American Mineralogist, Volume 67, pages 90-96, I9E2

A contribution to the crystal chemistry of ellestaditeand the silicate sulfate apatitest

RolnNo C. RousB Departmentof GeologicalSciences University of Michigan Ann Arbor, Michigan 48109

euo PEtp J. DUNN Departmentof Sciences Smiths o nian Inst itution Washington,D. C. 20560

Abstract

A seriesof silicate sulfate apatitesfrom Crestmore,California, which contain the coupled substitutionSilvsvl for 2Pv, has been investigatedusing electron microprobe, powder diffraction, and single-crystal diffraction methods. Chemical analysis of eighteen specimensof differentphosphorus contents proves that the Si:S ratio is essentiallyI : I and yieldsthe idealizedgeneral formula Caro(SiOn):-*(SO4)3-^@O4)2.(OH,F,CD2,where x : 0 to 3. The membersof this seriesfor which x : 0 and 3l2 have beenlabelled "ellestadite" and "wilkeite", respectively, by previous workers. "Ellestadite" is actually a solid solutioninvolving the end-membersCale(SiO a,){SOq)tZz, where Z: OH (hydroxylellesta- dite), F (),or Cl (chlorellestadite).The term ellestaditeis redefinedto make it a group name for all compositions having >(Si,S) > P. Wilkeite is not a valid mineral species,since it is only one of many solid solutions involving the six end-members fluorapatite,hydroxyapatite, chlorapatite, fluorellestadite, hydroxylellestadite, and chlor- ellestadite. Although natural hydroxylellestadite is monoclinic, precession photographs of type "ellestadite" and "wilkeite" show hexagonalsymmetry and no evidenceof Si-S ordering as suggestedby the Si: S ratio of I : I . The silicate sulfate from Crestmore show a strong linear relationshipbetween their P and F contents,such that these two variables simultaneouslygo to zero. Linear relationshipsalso exist betweentheir unit cell parame- ters and their P, F, and (Si+S) contents.These correlations imply a convergenceof the Crestmore series towards a hypothetical member-of composition Caro(SiOa)r (SO4)3(OH,CI)zand cell constantsa:9.543 and c = 6.9174.

Introduction ellestadite,conventionally formulated as Ca5(SiOa, Crystal structures of the apatite type, general SO4,PO4)3(OH,F,Cl)and Ca5(SiO4,SO4)3(OH,F,CI), formula Ato(XO)622, &r€ rot?ble for their ability to respectively.Since these phases are of someindustri- (Pliego-Cuervo accommodatea greatvariety of atomsin their A, X, al and Glasser, 1978; Trivino Yaz- and Z sites by means of simple or coupled ionic quez, 1979)as well as crystal chemical and mineral- substitutions (Dihn and Klement, 19421'Schwarz, ogicalinterest and since very little modern analytical 1967a,b, c; Cockbain, 1968;Ito, 1968;Kreidler and data exists for either the natural or their Hummel, 1970;McConnell, 1973).An example of syntheticanalogues, we have undertakena re-exami- the latter mechanismis the substitutionSirvsvr - nation of ellestadite and wilkeite using type and 2Pv, which occurs in the minerals wilkeite and cotype materialsfrom Crestmore,California. In this report we present new analytical chemical, powder difraction, and single-crystaldiffraction data, which tContribution No. 370,The MineralogicalLaboratory, Depart- bear upon the crystal chemistry of the silicate sulfate mentof GeologicalSciences, The University of Michigan. apatitesand the interpretationsmade of it by previous 0003-0M)vE2l0I 02-0090$02.00 ROUSEAND DUNN: ELLESTADITE investigators.Nomenclatural changes were approved The term "hydroxylellestadite" was introduced prior to publication by the Commission on New by Harada et al. (1971)to denote an impure form of Mineralsand Mineral Names,I.M.A. Cals(SiOa)3(SO4)3(OH)2,which occurs at Doshin- kubo in the Chichibu mining district of Japan.The Historical review new mineral was found to be crystallographically The replacementof a substantialportion (-50%) and chemically similar to ellestadite, except for the of the POaby (SiO4,SO4)in a new mineral of the fact that OH > (F + Cl), and the spacegroup was apatite group was reported by Eakle and Rogers assignedas P63lmon the basisof precessionphoto- (1914) in their description of wilkeite from Crest- graphs.More recently,however, Sudarsanan (1980) more. This was the first example of a coupled has reportednatural hydroxylellestaditeto be mono- atomic substitution in an apatite structure to be clinic-pseudohexagonalwith spacegroup PZ/m and,a : recognizedas such. Additional data on wilkeite 9.476(2),b : 9.508(2),c : 6.919(1)4,and 1 = were given by Rogers (1929), McConnell 093j. 119.53(2).The monocliniccell is not of the sametype 1938),and Taylor (1953),who showedthe supposed as that found in monoclinic phosphate apatites, mineralcrestmoreite to be a submicroscopicinter- sincethe latter is a supercell having b : 2a &nd 1 : growth of wilkeite and the calcium silicate hydrate, 120.0'(Young, 1975).Whether all hydroxylellesta- tobermorite.More recently, wilkeite has been re- dites are monoclinic remains an open question. ported from Timna', Israel (Wtirzberger, 1970), Sudarsanan'sresults were obtainedfrom one crys- Takiue,Japan (Harada et al.,l97l), and the Adiron- tal and it is possiblethat,like hydroxyapatite,some dack Mountains of New York (J. W. Valley, per- crystals are monoclinic and others hexagonal. sonal communication).Apatites containing a few Trivino Yazquez (1979)reported ellestadite as a percentSiO2 and SO3occur at other localities,e.g., constitutent of the incrustations formed in cement the Rhineland (Brauns, 1916) and the Ural and heat-exchangecyclones. He further stated that the Aldan regionsof the Soviet Union (Vasileva,1958), thermaldecomposition of ellestaditein the presence but these are better termed silicatian sulfatian apa- of potassiumyields a "K ellestadite" of probable tites rather than wilkeite. In fact, as we shall later formula Ca1sKSi3S2O22F.The existence of this show, the name wilkeite is itself superfluousand compoundremains unproven as of this writing. shouldbe discarded. Ellestaditewas also first describedfrom Crest- Experimental more by McConnell (1937), who distinguishedit Chemicalanalysis of eighteensilicate sulfate apa- from wilkeite chiefly on the basis of chemical tites including type ellestaditeand cotype wilkeite composition. Whereas type wilkeite had 20.85% are presentedin Table 1. All analyseswere per- P2O5,McConnell found a specimenwith only 3.06% formed on an ARL-seuqelectron microprobe using P2O5and called it ellestadite. No difference in an operating voltage of 15 kV, a beam current of paragenesisor spacegroup symmetry was evident 0.15 pA, and the following standards:hornblende for ellestadite and the mineral is visually indistin- for Si; fluorapatitefor P, Ca, and F; scapolitefor Cl; guishablefrom wilkeite. The crystal chemistry of and celestinefor S. The data were conected using ellestadite,wilkeite, and intermediatecompositions Bence-Albeefactors. The low summationsin Table was further explored by McConnell (1938).Dihn I are due mostly to the fact that CO2 and H2O were and Klement (1942) synthesized the pure com- not determined. Published values for these two pounds Caro(SiO+)r(SO4)3F2and Caro(SiO+): componentsare 2.l0VoCO2 and a trace of H2O in (SO4)3(OH)2,which we shall call fluorellestadite wilkeite (Eakle and Rogers, l9l4) and 0.61% COz and hydroxylellestadite,respectively, along with a and 0.63VoH2O in ellestadite(McConnell, 1937). numberof other exotic apatiteisotypes containing Although McConnell reported minor quantitiesof PO4,SiO4, and SOain variousproportions, some of MgO, MnO, Al2O3 and Fe2O3(<0.5Vo each with which deviated from the ideal A19(XO)622 stoichi- the Fe and Al attributed to admixed ) in ometry. Pliego-Cuervoand Glasser(1978) synthe- ellestadite,our specimensshow tracesof Mn as the sized a third ellestadite-like phase, Cals(SiOa)3 only other detectable element of atomic number (SO4)3C12,which we shall call chlorellestadite.Fur- greater than nine. ther developmentsalong these lines appearin pa- Single-crystal fragments of type ellestadite pers by Schwarz (1967a, b, c), Ito (1968), and (NtrrNn #103072) and cotype wilkeite (NuNH Kreidler and Hummel (1970). #95685)were examinedby the precessionmethod. 92 ROUSE AND DUNN: ELLESTADITE

Table l. Electron microprobe analysesof silicate sulfateapatites NBS (a : 5.43088A)internal standard.The from Crestmore, California d-values represent a considerable improvement over those previously published for ellestadite and NTII{H# si02 Cao 50: PzOS F Cl o=F,Cl Total wilkeite. In addition, unit cell parametersfor five of

Bl72l9 3.8 55.5 4.2 33.1 2.6 0.5 1.2 98.5 the chemically analyzed specimens(but not for the R5256 4.1 55.5 5.r 33.0 2.8 0.5 1.3 99.7 9s68s 9.4 55.2 ll.7 r8.2 1.6 0.8 0.8 96.r samecrystals which had beenanalyzed) were calcu- 94033-3 9.6 55.9 12.1 l8.l t.3 0.8 0.7 97.7 94463 I 0. I 55.8 13.9 15.4 I .3 0.7 0.7 97-5 lated from the powder data using a silicon internal 94033-2 10.8 55.5 14.0 I 5.6 1.2 0.7 0.7 97.1 standard. These were refined by the method of 93417 10.8 54.7 ',14.5I 4.3 I 5.0 I .3 0.7 0.5 96.2 The 94415 11.4'n 55.2 14.2 I .l 0.8 0.5 96.6 least-squaresassuming hexagonal symmetry. c40s0 .I 55.I I 4.9 13.8 1.2 0.7 0.7 96.1 R1874? t1.5 s5.0 14.7 13.4 l.l 0.8 0.6 95.9 results appear in Table 3. 94033-1 12.9 55.2 17.3 9.9 0.8 0.8 0.5 96.4 c4105 t 2.8 54,7 16.5 9.8 0.9 0.8 0.5 95.1 87281 t4.l 55.8 17.1 8.8 0.7 1.0 0.6 97-5 Resultsand discussion R9217-l 15.5 55.7 20.4 4.7 0.4 0.4 0.3 96.8 R92',13 16.0 55.4 21.0 4.3 0.4 0.5 0.3 97.3 Chemicalcomposition R9217-3 16.2 55.3 19.9 3.8 0.3 0.5 0.2 95.8 R92l5 16.2 56.3 20.9 3.8 0,4 0.5 0.3 97.8 formula assigned to ellestadite in I 03072 '17.4iI 5.8 55.7 21.2 3.2 0.3 1.7 0.5 98.4 The chemical Theoryf 54.12 23.18 6.84 t.55 I 00.0 Theoryt+ 17,98 55.90 23.94 3.78 1.60 I 00.0 most referencebooks is Cas(SiO+,SO4)3(OH,F,CI)or Theoryt+f 18.05 56.12 24 04 I 00.0* somevariant thereof. All of them are derived from the *----lncludes 1.79 percent Hro. only published analysis of this mineral by Ellestad +----For end-nenber chlorel iEstadite. ++---For end-menberfluoreilestadite. in McConnell (1937).A new computationof the unit f+t--For end-nember hydrcxylellestadite. NIINH94463 is type wllkelte; Nl4l{H95585 is cotype wilkeite. cell contents using our refined cell parameters and Accuracy of data: i4 percent of the amount present for S. !3 percent of the amount present for other elenents. Ellestad's analysis -recalculated to l00Vo after deducting Ca, Mg, Fe, and Al due to vesuvian- ite yields [Caro.orMgo.06]>ro.o7[(Sior)z.ss(SO+)z.o+ Zero and upper level photographs of both minerals (PO4)0.44(COr)0.ral>o.oz[(OH)s.51Fs.31Cls.a7h1.2e. Our show no deviation from the expected 6lm Friedel partial chemicalanalysis of type ellestadite(Table 1) symmetry, indicating P63 or PQlm is consistentwith this result. In the aboveformula all when the extinctions are taken into account. and no of the carbonhas been allotted to the phosphorussite, evidenceof monoclinic superstructurereflections. McConnell's allotmentof someto the calcium site as The Debye-Scherrer powder data in Table 2 were Ca* being chemicallyunrealistic. Carbonate ion may obtainedwith a 114.6mm diametercamera and an also be situated in the Z (hahde) site according to Young (1975),but this would not alleviate the appar- ent deficit of 0.7 atom in I,(OH,F,CI). The sourceof Table 2. Powder X-ray diffractiofl"Ou,utor type ellestadite and the deficit may be an elror in the determinationof H2O, or the presenceof 02- ions in the Z site as E'llestadite l,|ilkeite Ellestadlte liilkeite proposedby McConnell (1937),or the existence of I dobs dcal dobs dcul hkc I dobs dcal dobs dca'l hk' atomic vacanciesat this site. In fact, a substantial

',l fraction (up to 50%) of the hydroxyl sites may be 2 8.28 8.25 8.26 8.21 100 4 I .855 .854 I .8471.849 213

Sanples used: fiMNH#103072 (ellestadite) and NMNH#95585 (wilkelte) Estimated standard deviations in parenthesesrefer to 114,5 nm Debye-Scherrercanerar Si internal standard, CuKdradiatlon' internal standard. visually estinated intensities, b = broad I ine last digit. NBSsJlicon ROUSE AND DUNN: ELLESTADITE

A mostimportant feature of the ellestaditeformu- contents decreasedand increased, respectively. la is the Si:S ratio, which is 28:26or nearly l:1. The converse would, of course, be true for the Referenceto Figure I shows the same to be true of variation of and with >(Si,S). all analysesin Table I, except perhaps for the Referenceto Table I showsthat thereis, in fact, no silicatiansulfatian apatite Bl72l9.If this specimen correlationbetween and chlorine con- is excluded,the averagedeviation from l: I stoichi- tents. Indeed, the latter element remains nearly ometryis only 4Voand in no casedoes it exceedgV. constant for all specimens analyzed except for Ignoringthe minor CO2content and possibleZ site #103072,which is type ellestadite.The meanchlo- vacancies,the correct structural formula for natural rine content,excluding #103072, is 0.7 wt.Vowith a ellestadite is therefore Cals(SiOa)3(SOa)3rangeof t-0.3 wt.Vo. (OH,F,CI)2,which is in accordwith the tetrahedral There is, however, a strong positive correlation anion ratio in the synthetic phases of Dihn and betweenfluorine and phosphorusas shown in Fig. Klement (1942). Extending this to the whole solid 2. A linear regressionof wt.VoF on wt.Vop vields solution seriesin Table 1, the general formula is the relationship l Cals(SiOa)3-*(SO4)3-*(PO4)2*(OH,F,Cl)2,where x wt.VoF: 0.03+ 0.18wt.VoP r : varies from 0 (ellestadite)to 3 (apatite).The con- 0.994, (0.04) (0.02) stancy of the Si:S ratio at 1:l is, of course, a consequenceof the necessityto maintain overall where r is the correlation coefficient and the num- electrostaticneutrality in the ,j.e., bers in parenthesesare the standarderrors of the the averagevalence of SiOX- and SO?- is the same regressioncoefficients. The two variablesgo to zero as the valenceof POl-. Moreover, the mean bond simultaneously,the apparent residual of 0.03% F lengthsSi-O : l.62Ain silicates(Liebau, 1972)and being statistically insignificant. It is tempting to S-O = 1.47A in sulfates(Wuensch, 197 2) av erageto speculatethat the regressionline in Fig. 2 might 1.544, which is equal to the mean p-O bond become somewhat nonlinear in the region above distancein monophosphates(Liebau, 1970). l0% P curvingupwards to terminateat the composi- The presenceof SiOaand SOagroups in consis- tion point for fluorapatite,which is the theoretical tently equal numbers suggeststhe possibility of limit for fluorine content. In any case, Figure 2 silicon-sulfurordering in the ellestaditestructure if shows that the compositionof the Crestmoresili- not in its more phosphate-richcongeners. If the cate sulfateapatite seriesconverges toward a fluo- spacegroup is really P63or PQlm as indicated by the precessionphotographs, no such ordering can occur since all tetrahedrally coordinated atoms must be situatedin the same6-fold equipoint,6c in P63or 6h in Pglm, to be consistentwith the apatite structure.However, in the structureof monoclinic hydroxylellestadite(Sudarsanan, 1980), the tetrahe- drally coordinated atoms (X) are divided among o20 three 2-fold equipoints and from the variation in X- O distancesthere doesappear to be somedegree of silicon-sulfurordering. Whether the samephenom- enon obtains in ellestadite, wilkeite, any of the intermediatecompositions, or even in all hydroxy- lellestaditescannot be determined from the data presentlyavailable.

EIementalcoruelations Si McConnell (1937)found an unexpectedcorrela- tion betweenthe elementsin the X and Z sitesin the Fig. I. Variation of the atomic proportions of silicon with those silicatesulfate apatites from Crestmore. of . The slope of the regression line is 0.970 Specifical- (correlation coeficient r : ly, as the phosphorus 0.995) indicating that Si:S is contentdecreased toward the essentiallyI : l. W and E are compositionpoints of type wilkeite ellestaditecomposition, the fluorine and chlorine and type ellestadite,respectively. 94 ROUSEAND DUNN: ELLESTADITE

fortunately, these values cannot be compared to FAP those of pure hydroxylellestaditeand chlorellesta- dite, since cell parametersfor the latter compound are unavailable. E The cell constantsof the apatite group are known ie2 = to be affectedby substitutionsin the A (cation),X

I (tetrahedral), and Z (halide) sites. According to McConnell (1973), an increasing chlorine content increasesa but decreasesc, while increasinghy- droxyl increasesboth parameters.LeGeros (1965) r. u-",, olo,*r.r"rffi: withwt.vo'lnorot'o.ur.*, found that increasingthe amount of carbonateion in E,","*. and FAP are the composition points of type wilkeite, type the X site shortensa by about 0.014 for each 1.66 ellestadite,and pure fluorapatite, respectively. wt.VoCO3 but lengthensc only slightly. Among the apatiteisotypes listed in Table 1, chlorine remains rine- and phosphorus-free entity, Cals(SiOa)3 approximately constant and will therefore contrib- (s04)3(oH,cl)2. ute negligiblyto the observedsystematic changes in a andc . The only availablevalues for CO2(0.61% in p Ce ll aramet er variatio ns type ellestaditeand 2.l0Voin type wilkeite) corre- Comparisonof the data in Tables 1 and 3 shows spondto decreasesof 0.005and 0.017,{in a relative somesystematic relationships between the unit cell to the samecompositions without carbonate.Since parametersand chemical composition. Both a and c two CO2determinations are not enough to establish increaselinearly as wt.VoP decreases(Figs. 3 and any systematicchanges, if any exist, in the carbon- 4), althoughthe latter parameterdoes so at a much ate content of the Crestmore silicate sulfate apa- slower rate. Vegard's Law is therefore obeyed by tites, the effect of CO3ion on their cell dimensions this solid solution series.Here a and c have been cannotbe evaluatedat this time. plotted against wt.Vo P for convenience; plots As noted previously,the Si:S ratio is essentially againstwt.% (Si+S) andwt.VoF are alsolinear. The constantat l:1 and the meanof the effectiveradii of regressionequations are SiOaand SOqis equalto the effectiveradius of POa. Thus changesin the (Si,S):P ratio shouldnot affect a : 9.543- 0.0087wt.Vo P r : -0.990 for (0.006) (0.0007) the sizeof the unit cell. Moreover. substitutions calciumin the A site are negligiblein the specimens c : 6.917- 0.0014wt.Vo P r : -0.985 increasingdegree (0.001) (0.0001) analyzed.This leavesonly the of OH for F substitutionin the Z site and a possible, which at DVoP predict parameters a = 9.543 and but unproven, systematic decreasein carbonate c : 6.9174 for the hypothetical series member contentas the principal causesof the increasein a Caro(SiO+)g(SO4)3(OH,Cl)2mentioned earlier. Un- and c as the ellestadite composition is approached.

o gGsoo I HAP MP

9 400 6 a50 FAP

6mo o 5 ol52o *tLt wr%P Fig. 3. Variation of the cell parameter a with wt.Vo Fig. 4. Variation of the cell parameter c with wt.%o phosphorus. FAP and HAP represent pure fluorapatite and phosphorus. FAP and HAP represent pure fluorapatite and hydroxyapatite,respectively. hydroxyapatite,respectively. ROUSE AND DUNN; ELLESTADITE 95

Nomenclaturereyisions also be discardedin favor of fluorellestadite,hy- From the foregoingdiscussion it is apparentthat droxylellestadite,and chlorellestadite.We believe ellestaditeand wilkeite are membersof severalsolid that "ellestadite" still has utility as a group name solution series involving (Si,S,P,C)and (OH,F, for all of the apatite isotypes of composition Cl,O?). Neglectingthe two minor components,the (Ca. . .)s[(Si,S,P,C)Oa]3(OH,F,Cl)when that com- carbonateand oxide ions (the presenceofthe latter positiondoes not approximateany of the pure end- is problematical),the solid solution relationships membersnamed above. The only restrictionis that amongthe calcium silicatesulfate apatites are best >(Si,S) must exceed P in atomic percent. Speci- displayedon a trigonal prismatic compositiondia- menshavlng >(Si,S) < P are best termed silicatian gram. The upper triangular face consists of the sulfatian apatites. three pure end-members Caro(SiO+)r(SOa\22, whereZ: OH (hydroxylellestadite),F (fluorelle- References stadite), or Cl (chlorellestadite),while the lower Brauns,R. (1916)Uber den Apatit aus dem LaacherSeegebiet. triangular face consists of the end-members Sulfatapatitund Carbonatapatit. Neues Jahrbuch fiir Minera- 41, : (hydroxyapatite), logie,Geologie und Palaeontologie,Beilage Band, 60-90. Carc(PO)622,where Z OH F Cockbain,A. G. (1968)The crystal chemistry of the apatites. (fluorapatite),or Cl (chlorapatite). MineralogicalMagazine, 36, 654-660. Since the atomic ratios in type ellestaditeare Dihn, P. and R. Klement (1942) Isomorphe Apatitarten. Zeit- Si:S:P:28:26:4 andOH:F:Cl : 5:3:5,the com- schrift ftir Elektrochemieund AngewandtePhysicalische Che- positionof this mineralfalls just below the approxi- mie,48, 331-333. Eakle,A. S. and A. F. Rogers(1914) Wilkeite, a new mineralof mate center of the ellestaditeface of the composi- the apatite group, and , its alteration product, from tion prism. The tetrahedral atom ratio in type SouthernCalifornia. American Journal of Science,4th Series, wilkeiteis Si:S:P : 16:15:30(F and Cl were not 37,262-267. determinedby Eakle), which places this material Harada, K., K. Nagashima,K. Nakao, and A. Kato (1971) somewherein the medianplane of the prism. Note Hydroxylellestadite,a new apatitefrom Chichibu mine, Sai- tama Prefecture, Japan. American Mineralogist, 56, 1507- that the compositionof wilkeite does not approach l518. that of any of the end-membersdefined above. Ito, J. (1968)Silicate apatitesand oxyapatites. American Miner- Wilkeiteis, in fact, merely one of numerouscompo- alogist,53,890-907. sitions intermediate between the ellestadite and Kreidler,E. R. and F. A. Hummel (1970)The crystalchemistry apatitefaces. There is thereforeno chemicaljustifi- ofapatite: Structure fields offluor- and chlorapatite. American Mineralogist,55, 170-184. cation for regarding this compound as a distinct LeGeros,R. Z. (1965)Effect ofcarbonate on the latticeparame- mineral species.Moreover, the single-crystaland ters of apatite. Science, 206, 403-404. powder diffraction data indicate that there is noth. Liebau,F. (1970)Phosphorus. l5-A Crystalchemistry. In K. H. ing in the known crystallographyof wilkeite to set it Wedepohl,Ed., Handbook of Geochemistry,Vol. lll2, p. apart from other intermediatecompositions in the l5Al-15,4.6.Springer-Verlag, Berlin. Liebau, F. (1972)Silicon. l4-A Crystal chemistry. In K. H. series.Its cell parametersfollow the same trends Wedepohl,Ed., Handbook of Geochemistry,Vol. lll2, p. illustrated in Figures 3 and 4 as those of other 14Al-14432. Springer-Verlag,Berlin. compositions. McConnell,D. (1937)The substitutionof SiOa-and SOa-groups There is one additionaldifficulty with attempting for POa-groupsin the apatite structure; ellestadite, the end to maintain the validity of wilkeite as a distinct member. American Mineralogist, 22, 977-986. McConnell,D. (1938)A structuralinvestigation of the isomor- species.Whereas in the caseof ellestaditethe type phismof the apatitegroup. AmericanMineralogist,23, l-19. materialis well-defined,there exists no exact type McConnell,D. (1973)Apatite, Its Crystal Chemistry,Mineral- specimenof wilkeite. The one designatedas the ogy, Utilization and Biologic Occurrences. Springer-Verlag, type in Table I (Ntr.lNu#94463) is only one speci- New York. men from the type lot. Documentationof the type Pliego-Cuervo,Y. and F. P. Glasser(1978) Phase relations and crystal chemistry of apatite and silicocarnotite solid solutions. statusof #94463is inadequateand the correspon- Cementand ConcreteResearch. 8.519-524. dencebetween our chemicalanalysis of this speci- Rogers, A. F. (1929) from Crestmore near Riverside, men and the analysisin Eakle and Rogers(1914) is California, with a from this locality. American poor. The compositionof nnaNu#95685 is in better Mineralogist, 14, 462-469. agreementwith Eakle's analysis,but the connec- Schwarz, H. (1967a) Verbindungen mit Apatitstruktur. I. UngewdhnlicheSilikatapatite. Inorganic and NuclearChemis- tion betweenthis specimenand his is very tenuous. try Letters, 3, 231-236. In view of the foregoing arguments,it may be Schwarz,H. (1967b)Verbindungen mit Apatitstruktur.II. Apa- asked whether the term "ellestadite" should not tite des Typs PbeK+(XvO4)4(XvIOo)2(Xv : P,As; Xvr : 96 ROUSE AND DUNN: ELLESTADITE

S,Se). Zeitschrift fiir Anorganische und Allgemeine Chemie, Wuensch,B. J. (1972)Sulfur. 16-,{Crystal chemistry.In K. H. 356,29-35. Wedepohl,Ed., Handbook of Geochemistry,Vol. lll2, p. Schwarz,H. (1967c)Verbindungen mit Apatitstruktur.IIL Apa- 16Al-l6.Al9. Springer-Verlag,Berlin. tite des Typs MlfiXvto4)3(Xrvo4)3F2(Mrr : Sr,pb; Xvt : Wiirzbuiger,U. (1970)Copper silicates in the Timna' ore depos- S,Cr; Xrv = Si,Ge). Zeitschriftfiir Anorganischeund Allge- it. Israel Journal of Chemistry, 8, 443-457. meineChemie. 356. 36-45. Young,R. A. (1975)Some aspects of crystalstructural modeling Sudarsanan,K. (1980)Structure of hydroxylellestadite.Acta in biological apatites. In Physico-Chemieet Cristallographie Crystallographica,836, 1636-1639. des Apatites d'Int6r€t Biologique,Colloques Internationaux Taylor, H. F. W. (1953)Crestmoreite and riversideite.Mineral- CNRS, No. 230, p. 2l-40. Centre National de la Recherche ogicalMagazine, 30, 155-165. Scientifique,Paris. Trivino Vazquez, F. (1979)Identification of the incrustations in heat-recovery cyclones of the oopol kiln. Materiales de Construccion(Madrid), 176,93-101 (not seen;extracted from Chemical Abstracts, 93, 244300). Vasileva,Z. V. (1958)Sulfur-bearing apatites. Geokhimiya, 368- Manuscriptreceived, May 20, 1981; 373 (transl. Geochemistry, 464-470, 1958). acceptedfor publication,September 23, 196l,.