uvarovite garnet and south african jade … · 2007. 4. 5. · transvaal, shoi,ving the extent of...
TRANSCRIPT
THE AMERICAN MINERALOGIST, VOL 11, MAY-JUNE, 1959
UVAROVITE GARNET AND SOUTH AFRICAN JADE(HYDROGROSSULAR) FROM THE BUSHVELD
COI,{PLEX, TRANSVAAL
J. J. Fnarvxr.r, Llniversity of Xtatal, Dwrban, Sowth Afr,ica.
ABSTRACT
Uvarovite garnet is an uncommon silicate matrix in chromitite seams in the Noritebody of the Bushveld rgneous complex, Transvaal, union of South Africa. separateduvarovite samples were analyzed chemicaily, and unit cell, specific gravity and refractiveindex values determined. uvarovite, grossularite ancl andra4ite are the dominant ,,end
combined r'vater. The mineral replaces basic feldspar through zoisite, in anorthosite andfeldspar-rich pyroxenite horizons near the Lower gioup of chromitite seams in the westernBelt of the Norite. The mineralogical changes aie thought to be due to the addition ofcalcium in heated waters.
INtnoluct toN
Uvarovite, the bright green chromian garnet, was first reported associ_ated with chromitite seams, from the eastern part of the Bushveld com-plex (Hal l , 1908, p. 60; Hal l and Humphrey, 1909, p.74) .Grossular i te ,calcium aluminium garnet, in a massive form associated with anorthosites
garnet in the Complex were given by Hall (1932) and Kupferbiirger(re37).
565
566 T. J. FRANKEL
vite and of South African Jade which could add to known mineralogical
data was then possible.The localities of the specimens investigated are shown on the map
(Fig. 1). In addition to these localit ies, several others are known'
The chromium-bearing garnet is less abundant and more restricted in
its associations than other members of the garnet family. uvarovite is
usually found as a matrix to closely-packed chromite grains, in thin veins
or scattered in metamorphosed sediments. The crystals are generally
small and intimately associated with other minerals. This makes a clean
separation of uvarovite difficult and may account for the paucity of good
chemical analyses coupled with adequate physical data. Thus correlation
of chemical composition and physical properties in the chrome-bearing
varieties of the ugrandite series has been based upon sparse and incom-
plete information.The distribution and mineral associations of uvarovite and grossularite
in the Bushveld Complex are dissimilar in many respects, although they
do occur more or less in the same zone above the base of the Norite body.
The uvarovite is rarely associated with grossularite and vice-versa. For
this there must be some fundamental reason. Aims of the present study,
therefore, were to clarify the relationships between these garnets and to
contribute towards an understanding of their origins in the Bushveld
Complex.
t
, - cB?or l r t xor rzo |s
Frc. 1. Geological sketch map of the Central
Norite body of the Bushveld Complex and the
(Modified from Union Geological Survey Maps.)
Transvaal, shoi,ving the extent of the
localities of specimens here described.
UVAROVITE GARNET AND SOUTII AFRICAN TADE 567
TnB Frnr,o SnmrNc oF THE Uvenovrrn ,qtrn Souru Annrcax JannThe field occurrence of the garnetiferous rocks will be made clear from
the following brief account of the Bushveld Igneous Complex. This basin-l ike structure occupies an area, elongated along an east-west axis, in thecentral part of the Transvaal Province of the Union of South Africa. Thesurrounding, underlying sediments dip towards the centre of the Com-plex of which much is covered by later formations. The main continuousexposures are preserved in two areas, known as the Eastern and WesternBel ts (F is . 1) .
The Complex comprises two major rock groups; an upper Red Graniteoverlying and bordered by a lower Norite-gabbro body. The Norite bodyexhibits "pseudo-stratification" or rifting which agrees in a general waywith the low to moderate dip of the floor strata. It has been sub-dividedinto several zones, of which the Crit ical zone rather Iow down in thebody, is composed of a well pseudo-stratified, diversified, Iayered succes-sion of peridotite, anorthosite, pyroxenite and chromitite "seams."
The chromitite seams consist of chromite crystals with little intersti-tial silicate mineral. They range from thin partings to robust seams sev-eral feet thick that persist for long distances along strike and parallel tothe pseudo-stratif ication of the enclosing rocks. Disseminations ofchromite crystals are common in the pyroxenites.
There are two well-defined groups of seams in the Eastern Belt, desig-nated Lower and Middle groups. They are weil exposed in the hil ly andmountainous country. There is also an Upper group in the Western Beltwhere the country is flat. Faulting or black turf overburden often makeoutcrop correlation uncertain.
The Bushveld uvarovite occurrences are all associated with chromititeseams, at the contacts with surrounding rock, as cross-cutting veins, aslenses and bands within the seams, or in lenticular bodies with dissem-inated chromite not far from a true chromitite seam.
In the field the uvarovite appears to take the place of pyroxene andfeldspar within the chromitite seams or at the contacts with anorthositicor pyroxenitic rocks. Sometimes, masses of almost pure uvarovite asso-ciated with crystals of diopside and zoisite and only very small amountsof disseminated chromite grains, are marginal to chromitite seams. Mostof the uvarovite in the Eastern Belt is associated with the Lower groupchromitite seams; the examples from the Western Belt may come fromthe Nliddle or Upper groups.
The rock prized as an ornamental stone and known as South African
Jade occurs in the almost featureless flat country about 40 miles due westof Pretoria. Although poorly exposed through a black turf overburden,the field relationships have been clearly revealed in trenches and pits.
T. J. FRANKEL
Two or three bands of massive pale green rock, from a few inches totwo feet thick are separated by bands of l ight-colored massive anortho-site. An anorthosite-norite with jade, rests on a chromitite seam belowthe main jade seams. Scattered thin bands or streaks of chromite crystalsalso Iie within the jade. The jade layers and the associated rocks conformto the characteristic pseudo-strati l ication of the Norite body and dip at15 to 20 degrees towards the north.
According to Hall (1932,p.415), the jade horizon can be traced at in-tervals along many miles of strike, and where it leaves the chromititehorizon it is a white l ime sil icate hornfels. Occurrences of similar rockhave also been reported from the Potgietersrust area in the northernpart of the Complex and from the eastern part as well.
Uvenovrrp
Er p erimentol P r o cedur e
The uvarovite is generally the matrix around line chromite crystals sothat separation cannot be achieved by simple hand-picking; similarityin grain size of garnet and chromite does not allow of separation by siev-ing. In most of the specimens garnet was only l iberated from the associ-ated minerals after crushing to minus 200 mesh Tyler. The followingprocedure for separating the garnet from other minerals was adopted.
Gentle crushing in a diamond mortar released larger chromite grains,so that a fair percentage could be eliminated by hand-picking and coarsesieving. The minus 200 mesh powder was gently deslimed, dried, and re-peatedly centrifuged in Clerici solution S.G. 3.8 to 3.9. In this way freeparticles of garnet and other l ighter minerals were separated from thechromite and aggregate chromite-garnet grains. The lighter mineralswere removed from the garnet by centrifuge treatment in methylene io-dide or suitably adjusted Clerici solution. A1l samples were carefuliychecked microscopically for purity. The minute amount of chromite inscattered aggregate grains sti l l present after purif ication cannot have asignificant efiect on the chemical analyses of the garnet. The garnet pow-ders were thoroughly washed to remove adhering heavy Iiquid and werefinally oven-dried.
The determination of specific gravity on the fine powders proved diffi-cult, but repeated measurements using Ellsworth's vitreosil pyknometersgave reasonably consistent results and all values were then calculated to20' C. A few values were checked by suspension of grains in Clerici solu-tion suitably diluted. The refractive index measurements were made inWest's high refractive index l iquids in sodium light.
The color of the garnet was measured on a Donaldson TrichromaticColorimeter which I had previously used with some success (Frankel.
UVAROVITE GARNET AND SOUTH AFRICAN TADE 569
1953). colors are also described by comparison with a standard colorchart.
Full chemical analyses were not made on all specimens, mainry owingto lack of sufficient material. The accuracy of ferrous oxide determina-tions in garnets is often questioned and considerable attention was paidto this matter by the analysts, who used several methods, all of whichgave reasonably consistent results. l,Ir. Schutte used Hey,s method(1941)' while n'Ir. Herdsman carried out the determination by simplesolution in phosphoric acid in constant f low of carbon dioxide andchecked against a blank and a standard powder. particular attention wasalso paid to the determination of combined water in the uvarovite.
The National Physical Laboratory reported that the Debye-scnerrerphotographs were taken in a camera of 114.59 cm. diameter according tothe Straumanis principle, using either Cu K. or iron-fi l tered Co radiation.Exposure times were approximately 7 and 16 hours respectively.
Chemicol and Physical Data on Llraroaite
The chemical analyses (Table r) of the garnet can be recast as iso-morphous mixtures of "end members"; mainry uvarovite, grossulariteand andradite which, except for Specimens Nos. 1 and 2, range in approx_imate proport ions f rom 5:5:1 through 4:2: I to 2;2:1, respect ive ly .chromium is the dominant trivalent ion in most of the specimens whichare, therefore, called "uvarovite." There are also smal amounts of py-rope and almandine. rn specimen No. 1 only, is grossularite the dom-inant "end member."
The molecular ratios show considerable divergence from the idealf o rmu la (3 : 1 :3 ) .
ROz
2 . 9 12 . 8 93 .483 . 0 92 9 7
RrOsAnalysis 4
5o
710
RO
3 . 0 63 .333 . 9 13 6 53 . 3 3
The unit celi contains an "excess" of ca ions (Table rr), and recastingthe composition into "end members" arso gives excess cao that makesthe true proportions of "end members" diff icult to estimate. Betekhtin(1946, p. 69) also found calcuiation of uvarovite analyses unsatisfactory.
vermaas (1952) experienced similar diff icurties with manganese-irongarnet. He suggests that near agreement between determined and calcu-Iated specific gravity values in such garnets, indicates that excess atomsoccupy positions within the unit cell. since the measured and calculatedspecific gravity values of the uvarovite specimens are in good agreement,
570 T. J. FRANKEL
Tler-n I. Cneurcer- ANar-vsrs ol Uvenovrrn
1 2 3 7 8 9 1 0
sioz 38 .67TiOz 0.38 n.d. 0.30AlzOa n.d.Fe:Os 4 .94 4 .86 9 .38CrzOa 3 .93 9 .44 10 .56F e O 1 . 9 6 2 . 7 OMnOM e O 1 . 5 1CaO 33.97
33.50 34 .64 36 861 . 9 7 2 . 2 4 0 . 4 09 . 1 2 9 . 4 4 8 . 9 86 1 3 6 . 3 0 2 . 2 2
l l . t6 11.49 11.542 . 7 3 1 . 7 2 2 . 9 20.03 n .d . n .d .| .67 0 . 78 0 .62
33.70 33 26 35.93
3 3 . 3 3 3 5 . 1 8 3 5 . 6 9 3 4 . 5 22 5 0 1 . 3 1 0 . 2 4 1 . 6 87 . 50 n.d. n.d. 8 .056 . 2 6 3 . 9 6 5 . 9 7 3 8 0
t t .70 11 .96 13 .62 14 .832 . 5 r 2 . 5 5 2 . 3 4 2 . s 3o.02 0 03| . 2 8 1 . 8 6
3 5 . r 2 3 3 . 1 1 3 3 . 0 7 3 2 . 8 r
100.01 99.87 99.47 100.22 t00.22
1. Lydenburg district, Eastern Belt. Partial Analysis. Anaiysts w. H. Herdsman
(FezOs, CrzO:, FeO). J. J' Frankel (SiOz, TiOz, MgO, CaO)'
2. Klipfontein, 119, Lydenburg district, Eastern Belt. Partial Analysis. Analyst w. H.
Herdsman.3. Winterveld, 424, Lydenburg district, Eastern Belt. Analyst Mclachlan & I'azar
(total iron as FezOr).
4. Vygenhoek, 209, Lydenburg district, Eastern Belt. Analyst C' E' G Schutte'
5. Uitvalgrond, 71, Brits area, Western Belt' Analyst W. H' Herdsman'
6. Derdegelid, 141, Lydenburg district, Eastern Belt. Analyst W' H' Herdsman'
7. De Kafierskraal, 359, Lydenburg district, Eastern Beit. Analyst c. E. G. Schutte.
8. Kafierskraal, 915, Rustenburg district, western Belt. Partial Analysis. Analysts
W H. Herdsman (Fe2O3, 61rQ3, FeO). J. J. Frankel (SiOz, TiO:, CaO)' Mclachlan
&Lazar (HrO+0.487;.
9. Hendriksplaats, 357, Lydenburg district, Eastern Belt. Partial Analysis. Anaiysts
W. H. Herdsman (TiOz, FezOr, CrgOs, FeO). J' J. Frankel (SiOz, CaO)'
10. Doornbosch, 423, Lydenburg district, Eastern Belt. Total includes sro 0.ll/6.
Analyst W. H. Herdsman. FeO by F. J de Wet'
the excess of RO group may also be accommodated in cavities in the
garnet. Packing index is close to that given for garnets by Fairbairn
(re43).Alderman (1935) suggested that Si and Fe2+ ions are replaced by Al
with Fe3+ entering the RO group. This replacement would account for
excess RzOs in almandine. Fieischer (1937) who showed that most gar-
nets conform closely to the ideal ratio, thought that Alderman's ob-
served discrepancies might be due to under-estimation of ferrous oxide,
so that more iron would be reported in the RrOs group. Fleischer did not,
however, deal specifically with uvarovite nor with garnets containing
more than 0.5 per cent TiOz and it is in such garnets containing much
TiOz or FeO that deviations from the ideal ratio are found'
Although, according to Kunitz (1936), Ti replaces Si, the exact be-
havior of Ti in garnet is not known with certainty and Zedlitz (1933) has
UVAROVITE GARNET AND SOUTH AFRICAN TADE 571
TasLE IL NuMsrn or Arous (0:96) rN UNrr CrLL oF UvARovrrE
Theoretical
SiTi
96
2 1 . 8 6o . 9 7
96
22.431 .09
96
2 3 . 8 70 . 2 0
2 1 . 8 8 2 2 . 4 3| .23 0 .82
22 .83
7 .023 . 0 1s . 7 5
23 .52
7 .203 . O 75 . 8 8
24.O7
6 . 8 51 . 0 65 .90
2 3 . 1 1 2 3 . 2 5
5 . 8 0 6 . 1 73 . 0 9 1 . 8 66 . 0 7 7 . 6 2
24
t o14.96 15.65
1 . 3 8 t . 3 70 .01 0 .021 .25 1 . 80
24.69 22 .830 .04
26.68 24.7 4 27 . t 0 27 .33 26.06
suggested that Ti may possibly be in the trivalent state and so cause acorresponding amount of FerOa to be reported as FeO. In calculations ofthe molecular ratios of Bushveld uvarovite given above, all Ti has beenadded to Si, although all or part can occur as TizOe and AI can replace Si.Further calculations were made in which all the Ti was assigned to theRrOa Broup with an equivalent amount of ferric iron subtracted from theiron of the RO group, and excess Al was added to Si. These results ap-proached the ideal formula more closely, but the RO group still showedexcess Ca.
Attempts to correlate chemical composition of garnet with physical
Telrn III. Sprcrrrc Gnevrrv, Oprrcnl ano X-Ray Dern lon Ilvanovrrn
A1Fe3+
Cr
r e 'M2nMgCaSr
1 5 . 7 8
1 . 4 90 . 0 21 . 6 2
2 3 . 5 5
1 6 . 1 5
0 . 9 3
0 . 7 523.06
1 3 . 8 1
1 . 5 8
0 . 6 02 4 . 9 2
1 0
SG1"2o"c 3 709( + . 0 0 s )CalculatedS.G.RI (Na l ighQ 1 .783- 1 .817- 1 824( + . 0 0 2 ) 1 . 7 8 6 1 . 8 2 0o o ( t . 0 o t A ; r r . o r s 1 1 . 9 4 6 1 1 . 9 5 5Mol wt.Packing Index
3 ?58 3 7671.830 1 801 1 .832
11 961 1 I .923 11 989485.4 482.2 493.4
6 3 6 . 2
1 828- 1 830- 1 .8371 .830 1 .833
1 1 . 9 5 5 1 1 . 9 6 9 [ L . 9 7 5488 8
3 . 7 4 2 3 7 6 3 3 . 7 7 2 3 7 4 0 3 7 7 2 3 7 5 4 3 . 7 8 9 3 . 7 9 8
3 . 7 6 71 833
1 1 9 7 3490.0
6 . 2
572 J. J. FRANKEL
Taslr lV. X-R,rv PowlrR SpAcrNG D.rra lon Uvenovrtr
d Ad Ad A
2.969+2.65412 . s 3 3 +2.42682.32932.1708| .93171 .87681 . 7 1 7 5I .65151.59281.49031.33431.30211 . 2 7 2 61.20621 .10951.09051 . 0 5 5 90.996010.982370 .969850.901880 .891560.881790 .881370 .863460.82971o.822080 .81436
2 . 9 7 5 12.65942 . 5 3 5 12.+2802.30122 . 1 7 2 81 . 93111 .8851| 71911 .65361.59431.499rr .33521.30351 . 2 7 3 81 .2058| 11401.09221 05750.998390 .983800 .97 108
0 892530. 88280
0 .864600 .83087o.923240 .81582
mS
mm-wm-wmm-wm-wmS
m-wmmm
mmmm-w
m
mm-w
m-wm-wms
m-ss
mmm-s
mm-ssmmmmwmmm
mm-wm-wmm
S
4 2187 ?2.9683 m-s2 . 6 5 3 5 s2 .5300 w2.4268 m s2 .3322 m-w2.1701 m-w1 9313 m1.8841 w1 7 2 2 3 m w1 6529 m-s1 . 5 9 1 0 s1 .4918 m-w1.3343 m w1.3023 m| .2720 m-w1.2063 w1 . 1 0 8 3 m1.0902 m1 0-560 m-w0.99564 w0.98236 w0.97010 m0 . 9 + ?0 .89148 m0 . 88173 m-w
0 86363 m w0.82995 w0 82197 m0.81453 s
,';J": r""";J;;.,".';,".":end members" (Ford, 1e15), andcorrelation involving four "end members" has been reasonably success-ful. Levin (1949, 1950) has shown that f ive components can account for99 per cent or more of the composition of a garnet. By means of partialchemical analysis and the three physical properties, refractive index(R.I.), specific gravity (S.G.) and unit cell dimension (oe), the molecularproportions of the five compositional molecules can be obtained by set-ting up and solving five simultaneous equations. His investigation didnot include garnets containing the uvarovite "end member."
UVAROVITE GARNET AND SOUTH AFRICAN JADE 573
% Cr.O,Frc 2' variation of unit cell with chromium content. Bushveld uvarovite specimens-
Arabic numerals. von Knorring's specimens_Roman numerals.
samples avaiiable in the present investigation wourd confirm and ex-tend his findings.
nlum was not reported in von Knorring's analyses, so that there is, pre*sumably, I itt le or none.
III
20
57+ T T. FRANKEL
has a larger ionic radius. When the unit cell values of the Bushveld
uvarovite are adjusted using this "correction," the amount of scatter of
the plotted points of Fig. i is not appreciably reduced' The deviations
from the l inear relationJip b"t*"en unit cell dimension and CrzOs con-
tent are not due in any degree to titanium but rather to iron'
illsG.
t 5 20
% CcrO"
Frc. 3. Variation of specific gravity with chromium content Bushveld uvarovite
specimens-Arabic numerlls' von Knorring's specimens-Roman numerals'
Thee f fec to f i r on i sshownby thed i f f e rencebe tweenun i t ce l l d imen-sions of von Knorring's low iron specimens and those of Bushveld uvaro-
vite of similar crrc)g content. By calcuiation (total iron regarded as ferric
o * i d e ) o n e p e r c e n t o f f e r r i c o x i d e i n c r e a s e s t h e u n i t c e l i d i m e n s i o n b y0.09 A approximatelY '
Figure J shows a general increase in S'G' as CrzOa content Increases'
Where CrzOa contents are similar, S'G' differences are obviously due to
variations in the amount of iron. In Specimen No' 9, for example' which
con ta inson l yasma i l amoun to f t i t an ium, theh ighe r i r oncon ten to f f se t spossible lowir values of physical properties' High refractivity of iron
raises the R.I., but titanium which has a low refractivity appears aiso to
have some elevating effect on R'I ' in many minerals for reasons not
c lea r l yknown ;asF^ i rba i " ' has remarked ,o the r fac to rsmus tbeope ra -tive.
Figure 4 shows an almost
content, roughlY Parallel tolinear relationship between R'I ' and CrzOa
the straight line drawn by von Knorring, or
UVAROVITE GARNET AND SOUT:H AFRICAN JADE 575
the l ine l inking pure grossularite and uvarovite. It is clear that while theTiO: content has modified the R.L-CrzO: relationship, the appreciableiron content has also raised the R.I. value.
The refractivit ies of Fe3+ and Cr3+ ions are probably similar, so thatferric oxide can be considered equivalent to chromic oxide. Accordinglywhen the amounts of ferric oxide and chromic oxide are combined foreach analysis, and the values of "chromic oxide" plotted against R.I., the
% Cr.O.
Frc. 4. Variation of refractive index rvith chromium content. Bushveld uvarovitespecimens-Arabic numerals. von Knorring's specimens-Roman numerals.
curve so produced is nearly parallel and closer to the curve obtained byvon Knorring and also to that of grossularite-uvarovite.
The replacement of small by larger radius ions expands the unit cellwhile large ions l ike Ca in grossularite are surrounded by distorted poly-hedra (Menzer, 1929). The present study supports the theory that thegarnet structure is probably "elastic" l ike that of micas (Jaffe, 1951), andcapable of taking up a wide range of ions.
Statistical methods of correlation of chemical composition and physical properties could not be applied to the small number of uvaroviteoccurrences for which both these factors are accurately known. It wasnot possible to use Levin's algebraic method for the calculation of chemi-cal composition from partial analysis and physical properties.
Quantitative color measurements were made on three specimens ofminus 200 mesh uvarovite powder using a Donaldson TrichromaticColorimeter. There is a slight increase in the red component for uvaroviteof higher titanium and iron contents.
,,''u'lt
2/
II
s r o / t s 2 0
R.I.ilI
lo
c. I
t ogrossulcritc
. l / f )
Specimen No. 4
No. 7No. 10
I. T. FRANKEL
f i y
0.382 0.4390.397 0.4370 .375 0 .4s2
Limpet green
Georgian green
Bronze green
Leaf green
Woodpecker green
Moss green
Rockingham green
z
a .1790 . 1 6 60 . t73
The colors of minus 200 mesh powder of a representative number of
specimens were compared with the matt surface of the British Trade
Council Color Chart. The followins are some of the closest color match-
inqs.*
Specimen No. 1
N o . 3No. 5N o . 6No. 7
No. 8N o . 9
Reference C C 97C C 2 4 9C C 8 4C C 9 2C C 2 MC C 8 3C C 9 3
Both the qualitative and quantitative measurements show that with
increase in iron content the green color becomes deeper' while the
uvarovite of higher titanium content has a "rusty" green color due to an
increase in the amount of red or purple. The purple color of titan-augite is
considered to be due to a similar effect.The D.T.A. curve (Fig. 5, No. 1) shows that the uvarovite undergoes
no change on heating to 1100o C.
too i loo 'c,
Frc. 5. Difierential thermal analysis curves of (1) Uvarovite, de Kafferskraal, No' 3'59'
Table I, Analysis No. 7). (2) South African Jade, Buffelsfontein, No. 205. (Table VI'
Analysis No. 1)
Trrn PrtnocuPgY oF TrrE Uvanovrra-BnanrNc Rocrs
Thin section examination showed that the relationship between uvarovite and chromite
is always very much thesame. Thefollowing account describes the features of those rocks
from which the garnet had been separated and analyzed.
* All the color measurements were kindly made b1' 
UVAROVITE GARNET AND SOUTH AFRICAN JADE 577
No. 1. Lydenbu'g d.istri.ct. There is no chromite in the section which is made up of largeplates (up to 1.8 mm. across) of slightly pleochroic, pale green orthopyroxene holding ex-solution lamellae, and diopside (2V7 58", c^24O",^t 1.706). The diopside may have re-placed an earlier pyroxene. Very pale green garnet grains fully isotropic, 0.4 to 0.6 mm.diameter, replace both pyroxenes rvhich are cloudy near the garnet.
No. Z. Klipfontein, No. 119.Large plates of colorless to pale green diopside 0.9 to 1.8 mm.across (2V" 58", c^Z 36", y 1.706), and markedly zoned orthopyroxene 2V, 68., with apale pink to green, pleochroic core rimmed by diopside, are partly replaced by uvarovite.
t-r--r--l--!
Ct-J--l--r-!
B,,,O O . 2 m m
Ftc. 6. A and B. Highly corroded chromite remnants (black) in uvarovite (stippled).Clear areas are mainly zoisite. Vygenhoek, No. 209. C. Chromite grains linked in shortchains in uvarovite. A few small serpentinized olivine grains are present. de Kafierskraal,No. 359. D Corroded chromite grains in uvarovite. de Kafierskraal, No. 359.
Highly corroded chromite grains 0.07 to 0.45 mm. across, lie in the irregular patches ofgarnet r,vhich are aLso peppered b1' chromite granules 0.01 mm. or less. The garnet is deepergreen where it surrounds chromite grains. There is also a little interstitial zoisite.
No. 4. Vygmhoeh, No.209. The chromite grains 0.15 mm. average diameter, are isolatedor lie in short, straight to curved chains in the deep green, granular uvarovite. The chromiteis rounded and embayed and almost completely absorbed by the garnet (Fig.6, A and B).Patches of zoisite (0.12 mm. across) partly replace the tittle diopside present. Small, brown-stained inclusions 0.02 mm.2 in the garnet may be serpentinized olivine.
No. 5. Uitvalgrond., No.71. The chromite is scattered as individual slightly roundeddodecahedral or octahedral grains 0.04 to 0.3 mm. in diameter, or as small clusters in atranslucent, pale olive green garnet. The subhedral uvarovite grains range from 0.08 to0.5 mm., with an average diameter of 0.15 mm. Small irregular patches of diopside andserpentine are interstitial to, or lie within the garnet or surround the chromite grains.
578 I. J. FRANK'EL
No.6 Derdegelid., No. 111. The chromite grains (0.08 to 0.4.5 mm. diameter) are slightly
roundecl and show {ew corrosion eflects even under high po'lver. They lie in subhedral to
euhedral grains of uvarovite. Patches of a lath-like mineral, probably zoisite, and a little
diopside are interstitial to the garnet. small, serpentinized granules, probably forsterite
originally, are imbedded in the garnet.
No 7 d,e Kafershraal, No. 359.In addition to isolated or small clusters of grains, the
chromite crystals form short, roughly parallei chains (Fig. 6, C)' Although the crystals
are frequently euhedral dodecahedrons, 0 05 to 0.4 mm. in diameter, ragged grain margins
demonstrate corrosion @ig. 6, D). The dull green uvarovite of the same grain size as the
chromite, contains a noticeable amount of small, elongate grains from 0.004X0.012 mm.
to 0027X0.054 mm. in size. The colorless grains generally have a brown surface stain
and show low interference colors. They are considered to be serpentinized olivine. A small
amount recovered during the uvarovite cleaning process gave a faint r-ray pattern re-
sembling that of chrysotile and a spectrographic analysis showed magnesium and silicon
as main constituents.* A few serpentine patches interstitial to garnet are present'
No.8. Kafierskraol, No.915. Euhedral chromite grains 0.07 to 0.25 mm. in diameter lie
in curved or short, parallel chains in the green uvarovite matrix of similar grain size. There
is little corrosion of the chromite grains which are sometimes surrounded by granules of
diopside (?) 0.01 mm or less in size. There are small inclusions resembling augite in the
garnet and some interstitial serpentine.
No.9. Hendriksblaats, No.357.In the field this occutrence is vein-like and cuts through
a chromitite seam (Hail, 1908, p. 60). The hand specimen is made up of patches of chromite
crystals in the bright green uvarovite, and irregular, grey veins up to 5 mm. wide merge
with larger aggregates of diopside crystals several cms. across.
Specimen No. 3 Winterveld, No. 424 is so similar to this material that the foilowing
account serves to describe both.
The pale green uvarovite grains average 0.1 mm. diameter and larger grains up to 0.3
mm. diameter are linked in distinct chains. Some grains which have coior zoning show
slight anomalous birefringence. The large garnet grains are intimatelv associated u'ith
colorless diopsicle (2v7 .58', I 1.70+) in plates several mm. across Fragments of partly
repiaced diopside in the garnet are in optical continuity '
Chromite is scattered throughout the garnet as residual granules of a few microns and
as larger, angular and embayed grains 0.25 mm. in diameter. The margins of the larger
chromite grains are bordered by brown, translucent zones of replacing garnet lr''hich still
contains some of the chromite substance (Plate 1 (1)). Chromite grains within pyroxene
plates are bordered by granules of uvarovite (Plate 1 (2)) Small circular areas of augite
0.01 mm. diameter lie in the chromite grains. There are also interstitial patches of ser-
pentine.
No. 10. Doornbosch, No. 423. Within the bright green garnet matrix, large, anhedral
chromite crystals from 2 to 7 mm. across are linked in chains (FiS.7), while most of the
subhedrai chromite grains 0 2 to 0.4 mm. in diameter lie between the chains in haphazard
fashion. A colorless mineral of low R.I. (1.450) and birefringence about 0.002, fills some
interstitial areas between, or forms extremely thin borders to chromite grains. It was de-
termined as gmelinite from an :r-ray photograph.t A little calcite interstitial to chromite is
also present. Polished section shor.s a few chaicopyrite particles and silvery grains of 2
microns diameter may be hematite or platinoids.
+ Kindly determined by Dr. S A. Hiemstra, Union Geological Survey.
t Kindly determined by Dr. H. Heystek, National chemical Research Laboratory.
AVAROVITE GARNET AND SOUTH AFRICAN JADE 579
Other uttarottite-beoringrocks.In the Winterveld-Hendriksplaats area of the Eastern Belt,thin, fine-grained chromitite seams of the "hard lumpy" variety of ore are separated byirregular patches and bands of green compact rock 0.3 to 5.0 cms. thick. They are similarto South African Jade in appearance but are duller and the hardness is frecuentlv lessthan 6.5.
Under the microscope the green bands and irregular patches are translucent, colorlessto grey, and composed mainly of laths of stocky crystals either in stellate or haphazardarrangement. There is a faint, ghost outline suggestive of original pyroxene plates. Thethin, lathJike crystai.s, 0.1X0.02 mm., identified as diopside,* are associated with granular
o 5mm
Frc. 7. Chain of large chromite grains in uvarovite-bearing chromitite.Locality Doornbosch, No. 423.
zoisite. Smal], almost equant, colorless patches with radiating cracks and sheafJike aggre-gates may be prehnite.
Small, pale green or yellow-green, subhedral uvarovite crystals (z: 1.g30) are scatteredthroughout the rock. The amount of garnet increases at the contact with the chromititewhere replacement of the chromite is cleariy demonstrated (plate 1. (3A)). There is alsoa blue-green, anisotropic mineral of low birefringence which resembles the green uvarovite.rt is probably a chromium-bearing hydrogrossular. rt replaces chromite but is itself re-placed by colorless chlorite and uvarovite which surround it (Fig. S).
Trvo specimens of the green bands were partially analyzecl. The material for AnalysisNo. 2 was treated in hot concentrated h1'd166h1ori. acid and then after water washing,strongly heated.
* Confirmed by an r-ray powder photograph taken by Dr. Heystek.
580 J. T. FRANKEL
Frc. 8. Contact between diopside-rich band (di) and chromitite, showing corroded and
replaced chromite (black) in green uvarovite (u), dark blue-green birefringent garnet
(heavy stipple) and colorless chlorite (c)'
SiOzRzO:CaOMgoIgnition loss
(1 )3 9 . 812.02 4 . 01 7 . 27 . 2
(2)49 .2
25.31 3 4 Analyst J. J. Frankel
100 .2 100.3
1. Pale green band between chromitite seams, Winterveld, No' 424'
2. Acid and heat treated material from green band, Winterveld, No' 424'
Both analyses cast into "norms" give dominant diopside with zoisite, and for Analysis
No. 1, some magnesium carbonate as well'
CnnounB AssocrnrBo wrrH UvARovrrE
The chromite crystals of the Doornbosch specimen (No' 10)' have a
wide grain size range known only in a few of the many chromitite out-
crops.-Because the crystals are not interlocked, this rock is very friable,
and it disintegrates when lightly tapped with a hammer. This treat-
of the grades so produced were ana'Iyzed (Table V).
UVAROVITE GARNET AND SOUTI{ AFRICAN JADE
Tlsra V. Cnrlrrcar ANalvsns or CnnoMrrss
581
SiOzTio,AlzorFeso;CrzOrFeOMnOMgoCaO
2 20 . 1 8
1 1 . 9 5 *1 1 047 .8513.20 .05
l J . o
n .d .
2 . 00 . 2 5
12.46*1 0 . 347 .051 3 . 9 50 . 0 4
1 3 7n.d .
2 . Oo . 2 2
13 .01 0 947 .1013 .35n .d .
13 .3sn .d .
1 . 80 . 1 0
13 .501 1 . 247 .2512.450 . 0 1
13 .65n.d .
0 . 6 00 _ 7 0
1 3 . 1 61 8 . 1 639.92t4 720 . 2 9
l z . + J
0 .50
100 .03 99 75 99.92 99.96 100.50
S G.4"20"0 + .46 4 . 4 6
* By difierence.Chromite crystals separated from uvarovite-bearing chromitite, Doornbosch, No. 423,
Eastern Bel t . No. 1. f 14 mesh Tyler . No. 2. -35+48 mesh. No 3 -48f 100 mesh.No.4. -100t200 mesh. Analyst J F de Wet (MnO by J J.Frankel) . No. 5. Chromiteseparated from uvarovite bearing chromitite, de Kafferskraal, No. 539, Eastern Belt.Analyst C E. G. Schutte.
There is no well-defined grading in crystal size which would be ex-pected had the chromitite been formed by crystal settl ing. The largecrystals l ie in scattered clusters or chain structures in the rock. It hasbeen suggested on geochemical grounds that early formed chromitecrystals are richer in Cr, A1 and I'Ig, while Iater crystals have moreps2+, ps3+ and Ti (N{alhotra and Prasada, 1956). This appears to betrue for some chromitite seams where crystals at the base have moreCrzO: than those at the top (Frankel, 1949). Except for the veryslightlyhigher percentage of CruOe in the coarse crystals, there are no distincttrends in the four graded fractions. ft is possible, therefore, that thechromitite crystall ized slowly from a re-fused chromitite (Sampson,1932, p. 143); the larger crystals having formed first.
The analyses Nos. 1 to 4 resemble those quoted for the l,Iain Seam ofthe Lower group on adjacent farms (de Wet, 1952, p. 150), but theydiffer from an analysis from Doornbosch farm itself (Kupferbiirgeret a"1. , 1937 , p. 25) .
Chromite from the de Kafferskraal uvarovite-bearing chromitite has aiower CrrOa content than the majority of seams in the two groups identi-f ied in the Eastern Belt. The field relationships of this rock are obscure-it might belong to the Middle group in which chromite generally has lessCrrOs than the underlying group of seams.
+ . 4 J4 . 4 5+ .46
582 J. J. FRANKEL
Pr,,q.rn 1. (1) Brown translucent zones arouncl chromite grains in uvarovite. The zonesconsist of uvarovite and brown remnants of chromite. Hendriksplaats, No. 357. Planepolarized light.
(2) Uvarovite granules around corroded chromite crystals in diopside. Winterveld,No. 424. Plane polarized light.
(3A) Uvarovite crystals in fine-grained diopside rock. Slightly corroded chromitegrains are bordered by uvarovite and chlorite. Winterveld, No. 424. Plane polarized light.
(3B) Chromite grains in diopside rock replaced by colorless chlorite, blue-green hydro-grossular (?) and uvarovite. Winterveld, No. 424, Plane polarized light.
(4) SouthAfricanJade. Gabbrotextureasseeninplanepolarizedlight. Between crossednicols this slide is fully isotropic.
TnB OnrcrN oF THE UvanovrrB
Although some chromitites hold significant amounts of plagioclase, itwas not the dominant constituent in any of the uvarovite-bearing chro-mitite specimens. Uvarovite formation, therefore, required the addition of
UVAROVITE GARNET AND SOUTII AFRICAN IADE
Plerr 2. (1) South African Jade in plane polarized light.(2) South African Jade betu'een crossed nicols. Same field as fig. 1. Note anomalous
birefringence and twinning efiects.(3) Partly aitered chromite-bearing pyroxenite. Clear areas original feldsparwith some
pyroxene. A few striated orthopyroxene crystals also visible. Plane polarized iight.(4) Same field as in fig. 3. Clear areas of fig. 3 are isotropic or anomalously birefringent.
The orthopyroxene is also partly altered to hydrogrossular. Crossed nicols.
lime and it is considered to be a product of lime metasomatism with at-tendant metamorphic effects.
Interstitial serpentine could be derived from orthopyroxene, whereasthe isolated small grains of serpentinized olivine, are similar to thosefound in impure magnesium limestones that have undergone low gradethermal metamorphism. This suggests that some magnesium carbonatewas present long before uvarovite was formed. It could have been pro-duced by the action of carbonated waters, perhaps of deuteric origin,
583
J. J, FRANKEL
on enstatite. Forsterite would develop early with increase in tempera-ture, and the introduction of calcium in heated waters and an increasedmetamorphic grade, would assist uvarovite formation. The moderateamount of aluminium required would be available from feldspar dissocia-tion.
Thin bands of orthopyroxene between chromitite seams are also con-sidered to have been partly converted to carbonate. Lime metasomatismand heat increase, assisted the formation of diopside and "grossularite,"and uvarovite was a final product.
Uvarovite is also associated with diopside and zoisite in veins cuttingacross chromitite seams. Betekhtin (1946, p. 70) reported a similar asso-ciation which he attributed to pneumatolytic-hydrothermal agencies.Fortier (1946) has also described the formation of chlorite and uvarovitefrom chromite, and many other examples are on record.
Synthetic uvarovite is readily produced in an anhydrous environmentat 855" C. (Hummel, 1950) . This temperature is obviously too high forthe metamorphic/metasomatic processes outl ined above for Bushvelduvarovite and associated minerals. Perhaps high pressure and the pres-ence of water, lower the formation temperature. Nevertheless, it is clearthat when chrornium is available, the ugrandite garnet, uvarovite, formsrelatively easily in nature in either hydrous or anhydrous environment.
It is suggested that in a magmatic body l ike the Bushveld Norite therewould be local surges of heat coupled with slight dynamic effects, due tosagging in the body. These surges would give rise to simple metamorphicefiects, contemporaneous or slightly later than the introduction of heatedIime-bearing solutions.
There is also the possibil i ty that the calciurn came from small amountsof carbonate xenoliths derived from the floor of the Complex and whichwere completely digested in the Norite body in the vicinity of the Lowergroup of chromitite seams. This would account for the erratic and irregu-Iar nature of the uvarovite-bearing chromitites.
Souru Arnrc.cN Jeon (HvnnoGRossur-AR)
Uvarovite holding a large amount of the grossularite "end member"occurs with diopside and zoisite and a l itt le disseminated chromite (Speci-mens Nos. I and 2 described above). Garnet, essentially grossularite, hasbeen reported as a metamorphic mineral at the contacts of the Noritewith underlying carbonate rocks from many localit ies in the Eastern andWestern Bel ts (Kynaston, 1909, p. 20; Wagner, t929, p. 175) .
Hall (1925) described the rock prized as an ornamental stone, andknown as South African Jade from the farms Turffontein No. 356 andBufielsfontein No. 205 in the Western Belt, as a massive fine-grained
UVAROVITE GARNET AND SOUTH AFRICAN JADE
variety of grossularite. Although it contains no nephrite or jadeite, thismaterial wil l be referred to as Jade, using the term in the general sensefor compact tough minerals suitable for the making of ornaments. Inhand specimen the rock is pale green to grey, massive and translucentwith scattered chromite grains. It is quite distinct from the sparklingcrystall ine granular uvarovite in chromitite.
Thin sections are mainly colorless or very pale cream-brown and faintlycloudy. Except for faint straight "cleavage" traces or crystal faces, thegarnet shows no crystal outl ines, and between crossed nicols is either iso-tropic or gives grey interference colors of anomalous birefringence. Zon-ing and the remarkable twinning effects described by Hail, simulatetwinning of albite.
In several slides, however, outl ines of stout Iaths arranged with some-rvhat cloudy interstit ial patches are seen in plane polarized l ight; betweentrossed n icols the whole rock may be fu l ly isotropic (Plate | (4) ;PIate2(1,2)) .This ghost texture is ident ica l to that of the fe ldspar-pyroxeneassociations commonly seen where the norit ic rocks grade into anortho-site. A particularly good comparison can be made with the anorthositiclayer associated with the Nlerensky Reef chrome band.
Partial transformation of pyroxenite to jade at the contact of a chro-mitite seam was observed. The small amount of ieldspar and larger areas
l.robably of clinopyroxene, were completely replaced by isotropic jade;the remaining pyroxene, dominantly enstenitic, is fairly fresh (Plate 2(3,4)). Only a Iew of the chromite grains show green chlorite-l ike mar-gins. No zoisite was seen. However, a l itt le zoisite is common as a welldistributed mineral in many specimens. It is dominant in saussurit izedanorthosite where feldspar (An62) shows partial or complete alteration tosaussurite which grades into a more clearly defined aggregate of smallgarnet crystals in Lhe zoisite. This in turn grades into clear isotropic jade,in which no trace of originai minerals is preserved.
Green isotropic rims around chromite are probably a true uvarovite.l,Iany chromite grains, however, are either siightly corroded or com-pletely replaced by jade.
Slightly wavy thin transparent bands alternate with somewhat lesscloudy zones in many specimens. In thin section the white bands areslightly darker than the rest of the isotropic jade and they Iie close to theclear-cut margins of saussurit ized zones. Rhythmic replacement as jadedeveloped from saussurite may account for the banding.
Hall 's Analysis No. 5, of his specimen No. 616-c5 (Slide 1128) in spiteof its high total, approaches the theoretical analysis of grossularite moreclosely than any of his other analyses. In thin section, this rock is color-less, clear and fullv isotropic with no relict structures. Variations in the
585
586 I. J. FRANKEL
other analyses may be attributed to small and variable amounts of zoisiteand residual pyroxene and feldspar. The trace of chromium reported isdue, either to the chromium in the garnet, or to scattered chromite grains.
Hall gave some physical data but made no comment thereon. In 1928van der Lingen found that the transmission spectrum of the SouthAfrican Jade showed the presence of OH groups in the molecule itself,and not in water of crystall ization. He concluded that the material isnot grossularite. Herbert Smith (1949, p.312) stated that the specificgravity and refractive index of the jade have slightly Iower values thanthose usually recorded for grossularite.
There is more combined water in all Hall 's analyses than is requiredfor the small amounts of zoisite. In the present investigation combinedwater and physical properties were determined on several different speci-mens. A complete chemical analysis was also carried out on clear iso-tropic jade which contains few scattered chromite grains and traces ofzoisite.
It is concluded from the chemical analyses and physical propertiesthat South African Jade is not true grossularite but hydrogarnet that
Tasr,B VI. Cnnurcer AN,tr,vsrs or T.lpr
SiOrTiOzAlzOsFe:orcrzo:FeOMnOMeoCaOHrO+COz
36 .5s0 . 1 2
2 3 . Mt . z l
o . 2 50 . 5 20.040 .79
36.061 . 1 6
none
37 .600 . 1 0
2 2 . 2 50 . 5 00 . 1 00 . 5 5
tracetrace
38.401 . 2 0
39 .30nil
2 t .930 .800 . 13o . 2 8nil
traces37 . ro0 . 3 0nil
100 .20 100 .95 99.84
S.G.R.I.
3.4881 .728
(+ .002 )11 .S59 A
( + . 0 0 1 )
3 . 5 2 3.5062r . 7 4 r
1 1 . S s ( + c l ) A
1. South African Jade, Bufielsfontein, No. 205, Dried at 105o C. Analyst J. J. Frankel(H2Of Messrs. Mcl,achian and. Lazar).
2. "Green Jade," Buffelsfontein. Total includes PzOb 0.O57o, HzO- 0.2O/s (IJall,r92s).
3. Grossularite from near Georgetown, California (Pabst, 1936).
UVAROVITE GARNET AND SOUTH AFRICAN JADE 587
has slight variations in composition. Hutton (1943) gave data for severalspecimens of the grossularite-hydrogarnet series which he named "hydro-grossular." Both Hutton and Yoder (1950) have suggested that the ma-jority of garnets described as grossularite are really hydrogarnet, andHutton says that the usual constituent of altered gabbros and relatedbasic intrusive rocks is a member of the garnet-hydrogarnet series nearthe grossularite end. The South African hydrogrossular has the samevariation trends of physical properties with change in water content asthe New Zealand specimens.
Tasre VII. Vanrlrrox ol Sprcrrrc Gnavrrv.lxo Rernecrrvr ftorxwrrn CoMsrNno Wl.rnn CoNrnnr rr.r Sourn ArnrceN Jeon
ToHrO+ (samplesdried at 105" C.)
0 . 8 3 | . 0 2 1 . 5 6 2 . 3 8 2 . 7 3
R. I .S. G.7a sio,
| .731 1 .730 | .7283 . 5 2 3 3 . 5 0 5 3 . 4 8 8
36. 55
r . 7 2 5 1 . 7 1 2 1 . 7 0 83.468 3.430 3.401
34 .82
The D.T.A. diagram (Fig. 5, No. 2) shows that no change takes placeduring heating until the temperature reaches 1100' C. approximately.At this temperature the mineral probably begins to decompose in thesolid state (c.f. Yoder, 1950, p. 239). There is no clear indication of thetemperature at which water is lost. The diagram affords no evidence ofan inversion from birefringent to isotropic jade.
Hall put forward two modes of origin for the jade-(1) magmatic and(2) metamorphic.
(1) The material of the jade, derived from the norite magma, waslater concentrated into bands, in a manner analogous to that of magnetiteor chromite seam formation.
(2) The garnet is a metamorphic-metasomatic product resulting fromintensive alteration of an original calcareous aluminous sediment (like acalcareous/dolomitic marl) upstoped from the sedimentary floor of theComplex and brought into the present position as xenoliths within theNorite body.
Alteration into massive lime garnet from an original aluminous calcare-ous rock implies an increase in sil ica and alumina, and removal of carbondioxide. Hall suggested that carbon dioxide could be lost because thethin bands of the rock were engulfed in a great thickness of magma andthe addition of other elements could be due to a metasomatic replacementcaused by magmatic infi l tration.
van Bil jon (1955, p. 132) stated that the composition of the jade sug-gests that it is an intermediate product of the metasomatism which trans-
J. I. FRANKEL
formed sedimentary carbonate rocks to anorthosite. Thin section, how-ever, has convincingly reveaied that feldspar laths (Anoz) are transformedprogressively with crystal outl ines preserved, through zoisite to hydro-grossular. If the feldspar had formed Jrom the hydrogrossular it is un-Iikely that weli-defined crystal boundaries would develop before thewhole of the enclosed volume had been converted to feldspar. In anyevent, the palimpsest textures of anorthosite and norite do not supportvan Bil jon's theory.
The fairly uniform thickness of the jade Iayers "interbedded" withanorthosites, pyroxenites and chromite horizons, points simply to theirbeing metasomatized equivalents of such rocksdn sllzz; access for calcium-bearing solutions being provided by the planes of pseudo-stratif icationor rift ing. The dominant cations in the original minerals were Al, Cawith less of Na in the feldspar, and NIg with some Fe2+ in the ortho-pyroxene. Additional l ime with a small amount of water could displacemagnesia and sil ica. Both Yoder (1950) and l{ummel (1950) stress thatwater acting as a catalyst may possibly assist the formation of evenanhydrous grossularite in hydrothermal transformations.
\{any authors have pointed out (Harker, 1.939, p. 17a) that saussuri-t ization of feldspar belongs to a late stageinthe coolingdownof anigne-ous rock, so that grossularite and zoisite may have formed during thisperiod when some stress may also have operated. This would afford asimpie explanation for jade formation. However, it is diff icult to under-stand why only certain zones of the anorthosiles and pyroxenites shouldhave been saussurit ized, and a source of additional l ime would sti l i benecessary for the production of jade from pyroxenite and norite.
Several kinds of l ime metasomatism are l isted by Turner and Ver-hoogen (1951, p. 488) and the type described in the present paper hasbeen previously recorded from New Zealand. Dykes containing diallage,prehnite and grossularite have been shown by Grange (1927) to be alteredgabbros really, with grossularite and prehnite secondary after feldspar.I{e suggested that grossularite-diopside veins were formed by solutionwhich took up l ime, magnesia and alumina from pyroxene, but directevidence of the transformation to grossularite was lacking.
Turner (1933) suggested that in pyroxenites the reaction products arestress minerals, and with aqueous solutions OH group minerals wouldform. Turner quotes Benson who thought that "rodingite" was due tothe action of concentrated magmatic water under high pressure, and thatl ime-sil icate rocks from New South Wales were transformed by calcium-rich aqueous solutions which came from a cooling peridotite mass; mono-clinic pyroxene was suggested as the source of the calcium.
Carbonate rocks in the floor of the Bushveld Comolex below the iade
UVAROVITE GARNET AND SOUTH AFRICAN TADE
occurrence on Buffelsfontein are a nearby source for the additional cal-cium required in the transformation to "garnet." This calcium couldhave been introduced by heated waters.
It is considered, therefore, that the South African Jade is a metaso-matic replacement of anorthosite, pyroxenite, and of the gangue min-erals of the chromitite seams.
CoNcr,usroNs
Both uvarovite and hydrogrossular have developed within the samezone oI the Norite body, although the modes of occurrence differ.Uvarovite is intimately associated with chromitite and pyroxenitewhereas the jade lies in anorthosite. It would appear, therefore, that theoriginal composition of the rocks and the subsequent metasomatic f meta-morphic processes have determined the variety of garnet developed.
Yoder failed to produce grossularite synthetically, however, and foundas other workers had, that only hydrogrossular formed relatively easilyin a hydrous environment. Hummel was able to prepare anhydrousuvarovite without difficulty. Undoubted anhydrous grossularite does ex-ist in nature; it is possible that at high temperatures and elevated pres-sures hydrogrossular may be made over to grossularite.
In the formation of the jade, fairly uniform thicknesses of anorthositehave been saussuritized and converted to hydrogrossular 'in situ. Thepale green tint is due to chromium from small amounts of dissolvedchromite grains. In rare instances chromite is rimmed by true uvarovitewhich suggests local temperature increase. Jade formation is not de-pendent on the presence of chromium.
It is unlikely that, as suggested by Hall, the fairly uniform jade layersrepresent actual l imestone horizons torn up from the floor of the Noritebody and brought to their present position where they were transformed.Rather, it is visualized that anorthosite bands were transformed byheated calcium-bearing solutions.
Hummel f ound that a hydrous environment is not necessary for uvarov-ite formation, and none of the Bushveld uvarovite contains a significantamount of combined water. The associations show that the uvaroviteformed essentially in pyroxenite and pyroxene-bearing chromitite as aresult of deuteric alteration and lime metasomatism, followed by thermalmetamorphism. IJvarovite formation was entirely dependent upon thepresence of adequate amounts of chromrum.
The chromite in the jade is relatively unattacked whereas it is generallycorroded in the uvarovite-bearing rocks. It is suggested that this alsopoints to higher temperatures at the time that the uvarovite was formed.
The complete digestion of small carbonate xenoliths within the chro-
590
mitite zone may explain the
rovite and such a possibi l i ty,
of ] ime.
AcrNowr,noGMENTS
Considerable assistance in chemical, differential thermal and r-rayanalyses was given by Dr. J. F. de Wet, National Chemical ResearchLaboratory, Dr. T. L. Webb, National Building Research Institute, Drs.
J. N. van Niekerk and F. Herbstein and N{r. F. T. Wybenga, NationalPhysical Research Laboratory; all of the Council for Scientif ic and In-dustrial Research, Pretoria. Through the good offices of the Director ofthe Union Geological Survey, some chemical analyses were carried outby the Division of Chemical Services, Pretoria. Other analyses weremade possible by the Research Grant Board of the Union of South Africaand the University of Natal Staff Research Fund. The Department of
Geology, Witwatersrand University made a number of Hall 's originalthin sections available. All other slides were prepared in this Departmentby N{r. C. N. Govender.
A short note by C. tr. Til leyhydrogrossular appeared in Geol.
RlrrnnNcrs
Ar.rrnrrraw, R. A., 1935, Almandine from Botallack, Cornwall: Mineralog. Mag.,24, 42-
48.Bntrr<ntrN, A. G., 1946, On chrome-garnets from the Nizhne-Tagilsk dunite massif:
D. S. Belyankin jubilee vol. Acad.. Sci.. Lr.S.S.R., 68-73.Bmrrsrr Cor.oun Couxcrr,, 1949, Dictionary of colours for interior decoration:3 vols',
London.nnWor, J.F. ,1952, Chromite invest igat ions pt .3:Chetn.Metal l Mi 'n.Soc.South
four ,52,143-153.Frrnrarnx, H. W., 1943, Packing in ionic minerals: Geol'. Soc. Americo Bul'|,.,54,
1374.Frrrscnnn, M.,1937, The relation between chemical composition and physical properties
in the garnet grotp: Am. Mineral . ,22,751-759.Fono, W. E., 1915, A study of the relations existing between the chemical, optical and
other physicai properties of the members of the garnet grotp: Am Iour. Sei.. ltlt ser-,
4C, 33-49.Fontrnn, Y. O., 1946, Some observations on chromite: Am. Jour.lci.,2441 649-657 '
FneNrnl, 1.1., 1949, Bushveld chromite investigation: South AJr'ican Min. Eng. Jot'tr.,
60,417-419.19.53, South African asbestos frbres: Mini'ng Mag.,89,73-83, 142-149.
GnaNcr, L.1.,1927, On the "Rodingite" of Nelson: New Zeolond' Inst Trons.,58 160-166.
Her,r-, A. L., 1908, Geology of the Haenertsburg goldfields and the adjoining parts of
Secoecoeniland east of the Lulu Mountains: Geol. Surv. Transttaal'. Ann. Rept.,
(leo7), 31-60.1925, On "Jade" (massive garnet) from the Bushveld in the Western Transvaal:
Geol Soc. Soul'h Africa Trans.,27 (for 1924), 39-55.
J. I. FRANKEL
irregular and erratic distribution of the uva-of course, would allow of an additional source
on the replacement of anorthosite bySoc. Sowth Afr ica Trans. ,60 ' 15-17.
Africa
1305-
UVAROVITE GARNET AND SOUTH AFRICAN JADE 591
1932, The Bushveld Igneous Complex of the Central Transvaal; Geol'. Su'rtt' South
AJri,ca Mem.,28.Hlrr,, A. L., eNo Huurrrnrv, W. A., 1909, On the occurrence of chromite deposits along
the southern and eastern margins of the Bushveld Plutonic Complex: GeoI. Soc. South
AJrica Trans., ll, (for 1908), 69-77.Hanxrn, A., 1939, Metamorphism: London, Methuen.Hrv, M. H., l9+1, Determination of ferrous iron in resistant silicates: Minualog' Mag.,
26,176-118.Hunuot, F. A., 1950, Synthesis of uvarovite: Am Minerol ,35r324-325.Hurrom, C. O., 1943, Hydrogrossular, a new mineral of the garnet-hydrogarnet series:
Royal Soc. Neut Zealand Trans ,73,174-180. ,
Jerrr, H. W , 1951, Minor elements in the garnet group: Am. Mine.ral.,36' 133-155.
KuNrrz, W., 1936, Beitrag zur Kenntnis der magmatischen Assoziationen. III. Die Rolle
des Titans und Zirkoniums in den gesteins bildenden Silikaten: lfezres Johrb., Mineral.,
Beilage-Bond 70, Abt. A., 385-436.Kupronetincnn, W., Lolrsaano, B V , et a1.,1937 , The chromite deposits of the Bushveld
Igneous Complex, Transvaal: GeoI. Surl. Soutlt Africa BUII ,No. lO.KvwasroN, H., 1909, The geology of the country north and north-west of Potgietersrust:
Geol . Surt . Ann. Rept. , (1908), 11-23.LrvrN, S. 8., !949, The physical analysis of polycomponent garnet: Presented 30th Ann.
Meeting. GeoI. Soc. America Bull.,6O, 1905.1.950, Genesis of some Adirondack garnet deposits: Geol. Soc. America Bull'.,61,
519-565.M,ttHorna, P. D., axn Pneseu Reo, G. H. S. V, 1956, On the composition of some
Indian chromites: Am. Mineral.,4l, 460-473.MrNztn, G.,1929, Die kristallstruktur der granate: Zei.ts Krist.,69, 300-396.
Prnsr, A, 1938, Garnets from vesicles in rhyolite near Ely, Nevada: Am Mi'neral',23,
101-103.S.tunsoN, E ,1932, Magmatic chromite deposits in Southern Afrtca: Econ Geol.,27,ll3-
144.Surur, G. F. H., 1949, Gemstones, London, Methuen.TunNnn, F. J., 1933, The metamorphic and intrusive rocks of Southern Westland, Part
lI: New Zealand. Inst. Trans.,63,178-284.Tunmr;n, F. J., ,lNo VnnnoocrN, J , 1951, Igneous and Metamorphic Petrology: New York,
McGraw-Hill Book Co.vAN BrLJoN, S., 1955, l'Origine des structures rubanies dans le partie basique du complexe
du Bushveld: Sciences de la terre, No. hors serie, one vol., Nancy. 131-148.
vAN DER LrNcrN, J S., 1928, Garnets: South AJricon Jow. Sci.,25, 10-15.
Vonu.Las, F. H. S , 1952, Manganese-iron garnet from Otjosondu, South-West Africa:
Mi.neralog. M ag, 29, 946-951.voN Kxonnn+c, O , 1951, A neu' occurrence of uvarovite from northern Karelia in Finland:
M,ineralog. M og., 29, 594-601.Wacxnn, P. A., 1929, The Platinum deposits and mines of South Africa: Edinburgh,
Oliver and BoydYoonn, H S., 1950, Stabitity relations of grossularite: Jour GeoI ,58,221-253.Ztotrrz, O., 1933, Ueber titanreichen Kalkeisengranat: Centralbl. Mineralogie, Abt. A,
225-239.1935, Ueber titanhaltige Kalkeisengranate: Centralbtr. Minerologie, Abt. A,68-78'
Manuscripl received July 31, 1958 .