33. CHROMIUM-BEARING SPINELS IN SOME ROCKS OF LEG 45: PHASE CHEMISTRY,ZONING AND RELATION TO HOST BASALT CHEMISTRY

A. L. Graham, R. F. Symes, J. C. Bevan, and V. K. Din, Department of Mineralogy, British Museum (NaturalHistory), London, United Kingdom

INTRODUCTION

The basalts obtained during Leg 45, the inauguralcruise of the International Phase of Ocean Drilling(IPOD), may be divided most simply into phyric andaphyric types. The lavas obtained from Holes 395 and395A (which penetrated 576 m sub-basement) havebeen further subdivided, on chemical grounds, into atotal of 10 principal units. The phyric units are desig-nated Pi-P5, the aphyric units A!-A5. This aspect of thebasalts is discussed elsewhere in this volume and willbe mentioned here only briefly.

This paper is concerned with the Cr-bearing spinelsthat constitute a rare phenocryst phase in many oceanfloor basalts. In Hole 395A, they are more common inthe phyric units than in the aphyric basalts. Chromium-rich spinel occurs (about one grain per thin section) inonly a small portion of the 207 meters of the majoraphyric unit (A3); generally the aphyric basalts do notcontain Cr-rich spinels. Only 34 of the 174 thin sectionsof Hole 395A basalts contain Cr-bearing spinel pheno-crysts, and only 4 of these are from an aphyric unit(A3, Cores 43 and 46). This discussion is restricted toHole 395A, since Cr-rich spinels have not so far beenfound in the 20 thin sections of basalts from Hole 395.In Hole 396, Cr-rich spinels are much more common,so much so that nearly all the thin sections examinedcontain at least two or three crystals. These spinels areoften associated with the phenocryst phases, olivineand Plagioclase, either included within them or at-tached to their rims. Free Cr-rich spinels as phenocrystsare less common than in Hole 395A. Data in this paperare restricted to those obtained from samples fromHoles 395 and 395A.

Cr-rich spinel is an early-crystallizing phase from abasaltic fluid containing around 300 ppm Cr; ocean-floor basalts generally have between 200 and 400 ppmCr, and might be expected to contain this phase. In thebasalts of Hole 395A, Cr-bearing spinels occur in-cluded within magnesian olivine ( - Fo88), calcic Plagio-clase (~AngQgg), or as a free, often subhedral, pheno-cryst. They range in size from 0.02 mm to 0.4 mmacross; the free phenocrysts are generally at the largerend of the range. Compositional zoning is common,and this is observed whether or not the spinel isapparently within a phenocryst.

RESULTS

Table 1 gives representative analyses of Cr-bearingspinels from Holes 395 and 3 95A, and each analysis is

related to the magma type of its host rock (P2-P5, A3;see elsewhere in this volume for magma chemistrygroups and their characterization). Generally, core andrim data quoted were obtained from single spinelgrains, but in two cases (columns 6a, 6b and 1 la, 1 lb),the composition range is not adequately represented byanalyses from a single grain, and the maximum rim-to-core variation in a particular lava type is indicated byanalyses from different grains. The cations have beencalculated to 32 oxygens, with Fe 3 + :Fe2 + proportionedso that Y = 16 (where the ideal spinel is XY2O4). Cau-tion is needed in interpreting the Mn data: the electronmicroprobe used was not able to resolve the MnKα peakfrom the large CrK/3 peak, and a portion of the Mnsignal from the spinels was due to Cr. Thus a portion ofthe MnO reported in Table 1 is due to Cr interference onthe Mn peak. One estimate of addition to the Mn signalsuggests that it is about 2 per cent of the Cr present inthe spinel.

DISCUSSION

The concentration of Cr in the lavas of Hole 395Avaries with lava type: between 200 ppm (P2) and 370ppm (P3, A3). In the P 3 and P4 lavas, the Cr-rich spineloften is accompanied by Cr-rich diopside (-1.9%Cr2O3), so in this part of the drilled core a proportionof the bulk Cr is in the pyroxenes. Indeed, the Cr-richspinels occasionally contain siliceous vermicular "inclu-sions" in their outer parts, which suggest that thespinel is actively reacting with the liquid. This seems tooccur only where a Cr-rich diopside is also present,though all the phenocryst Cr-rich spinels are roundedto some extent.

The partition coefficient for Cr between spinel andbasaltic liquid so favors spinel that under experimentalconditions (Hill and Roeder, 1974) Cr-rich spinelcrystallized from melts containing only 80 ppm Cr. It isodd therefore that mid-ocean ridge basalts, with char-acteristically between 200 ppm and 400 ppm Cr,should not contain greater modal proportions of spinel.In the Leg 34 (Nazca plate) basalts, Cr-rich spinel isapparently very rare, even though the basalts contain^300 ppm Cr, whereas in Leg 37 basalts from theMid-Atlantic Ridge they are much more common(Sigurdsson, 1977). Sigurdsson and Schilling (1976)suggest that lavas containing less than 400 ppm Cr donot in general contain Cr-rich spinel: this does notseem to be the case in this work, nor does there seemto be any chemical reason why it should be so. Thedata of Hill and Roeder (1974) suggest a maximum

581

A.L. GRAHAM, R.F. SYMES, J.C. BEVAN, M.K. DIN

TABLE 1Electron Microprobe Analyses of Cr-Bearing Spinels in Rocks From Holes 395 and 395A (compositions in wt %)

Magma Type

Hole 395

Specimen

TiO2A12O3Cr2O3FeOMnOMgOTotal

TiAlCrFe3 +

Fe2+

MgMn

Mg+Fe2+

1

0.1133.5334.0015.570.68

15.4399.32

0.0209.2736.3080.3992.6565.3960.135

0.67

P2

395A

2

Core

0.3740.6323.9814.290.37

19.2798.91

0.06310.8074.2780.8521.8456.4800.071

0.78

Rim

0.3740.0924.2517.080.46

17.0999.34

0.06310.7754.3710.7912.4665.8080.089

0.70

P3

3Core

0.6324.5440.1719.550.73

14.3599.97

0.1167.1027.7970.9853.0305.2510.152

0.63

Rim

0.9622.6138.1424.11

0.8111.5498.17

0.1856.8397.7381.2383.9374.4140.176

0.5 3

P3

4

Core

0.7228.7435.2320.18

0.5814.62

100.07

0.1308.1586.7111.0013.0655.2520.118

0.63

Rim

0.3525.9537.3120.730.67

14.1699.17

0.0657.5467.2811.1083.1715.2110.140

0.62

P3

5Core

0.6627.3837.4818.550.66

15.2599.98

0.1207.7867.1480.9462.7975.4830.135

0.66

Rim

0.8326.1436.1621.410.68

13.3298.54

0.1687.6567.1031.0733.3764.9330.143

0.59

P4

6a

Core

0.6226.4939.1418.180.60

14.98100.01

0.1117.4727.4061.0112.6275.3420.030

0.67

P4

6b

Rim

1.4523.3536.3031.160.766.79

99.81

0.2837.1417.4461.1305.6312.6260.167

0.32

P4

7

0.7231.0034.4518.040.51

15.0699.78

0.1288.6776.4710.7242.8605.3350.103

0.65

P5

8

Core

0.5638.8826.3916.630.21

16.2598.920.097

10.5364.7970.5702.6285.5680.041

0.68

Rim

0.6336.0728.4116.980.23

15.9798.290.1119.9595.2610.6692.6585.5750.046

0.68

Core

0.5927.4837.8816.920.61

16.4399.91

0.1057.6487.0721.1752.1675.7800.047

0.72

P5

9

Rim

0.7925.6437.4123.540.64

12.53100.55

0.1447.3287.1731.3553.4184.5290.052

0.57

Core

0.5831.9432.1519.000.66

14.6799.00

0.1049.0026.0810.8132.9885.2320.134

0.64

P5

10

Rim

4.7321.0231.6933.550.818.39

100.190.9256.4406.5122.1235.1713.2500.178

0.39

A3

l l a

Core

0.6128.7139.4614.100.67

16.93100.48

0.1087.9597.3370.5962.1755.9340.133

0.73

l i b

Rim

0.7528.4139.2218.680.72

13.55101.33

0.1347.9797.3880.4993.2244.8120.145

0.60

Note: 1 - 1 8 - 1 , 74-75 cm (#2D); 2 = 14-1, 90-100 cm (#12); 3 20-1,142-144 cm (#11); 4 = 22-1, 92-94 cm(#13B);5 = 22-1, 87-92 cm (#13A); 6a = 26-2, 25-33 cm (#1C); 6b = 22-2, 72-76 cm (#4D);7 = 61-1,120-122 cm (#2A); 8 = 31-1, 83-87 cm (#6); 9 = 29-1, 143-148 cm (#5); 10 « 31-1, 70-70 cm (#4); 31-1, 70-76 cm (#4); l l a = 43-1, 132-134 cm (#3); l i b = 46-1,4143 cm (#1).

solubility of Cr in the liquid of between 200 and 300 ppm,and also that the lower the Cr content of the liquid thelower the liquidus temperature of the Cr-rich spinel (ata fixed Po2). A possible result of the latter finding isthat the clinopyroxene liquidus may be intersectedbefore that of the Cr-rich spinel. Once this has hap-pened, it is less likely that the spinel will be precipi-tated, since the clinopyroxene contains more Cr thanthe liquid, and the nucleation of Cr-rich spinel isinhibited.

Petrographic observations indicate that the Cr-richspinels are not inert, and that once formed they reactwith the residual liquid. This is suggested by thesubhedral outline of the crystals, whether or not theyare within other phenocryst phases, either olivine orPlagioclase. Analyses (Table 1) show that the early-formed spinels have reacted and exchanged divalentand trivalent cations with the liquid. The amount oftrivalent ion exchange is more limited than that of thedivalent ions; indeed, the range in Yg (= molCr: (Cr + Al + Fe3+) in the spinel) is restricted (0.51 to0.27; see Figure 2), despite large changes in the X^g(molMg: (Mg + Fe2+) in the spinel, (0.78 to 0.23; see Figure 2and 3[b]). There seems to be no trend in Y^f* of thespinel cores which suggests for them a primitive char-acter. The Ysp variation is, in general, less than 2 percent, with only rare excursions beyond this; see Figure4. The variation in Ti, however, is often a sensitiveindicator of Mg/Fe zoning, so that a fraction of the Ycations must be exchanging with the liquid to relativelylow temperatures. It appears that the two exchangereactions, one involving divalent ions, Mg and Fe2+,and the other involving trivalent ions, Fe3+, Al, Cr, canoccur independently of one another. The trend of thezoning in Y3+ is to higher Cr contents with decreasingXMs, though this is not invariably so. Rarely thiscovariation is reversed, particularly in the "cores" ofthe spinels. Here the Al content rises slightly before thedominant trend of Cr increase asserts itself. Generallythe major zonation is in Mg:Fe.

Despite the similarity in the major element chemis-try of the phyric lavas of Hole 395A (Table 2 and

Figure 1. Ternary plot of the molecular proportions ofY^+ (Al, Cr, Fe^+) in the analyzed spinels. The lines joincore and rim compositions; the more aluminous of thetwo is that of the core. The symbols designate themagma type.

elsewhere in this volume), the spinel composition canbe related to that of the host lava in terms of theproportions of the trivalent ions (e.g., Y§p, in thefollowing order:

P _ p _ p pr2 r5 r 3 ' r 4increasing Y91

However, since the bulk alumina contents of theselavas are very similar [P2 ^ 18.0% A12O3 (wt), P5 ^18.3%, P3 ^ 17.7%, P4 ^ 17.1%], the correlationbetween A12O3 in the spinel and A12O3 content ofthe host lava is weak, but the trend is for the leastaluminous spinels to occur in the least aluminous lava.Since this trend, P2—P5—P3, P4, is also that of increas-

582

CHROMIUM-BEARING SPINELS

[= molMg: (Mg+Fe2+)] plotted against Y%r: ( Cr+Al+Fe3+)] for the Cr-bearing spinels

Figure 2. ^ [ g ( g ) ] p g %[ mol Or: ( Cr+Al+Fe3+)] for the Cr-bearing spinelsoccurring as phenocrysts in basalts from Hole 395A. Thelines join core and rim compositions; the more magne-sian of the Wo is that of the core. The symbols designatethe magma type, as indicated.

ing Cr concentration in the lava, it may be that theprime factor dictating the spinel Y3+ proportions is Cravailability, not the A12O3 content, which, in the lava,is much in excess of that needed for spinel formation.

Sigurdsson (1977) has shown that for Cr-rich spinelsin Leg 37 rocks there is a correlation between theA12O3 content of the rock and that of the spinel; but itis more likely that the Cr content of the lava controlsthe spinel composition, and that aluminum takes apassive role. For the Hole 395A lavas, the correlationbetween the Cr content of the lava and Y g is muchbetter than the alumina correlation.

There is no correlation between X p̂g and Mg:(Mg

+ Fe2+) of the host basalt, which might suggest thatthe Cr-rich spinels are xenocrysts in their present hostlavas. In fact this is only rarely the case. Some coexist-ing olivine-Cr-rich spinel pairs have been analyzed,and a plot of lnKD against Y ^ (following Irvine,1965) is given in Figure 5. In general, the majority ofthese pairs plot in a restricted area. The line drawn isthat of maximum slope passing through the origin andthe center of the field of plotted points. For equilib-

rium, the intercept on the lnKD axis must be greaterthan 0 (see Irvine, 1965), but the scatter of points andthe lack of definitive lineation does not define aparticular intercept on lnKD. Two points plot wellaway from the main cluster and are labeled A and B.Point A is that for the olivine-spinel pair in thetectonized harzburgite from Hole 395. Point B, from aspinel-olivine pair in a P2 basalt, plots well below theline and in a position which indicates that the equilib-rium conditions were very different from those of themajority of the spinel-olivine pairs. This particularspinel is included within a Plagioclase phenocryst: it iszoned, and has certainly attempted to equilibrate inpart with a host liquid. Although the core is toomagnesian to be in equilibrium with the coexistingolivine, the rim is not. The average rim composition(triangle in Figure 5) plots close to the line drawnarbitrarily on the diagram. This may well be fortuitoussince most of the rim compositions give lnKD valueswhich plot well above the line and are in no lineararray, as would be expected if the spinel continued toequilibrate with the liquid after olivine had ceased tocrystallize. In the case of the spinel-olivine pair plottedat B in Figure 5, the spinel rim is still in equilibriumwith olivine, since it is included in a Plagioclasephenocryst which has prevented re-equilibration withthe later stage liquids. The zoning in the olivinesshould also reflect this continuing equilibration withfalling temperature, but so far data are not availableon these samples to test this, partly because most of theolivines have altered rims.

With a few exceptions the core compositions of thespinels in the lavas of Hole 3 95A appear to be inequilibrium with their coexisting olivine. Unfortunately,this does not exclude a xenocryst origin for bothphases, initially at equilibrium under conditions verydifferent from those in which they were incorporatedinto their present host. The olivine composition variesbetween Fo83 and Fo§8, with occasional values of Fo90.All these olivines have a CaO content (^0.35% byweight) quite distinct from that of the olivine in thetectonized ultramafic of Hole 395 (0.06% CaO), al-though the bulk composition is very similar: Fo90.Thus, the olivines in the lavas are distinguishable fromolivines of ultramafic origin, as seen in Hole 395 and inthe ultramafics of Leg 37 (Symes et al., 1977), by theirCaO content. No olivine in the lavas has been found tohave a CaO content of less than 0.28 per cent; thisimplies that they all crystallized from a basaltic fluidand not from an ultramafic source. By inference, thesame considerations apply to the Cr-rich spinels, whichmust have crystallized from their basaltic hosts and arenot xenocrysts in their present environment. They havenot been derived from a source of ultramafic composi-tion.

Since the spinel core compositions appear to reflectthe bulk chemistry of their host lavas, the zoningobserved in these spinels does depend on the residualliquid composition at various stages of crystal fraction-ation. The modal abundance of Cr-rich spinel is small,so that crystallization of this phase will have little effect

583

0 0

FeAI204

HERCYNITE

90 70

CHROMITE

FeCr2O4

PICOTITE

"V̂\ \\\\50 30

/

BERESOFITE

90

70

50

30

\ ", \10 /

/ 1 0

30

50

70

90

Y

-

CEYLONITE

-

SPINEL

MgAI2O4

/ ' MAGNOCHROMITE

" /

\ 10 30 50

. \\

\

: X..CHROMEPICOTITES

-

-

70

PICRO-CHROMITE

90 X

MgCr2O4

100(Mg-Fe2÷)

Mg+Fe2t

I I

b.

J I I I i I I

typicalp error

bar

^ ^

P 5

-

Figure 3. (a) A spinel composition plot after Simpson (1920), showing the derivation of fig. 3b. (b) Spinels plotted in the manner of Simpson (see fig. 3a andSimpson [1920] ) , and separated according to magma type. Filled circles spinels within phenocrysts; open circles spinels in groundmass; open star = spinelfrom tectonited harzburgite of Hole 395. Arrows indicate core-to-rim variation. Scale units on (b) same as in (a).

CHROMIUM-BEARING SPINELS

Figure 4. A three-vector plot in two dimensions of themolecular proportions of Y3+ in the spinels, showing theextent of core-to-rim zoning in Y3+. The cores are allplotted at the intersection of the axes, and the distanceof the plots from this intersection indicates the extent ofzoning. Most of the spinels show a Y3+ variation of lessthan 3 per cent (molecular). Symbols same as in Figure 2.

on the liquid composition with respect to the majorelements, Mg, Fe, Al, Ti, but changes in the liquidcomposition will be reflected in the composition of thespinel. These considerations suggest that Cr-rich spinelis a sensitive indicator of liquid compositions duringthe fractionation of a basaltic fluid.

REFERENCES

Hill, R. and Roeder, P., 1974. The crystallization of spinelfrom basaltic liquid as a function of oxygen fugacity: /.Geoi, v. 82, p. 709-730.

Irvine, T. N., 1965. Chromian spinel as a petrogenetic indica-tor, 1. Theory: Canadian J. Earth Sci., v. 2, p. 648-672.

Sigurdsson, H., 1977. Spinels in Leg 37 lavas: Phase chemis-try and zoning. In Aumento F., Melson, W.G., et al.,Initial Reports of the Deep Sea Drilling Project, v. 37:Washington (U.S. Government Printing Office), p. 883-892.

Figure 5. A plot of InKj) =

yMg yFβAol ASpyFe yMg

Aol ASp

against Y%r

n.Sp

The line is arbitrarily drawn through inKp = 0, Y^r

p = 0and the cluster of data points. The black star A is fromthe tectonized harzburgite of Hole 395. For an explana-tion of B and the open triangle, see text.

Sigurdsson, H. and Schilling J.-G., 1976. Spinels in Mid-Atlantic Ridge basalts: Chemistry and occurrence: EarthPlanet Sci. Lett., v. 29, p. 7-20.

Simpson, E. S., 1920. A graphic method for the comparisonof minerals with four variable components forming twoisomorphous pairs. Min. Mag., v. 19, p. 99-106.

Symes, R. F., Bevan, J. C, and Hutchison, R., 1977. Phasechemistry studies on gabbro and peridotite rocks from site334, DSDP Leg 37. In Aumento, F., Melson, W. G., et al.,Initial Reports of the Deep Sea Drilling Project, v. 37:Washington (U.S. Government Printing Office), p.841-845.

585

A.L. GRAHAM, R.F. SYMES, J.C. BEVAN, M.K. DIN

TABLE 2Chemical Analysis of Samples From Sites 395 and 395A (compositions in wt %)

SiO2

TiO2

A12O3

Fe 2O 3FeOMnOMgOCaO

Na2θK2O

H 2O+

H2O"P2O5(Total C)"Others"TotalTrace ElementsBBeCrLiNbNiCoCuZn

VZrYSrBaRb

q

orabandihyolmtilap

1

48.81.58

14.53.537.690.178.43

10.33.050.080.790.520.140.220.16

99.96(ppm)

5<2

2707

<5145

402795

300105

39110

<50<5

0.4725.8125.6419.7812.28

5.845.123.000.33

2

48.91.60

14.82.828.250.178.54

10.42.870.090.650.480.140.220.16

100.09

7<2

27076

1404517

105290115

37105<50

<5

0.5324.2827.2418.9713.59

6.514.093.040.33

3

49.01.62

14.73.267.810.178.11

10.42.950.130.630.360.140.160.16

99.60

2<2

25077

155325590

280115

38100<50

<5

0.7724.9626.4919.5414.114.294.733.080.33

4

48.81.60

14.62.888.200.178.36

10.42.940.090.620.840.130.170.15

99.95

7N.F.

2707

12125

351395

260105

37105

<50<5

0.5 324.8726.3819.7312.62

6.524.183.040.31

5 a

48.00.37

16.61.524.460.11

11.48.863.510.174.480.670.010.140.12

100.42

7<2

14020<5

165391134

190<511

280<50

<5

0 481.00

28.8129.0411.81

_

20.932.200.700.30

6

40.80.021.094.324.530.10

40.80.860.07

N.F.6.600.55

<0.010.400.63

100.77

50<2

2180< l

62187

70286036

9<5<5

<50<5

_

0.592.661.26

24.3357.45

6.260.04-

7

48.81.62

15.03.507.750.188.27

10.32.870.160.590.690.140.140.16

100.17

3<2

25566

165335590

270120

39105

<50<5

0.9524.2827.5818.2414.57

4.505.073.080.33

8

49.21.63

14.82.259.010.198.41

10.42.750.150.420.100.140.170.17

99.79

1<2

2759

<5160

557595

280110

35115<50

<5

0.8923.2727.6018.7815.22

6.503.263.100.33

9 a

48.81.34

17.32.076.580.147.26

11.72.730.110.530.670.130.130.13

99.62

3<2

20576

100501475

2309528

140<50

<5

0.6523.1034.6318.2410.49

5.213.002.540.31

Sample

1 0 s

48.71.12

18.91.935.750.127.00

11.82.730*071.100.570.110.170.12

100.19

8<2

2107

N.F.75411065

2107532

140<50

<5

0.4123.1039.1215.08

9.745.612.802.130.26

11

49.10.98

17.73.044.590.127.84

12.42.370.090.850.790.090.070.13

100.18

9<2

31025<5

145476570

2006031

110<50

<5

0.5320.0537.4018.7215.110.034.411.860.21

12

49.31.10

16.02.485.930.148.35

12.12.380.110.880.710.120.110.15

99.86

6<2

36011N.F.

125444870

240703395

<50<5

0.6520.1432.6521.3415.79

1.483.602.090.28

13

49.61.10

16.63.404.860.157.65

12.02.530.151.000.720.090.100.16

100.11

N.D.<2

31017

<5135

906050

2709528

110<50

7

0.86

0.8921.4133.5020.2514.00

_

4.932.090.21

14

48.70.97

16.62.954.790.148.65

12.42.300.101.200.870.070.630.17

100.54

N.D.N.F.

41018

<5165

956545

2108528

120<50

<5

0.5919.4634.6820.9913.38

2.294.281.840.17

15

49.61.11

16.62.036.050.148.24

U.S2.510.060.740.810.110.070.14

100.01

5<2

285

7<5

110492870

2008528

115<50

<5

0.3521.2433.8619.2715.93

2.302.942.110.26

16

48.51.08

18.03.354.560.137.59

12.02.670.160.981.010.100.040.13

100.30

8<2

24018

<5135

446070

1807525

140<50

<5

0.9522.5936.6617.65

9.563.594.862.050.24

17

48.61.57

14.64.305.990.188.48

10.92.750.231.201.280.140.180.17

100.57

9<2

30513<5

160437090

270110

35125

<50<5

1.3623.2726.8220.9814.39

1.396.232.980.33

18

49.31.12

16.61.926.170.148.79

11.72.530.081.030.680.100.040.13

100.33

4<2

3008

<59040

670

2207529

115<50

<5

0.4721.4133.7119.0312.99

5.702.782.130.24

19

49.71.15

16.13.255.100.158.67

11.72.650.110.850.710.100.040.15

100.43

N.F.<2

28518N.F.

115506570

2307028

120<50

<5

0.6522.4231.7220.4314.82

1.524.712.180.24

2 0 a

49.41.05

17.01.895.920.148.88

11.72.440.071.180.430.090.100.14

100.43

8<2

31586

105551460

2107029

120<50

<S

0.4120.6435.2317.8414.68

4.832.741.990.21

Note: 1 = 395-11-2, 62-64 cm; 2 = 395-12-2, 109-111 cm; 3 = 395-14-1, 131-132 cm; 4 = 395-15-1, 74-76 cm; 5 - 395-17-1,56-59 cm; 6 = 395-18-1,61-70 cm; 7 = 395A-7-1, 76-82 cm; 8 395A-8-1,50-52 cm; 9 =395-A-15-3, 128-142 cm; 10 = 395A-15-5, 0-11 cm; 11 = 395A-22-1, 87-92 cm; 12 » 395A-22-2, 72-76 cm; 13 • 395A-25-1, 38-43 cm (#5); 14 = 395A-26-2, 24-33 cm (#1C); 15 = 395A-27-1, 127-131 cra;16 =395A-33-2, 9-13 cm; 17 = 395A-57-1, 125-131 cm; 18 = 395A-61-2, 3745 cm; 19 = 395A-62-1, 80-87 cm; 20 = 395A-63-1,108-116 cm. N.F. = Not Found. N.D. = Not Determined. "Others" includes trace elementscalculated as oxides. Analytical Methods; Si; Ti; Al; Fe (Total); Mn; Mg; Ca; P; Nb; Zr; and Y: XRF analysis of fused beads prepared from 500 mg of sample + 2.000 g of lithium metaborate. FeO: Hydrofluoric/sulphuric acid dissolution followed by titration with standardized potassium permanganate [Reference: French and Adams, Analyst (London), v. 97 p. 828, 1972]. H2O~: Weight loss after 1-2 hours at 110°C.H 2O+ and CO2 (= total carbon): using a Perkin-Elmer 240 Elemental Analyzer, samples dried at 110°C, flux added. Results corrected to "Sample as received" basis. B: Colorimetric method using curcumin(Analyst: C. J. Elliott). Remainder: Atomic absorption analysis of hydrofluoric boric acid solution of samples after Langmyhr and Paus (Reference: Anal. Chim. Acta., v. 43, p. 397, 1969).

a = Interlaboratory Standards

586