sign in sign up
<<
Home , Alleghanyite, Humite, Jerrygibbsite, Leucophoenicite, Sonolite

American Mineralogist, Volume 69, pages 546-552, 1984

Jerrygibbsite, a new polymorph of Mne(SiOa)c(OH)zfrom Franklin, New Jersey, with new data on

Prrr, J. DUNN

Department of Sciences Smithsonian Institution, Washington, D.C. 20560

DoNer-o R. Peacon Department of Geological Sciences University of Michigan, Ann Arbor, Michigan 48109

Wrlrreu B. SruuoNs Department of Earth Sciences University of New Orleans, New Orleans, Louisiana 70148

aNo Enrc J. Essexe Department of Geological Sciences University of Michigan, Ann Arbor, Michigan 48109

Abstract

Jerrygibbsite, ideally Mnq(SiO+)+(OH)2,is polymorphous with the Mn- , and a probable member of the leucophoenicitegroup. It occurs as intergrown grains with a typical metamorphic texture at Franklin, New Jersey, intimately associatedwith leuco- phoenicite. It is orthorhombic with spacegroup Pbnm or Pbn21and unit cell parametersa : 4.85(1),b : 10.70(l),and c = 28.17(3)4.The strongestlines in the powder diffraction pattern ared (I): 2.557(lN), 1.806(100),2.869(78),2.752(49),2.702(46),2.362 (39). Optical parametersinclude: biaxial negative,2V = 72", a = 1.77?(4),F = 1.7$G), y: 1.789(4);Z : a, X : b, amd Y: c. The color is violet-pink;the Mohs' hardnessis approximately5.5; the observed density is 4.00(2) and the calculated density is 4.045 g/cm3.The name is in honor of ProfessorGerald V. Gibbs of Virginia Polytechnic Institute and State University. Chemical analytical data for leucophoeniciteindicate that minor amounts of Ca andlor Zn may sel'veto stablilize leucophoenicite-groupminerals relative to humite-groupminerals at Franklin, New Jersey. Introduction We take great pleasure in naming this mineral jerry- During a systematic survey of the chemical composi- qibbsite in honor of Dr. Gerald V. Gibbs, Professor at tion of specimensof leucophoenicite by electron .i"ro- Virginia Polytechnic Institute and State University, in probe analysis, several sampleswere found to be another, recognitionofhis outstandingcontributions to the science but closely related phase. These samples gave X-ray of.mineralogyandthesocietyof mineralogists.Thenew prior publica- powder diffraction fatterns which, although generally mineral afld the name were approved, to similar to those of leucophoenicite and memUe-rsof tne tion, by the Commission on New and Mineral humite group, were unique and did not match any known Names, IMA. Type material is preserved at the Smithso- species. The powder X-ray diffraction data, combined nian Institution under catalogue numbers C3209 and with the chemical analyses,implied that this material was R.18772(holotypes) and 149037(cotype). a new manganesemember of either the humite or leuco- phoenicite groups. The distinctions between these two Description and physical properties series or families of structures has been described by Jerrygibbsite occurs as a massive mineral in interlock- White and Hyde (1983)and is summarizedin Table l. Our ing anhedralcrystals, up to 0.5 x 2.0 mm, which display a subsequentinvestigation has confirmed that this phaseis typical metamorphic texture; there are no euhedral crys- a new mineral. tals. The hardness (Mohs) is approximately 5.5. The 0003-(MvE4/0506-0546$02.00 DUNN ET AL,: JERRYGIBBSITE 547

Table l. Crystallographicdata and formulae for the leucophoeniciteand humite groups

Speci es Ideal formula Spacegroup a D Reference

LEUCOPHOEI{ICITEGROUP

Jerrygi bbsi te l.lnn(si 0u )o ( 0H ), Pbnm/Pbn2, 4.85 10.70 28.17 90o Presentstudy Present Leucophoenici te Mn7(si04)3(oH)2-tP?,/b - 4.828 I 0.85 11.380 l03.73o study

Htl,lITE GRoUP

Norbergi te Itsr(si0o)(0H)2 Pbnm 4.7104 10.2718 8.7476 90o Gibbsand Ribbe(1969)

Synthetic phase Mnr(si0n)(0H)2 Pbnm 4.869 I 0.796 9.179 900 Francisand Ribbe (1978)

( (1970) l'lgssi o+)z (oH )z -t-P2.lb 4.7284 10.2539 7.8404 109.060Gibbs et al., (1970) Al I eghanyite Hn5(si04)2(0H)2-tP2,/b - 4.8503 I 0.7l98 8.2747 108.640Rentzeperis

Hurnite ilsr(si0o)r(0H)2 Pbnm 4.7408 10.2s80 20.8526 9oo Ribbeand cibbs (]97]) Francisand Ribbe (1978) lihnganhunite Mn7(si 04)3( 0H ) z Pbnm 4.815 10.580 21.448 900

Cl i nohumite tlse(si 04 (0H P2,/b 4.7441 r0.2s0r 13.6635l0o.79oRobinson et al. (1973) )4 )2 -t-

Sonoli te ilnr(si0o)u(0H)2 -P2. /b 4.872 r0.669 14.287 loo.30 Kato(1978) in t- Ribbe(1980)

is light pink. The luster is vitreous on both cleav- very similar compositions, but differ slightly in Zn-con- age and surfaces. There is one imperfect cleav- tent. The ratio of lt'f+ ions to siliconis 8.85:4.00(sample age, parallel to {001}. The density, determined using R18772)and 8.78:4.00(sample C3209),Z : 4, both in heavyliquid techniques,is 4.00(2)g/cm3, which compares reasonable agreement with the 9:4 ratio of sonolite, favorably with the calculated value of 4.045 glcm3. which is consistentwith the relation thatjerrygibbsite is a Jerrygibbsite is medium violet-pink in color and light polymorph of Mnq(SiO+)+(OH)2.Although the M2+:Si pink in thin-section. It does not fluoresce when exposed ratio is less than the ideal 9:4, we attribute this to to ultraviolet radiation and there is no cathodolumines- analytical error and not to admixed leucophoeniciteinas- cenceunder the beam ofthe electron microprobe. Grains much as leucophoenicite impurities would increase the typically exhibit lamellar structure, parallel to {001}which ratio to >9:4. and not decreaseit. is defined by lamellae difering in opacity; that is, more transparentlamellae alternate with relatively translucent Table 2. Chemical analysesof jerrygibbsite and lamellae. Aside from this diference, the lamellae are leucophoenicite optically identical. The cause of the ditrerence is not definitely known, but is ascribed tentatively to differ- JERRYGIBBSITE LEUCOPHOENICITE encesin fluid inclusion densities. Optically, jerrygibbsite l,leight Atons per Ueight Atomsper fleight is biaxial negativewith 2V : 72". The indices of refraction percent unit cell percent unit cell percent are a: r.772(4),B:1.783(4) and y: 1.789(a).Jerry- gibbsite pleochroic; si 02 27.1 16.00 26.6 ls.7l 26.2 is not is moderate to Fe0 0.3 0.15 0.3 0.r5 0.3 strong, r ) v. The orientation of the indicatix is Z = a, t'lgo 1.4 1,23 'II .l 0.97 ?.6 Ca0 0.4 0.25 .0 0.63 5.2 X:b,andI:c. l4n0 64.I 32.05 62.1 3t 05 57.6 Zn0 3.9 1.70 5.3 2 31 5.6 Chemical composition Hzo 2.13* 8.38 2.25* 8.86 2.5+* F 0.0 n.o, n.o. Jerrygibbsite was chemically analyzed with an anr- Total 99.3 98.6 100.0 sruq electron microprobe utilizing an operating voltage * - l{ater detemined by the Penfield rethod. **- l{ater calculated by difference. of 15 kV and a sample current of 0.025 pA, measuredon Accuracyof data: r3t of the amunt present for mior elerents. brass. Water was determined using the Penfield method. standards: hornblende (Si, Fe,lilg,Ca) ; mnganite(Mn); synthetic Zno(Zn); and fluorapatite(F) for sanp'lesRI8772 and 149543. The resultant analyses,together with standardsused, are : synthetic (Si,l'ln); hornblende(Fe,l'lg,Ca); and (zn). presentedin Table 2. The two jerrygibbsite sampleshave synthetic ZnO 548 DUNN ET AL,: JERRYGIBBSITE

X-ray crystallography diffraction patterns were unsuccessful.We subsequently determined that this was due to contamination by other Cleavagefragments from specimensR18772 and C3209 phases.The principal contaminatingphase is leucophoen- have been studied using Weissenberg and precession icite although other Mn-humites such as alleghanyitealso single-crystal techniques,and the results are identical for are found with jerrygibbsite. Indeed, we have been un- all photographs jerrygibbsite crystals. The show that is able to obtain a jerrygibbsite powder pattern which is orthorhombic with space group Pbnm or Pbn21, with entirely free of leucophoenicite contamination. The pat- refined parameters : = unit c-ell a 4.85(l), b 10.70(l),and terns for jerrygibbsite, leucophoenicite, and the Mn- c = 28.17(3)4. This setting of the unit cell translations is humites have many features in common. Becausemulti- used so as to emphasize the relation to the related phase patterns appear to be the rule and not the members group, : of the humite all of which have a 4.9 exception, and because it is difficult to interpret these = andb 10.5A translationsin common; this settingwas patterns even after considerable experience with them, recommended (1969). by Jones The spacegroup identifi- we have gone to some lengths to try to obtain standard, cation is somewhattentative as extinction rules are not as single-phasepatterns of each phase. These patterns were well defined as usual due to the presence of classes of shown to be single-phasein two ways: (l) d-valueslead to systematically weak reflections. For example, very few satisfactory indexing of unit cell parameters, and (2) hOl reflections are observable. We feel that the space reflections known to be characteristic ofother phasesare group as probably determined'is very correct, although found to be very weak (in which case appropriate reflec- there is a possibility that intensities as determined in the tions were subtracted from the pattern) or absent in the course of a structure determination may show that the pattern chosen to be representative of a given single apparent glide-planes are not required in the detailed phase. We have separatelydescribed patterns for allegh- structure. anyite, sonolite,and manganhumite(Winter et al., l9E3) Individual lamellae from the zones defined by translu- and report the leucophoeniciteand jerrygibbsite patterns cency differences noted above were separately studied here (Table 3). It should be possible to characterize any using single-crystalX-ray diffraction and chemical analyt- mixture of leucophoenicite-humite family phases with ical techniques. The results for all lamellae studied were these patterns as a guide. identical, showing that the lamellae are not a function of No jerrygibbsite pattern is entirely free of peaks of diferences in the structure. Because the lamellae have leucophoenicite. We therefore first separatelycharacter- as the {001} interface and becauseit is this plane which is ized the diffractometer pattern of leucophoenicite using the common interface between intergrowths of two or several samples from Franklin, New Jersey, including more members of the humite or leucophoenicite groups type material, and then substracted the peaks of leuco- (White and Hyde, 1982),it was especially important to phoenicite from the jerrygibbsite pattern. The reflections establish (1982) this relation. White and Hyde have displayed in the Debye-Scherrerand Gandolfi patterns of shown, using high-resolution TEM lattice-fringe-imaging powdered, polycrystalline samples of the same leuco- techniques, that members of the humite family may phoenicite and jerrygibbsite samplesare invariably weak exhibit fine-scale, mixed layering parallel to {001}. We and diffuse. We have no explanation of this phenomenon, therefore examined the X-ray photographs especially inasmuch as the diffractometer patterns of these same carefully in order to determine the presence of any samplesare of high quality. We note this becausein such streakingparallel (001). to None was observed, indicating Debye-Scherrerand Gandolfi powder patterns the weak- that if there is random mixed layering of more than one er peaks of contaminating leucophoenicite are generally species, it is below levels of detection by such X-ray not detectable. techniques. Powder X-ray diffraction data are given in Table 3. These were obtained using a powder X-ray difractom- Occurrence eter, CuKo X-radiation monochromatedwith a flat graph- Jerrygibbsite occurs in specimens from the Franklin ite crystal. The powder was mounted on a glass slide and Mine, in Franklin, SussexCounty, New Jersey. Three Si was used as an internal standard. The lattice parame- specimenshave been characterized. Two are the holo- ters D and c are relatively large, resulting in a large types and the third, a cotype (#149037) was generously number of symmetrically non-equivalent d-values; thus donated to the Smithsonian Institution by its owner, Mr. few reflections could be unambiguously indexed. The Charles Key, when he was informed of its significance. limited number of unambiguously indexable reflections No data have been preserved as to the geological occur- occurred despite the elimination of the classesof system- rence(s)within the Franklin orebody, as is the case with atically weak reflections. We have therefore given indices most specimens collected in past years. A label with for only a limited number of reflections and only 6 could specimen #C3209 suggests it was collected in 1907. be used in least-squaresrefinement of the lattice parame- Inasmuch as the three known specimenswere all labelled ters. leucophoenicite, and becausejerrygibbsite also resem- Our original attempts to obtain high-quality powder bles the Mn-humites from Franklin, it is quite possible DUNN ET AL.: JERRYGIBBSITE 549

Table 3. X-ray powder diffraction data for jerrygibbsite and ship. The analysesare low by 2-3 weight percent and are leucophoenicite.The d-values are listed so as to facilitate not given here, but they nonethelessconfirm equality of direct comparison of the patterns, composition for thesephases. The sonolite andjerrygibb- site composition, with (OH) calculated and F undeter- l' Jcrryglbbslte LeucoDhoenlclte mined, yields the formula calculated on the basis of Si : d(obs) d(calc) hkl Uto d(obs) d(calc) hkl Uto 4: (Mne.zZno.asMg6.22Caa.11Fes.e3)1e.65(SiOa)+(OH)z.lo. 5,25 5,26 021 9 5.26 5,27 021 13 The compositions of both the sonolite and jerrygibbsite J.ZT fi2 4.40 4.41 110 II 4.40 4.39 110 19 are identical to two decimal places. Both species were 4.37 022 4.37 021 also characterized by X-ray powder diffraction tech- 4.25 4.26 024 9 3.95 3.95 lll l8 niques. 3.74 3.74 114 24 3.62 3.64 102 41 A notable characteristic of the jerrygibbsite assem- 3.62 tI2 3.55 3.56 121 5 3.56 3.56 tzL 10 blages, compared with those of most leucophoenicite, is 3.53 026 3.56 120 3.48 3,52 008 29 the virtual absence of -bearing minerals, which 3.48 115 3.28 3.28 111 ln are almost invariably found associatedwith leucophoeni- 3.2t 3,22 116 14 3.24 3.24 Lzz 6 leucophoeni- 3.22 027 3.24 t21 cite. This is, however, only a rule-of-thumb; 2.97t 2.971 1T3 15 cite has been found, albeit rarely, without associated 2.869 2.873 r30 78 2.890 2.889 m 100 2.859 l?t calcium-bearingphases. These latter observations are of 2.813 2.8t7 00,10 9 2.840 2.842 2,815 132 a general nature, based on the examination of a large 2.752 2.753 118 49 2.751 2.751 Liz 32 2.748 133 2.164 004 numberof specimens. 2.702 2.702 029 46 2.722 2.728 024 54 2.714 0:n 2.651 2.664 041 34 2.695 2.694 l3l 94 2.661 I34 2.531 2.629 M23 2.628 2.635 M2 32 Discussion 2.636 040 2.633 131 The chemical, X-ray diftaction, and other data for 2.557 2.560 135 100 2.554 Il9 jerrygibbsite clearly showsthat it belongsto the humite or 2.493 2.493 133 20 2.450 2.450 lT4 52 leucophoenicitegroup ofminerals, but the precise nature 2.416 16 2.426 25 of the relation is ambiguous. There appear to be two 2.377 11 2.371 29 2.362 39 possible structural relations for jerrygibbsite, and its 2.342 28 2.225 4 2.202 12 classification depends on which of these two structure 2.118 q 1.973 9 1.806 I@ 1.811 96 types it possesses. 1.730 l8 L,752 28 jerrygibbsite 1.7t2 l0 1.709 20 First, because orthorhombic and mono- 1.692 I1 clinic sonolite are both polymorphs of Mnq(SiO+)r(OH)z 1.661 9 1.642 7 1.619 I and because the relationship between their unit cell 1.598 t4 1.576 27 dimensionsis that of a classic unit-cell twinning relation: 1.567 26 1.566 28 1.550 34 a (errygibbsite}: o (sonolite); b fie11ygibbsite;- 6 1.458 18 1.473 19 (sonolite); and c (jerrygibbsite) - 2 d (lM) (sonolite), it initially appeared to us as if jerrygibbsite is simply the Data obtained uslng porder dlffractometer (1o7mln.scan rate) rith graphite mnochro0Etor, using an internal strndard. unit-cell twinned equivalent of sonolite, where the term = = = r. q 4.85,!. l().70,I 28.17R. "unit-cell twinning" is used in the senseof Ito's defini- 4.828, b = 10.85, = 1,|.3801,o = 103.730. z. L= 9. tion (1950).Single-crystal diffraction photographsof both twinned and untwinned crystals of sonolite were directly compared with those of jerrygibbsite in order to investi- that other samplesmay repose misidentified in public or gate such a relation. These comparisons verified the private systematic collections, particularly those of unique diffraction characteristicsof jerrygibbsite relative Franklin minerals. to sonolite and were at least consistent with a unit-cell Jerrygibbsiteoccurs in contact with , willem- twinning relation and not twinning by pseudo-merohedry ite, , and sonolite in a very colorful and uncommon as observed in sonolite. Such evidence is merely non- assemblage.The mineral resembles some leucophoeni- exclusive of unit-cell twinning, however, and does not cite, but is decidedly more brown and violet-colored than exclude other possibilities. It is interesting that where most leucophoenicite.As with leucophoeniciteand mem- unit-cell twinning is found in other mineral systems, bers of the Mn-humites, jenygibbsite is not easily recog- disorder exemplified by streaking of X-ray reflections is nized without X-ray or chemical data. In one of the type @mmon. None has been observed in this case, implying specimens(R18772),jerrygibbsite is associatedwith, but that such a relation may not exist. not seen to be in contact with, leucophoenicite. On A second possible structural relation for jerrygibbsite specimen #149037,jerrygibbsite occurs in contact with has been pointed out by White (personalcommunication, dark brownish-pink sonolite; both were analyzed and 1982)and is based on the nature of structural relations have essentially identical compositions within analytical within the leucophoenicite group vis-a-vis those of the error, which further supports their polymorphic relation- humite group as defined,for example, by White and Hyde 550 DUNN ET AL,: JERRYGIBBSITE

(1983). Their view of the structures focuses on the polymorphs of Mnz(SiO+):(OH)zand Mne(SiO+)r(OH)2, closest-packingof the cations, rather than ofthe anions, respectively. They occur at Franklin, New Jersey, where as is conventional (Moore, 1970). Sequences of Mn sonolite, , and tephroite also occur, in para- cations are designated (1,2^) for leucophoenicite-group geneses similar to those of the Mn-humites. Neither phases, and (2",3) for the humite-group phases. This jerrygibbsite nor leucophoeniciteis known to occur in the is related to the twinned cubic closest packed array very similar deposit at Sterling Hill, New Jersey, suggest- of Ni atoms (. . . .2,2. . .) in Ni2In and of Cr atoms ing very localized conditions might be needed for the (. . . . I,l . . .) in CrB, with leucophoenicite formation of these phases. having the twinned c.c.p. sequence for Mn of In the next section we discuss some compositional (. . . . 1,2,2,2,1,2,2,2. . . ) or (1,2\. White (pers. factors which may be related to the formation of leuco- comm., 1982)points out that jerrygibbsite could well be phoenicite. We show that it is characterized, in part, by a member of the leucophoenicite-group with sequence the presence of Ca and/or Zn, implying that these ele- Q,2\2. White and Hyde (1983)have carried out a TEM ments stabilize leucophoenicite relative to manganhu- study of leucophoeniciteand they report that other mem- mite, and that the two phasesmay not be, sensustrictu, bers of the leucophoeniciteseries, such as (1,24),and (1,2s) polymorphs. Assuming that jerrygibbsite has a leuco- do occur within the prevalent (1,23)units which define phoenicite-grouptype structure, we tentatively postulate leucophoenicite. Whether jerrygibbsite is a member of that the stable polymorph of pure Mne(SiOr)+(OH)zis the leucophoenicite or humite groups cannot be proven sonolite under the P-T conditions of formation at Frank- without a analysis or high-resolution lin. Because jerrygibbsite always has been found inti- TEM study.In this regard,Dr. T. Kato (pers.comm.) has mately intergrown with leucophoenicite, we postulate undertakena structure analysis and we are carrying out a that the initial formation of leucophoenicitemay serve as TEM analysisofjerrygibbsite, for which only preliminary the nucleus which requires further growth of a coherent results are described below. leucophoenicite-group phase, which under the proper There is strong circumstantial evidence that jerrygibb- bulk composition conditions is jerrygibbsite. Alternative- site is a member of the leucophoenicitegroup becauseof ly, some Zn may serve to stabilize the jerrygibbsite the close associationof these minerals. Indeed, we noted structure relative to sonolite, in which case these miner- above that we have not been able to isolate enough als may not be polymorphs,sensa strictu.If the former is jerrygibbsite for a powder diffractometer pattern without correct, sonolite should prove to be the common form of including some leucophoenicite. In order to understand Mnq(SiOa)e(OH)2and it should generally be the phase this association, an ion-thinned, single-crystal of jerry- observed in synthesis experiments, but jerrygibbsite gibbsitewas studied using conventional TEM techniques, should continue to be found only rarely and always in with (001) (the interface common to members of the intimate associationwith leucophoenicite. leucophoeniciteand humite groups) parallel to the elec- tron beam. Parallel lamellar features, approximately Leucophoenicite 10004 in width, were found to be randomly distributed In order to obtain additional samplesof jerrygibbsite, throughout the jerrygibbsite which was otherwise homo- we searchedin several private collections. One sample, geneous.We tentatively assumethat thesefeatures corre- labelledleucophoenicite, was found in the mineral collec- spond to interleaved leucophoenicite. The close spatial tion of Mrs. Alice Kraissl, and evidenceda markedly relation between leucophoenicite and jerrygibbsite lamellar texture. Inasmuch as this texture is similar to strongly implies that they are closely related structurally that ofjerrygibbsite, but unlike that of other leucophoeni- and that the structure tentatively suggestedby White is cite samples, the specimen was subjected to a detailed the correct one for jerrygibbsite. These very tentative investigation. That work indicated that this lamellar min- results also imply that physical and chemical properties eral is indeed leucophoenicite, albeit one of uncommon as reported for jerrygibbsite may include a small (<5Vo) habit; most leucophoenicite is glassy with no evident contribution from included leucophoenicite. textural features. The identification of this leucophoeni- The limited occurrence of leucophoenicite and jerry- cite (NMNH 149543)was verified using single-crystal gibbsite relative to the well-known Mn-humites alleghan- methods and powder difractometer techniquesas previ- yite, sonolite, and manganhumite,is somewhat puzzling. ously described.The powder diffraction data are given in Winter et al. (1983)have shown that these humites plus Table2. The refinedunit cell parametersare a:4.828(5), the Mn-analogue of olivine, tephroite, all occur at the D : 10.85(l),c = 11.380(9)4,a = 103.73(5)".The powder Bald Knob manganesedeposit in Sparta, North Carolina. diffraction data are free of peaks from other phases, These phases are shown to exist in equilibrium assem- including those of jerrygibbsite. This texturally atypical blages as a function of the ratios of SiO2 and H2O leucophoenicitesample was found to give difractometer activities, under amphibolite P-T conditions. These min- data identical to that obtained from type leucophoeni- eralsare also not uncommon throughout the world, where cite. We obtaineda microprobeanalysis of this sample, bulk-rock compositions are appropriate. using the operating conditions and standardsused for Leucophoenicite and jerrygibbsite are relatively rare jerrygibbsite (R18772);the analysisis presentedin Ta- 551 ble 2. This yields the chemical formula, caleulated e. No known leucophoenicite approaches end-member on the basis of 14 (O+OFI) with water by difference: composition; all have some Zn and/or Ca replacing Mn. (Mn5.5eCaa6aZn6 aTMg6.aaFeo or)>z rzSiz.s7O12.11(OH)1 se,in The above observations suggest that special composi- excellent agreementwith the ideal composition of leuco- tional factors, at least in part, may be responsible for the phoenicite,Mnz(SiO+h(OH)2. formation of leucophoenicite and, perhaps, jerrygibbsite. As mentionedpreviously, all the Mn-humites are found This suggestion is prompted by the lack of evidence for in diverse localities except for leucophoeniciteand jerry- other special conditions; the presence of Ca or Zn in all gibbsite. Jerrygibbsite is so far restricted to the Franklin samples; the absence of any pure end-member; and the deposit, as was leucophoenicite until recently, when it very localized occurrence at Franklin of these minerals in was reported from Pajsberg,Sweden, by White and Hyde bulk quantities. (1983).The Pajsbergocculrence is one of very small The conditions necessary for the formation of jerry- quantity and thesetwo speciesare known in bulk samples gibbsite and leucophoenicite remain enigmatic. Although only from Franklin. This localized restriction is not easily the solution to this dilemma may be a complex one, explained in terms of assemblagesor P-I conditions. involving many factors, it seems likely to us that the Accordingly, we have examined a large number of leuco- above mentioned compositional factors may play some phoenicite samplesin an attempt to make some observa- role in stabilizing these species. tions of a generaland specific nature which might help to explain the occurrence at Franklin of these species. Acknowledgments Unlike many of the 34 species unique to Franklin and We are grateful to Dr. T. J. White of the Australian National Sterling Hill, leucophoenicite does not occur in a very University for providing descriptions of his unpublishedresults restricted assemblage.An attempt to sort over 70 speci- regardingthe humite and leucophoenicitegroups. We thank Mrs. mens into assemblagesbased on associatedminerals, Alice Kraissl and Mr. Charles Key for the doriation of critical textures, ore-relations, and available chemical compo- samples.The Penfield water determinationswere made by Ms. nents failed to limit these samplesto a restricted number Julie Norberg. This project was supported, in part, by grant from ofparageneses.The opposite,in fact, was evident;leuco- Mrs. E. Hadley Stuart, Jr. PJD expresseshis gratitude to the for their assistance. phoenicite occurs with a broad and diverse array of Trustees of the Franklin Mineral Museum This manuscript was greatly improved by the reviews by Dr. T' speciesboth rare and common, and occurs both within J. White, Dr. Gerald V. Gibbs, Dr. Paul Ribbe, and Dr. Malcolm vein assemblages. ore and in Ross and we are grateful for their helpful suggestions. Microprobe analysesof 30 samplesof leucophoenicite, representinga wide variety of parageneseswere carried out. Data from these analysesindicates: References samplesanalyzed,28 have CaO values be- a. Of 30 Francis. C. A. and Ribbe, P. H. (197E)Crystal structures of the tween3.8 and 6.9 weight percent,and only two have Mn humite minerals: V. Magnesian manganhumite. American in excessof6 atomsofthe 7 possibleoctahedral cations. Mineralogist, 63, 874-877 . The Ca content of the 28 calcic samplesvaries from 0.47 Gibbs, G. V. and Ribbe, P. H. (1969)The crystal structures of to 0.85 Ca atoms per 7 octahedrally coordinated cations. the humite minerals: I. . American Mineralogist' b. The two sampleswith very low Ca content and Mn 54,376-390. contentsof 6.0 and 6.3 atoms per 7 octahedralcations Gibbs, G. V., Ribbe, P. H., and Anderson, C. P. (1970)The were studied using single-crystal and powder diffraction crystal structures of the humite minerals. II' Chondrodite. methods and found to be leucophoenicite, but multiply American Mineralogist, 55, 1182-1194. Ito, T. (1950) X-ray studies on polymorphism' Maruzen Co., twinned. Tokyo. c. All the analyzedleucophoenicites contain which Jones,N. W. (1969)Crystallographic nomenclature and twinning varies from 0.23 to 0.47 Zn atoms per 7 octahedrally in the humite minerals. American Mineralogist, 54,3W-312. coordinated cations. There is no apparent systematic Moore, P. B. (1970) Edge-sharing silicate tetrahedra in the variationbetween Ca and Zn, Mg or I(Zn+Mg). The Zn crystal structure of leucophoenicite. American Mineralogist, content is highest in those samplesassociated with zinc- 55, 1146-1166., ite, although these samples comprise a minority among Rentzeperis,P. J. (1970)The crystal structure of alleghanyite, known leucophoenicite assemblages.In the majority of Mnst(OH)zl(SiOn)zl.Zeitschrift fiir Kristallographie' 132,l-18' leucophoenicite assemblages;zincite is usually absent Ribbe, P. H. (1980)The humite seriesand Mn-analogs.Reviews 4. and is the dominant Zn-bearing phase. in Mineralogy, 5, Orthosilicates. 231-27 H. and Gibbs, G. V. (1971)Crystal structuresof the d. Almost all the leucophoenicite assemblagesstudied Ribbe, P. humite minerals: III. Mg/Fe ordering in humite and its relation high in calcium. The associ- have bulk-rock compositions to other ferromagnesiansilicates. American Mineralogist, 56, ated minerals vary, but and/or andradite are the l 155-l173. dominant Ca-bearing species. Although a few samples Robinson,K., Gibbs,G. V., and Ribbe,P. H. (1973)The crystal were found which were not associatedwith Ca minerals, structuresof the humite minerals: IV. and titano- these were very small samples,perhaps not at all indica- clinohumite. American Mineralogist, 58, 4349. tive of the complete rn sitrz assemblage. White, T. J. and Hyde, B. G' (1983)An electron microscope s52

sludy of leucophoenicite. American Mineralogist, d8, l0@- Chemistry of Minerals, E, 167-174. lwr. Winter, G. A., Essene, E. J. and Peacor, D. R. (19E3)Mn- White, T. J. and Hyde, B. J. (1983) A description of the humites from Bald Knob, North Carolina: mineralogy and leucophoenicite family of structures and its relation to the phase equilibria. American Mineralogist, 58, 951-959. humite family. Acta Crystallogr.aphica, 839, lfl7. White, T. J. and Hyde, B. G. (19E2)Electron microscope study Manuscript received, March 18, 1983;. of the humite minerals. II. Mn-rich specimens. physics and accepted for publication,. October 2il, 1983.