34. PETROLOGY OF ULTRAMAFIC ROCKS FROM SITE 395

S. Arai,1 Geological Institute, University of Tokyo, Tokyo, Japanand

T. Fujii2, Carnegie Institute, Geophysical Laboratory, Washington, D.C.

ABSTRACTSerpentinized peridotites recovered at DSDP Site 395 preserve

the original minerals: olivine, orthopyroxene, clinopyroxene, andchromian spinel. The estimated equilibration temperature indicatesthat they were equilibrated under upper-mantle conditions, and thecompositions of constituent minerals indicate that they are residuesleft after removal of more fusible components as basaltic magmas.It is suggested, therefore, that they are the fragments of residualoceanic upper mantle, which were emplaced into oceanic crust assolid intrusion along faults.

They are mineralogically similar to the ultramafic tectonites ofophiolite complexes; this supports the hypothesis that ophiolitecomplexes are generated at or near the mid-oceanic ridges and thattheir ultramafic tectonites represent the oceanic upper mantle.

INTRODUCTION

On the basis of the similarity between the oceaniccrust-mantle section, derived mainly from geophysicaldata, and the stratigraphic sequences of ophiolite com-plexes in orogenic regions, it is widely believed thatophiolite complexes are slices of the ancient oceanicplates (e.g., Coleman, 1971). Geochemical studies ofbasic rocks of ophiolite complexes support this hypoth-esis, and suggest that ophiolite complexes were formedat mid-oceanic ridges (e.g., Moores and Vines, 1971),although some ophiolite complexes might have beengenerated in other tectonic environments (Miyashiro,1973).

The ultramafic rocks from oceanic regions may pos-sibly be equivalent to the ultramafic rocks of ophiolitecomplexes (e.g., Prinz et al., 1976), but, detailed petro-graphical and mineralogical comparisons are stillscarce. It is important, therefore, to compare the petrol-ogy of ultramafic rocks from the mid-oceanic ridgeswith that of ultramafic rocks in ophiolite complexes oforogenic regions. It is also important to understand thepetrochemical and mineralogical characteristics of oce-anic peridotites for further understanding of the oce-anic crust and mantle.

Peridotites from Hole 395, Core 18, may be suitablefor the above purposes, since they are relatively fresh,with original minerals preserved.

1 Now at Geoscience Institute, Shizuoka University, Shizuoka,Japan.

2 Now at Geological Institute, University of Tokyo, Tokyo,Japan.

PETROGRAPHIC DESCRIPTION

Two samples were available from Core 395-18: 18-1, Piece 2C and 18-2, Piece 17B. Although accordingto the Leg 45 Initial Core Descriptions Samples 18-1,2C is from a harzburgite zone and Sample 18-2, 17B isfrom a lherzolite zone, both seem to be lherzolitewithin the small available samples, and contain olivine,orthopyroxene, clinopyroxene, and spinel.

Sample 18-1, 2C is relatively fresh; 40 to 50 per centserpentinized. Yellow to brown serpentine (lizardite-crysotile) has selectively replaced olivine, whereas py-roxenes are almost intact. Carbonate, magnetite, pyr-rhotite, and clay minerals are present as alteration pro-ducts. Brucite and. amphibole are absent. Olivine con-tains lamellae of thin, parallel flakes of yellow-brownspinel. No definite relationship between the orientationof spinel flakes and the crystallographic direction ofhost olivine is evident, however. Olivine also includesirregularly shaped chromian spinels. Symplectite, possi-bly composed of chromian spinel and clinopyroxene, isoccasionally present as thin plates in olivine. Orthopy-roxene contains numerous fine lamellae of clinopyrox-ene in its central part. Clinopyroxene is round or irreg-ular in shape, and in some cases interstitial to orthopy-roxene and olivine. Conspicuous parting is commonlydeveloped. Fine-grained irregular clinopyroxene inter-grown with fine-grained chromian spinel is occasional-ly present. Large grains of brown chromian spinel alsooccur.

Sample 18-2, 17B is moderately to intensely serpen-tinized, and markedly deformed. Olivine rarely con-tains rods of chromian spinel. Pyroxenes are stronglybent and kinked, and show remarkable wavy extinc-tion. Conspicuous parting is also evident. Exsolution la-

587

S. ARAI, T. FUJII

mellae seem to be thicker than those in Sample 18-1,2C, and often grow to exsolution blebs. Thin clinopy-roxene occupies the crack produced by deformation ofthe orthopyroxene grain. The bent pyroxene grains aresurrounded by fine-grained granular aggregate com-posed of pyroxenes, olivine, and spinel (mainly clino-pyroxene). These fine-grained zones seem to be recrys-tallized from bent pyroxene. Coarse-grained olivine isless deformed than pyroxenes, and only shows kinkbands and weak wavy extinction. Coarse-grained spinelis very rare.

MINERAL CHEMISTRYMinerals were analyzed using the JXA-5 type micro-

analyzer of the Geological Institute, University of To-kyo. Correction methods are the same as those ofNakamura and Kushiro (1970). Except for spinel, ironwas treated as ferrous. Ferric iron in spinel was calcu-lated to bring RO/R2O3 to unity.

Olivine. The Mg/(Mg + Fe*) ratios of olivines rangefrom 0.899 to 0.913 in Sample 18-1, 2C, and from0.902 to 0.913 in Sample 18-2, 17B. As shown in Fig-ure 1, olivines of Sample 18-1, 2C tend to be less mag-nesian than those of Sample 18-2, 17B. NiO contentsare about 0.4 wt per cent in both samples. The compo-sition of fine-grained granular olivine (the product ofrecrystallization) in Sample 18-2, 17B is nearly identi-cal to that of the coarse-grained olivine (Table 1).

Orthopyroxene. The chemical compositions, care-fully determined to avoid contamination of visible cli-nopyroxene lamellae, show that the Mg/(Mg + Fe*)ratio ranges from 0.899 to 0.912 (average 0.907), andthe Ca/(Ca + Mg + Fe*) ratios are around 0.015 (Fig-ure 2) in both samples. The averaged compositions ofthe central part of large orthopyroxene ("homogenizedcore" compositions), which may represent the compo-

TABLE1Selected Analyses of Olivines

10

5

f an

alys

esN

umb

ers

c c

5

0

-

-

i , , m I

πI

-

i , , r

J"

h π i

18-1,

i i i

18-2,

i i i i

2C

-

1

17B

-

I.92 .89.91 .90

Mg/(Mg+Fe)

Figure 1. Histograms showing compositions of olivines.

SiO2

A12O3

Tiθ2FeO*MnOMgOCaONa2θCr2θ3NiO

Total

Fo

1

39.80.03nd9.520.10

50.50.070.000.000.42

100.490.4

2

39.70.00nd9.700.12

50.40.080.000.010.38

100.4

90.2

3

40.40.010.009.850.12

49.10.060.000.00nd

99.5

89.9

4

40.60.020.008.580.13

50.60.030.000.12nd

100.191.3

5

40.30.050.009.060.12

50.60.030.000.07nd

100.2

90.9

6

40.10.040.008.620.11

50.80.010.000.11nd

99.8

91.3

7

39.8ndnd9.740.08

50.20.05ndnd0.40

100.3

90.2

8

39.5ndnd9.210.11

50.00.02ndnd0.45

99.3

90.6

Note: nd: not determined; (1) 01-2, 18-1; (2) 01-3, 18-1; (3) in contact with cρx,18-l;(4) in contact with spinel, 18-1; (5) in contact with spinel, 18-2; (6) in contact withspinel, 18-2; (7) large olivine, 18-2; (8) fine-grained, granular, 18-2.

;o. ;

O18-1

• 18-2

.90 .89

Figure 2. Compositions of orthopyroxenes plotted in theMg-rich portion of the triangle Ca-Mg-Fe* (total iron asFe). Those of "homogenized core" of the large grainsare enclosed by a broken line.

sitions before exsolution, were estimated by averagingmany broad-beam analyses (more than 20 points).Their Ca/(Ca + Mg + Fe*) ratios are around 0.04 (Fig-ure 2), and the Mg/(Mg + Fe*) ratios range from0.906 to 0.909 (average 0.908) in both samples.

The Al and Cr contents are higher in the "homoge-nized core" of large orthopyroxene and in orthopyrox-ene exsolution blebs in clinopyroxene than in the otherorthopyroxene grain.

Clinopyroxene. The Ca/(Ca + Mg + Fe*) ratios ofclinopyroxene in direct contact with orthopyroxene areall higher than 0.465. Some clinopyroxenes from Sam-ple 18-2, 17B have slightly higher Ca/(Ca + Mg + Fe*)ratios than those from Sample 18-1, 2C (Figure 3).The Mg/(Mg + Fe*) ratios are higher than 0.92 inboth samples. "Homogenized core" of clinopyroxene(host clinopyroxene + orthopyroxene lamellae) hasnoticeably low Ca/(Ca + Mg + Fe*) ratios (lower than0.42). Its Mg/(Mg + Fe*) ratio is generally lower than0.915 (average 0.913) in both Sample 18-1, 2C andSample 18-2, 17B. The Na2O content in clinopyroxeneof Sample 18-1, 2C averages 0.2 wt per cent, and isabout twice as high as that of Sample 18-2, 17B (Tables2A and 2B).

The Al and Cr contents show a highly positive cor-relation with one another, and are especially high in

588

PETROLOGY OF ULTRAMAFIC ROCKS

C a 5 M 95 ,, .03 .04 .05 .03 .04 .05

Figure 3. Compositions of clinopyroxenes plotted in thetriangle Ca-Mg-Fe*. The large circles are the clinopy-roxenes in direct contact with orthopyroxene. Compo-sitions of "homogenized core " of the large grains areenclosed by broken lines.

the "homogenized core" of the large clinopyroxene(average 4.7 wt % A12O3 and 1.5 wt % Cr2O3) and inclinopyroxene exsolution lamellae in orthopyroxene(Tables 2A and 2B).

Chromian spinel. Large spinel grains tend to bemore chromian than the fine-grained ones intergrownwith clinopyroxene in both samples. Chromian spinelin Sample 18-1, 2C is slightly more chromian than thatin Sample 18-2, 17B; the Cr/(Cr + Al) ratio of theformer is from 0.32 to 0.43, whereas that of the latter is

from 0.28 to 0.41 (Figure 4A). Mg/(Mg + Fe2 +) ratioof chromian spinel ranges from 0.6 to 0.7 (Figure 4A;Table 3) in both samples. Ferric iron content is gener-ally low, and the Fe3+/(Cr + Al + Fe3 +) ratio never ex-ceeds 0.05 (Figure 5A; Table 3). Chromian spinel rodsin olivine, common in Sample 18-1, 2C, have composi-tions almost identical to the other types of spinel.

Chromian spinels of these samples are enriched inAl compared with those from harzburgite dredged at

TABLE 2ASelected Analyses of Pyroxenes From 18-1 Lherzolite

SiO2

AI2O3TiO 2

FeO*MnOMgOCaONa?OC r 2 O 3

Total

1

51.14.360.133.330.09

19.620.7

0.161.48

101.0

2

50.04.040.122.470.04

17.923.5

0.291.41

99.8

Numbers of atoms on 0=6

SiAl4Al*TiFe*MnMgCaNaCiMgCaFe

1.8440.1560.0300.0040.1010.0031.0530.8010.0110.042

53.941.0

5.1

1.8380.1620.0130.0030.0760.0010.9810.9240.0200.041

49.546.6

3.8

3

54.23.100.036.160.11

35.80.830.020.89

101.1

1.8620.126

0.0010.1770.0031.8370.0310.0010.024

89.81.58.7

4

53.63.410.076.170.12

34.52.280.030.95

101.1

1.8530.139

0.0020.1780.0031.7770.0850.0020.026

87.14.18.7

5

51.44.150.162.580.08

16.723.1

0.211.40

99.8

1.8790.1210.0580.0040.0790.0030.9110.9040.0150.041

48.147.7

4.2

6

54.93.300.056.440.10

33.20.870.000.86

99.7

1.9080.0920.0370.0010.1870.0031.7220.0320.0000.024

88.71.79.6

7

51.13.020.102.750.04

18.923.1

0.180.85

100.0

1.8690.130

0.0030.0840.0011.0290.9040.0130.025

51.044.8

4.2

8

55.22.62nd6.38nd

33.50.85nd0.66

99.2

1.9250.0750.033

0.186

1.7430.032

0.01888.9

1.69.5

Note: nd: not determined; (1) 3-cpx "homogenized core"; (2) 3-cpx rim; (3) 3-opx rim;(4) 3-opx "homogenized core"; (5) cpx lamella of No. 6.; (6) opx host of No.5.;(7)2-cpx,fine-grained, intergrown with spinel; (8) 4-opx, in contact with chromian spinel.

TABLE 2BSelected Analyses of Pyroxenes From 18-2 Lherzolite

Siθ2AI2O3TiO2

FeO*MnOMgOCaO‰ 2 θCr 2 O 3

Total

1

51.14.330.122.650.07

17.023.4

0.141.44

100.3

2

54.53.560.006.110.16

33.91.030.000.89

100.2

Numbers of atoms on 0=6

Si

Al4Al6TiFe*MnMgCaNaCrMgCaFe

1.8620.1380.0480.0030.0810.0020.9240.9160.0100.042

48.147.7

4.2

1.8880.1130.0330.0000.1770.0051.7510.0380.0000.024

89.11.99.0

3

53.74.440.036.610.14

33.10.900.001.08

100.0

1.8680.1320.0500.0010.1920.0041.7150.0340.0000.030

88.41.79.9

4

50.15.480.132.710.08

16.223.8

0.131.61

100.0

1.8330.1670.0690.0040.0830.0030.8810.9330.0090.047

46.549.2

4.4

5

55.52.780.056.380.18

33.70.790.000.59

100.0

1.9220.0780.0360.0010.1850.0051.7390.0300.0000.016

89.01.59.5

6

52.53.270.162.360.08

17.024.0

0.090.95

100.4

1.9050.0950.0450.0040.0720.0020.9210.9340.0070.027

47.848.5

3.7

7

55.82.690.016.320.17

34.50.740.000.42

100.7

1.9190.0810.0280.0000.1820.0051.7670.0270.0000.015

89.41.49.2

8

52.23.010.062.540.09

17.723.4

0.050.66

99.7

1.9070.0930.0360.0020.0780.0030.9620.9170.0040.019

49.246 9

4.0

9

54.73.280.076.050.13

33.12.130.020.87

100.4

1.8950.1050.0290.0020.1750.0041.7110.0790.0020.024

87.14 08.9

10

51.74.920.143.540.12

19.619.3

0.101.38

100.8

1.8580.1420.0670.0040.1070.0041.0490.7430.0070.039

55.339.1

5.6

Note: (1) 7-cpx, lamella in No. 2; (2) 7-opx, host of No. 1;(3) 9-opx, bleb in No. 4, (4) 9-cpx, host of No. 3; (5) rim of adiscrete grain, in contact with No. 6; (6) rim of a discrete grain, in contact with No. 5; (7) 5-opx, granular, in contactwith No. 8; (8) 5-cpx, granular, in contact with No. 7; (9) "homogenized core" of a large grain; (10) "homogenizedcore" of a large grain.

589

S. ARAI, T. FUJII

. 8 -

1

-

-

1

•

•é

G

1

1

1

1

1

(A]

1

1

O

O

1

1

00

(9

I

i

°o

c

o

1

1

(B)

-

.2 -

1.0 .8 .6 .4 .2 1.0 .8 .6 .4 .2 CMg/(Mg+Fe2+)

Figure 4. Compositions of spinels in the diagram Cr/(Cr + Al) - Mg/(Mg + Fe2+). (A) Spinels fromultramafic rocks recovered near the Mid-Atlantic Ridge. Symbols: open circle, Sample 18-1,2C; solid circle, Sample 18-2, 17B; open square, Iherzolite from the Vema fracture zone (Prinzet al, 1976); solid square, harzburgite from 45°N on the Mid-Atlantic Ridge (Aumento andLoubat, 1971). (B) Spinels from the Newfoundland ophiolite complex. Symbols: open circle,harzburgite; solid circle, Iherzolite.

45 °N on the Mid-Atlantic Ridge (Aumento and Lou-bat, 1971), and are similar to or more chromian thanthose from lherzolite obtained from north of Vemafracture zone (Prinz et al., 1976) (Figures 4A and 5A).

When compared with spinels from the ultramafic tec-tonite of an ophiolite complex (Malpas and Strong,1975) and those from lherzolite xenoliths in alkalinebasalts and worldwide alpine peridotites, chromian spi-

Lherzolite(Xenolith & Alpineperidotite

Dunite & Harzburgite(Alpine peridotite)

3+Al Al Al Fe

Figure 5. Compositions of spinels in the triangle Cr-Al-Fe3+. (A) Spinels from ultramaficrocks recovered near the Mid-Atlantic Ridge. Symbols are the same as for Figure 4A.(B) Spinels from the Newfoundland ophiolite complex. Symbols are the same as forFigure 4B. (C) Compositional fields of spinels from Alpine-type dunite-harzburgiteand from Iherzolites of Alpine-type and of xenoliths in basaltic rocks. References arein Appendix.

590

PETROLOGY OF ULTRAMAFIC ROCKS

nels from oceanic peridotites show the transitionalcompositions between those of lherzolite and harzbur-gite in the Cr/(Cr+Al) and Mg/(Mg + Fe2+) ratios(Figures 4B, 5B, and 5C).

TABLE 3Selected Analyses of Chromian Spinels

SiO2

A12O3

TiO2

FeO*MnOMgOCaONa2OCr2O3

NiOTotal

1

0.0532.60.07

17.20.35

14.60.19nd

34.80.11

100.0

Numbers of atoms on 0=4

SiAlTiFe*MnMgCaCrNiMg/Mg+Fe"CrAlFe'"

SiO2AI2O3TiO2

FeO*MnOMgOCaONa2OCr2O3

NiOTotal

0.0021.1310.0020.4230.0090.6410.0060.8110.0030.6400.4050.5640.031

6

0.0233.70.13

15.70.35

15.90.000.00

34.7nd

100.5

Numbers of atoms on 0=4

SiAlTiFe*MnMgCaCrMg/Mg+Fe"CrAlFe'"

0.0011.1510.0030.3800.0090.6850.0000.7960.6830.3960.5730.030

2

0.0433.20.14

15.80.28

15.50.010.00

35.70.15

100.8

0.0011.1380.0030.3850.0070.6720.0000.8220.0030.6680.4090.5660.025

7

0.0435.70.08

16.10.33

15.60.000.00

32.0nd99.9

0.0011.2180.0020.3910.0080.6710.0000.7310.6690.3650.6090.026

3

0.1031.70.12

17.90.31

14.20.010.02

35.40.08

99.8

0.0031.1130.0030.4450.0080.6310.0000.8330.0020.6270.4130.5520.034

8

0.0939.60.05

16.80.32

15.30.010.04

27.7nd99.9

0.0031.3310.0010.4010.0080.6490.0000.6250.6480.3120.6640.024

4

0.0038.00.05

18.50.33

15.30.01nd

28.40.15

100.7

0.0001.2820.0010.4440.0080.6530.0000.6450.0040.6480.3200.6360.044

9

0.1642.1

0.0415.90.24

15.90.020.00

25.0nd99.4

0.0051.4010.0010.3760.0060.6710.0010.5570.6700.2780.7000.022

5

0.1436.10.04

18.20.31

14.80.010.04

30.6nd

100.2

0.0041.2320.0010.4410.0080.6410.0000.700

0.6380.3480.6130.039

10

0.1240.9

0.0115.80.27

16.10.090.02

27.4nd

100.7

0.0031.3510.0000.3710.0060.6750.0030.6070.6740.3030.6750.022

Note: nd: not determined; (1) fine-grained spinel intergrown withcpx, 18-1; (2) core of a large anhedral grain, 18-1; (3) protruded"arm" of No. 2, 18-1; (4) 2-spinel, in contact with secondarymagnetite, 18-1; (5) fine-grained spinel intergrown with cpx, 18-1; (6) core of a large anhedral grain, 18-2; (7) protruded "arm"of No. 6, 18-2; (8) 11-spinel, in contact with opx, 18-2; (9) 12-spinel, in contact with olivine, 18-2; (10) 1-spinel, granular, fine-grained, and associated with clinopyroxene, 18-2.

ESTIMATION OF TEMPERATURES

The partition of Mg and Fe between clinopyroxeneand orthopyroxene is sensitive to equilibration temper-ature, and the partition coefficient, i:D

cPx-oPx = (xMg

cpx/XFe

cPx) (XFe°Px/XMg°Px), decreases with increasingequilibration temperature (e.g., Kretz, 1961). Twokinds of clinopyroxene-orthopyroxene pairs are ob-tained from Sample 18-1, 2C and 18-2, 17B perido-tites: one is the pair of two discrete pyroxenes in con-tact with one another, and the other is the pair of cli-nopyroxene lamella (bleb) and orthopyroxene host, orvice versa. Obtained KD values (Table 4) are around1.2 or 1.3 in both Samples 18-1, 2C and 18-2, 17B, al-though KD values of the lamella-host pairs in Sample18-2, 17B (average 1.16) seem a little smaller thanthose of other pairs. In Sample 18-2, 17B, KD values ofdiscrete pairs both from the fine-grained granular partand from the coarse-grained part give the nearly samevalue. It is noteworthy that the KD values of the core-core pairs of large discrete pyroxene ("homogenizedcore") are considerably lower (1.05 in Sample 18-1,2C and 1.07 in Sample 18-2, 17B), indicating the hightemperature of equilibration (e.g., Kretz, 1961).

The KD values for experimental magnesian pyroxenepairs are around 1.1 at about 1100°C (Akella andBoyd, 1973; Hensen 1973) and about 0.9 at 1410°C(Hensen, 1973), assuming no ferric iron in pyroxenes.On the basis of the above experimental results, the KDvalue of 1.2 or 1.3 probably denotes temperaturesaround 900 °C. On the other hand, the KD value

TABLE 4Determined Kjj"s of Partitioning of Mg-Fe (total iron)

Between Clinopyroxene and Orthopyroxene, andBetween Clinopyroxene and Olivine

Samplecpx-opx(Mg-Fe) Occurrence of the Pair

18-1

18-2

Sample

18-118-2

1.221.25

(1.05)

1.251.271.121.331.37

(1.07)

1.27

1.31

1.151.161.161.161.19

cpx-olKD (Mg-Fe)

1.18,1.24,1.19,1.21,

discrete-discretediscrete-discrete"homogenized core"discrete-discretecpx lamella in opxcpx lamella in opxdiscrete-discretediscrete-discretediscrete-discrete"homogenized core"discrete-discretediscrete-discrete

fine-grained, granulardiscrete-discrete

fine-grained, granularcpx lamella in opxcpx lamella in opxcpx lamella in opxcpx transudation around opxopx bleb in cpx

1.25,1.29 and 1.321.22, 1.27 and 1.28

591

S. ARAI,T.FUJII

smaller than 1.1 of "homogenized core" pairs indicatestemperature in excess of 1100°C, and may indicate thetemperature of igneous activity prior to recrystalliza-tion.

Partition of Fe and Mg between olivine and clinopy-roxene is also sensitive to temperature. The partitioncoefficient, KD<v*•«p* = (XMgcp*)/XFe<Φ*).(XFe°VXM Aincreases with decreasing temperature (e.g., O'Haraand Mercy, 1963); a simple relation between the KD

value and temperature cannot be obtained, however,probably because of the composition dependence ofthe KD value (Obata et al., 1974; Wood, 1976). Whenthe Mg/(Mg + Fe*) ratios of olivine and clinopyroxenepairs are plotted in a Roseboom diagram, (Figure 6)however, it is possible to deduce approximate equili-bration temperature by comparison with other pairs ofknown temperatures (Mori and Banno, 1973). The oli-vine-clinopyroxene pairs from both Samples 18-1, 2Cand 18-2, 17B peridotites fall in the region betweenthose from Seiad and Finero ultramafic complexes(Medaris, 1975) and the line which represents a parti-tion coefficient of unity, indicating they may have beenequilibrated at temperatures higher than 750°C (Me-daris, 1975). The olivine-clinopyroxene pairs fromlherzolite inclusions in an alkaline basalt (Frey andGreen, 1974) fall in almost the same region as thosestudied here. Frey and Green (1974) argued that thelherzolite inclusions were equilibrated at temperatureslower than 1050°C. From the above discussion, wesuggest that olivine-clinopyroxene pairs from the perid-otites of Samples 18-1, 2C and 18-2, 17B were equili-brated at temperatures between 750 °C and 1050 °C,and probably at 900 °C.

.90

Mg/(Mg+Fe) Olivine

Figure 6. Mg-Fe partitioning between clinopyroxene andolivine. Symbols: open circle, Sample 18-1, 2C; solidcircle, Sample 18-2, 17B; open square, lherzolite xeno-liths in an alkaline basalt (Frey and Green, 1974); soliddiamond, Finero complex (Medaris, 1975); open dia-mond, Seiad complex (Medaris, 1975).

Although an olivine-spinel geothermometer was de-veloped by Irvine (1965) and Jackson (1969), temper-atures estimated by their methods seem too high andno experimental calibration is available. For compara-tive purposes, however, Fe-Mg partition between oli-vine and spinel should be useful. As shown in Figure 7,the olivine-spinel pairs from both Samples 18-1, 2Cand 18-2, 17B peridotites fall in the region betweenthose from Seiad and Finero ultramafic complexes andthose from lherzolite inclusions in the west Victoria al-kaline basalt (Frey and Green, 1974), which suggeststhat the equilibration temperature of these oceanic pe-ridotites is between 750°C and 1050°C (Medaris,1975; Frey and Green, 1974).

DISCUSSION

It is not clear whether these peridotites were em-placed into their present position in the oceanic crustby intrusion along a fault or were derived as sedimen-tary fragments from a peridotite body first exposed atthe surface along a fault scarp, then covered by laterlava flows. It is important, however, that serpentinizedperidotites were recovered at the shallower portion ofoceanic crust at Site 395, where no major fracture zoneexists nearby. This may support the theory that serpen-tinized peridotite is a significant component of the nor-mal oceanic crust, where it is emplaced by vertical in-trusion into a fault zone (Bonatti and Honnorez, 1976).

The internal consistency of the estimated equilibra-tion temperatures implies that peridotites in Samples18-1, 2C and 18-2, 17B were well equilibrated at sub-solidus temperature (less than 1000°C, probably900 °C). Such temperatures are too high for oceaniccrust, and it is most likely that these peridotites wereequilibrated in the oceanic upper mantle, then intrudedinto the oceanic crust along faults as solid intrusions.

There is no evidence that these two samples are cu-mulates from mafic or ultramafic magma. Considering

.4 .6Cr/(Cr+AI) in Spinel

Figure 7. Mg-Fe partitioning between spinel and olivine.Symbols are the same as for Figure 6.

592

PETROLOGY OF ULTRAMAFIC ROCKS

that olivine and pyroxene are very magnesian [(Mg/(Mg + Fe*) ratios are higher than 0.9], that K-bearingminerals are absent, and that Na content in clinopyrox-ene and consequently in bulk compositions is very low,Samples 18-1, 2C and 18-2, 17B may be residual oce-anic upper mantle left after removal of more easily fu-sible components as basaltic magmas.

Among the ultramafic rocks recovered near the Mid-Atlantic Ridge, an entire spectrum exists, from thelherzolite to the harzburgite subtypes (e.g., Aumentoand Loubat, 1971; Prinz et al., 1976), and most, in-cluding two samples studied here, show the prominentdeformation and recrystallization textures and a resid-ual bulk compositional character commonly recognizedin ultramafic tectonites of ophiolite complexes. Further-more, ultramafic rocks corresponding to the ultramaficcumulates of the ophiolite complex have been also ob-tained at DSDP Site 334 (Hodges and Papike, 1976).It is probable, therefore, that the ophiolite complexesfor the future are produced at or near the Mid-AtlanticRidge, and the peridotites studied here are of such anophiolite complex of the future.

Although the similarity between the tectonites of theophiolite complexes and the ultramafic rocks from theMid-Atlantic Ridge was emphasized above, there aresome differences. In the tectonites of the ophiolite com-plexes, dunite and harzburgite subtypes are predomi-nant and lherzolite and lherzolitic harzburgite are rare(e.g., Malpas and Strong, 1975), whereas lherzolite orlherzolitic harzburgite seem to be more common in theMid-Atlantic Ridge region (e.g., Prinz et al., 1976). Inthis sense, the stratigraphic section of the Mid-AtlanticRidge region may be different from that of the ophio-lite complex. The difference may be only apparent,however, because all the ultramafic rocks of the Mid-Atlantic Ridge region were recovered from the shallowportion of the oceanic crust or from fracture zones, andconsequently they do not always represent the spec-trum of the normal oceanic upper mantle.

Another difference is the equilibration temperature.Most peridotites from ultramafic tectonites of ophiolitecomplexes are equilibrated near 700 °C (e.g., Medaris,1975), whereas the samples studied here show higherequilibration temperatures. The low equilibration tem-peratures of ultramafic tectonites are believed by some(e.g., Medaris, 1975) to result from the effect of re-gional metamorphism which took place after they mi-grated with oceanic crust at the converging plateboundaries. Considering that ultramafic tectonites indi-cate higher equilibration temperature (e.g., granulitefacies) than the surrounding basic rocks (e.g., greenschist or amphibolite facies), however, such interpreta-tion may not be correct. It is more likely that ul-tramafic complexes generated near the mid-oceanicridges cool to about 700 °C during transport from themid-oceanic ridges to converging boundaries, becausethe thermal gradient of oceanic upper mantle becomesmore gentle toward the marginal part (e.g., Solomon,1976). The peridotite complexes then are emplaced ascold tectonites at the converging plate boundaries toform ophiolite complexes. If regional metamorphism

continues after emplacement, those ultramafic tecton-ites may be equilibrated at a temperature similar to thesurrounding basic rocks, as is the case of Mt. Higashi-Akaishi (Mori and Banno, 1973). If the metamorphismdoes not continue, those ultramafic tectonites show anequilibration temperature higher than the surroundingbasic rocks, as is the case of Finero complex (Medaris,1975).

ACKNOWLEDGMENTS

The authors are indebted to Drs. B. O. Mysen and N. T.Arndt for their critical review and useful suggestions.

REFERENCES

Akella, J. and Boyd, F. R., 1973. Effect of pressure on thecomposition of coexisting pyroxenes and garnet in the sys-tem CaSiθ3-MgSiθ3-FeSiθ3-CaAlTi2O6, Carnegie Inst.Wash. Yearbook, v. 72, p. 523-526.

Aumento, F. and Loubat, H., 1971. The mid-Atlantic Ridgenear 45 °N. XVI. Serpentinized ultramafic intrusions, Ca-nadian J. Earth Sci., v. 8, p. 631-663.

Basu, A. R., 1977. Olivine-Spinel equilibria in lherzolitexenoliths from San Quintin, Baja California, Earth PlanetSci. Lett., v. 33, p. 443-450.

Bonnatti, E. and Honnorez, J., 1976. Sections of the earth'scrust in the equatorial Atlantic, J. Geophys. Res., v. 81, p.4105-4116.

Coleman, R. G., 1971. Plate tectonic emplacement of uppermantle peridotites along continental edges, J. Geophys.Res.,v.76,p. 1212-1222.

Frey, F. A. and Green, D. H., 1974. The mineralogy, geo-chemistry and origin of lherzolite inclusions in Victorianbasanites, Geochim. Cosmochim. Acta, v. 38, p. 1023-1059.

Hensen, B. J., 1973. Pyroxenes and garnets as geothermome-ters and barometers, Carnegie Inst. Wash. Yearbook, v.72, p. 527-534.

Hodges, F. N. and Papike, J. J., 1976. DSDP Site 334: Mag-matic cumulates from oceanic Layer 3, /. Geophys. Res., v.81, p. 4135-4151.

Irvine, T. N., 1965. Chromian spinel as a petrogenetic indica-tor. Part 1—Theory, Canadian J. Earth Sci., v. 2, p. 648-674.

Jackson, E. D., 1969. Chemical variation in coexisting chro-mite and olivine in chromitite zones of the StillwaterComplex, Econ. Geol. Monogr., v. 4, p. 41-71.

Kretz, R., 1961. Some applications of thermodynamics to co-existing minerals of variable composition. Examples: or-thopyroxene-clinopyroxene and orthopyroxene-garnet, /.Geol.,v. 69, p. 361-387.

Malpas, J. and Strong, D. F., 1975. A comparison of chrome-spinels in ophiolites and mantle diapirs of Newfoundland,Geochim. Cosmochim. Acta, v. 39, p. 1045-1060.

Medaris, L. G., 1975. Coexisting spinel and silicates in alpineperidotites of the granulite facies, Geochim. Cosmochim.Acta, v. 39, p. 947-958.

Miyashiro, A., 1973. The Troodos ophiolitic complex wasprobably formed in an island arc, Earth Planet. Sci. Lett.,v. 19, p. 218-224.

Moores, E. M. and Vine, F. J., 1971. The Troodos massif,Cyprus and other ophiolites as oceanic crust: evalutionand implications, Phil. Trans. Royal Soc. London, ser. A,v. 268, p. 443-466.

Mori, T. and Banno, S., 1973. Petrology of peridotite andgarnet clinopyroxenite of the Mt. Higashi-Akaishi mass,central Shikoicu, Japan—subsolidus relation of anhydrousphases, Contrib. Mineral. Petrol, v. 41, p. 301-323.

593

S. ARAI, T. FUJII

Nakamura, Y. and Kushiro, I., 1970. Compositional relationsof coexisting orthopyroxene, pigeonite and augite in atholeiitic andesite from Hakone volcano, Contrib. Mineral.Petrol, v. 26, p. 265-275.

Obata, M, Banno, S., and Mori, T., 1974. The iron-magne-sium partitioning between naturally occuring coexisting ol-ivine and Ca-rich clinopyroxene: an application of thesimple mixture model to olivine and solution, Bull. Soc.franc. Mineral. Crist., v. 97, p. 101-107.

O'Hara, M. J. and Mercy, E. L. P., 1963. Petrology and pe-trogenesis of some garnetiferous peridotites, Trans. Roy.Soc. Edinburgh, v. 65, p. 251-314.

Prinz, M., Keil, K., Green, J. A., Reid, A. M., Bonatti, E., andHonnorez, J., 1976. Ultramafic dredged samples from theequatorial Mid-Atlantic Ridge and fracture zones, / . Geo-phys. Res., v. 81, p. 4087-4103.

Solomon, S. C, 1976. Geophysical constraints on radial andlateral temperature variations in upper mantle, Am. Min-eralogist, v. 61, p. 788-803.

Wood, B. J., 1976. An olivine-clinopyroxene geothermome-ter: A discussion, Contrib. Mineral. Petrol., v. 56, p. 297-303.

APPENDIXReferences Not Cited in the Text but Used to Construct

Figure 5C

Aoki, K. and Prinz, M., 1974. Chromian spinels in lherzolite inclu-sions from Itinome-gata, Japan, Contrib. Mineral. Petrol, v. 46,p. 249-256.

Arai, S., 1976. Petrology of Alpine-type ultramafic complexes inSangun zone, western Japan. Unpublished thesis, University ofTokyo.

Basu, A. R., 1975. Hot-spots, mantle plumes and a model for the ori-gin of ultramafic xenoliths in alkali basalts, E.P.S.L., v. 28, p.261-274.

Basu, A. R. and MacGregor, I. D., 1975. Chromite spinels from ul-tramafic xenoliths, Geochim. Cosmochim. Acta, v. 39, p. 937-945.

Best, M. G., 1974. Contrasting types of chromium-spinel peridotitexenoliths in basanitic lavas, western Grand Canyon, Arizona,E.P.S.L., v. 23, p. 229-237.

Challis, G. A., 1965. The origin of New Zealand ultramafic intru-sions, / . Petrol, v. 14, p. 113-131.

Green, D. H., 1964. The petrogenesis of the high-temperature perid-otite intrusion in the Lizard area, Cornwall, J. Petrol, v. 5, p.134-188.

Hamad, S. el D., 1963. The chemistry and mineralogy of the olivinenodules of Calton Hill, Derbyshire, Min. Mag., v. 33, p. 483-497.

Himmelberg, G. R. and Coleman, R. G., 1968. Chemistry of primaryminerals and rocks from the Red Hill-Del Puerto ultramaficmass, California, U.S.G.S. Prof. Paper, 600-C, p. 18-26.

Himmelberg, G. R. and Loney, R. A., 1973. Petrology of the VulkanPeak Aipine-type peridotite, southwestern Oregon, Geol. Soc. Am.Bull, v. 84, p. 1585-1600.

Irving, A. J., 1974. Pyroxene-rich ultramafic xenoliths in the newerbasalts of Victoria, Australia, N. Jb. Min., v. 120, p. 147-167.

Komatsu, M., 1975. Recrystallization of the high-alumina pyroxeneperidotite of the Uenzaru area in Hidaka province, Hokkaido, Ja-pan, / . Geol. Soc. Japan, v. 81, p. 11-28.

Kornprobst, J., 1966. A propos des peridotites du Massif des Beni-Bouchera (Rif septentrional, Maroc), Bull. Soc. franc. Mineral.Crist, v. 89, p. 399-404.

Littlejohn, A. L. and Greenwood, H. J., 1974. Lherzolite nodules inbasalts from British Columbia, Canada, Canadian J. Earth Sci.,v. 11, p. 1288-1308.

Loney, R. A., Himmelberg, G. R, and Coleman, R. G., 1971. Struc-ture and petrology of the Alpine-type peridotite at Burro Moun-tain, California, U.S.A., / . Petrol, v. 12, p. 245-309.

MacGregor, I. D. and Smith, C. H., 1963. The use of chrome spinelsin petrographic studies of ultramafic intrusions, Canadian Min., v.7, p. 403-412.

Medaris, L. G., 1972. High-pressure peridotites in southwestern Ore-gon, Geol. Soc. Am. Bull, v. 83, p. 41-58.

Miller, R. Ill, 1953. The Webster-Addie ultramafic ring, JacksonCounty, North Carolina, and secondary alteration of its chromite,Am. Mineralogist, v. 38, p. 1134-1147.

Monchoux, P. and Besson, M., 1969. Sur les compositions chimiquesdes minéraux des lherzolites pyrénnéennes et leur significancesgénétique, Bull. Soc. franc. Min. Crist., v. 92, p. 289-298.

O'Hara, M. J., Richardson, S. W., and Wilson, G., 1971. Garnet-pe-ridotite stability and occurrence in crust and mantle, Contrib.Mineral. Petrol, v. 32, p. 48-68.

Rio, M., 1968. Quelques precisions sur la composition minéralogiquede la lherzolite de Moun Caou (Basses-Pyrenees), Bull. Soc.franc. Mineral Crist., v. 91, p. 298-299.

Ross, C. S, Foster, M. D., and Myers, A. T., 1954. Origin of dunitesand of olivine-rich inclusions in basaltic rocks, Am. Mineralogist,v. 39, p. 693-737.

594