sign in sign up
<<
Home , Bronzite, Chromitite, Spinel group, Titanium in Africa

American Mineralogist, Volume 64, pages 140_t50, 1979

Titanium-bearingoxide of the critical zoneof theEastern BushveldComplex

EuceNr N. CeNaenoN Department of Geology and Geophysics Uniu e rsity of Wis consin- M adison M adis on, IVis consin 5 3 706

Abstract

The principaltitanium-bearing minerals in theCritical Zone of theEastern Bushveld Complexare ,the most abundant,and rutile, widelydistributed as an accessory . and a complexTi-rich oxide are rareminerals in the zone. The TiO, contentof one chromiteis 5.17percent, but the rangefor all otherchromites analyzedis 0.31percent ro 2.47percent. Rutile occurs as aciculaicrystalsin plagioclaseor bronzitein somelayers of the rock sequence,but mostly it is associatedwith chromite. Orientedacicular inclusions in chromiteand anhedramarginal to chromiteappear to have formed by subsolidusoxidation-exsolution from chromite.Prismatic included in chromite,straddling chromite-silicate boundaries, or embeddedin plagioclaseor bronziteare interpretedas cumuluscrystals formed more or lesssimultaneously with chromite.Rutile interstitialto chromitecrystals in two chromititesapparently formed during the postcumulus stage. Rutile and chromitewere evidently stable as coexistingphases during accumulation of manyunits of theCritical Zone. The low TiO, contentsof chromitethus represent saturation levels at time of formation.A miscibilitygap may thereforehave existedin the series chromite-titanomagnetiteunder the conditionsof crystallizationof the Critical Zone.The occurrenceof cumulusrutile in a chromititefrom the Muskox intrusion and another from the Fiskenaessetintrusion suggests a similarconclusion for thosebodies.

Introduction Previous work Two majorgroups of oxidemineral concentrations Available information on in oxide miner- arepresent in thelayered rocksof theBushveld Com_ als in the Critical Zone is of two kinds: (l) analyses plex.One group consistsof vanadium-bearingtitano- reporting TiO, contents of from various magnetitedeposits in the upperpart of theComplex. horizons, and (2) descriptionsof titanium-rich oxide Titanium-bearing oxidesin thosedeposits have been phasespresent in various rock units, mostly in chro- studiedin somedetail (Molyneux, 1972).The second mite-bearing rocks. All complete analysesof chro- group consistsof chromititesin the Critical Zone. mites from the Eastern Bushveld Complex show Titaniumis knownto be present in chromitein these TiOr, the amounts reportedranging from 0.20to 2.40 rocksand in associated silicate-richrocks, and other percent.Most analyses(cl DeWet and Van Niekerk, titanium-bearingoxide phases haveoccasionally been 1952; Cameron, 1964;Heiligman and Mikami, 1960; reported,but the occurrenceof titanium in oxide Cousins and Feringa, 1964;Waal, 1975)are of sam- mineralshas not been systematicallyexamined. ples from various , in which chromite is This paper describesthe occurrenceand composi- the principal mineral but silicatesare presentin vari- tionsof titanium-bearing oxideminerals in the Criti_ ous amounts. The exact significanceof the reported calZone of the Eastern BushveldComplex. The data values of TiOz is not indicated for any of the pub- help to explainthe compositional and stratigraphic lished analyses.Purification procedures used have gap betweenthe -rich spinelsof the lower been designedprimarily to separatechromite from part of the Bushveldand the -titaniumspinels of associatedsilicates by crushing, elutriation, or mag- the upperpart. netic separation,and solution of remaining silicates 0003-004x/ 7 9 /0 | 02_0I 40$02.00 CAMERON: BUSHVELD COMPLEX t4l in hydrofluoricacid with or without sulfuricacid. diameter, but crystals 20 to 50 pm in length are The presenceor absenceof separateTiO, oxidesin found. the analyzedconcentrates has generally not beende- 2. As stout to elongate prisms included in chro- termined. mite, I to 4 inclusionsper chromite grain (in section). Thereare a few reportsof titaniumoxide minerals Control of orientation by the lattice of host chromite in various chromites of the Eastern Bushveld. may or may not be evident (Fig. 3B). Crystals range Younger(1958) described bodies of ilmenitein chro- up to 0.73 mm in length. mite in chromititesfrom Dwars River (Fig. l) and 3. As anhedra along margins of chromite crystals from the M unit (Fig.2) on Winterveld417 KS, and (Fig. 3C) or at intergrain boundariesof chromite. At tiny irregulargrains of rutilealong grain boundaries triple junctions of chromite grains, rutile may appear of chromite.Cameron and Emerson(1959) men- to occupy grain interstices(Fig. 3D). tionedthe occurrenceof rutilein certainchromitites, 4. As euhedraalong margins of chromite crystals; and Guilbert (1962)reported rutile as an accessory the chromite is molded against rutile. Some rutile mineral(trace amounts to 0.2percent) in rocksof the crystals straddle chromite-silicate boundaries (Fig. C and D units(Fig. 2) on Winterveld417 KS and 44). Jagdlust418 KS. Rutilewas also found in acicular 5. As independentcrystals up to 0.32mm in length crystalsin labradoriteand bronzitein feldspathic (Figs. 48, 4C), mostly associatedwith chromite but .Cameron and Glover (1973) briefly com- largelyor completelydetached from it and embedded paredthe TiOz contents of spinelsin maficpegmatites in postcumulusplagioclase, less commonly biotite. with TiOzcontents of spinelsin the normalrocks of 6. As aggregatesinterstitial to chromite (Fig. 4D). the EasternBushveld. 7. As acicular crystals in ,less com- monly in . Materialsstudied Acicular inclusions in chromite and marginal an- Materialsstudied are outcrop samplesand drill hedra (l and 3 above) are commonly associatedand coresrepresenting the full 1600-meterthickness ofthe are found in various chromitic rocks from the B unit CriticalZone on FarmsJagdlust 418 KS, Umkoanes to the top of the Critical Zone. Prismatic euhedral Stad412 KS, andWinterveld 417 KS (Figs.I and2), inclusionsin chromite, marginal anhedra, and inde- plus collectionsfrom numerousother farms in the pendentcrystals (2,3, and 5 above) occur togetherat EasternBushveld. Polished sections representing all a number of horizons from the D unit upward. They major units of the Critical Zone, as well as many are especially abundant from the F unit upward. individuallayers within units,have been examined in Acicular and prismatic inclusions may occur in a the courseof the presentstudy. single grain of chromite; in such grains the portions adjacent to large inclusions are devoid of acicular Occurrenceof titaniumin oxide minerals inclusions.Types 1, 2,4,5, and 6 may alI occurin the same rock, but type 6 is the least common and has Generalslqtement been noted thus far only in two chromitites. In rocks of the Critical Zone, tiranium occursin The acicular, oriented inclusions in chromite are four oxideminerals-rutile, chromite, ilmenite, and presumablydue to separationof TiOz from the chro- an oxideof complexcomposition (Cameron, 1978). mite structure.The processis not simple exsolution, Chromiteand rutile are by far the mostwidely dis- however,since charge imbalance in the (cation tributed,and chromiteis by far the mostabundant. excess)would result. More likely is an oxidation- exsolution similar to that proposed as the origin of Rutile ilmenite-titanomagnetite intergrowths (Verhoogen, Rutile is readilyrecognized by meansof its white 1962;Buddington and Lindsley, 1964).ln terms of a color, reflectivity,anisotropism, bireflectance, twin- hypothetical ulv6spinel component of the chromite ning in largercrystals, habit, and internal , the reaction reflections(yellow to amber,less commonly reddish). 3 FerTion * 02 - 2 Feso^+ 3 Tio, Rutile occursin the CriticalZone in sevendifferent ways: may be written. l. As acicular inclusions in chromite, parallel to The separationof TiOz from chromite appearsto (l1l) of chromite (Fig. 3A). Most suchinclusions are have beenvery uneven.Chromite-rutile intergrowths 5 pm or less in length and are less than I pm in have beenobserved in polishedsections of rocks rep- t42 CAMERON: BUSHVELD COMPLEX

LISI OF FARMS 1 Scheiding 4O7 KS 2 Zeekoegot 421 KS 3 Jogdlusi 418 KS 4 Wrntenveld 417 KS 5. Moeijelijk 412 KS KS 6, Umkoones Stod 412 KS 7. Zwortkoppies 413 KS I Woterkop 113 KT 9 Poschos Krool 466 KS 1O Mecklenburg ll3 KT 11, Hockney 116 KT i2 Surbiton t15 KT 13, Fofest Hill 117KT 14. Clophom 118 KT 15- Winnoorshoek 25O KT 16 Groothoek 256 KT 17 Driekop 253 KT 18. Mdondoogshoek 254 KT 19 Mooihoek 255 KT 20. Hendriksploqts 28t KT 2l Onverwocht 292 KT 22. Doordbsch 294 KT 23. Wlnte.veld 293 KT 24, Eee.ste Gcluk 322 25. Goudmyn 337 KT 26. De Goede Verwochtinq J32 KT 27 Annex Gnootboom 335 KT 28, Grootboom 336 KT 29. Spitzkop 333KT 30. Kennedys Vole 361 KT 3l Tweefontein 360 KT 32. Frischgewoogd 362 KT 33. Kolkfontein 367 KT 34, Dwors River 372 KT 35. Dc Grooteboom 373 KT 36 Rietfontein 375 KT 37 Richmond 37O KT 38 Thorncliffe 374 KT 39. Heleno aJT

M)Z\Z NePhclire syenrte

BushwH groniie ond oss@ioted nocks

E6tem Bushveld Complex F_l Moan_ond Upper zones

Criticol Zone

fL_-.l Lorcr Zone

Merensky Reef vr7 Pfetorio series

or----:-to Kilomete.s

Fig. I Generalized geologic map of the Eastern Bushveld Complex. CAMERON.' BUSHVELD COMPLEX 143

u t ii-- uJ o o- z6 intenloyered with ononthosite(o) ZN

363 Noritc, onorthosite(o), ond feldspothic bronzitlte (b)

47 Gobbro (obove) ond norite r.il I l5 Feldspothlc bnonzltite z Anorthosite, norite(n). gobbro (g), ond feldsPothic o bronzitite (b), with thin chnomititcs in uPPenPont N

J I L ( r) 1O4 Feldspothic bronzitite interloyencd with chromititcs Steelpoort ond Leoder chromlte ii. ihl..ldvcned with chromitites

Fcldspothic ond hot'zbungite(dr) ond feldspothic bronzitites, with thin chromitltcs

273 Feldsoothic bnonzitite with thin chromitites

LOWER ZONE

Fig. 2. Section of the Critical Zone on Farms Jagdlust, Winterveld 417 KS, and Umkoanes Stad. resentingmore than 100different layers in the Criti- homogeneouschromite may thus contain sub- cal Zone. In everysection, there are numerouschro- microscopicrutile. mite grains devoid of rutile; e.9., in a chromitic In rocks in which acicularrutile forms the only anorthositefrom the lower part of the F unit, 79 inclusionsin chromite,marginal rutile anhedraor percent of the grains are rutile-free. Also, inter- subhedraare almostinvariably present. In part, at growthstypically are presentonly in partsof grains least,the latter may represent "exsolution" of TiO, to andare randomly disposed. In somerocks, this heter- the marginsof chromitegralns. ogeneitymay be apparent,not real,for exsolvedru- Prismaticinclusions (Fig. 38) appearfirst in a tile crystalsin chromiteare mostlyvery small,with chromiticpyroxenite layer in the upperpart of the D sizesextending to thelimits of opticalresolution with unit and are found in chromiticrocks at intervalsto a l05X oil immersionobjective. Some apparently the MerenskyReef. The distinctionbetween acicular CAMERON: BUSHVELD COMPLEX

those that are only partially enclosed,as in Figure 44, to those that are completely detachedfrom chro- mite. Some apparent inclusionsundoubtedly are the ends of crystalsthat are partly embeddedin silicates but project into the margins of the host chromite crystals. The textural relations,taken together,strongly in- dicate that in such rocks rutile precededboth chro- mite and silicateson the liquidus. Crystallization of rutile may have ceasedwhen chromite crystallization began, but the sharp outlines of the inclusions in- dicate that rutile remained stablein the presenceof chromite. The bulk TiO, contents of the chromites thus representsaturation for TiO, under the condi- tions prevailing. The textural relationsof rutile of type (6) are also pertinent to the question of saturation. This type consists of aggregatesof rutile crystals interstitial to

Fig. 3 Photomicrographs, reflected light, oil immersion. (A) Acicular inclusions of rutile (light gray) in chromite (medium gray). F unit, Jagdlust.(B) Prismaticinclusions of rutile (white) in chromite (gray). H unit, Jagdlust. (C) Anhedral and subhedral rutile (white) at margins of chromite grains. in , Umkoanes Stad. (D) Rutile (white) at triple junctions of chromite grains (gray). Steelpoort chromitite, Jagdlust. and prismatic inclusionsis clearcut in most rocks in which the two types of inclusions occur. In a few there is a gradation between the two in size and in habit, and in such rocks the larger inclusions,like the smaller, may be due to exsolution. However, small crystals of chromite are common as inclusions in cumulus and bronzite, and these inclusions are bestinterpreted as crystalstrapped during growth of the host crystals (Cameron, 1975). In the same way, most of the large inclusionsof rutile in chromite appear to representtrapped crystals. Three features support this interpretation. First, there seemsto be no systematicrelation betweensize of inclusion and size of host crystal. Second,in any given sample, most chromite grains lack such in- Fig. 4. Photomicrographs, reflected light. A, C, and D in oil clusions, polished in a sectionof chromitic anortho- immersion.(A) Two rutile crystals(light gray),one partly enclosed site from the H unit, for example, only 15 of 321 in chromite (medium gray), one completely enclosed.H unit, chromite crystals intersected by parallel traverses Jagdlust (B) Cumulus rutile crystal (white) associated with contain large rutile inclusions.Third, rocks in which chromite (light gray) and bronzite (medium gray). Base of F chromitite, Annex Grootboom. (C) Rutile crystal (light gray) prismatic inclusionsare presentin chromite also con- in plagioclase(black). Medium gray-chromite.F unit, Jagdlust.(D) tain independent rutile, and commonly there is a Postcumulusrutile (white) interstitialto chromite (medium gray). range from crystals that appear totally enclosedto Footwall chromitite, Winterveld. CAMERON BUSHVELD COMPLEX chromite (Fig. aD). The aggregatesare very irregu- larly distributed. Rutile is very abundant in places, absentin others.The erratic distribution ofrutile and the local high rutile/chromite ratios cannot be ex- plained in terms of exsolution. Rutile of this type evidently formed at the postcumulus stage.In both chromititesin which such rutile occurs,the chromites contain less than I percent TiOr, and saturation of chromite at low levelsis again indicated. lrvine (1974) reported that in chromite from the Muskox intrusion rutile occursas inclusionsin chro- mite, but more commonly as primary rodlike crystals scatteredamong the chromite grains. A sample of chromite collectedby the writer from unit 17south of Transition Lake (see Smith et al., 1967) shows the two types of rutile occurrencedescribed by Irvine. In addition, rutile occurs as anhedra marginal to chro- mite grains, and as minute oriented needleswithin chromite grains,both types presumablyrepresenting oxidation-exsolution.Finally there is one occurrence Fig. 6. Photomicrographin reflectedlight, oil immersion.Two (Fig. 5) of what appearsto be postcumulusrutile. cumulus crystalsof rutile (white) associatedwith chromite (gray) and plagioclase(black). Minute inclusionsof rutile in chromitite The normal TiO, content of chromite in the appear due to oxidation-exsolution. Chromitite, Fiskenaesset Muskox sample is 4.6 percent. Crystals in contact complex, Greenland with the postcumulusrutile, however, have only 3.4 percent TiOr. Partial depletion during formation of (lrvine, 1974, p. 1009) as inclusions in Stillwater rutile is indicated,but it is not enough to accountfor chromite. From Campo Formoso, Brazil,Hedlund er all the rutile present. al. (1974) reported rutile and ilmenite in chromitite, Rutile has also been reported by E. D. Jackson apparentas unorientedinclusions in chromite grains. Ghisler (1970)found rutile in chromititesin the Fis- kenaessetcomplex, Greenland, in intergrowths with chromite, as small marginal anhedra, and as large anhedrabetween chromite grains. He concludedthat rutile was produced by exsolution during meta- morphism, although the evidencefor this does not appear persuasive.In a section of chromitite from Fiskenaesset,the writer found rutile that appearsto have formed independentlyof chromite (Fig. 6). As- sociated chromite contains only 0.07 percent TiOr. Chromite from the Sittampundi Complex, India (Subramanian,1956) contains both exsolvedoriented rutile, and separatelarge anhedralgrains. Both were regardedby Subramanianas metamorphic in origin. The complex is metamorphosed; the anorthositic rocks enclosing the chromitites now fall in the am- phibolite facies.An analysisof a chromite from the complex shows 0.36 percentTiOr. It thus appearsthat the stableassociation of rutile with chromite low in TiO, is not unique to the Bush- veld Complex. In the Bushveld,in the Muskox intru- sion, and possiblyin the Fiskenaessetintrusion, rutile Fig. 5. Photomicrographin reflectedIight, oil immersion.Rutile (light gray) interstitial to chromite (medium gray). Chromitite, and chromite are associatedin part as cumulus min- Muskox intrusion. N.W.T.. Canada. erals. 146 CAMERON: BUSHVELD COMPLEX

Table l. Electron microprobe analysesof rutile and ilmenite rutile of type 6 appears to represent still another generation formed very locally from the magma dur- 1. 2. 3. 4. 5. ing the postcumulusstage. lio 98.51 98,9L 2 99.24 98.37 52.44 Analysesof rutiles from various samplesare given Ar203 0.05 0.04 0.08 0.05 0.01 in Table l. Small amounts of iron and chromium are C.203 0.87 0.6s 0.69 0.86 0.49 invariably present.

UgO 0.06 0.06 0.04 0.05 2.77

MnO n. d. n,d. n.d. n.d. 2.36 Ilmenite

Fe0* 0,41 0.17 0.23 0.47 41.38 Ilmenite associatedwith chromite,locally in a reac-

99.90 99.83 100.28 99.75 99.4s tion relation, has beenreported by Legg (1969)from

*Tota1 Fe a6 FeO. chromititesin the Merensky Reef at the Union Mine, Western Bushveld Complex. Macdonald (1967) re- I-4, Rutile. 1. Chromitic pyroxenite, upper pert of D unit, ported the occurrence of ilmenite and rutile in Jagdlust. Independent crystal. 2. Chromitite, top of H unit, chromitites on Farm Ruighoek in the WesternCom- Jagdlust, Indepef,dent crystal. 3. As for 2, inclusion in plex. Ilmenite is an uncommon mineral in the normal chromite. 4. Chronitite, F uniE, itrclusion in chtomitite. layered rocks of the Eastern Complex. I have not 5. Ilnenite fron norite belov llain Chromite Subzone, De Gtooteboom. found it in any of hundreds of samplesof chromitic rocks from various localitiesnorth of the Steelpoort Although in part rutile and chromite have crystal- River, including samples of chromitite from the lized as separatephases, the two minerals are closely Merensky Reef on farms Winnarshoek and Um- associated.In the part of the Critical Zone from the koanesStad. Reexaminationof the ilmenite reported baseto the top of the M unit, rutile is lacking or rare by Younger(1958) indicates that it is rutile.South of in non-chromitic rocks thus far examined.It is rela- the SteelpoortRiver, however,ilmenite has beenob- tively abundant in chromitic rocks of the F, H, and L servedin a few samplesof chromitic rocks from the units but is notably absentfrom non-chromitic rocks Main Chromite Subzoneon Annex Grootboom, and of the G and J units and the lower two-thirds of the in two thin chromitites below the Main Chromite M unit. Subzoneon De Grooteboom. In all thesesamples it Acicular rutile in bronzite or plagioclaseis found occurs as tiny grains interstitial to chromite. In mostly in chromite-free rocks of the O and higher chromitite from the Merensky Reef on Grootboom, units. Swarms of rutile crystalsappear to have been it forms scatteredgrains interstitialto chromite, with trapped in crystalsof the two silicates. or without associatedrutile, and a few inclusionsin The history of crystallizationof rutile in the chro- chromite. Rutile is much more abundant in the mitic rocks of the Critical Zone is analogousto that chromitite. however.than ilmenite. of ilmenite in many titanomagnetite-bearingrocks The chemistryof ilmenite hasnot beeninvestigated (Vincent, 1960;Buddington and Lindsley, 1964).In fully. One analysisis given in Table l. The MgO and these,a first generationof ilmenite was depositedas MnO contentsare noteworthy. discrete(independent) crystals, more or lesscontem- The great preponderanceof rutile over ilmenite in poraneously with titanomagnetite solid solution. the chromitic rocks is somewhat surprising in the Subsequently,at relatively high temperatures,a sec- light of experimentalinvestigations. Lindsley (1976) ond generationof ilmenite was formed by oxidation has recently reviewed phase relations in the system of titanomagnetite""and exsolutionof ilmenite to the FeO-FerO'-TiOr. Muan et al. (1972) summarized margins of the titanomagnetite grains. Finally, at known phase relations in the systemsFeO-AlrOr- lower temperatures,hence lower diffusionrates, ilme- TiOr, MgO-Al2Os-TiO2, FerTiOa-FeAlzOn-FeCrzOr, nite lamellae within the grains were pro- and MgzTiOo-MgAlrOn-MgCrrOo. Muan et al. duced by further oxidation-exsolution. In the rocks (197l) describedstudies of simplifiedternary systems of the Critical Zone, the independentrutile crystals involving TiO2, Al2Osor CrrOr, and MgO or FeO as and most of the large inclusionsin chromjte corre- components.In all systemsstudied thus far, at tem- spond to the first generationof ilmenite. The rutile peraturesof 1000' to 1300oC,the spinelfield is sepa- anhedramarginal and interstitialto chromite and the rated from the rutile field by a pseudobrookitefield, a acicularoriented inclusionsof rutile in chromite cor- rhombohedral oxide""field, or both. In systemscon- respond,respectively, to the secondand third genera- taining FeO or Fe2Or,ilmenite is the stable phase tions of ilmenite. In a few rocks. however.interstitial coexisting with spinel. The assemblagemagnetite- CAMERON. BUSHVELD COMPLEX rutile is apparently stable only at temperatureswell below the magmatic range (Lindsley, 1976). Addi- tional experimentalwork is evidently neededto un- derstandrelations in the complex systemFeO-MgO- Fe2Or-AlrOs-CrrOr-TiO2 represented by Bushveld FeO chromites.

Complex Ti-rich oxide At certain horizons,tiny grainsof a Ti-rich mineral have beenfound. The mineral is either a member of 20L the pseudobrookite seriesor the senaite-crichtonite 40 series,containing Cr, Fe, Al, Mg, Mn, Zr,Sc, Y, and Ti02 rare earths. Description and analysesare given in a Fig. 7. Weight percent FeO (total iron calculated as FeO) separatepaper (Cameron, 1978a). plotted against weight percent TiO, for chromites in 26 chromitites of the Critical Zone. For complete analysesof the chromites see Titanium in chromite Cameron (1977) Chromites in samples representing 129 strati- graphic horizons from top to bottom of the Critical 1977) occurring in trace amounts in norite on Farm Zone of the sectorof the EasternComplex north of Grootboom. It is an aluminian ferrian chromite con- the Steelpoort River have been analyzedby the writer taining 5.17 percentTiOr. by means of the electron microprobe. Analyses of The TiOz content of chromite might be expectedto chromitesfrom 64 horizonsare reportedin a separate vary with its bulk composition, but plots show no paper (Cameron, 1977).In addition, chromitesfrom correlation betweenTiO, content and proportions of 52 horizons in the Critical Zone south of the Steel- various cations.There is, however,some correlation poort River have been analyzed by D. B. Furgason betweenTiO, content and total iron content of chro- (1917) and chromites from three other horizons by mite, as displayedin Figures 7 and 8. the author. Resultsare summarizedin Table 2. Only Discussion ranges are given. A more detailed presentation is hardly justified becausethe TiOz content of chromite Studies of containing various proportions does not vary systematicallywith stratigraphicposi- of Ti and Cr, along with other cations,have now been tion, and the TiO, content of a given chromite is not a made for two principal groups of rocks-lunar ba- reliableindicator either of stratigraphicposition or of salts and associatedrocks, and terrestrial rocks. In differentiationtrends. In general,chromites from the some lunar rocks a complete range of compositions Upper Critical Zone north of the SteelpoortRiver are from ulvdspinel to Cr-rich spinels has been found, higher in TiO, than chromites from the Lower Criti- with no evidenceof a miscibility gap in the series cal Zone, but in each of the two zones there is a considerablerange of TiO, contents. It is also true that in general chromite in chromitites is lower in TiO, than chromite in associatedsilicate rocks, as indicatedin the table, but again rangesoverlap. Not included in Table 2 is a spinel (Furgason,

Table 2 TiO, contents of chromites, in weight percent

Chronites in Chromites in Chromitites Silicate Rocks

North of Steelpoort val1ey Ti02

Upper Critical zone o.52-I.29 o.64-2,24 Fig. 8. Weight percent FeO (total iron calculated as FeO)

Lover Critical zone 0.48-0.90 0. 31-2. 10 plotted against weight percent TiO, for chromites in 38 layers of chromitic and chromitites in the upper part of the D unit. The sequence includes the Steelpoort and Footwall Souch of Sreelpoort va11ey 0.50-2.35 o.42-2.47 chromitites. r48 CAMERON, BUSHVELD COMPLEX

(Gibb el al., 1970;Cameron, l97l; Champnesset al., clinopyroxeneand was resumedonly when the FeO, l97l; El Goresy et al., l97l; Keil et al., l97l; El FerOs, and TiO, contents of the residual liquid in- Goresy, 1976).In others, compositional gaps (Hag- creasedsufficiently to causeprecipitation of titano- gerty and Meyer, 1970)have been found. magnetite. In terrestrialextrusive rocks, spinelsranging from Hill and Roeder concludedthat the crystallization chromite to titanomagnetitehave been reported. In of spinel from basaltic liquids in near-surfaceenvi- basalticlavas of Makaopuhi, Hawaii, Evans and ronments is a function of a combination of factors, Moore (1968) found a complete seriesbetween the the most important being Cr content of the liquid, two end membersand inferredcomplete miscibility at oxygencontent or ferrous-ferricratio, MgO content, temperaturesnot above 1200'C. Spinelsintermediate TiO, content, temperature, and the presenceor ab- betweenchromite and titanomagnetitehave also been senceof clinopyroxeneas a stable phase.For intru- found in volcanic rocks by Gunn e/ al. (1970) and sive rocks total pressureshould be added to these, Thompson (1973), and a nearly complete serieshas becauseOsborn and Watson (1977)have shown that been reported by Arculus (1974). total pressuresignificantly influencesthe composi- Chromite-titanomagnetite seriesin true intrusive tions of spinels. rocks are apparently rare (Haggerty, 1976)but have As pointed out by Cameron and Glover (1973), been reported from (Mitchell and Clark, Irvine's explanation seemsapplicable to the Eastern 1976). A continuous serieswas found by Cameron BushveldComplex in a broad sense.Bronzite was the and Glover (1913)in mafic pegmatitesin the Eastern stablecumulus pyroxeneduring crystallizationof the BushveldComplex, but those bodieswere formed by Lower Critical Zone and most of the Upper Critical replacementof the normal layered rocks. In layered Zone. ln the few units of the Upper Critical Zone in mafic intrusions in general,spinels are either Ti-rich which cumulus Clinopyroxeneis present,there is no and Cr-poor, or Cr-rich and Ti-poor, as in the Bush- spinel.Chromite crystallizationterminated with for- veld. Where both are present,they occur at widely mation of the Merensky Reef, above which cumulus separatedstratigraphic horizons. clinopyroxene is prominent. Spinel crystallization The reasonsfor discontinuitiesin titanomagnetite- was not resumeduntil thousandsof feet of rock had chromite and ulvdspinel-chromite serieshave been accumulated,and Fe and Ti contentsof rest magma discussedby various authors (e.g. Kushiro et al., permitted crystallizationof titanomagnetite. 1970;Cameron, l97l; Buscheet al., 1972;Haggerty, The presentstudy indicates,however, that this gen- 1972;Nehru et al., 1974;El Goresy, 1976).A mis- eral explanation may not account for the low con- cibility gap in the series is obviously one possible tents of TiO, in chromites of the Critical Zone.The explanation. Irvine (1967), however, suggestedthat presenceof coexisting cumulus rutile, and, locally, the onset of clinopyroxenecrystallization in a mafic postcumulusrutile indicatesthat chromite was satu- magma could terminate the crystallizationof chro- rated with TiOr, hence under the conditions pre- mite, the incorporation of Cr in the clinopyroxene vailing, a miscibility gap may have existed in the reducingthe concentrationof Cr in the magma below series chromite-titanomagnetite. Furthermore, the the saturation level for chromite. Spinel crystalliza- presence of oxidation-exsolution intergrowths of tion would not be resumeduntil the concentrationsof chromite and rutile indicatesthat the solubility of Ti Ti and Fe rose to levelsat which titanomagnetiteis a in chromite decreasedduring the subsolidusstage, stable phase. This explanation has been applied to presumablybecause the systembecame more oxidiz- spinels in certain lunar rocks by Cameron (1971), ing with fall in temperature. Buscheet al. (1972),and Nehru et al. (1974).Support The nature of conditions in the Bushveldmagma is is given by the experimentalwork of Hill and Roeder partly indicated. Using the Sato cell, Flynn and Ul- (1974), who studied the crystallizationof a basaltic mer (1972) obtained 1075'C and l0-12'2as the equili- melt under varying conditions of temperature and bration temperatureand fugacity for the as- oxygen fugacity. In the temperaturerange 1050oto semblageat the contact of the Lower Critical Zone 1200oC,at oxygen fugacitiesin the range l0-o' to and Upper Critical Zone. lt is not certain, however, 'g, l0 they obtained a continuous series of spinels whether these values represent cumulus, post- from chromite to titanomagnetite,the Cr-rich mem- cumulus,or subsolidusequilibration. Guilbert (1962) bers forming at high temperatures and low oxygen studied the crystallization of melts of mixtures of fugacity. At still lower oxygen fugacities, however, labradorite,bronzite, and chromite with proportions spinel crystallization ceasedwith the appearanceof and compositions appropriate to the D unit of the CAMERON: BUSHVELD COMPLEX t49

CriticalZone. The temperaturerange of the experi- References mentswas 1100"to 1550"C.Oxygen fugacity was Arculus, R. J. (1974)Solid solutioncharacteristics ofspinels: pleo- controlledover a rangefrom 10-12atm at I 100' to naste-chromite-magnetitecompositions in some island-arcba- l0-? atm at 1550'C,so as to avoid changesin the salIs. Carnegie Inst llash. Year Book,73,322-327. (1964) ps2+/ps8+ratios of the phases.A labradorite-ortho- Buddington,A. F. and D. H. Lindsley Iron-titanium and syntheticequivalents. J Perrol, J, 310-357. -plagioclaseeutectic was found at 1215'C, Busche,F. D., M. Prinz, K. Keil and T. E. Bunch (1972)Spinels at 36 percentorthopyroxene, 6l percentplagioclase, and the petrogenesisof some Apollo 12 igneous rocks. Am. 3 percentchromite. In viewof uncertaintiesabout the M ineral..57. 1729-1747. compositionof the Critical Zone magma,these val- Cameron, E. N. (1971) Opaque minerals in certain rocks from ues Apolfo 12. Proc Second Lunar Conf. I, 193-206. The M.l.T. cannotbe takenliterally, but theyprobably are Press. not far from the actualvalues. Guilbert calculated - (1964) Chromite depositsof the easternpart of the Bush- that for precipitationof chromitealone, a minimum vefd Complex. ln S. H. Haughton, Ed., The Geology of Some contentof about 1200ppm Cr was requiredin the Deposits of Southern , 2, l3l-168. melts.Cameron (1978b) has calculated that the Cr - (1975) Postcumulusand subsolidusequilibration of chro- contentof the Bushveldmagma was mite and coexistingsilicates in the EasternBushveld Complex. at least 1000 Geochim Cosmochim Acta. 39. l02l-1033. ppm.This is within the rangeof 540-1600ppm calcu- - (1977)Chromite in the central sectorof the EasternBush- latedby lrvine (1977) for the Muskox magmaduring veld Complex, . Am. Mineral.,62, 1082-1096. crystallizationof chromiteand olivine. - (1978a)An unusual titanium-rich from the Calculationsof total pressurerest on knowndata Eastern Bushveld Complex. Am Mineral., 63, 37-39. - (1978b) for thicknessof roof rocks, The Lower Zone of the EasternBushveld Complex about 14,000feet, and in the Olifants River Trough. J. Petrol., 19, 437-462. uncertaininferences as to the heightof thecolumn of - and M. E. Emerson (1959)The origin of certain chromite magmapresent between the roof and the top of the deposits in the eastern part of the Bushveld Complex Econ. cumulatepile at the timethe Critical Zone was form- Geol.54.ll5l-1213. ing. The problemwith the latteris that recentwork - and E. D. Glover ( 1973)Unusual titanian-chromianspinels (Gruenewaldt,1973) has indicated from the EasternBushveld Complex. Am Mineral.,J8, 172-188. a heaveof magma Champness,P. E., A. C. Dunham, F. G. Gibb, H. N. Giles, W. S. during the formationof the Main Zone. The best Mackenzie, E. F. Stumpfl and J. Zussman (1971) Mineralogy estimateis that total pressurewas between 3 and 4 and petrology of some Apollo I I samples Proc. Secowl Lunar kbar. Science ConJ., l, 359-376. M.l.T. Press. Cousins,C. A. and G. Feringa(1964) The chromite depositsof the westernbelt of the BushveldComplex. ln S. H. Haughton, Ed., Conclusions The Geology of Some Ore Deposits of Southern Africa, 2, 183- 202. 1. Chromiteand rutile were stableas coexisting DeWet, J. F. and J N Van Niekerk (1952) Chromite investiga- phasesduring accumulation of manychromitic units tions-Pt. V: The total analysis,composition, and structureol Transvaal chromite. J. Chem. Metall. Mineral. Soc. South Af- of the Critical Zone of the EasternBushveld Com- rica.52. 10-18. plex. Ef Goresy, A. (1976\ Oxide minerals in lunar rocks. Mineral. Soc. 2. Chromitewas saturatedwith TiO, at levelsof Am Shorr CourseNotes, J, EG-l to EG-46. 2.5 percentTiO, or less,hence a miscibilitygap may -, P. Ramdohr and L. A Taylor (1971)The opaqueminerals existin theseries chromite-titanomagnetite under the in the lunar rocks from Oceanus Procellarum. Proc Second Lunar Science Conf., 1,219-236. M.l.T. Press. conditionsof crystallizationof the CriticalZone. Evans, B. W. and J. G. Moore (1968)Mineralogy as a function of 3. In general,rutile contentand TiO, contentof depth in the prehistoricMakaopuhi tholeiitic lava lake, Hawaii. chromiteincrease upward in the CriticalZone. Conrrib Mineral Perrol., 17, 84-115. 4. During the subsolidusstage, there was lossof Flynn, R. T and G. C. Ulmer (1972)Petrogenesis of the Bushveld TiO, from chromitesdue to oxidation-exsolution. Complex: pyroxene-spinel-feldspar peritectic? (abstr.) Geol Soc Am. Abstracts with Programs, 4, 507. Furgason,D. C. (1977) Petrologyand geochemistryof the south- ern sector,southeastern Bushveld Complex. M.S. Thesis,Uni- Acknowledgments versityof Wisconsin. Investigationsof which this work is a part havebeen supporled Chisler, M. (1970) Premetamorphicfolded chromite depositsof largelyby NSF grants cP-5523, cP-2901, and DES75-14829and stratiform type in the early Precambrian of West Greenland. by grantsfrom the WisconsinAlumni ResearchFoundation. This Mineralium Deposira,5, 223-236. support is gratefully acknowledged. Gibb, F. G. F., E. F. Stumpfl and J. Zussman (1970) Opaque The writer is also indebted to Drs. G. Allard, C. V. Guidotti, minerafs in an Apollo 12 rock Eanh Planet. Sci.Leu., 9, 217- T. N. Irvine, and W. C. Shanks for helpful comments based on 224 their critical reviewsof the paper. Gruenewaldt, G. von (1973) The Main and Upper Zones of the t50 CAMERON: BUSHVELD COMPLEX

Bushveld Complex in the Roossenekal area, Eastern Transvaal. Lindsley, D. A. (1976) Experimental studies of oxide minerals. Trans Geol. Soc. South Africa,76,207-221. Mineral. Soc. Am. Short Course Notes, 3, L-61 to L-88. Guilbert, J. M. (1962) The origin of the chromitires below the Macdonald, J. A. (1967) Evolution of part of the Lower Critical SteelpoortMain Seamof the BushveldComplex. Ph.D. Thesis, Zone, Farm Ruighoek, Western Bushveld."/ Petrot., 8, 165-209. U niversityof Wisconsin. Mitchell, R. H. and D B Clark (1976)Oxide and sulfidemineral- Gunn, B. M., R. Coy-Yll, N. D. Watkins, C. E. Abranson, and J. ogy of the Peoyuk , Somersetlsland, N.W.T., Can- Nougier (1970)Ceochemistry of an oceanite-ankaramanite-ba- ada. Contrib. Mineral. Petrol.. 56. 157-172. salt.suite from East Island, Crozet Archipelago. Contrib Min- Molyneux, T. G. (1972)X-ray data and chemicalanalyses of some eral. Petrol . 28. 319-339. titanomagnetite and ilmenite samples from the Bushveld Com- Haggerty, S. E. (1972) Solid solution, subsolidusreduction, and plex, South Africa. Mineral Mag ,38, 863-871. compositionalcharacteristics of spinelsin some Apollo l5 ba- Muan, A., J. Hauck and E. F. Osborn(1971) Equilibrium relations salts. Meteoritics, 7, 353-370. among phasesoccurring in lunar rocks. Proc. SecondLunar Sci. - (1976) Oxidation of opaque mineral oxides in basalts Min- Conf., l, 497-505. M.I.T. Press eral. Soc. Am Short CourseNotes, J, Hg-l to Hg-100, Hg-l0l to and T. Ldfall (1972) Equilibrium studies with a Hg-300. bearingon lunar rocks.Proc Third Lunar Sci.Conf.,1, 185-196. - and H. O. A. Meyer (1970) Apollo l2: Opaque oxides. M.l.T. Press. Earrh Planet. Sci. Lett., 9, 379-387. Nehru, C. 8., M. Prinz, E. Dowty and K. Keil (1974)Spinel-group Hedlund, D. C., J. F. de Couto Moreira, A. C. F Pinto, J C. G. mineralsand ilmenite in Apollo l5 rock samples.Am. Mineral., Da Silva and G. V. V. Souza (1974) Stratiform chromitite at s9,1230-t23s. Campo Formoso, Bahia, Brazil. J. Res. U. S. Geot. Suru,2, Osborn,E. F.and E. B. Watson (1977) Studies of phaserelations 55t-562. in subalkafine volcanic rock series. Carnegie Inst. Wash. Year Heiligman,H. A. and H. M. Mikami (1960)Chromite. In J. L. Book. 76. 472479. Gif fson, Ed., Industrial Minerals and Rocks,3rd ed.. 243-258. Smith, C. H., T. N. lrvine and D. C. Findlay (1967) Muskox Hill, R. and P. Roeder (1974) The crystallizationof spinel from Intrusion. Geol. Suru. Can., Map l2l3A. basalticmagma as a function of oxygen fugacity.J. Geol ,82, Subramanian,A. P. (1956) Mineralogy and petrology of the Sit- 709-729. tampundi Complex, Salem district, Madras State, India. Bill/. lrvine, T. N. (1967) Chromian spinelas a petrogeneticindicator, Ceol Soc Am .67.317-390. Part 2 Petrologicapplications. Can J. Earth Sci,4,7l-103. Thompson, R. N. (1973)Titanian chromite and chromian titano- - (1974) Crystallizationsequences in the Muskox intrusion magnetite from a Snake River Plain basalt, a terrestrialana- and other layeredintrusions-ll. Origin of chromititelayers and logue to funar spinels,Am Mineral , J8, 826-830. similar deposits of other magmatic . Geochim. Cosmochim. Verhoogen,J. (1962)Oxidation of iron-titanium oxidesin igneous Acta. 39.991- 1020. rocks."/. Geol..70. 168-181. - (1977)Origin of chromitite layersin the Muskox intrusion Vincent, E. A. (1960) Ulvdspinel in the Skaergaardintrusion, and other stratiform intrusions: a new interpretation. Geology,5, Greenland NeuesJahrb. Mineral Abh., 94,993-1016. 273-277. Waal, S. A. de (1975) The mineralogy, chemistry, and certain Keil, K., M. Prinz and T. E. Bunch (1971) Mineralogical and aspects of reactivity of chromitite from the Bushveld igneous petrologicaf aspectsofApollo l2 rocks. Proc. SecondLunar Sci complex. National Institute for Metallurgy, Rep. 1709.Johan- Conf., I,329-342. M.l.T. Press. nesburg Kushiro, K., Y. Nakamura and S. Akimoro (1970)Crystallization Younger,J. A. ( 1958)Chromitites in the lower part of the anortho- of Cr-Ti spinelsolid solutionsin an Apollo l2 rock, and source site series,Bushveld Complex, Union of South Africa. M. S. rock of magmas of Apollo 12 rocks (abstr.). Trans Am Thesis.Universitv of Wisconsin. Geophys. Union, 5 l, 585. Legg, C. A. (1969) Some chromite-ilmenite associationsin the Manuscript receioed, February I3, 1978; Merensky Reef, Transvaal. Am. Mineral , 54. 1347-1354. acceptedfor publication, May 18, 1978