centennialfeature - Sglavo manufacturing.pdf · (v) To use metal molds. Ashley in Yorkshire undertook the first trials that led to success. His original machine attempted to use only

Mechanization of Glass Manufacture

Michael Cable

Glass Research Centre, University of Sheffield, Sheffield S1 3JD, United Kingdom

Mechanization of glass manufacture had just begun in 1898but was accelerating. Notable advances were soon to bemade in all branches of the industry, many of them in theUnited States, and have continued to the present day. M. J.Owens occupies a unique position as the inventor of thefirst successful fully automatic container machine in 1903;he was also involved in several other important develop-ments. However, within 20 years, machines fed with gobssupplied by the Peiler forehearth and gob feeder, alsoAmerican inventions, became successful competitors. Fromabout 1930 to 1960, a wide variety of machines was used tomake containers, but these have been superseded nearlyeverywhere by the individual section machine, anotherAmerican invention. The Lubbers process for sheet glasscylinders was an early success in 1903. About 12 years laterthe direct drawing of flat sheet was developed almost si-multaneously by Fourcault in Belgium and Colburn in theUnited States. The Pittsburgh process followed a few yearslater. Ford was the first to succeed in making continuouslycast plate, but Pilkington in the United Kingdom soon ex-ploited that process to make much wider cast plate. At thattime Pilkington also developed the simultaneous grinding ofboth faces of the continuous ribbon. The invention anddevelopment of float glass by Pilkington in the 1950s hasmade all other flat glass processes obsolete for most pur-poses. Owens–Corning has played a major role in the glass-fiber industry since its beginnings in the 1930s. The mostimportant advance in manufacture of optical glass was thesmall continuous tank with vigorous stirring developed byCorning Glass Works. Corning also made other notableadvances, such as the ribbon machine for lamp bulbs andthe lamination process used for Corelle ware.

I. Introduction

THE techniques of manufacturing glass products changedvery little between the invention of glass blowing around

50 AD and the last few years of the 19th century. However,developments then became rapid, and many were the work ofNorth American inventors and engineers. The centennial ofThe American Ceramic Society is thus a natural occasion onwhich to review some of those innovations. Enormous changeshave occurred in methods, scale of production, range of prod-ucts, and quality of glassware during the past century. By 1890,the Siemens regenerative furnace was already able to providethe continuous supply of molten glass needed for mechanizedproduction.

The past century has been a turbulent one. The United Stateswas expanding vigorously when The American Ceramic Soci-ety was founded, and there was much political unrest; employ-ers and unions were often in conflict, sometimes physically.European culture was also in turmoil that led to World War I,to be followed only 20 years later by World War II. Rapidpolitical and economic changes have continued ever since; allof these have inevitably affected progress in glassmaking.Many historians have written about these times; a particularlyvivid account of the 25 years ending with the outbreak ofWorld War I is by Tuchman.1

J. J. Thomson discovered the electron in 1897; the first com-mercially successful American automobile, the Oldsmobile,was put on the market in 1901; Planck’s theory of black-bodyradiation was published in 1910. Coast-to-coast telephone callsbecame possible in 1914; the first primitive television broad-casting system began transmitting in England in 1929. Theatomic nucleus was split in 1945; Crick and Watson unravelledthe structure of DNA in 1953; the transistor was invented in1957.

Davis2 discussed the history of the glass industry in theUnited States, from the earliest times up to 1929, but his maininterest was economics rather than technology. Scoville3 alsodescribed the early growth of the American glass industry andexamined in detail the development of glassmaking companiesin Toledo, OH, between 1880 and 1920, again as an economichistorian.

R. K. Brow—contributing editor

Manuscript No. 189571. Received February 8, 1999; approved March 30, 1999.

J. Am. Ceram. Soc., 82 [5] 1093–1112 (1999)Journal

centennialfeature1093

II. Pressing

Pressing requires only suitable molds and muscle power withsimple mechanisms; therefore, it has been exploited from veryearly times but can be used only for a limited range of shapes.Simple side-lever presses giving a considerable mechanicaladvantage were widely used a century ago. In the early years ofthe 20th century, pressing was widely exploited to make dishes,plates, cups, saucers, etc., often of rather low quality. In thosedays, pressing was the only process available for making largevessels, such as 100–200 L electric (lead–acid) accumulatorjars.

The two major advances in pressing were to operate themachine pneumatically and to put several molds on a rotatingtable; these much increased the rate of production and de-creased the labor needed. Such presses often used a spring cageand toggle action. W. J. Miller4 was one of the pioneers inmanufacturing presses; he introduced a “one-man” press. FrankO’Neill,5 whose first machine was a power-driven press ca-pable of making 27 pressings a minute (Meigh6), was anotherpioneer. The quality and range of pressed domestic wares thenimproved rapidly; pressed kitchen and ovenware of Pyrex be-came familiar by 1940.

Although pressing can use relatively high pressures to ensurevery accurate shaping, surface quality is poor unless the moldsare of the highest quality and conditions at the glass–moldinterface are very closely controlled. A common defect causedby poor control is a series of parallel or concentric waves calledwashboard.Much of the early output bore deliberately pat-terned surfaces, partly to obscure mold seams and defects,partly to imitate expensive cut glassware. Pitt7 reported in 1918that American decorative pressed ware was of much higherquality than that being made in Britain. Some high-qualitypressed ware, difficult to tell apart from cold worked, wasbeing produced in the United States during the 1930s.

Much higher standards of surface finish were being de-manded by the 1950s for large items, such as television tubefaceplates, and for complex surfaces, such as vehicle headlightlenses. Quality has continued to improve as customer demandshave also increased. More sophisticated standards of engineer-ing and better control of temperatures, pressing velocities, andforces, as well as mold surface quality have allowed thesedemands to be met. Valstar8 described some of the technologi-cal problems of making large color television faceplates.Lenses of good quality can now be pressed, and flat sheets ofglass are pressed to form complex three-dimensionally curvedvehicle glazing components.

The Corning Hub machine, invented by Giffen,9 is unlikeany other press. It looks rather like a miniature Ferris wheeland is used to press Corelle ware from a ribbon of glasspassing below the wheel.

III. Centrifugal Casting

In the late 1940s, the existing technique of pressing thecones for television tubes proved unable to keep pace withdemand. Giffen10 at Corning solved that problem by develop-ing centrifugal casting, although many experts thought it im-possible to use that method. By 1949 he had produced a ma-chine that could make 800 cones/d, and, thereafter, he alsomodified the process to make the bodies for tubes with rect-angular screens. Other manufacturers, including some makingpatterned decorative domestic and artistic wares, then began touse this technique.

IV. Container Manufacture

(1) First AttemptsThe mechanization of bottle and jar manufacture required

several things.(i) To begin by making the mouth or “finish” of the

container.

(ii) To use a suitable parison mold, thereby reducing to twothe skilled glass blower’s numerous operations in forming thebody.

(iii) To use pneumatic or electric power to drive the ma-chine and to provide compressed air, perhaps vacuum as well.

(iv) To have an efficient mechanical method for supplyinggobs of glass to the machine.

(v) To use metal molds.Ashley in Yorkshire undertook the first trials that led to

success. His original machine attempted to use only one in-verted mold with a loosely fitting movable base plate that couldforce the glass down into the neck ring and then serve to formthe base of the bottle, using compressed air for blowing (Fig.1). Ashley adopted this idea by Arnall, who had conceived itabout 20 years earlier, and the two of them obtained patents in1886.11,12

The first machine was too crude to work well, but Ashleysaw that it could be refined. His next single-unit machine (Fig.2) had all the essential elements of neck ring, parison mold, andblow mold. The inverted parison mold was filled through itsbase; then the parison turned upright before the final blowing.Ashley’s second patent included the use of suction around theoutside of the parison for “blowing” it in the final mold as analternative to blowing in the normal way; he also consideredusing materials that evolved gas when heated rather than anexternal air compressor. With two of those machines, one gath-erer could make 1560 bottles in a 10.5 h working day.

In 1889, Ashley13 obtained further patents, the last of whichwas for a rotary machine with several (four) parison molds thatused a pneumatically operated piston to open and close theparison molds as well as compressed air for blowing. A Leedsnewspaper report described in awe-struck terms the manufac-ture of bottles on Ashley machines at Castleford, Yorkshire, in1887. Compressed air is said to have first been used in a glasshouse at Baccarat in France and was sufficiently novel in 1883to be reported to the Academy of Science in Paris (Appert14).

At one time Ashley had 10 of these machines in operationand two workers could make 180 dozen bottles/d on one ma-

Fig. 1. Ashley’s first method for blowing bottles. Glass is loaded intothe inverted mold and pressed down using the movable base plate.Base plate is raised to the end of the mold and the bottle blown. Hingedmold is then opened to remove the bottle. (After British Patent No.8677, 1886.)

1094 Journal of the American Ceramic Society—Cable Vol. 82, No. 5

chine (English15). However, various problems, including hos-tility of the trade unions, led to failure of the firm some yearslater. These machines had the major disadvantage of needing askilled worker to gather the glass and fill the parison mold,which limited the rate of production. This shortcoming heldback the development of the rotary blow–blow or press-and-blow machine for more than 20 years.

Horne, who had built machines for Ashley, patented his ownimprovements and, by 1917, had sold more than 200 of hismachines in Britain (Meigh6).

A coherent history of the development of container machinesbecomes difficult from this point, because so many differentglassmaking companies and numerous talented engineers de-voted their skills and ideas to the problems of mechanizingproduction. Several different approaches were developed, andthere are many tangled skeins in the history of container manu-facture. Given the small size of the glass industry comparedwith the whole field of mechanical engineering, it is pleasantlysurprising to see how many outstanding talents have servedglass manufacture, but their achievements cannot all be in-cluded in a brief review. Fortunately there have been a fewextended reviews, some chiefly about British practice (whichhas tended to follow that of the United States), such as Meigh16

and Clark,17 who reviewed his 50 years with his family firm,Beatson Clark of Rotherham, Yorkshire, which had beenfounded in 1751. During his career, Clark saw changes fromsemiautomatic machines with hand gathering to fully automaticproduction on individual-section (IS) machines. Giegerich andTrier18 produced a notable book describing all types of glass-working machines in use 30 years ago.

(2) Press-and-Blow FormingPressing the parison then blowing the body of the ware, with

the finish still held in the neck ring, relaxes some of the re-strictions on shape imposed by one-stage pressing. J. S. and T.B. Atterbury19 of Pittsburgh, PA, patented this method in 1873for making molasses pitchers with a spout and handle. Thesame press-and-blow method was patented by Arbogast20 in1882 for making bottles (Fig. 3), but it was not used commer-cially until 1893, when it was used by the Enterprise GlassCompany for mass production of Vaseline jars.21 At that time,the glass still had to be gathered by hand. The process was soonbeing used for a considerable range of wide-mouth ware, in-cluding Mason jars. However, no machines for narrow-neckbottles were made in the United States until 1908, when copies

of English machines were introduced (Meigh6); some of thesewere called “Johnny Bull” machines.

Many semiautomatic press-and-blow machines were avail-able between about 1900 and 1940. In the early days, theseusually required a gatherer, one worker at the parison table, andanother at the blow mold table, sometimes with a worker totransfer the parisons from one to the other. It soon becamepossible for the gatherer to control the whole process by op-erating a pedal as soon as the mold had been filled. W. J.Miller22 of Swissvale, PA, was an early manufacturer of such“one-man” two-table press-and-blow machines. Rather primi-tive feeders began to replace the gatherer around 1910, and gobfeeders improved rapidly from about 1918. Semiautomatic ma-chines were all eventually superseded by automatic ones, gen-erally similar to two-table blow–blow machines.

Edward Miller,23 of Columbus, OH, was a pioneer in manu-facturing press-and-blow machines on a large scale; his JP andPB machines were very widely used. He sold his business tothe Lynch Corporation in 1933.

(3) The Suction ProcessMichael J. Owens24 (1859–1923) invented the first success-

ful automatic machine for narrow-neck containers. Owens, theson of Irish immigrants, entered the glass industry in Wheeling,WV, as a boy aged 10 years and was already a blower at the ageof 15. Over the next 10 or so years, he managed to acquiresome education and to display undoubted ability.

In 1888, Edward D. Libbey, the owner of a Massachusettsglass company, was beset by strikes. The glassmakers’ unionswere very powerful at that time in both the United States andBritain: in 1919, the British bottle-making industry still had 13unions. Libbey therefore closed that factory and moved hisoperations to Toledo, OH, where natural gas and good sandwere available. Owens there became one of a team of blowersmaking lamp chimneys. In 1892, Owens obtained his first pat-ent25 for a mechanical device to replace the worker who

Fig. 2. Ashley’s early machine. Inverted parison mold (A) and neckring (B) are fitted to hand-held tongs. After turning the parison upright,the blow mold (C) is raised by a treadle (D) and the bottle blown.

Fig. 3. Illustrations from his patent of Arbogast’s press and blowprocess for containers.

May 1999 Mechanization of Glass Manufacture 1095

crouched at the glassblower’s feet to open and close the mold.Within two years, Owens had become superintendent of thefactory and two of his next four patents were concerned withimproving the glassblower’s tasks.26

Owens later stated that the basic idea for a bottle-makingmachine occurred to him in 1898, but it took five years ofextensive experimentation, generously supported by Libbey, toproduce a useful machine. Owens proved himself a toweringbut rather abrasive genius of practical engineering insight andskill in glassmaking who could devise an effective solution tonearly every problem that arose, thus justifying the unstintingsupport of Libbey. Together they brought revolutionary changeto several branches of the glass industry, most of all to themanufacture of containers.

Owens recognized that gathering the correct quantity ofglass was the first important problem, and he solved it in anovel way, by using suction to fill a gathering mold (Fig. 4). By1903, he had invented the rotating pot (a shallow dish fed froma tank27). This allowed each gather to be made from fresh glassnot chilled by the previous gather (Fig. 5), but it increased thefuel consumption of the furnace. In 1903, he also built hisfourth machine with six arms. Trials with this machine at theToledo Glass Company proved that automatic production ofbottles was possible, and the Owens Bottle Machine Companywas established that year. Further improvements followed; thenext experimental machine still had six arms but had becomecontinuously rotating. The whole rotating mass of several tonswas, however, still lowered and raised each time that glass hadto be gathered. This became the type-A machine, several vari-ants of which were developed during the next few years. Thedipping-head machine (type AN) was introduced in 1912. Eachhead now dipped by itself for gathering and was a self-contained bottle-making unit. The final stage of forming natu-rally required compressed air for blowing and for mold cool-ing. Many of the operations were controlled by cams mountedon the central pillar of the machine. This gave very precisemovements but permitted only minor adjustments of many ofthe individual operations. The Owens machines were outstand-ingly beautiful pieces of engineering; some data about the vari-ous models are given in Table I. The OS six-arm machine was

designed and made by Schwartzkopff, the Owens Company’sEuropean agents.

The 15-arm machines (AQ and AV), first licensed in 1914,could produce as many bottles as 50 blowers working by hand.The 10-arm CA and CB machines, fitted with 20 sets of molds,rotated at∼4 rpm and could produce 80 bottles/min. Owensalso introduced multiple-cavity molds (up to four per mold), arefinement much easier to fit on a suction-fed machine than onesupplied by a gob feeder. The AT carboy machine was hugeand weighed more than 100 tons. It could make 5 gallon bottlesat five or six per minute (Meigh6). The number of Owensmachines working in the United States increased from 8 in1906, to 103 in 1911, and to 187 in 1916 (Weeden28). The firstOwens machine in Britain was installed at Alloa, in Scotland,in 1919, and led to a bitter dispute with the blowers and furnaceworkers, but that was not typical of the response of the workersto machines.

Fig. 4. Main stages of the Owens suction process: (1) the suctionmold dips into the melt; (2) the mold is raised and the glass sheared;(3) the parison mold opens and the blow mold rises from below; (4) theblow mold closes around the parison; (5) the plug retracts and thebottle is blown; (6) all mold parts open to allow the bottle to beremoved.

Fig. 5. Details of the construction and operation of the Owens ma-chine. Unit carrying the parison mold (A) is shown gathering from therotating pot (B). Blow mold is carried on a radial arm that drops downto pass below the rotating pot.

Table I. Owens Series of Machines

TypeNo. ofarms

Yearintroduced

Capacity ofcontainers

(fl oz)‡Production

(min−1)

A 6 1903–1904 4–40 10–20AC 6 1908 4–40 11–22AE 6 1909 4–64 12–24AD 10 1909 4–64 24–26AL 8 1912 16–160 8–12AN 10 1912 0.125–11.5 20–60AR 10 1912 4–80 12–45AQ 15 1914 4–80 18–60AV 15 1917 0.125–11.5 30–90CA 10† 1920–1923 #8 80–320CB 10† 1920–1923 8–32 80–160AT 6 1915 96–1952 6–7OS 6 1954 7–35 22–42

†Having two heads each.‡1 fl oz (U.S.)4 30 mL.


The Redfern29 machine, designed to make whisky bottles,was an initially successful British competitor for the Owensmachine; it was first installed in 1920 at the works of JohnLumb in Castleford, Yorkshire. The two machines appearedsimilar, gathering and working the glass in similar ways, butthey differed greatly in many details of the engineering. Forexample, the main rotary drive of the Redfern ran in an oil-filled channel, complete individual parison or blow mold unitscould be changed very quickly, and all main-drive links werespring loaded to minimize risk of damage. According to Dralleand Keppeler, an Owens machine needed to be stood down fortwo or three hours at least once every two weeks for cleaning,adjustment, and lubrication, but a Redfern machine could runfor months without stopping for those purposes. The 15-armRedfern machine was 5.1 m in diameter and weighed∼60tonnes; there were also six- and ten-arm versions.

An Owens machine used∼6 hp to drive the machine itself,only twice as much as the rotating pot, but vacuum, com-pressed air for blowing, and cooling air added another 47 hp(Dralle and Keppeler30). U.S. government statistics for glass-container manufacture in 1927 showed that machines increasedproductivity by factors of from 6 to 41 over hand methods andthat labor costs fell by more than 90% for many products.

The Owens–Illinois Company, one of the most famous incontainer manufacture, was formed in 1929. Despite their bril-liant engineering, Owens machines were not popular with allglassmakers. They were large (the CA was more than 5 m indiameter) and costly: A complete Owens unit cost $80 000 butcould work out 70 tons of glass and make more than 100 000containers/d. Owens machines were best suited to making longruns of one type of container; a very heavy investment in moldswas required when several types of container were to be made.Moorshead31 reported that the CA machine needed a largerrotating pot (∼4 m in diameter) than the other models and, thus,an even greater amount of fuel.

The quality of the containers was outstandingly good forthose days, but they had one technical flaw that proved impos-sible to cure: the “cut-off scar.” This was a fold in the bottomof the base where a chilled tail had been formed as the knife cutthe stream of glass on lifting the parison mold from the rotatingpot. The scar was unsightly but could be placed where it did notbadly affect the strength of a container with a thick base. Oneof Peiler’s32 patents attempted to solve that problem by havingan additional shallow mold with only a small aperture that wasplaced over the bottom of the parison mold during gatheringand removed as soon as the glass had been sheared.

There were commercial disadvantages over the way that li-censes were granted, as well as the cost. Initially, strict condi-tions over the types of ware that could be made were imposedon licensees, who, of course, also often complained about theroyalties they had to pay. Biram33 recorded details of the legalcontracts under which Owens machines were used in Europefrom 1907 to 1962. Opportunities therefore remained for othermachines that were smaller, less expensive, and more flexiblein operation. The major competitors were blow–blow or press-and-blow machines fed by gob feeders. Such machines wereeventually produced, but their success depended on first in-venting an effective gob feeder to supply them.

Other inventors of machines using suction gathering oftenfollowed different ideas or ones discarded by Owens. Theyusually wanted to produce smaller, simpler machines not need-ing a special gathering pot that was almost as big as the ma-chine itself and used excessive fuel.

One of the successful European manufacturers was EmileRoirant34 (1882–1955), in France. Roirant was an engineer ofoutstanding talent, one of whose favorite sayings was, “Alwaysremember that any idiot can make a complicated machine.”The Roirant B machine, introduced in 1922, had one pair ofmolds, and many of its operations were controlled by a largecam rotating about a horizontal axis (Fig. 6). The machinecould be wheeled up to a furnace to gather from a normalworking hole. The Roirant A6,35 which followed in 1928, was

a six-station, intermittently rotating machine using cams tocontrol the opening and closing of the molds. The BB2 ma-chine, patented in 1935, could make good-quality containers ofcapacity from 10 to 60 L at a rate better than one per minute.Roirant’s last machine was the single-table, continuously ro-tating, seven-arm feeder-fed R7 introduced in 1950.

The other notably successful machine using suction gather-ing was the paste-mold Westlake machine. It is discussed in thesection on lamp bulbs and continues in use for thin-walleddrinking glasses and stemware.

The Monish was another quite widely used, small, intermit-tently rotating, three-mold suction machine, made by Mon-crieff in Scotland. It gathered the glass from a small horseshoe-shaped channel and used a paddle in the forehearth to avoidgathering glass chilled by the previous gather. It could makecontainers with internal screw finishes having a wide range ofcapacities and needed only 2 hp to run it. Weekly productionwas up to 1100 gross of 3 or 6 fl oz (90 or 180 mL) bottles indouble-cavity molds and 200 gross of 80 fl oz (2.4 L) Win-chesters (Creaser, Creaser, and Hodkin36). Monish machineswere still being used in the 1950s (Clark17).

(4) Gravity Feeder-Fed MachinesMost machine designers wanted to develop simpler, feeder-

fed, two-table machines, with parison molds on one table andblow molds on the other. Many designs made blown parisons,but some were press-and-blow machines. At first, these ma-chines generally moved intermittently, and operations weredone only while stationary. Many machines had six molds oneach table, so that a maximum of six operations could be per-formed in forming the parison or in blowing the bottle. Atwo-table machine need not have equal numbers of parison andblow molds, but that arrangement was easier to engineer (Fig.7). These machines initially often needed a “boy” to transferthe parisons to the second table as well as to remove the fin-ished ware. Such semiautomatic machines (sometimes withonly three or four stations) could still be seen in many Euro-pean factories during the 1950s, mostly for short productionruns.

Within a few years, it became possible to eliminate the“boy;” the first machines to do so, introduced by Lynch37 andMiller38 in 1917, were therefore named “No-Boy.” By mount-ing other components besides the molds on the tables, it soonbecame possible to execute operations while in motion. Thisaccelerated the whole process and opened the way for somemachines to become continuously rotating. That saved time, agreat deal of mechanical energy, and much decreased wear andtear.

Fig. 6. Roirant B machine: (A) large cam controls many of theoperations; (B) the parison mold; and (C) blow mold. (B8) shows theposition of the parison mold when gathering.


Two of the most enduring names among those making thesetwo-table machines were Lynch and O’Neill, both Americanengineers. Their machines of several types were widely usedfrom the 1920s until around 1960, as were press-and-blowmachines by Miller. The blow–blow Lynch 1039 was one of themost popular.

Notable European pioneers included Boucher in France, whoobtained patents in 1894 and 1895. He was a bottle manufac-turer in Cognac, and his machines were widely used in France.One of them was installed by Forsters of St Helens, Lancashire,in 1901. Schiller in Germany introduced a press-and-blow ma-chine in 1905 and a blow–blow machine in 1906. Schillermachines were very widely used in Europe for many years;more than 1000 semiautomatic blow–blow machines were soldworldwide. The ability of early machines working with gobfeeders to compete with the Owens machines, once reliable gobfeeders were available, is shown by Fig. 8.

(5) The Gob Feeder and ForehearthMachines able to compete successfully with the Owens ma-

chine depended on the invention of a feeder able to delivergobs of uniform weight and temperature to the machines atexactly the right intervals. The lack of a good gob feeder in theearly days fortunately did not stop machine manufacturers con-tinuing to develop small machines intended to work with suchfeeders.

The earliest gravity feeder to be used, in 1903, was inventedin New York by Homer Brooke,40 who had been born in York-shire. His feeder used a steady stream of relatively fluid glass.When the mold was full, this stream was sheared and gatheredby a cup that was tipped up to hold the glass until the next moldwas in place; the cup then dumped its contents into that mold(Fig. 9). Four sets of cups and shears were mounted on arotating table. The main defect was poor thermal homogeneityof the glass in the parison mold caused by chilling of the ratherthin stream of glass as it collected in the cup. However, it wasadopted by several companies, including Hazel–Atlas and BallBrothers, for cheap wares. Hazel–Atlas obtained a patent forsuch a device, but with the shears above the cup, as late as 1925(Stenhouse41).

The Hartford–Fairmont Company had been established in1912 to build new and improved glassmaking machinery tocompete with Owens. Finding another way of supplying good-quality glass to forming machines was one of its importantobjectives, and Karl Peiler,42 a 1904 graduate from Massachu-setts Institute of Technology, was already working on this task.Peiler remained with the company for 42 years.

Peiler decided to work with glass at about the viscosity usedfor hand gathering (more viscous than for a suction machine)and chose two approaches. The first was to attempt to mecha-nize the manual operation of a gathering iron, as shown in Fig.10. The gathering iron rotated continuously while inside thefurnace but stopped when retracted so that gravity would forma pendant mass that was cut with shears; unfortunately, thatproved difficult to control accurately. The second attempt43

was to produce the gobs intermittently using a paddle havingboth vertical and horizontal movements that made the glasssurge over the rim of a refractory bowl. Peiler also conceivedanother way of achieving this result using two adjacent refrac-tory bodies dipping into the glass but moving only vertically;by reciprocating more or less alternately, they could achievethe same result.44 The hanging mass was again severed byshears at the right moment to allow the gob to fall into a chutethat led it into the mold. This worked rather better, and someHartford paddle feeders were used in the industry, from 1915.Those feeders allowed both O’Neill and Miller to make useful,fully automatic machines from 1916.

However, those feeders also proved unable to give the con-trol desired. Peiler therefore added a bowl containing a recip-rocating needle into which the glass dropped, to give the re-quired finer control and some shaping of the gob (Fig. 11). Thispaddle–needle45 feeder was first used in 1915. By 1918, 120 ofthese feeders were already in use (Meigh6). However, it turnedout that they only gave good control of gob weight and shapeat high speeds of operation. They were therefore often used tosupply two or three machines using a swinging distributor, aprecursor of that later essential for the IS machine.

Peiler then turned to another way of controlling gob forma-tion. That method used the reciprocating vertical motion of alarger plunger above a concentric orifice through which theviscous glass flowed. A concentric tube around the plungerrestricted the effect of the plunger to the glass in and immedi-ately above the orifice. Peiler produced the first model of whatis now the standard type of feeder in 1922 (Fig. 12). This feederuses cams to give positive and reproducible motion of thevarious elements; rotation of the tube improves thermal homo-geneity in the nose of the bowl. The company was renamedHartford Empire in 1922, after taking over the Empire Machin-ery Company from Corning Glass Works, and again in 1951,when it became Emhart.

Dowse and Meigh46 reviewed in detail the state of feederdevelopment in 1921. Another generously illustrated review,taking the story rather further, appeared in 1932 (Swain47), butthat strangely does not mention the paddle–needle feeder. Peil-er’s feeder has since undergone many improvements, but thebasic design and operation remain remarkably similar to theoriginal model; the essential elements remain the orifice, thereciprocating plunger, the rotating tube, and the shears. A no-

Fig. 7. Typical schematic layout of a two-table blow–blow containermachine with turnover of the parison before transfer.

Fig. 8. Growth of the annual production of containers on automaticmachines in the United States, 1905–1925. Up to 1917, production wasentirely by Owens machine; beyond 1920, the increase was largelyfrom gob-fed machines. Production reached 10 billion/year in the1940s. (After Meigh.6)


table modification enabled it to deliver double, then latertriple or even quadruple gobs to the forming machines. Thatallowed simultaneous production of more than one container inone mold body having more than one cavity; the process hadbeen an early and rather simpler development with suctionmachines.

Successful operation of the feeder required a constant supplyof glass of a particular uniform viscosity so that the glass hadto be cooled and made thermally homogeneous between theworking end of the furnace and the feeder bowl. The construc-tion of furnaces and machines also made it necessary to dis-tribute the glass to machines some distance apart. Much longerforehearth channels than originally used thus became desirable,and much work was done to develop forehearths that fulfilledthe need for constancy of temperature and thermal homogene-ity despite variations within the furnace. Hartford took a lead-ing role in that task. Hartford feeders were brought to Britain in1917 by W. A. Bailey, who set up the British Hartford–Fairmont Syndicate, which, in 1918, began to operate a paddle–needle feeder with a Hartford milk bottle machine at Key GlassWorks, New Cross.

The basic features of the modern forehearth were introducedin 1922. Forehearth design and control have improved hand inhand with the gob feeder. Relatively early patents includedsealing the combustion space of the forehearth from the work-ing end of the furnace to enable much better control of thetemperature distribution along the forehearth and around thegob feeder (Peiler48).

As often happens, many ingenious devices were patented toovercome particular problems, and some proved crucial, butmany were later abandoned. One of the former was the mov-able chute that allows one forehearth and feeder to supplyseveral sections of an IS machine (Ingle49). One of the latter

was “a guide or abutment mounted below the shear blades” tostop the gob being pushed sideways as it was sheared(Peiler50). One of the most notable improvements was the gas-fired K-type forehearth introduced in 1944, many of whichremain in operation.

Considerable ingenuity was used to devise feeders that couldsupply gobs of two or more different controlled weights suc-cessively to different machines (Chalmers51), but these neverbecame widely used. As soon as the successful gob feeder wasavailable, the blow–blow and press-and-blow machines beganto compete with the Owens machine.

In recent years, advances in mechanisms capable of complexoperations as well as the greatly improved methods of controland temperature measurement have allowed the developmentof successful gathering machines that returned to earlier meth-ods. Some modern mechanical ball gatherers have essentially

Fig. 9. Operation of a Brooke gob feeder: (1) continuous stream of glass falls from the orifice into the mold; (2) mold is full so the cup and theknife sever the stream; (3) cup is tilted to collect the still flowing stream while a new mold moves into place; (4) cup is tilted further to empty intothe mold; and (5) is moved aside until the mold is full, then the cycle begins again.

Fig. 10. Peiler’s first automatic gathering device, which attempted toreproduce the actions of a skilled worker. (After U.S. Patent No. 1 324464, 1919.)

Fig. 11. Peiler paddle–needle feeder (paddle is P and the needle N):(1) glass at its maximum height in the bowl; (2) backward motion ofthe paddle, then upward movement of the needle neck, the streamflowing through the orifice just before shearing.


automated the operations of hand gathering, as Peiler first at-tempted to do, or returned to gathering by suction. These areespecially useful for very large gobs (Lindner52), such as areneeded for large television screens.

(6) The Individual Section MachineThe IS machine has now displaced almost all other container

machines. It was invented in 1924 by Henry Ingle53 of theHartford Empire Company, but the patent was not granted until1932. Ingle himself named his machine “Individual Section.”However, Meigh54 reported that the abbreviation “IS” was al-ready widely used when he saw the first four-section machinesin 1927. It was assumed to have been adopted to recognize boththe inventor, Henry Ingle, and the President of the Company,Charles Goodwin Smith. Smith supported Peiler and Ingle inthe way that Libbey had supported Owens. Many of the devel-opments of the IS machine up to 1960 were guided by GeorgeRowe, who joined the company in 1924.

The IS machine introduced several novel features; a 1923patent granted to Ingle55 indicates how big a step forward theIS machine represented. A well-illustrated description of it waspublished in 1928.56 The most revolutionary features were thateach section could operate independently and that the arm car-rying the neck ring moved in a vertical half circle to transfer theparison to the blow mold. Both parison and blow molds re-mained stationary; the only essential horizontal motions werethe opening and closing of the molds and neck rings, and theremoval of the finished ware (Fig. 13). The machine was builtfrom as many sections as desired, mounted side by side, andprovided with one long horizontal main drive shaft. In the1950s, a five-section machine was considered large, but asmany as 16 sections are now in use. Although all sectionsreceive gobs of the same weight, it is possible to make con-tainers of somewhat different designs on the various sections.

The IS machine is compact, each section is only∼0.7 mwide, but tall. Another very attractive original feature was theuse of an easily accessible large-diameter timing drum to carrythe on–off controls for all the essential pneumatically con-trolled operations (at least 19). This gave accurate control andvery wide flexibility in settings, something not possible withmany of the operations on the Owens machine. Electronic con-trol was introduced in 1974.

The IS machine did not achieve its present dominance formany years. It was expensive and, like the Owens, for someyears, could only be licensed. Although much more flexiblethan the Owens machine, it too was generally perceived as bestfor very long production runs. Double-gob operation (making apair of containers simultaneously in one mold body) was in-troduced in 1939 and triple gob in 1967. The technical supe-riority of the IS machine has been supported by a first-classcommercial and engineering organization that expanded tocover almost the whole world.

Ingle also recognized the need for a good alternative toheavy iron boxes or solid iron slats for carrying the ware in thelehr. He realized that heating and cooling large amounts of ironconsumed a lot of energy and limited the rates of heating andcooling that could be achieved. Two of his patents57,58are forthe use of lightweight metal mesh for the lehr belt. This belthad alternate sections woven right-handed and left-handed toavoid creep of the belt to one side. The use of the mesh belt alsopermitted much better gas circulation in the lehr and, therefore,better temperature control.

As machine performance improved, the rate at which heatcould be extracted from the metal molds became a limitingaspect of machine operation. The basic design of the IS ma-chine allowed better mold cooling than other machines becausethe molds remain in the same place. Giegerich59 published aninteresting study that compared the relative efficiencies of sev-eral types of bottle machine and showed that the IS machineperformed best from most points of view. The Vertiflow60

system of making air flow through vertical passages drilled inthe mold bodies was a major advance in mold cooling. Thisgave better control of temperature distributions in the molds,

Fig. 12. Essential elements of the modern Hartford feeder: (A) fore-hearth brings thermally homogeneous glass to the feeder bowl; (B)reciprocating plunger controls flow of glass in the orifice; (C) rotatingtube makes the glass circulate into the nose of the bowl, maintainingits uniform temperature, and helps to control the flow; (D) the inter-changeable orifice; and (E) the shears.

Fig. 13. Basic operations of the IS machine: (1) gob drops into themold; (2) neck ring is filled; (3) parison is blown; (4) parison moldopens and parison is swung over to be enclosed by the blow mold; (5)neck ring is released and the parison reheats; (6) bottle is blown; and(7) bottle is released.


decreased the consumption of compressed air, and reducednoise levels.

The 62 process for making pressed parisons for wide-mouthed ware, such as jam jars, was introduced in 1940. Press-ing of the parison makes it much easier to ensure completefilling of the neck ring and accurate formation of the finish. Byalso guaranteeing the internal shape of the parison, it improvescontrol of wall thickness in the finished article. The most im-portant advance of recent years has been the widespread use ofpress-and-blow for narrow-mouth containers. This uses ma-chines and materials very close to the limits of what is possible.The narrow-neck press-and-blow process is now consideredessential for further progress toward thinner walls and lighterware. It is accepted that ensuring a minimum wall thickness of0.8 mm requires a designed wall thickness of about 2.0 mm forblow–blow but only 1.3 mm for the press-and-blow process.

During the past 40 years, much research has been done andvarious technically advanced and promising machines havebeen developed by other engineers. Owens–Illinois, for ex-ample, spent many years investigating the crucial factors incontainer production that could help to improve performanceand might lead to better machines. These potential competitorsincluded the Lynch 44,61 which had individual sections, likethe IS machine, but the basic sequence of operations used threestations (parison forming, blowing the container, and take out)with three neck rings rotating about a vertical axis.

Another machine of great promise in the 1970s was the Heye6-12.62 This was a two-table narrow-neck press-and-blow ma-chine with an unusual layout. The tables were some distanceapart and rotated about vertical axes; the blow mold table wasbigger and carried more molds than the parison table. A con-tinuous belt bearing multiple neck rings carried the parisons tothe blow molds and the finished containers away to the dis-charge point. A machine having different numbers of parisonsand blow molds recognizes that the two parts of the wholeoperation do not necessarily need the same times, and that thiscould improve efficiency.

However, no other new machine has to date proved able tocompete with the IS machine. This has been partly due to theheavy investment in capital, spares, operating experience, andso on, that most manufacturers have already made in IS ma-chines. Radical changes to the IS machine itself would faceobstacles similar to those of other competitors, but it is difficultto believe that further major advances will not occur. A recentEmhart patent (Jones63) adapts the IS machine to transfer theparison to an intermediate station where the blow mold en-closes the parison and the neck ring is released; after reheating,the blow mold is moved linearly to the final blowing position.Edgington64 has written a detailed history of Emhart.

V. Flat Glass

(1) Sheet GlassIn the 1890s, two processes existed that had long been used

to make window glass. Broad, or cylinder, glass was made byblowing uniform cylinders as large as possible, cutting off theends, then splitting them longitudinally. These were reheated ina special furnace and flattened, using a special rake to openthem out. Cylinder glass nearly always had some surface dam-age, with waves or ripples on both faces, caused by the actionof the rake and contact with the substrate. Cylinder glass wasunsuitable for mirrors or for windows intended for undistortedvision rather than merely admission of daylight. However, theintroduction of some simple mechanical aids during the secondhalf of the 19th century had led to improvements, as much inthe size of the cylinders and productivity, as in the quality ofthe glass. These aids supported the horizontal blowpipe whilereheating at the glory hole and provided a pivot for swingingthe cylinder to elongate it after reheating.

Crown glass was made using centrifugal force to spin acarefully formed shallow bowl of glass into a circular disk.

Both faces of the disk were therefore fire polished, but the diskwas always slightly saucer shaped and had a central thickenedregion, the bullion, where the punty had been attached. Thenatural fire-finished surface quality was much better than withcylinder glass, apart from the inevitable curvature; however,any inhomogeneities in the glass formed slight imperfectionsvisible as arcs of circles on the surfaces. The largest panes thatcould be cut from crowns were smaller than could be producedfrom cylinders. Nineteenth century English glassmakers be-came especially skilled in the crown process. The natural cur-vature of the crown was less than that of mirrors commonlymade by cutting pieces out of large blown spheres. However,crown glass had become obsolete by 1890, except for specialpurposes, including microscope cover slips and colored sheetfor stained-glass windows.

The first important step toward mechanization was the Lub-bers65 process, on which development work began in 1894. Itwas patented in 1902 and brought into commercial operationby the American Window Glass Company in 1903. Thismechanized the vertical drawing of cylinders. Initially, glasswas cast into a special preheated bowl, and a vertical blowpipecarrying a circular lidlike bait lowered into the glass. As soonas the glass had adhered to it, the bait was raised. Compressedair was blown into the emerging cylinder; both the upwarddrawing velocity and the blowing air pressure were carefullycontrolled to produce a cylinder of uniform diameter and wallthickness over most of its height (Fig. 14). Cylinders up to 13m long and 1 m diameter were produced, and they needed15–18 min to be formed (Turner66). However, the cylindersstill had to be cut in shorter sections, then split and flattened, so

Fig. 14. Drawing a cylinder of glass by the Lubbers process. Glassis cast into a (A) reversible preheated bowl sitting on top of a (B)heated muffle. (C) Bait is dipped into the glass; when the glass hasadhered, the blow-pipe unit is raised up the (D) vertical track.


that optical properties generally were little, if any, better thanwith handmade mouth-blown glass.

The Lubbers process was an immediate commercial success.In the years from about 1895 to 1903, the United States annu-ally imported more than 20 000 tonnes of sheet glass fromBelgium: in 1904, this abruptly fell to 12 700 tonnes, andimports never again reached former levels. By 1923 more than60% of the window glass produced in the United States wasmade by the Lubbers process.

Numerous inventors attempted to draw flat sheets directlybut had to attempt to overcome two natural laws: first, that anyjet of liquid is inherently unstable and, second, that it tendstoward a circular cross section. A British patent by Clark67 in1857 recognized that forming a flat sheet by upward drawingrequired some way of stabilizing the width of the sheet (heclaimed the use of immersed metal hooks at each side) andrapid chilling of its edges. A French patent of 1871 by Vallin,a cylinder-sheet manufacturer, had very similar ideas (Borel68).Success eventually attended the efforts of two inventors, one inBelgium and one in the United States.

Emile Fourcault (1862–1919) of the Dampre´my glass worksin Belgium obtained French69 and U.S.70 patents for his pro-cess in 1902 and later obtained several British patents.71–73Hehad to overcome many difficulties before managing to makesaleable glass in 1916. One of the major customers in war-tornBelgium was the German occupying power; Fourcault thensuffered grievously for being considered to have traded withthe enemy, and this may have contributed to his early death.

Fourcault’s major invention was to stabilize the foot of thesheet using a “de´biteuse,” a long clay boat floating in the glassand having a carefully shaped long slot in its base. When thedebiteuse was slightly depressed, glass flowed up through theslot. If the glass was immediately gripped, pulled up, andcooled rapidly, it formed a vertical sheet (Fig. 15). The emerg-ing sheet passed between water-cooled iron boxes, placed justabove the de´biteuse, quickly making it sufficiently viscous toavoid the development of catastrophic instability. The methoddescribed in the first patent drew the sheets up a vertical towerby a simple apparatus remarkably similar to a guillotine work-ing in reverse; the tower also served as the lehr. The use ofseveral (typically 12–14) pairs of asbestos-covered rollers thatpressed very gently on both faces of the sheet, to make theprocess continuous, came later. Fourcault’s first patent in-cluded the options of drawing cylinders or bending a continu-ously drawn sheet over a roll and leading it away horizontallyinto the lehr.

The stream of glass flowing through it eventually corrodedthe debiteuse. Long threads of corrosion products then con-taminated the surfaces of the sheet, causing obvious opticaldistortion, and the de´biteuse had to be replaced. The glass inthe drawing chamber had to be held so near the liquidus thatdevitrification inevitably developed, and production also regu-larly had to be interrupted (approximately every 10 days) be-cause of this.

Irving W. Colburn (1861–1917), an inventor from Fitchburg,MA, was inspired by seeing a paper mill. His father hadworked for Owens (Owens and H. J. Colburn had a joint patentin 1897). He began to experiment with an upward drawingprocess at about the same time as Fourcault. Colburn’s74 suc-cessful process differed from Fourcault’s in two essential ways.First, the glass was taken directly from the free surface of themelt, and the edges of the sheet were stabilized by driving theglass upward using rotating bodies. Second, once the sheet hadbeen formed, it was bent over a polished metal roll and carriedaway horizontally into a lehr, clearly more convenient than avertical tower if the space was available (Fig. 16); however,this was not included in the first patent.

Colburn continued trials at his own expense until his fundswere exhausted in 1912, and all his assets, including plans andpatents, were put up for sale. Colburn then appealed directly toOwens and convinced him that the process could succeed. TheToledo Glass Company then took over the project, employingColburn as a consultant, and the success of their initial trials ledto formation of the Libbey–Owens Sheet Glass Company in1916; unfortunately Colburn, like Fourcault, did not live to seehis process achieve real commercial success (Clarke75).

Initially, Colburn used partially immersed spheres to attemptto form and stabilize the edges of his ribbon but, after numer-ous trials, adopted knurled rolls on the ends of long shafts.76,77

Colburn used a refractory drawbar, somewhat similar to a de´bi-teuse but totally immersed, to help stabilize the foot of thesheet. Strings of corrosion products could eventually be drawnup from this drawbar, but they lay buried inside the drawnsheet and did not impair optical quality as badly as the surfacedefects produced by the Fourcault process. One of the earlydifficulties was placing the bending roll too far (1.5 m) abovethe point of draw, something apparently quickly recognized byOwens. At least one patent was then obtained in the names ofboth Colburn and Owens. A factory making 63 in. (∼1.6 m)wide sheet was in operation at Charleston, WV, in 1918. Atthat time, and for a number of years afterward, the quality wasnot very good. It took about another 10 years for the quality ofeither Fourcault or Libbey–Owens glass regularly to match thatof good cylinder glass and render the Lubbers process obsolete.

The Pittsburgh Plate Glass Company (PPG) process for ver-tical drawing of sheet formed the foot of the sheet in the sameway as in Colburn’s process but drew the sheet up a verticaltower similar to Fourcault’s. This combination presumablyavoided infringement of the patents of either. The PPG process,largely the work of Slingluff,78 was patented and came intooperation in 1926 but, similar to the other processes, tookseveral years to produce top-quality glass regularly.

In 1927, Gross79 provided data about the relative costs ofmaking sheet glass by the different processes (Table II). TheLubbers process used extravagant amounts of fuel and theLibbey–Owens (Colburn) process needed the least labor. TheLibbey–Owens–Ford Company was formed in 1930 by themerging of Libbey–Owens with the Edward Ford Glass Com-pany of Rossford, OH; Edward Ford’s father was one of thefounders of PPG (see below).

Having remained committed to the cylinder process, Pilk-ington,80 who had interests in Canada as well as Britain, ini-tially failed to obtain licenses for either the Fourcault or Libbey–Owens process, after first backing an unsuccessful downwarddrawing process. Pilkington was, however, able to take up thePPG process in 1931, when four machines making glass 94 in.(∼2.4 m) wide were installed. These performed so well thatanother four machines making glass 132 in. (∼3.4 m) wide, thewidest in the world, were installed the next year. In 1933, theyear that they abandoned the cylinder process, Pilkington atlast acquired rights to use the Fourcault process and, in 1934,was using it to make 80 in. (∼2.0 m) wide glass only 1.2 mmthick.

All three of these processes coexisted for several decades,each having somewhat different advantages and limitations.With all of them, devitrification in the drawing chamber peri-Fig. 15. Essential component of the Fourcault process, the de´biteuse.


odically forced a shut down. Rates of production varied; forexample, the drawing speed of 4 mm sheet was∼0.50 m/minfor the Fourcault process but 1.0 m/min for the Libbey–Owens,and 0.80 m/min for PPG (Goerk81). Annual production of aLibbey–Owens machine was∼13 × 106 ft2, (1.2 × 106 m2) fourtimes that of a Fourcault machine. The Fourcault process, atleast, continues to be used for thin sheet and for short runs of,for example, special colors. However, all have been replacedby float glass for most purposes.

Asahi,82,83 in the 1970s, patented a new sheet-drawing pro-cess that could be installed in an existing Fourcault installation.It was capable of making good-quality sheet 0.7–6 mm thick.Its features included two immersed parallel rotatable cylindri-cal bodies forming a slot through which the glass flowed up,with edge blocks at the ends of these bodies. These in effectformed a de´biteuse of adjustable width. There were also rolls,

similar to those of the Colburn process, to grip the edges of thesheet just above the surface of the melt.

Thin glass for microscope cover slips, liquid-crystal devices,and other applications in modern electronics is the one type ofglass not easily made by the float process (described below).Numerous methods of drawing such glass have been proposedin the past 60 years. A Corning down-draw process for micro-scope cover glasses using a platinum drawing die was granteda patent in 1947 (Brown84). A very elegant later Corning down-draw process (Dockerty85) allows a steady stream of glass toflow over both sides of a special platinum channel to form athin sheet as it falls away from the keel along the lower edgeof the boat (Fig. 17). Both surfaces thus have a natural firepolish. That patent is unusual in that it is little more than adetailed mathematical analysis of the critical requirement of theapparatus, that is, how to vary the internal depth of the channel,of otherwise constant dimensions, to obtain the same thicknessof the overflowing layer along the entire channel.

Some other processes rely on the Newtonian behavior of apreform to maintain its cross-sectional shape when extruded ata viscosity sufficiently high to make stresses due to surfacetension effects much smaller than the other stresses involved(Humphrey,86 and Roeder and Egel-Hess87).

(2) Plate GlassIn the 1890s, plate glass with perfectly flat polished surfaces

could only be made by the method invented in France around1695. Plate glass was first made in the United States, in 1870,

Fig. 16. Early embodiment of the Colburn (Libbey–Owens) process. Edges of the glass ribbon are gripped by knurled rolls and the sheet at oncepasses between water-cooled boxes. Sheet is slightly reheated before passing over the bending roll and is then drawn away horizontally into the lehr.

Table II. Relative Costs of ProducingSheet Glass by Various Processes

in 1927†

Process Wages Fuel

Mouth blown 100 100Lubbers 38 146Fourcault 25.3 60Libbey–Owens 9.5 65

†After Gross.79

Fig. 17. Dockerty’s overflow process for thin sheet. Cam A allows fine adjustment of the slight inclination of the vessel. (After U.S. Patent No.3 338 696.)


by J. B. Ford88 (1811–1903), a riverboat captain, who was laterone of the two founders of PPG, and lived for many years atCreighton, PA. In this classic procedure, the contents of aspecial large pot were cast onto a metal table and the glassrolled out by a heavy metal roller running on metal guides ofappropriate thickness placed along each side (Fig. 18). The slabwas then pushed into a lehr. After annealing, it was cut intosmaller pieces for grinding and polishing. Although the basicprocess remained the same, the quantity of glass cast had in-creased during the 19th century, and the details of the apparatusimproved, especially for grinding and polishing. However, thecast plate still had to be almost twice the intended final thick-ness if all low spots were to be eliminated during grinding.Grinding with sand and then polishing with rouge were veryexpensive in time, mechanical power, and labor. In 1924,Fick89 reported that a grinding table, 36 ft (∼11 m) in diameterneeded 500 hp to operate it, and a polishing table of the samesize, 700 hp. These processes used 80% of all the electricalenergy consumed in a plate glass plant.

The first important advance was the second version of theBicheroux process, introduced in Germany around 1920.90

That process made some important changes in the casting ofglass from a pot of 0.85 m3 capacity. The glass was pouredfrom the long side of an oval pot onto a casting tray that couldbe tilted; the glass flowed down this tray to run between twohorizontal rollers of equal diameter. These rollers extruded theglass to the required thickness and delivered it on to a tablemoving forward below the casting unit at the appropriate rate(Fig. 19). One of the important improvements was to rotate thepot about the lower edge of its lip, rather than its center ofgravity, during casting. Powerful cutting knives chopped thecast plate into several pieces as the table moved forward. Thisprocess increased the maximum size of plate that could be castfrom 31.5 to 60 m2. The Bicheroux process continued to beused for some purposes by Pilkington at St Helens, Lancashire,until 1958 (Pilkington91).

By the end of World War I, the Ford Motor Company wasexperiencing difficulty in buying the large quantities of plateglass that they needed (Fig. 20); therefore, Henry Ford decidedthat the company should make its own. One of Ford’s seniorengineers, C. W. Avery, was put in charge of this project al-though he knew nothing about glassmaking. After about twoyears of unsuccessful experimentation and expenditure of $1.5million, Avery, in 1921, at last succeeded in solving all theproblems of continuous casting (Fig. 21) but still did not havea furnace capable of providing glass of sufficient quality. Atthat stage, Pilkington heard about this work, and a visit wasarranged to see the Ford plant. One of the Pilkington represen-

tatives, R. F. Taylor, was greatly impressed by what Avery hadachieved and recognized its potential. Pilkington quickly de-cided to make their own trials to demonstrate that glass 6 ft(∼1.8 m) wide, considerably wider than Ford needed for auto-mobile windshields (40 in. (∼1.0 m)), could be made by theFord process. These trials were successful, but Ford was stillnot melting glass of good enough quality. In 1923, Ford’soriginal plant at Highland Park was producing only 7% of theglass that they needed, but a new factory to have four furnaces

Fig. 18. Late 19th century French plate-glass casting hall. There should be two workers at the far end of the roller, just as there are at the nearend. (After Peligot.171)

Fig. 19. Improved Bicheroux process: (1) pot is brought to the hori-zontal tray; (2) pot is rotated about the lowest point of its lip to beginpouring onto the tray; (3) as pouring continues, the tray is raised andthe glass extruded between the rolls and carried away on the castingtable.


fired by coke oven gas was built at Rouge River. In that year,∼40 × 106 ft2 (3.7 × 106 m2) of plate glass were used forautomobiles.

Huntley92 published an account of the new Ford plant thatincluded photographs of two prototype rolling machines andthe final successful unit as well as of the grinding and polishingmachines. The lehr was said to be 442 ft (∼135 m) long. An-other account by Huntley93 said to be “the first authorized

description” mentioned that one of the units produced glass 42in. (∼1.3 m) wide, rather than 40 in. Some of the rollers that theglass passed under were said to be of nichrome.

In 1922, Pilkington, having developed an improved continu-ous grinding machine that they offered to Ford, was able tonegotiate a mutually satisfactory contract for exchange of ex-perience. Pilkington staff were sent to Detroit to assist Fordwith furnaces and grinding equipment while continuous castingwas put into operation at Cowley Hill, St Helens. By 1924,Pilkington had made a ribbon 72 in. (∼1.8 m) wide, but theircontinuous grinder could handle only 60 in. (∼1.5 m) width; by1927, the grinder had become capable of dealing with glass 100in. (∼2.5 m) wide (Barker94). Turner,95 in 1924, mentioned afew facts about the process. However, neither Ford nor Pilk-ington made any further public announcement about continu-ous casting until a report by Partridge96 appeared in 1929 andadded a few more details. The casting rolls were both of castiron. The larger roll, 48 in. (∼1.2 m) in diameter, had a millingcutter mounted on its frame so that its surface could be dressed,when necessary, without having to remove the roller from thecasting machine. The elapsed time between the glass emergingfrom the furnace and the polished plate being ready for dis-patch was 7.5 h.

A lecture by Turner66 in 1930 was the occasion for publicannouncement of the history and then current state of continu-ous casting. In 1931, Libbey–Owens–Ford acquired some otherplate glass plants and obtained a 7 year contract to supplyGeneral Motors Corporation with nearly all the glass that theywould need.

Presumably because of widespread interest in rolled plate atthat time,Glass Industrytook the unusual course of publishingan English translation, by F. W. Preston,97 of the entire chapteron that subject written by E. Lutz98 for Dralle and Keppeler’sbook. It appeared in 25 parts inGlass Industryfrom October1930 to October 1932.

The Ford process allowed a controlled stream of glass 14 in.(∼0.36 m) wide to flow from the furnace onto the upper surfaceof a large water-cooled roller, where it spread sideways. Afterit passed under a 9 in. (∼0.11 m) diameter upper roll, also watercooled, the ribbon of glass was carried into the long lehr (Fig.21). At the end of the lehr, the glass was cut into 113 in. (∼2.9m) lengths for grinding then polishing on two lines, each aboutas long as the lehr. The first of these ground and then polishedone side, the second doing the same with the other side. Therewere more than 40 grinding heads (Partridge reports 29 usingsand and 14 using garnet) and 35 or 36 polishing heads (usingrouge). The glass being ground and polished was carried oncars running on rails and clamped together after the slabs ofglass had been set in plaster of Paris. The blanks for 0.25 in.(∼6 mm) plate were cast 0.375 in. (∼9 mm) thick and weighed75 lb (∼34 kg): after grinding and polishing the weight haddecreased to 51 lb (∼23 kg), a 32% loss. The Pilkington versionof the Ford process made a considerably wider ribbon of glassand a range of thicknesses.

The next important advance was a twin grinder able simul-taneously to grind both faces of the still continuous cast ribbon.Pilkington engineers had made sufficient progress with this by1932 to begin to turn their attention to the problems of a twinpolisher. The first production line with a twin grinder wascommissioned at Doncaster, Yorkshire, in 1935. Its lengthfrom the doghouse to the end of the twin grinder was 1400 ft(427 m). The grinder itself occupied 300 m and consumed 1.5MW of electric power. The glass originally moved through thattwin grinder at 66 m/h, but, as improvements were made, thiswas later increased to 300 m/h.

The twin polisher was eventually developed but never ex-ploited on a large scale. The dissipation of the heat generatedwhen polishing both faces simultaneously was one importantproblem when there was no generous supply of coolant, as ingrinding. For example, patents concerning the cleaning of thelower polishing heads were granted to Waldron and Griffin,99

Waldron,100 and Baillie in 1947.101

Fig. 20. Growth of plate glass production in the United States, 1895–1930, showing the importance of its use in automobiles after 1915.

Fig. 21. Ford method for continuous casting of plate glass. Stream ofglass 14 in. (36 cm) wide falls onto the top of the 48 in. (123 cm)diameter main roll, which rotates at∼1 rpm. Stream spreads out to 40in. (102 cm) width before it passes under the upper 9 in. (23 cm) roll.Glass ribbon is then carried away horizontally into the lehr.


After 1930, some plate glass was produced from thick drawnsheet. This practice was at least a century old with thick cyl-inder glass: J. T. Chance102 had obtained a patent for such aprocess in 1838. This possibility having been foreseen andconsidered undesirable, Fourcault licensees were prohibited fora number of years from making sheet more than 4 mm thick.Photographic plates,∼1.5 mm thick, which needed to be ex-ceptionally flat but mostly not very large, continued to be madefrom top quality cylinder glass for a considerable time. Pilk-ington abandoned manufacture of plate glass in 1967.

The continuous casting of plate glass was also developedindependently by Saint Gobain in France, largely by Boudin.Their first patent,103granted in 1926, shows the use of two rollsof equal diameter positioned so that the glass has to be carriedslightly uphill by the lower roll before passing under the upperone. That patent envisages the upper roll being used to imprinta design on the surface of the glass. Boudin subsequently ob-tained a considerable number of other patents for refinementsof the method.

(3) Float GlassIn the late 1930s, Pilkington was consciously seeking ways

of improving the very costly and time-consuming grinding andpolishing; other plate glass makers were doubtless thinkingabout the same problems. Henry Bessemer104 had patentedseveral of the main features of the float process in 1848, in-tending to support the glass on molten metal during annealing,but obviously lacked the detailed understanding and resourcesnecessary to bring it to success. Since 1902, various inventors,notably Heal105 and Hitchcock106 had suggested floating glasson molten metal for a variety of reasons, some rather strange,but no one had developed a useful process. In the 1920s, PPGmade one unsuccessful experiment to float glass on moltenmetal, using antimony (Edge107). World War II caused furtherdevelopment work to be deferred.

Around 1950, when grinding and polishing still removed∼20% of the glass, a Pilkington engineer, K. Bickerstaff, wasexperimenting with using molten tin to support a ribbon ofglass at∼600°C. However, the crucial idea for the float pro-cess, smoothing of the glass to give mirrorlike surfaces whilesupported on molten tin, was developed by Alastair Pilkington(1920–1995). Alastair was a very distant relative of the glass-making Pilkingtons who had come to their notice. He hadjoined the company as a trainee in 1948 after graduating fromCambridge and quickly showed outstanding ability.

In 1952, Alastair suggested that it might be possible to firefinish both faces of glass if it were floated on molten tin at∼1000°C then cooled to∼600°C as it passed along the bath.Small-scale pilot plants were built between 1952 and 1955. Thefirst Pilkington float glass patent108 in 1953 names both Alas-tair Pilkington and Kenneth Bickerstaff as inventors. It de-scribes the flotation of the ribbon of glass on a bath of moltenmetal and its lifting off at the cool end of the float bath. Thenext two patents109,110concern methods of immersing the glassribbon in a molten metal bath and, thus, suggest that seriousproblems were encountered with interactions between theglass, the tin, and the atmosphere in the tin bath.

An unwelcome discovery was that the original process couldonly make acceptable float glass of one thickness (∼6.5 mm),fortunately very close to that which was in greatest demand(0.25 in.). The problems of varying the thickness were there-fore studied; another patent granted in 1958111 mentions bothlateral spreading to equilibrium thickness and the use of anglededge rolls to control width further along the float bath. The useof nonwetted (graphite) edge guides to prevent lateral spread-ing is given in two patents of 1962.112,113

In 1955, the company had to decide whether to risk buildinga full-sized float plant or another very expensive twin-grinderinstallation: the float plant was chosen. This was built at Cow-ley Hill, St Helens, and brought into operation in May 1957 butdid not produce saleable glass until July 1958. That period wasa sore trial to Alastair and his colleagues, who had to solve

many completely new types of technical problems, Of course,it also was a drain on Pilkington finances, but perseveranceeventually paid off. The further work needed to improve theprocess and find ways of varying the thickness at will provedexpensive, and it was 1963 before the process became unin-terruptedly profitable (Barker114).

At that stage, reaction of the tin that penetrated a very smalldistance into the bottom surface of the glass caused a micro-scopic wrinkling of the surface. That in turn gave the surface aslight “bloom,” which had to be removed by a light polish withrouge. Several more years passed before control of the chem-istry of the float bath cured that defect and the float processcould regularly produce glass with both surfaces having thelong-desired perfectly flat fire-polished quality. Pilkington wasgranted 37 British patents, and numerous foreign patents, forvarious aspects of the float process in its first 10 years.

The first licensee was PPG, in 1963, and other companiessoon followed. Ford, who took a license in 1966, investedmuch effort in experimental work to reproduce and develop theprocess as soon as it was announced. For example, Ford, in1965, applied for a patent115 on the use of knurled rolls press-ing on the upper edges of the ribbon to make thinner glass bystretching (Fig. 22). This work may have helped them in theirbargaining when they took out a license from Pilkington. Fordappears to have been the first manufacturer to convert entirelyto the float process.

Pilkington quickly discovered that the spout or lipstone, overwhich the glass flowed onto the tin bath, could not be im-mersed in the tin because that caused very rapid corrosion atthe refractory–tin–glass line of contact. A method of control-ling the flow of the glass onto the tin bath from a nonimmersedspout, to avoid corrosion products forming the lower surface ofthe ribbon, was soon developed (Robinson116).

PPG devised a method of making float glass,117 announcedin 1975, that did not infringe the Pilkington patents. In the PPGprocess, there is direct contact between the lipstone, the glass,and the tin (McCauley118): solid silica is probably the onlyrefractory having the required properties for that purpose. Inthis process, the glass flows into the tin bath already close to itsfinal width and thickness, which may decrease the time neededto smooth the surfaces.

Alastair Pilkington was elected a Fellow of the Royal Soci-ety in 1969 and knighted in 1970; he published three interestingpapers about the development of the float process.91,119 By1976, 21 manufacturers were paying royalties to Pilkington forthe use of the float process, and the output of a typical float linehad increased from 1000 to 5000 tonnes a week. There are nowseveral dozen float glass plants worldwide, and there are nosigns of the float process being superseded. Other interestingaccounts are by Hynd120 and Edge.107,121 Schott,122 in Jena,

Fig. 22. Decreasing the thickness of float glass by both lateral andlongitudinal stretching, with knurled wheels pressing on the edges ofthe ribbon.


Germany, has a small float line making low-expansion boro-silicate float glass; this uses mechanical stirring to achievehomogeneity and needs only a very short lehr.

(4) Wired and Patterned GlassA comprehensive review of the flat glass industry should

also include the production of wired glass and rolled patternedglass, both of which have been extensively used in buildingsfor many years, but these are not treated here. Dralle and Kep-peler123 discussed their early development.

VI. Electric Lamps

(1) BulbsAt the end of the 19th century, the glass chimney for oil

lamps was still a very important domestic article. Besides sta-bilizing the flame to stop it being blown out by draughts, thechimney controlled the air flow and kept the burner warmenough to ensure vaporization of the oil. However, the oil lampwas becoming obsolete in towns. Gas lamps were widespreadbut not very efficient and had other disadvantages, such as riskof explosion and toxic fumes.

The possibility of using an electrically heated wire as a lampwas known early in the 19th century. For example, Grove124

developed a lamp using a platinum filament, platinum havingthe advantage of not needing protection from oxidation by theair. However, the search for more-efficient lamps using higherfilament temperatures continued. J. W. Swan, in England, haddemonstrated in 1860 that a carbon filament inside an evacu-ated glass envelope could make an effective incandescent elec-tric lamp, but he could not then produce a practical product. Inthe 1870s, Swan turned to photography and invented the dryphotographic plate and bromide printing paper. However, hethen returned to the problem of electric lamps, and both he andThomas Edison produced lamps capable of commercial exploi-tation in 1879. Edison installed the first successful electriclighting system in 1880. The tungsten filament was introducedby General Electric Corporation (GE) in 1913.

There was soon a rapidly increasing market for good quality,thin-walled, seamless bulbs to make incandescent electriclamps and, of course, for tubing to make the other components.The bulbs, like oil lamp chimneys, were therefore made inpaste molds, where a film of water, turned to steam by the hotglass, prevented close contact between glass and mold surface.The long-established technique of rotating the glass in a wetwooden mold gave a circular section and virtually undamagedfire-polished surface without a seam. As the design of theelectric lamp bulb soon became standardized, the bulb enve-lope was a natural product for mass production.

A semiautomatic paste mold machine for blowing lampbulbs was developed by Chamberlin125 in 1912 and used byCorning Glass Works. An alternative, the four-arm Empiremachine, was patented in 1914 and widely used (Pitt7).Turner126 reported in 1919 that the annual production of lampbulbs in the United States already exceeded 200 million. Manybulbs were then still made by hand. The American system,using a team of three—a gatherer, a blower, and an unskilledassistant—produced 1200 to 1300 bulbs in 9 h. The Britishsystem, each blower working individually, produced about 800bulbs per worker in 8 h. The Westlake127 machine, anotherinvention from Toledo, OH, was becoming popular. In 1919,Turner saw, in the United States, 12-arm Westlake machinesmaking 65 000 bulbs/d.

The Westlake machine was intended specifically for electriclamp bulbs. The operation of both the Empire and the Westlakemachines is surprisingly similar to that of a glass blower (Fig.23). At the top of each unit an arm (a pair with the Westlake)can shoot forward into the furnace to gather gobs of glass bysuction. On retraction, the gobs are dropped onto the ends ofupturned vertical blowpipes and gripped by neck rings. Somepreliminary shaping is then done before the blow-pipe unit

swings through 180° so that the parisons hang down. The blowmolds at the bottom of the machine are raised, closed aroundthe parisons, and the bulbs blown, the blow pipes rotating theentire time. After release from the machine, the tops of thebulbs need to be burned or cracked off.

Unlike the Owens machine, the Westlake machine does notrequire a rotating pot, because the smaller size of the headsdipped into the glass, the smaller quantities of glass gathered,and the shorter contact times avoid serious chilling of the melt.Another report in 1923 claimed only 45 000 bulbs/d for aWestlake machine with 12 units (24 pipes) and stated that only11 machines made more than 100 million bulbs/year in theUnited States (Dralle and Keppeler128). Westlake machinescontinue to be used to produce thin-walled drinking glasses,including stemware.

The ever-expanding market for electric lamp bulbs led Corn-ing to develop a completely new machine for their production.The linear form of Corning ribbon, or 399, machine is unlikeany other used for making hollowware. The machine was con-ceived by W. J. Woods but largely developed by D. E. Gray,Corning’s chief engineer, around 1926. The machine (Fig. 24)was so unusual that the patent129 includes 18 sheets of dia-grams as well as 15 pages of text. A continuous stream of glassfalls from a forehearth, passes between two rollers, and isturned horizontally as it falls onto a band made up of a seriesof plates, each with a hole in its center. This method of allow-

Fig. 23. Basic operations of the Westlake machine: (1) gatheringmold is filled by suction; (2) gob is dropped onto the upturned blow-pipe; (3) gob is gripped; (4) plug is retracted and the parison blown; (5)parison is allowed to sag; (6) bulb is blown after blow-pipe has beenturned so that the parison hangs down.


ing glass held by a circular ring to sag under gravity had earlierbeen exploited by Sievert. Blow heads mounted on a continu-ous caterpillar track above this band then engage the uppersurface of the ribbon and begin to blow the glass hanging downthrough the holes. Wetted paste molds mounted on a similartrack then rise from below and close around the parisons, andthe bulbs are blown as the molds rotate. When the molds openand drop away, the bulbs hang from the ribbon. After coolinga little more, they are cracked off and gathered up by a rotatingcircular table, then dropped onto the lehr belt. The rest of theribbon is broken up and returned to the furnace as cullet. Pres-ton130 mentioned that, in its early days, the machine couldmake up to 300 bulbs/min and that the ribbon moved at 0.6 m/s.A version intended to make containers was also patented butnot exploited. The first ribbon machine outside the UnitedStates was installed in Yorkshire in 1950. There are no signs ofthe ribbon machine being superseded (Suey131).

(2) TubingTubing has long been made and used as the starting material

for many glass products from flashlight bulbs to laboratorywares and industrial chemical plant products. Considerablequantities were needed for the internal parts of incandescentbulbs. Currently, the most familiar use of tubing is in fluores-cent lamps. The potential of using luminous discharge in gases,instead of tungsten filaments, for lighting was recognizedaround 1920. Much effort was expended in developing suchlamps from about 1925. This created a rapidly growing marketfor tubing.

The earliest successful mechanized process, invented byDanner,132 in 1917, is unusual in being obviously based onreproducing the actions of a glass blower but making the pro-cess continuous. A stream of glass flows from the furnace ontothe upper end of a rotating mandrel inclined at 15°–20° to thehorizontal, the mandrel being inside a muffle furnace. Theglass forms a uniform layer as it flows along the mandrel, thenthe rod or tube is drawn away from the lower end. When tubingis being made, the mandrel has low-pressure compressed airblown through its hollow center. The glass is drawn away andcarried by rollers along a horizontal track∼30 m long. Towardthe end of the track, two caterpillar belts gently grip the tube(or rod) above and below to pull it along the track. Thesedriving belts are usually set at slight opposing angles to thedirection of draw to give the tubing a slight rotation to coun-teract the rotation of the mandrel. As a result, the molten lumpof glass hanging from the end of the mandrel may look alarm-ingly off-center and asymmetrical. The Danner process was

described by Gehlhoff133 and by Sibilia;134 Bajorat andWeiss135 reported some investigations of its control.

Several other tube-drawing processes are used. A processcredited to Philips in the Netherlands, otherwise very similar tothe Danner process, draws the tubing from the inside of ahollow mandrel. This decreases the possible contamination ofthe inner surface of the tubing to which phosphors or othercoatings may be applied. Several methods for drawing eithervertically upward or downward have been patented and ex-ploited, for example by Corning and by Schuller (see Giegerichand Trier18).

Since the 1920s, highly refractory tubing has been in de-mand for special types of lamps. In the early days, high-pressure mercury lamps made severe demands, then high-power tungsten filament lamps were needed for slide and kine´projectors as well as searchlights. More recently, the develop-ment of quartz–halogen lamps required higher operating tem-peratures for lamp envelopes than most glasses can provide.

Effective methods of making good quality quartz glass tub-ing were therefore being sought as early as 1930. The Ha¨n-lein136 process became the standard method.

VII. Optical GlassProduction of optical glass may be considered chiefly a

matter of special melting procedures but was not dealt with inthe author’s previous centennial review.137 It deserves inclu-sion, because it is a separate and very important branch of theindustry.

The classic process for making good optical glass by stirringthe melt in a pot to homogenize it was invented by a Swiss,P.-L. Guinand, around 1790. On completion of melting, the potwas cooled very slowly, then the cold pot broken open and thebest lumps of glass taken for further processing. The processwas first brought to commercial success around 1805 by Gui-nand and Fraunhofer in Bavaria. Guinand soon left that col-laboration and resumed manufacture on his own account, con-tinuing until his death in 1824. The manufacture was latercarried to France by members of Guinand’s family. One oftheir collaborators, Georges Bontemps, took the knowledge toChance Brothers in England in 1848, and Chance Brothersremained a major manufacturer until about 1960, some timeafter they were acquired by Pilkington; the Chance–Pilkingtonworks has continued to make optical glasses in Britain. In1931, Hampton and Wheat138 reported that, despite many in-vestigations of more-complex stirrers, Guinand’s original pro-cedure remained as good as any. That remained the onlymethod until the 1940s.

Until about 1880, the main requirement was for crown andflint glasses to make achromatic doublets. However, the workof Abbe, Schott, and Zeiss in Germany, beginning around 1885(Everett139), soon produced a much wider range of glasses formicroscope lenses; Schott continues to be one of the leadingmanufacturers. World War I made several countries initiate orexpand their facilities to replace glasses previously importedfrom Germany. World War I also increased demands for awider range of glasses for instruments, such as range findersand binoculars. The rapid development of photography andkinematography soon made further demands on optical glassmanufacturers.

The first successful attempt to make optical glass in theUnited States was, according to Glaze140 and Chance,141 un-dertaken by E. Feil (great-great-grandson of P.-L. Guinand)and G. A. Macbeth in 1888, but their production ended before1902. Manufacture did not begin again until 1912, whenBausch and Lomb, who had made an agreement with Zeiss in1908, began their own trials. By 1914, they had achieved somegood results, and the National Bureau of Standards also beganto work on the manufacture of optical glasses in that year.These two organizations began a close cooperation in 1917and, from the beginning of 1918, achieved a production of∼40000 lb/month. In England, sales by Chance Brothers averagedalmost 20 000 lb/month in 1917.

Fig. 24. Operation of the Corning ribbon machine: (1) stream ofglass falls from the forehearth, is slightly shaped by a pair of rolls andcarried forward on (2) a series of plates with circular apertures throughwhich the glass begins to sag; (3) blow heads engage the top of theribbon and begin to blow; (4) blow molds rise from below and closearound the parisons; (5) at the other end of the machine, the blowmolds open and drop away; (6) the blow heads are disengaged; and (7)bulbs are then cracked off. Ribbon is crushed and returned as cullet.


Two of the least attractive features of the classic method ofmanufacture were the very slow cooling of the whole pot,which sometimes caused devitrification, and the randomlysized and oddly shaped fragments produced by breaking thepot. The possibility of casting and rolling optical glass in thesame way plate glass was investigated by both Bausch andLomb and by PPG. Such glass, up to 2 in. (∼5.1 cm) thick, didnot meet the highest standards but was found suitable for manypurposes and extensively used by the end of World War I.Because permissible rates of cooling depend inversely on thesquare of thickness, such cast slabs could be annealed muchmore quickly than large pots. The Bicheroux method could beused to make slabs up to∼1.5 in. (∼3.8 cm) thick, but thesecontained some striae. However, because the striae werealigned parallel to the surfaces, the glass was suitable for oph-thalmic purposes.

The methods of making optical glass remained unchangeduntil the end of World War II. Chance142 reported that hisfirm’s yearly production had reached 270 tons in 1942. He alsogave some interesting data about average losses at differentstages of manufacture, showing that an average of only 27% ofthe glass melted was finally packed and dispatched (Table III).

Because 45% of the glass was lost by the time that the lumpshad been selected, manufacturers attempted to find an alterna-tive to breaking the slowly cooled pot. Considerable effort wastherefore put into various techniques of casting and improvingboth the range and the quality of molded blanks.

Much ingenuity was exercised in attempting to avoid striaecaused by pot corrosion products, which almost inevitably con-taminated the glass cast from a pot; a Morey143 patent of 1939mentions the use of platinum-lined pots. Glasses for some pur-poses were melted in platinum by Bausch and Lomb between1945 and 1950 (Glaze140). Platinum pots were also used inBritain at that time for some refractory glasses made by theGeneral Electric Company (GEC), such as those needed forpowerful projector lamps.

Some experiments were done at that time on melting opticalglass in tanks, but achieving the necessary homogeneity re-quired the development of a sufficiently effective stirrer. Therewere also doubts about the ability to sell and use the quantitiesof glass that a tank would produce. It was at first assumed thatchanging from one type of glass to another would require com-plete draining of the tank and cause unacceptably large losseson each changeover to another type of glass.

Various attempts were made to avoid contamination by re-fractory corrosion products and to achieve good homogeneityby stirring. Thus, in 1941, Hicks and Henkel144 applied for apatent for a melting unit made of platinum and comprising aV-shaped vessel that could be rotated to mix the glass and thentilted to collect all the glass in one arm for discharge.

The use of a small electrically heated tank, partly lined withplatinum and provided with means for vigorous mechanicalstirring, seemed to be very attractive. Work at Corning byErickson and DeVoe led to the design and construction of such

a furnace, which was in operation by 1943; that furnace couldproduce 50 lb/h (∼22 kg/h) of optical-quality glass.

A few years later the U.S. Air Force asked Corning to pro-duce much larger blanks of optical glass. That led to furtherwork on a larger furnace having a more complex stirrer devel-oped by De Voe145 that was fitted at the end of the furnace justbefore the stream of glass was delivered. Data published byPlatt146 indicate that such a furnace may contain 2500 lb (1.14tonne) and that the average residence time is only∼7 h. Thispermits rapid conversion from one glass to another withoutdraining the tank. This type of furnace has since then been thestandard method of making optical glasses as continuously castslab, usually about 600 mm wide and up to 100 mm thick.

Wright147 described in detail the development of opticalglass manufacture in the United States during World War I. Aseries of articles by Glaze140 covered the period up to 1953.The centenary of making optical glass in England occasioned areview by Chance148 in 1947. A more recent review of modernoptical glasses is by Deeg.149

VIII. Glass Fibers

Reaumur recognized some of the interesting properties ofglass fibers more than 250 years ago. In the 19th century, slagwool fibers were blown using steam jets and small quantities ofnovelty fabrics were woven from silk and glass fibers. The firstscientific application of very fine silica fibers, extruded by bowand arrow, was by Boys150 (1855–1944), who needed verydelicate but strong perfectly elastic torsional suspensions forsensitive galvanometers. However, successful commercial ex-ploitation of glass fibers seems to have begun in Germanybefore 1914. A detailed account by von Reis151 described theneed for fiber products in Germany in 1937, their methods ofproduction, and their properties.

In the 1930s,∼4000 tonnes of glass-fiber insulation werebeing produced annually. As processes improved, so did de-mand. Industry currently produces millions of tonnes everyyear, and the range of uses for insulation, filtration, reinforce-ment, and fabrics keeps steadily increasing.

One of the classic investigations of glass technology, that ofThomas,152 proving that glass strength does not inherently de-pend on sample size, was done by drawing and testing un-touched fibers in a clean atmosphere of low humidity.

Fibers for optical communication systems, now a very im-portant application, were discussed by Kurkjian and Prindle153

in another Centennial Feature

(1) Glass Wool for InsulationGlass fibers were used in Germany as an asbestos substitute

during World War I. The Gossler process for drawing continu-ous fibers was used by several European companies during the1920s (von Pazsiczky154), but the hanks of straight fibers wereunsuitable for bulky, low-density insulation.

Chance Brothers, in England, decided to begin the manufac-ture of glass silk in 1930 and licensed a process developed byPollak in Vienna. Cullet was heated in an electric furnace fromwhich continuous fibers were drawn out horizontally onto a 3ft (∼0.9 m) diameter rotating drum. This was operated for a fewyears but proved unsatisfactory for several reasons; corrosionof the refractory drawing orifices was one of the serious prob-lems. Gossler155 patented the use of a metal lining, clearlyintending to use steel, and used downward drawing of thefibers.

In 1933, Slayter and Thomas156 applied for a patent, notgranted until 1938, for the Owens steam-blowing process thatused the turbulence of steam jets to extrude the streams of glassfalling from orifices and curl the fibers. Such an action followsfrom the fact that gases and glasses have similar kinematicviscosities at glassmelting temperatures and, therefore, interact.Those fibers naturally formed low-density fine-fiber-wool suit-able for insulation or filtration. A related patent157 describesthe use of a metal bushing, winder, and the mechanical drawingof fibers. The Owens process remained the major one for two

Table III. Typical Losses in AllStages of Manufacturing Optical Glass†

OperationPercentage

loss

Melting 2.5Breaking down 25.0Trimming 22.0Molding 4.5Grinding and polishing 11.0Sorting 16.0Cutting 24.5Spindle grinding 2.0Remolding 5.0Fine annealing 1.0Packing 2.0

†After Chance.142


decades. The most important developments in this processwere the use of platinum-lined bushings and a platinum con-tainer that could be heated electrically.

Chance Brothers in England obtained a license for that pro-cess in 1935. Early trials by Chance Brothers with fibers laterled to establishment of the main British makers of glass-fiberproducts as a Pilkington subsidiary, Fibreglass, Ltd. Owens–Illinois and Corning continued to cooperate closely during the1930s, and their success led to the setting up of Owens–Corning in 1938.

Other more efficient methods of producing low-densityglass-wool were sought. Both Harford and Stafford,158 em-ployed by A. D. Little in the United States, and Rosengarth andHager, in Germany, developed ways of making fine curledfibers by dropping a stream of glass onto the center of a rapidlyrotating disk. The centrifugally formed fibers were then blowndownward and further extruded by a surrounding ring of gasjets. However, that method was not at first as efficient as theOwens steam-blowing process, and it did not initially prosperin the United States, although the Hager159 process was usedwidely in Europe. The markets for fiber insulation and otheruses grew rapidly on both sides of the Atlantic.

After World War II, the search for higher throughputs con-tinued. Both Saint Gobain in France and Owens–Corning ap-plied for patents to cover processes that used centrifugal forceto form fibers from a multiplicity of holes around the rim of ametal spinner. The streams of glass leaving the spinner werefurther extruded by encountering annular gas flames directeddownward. The Saint Gobain patent of Peyche`s160 was even-tually decided to have priority, and Owens–Corning negoiateda license for that process in 1956. Further French patents forrefinements of this process were granted to Peyche`s, Heyme`s,Levecque, and Charpentier.161 However, by 1958, Kleist,162 atOwens–Corning, had developed a process that used radiant orradio-frequency heating of the spinner and an annular airblower for attenuation of the fibers. Another patent by Kleistand Snow163 shows all the elements of the successful Owens–Corning process. By the early 1970s, these two processes andvariants of them had come to dominate fiber insulation manu-facture throughout the world. Although quality and productiv-ity have improved since then, the processes remain essentiallythe same.

Owens–Corning patented and introduced an important de-velopment in 1994 (Houptet al.164) when a new fiber wasmarketed under the trademark Miraflex. This material of verylow bulk density feels similar to natural cotton wool; it isalmost entirely free from the irritation of the skin often expe-rienced when glass-fiber insulation is handled. These naturallycurled fibers comprise two semicircular sections of glasseshaving different coefficients of thermal expansion that aresealed together. The fibers are produced by a novel rotaryspinner that feeds streams of both glasses to each of the severalthousand holes in the spinner.

(2) Continuous Filament and TextilesBecause some hand-woven glass-fibers had been found very

useful as electrical insulation, Slayter at Owens–Illinois, in1933, began to explore the possibilities of weaving glass-fibercloth. By the next year, a method had been developed of gath-ering steam-blown fibers from a perforated drum or conveyorto form a staple sliver that was wound onto a spool. This wasthen twisted on a standard textile machine and woven on a handloom. In 1936, Slayter and Thomas found that they could formsteam-blown fibers into continuous strands if the steam jetswere used only to cool the fibers, not to extrude them. Thisprocess gave much finer fibers than the Gossler155 methodcould produce. Subsequent developments led to the applicationof coatings during fiber drawing and improved winding of thestrand onto the spool.

The modern process using multiple orifices in platinumbushings was invented by Russell165,166at Owens–Corning andpatented in the 1950s. One of the key features was the use of

chilled-metal cooling fins to stabilize the hot glass emergingfrom the orifices. This permitted the number of orifices or tipsper bushing to be increased from 102 in the 1930s to 4000 in1999.

Several books have been devoted to the manufacture anduses of fibers; the most useful are Carroll-Porcyzynski167 andLoewenstein.168 Recent chapters about fibers are by Aubourgand Wolf169 and by Gupta.170

Acknowledgments: Numerous friends and colleagues have generouslygiven advice, encouragement, and information, always gratefully received, evenif not incorporated into this article. The assistance of the following is particu-larly acknowledged: Jim Barton, Ed Boulos, Neil Cameron, John Clark, ChuckEdge, John Edgington, Robert Gardon, David Grossman, Prabhat Gupta, An-thony Halliwell, Peter Jones, Russell Jordan, Dieter Kaboth, Roger Kent, DavidMartlew, Jeffrey Mills, Bill Prindle, Dennis Roberts, Helmut Schaeffer, Micha-ela Valery, and the staff of Sheffield City Library.

References1B. W. Tuchman,The Proud Tower.Ballantine Books, New York, 1996.2P. Davis,The Development of the American Glass Industry; p. 316. Harvard

University Press, Cambridge, MA, 1949.3W. C. Scoville,Revolution in Glassmaking. Harvard University Press, Cam-

bridge, MA, 1948.4W. J. Miller, “Apparatus for Pressing Glass Articles,” U.S. Pat. No. 1 130

918, 1915.5Anonymous, “Frank O’Neill, Pioneer in Glass Working Machinery,”Natl.

Glass Budget, 59 [10] 11–13 (1943).6E. Meigh, The Story of the Glass Bottle. Ramsden, Stoke-on-Trent, U.K.,

1972.7H. H. Pitt, “Some Notes on American Methods and Practice in the Glass

Industry,” J. Soc. Glass Technol., 2, 19–29 (1918).8P. Valstar, “Pressing Large Colour Television Screens to Tight Dimensional

Tolerances,”Glass Technol., 20, 252–56 (1979).9J. W. Giffen, “Apparatus for Forming Glass Articles,” U.S. Pat. No. 3 231

356, 1966.10(a)J. W. Giffen, “Glass Forming,” U.S. Pat. No. 2 696 699, 1954. (b)J. W.

Giffen, “Centrifugal Casting of Glass Articles,” U.S. Pat. No. 2 778 162, 1957.11J. C. Arnall and H. M. Ashley, “Improvements in the Manufacture of

Bottles and Other Articles in Blown Glass,” Br. Pat. No. 8677, 1886.12H. M. Ashley, “Improvements in the Manufacture of Bottles and Other

Articles in Glass,” Br. Pat. No. 14727, 1886.13(a)H. M. Ashley, “Improvements in Apparatus for Manufacturing Bottles

and Other Hollow Glass Articles,” Br. Pat. No. 3673, 1889. (b)H. M. Ashley,“Improvement in Machinery for Making Bottles and Other Like Hollow Glass-ware,” Br. Pat. No. 3686, 1889.

14L. Appert, “Sur le Soufflage du Verre par l’Air Comprime´ Mecanique-ment,” C. R. Acad. Sci. Paris, 96, 1635–37 (1893).

15S. English, “The Ashley Bottle Machine. A Historical Note,” J. Soc. GlassTechnol., 7, 324–34 (1923).

16E. Meigh, “The Development of the Automatic Glass Bottle Machine,”Glass Technol., 1, 25–50 (1960).

17A. W. Clark,Through a Glass Clearly.Golden Eagle, London, U.K., 1980.18G. Giegerich and W. Trier,Glass Machines(English translation). Springer,

Berlin–Heidelberg, Germany, 1969.19J. S. Atterbury and T. B. Atterbury, “Process for Molasses Pitcher,” U.S.

Pat. No. 140 236, 1873.20P. Arbogast, “Machine for Blowing Bottles,” U.S. Pat. No. 260 819, 1882.21Anonymous, “Machine Production of Containers in America,”Diamant

Glasind. Z., [3] (1911).22(a)W. J. Miller, “Automatic Glass-Blowing Machine,” Br. Pat. No. 142156,

1918. (b)W. J. Miller, “Apparatus for Forming Glass Articles,” Br. Pat. No.152597, 1918; and Br. Pat. No. 174645, 1922. (c)Anonymous, “W. J. Miller,”Am. Ceram. Soc. Bull., 25, 2–4 (1946).

23A. Wendler, “Pressblasemaschine von Edward Miller, Columbus, Ohio”;pp. 857–63 inDie Glasfabrikation, 1st ed., Vol. 2. Edited by R. Dralle. Old-enbourg, Munich and Berlin, Germany, 1911 (no later detailed account ofEdward Miller’s work has been found).

24Anonymous, “Michael Joseph Owens,”Am. Ceram. Soc. Bull., 17, 148–51(1938).

25M. J. Owens, “Apparatus for Mechanically Operating Paste Glass Molds,”U.S. Pat. No. 482 526, 1892.

26(a)M. J. Owens, “Apparatus for Operating Paste-Glass Molds,” U.S. Pat.No. 489 543, 1893. (b)M. J. Owens, “Receptacle for Leers or Annealing Ov-ens,” U.S. Pat. No. 506 719, 1893. (c)M. J. Owens, “Machine for BlowingGlass,” U.S. Pat. No. 515 838, 1894. (d)M. J. Owens, “Device for RemovingMoils from Blowpipes,” U.S. Pat. No. 515 838, 1894.

27M. J. Owens, “Glass Melting and Shaping Device,” U.S. Pat. No. 774 690,1904.

28C. E. Weeden, “The Glass Industry, 1916–1976,”Glass Technol., 17, 165–81 (1976).

29F. Redfern, “The New British 15-Arm Automatic Suction Bottle Machine,”J. Soc. Glass Technol., 11, 257–65 (1921).

30R. Dralle and G. Keppeler (Eds.),Die Glasfabrikation, 2nd ed., Vol. I; p.662. Oldenbourg, Munich, and Berlin, Germany, 1931.

31T. C. Moorshead, “The Glass Bottle Industry and Future Developments,”J.Soc. Glass Technol., 19, 282–316 (1925).


32K. E. Peiler, “Improving the Action of the Knife in the Owens Machine,”U.S. Pat. No. 1 785 209, 1931.

33R. S. Biram, “The Introduction of the Owens Machine into Europe,”J. Soc.Glass Technol., 42, 19N–45N (1958).

34A. Wyss, “Zur Entwicklungsgeschichte der Roirant-Maschinen,”Glastech.Ber., 35, 79–84 (1962).

35(a)Anonymous, “The Development of the Roirant Machines,”Glass, 13,280–82 (1936). (b)H. Severin “Die Entwicklung der Roirant-Maschine A6,”Glastech. Ber., 20, 65–71 (1942).

36J. Creaser, L. G. Creaser, and F. W. Hodkin, “Suction and Feeder-fedBottle-making Machines,”J. Soc. Glass Technol., 37, 48T–56T (1953).

37Lynch Machinery Co., “Improvements in Glass Blowing Machines,” Br.Pat. No. 160366, 1921.

38(a)W. J. Miller, “Improvements in Apparatus for Forming Articles ofGlass,” Br. Pat. No. 152597, 1922. (b)W. J. Miller, “Apparatus for FormingGlass Articles,” Br. Pat. Nos. 174644 and 174645, 1922.

39Anonymous, “Details of New Bottle Machine by Lynch,”Glass Ind., 12,119–24 (1931).

40(a)H. Brooke, “Dividing Molten Material,” Br. Pat. No. 24324, 1901. (b)H.Brooke, “Device for Cutting Molten Material and Distributing Same,” U.S. Pat.No. 723 983, 1903. (c)H. Brooke, “Automatic Feeding Devices for Glass-making Machinery,”J. Soc. Glass Technol., 14, 296–98 (1920).

41T. Stenhouse, “An Automatic Feeder,” U.S. Pat. No. 1 563 935, 1925.42Anonymous, “Karl E. Peiler,”Am. Ceram. Soc. Bull., 19, 105–106

(1940).43K. E. Peiler, “Feeder for Molten Glass,” U.S. Pat. No. 1 277 254, 1918.44K. E. Peiler, “Feeder for Molten Glass,” U.S. Pat. No. 1 277 255, 1918.45K. E. Peiler, “Improvements in Apparatus for Feeding Molten Glass,” Br.

Pat. No. 142786, 1920.46G. Dowse and E. Meigh, “Automatic Glass Feeding Devices,”J. Soc. Glass

Technol., 5, 134–55 (1921).47R. E. Swain, “A Study of Pulsating Feeders for Molten Glass,”Glass Ind.,

13, 99–102, 115–17, 133–36 (1932).48K. E. Peiler, “Improvements in or Relating to Apparatus for Feeding Mol-

ten Glass,” Br. Pat. Nos. 227078 and 227079, 1925.49H. W. Ingle, “Delivering Charges of Glass to Forming Machines,” U.S. Pat.

No. 1 670 770, 1928.50K. E. Peiler, “An Improvement in Automatic Feeders,” U.S. Pat. No. 1 680

391, 1928.51H. B. Chalmers, “Improvement in Automatic Feeder,” U.S. Pat. No. 1 558

790, 1925.52(a)W. Lindner, “Suction Gatherer Type 301,” and “Ball Gatherer Type

302,” Linder Maschinen GmbH. (b)Walter “Robot Gatherer.”53H. W. Ingle, “Glass Blowing Machine and Method,” U.S. Pat. No. 1 843

159, 1932.54E. Meigh, “The Development of the Automatic Glass Bottle Machine,”

Glass Technol., 1, 47 (1960).55H. W. Ingle, “A Method of Blowing Glassware,” U.S. Pat. No. 1 680 544,

1923.56Anonymous, “The Hartford I.S. Machine,”Glass, 5, 528–32 (1928) (the

intended continuation did not appear).57H. W. Ingle, “Leer for Annealing Glassware,” U.S. Pat. No. 1 583 046,

1926.58H. W. Ingle, “Lehr for Annealing Glassware,” Br. Pat. No. 250200, 1926.59G. Giegerich, “Uber technische Grundlagen der vollautomatischen Flas-

chenherstellung,”Glastech. Ber., 33, 441–49 (1960).60(a)T. V. Foster and S. P. Jones, “Mould Arrangement for a Cyclicly Op-

erating Glassware Forming Machine,” Eur. Pat. No. 0 153 534, 1987. (b)T. V.Foster and S. P. Jones, “Method of Cooling a Mould of a Cyclicly OperatingGlassware Forming Machine,” Eur. Pat. No. 0 167 871, 1989. (c)F. A. Fenton,T. V. Foster, S. P. Jones, and J. H. Williams, “The Application of a CoolingSystem to General Glass-Works Use with the Aid of Mathematical ModellingTechniques”; pp. 80–87 inProceedings of XIV International Glass Congress,New Delhi, Vol. III. Indian Ceramic Society, Calcutta, India, 1986.

61A. Stein, “Konstruktion und Arbeitsweise der Lynch 44 Maschine,”Glastech. Ber., 36, 259–65 (1963).

62L. Schaar and H. G. Seidel, “Stand der Heye-Maschinenentwicklung,”Glastech. Ber., 48, 43–50 (1975).

63S. P. Jones, “Method of Manufacture of Glass Container in a Section of anIS Machine,” U.S. Pat. No. 5 649 989, 1997.

64J. H. Edgington,Emhart Glass and the Story of Glass Packaging; p. 307.Emhart, Doncaster, U.K., 1995.

65(a)J. H. Lubbers, “Apparatus for Drawing Glass,” U.S. Pat. Nos. 702 013and 702 014, 1902. (b)J. H. Lubbers, “Glass Drawing Apparatus,” U.S. Pat.Nos. 702 015 and 702 016, 1902. (c)J. H. Lubbers, “Apparatus for DrawingGlass Articles,” U.S. Pat. No. 758 988, 1903.

66W. E. S. Turner, “Machinery and Methods of Manufacture of Sheet Glass,”Proc. I. Mech. E., 84, 1077–27 (1930).

67W. Clark, “Improvements in the Manufacture of Sheet Glass,” Br. Pat. No.489, 1857.

68E. Borel, “A Drawn Sheet Glass Process of 1871,”Glass Ind., 39 [9] 482,483, 509 (1958).

69(a)E. Fourcault, “Proce´de et Appareil pour la Fabrication de Verre parEtirage,” Fr. Pat. No. 320.631, 1902. (b)R. Riez, “History of the Dampre´myGlassworks,”Ver. Silic. Ind., 5, 9–11 (1934). (c)H. Goerk, “Fu¨nfzig Jahremechanische Tafelglaserzeugung nach dem Fourcault-Verfahren. Zum fu¨nfzig-sten Todestag Emile Fourcaults,”Glastech. Ber., 42, 448–51 (1969).

70(a)E. Fourcault, “Apparatus for the Manufacture of Plate Glass by Stretch-ing,” U.S. Pat. No. 717 378, 1902. (b)J. H. Lubbers, “Apparatus for the Manu-facture of Glass Sheets or Plates,” U.S. Pat. No. 901 800, 1908.

71E. Fourcault, “Improvements in and Apparatus for the Manufacture ofGlass Sheets or Plates,” Br. Pat. No. 28790, 1903.

72E. Fourcault, “Manufacture of Glass Sheets or Plates,” Br. Pat. Nos. 12431and 12603, 1906.

73(a)S. A. Brevets Fourcault, “A Method and Means for Controlling theTemperature of Molten Glass,” Br. Pat. No. 202576, 1923. (b)S. A. BrevetsFourcault, “Method and Means of Controlling or Regulating the Thickness ofSheet Glass,” Br. Pat. No. 205779, 1923.

74I. W. Colburn, “Method and Apparatus for Drawing Sheet Glass,” U.S. Pat.No. 860 528, 1903.

75W. P. Clarke, “Irving Wightman Colburn,”Am. Ceram. Soc. Bull., 22,31–34 (1943).

76(a)I. W. Colburn, “Apparatus for Drawing Sheet Glass,” U.S. Pat. No. 823581, 1906. (b)I. W. Colburn, “Method and Apparatus for Drawing Sheet Glass,”U.S. Pat. No. 840 433, 1907.

77I. W. Colburn, “Appareil pour l’Etirage du Verre,” Fr. Pat. No. 356.030,1905.

78(a)H. G. Slingluff, “Drawing Glass Sheets,” U.S. Pat. No. 1 339 229, 1920.(b)H. G. Slingluff, “Glass Drawing Process,” U.S. Pat. No. 1 364 895, 1921.(c)H. G. Slingluff, “Drawing Sheet Glass,” U.S. Pat. No. 1 376 975, 1921. (d)H.G. Slingluff, “Edge Holding Device for Sheet Glass Machines,” U.S. Pat. No.1 492 942, 1924. (e)H. G. Slingluff, “Edge Holding and Forming Device forSheet-Glass Machines,” U.S. Pat. No. 1 503 608, 1924. (f)H. G. Slingluff,“Apparatus for Drawing Glass,” U.S. Pat. No. 1 544 947, 1925. (g)H. G.Slingluff, “Apparatus for Making Sheet Glass,” U.S. Pat. No. 1 549 513, 1925.(h)H. G. Slingluff, “Apparatus for Drawing Sheet Glass,” U.S. Pat. No. 1 565821, 1925. (i)H. G. Slingluff, “Apparatus for Drawing Sheet Glass,” U.S. Pat.No. 1 598 751, 1926.

79F. Gross, “Neuzeitliche Verfahren zur Herstellung von Tafelglas,”Z. Ver.Dtsch. Ing., 71, 213–17 (1927).

80T. C. Barker,Pilkington Brothers and the Glass Industry; Chs.18 and 20.George Allen and Unwin, London, U.K., 1959.

81H. Goerk,Tazenı´ Plocheho Skla(The Drawing of Sheet Glass) (446 pages).State Publishing House for Technical Literature, Prague, Czechoslovakia,1966.

82(a)S. Takahashi and M. Ichinose, “Method and Apparatus for FormingContinuous Sheets of Glass,” U.S. Pat. No. 3 740 208, 1973. (b)S. Takahashiand M. Ichinose, “Method and Apparatus for Forming Continuous Sheets ofGlass,” Br. Pat. No. 1329433, 1973.

83S. Takahashi and M. Ichinose, “New Vertical Drawing Process for SheetGlass,”Glass Ind., 61 [4] 24, 29, 30, 32 (1980).

84D. N. Brown, “Making Sheet Glass, e.g., Microscope Cover Glass, byDownward Flow Method,” U.S. Pat. No. 2 422 466, 1947.

85S. M. Dockerty, “Sheet Forming Apparatus,” U.S. Pat. No. 3 338 696,1964.

86R. A. Humphrey, “Forming Glass Filaments with Unusual Cross Sections”;pp. 1–8 inProceedings of VII International Congress on Glass, Brussels, PaperNo. 77. Institut National du Verre, Charleroi, Belgium, 1965.

87E. Roeder and W. Egel-Hess, “Extrusion of Complicated Inner Profiles ofGlass,”Glastech. Ber., 60, 177–81 (1987).

88Anonymous, “Capt. John B. Ford Founded Plate Glass Manufacture inAmerica,” Am. Ceram. Soc. Bull., 18, 263–67 (1939).

89C. W. Fick, “Power and Methods of Drive for Plate Glass Grinding andPolishing Tables,”Glass Ind., 5, 141–44 (1924).

90Anonymous, “Bicheroux Process of Making Plate Glass,”Glass Ind., 8,207–208 (1927).

91(a)L. A. B. Pilkington, “Float: An Application of Science, Analysis, andJudgement,”Glass Technol., 12, 76–83 (1971). (b)L. A. B. Pilkington, “FlatGlass: Evolution and Revolution over 60 Years,”Glass Technol., 17, 182–93(1976).

92F. J. Huntley, “Making Glass for Ford Windshields,”Glass Ind., 4, 1–5(1923).

93F. J. Huntley, “The River Rouge Plant of the Ford Motor Company,”GlassInd., 5, 122–24 (1924).

94T. C. Barker,The Glassmakers. Pilkington: The Rise of an InternationalCompany; pp. 276–80. Weidenfeld and Nicolson, London, U.K., 1977.

95W. E. S. Turner, “The Glass Industry in North America,”J. Soc. GlassTechnol., 8, 286–302 (1924).

96E. P. Partridge, “Continuous Process for Plate Glass at the Ford RiverRouge Plant,”Ind. Eng. Chem., 21, 1168–71 (1929).

97(a)F. W. Preston (translator), E. Lutz, “The Manufacture of Rolled Plate,”Glass Ind., 11, 227–34, 255–61, 277–83 (1930). (b)ibid., 12, 6–11, 30–35,52–57, 84–87, 103–107, 128–31, 144–49, 166–71, 184–89, 204–207, 222–25,242–45 (1931). (c)ibid., 13, 6–11, 23–28, 42–46, 63–69, 84–89, 104–109, 118–23, 136–39, 149–51, 162–67. 169 (1932).

98E. Lutz, “Gewalztes Spiegelglas”; Ch. 11, pp. 829–1061 inDie Glasfab-rikation, 2nd ed., Vol. II. Edited by R. Dralle and G. Keppeler. Oldenbourg,Berlin and Munich, Germany, 1931.

99F. B. Waldron and J. H. Griffin, “Cleaning Lower Pads during Simulta-neous Polishing of Both Faces of a Travelling Strip of Glass,” U.S. Pat. No. 2419 925, 1947.

100F. B. Waldron, “Cleaning Lower Pads during Simultaneous Polishing ofBoth Faces of a Travelling Strip of Glass,” U.S. Pat. No. 2 419 926, 1947.

101G. H. Baillie, “Cleaning Lower Pads during Simultaneous Polishingof Both Faces of a Travelling Strip of Glass,” U.S. Pat. No. 2 419 927,1947.

102J. T. Chance, “Improvements in the Manufacture of Glass,” Br. Pat. No.7618, 1838.

103Compagnie de Saint-Gobain, “Dispositif de Fabrication du Verre par Cou-lage de Laminage,” Fr. Pat. No. 600.722, 1926.


104H. Bessemer, “Certain Improvements in the Manufacture of Glass,” Br.Pat. No. 12101, 1848.

105(a)W. Heal, “Manufacture of Window and Plate Glass,” U.S. Pat. No. 710357, 1902. (b)W. Heal, “Method of Making Glass Plates,” U.S. Pat. No. 1 553773, 1925.

106(a)H. K. Hitchcock, “Apparatus for Manufacturing Glass Sheets or Plates,”U.S. Pat. No. 789 911, 1905. (b)H. K. Hitchcock, “Process and Apparatus forMaking Sheet Glass,” U.S. Pat. No. 1 564 240, 1925.

107C. K. Edge, “Float Glass”; pp. 43–49 in Advances in Ceramics, Vol. 18,Commercial Glasses. Edited by D. C. Boyd and J. F. Macdowell. AmericanCeramic Society, Columbus, OH, 1986.

108L. A. B. Pilkington and K. Bickerstaff, “Improvements in or Relating tothe Manufacture of Flat Glass,” Br. Pat. No. 769692, 1957; and U.S. Pat. No. 2911 757, 1959.

109L. A. B. Pilkington, “Improvements Relating to the Manufacture of FlatGlass in Continuous Ribbon Form,” Br. Pat. No. 797101, 1958.

110L. A. B. Pilkington, “Improvements Relating to the Manufacture of FlatGlass in Continuous Ribbon Form,” Br. Pat. No. 797102, 1958.

111L. A. B. Pilkington, “Improvements in or Relating to the Manufacture ofFlat Glass,” Br. Pat. No. 893663, 1962.

112L. A. B. Pilkington, “Improvements in or Relating to the Manufacture ofSheet Glass,” Br. Pat. No. 1010914, 1965.

113L. A. B. Pilkington, “Improvements in or Relating to the Manufacture ofSheet Glass,” Br. Pat. No. 1010915, 1965.

114T. C. Barker; see Ref 94, pp. 419–20.115Ford Motor Co., “Float Method of Manufacturing Glass,” Br. Pat. No.

1085010, 1967.116A. S. Robinson, “Improvements in Flat Glass Manufacture,” Br. Pat. No.

1068345, 1963.117C. K. Edge and G. E. Kunkle, “Manufacture of Sheet Glass by Contiguous

Float Process,” U.S. Pat. No. 3 843 346, 1974.118R. A. McCauley, “Float Glass Production: Pilkington vs. PPG,”Glass

Ind., 61 [4] 18, 20–22 (1980).119L. A. B. Pilkington, “The Float Glass Process,”Proc. R. Soc. London,

A314, 1–25 (1969).120W. C. Hynd, “Flat Glass Manufacturing Processes”; Ch. 2, pp. 45–106 in

Glass: Science and Technology, Vol. 2,Processing I. Edited by D. R. Uhlmannand N. J. Kreidl. Academic Press, New York, London, Toronto, 1984.

121C. K. Edge, “Flat Glass: Update”; pp. 714-3–714-21 inHandbook of GlassManufacture, 3rd ed., Vol. 2. Edited by F. V. Tooley. Books for Industry,Ashlee, New York, 1984.

122K. Schneider, G. Lautenschla¨ger, T. Kloss, T. Kimura, Y. Takahashi, P.Keller, and W. Linz, “Microfloat Technology—A New Challenge for the Pro-duction of High-Tech Borosilicate Flat Glass,”Glastech. Ber., 68C [2] 14–23(1995).

123See Ref. 98, Ch. 12.124W. R. Grove, “On the Application of Voltaic Ignition to Lighting Mines,”

Philos. Mag. J. Sci., 27, 442–46 (1845).125B. D. Chamberlain, “Apparatus for Production of Blown-Glass Articles,”

U.S. Pat. No. 1 124 702, 1915.126W. E. S. Turner, “The Glass Industry of North America,”J. Soc. Glass

Technol., 3, 166–200 (1919).127(a)A. Kadow, “Transmission Apparatus,” U.S. Pat. No. 1 142 437, 1915.

(b)A. Kadow, “Raising and Lowering Mechanism,” U.S. Pat. No. 1 151 795,1915; (c)A. Kadow, “Glass-Forming Machine,” U.S. Pat. No. 1 251 671, 1918.

128See Ref. 30, pp. 632–33.129W. J. Woods and D. E. Gray, “Glass-Working Machine,” U.S. Pat. No. 1

790 397, 1931.130F. W. Preston, “New Lamps for Old,”Glass Ind., 12, 159–65 (1931).131N. Y. Suey, “The Ribbon Machine—Half a Century On,”Glass Ind., 60,

30–31 (1979).132Anonymous, “Edward Danner, Inventor of Tube-Drawing Machines,”

Natl. Glass Budget, 73 [48] 6, 16 (1958).133G. Gehlhoff, “Uber maschinelles Rohrziehung,”Glastech. Ber., 3, 205–13

(1925/1926).134V. E. Sibilia, “The Automatic Production of Glass Tubing and Rod on the

Danner Machine,”J. Soc. Glass Technol., 23, 292–307 (1939).135H. Bajorat and W. Weiss, “Einige Versuche u¨ber die Ursachen von

Dimensionelschwankungen am Dannerrohr,”Glastech. Ber., 38, 147–52(1965).

136(a)W. Hanlein, “Schmelzen und Verarbeitung von Quartzglas und a¨hnli-chen hochschmelzen Gla¨sern,”Glastech. Ber., 18, 308–14 (1940). (b)W. Ha¨n-lein, “Apparatus for Drawing Pipes from Quartz or Glass Having a High Con-tent in Silicic Acid,” U.S. Pat. No. 2 155 131, 1939.

137M. Cable, “A Century of Developments in Glassmelting Research,”J. Am.Ceram. Soc., 81, 1083–94 (1998).

138W. M. Hampton and W. N. Wheat, “Recent Developments in OpticalGlass Manufacture,”J. Soc. Glass Technol., 15, 306–20 (1931).

139J. D. Everett and A. Everett (translators), H. Hovestadt,Jena Glass and ItsScientific and Industrial Applications. Macmillan, London, U.K., 1903.

140F. W. Glaze, “History of Optical Glass Production in the United States,Part I,” Am. Ceram. Soc. Bull., 32, 242–47 (1953); “Part II,”32, 281–84 (1953);“Part III,” 32, 313–17 (1953).

141W. H. S. Chance, “Edmond Feil. The First Successful Manufacturer ofOptical Glass in America,”J. Soc. Glass Technol., 27, 113–32 (1943).

142Sir H. Chance, “Optical Glass in War,”J. Inst. Civil Eng., 27 [2] 93–112(1946/1947).

143G. W. Morey, “Optical Glass,” U.S. Pat. No. 2 150 694, 1939.144J. G. F. Hicks and C. F. Henkel, “Melting High-Quality Optical Glass in

Tilting Furnace and Casting into Plates,” U.S. Pat. No. 2 451 086, 1948.145C. F. De Voe, “Method and Apparatus for Stirring Glass,” U.S. Pat. No.

2 569 459, 1951.146D. H. M. Platt, “Response of a Glass Tank to Variations in the Batch

Composition,”Glass Ind., 49, 428–32 (1968).147F. E. Wright,The Manufacture of Optical Glass and Optical Systems; p.

309. U.S. Government Printing Office, Washington, DC, 1921.148Sir H. Chance, “The Centenary of Optical Glass Manufacture in England,”

Chem. Ind., 795–803 (1947).149E. Deeg, “Optical Glasses”; see Ref. 107, pp. 9–34.150C. V. Boys, “On the Production, Properties, and Some Suggested Uses of

the Finest Threads,”Philos. Mag., Ser. 5, 23, 489–99 (1887).151L. von Reis, “Herstellung und Verwendung von Glasgespinst,”Glastech.

Ber., 15, 219–28 (1937).152W. F. Thomas, “An Investigation of the Factors Likely to Affect the

Strength and Properties of Glass Fibres,”Phys. Chem. Glasses, 1, 4–18 (1960).153C. R. Kurkjian and W. Prindle, “Perspectives on the History of Glass

Composition,”J. Am. Ceram. Soc., 81, 795–813 (1998).154(a)G. von Pazsiczky, “Ueber die Herstellung und Verwendung von

Glasgespinst,”Keram. Rundsch. Kunst.–Keram., 39, 244–46 (1931). (b)G. vonPaziczky, “Herstellung, Verarbeitung und Verwendung von Glasfaden,”Glastech. Ber., 14, 206–11 (1936).

155O. Gossler, “Method and Apparatus for Making Glass Yarn,” U.S. Pat. No.1 954 732, 1934.

156G. Slayter and J. H. Thomas, “Glass Wool and Method and Apparatus forMaking the Same,” U.S. Pat. No. 2 133 236, 1938.

157G. Slayter and J. H. Thomas, “Mechanically Drawing Fibers,” U.S. Pat.No. 2 234 986, 1941.

158C. G. Harford and E. Stafford, “Method for Spinning Glass,” U.S. Pat. No.2 156 982 (1939).

159G. Hager, “Verfahren zum Herstellen von Fasern oder Gespinst aus Glas,Schlacke und a¨hnlichen in der Hitze plastischen Stoffen,” Ger. Pat. No. 539738, 1931.

160I. Peyche`s, “Method and Apparatus for the Production of Fibers,” U.S. Pat.No. 2 497 369, 1950.

161(a)I. Peyche`s and P.-R. Heyme`s, “Procede et Dispositifs pour la Fabrica-tion de Fibres a` partir de Matieres Thermoplastiques telle que le Verre,” Fr. Pat.No. 994.865, 1951. (b)M. Levecque and M. Charpentier, “Perfectionnementsaux Dispositifs pour la Fabrication de Fibres a` partir de Matieres telles que leVerre,” Fr. Pat. No. 1127.561, 1956.

162D. Kleist, “Method and Apparatus for the Attenuation of Heat SoftenableMaterials into Fibers,” U.S. Pat. No. 2 936 480, 1960.

163D. Kleist and H. J. Snow, “Apparatus for Centrifugally Forming Fibers,”U.S. Pat. No. 2 949 632, 1960.

164R. A. Houpt, R. M. Potter, T. D. Green, D. P. Aschenbeck, and C. Burdan,“Dual-Glass Fibers and Insulation Products Therefrom,” U.S. Pat. No. 5 431992, 1995.

165R. G. Russell, “Apparatus for Forming Glass Fibers,” U.S. Pat. No. 2 634553, 1953.

166R. G. Russell, “Apparatus for Production of Glass Fibers,” U.S. Pat. No.2 908 036, 1959.

167C. Z. Carroll-Porcyzynski,Inorganic Fibers. Academic Press, New York,1958.

168K. L. Loewenstein,The Manufacturing Technology of Continuous GlassFibres, 3rd ed. Elsevier, Amsterdam and London, 1993.

169P. F. Aubourg and W. W. Wolf, “Glass Fibers”; see Ref. 107, pp. 51–63.170P. K. Gupta,Fiber Reinforcements for Composite Materials; Ch. 2, pp.

19–71. Edited by A. R. Bunsell. Elsevier, 1988.171E. Peligot,Le Verre, son Histoire, sa Fabrication. Masson, Librairie de

L’Academie de Medecine, Paris, France, 1877. h

Editor’s Note: A photograph and biographical sketch of author Michael Cable can be found as part of an earlier CentennialFeature he authored, “A Century of Developments in Glassmelting Research,” for the May 1998 issue of theJournal of theAmerican Ceramic Society(page 1094). Since publication of that biographical sketch, Dr. Cable received the President’sAward at the 1998 San Francisco International Glass Congress.


Documents

centennialfeature - Sglavo manufacturing.pdf · (v) To use metal molds. Ashley in Yorkshire undertook the first trials that led to success. His original machine attempted to use only