ALPINE-TYPE CHROMITE in NORTH BORNEO, with SPECIAL REFERENCE to DARVEL BAY1 Cuenr,Ps S. Hurchrsox, Department Ol Geology, Uruiue

Ameriean Mi*eralogist I Vol. 57,pp. eas-dSZ irszzl

ALPINE-TYPE CHROMITE IN NORTH BORNEO, WITH SPECIAL REFERENCE TO DARVEL BAY1 Cuenr,psS. HurcHrsox, Departmentol Geology,Uruiuersitg ol MalaEa, Kuala Lumpwr, Malaysia.

Assrnesr chromite occursas layers and pods in dunite and serpentinitelenses within most of the peridotite outcrops of saibah.The crystals are anhedral,rounded, cataclastic,and unaltered.

INrnooucrroN

volume of chromite for economic exploitation. rn all the investigated areas the chromite occurs in dunite and serpentinized dunite bodies and lenses (forming about 5/o of the ultramafite total) which occur always within Iarger peridotite outcrops.

lPresented at the Twelfth pacific science congress, canberra, Australia: August24th, 1971. [I{utchison, l97l, (abstract)l. .: &35 836 CHANLES S. HATCHISON

Legend

f.owER ^tPtrt //ffieow,oic- \ lERtARv l-xAFlc/ IO I ULTRAI'AFIC cRETAcEousl,uo'ot't-,1 ggrp,gt65

Fi:IIil I€lonorPhicoc.ooic i.:r:;::l bosemtnl

a Moinoccut,nca of ItJl chromite

SUIU SEA

Sondolon

-lt!{ -**o' Xino botongot . Rt ... DE llr PEiIINSUL A Voll.t Lohod

OARVEL 8AI

KALIIIAI{TAN

occurr^enceof Frc. 1. Outline geologic map of Sabah showing the main ultra,ma^fic,/mafic chromite in relation to thle crystalline basement and the alpine ere left unshaded' complexes (modified a1ter Kiri, 1968). Sedimentary formations

(Fig' 2)' They emplacementthan the other ultramafites of the region the have also shown (Hutchison and Dhonau, 1969 and 1971) that peridotite of Tabawan island was of a higher emplacementtempera- the ture and has causeda thermal auleole in the Silumpat Gneissfrom regional greenschist,through almandine amphibolite,to hornblende ALPINE CHROMITE IN BONNEO &]7

AJ =90- E.

3 c-

d a

-N

;@ HO

;: AX

9s a bo :- ;*c / o -7 - )t1.[.

g 11

6 83t3 CHARLES S. HUTCHISON granulite faiies. By contrast,the younger Saddleisland group ultramafite was emplacedin equilibrium with the greenschistfacies condi- tions, and it is this emplacementalone in the Darvel Bay area that containschromite ore. Texruns The chromite occurs as layers and bands, ranging in thickness from a few millimeters to several meters, and as lenses,pods, and layer-zonesof disseminatedcrystals. The layershave beenfolded and boudinaged,but in several instances original cumulate textures are preserved.On Katung-Kalungan island (Fig. 2), for example,the chromite bands are sharply folded, but even after folding they pre- serveexcellent graded layering reminiscentof an originally stratiform origin (Fig. 3). In the Labuk Valley, 800 meters east of Katai on Bukit Lumisir, the distribution of narrow bands of fine grained chromitein severaldunite bodiessuggests crystal settling (Collenette, 19M). Generally, however, the distribution of chromite is tectonically

Frc. 3. Left: sharp fold in chromite-serpentinite layers. The chromite layers show high relief on the natural outcrop surface. (Specimen J 1173 from Kalung- Kalungan island).

Center: End view of the same specimen on a polished surface showing graded layering of the chromite, inverted by folding.

Right: Chromite disseminations and segregations in serpentinite. The distribution probably resulted from tectonic emplacement, and the original layering has been largely obliterated. (Polished surface of specimen J 1187 from Laila island). The white bars represent 3 em on each adjacent specimen. ALPINE CHROMITE IN BORNEO 839

controlled,and original stratiform structuresIargely obliterated (Fig. 3). In some areas,for exarnpleat Bidu-Bidu in the Labuk Valley area, the ultramafic foliation is not parallel to, and is superimposed on an original stratiform layering.Newton-Srnith (1962) took this to indicate that the chromite had a magmatic-settling origin before tectonic emplacement. Individual localitiesare listed under Table l. The chromite crystals from all localities in Sabah have rounded anhedral outlines (Fig. a). All are cataclasticto some degree.The massivepodiform bodiesare also brecciatedand healedby the serpentinite matrix (Fig. a). Sabah chrorniteis remarkably free from hydrothermally altered rims as have been describedby Mih6,lik and Saager (1968). The absenceof compositionalzoning is indicated by similar reflectivity values across crystals and by sharp and single X-ray diffraction peaks.Irregular patchesand zonesof ferritchromite as describedby Engin and Aucott (1921) have been observedvery rarely in Sabah. From the whole collection,only two examplesof higherreflectivity ferritchromitehave beenfound (A and B in Fig, 4).

Tnecn Er,nlrnrrs rN Couurny Rocrs

Method

Ana,lyseswere by comparison with U. S. Geol. Surwey PCC-I, taking Ti = 44, - Cr 2600, and Ni - 2300 ppm and W-t taking K - 5310, Ti - 6400 ppm. Whole rock specimens,pulverised in a tungsten carbidelined Tema mill vessel to avoid contamination, were made into pressed pellets backed with boric acid after the manner of Norrish and Chappel (1g67) and analysed on a Philips X-ray vacuum spectrometer. K' lines were used throughout and the pulse height analyser carefully set to ensure maximum peak to background ratios. Operati,ngcondttior*

Counting time Sensitivity Element Tube Crystal kV mA sec. Path Detector per l0 ppm K Cr PET 44 26 40 vacuum flow c. 2.5 counts/ 4 sec. Ti Cr LiF 44 26 10 vacuum flow c. 49 counts/ I sec. Cr W LiF 44 26 10 air flow c. 11 counts/ 1 see. Ni W LiF 50 20 10 air scintill- 31 counts/ ation 1 sec. counter

Background meafltrements were made by averaging the two readings made at -+ 1.7"2 0 840 CHARLES S. HATCHISON

Ils€ €i t rir* fI{*qa si f,,tE;i is +,;e ;*ff f,i+;e i; li++ ru;iii t* is ea;tiet* t*esiess EEE.EE.E.E. g EEEEE ;;;E ! ; Y fi ; P jji.:H ! r ,E ; .1 fl ; h H : I iE fl # f HHH;e- fr H'*cH-*aH E;d e H I H$ H $ $ 3 EEEff E E Ee! !eE 5e.Ef ! F EEE E E

ESR ddd

@oooooo$d@ ddNos s @ @ < !) s o

oo iddct.; Ai oi d d .i A; oi d Ai oi doi d d o; "i o<

E s=!s: P i € : € e ? b o @ d ;": d N .t

+c'c;dd d ci d ct d r; di d d N o +qi d; d; oi

ocN6@ N FN+{O s + r

s:3., 3 R ts

RSES $o qJ9?t "l ": t ; +zo@o @

o06 + r o. g E R b F * 3g 3 K t e i si d e P N ft sN * K: e REB'' D g S q q q C e t { e qq q Q 1 s s $ s g s 3 g 33 s g ;

=d E** F"EIIP g $ o N N E .s a$ q B g E==9- A 3 2: I ! P ; - Ev I I 2 ALPINE CHROMITE IN BONNEO B4lr s

D F

0mml F: |ERR|TCHR0XITE lsJ+J xtr.ntx sEnPEilililrE

Frc. 4. Drawings from polished sections of chromite occurrerrcesin Sabah. The matrix is generally serpentinite. F - ferritchromite with considerably higher reflectivity than chromite, which is shown hatched. A : specimen 7182, accessory hydrothermally altered chromite from the Ranau Luhan road, Kinabalu area. B - J 1186, a,ccessorychromite with small areas of ferritchromite from Laila island. C - J t27, rounded chromite grains from Bidu-Bidu in the Labuk Valley. D - BB 120, massive fractured chromite from Banggi island. E and F - NB 4169, massive fractured chromite from the Kinabalu area, E from the edge and F from the interior of a layer.

In view of the apparent absenceof chromite, except as a rrery minor accessory,in the ultramafic bodies of Sakar and the outer islandsof Tabawanand Maganting (Fig. 2), a selectionof rock specimens from all the Darvel Bay ultrarnafic outcrops$ras analysed for chromium and related elements.The results &re shown in Figure 5. SpecimensI through 17, and 19 and 27 are ultramafites.At the same time a number of Eilumpat Gneissspecimens, 18 and 20 through 26 wereanalysed. The roek contentsof chromium,nickel, and titanium, when com- paredin relation to the potassiumcontents, show the sharp distinction betweenthe ultrarnafic rocks and the Silumpat Gneiss (Fig. 5). Ex- cluding number27 f.romthis study becauseit is abnormally enriched in potassium,we find that the Darvel Bay ultramafic rocks have the f ollowingaverage contents: u2 CITARLES S. HUTCHISON

ULTRAMAFICROCKS

I o5 o4 o6 oo? o2 o I ol A rl I

THOTEITICBASALTS l5 l5 II l7 lr]

I I 2l o la

?o ^ AI I I o

THOLEITICBASALTS IJLTRAMAFICROCKS

t0 600 1,000 Polossiump0rls per mrllion by qerqhl 000t% 00 % 0l%

Frc. 5. Plot on logarithmic scales of the Cr, Ni, Ti, and K contents of some Darvel Ba.y rocks. The dashed vertical line is an arbitrary division between the potassium-based fields of ultramafic rocks and tholeitic basalts. The Silumpat Gneiss extends to the right from number 18, but 27 is an ultramafic rock anoma- lously rich in potassium. The Silumpat Gneiss is characterised by its richness in potassium and titanium; the ultramafites by their richness in chromium and nickel. 1 = J 1080, Dunite from Tabawan.z - J 1095, Serpentinite from Tabawan. 3 - J 1191,Pyroxenite from Nipa-Nipa. 4 - J 1168, Peridotite from Maganting. 5 - J 1103,Peridotite from Tabawan. 6 : J 1083,Peridotite from Ta- bawan. 7 = J 1101, Peridotite from Tabawan. 8 - J 1175, Peridotite from Katung-Kalungan. I = J lzm, Serpentinite from Saddle. 10 - J 1245, :ullrw- basite from Silam. 11 - J 1329Peridotite from Sakar. L2 - J 1182,Serpentinite from Baik. 13 - J 1200, Pyroxenite from Giffard. 14 - J 1319, Peridotite from Saka.r. 15 - J 1f96, Pyroxenite from Enam. 16 = J 1088,Dunite from Tabawan. 17 - J 12i5, ultrabasite from Sakar. 18 - J 1344,Silumpa,t Gneiss from Silam. 19 - J 1185,Py'roxenite from Laila. 20 : I 1075,Silumpat Gneiss from Tabawan. 2l = J 1324,garnet amphibolite from Sakar. 22 = J 1050,Silumpat Gneiss from Silumpat. tJ - J 1060, Silumpat Gneiss from Silumpat. 2t4 - J 126L, Silumpat Gneiss from Silam.25 - J !n, Silumpat Gneiss from Tabawan.26 - J 1337, Silumpat Gneiss from Sakar. 27 - J 1215.,Pyroxenite from Enam. ALPINE CHBOMITE IN BORNEO 843

K 74 ppm, ranging from l0 to 240 (excluding2Z) Ti 412ppm, ranging from 60 to ZB0 Cr 2770ppm, ranging from t4b0 to 56g0 (excluding1g and2T) Ni 1531ppm, ranging from 120to 2Bb0(excluding lg and,27). When comparedwith the worldwide averageultramafic values of - Goles(1967): K 200:t, Ti = 300,Cr = 2400,and Ni: 1500=L ppm, the Darvel Bay ultramafic rocks seemremarkably low in potassium and high in chromium.This might be taken to indicatethat they representa suite of rocks strongly residual in characterand to have come accordingly from a Mantle region from which much basaltic magma has already beenwithdrawn by partial melting. It is interestingto note that althoughno segregationsof ore occur on the islandsof Sakar,Tabawan, and Maganting,yet the ultramafic rocks from theseislands are often as rieh in chromium as the rocks from the Saddleislands. Saddle island Group specimensin Figure b are 3, 8, g,10,12,13, 15,19, and 22. Thesedo not differ significantly in their distribution of Cr, Ni, Ti, and K from the ultramafic masses of other Darvel Bay areas.Of course,the speciniensthat were selected for analysisdid not contain layers and segregationsof chromite.Such layers and segregationsoecur only in the Saddleisland Group and on Mount Silam. Hutchison and Dhonau (1969) showedby structural data (summarisedon Fig. 2) tha| the deformationof the Saddleisland Group ultramafic body is later than elsewherein Darvel Bay. Ac- cordingly it must be concludedthat not all ultramafic emplacements are of the sameage and presumably not all from the sameparb of the llantle. The Saddle island Group ultramafites have come from a Mantle area that had a considerabledevelopment of ehromite layering, and Hutchison and Dhonau (196gand 1gZ1)showed that this was a cold emplacement.On the other hand the ultramafite of Tabawan was of sufficienttemperature to have raisedthe Silumpat Gneissin a thermal aureoleto as high as hornblendegranulite facies. Specimens18 and 20 through 26 arc of Silumpat Gneiss,which Hutchisonand Dhonau (1g6gand 1gZ1)believe to be metamorphosed oceanicbasaltic crust. For the specimensanalysed, the following aver- agecontents have beenfound: K 653 ppm, ranging ftom 220 to 1500 Ti 5778 ppm, ranging from 1760to 12690 Cr 232 ppm, ranging from 50 to 240 Ni 96 ppm, ranging from B0 to 2G0 These values cornparereasonably well with the averageworldwide U4 CHARLESS: HUTCHISON tholeitic basalt: K 5400 ppm with a lower limit of about 500 ppm (Manson,1967) ; Ti 9900 ppm (Manson, 1967); Cr 218 ppm (Prinz, 1967); Ni 85 ppm (Prinz, 1967).From a eomparisonof thesevalues, the Silumpat Gneiss can well be regardedas typically tholeitic in character. Metamorphism to amphibolite and granulite faeies has causeda partial depletionin potassiumand perhapstitanium. One of the specimensJ1324, is of a Eilu,ntpatGneiss inclusion in the ultramafic mass of Pulau Sakar. It is now a garnet amphibolite, and although it must have been considerablymodified by the enelosing ultramafic mass,its charactersas shownin Figure 5 (No. 21) are still that of the Silumpat Gneiss.It is hard thereforeto escapethe con- clusion that the Silumpat Gneiss is distinct both chemieally anil mineralogicallyfrom the ultramafic bodiesof the Darvel Bay area. Dr T. P. Thayer (personalcommun.) has questionedour interpreta- tion that the Silumpat Gneissis divorced in age and consanguinity from the ultramafic bodiesbecause as he has noted elsewhere,Thayer (1967), he regards "the ultramafic and gabbroic parts of ophiolite complexesas beingidentical with, and part of, the alpine igneousrock suite." However the problem has beento a large extent resolvedby Leong (1971) who has reeentlyre-mapped large areasof the hinter- land of Darvel Bay. He has found that there is a distinct Crystalline Basementwhich includesthe Silumpat Gneissand is, by radiometric and stratigraphic evidence,of pre-Lower Triassic age' In addition there are distinct occurrencesof banded gabbrosassociated with the ultramafic rocks and these form the peridotite-gabbro complexes which post-datethe Crystalline Basement.The problemin Sabahhas arisenfrom the fact that the bandedgabbros of the alpine complexes have in the past beenconfused, especially by Fitch (1955,and 1970)' with the banded amphibolitegneisses which constitutethe Silumpat Gneissof the CrystallineBasement. Crrnourrp Crrplrrstnv A numberof analysesof chromitesseparated from ultramafic rocks of Sabahhave beenfound in the literature.These are given in Table 1 togetherwith a few analysesby the presentauthor by X-ray fluorescence.The conversionto unit cell cationic proportionsis best illus- trated by an example:NB5452 Weight Molecular Cationic Oxide percent weight proportions CrO, ,7, 54.5 75.0r 0.7r70 Alo, 1/2 8.9 50.98 0.1745 ALPINE CHROMITE IN BORNEO u5

Total iron expressed as FeOr rz, 16.0 79.85 .2003 less .0003for ilmenite imPurity : .2000 Mgo 16.0 40.32 .3968less t+ X .0248for serpentine impurity : .3968- .L122: .2846 Total cations 1.3761 sio, 4.5 60.09 .0748 Tio, 0.03 79.90 .0003

For the basic formula Rs*16R,*sO32, then niultiply by 24/l.816l, giving (crp.6Al3.6Feo.s)ro (Fq.eMg5.6)so32. The total iron of B.b is first

P urifi catio n B eI ore Analg si.s

samples were crushed and sieved -120 + 280 u. s. standard, washed as often as necessary with water to remove fine powder, then dried. The very magnetic fraction was collected in a Frantz Ferrofilter and. discarded. The powdlr was now slowly passed through a Frantz isodynamic separator with a side tilt of 15" and a forward tilt of 25. (Rosenblum, lgbg) and the chromite was collected within the range + 0.1 A, and -0.5 A. The non magnetic fraction was usually serpentine or olivine. If the chromife fraction visibly contained some green maierial, it *u, re-nrn at 0.4 or 0.45A.

X-ray F luorescenceAnalgtical M ethod,

'obtainable from crescent Dental Manufaituring co., 77ffi west 4zth st., Lyons, Illinois, 60534,USA 846 CHARLES S. HATCHISON then the simple X-ray dilution procedure is capable of giving satisfactory results. This is especially important because of the difrculty of putting chromite into solution. one important point should be noted and that is that the Fe,Oe calibration curve between the two standards had a negative slope, showing the overall importance of ma,trix effects (Mitcheu and Kellam, 1968). Accordingly it is possible that iron determinations by this method are less than perfect. But it appears that Cr, Al, Mg, Si, and Ti determinations are good.

Instrumental Settings for the Elemental Determinations

Counting time Average sensitivity |ot l/s in Element Tube Crystal kV mA sec. chromite

CrzOa wLiF448 10 603 counts Per sec. AbOa Cr PET M 26 20 20 counts Per 2 sec. Mso Cr KA? M 26 40 6 counts Per 4 sec. FerOs wLiF508 10 790 counts Per sec. TiOr Cr LiF 44 16 10 6700 counts Per sec. SiOz Cr PET M 26 20 88 counts Per 2 sec. AII determinations were with a flow counter and vacuum path. The background of each peak was determined at -F1.6' 20from the peak position for each element except in thq case of iron where it had to be at -+2.3' because of the close position of the m&nga.DeseK F peak.

Discussiwt, The analysed chromitesof Table I are shown on Figure 6. The majority of sabah chromitesfall in the field of aluminianchromite but a few are in the chromian spinel compositionalrange of Stevens (1944),and a few can be consideredas of metallurgicalgrade. Figure 7 showsthat with only three exceptions(BW 68, MW 12(32)' and BB 120) the Mg:Fe ratio in R'* lies betweenthe small range of 4.7:3'3 and 6.0:2.0.This remarkably small range in Ri- while Cr:Al ranges from 12.5:3.0to 6.8:9.1(Table 1) is remarkable.Thayer (1970)has brought attention to this very fact that as alpine-type chromites becomeless and lessrich in chromium,they becomericher in aluminium and not in iron; indeedthe iron characteristicallyremains ap' proximately constant.In contrast, stratiform chromitesbecome pro- gressivelyricher in total iron as their chromium content falls. The sabah chromites are in this respect typically Alpine. Another inter- esting differenceis that stratiform chromiteschalacteristically have a unimodal and restricted chromium range. Alpine type, or podiform chromites,have a much wider range of chromium content which is characteristicallybimodat (Thayer, 1970). The range of chromium in the Sabah chromites (fig. 6) is considerablywider than in any stratiform body and there is a suggestionof a bimodal chromiumcon- ALPINE CHROMITE IN BORNEO u7

P ifr

,( 3

Cr,5( Mg,Fc lo Ora {"" .-rrn,ro.-l GRADES ' /l ----F\i i',lr',iliJ - 7,.61','ir6i ox_ Fcr6( lrg, Fc)s 052 \ z (1t19, cHR0ItAtl Alrs Fcls032 SPIIIEL

r0 9 E ? 6 5 4 3 Al+3 cATtol{s pER uNtT cELL Frc. 6. Distribution of Sabah chromites in the Cr,-Alr-FeruRt*ru triangular classification. The actua^lcationic proportions are plotted in only a small portion (Ieft) of the complete triangle (shown right).

+1 (Cr,Al),UMgrO,a t" 0.. iJ "n. MAGNESIOCHROMITE (Cr * Al) CATIONSPER UNIT CELL MAGNESIOFERRITE t2r0864

J J U U () (J F F6 2z z. f G, E. U gr4 4' at o T\EVENS, /944) z z. o o F |-. A< (.) o o + N E^ U afi o2468t0t2 t4 t6 CHROMITE + -3 MAGNETITE (cr,Ar)t6 Fe8'oj2 Fe' CATIONSPER UNIT CELL +a +) Fe,j Fea- 0 ,a Frc.7. Diagram to show the rema,rkable lack of variation in Mg:Fer* ratio for the same chromitm. Stratiform chromites show a much wider range, 848 CHARLES S. HUTCHISON tent with one mode around 7.9 and the other around 11.3 chromium cations per unit cell.

Pnvsrcer,Pnopnnrrus Unit CeUEdge Stevens(1944, Fig. 2) plotted percentCreOs against cell edgeo in A and found, with a few significant exceptions,a good linear fit. His data can be recalculatedto give the following regressionequation: percentCrzO's in chromite : 301.85(oA) - 2445.66 His data consistedof 25 measurementsof chromites ranging from CrsO30.0 to 59.5percent but excludingtwo sampleswhich had high ferric iron contents.But I do not feel that this was a valid justification for their exclusion,for the difficulty of putting chromite into solution for analysismust cast doubt on chromiteFe3*:Fe2* determinations. Stevens(1544) must have beenfortunate in his choiceof data, for the Sabahchromites do not fit well on his curve.Indeed it is surprising that a linear relationship between Cr2O3 and o exists, for in the chromiteunit cell (Cr, Al, Fet*)ro (Fe2-,Mg)3O32 there are five var- iableswith Cr, Fe3*,and Mg of roughly equivalentionic radius (0.63, 0.64, and 0.66 A respectively) and Al (0.51 A) and Fe2* (0.74 A). Hence it should be expectedthat variation in Al and in Fe*2would havemuch more controlon the unit cell edgethan Cr. Again we would expect to find a different relationship in stratiform chromites, where total iron increasesas Cr decreases,than in podiform chromites,in which aluminiumincreases as chromiumdecreases. In the presentstudy severalchromites listed in Table 1 were measuredfor unit cell edse (o) on a diffractometer.

Chronrite mixed with an almost equal amount of potassiurn bromate was smeared on a glass slide and run with CuKa radiation at 1/8" per minute from 61.5 to 63.5' 2e. The 2e difference between KBrOe K at22l at 61.61 and the chromite K a,M4 at approximately 63.16' is added to 61.61 to give the preeise chrdmite 20 K at044.From this, a2 = 32 ffo*. The results are tabulated in Table 1. Chromite peaks for a.ll Sabah specimens are sharp and well resolved indicating the homogeneous nature of each specimen and the gene.ral absence of hydro- therrnal alteration to ferritchromite. The variation of o in A with Al and Cr cations per unit cell is shownon Figure 8. The Al contentin R*3is consideredto be the most important variant becauseof the small sizeof the Al cation.The ratio of Mg:Fe in R*' is likely to have a pronouncedeffeet also on o and the value given againsteach point on Figure 8 is the number of Fe*2 ALPINE CHROMITE ]N BORNEO 849:: computedin the unit cell. Sincethis has a large ionic radiusit is likely to affect the size of o. The points in solid circlesare from this study and those in open circles are from Stevens (1944). A considerable spreadof valueshas beenobtained (Fig. 8) and the spreadfrom the computedbest-fit linear correlationsis not systematicallyrelated to Fe2ncontents as hoped.The regressionequations for thesedata, bearing.in mind that the spreadis considerable,are: .

Number of aluminium cations per unit cell : 407.30 - 48.60 (o in A) Number of chromium cations per unit cell : 58.46 (o in A; - 473.88 But theseequations can be usedonly for very approximatedetermina- tions. Ferhaps the reason for the imperfect nature of the data of Figure 8 is the oxidation state of iron. Many chromitesmay have some of the iron in R2* oxidisedto ferric iron. The changein ionic size would be expectedto distort the lattice. Be this as it may, the results of Figure 8 show that at this time X-ray diffraction determina- tion of chromite unit cell edge (a) cannot be with certainty related to the compositionof the chromiteand accordinglydoubt must be caston the validity of Steven's(1944) relationship between CrzOe and o. Reflectiuity Polishedspecimens of 11 of the chromiteslisted oq Table 1 were comparedwiih SiC No. 88 (reflectivity given as 20.1GpercenL at, a 590 nm) on a Vickers microscopefitted with an EEL digital-readout microphotometer.The results of several determinationson different mystals within each polishedspecimen are given in Table 2. An ex- cellent correlation was obtained of reflectivity with the number of aluminium cationsper unit cell (Fig. 9) giving a computedregression equation:

Number of aluminium cations per unit cell : 27.93 - L.82 X reflectivity (at 590 nm) The relationship given in Figure g appearsto be a precise and quick method of determiningAl content in chromites.However this relationshiphas been determinedon alpine-type chromitesin which, as chromium content drops, aluminium rises progressively.This same relationship is not expectedto apply exactly to stratiform chromites in which the overall iron content rises with chromium fall-off. For such chromitesthe slope would be expectedto be ]esssteep or reven reversed. A wavelengthof 590 nm was used,so that a comparisoncould be 850 CHANLES S. HUTCHISON

\ z.j CHROMIUMCATIONS \ : 58.46x of8)-473.88 ,.r ,'to )\ o.o^ 2.5'\

\ E \ ^c) 2,5 I \ c = qb \\ '''. c 3'oo \ o roE 3'5oor.r\a 2'4 O E .= \ rr E rr o 3.5^ \ - " I't \ (-) 2.8 a\ E 2.5 3'2o^ \ 3.1 "3.1 \ a tz.s =8 (J \.2-s .= \ =7c \ \o

8'20 8.24 8'28 8.32 oA o Sleveds(1964) o ThisStudy ALPINE CHNAMITE IN BORNEO 85t

Tenru 2. Rprr,ncrrvrrv er wavnr,rwcrn EgOr.rnr (coupennn wrrH src no. gg) elvn VH NunrssnsFoB soME SeseH.Cgnordrrng.

SPEC]I\,IEN Ftrtrr Farrwrv oI VICKERS NUIIIBEB 59ONM tlABDNESS

CHBOMITE

NB s452 1380 * 90 J 1186 a? 6 r ? 14EO ! 20 J 11A7 13,4 + ,4 1510I 60 J 224 42AL 1 1430 + 40 '12.5 BB 12O + ,1 1370 + 60 J 22? 1290 ! 24O 'tzl EW 105 1'1,9+ .1 14lO ! J 8090 11.4 + ,3 1320 + 80 H 134 1O,7t .1 14UO+ 90 NB 4169 al 2L a 1450 + 170 't J 142 1420 t 60

FEFRITCHBOMlTE

'l J 186 10 t r O 10?0 + 50 made with the values of Mih6,lik and Saager (196g), who quoted a similar range of values but did not try to rerate the reflectivity rc composition.They notedthat chromitegrains in the Basal Reef of the witwatersrand system often show higher reflecting bordersas a result of alteration in situ. The alteration results in an increasein iron and chromium,and a decreasein aluminium and magnesium:This artera- tion is now regardedby most workers as being hydrothermal (see Engin and Aucott, 1971) and it givesrise to higher reflectivity varues. rndeedit resemblesmagnetite in polishedsection, and if we follow the schemeof Stevens (19,M) as shown in Figure 6, the hydrothermal alteration of the sabah chromiteswould be expectedto put them into the field of ferrian ehromite,which is now called by most people ferritchromite. Reflectivity values of the Sabah ferritchromite are higher than those given by Mih6,lik and Saager(1968). They given 13.6 percent

Frc. 8. Relationship between the unit cell edge (a, A) and the number of Al and cr cations per unit cell for 6 sabah chromites and l0 from stevens (1g44). The number beside each point is the number of iron cations in R*. 852 CHARLES S. HUTCHISON

+a' NUMBEROF Cr*Fe " CATIONS PER UNIT CELL t2 ll 10987 k

E t3

O) tr)

F l2 "-e t

F (-) rr ul" I LL U.J E.

345678910 NUMBEROF ALUMINIUMCATIONS PER UNIT CELL ' Fri. 9, Variation of chromite reflectivity at 590 nm wa,velength (compared with Sic No. 88 of given 20.16 percent reflectivity) with change in aluminium cont€nt.

as.their highestdetermination, but this study indicatesa value of the order of 20 percent (Table 2). This reflectivity is more characteristic of magnetite,and indeedRamdohr (1968) shoIMSan alpine chromite (p. 929) rimmed progressivelyby ferritchromite and then by magnetite. Hardness Eaidness-wasdetermined using a 200 g load and 25 sec. contact time on a teitl.Ainiload hardnesstester calibratedagainst a hardened ALPINE CHRAMITE IN BORNEO 853

steelstandard. The valuesobtained are listed in Table 2. The figures are considerablyhigher than those quoted by Mih6lik and Saager (1968),but hardnesscan be of little determinativeuse for indicating chromite compositionbecause at these high values the margin of error is greater than the range of variability sought.

Ur,rnltrerrrm-Onrcrx eun EuplecEMENT The wide range of chromium content of Sabah chromitesin common with all podiform chromites,the tendencyto a bimodal chromium content, and the absenceof iron enrichmentas chromium falls, all point to major and essentialdifferences between Alpine-type and stratiform ultramafites. These differencesmust reflect a radically different ultimate origin. It seemsreasonable therefore to presumethat Alpine-type ultramafites have evolvedin the Mantle just as stratiform ultramafites have differentiated from a gabbroic magma in the crust. For these reasonsthe hypothesisof McTaggart (1971) that Alpine-type ultramafites were originally identical to stratiform types as cumulatesin basic magma ehambershigh within the crust, sub- sequently subsidingduring tectonism to form oold, fault-controlled intrusionsdownward into lower tectonic levels,cannot be accepted. The layering of chromite within dunite, showingexcellent grading (Fig. 3) indicatesthat the chromitelayers owe their origin to "igneous sedimentation" just as in stratiform deposits,but this differentiation must havetaken placein the Upper Mantle. Partial melting of Mantle material to give basalticmagma, which eventuallyextrudes at oceanic ridges, must, leave behind a strongly ultramafi.c residuum. The very low potassiumcontents of the ultramafic rocks of Darvel Bay (Fig. 5) indicatethis residualcharacter. The associationof the Sabah chromite-bearingultramafic rocks with gabbro bodiesand highly metamorphosed(generally in almandine amphibolite but occasionallyin hornblende grairulite facies) tholeitic metabasaltscould be accomodatedwithin an oceaiiicspread- ing zonewith rising Mantle current and accompanyinghigh heat-flow. The work of Karig (1971), further elaboratedby Dewey and Bird (1971),shows that such oceanspreading zones not only occur in the main oceanridges, but also in ridges behind the island arcs of the westernFacific in marginal or inter-arc basins.Indeed Karig (1971) has designatedthe Sulu basin, which lies in the top right corner of Figure 1, as an inactive or fossil marginal basin with high heat-flow. The curvilinear outcrop pattern of the ultramafite bodiesof Nor[h. Borneo (Fig. 1) suggestsheeL-like emplacements from the Mantle into the crust, and the convex westwards pattern, from Labuk Valley to 854 CHARLES S. HUTCHISON

Darvel Bay, must be taken to indicate that the ultramafite sheet dips eastwards,implying a westwardsand upwards thrusting of Mantle material into the Crust.

Acrlqowr,nocrrvrpwrs I am grateful to Mr Leong Khee Meng of the Geological Survey of Malaysia, Kota Kinabalu, for making available several chromite specimens.I wish to thank Mr Jaafar bin Eaji Abdullah for the photography, Mr S. Sriniwass, Mr Ching Yu llay, and Mrs Eaidar binte Ludin for the drawings, Tengku fsmail bin Tengku Mohammad for assistance,in X-ray preparations, and Miss Anne Chong for typing the manuscript.

Appendix. Description of chromite localities of Table 1 Banggi Island (Kapitangan)-as thin lenses of 5 to 10 cm thickness in a 1 m band of serpentinized dunite within peridotite. (Sungei Tengah)-as bands and pods up to l0 crn width in strongly faulted and sheared serpentinized dunite (Collenette,1964) . Malawali Island: strongly brecciated chromite-bearing serpentinite. Labuk Valley: (Porog)-as veins and disseminations, striking 40"/-90' in a dunite belt 61 m wide. One of the larger veins is banded and 4.6 m thick, passing into solid chromite 3.7 m thick. It is locally brecciated. fndividual bands vary up to 23 cm thick and crystals seldom exceed 3 mm (Collenette, 1964). (Bukit Lumisir)-as sheared and mylonitized veins. One ore-body is of 4.6 X 1.8 m of coarse grained chromite. Dunite lenses contain layers, veins, and accessory chromite (Collenette, 1964). (Sungei Taguuk)-as pods and bands up to 15 cm thick in dunite lenses.(Tangulap and Kuun-Kuun)-as a 6 mm layer and disseminations in dunite lenses. (Ruku-Ruku)-as a pod, 80 X 50 cm, and as thin stringers in dunite within peridotite. Saddle Island group: The thickest dunites and serpentinites of Sabah occur here. As tatrular layers grading into disseminations, and sometimes as irregular bodies with sharp contacts against dunite (Collenette, 1964). (Katung-Kalungan island)-as trumerous discontinuous bands a few mm to 6 cm thick in serpentinite' (Laila island)-as common discontinuous bands, 3 to 13 mm thick, loeally swelling to 60 cm. (Saddle island)-as a fractured band, 30 cm thick, increasing to 1 m, outcropping over a 3 m distance. (Nipa-Nipa)-as bands 13 mm to 5 cm thiek in an ore zone5O cm wide. Mount Silam: An ore body, 13 X 7 m X 60 em maximum thickness, dips irregularly N.E. at 15". It has a skin (after sheared dunite) of schistose serpentinite 2 to 30 cm thick, weathering to a lateritic clay. The ore body and serpentinite envelope is thought to have bden detached from its parent dunite and intruded upwards into the overlying peridotite (Collenette, 1964).

Repnnnrvcns Cor,r,nlvnrro, P. (1958) The geology and mineral resources of the Jesselton- Kinabalu area, North Borneo. GeoI. Suru. Dep. Bri,t. Territ. Borneo, Mem. 6, 194p. ------(1964) Chromite prospecting in Sabah 1959-63,Borneo Reg. Malaysi'a GeoI.Swu. Ann. Rep.1963,4&58. Drwoy, J. F., exn J. M. Brnn (1971) Origin and emplacement of the ophiolite ALPINE CIIROMII'E IN BORNEO 855

suite: Appalachian ophiolites in Newfoundland. J. Geophas, Res. 761 3179- 3206. ENcrw, T., exo J. W. Aucom (1971) A microprobe study of chromites from the Andizhk-Zimparalik area, south-west Turkey. M ineral. M ag. 38, 76-82. Frrcu, F. H. (1955) The geology and mineral resources of part of the Segama Valley and Darvel Bay a.rea,Colony of North Bomeo. Geol. Swu. Dep. Brit. Terit. Bomeo, Mem,4, L42p. - (1970) Peridotite-gabbro problems: a comment. Geol. Soc. Malagsia Neuslett.27, t-2. Gor,ns, G. G. (1967) Trace elements in ultramafic rocks. In P. J. Wyllie, ed. Ultramafic amd relnted roclcs. John Wiley and Sons, New York, p. 352-362, IfurcutsoN, C. S. (1971) Alpine-type chromite deposit in the Da,rvel Bay area of North Borneo (abstr.). 12th Pacifi,c Sci,.Congr., Rec. Proc. yor. 1, (Abstr.) Australian Academy of Science, Canberra, p. 415. -t eNl T, J. Dsortrau (1969) Deformation of an alpine ultramafic association in Darvel Bay, Eiast Sabah, Malaysia. GeoI. Mijnbouw 4E, 481494. --, AND - (1971) An alpine association of metabasites and ultrabasic rocks in Darvel Bay, East Sabah, Bomeo. Ouerseas Geol. M'ineral Resow. 10,289-308. Kenrc, D. E, (1971) Origin and development of marginal basins in the wegbern Pacific. J. Geophys. R es.76, 2542-2561. Krnr, H. J. C. (1968) The igneow rocks of Sarawak and Sabah. Geol. Swu. Malays,in,Bomeo Reg, Bull.S, 210p. LnoNc, K. M. (1971) Peridotite-gabbro problems, with special reference to the Sega,ma Valley and Da,rvel Bay area, Sabah, East Malaysia. GeoI. Soc, M alaysia N ewsl,ett.28, 4;L3, MeNsoN, V, (1967) Geocbemistry of basaltic rocks: Major elements. In H. H. Hess, and A. Poldervaarb, (eds.) Basalts: I'he Pold"eruaart Treat'ise o'l Rocks o't Basaltic 'Compositinn. John Wiley-Interscience, New York, p.215-269. McTeccrnr, K, C. (1921) On the origin of ultramafic rocks, GeoI. Soc. Amer. Bull.g2,W42. Mrn.{r,rr, P., ern R. Saacon (1968) Chromite grains showing altered borders from the Basal Reef, Witwatersrand System. Amer, Mineral.53, 1543-1550. Mrrcnnr.l, B. J., axo J. E. Kmr,ervr (1g68) Unusual matrix effects in X-ray spectroscopy: a study of the range and reversal of absorption enhancement. Appl,. Spactro s c. 22, 7 42-748 Nnwrow-Snrrru, J. (1967) Bidu-Bidu llills area, Sabah, East Malaysia: explana.- tion of sheet 5-117-2 and part of 5-117-1.GeoI. Suru. Bornio Reg. Mala'gsia, Rep.4,t@p. Nonnrsu, K., aNo B. W. Cseppnr,l (1g67) X-ray fluorescence spectrography, In J. Zussman, (ed.) PhAsi,calMethoils in Detqrmtirwti,ue Mineralagg. Academic Press,London, p. 161-214. PnrNz, M. (1967) Geochemistry of basaltic rocks: Trace elements. 1n E. E. Eess, and A. Polderva,art, (eds.) Bosalts : tlte Pold,eruaart Treatise on Bochs ol Basaltin Com,positinn. John Wiley-Interscience, New York. p. 27L-323. R"rrtoour, P. (1969) The Ore Minerals and, Thei,r Imtergrowths. Pergamon Press, Oxford, 1174p. Rosnrnlurrr, S. (1958) Magnetic zusceptibilities of minerals in the Frantz Iso- dynamic Separator. Amer. M,i,nerol. 4, lZ0-lZB. 856 CHARLES 8. HUTCHISON

SrnvoNs, R. E. (1944) Composition of some chromites of the westqm hemisphere. Amer. M ineral. 29, L-34. Tnernn, T. P, (1967) Chemical d,nd structural relations of ultramafic and felds- pathic rocks io alpine intrusive eomplexes. In P- J. Wyllie, (ed.) Ultramartc cmd,Eelated,Roclcs. John Wiley, New York, p. t2Z-235. - (1970) Chromite segregations as petrogenetic indicators. Geol. Soc, S. Alr, Spec.Publ.1, 380-390. Wrr,soN, R. A. M. (1961) The geology and mineral resorlrcesof the Banggi island and Sugut river area North Borneo. GeoI. Suru. Dep. Brit, Ten. Bomeo, Mem. 15,143p.

Mantrccript recei,ued,,September 20, 1971; accapted lor publicati,on, Nouember 11. lnI.