zAmerican Mineralogist, Volume 82, pages 1019-1037, 1997
Nomenclatureof amphiboles:Report of the Subcommitteeon Amphibolesof the International Mineralogical Association,Commission on New Minerals and Mineral Names BnnN.q.nnE. Lrlrntt (Crumnanr),Ar-aN R. Woor-r-nv(Sncnnrlnv)r2 Crunr-BsE.S. Anrs,xf Wrr-r,rlrrD. BmcH,x$M. Cnanr,BsGrr-nrnrrll Jonr, D. Gnrcnrs# FnlNr C. H.q,wrHoRNEr4Arrnl Karort HeNaN J. Krscnru Vr,.a.rrnrrnG. KnrvovrcHEvrTKnns LrNrHourrs Jo L,lrnure Josnpn A. MlNnaRrNor*'xx War-rnn V. MannscHrr0ERNEST H. NrcrElrttx Nrcnor-as M.S. Rocrrft JonN C. ScnurmcHERrrrDlvru C. Svrrrnr$$Nrcr C.N. SrnpHENSoNrr2 LucrlNo Uuc.lnnrurllll Enrc J.W. WnrrrAKERrl3.lNoGuo Youznrl{
'Departmentof Geology and Applied Geology, University of Glasgow, Glasgow G12 8QQ, U K 'zDepartmentof Mineralogy, Natural History Museum, Cromwell Road, London SW7 5BD, U.K. rMineral SciencesDivision, CanadianMuseum of Nature, PO. Box 3443, Station D, Ottawa, Ontario K1P 6P4, Canada aDepaftmentof Earth Sciences,University of Manitoba, Winnipeg, Manitoba R3T 2N2, Canada 5Departmentof Geology, Natural ScienceMuseum, 2-23-7 Hyakanin-cho,Shinjuka, Tokyo 160, Japan 6Departmentof Geology and Mineralogy, Ben Gurion University of the Negev, Beer Sheva 84105, PO. Box 653, Israel ?Facultyof Geology, St PetersburgUniversity, University Em.719, 199034, St. Petersburg,Russra EDepartmentof Ore Geology, Petrology and Mineralogy, Institute of Earth Sciences,Free University, Boelelaan 1085, 1081 HV Amsterdam, The Netherlands eDepartmentof Earth Sciences,College of Engineering and Physical Sciences,University of New Hampshire, Durham,New Hampshire03824, U.S.A. 'olnstitutefor Mineralogy, WestfalischeWilhelms, University of Mtinster, Correnstrasse24, D4400 Miinster, Germany 'llnstitut fiir Mineralogie-Petrologie-Geochemieder Albert-Ludwigs Universitiit Freiburg, Albertstrasse23b, D-79104 Freiburg i. Br., Germany r2Departmentof Geology and Geophysics,University of New England, Armidale, New South Wales 2351, Australia "60 Exeter Road, Kidlington, Oxford OX5 zDZ, U K raCentralLaboratory, Bureau of Geology and Mineral Resourcesof Hunnan Province, Dashiba, Kunming, China
Ansrnncr The International Mineralogical Association's approved amphibole nomenclature has been revised to simplify it, make it more consistentwith divisions generally at 5OVo,define prefixes and modifiers more precisely, and include new amphibole speciesdiscovered and named since 1978, when the previous scheme was approved. The same reference axes form the basis of the new schemeand most namesare little changed,but compound species nameslike tremolitic hornblende (now magnesiohornblende)are abolished,as are crossite (now glaucophaneor ferroglaucophaneor magnesioriebeckiteor riebeckite), tirodite (now manganocummingtonite),and dannemorite (now manganogrunerite).The 50% rule has been broken only to retain tremolite and actinolite as in the 1978 scheme;the sodic-calcic amphibole range has therefore been expanded.Alkali amphiboles are now sodic amphi- boles. The use of hyphens is defined. New amphibole names approved since 1978 include nybriite, leakeite, kornite, ungarettiite, sadanagaite,and cannilloite. All abandonednames are listed. The formulae and source of the amphibole end-membernames are listed and proceduresoutlined to calculate Fe3* and Fe2* where not determined by analysis.
* Indicates a non-voting official of the CNMMN. ** Canada;retired December 1994. f E-mail address:[email protected] tt Australia. f The Netherlands:retired December 1994. *f Australia; died February 1992. $ Australia;from January1995. gg France;resigned 1994 ll U S.A.; resigned1994. llllItaly; resignedApril 1993. # Canada;non-voting official since January 1995.
0003-004x/97l09 10-1 0 r 9$05.00 1019 1020 LEAKE ET AL.: AMPHIBOLE NOMENCLATURE
InrnonucrroN names, such as crossite. End-member formulae defined generally This report was produced in responseto a motion at and approvedin IMA 78 are retained,although the IMA 1986 meeting in Stanford, California, asking the the ranges to which they apply have commonly been Commission on New Minerals and Mineral Names changed.Information on the etymology, the type locality, parameters pro- (CNMMN) to produce a more simplified amphibole no- and the unit-cell of thirty end-membersis menclature than that currently approved, which dates vided in Appendix 1. from 1978. The 1978 nomenclature(IMA 78) took over The principal reference axes of IMA 78, namely Si, 13 years to formulate; a quicker responsewas attempted Nau, and (Na + K)" (see below), are retained, but the this time. primary divisions between the calcic, sodic-calcic, and To ensurea fresh look at the nomenclaturescheme the alkali (renamedsodic) amphiboleshave been adjustedto ( = Chairman of the Amphibole Subcommittee,B.E. Leake, divisions at Na" 0.50 and Na" 1.50,instead of Nau > with the agreementof the CNMMN officials, completely < 0.61 and Nau 1.34. (Here, and elsewherein this reconstitutedthe committee so that (1) representationwas report, concentrationsare expressedin atoms per formula more international; (2) more than 80Vo of the voting unit of the standard formula of an amphibole given be- membersof the committee were not membersof the com- low.) Previously, the amphibole "box" was divided into mittee that produced the 1978 report; in addition, none three equal volumes with respectto Na". The new scheme of the CNMMN officials was on the 1978 committee; (3) enlargesthe sodic-calcic amphibolesat the expenseofthe three memberswere retained from the 1978 committee to calcic and sodic amphiboles(Fig. 1) to make the divisions ensure that there was some continuity and collective at 50%opositions. memory of the main problems that had been dealt with As with the 1978 scheme,the problem of what to do previously; (4) representationincluded the principal pro- with analysesin which only the total Fe is known (and poser to the CNMMN of an improved schemeof nomen- not its division into FeO and FerO.) has been left to in- clature; and (5) representation was sought across the dividual judgement, although a recommendedprocedure various fields concerned with amphibole nomenclature, is given. This means that again an analysis may yield from crystal chemists, metamorphic and igneous petrol- different names depending on the procedure used to es- ogists to computer experts and ordinary broad-basedpe- timate Fe3* and Fe2*. It clearly would be advantageous, trologists. There were 18 voting memberswhen the major for purposes of naming an amphibole, if the recommend- framework of the revised schemewas approved. ed procedure were followed even if other procedures The committee circulated over 1000 pages over nine were used for other purposes. years and consideredin detail all proposals made to it. General works dealing with the amphiboles include Views were expressedthat, becausethe amphibole system Deer et al. (1963,1997),Ernst (1968),Chukhrov (1981), is so complicated, adequate representation cannot be Veblen (1981), Veblen and Ribbe (1982), and Anthony et made with two- and three-dimensionaldiagrams, whereas al. (1995), from which adequategeneral background sum- four variables can representthe system adequately.How- mariescan be obtained. ever, the committee, by a very large majority, wanted to retain conventional nomenclaturediagrams becausethey GnNBn-q.LcLASsrFrcATroN oF THE AMpHIBoLES are easier for most scientiststo use. The committee con- As with the IMA 78 scheme,the proposed nomencla- sidereda range of different schemesof nomenclature,but ture is based on chemistry and crystal symmetry; where none was judged overall to be sufficiently justify better to it is necessaryto distinguish different polytypes or poly- abandoning the main basis of IMA 78, which has been morphs, this may be done by adding the space group widely acceptedand is capable of simplification to pro- symbol as suffix. Anthophyllite with Pnmn symmetry (as vide an improved scheme. It must be remembered that distinct from the more usual Pnma symmetry) may be over 95Vo of all amphibole analyses are currently ob- prefixed proto. tained by electron microprobe, with no structural infor- The classification is based on the chemical contents of mation, no knowledge of the oxidation statesof Fe, Ti, the standard amphibole formula AB2vIC.IvT8Orr(OH)r.It Mn, the HrO content, or how the site populations are is to be noted, however, that possessionof this formula derived. What follows is a schemeof nomenclature, not does not define an amphibole. An amphibole must have one to determine at which position the ions really are a structurebased on a double silicate chain: A biopyribole located. consisting of equal numbersof pyroxene chains and triple The proposedscheme involves reducing the number of chains would have this formula but would not be an subdivisions,especially in the calcic amphiboles,making amphibole. the divisions generally follow the 50Vo rule (whereas The components of the formula conventionally de- IMA 78 usesdivisions at 90, 70,66,50, 33, 30, and lOVo) scribed as A, B, C, T and "OH" correspondto the fol- and making the use of adjectival modifiers (additional to lowing crystallographicsites: prefixes that are part of the basic names) optional. The new schemehas over 20 fewer names than IMA 78 and A one site per formula unit; involves the abolition of only a few commonlv used B two M4 sites per formula unit; LEAKE ET AL: AMPHIBOLE NOMENCLATURE t021
NaNa2(LaM)Si8022(OH)2 NaNa2(kM2)Si7Al022(OH)2 Eckermannite Nyboite NaNa2(LMa)S'5A13O22(OH)2 1.00 (Na+K)o l\4agnesio-arfvedsonite Fetric-nyboite NaNa2(L2M3)Si6A|2O22(OH)2 (Na+K)o 8.00si 1 00 2 00 NaB 500si
3o22(oH)2 0 00 (!"!a+K)o .00 Na, 1.50 NaB ) 500si
| 00 Na-
Na(NaCa)L2l,43SisAl 30 22(OH)2
0 50 NaB
500si 1.00(Na+K)o 0.00Na^ NaCa2(LsM2)Si5AI3022(OH)2 Magnesiosadanagaite
5.00 si 0.00 (Na+K)o A trca2lssiso22(oH)2 trCa2(LaN.4)Si7AlO22(OH)2 trCa2(L3Nr2)Si6Al2022(oH)2 0.00 NaB '{-\r L4agnesiohornblende Tschermakite Tremolite trCa2(L2N.43)SisAl3O22(OH)2 @ (U ,$g z \ M = Alvr,Fq3* | Mg endmembers are named L = Mg,Fe2*, Mn AIIV OH = 911,O, F, Cl O hypotheticalend members per cations 24 O, OH, F, Cl ! = emptyA-site
Frcunn 1. General classificationof the amphiboles,excluding the Mg-Fe-Mn-Li amphiboles
C a composite of five sites made up of two Ml, Throughout this report, superscriptedarabic numerals two M2, and one M3 sites per formula unit; refer to ionic charge (oxidation state), e.g., Fe2*, super- T eight sites, in two sets of four, which need not scripted roman numerals to coordination number, e.9., be distinguished in this document; vIAl, and subscriptednumerals to numbers of atoms, e.g., OH two sites per formula unit. Ca.. of these facts, it is recommendedthat The ions considerednormally to occupy thesesites are To take account in the following categories: the standardamphibole formula be calculated as follows, though it must be clearly appreciatedthat this is an arith- n (empty site) and K at A only metic convention that assignsions to convenient and rea- Na atAorB sonable site occupancies. These cannot be confirmed Ca at B only without direct structural evidence. L-type ions: Mg, Fe2*,Mn2*, at C or B Li and rarer ions of similar (l) If HrO and halogen contentsare well established,the size such as Zn, Ni, Co formula should be calculated to 24(O,OH,ECI). Mtype ions: A1 at C or T (2) If the HrO plus halogen content is uncertain, the for- Fe3* and mofe rarelv Mn3+.Crr+ mula should be calculated on the basis of 23(O) with at C only 2(OH,ECI) assumed,unless this leads to an impos- High-valency ions: Tio* atCorT sibility of satisfying any of the following criteria, in Zl"* at C only which case an appropriate change in the assumed Si at T only number of (OH + F + CD should be made. Anions, OH, Cl, O at "OH" E (3) Sum T to 8.00 using Si, then Al, then Ti. For the sake M-type ions normally occupy M2 sites and so are nor- of simplicity of nomenclatwe, Fe3* is not allocated mally limited to two of the five C sites. Exceptions may to T. The normal maximum substitution for Si is 2, occur to the above "normal" behavior but are ignored for but this can be exceeded. the presentpurposes of nomenclature. (4) Sum C to 5.00 using excessAl and Ti from (3), and 1022 LEAKE ET AL.: AMPHIBOLE NOMENCLATURE
then successivelyZr, Cr3*, Fe3*, Mn3*, Mg, Fe2*, Mg-Fe-Mn-Liamphiboles Mn2*, any other L,"-type ions, and then Li. Diagram Parameters: (Ca + NaB) < 1 oo; (M9, Fe2*, Mn, Li)B > 1 00; LiB <'l 00 (5) Sum B to 2.00 using excessMg, Fe2*,Mn2*, and Li from (4), then Ca, then Na. Orthorhombic Monoclinic (6) ExcessNa from (5) is assignedto A, then all K. Total A should be between 0 and 1.00. The most cofllmon uncertainty results from lack of I analysesfor HrO, Fe3*, and Fe2*. The procedure adopted + ^- o uc to divide the Fe into Fe3* and Fe2* can influence the re- sulting name, especially if a composition is near Mg/(Mg + Fe2*) : 0.50 0r Fer*/(Fe3*+ vrAl) : 0.50, i.e., the same bulk composition may give rise to two or more names depending on the allocation of the Fe. The com- 80 70 mittee was almost unanimous in not wanting to specify Si in formula Si in formula one compulsory procedure for allocating Fe3* and Fe2*, but in recommendingthat a common procedure be used Diagram Parametes: (Ca + NaB) < 1 00; ([/9, Fez+, l!{n, Li)B > 1 00; LiB > 1 00 for purposes of naming the amphibole. Rock and Leake Orthorhombic l\4onoclinic (1984) showed that, on the basis of processingover 500 1.0 amphibole analyses,the IMA-favored procedure of ad- justing the sum (Si + Al * Cr 4 Ti + Fe + Mg + Mn) to 13, by varying the Fe3* and Fe2* appropriately, gave L Fe3* and Fe2* values reasonably close to the true deter- + ^- mined values in 8O7oof the compositions studied, ex- cluding those of kaersutite, for the calcic, sodic-calcic, and sodic amphiboles.If this sum is adjusted to include Li and Zr, i.e.,(Si + Al * Cr + Ti + Zr + Li + Fe + 00 Mg + Mn) : 13, and if for the magnesium-iron-man- 80 70 ganese-lithiumamphiboles the sum of (Si + Al + Cr + Ti + Zr + Li + Fe + Mg + Mn * Ca) : 15 isused, Frcunp2. Classificationof theMe-Fe-Mn-Li amphiboles. then only the Ti > 0.50 amphiboles need special treat- ment, although it is recognized that Mn-rich amphiboles pose problems with the variable valence stateof both the amphiboles.The new name is more precise,since Na Fe and Mn and that, as shown by Hawthorne (1983, p. is the critical element. not anv other alkali element 183-185), both in theory and practice, any calculation of such as K or Li. Fe3* and Fe2* values is subject to considerableuncertain- Within each of these groups a composition can then be ty. A full discussionof the problem and a recommended named by reference to the appropriate two-dimensional procedure,both by J.C. Schumacher,are given in Appen- diagram (Figs. 2-5). These are subdivided with respect dix 2. Some analyseshave given HrO* contentsthat lead to Si and Mg/(Mg + Fe2*) or Mg/(Mg + Mn'z*) with to more than (OH), in the formula, but the structurecon- prefixes to indicate major substitutionsand optional mod- tains only two sites for independentOH ions, and the ifiers to specify less important substitutions. structural role of the extra H ions is uncertain. Within the groups, the amphiboles are divided into in- The amphiboles are classified primarily into four dividually named speciesdistinguished from one another groups dependingon the occupancyof the B sites.These on the basis of the heterovalentsubstitutions: Si : r"Al, four principal groups of amphibole are slightly redefined n : (Na,K)",Cau : Nau,Li: L'*, M. : L'z.*,(Ti,Zr) as compared with IMA 78: : L", O : (OH,E Cl). These substititions necessarily (1) Where (Ca + Na)u < 1.00 and the sum of L-type occur in pairs or multiplets to maintain neutrality. The ions (Mg,Fe,Mn,Li)" is = 1.00, then the amphibole speciesdefined on this basis are shown in Figure 1 and is a member of the magnesium-iron-manganese-lith- along the horizontal axes of Figures 2-5. Different spe- tum group. cies defined in this way correspondto different distribu- (2) Where (Ca + Na)" = 1.00 and Nau < 0.50, then the tions of charge over the A, B, C, I and "OH" sites. amphibole is a member of the calcic group. Usually, Discovery of amphiboles with new or quantitatively ex- but not in every case,Ca" > 1.50. tendeddistributions of chargeover thesesites would mer- (3) Where (Ca + Na)" = 1.00 and Nau is in the range it the introductionof new speciesnarnes. 0.50-1.50, then the amphibole is a member of the Within the species there occur homovalent substitu- sodic-calcic group. tions, most commonly Mg : Fe'*, vrAl = Fet* and OH (4) Where Nau > 1.50, then the amphibole is a member : E The end-membersof theseranges of substitutionare of the sodic group, previously referred to as alkali distinguished by the use of prefixes, one or other end- LEAKE ET AL: AMPHIBOLE NOMENCLATURE 1023
member usually having a traditional name without a pre- Trele 1. Prefixesin additionto those in the fiqures fix. These substitutionsusually correspond to independent Meaning Applicableto binary systemsX-Y: The name of the X end-memberap- vrAl plies over the range 1.00 > X/(X + Y) > 0.50, and the Alumino > 1 00 Calcicand sodic-calciconly Chloro cl>100 All groups name of the Y end-memberto 1.00 >Y/(X + Y) > 0.50. Chromio Cr>100 All groups For the boundariesof substitution ranges in ternary sys- Ferri Fe3t > 1 00 All groupsexcept sodic (1992). Fluoro F>100 All groups tems, see Nickel Mangano Mn'z+:100-299 All groupsexcept kozulite The discovery of amphiboles with new or exotic hom- and ungarettiite ovalent substitutionsnever requires a new speciesname. Permangano Mn,* - 30G-499 All groupsexcept kozulite Mangani Mn3*> 1 00 All groupsexcept kornite They can always be named by use of an appropriate pre- and ungarettiite fix. In future one root or trivial name ONLY should be Potassic K > 0.50 All groups approved each charge arrangement in each Sodic Na > 050 Mg-Fe-Mn-Lionly for amphi- Titano Ti > 050 All groups except kaersutite bole group, and all species defined by homovalent sub- Zinco Zn > 1.00 All groups stitutions should be designated by the relevant prefix. /Vote:The prefixes in the figures are ferro (Fe'?-> Mg) and magnesio New species defined by heterovalent substitutions [in- (Fe'z-< Mg) and in Figure 5a only ferric-nyboitewith vrAl< Fe3* (not cluding major replacementof (OH, E CU by O, and ma- ferricnyboite,which is not clear). jor entry of high-charge(>3+) cations into A, B, or Cl result in new root or trivial names. The principal referenceaxes chosen for the calcic, sod- to alumino-ferrohornblende,chloro-ferro-actinolite, and ic-calcic, and sodic amphibolesare as in IMA 78, namely fluoro-ferri-cannilloite. Most (>907o) nameswill lack any Na", (Na + K)o, and Si, as shown in Figure l, but the hyphens and less than 5Vowill have more than one prefix. subdivison into the sodic-calcic group is now at Na" : In general, excluding juxtaposed vowels, the prefixes 0.50 (insteadof Na" : 0.61) and Na" : 1.50 (insteadof (Table 1), which have o, i, or ic endings, are either at- Nau : 1.34r. This increasesthe volume, and thereforethe tached directly to the root name (without a spaceor hy- number of compositions,assigned to the sodic-calcic am- phen) or to a following prefix with a hyphen. All these phiboles at the expenseof the calcic and sodic amphibole charactersdistinguish them from modifiers. groups, but is a logical consequenceof applying the 5OVa All modifiers (Table 2) have ian or oan endings to in- rule for all divisions rather than dividing the Nau, (Na + dicate moderate substitutions as listed by Nickel and K)^ and Si box into equal volumes, as in IMA 78. The Mandarino (1987). Modifiers are not accompaniedby a cornmittee consideredat length various proposalsfor the hyphen and are invariably followed by a spaceand then use of axes other than the three chosen, including four the remainder of the name.The excludedapplications fol- components,but eventually agreed,by a significant ma- low from the fact that thesegroups will usually have sub- jority, that the IMA 78 axes be retained despite their in- stantial contents of these elements as part of the param- ability to representR2* and R3* (i.e., usually L- and M- eters that define them. The use of modifiers is optional type ions) separatelyin the C group. The importance of and strictly qualitative (i.e., they can be used in other the difference between R2* and R3* in the C group has, sensesthan in Table 2, but use as in Table 2 is strongly however, been recognized rather more formally than pre- recommended). viously by the way in which the abundanceof Fe3*,Al3*, Cr3*, or Mn3* has been defined with prefixes, not modi- The naming of amphibolesin thin section and hand fiers, when they occupy 5OVoor more of the normal max- specimen imum of 2Rl* as shown in Table l. For amphiboles of which the general nature only is Following Nickel and Mandarino (1987), prefixes are known, for instance from optical properties without a an essential part of a mineral name (e.g., fenoglauco- chemical analysis, it is not generally possible to allocate phane and ferro-actinolite), whereas modifiers indicate a a precise name. The nearest assigned amphibole name compositional variant and may be omitted (e.g., potassian should then be made into an adjective, followed by the pargasite).Modifiers generally representsubsidiary sub- word amphibole, e.g., anthophyllitic amphibole, tremoli- stitutions whereasprefixes denote major substitutions.To tic amphibole, pargasitic amphibole, glaucophanic am- reduce the number of hyphens used, a single prefix is phibole, and richteritic amphibole. The familiar word generally joined directly to the root name without a hy- hornblende can still be used where appropriatefor calcic phen (e.g., ferrohornblende) unless two vowels would amphiboles in both hand specimen and thin section, be- then adjoin (e.g., ferro-actinolite) or "an unhyphenated cause hornblende is never used without a ferro or mag- name is awkward and a hyphen assistsin decipheringthe nesio prefix in the precise classification, such that con- name" (Nickel and Mandarino 1987) e.g., ferric-nybdite. fusion should not arisebetween colloquial use and precise For all amphibole names involving multiple prefixes, a use. hyphen shall be inserted between the prefixes, but not As in IMA 78, asbestiform amphiboles should be between the last prefix and the root name, unless two named according to their precise mineral name, as listed vowels would be juxtaposed or the name would be dif- in this report, followed by the suffix-asbestos:e.9., an- ficult to decipher or awkward. This convention gives rise thophyllite-asbestos,tremolite-asbestos. Where the nature ro24 LEAKE ET AL.: AMPHIBOLE NOMENCLATURE
Tnele 2. Modifiersand their suggestedranges
Modifier Meaning* Applicableto Barian Ba> 010 All groups Borian B > 0.10 All groups Calcian Ca>050 Mg-Fe-Mn-Ligroup Chlorian 0.25
of the mineral is uncertain or unknown, asbestosalone or Si < 7.00, this being the distinction from anthophyllite. amphibole-asbestosmay be appropriate. If the approxi- Li < 1.00. mate nature of the mineral only is known, the above rec- End-members: ommendationsshould be followed, but with the word am- Gedrite EMg.Al,Si.Al,O,,(OH), phibole replacedby asbestos,e.9., anthophyllitic asbestos, Fenogedrite nFel.Al.Si"Al,O,,(OH). tremolitic asbestos. Sodicgedrite NaMguAlSiuAl,O,,(OH), Sodic-ferrogedrite NaFe!*AlSiuAl,O,,(OH), Mg-Fe-Mn-Li amphiboles Limits for the use of end-membernames: The group is defined as possessing(Ca + Na)u < 1.00 + = and (Mg,Fe,Mn,Li)B > Gedrite Mg/(Mg P"'*1 050 1.00 in the standardformula; the Fetrogedrite Mg/(Mg+Ps,*)<0.50 detailed classification is shown in Figure 2. The main Sodicgedrite Mg/(Mg + Fe'z+)> 0.50; Na > 050 changes from IMA 78 are the adoption of divisions at Sodic-fenogedrite Mg(Mg + Fe'*) < 050; Na > 0.50 Mg/(Mg + Fe'z*): 0.50, the reduction of adjectives,and It should be noted that gedrite and ferrogedrite, with the abolition of tirodite and dannemorite. or without sodic prefixes, extend down to at least Si = Orthorhombic forms 5.50. Discovery of homogeneous Na(Fe,Mg)5Al,Si5 AI,O,,(OH), will justify a new name. Anthophyllite series.N4Li. (Mg,Fe,*,Mn), _"-.A1, Holmquistite series. n(Li,(Mg,Fe'z*).(Fe3*,Al),)Si8O,, (Si, whereSi > 7.00 (oth- , ,*.A1"*,, .)O2,(OH,RCI), (OH,ECI), Li > 1.00 is critical. erwise the mineral is gedrite) and Li < 1.00 (otherwise the mineral is holmquistite). Most anthophyllites have End-members: the Pnma structure; those with the Pnmn structure may Holmquistite n(Li,Mg.Al,,Si,O,,(OH), be prefixed proto without a hyphen. Ferroholmquistite n(Li,FeltAl,)Si,O,.(OH),
End-members: Limits for the use of end-membernames: Anthophyllite nMg, Si,O,,(OH), Holmquistite Mg/(Mg + P":*; > 050 Feno-anthophyllite nFel-Si*O22(OH), Femoholmquistite Mg/(Mg + Fez*)< 050 Sodicanthophyllite NaMg,Si,AlO,,(OH), Sodic-ferro- Monoclinic forms of the Mg-Fe-Mn-Li amphiboles anthophyllite NaFq.Si?AIO,,(OH), Cummingtonite-Grunerite series. I (Mg, Fe2*,Mn, Limits the use of end-membernames: for Li),Si,O,,(OH),. Li < l.fi). Most members of this series > Anthophyllite Mg/(Mg+Fe'?t) 0 50 have space group CUm; those with Wm may optionally Feno-anthophyllite Mg/(Mg + Fe,*) < 0.50 Sodicanthophyllire Mg/(Mg + Fe,*) have this symbol as a suffix at the end of the name. >050;Na>050 End-members: Sodic-feno- anthophyllite Mg/(Mg + Fe.*) < 0.50; Na > 0.50 Cummingtonite nMg, Si,O,.(OH), Grunerite nFel-Si,O.,(OH), Gedrite series. NaJ-i.(Mg,Fer*,Mn),, . Alr(Sir_, ,*. Manganocummingtonite nMn,Mg.SirO,,(OH), Al,*,-.)O,,(OH,ECD, where (.x + y - z) > 1.00 so that Permanganogrunerite lMnnFel-SirOrrOH), LEAKE ET AL.: AMPHIBOLE NOMENCLATURE t025
calcic amphiboles compound names like tschermakitic hornblende have been abolished, sadanagaite(Shimazaki et al. 1984) and rNa-h)A 0.50 ,DtagnnPannetere:CaB,1.50 | l--ffi1 cannilloite (Hawthorne et al. 1996b), have been added, 10 and the boundariesof the group have been revised. Horn- blende is retained as a generalor colloquial term for col- ored calcic amphiboleswithout confusion with respectto I the precise range shown in Figure 3, becausehornblende is always prefixed with an adjective in the precise no- menclature. Because of the strong desire especially, but hastingsile not solely, expressedby metamorphic petrologists to re- 00 the distinction of green actinolite from colorless J tain 70 60 so 60 7s 65 55 45 65 55 tremolite, the subdivisionstremolite, actinolite, and ferro- Si in formula Si in formula actinolite of IMA 78 are retained, as shown in Figure 3. End-members: Tremolite nCa,Mg,SiO"(OH)' 10 Ferro-actinolite nCa,Fel*Si,O,.(OH)' 09 Edenite Naca,Mg, Si,AIO,,OH), Ferro-edenite NaCa.Fel*Si'AlO"(OH), Pargasite NaCa'(MgoAl)SiuAl'0"(OH). L Final nanes rqui.e the relevenl Fefixes which de lbd in Ferropargasite NaCa,(Fel*Al)SiJ.l,O"(OH)' Table I and may opdondly Dclude denod'fEF rhare Magnesiohastingsite NaCa,(MgFe3*)SiuAl,O,'(OH). Hastingsite NaCa,(Fel*Fer*)Si6Al,O,,(OH), Tschermakite nCa,(Mg.AlFe3t)SiuAl,O"(OH). 00 Ferrotschermakite !Ca,(Fel-AlFe3-)Si"Al.O,,(OH), 7o Aluminotschermakite nCadMg.Al,)Si"Al'O"(OH)' 80 7s 65 60 55 Si in formula Alumino-ferrotschermakite nCa,(Fe3*Al')SiuAl,O.,(OH), Ferritschermakite nCa'(Mg.Fel.)SiuAl,O"(OH)' Frcunn 3. Classificationof the calcic amphiboles. Ferri-ferrotschermakite trCa,(Fe3-Fq*)Si6Al,O,,(OH), Magnesiosadanagaite NaCa,[Mg.(Fe'-,Al),]Si,Al.O,'(OH)' Sadanagaite NaCa,[Fe'?*.(Fer*,Al),]Si,Al.O,,(OH), Manganogrunerite flMn,Fe3.siso,,(oH), Magnesiohomblende !Ca,[Mg"(Al,Fe3*)]Si?AIO,,(OH), Fenohornblende !Ca,[Fe'z*n(Al,Fe3*)]SirAlO,,(OH), Limits for the use of end-membernames: Kaersutite NaCa,(Mg,Ti)SiuAl.O'.(OH) Cummingtonite Mg/(Mg+P.'*;=0.50 Fenokaersutite NaCa'(Fe.2-Ti)Si"Al'O,.(OH) Grunerite Mg/(Mg + Fe'z.)< 050 Cannilloite CaCa,(Mg,Al)Si.Al.O,,(OH)' Manganocummingtonite Mg/(Mg + Fe't) - 0.50 1.00< Mn < 3.00 Permanganogrunerite Mg/(Mg + Fe,*) < 050; 3.00 < Mn < 500 Limits for the use of the end-membernames: Manganogrunerite Mg(Mg + Fe't) < 0.50;100 < Mn < 300 These are summarizedin Figure 3 with respect to Si, It should be noted that the names given extend down (Na + K)o, Mg/(Mg + Fe'z*),and Ti. The prefixes ferri to 7.00 Si. If a mineral with less than Si : 7.00 is dis- and alumino are only used when Fe3" ) 1.00 and "IAl > covered,then it will justify a new name basedon the end- 1.00 (Table 1). For kaersutite and ferrokaersutite,Ti > member MgrAlrSiuAlrOrr(OH), 0.50; any lesser Ti content may optionally be indicated = Clinoholmquistite series: nGi, (Mg,Fe2*,Mn). as in Table 2. Cannilloite requires Cao 9.56. (Fe3*Al),)SirO,,(OH,ECl),Li = 1.fi). Sodic-calcicamphiboles End-members This group is defined to include monoclinic amphiboles Clinoholmquistite !(Li,Mg.A1,)Si. O,,(OH). in which (Ca + Na)u = 1.00and 0.50 ( Nau ( 1.50.The Clinofenoholmquisrite nGi,Fe?*Al,)Si*O,,(OH), is shown in Figure 4. There are no Ferri-clinoholmquistite !(Li,Mg,Fei*)Si,O..(OH), detailed classification Ferri-clinoferroholmquistite n(Li, Fe'?*rF{t)SisO,,(OH), significant changesfrom IMA 78 except for the 5OVoex- pansion of the volume occupied by the group in Figure Limits the use of end-membernames: for 1. Becauseof the concentrationof compositionsrelatively = Clinoholmquistite Mg/(Mg + Fe'*) 050 near the end-members,the increasein the number of anal- + ps,t) < Clinoferroholmquistite Mg/(Mg 050 the number classified Ferri-clinoholmquistite Fe3* > 1; Mg/(Mg + Fe'*) = 0 50 yses in this group compared with Ferri-clinoferroholmquistite Fer* > 1; Mg/(Mg + p"z*1 < 0 50 in IMA 78 is quite small (much less than 50%). Never- theless,a number ofpreviously classifiedcalcic and alkali Calcic amphiboles amphibolesnow become sodic-calcic amphiboles. The group is defined as monoclinic amphiboles in End-members: which (Ca + Na)u > 1.00 and Nau is between 0.50 and Na(CaNa)Mg,Si,O.,(OH), 1.50; usually Ca" > 1.50. The detailed classification is Richterite Ferrorichterite Na(CaNa)F4*Si.O,,(OH), shown in Figure 3. The number of subdivisions used in Winchite ntCaNa tMgotAl.Fe' )Si*O,,(OH), IMA 78 has been more than halved; silicic edenite and Ferrowinchite tr(CaNa)Fe?-(Al Fe3*)Si,O.,(OH)' 1026 LEAKE ET AL.: AMPHIBOLE NOMENCLATURE
sodic-calcicamohiboles cipal changesare the introduction of nybdite with Si close Diagram Parameters: to 7, as approvedin 1981 (Ungarettiet al. 1981),ferric- (Na+K)A>050;(Ca+NaB)>1 00; 050
yn2+; DiagramParameters: Nas > 1.50;(Mg + Fe2++ Mn2+)> 2 5; Diagtam Parameters: Nag > 1.50; (MS * 5"2+ a > 2 5; (Alvlor Fes+)> Mn3*;Li < 0.5;(Mg or Fe2+)> Mn2+ (AIvl or Fe3+1, yn3+; Li < 0 5; (Mg or Mn2+)> Fe2+ (Na+K)a>050 (Na+K)A>050 1.0 nybdite .'ckermannil€ (Atvl> Fe3.) (Atvr> Fe3r) (Arvr> Fe3+)
magneslo- magnesro- + + arfvedsonite N arfvedsonite feric-nyboite N o) (Atvr< (Atvl< c (Atvr< Fe3+) II Fe3+) Fe3+)
T nq lo.s o"" lerro- terronyboite (Atvr (Atvr> Fe3+) > Fe3+) kozulite
arfuedsonite lerric-ferronyboite (Atvr< Fe3+) (Atvr< Fe3*) 00 70 g.o 75 7.0 7.5 70 8.0 7.5 Si in formula Si in formula Si in formula b sodicamphiboles
NaB>1 50;(Na+K)A>o50; (Mg+Fe2++ Mn2+)<25; lDiagramParametersj Li > 0.5
(Mg or Fe2+) > Mn2+ (Mg or Mn2+) > Fe2+
leakeite leakeite o Fe3+> [Atvror Mn3+ Fe3+> [AlVlor Mn3+] l! c + + > 05 ovJ
terroleakeite kornite (Fe$ > [Al or Mn3+l) Mn3+ > fAtvl or Fe3+l
75 75 Si informula Si in formula
DiagramParcmeters: NaB> 1 50, (Na+ K)A>O 50; ([/g + Fe2++Mn2*) <25;
(Mg or Mn2+) > Fe2+
+ oo5
ungar€ttiiter Mn3+ > [Atv] or Fe3+l
8 o 7'o si intlismura ''*'j"",'fi':ffi ;l,ffif,l'?;l^3, Frcunr 5. (a) Classification of the sodic amphiboles with (Mg + Fe,* + Mn'*) > 2.5 apfu. (b) Classification of the sodic amphiboleswith (Mg + Fe2' + Mn.*) < 2.5 apfu. 1028 LEAKE ET AL,: AMPHIBOLE NOMENCLATURE
Actinolitic hornblende magnesiohornblende amphibolefrom TanohataMine, Iwate Prefecture,Japan Journal ofthe Ferro-actinolitic JapaneseAssociation of Mineralogists,Petrologists and Economic Ge- -328. hornblende ologists, Sendai,62, 3 11 ferrohornblende Nickel, E.H. (1992) Nomenclaturefor mineral solid solutions American Tschermakitic Mineralogist, 77, 660-662. hornblende tschermakite Nickel, E.H. and Mandarino, J.A (1987) Proceduresinvolving the IMA Ferro-tschermakitic Commission on New Minerals and Mineral Names and guidelineson hornblende ferrotschermakite mineral nomenclature.American Mineralogist, 72, 7031-1042. Rock, NMS and Leake, B.E. (1984) The Intemational Mineralogical Edenitic hornblende edenite Association amphibole nomenclaturescheme: computerization and its Ferro-edenitic consequencesMineralogical Magazine, 48, 211-217 hornblende ferro-edenite Shimazaki,H, Bunno, M, and Ozawa,T. (1984) Sadanagaiteand mag- Pargasitichornblende pargasrte nesio-sadanagaite,new silica-poor membersof calcic amphibole from Ferroan pargasitic Japan American Mineralogist, 69, 465-47 1. Ungaretti,L, Smith, D.C., and Rossi, G (1981) Crystal-chemistryby hornblende pargasrteor X-ray structurerefinement and electron microprobe analysisof a series ferropargasite of sodic-calcicto alkali amphibolesfrom the Nyb6 eclogite pod, Nor- Ferro-pargasitic way Bulletin de Mineralogie, 104,4OO-412 hornblende ferropargasite Veblen, D.R. (1981) Amphiboles and other hydrouspyriboles-mineralogy In Mineralogical Society of America pargasite Reviews in Mineralogy, 9,A, 372 Ferroan pargasrteor Veblen, DR and Ribbe, PH. (1982) Amphiboles: petrology and experi- ferropargasite mental phase relations In Mineralogical Society of America Reviews Silicic edenite edenite in Mineralogy, 9B, 390 Silicic ferro-edenite ferro-edenite Magnesio-hastingsitic AppBNorx 1. hornblende magnesiohastingsite Information concerning the etymology, the type local- Magnesian hastingsitic ity, and the unit-cell parametersof thirty amphibole end- hornblende magnesiohastingsiteor members hastingsite Hastingsitic hornblende hastingsite Actinolite Magnesian hastingsite = magnesiohastingsiteor From the Greek, ahin, a ray, and lithos, a stone,alluding hastingsite to the radiating habit. Type locality: None. RrrnnnNcns crrED X-ray data: a 9.884A, U tS.t+S A, c 5.294A, B tM.t'. Anthony, J W., Bideaux,R.A, Bladh, K.W., and Nichols, M.C (1995) (PDF 25-151 on specimen from Sobotin, Czech Handbook of Mineralogy, 2, pt. I Mineral Data publishing, Tircson, Republic). Aizona,446 p References: R. Kirwan (1794): Elements of Mineralogy, Armbrustet T,, Oberhiinsli,R, Bermanec, V., and Dixon, R (1993) Hen- (actynolite). nomartinite and komite, two new Mnr* rich silicatesfrom the Wessels l, 167 Modified by J.D. Dana [1837. Sys- Mine, Kalahari, South Africa SchweizerischeMineralogische und pe- tematicMineralogy (1st ed.), 3091. trographischeMitteilungen, 73, 349-355. Chukhrov, FV, Ed. (1981) Minerals: a handbook Silicateswith multiple Anthophyllite chains of Si-O tetrahedra,397 p Nauka, Moscow Named from anthophyllum 'clove' referring Deer,WA, Howie, R.A., andZussman, J (1963)Rock-forming minerals. to its char- 2 Chain silicates,379 p Longmans,London acteristic brown color. -(1997). Rock-forming minerals 28 Double-chain silicates. Geo- Type locality: Described by Schumacher(1801, p. 96) logical Society of London (in press). as from the Kongsberg arca, Norway, the exact locality Ernst, W.G (1968) Amphiboles, york 119 p Springer-Verlag,New being kept secret, but later (Mdller 1825) as from the Hawthorne, FC. (1983) The crystal chemistry of the amphiboles Cana- dian Mineralogist, 21, l'l 3-480. KjennerudvannLake near Kong"sberg. Hawthorne,EC , Oberti, R, Cannillo, E, Sardonne,N., Zanetti, A., Grice, X-ray data: d 18.5A, b I7.9 L, c 5.28 A. (PDF9-455 JD., and Ashley, PM. (1995) A new anhydrousamphibole from the on specimenfrom Georgia,U.S.A.). Hoskins mine, Grenfell, New South Wales, Australia: Description and References; N.B. Mdller (1825. Magazin for Naturved- crystal structure of ungarettiite,NaNa,(Mni*Mn:*)SisO,,O,. American Mineralogist, 80, 165-172. enskaberne.Christiania. 6: 174). C.F Schumacher(1801. Hawthorne,EC., Oberti, R., Ungaretti,L., and Grice, J.D (1992)Leakeite, Versuch Verzeich. Danisch-Nordisch Staat, einfach Min. NaNa,(Mg.Fel*Li)SisO.,(OH),,a new alkali amphibole from the Kajli- 96 and 165). dongri manganesemine, Jhabuadistrict, Madhya Pradesh,India. Amer- ican Mineralogist, 77, ll12-1 115 Arfvedsonite -(1996) A new hyper-calcicamphibole with Ca ar the A site: fluor- cannilloite from Pargas,Finland American Mineralogist,8 l, 995-1002 Named for J.A. Arfvedson. Hawthome, FC, Oberti, R, Ungaretti, L., Ottolini, L., Grice, J D, and Type locality: Kangerdluarsuk,Greenland. _ Czmarske, G.K. (1996) Fluor-fenoleakeire,NaNa,(Fq*Fer-LDSisO,,q, X-ray data: ag.g4 A, u ts.tl A, c 5.34 A. B tO+.+O'. a new alkali amphibole from the Carada Pinabete pluton, euesta, New (PDF 14-633 on specimen from Nunarsuatsiak, Mexico, U S.A American Mineralogist,81,226-229 IMA (1978) Nomenclature of amphiboles American Mineralogist, 63, Greenland). r023-t052. References:H.J. Brooke [823. Ann. Phtl. 2l: (2nd ser., Nambu, M, Tanida, K., and Kitamura, T. (1969) Kozulite, a new alkali vol. 5), 3811 (arfwedsonite): amended by T. Thomson LEAKE ET AL.: AMPHIBOLE NOMENCLATURE 1029
(1836. Outlines of Mineralogy, Geology, and Mineral by N. Sundius(1946. Arsbok Sver.Geol. Unders.40: no. Analysis, 1: 483). 4). Nearestanalysis to end-membermay be that of Leake (1971.Min. Mag. 38: 405). Barroisite Origin of name not found. Gedrite Type locality: Not traced. Named from locality. References:G. Murgoci (1922. C.R. Acad. Sci. Paris, Type locality: H6as Valley, near Gddres,France. l15A: 373 and 426).Now definedby B.E. Leake (1978. X-ray data: a 18.594A, b ll.890 A, c 5.3(X A. (PDF Min. Mag. 42:544). 13-506 on specimen from Grafton, Oxford County, Maine,U.S.A.). Cannilloite References:A. Dufr6noy (1836. Ann. Mines, ser.3, 10: Named for Elio Cannillo of Pavia, Italy. 582). Definedby B.E. Leake (1978.Min. Mag. 42:539). Type locality: Pargas,Finland. X-ray data: (Fluor-cannilloite) a 9.826 A. U tl.SOl A, c Glaucophane 5.301A. B 105.41'. Named from the Greek glaukos, bluish green and phai- Reference: FC. Hawthorne, R. Oberti, L. Ungaretti and nesthai, to appear. J.D. Grice (1996. Am. Mineral. 81: 995). Type locality: Syra, pyclades, Greece. X-ray data: a9.595 A, b 17.798A, c 5.307A. B 103.66'. Clinoholmquistite (PDF 20-453 on specimenfrom SebastopolQuadrangle, Named as a monoclinic polymorph of holmquistite. California,U.S.A. Seealso PDF 15-58 and 20-616). Type localityz Golzy,^SayanyMountain. Siberia,Russia. Reference: J.EL. Hausman(1845. Gel. Kijn Ges. Wiss. X-ray data: a9.80 A, b I7.83 A. c 5.30 A. B 109.10'. Gcittingenp. 125) (Glaukophan). (PDF 25-498 on specimenfrom Siberia, Russia). References: I.V. Ginzburg (1965. Trudy Min. Muz. Grunerite Akad. Nauk. SSSR,16:73).In B.E. Leake(1978. Min. Named for E.L. Gruner. Mag. 42: 540) defined in a series with magnesio-clino- Type locality: Collobribres, Var. France.. holmquistite and ferro-clinoholmquistite. X-ray data: a 9.57 A, b 18.22A, c 5.33 A. (PDF 17-145 on specimenfrom White Lake, Labrador, Canada). Cummingtonite References: Describedby E.L. Gruner (1847. C.R. Acad. Named for locality. Sci. Paris, 24: 794) but named by A. Kenngott (1853. Type locality: Cummington.Massachusetts. U.S.A. Mohs'scheMin. Syst.69). Definedby B.E. Leake (1978. X-ray data: a 9.534 A, U ft.ZZt A, c 5.3235 A. B Min. Mag. 42:549). 101.97".(PDF 31-636 on specimenfrom Wabush iron formation, Labrador, Canada). Hastingsite References:C. Dewey (1824.Amer. J. Sci. ser.l, 8: 58). Named for locality. Definedby B.E. (1978. Leake Min. Mag. 42:549). Type locality: Hastings.County, Ontarip, Canada. X-ray data: a 9907 A, b 18.023 A, c 5.278 A. B Eckermannite 105.058'. (PDF 20-378 on specimen from Dashkesan, Named for H. von Eckermann. Tianscaucasia,Russia. Also PDF 20-469). Type locality: Norra Kiirr, Sweden. References: ED. Adams and B.J. Harrington (1896. X-ray data: a 9j7652 A, U tl.SgZ A, c 5.284 A. g Amer.J. Sci.4th ser.,l;212;1896.Can. Rec. Sci. 7: 8l). 103.168'.(PDF 20-386 on syntheticmaterial). Defined by B.E. Leake (1978.Min. Mag. 42:553). References: O.J. Adamson (1942. Geol. Ftjr. Stockh. 64: 329;1bid' I9M. 66: 194). Definedby B.E. Leake (1978. Holmquistite Min. Mag. 42:546). Named for PJ. Holmquist. Type locality: Utci, Stockholm, Sweden. Edenite X-iay data: a 18.30 4,, U n.AS A, c 5.30 A. Gnp Named for locality. I3-4OI on specimenfrom Barrante, Quebec, Canada). Type locality: Eden (Edenville), New York, U.S.A. References: A. Osann (1913. Sitz. Heidelberg Akad. X-ray data: a9.837 A, U n.gS+ 4,, c 5.307A. B tOS.tS'. Wiss., Abt. A, Abh. 23). Dimorphous with clinoho- (PDF 23-1405 on specimenfrom Franklin Furnace,New lmquistite.Defined by B.E. Leake (1978.Min. Mag. 42: Jersey,U.S.A.). s49). References: Not analyzed in original description. Two analyses of topotype material by C.E Rammelsberg Hornblende [1858. Ann. Phys. Chem. (Pogg), lO3: 44I] and G.W. The name is from the German mining term horn, horn, Hawes (1878. Amer. J. Sci. Ser. 3, 16: 391) differ con- and blenden, to dazzle. siderably, and neither falls within edenite range of Leake Reference: Use of the term hornblende and relationship (1978. Min. Ma5 42:542). Cunent definition proposed to other calcic amphibolesdiscussed by Deer et al. (1963. 1030 LEAKE ET AL.: AMPHIBOLE NOMENCLATURE
Rock-forming minerals. 2. Chain silicates. 265. Long- Reference:L. Ungaretti,D.C. Smith,and G. Rossi(1981. mans,London). Defined by B.E. Leake (1978.Min. Mag. Bull. Min. 104: 400). 42: 551). Pargasite Kaersutite Named for locality. Named from locality. Type locality: Pargas,Finland. Type locality: Kaersut, Umanaksfjord, Greenland. X-ray data: a9.870 A, b 18.006A, c 5.300A. B 105.43". X-ray data: a 9.$ A, b 11.89A, c 5.30 A. B tOS.ts'. (PDF 23-1406 and PDF 4I-1430 on synthetic material). (PDF 17-478 on specimen from Boulder Dam, Aizona, References:F von Steinheil[1814 in Tasch.Min. (1815) u.s.A.). Jahrg. 9, Abt 1, 3091. The name was widely used for References: J. Lorenzen (1884. Medd. Grgnland 1: 21). green hornblendebut was redefinedby N. Sundius (1946. Defined and given speciesstatus by B.E. Leake (1978. Arsb. Sver. Geol. Undersdk. 40: 18) and B.E. Leake Min. Mag. 42:551). (1978.Min. Mag. 42:550 and 552).
Katophorite Richterite Named from the Greek kataphora, a rushing down, in Named for T. Richter. referenceto its volcanic origin. Type locality: L Angban, Vdrmland, Sweden. Type locality: Christiana District (now Oslo), Norway. X-ray data: a 9.907A, b I7 .979A, c 5.269A. B 104.25'. References: W.C. Brdgger (1894. Die Eruptivgest. Kris- (PDF 25-808on syntheticmaterial; see also PDF 31-1284 tianiagebietes,Skr. Vid.-Selsk I, Math.-natur. Kl. 4: 27). for calcian and 25-615 and 31-1082for potassian). Frequently termed catophorite,and other variants,but ac- References: An imperfect description by A. Breithaupt cepted IMA spelling is katophorite. Defined by B.E. (1865. Bergmann Huttenmann.Zeit. 24: 364) was shown Leake (1978.Min. Mag. 42: 544). by H. Sjcigren(1895. Bull. Geol. Inst. Univ. Upsala,2: 71) to be an amphibole.Defined by B.E. Leake (1978. Kornite Min. Mag. 42:544). Named for H. Korn. Riebeckite Type locality: Wessels Mine, Kalahari Manganese Named for E. Riebeck. Fields, South Africa. Type locality: Island"of Socotra, Indian Ocean. X-ray data: a 9.94(l) A , U n.eOQ) i+, c 5.302(4)A. B X-ray data: a 9.769A, b 18.048A, c 5.335A. 103.59". 105.52'. B (PDF 19-1061on specimenfrom Doubrutscha,Romania). Reference: T. Armbruster, R. Oberhiinsli, V. Bermanec, References:A. Sauer (1888, Zeit. deut. geol. Ges. 40: and R. Dixon (1993.Schweiz. Mineral. Petrogr.Mitt. 73: 34e). 138). Definedby B.E. Leake (1978. Min. Mag 42: 546). Sadanagaite Kozulite Named for R. Sadanaga. Named for S. Kozu Type locality: Yuge and Myojin.Islands, Japan. Type locality: Tanohatamine, Iwate Prefecture,Japan. X-ray data: a9.922 A, b 18.03A, c 5.352A. B 105.30'. X-ray data: g a 9.991A, A tS.tt A, c 5.30 A. to+.6'. Reference: H. Shimazaki. M. Bunno. and T Ozawa (PDF 2s-850) (1984.Am. Mineral. 89: 465). References: M. Nambu. K. Tanida. and T. Kitamura (1969. J. JapanAssoc. Min. Petr.Econ. Geol. 62: 3lI). Taramite Defined (1978. by B.E. Leake Min. Mag. 42: 557). Named for type locality. Type locality: Walitarama,Mariupol, Ukraine. Leakeite X-ray data: a 9.952A, b 18.101A, c 5.322A. B 105.45". Named for B.E. Leake. (PDF 20-734 on specimen of potassian taramite from Type locality: Kajlidongri manganesemine, Jhabuadis- Mbozi complex, Tanzania). trict, Madhya Pradesh,lndia. References:J. Morozewicz [1923. Spraw. Polsk. Inst. X-ray data: a 9.822A, tr t1.836A, c 5.286A. B tO+.:Z'. Geol (Bull. Serv. Geol. Pologne), 2: 61. Redefined by Reference: EC. Hawthorne, R. Oberti, L. Ungaretti, and Leake(1978. Min. Mag. 42:544). J.D. Grice (1992. Am. Mineral. 77: lll2). Tremolite Nybiiite Named for locality. Named from locality. Type locality: Val Tremola, St. Gotthard, Switzerland. Type locality: Nybii, Nordfjord, Norway. X-ray data: a 9.84 A, b 18.02A, c 5.27 A. B 104.95'. X-ray data: In Ungarettiet al. (1981) X-ray data given (PDF 13-437on specimenfrom San Gotardo, Switzerland for many specimensand a single "type" specimen not and PDF 3l-I285 on synthetic material). distinguished. References:E. Pini (1196.in H.-8. Saussure,1923.Yoy- LEAKE ET AL,: AMPHIBOLE NOMENCLATURE t03I ages dans les Alpes, 4: sect). Defined by B.E. Leake Empirical estimatesof ferric iron are not just poor ap- (1978.Min. Mag.42: 542). proximations that suffice in the absenceof analytical de- terminations of ferric-ferrous ratios. Empirical estimates Tschermakite yield exactly the sameresults as analytical determinations Named for G. Tschermak. Originally described as a hy- of ferric iron, if (1) the analysisis complete (total Fe plus pothetical "Tschermak molecule." all other elements), (2) the analytical determinationsare References A.N. Winchell (1945. Am. Mineral. 30: 29). accurate,and (3) the mineral stoichiometry (ideal anion Defined by B.E. Leake (1978. Min. Mag. 42: 550 and and cation sums) is known. In the case of amphiboles, ssz). condition (3) cannot be uniquely determinedbecause the A-site occupancyvaries. However, knowledge of amphi- Ungarettiite bole stoichiometry and element distribution can be used Named for L. Ungaretti. to estimatea range of permissible sffuctural formulae and Type locality: Hoskins mine, near Grenfell, New South ferric contents. Wales, Australia. The most welcome circumstanceswill be those where X-ray data: a 9.89(2) A, A tS.04(:) it, c 5.29(Il A. B the difference between the limiting structural formulae 104.6(2)". are trivial and the entire range plots within the sameclas- Reference: FC. Hawthorne, R. Oberti, E. Cannillo, N. sification field. However. there will also be caseswhere Sardone,and A. Zanetti (1995.Am. Mineral. 80: 165). the range of stoichiometrically allowable formulae is broad and spanstwo or more fields in the classification. Winchite Some users of the amphibole nomenclaturemay consider Named for H.J. Winch, who found the amphibole. this a less than satisfactory solution, but, until it is pos- Type locality: Kajlidongri,Jhabua State. India. sible to determine ferric contentsroutinely with the same X-ray data: a 9.834A, b t8.062 A, c 5.300A. g 1(X.4'. ease and convenience of electron microprobe analyses, (PDF 20-1390) empirical estimatesare probably the best alternative. References: L.L. Fermor (1906. Trans. Mining Geol. The procedureof estimating ferric iron requires at least Inst. India, 1: 79) naming the amphibole described in one recalculation of the all-ferrous analysisto a different 1904 (Rec. Geol. Surv. India, 31: 236). Topotypematerial cation sum. Consequently,familiarity with calculation of found by B.E. Leake, C.M. Farrow, F Chao, and V.K. mineral formulae is highly recommendedfor a fuller un- Nayak (1986.Min. Mag. 50: 174)proved to be very sim- derstandingof the ferric estimation procedure.Thorough ilar in composition to that originally found by Fermor discussionsof the calculation of mineral formulae can be (1909.Mem. Geol. Surv. India,37: 149). found in the appendicesof Deer et al. (1966, 1992). The topic of ferric estimatesin amphiboleshas beendiscussed GnNnn-tL REFERENCES by Stout (1912), Robinsonet al. (1982,p. 3-12), Droop Chukhrov, FV, Ed (1981) Minerals: a handbook Silicateswith multiple (1987),Jacobson (1989), J. Schumacher(1991), and Hol- chains of p Si-O tetrahedra,397 Nauka, Moscow land and Blundy (1994). An example of the recalculation Clak, A.M. (1993) Hey's mineral index, 852 p Natural History Museum and Chapman& Hall, London of an electron microprobe analysis and the procedurefor Deer, W.A., Howie, R A, and Zussman,J 1963 Rock-forming minerals estimating minimum and maximum ferric contents are 2 Chain silicates,p 2O3 374 Longmans,London. given at the end. Leake, B.E (1978) Nomenclatureof amphiboles American Mineralogist, 63. tO23 tO52. Euprnrca.r, FERRTCrRoN ESTTMATESFoR Appp,Nnrx 2: Tnn EsrrMATroN oF FERRICrRoN rN AMPHIBOLES ELECTRON MICROPROBE ANALYSN OF AMPHIBOLES* The basic formula
INrnooucrroN Present knowledge of amphibole crystal chemistry suggeststhat many amphiboles contain essentially ideal Most users of the amphibole nomenclature will want stoichiometric proportions of 2 (OH) and 22 O atoms. to classify amphibole compositions that have been deter- These anions can be rearranged to give the anhydrous mined with the electron microprobe, which cannot distin- formula basis 23 O (+ HrO), and calculation of the an- guish among the valence statesof elements.This is un- hydrous formulae on this basis is the first basic assump- fortunate becauseit is clear that most amphibolescontain tion necessaryto estimateferric Fe. The ideal cation sums at least some ferric iron; see compilations of Leake in amphibole formulae are not fixed and can vary be- (1968) and Robinsonet al. (1982), for examples.Con- tween 15 and 16 cations per 23 O (anhydrous). Conse- sequently,the typical user of the amphibole nomenclature quently, it is not possible to arrive at a unique ferric es- will need to estimate empirically ferric contents of timation based on stoichiometry, as can be done for amphiboles. minerals with fixed ratios of cations to anions (e.9., py- ilmenite-hematite series). Nevertheless, * Preparedby John C. Schumacher,Institut fiir Mineralogie- roxenes or the Petrologie-Geochemie,Universitiit Freiburg, Freiburg i. Br., based on our present understanding of permissible and 79104.Germanv. usual site occupancies,limits can be placed on the max- t032 LEAKE ET AL.: AMPHIBOLE NOMENCLATURE
Summaryof siteassignments therein). In amphiboles, this can result in violation of at andstoichiometric constraints least one of the criteria Si = 8. lCa = 15. or lK < 16 (Appen. 2 Fig. l). Violations of the other three criteria, Sile and Stoichiometric = 8, > 13, and = 15 (Appen.2Fig.l), Occupancy Calion' Limil lAl lMn lNa cannot be due to failure to account for ferric iron and -1.. si 6 I-slte Si < 8 usually indicate an analytical problem (too few cations at n Ar-24;;;- s fr > g some of the sites'). These analysesshould not be used for $ ll empirical ferric estimates. C-site Cr Fe3* Minimum and maximum estimates For many amphibole analyses,none of the criteria Si = 8, I,Ca < 15, and )K = 16 will be exceededby the $f;H all-ferrous formula; the minimum ferric estimate is the B-sife $ Fe2* all-ferrous formula (i.e., Fe3* : 0 and the site occupan- cies of the all-ferrous formula are all allowable). If one or more of the three criteria Si = 8, )Ca = 15, and )K < 16 are exceeded,Fe3* may be present,and a minimum ^-*'" ferric estimate can be made that yields a formula with < 16 [_rK acceptablestoichiometry. The condition that is exceeded by the greatest amount determines the basis for the re- ' calions arrangedaccording to increasingbnic radius calculation.For example,if Si : 8.005,ICa : 15.030, (smallest,Si to largest,K) and I,K : 15.065,then the lSi limit is exceededby 0.005 = E calion total or subtotal(e. g EMn = sum of all cationstrom greatest Si throughMn in the list) and the ICa limit by 0.030. Because )Ca is in excess, the minimum ferric estimate is obtained by re- tr = vilCdncyal the A-site calculating the formula so that lCa : 15.000 (15 eNK Appen. 2 Fig. l). AppENDrx2: rrcunn 1. Summaryof ideal site assignments,estimate, limitsof variouscation subtotals, and the type of conection(min- The maximum ferric estimates are obtained from the imum or maximum)that canbe obtainedby calculatingthe for- stoichiometric limits >Al > 8. IMn > 13. and INa = mulaeto thesestoichiometric limits (afterJ. Schumacherl99l). 15 (Appen. 2 Fig. l). The condition that is nearestto the Abbreviationsof normalizations:8Si : normalizedsuch that to- minimum value of one of thesesums gives the maximum tal Si : 8; 8SiAl : normalizedsuch that total Si +Al : 8; 13 ferric estimate. For example, if lAl : 9.105, IMn : eCNK : normalizedsuch that total the sum of the cationsSi 13.099, and INa : 15.088, then IAI is exceededby : throughMn (i.e, all cationsexclusive of Ca,Na, K) 13; 15 1.105, lMn by 0.099, and lNa by 0.088. The lNa is : eNK normalizedsuch that total the sum of the cationsSi nearestthe minimum value, and recalculatingthe formula throughCa (i.e., all cationsexclusive of Na andK) : l5; I6CAT so that INa : 15.000(15 eK estimate,Appen. 2Fig. l) : normalizedsuch that total the sum of all cations: 16 (see give alsoRobinson et al 1982,p. 6-12). will the maximum ferric formula. Recalculation of the formulae imum and minimum values of ferric contents, and these The recalculation procedure is described step-by-step limits yield a range of acceptablemineral formulae. at the end of this discussion, but some general aspects are discussedhere. Appendix 2 Table I lists a hypothet- Critical examination of electron microprobe analyses ical amphibole analysis (in weight percent) and four for- The suitability of an electron microprobe analysis of mulae that are based on 23 O atoms. Formulae were cal- an amphibole for a ferric estimation requires the evalua- culated for the two chemical limits (all Fe as FeO or tion of the all-ferrous, anhydrous formula that is calcu- FerO.); the other two are the stoichiometric limits (see lated on a 23 O atom basis. The site assignmentscan be Appen. 2Fig.l), which give the minimum (15 eNK) and used to evaluatethe analyses,and these are given in Ap- maximum (13 eCNK) ferric estimates.All of the stoichi- pendix 2 Figure 1. From the site assignmentdata, it is ometric limits except lCa = 15 (here lCa : 15.029) ne possibleto define the important stoichiometriclimits (cat- met by the all-ferrous formula, which means that the min- ion subtotals) for the amphiboles (column 3, Appen. 2 imum ferric formula is given by the 15 eNK estimate Fig. 1). Acceptable amphibole formulae satisfy all six of (Appendix 2 Table l). these criteria. Exceeding one or more of these stoichio- BecauselMn (13.201)is nearestthe lowest allowable metric limits indicates that problems exist with the struc- sum, the maximum ferric estimate values, and the ferric tural formula, and the identity of the unfulfilled condition formula is obtainedby recalculatingas before, but, in this suggeststhe cause. For minerals that bear ferric iron, the all-ferrous struc- 'Exceptions do exist: Potassiumtitanian richterite (Oberti et tural formulae have cation sums that are too high (for al. 1992) has Ti at tetrahedralsites, and cannilloite has one Ca further discussionsee J. Schumacher1991. andreferences at the B (M4) position. These exceptionsare rare. LEAKE ET AL: AMPHIBOLE NOMENCLATURE 1033
Appenorx2: Taele 1. A hypotheticalamphibole analysis and the structuralformulae that are based on the chemicaland stoichiometric limits All ferrous Fe (chemi€llimit) l5eNK Formulae (stoichiomelric Analysis Atl All ferric Fe l3eCNK 16CAT (chemlcalllmit) (stoichiometric (stolchiomet.iclimil) (wt%) ferrous 15eNK 13 eCNK All fenic 15eK (stoichiomet.ic limil) sio, 39.38 DI 6.093 6 081 6 000 5.714 16 Al2o3 1670 AI 1.907 ---- 1.919 2_000 2.246 tca$ons - - tcatlons eNK FeO 23.54 8.000 I 000 8.000 8.000 14 Mgo 4.40 - 11.03 AI 1.139 1 122 1.000 0.571 - tcations eCNK Naro Fe3* 0.000 0088 0700 2.857 ,12 Mg 1.015 1014 1000 0.952 o C" Total 97.42 Fe2* 2.845 2777 2.300 0.000 ?10 t 5.000 5.000 5 000 4.380 o (t Fer* 0.201 0.176 0 000 0.000 l8 Ca 1.799 1 A24 1 800 1.714 o NA 2.000 2 000 2 000 2.000 E6 G Ca 0.029 0 000 0 000 0.000 o4 Na 0.711 0.709 0 500 0.381 Sum 15.740 15 709 15 500 14.761 2 Note.'Theferrous formula assumes total Fe as FeO, the ferricformula j:i'-.:'-'i assumestotal Fe as Fe2Os,the 13 eCNK and 15 eNK formulaeare based 0 on stoichiometriclimits. See text for discussion 14.5 15.0 15_5 TotalCations per 23 Oxygens
AppsNDrx2: Frcuno2. Plot ofvariouscation values and sums case, the normalization must insure : that lMn 13.000 vs. total cationsthat illustratesthe continuousvariation of these (here : the normalization factor is 13 + 13.201 0.9848). valuesrelative to chemicaland stoichiometriclimits. The stoi- The minimum values for IAl, IMn, and I,Na are, re- chiometriclimits aregiven in Appendix2 Figurel, andthe val- spectively, 8.000, 13.000,and 15.000,and the actual val- uesare based on theamphibole example in Appendix2 Tablel. uesare 9.139.l3.20L and 15.740. These formulae for the minimum and maximum ferric estimatescan be calculatedin either of two ways: (1) by Discussion of the recalculation results normalizing all the cations of the all-ferrous formula that The variation in some cation values within the ranges were calculated on a 23 O atom basis such that : XCa of possible formulae (Appendix 2 Table 1) that are de- 15.000and I,Mn : 13.000(i.e.. cationsof each element fined by the chemical and stoichiometric limits are com- multiplied by 15 + or 13 + here + lCa lMn, 15 15.029 pared in Appendix 2 Figure 2. In general, the range of : 0.9981 and 13 + : 13.201 0.9848,respecrively), or possible formulae that are defined by the stoichiometric (2) by using the normalization factor to determine the limits will be much narower than the range obtained new cation sum and then recalculating the entire formula from the two chemical limits. A diagram like Appendix : = on cation bases that set lCa 15.000 and )Mn 2 Figure 1 could be constructed for every electron mi- 13.000.The secondmethod requiresmore calculation,but croprobe analysis, and on such a diagram, the range of (1991) J. Schumacher has shown that this method leads both the chemical and the appropriateset of stoichiomet- to fewer rounding errors than normalizing the cations in ric limits can be considered.These will vary greatly from the 23 O atom-basedformula. example to example. It can be inferred from Appendix 2 The formula obtainedfrom either recalculationmethod Figure 2 that the range of permissible amphibole formu- has less than 23 O atoms. The number of Fe3* cations is lae could be, and commonly is, bounded by one of the found by calculating the number of moles of FeO that chemical limits and one of the stoichiometric limits. must be converted to FeO,, to bring the sum of the O The relationships among cation sums that are illustrat- - atoms to 23 and equals (23 XOx) x 2, where lOx is ed in Appendix 2 Figwe 2 show that comparisonof some the sum of the O atoms in the normalized formula (lOx of the possible normalization factors, which are obtained -- )Ro* x 2 + >R3* x 1.5 + tRr* + tR'* x 0.5, where from the stoichiometric limits, can be used to (1) check lR : the sums of cations with the same valence). The the applicability of a specific estimate of the proportion moles of FeO equal Fe, - Fe3*, where Fe, : total Fe in of ferric iron, and (2) determine limits, chemical or stoi- the normalized formula. Following any recalculation,it is chiometric, that give the minimum and maximum esti- good practice to recheck that all six stoichiometric limits mates of the proportion of ferric iron. To accomplishthis, are also satisfiedby the new ferric formula. all the normalization factors for all stoichiometric con- 1034 LEAKE ET AL.: AMPHIBOLE NOMENCLATURE
Appelotx 2: Teele 2. A worked exampleof an amphiboleanalysis that appearsin Deer et al. (1992, p. 678)
4 Atomic Atomic proportions proportions 56 Molecular (cations) (O atoms) Anions on the basis Cations on the basis 'I proportions col 2 x cations col.2 x oxygens ot 23 O atoms of 23 O atoms wt% wtTo + mol Wt in oxide in oxide col.4 x 8.45012 col.3 x 8.45012 sio, 51 63 0 85928 0.85928 1.71857 14.52208 7.261 Tio, 0.00 0 000 00 0 000 00 0.00000 0.00000 0 000 Alro" 7.39 o 07248 0 14496 0 21744 1.83736 1 225 Cr,O. 000 0 00000 0 000 00 0.00000 0.00000 0 000 FeO 7.55 0.10509 0.105 09 0 10509 0.88799 0 888 MnO 017 0 00240 0.00240 0.00240 0.o2025 0 020 Mgo 1809 o.44884 o 44884 0 44884 3.79274 3 793 tz oz 021969 0.21969 021969 1 85641 1.856 Na"O 061 0 00984 0.01968 0 00984 0 08317 0.166 KrO 0.00 0 00000 0.00000 0 000 00 0 000 00 0 000 Sum 1.79994 2 72185 23.0000 15210
Factor tor the recalculationof alomic proportions to 23 O basis:23 + 2.72185 : 8.45012
Notes. See the end of the text for a step-by-stepdiscussion of this table Molecularweights are from Robie et al (1978). straints and the chemical limits must be compared (see In addition to the stoichiometric constraints listed in Appen. 2 Fig. 1). The normalization factors for the stoi- Appendix 2 Figure 1, another constraint on maximum chiometric constraintsare calculated from the all-ferrous Fe3* can be defined if the C site in the formulation of the formula using the data in Appendix 2 Table I and are: amphibole nomenclature is further subdivided. The five Minimum ferric estimate: C positions consist of three mica-like octahedra(two Ml 8Si = S/Si : 816.093= 1.313, (1) and one M3) and two pyroxenelike M2 octahedra.The cations Al, Fe3*,Ti, and Cr3* are strongly partitioned into l6cAT : 16DK : 16n5.740: 1.017. Q\ the M2 octahedra.Consequently, an additional maximum all ferrous (no change) : 1.000, (3) ferric estimate can be obtained by assuming all the tet- rahedral and M2-octahedral sites are completely filled : : = 15eNK 15DCa l5lr5.O29 0.998, (4) with cations of valences of 3+ and 4+. This normaliza- Maximum ferric estimate: tion factor (N) can be calculated by solving the two si- multaneous equations for N: (1) N x (Si +Ti + Al + l3eCNK : 13DMn : l3lI3.20I: 0.985, (5) Cr) + Fe3* : 10, which describesthe desired resulting 15eK : 15DNa : 15115.740:0.953, (6) stoichiometry and (2) Fe:+ : (23 - 23 x N) x 2, which gives the Fe3* for this normalization. The solution is N all ferric : 0.938, (7) : 361(46 - Si - Ti - Al - Cr), where Si, Ti, Al, and SSiAl : 8DAl : 8/9.139: 0.875. (8) Cr are the amounts of these cations in the all-ferrous for- For the normalizations that yield minimum estimates(1- mula. For the analysis in Appendix 2 Table 1, this nor- (here 4), the recalculation that requires the lowest normaliza- malization factor abbreviated 10)Fe3*) is 0.9'77, tion factor will be the minimum ferric estimate. For the which is less than the value 0.983 of the 13 eCNK factor; normalizations that yield maximum estimates (5-8), the therefore the 10lFe3* normalization does not give the recalculationthat requires the largestnormalization factor maximum ferric estimatein this case. will be the maximum ferric estimate.All normalizations Most usersof the nomenclaturewill want to report only that lie between these values (in this example, 0.998 and a single mineral formula and name for each amphibole 0.985) will give stoichiometrically acceptableformulae. analysis; consequentlythe overriding question is: Which If any of the normalization factors for the maximum es- correction should be used?Unfortunately, there is no sim- timate (5-8) are greater than any of those for the mini- ple rule, and each group of similar analysesmay require mum estimate (1-4), then the analysis is not suitable for individual treatment.Robinson etal. (1982,p. l1) and J. empirical ferric estimations.Note that normalization fac- Schumacher(1991, p. 9-10) discusssome of thesepos- tors greaterthan 1.000 or less than the normalization fac- sibilities for Mg-Fe-Mn-Li, calcic, sodic-calcic,and sodic tor for the all-ferric formula would yield impossibleferric amphibolesin greater detail. The 10IFe3* correction dis- estimatesthat lie outside the chemical limits. cussed in the preceding paragraph will probably not be LEAKE ET AL.: AMPHIBOLE NOMENCLATURE r035
Apperorx 2: Taele 2.-Extended 'I 1 I 10 Formula 12 ldealsite 7 8 Col 8x Formula (15 eK) Formulaave 13 assign- Min formula Col 6x oxygenper ldeal site (15 eNK) maximum o{ min and Formula ments trom col 6 Cations 0 997 14 cation assignments minimumFe3* Fe3t max. Fe3* from DHZ Si 7 261 Si 7 240 7.161 7 201 7 196 AIIV 0 739 Alv 0 760 0.839 0 799 0 804 Sum T 8 000 Si 7 2401 14 4802 Sum f 8 000 8.000 I 000 8 000 Alvl 0 486 AI 1 2214 1 8321 AIVI o 462 0.369 0 416 0 410 Fe3t 0.000 Ti 0 0000 0.0000 Fe3* 0.133 0.634 0 383 0 263 Cr 0 000 Cr 0 0000 0 0000 Cr 0.000 0 000 0.000 0 000 Mg 3 793 Mg 3.7818 3 7818 Mg 3.782 3 740 3.761 3.759 Fe2* 0 721 Fe2* 0 8854 0.8854 Fe2* 0.624 o 242 0.440 0 618 Mn 0.000 Mn o0202 o 0202 Mn 0 000 0 015 0.000 0 000 Sum C 5.000 1.8511 1.8511 Sum C 5.000 5 000 5.000 5 000 Mg 0.000 Na 0 1659 0.0829 Mg 0.000 0.000 0.000 0 000 Fe'?* 0.167 K 0 0000 0.0000 Fe2* 0.129 0 000 0.057 0 050 Mn 0.020 Sum 15.1659 22.9337 Mn 0.020 0 005 0.020 0.020 UA r.856 UA 1.851 1 831 1.841 1 840 Na 0.000 ICa (col.7) Na 0.000 0 164 0.082 0.090 SUm t' 2 043 15; 15.043:099714 Sum B 2 000 2 000 2.000 2.000 Na 0 166 (23 229337)x 2: O1325 Na 0 166 0 000 0.083 0.074 K 0 000 0885 0.133- 0.753 K 0 000 0 000 0 000 0.000 Sum A 0.166 Sum A 0 166 0.000 0 073 0.074 Total 15.210 Total 15.166 15.000 15.083 15.074
important in calcic amphiboles,but in sodic amphiboles portions of OH, F, or Cl that are present. The critical (e.g., riebeckite, glaucophane)this correction may com- assumption is that exactly two anions (OH, R Cl) are monly yield the maximum ferric estimate. presentfor every 22 O atoms. Choosing a single representativeferric formula from substitutionsinvolving anions the range of possible formulae requires further justifica- Coupled tion or additional assumptions.One solution is to use the The validity of a basic 23 O atom anhydrousamphibole mean value between maximum and minimum ferric con- formula [i.e., exactly two (OH + F + Cl)] is an under- tents (Spear and Kimball 1984). Other solutions can be lying assumptionin the procedureto estimateFe3* in am- obtained for restricted types of amphibole. For example, phiboles. Any variation in these values will have a tre- R. Schumacher (1991) derived a normalization scheme mendous effect on the Fe3* estimation. The partial that yields formulae intermediate to maximum and min- replacementof (OH + F + Cl) by O in the amphibole imum ferric formulae for Ca-saturated, metamorphic structure is an example of this kind of variation and has hornblende and is based on regressionanalysis of horn- long been recognized.Amphiboles that are referred to in blende compositions for which ferric-ferrous determina- numerous mineralogy and optical mineralogy textbooks tions were known. as "basaltic hornblende" (Deer et al. 1966), or the kaer- Generally, it is desirable to determine the extent to sutite end-memberof the IMA amphibole nomenclature, which the minimum and maximum ferric estimations af- can show this type of compositional variation (see also fect the classification of the amphibole in question by Dyar et al. 1993). inspecting the formulae of both the maximum and mini Intuitively, one would expect analytical totals to be af- mum feffic estimates. If the entire range of formulae fected by variable O/OH; however, since these amphi- gives a wide spectrum of possible names, this should boles tend to be richer in Fe3*, the increase in the sum probably be mentioned wherever the amphibole is from the partial exchange of O for OH tends to be offset described. by treating the larger amounts of FerO. as FeO. Conse- quently, even in anhydrous amphiboles with significant Drvrarrons FROM THE BASrc ASSUMpTToNS Fe3*, no compelling evidence of these substitutionswill F and Cl substitutions necessarilybe seenin the analyses.Ferric estimation can still be carried out on analyseswith variable O/OH, but Both F and Cl may substitutefor OH in the amphibole an additional estimate of the HrO and halogen contents structure,and theseelements are not routinely determined is an essentialadditional requirement. at all electron microprobe facilities. Although it is highly recorffnended that these elements also be determined, Wonxro EXAMpLE: CALCULATToNoF A MINERAL their presencehas no effect on the ferric estimation pro- FORMULA AND A FERRIC ESTIMATE FROM AN cedure.Exchange of F or Cl for OH does not change the ELECTRON MICROPROBE ANALYSIS OF AN total number of negative charges(-46) in the anhydrous AMPHIBOLE amphibole formula, so the proportions of cations required The analysis that appearsin Deer et al. (1992, p. 678) to give 46 positive charges are independentof the pro- was chosen as an example (Appendix 2 Table 2). To sim- 1036 LEAKE ET AL.: AMPHIBOLE NOMENCLATURE
AppeHorx2: Taele 3. Calculationof normalizationfactors (a) the eight tetrahedral (T) sites: (i) < Normal- Place all Si here; if Si 8 fill the remain- ization ing sites with Al. Limit Calculationmethod Calculation factor (ii) If Si + total Al < 8, then place all Si + Calculationsfor minimum ferric estimates Al here. 8Si 8+Si 8+7.261 11018 (b) Thefive octahedral (C) sites (INIZ,iN{l, N43) 16CAT 16 + I,K 16 + 15210 1.0519 (i) Place Al remaining from step (a), Ti, Fe3* All ferrous 1.0000 15eNK 15 + >Ca 15 + 15 043 0.9971" (initially : 0), and Cr here. In the follow- Calculationsfor maximum ferric estimates ing order, place enough Mg, Fe2*, and Mn 15 eK 15 + I,Na 15 + 15.210 0 9862. here to bring the total to 5. 13 eCNK 13 + >Mn 13 + 13 187 0.9858 (ii) If >("'Al...Mn) < 5, then place all these AII ferric 23 + 123+ (0.5 x Fe,*)l 23 + 23.444 O9811 10>Fe3* 36 + (46-Si-Al-Ti-Cr) 36 + 37.5141 0.9596 elementshere. SSiAl 8+>Al 8+8.486 0.9427 (c) The 2 (B) sites (M4): * (i) Indicatesnormalizations that Vieldeither the minimumor maximum Place any Mg, Fe2*, or Mn and Ca re- ferricestimates. maining after step (b) here. (ii) If I(Mg. . .Ca) at B < 2, fill the remaining sites with Na to bring the total to 2. (d) The single large (A) site: ulate analysis by elecfion rmcroprobethe Fe3*was recast as (i) Place any remaining Na and K here. Fe2* and the HrO analysis was ignored. The ferric estimate (6) Evaluating the structural was made assuming2(OH) are presentrather than the2.146 formula If any site has less than its ideal value (T : 8, C suggestedby the actual HrO determination. Any discrep- : 5. B : 2. A : 0-1). then a ferric estimate is anciesin the final decimal placesof the mrmbersthat appear either impossible or only possible with additional below and in Appendix 2 Table 2 are due to rounding ef- constraininginformation. This could also indicate an fects.Molecular weights are from Robie et al. (1978). analytical problem. (1) Divide eachwtVovalue (column 1) by the molecular The suitability of the analysis for a ferric esti- weight of the oxide to yield the molecular propor- mation and the normalizations that yield the maxi- tion of eachoxide (column2). Ie.g.,for SiO,: 51.63 mum and minimum estimatesof ferric iron can be + 60.085: 0.859281 determined by calculating the normalization factors (2) Obtain atomic proportions of the cations (column 3) for all the various stoichiometric and chemical lim- and atomic proportions of the O atoms (column 4) its. These are given in Appendix 2 Table 3 and are by multiplying each molecular proportion value by obtained from columns 6 or 7. the number of cations and O atoms in the oxide. If the all the normalization factors (8Si, 16CAI, [e.g.,for SiO,: 0.85928x 1 : 0.85928and 0.85928 and 15 eNK) are greater than all the normalization x 2 : t.lI857l factors (8 SiAl, 15 eK, 10)Fe'*, and 13 eCNK), Nofe.' Assuming 2 (OH) groups are present, 1 O then a minimum and a maximum ferric estimation atom is balanced by 2 H (i.e., H,O) so the cation can be calculated; if not, then no estimation is chargesare balancedby the remaining 23 O atoms possible. that are the basis of the anhydrous amphibole for- (1) Minimum fercic estimate s mula (see text for discussion:it can be shown that, The lowest normalization factor among the four even if F and Cl have not been determined,as long choices, SSi, 16CAT, 15eNK, and all ferrous, de- as OH + F + Cl : 2, the 23 O atoms formula gives termines the formula that yields the minimum fer- the correct mineral formula). ric estimate.If the factors 8Si. 16CAT and 15eNK (3) Obtain the anions based on 23 O atoms (column 5) are all greater than 1.0000, then the all-ferrous by multiplying each value in column 4 by (23 + the formula (Fe3* : 0) is the lower limit. In this ex- sum of column 4) [e.g., 23 + 2.72185 : 8.45012; ample, the 15 eNK normalization factor is the for SiO,: 0.85928 x 8.45012: 14.522081 lowest. (4) Obtain the cations on the basis of 23O atoms (col- To obtain the formula that gives the minimum fer- umn 6) by multiplying each value in column 3 by ric estimate (column 8), multiply the cations from (23 + the sum of column 4) [e.g.,for SiO,: 0.85928 column 6 by the 15 eNK normalization factor, x 8.45012: 7.261) 0.99714(15 + 15.043). Noter Column 6 is the all-ferrous mineral formula (8) Find the sum of O atoms Q2.933T in the normal- for the amphibole. Assigning the cations to sites ized formula by multiplying each single cation value shows if any deviations from ideal stoichiometry (column 8) by the number of balancing O atoms can be explained by failure to account for ferric [e.g., for SiO,,7.24Ol x 2 : 14.4802;for AlO,,, rron. I.22I4 x 1.5 : 1.8321;for MgO, 3.7818 x 1 : (5) Ideal site assignments(column 7) are made from the 3.7818;for NaOo.,0.1659 x 0.5 : 0.08291. cation values in column 6: a general procedure is: (e)Ferric Fe equalsthe amount of ferrous Fe that must LEAKE ET AL.: AMPHIBOLE NOMENCLATURE 1037
be converted to bring the total O atoms up to 23. eNK) in column 10 and maximum (15 eK) in col- The amountis (23 - 22.9331)x 2 : 0.133. umn 11, but is nearer to the minimum. (10) The new fenous Fe vahte is the total Fe from col- RnnnnrxcBs cITED umn 8 minusthe ferric Fe. [e.g.,0.885- 0.133 : o.1s3l Deer, W.A , Howie, R A, and Zussman,J. (1966) An Introduction to the (11) Recast the normalized cations as in step 5 (column Rock-forming Minerals, 528 p Longman, London, -(1992) to the Rock-forming Minerals, 2nd ed, l0). This should yield a formula An Introduction with no violations 696 p Longman, London of the ideal stoichiometry. Droop, G T.R ( i 987) A generalequation for estimatingFer* in fenomag- Nofe: Step 11 double checks the correctnessof nesian silicates and oxides from microprobe analyses,using stoichio- your calculations. It also is a check of whether metric criteria. Mineralogical Magazine, 5I, 431 437 correcting the initial stoichiometric violation will Dyar, M D , Mackwell, S.J, McGuire,A V, Cross,L.R., and Robertson, J D (1993) Crystal chemistry of Fe3* and H* in mantle kaersutite:Im- produce another violation [Here, insufficient cat- piications for mantle metasomatism American Mineralogist, 78, 968- ions to fill T or C could result from the 15 eNK 9'79 normalization. Such analysescannot be used for Holland, T and Blundy, J. (1994) Non-ideal interactionsin calcic amphi- ferric Fe estimates (unfortunately, much of cal- boles and their bearing on amphibole-plagioclasethermometry. Contri- to Mineralogy and Petrology, 116, 433-447. culation is involved in determining this)1. butions Leake, B E (1968) A catalog of analyzedcalciferous and sub-calciferous (12) Maximum ferric estimates amphiboles together with their nomenclatureand associatedminerals, The largest normalization factor among the four 210 p Geological Society of America Special Paper 68. choices, 8SiAl, 15 eK, 13 eCNK, and all ferric, de- Jacobson,C E. (1989) Estimation of Fe3* microprobe analyses:observa- termines the the formula that yields the maximum tions on calcic amphibole and chlorite Journal of Metamorphic Geol- ogy,7, 507-513. ferric estimate.If the factors SSiAl, 15 eK, and 13 Oberti, R., Ungaretti,L, Cannillo,E, and Hawthorne,EC (1992)The eCNK are all less than the all-ferric value, then the behaviour of Ti in amphiboles:I Four- and six-coordinatedTi in ri- all-ferric formula would give the maximum Fe3*.In chterite EuropeanJournal of Mineralogy, 4,425-439. this example, the 15 eK normalization factor is the Robie, RA., Hemingway, B.S, and Fisher, J.R. (1978) Thermodynamic 298 15 K and 1 bar largest and can be gives properties of minerals and related substancesat used to the formula with (105 Pascals)pressure and at higher temperatures,456 p. U S. Geolog- maximum Fe3*. ical Survey Bulletin 1452. To obtain the formula that gives the maximum Robinson, P, Spear,FS , Schumacher,J C., Laird, J, K-lein, C , Evans, ferric estimate (column l1), repeat steps 7 through B W, and Doolan, B.L (1982b) Phaserelations of metamorphic am- 10 using the 15 eK normalization factor, 0.98621 phiboles: natural occurrenceand theory In Mineralogical Society of 1-211 (15 + AmericaReviews in Mineralogy,98, 15.210). Schumacher,J C (1991) Empirical ferric iron corrections:necessity. as- (I3) Average of the maximum and minimum ferric sumptions,md effects on selectedgeothermobarometers Mineralogical esttmates Magazine,55, 3-18 To obtain the formula that gives the average of Schumacher,R (1991) Compositions and phase relations of calcic am- phiboles and clinopyroxene-bearingrocks of the amphibolite the maximum (col- in epidote- and minimum ferric estimates and lower granulite facies,central Massachusetts,U.S A. Contributions umns 10 and 1l), repeatsteps 7 through l0 using to Mineralogy and Petrology, 108, 196-2ll the average of the normalization factors that were Spear,FS and Kimball, C. (1984) RECAMP-A FORTRAN IV program obtainedin steps7 and 12. This normalization factor for estimatingFe3* contents in amphiboles.Computers and Geoscience, is 0.99167 + 098621) + 21. to.3t7-325 l(0.99114 Stout,J.H (1972)Phase petrology and mineral chemistryofcoexisting (14) The actual formula (column 12) given in Deer et al. amphibolesfrom Telemark,Norway Journal of Petrology, 13, 99 (1992) lies approximatelybetween the minimum (15 145