Canadian Mineralogist Vol. 19, pp. 381-387(1981)

GORMANITE,Fe2*,Al*(POo)r(OH)u.2H,O, THE FERROUS ANALOGUE OF SOUZALITE,AND NEW DATA FOR SOUZALITE

B.D. STURMAN eNp J.A. MANDARINO Departmentof Mineralogy & Geology,Royal Ontario Museunr, 100 Queen'sPark, Toronto, Ontario MSS 2C6

M.E. MROSE U.S. GeologicalSurvey, National Center,Stop 959, Reston,Virginia 22092,U.S.A.

P.J. DUNN Deportnrciltol Mineral Sciences,Smithsonian Institution, llashington, D.C. 20560,US.A.

ABSTRACT nent a la composition (Mg1.rsFe2*r.rrMno.or)"r.oo (Alr.s:Fe3+o.zr)"o.or(POn)r.r.(OH)e.gs'1.82 H2O. Par Souzalite,ideally Mg3Al4(PO4) {( OH ) 6.2H,O, and contre, nous obtenons (Fez+r.soMgr.r6Ca6.6aMn6.ns) , +o.ru) ideally Fe2+rAl4(PO)4(OH)s.2H2O, are s2.6s( Al3.s2Fe3 POn)r.*( OH )..rr. I .69 HnO isostructural and form a solid-solution series. New pour la gormanite"o.,r( du Nord-Est du Yukon. Les analytical data for souzalite from the Corrego cristaux de gormaniten en lames allong6es, rem- Frio pegmatite, Minas Gerais, Brazil, yield the plissent des fissures dans une formation de roches composition (Mg1.7sFe!+1.1sMns.d)1"...(Alr."rFea+ferrugineuses. Duret6 4 ?r 5, plan de s6paration ,.r.) POr)g.ru(OH ),.31. 1.82HrO, whereas the (calc.), "n.or( [001], clivage {001} indistinct, densitd 3.13 composition of analyzed gormanite from the north- 3.12 (mes.). La gormanite est bleu-vert et donne eastern Yukon Territory is (Fes+1.s6Mg1.2rCa...nune rayure vert pile; optiquement, l'6chantillon Mnn.m) 12.sx(Alg.82Fe3+s.36) la.tr(PO, ) n.oo(OH ) 1.69 analys€ est biaxe n6gatif, 2V 53o (mes.); a 1.619 ".r2. HrO. Cormanite occurs as elongate blades in the (3), p 1.653(3), z 1.660(3); pl6ochroique,absorp- fractures of an formation. Gormanite crvstals tion X - Z 1 Y; X incolore, Y bleu, Z incr.lore, have a hardness of 4-5, parting along planes Z L b (direction d'allongement) = 14o. La gor- {001}, poor {001} , density 3.1il gZcms manite est triclinique, Pl ou PI, a 11.79(l), b (calc.) and 3.12 g/cms (meas.). The mineral is 5.ll(l), c 13.61(1)A, o 90050(5)',p 99o00(5)', blue-green and has a pale green . Analyzed r90o05(5)'; Z - 2. trs si6 raies les plus intenses gormanite is biaxial, optically negative, with- 2V du clich6 de poudre fd en \(I)(hkl)l sont:4.?61 530 (meas.);a 1.619(3),p 1.653(3),r 1.660(3); (60)(0ll), 3.3e5.(100)(013), 3.r54 (60)(113, - absorption: X Z ( Y; pleochroism: Xcolorless, 302), 3.062(40)(213), 2.925(80)(311) et 2.5s4 Y blue, Z colorless; Zltb (eloneation) - 14o. (e0)(020,4r r ). Gormanite is triclinig, pl or p7, a 11.79(l), b (Iraduit par la Redaction) 5.11(l), c 13.61(1)A, a 90050(5)',p 99"00(5),, - z 90005(5)'; Z 2. fhe strongestsix lines in the Mots-clds: gormanite,souzalite, Yukon, phosphate, powder pattern td in A o&k0l are: 4.261(60) min6ral nouveau. (0rr), 3:395(100)(013), 3.154(60)(113,302), 3.062(40)(2r3), 2.92s(80)(3ll) and 2.554(90) (020,4rD. INrnopucnoN

Keywords: gormanite, souzalite, Yukon, phosphate, Several new minerals have been described new mineral. from a phosphate-ironstone occurrence in the northeastern corner of the Yukon Territory. These are kulanite (Mandarino & Sturman So\ar'rlrns 1976), baricite (Sturman & Mandarino L976), La souzalite MgsAlr(POa)a(OH1u.211,9 ,u penikisite (Mandarino et al. 1977), gormanite, son analogue ferreux, sont isostructurales"t (Sturman et al. 1977) and (Manda- et forment une solution solide. Les nouvelles don- rino et al. 1978). One of the first phosphate n6es analytiques obtenues sur la souzalite de la peg- minerals examined from this occurrence is a matite Corrego Frio (Minas Gerais, Br6sil) md- green radiating speciesthat is common in many 381 382 THE CANADIAN MINERALOGIST outcrops in the deposit, but whose identification name gormanite-souzalite for crystals that have proved to be extremely difficult. A semiquanti- not been analyzed. tative X-ray-fluorescence analysis and micro- scopic study indicated that although the optical OccunnBNce properties and composition are similar to those of souzalite, the X-ray powder-diffraction data Gormanite-souzalite crystals are very com- differ completely from those published for sou- mon in both the Rapid Creek and Big Fish zalite. areas in the northeastern part of the Yukon * Souzalite, ( Mg,Fe"* ) r(Al,Fee ) a( POa)a(OH ).' Territory. The mineral is seen in thin sections 2HzO, was describedby Pecora & Fahey (1949) of the phosphatic ironstone beds as green, from the Corrego Frio pegmatite, north of Di- elongate crystals in radial aggregates, and in vino, Minas Gerais, Brazil. One of us (M.E.M.) open fractures as elongate, blade-like crystals. discovered that the X-ray powder-diffraction In thin section, the mineral is similar to some data given in the description are insorrect. We of the chlorites, for which it can be easily also had trouble reconciling the crystallographic mistaken. data for souzalitegiven by Moore (1970) with In many fractures, gormanite-souzalite crys- those found in this study. Consequently,we have tals constitute the only phosphate mineral and established new powder data and determined are generally accompdnied by and sider- new optical and crystallographic data for sou- ite. Elsewhere, gormanite-souzalite has been zalite from Brazil for the crystals from the type found in close associationwith ludlamite, ar- specimen,NMNH C5863 and ROM M34O|O rojadite, kryzhanovskiteand oxidized vivianite' (NMNH refers to the National Museum of In places, needles of gormanite-souzalite are Natural History, Smithsonian Institution, Wash- included in quartz, giving the quartz crystals a ington, D.C., and ROM, to the Royal Ontario green color. Phosphate minerals from the iron Museum, Toronto). These new data for sou- deposit in the northeastern Yukon, and the de- posit briefly described in the paper zalite from Brazil are in very good agreement'the itself, are with those observed for crystals from on kulanite by Mandarino & Sturman (1976). Yukon. The main difference is the great varia- The absence of any sign of metamorphism in tion in the Mg:Fe ratio that was found in the surrounding rocks indicates that these have some crystals from the Yukon. Ideally, souzalite crystallized at a low temperature. is MgsAla(POn)n(OH)u'2HzO,but many Yukon crystals have Feo* predominanl. This new min- CnvsterlocnePHv eral, the ferrous analogue of souzalite, has an ideal composition Fe2+rAlo(POn)n(OH) 6'2HnO; The largest gormanite crystals observed it is named gormanite (GAWR.MENAIT) in (ROM M35124) are elongate blades in frac- honor of Professor D.H. Gorman, who has tures in the Big Fish area. Crystals are as much inspired interest in minerals and taught min- as 3 mm long, 0.5 mm wide and 0.1 mm thick. eralogy to several generations of mineralogists Crystals from other specimensare smaller and and geologists at the University of Toronto. form radial aggregatesin which the individual The name and the mineral were approved by the Commissionon New Minerals and Mineral TABLEi. AIIGLETABTE FOR GORI.TANITE (M35123)

Names, I.M.A. Type material is deposited in UNITCELL: Spacegroup PL or PIt z ' 2 droo 'Museumthe mineral collections of the Royal Ontario a = 11.79(t) = n.oa(r) doro (ROM M35123 and M35124) and the D = 5.11(1) = s.n(r) door National Museum of Natural History (NMNH c = 13.6i(r) = rs.qe(r) 137494and 137495). c = 9oo5o(5), oi = gsors(s)' Many crystalsin the specimensfrom the Yukon I = 99000(5)', B* = 81000(5)' show strong chemical zoning wherein the com- y = 9ooo5(5)' yr = 89050(5)' position changesfrom gormanite (Fe'* predom- C00RDINTCTES: 9e inant) to souzalite (Mg predominant) in a (100) ggo47' goooo' single crystal. Only a microprobe analysis shows (010) oooo, goooo' goo2' whether gormanite or souzalite zones predom- Foms (001) a5oog' inate. We could not determine any differences (102) ea9qo' 36036' ( -Bao28, 23004' in appearance (color, , erc.) be- 102) .-^o .0 x loJ tween souzalite and gormanite in these speci- Principal the specimens vibratlon Y -77r: 86!0 mens. Therefore, in describing dlrections from the Yukon, we recommend using the Z tzr"t 840 GORMANITE, THE FERROUS ANALOGUE OF SOUZALITE 383

x odi ).(

Ftc. l. Stereographic projections (with the b axis vertical) of the twinned lamellae in a gormanite crystal. crystals are always bladed. axis. Figure I is a stereographicprojection of a The bladesare elongate[010], and the largest twinned crystal of gormanite with the b axis face (plane of the blade) is {001}. Other foims placed vertically to illustrate more clearly the observed by optical goniometer and universal relationship between the individuals. The com- stage are small faces {100} and minute faces position plane is commonly irregular but is {102}, {I02} and {010}. The calculated@ and p generally parallel to {@1}. values for these forms are given in Table l. The gormanite crystal used in this study Pecora & Fahey (1949) described souzalite was first examined on the optical goniometer as probably monoclinic because of its optical and then was studied by means of the polarizing properties. After a Weissenberg study of sou- microscope, using a spindle stage modified to zalite, Moore ( l97O) proposed a monoclinic allow a goniometer head to be mounted. The cell that has the b axis (twofold axis) parallel optical properties show that the mineral is not to the elongation. Our study shows that all monoclinic, with the D axis (two fold axis) crystals of souzalite from Brazil and gormanite- parallel to the elongation, becausenone of the souzalite from the Yukon are triclinic and are principal vibration directions parallels the axis polysyntheticallytwinned with [010] as the rwin of elongation. Furthermore, examination of

TABLE2, UNIT-CELLPARAMETERS OF GORJ4ANITEAI{D SOUZALITE

Gomanlte Souzal ite YukonTerrltofy. Canada Mlnas Gerais, Brazll This study li|oore m5r24 M35123 M34010 ( 1970) Slngle Reflned from Reflned frm Reflnedfrm Slngle crystal porder data porder data poilCer data crJstal 11.76(l)l 11.77(1)8 u.7e(1)8 u.74(t)R re.seI 5.10(I ) 5.11(r ) 5.11(r) 5.11(1) 5.10 13.57(1) 13.57(1 ) rJ.ortU 13.58(1) 13.48 gooco(s), soo+s1s1, gooso(s), sooss(s)' B ggoro(s)' sgors(s), ssooo(s)' ggoos(s)' 1t:. oo gooro(s), ooooslsy, sooos(s), goozo(s), go:.+t i3 sos.4sR3 eog.rg13 eoc.srA3 796.10 Space group Pl or --:- Azln or Az "I z 2 384 THE CANADIAN MINERALOGIST other crystals on the universal stage showed tlte X-Rev Pownrn-DrrFRAcrIoN DATA twin axis to be parallel to the elongation, another indication that the elongation axis can- The X-ray powder-diffraction data for sou- not be tlte 6 axis of the monoclinic cell. Finally, zalite given by Pecora & Fahey (1949) were a single-crystalX-ray-diffraction study, by Weis- incorrectly calculated. The powder pattern was senberg and precession methods, proved that prepared with Fe radiation, but the d values the mineral is triclinic, although the twinned were calculated as if Cu radiation had been crystals show pseudomonoclinic symmetry. used. The d values for souzalite given in the The triclinic unit cell given in Table 2 reflects right column of Table 3 were recalculated from the pseudosymmetry of the twinned crystals. the data reported by Pecora & Fahey (1949). The spacegroup is either Pl or PI. The mono- These powder data are comparable to our clinic unit cell of Moore (1970) can be derived Guinier powder data (Table 3) for Brazilian from the triclinic cell in the following way: souzalite. Guinier powder data for gormanite *6*r"; -q*61o; D-oo = A'!-oo : c*aon : [T02]r.t" from the Yukon (ROM M35124) also are given (seeFig. 1). in Table 3. The powder patterns of gormanite A search for an untwinned crystal of either and sonzaliteare similar; small differencesin d souzalite or gormanite suitable for single-crystal X-ray study proved fruitless, even though many crystals from several specimensfrom the Yukon 3. X-RAYPOI{DER D1FFRACT1ON DATA FOR SOUZALITE AND GORXANITE were examined under the microscope in immer- TABLE sion liquids and, later, on the precessioncamera. Gomanlte,Yukon Souzalite, Brazi] Finally, a small crystal fragment composed of ffi5123 1i134010 one large individual and a small one adhering Recalculated Gulnlercmra Gulnler cMera fr@ PecoraE to it was chosen for study. Because of the thls study (Fahey(1949) dca1c. dobs. dobs. greatly differing sizes of the individuals, the t dobs. wx I t respective diffraction spots were readily dis- 6.72 6.72 'OO2 30 6.72 S 6.72 tinguishable. This crystal was taken from ROM 4.79 4.80 0I1 5 4.74 specimen M35124; it was impossible to separate 4.79 202 4.76r 4.754 'lll 60 4.760 s 4.76 sufficient pure material from this specimen for I 4.495 4.495 'I1l 1 4.490 F 4.48 4.094 4.095 '202 5 4.080 !l 4.08 a full chemical analysis;only a microprobe anal- 1 3.802 3.803 21r r 3,795 3.794 zrl ysis was done. A complete shemical analysis of 10 3.614 3.6L2 '2lr l0 3.614 gormanite was made of material from ROM t0 3.580 3.58r r211 10 lil 3.584 100 3.394 r0I3 100 3.386 v5 3.381 specimen M35123, which is slightly different I I 10 30 3.341 chemically and crystallographically from that d60 3.154 3.161 1I3 50 ( 3.148 in ROM M35124. Chemical data for crystals 20 3.131 *204 10 , l r.ro, given 20 3.117 3.118 *113 x0 3.104 from these two specimens are in Table 30 3.093 r213 20 3.086 40 3.062 3.063r213 40 3.060 \ 3.061 5, and their unit-cell parameters in Table 2. '3Lr 2.9M 2.944 d80 2.92r 14 2.929 Although we were unable to find a suitable 80 2.925 2.924 '3L1 I 2.838 2.s38 40? 1 2.836 F 2.442 crystal for the single-crystal study of the type d90 2.554 2.5s5 r020 d90 2.547 2.551 4rr souzalite from Brazil. we conclude that it also is 2.417 2.400 ll 2.380 for (l) l0 2,362 10 2.359 triclinic, the following reasons: Many 10 2.348 l0 I,I 2,343 crystals from the Yukon localities show variable F d1 2.W8 2.092 compositional zoning from gormanite to sou- 10 2.051 d10 2.045 I'Si 2.047 t0 2.0u zalite; souzalite and gonnanite are isostructural 5 2.023 5 2.024 I L.976 1 1.976 1.978 and form a solid-solution series. The triclinic dto 1.925 d10 r.924 W !.922 unit cell determined for gormanite is certainly d10 1.901 d10 1.900 ! 1.899 10 1.820 5 1.820 !l 1.819 valid for the souzalite zones on the crystals 5 1.805 5 l.8M 5 1.783 1 I.782 F 1,742 from the Yukon. (2) Souzalite from Brazil and L !.732 I r,732 F !.729 1 r.696 1 1AO2 F r.694 gormanite-souzalite crystals from the Yukon 15 1.660 20 1.658 Ill 1.659 have practically powder patterns; l0 1.545 10 1.543 I'l 1.543 identical small 10 1.528 10 r.529 fiI differences in d values represent small differ- I 1.491 I 1.491 ll 1.486 l0 r.471 10 1.470 ences in unit-cell dimensions caused by small 5 1.281 5 1.274 variations in chemical composition. (3) Sou- Intensltjes were estlmated by eye: d-double llne. CuKs radlatlon zalite from Brazil and gormanite-souzalite was used with Gulnier c@eras, and Fe-radlation and Debye-Schefrer canera were used by pecora & Fahey (1949). Intens{ty of the line crystals from the Yukon show twinning and 3.346 and 3.341 probably depends on the ilount of the adf,lxed quartz. have optical propertiesindicating a triclinic unit {Lines used for the least-squares reflnoent of the unit cell. Three of these (6.72, 4.495 and 3.361) were glven lower weightlngs than cell. other llnes ln the refjnment progrme. CORMANITE,'THE FERROUS ANALOGUE OF SOUZALITE 385 values can be observed only when the two sition (Table 5). Crystals from specimen Guinier patterns are placed side by side. The M35123 showed only small variation in chemi- small differences in the intensities of the lines cal composition, whereas strong zoning in crys- in the two patterns are within the range of tals from M35124 indicates considerable varia- experimental error. Many spacingscould be mul- tion in composition. Table 5 gives the average tiply indexed; single-crystal films were used to composition f.or M35124 that was used to cal- determine the appropriate Miller indices. culate the density of the whole crystal. The value of 2V and the orientation of the Pnystcel eNp Oprrcel Pnornntrcs indicatrix, crystal faces and composition plane were determined by means of the universal Gormanite and souzalite are blue-greenoand stage. Becauseof small grain-size and twinning, the streak is pale green. Changes in the color neither the spindle stage nor the methods of of zoned crystals are not related to the varia- Sturman (1973) could be used to determine tion in chemical composition. The is the indices of refraction. Consequently, a was vitreous, and the mineral is nonfluorescent un- determined as the smallest observed index of der short- and long-wave ultraviolet light. A many grains in immersion mounts, and 7 as the hardness of 4 to 5 was determined on a sawn largest index. The value of B was measured surface of a gormanite-souzalite aggregate.This directly from crystals lying on blade faces be- is less than the range of. 51/z-6 given for sou- cause Y lies almost in that plane. The previ- zalite from Brazil by Pecora & Fahey 1949) ously determined orientation of the indicatrix and is probably lower owing to the breaking and pleochroism were useful in choosing suitably of individual blades in the aggregate. oriented grains. The complete optical data for Poor cleavageparallel to {001} was observed gormanite M35123 are given in Table 4, where in a few grains in thin sections of the gor- they are compared with the optical data for manite-souzalite crystals from the Yukon. Pe- Brazilian souzalitedetermined in this study and cora & Fahey (1949) reported two cleavages by Pecora & Fahey (1949). in souzalite, a good one at about 90o to the This method probably gave reasonably ac- poor one. In the crystals from the Yukon and curate optical data for crystals from M35123 Brazilo we could find only the one poor cleavage (Yukon) and M340lO (Brazil) that show only parallel to {001}. When souzalite or gormanite weak zoning. On the other hand, the small, crystals are crushed, however, they break along irregular size of the zones and their variation the composition plane of the twins, thus giving in chemical composition and twinning made the false impression of a good cleavageparallel correlatiori between optical data and composi- to the composition plane {001}. This may ex- tion of the zones irnpossible for material from plain the good cleavage reported by Pecora & M35124; they are not included in Table 4. Fahey (1949), but we cannot account for the Dispersion of the optic axes is very strong, poor cleavage at 90o to this. with r ) v. Principal vibration direction Z for Pecora & Fahey (1949) gave a measured Na-light makes an angle of about 14o with the density of 3.09 g/cma for souzalite from Brazil, elongation in gormanite M35123. Instead of and we have calculated a value of 3.07 g/cm8, showingan extinction angle of 14o on the blade using their chemical analysis and our refined face, this extinction is replaced by the change unit-cell dimensions. Determination of the den- in color from reddish to blue in such a way as sity of analyzed gormanite by means of the to indicate that Zltb is much smaller for red Berman microbalance gave values from 2.98 to light than for violet lighl; Z,lrb < Z,Lb. 3.02 g/cm', but examination of thin sections Gladstone-Dale calculations were made for of gormanite aggregatesshowed that fragments both gormanite and souzalite to determine the weighing only 5 to 10 mg contain admixed compatibility of the data as outlined by Man- quartz, and voids are present between grains. darino (1976, 1979). The data used for gor' To avoid the problem caused by impurities manite are the chemical analysis of specimen and voids, we crushed the samples; the density M35123. the mean index of refraction 0.644) of small grains of pure material was determined and the calculated density (3.12 e/ims). For by immersion in calibrated heavy liquids. The souzalite, the chemical analysis given by Pecora values thus obtained are 3.10(3) g/cme for & Fahey (1949) was used, with the mean index M35124 and 3.13(2) g/cms for M35123. T'he (1.637) determinedin this study and the den- density calculated f.or M35124 is 3.10 g/cm9, sity calculatedin this study (3.O7 g/cm'). The and for M35123 is 3.12 g/cmsnusing the refined results for gormanite are Kc - 0.212 and Kr unit-cell data (Table 2) and chemical compo- = 0,206. which give a value for 1-(K./K.) of 386 THE CANADIAN MINERALOGIST

TABTE4. OPTICALPROPERTIES OF GORI4ANITEAND SOUZALITE

Gomnlte Souzallte YukonTemltory, Canada l,linasGerals, Brazll t'{35123 I'134010 Pecora & Fahey Thls study (1e4e)

1.6re (3) 1.617(3) 1.618 1.653(3) i.642 (3) L.642 T 1,660(3) 1.653(3) L.652 zvx(calc. ) oo 680 zvx(meas.) $(2)0 Dlsperslon pv, symetrlc Pleochroi s x colourless colourless green Y blue blue blue z colourlesr colour'less-paleyellm yel l0 Absorpt'lon X-Z

TABLE5. RESULTSOF AI1ALYSESOF GORI'IANITEND SOUZALITE 0.028, indicating excellent compatibility of the data. The resultsfor souzaliteare Kc - O.2ll, Gomanite souzallte - = 0.019, which YukonTerfitory, canada I'ilnasCerals, K, 0.207 and l-(Kp/Kc) Brazi I indicates superior gompatibility of the data. M35X23 M35L24 /rt electron /r\ electrcn /21 (3) chmica]\" n'lcroprober-' microprobe'-' chml cal CHevrcel CovPosttlott

MSo 6.65 6.9 7.4 9.62 A chenrical analysis was made of material Cao 0,26 0.02 from ROM specin'tenM35123, and electron- Mno 0,31 0.4 0.4 0.31 microprobeanalyses of materialfrom this speci- ro.o(a) 11.49 FeO 14.68 i.q?l men and from ROM specimenM35124 (Table A1203 25'5r 26,7 27.0 26.O7 2.65 5). The chemical analysisof M35123 has been Fer0, 3.82 present Pz0s 37.23 38.4 38.5 37.70 corrected for 7.0 wt, Vo SiOz,which was HzO 11.45 n.d. L2.04 as admixed qtlartz. Magnesium, calcium and Sn0 0.04 aluminum were determined by atomic absorp- Tl02 r 0.07 tion, manganese,total iron and phosphorusby total 99.91 100.01 X-ray fluorescence. ferrous and ferric iron by Nmber of lons titration with potassium permanganate, and (5) (6) (6) (5) HrO by thermogravimetric analysis. The other Ms r.26 1.27 1.42 1.78 samples were analyzed by means of an ARL- Ca 0.04 0.00 SEMQ clectron microprobe, using an operating Mn 0.03 0.04 0.04 U. UJ current of 0.15 I to voltage of 15 kV and a beam FeZ' 1.56 1.69 examination 3.90 3.42 rr,A. The samples for microprobe A1 3.42 3.88 examined in Fe"' 0.36 0.11 0.11 0.25 were mounted in epoxy and P 4,00 4.01 toa polished thin sections. Unmounted samples also H 9.69 ooi 10.03 9.97 were examined to check on possibledehydration 0 24.00 23.94 24.01 24.00 during the sample preparation. No evidence of (1) Analyst: Dr. E. J. Brooker, X-ray AssayLaboratorles, dehydration was observed.Microprobe standards Toronto (see tqt for details) used were montgomeryite for Al and P, and (2t nnaiyit,'pete J. Dunniaccura;y of data I 3%of the ilount Dresenr arrojadite for Mn, Ca, Mg and Fe. Composi- (3) Peora & Fahey(1949) (4) Total Fe expressed as Feo tions were verified using different standards for (5) calculated on the basls of 0 = 24 not given here' The data (6) of Mgr Mn+ Fe i Al + P = 11. additional analyses Calculatedon the basis fluo- Fe partltioned betweenFe2+ and Fe3+ to make were corrected for absorption" backscatter, Mgr liln+ Fe2* - 3. H calculated to glve 2HrOand rescence and background effects using several enough0H to balancecharges. different computer programs, with similar re- GORMANITE. THE FERROUS ANALOGUE OF SOUZALITE 387 sults. The analytical results given in Table 5 mic peak at 850'C. The TGA showed weight were obtained using the standard Bence-Albee loss beginning at 39O'C and continuing to correction factors. Wavelength scans on the mi- 700"c. croprobe failed to indicate the presence of other cations. AcKNowLEDcEMENTS The formula derived from the chemical anal- ysis of M35 123 is ( Fea*r.uoMgr.zoCoo.orMno.or) The authors are grateful to the following (AL.erFeuno.au) o.oo(OH),.rr' L69 HrO."r.r, staff members at the Royal Ontario Museum for From the electron-microprobe"n.tr(PO.) analysis of their help: Mr. Donald McKinnon, for DTA- M35123, the cation portion of the formula may TGA analyses, Mrs. Cynthia Peat, for prepara- be written (Fe2*r.ueMgt.rrMno.oa)rg.oo(Alr.rrFe'*tion of X-ray powder-diffraction patterns, and o.,t)"r.n, if Fe is partitioned between the two Mrs. Gloria Nelson, for typing the manuscript. valence states to make the summation of the divalent ions equal to three and that of the tri- RrrEneNcr,s valent ions equal to four. Similarly, the cation portion of the formula derived from the elec- MeNoanrNo, J.A. (1976): The Gladstone*Dale tron-microprobe analysis of M35124 may be relationship. I. Derivation of new constants. Can. written (Fes*t.lrMgt.nlMno.on) Mineral. 14, 498-502. 'We "r.*(Alg."oFe"*o.r) have derived the following formula (1979): "0.*. The Gladstone-Dale relationship. from the souzalite chemical analysis given by IIL Some general applications. Can. Mineral. 17, Pecora & Fahey (1949) based on 24 oxygen 7l-76. atoms: ( Mgt.r*Fe'*t.teMno.oa) >s.oo (Al".rdFet*o.uu) B.D. (1976): Kulanite, a new It & S'runrueN, ro.o'(POa)a.ge(OH)u."'1.82HrO. is evidentthat iron phosphate from Yukon general barium aluminum the formula for the souzalite-gormanite Territory, Canada. Can. Mineral. 74' 127-131. seriesis R'*rRt*o(POn)o(OH)r.2H"0 where R2* is mainly Mg (souzalite) or Feu* (gormanite), & ConrErr, M.I. (1977): Peniki- and R8* is mainly Al with someFe3*. site, the magnesium analogue of kulanite, from In the electron-microprobe analytical data for Yukon Territory. Can. Mineral. 15, 393-395. M35123, the results from several crystals show & - (1978): Satterlyite,a new little variation in composition. On the otier hydroxyl-bearing ferrous phosphate from the Big hand, M35124 shows variation in the Mg:Fe Fish River area. Yukon Territory. Can. Mineral, ratio from 0.60 to 1.30. Thus, in this speci- 16, 411-413. men, the composition varies from ferroan sou- MoonE, P.B. (1970): Crystal chemistry of the basic zalite to magnesian gormanite. No relation was iron phosphates.Amer. Mineral. 55, 135-170. observedbetween this compositional zoning and crystallographic orientation, although many PEcoRA,W.T. & FeHEv, l.J. (1949): The Corrego crystals were examined. Density maps of cation Frio pegmatite, Minas Gerais: scorzalite and sou- zalite, two new phosphate minerals. Amer, Min- distribution for Mg and Fe confirmed the ran- aral. 84, 83-93. domness of the chemical inhomogeneity. The zones are small, and no attempt was made to SrunueN, B.D. (1973): Determination of the prin- determine the indices of refraction or other cipal refractive indices of biaxial minerals from physical properties of the analyzed zones. The any randomly oriented grain. Can. Mineral, 12, analysis of. M35124 in Table 5 is an averageof 147-148 (abstr.). more than 25 sample points on many crystals. & MeNoemNo, J.A. (1976): Bai€ite, the magnesium analogue of vivianite, from Yukon Tuenvrer PnopentrEs Territory, Canada. Can. Mineral. 14, 443-406. -, Conr-rrr, MI. (1977): Maric/ite, A 350 mg sample of gormanite (M35123) & a sodium iron phosphate from the Big Fish was subjected to simultaneous differential ther- River area, Yukon Territory, Canada. Can. Min- mal and thermogravimetric analysesin a Mettler eral. 15, 396-398. Thermal-Analyzer using a nitrogen atmosphere and heating rate of 8'C/min. The DTA showed a strong endothermic peak at 510'C, a broad Received February 1981, revised manuscript ac- exothermic peak at 720"C and a small exother- cepted Mat 1981.