable polyethylene tubing(HST). Pinning was not suitable for attachingheavy objects (and would not be suitable forattaching heavy objects in drawers without abraking system). Some glass objects weresewn onto the Plastazote using Nylon fishingline (10 lb or 20 lb breaking strength). A pareddown piece of Correx board placed beneaththe foam acted as an anchor and minimizeddistortion of the foam and movement of theobjects. Where sewing was used to secureobjects with a particularly fragile surface, thefishing line was covered with HST. Somemedallions were too small to pin or sew inplace and were placed in cut out shapes inthe foam and secured with a dot of Paraloid


Figure 7.51 Securing glass fragments and small objectsto a substrate for storage. (a) Part of a wine glass stemembedded in Plastazote (see also Figure 7.11). (b)Pinning; a fragment and threaded hollow beads pinnedto a substrate. (c) Sewing: a heavy object secured withNylon fishing line; the buttons made from Correx,prevent the Plastazote from buckling. (After Navarro,1999).

B-72 (methyl methacrylate copolymer) appliedto their reverse side, and avoiding the acces-sion numbers.

Handling

Glass, being light, is easily knocked over, andin the case of clear, transparent glass, is easilydamaged by misjudging distances when settingit down. Some objects are particularly proneto damage, e.g. figures with projecting limbs,tall top-heavy objects that can topple, objectsmade in two or more parts or having weakoriginal joints such as handles, and objectswith rounded bases. When items have to bemoved out of the way in order to reach anobject, they should be placed on anotheradjacent surface and not merely pushed to oneside, especially if the shelf is crowded. On noaccount should one reach over glass vesselsto move others at the back of a shelf.

Before moving a glass, a check should bemade that it does not consist of more than onepiece, or that any previous restoration is notfailing. Glass objects should never be pickedup by the rim, but the bowl should be cuppedin one hand and the base supported in theother in order to cradle the glass againstknocks. Vessels should be carried one at atime unless a carrying basket or tray is beingused, in which case the trays should be linedwith cotton wool covered with neutral tissuepaper, and the objects separated from oneanother by twists of tissue paper.

One should never turn to talk to anyonewhile setting a piece of glass down, since thedistance between the base of the object andthe table-top will almost certainly bemisjudged and the glass may be broken. Thebase of a glass should be set down flat andnot heavily at an angle. Glass objects shouldnever he left near the edge of a table as theycould easily be brushed against and knockedover, particularly as they can be difficult tosee. Adherence to these simple rules will helpto prevent accidents.

Packing

Temporary packing of glass vessels for transitwithin a building has been dealt with above.Before objects are sent on loan, detailedcondition reports, which include photographs

of any damage, should be made. On arrival attheir destination, and again on their return, theobjects can be checked against the originalphotographs.

For transport to other institutions at homeor abroad, glass vessels should be extremelywell packed in sturdy wooden boxes filledwith sheets of solid polyester or polyethylenefoam such as Plastazote (Figure 7.52). Thepacking operation should be planned so that,if possible, glass is separated from othermaterials, and that in any event, heavierobjects are placed at the bottom of thecontainer. Enough squares of 50 mm thickfoam should be cut to fill the box being usedto transport the glass. (Other materials whichcan be used are Kempac, Bubble-Wrap andthick wads of twisted acid-free tissue paper:Wakefield, 1963; Lindsey, 2000.) The vesselsshould be placed on a sheet of foam (on theirsides if they are tall), their shapes being firstmarked on the sheet and then cut out usinga sharp knife, through enough layers of foamto encompass their depth. There should be atleast 50 mm of solid foam between twoobjects. The vessels should then be wrappedin acid-free tissue paper, and the foam cut-outs trimmed down and fitted inside the glass.They are thus fully supported inside and out,and cushioned from shock. Should an accidentoccur, or a repair break down in transit, allthe fragments will be retained in the tissuepaper in their relevant positions.

The lids of boxes are best secured with ahook-lock, which can be easily opened for


Figure 7.52 Glass objects packed for transport outsidea building. Each object is wrapped in acid-free tissuepaper and set into a shape cut out of dense Plastazotefoam, laid in a strong wooden crate, secured with metalclasps.

Customs inspection if necessary, and whichwill enable the box to be re-used. If boxeswith screwed-on lids are re-used the screw-holes eventually become enlarged and there isa possibility of the lid becoming loose duringtransit. Securing the box with string is not tobe recommended since string can easilybecome chafed or cut. Metal handles may bescrewed on to the box. The box should beclearly labelled ‘Fragile’ (using the interna-tional glass symbol) and with its origin anddestination. Loans should always be accompa-nied by a member of the staff from the lendinginstitution, preferably by whoever has checkedthe condition of the objects, packed them andwill place them on exhibition at their destina-tion. This will ensure that the glass does notundergo rough handling.

Glass requiring specializedconservation treatment

Objects made of Egyptian faience and ofenamel will require different treatments tothose commonly applied to glass conservationand restoration. In addition, there are severaltypes of glass objects, such as those which areunstable (weeping or crizzling), whichthemselves will require specialized treatment;and yet others which will involve experts in aparticular field of conservation or otherconservation disciplines. This last group willinclude glass found in association with othermaterials, such as metals, textiles and ethno-graphical materials, chandeliers, mirrors, andpaintings on or composed of glass.

Deterioration and conservation ofEgyptian faience

The friability of Egyptian faience (Figure 7.53)can be attributed to its manufacture (Smith,1996). It is subject to the leaching out of itsalkali content, in the same way as true glass,to crumbling of the friable core material, andto damage caused by the action of solublesalts. Leaching of the alkali content may leadto an overall whitening of the glaze, anddeterioration of the metallic salts may result ina change of glaze colour. The strength of thequartz core depends upon the amount ofinterstitial glass present. Where interstitial glass

content is low, the body is brittle. Only inHarappan faience (2300–1800 BC) and later‘glassy’ faience, where frit was intentionallyadded to the quartz in order to produce ahomogenous glassy phase throughout theobject, is the body strong enough to withstandabrasion once the surface glaze has been lost.

Soluble salts can be present in Egyptianfaience either as a result of the manufacturingprocess, or as a result of burial in a salt-ladenenvironment. Their action can result in thedisruption and loss of the glaze surface, or thetotal disintegration of an object. If the outerglaze layer rather than any interstitial glass isholding the majority of the quartz grainsforming the body in place, the introduction ofwater to remove soluble salts may cause thedisintegration of the core and total collapse ofthe object. Where soluble salts are crystalliz-ing from an intact glaze, it may be possible todesalinate without causing loss of materialfrom the core. The danger is that as the coreof the faience takes up water during saltremoval, pressure can build up within theobject, which may cause cracking or totaldisintegration. If salts cannot be removed,Egyptian faience objects should be stored withsilica gel (Figure 7.54) or kept in a constantrelative humidity.

If the surface permits, any necessary conser-vation work is limited to dry cleaning methodsand repair with a relatively weak and readilyre-dissoluble adhesive such as Paraloid B-72(methyl methacrylate copolymer) or HMG(cellulose nitrate). Repair may have to bepreceded by local consolidation of thecrumbling edges; it should be remembered


Figure 7.53 An Egyptian faience ushabti figure,broken as the result of having an unstable mount.

that the consolidation can trap soluble salts inthe faience body and that the solvents canmobilize them, resulting in internal disruptionand damage to the objects. This risk can beminimized by using dried distilled solvents,which contain less water than standard labora-tory solvents. The consolidation will result insome colour change and will be irreversible.Thus a system with proven long-term ageingproperties must be used, such as Paraloid B-72, B-67 and B-99. All have good adhesiveproperties, but the system chosen will have tobe decided by local climatic conditions, sincetheir glass transition temperatures vary.Dismantling adhesive repairs also poses a riskof damage. If this should become necessary,the safest option may be to soften the joinsby placing the object in a solvent vapourduring the day when a conservator can bepresent to dismantle the joins as soon as theadhesive softens. Should gap-filling berequired, Smith (1996) suggests using a stiffpaste composed of glass micro-balloons in a30 per cent solution of Paraloid B-72 in 50:50acetone/IMS applied directly onto the consol-idated core. The filling material is strong butlightweight, and on drying forms a texturedsurface compatible with the faience. (If neces-sary, the filling can be easily removed withacetone.) Fine-ground dry pigments in anacrylic gloss medium can be used for colour-ing the restored areas. It is probably best not

to attempt the conservation of faience unless,by doing nothing, the object will be lost.

Chemically unstable (weeping orcrizzling glass)

Some of the past attempts to preserve chemi-cally unstable glass have severely interferedwith the glass itself. An investigation by Bohm(1998) into the treatment of crizzled glass inSwedish collections revealed a number ofearly attempts to stabilize the glass. HeribertSeitz, an art historian, conducted conservationexperiments when curator of the NordiskaMuseet (National Museum of Cultural History),Stockholm. The work, dating from 1935, 1938and 1942, was documented on the objectrecord cards. Protective coatings of Zaponlacquer (cellulose nitrate) thinned withacetone or Zapon lacquer thinner wereapplied to some glasses. Others were coatedwith chlorinated rubber thinned with xylene,which discoloured to a dark brown colourwithin a few years. It was removed andreplaced with Zapon lacquer. Seitz recom-mended storage of the glass at 40 per cent RHusing air-tight cases containing silica gel, or asaturated solution of calcium chloride. Theelimination of atmospheric carbon dioxideusing sodic calcium (soda lime) was alsoconsidered.

Hedvall and Olson developed a treatmentfor weeping glass by impregnating it withpolymethyl methacrylate under vacuum. Therecords of this treatment are scanty. It seemsto have been in use in the 1960s, and treatedglass is in several Swedish collections. Theglass was degreased in carbon tetrachlorideand leached alkaline products were removedwith 5 per cent nitric acid, followed bythorough rinsing in distilled water and airdrying. The glass was then placed in a vacuumchamber and heated to 60–70°C (apparently tolower the vapour pressure in the capillaries ofthe glass and to improve impregnation with 50per cent polymethyl methacrylate resin).Excess resin was allowed to drain out of theglass before it was removed from the chamber.In 1956, Bostrom invented an apparatus topreserve a pair of seventeenth-century, deteri-orated Venetian glass goblets in Perspex cylin-ders from which the air was evacuated andreplaced with argon gas. The glass was


Figure 7.54 A repaired Egyptian faience scarab, whichhad split due to the action of atmospheric moisture onthe hygroscopic interior, shown after repair and storagewith silica gel in a (lidded) box. (© Copyright TheBritish Museum, London).

removed from the cylinders in the 1980s, aftersome 25 years, for aesthetic reasons.

The application of surface coatings asprotection for unstable glass has been unsuc-cessful as they were prone to failure anddiscoloration and were not entirely imperviousto moisture. Water thus penetrated the lacqueror entering the glass where the coating wasnot continuous, producing a corrosive high pHconcentration, and allowing the leachingprocess to continue with inevitable results.The required level of RH may be achieved bythe inclusion, in a well-fitting display case, oftrays of silica gel which have been previouslyequilibrated to the target RH (Brill, 1975, 1978;Thomson, 1998). However, the silica gel hasto be changed at regular intervals (i.e. whenit has turned from blue to pink due to theabsorption of water), and unless it has beenequilibrated to the required level prior to use,it may over-dry the atmosphere and causedehydration of the glass.

More recently, however, experiments tostabilize glass by introducing ions into thenetwork have been undertaken. (This workshould not be confused with the applicationof calcium ions to deteriorated glass from amarine environment: Corvaia et al., 1996.)

Between 1987 and 1990 an in-house surveywas made of glass in the collections of theVictoria and Albert Museum, London.Following this survey, research began withImperial College, London to investigate vesselglass deterioration in a museum environment,and to try to gain a greater understanding ofthe deterioration process. By using sophisti-cated analytical techniques to investigate themechanisms by which glass decays, theresearch has given an insight into why someglasses are more susceptible than others. In1992, Ryan investigated the mechanisms ofactive corrosion (Ryan, 1995; Ryan et al., 1996)and in 1995, Hogg began experimenting withactive consolidation techniques. Hogg et al.(1998, 1999) treated susceptible glasses withmono-functional organo-silanes in order toinhibit the deterioration process and suggestthat, by this process, the lifetime of the glassmay be extended by a factor of ten. Theprotection is probably mostly due to the actionof the silane bonding to the glass surface. Thereplacement of Si–OH groups with Si–O–Sibonds removes the possibility of hydrogen

bonding between atmospheric moisture andthe glass surface. There is the ethical questionof reversibility to consider. Since organo-silanes are only monofunctional and do notcross-link, they exist only as a molecularmono-layer on the surface. (This does notimply reversibility, but then neither is it possi-ble to remove an entire coating of consolidantor lacquer from a fragile glass surface.) Theorganic groups on the silane are known to bestable, and breakdown of the silane will notproduce corrosive by-products. If the coatingwere to degrade, being only bound to one siteon the surface, it would not cause damageover a larger area, as is often the case withlacquers and other surface coatings. Theresults of the research have enabled conser-vation scientists to give recommendations forthe care and display of unstable glass. Thework has also extended to the developmentof new methods of active surface treatment toarrest deterioration.

Unstable glass (Egyptian faience andenamels) is best stored in a stable relativehumidity of 35–40 per cent, with 42 per cent(the point at which potassium carbonatedeliquesces) being the upper limit, to preventalkali being leached by moisture. Periodiccleaning to remove the build-up of alkali onthe surface of weeping glass may be advisable.The frequency of this treatment needs to bebalanced against the risk of causing furtherdamage and disruption of the surface. De-ionized or distilled water should be used forwashing and rinsing the glass, and glovesshould be used when handling and washingthe glass, to avoid the transfer of salts and acidpresent on the skin. Such a cleaning processmay be complicated by the fact that it may notbe advisable to immerse objects, and that theuse of cotton wool swabs must be avoided ifa glass surface is cracked, as cotton filamentswould snag and detach fragments. It may bepossible to use a fine spray of water adminis-tered from a hand-held spray held at somedistance from the object. A few drops of anon-ionic detergent may be added to thewater to aid the removal of grease and dirt.

When washing collections of glass, thewashing water should be changed frequently.Since the glass surface is slightly acidic, a rinsein de-ionized water should help to neutralizethe surface in addition to removing traces of


detergent. The objects should be thoroughlydried either by gently blotting with acid-freepaper or by placing them in a flow of coolair. No form of heat drying should be used,as this would cause rapid crizzling to occur,sometimes instantaneously. Alcohols are effec-tive at removing grease and dirt, and it hasbeen suggested that their use might be advan-tageous (Newton and Davison, 1989), in thatalcohols displace surface water held in cracksand pores. However, the application ofalcohol to the surface of a hydrated silica gelmay cause water held in cracks and pores tobe replaced to such an extent that the glass isdehydrated.

Washing will help to retard the deteriorationof weeping glass for a short time and will alsoimprove its appearance by removing dirtadhering to the slippery surface; it does not,however, offer any long-term protectionagainst recurrence of weeping.

Severely deteriorated (crizzled) glass objectscannot be washed without endangering them,and will require to be stored in carefullycontrolled environmental conditions. However,if the humidity is too low, the hydrated glasssurface may dehydrate: removal of watermolecules will generate tensile stresses in thesurface layer, which are then relieved by theformation and propagation of surface cracks.The temperature and relative humidity ofcontrolled storage and display cases should beregularly monitored by the installation of athermometer and a hygrometer, or a record-ing thermo-hygrograph.

The provision of environmental control andsuitable lighting levels in both storage andexhibition conditions is the safest method ofslowing down an inevitable worsening of thecondition (see Figure 7.50). This is usuallyachieved by means of establishing micro-climates in cases rather than by controlling anentire gallery (Weintraub, 1998). In general,and also specifically for unstable glass, theenvironmental parameters that need to becontrolled include temperature, relativehumidity and gaseous – and particulate pollu-tants. Within a showcase there are a variety ofpassive and active methods which can be usedto effect control. Passive conservation in theform of high-grade, air-tight case design is theleast expensive option in the long term. Inaddition to being well-sealed, any materials

used within the interior display areas, such astextiles, wood or paint, must not emit harmfulgaseous pollutants. Provisions must beincluded for the use of passive environmentalcontrol materials such as silica gel and pollu-tion scavengers. Specialist advice from aconservation scientist and companies special-izing in museum case design will be required.Active systems for environmental control aregenerally expensive and time-consuming tomaintain, especially if applied to a largenumber of cases. Here again, a conservationscientist will normally be required to giveadvice on the type and amount of silica gel touse, its method of installation and frequencyof its renewal. In reality, temperature controlis difficult to achieve without specialist equip-ment and case design. The use of temperaturecontrol equipment raises concerns regardingthe safety of collections within a case.However, the heat generated within a case bylighting can be addressed by the use of fibreoptic transmitters and fluorescent ballasts, andby isolating lighting systems by the use ofdiffusers. A passive relative humidity controlsystem in storage or display cases involves theuse of conditioned silica gel. This serves as abuffer to offset changes in relative humiditycaused by air infiltration or through tempera-ture variations.

Artsorb silica gel equilibrated to 38 per centRH has been used to condition display casesin the glass gallery of the Victoria and AlbertMuseum, London since 1993. Its use hasproved to be a relatively simple and inexpen-sive method for protecting a large number ofsusceptible glass objects. The large-scalerenovation of the exhibitions at the CorningMuseum of Glass has enabled the design forthe galleries and cases to address the specialrequirements of crizzled glass objects. Methodsof stabilizing glasses in various stages ofcrizzling and of displaying them safely (struc-tural supports, suitable lighting and humiditycontrol) have been investigated (Page, 1998).Similar investigations have been carried out inconnection with crizzled glass vessels of theChinese Qing Dynasty (1644–1911) in theFreer Gallery of Art and Arthur M. SacklerGallery, Smithsonian Institution (Koob, 2000).The glasses are extremely hygroscopic, havinga high potassium content and (presumed) lackof sodium. Both galleries now have environ-


mental control throughout the buildings,maintained at 21°C (70°F) and 50 per cent RH.

Further analysis will identify the exactcompositions of the glasses. If the potassiumis found to be the most reactive component,a lower humidity (40–42 per cent) would berecommended for storage, to which end itmight be necessary to provide a custom-builtcase with a higher air exchange to ensuregreater stability. Koob (2000) describes theconservation treatment of two of the glassesfrom these galleries, which had previouslybeen repaired. The subsequent re-repair wascarried out using Paraloid B-72 adhesive.Schack von Wittenau (1998) describes acollaboration between the Kunstsammlungender Vest Coburg with the Institute for MaterialResearch of Glass and Ceramics of theUniversity of Erlangen-Nuremberg, to researchthe problems of crizzled glass, and later withthe Fraunhofer Institute for Silicate Researchand to formulate a design for environmentallycontrolled display cases for the GlassCollections of the Veste Coburg in Germany.

There are 22 controlled cases, seven ofwhich are wall cases with fabric-covered woodbacking; the others are free-standing. The glassand steel cases are constructed using neutralsealing compounds and appropriate adhesives.The base of each free-standing case is formedof three high-grade stainless steel cupboards,which contain plastic containers filled with asaturated solution of magnesium chloride. Thecupboards can be opened from the exterior,which allows for maintenance, as every threemonths the salt levels have to be checked andreplenished, and the surplus salt solution thatresults from the climate control process,removed. A lattice is fitted across the top ofthe base and this allows an exchange of gaseswithin the cases. Seven layers of silica gel inthe form of absorption strips about 4.0 cmwide are placed in a criss-cross pattern on thelattice, about 4.0 cm apart from one another.These are covered by a perforated panelcovered with textile. The absorption layer isconditioned by a saturated solution of magne-sium chloride to a relative humidity of 37–42per cent. Whilst guaranteeing the desiredrelative humidity, the saturated salt solutionreacts very slowly to fluctuations in humidity.This is balanced by the silica gel, which as afast buffer, cushions the effects of variations in

humidity. Ventilation holes in the top of thelight box prevent the build-up of heat. Glassdust shields installed between the light fittingsand display area are fitted with a PVP ultravi-olet filter. Textiles and wooden backings,which do not emit pollutants, were chosen forlining the cases. The condition of the glasseson display will continue to be monitored andrecorded, as will the condition of modelglasses (glass sensors) composed of sensitivepotassium lime silicate glasses, which wereplaced in four of the cases.

Glass in association with other materials

MetalsMetal is usually found in association with glassin the form of supports:

(i) a wire armature around which glass isshaped when hot to produce figures suchas those from Nevers;

(ii) wires used to outline a design made inenamel (cloisonné) (see Figure 3.33) orcloisonné glass (see Figure 7.68);

(iii) a flat support to which a vitreous mater-ial, glass or enamel, is fused (see Figures3.3, 7.57 and 7.58);

(iv) a support to which glass, glass beads orfragments are adhered as in the case ofglass bead pictures and micromosaics;

(v) a structural support for glazing andchandeliers (see Figure 7.61c).

The metal oxides of tin and silver have beenused in the form of amalgams with mercury,and lead oxide, to form mirrored surfaces onglass. In addition to the glass itself requiringconservation, problems of deterioration maybe associated with a metal support, mirroredsurface or materials such as adhesives andresins, which have been used to secure theglass.

Archaeological glass may be coated withcorrosion products from metal objects buriedadjacent to them (Figure 4.9), or into whichthe glass has been set (Plate 3) (Dove, 1981;see also above Archaeological glass). Glasscan also be associated with metals, in the formof previous restorations, armatures, supportsand stands (Figure 7.55), mounts and lids anddecoration such as electro-plating, and of


course in the form of enamel (see Figures 7.57and 7.58). Metal mounts etc. are often securedto the glass with gesso, with or without theaddition of threaded metal rods, nuts andbolts. If they have to be removed for anyreason, great care should be taken not to exertpressure on the glass or to scratch either theglass or the metal with tools such as pliers.Metal fixings may have corroded and willrequire careful use of chemicals such as rustremovers, and thin penetrating oil to releasethem. If necessary, a metals conservatorshould be consulted before work begins.

Small glass figurines such as flowers invases, were used as table decorations, chande-lier trimmings, table altars, hat-pins, cigaretteholders and as scientific models (mostlyflowers and insects such as beetles).

Glass figures, constructed over wire arma-tures, may become broken and misshapen.Conservation problems may arise if metalshave transferred corrosion products to theglass, or worse, if a metal armature has beenexposed to damp, corroded and expanded,causing the glass to split. It will not normallybe possible to remove the armature since theglass is fused around it. Where exposed, themetal can be cleaned and stabilized, the glassfragments replaced, and the object kept in aconstant temperature and humidity.

The Ware Collection in the BotanicalMuseum of Harvard University in Boston,Massachusetts, comprises almost 3,000 botani-cally accurate plant models, made primarily of

glass. They represent more than 850 speciesof flowers, fruits and leaves. The models arethe sole production of the Dresden glass-workers Leopold Blascka (1822–1895) and hisson Rudolph (1857–1939). Leopold Blasckahad begun his career making glass eyes andfloral jewellery. He then became famous forthe production of models of invertebratemarine animals, such as jellyfish, primarily formuseum collections (Figure 7.55). The modelsare in fact of mixed media. Many are formedover a wire armature. Sometimes the glasselements are fused to one another, and atother times they are glued together. The earlymodels were made of clear glass, painted withhide glue or isinglas (fish glue obtained fromsturgeon). Striving for greater subtleties,Rudolph began to produce coloured glassformulated to melt at differing temperatures.Thus coloured glasses could be ground to apowder and then dusted onto a base glass andfused by firing, without distorting the base.The most remarkable effects can be seen onthe ‘diseased fruit’ series of models, completedin 1929. However, these now exhibit a corro-sion ‘bloom’ which suggests that the glassformulae were not stable. It is hoped thatmethods of conservation can be devised toensure the preservation of these uniquemodels (Pantano, Rossi-Wilcox and Lange,1998). Apart from the difficulties presented bythe mixed media, conservation problems resultfrom the glass being very thin, and from thelack of access to the interior of small hollowobjects such as the sea creatures. Methodsused by the author to support fragments arethe application of the dental cement; StickyWax melted across joins in the glass; or abacking made of fine glass-fibre tissue impreg-nated with HMG (cellulose nitrate adhesive ofParaloid B-72 (methyl methacrylate copoly-mer)).

The Czech sculptor Jaroslav Brychta(1895–1971) took up glassmaking in ŽelenyBrod, where ancient glassmaking technologywas being used to produce glass beads andbuttons. This formed the basis for Brychta’scharacteristic glass figurines, fanciful creaturesmade of beads and glass rods, the first ofwhich were exhibited at Želeny Brod in 1921.Brychta designed the figures, produceddetailed drawings of them, and researchedglass technology, but the figures were actually


Figure 7.55 Cephalopod made by Blaschka, of glassand mixed media. Nineteenth century. Germany.(Zoology Department, The Manchester Museum).

made by glassmakers. Further techniques weredeveloped, and in 1927 figures of blown glassbegan to be made.

Thousands of small figurines were produced(Volf, 1977) (Figure 7.56). The figurines wereextremely fragile. Brychta recommended thatthey should be given a protective coat ofcellulose nitrate lacquer. It is common to findfigurines covered with a thick layer of lacquer,which has discoloured to yellow or darkbrown. Some of the figurines were formed by applying the molten glass over a wirearmature (probably an alloy of iron,chromium, aluminium and cobalt), which hadthe same coefficient of expansion as the glass.However, the wire is very fragile and is easilybroken and its shape distorted. Great care hasto be taken when distorted wires are straight-ened and in relaying it in its original positionagainst the broken glass. Broken wire can beconnected with tin solder. Replacing wires isextremely difficult and usually requires anindividual approach to each figurine. Thin

brass wire laid onto the glass with clearpolyester resin has been used successfully.Repairs are carried out with clear epoxy resinhidden on the reverse side of the figurine.Brass or steel wires were used in the originalconstructions for mounting beads, i.e. notembedded in the glass. These are not sofragile and can easily be replaced if necessary.If it becomes necessary to remove the lacquercoating, it can be dissolved in acetone, or, ifa thick layer has cracked, by soaking thefigurine in water for 24 hours. Water seepsbetween the lacquer and the glass, after whichthe lacquer is easily peeled away. Some smallmissing glass parts can be made of clear epoxyresin, others such as horse bridles have to bereproduced in glass and adhered to the origi-nal with epoxy resin, cyanoacrylate being usedto hold pieces in position temporarily. A smallgroup of the figures is displayed in themuseum in Zeleny Brod (Czech Republic).

Glass beads as decorative elements ontextiles and ethnographic materialsGlass beads are often found as decoration ona wide variety of objects from most periodsand cultures (Plate 6). Their use on textilesand ethnographic material is particularlyfrequent. They are often applied to costumeand may actually form the major part of it.Baskets, bags, boots, shoes and bead-dressesare just some of the other items which can bedecorated with beads. They may even befound on decorative or ritual statuary.

Beads have been made from a very broadrange of materials, and it is not uncommon tofind a mixture of different types of bead onone item. Whilst some of these are easilydistinguished from glass, others are not soobvious. Fake pearls could be made frompearlized glass or cellulose acetate and fishscales. Other materials such as metals, plastics,stone (for example obsidian, a naturally occur-ring form of glass), bone ivory, amber, shellor Egyptian faience may also be mistaken forglass. As beads are small and easily trans-ported, they may be found far from their placeof manufacture. Trade beads are a well-knowncategory; made in Europe but exported toAmerica and Africa where they were madeinto distinctive native artefacts. Glass may alsobe recycled, at much lower temperatures thanthat required to melt glass. For example, beads


Figure 7.56 Small glass lamp-worked figure in theform of a rabbi, in the style of Jaroslav Brychta.Twentieth century. Czech Republic.

made from melted bottle glass are known fromInuit and African cultures: the Krobo peopleof West Africa grind glass into a powder whichis poured into a mould and fired in a simplefurnace which fuses the granules together.

When approaching the conservation of glassbeads, it is important to consider the othermaterials present. Beads may be attachedindividually to a substrate such as fabric orleather, strung as a necklace or fringe, or bewoven into panels using various types ofthread. The thread may be made of animal,plant or synthetic fibre, but sinew, leather andwire are also used. Often a mixture of threadsis found, especially where an object hasundergone repair from the time it was still inuse, or previous conservation treatment.Sometimes beads are embedded in resin orwax applied to a surface such as wood.

There may be health and safety implicationsto be considered when dealing with glassbeads. Pearlized finishes on glass beads maycontain lead, some trade beads from the mid-nineteenth century contain arsenic. Thesesubstances will be released as the glassdegrades. In Namibia bead collars have mudand blood incorporated, used to bind thethreads attaching the beads, resulting in apossible biohazard. If the beads are attachedto an organic substrate then there may beresidual pesticides from a previous fumigation.

In many cases the deterioration of beadworkis more dependent on the condition of thethreading than of the beads themselves. Glassbeads suffering from glass disease (Fenn,1987) are an obvious exception, which is aphenomenon described elsewhere. Beadsmade of materials other than glass, particularlyplastics, are subject to their own decay mecha-nisms, some of which may evolve substancesdamaging to glass and other materials. Openstorage of sensitive objects is recommended toallow any noxious gases formed during degra-dation processes to dissipate, rather then buildup to dangerous levels.

Many problems arise from the makers’ useof incompatible or poor quality materials.Beads with rough or sharp edges may damagethread and cause bead loss. Thread which istoo strong for the beads may cut throughthem, whilst beads with glass disease maydamage their threads and support fabrics.Many beads are heavy and their weight alone

may cause damage to the support, thread orfabric, thus their handling and display must becarefully considered. Beaded dresses, sopopular in Europe in the 1920s, often had abase of silk chiffon, which has now becometoo fragile to support the weight of the beadsattached to it. Older African beadwork isusually strung on vegetable fibre, whichbecomes very brittle with age, particularly if ithas been dyed. For this reason water shouldnever be used for cleaning, as the fibre willbecome weaker and liable to break.

Threads, sinew and leather will act as wicksif liquid comes into contact with them, sodamage to the thread or interior of the bead,caused by damp conditions or previous clean-ing involving the use of aqueous solutions,may not be immediately determined. In thecase of beads attached to skin or leathersubstrates, the use of leather dressings shouldbe avoided as they can wick along the thread-ing, causing deterioration of the thread andthe interior of the beads.

When the original threading has broken, theintroduction of a new thread to support theweight of the beads is usually the best option,retaining the original thread as far as possible.Thread used in conservation work should be waxed to prevent the wicking actionmentioned above. The thread is treated withan inert paraffin wax rather than beeswax,which can provide food for insects. In caseswhere original thread fills the hole through thebeads, it is often not possible to re-thread the beads, and other methods of supportingthe weakened structure are needed. Thesemay involve the use of adhesive to secureindividual beads, or the stitching of thebeadwork to a support fabric.

A particular problem is encountered wheresinew has been used to thread beads whichhave deteriorated, e.g. as a result of glassdisease. It is very reactive to changes inhumidity; swelling in a high humidity causesfragile beads to fracture. However, humiditywhich is low enough to prevent further glassdeterioration may result in the sinew becom-ing brittle and liable to snap. Metal wires areprone to corrosion; corrosion products maycause beads to split. The combination of poorquality cotton thread and chemically unstableglass beads can be particularly difficult to treat.As the beads are often small and their holes


tiny, conservation options are very limited.Methods of immobilizing the beadwork, suchas securing them to a rigid support, will helpto prevent the flexing and abrasion responsi-ble for much of the deterioration of fragilebeads and weak thread.

If the support itself is under biologicaland/or insect attack and requires fumigation,the effect that the treatment will have on theglass beads must be taken into account.Berkouwer (1994) freeze dried nineteenth-century textiles decorated with glass beads.Inspection of the beads under high magnifi-cation showed that there appeared to be nochange in the condition of the glass. Wherebeads requiring conservation are encounteredin association with textiles and other organicmaterials, the advice of a textile conservatorshould be sought (Lougheed and Shaw, 1986;Hosforth and Davison, 1988; Sirois, 1999).

Enamels

Enamels suffer from physical damage due tohandling or to corrosion of the underlyingmetal whose corrosion products may force theenamel to fall away (Figure 7.57), or to thebreakdown of the vitreous enamel itself. Someenamel objects may exhibit the ‘weeping’phenomena seen in unstable glass. The degra-dation of Limoges enamels, produced between1480 and 1530, was investigated by Richter(1998). Severe deterioration of the enamels,especially the blue, mauve and violet enamels,has occurred since they were manufactured. Alarge selection of painted enamels wasstudied, using specific ion beam analyses forthe chemical characterization of the glass withthe accelerator AGLAE using PIXE (proton-induced X-ray emission) and PIGME (proton-induced gamma ray emission). The analyseswere carried out without sampling theenamels, using two X-ray detectors and onegamma ray detector. In addition, acceleratedageing tests were carried out on syntheticglasses, similar in composition to some antiqueenamels. The results were compared with thedegraded enamels.

If the damage is due to metal corrosion, andthe metal itself has become exposed in part,it may be cleaned and stabilized, before theenamel is readhered to its base. Where themetal is not exposed, the only course of action

will be to keep the enamel in a constant dryenvironment.

Early restoration of enamels was carried outusing coloured waxes, or by organic bindersin the form of natural resins such as shellacor animal glue and its derivatives. These werethickened by the edition of a filling material,such as whiting or gypsum, and coloured bya variety of pigments. Such restorations haveusually discoloured with age. They may alsohave shrunk and become brittle, even fallingaway, so that the enamel is left prone tofurther damage. Restorations can be under-taken by adapting the methods and materialsused for restoring porcelain. For example,opaque fillers such as epoxy resin thickenedwith an inert filler, or commercially availableepoxy putties were used to fill missing areasof enamel (Figures 7.58a,b). These, whencured, were abraded to form a smooth surface


Figure 7.57 A damaged enamelled candlestickshowing loss of the enamel surface and damage to thecopper substrate.

and then painted by hand with glaze andpigment. The risk of damaging the relativelysoft enamel, surrounding metalwork and setstones, when abrading small fills is great.Removal of old restorations may be under-taken by careful use of a scalpel. If solventsare necessary to soften fills these must be keptto an absolute minimum, and not be allowedto flow freely onto the metalwork. Great caremust be taken not to cause more damage tothe edges of the damaged enamel, particularlyby filaments of cotton wool from swabs catch-ing jagged edges. The modern approach to theconservation of enamels is outlined by Cronynet al. (1987).

Transparent enamels can be restored withclear epoxy resin containing a tiny amount ofdry pigment, which will not impair the trans-parency of the resin. Some colour will also bereflected from the surrounding enamel. Pliqueà jour enamel can be restored by providing themetal cells with a temporary backing of dentalwax lightly adhered to the metalwork, beforeintroduction of the clear coloured resin. Oncethe resin has fully cured, the dental wax isremoved. The resin which has been in contactwith the wax will have a dull surface. This iscleaned with white spirit, then water and leftto dry, after which a thin layer of epoxy resincan be painted over the surface and left tocure. Translucent enamels can be restored witha paste mixed to varying degrees of translu-cency by adding different amounts of drytitanium dioxide pigment and thickening it

with fumed silica. Depending upon the sizeand depth of the areas to be restored, the pasteis applied with a spatula, fine sable brush ora needle. In the latter case, the work may haveto be carried out beneath a microscope, inorder to apply the correct amount of filler, inorder that no further work will be requiredonce the resin has cured. Large, flat expansesof enamel are difficult to restore in terms ofattaining a perfect flat surface without beingable to abrade and polish it without the riskof damage to adjoining metalwork or otherdecoration such as set stones. Where it ispossible to imitate the original enamel withrestoration applied as layers of sprayedcoloured glaze, adjoining areas must bemasked off. Fisher (1991) describes the clean-ing and repair of cloisonné metalwork, andrestoration of the enamel with the epoxy resinHXTAL NYL-1, of a collection of seventeenth-century enamels. Some so-called enamel can infact turn out to be unfired lacquers or paints,or to have been restored with such materials,which will be damaged if solvents in conser-vation materials come into contact with them.

Micromosaics

Conservation problems associated with micro-mosaics are physical damage, and deteriora-tion of one or more of the components,particularly the disintegration of the adhesiveto which the glass tesserae are secured to thesupport (see Figure 2.17).


Figure 7.58 An enamelled clock dial: (a) damaged, with loss of enamel; (b) restored with coloured epoxy paste.

(a) (b)

Chandeliers

Conservators and restorers are becomingincreasingly involved in the work of preparingcondition reports on historic chandeliers andother fixtures and fittings in historic buildings.In some cases conservators are responsible forundertaking the work, in conjunction withspecialist firms where necessary, e.g. wherechandelier frames require rebuilding –especially of the metalwork – re-pinning, orthe manufacture of new glass components.There are commercial firms who specialize incleaning, repairing and restoring chandeliers.The larger companies are aware of conserva-tion issues and adopt a conservation approachto cleaning and refurbishment; others mayclean antique and modern chandeliers byidentical methods, using proprietary solutionsand abrasives, the contents of which arelargely unknown.

There are few published articles describingthe treatment of historic glass chandeliers(Davison, 1988; Reilly and Mortimer, 1998,Sommer-Larson, 1999). In the past, manychandeliers in private houses were hung andignored or ‘cared for’ by staff or handymen, norecords being kept of any work carried out. Thisseems to have comprised dismantling and clean-ing, replacing missing metal and glass sectionsand re-pinning the lustres which form pendantsand chains. Pieces were often incorrectlyreplaced, or replaced with unmatching glass,chains of lustres removed or incorrectly linkedtogether, and the pots of arms, or the holes inthe receiving bowl into which they fitted, fileddown when arms were wrongly placed.

Most historic chandeliers – some of whichmay be two hundred or so years old – willrequire conservation/restoration (Figure 7.59).The glass elements themselves might still be ingood condition. However, it is often the casethat some glass has been broken duringhandling and cleaning, by undue pressurehaving been exerted on arms and by carelessuse of ladders, during replacement of lightbulbs or decoration. Many chandeliers havechipped or missing lustres, broken candlenozzles, greasepans, arms, balusters and bowls.

During the life of a chandelier there mayhave been much rearrangement of its decora-tive elements, and even an exchange of partsbetween chandeliers of the same configura-

tion, (particularly if they were all dismantledfor cleaning at the same time). When glasselements were broken they may have beenrepaired with inappropriate materials, replacedwith new or contemporary pieces, which maynot be of the correct size, shape or cut. Inaddition, parts such as the dressings werechanged as fashion demanded. Pendants andchains may be entirely missing or alteredwithout regard for their correct configuration.The surface of metal parts will have oxidized,and may also have been affected by the useof chemical cleaning agents. It is also possiblethat water from cleaning or from burst waterpipes may have penetrated the chandelier andcaused the formation of corrosion productssuch as rust on iron elements, though thisdoes not usually cause safety problems unlessthe electrical wiring is affected. In extremecases part or all of a chandelier may havebeen damaged by heat (such as those in theBrighton Pavilion, UK) (Rogers, 1980), or mayhave fallen down, e.g. due to incorrecthanging or during a fire (such as those atHampton Court Palace and at Uppark, UK).

Drawn copper wire was used throughoutthe late eighteenth and most of the nineteenthcenturies to form linking pins between glasslustres. Over time, the links stretch, and manywill have been repeatedly unbent and re-bentduring dismantling for cleaning, actions whichoften cause them to break. Water and chemi-cals used to clean chandeliers can cause corro-sion and embrittlement of the wire. Fixings


Figure 7.59 Damaged glass items from a chandelier: aglass arm broken at its junction with the metal pot;broken glass lustres; broken brass pin, alongside a newpin, and small pliers and cutters used to shape it.

such as screws and nuts may have been lostand not replaced, and metal threads may havebecome stripped. Plaster of Paris or otherfixatives such as gutta percha, which secureglass to metal components, may have becomeweakened and fail.

Since the introduction of electric lighting inthe early nineteenth century, many periodchandeliers have been electrically wired andfitted with lamp housings to accommodatelight bulbs or imitation candles. The introduc-tion of wiring often causes damage to bothglass and metal parts; they were often glued,wired, soldered or twisted along the chande-lier arms. Occasionally holes were drilled onglass nozzles in order to pass the wire throughto the lamp housing. In order to hide thewiring and to be able to introduce it throughsmall spaces not originally intended for wire,small gauge wire was sometimes used, whichis below the current electrical code standards.Wires were sometimes forced into spaces toosmall to accommodate them, causing cracksand breaks in the glass and damage to thesleeve wiring. In general, wiring is often oldand worn, not having withstood repeatedmovement, adjustment, cleaning and handling.Exposed metal wires and wiring which is notcorrectly insulated can cause electrical shortsand subsequent fire. The safety requirementsfor the electrical wiring of lighting are verystringent; all wiring should by undertaken byan appropriately qualified electrician.

ExaminationPreliminary examination of the condition ofhanging chandeliers is normally undertakenfrom a metal stepladder. Furniture and anyother objects beneath or near the chandeliershould be removed to a safe distance. Thiswill allow the free movement of a stepladderaround the chandelier; and prevent damage inthe event of any glass falling down (not anuncommon occurrence in cases wherechandeliers have been neglected for manyyears). A small dictaphone or pad is requiredfor recording notes of the chandelier’s condi-tion. From time to time during examination,the stepladder should be moved around thechandelier, rather than rotating the chandelieritself. The method by which the chandelier issecured to the ceiling joists may not be knownat this point, and either the fixing or the

shackle, by which the chandelier issuspended, may become unscrewed withdisastrous results. In particular, rotating achandelier in an anticlockwise direction maycause it to unthread itself if the nut whichsecures the suspension loop to the centralsupport is missing. Particular care must betaken in cases where chandeliers have beenelectrified and not undergone regular andrecent safety checks. Sub-standard or sub-codewiring is common. Wear and use may haveloosened electrical connections, resulting inelectrical shorts causing shock to the conser-vator, or fire. Ideally, the electrical supplyshould have been turned off.

Chandeliers which have been dismantled forstorage or transport present differentproblems. There may be no record of theiroriginal configuration, or simply photographstaken as the chandeliers were hanging. Theseare usually confusing due to the amount ofglass present, the distance from which thephotographs were taken and/or their poorquality. Photographs showing chandelierswere often taken to show the entire roomrather than the chandeliers themselves. Theuse of a hand-held magnifying glass may aidtheir interpretation. The arrangement of theglass components may have to be hypothe-sized, especially if the chandelier is unique;measuring the total length of the stem piecesand comparing their total to the length of themetal support rod will determine whether allthe glass pieces are present. Age and stylecomparisons will help, although it must beremembered that broken pieces may havebeen replaced with incorrectly shaped compo-nents over time. Where there is doubt, a histo-rian of historic lighting should be consulted.The metal frame and all the brass linking pinsrequire to be checked for signs of weaknessand/or corrosion, and for the site of missingscrews, washers etc. to be noted; this mayrequire that the chandelier has to be disman-tled.

In addition to chandeliers that remain essen-tially intact, it is sometimes the case that all orpart of them falls to the ground. This mayhappen as a result of the failure of the fixingmechanism, failure of part of the metal frame,of old repairs and restorations, or as a resultof damage to the room or building which theydecorate, such as a fire. In such cases, a


conservator may be faced with trays of brokenglass belonging to one or more chandeliers. Apolicy decision must then be taken with theowner and insurers, as to how much of thehistoric glass can be repaired for re-use, or toserve as patterns from which new pieces maybe copied (the original remaining in anarchive); and how much is beyond economicor practical repair/restoration. Sorting thefragments alone is a time-consuming andtherefore costly operation. The first step is tosort the glass fragments by shape, thickness,colour, cut and design into the various compo-nents; especially difficult if there are, forexample, a number of canopies, which whilstbeing of a slightly different size to each other,will bear the same cut design. Once thefragments have been sorted into balusters,pendants, greasepans, bowls etc., the glassfrom the different components can be laid out.Sometimes, as is the case with greasepans, iswill be possible to say at this point whetherthere is sufficient glass to be able to effect asafe and aesthetical repair/restoration. In thecase of deeper three-dimensional componentssuch as bowls and balusters, fragments can betaped together, or supported on polystyreneforms.

If there seems to be sufficient glass to effecta repair, the fragments can be adhered withepoxy resin. These shapes are difficult torepair due to the cumulative weight of theglass fragments, and in the case of bowls, the

need to align the fragments correctly in curva-tures in two directions, i.e. circumference andfrom rim to base. The glass will usuallyrequire a support during the resin curing-time.It is not considered wise to re-hang bowls orcanopies if heavy glass pendants will besuspended from the rims. It must be borne inmind that people will walk beneath chande-liers, and that resin repairs such as those justdescribed could fail in time and cause injuryor damage furniture below the chandelier.Large broken pieces of glass, such as bowls,if repaired and re-used, must be inspectedannually to check the strength of the repairs.In the case of modern chandeliers it is prefer-able to replace the broken glass, especially ifheavy and/or load bearing. Replacing glasssections either with contemporary pieces orwith newly made sections can be expensivebut may be necessary for reasons of safety. Ifit is not possible or desirable to use periodreplacement parts, new ones can be made bytaking moulds from original examples, andglass cast or blown and then cut, and metalparts cast and painted or electrotyped. Smallglass elements can be reproduced morecheaply using clear polyester or epoxy resins(although these will discolour in time thediscoloration is unlikely to be visible from theground). It may be possible to re-use brokenglass arms, depending on the number andposition of the breaks in them. Arms brokenat the point at which they emerge from their


Figure 7.60 A broken glassarm from an épergne repairedby extending the metal potinto which it had fitted. Thepot extension replaces themissing end of the arm.Similar repairs are carried outon chandelier arms. (Courtesyof K. van LookerenCampagne).

metal pot can be repaired and the metal potextended to cover and support the repair(Figure 7.60). Mid-length breaks will need tobe repaired, and covered by a metal collar.Old lead and brass collars look unsightly, butwhite metal, e.g. silver, when polished is notimmediately visible, especially if the chande-lier is highly decorated. Glass elements shouldbe separated from one another by washersmade of chamois leather, or by metalwashers, which fit inside the glass.

DismantlingBefore beginning work, it must be ensured thatthe electricity supply has been turned off. Theconservator, accompanying the housekeeper orelectrician assigned to do this, should ensurethat there is no risk of the power beingrestored. It is advisable to attach a notice to thefuse board cupboard to the effect that thechandeliers are being worked on and mustremain isolated until further notice. The isola-tion must be confirmed by an electrical test at


Figure 7.61 Recording glasschandeliers: (a) section; (b)layout viewed from beneath;(c) dismantled centralsupport; (d) chains andpendants removed from achandelier and laid out forphotography.

(a)

each chandelier, i.e. the use of a voltmeter orother indicator to demonstrate the absence ofthe supply. Adequate space must be clearedbeneath and around each chandelier to allowfor movement of ladders and/or scaffolding.Soft padding such as decorators’ groundsheetsshould be placed beneath the chandelier tosoften the landing of any pieces which might

fall. A number of tables also padded will berequired nearby for laying out the componentsas they are dismantled. The components of achandelier take up a great deal of space whenthey are laid out. A chandelier can either bedismantled in situ, or lowered to a safe workingheight by means of a block and tackle systemattached by snap links to the suspension chain.


(b)

(c) (d)

Prior to and during dismantling and clean-ing, a photographic and written record ismade, by means of notes, sketches, black andwhite and colour photographs, so that there isa clear record of where each piece of achandelier belongs in relation to the others(Figure 7.61a,b). The amount of glass on achandelier makes photography difficult. It iseasier to photograph the arrangement ofchains and pendants by removing oneexample from the chandelier, and laying it outon a flat surface (Figure 7.61c). It is particu-larly important to note the way in which thebranches join the metal receiving plate and thearrangement of electrical wiring (Figures 7.62).Other items to note are damaged and repairedglass, missing glass and technological details.

Each chandelier is dismantled in sequence,beginning with the dressings and workingfrom outside inwards and from the bottomupwards (Figure 7.63a–e). During this processit is not uncommon for chains to break andfor lustres to fall as a result of weak linkingpins failing. Although the arrangement of thependants and chains will require to berecorded, it is not always necessary to notethe exact position of every one. Their arrange-ment should be regular around the chandelier,and the chandelier will have been dismantledand cleaned many times since its installationso that the components are unlikely to be intheir original positions. For safety reasons andease of identification, the pendants and chainsare removed in an orderly sequence, placed

in labelled Polythene bags and passed downfrom the working platform in plastic basketsor trays (Figure.7.63a,b).

Detachable glass nozzles, drip pans, opalglass or cardboard candle tubes (if present)and other ornaments such as spires onthreaded mounts should then be removed.The next stage involves the removal of thelight bulbs. In order not to put undue strainon the glass arms during this operation, theyare supported by placing a hand beneath thegreasepan to apply a slight upwards pressureto counteract the slight downward pressurerequired to release the light bulb, especiallyfrom a bayonet fixing (Figure 7.63c). It willusually be necessary to remove the lowersection of the chandelier next, i.e. up to andincluding the receiving bowl, which covers theelectrical connections joining the arms. Theelectrical supply already having been isolated,the wiring can be disconnected. If thisinvolves cutting the wires, the arms and/orcentral stem may have to be rewired beforethe chandelier is reassembled.

Once the electric wires have been discon-nected, the arms themselves can be removed.Glass arms, ending in square or round metalpots, may simply slot into the metal receivingplate; some round pots were locked inposition by a locating pin which fitted into aslot in the plate, or secured by a threaded nutbelow the receiving plate. Gentle upwardpressure applied from beneath the receivingplate and controlled sideways pressure isapplied to each arm in turn to ascertainwhether the arm is loose. If the arm is loose,it can be held firmly at both ends and drawnclear of the plate, taking care not to jerk thearm or to bring it into contact with any nearbyglass or metal components. If the arm is notloose, it may be necessary to apply thin lubri-cating oil around the pot, or even in somecases a small amount of rust remover. Theglass branches are removed one at a time, ifpossible working from alternate sides so thatthe structure remains hanging evenly, andhanded down from the working platform sincethere would be a danger of them tipping outof a tray (Figure 7.63d).

Once all the arms are removed, the metalsafety hawser (if present) can be disconnectedfrom below the receiving plate so that thereceiving plate and glass and metal sections


Figure 7.62 A receiving plate for chandelier arms. Inthis example the holes to receive the metal pots intowhich the glass arms are fitted, are square.


Figure 7.63 Stages in dismantling a chandelier: (a)removing the dressings (decorative elements); (b)handing pieces down in baskets; (c) supporting the armwhilst removing a light bulb; (d) handing large piecesdown to the ground; (e) a fully dismantled chandelier.

(a)

(b)

(e)

(c)

(d)

from the upper section of the central supportstem can be removed individually. This opera-tion will require two pairs of hands, one toremove the glass a section at a time, and theother to hold the remaining unsecured glassin place. The glass sections are passed downfrom the working platform in plastic baskets.Lastly, the central support is unhooked fromthe chain suspending the chandelier from theceiling. If the chandelier is relatively small andmeasures no more than about 2 metres inheight, it may be considered safer to pass thewhole of the central support stem down to the floor in one piece by unhooking it fromthe suspension chain. Once down, the suspen-sion ring and fittings and/or the circular metalreceiving plate are removed, in order torelease the glass balusters and bowls andattendant silver metal fittings (tubes, washersand nuts) (Figure 7.63e). A sturdy metal or

wooden frame on a rolling base is useful fortemporary movement and storage of an assem-bled shaft of a chandelier (Figure 7.64).

If the chandelier is to be transported fromsite or stored for any length of time, it ispreferable to dismantle it fully and pack thecomponents in acid-free tissue andbubblepack in a wooden crate labelled withits fragile contents.

ConservationCorrect installation is the primary factor inpreventative conservation. The appropriateweight and type of chain and ceiling hookmust be used to suspend the chandelier. Thehook itself must be secured into a load-bearing ceiling joist in such a way thatcomplete rotation of the chandelier cannotunscrew the hook or disturb its connection tothe ceiling. This is best arranged by a quali-fied structural engineer, especially in the caseof historically important, and extremely heavychandeliers. The chain from which a chande-lier is suspended should be made of closed-link steel or brass. It should be of sufficientstrength to bear more than the full weight ofthe chandelier. It is likely that the chandelierwill have to be weighed; easily done byweighing individual or groups of componentson a bench balance and adding the totalweights. The chandelier itself should hanglevel and the weight of the stem pieces on thecentral support rod evenly distributed by theinsertion of washers and bushings betweenthem.

Thought must be given as to how thechandelier will be accessed for routine mainte-nance and cleaning; and how often this willbe done. In the case of small chandeliers hungat no great height, this might be managedfrom a set of stable metal steps or a lightscaffolding tower. In the mid-eighteenthcentury, chandeliers were often hung oncounterpoises from a fixed hook. As thechandelier was drawn down, by grasping thefinial, compensating counterpoises rose up thechain. Some of these mechanisms remain and,if used, should be checked periodically toensure that they are in working order. Forchandelier installations at great height,purpose-built winches with reduction gear andlocking capabilities are desirable. Handwinches are to be preferred to those operated


Figure 7.64 Moving a chandelier suspended from aframe on a trolley.

by push buttons, as the latter may jolt to astop, causing movement and damage to theglass. Glass chandeliers are most prone todamage from mishandling and from knocks byscaffolding or ladders used to enable cleaningor the changing of light bulbs. The risk ofdamage could be greatly reduced if eachchandelier were to be fitted with a rise andfall mechanism, which could incorporate alimiting device restricting the drop of thechandelier. The replacement of light bulbswould be made safer, quicker and easier,thereby reducing maintenance costs after theinitial cost of installation.

If the chandelier is electrified it should beexamined annually by a qualified electrician.A record of the date when each chandelier iscleaned, dismantled, checked for the safety ofthe electrical wiring and ceiling connectionsshould be kept by the conservator and by theowner.

The breakage of opal glass ‘candle-tubes’often occurs as they are compressed duringthe insertion of the electric light bulbs. Thelight fittings are not fixed in the cut glassgreasepans or candle nozzles, they thereforerotate whilst the bulbs are being inserted,requiring a certain amount of pressure to beexerted. It should be possible to design adevice to prevent this from occurring in thefuture.

The period of time between each cleaningwill need to be determined by conservatorsand the owners. In deciding the frequency ofcleaning, finance will have to be consideredand the fact that there is a potential danger inhandling and dismantling the chandeliers. Ifthe chandeliers decorate rooms that are inconstant use, general cleaning may be neces-sary at yearly intervals, and in-depth cleaninginvolving dismantling, perhaps every ten years.If, on the other hand, the chandeliers are inrooms that are closed to the public for part ofthe year, the amount of dust and dirt thatsettles on them is reduced if they are protectedby large muslin bags.

The glass is cleaned with a 50:50 solutionof industrial methylated spirits (IMS) (US:ethanol) and distilled or de-ionised water.Reilly and Mortimer (1998) suggest adding afew drops of ammonia to increase the effec-tiveness of the cleaning solution; whilstwarning against raising the pH above 9.

(Above pH 9 an aqueous alkaline solutionpotentially dissolves glass, but not in the shorttime taken to clean a chandelier.) The additionof a few drops of a non-ionic detergent is tobe preferred. Care should be taken not toallow the cleaning solution to penetrate thehollow glass branches, or the metal lightfittings. The glass is then wiped over withdistilled water on a soft cloth and then gentlypolished with a lint-free, dry cloth to removesmears and finger-prints. Wax deposits canusually be easily dislodged using a chisel-ended bamboo stick, the residue beingremoved with warm water or white spirit(aliphatic hydrocarbon solvent).

Cleaning using chemicals, if absolutelynecessary, should be restricted to the substan-tial metal parts, and then only to those fromwhich the chemical residues can be safely andconveniently removed. If not removedthoroughly, chemical residues or by-productswill remain and eventually attack the glass,metal and metal link pins. The composition ofmaterials used for cleaning should be known,plus the short- and long-term effects on thesubstrates (and upon the persons undertakingthe work). Light deposits of rust (ferric oxideFe2O3) may be removed mechanically using aglass-fibre bristle brush or, if necessary, chemi-cally with a commercial rust remover based onphosphoric acid, which reacts with the ferricoxide to form a protective coating of blackiron phosphate, a positive corrosion inhibitor.The use of such a rust remover would alsoincur the removal of original paint on the ironpipes. Thus it may be necessary to analyse asample of the paint if any repainting is antic-ipated.

Light silver tarnish may be removed byimmersing the silver in a solution of Goddard’sSilver Dip (UK) (active ingredient ammoniumthiocyanate). A soft bristle brush is used toclean the silver, alternatively the solution maybe applied locally on cotton wool swabs. Thebath of solution should be renewedfrequently. After cleaning, the silver is rinsedin changes of distilled water. Severe tarnishmay necessitate the use of a mild abrasivesuch as Goddard’s Silver Foam, gammaaluminium in alcohol, or precipitated chalk ina distilled water/alcohol mixture. Coarserabrasives would damage the metal surface.Traces of the cleaning agent are removed with


copious amounts of warm distilled water andthe silver left to air-dry before being polishedwith a soft cloth impregnated with an anti-tarnishing agent. Lacquering the silver is notrecommended since it will easily be scratchedoff moving parts and thus the silver willtarnish in streaks; often much of the metal iscovered by glass and therefore tarnishes veryslowly. Lacquer would be time-consuming toremove and renew.

Broken metal parts, which are not structural,can be repaired using a two-part epoxy resin.However, for obvious safety reasons, structuralrepairs must be carried out by a metal worker,using soldering and brazing techniques. Weak,brittle wire pins should be replaced with newpurpose-made chandelier pins. Made of softbrass, they are easily bent to shape, whilsthaving sufficient strength to support glasslustres and pendants. Care must be takenwhen cutting away old pins and wiring, asmetal wire cutters can easily damage the glass.Reilly and Mortimer (1998) suggest immersingglass dressings having an enormous number ofpins in a bathe of dilute nitric acid to dissolvethe metal. The acid concentration and lengthof immersion are not specified, but this courseof treatment is not recommended.

During the life of a chandelier, drops,pendants and chains (dressings) will almostcertainly have become broken or lost. Theymay have been replaced at random with anyglass lustres to hand irrespective of whetheror not they are of the correct size or shape,and dressings incorrectly hooked up to fill agap. Thus it is important to remember whendismantling the chandelier for cleaning that ifpendants and chains are rearranged correctly,the final appearance may be worsened unlessnew glass lustres can be purchased.

Where individual lustres, pendants andchains have been misplaced, they can berearranged where possible. However, thereason for misplacement may be that lustreshave broken in such a way that there are nolonger holes through which to secure thelinking pins. New lustres will have to bepurchased. If, however, a full restoration is inorder, the dressings can be fully dismantledand patterns made up as guides for each setof pendants or chains required. Chains andswags are almost always graded by size, withthe larger drops at the bottom (or centre). If

the lustres are pear-shaped, the direction ofthe lustres changes on either side of thecentral lustre, which should be oval. Thelinking pins should be aligned to allow the lustres to hang freely and in plane withthe dressing. Correct assembly of dressings forneoclassical chandeliers requires that theyshould be permanently fixed at their righthand end to a four-way (i.e. a glass lustrepierced with four holes), while the left handend should be provided with a slightly largereye for attachment to the neighbouring four-way. The vertical chains suspended fromswags should remain permanently at the topand bottom of the four-way. These units (aswag, a four-way, a top vertical chain, and abottom vertical chain) can be removed andreplaced as a unit. The vertical drops gradefrom small to large as they fall. Drops aregenerally hung with their flat, cut or pressedside to the outside in order to catch and reflectthe available light (Reilly and Mortimer, 1998).

Dismantling of old repairs, cleaning, repairand restoration of the glass elements, can beundertaken using the methods describedelsewhere in this chapter. On completion of the work, a conservation report is pro-duced which should include instructions onassembly/dismantling, notes of any modifica-tions which may have been made, and safetycertificates for suspension hooks, chains, ceil-ing fixings and the electrical wiring. Theconservation of an iron and glass chandelieris described by Sommer-Larson (1999).

Mirrors

Deterioration of the glass itself is discussed inChapter 4. In addition, deterioration processeswill occur in any decoration such as verreeglomisé or reverse paintings, within themirrored surface itself. There may also beproblems with the frame (weakened by woodworm or dry rot; the latter also causes black-ening of the mirroring); the support (linseedputty drying out) or backboard (which mightbe flimsy); the backings (such as woolblanketing); and the fixings – pins often withdecorative glass heads, either pinned into theframe or secured in place with animal glueand or plaster of Paris. This method of attach-ment was often used to secure decorativeitems used to hide joins in the glass plates.


The first signs of corrosion in a tin–amalgammirror appears as small dark patches whichgive the mirror a dark and cloudy appearance(Figure 7.65). Massive corrosion shows as adark grey layer or as concentric colouredbands varying from dark grey to yellow-brownand white. As the amalgam deteriorates, thecrystal phase changes, in that the crystalsenlarge and cover a larger proportion of theglass surface, whilst the mercury slowly evapo-rates. In time, tiny voids appear between theglass substrate and the amalgam. The fluidphase also migrates to the bottom of themirror. Corrosion of the amalgam results in theformation of tin oxide (cassiterite) and tinmonoxide (romarchite) and releases liquidmercury from the solid phase. Deterioratingamalgam mirrors contribute a few microgramsof mercury per cubic metre of air in the roomin which they are situated, though this level isnormally far below the toxic limit of 50 μg.The frames, however, often trap drops ofmercury and special precautions are necessarywhen handling them.

Conservators may be asked to deal with thepreservation of mirrors in several differentcircumstances. Mirrors may be encountered inthe form of architectural elements, in glazeddoors, recessed alcoves, window bays, wallcoverings and even as floor coverings, e.g. theMirror Room in the Rosenborg Palace inCopenhagen, furnished for King Frederik IVcirca 1700 and inspired by Louis XIV’s palaceat Versailles. More commonly, mirrors areframed and fixed to walls, e.g. the pier glasses

at Hampton Court Palace (Jackson, 1984, 1987;Quinton, 1998); smaller framed mirrors arehung on nails or other metal fixings. Mirrorsare also found as decorative panels on furni-ture such as side tables and drawer fronts. Inthe nineteenth century, mirrors combined withpaintings on glass, executed in reverse (seeChapter 2), were produced in China for exportto Europe. Other types are hand-held mirrorsand those produced for scientific use.

Mirrors cannot be considered in isolationfrom their frame, structure or furniture withwhich they are associated. It is necessary toagree a level of conservation in order thatvisual harmony is maintained. As far as possi-ble, the original glass should be preserved;replacing mirror glass in antique furniturewould look aesthetically displeasing, and maysignificantly reduce the value of the piece.Almost all mirrors can be conserved to somedegree if the budget allows. Where replace-ment cannot be avoided, the original shouldbe preserved in an archive. As a general ruleglass should be treated in situ wherever possi-ble. However, in some cases it will be neces-sary to remove the glass from its support inorder to carry out the conservation. Forexample, repairs to a mirror frame or furnituremay be necessary, whilst the mirror itself maybe in good condition. The potential risk ofdamage to the glass during handling and trans-port must be considered. It is generally safestfor the mirror to be well padded and to travelwithin the object which it decorates.Depending on the work to be undertaken onframes or furniture, the mirrors may remain insitu, but the risk of accidentally scratching theglass with abrasive papers or tools, or ofactually breaking it, are considerable.

In addition to such physical damage, thereis the possibility of chemical reaction betweenthe amalgam and emissions from wood preser-vatives, varnishes and paints, adhesives andlacquers. Acetic acid is given off by PVACadhesives, and by manufactured boards, suchas plywood and medium density fibre-board(MDF), which contain adhesive. Materialsbased on polystyrene resin emit styrene, whichblackens tin and silver. Where the glass andfurniture are to be treated by different conser-vators, the object should first be delivered tothe glass conservator who will remove themirrors. Conservation grade materials should


Figure 7.65 Chemical and physical damage of amirrored glass: deterioration of the amalgam, beginningas dark spots, loss of amalgam, and cracks in the glass.

be used wherever possible, and wood used toeffect replacements chosen with minimalemission of acetic acid.

If mirrors have to be removed from theirmounts, it must be ensured that there isadequate room to manoeuvre the glass and tosupport it in a vertical position without risk ofit being broken. This is particularly importantin the case of large mirrors, where ideally anarea several times larger than the sheet ofglass, and several operators, are needed inorder to ensure maximum safety. It may, forinstance, become necessary to remove mirrorsduring structural building works or if thesupport has failed. As with any conservationprocedure, it is necessary to balance the riskof leaving the mirror in situ against the risksinvolved in moving it.

Consideration must be given to the size andweight of the glass, its condition, e.g. thepresence of cracks or breaks, the strength(thickness), the condition of the mirroredsurface and any painted decoration, and theease of removal of the fixings which securethe glass to the mount. Advance planningshould include measures to ensure safehandling, storage and, if necessary, packagingand transport and reinstatement. If the glass isto be removed and laid flat, it should beplaced on a surface cushioned with a layer ofclean, non-abrasive, non-slip material such asPlastazote (inert polyethylene foam). Themirror glass is unlikely to be completely flat,and undulations must be packed withPlastazote in order to prevent areas of stressconcentration. In order to prevent accidentaldamage, large areas of mirror can be enclosedby a shallow wooden frame; smaller pieces ofmirror can be supported in wooden orpolypropylene trays lined with Plastazote.

Before and during dismantling the mirrorshould be fully documented, includingphotographs. The orientation of the mirror andits backboard should be recorded, and allcomponents, including the fixings, should belabelled and marked on a diagram. Whenwork begins, it is important to ensure that allfixings holding the glass are removed other-wise they might crack or scratch the glass asit is lifted away. The most common fixings arenails, screws, metal clasps or frames, woodenwedges, linseed putty and paper or fabrictape. A soldering iron can be used with care

to soften putty which will have hardened overtime and is otherwise difficult to remove.Needless to say, care must be taken not tocrack the glass. The glass once, loosened, canbe removed and laid on a support, putty sideup. Heat can then be applied to any remain-ing putty in order to peel it away. During andafter removal of glass, care must be taken notto damage loose glazing bars or to supportframes and support components.

There is no way of preventing the changein structure of a tin–mercury amalgam, and thedeterioration is irreversible. Although notpossible to predict the outcome accurately, itmay be possible to slow the process down bykeeping the mirror at a constant low temper-ature. The amalgam will be adversely affectedif the mirror is cooled to the low stability levelof 17.5°C or heated to 58°C, at which temper-ature the crystal phase will change. A relativehumidity of about 5 per cent, the lowest toler-ated by wooden frames and furniture in whichmirrors may be set, is not low enough toprevent corrosion of the amalgam by somehygroscopic salts.

It may be thought logical to cure mirrorswith a softening of the amalgam at the loweredge where there is an excess of mercury, bylaying them flat horizontally, or even byturning the mirrors onto their sides to redis-tribute the mercury. However this course ofaction is to be avoided since it can result infurther damage caused by flooding withmercury parts of the mirror which have a tight-packed relatively ‘dry’ crystalline structure.Excess mercury can be sucked away on tinfoil, but removal of too much liquid phase cancause further damage to the two-phaseamalgam. Mercury can be collected usingmercury salvage kits obtained from chemicalsuppliers, and disposed of in accordance withcurrent health and safety legislative guidelines.

It is advantageous to remove loose dust anddebris such as spiders’ webs, although it maybe difficult if the amalgam is flaking or liquid.Any attached framework, or surrounding woodshould be protected with acid-free tissue. Inthe case of glass that is undecorated with paintor gilding, loose dust can be removed usingsmall pads made from silk or lint-free cotton(large cloths may snag on delicate elementsand should therefore be avoided), and usingthe absolute minimum of pressure. If it


becomes necessary to use wet cleaningmethods, solvents such as distilled water witha non-ionic detergent, a 50:50 mixture ofdistilled water and industrial methylated spirits(IMS), IMS alone, acetone and white spirit (US:Stoddard solvent) may be used on cotton woolswabs. The swabs should be dipped in thesolvent and then blotted on paper before usein order to minimize the amount of solvent inuse. This is especially important where themirror remains in situ, to prevent solventsleaking beneath frames or beneath the glass.The surround should be protected withMelinex (US: Mylar) sheet, which can often beinserted between the glass and the surround.Wet cleaning treatments should obviously beavoided around cracks, as dirt would becarried into the cracks.

The use of commercial glass cleaningsolutions is to be avoided, since their chemi-cal composition is unknown; as is the sprayapplication of solvents, as the spray area isdifficult to control. Fine abrasive pastes shouldnot be used to clean glass as the particles mayscratch it. A soft cleaning paste can be madeup of whiting and containing a drop ofammonia. In the case of mirrors havingpainted or gilded surfaces, which are welladhered, the surface can be cleaned using asoft sable brush. Paint must be checked todetermine whether it is cold (unfired) or fired.Gilding will often have been applied in theform of water gilding, which would beremoved by wet treatments.

The amalgam surface of tin–mercury mirrorsshould not be vacuum cleaned since themercury will vaporize and be released into theenvironment. It is particularly important towear a mask with a mercury filter to preventthe inhalation of vapour, and to wear glovesto prevent mercury from being absorbed bythe skin as mercury can cause damage to thenervous system. The wearing of gloves willalso prevent acid from the fingers beingdeposited on the amalgam, to form finger-prints, which eventually penetrate through theamalgam (Figure 7.66). Loose debris can begently removed with a soft sable brush, keptonly for use on mirrors and stored in a self-seal plastic bag. In cases where the amalgamis unbroken, a weak rubber-based contactadhesive can be applied in a thin layer, androlled off after a few minutes as a coherent

film. If gaps in the amalgam are to be maskedby laying metal foil behind it, the new metalmust be isolated from the amalgam (e.g. withlacquer) in order to prevent mercury reactingwith the metal foil (Hadsund, 1993). Themirror amalgam itself should not be paintedor lacquered since it will not be able toremove the coating at a later date, should thisbecome necessary, without risk of damagingthe soft amalgam. A better solution used byJackson is the laying of a reflective coatedsheet of Scotchtint Melinex (US: Mylar) behindthe damaged glass. This material, developedfor use in solar reflective windows, is securedto backboards, not to the amalgam surface(Figure 7.67).


Figure 7.66 Fingerprints formed by acid from fingers,in an amalgam surface.

Figure 7.67 Sheet of Scotchtint placed behind adeteriorated mirror, to impart a reflective appearance.

Consolidation of flaking amalgam orpainted/gilded surfaces may be necessary, andshould be referred to a conservation special-ist. Painted decoration can be particularlyproblematical since there is a wide variety ofmaterials and techniques (see the section onreverse paintings on glass). Consolidation offlaking amalgam is difficult since the flakes areoften brittle so that the traditional method oflaying paint flakes by flowing consolidantbeneath the flake and then securing it to thesubstrate by application of a heated spatula,cannot apply. In addition, the amalgam maybreak down when subjected to heat. In somecases, the application of a small amount of aviscous consolidant, e.g. 50/50 wt/vol solutionof Paraloid B-72 (methyl methacrylate copoly-mer) in acetone, to the thin edge of amalgamflakes may act as a bridge to the glass, therebykeeping it in position. The consolidant shouldnot flow beneath the flake. Any pressureapplied to the flake will break or completelyshatter it.

The frame and backboard act as a supportfor the mirror glass and afford protection fromdust, dirt and insect nests. The frame andbackboard should be reasonably well-fitting,but not trap potentially harmful accumulationsof hazardous corrosion products. Suitablebacking materials for the glass include conser-vation grade fabrics, an acid-free card andpaper. The backboard can be lined with asheet of Melinex (US: Mylar). New woodshould always be covered with Melinex toprotect the mirror from any acid vapours thatmight be given off. Where possible, the origi-nal wedges and nails should be re-used whenreinstating a mirror. Hadsund (1993) illustratesa system for mounting and supportingtin–amalgam mirrors, and Dowling (1999)discussed the removal and installation ofmirrors in a museum environment.

Conservation of historic chandeliers andmirrors is often undertaken on site, for whichconsiderable forward planning is required (seeAppendix I).

Plain glazing

Painted and stained glazing is immediatelyvisually appreciated, and can easily be dated;it has been the subject of much scientific

research and conservation/restoration. How-ever, plain quarried, or plain leaded glazing isnot so visually attractive or easily dated, anduntil recent years, broken clear glass window-panes have simply been replaced. Since thenineteenth century, old glazing has beenreplaced with hand-made window glass, orworse, with modern machine-made, clear-white glass, without regard for the resultingchanges in visual appearance. Old hand-madeplain glazings will normally not be entirely flatnor entirely clear, and its light bending proper-ties result in a surface movement. It may havea greyish appearance, or exhibit a huge rangeof tone, contain air bubbles or exhibit paral-lel curves (if cut from crowns). In order todate the glass, a study has also to be made ofthe leading, lead ties, ferramenta, timberframes and mortar.

The historical value of original plain andplain leaded glass is beginning to be appreci-ated. Where it survives it should be recorded,retained, preserved, repaired and re-usedwhere possible. Some specialist glaziers haveextensive stocks of original glazing, recoveredfrom old buildings (domestic, public andecclesiastical). It is important to match thetexture and type of replacement glass formissing areas. Hand-made glass is no longermade commercially in Britain, but can beobtained from the Continent.

Features which should be recorded andpreserved, either in situ or in an archive, areinscriptions on lead, or diamond cut into theglass surface, the survival of pieces of horn,lead ties and the use of pattern-stamped leadgrilles glazed in for ventilation. It is importantthat anyone involved with original glazing ina building, i.e. glazier, architect, should bemade aware of its historical value, and of theneed to retain as much as possible.Specifications for this should be clear, avoid-ing such misleading terms as reglaze, reform,replace, or repair as new. It is possible for anexperienced glazier to repair and reconstructthe characteristic irregularities of pre-nineteenth century plain glazing patterns. Alead rubbing is taken of every individual panelbefore dismantling the glass, then numberingeach piece of glass on the glass itself and therubbing, so that each can be reglazed into itsoriginal position. The same width of leadingas the original should be used so that the


proportional relationship and size of eachpanel remains the same. Any replacementglass can be cut to size using the lead rubbingas a cartoon (pattern). Nineteenth-centuryreticulated plain quarry glazing is frequentlyreplaced with modern plain glazing of contem-porary design. It should be remembered thatplain quarry glazing was so widely used as itwas practical, inexpensive and a successfulsolution to maximizing light and glass becauseit held together without distortion. Whencorrectly fixed, the interactive network of glassand lead of diamond quarry glazing ensuredthat the considerable combined weight of theglass and lead in the glazed interspace wascarried evenly from top to bottom withoutstress (Kerr, 1991).

Painted and stained glass

The museum approach to the conservation ofpainted and stained glass windows, usingepoxy resins, is described by Holden (1991).The conservation and restoration of stainedglass windows, especially those in situ,requires specialist skills (Lee et al., 1982;Newton, 1987; Newton and Davison, 1989;Sloan, 1993).

Resetting excavated window glass

The resetting of fragile, excavated windowglass has been undertaken, but ideas are stillundeveloped. Much excavated material, evenafter consolidation, is not strong enough to bereleaded, nor stable or transparent enough tobe placed in a window opening, due to thepresence of opaque surface crusts. However,fragments of the medieval glass found orexcavated from churchyards have beenreleaded and installed as windows, e.g.Kirdford and Chiddingfold in Sussex (UK). Agreat deal of English medieval church windowglass was deliberately destroyed during theCivil War in the seventeenth century. Remainsof the medieval glass windows, found buriedin the churchyard of St Dunstan’s Church,Monks Risborough in Buckinghamshire (UK)were re-leaded in a random fashion, andreinstated in a window in 1807. Undoubtedlythis has been done elsewhere, and the workprobably involved injudicious cleaning of the

glass itself. Cleaning of discoloured medievalwindow glass is discussed by Fitz (1981; Fitzet al., 1984). The use of selected chelatingagents to reveal painted decoration onexcavated medieval glass has been publishedby Barham (1999).

Fragments of Anglo-Saxon glass excavatedfrom Jarrow Monastery (Northumbria, UK),have been re-leaded for display in JarrowMuseum (Cramp, 1968, 1975). Fragile glasswould not withstand incorporation in heavyleading, even for museum display, and there-fore, the use of narrow copper foil, fine leadsor other materials or devices may have to beexplored. For example, a lightweight, rigidsynthetic frame could be used, and thewindow suspended vertically; or the glasscould be laid flat, or at a slight angle, andmirrors used to produce a vertical effect.Before resetting, it is essential that details suchas grozed edges, which will be hidden by thenew cames, or which will be too distant forexamination in the new position, arethoroughly recorded.

Cloisonné glass

The manufacture of cloisonné glass panelsdates from the turn of the nineteenth century.There is little information regarding themanufacturing technique; the following infor-mation was taken from publicity materialissued by The Cloisonné Glass Company.Cloisonné panels were made up, by firstoutlining the design with thin gilt or silveredmetal wire secured with translucent cement(adhesive) to a sheet of clear glass (backplate).The cells thus formed were filled with potmetal coloured glass in granular form(< 1 mm), either globules or squares, andsecured in place with adhesive. The panelscould be used as an alternative in any situa-tion where stained or leaded glass, mosaic,fresco or tempera would be appropriate.

Deterioration and conservationThe problems most likely to be encounteredwith this type of glass are damage of the glasscovers and to the tinfoil binding. If either isdamaged, the glass beads and even the metalmay become detached with disastrous results.Another form of damage occurs if waterpenetrates the interior of the panel, dissolving


the adhesive binding the glass beads, andcausing the metal strips to corrode. A greatdeal of restoration of cloisonné glass panelshas been carried out in Spain. However, theballotini were removed entirely and the panelsreformed (Martin, 1985).

The restoration of a lampshade, the sides ofwhich were formed of panels of cloisonnéstained glass, with a base of transparentrippled green glass, is shown in Figures7.68a–c). The edges of the cloisonné panelswere bound in lead foil and clipped into abronze frame. The lampshade was suspendedfrom the ceiling by four metal chains (Figure7.68a). It was in a deteriorated condition as aresult of the lead foil, which bound the edgesof each panel, being corroded or missing. Itwas also extremely dirty, and had beenimmersed in water by its owner, whereuponwater had entered between the glass panels,dissolved the animal glue holding the glassballotini in place, and initiated corrosion ofthe metal bands forming the cloisons or cells.One of the panels was so badly damaged, theglass itself being broken and having onecorner and parts of the design missing, that itwas decided to open the panel up in order toeffect the restoration.

The sheet of glass to which the metalcloisons had not been originally attached, wascarefully lifted, whilst watching for any disrup-tion of the design (Figure 7.68b). Metal orglass, which has become detached or whichhas become transferred to the cover glass,would have to reattached to the base glass,using the tracing of the original design placedbeneath the base glass as a guide to enablethe metal strips to be correctly positioned. Theminute glass beads and fragments (Figure7.68c) were retrieved with tweezers, and,working under magnification, sorted intocolours and sizes. In order to secure the glassbeads, the surface of the base plate wascoated with Arabian glue, which was also usedto consolidate the mass. Once the adhesivehad dried, the cover plate was replaced, andthe edges bound with copper foil strips withan adhesive backing.

Photographic images on glass

Broken glass plates bearing photographicimages can be repaired (Figure 7.69a,b and


Figure 7.68 Restoration of a cloisonné glasslampshade. (a) The lampshade after restoration. (b) Oneof the panels opened up showing details of thecloisonné design, and the fact that some of the wiresand glass beads have become transferred to the coverglass. (c) Detail of the brass wires forming the cloisons,and the coloured glass beads, which form the design.

(a)

(b)

(c)

7.70a–c). Great care has to be taken not toallow pressure-sensitive tape, adhesive orsolvent to come into contact with the photo-graphic image. The plate is placed, image sidedown on a clean surface, and the fragmentssecured by placing narrow strips of pressure-sensitive tapes across the breaks. A smallamount of optically clear epoxy resin isapplied along the breaks, after which the plateis turned image side up, placed on a sheet ofMelinex and the resin left to cure.

Paintings executed in reverse on glass

The causes of damage to paintings on glasscan be broadly divided into two categories: (i)those caused as a result of damage to theglass, frame, backboard and other associated

materials (Figure 7.69a–c; and (ii) thoseassociated with changes to the bindingmedium and paint layer (Figure 7.70a,b).

The most obvious cause of damage topictures created on glass sheets is the break-ing of the glass itself, often as a result of thepicture falling to the floor when the cord fromwhich it is suspended rots or frays; or whenthe picture is undergoing re-framing. Thewooden backing and frame may also becomeweakened by the action of woodworm; orbroken, and lead to damage of the glassand/or paint. If the uneven glass is too firmlyheld in a rigid framework, or a backboard issecured by hammering or pushing in pins, orby firing in staples, the glass cracks or splin-ters. Wooden supports such as small blocksplaced within the frame, directly against the


Figure 7.69 Photographic image on an opal glass plate, (a) before and (b) after restoration with epoxy resin.

(a) (b)

glass, can abrade the paint layer if the glass isloose within the frame, or if they becomedetached. An applied backing of paper orother material can cause damage, either as aresult from adhesive with which it was appliedmigrating into or beneath the paint layer, orby pulling the paint off the glass if theadhesion between the paper and paint isgreater than that between the paint and theglass. The paint layer itself can be damagedin a number of ways.(i) Chemical instability of the paint, which

can be changed by light etc., so that theoriginal colour can then only be foundbeneath the edge of the frame, where itwas protected from light.

(ii) Failure of adhesion of the binder with theglass. Loss of the paint layer arises primar-ily from the glass being broken, or fromfailure of adhesion between the paint and

the glass or the paint and the transparentpriming coat. This occurs due to theeffects of oxidation of the paint itselfand/or the effects of UV light and heat.These agents can cause powdering,blistering or peeling of the paint, particu-larly in the case of multi-layered paintings.The paint may even become detached sothat fragments are found beneath theframe. The most typical disfigurementoccurring to reverse paintings on glass,amelierung, eglomisé and mezzotints is thedistortion caused by the air-pockets (blindcleavage) between the glass and paintlayer, which seen from the front appear asa greyish, less saturated areas of paint(Figure 7.70a). The gold leaf of goldreverse engravings usually remains ingood condition. The cleavage is caused bydrying out (oxidation) of the priming coat


(a) (b)

(c)

Figure 7.70 A hand-coloured, photographic image onthe reverse side of a sheet of glass. (a) Before repair;(b) after repair and restoration; (c) missing areas of thedesign, painted in acrylic colours on a sheet of Melinexand placed behind the original glass.

materials from which the painting wascreated, and/or the contraction or othermovement of any backing materialspresent. Paint flakes may curl away fromthe glass or even become detached. Dirtcan become trapped under cleaved andlifted paint areas, where it is impossible toremove it without causing further damage.

(iii) Penetration of water/water vapour. Some-times an osmotic effect can be created bet-ween a painting executed in two layers,between a hydrophilic base (of, glue orgum) and a hydrophobic paint layer. Thepresence of water vapour then causesproblems of deterioration. Display or stora-ge environments, where there has been arising water table or excessive condensa-tion, can cause damage. If hydrophobicmaterial has been painted on the reverse ofthe painting, micro-organisms can grow.

(iv) Previous attempts at restoration whichhave proved unsatisfactory, especially ifthese have been carried out usingadhesive tape, gummed paper tape,postage stamps, medical plasters etc.

Conservation and restorationThere are many different types of paintings onglass, and each requires an individualapproach to its conservation. Brill et al. (1990)published a classification of zwischengoldglas,accompanied by a discussion of the conserva-tion problems associated with zwischengold-glas.

Where possible, dust and dirt can becarefully removed with a small soft paintbrush, or if necessary with a small amount ofappropriate solvent. The difficulties of repair-ing panels of glass bearing reverse glass paint-ings have been reported by a number ofconservators (Wallace, 1976; Davison andJackson, 1985; Tremain, 1988). All havedocumented solutions for supporting unevenglass panels following the taping together offragments and the introduction of a highlymobile form of epoxy resin from the non-painted side (see Figure 7.25a–c). Thornton(1990) describes a light-box apparatus for therepair of flat glass. Great care has to be takennot to allow the resin to flow between theglass and the paint layer where it would causean irreversible darkening of the painting.Although the procedure is not generally

recommended, as the less interference on thepainted side the better, coating the originalpaint on either side of the breaks with watersoluble PVA may enable excess resin to beremoved without disturbing the paint(Wallace, 1976; Davison and Jackson, 1985). Itmay, however, be useful in instances where alarge number of joins has to be secured at thesame time.

In general, the edges of the glass fragmentsare carefully cleaned to remove dirt, grease orpreviously used adhesives, mechanically (inthe case of adhesives) with small swabs ofcotton wool moistened with acetone. The useof solvent has to be severely limited, for fearof disturbing the paint layer. The fragments arethen held in place with thin strips of trans-parent pressure-sensitive tape on theunpainted side only. Occasionally it may benecessary to employ small spots of cyano-acrylate adhesive or Sticky Wax on badlyshattered glass. On large panels especially,taping on the one side may pull the entirepanel out of alignment, so that it may have tobe supported by clamps, in order that it canstand freely on one edge. A short time afterapplication of the adhesive, any excess isremoved from the painted side if necessarywith dry or slightly acetone-moist cotton woolswabs, discarding each one after use in ordernot to spread the adhesive over the paintlayer. Small flat panels can be laid glass downon a sheet of Melinex placed on a flat surface,in order that excess adhesive drains back ontothe glass surface from where it can beremoved after curing, with a scalpel blade.

Reattaching the paint layerDifferent approaches have been adopted forrefixing a paint layer that has become partiallydetached from the glass substrate. The conser-vation methods can be broadly grouped intothose which employ water-based fixatives,those which use solvent-based techniques andthose which use wax, resin or a wax-resinmixture (Figure 7.71b, compare to Figure7.71a).

Aqueous consolidation agents were used tosecure brittle paint by Schott (1999); Klucel E(hydroxyl-propyl cellulose), Tylose MH 300(methyl cellulose) and polyvinyl alcohol werefound to be elastic enough not to exert stresson the paint layer.


Solvents have been used by several conser-vators, to relax and reattach flakes of paint(the Pettenkofer process). These includeParaloid B-72 (methyl methacrylate copoly-mer) used as a 25 per cent solution in xylene,diluted to 15 per cent with ethylene glycolmonoethyl ether (Cellosolve) (Wallace, 1976),or dissolved with xylene alone (Tremain,1988) or in ethylene glycol monoethyl etheralone (Wallace, 1976; Caldararo, 1997);Paraloid B-67 (Wharton and Oldknow, 1987);Plextol D 466 (dispersion of acrylic resin)(Agnini. 1999) and Plexisol P550 (Caldararo,1997; Agnini, 1999); PVA-AYAC (polyvinylacetate) in diacetone alcohol (Graham, 1976);Vinac B-7 in methanol in toluol (Roth in

Caldararo, 1997). The advantages of low-molecular materials, such as hydrocarbonresins, e.g. Regalrez 1094, are their solubilityin non-polar solvents, and their light-stability,which can be enhanced by a light-stabilizingadditive (HALS – hindered amine light stabi-lizer) such as Tinuvin 292. Their brittleness canbe offset by the addition of plasticizers suchas Kraton G1650 (2 per cent), but these canimpair the light stability of the resin and maymigrate into the paint layer. Paraloid B-72 andRegalrez 1094 were used by Coppieters-Mols(1999). Seidler (1987) suggested impregnatingthe paint layer with silane (irreversible) andEmery (1991) used Penetrol, a linseed oildissolved in a slow-drying solvent mixture, butwarns against spraying the paint layer withdammar varnish. When resins are flowedunder cleaved areas of paint to consolidatethem, it is difficult not to trap air bubbles; andfurther air bubbles can form during evapora-tion of the solvent. Areas of intact paint filmexhibiting blind cleavage cannot have consol-idants flowed beneath them without firstbreaking the paint film. As the paint is gener-ally friable, extensive loss can occur in tryingto create an opening through to the glass.

Wallace (1976) describes a method of treat-ment, in which a 25 per cent solution ofParaloid B-72 (methyl methacrylate copoly-mer) diluted to 15 per cent with Cellosolve(ethylene glycol monoethyl ether), was usedto reattach loose paint on two Europeanportraits and a nineteenth century Americanpicture, all painted in with a thin oleaginousmedium. The portraits had an intermediatepriming layer of linseed oil; the Americanpainting had no priming layer. The reverse ofeach fragment was brushed with the ParaloidB-72/Cellosolve mixture and allowed to dryuntil tacky. The fragments were then placedbetween four sheets of wet strength paper(two on each side) and placed in a preparedvacuum envelope, exposed to evenlydispersed air-blown heat of circa 55°C (130°F).After consolidation, the wet-strength tissues,which had served as isolation during thevacuum and heat treatment, were removedmechanically without difficulty. The panelswere repaired, by taping the fragments of glasstogether on the unpainted side, and introduc-ing an epoxy resin (EPO-TEK 301) to the joins,by capillary action. On one portrait, areas of


Figure 7.71 (a) Detail of flaking paint on ahinterglasmalerei; (b) re-laid by the resin-wax method.(Courtesy of S. Bretz).

(a)

(b)

loss were in-painted directly on the glass usinga methacrylate paste using xylene as a diluent;on the other, in-painting was carried out on arag-board backing, using opaque watercolours. Both paintings were sealed betweena backing board of 100 per cent rag and anew piece of glass sealed with Scotch MagicTape 810 pressure-sensitive tape. In-paintingfrom the reverse is extremely difficult bothtechnically and because slight coloration of theglass itself will alter the colours when viewedthrough it.

Williston and Berrett (1978) describe amethod of setting down paint flaking from anAmerican reverse painted glass surrounding aclock face, by relaxation of the paint in asolvent vapour. Caldararo (1997) also reportsthe Williston and Berrett work in an articlewhich contains an extensive bibliography.Thornton (1981) states that cleavage laid downusing solvent-relaxing techniques tends torecur after a period of time, possibly becausethe dirt and oxidation products at theglass/paint firm interface have not beenremoved, thus preventing good adhesion ofthe paint to the glass.

Wax and wax-resin mixtures applied with aheated spatula have been used to secure paintfragments for the past 15 years. Cerowax(micro crystalline hydrocarbon wax with ahigh proportion of isoparaffin); paraffin wax;and wax–resin mixtures (beeswax with theaddition of 5–20 per cent resin) have beenused to secure paint fragments. Agnini (1999)used unbleached beeswax as well as lowmelting point microcrystalline waxes. Schott(1988b, 1999) used a mixture of bleachedbeeswax with 5 larch-turpentine resin.Coppieters-Mols (1999) used bleachedbeeswax with Regalrez 1094, but there is aquestion concerning the possible futureseparation of the resins. Tremain (1988) andKleitz (1991) used paraffin wax in petroleumether in equal proportions. Horton-James(1990) and Agnini (1999) tried using the hot-seal adhesive Beva 371 for conserving paint-ings.

Thornton (1981) gives a detailed account ofa transfer treatment for dealing with cuppedpaint flakes and blind cleavage and cosmeticcompensation for reverse paintings on glass.The process solves many of the problemsencountered with previously reported treat-

ments involving solvent relaxation of the paintand/or consolidation with synthetic resins. Inthe transfer treatment, a severely damagedpaint film is transferred off the glass virtuallyintact, then cleaned and compensated from thefront, prior to reattachment with a low viscos-ity epoxy resin. The process is extremelydelicate and time-consuming. Two importantconsiderations are the disturbance of the paint-ing from its original substrate (although thiscould be likened to the re-lining of a paintingon canvas), and the irreversibility of theattachment with epoxy resin in practice.However, the method could be useful in caseswhere to do nothing may result in the totalloss of an entire painting.

Retouching the paint layerRetouching missing areas of paint on hinter-glasmalerei is difficult because the thickness


Figure 7.72 Retouching the design on a Chinesepainting in reverse on glass, using acrylic pigments.

and colour of the glass can distort the colourswhen they are viewed through it. This problemcan be overcome by setting the painting up infront of or above a mirror; the restoration inprogress can be viewed in the mirror. (If theoriginal fixative remains on the glass, it will actas a ground for the new paint.) Alternatively,missing areas can be painted on acid-freepaper (Wallace, 1976) or on Melinex sheet,which are then placed behind the original(Figure 7.70c). Paints used include Cryla acrylicpaints (Davison and Jackson, 1985) (Figure7.72), Aquarel paints, Rebel 2000 in ParaloidB-72 (Coppieter-Mols, 1999)

StorageReverse painted glass pictures should remain intheir frames during storage with the glass sideplaced face down on a soft surface such as felt.The pictures should be wrapped in acid-freetissue and placed in an archival quality boxresistant to air and water vapour. If a paintingis in a poor state of preservation, which doesnot permit transport, the paint layer should begiven a coat of a volatile bonding agent(cyclododecane). The storage environmentshould be air conditioned, free from air pollu-tion and have a temperature range of 18–20°C, with a relative humidity of 50–55 per cent.


It would be impractical to attempt to producea definitive list of materials used for glassconservation throughout the world. The major-ity are not produced specifically for conserva-tion, but have been tested by conservationscientists, and been proven in use over anumber of years.

Generally speaking, it is preferable to obtainmaterials in the country in which they are to beused, in order to have the possibility of directcommunication with manufacturers and suppli-ers (although communication has been greatlyimproved by the use of e-mail); and to minimizestorage and transport time. It should be remem-bered that materials that perform well in onepart of the world (e.g. in a temperate zone),may not in another (e.g. tropical zone). In manycases, major manufacturers have outlets for theirmaterials in several countries and the addressescan be obtained from the parent company.

Other sources for locating materials andequipment are the conservation sections ofmuseums and other institutions, privaterestoration studios, manufacturers and suppli-ers of adhesives, resins, chemicals and suppli-ers of scientific laboratory equipment.

There are, however, a few specialist suppli-ers of conservation materials:

• Conservator’s Emporium (replacing Con-servation Materials Ltd), 100 Standing RockCircle, Reno, Nevada 89511, USA.

• Conservation Resources InternationalL.L.C., 8000-H Forbes Place, Springfield,Virginia 22151, USA.

• Conservation Resources (UK) Ltd, Units 1,2, 4 & 5, Pony Road, Horspath IndustrialEstate, Cowley, Oxford OX4 2RD. Con-servation Resources’ products (which areprincipally archival) are available from anumber of specialist distributors world-wide.

• Conservation Support Systems, PO Box91746, Santa Barbara, California 93190-1746, USA.

• Stuart R. Stevenson (Artists’ & GildingMaterials), 68 Clerkenwell Road, LondonEC1M 5QA, UK.

On-site glass conservation/restoration –checklist of materials and equipment

Archaeological excavationMuch of the equipment required is of ageneral nature, used on archaeological sites(Sease, 1988; Cronyn, 1990). (See also itemslisted below.) Where the retrieval of glass inlarge or small quantities is anticipated, specialprovision will have to be made, especially ifthe glass is damp or waterlogged. It may, forexample, be necessary to construct temporarystorage tanks on site.

Packaging materials should include self-sealpolythene bags, acid-free polyethylene foamand tissue, pressure-sensitive tapes, labels andmarking pens (which may need to be water-proof), rigid plastic boxes with lids, plasticbowls and trays, conservation record forms.

345

Appendix 1

Materials and equipment for glassconservation and restoration

Historic buildings, museum collectionsetc.Arrangements to be made in advance of thework commencing:• A contract stating what the conservator is

and is not responsible for.• Public liability and employer’s liability

insurance.• Security clearance, security passes, security

arrangements.• The provision or hiring and erection of

scaffolding.• Sources of electricity, water and means of

refuse disposal (including solvents).• Liaison with a qualified electrician (in the

case of lighting e.g. chandeliers).• Materials, tools and equipment.• Transport of equipment and personnel to

and from site.• Delivery points and temporary and perma-

nent parking facilities.• Times of access (particularly any restric-

tions on working outside normal workinghours, or at weekends).

• Facilities for tea-making, eating, location oftoilet facilities.

• Use of a telephone for official calls, a listof names and telephone numbers of allpersonnel involved with the project,including emergency contacts.

• Accommodation (if the site is a consider-able distance from base).

• Packaging, transport and secure storage ofobjects, e.g. dismantled chandeliers, ifrequired.

• Administration costs, including costs ofwriting and producing condition andconservation reports, and photography.

Tools and equipment• Adjustable spanners and wrenches, mole

wrench, screwdrivers, long-nose pliers andside cutters.

• Light-weight scaffold tower(s), tall steplad-der.

• Trestle tables, chairs.• Torches (large and small).• Weighing scales – digital platform type (for

chandeliers).• Buckets, bowls, trays, miscellaneous con-

tainers.• Portable vacuum cleaner, dust-pan and

brush, broom.• Hard hats, safety goggles, dust/fume/

mercury vapour masks, cotton and dispos-able vinyl rubber gloves, protective cloth-ing. First-aid kit (especially scissors andplasters).

• Camera(s) and accessories: lights, flashlight, film: black and white, colour print,transparencies, batteries, tripod.

Materials• Soft cloths, paper towelling, cotton wool,

satay sticks.• Non-ionic detergent.• Clear epoxy resin, rapid cure epoxy resin,

cellulose nitrate-based adhesive.• Paraloid B-72 adhesive and consolidant.• Acetone, IMS, distilled or de ionized water.• Fine lubricating oil (spray).• Rust remover (liquid or gel).• Silver cleaning solution.• Melinex sheet.• Fine abrasives.• Brass wire and pins.• Rubbish disposal bags.• Packaging materials: wooden crates, strong

cardboard boxes, Bubblewrap, acid-freetissue, Plastazote (acid-free polyesterfoam), self-seal Polythene bags; pressure-sensitive tapes (Magic Tape, Sellotape,masking tape, plastic parcel tape); markingpens, stencils for marking crates.

StationeryRecording/condition report forms, note-pads,paper, pens, pencils, erasers, tie-on andadhesive labels.


Information concerning all aspects of glass iscontinually being updated in journals andconservation proceedings, by research projectsand in (largely unpublished) university studenttheses. Information can be found in a numberof dictionaries, encyclopaedias and bibliogra-phies. For convenience, a list is given below,with publication details for those works notappearing in the Bibliography of this volume.

Dictionaries, encyclopaedias andbibliographies

ASH, D. (1975) Dictionary of British Antique Glass.Pelham Books, London.

BERLYE, M.K. (1963) The Encyclopaedia of Working withGlass. Oceana Publications, New York.

BLECK, R.D. (1967–79) Bibliographie der Archaologisch-Chemischen Literatur (Nature of ArchaeologicalMaterials). Three Volumes.

DIDEROT, D. and D’ALEMBERT, J. le ROND (1751–71)Dictionnaire Raisonné des Sciences, des Arts, et desMétiers. Contains entries on eighteenth-century glass-making, window and mirror manufacture. Well illus-trated with line drawings.

DUNCAN, G.S. (1960) Bibliography of Glass (from theearliest records to 1940).

FLEMING, J. and HONOUR, H. (eds) (1976) The PenguinDictionary of Decorative Arts. Penguin, London.Contains entries on glass and related materials.

HENCH, L.L. and McELDOWNEY, B.A. (1976) ABibliography of Ceramics and Glass. American CeramicsSociety International Commission on Glass (1972–1979),The Chemical Durability of Glass. Three volumes.

NEWMAN, H. (1977) An Illustrated Dictionary of Glass.Glass in general.

NEWTON, R.G. (1973, 1974a, 1982b) Three extensivebibliographies on the conservation of glass; the 1982revision contains a thirty-page introduction which alsoserves as an index.

NEWTON, R.G. and DAVISON, S. (1989) Conservation ofGlass. Butterworth Heinemann, Oxford. Contains anextensive bibliography. Also a section on painted(stained) glass, which is not included in this edition.

TRENCH, L. (ed.) (2000) Materials and Techniques in theDecorative Arts, John Murray, London. Contains entrieson glass and related materials (by S. Davison).

VANDIVER, P.B. DRUZIK, J.R., WHEELER, G.S. andFREESTONE, I.C. (eds) (1988 and cont.) Materials Issuesin Art and Archaeology. Materials Research SocietySymposium Proceedings, Pittsburgh, PA. In severalvolumes.

WHITEHOUSE, D. (1993) A Pocket Dictionary of TermsCommonly Used to Describe Glass and Glassmaking.New York, The Corning Museum of Glass.

Journals

Glass Engraver, published quarterly by the Guild of GlassEngravers, 8 Rathcoole Ave., London N8 9NA, UK.

Glass News, published by The Association for the Historyof Glass Ltd, Museum of London, 46 Eagle Wharf Road,London N1 7EE.

Glass Technology, published quarterly by the Society ofGlass Technology, Don Valley House, Savile Street East,Sheffield, S4 7UQ. Contains articles in English,comments, information, book reviews and abstractsfrom a wide range of journals.

Journal of Glass Studies (1959–), published annually byThe Corning Museum of Glass, One Museum Way,Corning, NY 14830-2253, USA. Includes an extensivechecklist of recently published articles and books on allperiods of glass, in its historical, economic and artisticaspects.

Verres, published by the Institut du Verre, 21 BoulevardPasteur, 75015 Paris, France.

Studies in Conservation (1952–), published quarterly bythe International Institute for the Conservation ofHistoric and Artistic Works, 6 Buckingham St, LondonWC2N 6BA, UK. Contains occasional articles on conser-vation of glass and associated materials.

The Conservator (1977–), published annually by theUnited Kingdom Institute for Conservation of Historicand Artistic Works. Contains occasional articles onconservation of glass and related materials.

Occasional articles on glass matters are published in theconference proceedings of various archaeological,historical association and conservation organizations,e.g. IIC, AIC, UKIC, ICOM–CC, AIHV.

Appendix 2

Sources of information

347

Conservation organizations

IIC – International Institute for theConservation of Historic and Artistic Works, 6Buckingham Street, London WC2N 6BA, UK.Details of conservation organizations can befound at: www.iiconservation.org.

Associations related to glass

Association for the History of Glass Ltd, Museum ofLondon, London Wall, London, EC2Y 5HG, UK.

AIHV – Association Internationale pour l’Histoire du Verre,Musee du Verre, quai de Maastricht 13, Liège B-4000,Belgium.

The Glass Association, Broadfield House Glass Museum,Compton Drive, Kingswinford, West Midlands, DY69NS, UK. Aims to promote the understanding andappreciation of glass and glassmaking methods.

The Glass Circle, Mrs J.M. Marshall (Hon. Sec.) 2 DowningCourt, Grenville Street, London, WC1N 1LX, UK.

The Glass Bead Society of Great Britain, c/o The HornimanMuseum, London Road, London, SE23 3PQ, UK. (Trust)

Crafts Council, 44a Pentonville Road, Islington, London,N1 9B, UK. Maintains a register of craftspeople of excel-lence in the UK.

British National Committee for the Conservation of StainedGlass (Corpus Vitrearum), c/o The British Academy,20–21 Cornwall Terrace, London, NW1 4QP, UK.

Glass and Glazing Federation, 44–48 Borough High Street,London, SE1 1XB, UK.

British Glass Manufacturers Confederation, (also knownas British Glass), Northumberland Road, Sheffield, S102UA, U.K.

Council for the Care of Churches, 83 London Wall,London, EC2M 5NA, UK.

Society of Glass Technology, Don Valley House, SavileStreet East, Sheffield, S4 7UQ, UK.

Selected glass collections

United KingdomThe British Museum, Great Russell Street, London, WC1B

3DG.The Victoria and Albert Museum, Cromwell Road, South

Kensington, London. SW7 2RL.The National Glass Centre, Liberty Way, Sunderland, SR6

0GL.The National Museum of Scotland, Edinburgh, EH1 1JF.The National Museums and Galleries of Wales, Cardiff,

CF11 3NP.The Fitzwilliam Museum, Cambridge, CB2 1RB.The Ashmolean Museum, Oxford, OX1 2PH.Broadfield House Glass Museum, Compton Drive,

Kingswinford, West Midlands, DY6 9QA.Centre for Glass Research, Sir Robert Hadfield Building,

Mappin Street, Sheffield, S1 3JD.The World of Glass (formerly The Pilkington Museum of

Glass), Chalon Way East, St Helens, Merseyside, WA101BX.

Many other towns and cities have glass collec-tions in museums, galleries and historic houses.Details of these can be found in annual publi-cations (handbooks) produced by The NationalTrust (for Places of Historic Interest or NaturalBeauty), 36 Queen Anne’s Gate, London,SW1H 9AS, and English Heritage, 23 SavileRow, London, W1X 1AB; and in Museums andGalleries in Great Britain and Ireland, andHistoric Houses, Castles and Gardens, widelyavailable in libraries and newsagents.

EuropeThe National and Applied Art Museums ofEuropean countries hold important collectionsof glass, as do some provincial museums. Onlythose dedicated specifically to glass are listedbelow.

BelgiumMusée du Verre, quai de Maastricht 13, LiègeB-4000.

Czech RepublicThe Glass and Costume Jewellery Museum,Jablonec nad Nisou.The Glass Museum, Novy Bor.The Glass Museum, Kamenicky Senov.

FinlandSuomen Lasimuseo (The Finnish GlassMuseum) (established in a former glasshouse),Tehtaankatu 23, FIN-11910, Riihimaki, Finland.(66 km from Helsinki).

ItalyMusee Vetrario (Museum of Glass), island ofMurano, Venice.

United States of AmericaThe Corning Museum of Glass, 1 Museum Way,Corning, NY 14830-2253.

Conservation and restorationdepartments and studios

The author decided against including a list ofconservation departments and studios whichundertake the conservation and restoration ofarchaeological and decorative glass, since thework of some are not known, and others may beinadvertently omitted. Many glass conservation/restoration departments and studios are attachedto museums and other institutions; the addressesof commercial organizations can be found inantique trade and telephone directories.


Abbreviations used to denote institutionsAIC = American Institute for ConservationBMP = British Museum PublicationsCCI = Canadian Conservation InstituteICOM–CCI– = International Committee of Museums -

Committee for ConservationIIC = International Institute for Conservation

and Restoration of Historic and ArtisticWorks of Art

INA = Institute of Nautical ArchaeologyUKIC = United Kingdom Institute for Conservation

ADAMS, P.B. (1984) Glass corrosion: A record of the past?A predictor of the future? J. Non-Cryst. Solids, 67,193–205.

ADLERBORN, J. (1971) Investigation of weathered glasssurfaces with the scanning microscope. OECD Report onScience Research on Glass. Ref: DAS/SPR/71.35.

AGNINI, E. (1999) Hinterglasbilder. Erfahrungen bei derKonservervierung, Restaurierung und Montierung.Restauro, 4, 258–65. Callwey Verlag, Munich.

AGRICOLA (1556) De re Metallica. Basle (1556).Translated by H.C. and L.H. Hoover, London (1912).

AIGNER, H. (1992) Die Hinterglasmalerei in Sandl/Buchers. Ein Beitrag zur Sozial- und Wirtschaftgeshicteds sudböhmisch-österreichischen Raumes (Forschungs-und Dokumentationsprojekt). Hinterglasmuseum Sandlo. J.

ALBRIGHT & WILSON LTD (1978) The Properties andApplication of Calgon.

ANDRÉ, J.M. (1976) The Restorer’s Handbook of Ceramicsand Glass. Van Nostrand, New York; also Keramik undGlas, Berlin.

ARRHENIUS, B. (1973) Teknisk verksamhet. Kungl.Vitterhets historie och antikvitets akademieus årsbok,176–82.

ARTAL-ISBRAND, P. (1998) A minimally intrusive fillingand stabilisation solution for very fragile archaeologicalglass objects. Proc. ICOM–CC Working Group, GlassCeramics, and Related Materials, (A. Paterakis, ed.),EVTEK Institute of Arts & Design, Vantaa, Finland.

ASAHINA, T-I., YAMAZAKI, F., OTSUKA, I., NAMADA, T.,SAITO, K. and ODA, S. (1973) On the colorless glassbottle of Tóshódaiji Temple ... [b]-ray backscattering.Sci. Pap. Japn Antiques, 6, 14–18. Abstract in Stud. inCons. I, 314.

ASHLEY-SMITH, J. (1999) Risk Assessment for ObjectConservation. Butterworth-Heinemann, Oxford.

ASHURST, D. (1970) Excavations at Gawber Glasshouse,near Barnsley. Post-Med. Archaeology, 4, 135–40.

ASMUS, J.F. (1975) The use of lasers in the conservationof stained glass. In Preprints of the Contributions to theStockholm Congress on Conservation in Archaeology andthe Applied Arts. IIC, London (D. Leigh, A. Moncreiff

and A. Oddy, eds), pp. 139–42.ASPINALL, A., WARREN, S.E., CRUMMETT, J.G. and

NEWTON, R.G. (1972) Neutron activation analysis offaience beads. Archaeometry, 14, 27–40.

AUGERSON, C.C. and MESSINGER II, J.M. (1993)Controlling the refractive index of epoxy adhesives withacceptable yellowing after ageing. JAIC, 32, 311–14.

BACON, F.R. (1968) The chemical durability of silicateglass. Glass Ind., 49, 438–9, 442–6, 494–9, 519, 554–9.

BACON, F.R. and CALCAMUGGIO, G.L. (1967) Effect ofheat treatment in moist and dry atmospheres on thechemical durability of soda-lime glass bottles. Bull. Am.Ceram. Soc., 46, 850–5.

BACON, L. and KNIGHT, B. (eds) (1987) From Pinheadsto Hanging Bowls; the Identification, Deterioration andConservation of Applied Enamel and Glass Decorationon Archaeological Artefacts. UKIC, London.

BACON, F.R. and RAGGON, F.K. (1959) Promotion ofattack on glass and silica by citrate and other anions inneutral solution. J. Am. Ceram. Soc., 42, 199–205.

BAKARDJIEV, I. (1977a) Korrosionsuntersuchungen anGläsern des 17. und 18. Jahrhunderts. DGG Meeting inFrankfurt, March, 1977 (Abstract No. 274 in CVMA NewsLetter 26).

BAKARDJIEV, I. (1977b) Stand der Untersuchungen an‘kranken’ Gläsern. Projektgruppe ‘Glas’ meeting inWürzburg, 14 September 1977 (Abstract No. 275 inCVMA News Letter 26).

BARHAM, E. (1999) The use of selected chelating agentsto reveal the painted decoration on excavated medievalwindow glass. Proceedings of Current Research on theHistory of Glass through Scientific Analysis. London.

BARNES, I.L., GRAMLICH, J.W., DIAZ, M.G. and BRILL,R.H. (1978) The possible changes of lead isotope ratiosin the manufacture of pigments: a fractionation experi-ment. Archaeological Chemistry, 2. Advances inChemistry Series 171 (G.F. Carter, ed.), pp. 273–7.American Chemistry Society, New York.

BARRELET, J. (1953) La Verrierie en France de l’epoqueGallo-Romaine à nos jours. Larousse, Paris.

BARRERA, J. and VELDE, B. (1989) A study of Frenchmedieval glass composition. Archéologie Médiévale,XIX, 81–130.

BARRERA, W.M. and KIRCH, P.V. (1973) Basaltic glassartefacts from Hawaii: their dating and prehistoric uses.J. Polynesian Soc., 2, 176–87.

BARRINGTON HAYNES, E. (1959) Glass Through the Ages.Pelican Books, Harmondsworth.

BASS, G. (1978) Glass treasure from the Aegean. NationalGeographic Magazine. Washington, DC.

BASS, G. (1979) The shipwreck at Serçe Limani.Archaeology, 32 (1).

BASS, G. (1980) The wreck of the Serçe Limani. INANewsletter, 7, (2/3) 1–6.

349

Bibliography

BASS, G. (1984) The nature of the Serçe Limani. Glass: J.Glass Stud., 26.

BAUER, W.P. (1976) Der Einfluss von Reinigungs-methoden auf die Glassoberfläche (Vorläufige Versucheund mikroskopische Untersuchungen). Verres etRéfract., 30, 62–-4. (A longer version, in English, isgiven in CVMA News Letters, 18, item 2, and 23,12–13.)

BAXTER, M.J., COOL, H.E.M., HEYWORTH, M. andJACKSON, C. (1995) Compositional variability in colour-less Roman vessel glass. Archaeometry, 37, 129–42.

BERKOUWER, M. (1994) Freezing to eradicate insect pestsin textiles at Brodsworth Hall. The Conservator, 18,15–22. UKIC, London.

BESBORODOV, M.A. and ABDURAZAKOV, A..A. (1964)Newly excavated glassworks in the USSR, 3rd-14thcenturies AD. J. Glass Stud. 6, 64–9.

BESBORODOV, M.A. and ZADNEPROVSKY, J.R. (1963)The early stages of glassmaking in the USSR. Advancesin Glass Technology., Part 2 (F.R. Matson and G.E.Rindone, eds), pp. 291–2. New York.

BETTEMBOURG, J-M. (1972) Nettoyage par voiechemique et par ultrasons des verres de vitraux. Compt.Rend. 8e Colloq. CVMA, York, p. 47.

BETTEMBOURG, J-M. (1973) Cleaning of medievalwindow glass (A statement of his practice in 1973 isgiven in CVMA News Letter 7, 3, 1974).

BETTEMBOURG, J-M. (1976) Composition et alteration des verres de vitraux anciens. Verres et Refract., 30,36–42.

BETTEMBOURG, J-M. and BURCK, J.J. (1974) Restaurationdes vitraux anciens. Methodes testées par la Laboratoirede Recherche des Monuments Historiques, 10 pp.typescript; abstracted on pp. 8-9 of CVMA News Letter13, 1975.

BETTEMBOURG, J-M. and PERROT, F. (1976) La restau-ration des vitreaux de la façade occidentale de la cathe-drale de Chartres. Verres et Réfract., 30, 92–5.

BICKERTON, L.M. (1971) An Illustrated Guide toEighteenth Century English Drinking Glasses. Barrie andJenkins, London.

BIMSON, M. (1975) Coloured glass and millefiori in theSutton Hoo grave deposit. In The Sutton Hoo ShipBurial, 3: The hanging bowls, silver vessels and domes-tic objects (A.C. Evans, ed.). BMP, London.

BIMSON, M. and FREESTONE, I.C. (1983) An analyticalstudy of the relationship between the Portland Vase andother Roman cameo glasses. J. Glass Stud., 25, 55–69,Corning Museum of Glass, Corning, NY.

BIMSON, M. and FREESTONE, I.C. (eds) (1987) EarlyVitreous Materials. British Museum Occasional Papers,56, London.

BIMSON, M. and WERNER, A.E. (1964a) The danger ofheating glass objects. J. Glass Stud., 6, 148–50.

BIMSON, M. and WERNER, A.E. (1964b) Scientific exami-nation of ancient glass. Ann 3e. Congr. Journ. Int. Verre,Damascus, AIHV, Liège, pp. 200–9.

BIMSON, M. and WERNER, A.E. (1967) Two problems inancient glass: opacifiers and Egyptian core material.Ann 4e. Congr. Journ. Int. Verre, Ravenne–Venise,AIHV, Liège, pp. 262–6.

BIMSON, M. and WERNER, A.E. (1969) Problems inEgyptian core glasses. Studies in Glass History andDesign, Sheffield, 121–2.

BIMSON, M. and WERNER, A.E. (1971) Notes on asuggested technique for the consolidation of fragileexcavated glasses. Proc. 9th Int. Congr. Glass, Versailles,pp. 63–5.

BIRINGUCCIO (1540) Quoted by Turner, W.E.S. (1956a);see also SMITH, C.S. and GNUDI, M.T. (1942) andChapter 14, Concerning Glass and All Other Semi-materials in General, pp. 126-33, in Pirotechnica.Cambridge, Mass., 1959 (a Renaissance treatise onsmelting and metalworking).

BLECK, R.D. (1967–1979) Bibliographie der Archaeo-logisch-Chemischen Literatur, 1 (1967), 2 (1968), 3(1971), 4 (1979), Weimar, Germany.

BLY, J. (ed.) (1986) Is it Genuine? Mitchell Beazley,London.

BOHM, C. (1998) Crizzling Glass [in Swedish collections]Revisited. Unpublished paper given at the ICOM–CCInterim Meeting of the Working Group on Glass,Ceramics and Related Materials, Vantaa, Finland.

BONTEMPS, G. (1868) Guide du Verrier, Paris.BOON, G.C. (1966) Roman window-glass from Wales. J.

Glass. Stud., 8, 41–5, Corning Museum of Glass,Corning, NY.

BORBELY-KISS, I., FUELOEP, Z., KISS, A.Z., KOLTAY, E.and SZABO, G. (1994) The PIXE–PIGE method for theclassification of late Roman glass sealings. NuclearInstruments and Methods in Physics Research B, 85,836–9.

BOVINI (1964) Les anciens vitraux de l’eglise Sainte-Vitala Ravenne. Ann. 3e Congr. Journ. Int. Verre, Damascus,AIHV, Liège, pp. 85–90.

BRAIN, C. (1992) Glass reconstruction materials andtechniques. In Glass and Enamel Conservation, UKIC,Occasional Papers, 11, 15–17, London.

BRAIN, C. (1999) Using glass luminosity to study 17thcentury English drinking glass. Proceedings of CurrentResearch on the History of Glass through ScientificAnalysis. UKIC, London.

BRANDA, F., LAUDISIO, G., CONSTARTINI, A. and PICCI-OLI, C. (1999) Weathering of Roman glass. A newhypothesis of pit formation on glass surfaces. GlassTechnol., 40 (3), 89–91.

BRETZ, S. and RYSER, F. (2000) Kleines Handbuch derHinterglasmalerei (Petit manuel de la peinture sousverre). In Glanzlichter. Die Kunst der Hinterglasmalerei.Exhibition catalogue. Museé Suisse du VitrailRomont/Museum in der Burg Zug.

BREWSTER, D. (1855) On the phenomenon of decom-posed glass. Trans. Brit. Assoc. Adv. Sci, 10.

BREWSTER, D. (1863) On the structure and opticalphenomena of ancient decomposed glass. Phil. Trans.R.. Soc. Edinb., 193–204.

BRILL, R.H. (1962) A note on the scientists’ definition ofglass. J. Glass Stud., 4, 127–38.

BRILL, R.H. (1963) Ancient Glass. Scientific American,November, pp. 120–30.

BRILL, R.H. (1965) The chemistry of the Lycurgus Cup.Proc. 7th Int. Congr. Glass, Brussels, Paper No. 223.

BRILL, R.H. (1967a) Lead isotopes in ancient glass. Ann.4e Congr. Journ. Int. Verre, Ravenne–Venise, AIHV,Liège, pp. 255–61.

BRILL, R.H. (1967b) Information for persons submittingsamples [for lead isotope analysis]. Ann. 4e Congr. Journ.Int. Verre, Ravenne–Venise, AIHV, Liège, p. 261.


BRILL, R.H. (1968) The scientific examination of ancientglass. Proc. 8th Int. Congr. Glass, London, 47–68.

BRILL, R.H. (1970a) Lead and oxygen isotopes in ancientobjects. Phil. Trans. R. Soc, Lond, A.269, 143–64.

BRILL, R.H. (1970b) Some chemical observations on thecuneiform glass making texts. Ann. 5e Congr. Journ. Int.Verre, Prague, AIHV, Liège, pp. 319–36.

BRILL, R.H. (1970c) The chemical interpretation of thetexts. In Glass and Glassmaking in AncientMesopotamia (Oppenheim et al., eds), pp. 105–28.Corning Museum of Glass, Corning, NY.

BRILL, R.H. (1971a) Chemical-analytical round-robin offour synthetic ancient glasses. Proc. 9th Int. Congr. onGlass, Versailles, pp. 93–110.

BRILL, R.H. (1972) Incipient crizzling in some earlyglasses. JAIC, 12, 46–7.

BRILL, R.H. (1973) Analysis of some finds from the Gnalicwreck. J. Glass Stud., 15, 93–7.

BRILL, R.H. (1975) Crizzling-a problem of glass cons.Preprints of the Contributions to the Stockholm Congresson Conservation in Archaeology and the Applied Arts (D.Leigh, A. Moncrieff and W.A. Oddy, eds), pp. 121–34.IIC, London.

BRILL, R.H. (1978) Use of equilibrated silica gel for theprotection of glass with incipient crizzling. J. Glass Stud.,20, 100–18.

BRILL, R.H. (1988) Scientific Investigations. In G.D.Weinburg, Excavations at Jalamie: Site of a GlassFactory in Late Roman Palestine. University of Missouri,Columbia, pp. 257–94.

BRILL, R.H. (1995) Appendix 3: Chemical analysis of someglass fragments from Nishapur in the Corning Museumof Art. In Nishapur: Glass of the Early Islamic Period,pp. 211–13.

BRILL, R.H. (1999) Chemical analyses of early glasses, 1The Catalogue. 2 The Tables. 3 The Interpretation of theResults, Corning Museum of Glass, Corning, NY

BRILL. R.H. and BARNES, I.L. (1988) The examination ofsome Egyptian glass objects. In Researches in theArabah, 1959–1984, I: The Egyptian Mining Temple atTimna (B. Rothenburg, ed.), pp. 217–23. LondonInstitute for Archaeometallurgy, London.

BRILL, R.H., AIKEN, C.A., NOVICK, D.T. and ERRETT, F.E.(1990) Conservation problems of Zwischengoldglas. Part1: Examination and Analyses. J. Glass Stud., 22, 12–35.

BRILL, R.H. BARNES, I.L. and ADAMS, B. (1974) Leadisotopes in some ancient Egyptian objects. Rec.Advances in Science and Technology of Materials, 3 (A.Bishay, ed.), pp. 9–25. Plenum Press, New York.

BRILL, R.H., BARNES, I.L. and JOEL, E.C. (1991) Leadisotope studies of early Chinese glass. In ScientificResearch in Early Chinese Glass. Corning Museum ofGlass, Corning, NY.

BRILL, R.H., FLEISCHER, R.L., PRICE, P.B. and WALKER,R.M. (1964) The fission-track dating of man-madeglasses: preliminary results. J. Glass Stud., 6, 151–5.

BRILL, R.H. and HANSON, V.P. (1976) Chemical analysesof Amelung glasses. J. Glass Stud., 18, 216–38.

BRILL, R.H. and HOOD, H.P. (1961) A new method ofdating ancient glass. Nature, 189, 12–14.

BRILL, R.H. and MARTIN, J.H. (1991) Scientific Researchin Early Chinese Glass. Corning Museum of Glass,Corning, NY.

BRILL, R.H. and MOLL, S. (1961) The electron beam

microanalysis of ancient glass. Recent Advances inConservation. Butterworth, Oxford, 145–51. (Also short-ened version in Advances on Glass Technology (1963),293–302.)

BRILL, R.H., SHIELDS, W.R. and WAMPLER, J.M. (1970)New directions in lead isotope research. In Applicationof Science in Examination of Works of Art (W.J. Young,ed.), pp. 73–84.

BRILL, R.H. and WOSINSKI, J.P. (1965) A huge slab ofglass in the ancient necropolis of Beth She’arim. Proc.7th Int. Congr. on Glass, Brussels, Paper No. 219.

BRUCE, J. (1979) Work in Progress. Glassblowing: aManual of Basic Technology. Crafts Council and JaneBruce, London.

BRUCE-MITFORD, R. (1975) The Sutton-Hoo Ship Burial,3, The Hanging Bowls, Silver Vessels and DomesticObjects (A.C. Evans, ed.). BMP, London.

BRUNDLE, C.R., EVANS, A.C. and WILSON, S. (1992)Encyclopaedia of Materials Characterization, MaterialsCharacterization Series, Surfaces, Interfaces and ThinFilms. Butterworth-Heinemann, Boston, Mass.

CABLE, M. (1998) The operation of wood fired glassmelting furnaces. In The Prehistory and History ofGlassmaking Technology, (P. McCray, ed.), Ceramicsand Civilisation, VIII, 315–29. American CeramicsSociety, Columbus, Ohio.

CABLE, M. and SMEDLEY, J.W. (1987) Liquidus tempera-tures and melting characteristics of some early containerglasses. Glass Technol., 28, 94–8.

CALDARARO, N. (1997) Conservation treatments of paint-ings on ceramics and glass: two case studies. Stud.Conserv., 23 (3), 157–64.

CAVAILLÉ, D., VIALLA, R., LEVELUT, C., VACHER, R., LeBOURIS, E. and GY, R. (1998) Internal stresses in silico-soda-lime glasses investigated by Brillouin scattering.Proc. 18th Int. Congr. Glass, San Francisco, CD-ROM,Paper No. 3 in Group D6, 17.

CHAMBON, R. (1963) L’evolution des procedes de fabri-cation manuelle de verre a vitres du dixieme siècle anos jours. In Advances in Glass Technology, Part 2,(Matson, F.R. and Rindone, G.E. eds), pp. 165–78.Plenum Press, New York.

CHANCE, J.F. (1919) A History of the Firm of ChanceBrothers and Co., Glass and Alkali Manufacturers.Privately printed, London.

CHARLESTON, R.J. (1958) Glass. In A History ofTechnology, III, From the Renaissance to the IndustrialRevolution, c.1500–c.1750 (C. Singer et al. eds), Oxford.

CHARLESTON, R.J. (1959) English glassmaking and itsspread from the XVIIth to the middle of the XIXthCentury. Ann Ier Congr. Journ. Int. Verre, AIHV, Liège,pp. 155–65.

CHARLESTON, R.J. (1960) Lead in glass. Archaeometry, 3,1–4.

CHARLESTON, R.J. (1962) Some tools of the glassmakerin medieval and renaissance times with special refer-ence to the glassmaker’s chair. Glass Technol., 3,107–11.

CHARLESTON, R.J. (1963) Glass ‘cakes’ as raw materialand articles of commerce. J. Glass Stud., 5, 54–67.

CHARLESTON, R.J. (1964) Wheel engraving and cutting;some early equipment: 1. Engraving. J. Glass Stud., 6,83–100.

CHARLESTON, R.J. (1965) Wheel engraving and cutting;

Bibliography 351

some early equipment: II Water power and cutting. J.Glass Stud., 7, 41–54.

CHARLESTON, R.J. (1967) The transport of glass in the17th to 18th centuries. Ann 4e. Congr. Journ. Int. Verre,Ravenne–Venise, pp. 183–92, AIHV, Liège, France, 1968.

CHARLESTON, R.J. (1972) Enamelling and gilding onglass. Glass Circle (R.J. Charleston, W. Evans and A.Polak, eds), pp. 18–32. Newcastle.

CHARLESTON, R.J. (1977) A brief survey of the history ofglass and glassmaking. An Illustrated Dictionary of Glass(H. Newman, ed.), pp. 9–14. Thames and Hudson,London.

CHARLESTON, R.J. (1978) Glass furnaces through theages. J. Glass Stud. 20, 9–33.

CHARLESWORTH, D. (1977) Roman window glass fromChichester, Sussex. J. Glass Stud., 19, 182.

CHASE, W.T. (1971) Egyptian blue as a pigment andceramic material. Science and Archaeology (R.H. Brill,ed.), pp. 80–91. MIT Press, Cambridge, Mass.

CHILD, D.G. (1990) World Mirrors, 1650–1990, London.CHIRNSIDE, R.C. and PROFFITT, P.M.C. (1963) The

Rothschild Lycurgus Cup; an analytical investigation. J.Glass Stud., 5, 18–23.

CHUN, F.S. FENG, S., SINGH, J.P., YUEH, F.Y., RIGSBY,J.T.III., MONTS, D.L. and COOK, R.L. (2000) Glasscomposition measurement using laser induced break-down spectrometry. Glass Technol., 41 (1), 16–21.

CLARK, D.E., PANTANO, C.G. and HENCH, L.L. (1979)Corrosion of Glass. Books for Industry, New York.

CLARKE, H.G. (1928) The Story of English Glass Pictures,1690–1810. London.

COLLONGUES, R. (1974) Sur le phénomène de corrosiondes vitraux. ENSCP Convention, 11 pp. typescript.

COLLONGUES, R. (1977) La corrosion des vitraux. LesMonuments Historiques de la France, Nr. 1, 14–16.

COLLONGUES, R. and PEREZ-Y-JORBA, M. (1973) Sur lephénomène des vitraux. Part 1 (Abstract No. 182 inCVMA News Letter, 13).

COLLONGUES, R., PEREZ-Y-JORBA, M., TILLOCA, G. andDALLAS, J-P. (1976) Nouveaux aspects du phénomènedes vitreaux anciens des églises françaises. Verres etRéfract. 30, 43–55.

COOPER, M. (1998) Laser Cleaning in Conservation.Butterworth-Heinemann, Oxford.

COPPIETERS-MOLS, J. (1999) Les peintures sur verre –propositions de materiaux et techniques de restauration.In Preprints of Triennial Meeting of ICOM-–CC, Lyon,France, pp. 787–92.

CORNING MUSEUM OF GLASS (1976) Conservation ofopus sectile panels from Kenchreai. Unpublished workreport.

CORNING MUSEUM OF GLASS (1992) Reverse Painting onGlass. The Ryser Collection: Exhibition Catalogue. NewYork: Corning Museum of Glass.

CORVAIA, C., MACLEOD, I.D. and HARLEY, C. (1996)Conservation of glass recovered from shipwreck sites.Preprints of the 11th Triennial Meeting of ICOM–CC (J.Bridgland, ed.). Edinburgh, 1, 884–90. James & James,London.

COSGROVE, M.G. (1974) The Enamels of China andJapan. Champlevé and Cloisonné. Robert Hale.

COX, G.A. and FORD, B.A. (1989) The corrosion of glasson the sea bed. J. Mater. Sci., 24, 3146–53.

COX, G.A. and FORD, B.A. (1993) The long term corro-sion of glass by ground water. J. Mater. Sci. 28,5637–47.

COX, G.A. and KHOOLI, A.R. (1992) The natural corro-sion of glass; the formation and structure of plugs. GlassTechnol., 33, 60–2.

COX, G.A., HEAVENS, O.S., NEWTON, R.G. andPOLLARD, A.M. (1979) A study of the weatheringbehaviour of medieval glass from York Minster. J. GlassStud., 21, 54–75.

COX, G.A. and POLLARD, A.M. (1977) X-ray fluorescenceanalysis of ancient glass: the importance of samplepreparation. Archaeometry, 19, 45–54.

CRAMP, R. (1968) Glass finds from the Anglo-SaxonMonastery at Monkwearmouth and Jarrow. Stud. GlassHist. Des., 16–19.

CRAMP, R. (1975) Window glass from the monastic site ofJarrow: Problems of interpretation. J. Glass Stud., 17, 88–95.

CRONYN, J.M. (1992) The Elements of ArchaeologicalConservation. Routledge, London and New York.

CRONYN, J.M., FOLEY, K. and HUNTER, K. (1987)Principles of conservation of enamels. In From Pinheadsto Hanging Bowls, the Identification, Deterioration andConservation of Applied Enamel and Glass Decorationon Archaeological Artefacts (L. Bacon and B. Knight,eds). UKIC, London.

CROSSLEY, D.W. (1967) Glassmaking in Bagot’s Park,Staffordshire, in the sixteenth century. Post-Med.Archaeol., 1, 44–83.

CROSSLEY, D.W. (1972) The performance of the glassindustry in sixteenth century England. Econ. Hist. Rev.,2nd series, 25, 421–33.

CROSSLEY, D.W. (1998) The English glassmaker and hissearch for materials in the 16th and 17th centuries. InThe Prehistory and History of Glassmaking Technology(P. McCray, ed.), Ceramics and Civilisation, VIII,167–79. American Ceramics Society, Columbus, Ohio.

CROSSLEY, D.W. (1999) Fragments from the field –aspects of sampling of glass-working residues. InCurrent Research on the History of Glass throughScientific Analysis, London.

CROSSLEY, D.W. and ABERG, F.A. (1972) Sixteenthcentury glassmaking in Yorkshire; excavations offurnaces at Hutton and Rosedale, North Riding. 1968–71.Post-Med. Archaeol., 6, 107–59.

DANDRIDGE, P. and WYPISKI, M.T. (1992) Preliminarytechnical study of medieval Limoges enamels. InMaterials Issues in Art and Archaeology, III, Pittsburgh.PA, The Materials Research Society, pp. 817–26.

DAVIDSOHN, A. and MILWIDSKY, S.M. (1978) SyntheticDetergents, 6th edn. Godwin, London and John Wiley,New York.

DAVISON, S. (1984) A review of adhesives and consoli-dants used on glass antiquities. Preprints ofContributions to the Paris Congress on Adhesives andConsolidants (N.S. Bromelle, E.M. Pye, P. Smith, P.G.Thomson, eds), pp. 191–4. IIC, London.

DAVISON, S. (1988) Cut glass chandeliers; dismantling,cleaning. recording and restoring. Preprints of thecontributions to the UKIC 30th Anniversary Conference(V. Todd, ed.), pp. 90–3. UKIC, London.

DAVISON, S. (1992) The conservation of a Louis XIV glasstable-top. Glass and Enamel Conservation, UKICOccasional Papers 11, 3–5, London.


DAVISON, S. (1998) Reversible fills for transparent andtranslucent materials. JAIC, 37, 35–47.

DAVISON, S. and JACKSON, P.R. (1985) The restorationof flat glass; four case studies. The Conservator, 9,10–13.

DAWSON, P.T., HEAVENS, O.S. and POLLARD, A.M.(1978) Glass surface analysis by Auger electronspectroscopy. J. Phys. C: Sol. Stat Phys., 11, 2183–93.

DEPASSIOT, M-C. (1997) Le Collage du verre, l’usure dutemps; la restauration des objects du patrimonie. Muséearchaelogique de Saint-Romain-en-Gal et les auteurs,pp. 108–9.

DIDEROT, D. et D’ALEMBERT, J. le ROND (1751–1771)Dictionnaire Raisonné des Sciences, des Arts, et desMetièrs. Diderot, Paris.

DODWELL, C.R. (1961) The Various Arts. Nelson, London(Translation of Theophilus’ De Diversis Artibus).

DOPPELFELD, A. (1965) Die Kölner Glasöfen vomEigelstein. Proc. 7th Int. Congr. Glass Brussels. NewYork, Paper No. 236.

DOUGLAS, R.W. and FRANK, S. (1972) A History ofGlassmaking. Foulis, Henley-on-Thames.

DOVE, S. (1981) Conservation of glass-inlaid bronzes andlead curses from Uley, Gloucestershire. The Conservator,5, 31–5.

DOWLING, J. (1999) Removal and installation of mirrorsfor the British Gallery Project. V&A Cons. J., 33, 8–9.(Victoria and Albert Museum, London.)

DOWMAN, E.A. (1970). Conservation in FieldArchaeology. Methuen, London.

DOWN, J.L. (1986) The yellowing of epoxy resinadhesives, report on high-intensity light ageing. Studiesin Conservation, 31, 159–70.

DUBIN, L.S. (1987) The History of Beads. Thames andHudson, London.

DUNCAN, G.S. (1960) Bibliography of Glass [to 1940].Dawsons, London.

EARL, N. (1999) The investigation of glass deterioration asa result of storage systems for waterlogged archaeolog-ical glass. In The Cons. of Glass and Ceramics: Research,Practice and Training. (N.H. Tennent, ed.), pp. 96–113.James and James, London.

EL-NADY, A.B.M., ZIMMER, K. and ZARY, G. (1985)Spectographical analysis of medieval glass by means ofinductively coupled plasma and glow discharge sources.Spectrochimica Acta, 40, 999–1004.

EL-SHAMY, T.M. (1973) The chemical durability ofK2O–CaO–MgO–SiO2 glasses. Phys. Chem. Glasses, 14,1–5.

EL-SHAMY, T.M., LEWINS, J. and DOUGLAS, R.W. (1972)The dependence on the pH of the decomposition ofglasses by aqueous solutions. Glass Technol., 13, 81–7.

EL-SHAMY, T.M., MORSI, S.E., TAKI-ELDIN, H.D. andAHMED, A.A. (1975) Chemical durability ofNa2O–CaO–SiO2 glasses in acid solutions. J. Non-Cryst.Solids, 19, 241–50.

EMELEUS, V.M. (1960) Beta-ray back scattering. A simplemethod for the quantitative determination of the leadoxide in glass, glaze and pottery. Archaeometry, 3, 5–9.

EMERY, M.J. (1991) Techniques in Reverse Glass Painting.Quality Printing, Pittsfield, Mass.

ENGLE, A. (1973a) 3000 years of glassmaking on thePhoenician Coast. Readings in Glass History, No. 1,1–26. (Phoenix Publications, Jerusalem.)

ENGLE, A. (1973b) A semantic approach to glass history.Readings in Glass History, No. 1, 81–4. (PhoenixPublications, Jerusalem.)

ENGLE, A. (1974) The De Gands of Ghent. Readings inGlass History, No. 4, 42–54. (Phoenix Publications,Jerusalem.)

ERNSBERGER, F.M. (1977) Mechanical properties of glass.Proc. 11th Int. Congr. Glass, Prague: Survey Papers, Vol.1, 293–321.

ERNSBERGER, F.M. (1980) The role of molecular water inthe diffusive transport of protons in glass. In Phys.Chem. Glasses, 21, 146–9.

ERRETT, R.F. (1972) The repair and restoration of glassobjects. JAIC 12, 48–9.

ERRETT, R.F., LYNN, M. and BRILL, R.H. (1984) The useof silanes on glass. In Preprints of the Contributions tothe Paris Congress on Adhesives and Consolidants (N.S.Bromelle, E.M. Pye, P. Smith and G. Thomson, eds), pp.185–90. IIC, London.

ESCRITT, J. and GREENACRE, M. (1972) Note on toxicgases in polyurethane foam. Studies in Conservation,17, 134.

ESHØJ, B. (1988) Glas: Historie, Teknologi, Opbyning,Nedbrydning, Konsiervering, Konservatorskolen. DetKongelige Danske Kunstakademi.

ESWARIN, R. (1982) Reverse Paintings on Glass: The RyserCollection. The Corning Museum of Glass, Corning, NY.

ETHRIDGE, E.C., CLARK, D.E. and HENCH, L.L. (1979)Effects of glass surface area to solution volume ratio onglass corrosion. Glass. Technol., 20, 35–40.

EVANS, W., ROSS, C., WERNER, A. (1995) WhitefriarsGlass: James Powell, and Sons, of London. Museum ofLondon.

FELLER, R.L. (1972) Problems in the investigation ofpicture varnishes. In Conservation of Painting and theGraphic Arts. Proceedings of the IIC Conference,Lisbon, pp. 201–9.

FENN, J. (1987) Deterioration of glass trade beads, or glassbeads in soapy bubbles. Proceedings of the ICOMTriennial Conference, pp. 189–91.

FERNANDEZ NAVARRO, J.M. and La IGLESIA, A. (1994)Estudio de la Coloracion Roja y Amarilla de Vidrios dela Catedral de Toledo. Boletin de la Sociedad Ceramicay Vidrio, 33, 864-8.

FERRAZZINI, J.C. (1977) Untersuchungen über eine neuein Chartres angewandte Methode zur Reinigung vonmittelelterlichen Glasgemälde. Maltechnik 83, 145–54.(Abstract in CVMA News Letter 23, item 3.2, 3 December,1976.)

FIELDS, J. (2000) Replacing Synperonic N. Cons. News, 72,3. (UKIC, London.)

FIORENTINO, P. and BORELLI, L.V. (1975) A preliminarynote on the use of fillers and adhesives in the restora-tion of glass. Stud. Cons. 20, 201–5. (IIC, London.)

FIORENTINO-RONCUZZI, I. (1967) Traditions et progrèsdans les materiaux vitraux de la mosaique. In Ann. 4e

Congr. Int. Journ. Verre, Ravenne–Venise, AIHV, Liège,pp. 230–8.

FIORENTINO-RONCUZZI, I. (1970) Dating and analysis ofmosaics. Ann. 4e Congr. Int. Journ. Verre, Ravenne–Venise, AIHV, Liège, pp. 353–6.

FISHER, P. (1988) Advances in the restoration of glass vessels.In Preprints of the Contrib. to the UKIC 30th AnniversaryConference (V. Todd, ed.), pp. 81–3. UKIC, London.

Bibliography 353

FISHER, P. (1991) The restoration of historical enamels.Les Arts du Verre, Historie, Techniques et Conservation.In Journées d’Études de la SFIIC, Nice, France, pp.181–91.

FISHER, P. (1992) HXTAL NYL-1, an epoxy resin for theconservation of glass. Glass and Enamel Conservation,UKIC Occasional Papers No. 11, pp. 6–9. UKIC, London.

FISHER, P. and NORMAN, K. (1987) A new approach tothe reconstruction of two Anglo-Saxon beakers. Stud. inCons., 32, 49–58.

FITZ, S. (1981) A new method of cleaning brownedmedieval glass. In Preprints of the 6th Triennial Meetingof ICOM–C.C., Ottawa, Canada.

FITZ, S., FITZ-ULRICH, E., FRENZEL, G., KRÜGER, R. andKUHN, H. (1984) Die Einwirkung von Luft-unvereinigungen auf ausgewälte Kunstwerke mittelalter-liche Glasmalerei. Deutsches Museum, Munich.

FLEMING, S.J. and SWANN, C.P. (1994) Roman onyx glass.A study of production recepies and colorants, usingPIXE spectrometry. Nuclear Instruments and Methods inPhysics Research B, 85, 864–8.

FLORIAN, M.L.E. (1979) Letter to R.G. Newton, dated 16February 1987.

FOLEY, K. and HUNTER, K. (1987) The Lincoln HangingBowl. In From Pinheads to Hanging Bowls; theIdentification, Deterioration and Conservation ofApplied Enamel and Glass Decoration on ArchaeologicalArtefacts, (L. Bacon and B. Knight, eds). UKIC, London.

FORBES, R.J. (1961) Glass throughout the ages. PhilipsTech. Rev., 22, 282–99.

FORBES, R.J. (1966) Studies in Ancient Technology.Leyden, The Netherlands.

FOWLER, J. (1881) On the process of decay in glass and,incidentally, on the composition and texture at differ-ent periods, and the history of its manufacture.Archaeologia, 46, 65–162.

FRANK, S. (1982) Glass and Archaeology. Academic Press,London.

FRANK, S. (1984) Gold ruby glass. In Glass Technol. 25,47–50.

FREESTONE, I.C. (1991) Looking into Glass, in Scienceand the Past (Bowman, S. ed.) pp. 37–56, BritishMuseum Publications, London.

FREESTONE, I.C. (1993) Compositions and origins ofglasses from Romanesque champlevé enamels. InStratford, N., Catalogue of Material Enamels in theBritish Museum, II Northern Romanesque Enamel, pp.37–45. BMP, London.

FREESTONE, I.C. (1999) Ancient Glass and ModernScience – Building on the legacy of W.E.S. Turner, 13thTurner Memorial Lecture, 1999. University of Sheffield.

FREESTONE, I.C. (2001) Post depositional changes inarchaeological ceramics and glass. In Handbook ofArchaeological Sciences (D.R. Brothwell and A.M.Pollard, eds). John Wiley & Sons Ltd, Chichester.

FREESTONE, I.C. and BIMSON, M. (1995) Early Venetianenamelling on glass: technology and origins. InMaterials Issues in Art and Archaeology IV, (P.B.Vindiver et al., eds), pp. 352, 415–31. Materials ResearchSoc. Symp. Proc., Pittsburgh.

FREESTONE, I.C. and GORIN-ROSEN, Y. (1999) The greatglass slab at Bet She’arim Israel. An early Islamic glass-making experiment? J. Glass Stud., 41, 105–16.

FREESTONE, I.C. and STAPLETON, G.P. (1998)

Composition and technology of Islamic enamelledGlass. In Islamic Gilded and Enamelled Glass of theMiddle East (R. Ward, ed.), pp. 122–8. BMP, London.

FREESTONE, I.C., BIMSON, M. and BUCKTON, D (1990)Compositional categories of Byzantine glass tesserae.Ann. 11e Congr. Assoc. Int. Hist. Verre, AIHV,Amsterdam, pp. 271–9.

FREESTONE, I.C., HUGHES, M.J. and STAPLETON, G.P.(1999) Compositional studies of some Saxon glass. InProceedings of Current Research on the History of Glassthrough Scientific Analysis, London.

FREMERSDORF, F. (1930) Die Herstellung der Diatreta,Schumaker Festschrift, Mainz, pp. 295–300.

FRENZEL, G. (1970) La conservation des vitreaux anciensII. Le nettoyage, 4 pp. typescript (Abstract No. 153 inNewton 1982b).

FRIEDMAN, I., SMITH, R.L. and CLARK, D. (1963)Obsidian dating. In Science in Archaeology. AComprehensive Survey of Progress and Research (D.Bothwell and E. Higgs, eds), pp. 47–58. Thames andHudson, London.

FRODL-KRAFT, E. (1967) Restaurierung und Erforschung,1, Die Südrose von Maria Strassengel. Osterr. Zeits.Kunst. Denkm. 21, 192–7.

FRODL-KRAFT, E. (1970) Konservierungsprobleme mitte-lalterlicher Glasmalerein. Ann. 5e Congr Int. Hist. Verre,Prague, pp. 357–70. AIHV, Liège.

FRODL-KRAFT, E. (1971) Restaurierung und Sicherung DieBildfenster der Wassenkirche. In Osterr. Zeits. Kunst.Denkm., 25, 70–3.

GABRIEL, J.H. (2000) The Gilbert Collection: Micromosaics.Philip Wison, London.

GAIROLA, T.R. (1960) Handbook of ChemicalConservation of Museum Objects. Department ofMuseology, Maharaja Sayajirao, University of Baroda.

GARLAKE, M. (1969) Recovery and treatment of fragileartefacts from an excavation. S. Afr. Archaeolog. Bull.,24, 61–2.

GASPARETTO, A. (1965) Les fouilles de Torcello et leurapport a l’histoire de la verrerie de la Vénétie dans leHaut Moyen Age. Proc. 7th Int.. Congr. Glass, Brussels,New York, Paper No. 239.

GASPARETTO, A. (1967) A proposito dell’ officina vetrariaTorcellana – forni e sistemi diffusione antiche. J. GlassStud. 9, 50–75.

GASPARETTO, A. (1973) Verres Venitiens du moyen age.Proc. 10th Int. Congr. Glass, Kyoto, pp. 9–21 to 9–29and 15–90.

GEILMANN, W. (1956) Beiträge zur Kenntnis alter GläserIV Die Zersetzung der Gläser im Boden. Glastech.Berlin. 29, 156–168.

GEILMANN, W. (1960) Beiträge zur Kenntnis alter GläserVI Eine eigenartige Verwitterungs-erscheinung aufRömichen Glass-scheiben. Glastech. Berlin., 33, 291–6.

GEILMANN, W. and BRUCKBAUER, T. (1954) Beiträge zurKenntnis alter Gläser. II. Der Mangangehalt alter Gläser.Glastech. Berlin, 27, 456–9.

GERMAIN-BONNE, D., BIRON, I. and TROCELLIER, P.(1996) The degradation of fifteenth- and sixteenth-century painted enamel. In Preprints of 11th TriennialMeeting of ICOM–CC, Edinburgh, Scotland, pp. 826–32.

GIBSON, P. and NEWTON, R.G. (1974) A study on clean-ing painted and enamelled glass in an ultrasonic bath.Br. Acad. CVMA Occasional Papers, Part 1, 70–8.


GILARD, P. and DUBRUL, L. (1937) The calculation of thephysical properties of glass III. The index of refraction.Trans. J. Glass Technol., 21, 476–88.

GODFREY, E.S. (1975) The Development of EnglishGlassmaking 1560–1640, Oxford.

GOLDSTEIN, S.M. (1977) Forgeries and reproductions ofancient glass in Corning. J. Glass Stud., 19, 40–62.

GOODMAN, C.H.L. (1987) A new way of looking at glass.Glass Technol., 28, 19–29.

GORIN-ROSEN, Y. (1993) Halera, Bet Eli’ezer; in EnglishEdition of Hadashot Arkheologiyot, No. 100, 42–3,Jerusalem.

GORIN-ROSEN, Y. (1997) Glass Workshop in the BetShe’an Excavation Project. 1992–1994. In (G. Mazor andR. Bar-Nathan, eds), in English Edition of HadashotArkheologiyot, No. 105 (Jerusalem), 27–9.

GRAHAM, K.W. (1976) Examination and treatment ofHinterglasmalerei or reverse paintings on glass. InConference Papers, Cooperstown Training Programme,USA, pp. 1–8.

GRATUZE, B. and BARRANDON, J-N. (1990) Islamic glassweights and stamps: analysis using nuclear techniques.Archaeometry, 32, 155–62.

GREEN, P. (1979) Tracking down the past. New Scientist,84 (1182), 624–6.

GREEN, L.R. and HART, F.A. (1987) Colour and chemicalcomposition in ancient glass: An examination of someRoman and Wealdon glass by means ofultravisible–infrared spectrometry and electron micro-probe analysis. J. Archaeol. Sci., 14, 271–82.

GROSE, D.F. (1977) Early blown glass. J. Glass Stud. 19,9–29.

GUIDO, M. (1977) The Glass Beads of the Prehistoric andRoman Periods in Britain and Ireland. London.

GÜNTHER, R. (1961a) Die Entwicklung der Glassch-meltzwanennofen. Glastechn. Berlin, 34, 417–82.

GÜNTHER, R. (1961b) Rauchgaskegel auf alten glashütten.Glastechn. Berlin, 34, 559–62.

HABEREY, W. (1963) Die Glasindustrie im RömichenRheinland. In Advances in Glass Technology, Part 2(F.R. Matson and G.E. Rindone, eds), pp. 349–58. NewYork.

HADSUND, P. (1993) The tin–mercury mirror: its manufac-turing technique and deterioration processes. Stud. inCons., 38, (1).

HALL, E.T., BANKS, M.S. and STERN, J.M. (1964) Uses ofX-ray fluorescence analysis in archaeology. Archaeo-metry, 7, 84–9.

HALL, E.T. and SCHWEIZER, F. (1973) X-ray fluorescenceanalysis of museum objects: a new instrument. 1. Anon-dispersive X-ray isoprobe. In Archaeometry, 15,53–7, 74–6.

HANCOCK, R.G., CHAFE, A. and KENYON, I. (1994)Neutron activation analysis of sixteenth and seven-teenth–century blue glass trade beads from the EasternGreat Lakes of North America. Archaeometry, 36,253–66.

HARDEN, D.B. (1936) Roman Glass from Karanis, Foundby the University of Michigan Archaeological Expeditionin Egypt (1924–1929). University of Michigan Press.

HARDEN, D.B. (1956) Glass vessels in Britain and IrelandAD 400–1000. In Dark Age Britain: Studies presented toE.T. Leeds (D.B. Harden, ed.), pp. 132–67. London.

HARDEN, D.B. (1959) New light on Roman and Early

Medieval window glass. Glastechn. Berlin, 32K, VIII/8– VIII/16.

HARDEN, D.B. (1961) Domestic window glass. Roman,Saxon, and Medieval. In Studies in Building History:Essays in Recognition of the work of B.H. St J. O’Neil(E.M. Jope, ed.). pp. 39–63. Odhams, London

HARDEN, D.B. (1963) The Rothschild Lycurgus Cup:addenda and corrigenda. J. Glass Stud., 5, 9–17.

HARDEN, D.B. (1968) Ancient glass I: Pre-Roman.Archaeolog. J,. 125, 46–72 (*).

HARDEN, D.B. (1969a) Ancient glass II: Roman.Archaeolog. J., 126, 44–77 (*).

HARDEN, D.B. (1969b) Medieval glass in the West. Proc.8th Int. Congr. Glass, pp. 97–111, Sheffield.

HARDEN, D.B. (1971) Ancient glass III: Post-Roman.Archaelog. J., 128, 78–117 (*) [The three articles marked(*) were re-issued in 1997].

HARDEN, D.B. (1973) In Excavations at Shakenoak Farm,near Wilcote, Oxfordshire, Part IV (A.C.C. Brodribb, A.R.Hands and D.R. Walker, eds), privately printed andavailable from Dr A.R. Hands, Exeter College, Oxford.

HARDEN, D.B. (1974) Window glass from the Romano-British bath-house at Garden Hill, Hartfield, Sussex.Antiquaries J., 54, 280–1.

HARDEN, D.B., PAINTER, K.S., PINDER-WILSON, R.H.and TAIT, H. (1968) Masterpieces of Glass. Trustees ofthe British Museum, London.

HARDEN, D.B. and TOYNBEE, J.M.C. (1959) TheRothschild Lycurgus Cup. Archaeologia, 97, 179–212.

HARTMANN, G., KAPPEL. I., GROTE, K. and ARNDT, B.(1997) Chemistry and technology of prehistoric glassfrom Lower Saxony and Hesse. J. Arch. Sci., 24, 547–59.

HARRINGTON, J.C. (1952) Glassmaking at Jamestown.Richmond Va.

HAWTHORNE, J.H., and SMITH, C.S. (1963) On DiversArts: the Treatise of Theophilus, 2nd edn, University ofChicago Press, 1979. [The foremost medieval treatise onpainting, glassmaking and metalwork.]

HAYNES, D.E.L. (1975) The Portland Vase, Trustees of theBritish Museum.

HEALTH & SAFETY EXECUTIVE (UK) (2000) INDG 307Hydrofluoric acid poisoning – Recommendations onFirst Aid Procedures. HSE, London.

HEDGES, R. and NEWTON, R.G. (1974) Use of the‘Isoprobe’ for studying the chemical composition ofsome 12th century glass from York Minster. Br. Acad.C.V.M.A. Occasional Papers, Part 1, 79–93.

HEDVALL, J.A., JAGITSCH, R. and OLSEN, G. (1951) Uberdas Problem der Zerstörung antiker Gläser. II.Mitteilung. Uber die Belegung von Glasoberfläschen mitSchutzfilmen. Trans. Chalmers Inst. Technol., No. 118,Göteborg.

HEJDOVÁ, J. (1965) Comments on archaeological finds inthe field of glassmaking in Czechoslovakia.Czechoslovak Glass Rev., 12, 353.

HENCH, L.L. (1975a) Characterization of glass.Characterization of Materials in Research, Ceramics,and Polymers (J.J. Burke and V.I. Weiss, eds), Ch. 8,pp. 11–51. Syracuse, NY.

HENCH, L.L. (1975b) Characterization of glass corrosionand durability. J. Non-Cryst. Sol. 19, 27–39.

HENCH, L.L. (1977) Physical chemistry of glass surfaces.In Proc. 11th Congr. Int. Comm. Glass, Prague, SurveyPapers, II, 343–69.

Bibliography 355

HENCH, L.L. (1982) Glass surfaces. J. Physique Colloque,C.9. Supplement to No. 12, 43, C9–625 to C9–636.

HENCH, L.L. and CLARK, D.E. (1978) Physical chemistryof glass surfaces. J. Non-Cryst. Sol. 28, 83–105.

HENCH, L.L. and SANDERS, D.M. (1974) Analysis of glasscorrosion. The Glass Industry, Feb. and Mar., 12, 13, 16,18, 19.

HENCH, L.L., NEWTON, R.G. and BERNSTEIN, S. (1979)Use of infrared reflection spectroscopy in analysis ofdurability of medieval glasses, with some comments onconservation procedures. Glass Technol., 20, 144–8.

HENDERSON, J. (1985) The raw materials of early glassproduction. Oxford J. Archaeology, 40, 267–91.

HENDERSON, J. (1988) Electron microprobe analysis ofmixed alkali glasses. Archaeometry, 30, 77–91.

HENDERSON, J. (1991) Chemical characterisation ofRoman glass vessels, enamels and tesserae.. In MaterialsIssues in Art and Archaeology II. Materials ResearchSociety Symposium Proceedings, 185, 601–7.

HENDERSON, J. (1993a) Technological characteristics ofRoman enamels. Jewellery Stud., 5, 65–76.

HENDERSON, J. (1993b) Aspects of early medieval glassproduction in Britain. Ann 12e Congr. Assoc. Int. Verre.AIHV, Liège, pp. 247–59.

HENDERSON, J. (1995) An investigation of early glassproduction at Raqqa, Syria. In Materials Issues in Art &Archaeology, IV, Mats. Res. Soc. Proc. 352, 433–443.

HEYWORTH, M.P., HUNTER, J.R., WARREN, S.E. andWALSH, J.N. (1989) I.C.P.S. and glass: the multi-elementapproach. In Neutron Activation and Plasma EmissionSpectrometric Analysis in Archaeometry (M.J. Hughes,M.R. Cowell and D.R. Hook, eds), British MuseumOccasional Paper 82, 143–154. BMP, London.

HOFFMAN, P. (1994) Analytical determination of colour-ing elements and of their compounds in glass beadsfrom graveyards of the Merowings’ time. In Fresenius’J. Analyt. Chem., 349, 320–33.

HOGAN, I. (1993) An improved method of makingsupportive resin fills for glass. Cons. News, 50, 29–30.UKIC, London.

HOGG, S.E.T., MCPHAIL, D.S., ROGERS, P.S. and OAKLEY,V.L. (1998) Mono-functional organo-silanes as candidatesfor treatments of crizzling glasses. In Preprints of theICOM-CC Working Group on Glass, Ceramics & RelatedMaterials (A. Paterakis, ed.), pp. 53–7. EVTEK Instituteof Arts and Design, Vantaa, Finland.

HOGG, S.E.T., MCPHAIL, D.S., OAKLEY, V.L. andROGERS, P.S. (1999) Modern instrumental methods forthe study of the deterioration of vessel glass. In TheConservation of Glass and Ceramics: Research, Practiceand Training (N.H. Tennent, ed.), pp. 42–55. James andJames, London.

HOLDEN, A. (1991) The museum approach to the conser-vation/restoration of stained glass. In Les Arts du Verre:Histoire, Technique et Conservation. Journées d’Étudesde la SFIIC, Paris, pp. 69–74.

HOLLISTER, P. (1974) The glazing of the Crystal Palace.J. Glass Stud. 16, 95–110.

HORIE, C.V. (1987) Materials for Conservation.Butterworths, London.

HORTON-JAMES, D., WALSTON, S. and ZOUNIS, S.(1990) Evaluation of the stability, appearance andperformance of resins for adhesion of flaking paint onethnographic objects. Stud. Cons., 36, 203–21.

HOSFORTH, J. and DAVISON, P. (1988) ConservingNdebele Beadwork. Am. Mus. Nat. Hist., 85–95.

HOWARTH, J.T., SYKES, R.F. and TURNER, W.E.S. (1934)A study of the fundamental reactions in the formationof soda–lime–silica glasses. Trans. J. Soc. Glass Technol.,18, 290–308.

HUDSON, A.P. and NEWTON, R.G. (1976) A means forthe in situ identification of medieval glass by the detec-tion of its natural radioactivity. Archaeometry, 18,229–32.

HUGGINS, M.I. and SUN, K-H. (1946) Calculation ofdensity and optical constants of a glass from its compo-sition in weight percentage. J. Soc. Glass Technol. 30,333–42.

HUGHES, M.J. (1987) Materials, deterioration and analy-sis. In From Pinheads to Hanging Bowls, theIdentification, Deterioration and Conservation ofApplied Enamel and Glass Decoration on ArchaeologicalArtefacts (L. Bacon and B. Knight, eds). UKIC, London.

HUME, I.N. (1976) Archaeological excavations on the siteof John Frederick Amelung’s New Bremen Glassmanufactury, 1962–1963. J. Glass Stud. 18, 138–215.

HUNTER, J.R. (1977) Glass fragments from the vicaragegarden, Brixworth. J. Brit. Arch. Assoc., 104–7.

HUNTER, J.R. (1981) The medieval glass industry. InMedieval Industry (D.W. Crossley, ed.), pp. 1430–50.CBA Research Report 40, London.

HUNTER, J.R. and HEYWORTH, M. (1999) The HamwichGlass. Council for British Archaeology, London.

HUNTLEY, D.J. and BAILEY, D.C. (1978) Obsidian sourceidentification by thermoluminescence. Archaeometry,20, 159–70.

HURST-VOSE, R. (1980) Glass. Collins, London.HUTCHINSON, M.E. (1981) An experimental method for

consolidating excavated medieval glass. Contributions tothe Seminar on Medieval Glass Technology andMedieval Glaziers, Urban Research Committee of theCouncil for British Archaeology and Corpus Medii Aevi.

IBRAHIM, L., SCRANTON, R. and BRILL, R.H. (1976) Thepanels of opus sectile in glass. Kenchrei – Eastern Partof Corinth, vol. 2. E.J. Brill, Leiden, The Netherlands.

ICCROM (1983) Conservation in situ.: mosaics. No. 3,Aquileia (Rome).

ICI (1977) Polyurethanes, General, Isocyanates, Hazardsand Safe Handling Procedures, 4th edn. ICI TechnicalInformation, U93, London.

ILIFFE, C.J.A. and NEWTON, R.G. (1976) Using triangulardiagrams to understand the behaviour of medievalglasses. Verres et Réfract. 30, 30–4.

INTERNATIONAL COMMISSION ON GLASS (1972–1979)The Chemical Durability of Glass; a BibliographicReview of the Literature. Institut National du Verre,Charleroi, 1, 1972; 2, 1973; 3, 1979.

JACKSON, C.M. and SMEDLEY, J.W. (1999) Raw materialsand glass recipes – glassmaking in Staffordshire in the14th–16th centuries. In Proceedings of Current Researchon the History of Glass through Scientific Analysis.London.

JACKSON, C.M., HUNTER, J.R., WARREN, S.E. and COOL,H.E.M. (1991) The analysis of blue–green glass andglassy waste from two Romano–British glass-workingsites. In Archaeometry ‘90, Basle (E. Pernicka and G. AWagner, eds), pp. 295–305. Birkhauser Verlag,Germany.


JACKSON, C.M., NICHOLSON, P.T. and GNEISINGER, W.(1998) Glassmaking at Tell El Amarna: an integratedapproach. J. Glass Stud., 40, 11–23.

JACKSON, L. (ed.) (1996) Whitefriars Glass: The Art ofJames Powell & Sons. Richard Dennis.

JACKSON, P.R. (1982a) A doweling technique for glassrestoration. The Conservator, 6, 33–5.

JACKSON, P.R. (1982b) Resins used in glass conservation.In Proc. Symp. Resins and Cons., 10.1–10.7, SSCR,Edinburgh University, Extra Mural Department,Edinburgh, Scotland.

JACKSON, P.R. (1983) Restoration of an Italic glassoinchoe with Technovit 4004a. The Conservator, 7, 44–7.UKIC, London.

JACKSON, P.R. (1984) Restoration of glass antiquities. InPreprints of the 7th Triennial Meeting of ICOM–CC.Copenhagen, Denmark, 84.20. 13–7.

JACKSON, P.R. (1985) Restoration of the Auldjo Jug. Ann.9e. Congr. Assoc.Hist. Verre. AIHV, Liège.

JACKSON, P.R. (1987) The Hampton Court fire – rescuingpier glasses. In: Recent Advances in the Conservationand Analysis of Artefacts. Proceedings of the Universityof London Institute of Archaeology Jubilee ConservationConference.

JACKSON, C.M., NICHOLSON, P.T. and GNEISSINGER, W.(1998) Glassmaking at Tell el-Amarna: an integratedapproach. J. Glass Stud., 40, 15–23.

JANTZEN, C.M. and PLODINEC, M.J. (1984) Thermo-dynamic model of natural, medieval, and nuclear wasteglass durability. J. Non-Cryst. Sol., 67, 207–23.

JOHNSTON, R.H. (1975) Master glassblowers of Heart,Afghanistan: some archaeological relationships. 77thGeneral Meeting of the Archeology Institute of America,22.

KARL, F. (1970) Behandlung korrodierter geschliffener Gläser.Arbeitsblätter für Restauratoren, 2, Gruppe 5, 17–19.

KEISER, H.B. (1937) Die deutsche Hinterglasmalerei.Munich.

KENYON, G.H. (1967) The Glass Industry of the Weald.Leicester University Press.

KENYON, I., HANCOCK, R.G.V. and AUFREITER, S.(1995) Neutron activation analysis of AD 1660–1930European copper-coloured blue glass trade beads fromOntario, Canada. Archaeometry, 37, 323–37.

KERR, J. (1991) The Repair and Maintenance of HistoricGlass. Church House Publishing.

KLEITZ, M.O. (1991) La peinture sur verre, technologie,degradation, problems de restauration de la couchepicturale. In Les Arts du Verre: Histoire, Technique etConservation. Journées d’Etude de la SFIIC, Paris,117–33.

KNAIPP, F. (1988) Hinterglas-Kunste. Eine Bildoku-mentation. Linz and Munich.

KNIGHT, B. (1999) Excavated window glass – a neglectedresource? In Preprints of the Contributions to theCopenhagen Congress on Archaeological Conservationand its Consequences (A. Roy and P. Smith, eds), pp.96–104. IIC, London.

KNOWLES, J.A. (1959) Decay of glass, lead and iron inancient stained glass windows. J. Brit. Soc, Master GlassPainters, 12, 70–6.

KOCK, J. and SODE, T. (1995) Glass, Glass Beads andGlassmakers in Northern India. Thot Press, Vanlose,Denmark.

KOEZUKA, T. and YAMASAKI, K. (1998) Investigation ofsome K2O–PbO–SiO2 glasses excavated in Japan. Proc.18th Int. Congr. Glass, San Francisco, CD-ROM, PaperNo. 1 in Group B3, 1–4.

KOOB, S.P. (1981) Consolidation with acrylic colloidaldispersions. In Preprints of the 9th Annual Meeting ofthe AIC, pp. 86–94. JAIC, Philadelphia and Washington.

KOOB, S.P. (1982) The instability of cellulose nitrateadhesives. The Conservator, 6, 31–4.

KOOB, S.P. (2000) New techniques for the repair andrestoration of ancient glass. In Preprints of theContributions to the Melbourne Conference on Traditionand Innovation: Advances in Conservation (Roy, A. andSmith, P. eds), pp. 92–5. IIC, London.

KOOB, S.P. BRILL, R.H. and THIMME, D. (1996) TheKenchrei opus sectile panels revisited: a comparison andassessment of previous treatments. In Preprints of theContributions to the Copenhagen Congress onArchaeological Conservation and its Consequences (A.Roy and P. Smith, eds), pp. 105–10. IIC, London.

KORDAS, G. (ed.) (2002) Hyalos–Vitrum–Glass. Pro-ceedings of the First International Conference on theHistory, Technology and Conservation of Glass andVitreous Materials in the Hellenistic World. Rhodes, 2001.

KRAWCZYK-BÄRSCH, E., DÄBRITZ, S. and HAUFFE, W.(1998) The characterization of the weathering processin medieval glass by scanning electron microscopy incombination with ion beam slope cutting. Proc. 18thInt. Congr. Glass, San Francisco, CD-ROM, Paper No. 11in Group B1, 61–6.

KUNCKEL, VON LOEWENSTEIN (1679) Ars VetrariiExperimentalis. Leipzig, Germany.

LABINO, D. (1966) The Egyptian sand core technique: anew interpretation. J. Glass Stud., 8, 124.

LAL, B.B. (1962–63) Chemical preservation of ancientglass. Ancient India, 18–19, 230–80.

LAMBERT, J.B., MCLAUGHLIN, C.D. and CARTER, G.F.(1978) Analysis of early Egyptian glass by atomicabsorption and X-ray photoelectron spectroscopy.Archaeological Chemistry. 2. Advances in ChemistrySeries. 171. American Chemical Society, New York.

LAMM, C.J. (1941) Oriental Glass of Medieval Date Foundin Sweden and the Early History of Lustre Painting.Kungl. Vitterhets Hist. och Antik. Akad. Handlingar, del50:1, Stockholm, Sweden.

LANFORD, W.A. (1977) Glass hydration: a method ofdating glass objects. Science, 166, 975–6.

LANFORD, W.A. (1986) Ion beam analysis of glasssurfaces: dating, authentication and conservation. Nucl.Inst. Meth., B14, 123–6.

LANG, J. and PRICE, J. (1975) Iron tubes from a lateRoman glassmaking site at Mérida (Badajoz), Spain. J.Archaeol. Sci., 2, 289–96.

LANMON, D.P. and PALMER, A.M. (1976) John FrederickAmelung and the New Bremen Glasmanufactury. J.Glass Stud. 18, 14–137.

LANZ, H. and SEELIG, L. (ed.) (1999) Farbig KostbarkeitenAus Glas, Kabinettstucke der Zurcher Hinterglasmalerei,1600–1650. Exhibition catalogue, Bayerisches National-museum, Munchen and Schwizerisches Landesmuseum,Zurich.

LANZ, H. and SEELIG, L. (2000) Glanzlichter. Die Kunstder Hinterglasmalerei. Exhibition catalogue. MuseéSuisse du Vitrail Romont/Museum in der Burg Zug.

Bibliography 357

LARNEY, J. (1975) Restoring Ceramics, pp. 100–2. Barrieand Jenkins, London.

LAVOISIER, A.J. (1770) Action of water on glass. Mem.Acad. Sci. (Paris), 73, 90.

LEE, L., SEDDON, G. and STEPHENS, F. (1982) StainedGlass. Mitchell Beazley, London.

LEWIS, G.D. (1973) The South Yorkshire Glass Industry,2nd edn. Sheffield City Museum.

LÎBIETE, J. (1998) Conservation of the Daugmale CastleMound Beads. In Preprints of the Interim Meetings of theICOM-CC. Working Group Glass, Ceramics and RelatedMaterials (A. Paterakis, ed.). EVTEK Institute of Arts andDesign, Vantaa, Finland.

LIEFKES, R. (1997) Glass. V&A Publications, London.LILLIE, H.R. (1936) Stress release in glass, a phenomenon

involving viscosity as a variable with time. J. Am.Ceram. Soc., 19, 45–54.

LILYQUIST, C. and BRILL, R.H. (1993) Glassy materials,glass and lead isotope analyses of some objects fromEgypt and the Near East. In Studies in Early EgyptianGlass. The Metropolitan Museum of Art, New York.

LINDSEY, H. (2000) Protective packaging: an introductionto the materials used to produce archival quality boxes,folders, sleeves, and envelopes. Journal of the Society ofArchivists 21, (1) 87–103.

LONGWORTH, G., TENNENT, N.H., TRICKER, M.J. andVISHNAVA, P.P. (1982) Iron-57 Moessbauer spectralstudies of medieval stained glass. J. Archaeol. Sci., 9,261–73.

LOUBENGAGER, A.W. (1931) Weathering of glass. J. Am.Ceram. Soc., 14, 833–6.

LOUGHEED, S. and SHAW, J. (1986) Deteriorating glassbeads on ethnographic objects. In The Care andPreservation of Ethnographic Materials. CCI, Canada,pp. 109–13.

MCCRAY, W.P., OSBORNE, Z.A. and KINGERY, W.D. (1995)The technology of Venetian girasole glass. In MaterialIssues in Art and Archaeology, II (Vandiver, P.B. et al.,eds). Pittsburgh, PA, The Materials Research Society.

MCGOVERN, P.E., FLEMING, S.J. and SWANN, C.P. (1991)The beads from tomb B10aB27 at Dinkha Tepe and thebeginning of glassmaking in the ancient Near East. Am.J. Archaeol. 95, 395–402.

MCGRATH, R., FROST, A.C. and BECKETT, H.E. (1961)Glass in Architecture and MACLEHOSE, L.S. (trans.)(1907) Vasari on Technique – Georgio Vasari, pp.253–7. Dover Inc., New York.

MACLEOD, C. (1987) Accident or design? GeorgeRavenscroft’s patent and the invention of lead-crystalglass. Technol. Cult., 28, 776–803.

MACLEOD, I.D. and BENG, E.H. (1998) Analysis and inter-pretation of glass artefacts from 17th–19th century Dutchand colonial Australian shipwrecks. In Proceedings ofthe ICOM–CC Working Group, Ceramics and RelatedMaterials (A. Paterakis, ed.). EVTEK Institute of Arts andDesign, Vantaa, Finland.

MACLEOD, I.D. and DAVIES, J.A. (1987) Deterioration ofglass, stone and ceramics. In Preprints of the ICOM–CC8th Triennial Meeting (J. Bridgland, ed.), pp. 1003–7.GCI, Los Angeles.

Decoration. Architectural Press, London.MADSEN, H.B. (1972) A new product for mending glass.

Stud. Glass Conserv., 17, 131–2.MAJEWSKI, E.J. (1973) Conservation of archaeological

material at Sardis, Turkey. Bull. Am. Inst. Cons., 13,99–104.

MANSFIELD, C.R., LAUGHLIN, A.W., CREMERS, D.A. andFERRIS, M.J. (1998) Real-time analysis of glass and glassraw materials using LIBS. Proc. 18th Int. Congr. Glass,San Francisco, CD-ROM, Paper No. 1 in Group A3, 1–9.

MÅNSSON, P. (circa 1520) Glaskonst. Modern translationby R. Geete, Stockholm, 1913–15.

MARCHINI, G. (1972) L’afflaiblissement de la grisaille, delong des plombs Compt. Rend. 8e Colloque, York,CVMA, 46.

MARSCHNER, H. (1977) Zur Lochfrass corrosion anMittelalterlichen Fenstergläsern. DGG Meeting inFrankfurt (Abstract No. 289 in CVMA News Letter 26).

MARTIN, J.H. (ed.) (1977) The Corning Flood: MuseumUnderwater. Corning Museum of Glass, Corning, NY.

MARTIN, M.G. (1985) Els vitralls cloisonné de Barcelona.Museum of Modern Art, Barcelona.

MARYON, H. (1971) Metalwork and Enamelling: APractical Treatise on Gold and Silversmiths’ Work andtheir Allied Crafts, 5th rev. edn. Dover Inc., New York.

MASS, J.L. (1999) Instrumental methods of analysis appliedto the Conservation of Ancient and Historic Glass. InThe Conservation of Glass and Ceramics: Research,Practice and Training. (N.H. Tennent, ed.), pp. 15–41.James and James, London.

MASS, J.L., STONE, R.E. and WYPYSKI, M.T. (1998) Themineralogical and metallurgical origins of Romanopaque coloured glasses. In The Prehistory and Historyof Glassmaking Technology (P. McCray, ed.), Ceramicsand Civilisation, VIII, 121–44. American CeramicsSociety, Columbus, Ohio.

MEHLMAN, F. (1982) The Phaidon Guide to Glass.Phaidon Press, London.

MEIGH, E. (1972) The Story of the Glass Bottle. C.E.Ramsden and Co., Stoke-on-Trent.

MELUCCO, A. (1971) Vetro-materiali vitrei. Problemi diconservazioni, (Urbain, ed) Sec. IV. Uff. de Min. per ilcoordinamento della ricerca scientifica e technologica,Bologna.

MERRETT, C. (1662) The Art of Glass (translation ofAntonio Neri, L’Arte Vetraria), London.

MERRIFIELD, M.P. (1849) Original Treatises on the Arts ofPainting, 2 vols. John Murray, London. (Reprinted byDover Inc., New York, 1966.)

MESSINGER II, J.M. and LANSBURY, P.T. (1989) Controllingthe refractive index of epoxy resins. JAIC, 28, 127–36.

MIRTI, P., CASOLI, A. and APPOLONIA, L. (1993)Scientific analysis of Roman glass from Augusta Pretoria.Archaeometry, 35, 225–40.

MOHYUDDIN, I. and DOUGLAS, R.W. (1960) Some obser-vations on the anelasticity of glasses. Phys. Chem.Glasses, 1 (3), 7–86.

MONCRIEFF, A. (1971) Polyurethane foaming resins. Stud.Cons., 16, 119.

MONCRIEFF, A. (1975) Problems and potentialities in theconservation of vitreous materials, In Preprints of theContributions to the Stockholm Congress: Conservationin Archaeology and the Applied Art, (D. Leigh, A.Moncrieff, W.A. Oddy and P. Pratt, eds), pp. 99–104.IIC, London.

MONEY, J.H. (1976) Fourth Interim Report on Excavationsin the Iron Age Hill Fort and Romano-British Iron-WorkingSettlement at Garden Hill, Hartfield, Sussex, Fig. 8.


MOODY, B.E. (1988) The life of George Ravenscroft. GlassTechnol., 29, 198–209.

MOOREY, P.R.S. (1994) Ancient Mesopotamian Materialsand Industries: The Archaeological Evidence. OxfordUniversity Press, Oxford.

MORGÓS, A., NAGY, L. and PALOSSY, L. (1984) Newsilicone rubber mould-making materials. The addition-type silicone rubbers. In Preprints of the 7th TriennialMeeting of the ICOM-CC. Copenhagen, pp. 84.20.19–84.20.20.

MORTIMER, C. (1993) Analyses of glass and glassworkingwaste from the collections of the Guildford MuseumAncient Monuments Laboratory Report, 106/93, EnglishHeritage, London.

MORTIMER, C. (1995) The glass working evidence fromOld Broad Street, London. In (D.R. Hook et al., ed.)Trade and Discovery. British Museum Occasional Paper109. BMP, London.

MORTIMER, M.M. (2000) The English Glass Chandelier.The Antique Collector’s Club.

MURGATROYD, J.B. (1942) The significance of surfacemarks on fractured glass. J. Glass Technol., 26, 155–71.

NAVARRO, J. (1999) Moving displays: four methods ofmounting ceramics and glass objects in drawers. InPreprints of the 12th Triennial Meeting of ICOM–CC,Lyons (J. Bridgland, ed.), pp. 793–8. EVTEK Institute ofArts and Design, Vantaa, Finland.

NERI, A. (1612) Dell ’Arte Vetraria (translated byChristopher Merrett, The Art of Glass). London.

NEWMAN, H. (1977) An Illustrated Dictionary of Glass.Thames and Hudson, London.

NEWTON, R.G. (1966) Some problems in the dating ofancient glass by counting the layers in the weatheringcrust. Glass Technol., 7, 22–5.

NEWTON, R.G. (1969) Some further observations on theweathering crusts on ancient glass. Glass Technol., 10,40–2.

NEWTON, R.G. (1970) Metallic gold and ruby glass. J.Glass Stud. 12, 165–70.

NEWTON, R.G. (1971a) The enigma of the layered crustson some weathered glasses, a chronological account ofthe investigations. Archaeometry, 13, 1–9.

NEWTON, R.G. (1971b) A preliminary examination of thesuggestion that pieces of strongly coloured glass werearticles of trade in Iron Age Britain. Archaeometry, 13,11–16.

NEWTON, R.G. (1971c) Glass trade routes in the Iron Age?In Proceedings of the 9th International Congress onGlass, Versailles. Art. Hist. Comm., pp. 197–204.

NEWTON, R.G. (1972) Stereoscan views of weatheringlayers on a piece of ancient glass. Glass Technol, 13, 54–6.

NEWTON, R.G. (1974a) Recovery of lost or faded decora-tion on painted glass. Brit. Acad. CVMA OccasionalPapers, Part 1, pp. 68–9, 87–9.

NEWTON, R.G. (1975) Arty Facts (Presidential Address).Glass Technol., 16, 103–6.

NEWTON, R.G. (1976) The effects of medieval glass paint.Stained Glass, 71, 226–30.

NEWTON, R.G. (1978) Colouring agents used by medievalglassmakers. Glass Technol., 19, 59–60.

NEWTON, R.G. (1979) Sulphur dioxide and medievalstained glass. In Sulphur Emissions and theEnvironment, pp. 311–13, 500–1. London, Society ofChemical Industry, Water and Environmental Group.

NEWTON, R.G. (1982a) Unusual effects of the weatheringof ancient glass. Crown in Glory (P. Moore, ed.), pp.73–80. Jarrold.

NEWTON, R.G. (1982b) The Deterioration and Conserva-tion of Painted Glass. A Critical Bibliography. BritishAcademy and Oxford University Press, OccasionalPapers, II.

NEWTON, R.G. (1985a) The durability of glass: a review.Glass Technol., 26, 21–38.

NEWTON, R.G. (1985b) W.E.S. Turner: recollections anddevelopments [The Eighth Turner Memorial Lecture]. InGlass Technol., 26, 93–103.

NEWTON, R.G. (1987) Air pollution damage. Profess.Stained Glass, 7, 14, 15, 17, 19, 21.

NEWTON, R.G. (1988) Who invented covered pots? GlassTechnol., 29, 49–50.

NEWTON, R.G. (1992) The future of the Ballidon glassburial experiment. Glass Technol., 33 (5), 179–80.

NEWTON, R.G. (1996) Fact or fiction: can cold glass flowunder its own weight, and what happens to stainedglass windows? Glass Technol., 37 (4), 143.

NEWTON, R.G. and BETTEMBOURG, J-M. (1976) Effectsof repeated condensation. CVMA News Letter, 21, 9–21.

NEWTON, R.G. and BRILL, R.H. (1983) A ‘weeping’ glassbowl at the Ashmolean Museum. J. Glass Stud., 23,93–9.

NEWTON, R.G. and DAVISON, S. (1989, 1996) Conserva-tion. of Glass. Butterworth-Heinemann, Oxford.

NEWTON, R.G. and FUCHS, D. (1988) Chemical analysesand weathering of some medieval glass from YorkMinster. Glass Technol., 29, 43–8.

NEWTON, R.G., HOLLOWAY, D.G. and HENCH, L.L(1981) A note on the ‘spontaneous cracking’ of ancientglass. Ann. 8e Congr. Assoc. Verre, London andLiverpool, pp. 355–67, 385. AIHV, Liège.

NEWTON, R.G. and RENFREW, C. (1970) British faiencebeads reconsidered. In Problems in European Prehistory(C. Renfrew, ed.), pp. 293–303. Edinburgh UniversityPress, Edinburgh.

NEWTON, R.G. and SEDDON, A.B. (1992) The durabilityof a silicate glass in the presence of a saturatedleachant. Corrosion Science, 23, 617–26.

NEWTON, R.G. and SEDDON, A.B. (1999) Organiccoatings for medieval glasses. In The Conservation ofGlass and Ceramics: Research, Practice and Training(N.H. Tennent, ed.), pp. 66–71. James and James,London.

NEWTON, R.G. and SHAW, G. (1988) Another unsolvedproblem concerning weathering layers. Glass Technol.,29, 78–9.

NEWTON, R.G. and WERNER, A.E. (1974) Definition ofthe term ‘devitrification’. Brit Acad. CVMA OccasionalPapers, Part I, No. 100.

NICHOLSON, P.T. (1993) Egyptian Faience and Glass.Shire Egyptology Series, 19. Shire Publications, PrincesRisborough, Bucks.

NICHOLSON, P.T. (1995) Glassmaking and glassworkingat Armarna: some new work. J. Glass Stud. 37, 11–19.

NICHOLSON, P.T. (1997) Early glass and glazing in Egypt:new excavations at Tell el Amarna. ArchaeologicalSciences, 1995 (Slater, Slater and Gowlett, eds). OxbowMonograph 64, Oxford.

NICHOLSON, P.T. and JACKSON, C.M. (1998) Kind ofblue. Glass of the Amarna period. In The Prehistory and

Bibliography 359

History of Glassmaking Technology: (P. McCray, ed.),Ceramics and Civilisation, VIII, 105–20. AmericanCeramics Society, Columbus, Ohio.

NICHOLSON, P.T., JACKSON, C.M. and TROTT, K.M.(1997) The Ulu Burun glass ingots, cylindrical vesselsand Egyptian glass. J. Egypt. Archaeol., 83, 143–53.

NIGHTINGALE, G. (1998) Glass and the Mycenaeanpalaces of the Aegean. In The Prehistory and History ofGlassmaking Technology (P. McCray, ed.), Ceramicsand Civilisation, VIII, 205–26. American CeramicsSociety, Columbus, Ohio.

NISHIMURA, S. (1971) Fission track dating of archaeolog-ical materials from Japan. Nature, 230, 242–3.

NOBLE, J.V. (1969) The technique of Egyptian faience.Am. J. Arch., 73, 435–9.

NOONAN, T., KOVALEV, R. and SHERMAN, H. (1998) Thedevelopment and diffusion of glassmaking in pre-Mongol Russia. In The Prehistory and History ofGlassmaking Technology (P. McCray, ed.), Ceramicsand Civilisation, VIII, 293–314. American CeramicsSociety, Columbus, Ohio.

NORTHWOOD, J. (1924) Noteworthy productions of theglass craftsman’s art. I. The reproduction of the PortlandVase. J. Soc. Glass Technol., 8, 82–92.

NOVIS, W.E. (1975) The lifting of mosaic pavements. InContributions to the Stockholm Congress onConservation in Archaeology and the Applied Arts (D.Leigh, A. Moncrieff, W.A. Oddy and P. Pratt, eds), pp.143–6. Stockholm, Sweden.

NYLÉN, E. (1975) Documentation and presentation.Fornvännen, 70, 213–23.

OAKLEY, V. (1999) Five years on: A re-assessment ofaspects involved in the conservation of glass objects fora new gallery at the Victoria and Albert Museum. In TheConservation of Glass and Ceramics: Research, Practiceand Training (N.H. Tennent, ed.), pp. 217–28. Jamesand James, London.

ODDY, W.A. and CARROLL, S. (1999) Reversibility – doesit exist? British Museum Occasional Paper No. 135.

OLIN, J.S., THOMPSON, B.A. and SAYRE, E.V. (1972)Characterisation of medieval window glass by neutronactivation analysis. In Developments in AppliedSpectroscopy, 10, pp. 33–55. Plenum Press, New York.

OLIVE, J and PEARSON, C. (1975) The conservation ofceramics from marine archaeological sites. In Preprintsof the Contributions to the Stockholm Congress onConservation in Archaeology and the Applied Arts (D.Leigh, A. Moncrieff, W.A. Oddy and P. Pratt, eds), pp.63–6. IIC, London.

OLIVER, R. (1999) The Stupendous Chandelier. V.&A.Cons J. (Victoria and Albert Museum, London).

OLSEN, D.A., JOHNSON, R.E., KIVEL, J. and ALBERS, F.C.(1969) Kinetics of leaching of lead glass by ethylene-diamine tetra-acetic acid. J. Am, Ceram. Soc., 52,318–22.

OPPENHEIM, A.L. (1973) A note on research inMesopotamian glass. J. Glass Stud. 15, 9–11. OPPEN-HEIM, A.L., BRILL, R.H., BARAG, D. and VON SALDERN,A. (1970) Glass and Glassmaking in AncientMesopotamia. Corning Museum of Glass, Corning, NewYork.

PAGE, J-A. (1998) Crizzling: a curatorial point of view.Proc. 18th Int. Congr. Glass, San Francisco, CD-ROM,Paper No. 2 in Group B2, 7–13.

PANNELL, J. (1990) Conservation of glass at BodrumMuseum of Underwater Archaeology. In Proceedings ofthe 1st Anatolian Glass Symposium, 1988 (TSCFAS,Information and Document Center, ed.) Istanbul,Turkey.

PANTANO, C.G., ROSSI-WILCOX, S. and LANGE, D.(1998) The Glass Flowers. In The Prehistory and Historyof Glassmaking Technology (P. McCray, ed.), Ceramicsand Civilisation, VIII, 761–78. American CeramicsSociety, Columbus, Ohio.

PAPE, T. (1933) An Elizabethan glass furnace. TheConnoisseur, 172 ff.

PETTANATI, S. (1978) I vetri graffiti et I vetri dipinti.Musie Civico di Torino, Turin.

PAUL, A. (1977) Chemical durability of glasses: a thermo-dynamic approach. J. Mater. Sci,. 12, 2246–68.

PAUL, A. (1978) Influence of complexing agents andnature of the buffer solution on the chemical durabilityof glass. Part 1. Theoretical discussion. Glass Technol.,19, 162–5.

PAUL, A. and YOUSSEFI, A. (1978) Influence of complex-ing agents, and nature of the buffer solution, on thechemical durability of glass. Part 2. EDTA, ethyl alcoholand sugar in the leach solution. Glass Technol., 19,166–70.

PAZAUREK, G.E. (1903) Kranke Gläser. Reichenberg.PEARSON, C. (1975) On-site conservation requirements for

marine archaeological excavations. In Proceedings of theICOM Sessions on Conservation, Venice, 75/13/2.

PEARSON, C. (1987) Conservation of Marine Archaeo-logical Objects. Butterworth-Heinemann, Oxford.

PELOUZE, T.J. (1867) Sur le verre. CR. Acad. Sci. (Paris),64, 53–66.

PELTENBURG, E.J. (1987) Early faience: recent studies,origins and relations with glass. In Early VitreousMaterials (M. Bimson and I.C. Freestone, eds), BritishMuseum Occasional Paper No. 56, 5–29.

PEREZ-ARRANTEGUI, J., CEPRIÁ, G., ABADÍAS, O.,USÓN, A., ARANDA, C. and CASTILLO, J.R. (2000)Voltammetry of immobilised microparticles: a non-destructive tool for materials characterisation. In The32nd International Archaeometry Symposium, AbstractNo. 299. Mexico City.

PEREZ-Y-JORBA, M., DALLAS, J-P., BAUER, C., BAHEZRE,C. and MARTIN, J.C. (1980) Deterioration of stainedglass by atmospheric corrosion and micro-organisms. J.Mater. Sci., 15, 1640–7.

PEREZ-Y-JORBA, M., DALLAS, J-P., COLLONGUES, R.,BAHEZRE, C. and MARTIN, J.C. (1978) Etude de l’alter-ation des vitraux anciens par microscopie électroniqueà bayalage et microsonde. Silicates Industrielles, 43,89–99.

PEREZ-Y-JORBA, M., TILLOCA, G., MICHEL, D. andDALLAS, J-P. (1975) Quelques aspects du phénomènede corrosion des vitraux anciens des èglises françaises.Verres et Réfract., 29, 53–63.

PERROT, P.N. (1971) A ‘tank’ furnace at Somelaria (Es-Samariya), Israel. Proc. 9th Int.. Congr. Glass, Art. Hist.Comm., pp. 51–61.

PETERMANN, R. (1969) Nachbildung antiker Gläser.Arbeitsblätter für Restauratoren, 1 (18), 9–14.

PETRICIOLI, S. (1973) The Gnalic Wreck: the glass. J.Glass Stud., 15, 85.

PETRIE, W.M.F. (1894) Tell el-Amarna. Methuen, London.


PEYCHES, I. (1952) The viscous flow of glass at lowtemperatures. Trans. J. Soc. Glass Technol., 36, 164–80.

PIERCY, R.C.M. (1978) Mombassa wreck excavation,Second Preliminary Report. Int. J. Naut. Archaeol.Underwater Expl., 7 (4), 310–19.

PILOSI, L. and WYPISKI, M.Y. (1998) Two ancient glasseswith modern decoration in the Metropolitan Museum ofArt. Proc. ICOM–CC, Working Group, Ceramics andRelated Materials (A. Paterakis, ed.). EVTEK Institute ofArts & Design, Vantaa, Finland.

PINDER-WILSON, R.H. (1968) Pre-Islamic, Persian andMesopotamian, Islamic and Chinese [glasses]. InMasterpieces of Glass (D.B. Harden et al. eds). Trusteesof the British Museum, London.

PLENDERLEITH, H.J. and WERNER, A.E.A. (1976)Conservation of Antiquities and Works of Art, 2nd edn.Oxford University Press, Oxford.

PLINY (AD 77) Natural History, 10, Book xxxvi (translatedby Eicholz (1962) Loeb Classical Library, p. 190, alsoWilliam Heinemann, London, p.65).

POLLARD A.M. (1979) X-ray fluorescence and surfacestudies of glass with application to the durability ofmedieval window glass. Unpublished PhD thesis,University of York.

PRICE, J. (1992) Retrieving the larger structure: ideas andcase studies in lifting technology. In Retrieval of Objectsfrom Archaeological Sites (R. Payton, ed.). ArchetypePublications, London.

QUINTON, J.C. (1998) The conservation of an 18thcentury pier glass mirror at Hampton Court Palace. InProceedings.of the ICOM–CC Working Group on Glass,Ceramics and Related Materials (A. Paterakis, ed.).EVTEK Institute of Arts and Design, Vantaa, Finland.

RAUCH, F. (1985) Applications of ion beam analysis to solid state reactions. Nucl. Inst. Meth., B10/11,746–50.

RAW, F. (1955) The long-continued action of water onwindow-glass: weathering of the medieval glass atWeoley Castle, Birmingham. Trans. J. Soc. GlassTechnol, 39, 128–33.

REILLY, J.A. and MORTIMER, M.M. (1998) The care andconservation of glass chandeliers. JAIC, 37, 149–72.

REISMAN, S.N. and LUCAS, D. (1978) Recommendationsfor the treatment of glass objects retrieved from anunderwater environment (unpublished report).

REUREN, T., PUSCH, E. and HEROLD, A. (1998) Glasscolouring works within a copper-centred industrialcomplex in Late Bronze Age Egypt. In The Prehistoryand History of Glassmaking Technology, (P. McCray,ed.), Ceramics and Civilization, III, 227–50. AmericanCeramics Society, Columbus, Ohio.

RICHEY, W.D. (1975) Chelating agents – a review. InPreprints of the Contributions to the Stockholm Congresson Conservation. in Archaeology and the Applied Arts(D. Leigh, A. Moncrieff, W.A. Oddy and P. Pratt, eds),pp. 229–343. IIC, London.

RICHTER, R.W. (1998) Crizzling in enamels – a museumslaboratory study of the relevant factors and phenomena.Proc. 18th Int. Congr. Glass, San Francisco, CD-ROM,Paper No. 1 in Group B1, p. 2.

RITZ, G.M. (1972) Hinterglasmalerei, Geschichte,Erscheinung, Technik. Georg D.W. Callwey, Munich.

ROBERTS, J.D. (1984) Acrylic colloidal suspensions as pre-consolidants for waterlogged archaeological glass. In

Preprints of the 7th Triennial Meeting of the ICOM–CC,Copenhagen, 84, 20–1.

ROGERS, J. (1980) The approach to the restoration of theMusic Room, Brighton Pavilion following arson in 1975.The Conservator, 4.

ROGERS, P., MCPHAIL, D.S., RYAN, J. and OAKLEY, V.L.(1993) A quantitative study of decay processes ofVenetian glass in a museum environment. GlassTechnol., 34, 67–8.

RÖMICH, H. (1999a) Historic glass and its interaction withthe environment. In The Conservation of Glass andCeramics: Research, Practice and Training (N.H.Tennent, ed.), pp. 5–14. James and James, London.

RÖMICH, H. (1999b) Laboratory experiments to simulatecorrosion on stained glass windows. In TheConservation of Glass and Ceramics: Research, Practiceand Training. (N.H. Tennent, ed.), pp. 57–65. Jamesand James, London.

RÖMICH, H., AERTS, A., JANSSENS, K. and ADAMS, F.(1998) Simulation of corrosion phenomena of glassobjects on model glasses. Proc. 18th Int.. Congr. Glass,San Francisco, CD-ROM, Paper No. 7 in Group B1,36–41.

ROOKSBY, H.P. (1959) An investigation of ancient opalglasses, with special reference to the Portland Vase.Trans. J. Soc. Glass Technol., 4, 285–8.

ROOKSBY, H.P. (1962) Opacifiers in opal glasses through-out the ages. Gen. Elec. Co. Sci. J., 29, 20–6.

ROOKSBY, H.P. (1964) A yellow cubic lead tin oxideopacifier in ancient glasses. Phys. Chem. Glasses, 5, 20–5.

ROOSMA, M. (1969) The glass industry of Estonia in the18th and 19th centuries. J. Glass Stud., 11, 70–2.

RYAN, J. (1995) Chemical stabilization of weathered glasssurfaces. A new approach to glass conservation. V.&A.Cons. J., 16, 6–9.

RYAN, J., MCPHAIL, D., ROGERS, P. and OAKLEY, V.(1993) Glass deterioration in the museum environment.Chemistry and Industry, 13, 498–501.

RYAN, J., MCPHAIL, D., ROGERS, P. and OAKLEY, V.(1996) Glass deterioration in the museum environment.A study of the mechanisms of decay using secondaryion mass spectrometry. In Preprints of the 11thICOM–CC. Triennial Meeting, Edinburgh, pp. 839–44.James and James, London.

RYSER, F. (1991) Verzauberte Bilder. Die Kunst derMalerei hinter Glas von der Antike bis zum 18Jahrhundert. Verlag Kilinkhardt and Bierman, Munich.

RYSER, F. and SALMEN, B. (eds) (1995) Amelierte Stuckuff Glas/Hinder Glas Gemäld. Historien und Gemald.Hinterglaskunst von der Antike bis zur Neuzeit.Exhibition catalogue. Schloßmuseum Murnau.

SALEM, A.A., GRASSERBAUER, M. and SHREINER, M.(1994) Study of the corrosion processes in glass fromScandinavia and Britain, PACT. Journal of the EuropeanStudy Group on the Physical, Chemical andMathematical Techniques Applied to Archaeology, 7,401–11.

SALMEN, B. (ed.) (1997) Glas Glanz Farbe. Viefaltbarocker Hinterglaskunst im Europa des 17 und 18Jahrhunderts. Exhibition Catalogue. SchloßmuseumMurnau.

SANDERS, D.M. and HENCH, L.L. (1973) Surface rough-ness and glass corrosion. Ceram. Bull., 52, 666–9.

SAYRE, E.V. (1963) The intentional use of antimony and

Bibliography 361

manganese in ancient glasses. In Advances in GlassTechnology, Part 2 (F.R. Matson and G.E. Rindone, eds),pp. 263–82. New York.

SAYRE, E.V. (1965) Summary of the Brookhaven programof analysis of ancient glass. Application of Science inExamination of Works of Art. 145–54, Seminar, Boston.

SAYRE, E.V. and SMITH, R.. (1961) Compositionalcategories of ancient glass. Science, 133, 1824–6 (No.3467).

SAYRE, E.V. and SMITH, R.W. (1974) Analytical studies ofancient Egyptian glass. In Recent Advances in Scienceand Technology of Materials, 3 (A. Bishay, ed.), pp.47–70. Plenum Press, New York.

SCHACK VON WITTENAU, C. (1998) ‘Sick’ glasses – a casefor the Veste Coburg. Proc. 18th Congr. Glass, SanFrancisco (CD-ROM).

SCHÄFER, F.W. (1968) Two pragmatic views on ‘VasaDietrata’. J. Glass Stud., 10, 176–7.

SCHÄFER, F.W. (1969) Wiederherstellung einesDiatrietetglases in Römicher Schlifftechnik. Studies inGlass History and Design, pp. 125–6. Sheffield.

SCHOFIELD, P.F., CRESSEY, G., HOWARD, P.W. andHENDERSON, C.N.B. (1995) Origin of colour in ironand manganese containing glasses investigated bysynchrotron radiation. Glass Technol., 36, 89–94.

SCHOLES, S.R. (1929) Density factors for soda limeglasses. J. Am. Ceram. Soc., 12, 753–5.

SCHOLZE, H. (1977) Glas, Natur, Structur undEigenschaften, 2nd edn. Springer Verlag, Berlin.

SCHOLZE, H. (1978) Characterisierung ‘kranker’ gläser.Contribution to the Projektgruppe ‘Glas’ Meeting inBonn (Abstract No. 336, in CVMA News Letter, 28).

SCHOLZE, H., HELMREICH, D. and BAKARDJIEV, I.(1975) Untersuchungen über des Verhalten vonKalk–Natron Gläsern in verdunten Säuren. Glastech.Berlin, 48, 237–46.

SCHOTT, F.L. (1998a) Zur Festigung der WurzburgerTischplatten und anderen Hinterglasmalerein. In Ausst.Kat. Gemalt hinter Glas, Bayerisches National Museum,Munchen. Munchen u. Wurzburg.

SCHOTT, F.L. (1998b) Eglomise – Technik und Konser-vervierung. Restauro, 9–17. Callweg-Verlag, Munich.

SCHOTT, F.L. (1999) Restaurierung von drei Hinter-glasmalereigefassen. In Ausst. Kat. Gemalt hinter Glas,Bayerisches Nationalmuseum, Munchen. Munchen u.Wurtzburg.

SCHREINER, M. (1987) Analytical investigations ofmedieval glass paintings. In Recent Advances in theConservation and Analysis of Artefacts (Proceedings ofthe Institute of Archaeology Jubilee ConservationConference, pp. 73–8. London.

SCHREINER, M.R., PROHASKA, I., RENDL, J. and WEIGEL,G. (1999) Leaching studies of potash–lime–silica glasswith medieval compositions. In The Conservation ofGlass and Ceramics: Research, Practice and Training(N.H. Tennent, ed.), pp. 72–83. James and James,London.

SCHREINER, M.R., STINGEDER, G. and GRASSERBAUER,M. (1984) Quantitative characterization of the surfacelayers on corroded medieval glass with SIMS. Fresnius’Zeitschr. Analyt. Chem., 319, 600–5.

SCHREURS, J.W.H. and BRILL, R.H. (1984) Iron- andsulphur-related colours in ancient glass. Archaeometry,29, 199–209.

SCHRÖDER, H. and KAUFMANN, R. (1959) Schutz-schichten für alte Gläzer. In Beiträge zur AngewantenGlasforschung (E. Schott, ed), pp. 355–61. Stuttgart.

SCHULER, F. (1959a) Ancient glassmaking techniques. Themolding process. Archaeology, 12, 47–52.

SCHULER, F. (1959b) Ancient glassmaking processes. Theblowing technique. Archaeology, 12, 116–22.

SCHULER, F. (1962) Ancient glassmaking processes. TheEgyptian core-vessel process. Archaeology, 51, 32–7.

SCHULER, F. (1963) Ancient glassmaking techniques.Egyptian fused miniature mosaics. In Advances in GlassTechnology, Part, 2 (F.R. Matson and G.E. Rindone, eds),p. 206. Plenum Press, New York.

SCHULZE, G. (1977) Chemisch–analytische Untersuch-ungen an Glasfunden aus der Kunckel Hütte. Contribu-tion to the DGG Meeting in Frankfurt (Abstract No. 297in CVMA. News Letter 26).

SCHUSTER, R. (1980) Hinterglasbilder aus Außergerfild inBöhmerwald. Grafenau.

SCHUSTER, R. (1983) Hinterglasbilder der NeukirchenerSchule. Grafenau.

SCHUSTER, R. (1984) Das Raimundsreuter Hinterglasbild.Geschischte der Raimundsreuter Hinterglasmalerei undihres Einflußgebietes. Grafenau.

SCHWARZ, M.D. (ed.) (1974) American Glass from thePages of Antiques, I. Blown and Molded. The PynePress.

SCHWEIG, B. (1973) Mirrors: A Guide to the Manufactureof Mirrors and Reflecting Surfaces. Pelham Books,London.

SCOTT, A.S. (1932) Apparent decay of ancient glass atWells Cathedral. J. Brit. Soc. Master Glass Painters, 4,171.

SCRANTON, R. (1967) Glass pictures from the sea.Archaeology, 20 (3), 165–73.

SEASE, C. (1981) The case against using soluble nylon inconservation work. Stud. Cons., 26, 102–10. IIC,London.

SEASE, C. (1988) A Conservation Manual for the Archaeo-logist. Archeological Research Tools, 4. Institute ofArcheology, University of California, Los Angeles.

SEELA, J. (1974) The early Finnish glass industry. J. GlassStud., 16, Fig. 3, 65.

SEIBEL, F. (2000) Technologie Römicher Glashütten imRheinland. Glastech. Berlin Glass Sci. Technol. 73, (1)N1–N6.

SEIDLER, R. (1987) The Use of Silanes for the Consolidationof Paintings on Glass. University of Delaware, USA.

SELLNER, C. (1977) Untersuchungen an Waldgläsern mitElektronen–spinresonanz. In Contribution to theProjektgruppe ‘Glas’ Meeting in Würzburg (Abstract No.298 in CVMA News Letter 26).

SELLNER, C., OEL, H.J. and CAMARA, B. (1979)Untersuchungen alter Gläzer (Waldglas) aufZusammenhang von Zusammenzetzung, Farbe undSchmelzatmosphäre mit der Elektronenspektroscopieund der Elektronenspinresonanz (ESR) Glastech. Berlin,52, 255–64.

SHAW, G. (1965) Weathered crusts on ancient glass. NewScientist, 21, 290–1.

SHORTLAND, A. (1999) Early Egyptian glass – evidencefor trade and tribute. In Proceedings of Current Researchon the History of Glass through Scientific Analysis.London.


SHORTLAND, A.J. and TITE, M.S. (1998) The interdepen-dence of glass and vitreous faience production atArmana. In The Prehistory and History of GlassmakingTechnology (P. McCray, ed.), Ceramics and Civilisation,VIII, 251–68. American Ceramics Society, Columbus,Ohio.

SIMPSON, H.E. (1953) Some factors affecting the testingof surface durability of glass. J. Am. Ceram. Soc., 36,143–6.

SINGLEY, K. (1998) The Conservation of ArchaeologicalArtefacts from Freshwater Environments. Lake MichiganMaritime Museum, South Haven, Michigan.

SIROIS, P.J. (1999) The deterioration of glass trade beadsfrom Canadian ethnographic and textile collections, InThe Conservation of Glass and Ceramics: Research,Practice and Training (N.H. Tennent, ed.), pp. 84–95.James and James, London.

SLOAN, J.L. (1993) Conservation of Stained Glass inAmerica: A Manual for Studios and Caretakers. Art inArchitecture Press, USA.

SMEDLEY, J.W., JACKSON, C.M. and BOOTH, C.A. (1998)Back to roots: the raw materials, glass recipes, andglassmaking practices of Theophilus. In The Prehistoryand History of Glassmaking Technology (P. McCray,ed.), Ceramics and Civilisation, VIII, 145–165. AmericanCeramics Society, Columbus, Ohio.

SMITH, C.S. and GNUDI, M.T. (1942) The Pirotechnica ofVannoccio Biringuccio. American Institute of Miningand Metallurgical Engineering, New York.

SMITH, R.W. (1963) Archaeological evaluation of analysesof ancient glass. In Advances in Glass Technology, Part2 (F.R. Matson and G.E. Rindone, eds), pp. 283–90.Plenum Press, New York.

SMITH, R.W. (1963a) Archaeological evaluation of analy-ses of ancient glasses. In Advances in Glass Technology,Part 2 (F.R. Matson and G.E. Rindone, eds), pp. 283–90.Plenum Press, New York.

SMITH, R.W. (1963b) The analytical study of glass inarchaeology. In Science in Archaeology (D. Brothwelland E. Higgs, eds), pp. 519–28. Thames and Hudson,London.

SMITH, S. (1996) The manufacture and conservation ofEgyptian faience. In Preprints ICOM–CC 11th TriennialMeeting, Edinburgh, pp. 845–50. ICOM, London.

SMITHSONIAN INSTITUTION (1969) Glass GobletExcavated from a Privy. Conservation–AnalyticalLaboratory, USA.

SODE, T. (1996) Anatolske Glasperler. Thot Press, Vanlose,Denmark.

SODE, T. and SCHNELL, U. (1998) Contemporary faiencemakers in Quarna, Egypt. In Preprints Interim MeetingICOM–CC Working Group on Ceramics, Glass andRelated Materials (A. Paterakis, ed.). EVTEK Institute ofArts and Design, Vantaa., Finland.

SOMMER-LARSON, A. (1999) Conservation of three glasschandeliers from Konigsberg Church, Norway. In TheConservation of Decorative Arts (V. Horie, ed.), pp.35–42. Archetype Publications, UKIC, London.

SPEEL, E. (1984) Popular Enamelling, London.STAFFELBACH, G. (1951) Geschichte der Luzerner Hinter-

glasmalerei. Luzern.STAHL, C.J. (1915) Dekorative Glasmalerei, Unter-

glasmalerei und mahlen auf Glas. In Chemisch undTecknisch Bibliotek, 354. Vienna and Leipzig.

STAUDE, H. (1972) Die Technik des Zussammensetzensund Ergänzens antiker Gläzer. Arbeitsblätter fürRestauratoren, 1, Gruppe 5, 20–7.

STEINBRUCKER, C. (1958) Artikel Eglomise. InReallexikon zur Deuytschen Kunstgeschichte, 4, 749ff.

STERN, E.M. (1998) Interaction between glassworkers andceramiscists. In The Prehistory and History ofGlassmaking Technology (P. McCray, ed.), Ceramicsand Civilisation, VIII, 183–204. American CeramicsSociety, Columbus, Ohio.

STEVELS, J.M. (1960) New light on the structure of glass.Philip’s Tech. Rev., 22, 300–11.

STONE, J.F.S. and THOMAS, L.C. (1956) The use anddistribution of faience in the ancient East and prehis-toric Europe. Proc. Prehist. Soc., 22, 37–84.

STRAHAN, D. (2001) Uranium in glass, glazes andenamels: history, identification and handling. Stud.Cons. 46, 181–195.

STRONG, D. and BROWN, D. (1976) Roman Crafts.Duckworth, London.

SWANN, C.P., MCGOVERN, P.E. and FLEMING, S. J. (1990)Colorants used in Ancient Egyptian glassmaking:specialised studies using PIXE spectrometry. NuclearInstruments and Methods in Physics Research, B, 45,199–209.

TABACZYNSKA, E. (1968) Remarks on the origin of theVenetian glassmaking center. Proc. 8th Int. Congr. Glass,Studies in Glass History and Design, pp. 20–3.

TAIT, H. (1968) Glass in Europe from the Middle Ages to1862. In Masterpieces of Glass (D.B. Harden et al., eds).Trustees of the British Museum, London.

TAIT, H. (1979) The Golden Age of Venetian Glass. BMP,London.

TAIT, H. (ed.) (1991) Five Thousand Years of Glass. BMP,London.

TAYLOR, T.H. (1984) In situ repair of architectural glass.In Preprints of the Contributions to the Paris Congresson Adhesives and Consolidants (N S. Bromelle, M. Pye,P. Smith and G. Thomson, eds), pp. 202–4. IIC, London.

TECHMER, K.S. and RÄDLEIN, E. (1998) TEM and AFMstudies on the alteration of natural glasses. Proc. 18thInt. Congr. Glass, San Francisco, CD-ROM, Paper No. 9in Group C4, 49–51.

TENNENT, N.H. (1981) Fungal growth on medieval glass.J. Br. Soc. Master Glass Painters, 17, 64–8.

TENNENT, N.H. and TOWNSEND, J.H. (1984a) The signif-icance of the refractive index of adhesives for glassrepair. In Preprints of the Contributions to the ParisCongress on Adhesives and Consolidants (N.S. Bromelle,E.M. Pye, P. Smith and G. Thomson, eds), pp. 205–12.IIC, London.

TENNENT, N.H. and TOWNSEND, J.H. (1984b) Factorsaffecting the refractive index of epoxy resins. Preprintsof the 7th Triennial Meeting of ICOM–CC, Copenhagen,pp. 84.20.26 to 84.20.28.

TENNENT, N.H., MCKENNA, P., LO, K.K.N., MCLEAN, G.and OTTAWAY, J.M. (1984b) Major, minor and traceelement analysis of medieval stained glass by flameatomic absorption spectrometry. N ArchaeologicalChemistry III (Advances in Chemistry Series), 205,133–50.

THEOBALD, W. (1933) Technik des Kunsthandwerks imzehnten Jahrhundert des Theophilus, Berlin.

THEOPHILUS, PRESBYTER. (circa 1140) Schedula Divers

Bibliography 363

Artium (Treatise on Diverse Arts) (translated as OnDivers Arts, Dodwell (1961) and by Hawthorne andSmith (1963)).

THOMSON, G. (1998) The Museum Environment, 2ndedn. Butterworth-Heinemann, London.

THOMPSON, R.G. (1925) On the Chemistry of the AncientAssyrians. London.

THORENSEN, L. (1998) The use of glass in the glyptikarts. In Preprints of the ICOM–CC Working Group:Ceramics, Glass and Related Materials (A. Paterakis,ed.). EVTEK Institute for Arts & Design, Vantaa, Finland.

THORNTON, J. (1981) A transfer treatment technique forHinterglasmalerei. JAIC, 20, 28–35.

THORNTON, J. (1990) A light box apparatus for the repairof glass. Stud. Cons., 35 (2), 107–9.

THORPE, W.A. (1938) The prelude to European cut glassTrans. J. Soc. Glass Technol., 22, 5–37.

THORPE, W.A. (1961) English Glass, 3rd edn, p. 99.London.

TITE, M.S. and BIMSON, M. (1987) Identification of earlyvitreous materials. In Recent Advances in the Conserva-tion and Analyses of Artefacts: Proceedings of the JubileeConservation Conference of the Institute of Archaeology,pp. 81–5. University of London Press.

TITE, M.S., BIMSON, M. and FREESTONE, I. C. (1983)Egyptian faience: an investigation of the methods ofproduction. Archaeometry, 23, 17–27.

TITE, M.S., BIMSON, M. and MEEKS, N.G. (1981)Technological characterisation of Egyptian Blue. InActes du XX Symposium International d’ArchaeometrieIII (et Analyse) Review d’Archaeometrie, Supplement,296–301.

TONINATO, T. (1984) Technology and tradition in Muranoglassmaking. Glass in Murano, 20–40. Venice.

TREMAIN, D. (1988) Reverse-glass prints: their history,techniques and conservation. Proc. Symp. 88, CCI, pp.143–52.

TURNER, W.E.S. (1930) The scientific basis of glassmelting. J. Soc. Glass Technol., 14, 368–93.

TURNER, W.E.S. (1954a) Studies in ancient glass and glass-making processes. Part 1. Crucibles and melting temper-atures employed in ancient Egypt at about 1370 BC.Trans. J. Soc. Glass Technol., 38, 436–44.

TURNER, W.E.S. (1954b) Studies in ancient glass andglass-making processes. Part II. The composition,weathering characteristics and historical significance ofsome Assyrian glasses of the eighth to sixth centuriesBC, from Nimrud. Trans. J. Soc. Glass Technol., 38,445–56.

TURNER, W.E.S. (1956a) Studies in ancient glasses andglass-making processes. Part III. The chronology of theglassmaking constituents. Trans. J. Soc. Glass Technol.,40, 39–52.

TURNER, W.E.S. (1956b) Studies in ancient glasses andglass-making processes. Part IV. The chemical compo-sition of ancient glasses. Trans. J. Soc. Glass Technol.,40, 162–86.

TURNER, W.E.S. (1956c) Studies in ancient glasses andglass-making processes. Part V. Raw materials andmelting processes. Trans. J. Soc. Glass Technol., 40,276–300.

TURNER, W.E.S. (1957a) Antichi vetri opali, il suppostoeffeto opalizzante del biossido di stagna. Vetro e Silicati,2, 27–30.

TURNER, W.E.S. (1957b) Ancient sealing wax red glasses.J. Egyptian Arch., 43, 110–12.

TURNER, W.E.S. (1959) Studies in ancient glasses andglass-making processes. Part VI. The composition andphysical characteristics of the glasses of the PortlandVase. Trans. J. Soc. Glass Technol., 43, 262–84.

TURNER, W.E.S. (1962) A notable British seventeenth-century contribution to the literature of glassmaking.Glass Technol., 3, 201–13.

TURNER, W.E.S. (1963) The tercentenary of Neri–Merrett’s‘The Art of Glass’. Advances in Glass Technology, Part2 F.R. Madson and G.E. Rindone, eds), pp. 181–201.Plenum Press, New York.

TURNER, W.E.S. and ROOKSBY, H.P. (1959) A study ofthe opalising agents in ancient opal glasses throughoutthree thousand four hundred years. Glastech. Ber., 32K,VIII/17–VIII/28.

TURNER, W.E.S. and ROOKSBY, H.P. (1961) Furtherhistorical studies, based on X-ray diffraction methods,of the reagents employed in making opal and opaqueglasses. J. Röm.–German. Zentralmuseums (Mainz),VIII, 1–6.

VANDIVER, P.B. (1992) Corrosion and conservation ofancient glass and ceramics. In Corrosion of GlassCeramics and Ceramic Superconductors (D.E. Clark andB.K. Zoitas, eds), pp. 393–430. Noyes Publications, ParkRidge, NJ.

VANDIVER, P.B. and KINGERY, W.D. (1986) Egyptianfaience. The first high-tech ceramic. In High TechnologyCeramics (W. Kingery, ed.), Ceramics and Civilization,III, 19–34. American Ceramics Society, Columbus, Ohio.

VANDIVER, P.B., GIBSON, MCG. and MCMAHON, A.(1995) Glass manufacture in the late third millenium BCat Nippur in Iraq. In The Ceramics Cultural Heritage:Proceedings of the International Symposium ‘TheCeramics Heritage’ of the 8th CIMTEC–World CeramicsCongress & Forum on New Materials, Florence, Italy.Monogr. Mater. Soc., 2, 331–41.

VANDIVER, P.B., SWANN, C.P. and CRANMER, D. (1991)A review of mid-second millennium BC Egyptian glasstechnology at Tell El-Amarna. In Material Issues in Artand Archaeology, II, 609–16. The Materials ResearchSociety, Pittsburgh, PA.

VERITÁ, M. (1985) L’invenzione del cristallo Muranese:una verifica analytica delle fonti storiche Rivista delleStazione Sperimentale del Vetro, 1, 17–36.

VERITÁ, M. (1995) Analytical investigation of Europeanenamelled beakers of the 13th and 14th centuries. J.Glass Stud. 37, 83–98.

VERITÁ, M. (2001) Technology and deterioration of vitre-ous mosaic tesserae. Rev. Cons., 1, 65–76.

VERITÁ, M., BASSO, R., WYPYSKI, M.T. and KOESTLER,R.J. (1994) X-ray microanalysis of ancient glassy materi-als: a comparative study of wavelength dispersive andenergy dispersive techniques. Archaeometry, 36, 241–51.

VINCENT, J. (1556) La Pyrotechnie. traduite d’Italien enFrançais par feu Maitre Jaques Vincent, p. 90. ClaudeFremy, Paris.

VOLF, M.B. (1977) Jaroslav Brychta. Zivé sklo, Prague.VON SALDERN, A. (1970) Originals – reproductions –

fakes. Ann. 5e Congr. Assoc. Hist. Verre, Prague, pp.299–318. AIHV, Liège.

VON SALDERN, A. (1974) Gläser der Antike, Museum fürKunst und Gewerbe, Hamburg.


VOS-DAVIDSE, L. (1969) Note of the reversible glueing ofbroken glass objects. Stud. Conserv., 4, 183.

WAKEFIELD, H. (1963) Methods of packing in the Victoriaand Albert Museum. In Recent Advances in Conservation(G. Thomson, ed.), pp. 16–18. Butterworths, London.

WALLACE, K. (1976) Examination and treatment ofHinterglasmalerei, or reverse paintings on glass.Unpublished paper, National Gallery of Art, Ottawa,Canada.

WALTERS, H.V. and ADAMS, P.B. (1975) Effects of humid-ity on the weathering of glass. J. Non-Cryst. Sol., 19,183–99.

WARD, C. (1998) Sadana Island shipwreck: final; season.INA Quarterly, 25, no. 3, 3–6.

WARREN, P. (1981) Irish Glass. Faber and Faber, London.WATKINSON, D. and LEIGH, D. (1978) Polyurethane

foam: a health hazard. Cons. News, IIC–UHG, No. 6, 7–8.WATKINSON, D. and NEAL, V. (ed.) (1998) First Aid for

Finds. Rescue/UKIC Archaeology Section and Museumof London, UKIC, London.

WATTS, D.C. (1999) Some London glasshouses in the 17thto 19th centuries, and their particular association withthe Thames South Bank. In The Guild of Glass EngraversSpring Newsletter 1999, pp. 4–9.

WEDEPOHL, K.H. (1993) Herstellung mittelalterlicher undantiker Gläser. Abh. Akad. Wissenschaften Lit. Mainz,Germany.

WEDEPOHL, K.H., WINKELMANN, W. and HARTMANN,G. (1997) Glasfunde aus der karolingischen Pfalz inpaderborn und die frühe Holzasche–Glasherstellung. InAusgrabungen und Funde in Westfalen–Lippe (Munster),Germany, 9/A, pp. 41–53.

WEIER, L.E (1973) The deterioration of inorganic materialsunder the sea. Bull. Inst Arch. (London), 11, 131–63.

WEINBERG, G.D. (1975) A medieval mystery: Byzantineglass production. J. Glass Stud., 17, 127–41.

WEINTRAUB, S. (1998) Crizzling: microclimate controlsolutions. Proc. 18th Int. Congr. Glass, San Francisco,CD-ROM, Paper No. 3 in Group B1, pp. 13–17.

WERNER, A.E., BIMSON, M. and MEEKS, N.D. (1975) Theuse of replica techniques and the scanning electronmicroscope in the study of ancient glass. J. Glass Stud.,17, 158–60.

WEYL, W.A. (1951) Coloured Glasses. Society of GlassTechnology, Sheffield.

WEYL, W.A. and MARBOE, E.C. (1967) The constitution ofglasses, a dynamic interpretation. In Surface ChemistrySilicate Glasses. 3 vols, 23, 1010–270. Wiley, New York.

WHARTON, G. and OLDKNOW, C. (1987) The conserva-tion of an eighteenth century glass medallion beaker.In The Conservator, 11, 42–5. UKIC, London.

WICKS, S. (1985) Jewellery Making Manual. Little, Brown,London.

WIECK, T. (1981) Glas als Bildtrager. HistorischeTechniken und ihre Anwendung ohne Berücksichtigungvon Scmelzvefahren. Diplomarbeit, Institut fur Tech-

nologie der Malerei, Staatliche Akademie der BildendenKünste. Stuttgart.

WIEDERHORN, S.M. (1967) Influence of water vapor crackpropagation in soda–lime glass. J. Am. Ceram. Soc., 50,407–14.

WIHR, R. (1963) Repair and reproduction of ancient glass.Recent Advances in Conservation, pp. 152–5.Butterworths, London.

WIHR, R. (1968) Möglichkeiten der Restaurierung undNachbildung der antiker Gläser, mittels giessbarerKunststoffe. Arbeitsblätter für Restauratoren, Gruppe 5,1–12.

WIHR, R. (1977) Restaurierung von Keramik und Glas.Callwey, Munich.

WILLIAMS, N. (1989) The Breaking and Re-making of thePortland Vase. BMP, London.

WILLISTON, S.S. and BERRETT, K.R. (1978) Preliminarynotes on setting down paint flaking from glass. JAIC,18, 46–8.

WILLS, G. (1965) English Looking Glasses, 1670–1820.London.

WINTER, A. (1965) Alteration des surfaces des verresanciens. Proc. 8th Int. Congr. Glass, Brussels, Paper No.229, 12 pp.

WOOD, E.S. (1965) A medieval glasshouse at Blunden’sWood, Hambledon, Surrey. The Surrey Arch. Collections,64, 54–79.

WOSINSKI, J.F. and BRILL, R.H. (1969) A petrographicstudy of Egyptian and other cored vessels. In Studies inGlass History and Design: Proc. 8th Int. Congr. Glass,Sheffield. 123–4.

WULFF, H.S., WULFF, H.W. and KOCH, L. (1968) Egyptianfaience, a possible survival in Iran. In Archaeology, 21,98–107.

YABUKI, H., YABUKI, S. and SHIMA, M. (1973) Fissiontrack dating of man-made glasses from Ali Tar Cavernvestiges. In Scientific Papers of the Institute of Physicaland Chemical Research, 67, 41–2.

YALOURIS, N. (1968) An unreported use of someMycenean paste beads. J. Glass Stud., 10, 9–16.

YPEY, J. (1960–1961) Das Bergen von stuck angegriffenGlas Während einer Ausgrabung mit Hilfe einerKunstharzlösung. Ber. van de Rijsdienst voor hetoudheikundig Bodeminders, 10–11, 363–9.

YPEY, J. (1965) The conservation of disintegrated glassduring the excavation. Proc. 7th Int. Congr. Glass,Brussels, paper No. 227, 3 pp.

ZACHARIASEN, W.H. (1932) The atomic arrangement inglass. J. Am. Ceram. Soc., 54, 3841–51.

ZECCHIN, L (1987) Il recettario Dardin. In Un CodiceVetrario del Seicento Transcritto e Commentato.Statzione Sperismentale del Vetro, Murano, Italy.

ZIMMER, K. (1988) Spectrographical investigation ofHungarian glass finds. Archaeometrical Research inHungary, pp. 131–40. Budapest National Centre ofMuseums.

Bibliography 365


AAS (atomic absorption spectrometry), 237Abbasid dynasty (Islamic glass), 27Abbé (optical glassmaker), 117Ablebond 342-1 (epoxy adhesive), 212, 276, 296Abies balsam (Canada balsam), 205Abrasion:

damage by, 170, Fig. 4.4, 4.5Accelerated ageing tests, 196–7Acetaldehyde acid, 309Acetic acid, 262, 309, 333, 334Acetates, effects of, 182Acetone, 256, 270, 272, 278, 314Acid, attack on glass, 170, 175, 190, 197

etching of glass, 101Acid-free tissue, 259, 312, 316Acre, Palestine (glassmaking in), 22Acrylic polymers, 205, 216–7, Fig. 5.10Acryloid B-72 (see Paraloid B-72)Active conservation, 242Adam, Robert and James (archtect designers), 70Adhesion (theory of), 205Adhesives, 204–6, Fig. 5.1AES (Auger Emission spectrometry), 235‘Agar Quf, Mesopotamia (Iraq), 93Aggry (aggri, see Beads)Airbrasive cleaning, 256, 262Air-twist stems or bubbles (decorative), 42, 106, Fig.

3.24Agate glass, 76Agricola, Georgius (writer of glassmaking), 147, 148,

149, 161Alabastron (vessel type), 19, Fig. 2.2Albert Memorial The, London (glass jewels on), 275Alcohol (solvent), 316Aleppo, Syria (glassmaking in), 27, 30, 101Alexandria, Egypt (glassmaking in), 22, 27, 30, 79, 98,

102Algae (removal of), 203Alkali, 73, 74, 139, 141, 176, 180–1, 193

attack on glass (see De-alkalization)Source of (plant ash), 74–5, 137, 143, 186

Alloa , Scotland (cone furnace), 166Alloway, New Jersey (USA) (glassworks), 46Aloplast (modelling material), 220, 294, 295–7Altare, Italy (glassmaking in), 80Aluminium, 319

foil, 246, 247, 248Aluminium oxide (alumina) in glass:

(emery, corundum, bauxite), 7, 76, 177, 223Amalgam, 58, 334, 335Amber glass, 8Amelierung, 57, 340

Amelung, New Bremen, USA, (glasshouse), 46, 158–9,Fig. 3.70

Amelung, Johann Frederick (glassmaker), 158–9Amiens, France (glassmaking in), 30Aminocarboxylic acids, 201, Fig. 5.3Amines, effects of, 182Ammonia, 225, 331Ammonium hydrogen carbonate, 202Amphorisk (vessel type), 21Amsterdam, Netherlands, 36Amsterdam, The (shipwreck), 187Ancient glasses (definition of), 1–3Anelasticity (internal friction), 11, 12Anglo-Venetian glass, 111Anions, 5Animal glues, 203Anisotropy, 229Annealing, 12, 86–87

annealing furnace, 160Antimony, 1, 77, 78, 79, 136, 237

as decolorizer, 9as opacifier, 10in mirror making, 134

Antioch (modern Antakya), Turkey, 18Antwerp, Belgium, 37Aqua regia, 9Aquamanile (vessel type), 33Aquarel paints, 343Aquileia, Italy (glassmaking in), 30Aphrodisias (Roman writer), 62Apollonia, Israel (glassmaking site), 143Application technique (Egyptian faience), 85Araldite adhesives (AY193/HY956), 225, 276, 306, 307Araldite 2020, 212, 276Arbogast machine, 48, 117Arsenic, 136

oxide (decolourizer), 9, 77Archae-o-Derm (PVC), 219Archetto (drill for engraving), 97Arrêt du Conseil d’Etat, 129Argon atmosphere, 314Argy-Rousseau, Gabriel (glass paste), 91Art Nouveau glass, 48, 49Ars Experimentalis (Experiments in the Art of Glass),

159Artsorb silica gel, 317Aryballos (vessel type), 21, Fig. 2.2Ascension windows, Le Mans cathedral, France, 64Ashurbanipal (cuneiform library of), 135Astarte (figurine of), 90, Fig. 3.7Atmospheric pollution, effect of, 171, 173, 174, 190, Fig.

4.24

367

Index

Atomic force microscopy, 241Augsburg Bible (illustration in), 163Augsburg Cathedral, Germany (windows in), 64Auldo Jug, The (cameo glass), 22, 98, 294–7, 304Aventurine glass (decorative red), 33, 76AYAF (polyvinyl acetate), 225–6

Baccarat , France (glass paperweights from), 94Back-matching corrosion (on windows), 192Badajoz, Spain (glassmaking site), 141Badarian civilization (Egypt), 18, 85Baghdad Iraq (glassmaking in), 27Bagot’s Park, Staffordshire (furnace sites), 155Baia, Naples, Italy (glass from), 187Bailey & Saunders (glassmakers)71Ballidon, Derbyshire (glass burial experiment), 198Ballotini (glass beads) (see Cloisonné glass windows),

338Barcelona, Spain (glassmaking in), 30, 57Barias, Israel (glassmaking site), 143Barilla (source of alkali), 39, 80Barium in glass:

barium oxide, 76barium sulphide (barytes or heavy spar), 76

Barovier, Ercole, (glassmaker), 49Basilica of S. Maria Assunta, Torcello, Italy (mosaics), 144Battledore (glass shaping tool), 104Baumgartner, Wolfgang (glass painter), 57Bauxite (aluminium hydroxide), 76Beads, 87–88, 235, 236, 237, 262, Fig. 3.3, 3.4Beauvais, France, (glassmaking in), 30Bedacryl 122X , 217Beechwood (fuel or source of alkali), 75, 79, Beijing, China (glassmaking in), 29Beilby (William & Mary, glass enamellers & gilders), 43,

113, 115Belfast, Northern Ireland (glassmaking in), 45, 166Belfast Newsletter, The166Belus River (R.. Na’aman, Palestine, source of sand), 16,

22, 73, 143Benckert, Hermann (glass decorator), 112Beta-ray back scattering test, 240Bet She’arim, Palestine (glassmaking site), 3, 102, 142–3Beva 371 (adhesive), 343Bianconi, Gian-Ludovico (author), 150Biedermeier glasses (decorative), 43, 44Bierschieben (celebration windows), 65Biocides, 199, 203Biringuccio (writer on glassmaking), 146, 147, 148, 149,

151Bismuth (in mirror making), 134Bishopp, Hawley, (glassmaker), 40Bisseval (medieval glassmaking family), 17Blaschka, Leopold, Rudolph, glass figures, 318, Fig. 7.55,

7.56Bloom on glass, 273, 274, 318, Fig. 4.30Blore Park, Staffs (furnace site), 155Blowpipe (blow iron, for glassblowing), 105Blue Vase, The (cameo glass, in Naples), 98Blunden’s Wood, Surrey, (furnace site), 154–156, Fig.

3.61, 3.62, 3.63Bocca/boccarello (furnace mouth), 145, 148, 149Bohemian glass, 34, 37, 77, 100Bone, Henry (glass enameller), 52

Bone beads, 319Bontemps (glassmaker), 117, 130, 165Boric oxide (boron, in glass), 4, 74Boshan, China (glassmaking site), 4, 28Boston Glass Manufacturing Co., 47Boulsovier, Thomas, 69Bracchio/bracchia (measurement), 146Brewster (investigator into glass), 174Bridging oxygen atoms, 4–5, 253Brighton Pavilion, Sussex (chandeliers in), 71, 323Brillouin light scattering (test), 241Bristol glass, 58Brittleness, 14Brixworth, Northamptonshire (glass from), 128Broad glass (see Cylinder glass)Broc à glaces (see Ice glass)Brown’s Hosptal, Canterbury (glass jewels in windows),

132Brunswick, Germany (glass engraving in), 36Brychta, Jaroslav (glass figure maker), 318–9Bubblewrap (packing material), 312Buckholt, Wiltshire, (furnace site), 158Burial experiments (of glass) (see Ballidon)Burmese glass (decorative), 44, 48Burne-Jones, Sir Edward (window designer), 65Butanone (hydrophilic solvent), 270Butteri, G.M., (wallpainting of glass furnace), 148, 149Butvar-B (polyvinylbutyral), 219Byzantine glass/mosaics, 26

Cage cups (see Diatreta)Caistor St. Edmund, Norfolk (furnace site), 142Calaton CA7, & CB (see Soluble nylon)Calcedonio (chalcedony glass, decorative), 76Calcium antimonate (opalising agent), 10Calcium chloride, 314Calcium fluoro-phosphate (opalising agent), 10Calcium gluconate (HF antidote), 203Calcium (lime, in glass), 17, 139, 176, 200, 202

oxide, 5, 60, 73, 81, 180, 309hydroxide (slaked lime), 86limestone, 89sulphate, 221

Calculi (quartz pebbles), 74Calgon (polyphosphate cleaning agent), 201, 202Calaton CA 7 CB (soluble Nylon), 211Cam, Jean le(glassmaker), 39Cameo glass, 21, 29, 98–100, 106–7, Fig. 3.19Cames lead strips for stained glass (see Leading)Campania, Italy (glassmaking centre), 22Canada balsam (abies balsama), 205Candelabra and candlesticks, 68–9Canosa, Apulia, Italy (glass from), 116, Fig. 2.3, 3.32, Canton, China (glasspainting in), 58Canqueray, Philippe de (glassmaker), 129Capella Palatine, Palermo, Italy (mosaics in), 60Carbon dioxide, 190Carbon disulphide, 262Carbon-sulphur amber glass, 8Carbonate precipitation (at sea), 171, 189–190Carbon tetrachloride, 314Carbowax 6000 (polyethylene glycol wax), 245, 266Carder, Frederick (glassmaker), 48 Carleon, S. Wales (Roman window glass), 62, 126

368 Index

Carlsberg, C.W, (illustrator), 167Carré, Jean le (glassmaker), 38, 63, 154Cartellina (thin sheet of glass), 60Cartoon, 65Casing, cupping, 106–7

(see also Cameo glass)Cassiterite (see Tin oxide)Cast glass sheets, 130Casting in moulds, 91–2Casting materials, 90Catcliffe, Yorkshire (cone furnace), 166Catechol (complexing agant), 263Cations, 5Cavalet (or crown), iron ring around Glory-hole, 149Cave relievo (engraving), 98Cellosolve (ethylene glycol monoethyl ether), 341, 342Celluloid (cellulose nitrate), 211Cellulose nitrate, 275, 319

(see also HMG, Durofix, Duco cement)Cementation (Egyptian faience), 84Chair (glassformer’s workplace), 102Champlevé (enamelling technique), 51, 52, 119, 122Chagall, Marc (artist), 65Chance Bros. (glassmakers), 117, 129, 130Chandeliers, 69–72

conservation, 320–332, Fig. 7.59, 7.60de crystal, 71dismantling, 326–330, Fig.7.62, 7.63, 7.64examination, 324–6recording, 328, Fig. 7.61

Chandelles (tallow candles), 69Chartres Cathedral, France (windows in), 63 Chelating (sequestering) agents,

for cleaning, 199, 200–1, 337, Fig. 5.1Chemical analysis, 239Chemical deterioration of glass, 173–182Chemical tests (for identifying soluble salts), 252–3Chichester, Sussex (early crown glass in), 128Chiddingfold, Sussex, 156, 158

church windows, 337furnace site, 46

Chinese glass, 28, 58, 59, 178, 231, 238, 316painting on mirrored glass, 333

Chinoiserie (decorative style), 37, 42, 43, 72Chloride ion (detection of), 252–3Chromium oxide (chromate & bichromate, in glass), 78Chrystoleum (backed photographs), 54Chunk glass (newly-made, made for export), 29, 82,

142, Fig. 3.1Cire perdue (see Lost wax process)Citrates (effect of), 182Clay (see Potters’ clay)Cleaning glass:

agents, 199on site, 257–8, Fig. 7.8, 7.9in the laboratory, 260–3, Fig. 7.12historic glass, 271

Clichy, France (glasss paperweights from), 94Clingfilm, 225, 247Cloisonné glass, 54, 66–7, 317, 337–8, Fig. 7.68Cobalt (in glass), 73, 76, 237, 319, Fig. 1.7Coefficient of expansion, 13, Table 5.1Colchester, Essex (Roman glass), 26Cold painting (decorative), 54, 113

Cold working (abrading, cutting, grinding), 95–8, 176,Fig. 2.3, 2.10, 3.15

Colemanite (hydrated calcium borate), 44Colloidal suspension (of metals), 8, 76, 138Cologne, Germany, 25, 30Colourants:

for glass, 6, 73, 74, 114, 115, 143, 177, 223for filling materials, 223

Commercial restoration, 308Complexing (chelating) agents, 182, 194Composition of glass, effect on decay, 176Conchae (shells), 74Condensation, effect on glass, 174, 183, 192Conductivity meter, 251Conqueror, J. (glassmaker), 130Consolidation, 204–6, 263–270, 314, 336, 341, Fig. 7.14,

7.15, 7.16, 7.17system, definition of, 264

Constantinople (Istanbul), Turkey, 26, 31, 32, 62Copper (in glass), 8, 73, 76, 77, 131, 139, 225

chloride, 86copper oxide, 86cuprous oxide, 10, 136, 138

Copper foil (for repair), 337, 338Copper-ruby glass, 8Cordoba, Spain (Roman glass), 26Core forming (technique), 18, 88–9, Fig. 2.1, 2.2, 3.5, 3.6Corinth, Greece (glassmaking site), 30, 31Cork, Eire (glassmaking in), 45Corning Mus. of Glass (USA), 254, 255, 316Corning Glassworks (USA), 116Correx board (packaging material), 226, 311Cotton wool (Surgical cotton; cotton buds), 259, 261,

278, 280, 315Counter enamelling (see Enamelling)Coupling agents (silanes), 205Coventry Cathedral, Warwickshire, 65Covered pots (for melting glass), 167Cracks in glass, 183, 233, 259, 280, 334Craquelure (decorated glass), 33, 49Cristallo (colourless glass), 30. 32, 80, 178Crystallum (crystallina), (Roman term for crystal glass),

79Cristobalite (type of silica), 1, 86Crizzling (of glass), 179, 191, 280, 310, 314–7, Fig. 4.17Cross, Henri & Jean (sculptors), 91Crown glass (Normandy-, spun-glass), 61, 63, 128–9Crucibles, 73, 82, 140Crutched Friars, London (glasshouse), 148Cryla acrylic paints, 344Crystal Palace, London (window glass in), 43, 63, 71, 117Culbin, Scotland (glass beads from), 85Cullet (re-usable glass), 73, 75, 80, 83, 86, 182Cuneiform tablets (in Assyria), 9, 82, 86, 135–7Cupping (see Cameo glass)Cut-line, 65Cyanoacrylate adhesives, 205, 217–8, 276, 280, Fig. 5.11Cylinder-, Broad- Spread-, Muff-glass, 61, 63, 126–8, 291,

Fig. 3.43Czechoslovakia (Czech Republic), 71, 154, 158

Dalle de verre (slab glass), 66Damascus, Syria (glassmaking), 27, 28, 30Dammar resin, 276

Index 369

Dammouse, Albert (glass paste), 91Damp storage of church windows, 191Daphne Ewer, The (cold painting on), 113–4, Fig. 3.30Daraa, Syria,Darduin, Giovanni (glassmaker), 80Darmstad, Germany,Darney Forest, France,Daruvar, Zagreb, Yugoslavia (cage cup from), 96Daugmale castle mound, Latvia (glass beads from), 262Daum glass (decorative)49Daumengläser (vessel vessel), 35, 110, Fig. 2.15Davenport’s Patent Glass (decorative), 111Davidson & Co. (glass manufacturers), 44De-alkalisation (leaching), 174, 175, 179, 196De Coloribus et Artibus Romanorum (On the Colours &

Arts of the Romans), 153Decorchement, Francois (glass paste), 91De Re Metallica (glassmaking text), 147, 148Decalomania (decorative technique), 54Decanters (conservation), 273, 304Decolorizers for glass, 7, 78–9Deka fabric dyes (use of), 306Density (of glasses), 14, 231, Fig. 1.8, 1.11Dental wax (use of), 220, 280–292Denton, Manchester (glassworks site), 77Der Blau Reiter (the blue rider, glasspainter’s group), 57Derm-O-Plast SG Normaal (PVC), 219, 245De la Pirotechnia, 146Desalination of glass, 250–253, Fig. 7.5

monitoring of, 251–2, Fig. 7.6Desolv 292 (solvent), 279Detergents, 200Deterioration of glass:

atmospheric pollution, 190, Fig. 4.13blackening of the surface, 186, 193, Fig. 4.24, 4.27deterioration layers, 240, Fig. 6.3dulling, 183, Fig. 4.19encrustations, 171, Fig.4.8, 4.9, 6.1flaking, 183iridescence, 260milky or enamel like weathering, 184, Fig. 4.22previous treatment, 194, Fig. 4.4, 4.10, 4.11, 4.12stains and residues, 171, Fig. 4.6surface pits and crusts, 190–1, 196spontaneous cracking, 183–4, Fig. 4.20, 4.21

Devitrification, 11Devitrite, 3Dextrins, 205Di pinto freddo (cold painting), 113Diamond point engraving (decorative technique), 31, 33,

42, 96, 115, Fig. 3.16Diatreta (cage cups), 96, 308Dichloromethane (solvent), 272, 279Dichroism (optical colour effect), 8, 237Diderot (author of Encyclopédie, illustrations of

glassmaking)150, 164, 165, Diploica lanescans Ach., (micro-organism), 193Displaying glass, 309–11Divalent alkaline earth cations, 5Dollond (optical glass maker), 117Domes (glass, repairing), 290–2Dor, Palestine (modern Israel ) (glassmaking site), 143Dow Corning Clear Seal (mastic), 257Dowelling (repair technique), 281, 284, Fig. 7.27

Dresden, Germany (glass engraving in), 36Drottengina af Sverige (shipwreck, glass from), 187Dry artists’ pigments, 284DSIMS (dynamic secondary ion mass spectrometry), 238Dublin, Eire (glassmaking in), 45, 68, 166Duco cement (cellulose nitrate), 211, 263, 275Due Lettere di Signor Marchese Scipione, 150, Fig. 3.9Maffei (furnace illustration in), 150Dunstable Swan, The (enameeled brooch), 52Durofix (cellulose nitrate), 211, 263, 275Durol-Polier-Fluat SD (fluoro-silicate), 269Dyers Cross, Sussex (glassmaking site), 64

Edkins, Michael (glass enameller & gilder), 113, 115Edison (glass light bulbs), 117Edrei , Syria,EDTA (ethylene diamine salt), 194, 200, 201–2, 263EDS (energy-dispersive X-ray spectrometry), 234Efflorescence (Egyptian faience), 84Egermann, Frederick ( glassmaker), 37Eglomisé, 54, 58, 340Egyptian Blue (pigment), 2, 3Egyptian:

beads, 82, 85–6, 319, Fig. 3.2, 3.3deterioration & conservation of, 313–4, 315faience, 2, 81, 83–5, 242, Fig. 7.53, 7.54glass, 19, 138–141, 191-Islamic glass, 7, Fig. 1.6

Ehlipakku (Ancient Egyptian raw glass), 139Eichensfeld, Germany (glassmaking in), 79Eigelstein, Germany (glassmaking site), 142El Djem, Tunisia (mosaic depicting glass), 102Electro-float (glassmaking technique), 130Electron beam analysis, 233, 234Elizabethtown, Pennsylvania, USA (glasshouse), 46Ellison Glassworks, 44Emery (polish), 97Emission spectrography, 239EPMA (electron probe micro-analysis), 235Enamels (vitreous coatings on metal), 1, 3, 51–3,

118–123, 234, 310, 317, 321–2, Fig. 3.33backing(counter -), 50

bassetaille, 51, 52, 121, Fig. 3.37champlevé, 51, 52, 119, 122, Fig. 3.36cloisonné, 51, 52, 62, 119, 122, 125, 322, Fig. 3.34counter, 50damage to, 321–2, Fig. 7.57, 7.58dipped, 119en plein, 52en resille sur verre, 52, 122en ronde bosse, 52, 120, 122en taille d’epergne, 119filigree, 120lavoro di basso relievo, 52–painted, 122, 237, Fig. 3.38plique à jour, 51, 52, 120, 322, Fig. 3.35Surrey enamelling, 52

Enamelling (vitreous paint on glass), 42, 43, 113decay of, 192, Fig. 4.29

Enclyclopédie (glass manufacturing), 150, 151, 164, 167Engraved glass, 37, 42

-gold glass, 54Ennion (Sidonian glassmaker), 23Ensell, George (glassmaker), 150

370 Index

Environment, effects on glass, 174, 182–194partially desiccated, 243dry, 243saturation, 243

Environmental control, 316–17Epotek 301-2 (epoxy adhesive), 212, 276, 343Epoxy resins, 205, 207, 211–15, 223, 276, 278, 280, 281,

284, 285, 286, 303, 319, 322, 343, Table 5.2, Fig.5.7, 5.8

EPR (electron paramgagnetic resonance) (see ESR)Eraclius (Medieval historian), 153Eridu, Mesopotamia, 138ESCA (electron spectroscopy for chemical analysis) (see

XPS)Escomb, Durham (glass from), 128ESR (electron spin resonance), 239Ethanol (ethyl alcohol), 203, 262, 270Ethazote (inert foam, packing material), 225Ethics of conservation, 242Ethyl alcohol, 262Etruria, Italy, 21Evelyn, John (diarist), 74Excavated glass, dry, wet, 243Examination of glass, 228–231, 26, Fig. 6.2Excise Act (see Glass-)

Faber, Johann Ludwig (glass decorator), 113Fabergé, Peter Carl (enameller), 51–52Facet cutting, 100–101Façon de Venise, 32, 34, 35, 37, 38, 39, 80, 96, 102,

110Façon de l’Angleterre, 41Faenza, Italy, 3Faience (see Egyptian faience)Fakes (see Replicas and forgeries)Faraday, Michael 193Ferric oxide, 6, 331Ferrous oxide, 6Ferrous sulphide, 193Fernfold, Sussex (furnace site), 154Ferramenta (in windows), 66, 336Fictive temperature, 11Filigrana (decorative technique)33, 107–109, Fig. 3.25Fire-cracking, 1.17Fission track dating, 240Flammability of solvents, 203Flashed glass, 8, 131Flat glass sheets, 130, Fig. 3.44Flint, 1Flint glass (see Lead glass)Float glass, 130Florence, Italy (glassmaking in), 30Fluorescence (in UV), 236Fluorspar (in glass), 78Fluxes (in glassmaking), 73, 118, 177Fondo d’oro (gold design), 54, 116, 124, Fig. 2.7Forest glass, 29, 37, 114, 160Forest of Darney, Vosges, France, 17Forgeries (see Reproductions and fakes)Formaldehyde, 309Formalose (moulding material), 221Formic acid, 309Fornax vitr (glass furnace), 145Fossiles harenae (Roman term for sandstones), 74

Fourcault machine (sheet glass), 118Fourier transform infrared microscopy, 253Fraches (conveyor belt), 150Fracture analysis, 232Fraunhofer Institute für Silicatsforchung317Freshwater environment, 186Freezing, soil block, 247

beads, 321Fritsche, William (glass engraver), 44Fritting, 81Fucus versiculosus (marine algae, Mn in), 78Fuel (for glassmaking furnaces), 73–40, 79, 80Fufeng, China (glass from), 28Fulgurite (natural form of silica), 1Fumed silica, 284Furnace:

ancient Mesopotamia, 137–8ancient Egyptian, 139141, Fig. 3.47Roman, 141–2, Fig. 3.49cone, 163–68, Fig. 3.21, 3.78, 3.79, 3.80northern, 151–63, Fig. 3.61, 3.62, 3.63, 3.64, 3.65,

3.74reverbatory, 146southern, 144–51, Fig.3.52, 3.53, 3.54, 3.55, 3.56,

3.57, 3.8, 3.59, 3.60tank, 142–43, 145, Fig. 3.50verrerie en bois, Fig. 3.57, 3.58

Fusing (powdered glass), 90–1Fustat, Egypt (glassmaking in), 27Fynebond (epoxy resin), 210, 212, 276

Gablonz, Czech Republic, 212, 276Gagiana (shipwreck), 30Galla Placidia, Ravenna (mosaics in), 60Galliano (pigment), 59Gallé, Emille, 48, 91, 100

-glass, 49Gallo, Andrea & Domenica (mirror makers), 62Gap filling (see Restoration)Garden Hill, Sussex (Roman window glass from), 126Gatchell, Jonathan (glass manufacturer), 45Gate, Simon (glass artist), 49Gatherer (part of glassmakers’ team), 102Gaul (Roman France, glassmaking in), 26Gawber, Yorkshire (glass cone), 166German sheet glass (see Cylinder glass)Gernheim, Germany (glass cone), 166Ghost images (of deterioration on windows), 193Giles, James (engraver & decorator), 42, 115Gilding, 23, 27–8, 33, 42, 115–7, 340Girandoles, 68, 69Girasole glass (Venetian), 235Glasborn, Spessart, Germany (glassworks site), 7Glass:

abrading95–8beads, 87–8on textiles & ethnographic materials, 319–21blowing (invention of ), 101, Fig. 3.48, 3.51bull’s eye, 62bulk (mass) glass, 176–80, 229–30crizzzling (see Weeping glass)179cold working, abrading, cutting, grinding, 95–8decoration:

abrasion, 176

Index 371

Glass (cont.)addition of glass blobs & trails, 110, Fig. 2.8, 2.9

tooling, 110, Fig. 3.27Davenport’s patent glass, 111enamelling (see Enamelling)gilding (see Gilding)grinding, 95–8ice glass, 111

definition of, 1, Fig. 1.3, 1.4, 1.5disease (see Weeping glass)Excise Act of, 42, 45, 70forming, 81, 86fritting, 81fusing, 319grinding, 95–8, Fig. 3.8making, 81, 86melting, 86of flint (lead glass), 40paste (pâte de verre, pâte de riz), 91seals (Roman), 236Sellers Company, 39, 40sick (see Weeping glass)surfaces, 176, 195–6, 228–9, Fig. 4.31transition temperature (see Viscosity)waste, 182, 234weeping179

Glassfibre surfacing tissue, 285matting, 295

Glassine (silicone paper), 265Glassmaking families (medieval), 17, 41Glassworkers’ tools, 102, Fig. 3.22Glastonbury, Somerset (glassworks site), 153Glaze, 2, 3, 139–41, 322, Fig. 1.2Glazing (windows), 63, 317Glenluce, Scotland (glass beads from), 85Glory Hole (part of furnace), 129, 144, 148Glyptic (see Cold-working techniques)88Gnalic Wreck (Yugoslavia), 30Goddard’s Silver Dip and Foam, 331Gold ruby glass, 8, 9Gordion, Turkey, 79Gob-fed machine (for glassmaking), 118Gotha, Germany (glass engraving in), 36Graal glass (decorative), 49Grand feu (enamelling furnace), 51Gravel Lane, London (glassworks site), 157, Fig. 3.66, 3.67Great Casterton, Leicestershire (Roman Site), 184Great Trade Exhibition (of 1851), 43Greenhouse effect, The, 43Greener & Company (glass manufacturers)44Greenwood, Frans (stipple engraver), 97Grey Eye Temple, Lebanon, 83Grinding (see Cold working)Grisaille (painting effect), 52, 56, 61, 65, 132Grünenplan, Germany (glassworks site), 159Grozing66Guide de Verrier (Guide for the glassmaker), 165Guildford, Surrey (glassmaking site), 64Guinand (glassmaker), 117Gumley, John (glass seller), 69Guilloche, 51Gurgan, Mesopotamia, Gypsum (weathering product), 190

Haaretz Collection of Glass, Israel, 27

Hackle marks (from fracture), 15Hadera, Israel, 142Haematinum (red glass), 10Hald, Edvard (glass artist), 49Hall-in-Tyrol, Austria (glass decoration in), 96, 115Halsinelle (hooks or supports), 147, 148Hallorenglas (type of medieval beer glass), 112Hama, Syria101Hambacher Forest area of Germany (glassmaking sites),

142Hamwic (Southampton), 75,Hampton Court Palace, Surrey (chandeliers in), 323, 333Han Dynasty, China (glassmaking in), 28, 178Hanover, Germany, 159, 166Hard glass (definition), 12Hardness of glass, 14Hartfield, E. Sussex (widow glass from), 62, 127Harrapan faience, 313Hartley, James (glassmaker), 130Hasanlu, Anatolia,Haughton Green, Manchester (furnace site), 164, Fig.

3.77Heating glass, Heemskirk, Willem von (glass engraver), 38Heindert, Luxembourg (furnace site), 157Helgö, Sweden,Helmback, Abraham (glass decorator), 113Hellenistic glass, Henley-on-Thames, Berkshire (glassworks site), 40Heart, Afghanistan (glassworking site)74Henriksdorp, Sweden (furnace site), 163Heraclius (Roman historian), 116Hilsborn, Grunenplan, Germany7Hill, John (glass manufacturer), 45Hinterglasmalerei (reverse painting on glass), 33, 53, 54,

55, 56, 57, 58, 114, 270, 39–344, Fig. 3.31, 7.71,7.72

HMG (cellulose nitrate adhesive), 211, 263, 267, 275,278, 279, 301, 307, 313, 318

HMS Sapphire (Newfoundland wreck), 193Hochschnitt (cutting technique), 65, 98, Fig. 3.18Hofkellereiglas (type of beer glass), 112Hope, Sir John (diarist), 116Hot melt preparations (moulding), 221HST (heat shrinkable tubing), 311Humidity (effect on glass), 174, 182, 191, 197, 315Humpen glasses (Bohemian beer glasses), 35, 112, 114,

162Hutton, Yorkshire (furnace site), 157HXTAL NYL-1 (epoxy adhesive), 212, 215, 307, 322Hyalith (black) glass, 37, 77, 88–90Hyaloplastic (see Hot working techniques)Hydration rind dating, 240Hydrochloric acid, 203, 262, 274Hydrofluoric acid (for glass etching), 101, 123, 203, 274,

308Hydrogen glass, Hydrogen peroxide, 202–3Hydrogen sulphide, 186Hydroxybiphenyl (biocide), 203Hydroxycarbolic acids (chelating agents), 202, Fig. 5.4Hyttekaer, Denmark (furnace site), 163

IBSCA (ion beam spectrochemical analysis), 240–1

372 Index

Ice glass (decorative technique), 49, 111, Fig. 2.14, 3.28ICP-OES (inductively coupled plasma optical emission

spectrometry), 237

ICPMS (inductively coupled plasma mass spectrometry),237

Igel (‘hedgehog’ shape) glass, 30Immersion techniques for refractive index, 232Impact cone (from glass fracture), 15, 233, Fig. 1.12Impact damage, 170, Fig. 1.12, 1.13IMS (Industrial Methylated Spirit), 225, 250, 261, 272,

314, 331, 334, 335Index of dispersion, 14Infra red (IR) photography, 231Inlays in furniture, 18Inorganic glasses, 1Insoluble salts (identification of), 252–3Institute of Nautical Archaeology (American), 82, 187,

257Intaglio engraving, 27, 98, 100Ions, and ion exchange (see De-alkalisation)Iridescence:

natural, 174–5, 183, 185, 260, 269produced artificially (decorative), 115

Irish glass, 44–6, 100Iron (in glass), 76, 186, 223, 319Iron-manganese amber glass, 8, 73IRRS (Infra-red reflection spectroscopy), 226, 240Ishgar (plant ash), 74Isinglas (fish glue), 318Islamic glass, 26, 27, 28, 115Isotopic ratio analysis, 238Ivory beads, 319Jacobite glasses, Fig. 3.17Jacobs, Lazarus (glass gilder), 115Jamestown , Virginia, USA (glasshouse site), 46, 147Japanese mulberry tissue, 270Jarrow, Northumbria (Saxon glassworking), 75, 130, 184,

337Jeanette Glassworks, Pennsylvania, USA, 48Jemdet Nasr phase (Mesopotamia), 18Jerash, Palestine (window glass from), 128Jesse Tree Window, Regensburg Cathedral, Germany,

132Jones, Robert (Regency designer), 71Josephus (Roman historian), 16Junkersfelde, Westphalia, Germany (furnace site), 163

Kaltmalerei (cold painting), 113Kandinsky, Wassily (glass painter), 57Kangxi, Chinese emperor (glassmaking under), 28Karanis, Egypt, 102Karlova Hut, Czech Republic, 158Karlsruhe, Germany, Kas, Turkey, (Bronze Age site), 187, 257Kassel, Germany (glass engraving in), 36Kempak (packing material), 312Kenchreai, Corinth, Greece (opus sectile), 61, 124–5,

187, 253–7Khorfush (Egyptian, vitrified material), 141Kimble Glass Company, 47Kimmeridge, Dorset (glasshouse site), 157–8, 163, 164,

Fig. 3.69King’s Lynn, Norfolk (source of sand), 40

King Sargon of Assyria (aryballos), 95Kirdford church, Sussex (windows in), 337Klucel C (hydroxyl-propylcellulose), 341Knightons, Alfold, Surrey (glassworks site), 155, Fig. 3.64Knops (decoration on stem of wine glasses), 32Knossos, Crete, 138Kny, Frederick (glass engraver), 44Kongsberg, Norway (chandeliers from), 71–2Kraron G1650 (plasticizer), 342Krautstrunk (cabbage stalk prunts), 30Kreybich, George Franz (glass engraver), 100Kubu images (Assyrian glassmaking), 136Kunstammlungen der Vest Coburg, Germany, 37, Kunckel, Johann (glassmaker), 8, 36, 150, 159, 160, 163,

164, 168kiln of, Fig. 3.71, 3.72

Kurfurstenhumpen, (decorated thin beer glasses), 35Kuttrof (German beaker), 35Kythera, Island of, Greece (shipwreck), 187

Labino, Dominic (glassblower), 49Lacrymae Batavicae (see PrinceRupert’s drops)La Farge, John (window glass maker), 65La Grandja de San Idelfonso, Spain (glass decorating

in), 115La Tene I Period, 88Lacquers, 204–6, 269, 319Lalique, René (glass maker), 49Lambeth, London (see Gumley, John)Lampshades, 67–8Lampworking, 108, Fig. 3.26Lanterns, 67–8Lapiz lazuli, 81Laser cleaning (ablation), 274Larmes de verre (see Prince Rupert’s drops)Latticino (latticinio) (decorative technique), 107Lattimo (decorative technique), 49, 57, 107Lavoisier, A .J.174Leaching (see De-alkalization)Le Mans Cathedral, France (windows in), 63Leading (in windows), 66, 131, 192, 336

lead-line (windows), 65Lead in glass, lead silica glass, 9, 10, 12, 40, 47, 134,

176, Fig.1.8arsenate, 235basic lead carbonate, 80lead crystal glass, 100lead oxide, 60, 74, 76, 80lead, red, 76, 80lead sulphide, 262lead, white (basic lead carbonate), 76, 80litharge, 9, 80

Lead isotopes in glass, 238Lechatelerite (a natural glass), 1, 2Lehman, Caspar (glass engraver), 36Lehr (or Leer) (annealing furnace), 149, 150, 151Lemington glass cone, Newcastle-on-Tyne, 166Libbey, E.D. (glass manufacturer), 47LIBS (laser induced breakdown spectroscopy), 241Lichens (on glass), 193Lepraria flavia Arch. (micro-organism), 193Liège, Belgium (glassmaking in), 30, 37Lifting materials (on excavations), 224Lime in glass (see Calcium)

Index 373

Limoges, France, 52, 122, 321Linnet-hole, 160Liquidus temperature, 3Liquitex (acrylic pigments), 257Lissapol (non-ionic detergent), 200Litharge (lead oxide), 9, 76Lithium (in glass), 8Littleton, Harvey (glassmaker), 49Lobmeyer, J.J. (glass manufacturer), 45Loetz glassworks, Czech Republic, 48London (Roman glass from), 26London Gazette, The, 166Longe, George (glassmaker), 45Lorraine, France, 38, 62, 63Lorsch, Germany,Lost wax process (cire perdue), 91, Fig. 3.9Low Countries (glass from), 36, 38, 62, 113Lubbers glassmaking machine, 128Lumella (occhio, furnace hole), 149Lustre painting, 27, 114–5Lycurgus Cup (diatreta), 8, 96, 237

Magnes lapis (Roman term for limestone), 74, 79Magnesia in glass, 5, 76, 77, 180, 186, 223

magnesium oxide, 5, 309chloride, 317

Malachite (copper corrosion), 83Mancetter (UK Roman furnace site), 26, 142Mandeville, Sir John (illustrator of glassmaking), 153,

157, 162, 168Manganese in glass, 75, 76, 77, 78, 86, 186, 223

as decoloriser, 7, 76dioxide, 77, 78

Mannheim, Pennsylvania, USA (glasshouse), 46Mansell, Sir Robert, 38, 39Månsson, Peder (writer on glassmaking), 143Maranyl (soluble Nylon), 211Manufacturing defects (in glass), 170Marble Hill House, London (glass fanlight in), 63Marcasite (in mirror making), 134Marking glass, 258–9, Fig. 7.10Marlik, Mesopotamia (mosaic glass from), 93Marine environment, glass in, 186–190, Fig. 4.26, 4.27

fouling organisms (on glass), 189, Fig.4.28Marmot, Maurice (glassmaker), 49Marvering (shaping hot glass), 89Masonite (support), 225, 226Mass spectrographic analysis, 233, 237Matting (painting technique), 192Mattioli, Alessio (mosaic maker), 50MDF (medium density fibre board), 333Medici, Francesco I de’ (studio wallpainting depicting

glass), 148Melinex (polyester film), 335, 336, 344Mekku (ancient Egyptian raw glass), 139Melting point, 3Menorah, 24, 25Mercury (amalgam for mirrors), 133–5Merrett (translator of Neri), 149, 156–7, 166, 167Mesopotamia (glassmaking in), 17–9, 27, 135–138, 177,

234transparent glasses, 137opal glasses, 138

Metal corrosion deposits on buried glass, 172, 317

Metal ‘staples’ (bridges for support), 278Metallic oxides (see Colourants for glass)Methanol, 203Methyl methacrylate, 216–7Metropolitan Museum of Art, New York (glass in), 308Mezzotints, 58–9Michelfeld, Austria (glass reliquary from), 30Micromesh (cushioned abrasives), 290Micro-mosaics, 49–51, 52, 322, Fig. 2.17Micro-organisms on glass, 193, 203Microscope (use of), 228, 308Middlewich, Cheshire (glassmaking site), 142Mildner, Johann Josef (glass decorator), 117Millefiori glass (decorative), 33, 52, 88, 94–5, 123, Fig.

2.13, 3.13, 3.14Milliput (epoxy putty), 284Millville, New Jersey, USA, 47Mineral acids, 203‘Mirror area’ (of fractured glass), 15Mirrors, 62, 133–5, 317, 332–6, Fig. 3.46, 7.65, 7.66

painting on, 54, Fig.7.67, 7.72Modelling materials, 220Modifiers (see Network modifiers)Mohn, Samuel (father & son, glass decorators), 113Mohs’ scale of hardness, 14Molar glass compositions, 180–1, Table 4.1, 4.2Monks Riborough, Buckinghmshire (church windows), 337Monkwearmouth, Durham (monastic window glass), 128Mombasa, S. Africa (ship-wreck, glass from), 249Monte Casino, Italy (library of), 144Monte Lecco, Italy (glassmaking site), 145, 146Morris, William (arts & crafts designer), 65, 100Mortar (glass embedded in), 172Mosaic glass, 18, 21, 22, 52, 91, Fig. 2.5, 3.10, 3.11Mosaico in piccolo (see Micromosaics)Mosaic tesserae, 30, 54, 59–60, Fig. 3.39, 3.42Mosque lamps, 27, 116, Fig. 2.11Mould:

blowing, 105–9flash lines, 308making materials,

dental wax, 286–92, Fig. 7.33, 7.34, 7.35, 7.36silicone rubber, 292, Fig. 7.37, 7.38, 7.39, 7.40

piece mould, 299–301, Fig. 7.43pressing, 286

Mount Carmel, Palestine, 16Mount Washington, USA (glass factory), 44Mounting fragmentary glass, 284, Fig. 7.28, 7.51Mowital B20H (PVB), 219Muff glass (see Cylinder glass)Muffle kiln, 52, 132

(see also Petit feu)Mulvaney, Charles (glass manufacturer), 45Münter, Gabriele (glass painter), 57Murano, Italy (glassmaking in), 30, 38, 49, 52, 80, 94Mutissu (Assyrian); mattara (Syrian), rake, 137Mycenaean glass, 18, 87

N-methylmethoxy Nylon (see Soluble Nylon)Na’aman, river (see Belus, river)NAA (neutron activation analysis), 236–7Naga plant (Assyrian source of alkali)136Nailsea Bristol (glassworks), 77Namur, France (glassmaking in), 74

374 Index

Navarre, Henri (glassmaker), 49Natron (Egyptian source of alkali), 74Nef (vessel type), 33, Fig. 2.12Nehou, Louis Lucas de (glassmaker), 62Nejsum, Denmark (furnace site), 163Neri, Antonio (glassmaker and author), 81, 149, 150,

160Netherlands (glass engraving), 115Network dissolution, 175–6Network formers and modifiers, 3, 4, 73, 176, 171, 177Netzglas (decorative),Nevers, Frances (glass figures from), 317New Bremen, Maryland (furnace site), Newcastle upon Tyne, Northumbria,

glass from, 37, 38, 41, 42, 43, 113source of sand, 40

Nickel oxide (millerite, in glass), 78Nimrud, Mesopotami (glassmaking in), 27, 79, 138Nineveh, Mesopotamia, (glassmaking in), 27, 174Nishapur, ancient Persia (Iran), (glassmaking in), 27Nitrate ion, detection of, 252Nitric acid, 203, 314Nitromors (solvent), 279

Superstrip, 279Nitrum (Roman term for soda), 74Non-bridging ions, 5, 175, 253Norland Optical (UV setting) adhesive, 217Norman slabs, 66, 129, Fig. 3.43North American glass, 46–8Northwood, Harry, 48

John, 44, 48, 99Nostangen, Norway (glass furnace), 167Notsjö, Finland (glass furnace), 159Nuppen (German, prunts), 30Nuremberg, Germany (glass engraving in), 35, 36, 37, 57Nylon fishing line, 310, 311

Obernkirchen, Germany (glass cone), 166Obsidian (volcanic glass), 1, 15, 16, 319, Fig. 1.1Ochsenkopfglas (type of enamelled beaker), 112Occhio (lumella), eye, 136–7OEL (Occupational Exposure Linits), 203Offhand blowing (glassforming procedure), 102Oil gilding, Oinochoe (jug), 21, 304–6, Fig. 2.2, 7.47Olympia, Greece (glass from), 90Ontario Glass manufacturing Co., Canada, 47Onyx glass (artificial opal), 76, 236Opacifying agents, 10, 223, Table 1.1Opalescent enamel, 51Opnaim’s red glass, Oppitz, Paul (glass engraver), 44Optical methods of analysis, 237Optical properties of glass, 13Opticon UV57 adhesive, 217Opus sectile, 23, 54, 58, 60, 123, 124, 125, 126, 187,

253-7, Fig. 3.40, 3.41Oratorio of St Rocco, Altare, Italy (picture of a

glassmaking furnace), 145Organic solvents, 199, 203Organo-silanes (see Silanes)Orrefors, Sweden, 49Osler, T&C (glass manufacturers), 43, 71Ostia, Sicily, 124

Overshot glass (see Ice glass)Owens suction glassmaking machine, 118Oxidation, 6Oxidation state methods of analysis, 233, 238Oxonium ion (H3O+), Oxyfume 12 (fumigant), 256Oxygen isotopes in glass, 238Packaging:

materials, 225–6methods, 244, 312–3, Fig. 7.52archaeological glass (see Storage)

Padua, Italy, 56Paillons (decoration beneath enamel), 51Painted enamel (on metal), 51Paintings on mirrored glass, 58Painted and stained glass windows, 130–3, 237Pala d’oro, St. Marks, Venice, Italy, 52Palm House, Kew Gardens, London (Glass panes in), 63Palmo (unit of measurement), 146Palmyra, Syria, 101Panacide (disinfectant), 310Paperweights, 94–5, 106Paraison, (gathering of molten glass), 103Paraloid:

B-48N, 311B-67314, 342B-72211, 219, 250, 266, 267, 268, 276, 285, 311, 312,

313, 314, 317, 318, 336, 342, 344B-99314

Particle methods of analysis, 236Passementerie, 69Passglas (vessel type), 30, 35Passive conservation, 242Pâte-de-riz and Pâte-de-verre (see Glass paste)Pattern moulding (technique), 105Paviken, Sweden,Peach Bloom (decorative), 48Peckitt, William (glassmaker), 133Pearlised glass (decorative), 319, 320PEG 4000 (polyethylene glycol) wax, 267Penarth, S. Wales (source of sand), 141?Penrose, George & William (glass manufacturers), 45Perrot, Bernard (glassmaker), 61Perry & Co. (chandelier makers), 70, 71Perspex (US Plexiglas, acrylic sheet), 242, 292, 302, 304,

314Pertusaria Leucosora Nyl (micro-organism), 193Petit feu (enamelling kiln), 50, 119, 132Pettenkoffer Process, 342Pfauen Island, Berlin, Germany (glasshouse), 9pH value, 175, 176, 189, 202, 263, 315, 331Phase-separated glass, 6Phidias (ancient glassmaker), 90Phiolarius (type of vessel), 31Phosphate (intermediate & effect of), 177Phosphorous, 4Photo-activated adhesives, 205Photographic images on glass, 54, 338–9, Fig. 7.69, 7.70Photography of glass, 231, 244Physical damage of glass, 169–71, 182, Fig. 4.2, 4.3PIGME (proton-induced gamma ray emission), 321Pilkington (glassmakers), 118, 130PIXE (particle induced x-ray emission), 236, 321Plain glazing (windows), 336–7

Index 375

Pit formation, 183Plant ash (source of alkali), 29, 86, 143Plastazote LD45 (inert polyester foam), 225, 245, 259, 3,

311, 334Plaster of Paris (gypsum), 221, 225, 247, 294, 299–302Plasticine (US Plastilina, modelling compound), 220, 301Plate glass (patent plate, improved cylinder glass), 130Plastogen G (polymethacrylate), 223Pleichroism (displays different polariscope colours), 229Plexigum 24 (polybutyl methacrylate), 217, 269Plexisol P550, 342Plextol D466 (acrylic resin dispersion), 342Pliny the Elder (Roman historian), 16, 22, 59, 74, 77, 79,

82Po, River, Italy, 20, 24Pocatky, Czech Republic, 154Pokal (vase type), Polariscope, 229Poli, Flavio (glassworker), 49Polyethylene glycol wax, 224Polyester resins, 205, 215, 220, 222, 276, 285, 286, 297,

Fig. 5.9Polythene (self-seal) bags, 244, 260, 265Polythene sheet, 248Polyethylene foam, 225Polymers, deterioration of, 208Polymethyl methacrylate, 208, 216–7, 314

resins, 222–3Polyphosphates, 201, Fig. 5.2Polyurethane foam, 224–5, 247Polyurethane resin, 205, 218, 247–8Polyvinyl acetate, 208, 219, 224, 225Polyvinyl alcohol, 287, 289, 295, 299Polyvinyl chloride, 208, 219, 224, 225Pompeii, Italy (glass windows from), 126Pontil (glassformer’s tool), 102Portland Vase (cameo glass), 99, 273, Fig. 7.20, 3.19Potash (in glass), 7, 73Potash glass, 118, 143, 175, 176, 177, 180, 183, 186, 233Potassium, 317Potassium carbonate, 81Potassium hydroxide, 262Pot-coloured glass, 76, 131Potash glass, 118, 143, 175, 176, 177, 180, 183, 186, 233Potichimanie (decorative technique), 54Potsdam, Berlin, Germany, 36Potters’ clay (modelling material), 220, 297–8Powell, J. (see Whitefriars Glassworks)Precipitation (effect of), 182Preissler, Ignatius (glass decorator), 37Press-moulded glass, 47–8Pressure sensitive tape,

damage by, 172use of, 225, 276, 284, 303, 343

Preussler, Christian (glass enameller), 162, Fig. 3.75Primal (acrylic emulsions),

AC-235, 267, 270WS-12, 267, 270WS-24, 269WS-50, 267, 270

Prince Rupert’s Drops, 12Proclamation Touching Glass (of 1614), 38, 63, 163Profilometer (flatness measuring device).240Propan-2-ol (hydrophilic solvent), 261, 270

Prophet windows (see Augsburg cathedral)Protons, 175Prudentius (Roman poet), 62Prunted beakers, 30Ptolomais (‘Akko), Palestine, 16PTFE (polytetrafluoroethylene), 222Pumice (volcanic glass), 1, 97Punty (pontil) (glassmakers’ tool), 102Purple of Cassius, 36PVA release agent, 303PVA-AYAA (conslidant), 270, 342PVAC (polyvinyl acetate), 208, 249–50, 255, 275PVAL (polyvinyl alcohol), PVB (polyvinyl butyral), 262PVC (polyvinyl chloride), 208, 219, 224, 225, Fig. 5.14Pyrex glass, 118

Qantir, Egypt, 139Qing Dynasty (glass from), 316Quarry (window pane), 337Quartz (source of silica), 1, 3, 4, 14, 29, 71, 81, 97, 139Quentglaze (polyurethane), 247

Radiation monitoring of potash glass, 238Radiocarbon dating, 240Ramla, Israel (glassmaking site), 143Ramos, Felix (glass engraver), 65Raqqa, Syria (glassmaking site), 143Ravenna, Italy (mosaics from), 123Ravenscroft, George (glassmaker), 10, 39, 40, 178, Fig.

2.16Reamer (glassworkers’ tool), 104Rebel 2000 (paints), 344Recording and documentation of glass, 227–8Red glass, 8, 9Reduced pressure (vacuum impregnation), 265Reduction, 6Refractive index, 207–8, Table 1.2, 1.11

measurement of, 231–2, 281Regency Period (in Britain), 70Reichadlerhumpen (type of beaker), 35, Fig. 2.15Reichenau, Austria (glasshouse), 162, 163, Fig. 3.76Regalrez 1094 (resin), 342Rejdice, Czech Republic (glasshouse), 154Relative humidity (RH), 183, 184, 315, 316, 317Release agents, 221–2, 300Removal of glass,

from the ground, 244–immobiliztion of glass, 245immobilization of a soil block, 246, Fig. 7.2, 7.3freezing a soil block, 247, 249from a marine or freshwater environment, 249

Removal of glass kilns, 248Removal of previous repair/restoration materials, 271–3,

Fig. 7.18, 7.19, 7.20, . 7.23, 7.24, 7.25, 7.26Repair:

of archaeological glass, Fig. 7.13of historic glass, 275

Replicas and fakes, 308–9, Fig. 7.48, 7.49Repton, Derbyshire (glass from), 128Re-setting excavated window glass, 337Restoration of glass, 271, 284–308

commercial, 308detachable gapfills304–306, Fig. 7.47

376 Index

filling losses in opaque glass284, Fig. 7.29, 7.30filling tiny losses284–5partial restoration285, Fig. 7.31gapfilling with casts from moulds taken off the

original glass, 286–290, Fig. 7.32, 7.33, 7.34gapfilling with casts from a mould taken from a clay

model or a previous restoration, 297–303, Fig.7.41, 7.42, 7.43, 7.44

gapfilling where the interior of a vessel is inaccessiblefor working, 303–304, Fig. 7.45, 7.46

Reticelli, 75, 94, Fig. 3.12Retouching paint, 343Retouching lacquers, 223–4Reverse paintings (on glass), 54–56, 339–344, Fig. 7.71Reverse foil engraving, 56, Fig. 2.18Reversibility of polymers, 208–9Rhineland, Germany (glassmaking in), 25Rhodes, island of, Greece, 83, 257Rhodorsil 11504A (silicone rubber), 296Rib marks (from fracture), 15Rice Harris & Sons (glass manufacturers), 44Richardsons (glass manufacturers), 44Rigger (sable tracing brush), 131Rishpon, Palestine, 102Riss (painting outline), 55River Euphtaes (Iraq), 137Rivenhall church, Essex, Roche, Captain Philip (glassmaker), 166Rock crystal (quartz), 13Roemer glasses (beakers with large bowl), 35, Fig. 2.15

Roman glassmaking, 141–2(see also Glassmaking in the Roman Empire)

Rosedale, Yorkshire (furnace site), 157, Fig. 3.68Rosenborg Castle, Copenhagen (mirrors in), 333Rouen Cathedral, France, 65Rousseau, Eugène, 48Royal Gold Cup, The, 121Rummer (large beer glass), 35Rüsselbecher (claw-beaker), 29Rustin’s Clear Plastic Coating (urea formaldehyde), 223

San. Apollinare in Classe, Ravenna (mosaics in), 60San. Cosmo, Rome (mosaics in), 60San. Damiano, Rome (mosaics in), 60St. Chapelle, Paris (windows in), 63St. Cuthbert’s Window, York Minster, 132St. Gobain, France (glass manufacturing), 62St. John’s Church, Stamford, Canterbury (jewels in

windows), 132St. Louis, France (glass paperweights from), 94St. Michael’s Church, York (window in), 132St. Weonard’s, Herefordshire (glassmaking site), 155Sabellico, Marc Antonio, 94Sadana Island, Red Sea, Egypt, 187Sala, Jean (glassmaker), 49Salicornia kali (source of alkali)74, 85Salsola kali (source of alkali), 85, 136Samara, Mesopotamia (Iraq, glassmaking in), 27San Idelfonso, Royal Palace of, Madrid, 65San Marco, Venice, 17San Sophia, Constantinople, 60San Vitale, Ravenna (mosaics in), 30, 60, 63, 129Sand blasting, 101Sand moulding109

Sandwiched gold leaf, 21, 116–7, Fig. 2.4, 3.32Sang, Jacob (glass engraver), 36, 38Santocel (fumed silica), 292Saranwrap (PVC film), 225, 247Sanfirico (see Zanfirico),Sardis, Turkey (glass from), 270Sargon, king of Assyria (alabastron of), 95Sarosel room (part of furnace complex), 150Sassanian period (Islamic glass), 97SA/V (surface area/volume ratio), effect of, 176, 197Schaper, Johan (schwarzlot enameller), 35, 112Schedula Diversum Artium (Treatise on Diverse Arts),

152Schurterre, John (glassmaker), 64Schwannhardt, George (glass engraver), 36Schwartzlot (black enamel decoration), 35Science Museum, London (model in), 151, 152Sconces (wall lights), 68Scopas (release agent), 222Scotch Magic Tape 810, 276, 343Scotchtint (solar refelcetive film), 335Scratches, effect of, 184Seawater (see Marine environment)Seed (bubbles in glass), 80, 81, 170Selenium in glass, 78Sellotape, 307SEM (scanning electron microscope), 234, 308SEM-EDS (energy dispersive X-ray spectrometry), 234–5SEM-WDS (wavelength dispersive X-ray analysis) (see

EPMA)SEM/EDXRA270Sequestering agents (see Chelating agents)Sepiolite clay, 201Serçe Liman wreck (Turkey), 28, 82, 83, 257Servitor (part of glassformer’s team), 102, 149Shading (glasspainting technique), 192Shakenoak Farm, Oxfordshire (Roman site), 127Shellac, 204, 205, 222, 330Sidon, Syria, 22, 26, 30, 102Siebel, Franz Anton, 113Siemens regenerative furnace, 117Silanes, 205–7, 218, 266, 315, Fig. 5.5, 5.13Silastic E RTV (silicone rubber), Silcoset 105 (silicone rubber), 296Silesia (glass cutting in), 171Silhouettes on glass, 54, Fig. 2.19Silica in glass, 3, 73, 74, 138, 176, 177, 223Silica gel, 315, 316Silicate deposits on glass during burial, 171Silicone release agent, 222Silicone rubber (moulding material), 218, 221, 292–303Silver nitrate, 133

in mirror making, 134Silver stain, 64, 65, 66, 130–3, 192SIMS (secondary ion mass spectroscopy), 236, 237Siraf, Persian Gulf (glassmaking centre), , 27Skenarice, Czech Republi (furnace site), 154Slagelse, Denmark (glass from), 268Smalti (cakes of glass), 50, 59Smalti filati (glass threads), 50Smear shading (painting technique), 66, 192Smithsonian Institution, Washington, 316Snake thread trailing (decorative technique), 25Soda crystals, 74

Index 377

Soda glass, 73, 74, 81, 83, 86, 118, 143, 175, 176, 177,180, 183, 231, 233

Sodium bicarbonate, 86carbonate, 74, 81, 82, 86, 137chloride, 74, 86nitrate, 74sesquicarbonate, 74sulphate, 74

Solarisation, 193Soluble nylon (consolidant), 211, 313Soluble salts, effects of, 187Song Dynasty, China (glassmaking in), 28Solvol Autosol (fine abrasive paste), 290Specific gravity, 14Specula (blown glass window), 61Springing (of glass), 170, 280Sprungli, Hans Jacob (glass painter), 57Stabilisers (see Network modifiers)Staffordshire, UK (glassmaking in), 79Stained glass, 130, 171, Fig. 3.45Stangenglas (German beaker type), 35, 112, Fig. 2.15Stassfurt, Germany (source of potash), 80Steigel, ‘Baron’ (glassmaker), 46Steatite, 81Steinkrug, Hanover, Germany (glass cone), 166Stenhult, Denmark (furnace site), 163Stevens & Willimas (glassmakers), 44‘Sticky Wax’ (see Waxes)Stipple engraving (decorative technique), 38, 97Stipple painting, 66Stonea, Cambridgeshire (Roman window glass from), 62,

126Stones (inclusions) in glass, 81, 170, Fig. 4.1Stones of Venice (John Ruskin), 308Storage of glass:

dry or treated, 259–60, Fig. 7.11and display of historic glass, 309–311of paintings in reverse on glass, 344weeping glass, Fig. 7.50wet glass, 244–5, Fig. 7.1, 7.4, 7.16

Stourbridge, Worcestershire (glassmaking in), 43, 44, 166Strabo (Roman historian), 78Strain and stress (caused by adhesives), 207Strain detection of, 232Striae (cords seen in glass), 170, 229, 230‘Striking’ (of colour in glass), 8Sumerian words for glassmaking (Akkadian trans.),

135–6abnu (glass mix)137ahussu(plant ash)136anzahhu (b-su)138bab kuri (door)137bilhu glass mix)137bit kuri (house of furnaces)136dabtu(crucible)137immanakku (ground quartzite pebbles)136kuri sa ibni (furnace), 136kuri sa siknat enait-pel-sa (furnace), 136kuri sa takkanni (furnace arch), 136mutirru (mattara, Syrian, rutabulum, glassmakers’

rake), 137naga plant (plant source of soda), 136nimedu (stilt), 137sipparuarhu (fast bronze)136

su’lu (ladle), 137taptu zakatu (crucible), 137urudu hi (slow copper), 136zuku (frit), 135, 136

Sulphate ion, detection of, 252Sulphur deposits on buried glass, 171Sulphur dioxide, 190Sulphur-reducing bacteris, 186, 188Sulphuric acid, 101, 203Superficial disfigurement of glass, 171–3Supporting glass during repair, 276, Fig. 7.21, 7.22Surrey enamelling (see Enamelling)Susa, Iran, 21Sussita, Palestine (ancient Hyppos), 102Sutton Hoo, Suffolk, 260Swann, Sir Joseph (glassmaker), 117Swiss enamelling (see Enamelling)Synperonic N (non-ionic detergent), 200Syriste, Czech Republic, 154Tacchardia lacca (see Shellac)Taffel offen (table oven), 162Tamerlane (Timur)112Tara, Eire,Tazza (type of bowl), Technovit 4004A (polymethyl methacrylate resin), 223,

304Tektites (natural glass), 1Tell al Rima, Iraq (glass mosaic from), 93, Fig. 3.11Tell el Amarna, Egypt, 18, 138, 139, 236Tell ’Umar, Mesopotamia?TEM (transmission electron microscope), Temperature, effect of, 197 316Tesserae (mosaic components), 23, 26, 30, 50, 59, 60,

117, 124en grisaille (micromosaic), 50

Thermal expansion, 13, Fig. 1.10shock damage, 170

Thermal history (of a glass), 10Thermodynamics, 197Theophilus (writer on glassmaking), 60, 113, 126, 132,

152– 53, 157Thetford, Essex (glass from), 129Thornham Hall, Suffolk (chandelier from), 69Thornton, John (window maker), 64Three-mould process, Thuret, André (glassmaker), 49Tiberias, Palestine, 102Tiefschnitt (engraving), 98Tiffany, Louis Comfort (glassmaker), 48, 100Tin amalgam (mirror making), 134, 333Tin oxide (cassiterite, opalising agent), 10, 77, 333,

Table 1.1Tinsholt, Denmark (furnace site), 163

monoxide (romarchite), 333Tiseur (part of the glassformers’ team), TL (thermoluminescence dating), 240TLV (Threshold Limit of Toxicity), 203Toledo cathedral, Spain (windows in), 239Toledo, Spain (glasspainting in), 57Toluene (solvent), 225Tongs (glassformers’ tool), 105Topkapi Palace, Istanbul (windows in), 64Torcello, Italy (furnace site), 123, 144Tourlaville, France (glass manufacturing), 57

378 Index

Toutin, Henri & Jean (enamellers), 62Toxicity of solvents, 203–4Trace elements in glass223Traprain Law, Scotland (glass from), 86Travels (illustrations of glassmaking), 153Trestenhult, Sweden (furnace site), 160, 162, Fig. 3.73Triangular diagram, 180–1, Fig. 4.18Trichlorethylene, 203, 225Tridymite (a form of silica), 1Triton (non-ionic detergent), 200Turquoise, 83Tut-ankh-amen (glass head-rest of), 95, 115Tuthmosis III (glassmaking under), 18, 19

glass jug of, 112, Fig. 2.1Twycross church, Leicestershire,Tylose MH 300 (methyl cellulose), 341Tyre, Syria (glasshouse site), 26

Ultrasonic cleaning, 262Ultra-violet light, 340

curing adhesives, 205, 217–8, Fig. 5.12examination of glass by, 230, 308UV-VIS (ultra-violet & visible spectroscopy), 239

Ulu Burun, Turkey (shipwreck, glass from), 82, 187,257, Fig. 3.1

Unguentarium vessel type), 25, 90, 102Universal indicator papers, 231Uranium glass, 49Uppark House, Sussex (chandeliers from), 323

Vacuum impregnation, 265, 269Vacuum tweezers, use of, 279Van Dyke (style of cutting), 70Vann Copse, Hambledon, Surrey (furnace site), 155, Fig.

3.65Vaseline (release agent), 298, 301Vauxhall, London (glassmaking in), 58Vel-Mix (gap filler), 256Venice, Italy, 17, 31–4, 65, 77, 96, 115Vermiculite (packing material), 296Verre:

craquelé111de fougère (bracken glass), 29de Nevers, 109églomisé, 54, 58, 116façon de Lorraine, 128, 158frisée, 109 verre en table, 128

Ververi Bityska, Czech Republic (furnace site), 153Verzelini, Jacob (glassmaker), 38, 39, 69, 96Vesuvius, Mount, Italy (eruption of), 24Vetro:@pi:a fili, 107

filigranato, 107a ghiacccio (ice glass), 111a reticello, 107, 108a retorti, 108corroso (modern ice glass), 111a redelexo (a redexin), 107de trina, 107

Victoria & Albert Museum, London (glass in), 310, 315,316

Vinac B-7, 342Vinamold (PVCmoulding material), 221

Vinamul 6815 (PVC emulsion), 267Vinyl polymers, 218–220Viscosity of molten glass, 10, 11–12, 81, Fig. 1.9

annealing point, 11, 12fictive temperature, 11liquidus temperature, 3softening point, 11strain point, 11, 73, 170transition point (Tg)/temperaure11, 13of adhesives, 219working point, 11

Visitello, Christopher (glassmaker), 167Visual examination of glass, 228Visscher, Anna Roemer (engraver), 38Vitre mania, 54Vitrum (Roman term for glass), 32, 79

in massa et rudus (crude lump glass), 32purum (pure glass), 79

Voltammetry of immobilized microparticles, 24Volturnus River, Italy, 16, 73Von Eiff, Wilhelm (glassmaker), 65Vosschemie (polyester) resin, 306

Wachsausschmelzfahren (see Lost wax process)Wadi Natrun, Egypt (source of soda), 74Waldglas (see Forest glass)Ware Collection of glass, The, 318Warmbrunn, Germany ( glass engraving in), 36Washing sound glass, 273Water:

cleaning with, 199–200content of glass, 243–4effect on glass, 174

Waterford, Eire (glassworks), 45, 68Waxes:

beeswax, 204, 276, 343cerecine, 276dental (for moulds), 303, 304, Fig. 7.33, 7.34, 7.35,

7.36natural, 204paraffin, 343polyethylene, 245, 266‘sticky wax’, 276, Fig. 7.21, 7, 22b

use of wax or wax/resin mixture, 343Webb, Thomas & Son (glass manufacturers), 44Wedgwood, Josiah (ceramic manufacturer), 99Weeping glasses, 242, 191, 309, 314–7, Fig. 4.15, 4.16,

7.50Weimar, Germany (glass engraving in), 36Wells Cathedral, Somerset (window glass in), 190Welwyn Garden City, Hertfordshire, 10Wemyss, Scotland (glasshouse), 166Wetland burial sites, 182Wheaton Glass Company, New Jersey, USA, 47Wheel cutting (wheel engraving), 38, 97–8, Fig. 3.17,

3.18, 3.20Whitefriers Glass (see Powell)44, 61White Spirit (solvent), 272Wilderspool, Cheshire (glassmaking site), 142Winchester House (cone furnace), 166Window glass, 61–7, 126–33, 155, 162, 235, 237, 317Windsor Castle, Berkshire, (window glass in), 62Winterthur, Delaware, USA (chandelier from), 70Wissembourg, Germany

Index 379

Wistar, Caspar & Richard (glassmakers), 46Wollastonite (devitrification product), 3Woodall, George (glass decorator), 44Woodchester, Gloucestershire, 156Woodhouse, George (glassmaker), 44Wroxeter, Shropshire (glassmaking site), 26, 197

X-ray examination methods (of glass), 233, Fig. 7.49X-radiographs, 227, 230–1XPS (X-ray photo-electron spectrometry), 255, 309

-ESCA (electron spectrometry for chemical analysis),236

XRD (X-ray diffraction analysis), 186, 235XRF (X-ray fluorescence analysis), 236X-ray tomography, 241Xylene (solvent), 272

Yellow Mesopotamian glass, 9York Minster (window glass in), 132, 133, 177, 191Yuan Dynasty, China (glassmaking in), 28

Zachariasen, W. H., 3Zanfirico (sanfirico, twisted stems), 107Zapon (cellulose nitrate), 314Zeilburg glasshouse, Bohemia, 162Zeiss, Carl (optical glassmaker), 117Želeny Brod Museum, Czech Republic (glass in),

318Zephiran chloride (antidote for HF), 203Zirconia (in glass), 78Zuytdorp (shipwreck) 270Zweizel (glasscutter), 96Zwischengoldglas, (decorative technique), 36, 117

380 Index