Petrographic, Chemical and B-Isotopic Insights into …hera.ugr.es/doi/15771210.pdfPetrographic, Chemical and B-Isotopic Insights into the Origin of Tourmaline-Rich Rocks and Boron

Petrographic, Chemical and B-IsotopicInsights into the Origin of Tourmaline-RichRocks and Boron Recycling in theMartinamor Antiform (Central Iberian Zone,Salamanca, Spain)

A. PESQUERA1*, J. TORRES-RUIZ2, P. P. GIL-CRESPO1 ANDS.-Y. JIANG3

1DEPARTAMENTO DE MINERALOGIA Y PETROLOGIA, UNIVERSIDAD DEL PAIS VASCO, 48080 BILBAO, SPAIN

2DEPARTAMENTO DE MINERALOGIA Y PETROLOGIA, UNIVERSIDAD DE GRANADA, 18002 GRANADA, SPAIN

3STATE KEY LABORATORY FOR MINERAL DEPOSITS RESEARCH, DEPARTMENT OF EARTH SCIENCES,

NANJING UNIVERSITY, NANJING 210093, CHINA

RECEIVED JANUARY 16, 2004; ACCEPTED DECEMBER 8, 2004ADVANCE ACCESS PUBLICATION FEBRUARY 9, 2005

Tourmaline in the Martinamor antiform occurs in tourmalinites

(rocks with >15–20% tourmaline by volume), clastic metasedi-

mentary rocks of the Upper Proterozoic Monterrubio formation,

quartz veins, pre-Variscan orthogneisses and Variscan granitic rocks.

Petrographic observations, back-scattered electron (BSE) images,

and microprobe data document a multistaged development of tour-

maline. Overall, variations in the Mg/(Mg þ Fe) ratios decrease

from tourmalinites (0�36–0�75), through veins (0�38–0�66) to

granitic rocks (0�23–0�46), whereas Al increases in the same order

from 5�84–6�65 to 6�22–6�88 apfu. The incorporation of Al into

tourmaline is consistent with combinations of xcAl(NaR)–1 and

AlO(R(OH))–1 exchange vectors, where xc represents X-site

vacancy and R is (Mg þ Fe2þ þ Mn). Variations in xc/( xc þNa) ratios are similar in all the types of tourmaline occurrences, from

0�10 to 0�53, with low Ca-contents (mostly <0�10 apfu). Based

on field and textural criteria, two groups of tourmaline-rich rocks are

distinguished: (1) pre-Variscan tourmalinites (probably Cadomian),

affected by both deformation and regional metamorphism during the

Variscan orogeny; (2) tourmalinites related to the synkinematic

granitic complex of Martinamor. Textural and geochemical data

are consistent with a psammopelitic parentage for the protolith of the

tourmalinites. Boron isotope analyses of tourmaline have a total

range of d11B values from �15�6 to 6�8%; the lowest correspond-

ing to granitic tourmalines (�15�6 to �11�7%) and the highest to

veins (1�9 to 6�8%). Tourmalines from tourmalinites have inter-

mediate d11B values of �8�0 to þ2�0%. The observed variations

in d11B support an important crustal recycling of boron in the

Martinamor area, in which pre-Variscan tourmalinites were remo-

bilized by a combination of mechanical and chemical processes

during Variscan deformation, metamorphism and anatexis, leading

to the formation of multiple tourmaline-bearing veins and a new

stage of boron metasomatism.

KEY WORDS: tourmalinites; metamorphic and granitic rocks; mineral

chemistry; whole-rock chemistry; boron isotopes

INTRODUCTION

Minerals that combine an extensive compositional rangewith chemical zoning can provide valuable informationfor understanding the degree of solid solution and pos-sible substitution mechanisms, the nature of reactionsinvolving these minerals, the physical–chemical condi-tions of formation and the genetic environment. This isthe case for minerals of the tourmaline group. Because ofits refractory nature and occurrence in a wide variety ofgeological environments, tourmaline has received great

*Corresponding author. E-mail: [email protected]

#The Author 2005. Published by Oxford University Press. All

rights reserved. For Permissions, please e-mail: journals.permissions@

oupjournals.org

JOURNAL OF PETROLOGY VOLUME 46 NUMBER 5 PAGES 1013–1044 2005 doi:10.1093/petrology/egi009

interest in recent years and is believed to be a usefulpetrogenetic indicator and a potential exploration toolfor ore genesis (Henry & Guidotti, 1985; Henry &Dutrow, 1996; Slack, 1996; Jiang, 2001). Experimentalstudies document a large P–T–X stability range for tour-maline, from surface conditions to granulite facies andanatexis (e.g. B�eenard et al., 1985; Vorbach, 1989; Henry& Dutrow, 1996; Werding & Schreyer, 1996; von Goerneet al., 1999). However, the stability of tourmaline is crit-ically dependent upon the composition of the coexistingfluid phase (Frondel & Collette, 1957; Weisbrod et al.,1986; Morgan & London, 1989).

Tourmaline is the main host for boron in crustal rocksand can play an important role in the behavior of boronduring metamorphism. In sedimentary and metamorphicrocks, boron abundances reflect interactions between thecoexisting fluid phase, sheet silicates and tourmaline(Shaw, 1996). Clay minerals are the main carrier ofboron in low-grade metamorphic conditions (Henry &Dutrow, 1996), but dehydration reactions may lead toboron depletion during prograde metamorphism (Moranet al., 1992; Bebout et al., 1993). Tourmaline, however,tends to remain unreactive up to the highest grades ofmetamorphism and provides an important control onboron concentration in anatectic melts (London et al.,1996). The genesis of B-rich anatectic melts is expectedto depend on the thermal history and metamorphic gradeof the source rocks with respect to the stabilities ofB-bearing host minerals (Leeman & Sisson, 1996). How-ever, B-bearing magmas may not ultimately yield tour-maline granites because the boron may be lost from thegranite in late-stage fluids to the proximal country rockswhere tourmaline-rich rocks can be developed. Tourma-line is typical of granite-related hydrothermal systems inwhich boron and other volatile components are involvedin a number of processes, including differentiation, fluidphase evolution, wall-rock alteration, and metal trans-port and deposition (e.g. Pollard et al., 1987). However,whether the association of boron with metals such as Snand W results directly from the influence of boron or issimply a consequence of geochemical processes is a mat-ter of debate (London et al., 1996). The presence of B andother volatiles such as F and P in granitic melts causes adepression of both the solidus and liquidus temperatures,and an increase of the solubility of H2O in the melt(Manning, 1981; London et al., 1993; Dingwell et al.,1996). Consequently, such fluxing components enhancethe solubility of Sn, W, Nb, Ta and related incompatibleelements, thus promoting the concentration of metalspecies in residual melts and the formation of magmatic-hydrothermal ore deposits (Pollard et al., 1987).

Tourmaline is abundant in the Martinamor antiformof western Spain, where tourmaline-rich rocks related toclastic metasedimentary rocks, veins of different styles,tourmaline-bearing granitic rocks, pegmatites and W

(� Sn) mineral deposits occur. Tourmaline in theseoccurrences has been described in relation to mica schists(Martınez-Garcıa & Nicolau, 1973), granites (Gonzaloet al., 1975), scheelite mineralization (Duong et al., 1980)and tourmaline-rich rocks adjacent to pegmatites (Garcıade Figuerola et al., 1983). However, there are no detailedstudies on the origin and significance of tourmaline-richrocks in this area. The geological setting and the relation-ship between the tourmaline-rich rocks and tectonicevents support the formation of tourmaline both beforeand during Variscan deformation and metamorphism.Hence, the presence of tourmaline in a variety of rocktypes suggests an important spatial and temporal redis-tribution of boron. In this paper, the authors reportpetrographic, mineralogical and geochemical data ondifferent occurrences of tourmaline in the Martinamorantiform (Central Iberian Zone, Salamanca). Interpreta-tion of the major and trace-element compositions of thetourmalines and their host rocks, combined with fieldand petrographic observations and the boron isotopiccompositions of the tourmaline, suggest that boron wasrecycled from pre-Variscan tourmalinites by meta-morphic processes and anatexis during the Variscan oro-geny. The authors believe that these pre-Variscantourmalinites formed during the Cadomian period bythe interaction of psammopelitic rocks with granite-derived B-rich magmatic fluids.

GEOLOGICAL SETTING

The Martinamor antiform is a small area within thelarger Central Iberian Zone, which is located on thenorthern margin of the Domain of Vertical Folds (DıezBalda et al., 1990) about 15 km south of Salamanca(Fig. 1). Three main lithological assemblages crop out inthis area: (1) a sequence of pre-Ordovician stratified rocksbelonging to the Schist–Greywacke Complex (SGC) ofUpper Precambrian to Lower Cambrian age (DıezBalda, 1986); (2) granitic rocks; (3) Tertiary sedimentaryrocks (Fig. 1). Strata of the SGC were grouped into theMonterrubio and Aldeatejada formations by Dıez Balda(1986). The former unit includes psammopelites, quart-zites and conglomerates of volcanic affinity (‘por-phyroids’), with subordinate calc-silicate rocks, whereasthe latter consists chiefly of pelitic rocks. The tourmaline-rich rocks occur in the Monterrubio Formation.

The tectonic evolution of the Martinamor antiform iscomplex. According to Dıez Balda (1986) and Dıez Baldaet al. (1995), three main phases of Variscan deformation(D1, D2 and D3) are superimposed on the rocks, accom-panied by regional metamorphism. Although D1 gaverise to the most important folds in the Domain of VerticalFolds, in the Martinamor antiform, the effects of D1 aresignificantly masked by D2, whose structural featuresand distribution are closely related to the metamorphic

1014

JOURNAL OF PETROLOGY VOLUME 46 NUMBER 5 MAY 2005

zonation. D2 is characterized by the development ofisoclinal to sheath folds with a subhorizontal penetrativefoliation (S2). It has been related to ductile-shearing pro-cesses of extensional origin that account for the myloniticfoliation and C–S fabric in the granitic rocks (Dıez Baldaet al., 1995). D3 produced open, upright folds and isbelieved to be responsible for the main antiformal struc-ture of the Martinamor area. During D1 and D2, therocks followed a clockwise P–T path that reached

pressures of >4 kbar at temperatures of 600–700�C,followed by a nearly isothermal decompression duringD2, leading to low-pressure conditions (�2 kbar) recor-ded by the stability of andalusite. Later fluid infiltrationproduced retrograde effects responsible for the trans-formation of biotite to white mica and chlorite. A moredetailed description of the structural and metamorphicevolution of the region has been given by Dıez Balda(1986), Dıez Balda et al. (1990, 1995).

Bt

GrtSt

St

Bt

Grt

Sil

Sil

Fig. 1. Simplified geological map of the Martinamor antiform (modified from Dıez Balda, 1986) showing the major lithological assemblages,metamorphic zones (Bt, biotite; Grt, garnet; St, staurolite; Sil, sillimanite) and the area that includes the principal tourmalinite occurrences.

1015

PESQUERA et al. ORIGIN OF TOURMALINE-RICH ROCKS

The granitic rocks, which mostly crop out in the east-ern part of the antiform within the Monterrubio forma-tion, have been described in several studies (e.g. Gonzaloet al., 1975; Saavedra et al., 1984). Two groups are recog-nized: (1) the biotitic orthogneiss of San Pelayo; (2) theMartinamor complex (Fig. 1). The orthogneiss of SanPelayo has a strong foliation parallel to S2. Based ongeological and available geochronological data [Rb–Srwhole-rock ages of 430 � 30 Ma (Linares et al., 1987)],Dıez Balda et al. (1995) argued in support of a Cadomianage for their intrusion. The Martinamor complex [Rb–Srand K–Ar whole-rock ages of 360 � 25 Ma and 334 �10 Ma, respectively (Linares et al., 1987)] comprises aswarm of lenticular tourmaline-bearing leucograniticand pegmatite bodies, synkinematic to D2, which altern-ate with metamorphic rocks parallel to S2. Within thesame area, a small granitic stock (1 km2), the two-micagranite of Sta. Genoveva intruded during D3. Galibert(1984) reported a zircon U–Pb age of 298 � 28 Ma forthis granite. Abundant quartz veins occur throughout theantiform.

ANALYTICAL TECHNIQUES

Chemical analyses of 30 representative bulk-rock samples(metasediments, tourmalinites, orthogneisses and gran-ites) from the Martinamor antiform and 14 tourmalinemineral separates from tourmalinites, granites andquartz veins were performed at Granada University.Rock samples were crushed in a steel jaw crusher andpulverized in a tungsten carbide shatterbox. Major ele-ments were determined by X-ray fluorescence (XRF)using a Philips PW1404/10 X-ray spectrometer by fusingwith lithium tetraborate and casting into glass discs. Thetrace element Zr was also measured by X-ray fluores-cence using pellets of pressed rock powder. X-ray countswere converted into concentrations by a computer pro-gram based on the fundamental parameters method of deJongh (1973). Precision was of 2–5% for major elements,except Mn and P (5–10%), and 2–5% for Zr. Traceelements, except Zr, were determined by inductivelycoupled plasma–mass spectrometry (ICP–MS), afterHNO3 þ HF digestion of 100 mg of sample powder ina Teflon-lined vessel at 180�C and 200 p.s.i. for 30 min,evaporation to dryness and subsequent dissolution in100 ml of 4 vol. % HNO3. Precision is better than �5%for analyte concentrations of 10 ppm. Calibration wasdone with international standards PM–S and WS–E(Govindaraju et al., 1994), dissolved in the same way(see Torres-Ruiz et al., 2003). Tourmaline separateswere prepared using the procedure of Pesquera &Velasco (1997). The purity of these separates is estimatedto be >99%; two separates (samples 3 and 4, Table 1)have accessory minerals such as zircon, xenotime and/ormonazite. Selected tourmalines were analyzed for their

boron isotopic composition following the analytical pro-cedure described by Jiang et al. (2003). During the ana-lytical period, about 20 analyses of the NIST SRM 951standard yielded an average 11B/10B ratio of 4�0089 �0�7% (2s).

Mineral compositions were determined at theUniversity of Granada with a Cameca SX50 electronmicroprobe, equipped with four wavelength-dispersivespectrometers, using both natural and synthetic stand-ards: natural fluorite (F ), natural sanidine (K), syntheticMnTiO3 (Ti, Mn), natural diopside (Ca), synthetic Fe2O3

(Fe), natural albite (Na), natural periclase (Mg), syntheticSiO2 (Si), synthetic Cr2O3 (Cr) and synthetic Al2O3 (Al).An accelerating voltage of 20 kV and beam current of30 nA, with a 1–2 mm focused electron beam, were usedto analyze tourmaline and associated minerals. Countingtimes on peaks were twice those of backgrounds, with 15 sfor Na and K; 20 s for Ti and Ca; 25 s for Fe, Si and Al;and 30 s for Mg. Data were reduced using the procedureof Pouchou & Pichoir (1985). Analytical errors are estim-ated to be on the order of �1–2% for major elements and�10% for minor elements. The mineral chemistry data-base (available as Electronic Appendix, which may bedownloaded from the Journal of Petrology website athttp://www.petrology.oupjournals.org) includes over700 analyses of all tourmaline types, and over 300 ana-lyses of associated minerals (micas, garnet, apatite, rutileand ilmenite).

Five tourmaline samples were also analyzed byMossbauer spectroscopy to assess the iron oxidationstate. The Mossbauer analyses of powdered tourmalinesamples were carried out in the Electricity and ElectronicDepartment of the University of the Basque Country.Mossbauer spectra were performed at room temperatureusing a standard spectrometer with Co–Rh source. Theisomer shift is reported relative to metallic iron. All of thespectra were collected in a multi-channel analyzer with512 channels, and the experimental data were evaluatedby means of a least-squares fitting.

TOURMALINE OCCURRENCES

AND PETROGRAPHY

Tourmaline is widespread throughout the Martinamorantiform. It has been found in three main geologicaloccurrences: (1) tourmalinites; (2) quartz veins; (3) gran-itic rocks.

Tourmalinites

The tourmalinites represent a conspicuous lithologywithin the Martinamor antiform. They occur in theUpper Proterozoic Monterrubio Formation and formdiscontinuous, ESE–WNW-trending stratiform bodieswithin psammopelitic rocks and quartzites in the

1016


http://www.petrology.oupjournals.org

staurolite zone (Fig. 1). Scheelite-bearing calc-silicaterocks are associated with tourmalinites in some areas.Unfortunately, the poor outcrop conditions, togetherwith the intense deformation, make it difficult to esti-mate the strike length and thickness of the tourmalinites.

Tourmaline-rich rocks also occur in contact withlenticular granitic and pegmatite bodies of theMartinamor complex.

Tourmalinites are fine-grained (usually <1 mm), darkbrown to black rocks, with variable amounts of

Table 1: Representative bulk-rock trace-element analyses (ppm) of tourmalinites, metasediments, granites

and orthogneisses

Tourmalinites Metasediments Granites Orthogneisses

Sample: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Li 323.7 283.0 67.4 68.0 99.8 69.7 129.1 212.9 224.7 51.6 61.7 21.0 19.6 48.9 133.7 127.8

Rb 108.1 73.1 8.0 41.9 11.2 55.2 71.4 345.5 164.7 122.0 177.7 217.9 227.3 226.8 211.7 193.7

Cs 23.0 12.8 0.4 10.2 1.4 9.2 30.0 196.2 15.9 9.1 17.0 43.6 9.7 28.4 23.3 15.0

Be 11.0 18.6 1.8 4.8 9.5 8.1 7.2 1.5 1.8 3.8 1.1 9.5 2.6 9.7 2.1 0.9

Sr 81.9 499.5 135.0 423.3 249.8 115.6 290.8 75.5 69.4 249.0 115.3 59.0 46.5 78.8 74.6 94.3

Ba 1177.7 389.2 23.0 108.4 41.2 867.2 91.9 992.3 845.6 1307.2 836.8 133.5 100.5 181.6 301.9 402.8

Sc 16.1 22.0 12.4 26.6 23.9 18.7 9.2 19.3 17.1 13.5 13.9 1.5 1.6 1.2 8.5 10.2

V 129.1 205.0 117.6 175.1 205.2 147.1 57.2 162.4 109.1 111.4 86.7 5.0 3.7 3.2 40.0 53.3

Cr 224.0 242.6 311.0 230.0 277.6 209.9 164.3 171.0 117.7 173.8 123.0 293.4 190.1 142.9 148.3 190.8

Co 2.6 3.4 7.9 7.3 11.4 3.3 5.6 9.3 7.1 7.0 10.9 2.9 1.8 1.1 4.4 6.8

Ni 13.4 23.7 25.6 23.3 51.5 6.4 9.4 34.6 12.3 12.6 27.0 3.1 1.8 0.0 9.3 11.4

Cu 24.7 88.8 10.8 83.0 14.5 5.0 5.9 48.6 19.6 37.0 30.8 1.8 118.1 14.6 6.5 16.3

Zn 72.6 287.8 144.9 156.5 270.1 132.1 49.0 142.5 98.8 62.7 102.2 21.1 21.0 17.8 50.1 53.3

Ga 28.1 47.7 27.1 32.6 41.8 29.3 13.6 26.5 26.9 21.8 22.4 18.8 16.3 17.0 19.9 19.9

Y 9.6 29.7 16.1 35.4 14.4 16.7 29.2 19.5 21.7 17.3 16.7 6.0 4.1 7.4 16.1 15.3

Nb 15.9 21.2 12.8 21.9 13.2 17.9 10.4 15.4 15.6 13.1 15.4 2.7 8.6 2.4 12.1 12.3

Ta 1.5 1.7 14.9 1.7 1.4 1.5 1.0 1.2 1.3 1.0 1.2 0.3 1.8 1.0 1.7 1.3

Zr 188.9 296.3 145.9 225.2 167.4 141.3 258.0 170.3 154.1 194.4 239.1 23.9 22.6 25.1 136.3 145.9

Hf 4.3 7.9 3.8 5.9 4.9 4.0 4.6 4.2 4.4 3.9 4.7 1.0 1.0 1.0 3.4 2.6

Sn 134.0 325.8 51.6 77.7 73.4 197.9 3.4 21.4 8.0 7.0 10.9 21.7 19.5 21.6 16.2 14.7

W 134.4 253.4 29.4 261.2 37.0 28.9 8.4 25.2 5.3 9.4 7.7 23.2 14.5 11.2 10.4 12.5

Pb 19.0 37.9 13.2 29.4 39.0 29.0 35.0 23.4 27.6 28.7 19.1 50.5 49.2 66.0 24.6 26.4

U 1.8 7.2 2.2 4.9 2.5 2.3 4.2 2.5 5.3 12.7 4.9 1.4 1.5 2.1 3.2 3.4

Th 3.39 33.1 8.84 22.1 20.18 13.30 13.97 16.02 15.72 14.48 13.06 1.10 0.66 1.23 13.07 10.64

La 8.33 44.43 14.56 52.76 18.14 32.31 40.98 47.16 41.46 41.76 43.21 3.15 2.19 4.05 26.65 25.96

Ce 17.27 84.66 31.91 95.93 42.54 61.98 84.58 93.73 85.00 82.79 87.03 7.04 4.33 8.43 57.41 55.33

Pr 2.00 9.68 3.72 12.85 5.56 7.91 9.84 11.05 10.09 9.64 10.43 0.78 0.49 1.06 6.59 6.45

Nd 7.24 34.96 14.65 47.48 21.81 28.65 37.83 42.26 37.35 34.67 39.51 2.72 1.95 3.95 24.40 23.92

Sm 1.26 6.46 3.02 8.87 4.51 5.06 7.32 7.74 7.08 6.39 7.10 0.80 0.62 1.17 4.89 4.81

Eu 0.64 1.52 0.81 1.99 1.53 1.00 1.70 1.43 1.35 1.35 1.35 0.32 0.23 0.46 0.63 0.81

Gd 1.17 5.75 2.92 7.57 3.93 4.18 7.22 6.38 6.10 5.15 6.08 0.86 0.65 1.28 4.46 4.43

Tb 0.23 0.93 0.48 1.16 0.58 0.60 1.00 0.87 0.88 0.72 0.84 0.18 0.11 0.23 0.63 0.64

Dy 1.55 5.40 2.86 6.29 2.87 2.96 5.22 3.98 4.41 3.70 3.96 1.04 0.71 1.40 3.34 3.21

Ho 0.33 1.14 0.62 1.35 0.57 0.60 1.02 0.70 0.82 0.68 0.71 0.20 0.16 0.26 0.61 0.58

Er 0.94 3.20 1.66 3.52 1.53 1.72 2.72 1.71 2.28 1.78 1.57 0.58 0.46 0.67 1.48 1.32

Tm 0.15 0.48 0.24 0.51 0.24 0.28 0.42 0.23 0.37 0.31 0.22 0.09 0.07 0.10 0.23 0.19

Yb 1.02 3.10 1.54 3.26 1.59 1.84 2.48 1.48 2.33 2.04 1.35 0.60 0.49 0.64 1.35 1.10

Lu 0.17 0.53 0.21 0.46 0.27 0.29 0.37 0.21 0.37 0.32 0.19 0.09 0.07 0.09 0.20 0.14

1017


tourmaline (up to >90% by volume). They are charac-terized by a prominent layering consisting of a finelylaminated sequence in which tourmaline-rich laminaealternate with quartz-rich laminae (Fig. 2a). The layering

is coplanar to the bedding and the main foliation (S2).Quartz and variable amounts of plagioclase, muscoviteand biotite are the most abundant minerals associatedwith tourmaline in the tourmalinites. Some layers contain

Fig. 2. Hand specimen photographs of tourmalinites: (a) finely layered quartz–tourmaline tourmalinite; (b) sheath folds in tourmalinite with cross-cutting quartz–tourmaline vein; (c) folds formed between closely spaced ductile shear zones that refold F1 isoclinal folds in quartz–tourmalinetourmalinite; (d) quartz–plagioclase–tourmaline mylonite showing tourmalinite porphyroclasts in a section perpendicular to the foliation andparallel to the stretching lineation; lenticular layers define the mylonitic foliation that wraps around tourmaline porphyroclasts; (e) same as (d) butin section, cut perpendicular to foliation and stretching lineation; (f ) quartz–tourmaline stretched veins in mylonitic tourmalinite; note foldsoverprinted by the mylonitic foliation. Tur, tourmaline; Qtz, quartz.

1018


appreciable amounts of apatite (up to 5% by volume) andrutile (up to 2% by volume). Fine-grained, subhedralgarnet was found exclusively in a tourmalinite adjacentto a thin albitite layer. In contrast, subhedral toanhedral garnet is relatively common in the surroundingpsammopelitic rocks. Small grains of zircon andmonazite are accessory minerals.

Most of the tourmalinites have been subjected to theregional deformation and metamorphism that affectedthe Martinamor area. Evidence includes features suchas: (1) sheath and isoclinal folds (Fig. 2b and c); (2)mylonites showing tourmaline porphyroclasts (‘augen’tourmalinite) (Fig. 2d and e); (3) overprinting relationsbetween foliations (Figs 2f and 3a); (4) stretching andboudinage of the tourmaline layers and crystals; (5) frac-turing and vein-filling processes (Fig. 2b, d and f); (6)extensive recrystallization. Augen tourmalinite containstourmaline porphyroclasts, separated by alternating lam-inae of recrystallized quartz � plagioclase and tourma-line, the laminae defining the foliation (S2) that wraps thetourmaline porphyroclasts and intrafolial folds (Figs 2dand 3b). It is difficult to distinguish the axial-planarfoliation associated with the isoclinal folding and themylonitic foliation, suggesting that both structures recordthe same deformational event. A preferred orientation ispresent in the tourmaline layers, as a result of the elong-ated shape of the grains. Tourmaline c-axis fabricpatterns for the tourmaline laminae indicate an orienta-tion of [001] parallel to the S2 foliation in a fan-likearrangement, and with a gentle inclination relative tothe stretching lineation (Fig. 4a).

In detail, the porphyroclasts are aggregates of numer-ous (typically >300 grains), very small (8–200 mm) tour-maline crystals that may be associated with quartz �plagioclase and rutile � pyrite along grain boundaries.They define a straight to curved internal fabric surroun-ded by the mylonitic foliation (S2) or in continuity with it.Tourmaline c-axis fabric patterns for the tourmaline por-phyroclasts display a preferred orientation of c-axes at ahigh angle to foliation (Fig. 4b). Some porphyroclastsexhibit deflection-fold and millipede microstructures(Figs 2d and 3c) like those of Bell & Rubenach (1980)and Passchier & Speck (1994). In sections cut parallel tothe mineral lineation and perpendicular to the foliation,some porphyroclasts appear as tourmaline ‘pull-aparts’wherein quartz � plagioclase have infilled between seg-ments of tourmaline (Fig. 2d). In this case, the porphyro-clasts are up to 6 mm thick and 15 mm long. Tourmalinec-axes are mostly oriented perpendicular to the longdimension of the ‘pull-aparts’. Porphyroclast-formingtourmaline may have formed by mechanical breakdownof single porphyroblasts or by coalescence processes.Cataclasis of porphyroblasts is unlikely to lead to asystematic preferred orientation of tourmaline grains. Amodel for complex garnet porphyroblast development by

multiple nucleation and coalescence has been postulatedby Spiess et al. (2001). In this case, although the growthand coalescence of tourmaline grains may be favouredalong certain orientations, it is difficult to imagine similarfabric patterns resulting from porphyroclast-formingcoalescence processes. The authors believe, alternatively,that tourmaline porphyroclasts evolved from old cleavagedomains of tourmaline into domains with a lesser degreeof alignment or microlithons (Barker, 1990) surroundedby zones of progressive shearing. Tails of tourmalineadjacent to the porphyroclasts that extend along themylonitic foliation may evolve into lens- or rod-shapeddomains of fine-grained tourmaline. Locally, a discretecrenulation cleavage defined by tourmaline-bearing darkseams of stylolitic appearance overprints the main folia-tion (S2). The dark seams are thought to result from theconcentration of insoluble material along dissolution sur-faces by solution transfer processes. Recrystallized quartzmay occur in the necks of the tourmaline boudins andcharacteristic reddish-brown tourmaline overgrowthstend to mend the separated boudins. The microstructuresdeveloped in the necks are similar to stretching veins inplaces.

Evidence for intracrystalline deformation and recoveryin tourmaline comes from the presence of crystalline-bending, deformation bands, undulose extinction andsubgrains (Fig. 3d). Aggregates of subgrains in associationwith new grains of tourmaline may account for dynamicrecrystallization processes that, in part, were controlledby subgrain rotation. Likewise, irregular lobes at grainboundaries suggest a bulge-assisted recrystallizationmechanism (Fig. 3d). Granoblastic–polygonal texturesin which undulose extinction is very rare indicate pervas-ive static recrystallization (Fig. 3e). Interstitial, atoll tosponge-like poikiloblastic tourmaline in quartzites seemsto represent a sequence of consecutive steps duringgrowth along the margins of quartz grains.

Tourmaline-rich rocks in which tourmaline is syn- topost-tectonic with respect to D2, and therefore formedduring the Variscan orogeny, are believed to be the resultof a new stage of tourmalinization related to the emplace-ment of granites and pegmatites of the Martinamor com-plex. Tourmaline crystals are oriented in the foliationplane defined by the preferred orientation of micas, mainlyaligned along two perpendicular directions (Fig. 4c).

Under the microscope, some tourmalines show adistinctive optical zoning with subhedral to euhedral,reddish-orange cores and green rims. Most of the tour-maline grains, however, display a weak optical zonationwith gradational changes in color from core to rim: chan-ging from greenish or reddish-brown to yellowish-brown.Irregular patches of green to blue color are also observed.Fine-growth lamellae, like those decribed by Taylor &Slack (1984), are observed in some grains and probablyrepresent a relict primary microstructure. Zonation in

1019


tourmaline is more clearly evident in SEM back-scatteredelectron images where more or fewer complex zoningpatterns are observed (Fig. 5). Oscillatory and sectorzoning, cellular textures, interior zones embayed by

outer zones, microfracturing and overgrowths are typicalfeatures of tourmaline in the tourmalinites (Fig. 5a–e).Fine-scale zonation is commonly truncated at the cor-roded edges and perturbed by cellular zones. Similar

Fig. 3. Photomicrographs of tourmaline-rich rocks: (a) crenulation cleavage (S2 subhorizontal) defined by preferred orientation of tourmalinegrains overprinting a previous foliation (S1); plane-polarized light; (b) tourmaline intrafolial fold wrapped by mylonitic tourmaline that shows theeffects of dynamic recrystallization; note presence of stretched veins; plane-polarized light; (c) tourmaline porphyroclast with a pattern similar to‘millipede’ microstructures; cross-polarized light; (d) recrystallized tourmalinite in which grains show a weak preferred orientation defining thefoliation; note different grain size of the tourmaline including large deformed grains (DG) with undulatory extinction, subgrains (SG) and bulgingof grain boundaries, and fine-grained dynamically recrystallized tourmaline (FG) that appears in discrete zones; cross-polarized light; (e)recrystallized tourmalinite showing effects of static recrystallization; cross-polarized light; (f ) elongated blocky vein with multiple inclusionbands that contain needles of tourmaline; plane-polarized light.

1020


replacive features provoked by the activity of reactivefluids during the deformation and metamorphism havebeen described in a tourmalinite vein from the TauernWindow, Eastern Alps (Henry et al., 2002) and in tourma-linites from the Sierra Nevada, Southeastern Spain(Torres-Ruiz et al., 2003). Tourmaline crystals that haveundergone stretching and boudinage may show multiplenarrow, bluish-green to reddish-brown bands (‘fuzzybands’), perpendicular to the crystalline elongation(Fig. 5f). In some cases, the back-scattered electronimages reveal possible coalescence phenomena in whichtourmaline grains include two or more subhedral to sub-rounded cores, mantled by a subhedral outer zone.

Tourmaline � quartz veins

An additional component of the tourmalinites involves avariety of tourmaline � quartz veins with variable shape,

orientation and dimension that range from submillimet-ric to metric scale. With existing data, it is very difficult toreconstruct the sequence of veining, but different genera-tions and styles related to the regional deformation can berecognized. For example, tourmaline occurs in: (1) syn- tolate-tectonic veins exhibiting a blocky texture, whichtruncate fold structures and are themselves unfolded(Fig. 2b); (2) brecciated veins; (3) necks of boudinagedcrystals; (4) tensional stepover sites associated with mylon-ites; (5) slickenfibres in which tourmaline fibres areslightly inclined to the vein margins; (6) fibrous, elongate,blocky and stretched veins (Figs 2f, 3b and f) showingsimilar characteristics to those described by Bons (2000).Elongate blocky veins of tourmaline and quartz showinclusion bands that suggest a crack-seal mechanism fortheir formation. Irregularly spaced surfaces (40–200 mm)rich in small tourmaline grains occur parallel tothe contacts. Needle-shaped crystals of tourmaline are

n=250

S

n=250

S

% Area

13579

111315

n=193

S

(a)

(c)

(b)

Fig. 4. Lower-hemisphere, equal-area stereoplots of tourmaline c-axis, measured using an optical universal stage: (a) c-axis fabric patternfor tourmaline defining the foliation in mylonitic tourmalinite (thin section taken perpendicular to foliation and parallel to stretching linea-tion); (b) c-axis fabric pattern for porphyroclasts in mylonitic tourmalinite [thin section orientated as in (a)]; (c) c-axis fabric pattern for syn- topost-D2 tourmaline in Variscan tourmalinite.

1021


arranged sub-perpendicular to the vein walls and to thesuccessive inclusion bands (Fig. 3f). Accordingly, if eachband represents one cycle of cracking and sealing, follow-ing the mechanism of Ramsay (1980), this means that theveins record several tourmaline-forming episodes, i.e. amultistage development of tourmaline at vein scale. Thisis consistent with formation via fluids in successive vein-forming cycles as proposed Yardley & Bottrell (1992) forquartz veins in Connemara (Ireland). Stretched tourma-line veins appear to have formed via multiple crack-sealevents, each crack being of the order of a few micronswide. Evidence in support of this comes from the occur-rence of stretched crystals displaying a fine-scale opticalbanding oriented normal to the elongation and parallel tothe vein walls, which accounts for the compositionalvariations. In thin section, orange–brown bands alternatewith bluish to pale green bands, with the boundaries

between them commonly marked by trails of solid inclu-sions. Tourmalines from late-tectonic veins commonlyshow an oscillatory zoning with numerous bands (Fig. 5g).

Gneisses, granites and pegmatites

Tourmaline is relatively abundant in the Martinamorgranitic complex. It occurs as black, fine-to-medium-grained crystals (<2 mm) individually disseminatedthrough the granite, but also concentrated in ovoid orlenticular clots (normally <1 cm) with long axes alignedaccording to the granitic foliation. In thin section, tour-maline crystals are unzoned or weakly zoned with euhed-ral to anhedral shapes, and are typically embayed byquartz � feldspar. In back-scattered electron images(Fig. 5h and i), many tourmalines display a zonationcharacterized by euhedral to subhedral cores and

Fig. 5. Back-scattered electron (BSE) images of representative tourmaline samples; BSE images of tourmaline from tourmalinites (a–f ), late-tectonic vein (g) and granites (h) and (i): (a) tourmaline crystal displaying an interior zone strongly embayed by slightly cellular outer zone withpatchy zoning; (b) spongy cellular inner zone, partially rimmed by non-cellular tourmaline; (c) a two-zone pattern in which a subhedral core isrimmed by corroded tourmaline; (d) c-axis section of tourmaline showing oscillatory zoning and cellular intermediate zones; (e) tourmaline crystaldisplaying microfractures and overgrowths (light gray) that are richer in Fe and Ti than the interior zone (dark gray); (f ) stretched tourmalinecrystal showing overgrowths (light gray) and transversal narrow bands (‘fuzzy bands’) that reflect compositional fluctuations in Al, Mg, Na, F andTi; (g) core (left)-to-rim (right) section in tourmaline displaying fine-scale chemical zonation within interior zones; (h) granitic tourmaline with arelatively homogeneous core surrounded by slightly zoned rim; (i) a two-zone pattern in granitic tourmaline diplaying a strongly corroded rim.

1022


subhedral to anhedral rims. Fine-scale zonation, as thinas a few microns, may be also superimposed on the mainzones, which suggests a formation mechanism related tothe kinetics of crystal growth more than to suddenchanges in P, T or melt composition. Pleochroism variesfrom pale or bluish-green to green. Small euhedral grainsmay be enclosed by muscovite and garnet. In the SanPelayo orthogneiss, pale green to green tourmaline is anaccessory mineral and the tourmaline clots are absent.It occurs as very fine-grained, subhedral crystals andas irregularly shaped clusters associated with quartzand feldspar. Tourmaline megacrysts surrounded bybiotite � muscovite and feldspars are also present.Small biotite crystals appear partially or completelyenclosed by tourmaline.

In tourmaline-bearing pegmatites, tourmaline is anaccessory or minor constituent and forms black, fine- tocoarse-grained crystals, in a matrix of quartz, feldsparand muscovite. In thin section, the tourmaline crystalsexhibit a subhedral shape with green to bluish-greencolors. Tourmaline-bearing quartz veins are also com-monly associated with the granitic rocks.

WHOLE-ROCK CHEMISTRY

Tourmaline-bearing rocks in the Martinamor antiforminclude granites, psammopelites, quartzites and tourma-linites. Major- and trace-element data for representativesamples of tourmaline-rich rocks, metasediments and

granites are given in Tables 1 and 2. According to theclassification scheme of Frost et al. (2001), the graniticrocks are ferroan, alkali-calcic and peraluminous (ASI ¼1�20–1�40), with Fe* (FeOtot/FeOtot þ MgO) values of0�76–0�86 and MALI (Na2O þ K2O � CaO) values of6�88–7�94. The orthogneiss samples also are peralumin-ous (ASI ¼ 1�37–1�40) but with a magnesian (Fe* ¼0�71) and calc-alkalic (MALI ¼ 5�78–6�00) character.Tourmaline-rich rocks show distinctive chemical com-positions, with B2O3-contents of 0�82–9�02 wt %.Overall, these rocks have large variations in Al2O3

(16�71–30�55 wt %) and Fe2O3 total (4�5–12�78 wt %).Smaller variations are observed for MgO (2�63–5�98 wt %), CaO (0�20–1�26 wt %), Na2O (0�74–3�16 wt %), K2O (0�04–1�84 wt %) and P2O5 (<1 wt %).Tourmalinites ‘sensu stricto’ are considered to containin excess of 1�5–2 wt % B2O3 (Slack, 1982). The meta-sediments have higher SiO2 and K2O, and lowerAl2O3, TiO2, Fe2O3 total, MgO, P2O5 and F than thetourmaline-rich rocks (Table 2). Some tourmalinitesamples have high CaO (3�34 wt %), P2O5 (2�51 wt %),Na2O (up to 4�86 wt %) or K2O (up to 4�82 wt %) as aresult of abundances of apatite, plagioclase or muscovite,respectively.

Variation diagrams of major elements plotted againstAl2O3 for the metasediments (Fig. 6) show linear trendswith high degrees of correlation for Al2O3 vs TiO2

(correlation coefficient, r ¼ 0�98), Fe2O3 total (0�94) andMgO (0�97). The linear trends of the data for the

Table 2: Representative bulk-rock major-element analyses (wt %) of tourmalinites, metasediments, granites and

orthogneisses


Sample: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

SiO2 61.47 36.36 58.45 47.64 35.53 62.90 71.00 60.72 60.96 59.32 66.34 73.10 74.71 73.17 71.36 70.50

TiO2 0.89 1.26 0.62 1.32 1.23 0.81 0.67 0.83 0.82 0.81 0.76 0.05 0.04 0.07 0.41 0.47

Al2O3 19.39 30.22 20.18 23.97 30.55 19.27 14.04 18.61 18.78 19.72 15.77 15.21 14.57 14.64 14.57 14.89

B2O3 2.61 7.70 5.38 6.02 9.02 4.38 <0.01 0.10 <0.01 <0.01 <0.01 0.45 <0.01 0.06 <0.01 0.03

Fe2O3 3.41 8.44 6.37 8.37 12.78 4.50 3.76 6.65 6.60 6.38 5.24 1.01 0.70 0.66 2.83 3.43

MnO 0.05 0.12 0.10 0.05 0.10 0.04 0.15 0.06 0.08 0.06 0.04 0.02 0.02 0.02 0.06 0.06

MgO 2.54 5.98 3.24 4.37 5.02 2.70 1.53 2.74 2.90 2.67 1.85 0.24 0.10 0.19 1.03 1.27

CaO 0.20 0.67 0.91 0.97 0.45 0.17 2.41 0.34 0.27 0.59 0.69 0.57 0.45 0.62 0.97 1.09

Na2O 1.18 1.80 2.11 3.16 1.79 1.00 3.89 0.90 0.82 1.15 1.25 4.09 3.99 3.74 2.92 2.62

K2O 3.85 1.61 0.12 0.68 0.21 1.46 1.15 5.44 4.71 4.19 4.50 4.14 4.40 4.82 4.05 4.25

P2O5 0.02 0.10 0.59 0.29 0.04 0.05 0.23 0.19 0.17 0.19 0.28 0.20 0.20 0.23 0.26 0.24

F 0.39 0.84 0.40 0.33 0.61 0.29 0.04 0.12 0.10 0.09 0.09 0.06 0.03 0.03 0.08 0.08

O¼F 0.17 0.35 0.17 0.14 0.26 0.12 0.02 0.05 0.04 0.04 0.04 0.03 0.01 0.01 0.04 0.04

LOI 2.83 3.50 1.73 2.53 3.05 2.12 1.14 3.00 3.77 3.92 2.74 0.65 0.94 0.52 1.17 1.06

Total 98.66 98.25 100.04 99.56 100.12 99.56 100.00 99.65 99.94 99.51 99.51 99.77 100.13 98.76 99.67 99.97

1023


metasediments are believed to reflect mixing lines inher-ited from clay-rich and quartz-rich sedimentary precurs-ors that have been preserved through metamorphism(Argast & Donnelly, 1987). Positive correlations betweenTiO2 and Al2O3, and TiO2 versus TiO2/Zr (not shown)suggest that Ti-bearing minerals were bound to the clayfraction of the primary sediments. A negative correlationof TiO2 versus TiO2/Zr would indicate that Ti waspreferably transported via heavy minerals during sedi-mentation (H€aaussinger et al., 1993). Chemical variations

in the tourmalinites show only linear trends, with positivecorrelations for Al2O3 vs Fe2O3 total (r ¼ 0�77) and MgO(r ¼ 0�88). TiO2 vs Al2O3 exhibits considerable scatter(Fig. 6) taking into account that these elements are tradi-tionally considered immobile during geological processes(Ague, 1991; Slack et al., 1993; Walter, 1997). This maybe related to: (1) differences in processes of detrital sedi-mentation between the protoliths of the tourmalinites andmetasediments; (2) the abundance of Ti in tourmaline-rich rocks being strongly affected by the distribution of

Fig. 6. Selected major- and trace-element variation diagrams for granites, metasediments and tourmalinites.

1024


rutile, whereas Ti in the metasediments is mostly har-boured in the micaceous fraction; (3) mobility of Al tosome degree in metamorphic fluids, which can contributeto the tourmaline present in the metamorphic veins. Thiscould explain the lack of correlation between Al2O3 andYb (r ¼ 0�05), and the weak positive correlation betweenAl2O3 and Th (r ¼ 0�40) (Fig. 6). Instead, better correla-tions for TiO2 vs Yb (r ¼ 0�43) and Th (r ¼ 0�63) (notshown) denote less mobility of Ti during hydrothermaland metamorphic processes. Evidence for Al or Timobility in such processes has been documented by someworkers (e.g. Tracy et al., 1983; Gier�ee, 1990; Kerrick,1990; Moine et al., 1998). Likewise, the solubility ofTi-bearing minerals appears to be considerable in somehalogen-rich fluids (Ayers & Watson, 1993). CaO, Na2Oand K2O also lack correlations (not shown) with Al2O3.Broad positive correlations (Fig. 6) between B2O3 andAl2O3 (r ¼ 0�70), Fe2O3 total (0�72) and MgO (0�83) arelargely controlled by the modal-proportion tourmaline inthe analyzed rocks.

The trace-element compositions of the tourmaline-richrocks are characterized by relatively low contents of Rb,Cs and Ba, enrichment of Be, Sr, Sc, V, Cr, Zn, Y, Zr andNi, and strong enrichment of Sn and W compared withthe psammopelites (Table 1). Although fluorapatite ispresent in the metasediments and tourmaline-rich rocks,the covariation of Al2O3 with B2O3 and F (r ¼ 0�70 and0�83, respectively) (Fig. 6) indicates that F is also con-tained in tourmaline and micas, as is evidenced from themicroprobe data (Tables 3 and 4). The covariation of Fwith Sn (r ¼ 0�86) (Fig. 6), W (0�66), Li (0�61) and Be(0�64) suggests a link to granite-derived magmatic-hydrothermal fluids. Whereas Sn and W are mainly acco-modated by rutile, Li mostly partitions into micas asindicated by the low Li-contents of the mica-free tourma-linites and the positive correlations of Li with K2O andBa (r ¼ 0�78 and 0�74, respectively). Partitioning of Liinto micas is favored by relatively high mica–fluid distri-bution coefficient for Li (Volfinger & Robert, 1980).

Contents of rare earth elements (REE) are low in thegranites compared with the orthogneisses, metasedimentsand tourmaline-rich rocks (Table 1). The Zr-contents inmetasediments and tourmalinites are within a range ofvalues (200 � 100 ppm) that suggests a minor or insigni-ficant influence of zircon on REE distribution (seeMcLennan, 1989). This feature, together with the poorcorrelation between Zr and HREE (r < 0�40) (not shown)and the covariation of Ca (r ¼ 0�73), Y (r ¼ 0�74) andHREE (r ¼ 0�50–0�60) with P (not shown), could indicatethat the REE distribution is dependent on other access-ory minerals (e.g. apatite, xenotime and monazite). Basedon the significant relationships existing between Ti–Nb(r ¼ 0�72) and Zr–Nb (0�75) (not shown) and the positivecorrelations among Th, Ce, Y and Yb (Fig. 6), it isprobable that the accessory minerals containing these

elements originally occurred in the clay-sized fractionrather than in the sand-sized fraction of the sediments(see Caggianelli et al., 1992). Both the metasediments andtourmaline-rich rocks show a good correlation between Yand Yb (r ¼ 0�81) (Fig. 6), which is consistent with acoherent geochemical behavior during hydrothermaland metamorphic processes. On a plot of Ce versus Th(Fig. 6), however, the greater scatter in the data for thetourmaline-rich rocks may be due to variable mobility ofCe, relative to Th, during such processes. Chondrite-normalized plots of REE concentrations are shown inFig. 7. In general, the granites display small positive Euanomalies (Eu/Eu* ¼ 1�06–1�16) and low La/LuCN

ratios (1�80–6�45), compared with the orthogneisses thathave uniformly negative Eu anomalies (Eu/Eu* ¼ 0�41–0�53) and high La/LuCN ratios (13�63–19�38). Most ofthe tourmalinites display high La/LuCN ratios (5�3–16�2),moderate negative Eu anomalies (Eu/Eu* ¼ 0�54–0�83)and total REE abundances ranging from 79�20 to 286�8.Two tourmalinites have the lowest REE abundances(30�31–42�31) with positive Eu anomalies and REE pat-terns similar to those of tourmaline separates from veins,suggesting high fluid/rock ratios during tourmalinization(Fig. 7). The positive Eu anomalies may record a hydro-thermal fluid component from a reduced, Eu2þ-bearingsolution at >200–250�C (Bau, 1991). Relative to tourma-linites, the metasediments show a narrow range of REEcontents (196�0–220�4), lower Eu/Eu* values (0�44–0�74)and higher La/LuCN ratios (7�3 to 24�6). REE patternsand Y/Ho ratios for the majority of the tourmalinites(av. Y/Ho ¼ 27 � 1�79) are similar to those of thepsammopelites (av. Y/Ho ¼ 26 � 1�91), which is consist-ent with a psammopelitic parentage for the protolith ofthese rocks.

MINERAL CHEMISTRY

Tourmaline composition

Average electron microprobe analyses of tourmalinefrom the Martinamor antiform are grouped in Table 3according to the host rock, i.e. tourmaline-rich rocks,granitic rocks and veins. Overall, significant variationsare observed for FeO (4�66–12�87 wt %), MgO(2�12–8�29 wt %) and TiO2 (0�11–2�66 wt %). Smallervariations occur for SiO2 (34�57–38�15 wt %), Al2O3

(29�52–36�07 wt %), Na2O (1�46–2�41 wt %) and CaO(0�01–1�29 wt %). Amounts of MnO (<0�30 wt % MnO)and K2O (<0�20 wt %) are very low. Fluorine ranges toover 1�23 wt %; contents of Cr and Cl are below detec-tion limit. Some significant components such as Fe3þ, Liand H2O cannot be directly determined by electronmicroprobe analysis.

Tourmaline is a complex borosilicate with a generalformula of XY3Z6(T6O18)(BO3)3V3W (Hawthorne &

1025


Henry, 1999), where X ¼ Na, Ca, K, vacancy; Y ¼ Mg,Fe2þ, Mn, Al, Li, Fe3þ, Ti4þ, Cr3þ, V3þ; Z ¼ Al, Mg,Fe3þ, Cr3þ, V3þ; T ¼ Si, Al, (B); B ¼ B, vacancy; V ¼OH, O; W ¼ OH, F, O. Several normalization proced-ures for tourmaline are possible (Henry & Dutrow, 1996).In this paper, structural formulae were calculated on thebasis of 15 (T þ Z þ Y) cations, assuming that boron has

a stoichiometric value of 3 apfu. This normalization givesa cation charge in excess of 49 that to a first approxima-tion suggests a negligible amount of Fe3þ. In general, themicroprobe data indicate that most of the tourmalinesare Al-saturated and belong to the alkali group(Hawthorne & Henry, 1999), with small amounts of Caand significant amounts of vacancies in the X site;

Table 3: Average compositions of tourmaline

Tourmalinites Veins Granitic rocks

Rock type: TM TQ TPQ Syn-tectonic Late tectonic Granites Orthogneisses

SiO2 36.56 (0.32) 36.69 (0.27) 36.76 (0.40) 36.59 (0.37) 37.12 (0.47) 36.23 (0.38) 36.51 (0.35)

Al2O3 30.95 (0.81) 32.89 (0.89) 33.15 (0.75) 32.42 (1.15) 34.05 (1.41) 34.44 (0.64) 33.73 (0.53)

TiO2 1.04 (0.41) 0.79 (0.36) 0.82 (0.41) 1.07 (0.40) 0.75 (0.42) 0.42 (0.19) 1.10 (0.33)

FeO 6.35 (0.71) 8.94 (0.86) 7.87 (1.42) 9.00 (0.85) 7.74 (1.81) 9.89 (0.86) 6.97 (0.32)

MgO 7.59 (0.41) 4.83 (0.59) 5.63 (1.02) 5.04 (0.48) 5.31 (0.76) 3.50 (0.53) 6.08 (0.23)

MnO 0.09 (0.04) 0.07 (0.02) 0.05 (0.03) 0.11 (0.04) 0.04 (0.02) 0.10 (0.06) 0.06 (0.02)

CaO 1.02 (0.20) 0.26 (0.19) 0.33 (0.24) 0.47 (0.15) 0.26 (0.23) 0.19 (0.08) 0.49 (0.08)

Na2O 2.07 (0.13) 1.89 (0.24) 2.10 (0.13) 1.89 (0.18) 1.92 (0.26) 1.88 (0.14) 2.11 (0.08)

F 0.70 (0.24) 0.41 (0.19) 0.39 (0.16) 0.38 (0.12) 0.32 (0.25) 0.31 (0.09) 0.28 (0.05)

O¼F 0.30 (0.10) 0.17 (0.08) 0.16 (0.07) 0.16 (0.05) 0.13 (0.11) 0.13 (0.04) 0.12 (0.02)

B2O3* 10.49 (0.08) 10.52 (0.06) 10.60 (0.09) 10.52 (0.09) 10.69 (0.11) 10.51 (0.09) 10.67 (0.06)

Si 6.06 (0.03) 6.06 (0.02) 6.03 (0.03) 6.05 (0.02) 6.04 (0.03) 5.99 (0.04) 5.95 (0.03)

Al 6.04 (0.12) 6.40 (0.15) 6.41 (0.12) 6.32 (0.18) 6.52 (0.22) 6.71 (0.09) 6.48 (0.09)

Ti 0.13 (0.05) 0.10 (0.05) 0.10 (0.05) 0.13 (0.05) 0.09 (0.05) 0.05 (0.02) 0.13 (0.04)

Fe 0.88 (0.10) 1.24 (0.12) 1.08 (0.20) 1.25 (0.13) 1.05 (0.26) 1.37 (0.12) 0.95 (0.05)

Mg 1.88 (0.10) 1.19 (0.15) 1.37 (0.24) 1.24 (0.12) 1.29 (0.18) 0.86 (0.13) 1.48 (0.05)

Mn 0.01 (0.01) 0.01 (0.00) 0.01 (0.00) 0.01 (0.01) 0.01 (0.00) 0.01 (0.01) 0.01 (0.00)

Ca 0.18 (0.04) 0.05 (0.03) 0.06 (0.04) 0.08 (0.03) 0.05 (0.04) 0.03 (0.01) 0.08 (0.01)

Na 0.67 (0.04) 0.61 (0.08) 0.67 (0.05) 0.61 (0.06) 0.60 (0.09) 0.60 (0.05) 0.67 (0.03)

F 0.37 (0.13) 0.21 (0.10) 0.20 (0.09) 0.20 (0.06) 0.16 (0.13) 0.16 (0.05) 0.15 (0.03)

X-vac 0.15 (0.04) 0.34 (0.09) 0.27 (0.05) 0.31 (0.09) 0.32 (0.12) 0.36 (0.05) 0.24 (0.03)

Na/(NaþCa) 0.79 (0.04) 0.93 (0.04) 0.92 (0.05) 0.88 (0.03) 0.93 (0.05) 0.95 (0.02) 0.89 (0.02)

Mg/(MgþFe) 0.68 (0.03) 0.49 (0.05) 0.56 (0.09) 0.50 (0.04) 0.55 (0.09) 0.39 (0.05) 0.61 (0.02)

xs charge 0.44 (0.05) 0.43 (0.06) 0.45 (0.06) 0.45 (0.05) 0.48 (0.07) 0.47 (0.06) 0.49 (0.05)

W(OH) 0.19 (0.13) 0.36 (0.14) 0.35 (0.11) 0.35 (0.09) 0.36 (0.09) 0.37 (0.06) 0.37 (0.04)

Schorl1 22.41 (3.61) 29.63 (3.78) 27.68 (6.66) 30.24 (5.28) 25.04 (9.78) 29.93 (3.73) 23.22 (2.32)

Dravite2 41.88 (2.22) 27.57 (6.40) 32.78 (4.93) 27.55 (4.13) 28.14 (4.99) 17.64 (3.26) 33.06 (2.48)

Uvite3 12.14 (2.61) 2.21 (1.68) 3.43 (2.70) 4.06 (1.42) 2.57 (2.59) 1.33 (0.70) 5.22 (0.98)

Feruvite4 5.84 (1.17) 2.44 (1.77) 2.31 (1.53) 4.19 (1.42) 1.95 (1.49) 2.01 (0.98) 3.22 (0.50)

Olenite5 2.25 (3.02) 3.46 (2.13) 6.23 (2.90) 2.89 (2.75) 7.30 (4.74) 12.67 (2.35) 10.48 (2.82)

Foitite6 4.84 (1.22) 17.98 (5.68) 12.25 (3.76) 15.62 (4.41) 15.34 (3.65) 22.34 (3.66) 9.76 (1.29)

Mg-Foitite7 10.64 (3.24) 16.71 (3.50) 15.32 (3.22) 15.44 (4.49) 19.66 (7.03) 14.08 (2.93) 15.04 (1.73)

No. 56 173 99 63 59 234 58

TM, TQ, and TPQ refer to tourmalinites containing, in addition to tourmaline, mica, quartz and plagioclase þ quartz,respectively. Structural formula on the basis of T þ Z þ Y cations ¼ 15. Numbers in parentheses are standard deviations (s).*Calculated. 1NaFe2þ3Al6B3Si6O27(OH)4;

2NaMg3Al6B3Si6O27(OH)4;3CaMg3(Al5Mg)B3Si6O27(OH)4;

4CaFe2þ3(Al5Mg)B3Si6-O27(OH)4;

5NaAl3Al6B3Si6O27[O3(OH)1];6c(Fe2þ2Al)Al6B3Si6O27(OH)4;

7c(Mg2Al)Al6B3Si6O27(OH)4.

1026


compositional variations are summarized in terms of anAl–Fe–Mg diagram (Fig. 8). As the concentrations ofLi are <190 ppm (Table 5), the tourmalines can bedescribed within Li-poor compositional space in whichseven components [schorl (S), dravite (D), uvite (U), fer-uvite (FV), foitite (F), magnesiofoitite (MF), olenite (L)]are required to specify the chemical variability. End-member molecules were calculated by fitting mineralcompositions based on the ideal formulae of IMA-approved natural tourmaline (Table 3) to the tourmalineanalyses, assuming

Mi ¼Xk

j¼1

CjXij ð1Þ

where Mi is the amount of the ith atom per unit formulain tourmaline consisting of k components, Xij is the atomconstituent for each component j, and Cj is theabundance of component j. Thus, we have established

the following equations:

U þ FV ¼ Ca ð2Þ

F þ MF ¼ ð1 � Na � CaÞ ð3Þ

FV ¼ Fe = ðFe þ MgÞ�Ca ð4Þ

F ¼ Fe=ðFe þ MgÞ� ð1 � Na � CaÞ ð5Þ

6S þ 6D þ 5U þ 5FV þ 9L þ 7F þ 7MF ¼ AlðtotalÞð6Þ

S þ D þ U þ FV þ L þ F þ MF ¼ 1 ð7Þ

ð3S þ 3FV þ 2FÞ=ð3S þ 3D þ 4U þ 4FV þ 2F þ 2MFÞ¼ Fe=ðFe þ MgÞ ð8Þ

Table 4: Average compositions of biotite (Bt) and muscovite (Ms)


Bt (n ¼ 15) Ms (n ¼ 43) Bt (n ¼ 16) Ms (n ¼ 8) Bt (n ¼ 15) Ms (n ¼ 6) Bt (n ¼ 15) Ms (n ¼ 9)

SiO2 36.55 (0.44) 46.75 (1.00) 36.50 (0.74) 47.20 (0.91) 37.70 (0.97) 46.63 (0.28) 35.74 (0.32) 46.83 (0.42)

TiO2 2.00 (0.33) 0.58 (0.19) 1.59 (0.18) 0.19 (0.17) 1.78 (0.25) 0.28 (0.24) 2.87 (0.33) 0.86 (0.09)

Al2O3 18.48 (0.46) 33.70 (2.06) 19.49 (0.59) 36.07 (1.48) 20.09 (0.52) 35.93 (0.48) 18.80 (0.26) 35.33 (0.57)

FeO 21.40 (0.57) 1.43 (0.50) 18.37 (0.64) 0.85 (0.50) 21.38 (0.90) 1.28 (0.16) 19.58 (0.41) 1.12 (0.08)

MnO 0.74 (0.04) 0.05 (0.04) 0.23 (0.11) 0.02 (0.02) 0.89 (0.13) 0.10 (0.02) 0.39 (0.02) 0.02 (0.01)

MgO 7.41 (0.29) 1.43 (0.82) 10.20 (0.48) 0.72 (0.51) 5.00 (0.33) 0.49 (0.15) 8.62 (0.15) 0.92 (0.13)

CaO 0.02 (0.03) 0.01 (0.01) 0.01 (0.01) 0.07 (0.10) 0.13 (0.04) 0.03 (0.06) 0.01 (0.01) 0.00 (0.00)

Na2O 0.04 (0.03) 0.47 (0.18) 0.06 (0.02) 0.72 (0.58) 0.06 (0.02) 0.39 (0.05) 0.09 (0.02) 0.46 (0.06)

K2O 8.85 (0.22) 10.40 (0.30) 9.09 (0.33) 10.02 (0.85) 8.73 (0.24) 10.91 (0.08) 9.62 (0.08) 10.79 (0.09)

F 1.11 (0.09) 0.56 (0.39) 0.51 (0.26) 0.19 (0.20) 0.53 (0.11) 0.23 (0.08) 0.57 (0.05) 0.11 (0.05)

O¼F 0.47 (0.04) 0.24 (0.17) 0.22 (0.11) 0.08 (0.09) 0.05 (0.03) 0.10 (0.03) 0.24 (0.02) 0.05 (0.02)

Total 96.16 (0.64) 95.17 (1.07) 95.88 (0.70) 95.99 (0.85) 96.10 (0.55) 96.18 (0.25) 96.08 (0.45) 96.43 (0.63)

Si 2.757 (0.019) 3.110 (0.051) 2.725 (0.042) 3.095 (0.047) 2.833 (0.053) 3.074 (0.017) 2.698 (0.015) 3.079 (0.020)

AlIV 1.243 (0.019) 0.890 (0.051) 1.275 (0.042) 0.905 (0.047) 1.167 (0.053) 0.926 (0.017) 1.302 (0.015) 0.921 (0.020)

AlVI 0.400 (0.034) 1.754 (0.112) 0.441 (0.032) 1.883 (0.078) 0.614 (0.049) 1.867 (0.027) 0.370 (0.027) 1.818 (0.017)

Ti 0.113 (0.019) 0.029 (0.009) 0.089 (0.010) 0.009 (0.009) 0.101 (0.014) 0.014 (0.012) 0.163 (0.018) 0.043 (0.005)

Fe2þ 1.350 (0.041) 0.080 (0.028) 1.147 (0.043) 0.047 (0.028) 1.345 (0.067) 0.071 (0.009) 1.236 (0.029) 0.062 (0.004)

Mn 0.047 (0.003) 0.003 (0.002) 0.015 (0.007) 0.001 (0.001) 0.057 (0.009) 0.005 (0.001) 0.025 (0.002) 0.001 (0.001)

Mg 0.833 (0.032) 0.142 (0.081) 1.136 (0.055) 0.071 (0.050) 0.560 (0.037) 0.049 (0.015) 0.970 (0.016) 0.091 (0.013)

Ca 0.002 (0.002) 0.001 (0.001) 0.001 (0.001) 0.005 (0.007) 0.010 (0.003) 0.002 (0.004) 0.000 (0.001) 0.000 (0.000)

Na 0.006 (0.004) 0.061 (0.024) 0.008 (0.003) 0.091 (0.073) 0.008 (0.003) 0.050 (0.006) 0.012 (0.003) 0.059 (0.008)

K 0.852 (0.019) 0.883 (0.022) 0.866 (0.029) 0.838 (0.073) 0.837 (0.029) 0.918 (0.007) 0.926 (0.008) 0.905 (0.011)

F 0.264 (0.021) 0.118 (0.083) 0.121 (0.059) 0.039 (0.042) 0.125 (0.025) 0.048 (0.016) 0.136 (0.012) 0.025 (0.009)

Mg/(MgþFe) 0.381 (0.014) 0.616 (0.113) 0.497 (0.017) 0.591 (0.043) 0.294 (0.015) 0.401 (0.039) 0.440 (0.009) 0.593 (0.027)

IV(F) 1.46 1.33 1.97 1.90 1.74 1.58 1.83 1.91

Structural formulae on the basis of 11 O. Numbers in parentheses are standard deviations (s).

1027


where S, D, U, FV, L, F and MF represent thepercentages of schorl, dravite, uvite, feruvite, olenite,foitite and magnesiofoitite components, respectively.

Tourmaline chemistry in tourmaline-richrocks

Tourmaline compositions from tourmalinites are sub-divided depending on the dominant mineral assemblage,

i.e. tourmalineþ quartz (TQ), tourmalineþ plagioclaseþquartz (TPQ) and tourmaline þ mica (TM) (Table 3).In general, the tourmalines show variable X/(X þ Na)ratios of 0�10–0�51 and Mg/(Mg þ Fe) ratios of 0�36–0�75, both ratios increasing and decreasing, respectively,from TM to TQ assemblages (Table 3, Fig. 9). Tourma-line from TQ and TPQ assemblages contain enough Alto fill the Z site. By contrast, many tourmalines fromTM assemblages have an Al-deficit and relatively high

10-1

100

101

102

103

Metasediments

10-1

100

101

102

103

Orthogneisses

Granites

Tourmaline from granites

Tourmalinites

10-1

100

101

102

103

La Pr Sm Gd Dy Er YbCe Nd Eu Tb Ho Tm Lu

Tourmaline from tourmalinites Tourmaline from veins

Roc

k / C

hond

rite

Roc

k / C

hond

rite

Roc

k / C

hond

rite

(a) (b)

(c) (d)

(e) (f)

La Pr Sm Gd Dy Er YbCe Nd Eu Tb Ho Tm Lu

Fig. 7. REE patterns for granites, orthogneiss, tourmalinites and metasediments, and REE patterns for selected tourmaline separates fromgranites, tourmalinites and veins. Chondrite data from McDonough & Sun (1995).

1028


Ti-contents, which reflect the influence of the uvite com-ponent and the substitution TiRAl�2 where R is (Mg þFe2þ þ Mn) (Table 3, Fig. 9). The lower Al-contents inthe TM tourmalines could also be controlled by somesubstitution of Fe3þ in accordance with the FeAl�1

exchange vector. However, the fact that the averagecharge excess of the TM tourmalines is similar to theTQ and TPQ tourmalines, together with the absence ofcompositions with charge deficiency on the xs charge vs(Fe þ Mg þ X) diagram (Medaris et al., 2003) (Fig. 9), arenot consistent with the presence of significant amounts of

Fe3þ. Mossbauer spectra support this contention. Thespectra show three Fe2þ doublets that are considered byDyar et al. (1998) to represent Fe2þ in three distinct typesof Y sites (Y1, Y2 and Y3) (Fig. 10). The proportion ofalkali-free tourmaline (foitite þ magnesiofoitite) increasesfrom TM (av. 15�18 mol %) to TQ (av. 34�68 mol %),whereas the average olenite proportion varies from2�24 mol % in TM to 6�23 mol % in TPQ. In theTQ and TPQ tourmalines, the xcAl(NaR)�1 andAlO(R(OH))�1 exchange vectors, where xc representsX-site vacancy, appear to be the dominant mechanisms

Fig. 8. Ternary X-site-vacancy–Na–Ca and Al–Fe–Mg plots [after Hawthorne & Henry, (1999), and Henry & Guidotti, (1985), respectively]showing fields of data for tourmalines from tourmalinites, veins and granites. TM, TQ and TPQ allude to tourmalinites containing, in additionto tourmaline, mica, quartz and plagioclase þ quartz, respectively. The terms Syn and Late refer to syn- and late-tectonic veins. Altot ¼ AlT þ AlZþ AlY.

1029


of Al-incorporation, judging from the Altot vs X-sitevacancy diagrams (Fig. 9). By contrast, the very shallowslope of the TM data in the Altot vs X-site vacancydiagram (Fig. 9) denotes a minor importance of thecAl(NaR)�1 exchange vector. Tourmalines from TMshow the highest Mg/(Mg þ Fe) ratios and Ca-contents. Accordingly, they have the largest dravite

(av. 41�88 mol %), uvite (12�14 mol %) and feruvite (av.5�84 mol %) components, each decreasing from TM toTQ assemblages (Table 3). Prevailing mechanisms of Ca-incorporation in TM tourmalines are consistent with oneor more of the substitutions CaMg(NaAl)�1, CaO-(Na(OH))�1 and CaMg2OHNa�1Al�2O�1 (Fig. 9). InTQ and TPQ tourmalines, the incorporation of Ca

Table 5: Trace-element contents (ppm) of tourmaline separates

Tourmalinites Veins Granites

1 2 3 4 5 6 7 8 9 10 11 12

Li 83.0 178.2 119.4 144.5 59.2 93.1 101.6 186.0 27.8 140.0 73.1 104.0

Rb 0.6 2.4 0.8 0.8 0.3 0.5 0.2 0.6 0.3 0.2 0.5 1.9

Be 6.0 10.4 5.3 10.3 8.0 7.1 6.3 6.8 9.5 10.2 4.3 5.4

Sr 273.2 423.9 232.6 243.9 248.7 251.8 163.4 405.6 114.6 332.3 10.1 15.8

Ba 1.3 3.9 3.1 8.6 3.6 2.1 1.9 3.3 1.2 1.3 0.4 1.2

Sc 21.8 16.7 23.0 29.3 28.9 24.4 16.3 20.3 34.0 23.8 6.7 8.3

V 211.8 200.1 236.2 221.1 236.4 187.2 233.1 210.6 419.2 208.2 7.1 8.5

Cr 167.0 192.5 248.8 165.7 197.9 125.9 173.5 5.3 46.0 35.4 6.9 7.4

Co 3.9 2.8 2.8 15.6 1.5 6.7 5.8 4.8 13.0 2.9 0.5 0.9

Ni 9.1 6.1 9.1 52.7 3.3 20.8 7.7 5.6 19.5 5.3 0.0 0.0

Cu 0.8 0.7 12.1 18.4 0.1 0.7 0.4 2.0 0.3 0.3 0.5 10.1

Zn 266.3 320.9 342.2 268.6 299.8 274.8 280.1 329.3 175.7 290.8 520.0 534.8

Ga 42.1 46.7 48.6 46.3 57.2 43.5 40.8 43.2 68.2 48.2 86.5 87.8

Y 2.0 11.0 18.7 16.4 7.4 3.3 9.2 0.2 0.4 0.6 3.5 2.8

Nb 0.7 2.8 3.5 5.7 0.9 0.8 3.7 0.4 0.7 0.4 1.2 0.9

Ta 0.1 0.4 0.5 4.5 0.1 0.1 0.4 0.1 0.2 0.2 0.2 0.3

Zr 44.1 22.4 165.6 188.2 47.1 30.8 41.5 0.3 2.0 3.5 6.2 19.1

Hf 1.4 0.7 5.1 6.1 1.5 1.0 1.3 0.0 0.1 0.1 0.3 0.8

Sn 76.9 116.0 78.8 83.5 152.5 67.7 79.3 94.6 50.5 143.6 11.1 13.9

W 0.3 0.6 3.4 12.6 1.3 0.7 2.0 1.4 0.1 0.4 0.6 0.9

Pb 17.8 20.8 21.1 23.6 17.2 28.9 20.6 22.4 6.7 20.8 4.6 9.9

U 0.31 0.21 3.25 4.92 0.65 0.21 1.15 0.01 0.04 0.04 1.15 1.88

Th 0.50 1.12 14.99 21.42 4.47 0.81 10.49 0.03 0.23 0.21 1.43 1.12

La 2.33 1.83 39.77 82.63 15.35 2.02 34.01 0.71 1.68 0.76 4.28 3.36

Ce 4.47 3.62 84.34 165.05 29.99 4.06 67.43 1.35 3.26 1.26 8.81 6.69

Pr 0.50 0.38 9.89 18.95 3.60 0.48 8.11 0.11 0.35 0.13 0.98 0.75

Nd 1.84 1.54 37.08 69.25 13.25 1.86 30.34 0.40 1.27 0.38 3.36 2.54

Sm 0.30 0.40 6.79 11.60 2.57 0.34 5.38 0.04 0.18 0.06 0.76 0.59

Eu 0.43 0.55 1.13 2.12 0.60 0.71 0.99 0.23 0.29 0.35 0.09 0.13

Gd 0.29 0.79 5.56 8.18 2.12 0.36 3.99 0.03 0.13 0.05 0.68 0.54

Tb 0.05 0.21 0.77 0.98 0.27 0.08 0.50 0.00 0.02 0.01 0.10 0.08

Dy 0.30 1.67 3.55 3.70 1.19 0.50 1.96 0.02 0.08 0.08 0.63 0.51

Ho 0.07 0.41 0.76 0.67 0.21 0.11 0.35 0.01 0.02 0.02 0.13 0.09

Er 0.23 1.22 1.87 1.77 0.57 0.37 0.88 0.02 0.07 0.08 0.35 0.27

Tm 0.04 0.20 0.30 0.28 0.09 0.06 0.12 0.00 0.02 0.02 0.06 0.04

Yb 0.27 1.24 1.97 1.90 0.62 0.49 0.88 0.03 0.20 0.19 0.33 0.28

Lu 0.05 0.21 0.29 0.31 0.11 0.10 0.13 0.01 0.05 0.07 0.05 0.05

1030


Fig. 9. Binary plots of X/(X þ Na) vs Mg/(Fe þ Mg), X-site vacancy vs AlTOT, X-site vacancy vs Ca, and xs charge vs (Fe þ Mg þX-site vacancy) for tourmalines from tourmalinites, veins and granites. The direction of several selected exchange vectors are shown forreference.

1031


could also be controlled by other mechanisms, such as thesubstitutions CaMgc�1Al�2 and CaMgO[(cAl(OH)]�1

(Fig. 9).Tourmaline records some compositional change as a

result of regional deformation and metamorphism(Table 6). Porphyroclasts have more Al (av. 6�46 apfu)and less Ti (av. 0�07 apfu), Fe (av. 0�92 apfu), Mg (av.1�52 apfu) and F (av. 0�16 apfu) than tourmalines definingthe S1 foliation (av. 6�39, 0�09, 0�95, 1�54 and 0�18 apfu,

respectively). Tourmalines that define the S2 foliationcontain lower Mg (av. 1�28 apfu) but higher Ti(av. 0�11 apfu), Fe (av. 1�17 apfu) and F (av. 0�20 apfu)than the tourmaline porphyroclasts and S1-tourmalines.Poikiloblastic, post-S2-tourmalines have lower Al(av. 6�36 apfu) and Mg (av. 1�16 apfu) but higher Ti(av. 0�15 apfu), Fe (av. 1�29 apfu) and F (av. 0�31 apfu).Like poikiloblastic- and S2-tourmalines, the overgrowthsthat tend to mend the separated boudins of tourmalinehave lower Mg (av. 1�25 apfu) but higher Fe (av. 1�30 apfu),Ti (av. 0�16 apfu) and F (av. 0�21 apfu) (Table 6).

The tourmalines in the tourmalinites are characterizedby three main types of chemical zonation: (1) a discon-tinuous core to rim zoning where Na, F and Ti increaseas Al decreases (common in TQ and TPQ tourmalines)(Figs 5c and 11a); (2) a discontinuous oscillatory zoningdefined by steep compositional gradients in Al, Ti, Mg,Fe, Na, Ca and F, however, with non-systematic chem-ical variations from core to rims of the crystals (Figs 5dand 11b); (3) compositional fluctuations, particularly inAl, Na, Mg, Ti and F, associated with overgrowths and‘fuzzy bands’ in elongated and boudinaged tourmalinecrystals (Figs 5e and f, and 11c and d). These composi-tional trends reflect the influence of various substitutions,in particular cAlNa�1R�1, AlOR�1OH�1, TiRAl�2,CaMg(NaAl)�1 and CaMgO(cAl(OH))�1.

Tourmaline chemistry in veins

Tourmalines from veins are divided into syn-tectonic(STV) and late-tectonic (LTV) groups, based on veinmorphology and relationship to regional deformation.Overall, these vein tourmalines display variable Mg/(Mg þ Fe) and xc/(xc þ Na) ratios of 0�38–0�66 and0�16–0�53, respectively (Fig. 9). STV tourmalines havehigher schorl (av. 30�24 mol %) than those from LTV(av. 25�04 mol %). Compared with the TM and TPQtourmalines, the content of alkali-free tourmaline in bothSTV and LTV is relatively high (av. 31�06 and 35 mol %,respectively) (Table 3). The olenite component is lower inSTV (av. 2�89 mol %) than in LTV (av. 7�31 mol %).Variations in Al are controlled by the cAl(NaR)�1 andAlO(R(OH))�1 exchange vectors (Fig. 9), the first beingthe dominant substitution (�70 and 60% for STV andLTV tourmalines, respectively). Among the minor com-ponents, uvite and feruvite together represent, on aver-age, 8�25 mol % in STV and 4�52 mol % in LTV. Themechanisms of Ca-incorporation in these tourmalines aresimilar to those that operated in TQ and TPQ tourma-lines (Fig. 9).

In BSE images, LTV tourmalines commonly displaychemical zonation (Fig. 5g). Figure 11e shows a typicaltwo-zone pattern in which the outer zone is richer in Fe,Na and F than the inner zone, whereas Mg and Ca showan inverse relation. In addition, an oscillatory chemical

93

94

95

96

97

98

99

100

% tr

ansm

itted

mm/seg

98

99

100

% tr

ansm

itted

85

88

91

94

97

100

% tr

ansm

itted

82

-4 -3 -2 -1 0 1 2 3 4

Fe2+ (1)

Fe2+ (2)

Fe2+ (3)

TM

TQ

Granite

(a)

(c)

(b)

Fig. 10. Mossbauer spectra of tourmaline samples from: (a) TMtourmalinites; (b) TQ tourmalinites; (c) Martinamor granite.According to Dyar et al. (1998), the two Fe2þ doublets with the highestD values are believed to represent Fe2þ in Y sites (Y1 and Y2), andthe Fe2þ doublet with the lower D values may be ascribed to anotherY site (Y3).

1032


zoning involving fluctuations in Fe, Mg, Al and F over-laps the two-zone pattern (Fig. 11e). The compositionalvariation of STV tourmalines has been evaluated inelongated blocky veins. These appear to have evolvedby a crack-seal mechanism and display successive acicu-lar tourmaline-bearing bands (Fig. 3f). The tourmalinethrough the bands shows significant compositional fluc-tuations in Al, Fe, Mg, Ti, Na and F.

Tourmaline chemistry in granitic rocks

In granitic rocks, the compositions of tourmalines plot inthe schorl field with Mg/(Mg þ Fe) and xc/(xc þ Na)ratios of 0�23–0�46 and 0�29–0�52, respectively. In con-trast, those of the SP orthogneiss are dravites (Mg/(Mg þFe) ¼ 0�55–0�63; xc/(xc þ Na) ¼ 0�19–0�35, with relat-ively high Ti (av. 0�13 apfu) (Fig. 9, Table 3). The amount

of Fe as Fe3þ in tourmalines from granites appears tobe negligible from crystal chemical considerations andMossbauer data (Fig. 9). In all cases, the tourmalineshave Al > 6 apfu, and the data in the Altot vs X-sitevacancy diagram indicate that both the cAl(NaR)�1

and AlO[R(OH)]�1 exchange vectors account for theincorporation of Al (Fig. 10). The alkali-defect substitu-tion seems to be more important in the granites (�60%)than in the orthogneiss (�40%), which is consistent withthe higher content of alkali-free tourmaline in granites(av. 36�42 mol %) than in orthogneisses (av. 24�80 mol %).In addition to these substitution schemes, someTschermak substitution could be operative as the amountof Si is <6�0 apfu (Table 3). Tourmaline in both typesof rock, however, has a similar olenite component(av. 12�67 and 10�48 mol %). Orthogneiss-hostedtourmalines are characterized by relatively high dravite

Table 6: Average compositions of tourmaline in relation to the regional deformation

Foliated tourmaline Porphyroclasts Poikilitic Overgrowths

S1 S2

SiO2 36.71 (0.20) 36.70 (0.40) 36.98 (0.27) 36.48 (0.28) 36.42 (0.21)

Al2O3 33.12 (0.71) 33.02 (0.90) 33.65 (0.43) 32.66 (0.39) 31.87 (0.53)

TiO2 0.72 (0.31) 0.89 (0.47) 0.54 (0.17) 1.23 (0.20) 1.26 (0.17)

FeO 6.97 (0.27) 8.46 (1.53) 6.69 (0.25) 9.31 (0.87) 9.36 (0.59)

MgO 6.32 (0.28) 5.24 (1.19) 6.25 (0.23) 4.69 (0.65) 5.03 (0.45)

MnO 0.04 (0.01) 0.08 (0.02) 0.04 (0.02) 0.04 (0.01) 0.08 (0.01)

CaO 0.37 (0.10) 0.39 (0.32) 0.33 (0.06) 0.09 (0.02) 0.52 (0.10)

Na2O 2.16 (0.09) 2.03 (0.14) 2.09 (0.08) 2.21 (0.09) 1.96 (0.10)

F 0.34 (0.12) 0.38 (0.17) 0.32 (0.09) 0.59 (0.06) 0.41 (0.06)

O¼F 0.15 (0.05) 0.16 (0.07) 0.13 (0.04) 0.25 (0.03) 0.17 (0.03)

B2O3* 10.62 (0.04) 10.58 (0.10) 10.67 (0.04) 10.52 (0.06) 10.47 (0.04)

Si 6.01 (0.02) 6.03 (0.02) 6.02 (0.03) 6.03 (0.02) 6.05 (0.02)

Al 6.39 (0.11) 6.39 (0.16) 6.46 (0.08) 6.36 (0.07) 6.24 (0.09)

Ti 0.09 (0.04) 0.11 (0.06) 0.07 (0.02) 0.15 (0.03) 0.16 (0.02)

Fe 0.95 (0.04) 1.17 (0.22) 0.92 (0.04) 1.29 (0.13) 1.30 (0.09)

Mg 1.54 (0.07) 1.28 (0.28) 1.52 (0.06) 1.16 (0.16) 1.25 (0.11)

Mn 0.01 (0.00) 0.01 (0.00) 0.01 (0.00) 0.01 (0.00) 0.01 (0.00)

Ca 0.07 (0.02) 0.07 (0.06) 0.06 (0.01) 0.02 (0.00) 0.09 (0.02)

Na 0.68 (0.03) 0.65 (0.05) 0.66 (0.03) 0.71 (0.03) 0.63 (0.03)

F� 0.18 (0.06) 0.20 (0.09) 0.16 (0.05) 0.31 (0.03) 0.21 (0.03)

X-vac 0.25 (0.03) 0.28 (0.06) 0.28 (0.03) 0.27 (0.03) 0.27 (0.05)

Na/(NaþCa) 0.91 (0.02) 0.91 (0.07) 0.92 (0.01) 0.98 (0.00) 0.87 (0.02)

Mg/(MgþFe) 0.62 (0.01) 0.22 (0.10) 0.62 (0.01) 0.47 (0.06) 0.49 (0.04)

xs charge 0.41 (0.05) 0.47 (0.05) 0.42 (0.05) 0.48 (0.04) 0.46 (0.04)

W(OH) 0.41 (0.07) 0.34 (0.10) 0.42 (0.04) 0.21 (0.04) 0.32 (0.06)

n 25 40 26 24 28

Structural formula on the basis of T þ Z þ Y cations ¼ 15. Numbers in parentheses are standard deviations (s).*Calculated.

1033


Fig.11.

Rep

rese

nta

tive

chem

ical

zon

ing

pat

tern

sfo

rto

urm

alin

es:

the

com

po

siti

on

alp

rofi

les

(a),

(b),

(c)

an

d(d

)co

rres

po

nd

toth

eto

urm

alin

es(c

),(d

),(e

)an

d(f

)fr

om

tou

rmalin

ites

inF

ig.

5;

(e)

com

po

siti

on

alp

rofi

lefo

rto

urm

alin

efr

om

late

-tec

ton

icve

in(F

ig.

5g);

(f)

com

po

siti

on

alp

rofi

les

that

corr

esp

on

dto

the

tou

rmalin

e(i)

fro

mgra

nit

esin

Fig

.5

.

1034


contents (av. 33�06 mol %) with appreciable uvite(av. 5�22 mol %) and feruvite (av. 3�22 mol %), whereasthose in the granites have averages of 17�64, 1�33and 2�01 mol %, respectively. The variations in Cafor the orthogneiss tourmalines may be influenced by oneor more of the substitutions CaMgc�1Al�2, CaMgO-[cAl(OH)]�1 and CaMg2(OH)c�1Al�2 (Fig. 9).

Tourmaline from granitic rocks may show a relativelyweak chemical zonation, commonly with a two-zonepattern (Fig. 5g and h). The transition from the core tothe outer zone is marked by a discontinuity across whichMg, Ti, Na and F increase as Al and Fe decrease(Fig. 11f). Oscillations in composition, particularly in Fe,Mg, Na and F, may be superimposed on this composi-tional trend.

Trace elements

Trace-element concentrations of representative tourma-line separates are presented in Table 5. Tourmalinesfrom tourmalinites have Li, Sn, and W in the rangeof 59�21–178�19 ppm, 67�67–152�52 ppm and 0�33–12�63 ppm, respectively. In general, the amounts of Liand Sn tend to be higher than those reported in theliterature for tourmalines from granites (Power, 1968;Neiva, 1974). In contrast, tourmalines from tourmalinitesassociated with metamorphic rocks in the Sierra Nevada(south-eastern Spain) have higher and lower contents ofLi and Sn, respectively (Torres-Ruiz et al., 2003). Exceptfor Rb, Cs, Zn and Ga, tourmalines from tourmaliniteshave higher Sr, Ba, Sc, V, Cr, Ni, Co, Y, Nb, Zr, Sn, W,Th, U and Pb, compared with the granitic tourmalines.The relatively high contents in Zr, Hf, Th, Y and REE offour tourmalines from tourmalinites (numbers 3, 4, 5 and7, Table 5) are believed to reflect the presence of zircon,xenotime and monazite in these samples.

Tourmaline from granites displays no, or negative, Euanomalies with REE-contents similar to the whole rock(Fig. 7). This is consistent with the formation of tourma-line from the melt, which is also in accordance with thetextural evidence. It appears that REE prefer melt relat-ive to coexisting aqueous fluids (Flynn & Burnham,1978), so that if tourmaline crystallized from fluids, itwould have a lower REE concentration. Tourmalinesthat derive from fluids rather than directly crystallizingfrom silicate melt are poorer in REE relative to corres-ponding melt-derived tourmaline ( Jolliff et al., 1987).

Tourmalines from tourmalinites show two differentchondrite-normalized REE patterns (Fig. 7): (a) relativelyhigh REE abundances and patterns similar to those of thewhole rock, probably influenced by heavy minerals;(b) low REE abundances and patterns characterized bypositive Eu anomalies and flat or slightly positive HREEtrends. This may be due to changes in the concentrationof fluids during the tourmalinization processes. A similarfeature was ascribed by Torres-Ruiz et al. (2003) to a

possible contribution of Eu from the hydrothermal fluids.Relative to the whole rock, tourmalines from tourmalin-ites exhibit: (1) depletion of low-field-strength elements(LFSE) such as Rb, Cs and Ba that are preferentiallypartitioned into micas and feldspar; (2) higher Li, Sc, V,Zn and Ga, but lower Cr, Ni, Co, Sn, W, Y, and high-field-strength elements (HFSE), including Zr, Hf,Nb and Ta, the latter probably controlled by thedistribution of zircon, rutile and monazite; (3) lowerHREE-contents.

Most of the trace-element and REE-contents in tour-malines from veins are lower than in tourmalines fromtourmalinites and granites (Table 5). It must be noted thattourmaline from veins displays REE patterns, particu-larly for the light REE, which are similar to those oftourmalines from tourmalinites with lower REE abund-ances (Fig. 7).

Chemistry of coexisting minerals

Garnets display a variable chemical composition, depend-ing on the host rock (Table 7). Tourmalinite-hosted gar-nets have the lowest almandine (av. 45�28 mol % Alm)and the highest spessartine (av. 52�24 mol % Sps)contents, whereas garnets from metasedimentary andgranitic rocks are higher in almandine (av. 63�79 and

Table 7: Average compositions of garnet

Tourmalinites

n ¼ 31

Metasediments

n ¼ 48

Granites

n ¼ 36

SiO2 36.20 (0.19) 37.02 (0.24) 35.95 (0.31)

TiO2 0.08 (0.07) 0.05 (0.05) 0.03 (0.02)

Al2O3 20.79 (0.10) 21.12 (0.13) 20.91 (0.09)

FeO 19.75 (1.51) 28.28 (3.74) 28.60 (1.49)

MnO 22.49 (1.37) 7.21 (2.59) 13.51 (1.60)

MgO 0.48 (0.08) 2.34 (0.56) 0.68 (0.10)

CaO 0.18 (0.12) 3.60 (2.28) 0.28 (0.03)

Total 99.96 (0.45) 99.62 (0.39) 99.97 (0.39)

Si 2.981 (0.007) 2.990 (0.009) 2.963 (0.013)

Al 2.018 (0.007) 2.011 (0.011) 2.032 (0.011)

Ti 0.005 (0.004) 0.003 (0.003) 0.002 (0.001)

Fe2þ 1.360 (0.103) 1.911 (0.256) 1.971 (0.098)

Mn 1.568 (0.097) 0.494 (0.176) 0.943 (0.115)

Mg 0.058 (0.010) 0.282 (0.067) 0.083 (0.012)

Ca 0.016 (0.011) 0.311 (0.196) 0.025 (0.003)

Pyrope 1.94 (0.32) 9.40 (2.24) 2.76 (0.39)

Almandine 45.28 (3.35) 63.79 (8.65) 65.22 (3.37)

Spessartine 52.24 (3.30) 16.47 (5.85) 31.19 (3.72)

Grossular 0.54 (0.36) 10.35 (6.49) 0.83 (0.10)

Structural formulae on the basis of 12 O. Numbers inparentheses are standard deviations (s).

1035


65�22 mol % Alm, respectively) and lower in spessartine(av. 16�47 and 31�19 mol % Sps, respectively). There isalso a significant difference in pyrope and grossular com-ponents. Garnets from the metasedimentary rocks havethe highest pyrope (av. 9�40 mol % Prp) and grossular(av. 10�35 mol % Grs), whereas those from tourmalinitesand granites show similar proportions (av. 1�94 and2�76 mol % Prp; 0�54 and 0�83 mol % Grs, respectively).Metasediment-hosted garnets show normal chemicalzoning in which Fe–Mg increases from core to rim,while Ca–Mn decreases—a pattern characteristic in gar-nets of the greenschist and amphibolite facies. Becausegarnet is known to fractionate Mn relative to other min-erals, the high content of the spessartine component inthe tourmalinite-hosted garnets probably is due to the lowmodal volume of garnet in these rocks.

Micas occur in variable amounts in tourmalinites ofthe Martinamor antiform; muscovite is dominant overbiotite. Distinct chemical compositions reflect composi-tional dependence on the nature of the host rock(Table 4). Muscovite is less aluminous and richer inMgO, FeO and F in tourmaline-rich rocks comparedwith muscovite in metasedimentary rocks, granites andorthogneiss (Table 4). Tourmalinite-hosted biotites areless aluminous but richer in F than those from the meta-sedimentary rocks, granites and orthogneiss. Amounts ofTiO2 are higher in biotites from orthogneisses than thosefrom granites, tourmalinites and metasedimentary rocks(Table 4). Ti-contents in biotite from tourmalinitesand metasedimentary rocks increase with decreasingMg/(Mg þ Fe) ratios, which seem to be controlled bycrystal-chemical factors (Henry & Guidotti, 2002). Tem-peratures calculated (�570–590�C) for biotites from thetourmalinites and metasedimentary rocks, using theTi-saturation surface method developed by Henry &Guidotti (2002), are consistent with the metamorphicconditions (staurolite zone) in the study area. Fe–Mgpartitioning between coexisting garnet and biotite in themetasediments mostly defines a lower range of temperat-ures (�520–560�C), using the calibrations compiled byReche & Martinez (1996). Such differences in temperat-ure may be due to a combination of factors: (1) specificcharacteristics of the thermometric methods; (2) differ-ences in pressure because an increase in this variabledecreases the Ti-content in biotite (Henry & Guidotti,2002); (3) Mg–Fe exchange on cooling or changes inTi by formation of rutile. The influence of pressure isexpected to be negligible relative to the 3�3 kbar isobaricsaturation surface of Henry & Guidotti (2002), taking intoaccount the pressure regime (�2–4 kbar) in the Marti-namor area during metamorphism. Fluorine enrichmentin muscovite and biotite is expressed as intercept valuesIV(F), as defined by Munoz (1984). Micas from tourma-linites have lower IV(F) values than those from metasedi-ments, granites and orthogneisses (Table 4). Such values

are in the range of intercept values reported in the liter-ature for micas associated with Sn–W–Be mineralizations(see Munoz, 1984). The calculated log( f H2O/f HF)values corresponding to biotites of the tourmalinites andmetasedimentary rocks are of the order of 4�0 and 4�5,respectively, for assumed temperatures of �550�C. Atthese temperatures, the log fugacity ratios for biotitesfrom granites and orthogneisses would be of the orderof 4�3–4�4. However, because the F exchange depends ontemperature (Munoz, 1984), it is expected that biotitesfrom granitic rocks have lower log( f H2O/f HF) values(�4�0–4�2), assuming that they stopped exchanging F athigher temperatures (600–700�C) than the biotites fromtourmalinites and metasediments.

Apatite is a common accessory mineral in all rock typesfrom the Martinamor antiform. However, in tourmaline-rich rocks, fine-grained euhedral to skeletal apatite mayoccur in significant quantities. Overall, Ca/P valuesrange between 1�60 and 1�65, and the Cl-contents arenegligible (<0�02 wt %). Amounts of SrO, FeO andMnO in apatites from tourmaline-rich rocks are higherthan those of the metasediments, granites and orthogneiss(Table 8). The F-contents of apatites are very similarin all rock types (av. 3�67–3�72 wt % F ). These apatite

Table 8: Average compositions of apatite

Tourmalinites

n ¼ 22

Metasediments

n ¼ 12

Granites

n ¼ 6

Orthogneisses

n ¼ 6

FeO 0.22 (0.13) 0.07 (0.07) 0.04 (0.03) 0.25 (0.06)

MnO 1.39 (1.27) 0.16 (0.10) 1.06 (0.46) 0.75 (0.15)

MgO 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.03 (0.01)

CaO 53.59 (1.21) 54.64 (0.68) 54.73 (0.70) 53.97 (0.42)

SrO 0.39 (0.21) 0.16 (0.12) 0.02 (0.02) 0.05 (0.03)

P2O5 42.37 (0.46) 42.66 (0.23) 42.41 (0.23) 42.53 (0.17)

F 3.71 (0.04) 3.69 (0.03) 3.72 (0.03) 3.72 (0.03)

Cl 0.01 (0.01) 0.00 (0.00) 0.00 (0.00) 0.01 (0.01)

H2O 0.02 (0.03) 0.04 (0.01) 0.03 (0.02) 0.02 (0.02)

O¼F,Cl 1.56 (0.02) 1.55 (0.01) 1.57 (0.01) 1.57 (0.01)

Total 100.15 (0.44) 99.87 (0.49) 100.45 (0.50) 99.75 (0.26)

P 6.034 (0.035) 6.060 (0.039) 6.014 (0.017) 6.054 (0.023)

Fe2þ 0.032 (0.018) 0.010 (0.009) 0.006 (0.004) 0.035 (0.008)

Mn 0.198 (0.182) 0.023 (0.014) 0.151 (0.066) 0.107 (0.022)

Mg 0.000 (0.000) 0.000 (0.000) 0.000 (0.000) 0.006 (0.003)

Ca 9.659 (0.196) 9.822 (0.098) 9.821 (0.101) 9.723 (0.072)

Sr 0.038 (0.021) 0.016 (0.012) 0.002 (0.002) 0.005 (0.002)

F 1.975 (0.033) 1.956 (0.016) 1.970 (0.019) 1.977 (0.019)

Cl 0.002 (0.001) 0.001 (0.001) 0.001 (0.001) 0.003 (0.002)

OH� 0.023 (0.033) 0.043 (0.016) 0.029 (0.020) 0.020 (0.017)

Structural formulae on the basis of 26 (O, OH, F, Cl).Numbers in parentheses are standard deviations (s).

1036


F-contents and the IV(F) parameter for coexisting biotitesin granitic rocks are consistent with peraluminous associ-ations, but they define lower temperatures than the sol-idus temperature (�650�C) of peraluminous silicic melts(Sallet, 2000). This may reflect a non-equilibrium relationbetween apatite and biotite, but, bearing in mind thepresence of tourmaline in the Martinamor granite, it ismore likely to be an effect of boron on the solidus tem-perature. In B-bearing systems, quartz and two alkalifeldspars may coexist with the melt below 600�C(Dingwell et al., 1996).

Rutile is also a minor component in the tourmaline-rich rocks. It mostly occurs as an intergranular,fine-grained (100–300 mm), yellowish-brown to opaquephase, typically containing tiny inclusions of tourmaline,quartz and/or mica. Its composition is characterized byrelatively high SnO2 (up to 2�80 wt %) and WO3 (up to4�70 wt %) in tourmalinites compared with rutile in themetasediments (Table 9), in which these elements arenegligible; manganiferous ilmenite (av. 5�06 wt %MnO) is the main Ti-bearing mineral. Nb and Ta arebelow the detection limit. Tiny grains of Sn- and W-freerutile included in tourmaline are also present in thetourmaline-rich rocks. The Sn- and W-free rutile isbelieved to be a relict detrital phase, whereas Sn- andW-bearing rutile probably records metasomatic andmetamorphic processes.

Plagioclase is the only feldspar in the tourmaline-richrocks, occurring in variable quantities as fine-grained,unzoned crystals, with an albite to oligoclase composition(Ab100–75). Plagioclase in metasedimentary rocks alsoshows a variable distribution and similar composition.

BORON ISOTOPES

Boron isotope analyses for representative tourmaline sep-arates from the major occurrences are given in Table 10.Overall, the data show a wide range of d11B values, from�15�60% to þ6�80%. In the Martinamor granite, fine-grained tourmaline has a lower value (�15�60%) thanthe tourmaline from the clots (�11�7%), both being inthe middle of the range of magmatic tourmalines (Barth,1993; Jiang & Palmer, 1998). A tourmaline from a dis-cordant pegmatite has a d11B value of �4�5%, which fallsat the heavier end of the range for pegmatitic tourmalines(Swihart & Moore, 1989; Jiang & Palmer, 1998). Tour-malines from premetamorphic tourmalinites have d11Bvalues that vary from �4�4 to þ2%, falling in the centralpart of the range for tourmalines from metapelites (�22to þ22%; Swihart & Moore, 1989). The highest d11Bvalues were found in tourmalines from syntectonic veins(þ2�60 and þ6�80%); tourmaline from a late-tectonicvein also has a relatively high d11B value of þ1�90%.

Boron isotope fractionation in tourmaline is mainlycontrolled by: (1) the composition of the boron source;(2) temperature; (3) fluid/rock ratios; (4) regionalmetamorphism (Palmer & Slack, 1989; Jiang, 2001).The main factor governing the behavior of boron iso-topes in tourmalines during metamorphism is consideredto be fluid activity (Palmer & Slack, 1989). In the Marti-namor area, there is clear evidence of fluid activity and,because boron isotope fractionation depends on temper-ature, it is logical that prograde metamorphism will giverise to isotopic differences between premetamorphic andrecrystallized tourmalines (Slack et al., 1993). Taking intoaccount that recrystallized tourmalines in the Marti-namor antiform (d11B ¼ �4�4% to þ2�0%) underwentmetamorphic conditions consistent with the staurolite

Table 10: Boron-isotope compositions of tourmalines

Sample type d11B (%)

Recrystallized tourmaline 2.00

Recrystallized tourmaline �1.80





Late tourmalinite �8.00

Syn-tectonic vein 2.60

Syn-tectonic vein 6.80

Late tectonic vein 1.90

Granite �15.60

Granite �11.70

Pegmatite �4.50

Table 9: Average compositions of rutile

Tourmalinites

n ¼ 68

Metasediments

n ¼ 11

TiO2 96.46 (1.83) 98.33 (1.75)

SnO2 1.41 (0.65) 0.00 (0.00)

WO3 0.74 (0.97) 0.01 (0.01)

SiO2 0.03 (0.09) 0.02 (0.05)

Al2O3 0.06 (0.11) 0.02 (0.05)

FeO 0.72 (0.47) 0.38 (0.45)

Total 99.43 (1.17) 98.77 (1.46)

Ti 0.985 (0.007) 0.998 (0.001)

Sn 0.008 (0.004) 0.000 (0.000)

W 0.003 (0.003) 0.000 (0.000)

Si 0.000 (0.001) 0.000 (0.001)

Al 0.001 (0.002) 0.000 (0.001)

Fe 0.004 (0.002) 0.002 (0.002)

Structural formulae on the basis of 2 O. Numbers inparentheses are standard deviations (s).

1037


zone (�550�C calculated from biotite–garnet geother-mometer), we can assume a fractionation of >3% basedon the experimental data of Palmer et al. (1992) for thefractionation of 11B/10B between tourmaline and fluid.The fluids released during metamorphism of tourmalin-ites account for the formation of many quartz–tourmaline veins in the study area. Because the heavyisotope, 11B, is preferentially partitioned into the fluidphase (Palmer & Slack, 1989; Palmer et al., 1992), it isexpected that tourmalines from veins will have higherd11B values than those of premetamorphic tourmalinites,in order to balance the fractionation effects of meta-morphic recrystallization. For example, two tourmalinesfrom syntectonic veins have d11B values of þ2�6 andþ6�8%, compared with the nearby tourmalinites inwhich tourmalines have d11B of �1�9 and þ2%, respect-ively. Hence, an approximate d11B range of 0–4% for thepremetamorphic tourmalinites can be inferred, whichfalls in the middle of the range for typical continentalcrust (Palmer & Slack, 1989). Reaction of tourmalinitesduring anatexis and migration of boron into a melt couldexplain the presence of tourmaline in the Martinamorgranite, which has lower d11B values (�15�6 and�11�7%) than the tourmalinite tourmalines (�4�4 toþ2�0%). Melting a tourmalinite and crystallizing tour-maline during cooling of the melt would produce similarisotopic effects to metamorphic recrystallization oftourmalinite (Slack et al., 1993). If minimal fractionationoccurs between melt and tourmaline (Smith & Yardley,1996), the isotopic composition of magmatic tourmalineshould approximate the boron isotopic signature of theparental granitic magma. Depletion of the granitic tour-malines in 11B and the isotopic difference between fine-and medium-grained tourmalines (�15�6 and �11�7%,respectively) suggest that 11B was incrementally removedby loss of fluids from the crystallizing magma. The effectof fluid loss on the boron isotopic composition can beevaluated by a Rayleigh process, using the expression ofTaylor & Sheppard (1986):

d11Bf � d11

Bi ¼ ð1000 þ d11Bf ÞðFa�1 � 1Þ

where d11Bi ¼ initial magmatic d11B value, d11Bf ¼ finalmagmatic d11B value, F ¼ fraction of boron remainingin the magma, and a ¼ fractionation factor. Assumingthat the isotopic shift between the initial magma (d11B ¼�11�7%) and the residual fraction of melt (d11B ¼�15�6%) is 4%, and assuming a fractionation factor of�3–3�5% at �550–600�C (Palmer et al., 1992), �70–75% of the boron in the original melt was removed intothe fluid phase. The late tourmalinite, in whichtourmaline has a lower d11B value (�8%) than the pre-metamorphic tourmaline, could represent an additionaltourmalinization event caused by boron-rich fluidsderived from the Martinamor granite. The isotopic

difference (�7%) between late tourmalinite and late-stage granitic tourmaline implies relatively high frac-tionation factors, involving separation of fluid from meltat relatively low temperatures (�350–400�C). Underthese conditions, the loss of boron would have been ofthe order of 50%. Nonetheless, repeated episodes ofphase separation may lead to large isotopic shifts in theresidual melt at relatively small fractionation factors(Smith & Yardley, 1996). We are aware that thesecalculations are a simplification; nevertheless, the boroncontents of the Martinamor granite (<0�47 wt % B2O3,Table 2) compared with the values required to attainsaturation of silicic melts in tourmaline (>2 wt % B2O3,Wolf & London, 1997) are consistent with significant lossof boron during magmatic evolution of this granite.

DISCUSSION

An important problem in the petrogenetic interpretationof the Martinamor tourmalinites is understanding therelationship of these rocks with the local granites andpegmatites. Deformation–crystallization relationships insome tourmaline-rich rocks indicate that tourmalinegrew syn- to post-tectonically with respect to D2, prob-ably linked to the evolution of the Martinamor graniticcomplex. However, strong evidence for an early, pre-Variscan development of other tourmalinites, which aretherefore unrelated to the granites and pegmatites, isprovided by field and petrographic observations. Theorigin of these older Cadomian tourmalinites, however,is not completely clear.

Inferences from mineral chemistry andboron isotopes

(1) Overall, the tourmaline compositions of the Marti-namor antiform belong to the alkali group and most fallwithin the psammopelitic domain (Fig. 9), suggesting animportant control by bulk-rock composition. Tourma-lines from tourmalinites that plot in the granitoid fieldmay reflect the chemistry and fluid/rock ratios of hydro-thermal systems. Compared with the tourmaline in TQand TPQ tourmalinites, tourmaline in TM tourmalinitesshows the highest contents of dravite and uvite þ feruvite,and the lowest contents of alkali-free tourmaline. Tour-maline in granites is characterized by low contents ofuvite þ feruvite with relatively high alkali-free tourmalineand olenite components (Table 3). No significant differ-ences in composition exist between tourmalines in early,Cadomian tourmalinites and those in late, Variscantourmalinites.

(2) Tourmaline in syntectonic veins is compositionallysimilar to those of late-tectonic veins and both are alsosimilar to those from TQ and TPQ tourmalinites (Figs 9and 10, Table 3). These veins, therefore, may have

1038


formed under closed-system conditions, assuming, in thiscase, that a closed system implies less than a meter scale(Cartwright et al., 1994; Oliver & Bons, 2001). An ancil-lary criterion in support of this is the fact that the vein-forming minerals correspond to those in the wall rocks. Incontrast, relatively large quartz–tourmaline veins and lesscommon feldspar-bearing veins are more consistent withopen-system conditions, probably related to the precip-itation of fluids of strongly magmatic affiliation.

(3) Compositional differences between recrystallizedtourmaline grains and tourmaline porphyroclasts suggestthat recrystallization took place by boundary migra-tion mechanisms [syndeformational recrystallization ofSt€uunitz (1998)]. Subgrains are observed in the micro-structures, but it appears that progressive subgrain rota-tion cannot explain the compositional differencesbetween porphyroclasts and recrystallized grains (St€uunitz,nitz, 1998).

(4) Tourmaline in tourmalinites shows different styles ofchemical zoning, even in the same sample, which mayinvolve steep compositional gradients in Al, Na, Mg, Fe,Ti and F (Figs 5 and 11). Oscillatory chemical zoning intourmalines from tourmalinites is probably a premeta-morphic feature that reflects mineral growth in opensystems. Taylor & Slack (1984) argued that multiple,oscillatory zoning in tourmaline involves a high fluid–rock environment and is generally uncommon in meta-morphic rocks. The early generation of tourmaline withoscillatory and sector zoning is very common in geologicsettings in which there are dynamically changing hydro-thermal fluid compositions (and, consequently, tourma-line compositions) (London et al., 1996; Slack, 1996). Theoscillatory zoning is a possible criterion to use in supportof extensive metasomatism (e.g. Yardley et al., 1991).Cellular textures, patchy zoning and embayed interiorzones in tourmaline (Fig. 5) are thought to reflect invas-ive, reactive fluids that replaced the tourmaline with alater generation of tourmaline. Intracrystal compositionalboundaries and other structural defects are relativelyhigh-energy sites that might have favoured dissolutionprocesses (Vance, 1965). Experimental studies indicatethat tourmaline is unstable under alkaline conditions(Frondel & Collette, 1957; Morgan & London, 1989)and von Goerne et al. (2001) demonstrated that theamount of Na in tourmaline depends on both temperat-ure and the Na-content of the coexisting aqueous fluid.At constant Na-concentration in the fluid, the Na-contentin the tourmaline should decrease as the temperatureincreases. Consequently, as Henry et al. (2002) claimedfor tourmaline partially replacing pre-existing tourmalinefrom a vein in the Eastern Alps, the relatively low-Natourmaline that forms part of some replacement zones,fractures and overgrowths (Figs 5 and 11) could be dueto the influx of reactive fluids of alkaline character dur-ing the metamorphism. Compositional fluctuations in

tourmalines from veins may reflect both cyclicity influid pressure related to fault valve action (Renard et al.,2000) and feedback processes associated with chemicaldisequilibrium (Ortoleva et al., 1987).

(5) The relatively high contents of Li, Be, Sn and W intourmaline separates from tourmalinites are consistentwith formation in a granite-related environment. Tour-malines in tourmalinites show two different REE patternsthat suggest differential fluid/rock ratios during metaso-matic processes; tourmalines with low REE abundancesand positive Eu anomalies are considered to reflectdeposition under relatively high fluid/rock conditionsin comparison with those that have high REE abund-ances (Fig. 7). These latter, however, might not berepresentative because of the influence of heavy mineralfractionations on REE patterns.

(6) Available boron isotope data are limited, but thed11B values give some insights into the origin and evolu-tion of the tourmaline-rich rocks. The observed vari-ations in d11B support an important recycling of boronin the study area, between the Cadomian and Variscancycles. Additional B-isotopic data are required in order toconstrain the boron source of the premetamorphictourmalinites, and to better understand the main stagesof tourmaline formation.

(7) Calculated values for log( f H2O/f HF) based on theF-content of biotite reflect equilibration with moderatelyF-rich fluids. In principle, the F-intercept values forbiotites from tourmalinites (�1�46) denote fluids richerin F than those from the granites and metasediments.However, because the halogen content of mica dependson temperature, it is possible that the difference in the logfugacity ratio accounts for different F-exchange temper-atures. Indeed, if the biotites from granites stoppedexchanging F at a higher temperature (e.g. 600�C) thanthe tourmalinite biotites, and assuming that they have notbeen reset by lower-temperature fluids, the log( f H2O/f HF) in both suites of biotites would have comparablevalues. In contrast, tourmalinites and metasedimentsunderwent similar P–T conditions, with a difference of�0�5 in the log( f H2O/f HF) ratio. This would indicatedifferences in the relative halogen acid activity during theformation and evolution of the tourmalinites. The F-enrichment, together with the higher Sn- and W-contentsin tourmalinites in comparison with the metasediments(Table 1), suggests that F, B, Sn and W are closely related.The F-contents of apatite coexisting with biotite in thegranitic rocks suggest that melt fractions may have per-sisted down to relatively low temperatures, probablycaused by abundant boron in the melt.

Bulk compositional constraints

In recent years, different theories have been postulatedto explain tourmalinite formation, from syngenetic to

1039


epigenetic models [see Slack (1996) for an overview]. Atthe Martinamor antiform, one can rule out a diagenetic–metamorphic origin on the basis of textural relationshipsand constraints from bulk-rock boron contents. Meta-morphism of clay-rich sediments with relatively highboron (�2000 ppm) would yield a maximum of 4 vol. %tourmaline, which represents an insignificant amountcompared with the tourmalinites in this area (up to90 vol. % or more). Likewise, an evaporitic source forthe boron seems improbable because of the lack ofknown evaporites or meta-evaporites in this region. Inthe Schist–Greywacke Complex, the occurrence ofbedded tourmalinites associated with scheelite-bearinglayers suggests a cogenetic relationship for B and W,which has been attributed to exhalative processes byArribas-Rosado (1986). Although the stratiform morpho-logy of the tourmalinite has been considered a character-istic feature of an exhalative origin (Slack, 1996), sucha configuration can equally result from selective replace-ment of favourable layers. In the Martinamor area, it isdifficult to support an exhalative model because there isno evidence for exhalites such as coticules (garnet-richquartzites), iron formations, bedded sulfides, etc.Premetamorphic formation of the tourmalinites byhydrothermal replacement of clastic sediments at orbelow the sediment–water interface may be a plausiblemechanism, based on the similar trends of some majorand trace elements and REE patterns for the tourmalin-ites and psammopelitic metasediments. Moreover, thefact that both types of rocks show similar Y/Ho ratios,close to the chondritic ratio (28), is consistent with thisinterpretation and indicates a coherent geochemicalbehavior during hydrothermal and metamorphic pro-cesses. Aqueous fluids and their precipitates displaynon-chondritic Y/Ho ratios (Bau, 1996). However, thehigh Li-, Sn- and Be-contents in some of the Martinamortourmaline-rich rocks suggest that tourmalinizing fluidshad a large igneous component.

Origin of the tourmalinites

Criteria in support of a contact metasomatic origin arenot individually definitive but, in combination, they offercompelling reasons to believe the tourmalinites affectedby Variscan deformation and metamorphism are theresult of metasomatic alteration of the metasedimentaryrocks by externally derived B-rich fluids of magmaticaffiliation. Besides boron, geochemical data suggest thatLi, Be, F, Sn and W were the main components intro-duced into the surrounding country rocks, which pro-vided a local source of the Al, Fe, Mg and Ca necessaryto precipitate tourmaline from boron-rich fluids. Thefelsic volcanic rocks (‘porphyroids’) that occur withinthe metasedimentary sequence are unlikely candidatesas a source for the boron because they lack tourmalineand tourmalinization effects on surrounding rocks.

A more probable candidate is the San Pelayo granitoid,taking into account that it: (1) is a pre-Variscan peralu-minous body, probably Cadomian; (2) crops out in thecore of the antiform; (3) contains tourmaline. The factthat the amount of tourmaline in this granitoid is smalldoes not mean that boron in the melt was necessarilyscarce, as the crystallization of tourmaline may be influ-enced by the formation of biotite. The B2O3-contentneeded to attain saturation of silicic melts in tourmalinevaries, depending on diverse factors. For example,whereas B�eenard et al. (1985) observed that tourmalinesaturation in silicic melts requires over 0�3 wt % B2O3,Wolf & London (1997) found that saturation of melt intourmaline, with or without ferromagnesian minerals,requires in excess of 2 wt % B2O3. Although a minimumamount of B is necessary for tourmaline crystallization, itappears that low Ti-contents in the melt stabilize tour-maline with respect to biotite (Nabelek et al., 1992; Wilkeet al., 2002). Tourmaline and biotite may coexist in someleucogranites, but tourmaline-bearing facies generally aredistinct from biotite-bearing facies (Nabelek et al., 1992).The TiO2-content of the Martinamor granitic rocks de-creases in the order: biotite-bearing SP orthogneiss,including accessory tourmaline (0�41–0�47 wt % TiO2);tourmaline-bearing granite with biotite subordinate(0�09–0�07 wt % TiO2); tourmaline-bearing granitewithout biotite (<0�06 wt % TiO2). This appears toindicate that the amount of TiO2 may have a significantinfluence on the tourmaline content in granites and, inturn, on the amount of B that can be expelled intocountry rocks.

Our proposed model for the origin and evolution of thetourmalinites in the Martinamor area can be summarizedas follows

(1) Formation of tourmalinites by boron metasomatismof psammopelitic rocks. The boron was probably derivedfrom magmatic fluids related to emplacement of theCadomian granitoid of San Pelayo, which escaped fromthe crystallizing melt to interact with surrounding coun-try rocks.

(2) Folding, mylonitization and recrystallization of thetourmalinites as a result of the Variscan regional meta-morphism and deformation. These processes, in additionto causing replacement and development of overgrowthson tourmaline, gave rise to multiple tourmaline � quartzveins, which represent important sinks of boron derivedfrom the Cadomian tourmalinites.

(3) Formation of boron-rich melts by partial melting ofhigh-grade metamorphic peraluminous rocks during theVariscan orogeny, the Cadomian tourmalinites beingthe main source of boron in the granitic magma. Theemplacement of Variscan granites and pegmatites led tonew tourmalinization processes. Early to late growth oftourmaline is believed to have occurred during crystal-lization of the Martinamor granite.

1040


(4) Fluids exsolved during the final crystallization oftourmaline-bearing melts and expelled into the wallrocks yielded a new stage of tourmalinization (Variscantourmalinites), as well as the formation of quartz þtourmaline veins. Albitization of psammitic schists andquartzites, and the development of albite veins wereconcurrent with these late tourmalinization processes.

This sequence of events suggests a multistage recyclingof boron in the Martinamor antiform from the Cadomianto Variscan orogenic cycles, which accounts for the pres-ence of tourmaline in a variety of metamorphic, hydro-thermal, granitic and pegmatitic rocks. Boron isotopiccompositions of tourmalines from these environmentsdepend on several factors (source region, metamorphicgrade, magma degassing, etc.) but, overall, the d11B val-ues indicate that redistribution of boron through time waslinked to recycling of crustal material. This model is inaccord with the idea that peraluminous leucogranites arethe only magmatic rocks that form entirely by crustalanatexis, without influx of mantle-derived material(Pati~nno Douce, 1999). In the Martinamor region, boronmay have been incorporated into peraluminous graniticmelts by partial melting of local B-rich metamorphicbasement with loss of B-bearing hydrothermal fluids.The recycling of pre-Paleozoic crust, including unusuallyboron-rich rocks, may have played an important role inthe generation of B-rich granitic melts in other areas ofthe Iberian massif. Granites that have elevated B-contents and contain magmatic tourmaline require anunusual B-enrichment in their source regions (Leemanand Sisson, 1996).

ACKNOWLEDGEMENTS

Support for this study was provided by the SpanishCICYT (project BTE 2002-01920). John Slack offeredconstructive comments and suggestions on a draft versionof this manuscript. The authors wish to thank BenitoAbalos from the Geodynamics Department of the BasqueCountry University for helpful comments on microtec-tonic aspects. Reviews of Darrell Henry, Martin Smith,an anonymous reviewer and the Executive Editor,Marjorie Wilson, greatly improved the manuscript.

SUPPLEMENTARY DATA

Microprobe data for tourmaline and coexisting mineralsare available at Journal of Petrology online (http://www.petrology.oupjournals.org).

REFERENCES

Ague, J. J. (1991). Evidence for major mass transfer and volume strain

during regional metamorphism of pelites. Geology 9, 855–858.

Argast, S. & Donnelly, T. W. (1987). The chemical discrimination

of clastic sedimentary components. Journal of Sedimentary Petrology 57,

813–823.

Arribas-Rosado, A. (1986). The significance of tourmaline in strata-

bound tungsten deposits in Spain. Tungsten deposits, Toulouse,

France. Abstract p. 33.

Ayers, J. C. & Watson, E. B. (1993). Rutile solubility and mobility in

supercritical aqueous fluids. Contributions to Mineralogy and Petrology

114, 321–330.

Barker, A. J. (1990). Introduction to Metamorphic Textures and Microstructures.

New York: Chapman and Hall, 162 pp.

Barth, S. (1993). Boron isotope variations in nature: a synthesis.

Geologische Rundschau 82, 640–651.

Bau, M. (1991). Rare-earth element mobility during hydrothermal and

metamorphic fluid–rock interaction and the significance of the

oxidation state of europium. Chemical Geology 93, 219–230.

Bau, M. (1996). Controls on the fractionation of isovalent trace

elements in magmatic and aqueous systems: evidence from Y/Ho,

Zr/Hf, and lanthanide tetrad effect. Contributions to Mineralogy and

Petrology 123, 323–333.

Bebout, G. E., Ryan, J. G. & Leeman, W. P. (1993). B–Be systematics

in subduction-related metamorphic rocks: characterization of

the subducted component. Geochimica et Cosmochimica Acta 57,

2227–2237.

Bell, T. H. & Rubenach, M. J. (1980). Crenulation cleavage

development: evidence for progressive, bulk inhomogeneous short-

ening from ‘millipede’ microstructures in the Robertson River

Metamorphics. Tectonophysics 68, T9–T15.

B�eenard, F., Moutou, P. & Pichavant, M. (1985). Phase relations of

tourmaline leucogranites and the significance of tourmaline in silicic

magmas. Journal of Geology 93, 271–291.

Bons, P. D. (2000). The formation of veins and their microstructures.

In: Jessell, M. W. & Urai, J. L. (eds) Stress, Strain and Structure: a Volume

in Honour of W.D. Means, Vol. 2. Journal of the Virtual Explorer 2, 12.

Caggianelli, A., Fiore, S., Mongelli, G. & Salvemini, A. (1992). REE

distribution in the clay fraction of pelites from the southern

Apennines, Italy. Chemical Geology 99, 253–263.

Cartwright, I., Power, W. L., Oliver, N. H. S., Valenta, R. K. &

McLatchie, G. S. (1994). Fluid migration and vein formation during

deformation and greenschist facies metamorphism at Ormiston

gorge, central Australia. Journal of Metamorphic Geology 12,

373–386.

Dıez Balda, M. A. (1986). El Complejo Esquisto-Grauv�aaquico, las Series

Paleozoicas y la estructura Hercınica al Sur de Salamanca. Universidad de

Salamanca, 162 pp.

Dıez Balda, M. A., Vegas, R. & Gonz�aalez Lodeiro, F. (1990). Central

Iberian Zone: structure. In: Dallmeyer, R. D. & Martınez Garcıa, E.

(eds) Pre-Mesozoic Geology of Iberia. Berlin: Springer-Verlag,

pp. 172–188.

Dıez Balda, M. A., Martınez Catal�aan, J. R. & Ayarza, P. (1995). Syn-

collisional extensional collapse parallel to the orogenic trend in a

domain of steep tectonics: the Salamanca Detachment Zone

(Central Iberian Zone, Spain). Journal of Structural Geology 17,

163–182.

Dingwell, D. B., Pichavant, M. & Holtz, F. (1996). Experimental

studies of boron in granitic melts. In: Grew, E. S. & Anovitz, L. M.

(eds) Boron: Mineralogy, Petrology and Geochemistry. Mineralogical Society of

America, Reviews in Mineralogy 33, 331–385.

Duong, P. K., Vache, R., Arribas, A. & Thadeu, D. (1980).

Yacimientos de tungsteno y metales asociados en Europa occidental.

Boletın Geol�oogico y Minero XCI–II, 391–416.

Dyar, M. D., Taylor, M. E., Lutz, T. M., Francis, C. A., Guidotti, C. V.

& Wise, M. (1998). Inclusive chemical characterization of

1041


http://www

tourmaline: Mossbauer study of Fe valence and site occupancy.

American Mineralogist 83, 848–864.

Flynn, R. T. & Burnham, C. W. (1978). An experimental determina-

tion of the rare-earth partition coefficients between a chloride-

containing vapor phase and silicate melts. Geochimica et Cosmochimica

Acta 42, 685–701.

Frondel, C. & Collette, R. L. (1957). Synthesis of tourmaline by

reaction of mineral grains with NaCl–H3BO3 solution, and its

implications in rock metamorphism. American Mineralogist 42,

754–758.

Frost, B. R., Barnes, C. G., Collins, W. J., Arculus, R. J., Ellis, D. J. &

Frost, C. D. (2001). A geochemical classification for granitic rocks.

Journal of Petrology 42, 2033–2048.

Galibert, F. (1984). G�eeochimie et g�eeochronologie du complexe

granitique de l’antiforme de Morille (Salamanca, Espagne). Unpub-

lished Rapport de stage de DEA, Lab. de G�eeochimie Isotopique,

Universit�ee de Montpellier.

Garcıa de Figuerola, L. C., Franco Gonz�aalez, P. & Castro Dorado, A.

(1983). Caracterısticas petrol�oogicas del complejo laminar peg-

matoide (Serie de Alamo) de las provincias de Salamanca y Avila.

Studia Geologica Salmanticensia XIX, 33–77.

Gier�ee, R. (1990). Hydrothermal mobility of Ti, Zr, and REE: examples

from the Bergell and Adamello contact aureoles (Italy). Terra Nova 2,

60–67.

Gonzalo, F. J., Saavedra, J., Garcıa S�aanchez, A., Pellitero, E., Arribas,

A. & Rodrıguez, S. (1975). Las rocas granıticas de la antiforma de

Martinamor (Salamanca). Actas II Reuni�oon Ibero-Americana de Geologıa

Econ�oomica 4, 227–246.

Govindaraju, K., Potts, P. J., Webb, P. C. & Watson, J. S. (1994).

Report on Whin Sill dolerite WS–S from England and Pitscurie

microgabbro PM–S from Scotland: assessment by one hundred and

four international laboratories. Geostandards Newsletter 18, 211–300.

H€aaussinger, H., Okrusch, M. & Scheepers, D. (1993). Geochemistry of

premetamorphic hydrothermal alteration of metasedimentary rocks

associated with the Gorob Massive Sulfide Prospect, Damara

Orogen, Namibia. Economic Geology 88, 72–90.

Hawthorne, F. C. & Henry, D. J. (1999). Classification of the minerals

of the tourmaline group. European Journal of Mineralogy 11, 201–215.

Henry, D. J. & Dutrow, B. L. (1996). Metamorphic tourmaline and its

petrologic applications. In Grew, E. S. & Anovitz, L. M. (eds) Boron:

Mineralogy, Petrology and Geochemistry. Mineralogical Society of America,

Reviews in Mineralogy 33, 503–557.

Henry, D. J. & Guidotti, C. V. (1985). Tourmaline as a petrogenetic

indicator mineral: an example from the staurolite-grade metapelites

of NW Maine. American Mineralogist 70, 1–15.

Henry, D. J. & Guidotti, C. V. (2002). Titanium in biotite from

metapelitic rocks: temperature effects, crystal-chemical controls, and

petrologic applications. American Mineralogist 87, 375–382.

Henry, D. J., Dutrow, B. L. & Selverstone, J. (2002). Compositional

asymmetry in replacement tourmaline: an example from the Tauern

Window, Eastern Alps. Geological Materials Research 4, 1–18.

Jiang, S.-Y. (2001). Boron isotope geochemistry of hydrothermal ore

deposits in China: a preliminary study. Physics and Chemistry of the

Earth (A) 26, 851–858.

Jiang, S.-Y. & Palmer, M. R. (1998). Boron isotope systematics of

tourmaline from granites and pegmatites: a synthesis. European

Journal of Mineralogy 10, 1253–1265.

Jiang, S.-Y., Yang, J.-H., Novak, M. & Selway, J. (2003). Chemical and

boron isotopic compositions of tourmaline from the Lavicky

leucogranite, Czech Republic. Geochemical Journal 37, 545–556.

Jongh de, W. K. (1973). X-ray fluorescence analysis applying

theoretical matriz corrections. Stainless steel. X-Ray Spectrometry 2,

151–158.

Jolliff, B. L., Papike, J. J. & Laul, J. C. (1987). Mineral recorders of

pegmatite internal evolution: REE contents of tourmaline from the

Bob Ingersoll pegmatite, South Dakota. Geochimica et Cosmochimica

Acta 51, 2225–2232.

Kerrick, D. M. (1990). Aluminium metasomatism. In Ribbe, P. H. (ed.)

The Al2SiO5 Polymorphs. Mineralogical Society of America, Reviews in

Mineralogy 22, 311–352.

Leeman, W. P. & Sisson, V. B. (1996). Geochemistry of boron and its

implications for crustal and mantle processes. In: Grew, E. S. &

Anovitz, L. M. (eds) Boron: Mineralogy, Petrology and Geochemistry.

Mineralogical Society of America, Reviews in Mineralogy 33, 645–707.

Linares, E., Pellitero, E. & Saavedra, J. (1987). Primeras edades radio-

m�eetricas en el �aarea estanno-wolframıfera de Morille–Martinamor

(Centro-Oeste de Espa~nna). Boletın Geol�oogico y Minero XCVIII–V,

640–646.

London, D., Morgan, G. B., VI, Babb, H. A. & Loomis, J. L. (1993).

Behavior and effects of phosphorus in the system Na2O–K2O–

Al2O3–SiO2–P2O5–H2O at 200 MPa (H2O). Contributions to Miner-

alogy and Petrology 113, 450–465.

London, D., Morgan, G. B., VI & Wolf, M. B. (1996). Boron in granitic

rocks and their contact aureoles. In: Grew, E. S. & Anovitz, L. M.



Manning, D. A. C. (1981). The effect of fluorine on liquidus phase

relationships in the system Qz–Ab–Or with excess water at 1 kbar.

Contributions to Mineralogy and Petrology 76, 206–215.

Martınez-Garcıa, E. & Nicolau, J. (1973). Los terrenos infraordovıcicos

de la antiforma de Martinamor (Salamanca). Boletın Geol�oogico y Minero

LXXXIV–VI, 407–418.

McDonough, W. F. & Sun, S.-S. (1995). The composition of the Earth.

Chemical Geology 120, 223–253.

McLennan, S. M. (1989). Rare earth elements in sedimentary rocks:

influence of provenance and sedimentary processes. In: Lipin, B. R.

& McKay, G. A. (eds) Geochemistry and Mineralogy of Rare Earth Elements.

Mineralogical Society of America, Reviews in Mineralogy 21, 169–200.

Medaris L. G., Jr, Fournelle, J. H. & Henry, D. J. (2003). Tourmaline-

bearing quartz veins in the Baraboo quartzite, Wisconsin: occur-

rence and significance of foitite and ‘oxy-foitite’. Canadian Mineralogist

41, 749–758.

Moine, B., Ramambazafy, A., Rakotondrazafy, M., Ravolomiandrin-

arivo, B., Cuney, M. & de Parseval, P. (1998). The role of fluor-rich

fluids in the formation of the thorianite and sapphire deposits from

SE Madagascar. Mineralogical Magazine 62, 999–1000.

Moran, A. E., Sisson, V. B. & Leeman, W. P. (1992). Boron depletion

during progressive metamorphism: implications for subduction

processes. Earth and Planetary Science Letters 111, 331–349.

Morgan, G. B., VI & London, D. (1989). Experimental reactions of

amphibolite with boron-bearing aqueous fluids at 200 MPa:

implications for tourmaline stability and partial melting in mafic

rocks. Contributions to Mineralogy and Petrology 102, 281–297.

Munoz, J. L. (1984). F–OH and Cl–OH exchange in micas with

applications to hydrothermal ore deposits. In: Bailey, S. W. (ed.)

Micas. Mineralogical Society of America, Reviews in Mineralogy 13, 469–493.

Nabelek, P., Russ-Nabelek, C. & Denison, J. R. (1992). The generation

and crystallization conditions of the Proterozoic Harney Peak

leucogranite, Black Hills, South Dakota, USA: petrologic and

geochemical constraints. Contributions to Mineralogy and Petrology 110,

173–191.

Neiva, A. M. R. (1974). Geochemistry of tourmaline (schorlite) from

granites, aplites and pegmatites from northern Portugal. Geochimica et

Cosmochimica Acta 38, 1307–1317.

Oliver, N. H. S. & Bons, P. D. (2001). Mechanisms of fluid flow and

fluid–rock interaction in fossil metamorphic hydrothermal systems

1042


inferred from vein–wall-rock patterns, geometry and microstructure.

Geofluids 1, 137–162.

Ortoleva, P., Merino, E., Moore, C. & Chadam, J. (1987). Geochem-

ical self-organisation I: reaction–transport feedbacks and modelling

approach. American Journal of Science 287, 979–1007.

Palmer, M. R. & Slack, J. F. (1989). Boron isotopic composition of

tourmaline from massive sulfide deposits and tourmalinites. Contri-

butions to Mineralogy and Petrology 103, 434–451.

Palmer, M. R., London, D., Morgan, G. B., VI & Babb, H. A. (1992).

Experimental determination of fractionation of 11B/10B between

tourmaline and aqueous vapour: a temperature and pressure-

dependent isotopic system. Chemical Geology 101, 123–129.

Passchier, C. W. & Speck, P. J. H. R. (1994). The kinematic

interpretation of obliquely-transected porphyroblasts: an example

from the Trois Seigneurs massif, France. Journal of Structural Geology

16, 971–984.

Pati~nno Douce, A. E. (1999). What do experiments tell us about the

relative contributions of crust and mantle to the origin of granitic

magmas? In Castro, A., Fern�aandez, C. & Vigneresse, J. L. (eds)

Understanding Granites. Integrating New and Classical Techniques. Geological

Society, London, Special Publication 168, 55–75.

Pesquera, A. & Velasco, F. (1997). Mineralogy, geochemistry and

geological significance of tourmaline-rich rocks from the Paleozoic

Cinco Villas massif (western Pyrenees, Spain). Contributions to

Mineralogy and Petrology 129, 53–74.

Pollard, P. J., Pichavant, M. & Charoy, B. (1987). Contrasting

evolution of fluorine- and boron-rich tin systems. Mineralium Deposita

22, 315–321.

Pouchou, J. L. & Pichoir, F. (1985). ‘PAP’ f(pz) procedure for improved

quantitative microanalysis. In: Armstrong, J. T. (ed.) Microbean

Analysis. San Francisco: San Francisco Press, p. 104.

Power, G. M. (1968). Chemical variation in tourmalines from south-

west England. Mineralogical Magazine 36, 1078–1089.

Ramsay, J. G. (1980). The crack-seal mechanism of rock deformation.

Nature 284, 135–139.

Reche, J. & Martinez, F. J. (1996). GPT: an Excel spreadsheet for

thermobarometric calculations in metapelitic rocks. Computers and

Geosciences 22, 775–784.

Renard, F., Gratier, J. P. & Jamtveit, B. (2000). Kinetics of crack-

sealing, intergranular pressure solution, and compaction around

active faults. Journal of Structural Geology 22, 1395–1407.

Saavedra, J., Pellitero, E., Garcıa S�aanchez, A. & Fern�aandez Turiel, J. L.

(1984). Rasgos petrogr�aaficos y geoquımicos de los granitoides y rocas

metam�oorficas del area de Morille—Martinamor (Salamanca, Centro

Oeste de Espa~nna). Estudios Geol�oogicos 40, 15–21.

Sallet, R. (2000). Fluorine as a tool in the petrogenesis of quartz-

bearing magmatic associations: application of an improved F–OH

biotite–apatite thermometer grid. Lithos 50, 241–253.

Shaw, D. M. (1996). Similarities and constrasts in lunar and terrestrial

boron geochemistry. In: Grew, E. S. & Anovitz, L. M. (eds) Boron:

Mineralogy, Petrology and Geochemistry. Mineralogical Society of America,

Reviews in Mineralogy 33, 745–769.

Slack, J. F. (1982). Tourmaline in the Appalachian–Caledonian

massive sulphide deposits and its exploration significance. Trans.

Inst. Min. Metall. 91 (B), 81–89.

Slack, J. F. (1996). Tourmaline associations with hydrothermal ore

deposits. In: Grew, E. S. & Anovitz, L. M. (eds) Boron: Mineralogy,

Petrology and Geochemistry. Mineralogical Society of America, Reviews in


Slack, J. F., Palmer, M. R., Stevens, B. P. J. & Barnes, R. G. (1993).

Origin and significance of tourmaline-rich rocks in the Broken Hill

district, Australia. Economic Geology 88, 505–541.

Smith, M. P. & Yardley, W. D. (1996). The boron isotopic composition

of tourmaline as a guide to fluid processes in the southwestern

England orefield: an ion microprobe study. Geochimica et Cosmochimica

Acta 60, 1415–1427.

Spiess, R., Peruzzo, L., Prior, D. J. & Wheeler, J. (2001). Development

of garnet porphyroblasts by multiple nucleation, coalescence and

boundary misorientation-driven rotations. Journal of Metamorphic

Geology 19, 269–290.

St€uunitz, H. (1998). Syndeformational recrystallization: dynamic or

compositionally induced? Contributions to Mineralogy and Petrology 131,

219–236.

Swihart, G. H. & Moore, P. B. (1989). A reconnaissance of the boron

isotopic composition of tourmaline. Geochimica et Cosmochimica Acta

53, 911–916.

Taylor, B. E. & Slack, J. F. (1984). Tourmalines from Appalachian–

Caledonian massive sulfide deposits: textural, chemical, and isotopic

relationships. Economic Geology 79, 1703–1726.

Taylor, H. P., Jr & Sheppard, S. M. F. (1986). Igneous rocks: I.

Processes of isotope fractionation and isotope systematics. In: Valley,

J. W., Taylor, H. P., Jr & O’Neil, J. R. (eds) Stable Isotopes in High

Temperature Geological Processes. Mineralogical Society of America, Reviews in


Torres-Ruiz, J., Pesquera, A., Gil-Crespo, P. P. & Velilla, N. (2003).

Origin and petrogenetic implications of tourmaline-rich rocks in the

Sierra Nevada (Betic Cordillera, southeastern Spain). Chemical Geology

197, 55–86.

Tracy, R. J., Rye, D. M., Hewitt, D. A. & Schiffries, C. M. (1983).

Petrologic and stable-isotopic studies of fluid–rock interactions,

south-central Connecticut: I. The role of infiltration in producing

reaction assemblages in impure marbles. American Journal of Science

283-A, 589–616.

Vance, J. A. (1965). Zoning in igneous plagioclase: patchy zoning.

Journal of Geology 73, 636–651.

Volfinger, M. & Robert, J.-L. (1980). Structural control of the

distribution of trace elements between silicates and hydrothermal

solutions. Geochimica et Cosmochimica Acta 44, 1455–1461.

Von Goerne, G., Franz, G. & Robert, J.-L. (1999). Upper thermal

stability of tourmaline þ quartz in the system MgO–Al2O3–SiO2–

B2O3–H2O and Na2O–MgO–Al2O3–SiO2–B2O3–H2O–HCl in

hydrothermal solutions and siliceous melts. Canadian Mineralogist 37,

1025–1039.

Von Goerne, G., Franz, G. & Heinrich, W. (2001). Synthesis of

tourmaline solid solutions in the system Na2O–MgO–Al2O3–SiO2–

B2O3–H2O–HCl and the distribution of Na between tourmaline and

fluid at 300 to 700�C and 200 MPa. Contributions to Mineralogy and

Petrology 141, 160–173.

Vorbach, A. (1989). Experimental examinations on the stability of

synthetic tourmaline in temperatures from 250 to 750�C and

pressures up to 4 kb. Neues Jahrbuch f€uur Mineralogie, Abhandlungen 161,

69–83.

Walther, J. V. (1997). Experimental determination and interpreta-

tion of the solubility of corundum in H2O between 350 and 600�C

from 0�5 to 2�2 kbar. Geochimica et Cosmochimica Acta 61, 4955–4964.

Weisbrod, A., Polak, C. & Roy, D. (1986). Experimental study of

tourmaline solubility in the system Na–Mg–Al–Si–B–OH: applica-

tions to the boron content of natural hydrothermal fluids and

tourmalinization processes. International Symposium on Experimental

Mineralogy, Abstract 1, 140–141.

Werding, G. & Schreyer, W. (1996). Experimental studies on

borosilicates and selected borates. In: Grew, E. S. & Anovitz, L. M.



1043


Wilke, M., Nabelek, P. I. & Glascock, M. D. (2002). B and Li in

Proterozoic metapelites from the Black Hills, U.S.A.: implications

for the origin of leucocratic magmas. American Mineralogist 87,

491–500.

Wolf, M. B. & London, D. (1997). Boron in granitic magmas: stability

of tourmaline in equilibrium with biotite and cordierite. Contributions

to Mineralogy and Petrology 130, 12–30.

Yardley, B. W. D. & Botrell, S. H. (1992). Silica mobility and fluid

movement during metamorphism of the Connemara schists, Ireland.

Journal of Metamorphic Geology 10, 453–464.

Yardley, B. W. D., Rochelle, C. A., Barnicoat, A. C. & Lloyd, G. E.

(1991). Oscillatory zoning in metamorphic minerals: an

indicator of infiltration metasomatism. Mineralogical Magazine 55,

357–365.

1044


Documents

Petrographic, Chemical and B-Isotopic Insights into …hera.ugr.es/doi/15771210.pdfPetrographic, Chemical and B-Isotopic Insights into the Origin of Tourmaline-Rich Rocks and Boron