26. GEOCHEMISTRY AND PETROGENESIS OF BASALTS FROM DEEP SEADRILLING PROJECT LEG 92, EASTERN PACIFIC1

J. A. Pearce, N. Rogers, A. J. Tindle, and J. S. Watson, Department of Earth Sciences, The Open University,Milton Keynes2

ABSTRACT

Basalts recovered on DSDP Leg 92 include all the major basalt types so far recovered from the ocean crust of theeastern Pacific. Basalts from Holes 597, 597A, 597B, 597C, and 599B are tholeiites exhibiting all the mineralogical andgeochemical characteristics of N-type mid-ocean ridge basalts (MORB). Fragments of ferrobasalts and alkali basaltswere also obtained, however, from Holes 60IB and 602B, respectively. Hole 597C, which penetrated 91 m into basementand is the deepest hole so far drilled in fast-spreading crust, yielded basalts that can be divided into three major litho-logic units. The lowest unit, Unit III, contains modal olivine and comprises basalts which, at about 8 to 10% MgO, areas basic as any sampled from fast-spreading crust. The middle unit, Unit II, is the most evolved; its basalts are olivinefree and contain between 6 and 7.5% MgO. The upper unit, Unit I, is intermediate in composition between Units II andIII; it is characterized by both modal olivine and glomerocrysts made up of plagioclase and rare olivine. Unit I is proba-bly a massive flow, whereas Units II and III may be massive flows or sills. The basalts appear to have undergone threestages of alteration ("deuteric," "relatively reducing," and "oxidizing"), the intensity of alteration decreasing markedlydowncore. Hole 597B, at 26.4 m of basement penetration the only other "deep" hole, contains just one lithologic unit,which closely resembles Unit I of Hole 597C.

Petrogenetic modeling reveals that the three lithologic units in Hole 597C are cogenetic and that they were derivedfrom a depleted mantle source similar to the source of the tholeiites and ferrobasalts sampled in other holes; the alkalibasalts are the only rocks derived from enriched mantle. Lavas of Unit III probably lay on the olivine-plagioclase cotec-tic, whereas the other lavas lay on an olivine-plagioclase-clinopyroxene peritectic. Some 60% of closed-system crystalli-zation is needed to generate the most-evolved from the last-fractionated tholeiite, and a further 50% crystallization(80% overall) is needed to generate the ferrobasalts. Xenocrysts of calcic plagioclase and pseudomorphosed olivine intholeiites from Hole 597B and Unit I of Hole 597C, and in the ferrobasalts from Hole 601B, provide evidence, however,that some magma mixing may have taken place.

INTRODUCTION

In this chapter we present whole-rock and mineral anal-yses for the basalts recovered during Leg 92 of the DeepSea Drilling Project. We then use these data for routineclassification of the basalts and for evaluation of theirpetrogenetic histories. Finally, we assess the contribu-tion that these results make to our understanding ofprocesses at fast-spreading ridges, with particular refer-ence to the continuing debate on the role of steady-statemagma chambers, the origin and setting of ferrobasalts,and the nature and extent of mantle heterogeneities.

The settings and locations of the Leg 92 sites aresummarized in Figure 1, together with details of the ba-salt recovery at each site. Site 597 (about 28.5 Ma) is lo-cated on crust formed at the now-extinct Mendoza Rise,which was transformed into the East Pacific Rise by aseries of ridge jumps between 20 and 6 Ma (Rea, 1981).Site 598 (about 17 Ma), Site 599 (about 8 Ma), and Sites600 to 602 (about 4.5 Ma) are all located on crust formedat the East Pacific Rise itself. Most significant as re-gards igneous processes is Hole 597C; this is now thedeepest hole (91 m basement penetration) in fast-spread-

Leinen, M., Rea, D. K., et al., Init. Repts. DSDP, 92: Washington (U.S. Govt. Print-ing Office).

2 Addresses: (Pearce, present address) Department of Geology, The University, Newcas-tle-upon-Tyne NE1 7RU, United Kingdom; (Rogers, Tindle, and Watson) Department ofEarth Sciences, The Open University, Milton Keynes MK7 6AA, United Kingdom.

ing crust, the previous deepest being Hole 319A of DSDPLeg 34 (59 m), which was also drilled into GalapagosRise crust (Yeats, Hart, et al., 1976). Core recovery inHole 597C was also high (48.5 m). This hole thus pro-vides the best opportunity so far to evaluate magmachamber processes at fast-spreading ridges. Holes 598 to602, by contrast, yielded only small quantities of basalt,but these do extend the geographic range of samples re-covered from the eastern Pacific (see Scheidegger andCorliss, 1981), and they should therefore provide fur-ther information on compositional heterogeneities with-in the crust.

SAMPLING PROCEDURE AND ANALYTICAL TECHNIQUES

A total of 103 minicore samples was chosen for analysis from thebasalts recovered during the leg. At least one sample was taken per1.5-m section, and areas of particular interest, such as possible cool-ing-unit boundaries, were sampled more intensively. Since the objec-tive was to study the primary igneous geochemistry, the freshest partsof each section of core were sampled, and veins were avoided wherepossible. A slice from each minicore was used to prepare a polishedthin section, and the remainder (about 50 g) was crushed, first by ajaw-crusher and then to about 240 mesh by agate "tema" mill. Blanksprepared by this method showed negligible contamination for the ele-ments of interest.

All samples were analyzed by Meca 10-44 energy-dispersive X-rayfluorescence (XRF) spectrometer for major elements (using fused discs)and for the trace elements Zr, Y, Nb, Cr, V, Ni, Rb, Sr, Cu, and Zn(using powder pellets). Full details of the technique are given by Pottset al. (1984); results obtained on three international standards, run asunknowns, and on a shipboard standard are listed in Appendix A. Theprecision of the data is considered to be less than 5% for all elements

435

J. A. PEARCE, N. ROGERS, A. J. TINDLE, J. S. WATSON

15°N

0° -

30°S -

150° 135° 120° 105° 75°W

Figure 1. Site locations of DSDP Leg 92 and summary of basement drilling. Inverted triangle denotes re-entry.

of interest, except Rb, Nb, P, and K, which are close to their detectionlimits in the analyzed samples and have precisions of 10 to 20%. Thefull table of XRF analyses is given as Appendix B.

A subset of 18 representative samples was also analyzed by instru-mental neutron activation analysis (INAA) for the rare earth elements(REEs), Th, Ta, Hf, Co, and Sc. Aliquots of 300 mg were irradiated atthe University of London reactor center at Ascot, and spectra wereprocessed at the Open University, as described by Paul et al. (1975)and Potts et al. (1981). Full analyses of these samples are given inTable 1.

Microprobe analyses of mineral phases were performed at The OpenUniversity on an automated two-spectrometer Cambridge InstrumentsMicroscan 9, using a defocused beam and operating conditions of20 kV and 30 nA. Natural and synthetic minerals were used as stan-dards. The data are summarized in Tables 2 to 4 and are listed in full inAppendix C (pyroxenes), Appendix D (feldspars), and Appendix E(oxides). Only a tiny quantity of fresh glass and olivine was found:analyses of this material are given in Table 4.

LITHOLOGIES

Detailed petrographic descriptions of the core are pre-sented by Goldfarb (this volume). Here, we summarizethe principal features relevant to the identification oflithologic units and to the construction of petrogeneticmodels.

Hole 597BAll core recovered from this hole consists of slightly

vesicular basalt. Neither glass nor pillow structures werefound, and the core is thus inferred to have sampled oneor more massive flows. No geochemical or mineralogi-cal boundaries were found within the core, but a num-ber of distinctive dark, fine-grained zones, typically about10 cm thick, do occur, and these have been tentativelyinterpreted as internal cooling zones within a single lith-ologic unit.

The primary mineralogy is typically plagioclase, cli-nopyroxene, and titanomagnetite, together with up to5% olivine and, in the finer-grained samples, some cryp-tocrystalline groundmass. The rocks are only rarely por-

phyritic, in terms of the grain size of individual crystals,but they do usually contain up to 10% glomerocrystsmade up of plagioclase with occasional euhedral oliv-ine. A number of plagioclase crystals within these glo-merocrysts are distinctly calcic; they also typically exhibitmore sodic rims and an internal zone of glass inclu-sions—evidence of a period of supercooling and injec-tion into magma more evolved than their parent mag-ma.

The degree of alteration is moderate throughout thecore; all olivine and cryptocrystalline groundmass havebeen pseudomorphosed by clay minerals and iron oxy-hydroxides. Studies of the sequences of vesicle and veininfilling and the mineralogy of alteration halos to inter-secting veins reveal an essentially threefold sequence ofalteration: an early (possibly deuteric) stage character-ized by blue smectite; a second (relatively reducing) stagecharacterized by dark green smectite, together with talc,chlorite, and rare sulfides and native copper; and a final(relatively oxidizing) stage characterized by brown smec-tite, calcite, iron oxyhydroxides and low-temperature ze-olites. This sequence is thus broadly similar to that de-scribed by Bass (1976) for basalts drilled during DSDPLeg 34. Owing to the high fracture- and vesicle-permea-bility of all recovered core, all the samples analyzed hadexperienced at least one episode of alteration.

Hole 597CAs in Hole 597B, no obvious flow boundaries were

identified in the core recovered. Significant mineralogi-cal and geochemical variations were observed, however,as illustrated in Figure 2. These provide the basis for theproposed subdivision into three lithologic Units: Unit I(the uppermost unit, 48 m thick); Unit II (the intermedi-ate unit, about 20 m thick); and Unit III (the lowermostunit, at least 13 m thick). Petrographically, Unit I can beseen to contain both plagioclase glomerocrysts and oliv-

436

GEOCHEMISTRY AND PETROGENESIS OF BASALTS, LEG 92

Table 1. "Complete" geochemical analyses for selected samples of basalts from DSDP Leg 92.

HoleCore-SectionLevel, cmUnit

597B2-155

597B3-227

597C3-152I

597C4-1126

I

597C4-5120

I

597C5-224I

597C6-1128

I

597C7-1121

597C7-5901

597C8-24311

597C9-1129II

597C10-1119II

597C11-193III

597C12-120III

599B4-114

601B2-143

601B2-1105

602B1-299

SiO2TiO2A12O3Fe2O3MnOMgOCaONa2OK2OP2O5L o r

Total

ZrYNbRbSrCrNiVCuZn

HfTaThScCo

LaCeNdSmEuTbHoTmYbLu

49.741.08

14.6910.630.168.12

12.732.640.220.120.52

100.65

6727

4.14.7

8533412733212381

1.590.150.30

45.550.0

2.57.78.02.450.970.721.200.472.890.48

50.101.22

14.6010.520.157.20

12.411.970.110.160.32

98.76

7631

4.12.5

92278151342124

89

1.920.200.40

35.550.0

3.09.37.32.831.130.761.100.483.19

0.53

49.681.13

14.4410.190.168.05

12.242.020.150.130.69

98.88

7026

3.74.3

8331211327712474

1.680.190.39

45.648.4

3.28.17.12.550.990.701.100.432.870.52

50.251.34

14.7611.410.167.26

11.802.560.100.110.19

99.94

8430

4.63.3

9119675

35213684

2.210.240.33

47.746.2

3.810.19.83.101.180.871.360.533.310.57

49.711.32

14.6211.620.167.36

11.772.340.090.120.41

99.52

8032

5.13.6

82211

7735312188

2.030.230.30

46.243.1

3.29.28.92.971.120.811.480.523.240.56

50.371.21

15.5011.250.167.94

12.112.500.090.140.53

101.80

7029

3.12.3

86283

8933011575

1.840.210.50

33.043.0

2.78.67.42.681.020.681.200.473.010.50

49.951.19

15.3010.980.166.69

12.412.190.450.110.65

100.08

72304.6

12.593

27968

28012585

1.850.200.26

44.445.0

3.09.08.02.731.030.711.130.463.010.51

49.791.15

15.0811.630.157.35

12.232.170.090.130.23

100.00

7427

4.32.6

84324

8830613180

1.890.210.27

45.444.5

3.28.78.22.671.030.690.980.502.880.50

49.961.20

15.1911.140.177.62

12.202.220.070.150.40

100.32

7228

3.83.2

9326610129810267

1.910.210.30

43.843.0

3.28.77.42.671.030.721.100.442.960.50

50.621.46

13.9512.670.187.45

11.262.840.190.170.17

100.96

8733

4.03.8

8214579

35211777

2.230.260.24

47.547.2

3.910.29.83.261.220.881.630.563.500.60

50.251.54

13.5213.350.197.25

11.142.610.200.160.09

100.30

9035

5.24.0

838856

40612180

2.450.280.25

47.249.9

4.210.510.23.371.240.871.170.563.620.60

50.261.38

13.7412.320.187.53

11.592.680.100.150.24

100.17

8030

3.93.0

8110782

361200

77

2.040.220.20

46.444.1

3.19.08.12.831.060.741.200.413.200.52

49.520.93

17.049.580.148.27

12.881.960.040.130.75

101.24

5420

3.31.4

85374112260

9560

1.320.150.22

38.341.5

2.16.36.01.940.790.530.800.352.170.37

49.331.05

14.4910.390.178.25

12.552.110.070.090.58

99.08

6225

4.12.5

7931910031813564

1.570.170.30

34.843.0

2.67.77.52.340.920.621.100.412.640.45

50.511.17

14.959.960.167.27

12.432.930.190.090.42

100.08

7632

2.17.2

94264

8225310284

1.950.000.60

44.642.0

2.37.97.32.841.130.781.200.453.090.52

49.752.58

12.6816 510.255.669.842.830.220.230.10

100.65

19269

4.93.0

1127457

42781

138

4.860.270.33

44.744.0

6.318.818.66.532.181.612.401.016.281.06

50.762.62

12.9314.550.215.289.702.780.480.280.61

100.20

19068

4.48.0

1245232

45055

116

5.000.240.40

33.639.0

5.219.220.6

6.602.231.602.201.006.641.10

47.523.00

15.5711.120.185.668.874.311.570.671.09

99.56

3115642.019.6

340135113257

53102

6.262.473.13

32.230.0

26.259.435.2

8.162.691.431.850.714.760.77

Note: Elements Si to Zn were analyzed by XRF, elements Hf to Lu by INAA. Nondetected la is given as 0. SiO2 to LOI in wt.%; Zr to Lu in ppm.a LOI = loss on ignition.

CorePI 01

Zr (ppm) Cr (ppm)

CO

500On3-

12-

21 -

-£ 3 0 -

.c

g• 3 9 -•σ

2 48-oo

57-

66-

75-84-

91 -

Figure 2. Geochemical stratigraphy of Hole 597C as exemplified bythe elements Zr and Cr. Core sections containing plagioclase glo-merocrysts (PI) and modal olivine (Ol) are indicated by stars inboxed column.

ine and thus resembles the core from Hole 597B. Unit IIbasalts are aphyric and olivine free. Unit III basalts aregenerally aphyric and contain up to 10% modal olivine.Geochemical profiles, exemplified in Figure 2 by the in-compatible element Zr and the compatible element Cr,also show significant inflections at these petrographicboundaries. It might also be argued, on the basis ofFigure 2, that the uppermost 5 m of Unit I belong to afourth unit; but we decided to restrict the subdivision tothe three major groups that can be identified by bothpetrographic and geochemical characteristics.

Interestingly, neither inferred boundary correspondsto an obvious textural or structural discontinuity. Theboundary between Units I and II is located in a regionof low core recovery between Cores 7 and 8, but that be-tween Units II and III is located within Core 10, whererecovery was high. The latter boundary was originallythought to correspond to one of the enigmatic fine-grainedzones, but can be seen in Figure 2 to be a transitionalrather than an abrupt boundary. It was also thoughtthat Units II and III could represent a single flow inwhich olivine accumulated at its base: chemical varia-tion diagrams (presented later) show, however, that ba-salts within these units do not lie along olivine controllines. We therefore suggest that this discontinuity—andperhaps the other one as well—represents an internalflow boundary whereby a new pulse of magma is inject-

437

J. A. PEARCE, N. ROGERS, A. J. TINDLE, J. S. WATSON

ed within a partly cooled flow. This hypothesis is evalu-ated in more detail in subsequent sections.

Alteration, vesicularity, and texture vary systematical-ly with depth. The basalt is slightly vesicular in the top45 m, below which it is nonvesicular. The three stages ofalteration identified in Hole 597B are also found here.Alteration intensity is greatest in the vesicular part ofthe core, below which it is restricted to halos aroundveins. In the lower half of the core, segments betweenveins can be very fresh, although olivine (if present) isinvariably altered. Textures range from spherulitic toophitic, except for a single 1-cm-wide glassy rind in theuppermost fragment. Ophitic textures are most commonin Units II and III, where zoned poikilitic crystals of au-gite up to 2 cm across are common. Feldspars some-times exhibit layering in this part of the core. The orderof crystallization is clearly (1) olivine, (2) plagioclase,(3) clinopyroxene, (4) Ca-poor pyroxene (Unit III), (5)titanomagnetite.

Holes 599A, 601B, and 602BAlthough the quantity of basalt recovered from these

holes is very small, the petrographic diversity is great.At Site 599, the basalts are almost nonvesicular, withspherulitic to intergranular textures. They are slightly pla-gioclase phyric and contain a groundmass of altered ol-ivine, plagioclase, titanomagnetite, and cryptocrystallinematrix. They thus strongly resemble Unit I of Hole 597C.

At Site 601, the basalts are typically made up of mi-crophenocrysts of plagioclase, clinopyroxene, and (al-tered) olivine in a groundmass of altered olivine, plagio-clase, clinopyroxene, and patches of yellow glass. In somesamples, coalesced spherulites of plagioclase fibers andclinopyroxene dendrites containing abundant interstitialmagnetite give the sample an opaque appearance in thinsection, a feature described by Natland (1980) as typicalof ferrobasalts. Sulfide inclusions in magnetite are com-mon. The microphenocrysts include isolated tabular andskeletal crystals of plagioclase and glomerocrysts con-taining plagioclase + olivine and plagioclase + clino-pyroxene. The largest olivine and clinopyroxene crystalshave euhedral outlines; the largest plagioclase crystalsare often calcic and occasionally exhibit resorbed rims.

At Site 602, the basalt samples are vitrophyric andcontain microphenocrysts of pink skeletal titanaugite,olivine, and plagioclase.

MAJOR-ELEMENT CHARACTERISTICSBoth primary and secondary major-element variations

can be recognized in the analyzed samples. Secondaryvariations are most obvious in the losses on ignition(LOIs) and in the abundances of elements of lowest ion-ic potential, notably K and Na, both of which vary sys-tematically with alteration intensity. Figure 3 shows atypical plot, that of K2O against LOI. In the most al-tered units (Hole 597B and Unit I of Hole 597C), thefreshest samples are characterized by K2O values of lessthan 0.2% and by LOI values of less than 0.5%; withincreasing alteration intensity (vesicle and open-space in-filling), both K2O and LOI increase. Alteration direc-tion 1 is most common, reflecting the replacement of ol-

ivine and cryptocrystalline groundmass and the infillingof vesicles by smectite; direction 2 is followed only byveined samples containing other (less potassium-rich) hy-drous and carbonated minerals. High K2O values in themore evolved basalts of Hole 60IB may, however, be ex-plained by fractional crystallization, as well as by altera-tion. In the freshest unit, Unit II of Hole 597C, almostall samples contain less than 0.2% K2O and 0.4% LOI.In Unit III of Hole 597C, however, LOI can be higheven though K2O concentration is generally very low.Since this unit differs from Unit II only by the presenceof modal olivine, it can be inferred that the alterationdirection (3) followed by these samples is due to the re-placement of olivine by the (potassium-free) hydratedoxides and hydroxides of iron.

From the evidence we have presented, it is possible touse the dashed line in Figure 3 to screen analyses fromHoles 597B and 597C before plotting them on petrologi-cal diagrams that are based on the alkali elements. Mo-bilities of other major elements—such as loss of Mg inUnits I and III of Hole 597C, resulting from olivinebreakdown—are less important, as demonstrated by thecoherent variations in the diagrams based on these ele-ments.

Figure 4 shows the distribution of screened data onthe projection of the normative basalt tetrahedron (Yo-der and Tilley, 1962). In common with most other MORBsuites, almost all samples are tholeiitic (olivine or quartztholeiites), the only exception being the basalt from Hole602B, which is Ne-normative and can hence be classi-fied as "alkalic." Of the tholeiitic samples, the mostevolved samples (i.e., those from Unit II of Holes 597Cand 60IB) plot predominantly in the quartz-tholeiite field,whereas other samples (from Hole 597B and Units I andII of Hole 597C) plot predominantly in the olivine tho-leiite field. Altered samples (not plotted in Fig. 4) showthe same intergroup relationships but a greater spreadtoward the olivine and nepheline apices.

2.00

1.00

0.00

× 597B

+ 597C

o 597C (I)

Δ 597C (II)

O 601 B (III)

~2

0.00 0.50 1.00LOI (wl.%)

1.50 2.00

Figure 3. Plot of K2O against LOI (loss on ignition) for tholeiitic ba-salts. Vectors, 1,2, and 3 indicate chemical changes due to altera-tion (see text); samples treated as "fresh" plot within the rectanglein the bottom left of the diagram.

438

GEOCHEMISTRY AND PETROGENESIS OF BASALTS, LEG 92

Ne-

602 B601 B

Ol Hy

Figure 4. Projection of analyses of "fresh" samples onto the diopside-hypersthene-olivine plane of the normative basalt tetrahedron. Theellipses represent the 90% probability contours for the lithologicunits indicated.

On the AFM diagram of Figure 5, all except the sam-ple from Hole 602B follow the iron-enrichment trendcharacteristic of the tholeiitic series. In Hole 597C, UnitIII defines the most basic end of the trend, and Units Iand II exhibit progressive enrichment in iron. The sam-ples from Hole 60IB show the greatest enrichment. Thelatter have total FeO contents between 12.66 and 14.06wt.% and TiO2 concentrations between 2.58 and 2.73wt.%, thus satisfying the criteria (>12.0% and >2.0%,respectively: Byerly et al., 1976; Natland, 1980) for clas-sification as "FETI" basalts or "ferrobasalts." Alteredsamples (not plotted on this diagram) show the same in-terrelationships but are displaced to higher alkali values(i.e., toward "A").

Figure 5. Projection of analyses of "fresh" samples onto the AFM tri-angle. As in Figure 4, the lithologic units are represented as 90%probability contours.

TRACE-ELEMENT CHARACTERISTICS

The principal trace-element characteristics are illus-trated in Figures 6 and 7 by means of geochemical pat-terns. In the conventional rare-earth-element (REE) pat-terns shown in Figure 6, all the tholeiitic basalts canbe seen to exhibit light-REE-depleted patterns with nor-malized La/Yb ratios generally between 0.6 and 0.8 andnormalized intermediate- to heavy-REE abundances be-tween 10 and 18, all within the 2-sigma range for N-typeMORB (Basaltic Volcanism Study Project, 1981). The

50

10

Hole 597CUnit I

Unit II

Unit III

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

50

10

La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Figure 6. Chondrite-normalized rare earth patterns. A. Hole 597C. B.Other Leg 92 basalts. Normalizing values are taken from Naka-mura (1974).

439

J. A. PEARCE, N. ROGERS, A. J. TINDLE, J. S. WATSON

ferrobasalts of Hole 601B exhibit patterns that are simi-lar in shape but are displaced to higher abundances andhave small negative Eu anomalies; these patterns strong-ly resemble the REE patterns described by Clague et al.(1981) and Perfit et al. (1983) for ferrobasalts from theGalapagos spreading center. The pattern for the alkalibasalt of Hole 602B shows an enrichment in the lightREEs comparable to that observed in E-type MORBsuites from the North Atlantic (Tarney et al., 1980).

"Complete" trace-element patterns, normalized to anestimated primordial mantle composition (Wood, 1979),are presented in Figure 7. The patterns of all tholeiiticbasalts and ferrobasalts show depletion in the most litho-philic elements, a feature characteristic of N-type MORB.Moreover, all patterns are subparallel, indicating deriva-tion from sources of similar composition, in this casethe incompatible-element-depleted convecting upper man-tle. The alkali basalt pattern, which shows a relative en-richment in the most incompatible elements, is compa-rable to that of E-type MORB. The negative Sr anoma-lies indicate the extent of plagioclase fractionation, whichincreases from Unit III of Hole 597 to Hole 601B.

MINERAL CHEMISTRY

Pyroxenes

Pyroxenes from Leg 92 are almost all groundmass crys-tals, which have been subdivided in Appendix C intovariolitic, intergranular, and ophitic types. Some repre-sentative analyses are given in Table 2. Figure 8 showsthe distribution of compositions on the pyroxene quad-rilateral. The most common trend is that exhibited byHoles 597B and 597C, in which Ca decreases and theFe/Mg ratio increases with fractionation. Unit III ofHole 597C additionally contains a significant group ofCa-poor pyroxenes. The few samples analyzed from theferrobasalts of Hole 601B have similar (i.e., augitic) com-positions, whereas the titanaugite microphenocrysts fromHole 602B are relatively calcic, plotting on the Di-Hedtie-line. The latter also plot, predictably, in the "alka-line" fields of pyroxene classification diagrams, such as

1OOr

8 1 0

E

601B* 602B

597C (II)597C (I)597C (III)

Rb Th K Ta Nb La Ce Sr Nd P Hf Zr Sm Ti Tb Y Yb

Figure 7. Primordial-mantle-normalized trace-element patterns for some"fresh" basalts representative of the various lithologic units. Nor-malizing values are taken from Wood (1979).

the SiO2-Al2O3 diagram of Le Bas (1962). The interpre-tation of these trends, which are typical of moderatelyundercooled MORB suites from other areas, is describedin detail in the literature (e.g., Muir and Mattey, 1982;Mattey and Muir, 1980). The final three columns of Ta-ble 2 also demonstrate the strong zonation found insome ophitic crystals: the rims of these crystals exhibitthe greatest iron enrichment of any Leg 92 pyroxenes.

Plagioclase

Figure 9 summarizes, in histogram form, the range offeldspar compositions listed in Appendix D. From this,it is evident that there is a bimodal distribution of com-positions in Holes 597B, 597C (Units I and III), and601B, the bulk of the analyses ranging from An50 toAn75 and a second, smaller group ranging from An80 toAn90. The latter compositions are restricted to glomero-crysts, where they form large, weakly zoned crystals withirregular, resorbed or thin, more sodic margins. Probetraverses across four such crystals are given in Table 3.Zones of glass inclusions sometimes occur between thecores and sodic margins of the crystals, indicating a sud-den period of rapid growth. These features have beenrecognized in a number of other MORB suites (see de-scriptions by, e.g., Dungan and Rhodes, 1978; Natlandet al., 1983; Stakes et al., 1984), where they have beenused as evidence that the calcic crystals are xenocrystspicked up from the margins of the magma chamber orintroduced into the magma chamber via a more primi-tive host. The less calcic group of analyses derives fromboth glomerocryst and matrix crystals, and it clearly crys-tallized in equilibrium with the host magma. These crys-tals are predictably most calcic, on average, in Unit IIIof Hole 597C and most sodic in Hole 601B.

Other Phases

Unfortunately, two of the genetically important phas-es, olivine and glass, are poorly represented in the Leg92 basalts, glass because of the paucity of pillow mar-gins and olivine because of ubiquitous alteration. Theonly data obtained in this study are presented in Table 4and are insufficient for making detailed interpretations.Titanomagnetite analyses are given in Appendix E andshow compositions similar to those described by Maz-zullo and Bence (1976) for Leg 34 basalts, plotting on orclose to the Fe2Tiθ4-Fe2θ3 tie-line on the TiO2-FeO-Fe2O3 triangle. Ilmenite does occur, primarily in Unit IIof Hole 597C; the composition of a typical ilmenite-magnetite pair is given in Table 4, and further analysesare listed by Nishitani (this volume).

PETROGENETIC MODELING

As we have demonstrated, the tholeiites from Holes597B and 599B and from the three lithologic units ofHole 597C, together with the ferrobasalts from Hole601B, have trace-element patterns that indicate deriva-tion from depleted mantle sources of similar composi-tions. We now carry out petrogenetic modeling to inves-tigate the partial melting of these mantle sources andthe subsequent evolution of the primary magma. Theprinciples are illustrated by the ternary diopside-anor-

440

GEOCHEMISTRY AND PETROGENESIS OF BASALTS, LEG 92

Table 2. Average electron microprobe analyses (wt.%) of pyroxenes from DSDP Leg 92 basalts.

HoleUnitPetrologya

No. of analyses

597B

V2

597B

I2

597B

P12

597CIV9

597CII9

597CIP16

597CIII6

597CIIP15

597CIIIV1

597CIIII13

597CIIIP6

597CIII*I6

599B

I3

601B

I3

602B

P2

597BRim

P1

597BInt.P1

597BCore

P1

SiO2TiO2A12O3FeOMnOMgOCaONa2OCr2O3

Total

CaMgFe

51.140.843.628.480.22

15.8819.650.260.24

50.470.974.089.760.26

14.9419.000.360.17

50.980.852.53

12.880.33

14.6717.340.270.09

51.200.873.349.600.25

15.7717.480.240.13

51.730.742.94

10.090.26

15.7118.630.250.17

51.350.793.119.210.24

16.0118.790.250.13

51.150.721.87

14.290.38

14.6216.980.260.01

51.620.672.11

12.060.34

15.5717.510.240.04

53.430.432.445.890.16

17.0720.740.190.45

52.520.502.018.210.23

16.6019.340.290.17

53.020.432.156.520.20

17.3919.800.220.34

53.370.410.78

18.920.49

21.214.850.080.00

51.330.853.448.800.26

16.4918.120.330.10

51.701.03

2.819.740.27

16.1318.250.340.05

44.104.787.759.070.18

11.0422.080.600.09

50.541.012.52

16.210.40

14.6614.750.320.00

51.750.772.818.810.23

15.9319.380.260.09

40.645.713.7

40.043.916.1

36.342.720.1

38.745.715.6

38.541.916.3

38.946.214.9

35.042.023.0

36.044.519.5

42.248.4

9.4

39.647.313.2

40.349.310.4

9.958.230.1

37.947.714.4

37.846.515.7

49.634.515.9

30.842.724.5

40.045.814.2

52.250.412,505.390.16

16.8621.02

0.250.73

100.30 100.00 99.94 98.89 100.52 99.89 100.28 100.17 100.80 99.86 100.07 100.10 99.72 100.31 99.69 100.41 100.03 99.57

43.248.2

8.6

I = intersertal; V = variolitic; P = poikilitic; p = microporphyritic; * = pigeonite.

Hed

En Fs

602B

Hed

B

En Fs

Figure 8. Pyroxene data from Leg 92 basalts plotted on the pyroxenequadrilateral (A) as data points and as (B) 90% probability ellipsesfor the individual lithologic units.

thite-forsterite diagram in Figure 10, on which phaseboundaries for the synthetic diopside-anorthite-forster-ite system have been superimposed (Osborn and Tait,1952; Shido et al., 1971). In this system, a source com-position, S, melts to give a primary melt, P, leaving aresidue, R. The composition P then evolves by olivinefractionation to the olivine-plagioclase cotectic ^t O andthen moves to the olivine-plagioclase-clinopyroxene peri-tectic Pe. Normative compositions of the freshest aphyr-ic rocks from Leg 92 (stars on Fig. 10) have also beenplotted to identify the phase boundaries of the "real"system. Graphic and numeric modeling enable the com-positions S, P, R, C, and Pe to be determined, with theproviso that the path C to Pe can be complicated byopen-system fractionation if magma of this compositionis repeatedly mixed with primitive magma of composi-

tion between C and P. It will be noted that petrographicevidence constrains the pathway by indicating that, incommon with other N-type MORB sequences, the crys-tallization history is (1) olivine, (2) plagioclase, (3) cli-nopyroxene. The following diagrams and calculationsenable this pathway to be investigated.

Least-Squares Modeling Based on Major-ElementAbundances

Least-squares modeling (linear programming) is prob-ably the most effective way to evaluate fractionation trendsby simultaneous consideration of all major elements.The technique used here closely follows that describedby Wright and Doherty (1970) and Mallet (1978), usingthe mini-max method of calculation and incorporatingweightings according to errors associated with the in-dividual measurements. The major-element oxides usedwere SiO2, AI2O3, total iron as FeO, MgO, CaO, Na2O,and TiO2. Statistically calculated weightings were, respec-tively, 10, 25, 40, 40, 100, 10, and 7500. K2O was not in-cluded because of its sensitivity to alteration; MnO andP2O5 were omitted because of the relatively large errorsassociated with their measurement and the fact that theyprovide information similar to that provided by FeO andTiO2, respectively. Phases used were forsteritic and fay-alitic olivine (analyses taken from Deer et al., 1966), cal-cic and sodic plagioclase (Samples 597C-6-3, 69 cm [anal-ysis 1-C in Table 3] and 597B-3-2, 27 cm [analysis 1from Appendix D]) and clinopyroxene (Samples 597C-11-1, 93 cm [analysis 3] and 597C-9-4, 93 cm [analysis1] from Appendix C). To obtain an underdetermined setof equations, it was necessary additionally to assume avalue for ÅT^Fe/Mg) (ol/cpx), the distribution coeffi-cient for the Fe/Mg ratio between olivine and clinopy-roxene. A value of 0.8 (see discussion and references inStakes et al., 1984) was used to model trends in whicholivine and clinopyroxene crystallized simultaneouslythroughout; otherwise, a higher value (nominally 1.0)was taken. Two sets of calculations were performed: thefirst examined the trend from the most basic unit (III)of Hole 597C to the most evolved unit (II) of Hole597C; the second investigated the trend from the mostbasic suite (Unit III of Hole 597C) to the most evolved(the ferrobasalts of Hole 60IB). More detailed break-

441

J. A. PEARCE, N. ROGERS, A. J. TINDLE, J. S. WATSON

100-1

% 50 -

597B n = 21

100 -i

% 50 -

597Unit I

n = 31

100-,

% 50 -

597CUnit I I

n = 17

100-,

% 50 -

597CUnit I I I

n= 20

k 5 0 -

0 -

599B

I ii 1ú

n= 6

i i i

1 0 0 1 601B

% 50 -

n = 13

I45 55 65 75 85 95

An (%)

Figure 9. Histograms of plagioclase compositions from the varioustholeiitic lithologic units, n = no. of analyses.

down of the fractionation trend gave poor results, as in-dicated by high sums of residuals.

The results of the first study are listed in Table 5.Samples 597C-10-5, 55 cm and 597C-9-4, 8 cm were cho-sen as the most suitable starting and residual composi-tions, respectively, both being aphyric and among thefreshest rocks sampled. To confirm the consistency ofthe results, several runs were performed using a range ofalternative compositions. It is apparent from Table 5that consistent results were obtained, bearing in mindlikely variations within each unit. This preliminary in-vestigation therefore indicates that Unit II could havebeen derived by closed-system fractional crystallizationfrom Unit III involving 48 to 69% crystallization (F =0.31 to 0.52) of an olivine (7.5 to 12.5%), clinopyroxene(32.1 to 40.9%), plagioclase (51.2 to 56.5%) assemblage.Calculated compositions of these phases are also givenin the table. Both proportions and compositions of phas-es are slightly dependent on the chosen Kd(Fc/Mg) (ol/cpx) value, as shown in the final four rows of Table 5;the removal of the oxide weightings can be seen to havenegligible effect on the solution, however.

The results of the model deriving the ferrobasalts fromUnit III of Hole 597C are shown in Table 6. Sample601B-2-1, 105 cm has been chosen as the most suitableferrobasalt for modeling, although the program was runusing each ferrobasalt in turn as the residual composi-tion. Results are again internally consistent, and indi-cate degrees of fractional crystallization between 76 and84% (F = 0.16 to 0.24) and crystallization assemblagesof olivine (7.7 to 10.4%), clinopyroxene (38.4 to 41.6%),and plagioclase (48.9 to 52.0%). This assemblage differslittle from that of Table 5, although F is predictablysmaller and the calculated phase compositions are rich-er in sodium and iron.

Graphic Modeling of Major-Element Variations

Oxide-MgO plots have been shown to be among themost effective for major-element modeling (e.g., Wright,1974). Figure 11 presents plots of CaO, FeO, and A12O3against MgO for the least altered samples (K2O < 0.20%,except Hole 601B); in each plot, the upper diagram showsthe observed variations and the lower diagram the inter-pretation in terms of the compositions of crystallizingphases. Two bulk crystallizing assemblages are shown ineach case: the first (XI) represents the assemblage forthe olivine-plagioclase cotectic, using the most calcic pla-gioclase xenocryst analysis and an estimate of the com-position of the coexisting olivine; the second (X2) repre-sents the three-phase assemblage (from Table 6) requiredto explain the fractionation from Unit III of Hole 597 tothe ferrobasalts of Hole 601B.

It is apparent from Figure 11, first, that the maincomponent of variation passes through X2, thus con-firming the result of the least-squares modeling. Unit IIof Hole 597C exhibits an additional component of vari-ation directed toward plagioclase, indicating that pla-gioclase cumulation has taken place, a result supportedby the petrographic evidence presented earlier. In addi-tion, there is a small inflection in the slope of the frac-tionation trends on both the Al2O3-MgO and CaO-MgO

442

GEOCHEMISTRY AND PETROGENESIS OF BASALTS, LEG 92

Table 3. Reconnaissance electron microprobe analyses (wt.%) of rims (R), intermediate locations (L), and cores (C) in fourfeldspar phenocrysts.

HoleCore-SectionLevel, cmPhenocryst no.-locationof analysis

SiO2A12O3FeOCaONa2OK2O

Total

An

597B2-2138

1-R

53.7628.440.86

12.684.370.04

100.15

61.5

597B2-2138

1-1

46.7633.290.35

18.141.400.00

99.94

87.8

597B2-2138

1-C

47.6833.100.38

18.131.530.00

100.82

86.8

597B2-2138

2-R

47.5832.920.51

17.561.710.00

100.28

85.0

597B2-2138

2-1

48.0232.620.53

17.511.790.01

100.48

84.3

597B2-2138

2-C

48.2332.180.49

16.972.030.01

99.91

82.2

597C6-369

1-R

49.7231.170.60

15.642.710.01

99.85

76.1

597C6-369

1-1

46.2633.370.37

18.181.310.00

99.49

88.4

597C6-369

1-C

45.8833.760.38

18.461.140.00

99.62

90.0

597C6-369

2-R

50.1730.880.65

15.492.860.01

100.06

74.9

597C6-369

2-1

45.9933.850.38

18.341.200.00

99.76

89.4

597C6-369

2-C

45.9133.920.37

18.431.200.00

99.83

89.5

Table 4. Additional microprobe data (wt.%) on phases from Leg 92basalts.

HoleCoreLevel, cm

SiO2TiO2A12O3FeOMnOMgOCaONa2OK2O

Total

Fo

597C3-13

50.910.65

14.527.900.199.33

14.711.950.05

100.21

Glass

597C3-13

50.950.78

14.487.410.179.24

15.551.910.02

100.51

597C3-13

50.650.72

13.089.510.238.16

15.501.960.06

99.87

602B1-299

40.070.05—

16.760.26

43.94———

101.35

82.4

Olivine

602B1-299

40.130.04—

16.260.26

44.36——

—

101.31

82.9

602B1-299

39.710.04—

18.410.31

42.67_—

—

101.40

80.5

Ti-magnetite/ilmenite

597C8-423

0.2619.511.60

73.190.440.270.11——

95.38

597C8-423

3.6746.920.38

44.980.461.860.07—

—

98.54

Figure 10. Diopside-anorthite-forsterite diagram showing phase fields(for 1 kb. pressure) for diopside (Di), anorthite (An), forsterite(Fo), and spinel (Sp), and the probable mantle source (S) and resi-due (R) compositions and probable crystallization paths from pri-mary magma (P) to the three-phase peritectic (Pe). Compositionsof fresh basalts (stars) have also been plotted on the basis of theirnormative pyroxene, plagioclase, and olivine proportions.

diagrams. The relative positions of XI and X2 indicatethat this inflection corresponds to the incoming of cli-nopyroxene as a crystallizing phase. If so, this takesplace at about 8% MgO, which corresponds well withthe value of 8.4% determined by Stakes et al. (1984) forthe FAMOUS area. It can also be seen from the FeO-MgO diagram that FeO decreases with decreasing MgOin the ferrobasalts, indicating that (unless this compo-nent is due to cumulation) titanomagnetite became acrystallizing phase at an MgO value of about 5.75%.

It may further be noted from these diagrams that thegood correspondence between the least-squares and graph-ic modeling, together with the linear fractionation trendsand clearly defined discontinuities, all argue for the dom-inance of closed-system fractional crystallization overthe open-system fractionation (repeated fractional crys-tallization and mixing with primitive magma) proposedby O'Hara (1977). Thus, although petrographic evidence(calcic plagioclase xenocrysts) suggests that open-systemfractional crystallization may have operated, it is likelythat few mixing cycles were involved, and unlikely that asteady-state composition was reached.

Study of Plagioclase CompositionsThe models presented in the foregoing sections ex-

plain variations observed in the erupted lavas, but donot provide information on the early history of crystal-lization in the olivine phase volume or on the olivine-plagioclase cotectic. The point at which plagioclase be-comes a crystallizing phase can, however, be estimatedfrom Figure 12, a plot of Xhxv against Mg number. Thehistogram of Figure 9 confirmed the presence, in Unit Iof Hole 597C and the ferrobasalts of Hole 601B, of cal-cic plagioclase xenocrysts (An8O-An9o). In Figure 12, twolines have been drawn to show the range of plagioclasecompositions that resulted from liquidus (upper line) tonear-solidus (lower line) crystallization from magmas ofgiven Mg number. Projection of the xenocryst composi-tions onto the linear extrapolation of the liquidus linegives a minimum range of 60 to 77 for the Mg numberof the host magma to the xenocrysts. The upper bound-ary of this range (which corresponds to an MgO valueof 11 % in the melt) gives a minimum value for the Mgnumber of the melt when plagioclase starts to crystallize

443

J. A. PEARCE, N. ROGERS, A. J. TINDLE, J. S. WATSON

Table 5. Linear programming model for derivation of Unit III lavas of Hole 597C from UnitII lavas of Hole 597C by closed-system fractional crystallization (see text for discussion).

Starting composition: 597C-10-5, 55 cm

Residue F

597C-9-2, 8 cm 0.35597C-9-3, 57 cm 0.31597C-9-4, 93 cm 0.38597C-10-2, 102 cm 0.36

Phaseproportions (%)

Ol Cpx

7.8 39.37.5 40.98.7 37.97.7 40.2

Residual composition: 597C-9-4, 93 cm

Startingcomposition F

597C-11-3, 128 cm 0.51597C-11-4, 73 cm 0.52597C-10-4, 94 cm 0.47597C-10-5, 55 cm 0.38

Phaseproportions (%)

Ol Cpx

10.7 36.311.4 32.112.5 36.38.7 37.9

Starting composition: 597C-1O-5, 55 cmResidual composition: 597C-9-4, 93 cm

Params. F

Phaseproportions (%)

Ol Cpx

PI

52.851.653.452.1

PI

53.056.551.253.4

PI

Fo

76.174.677.776.1

Fo

84.382.677.677.7

Fo

Phasecompositions (%)

Wo

38.838.239.338.8

En

48.147.648.548.0

Fs

13.214.212.213.2

Phasecompositions (%)

Wo

41.340.839.339.3

En

50.450.048.548.5

Fs

8.39.2

12.212.2

Phasecompositions (%)

Wo En Fs

An

77.977.279.779.7

An

88.886.179.479.7

An

Sum ofsquares ofresiduals

0.0090.0100.0000.007

Sum ofsquares ofresiduals

0.3520.0670.0290.000

Sum of

residuals

Model 1Model 2Model 3Model 4

0.390.380.370.39

9.3 37.08.7 37.98.3 38.79.3 37.0

53.753.453.053.7

76.177.779.075.9

40.0 49.2 10.839.3 48.5 12.238.7 48.0 13.340.0 49.1 10.9

80.079.778.779.8

Model 1: Fe/Mg(cpx):Fe/Mg(ol) = 0.8; statistical weighting of elementsModel 2: Fe/Mg(cpx):Fe/Mg(ol) = 1.0; statistical weighting of elementsModel 3: Fe/Mg(cpx):Fe/Mg(ol) = 1.2; statistical weighting of elementsModel 4: Fe/Mg(cpx):Fe/Mg(ol) = 1.0; no weighting of elements

0.0000.0000.0000.000

(point C in Fig. 10). If the fractional crystallization trendson the oxide-MgO plots in Figure 11 are extrapolatedbackward from this value, along olivine control lines, toan estimated primary magma composition at MgO =17% (Basaltic Volcanism Study Project, 1981), simpleuse of the "lever rule" then suggests that a maximum of15% of olivine crystallized before plagioclase became acrystallizing phase.

Trace-Element ModelingResults comparable to those from the major-element

modeling are obtained using logarithmic trace-elementcovariation plots with Zr as the index of fractionation(Fig. 13). These diagrams can be used to compare ob-served fractionation trends with theoretical vectors, cal-culated from the Rayleigh fractionation equation

CL = Co

for the likely crystallizing phases, olivine (o), plagioclase(p), clinopyroxene (c), and titanomagnetite (m). The vec-tors show how the melt composition will change in re-sponse to crystallization of these phases, the lengths ofthe vectors corresponding to 50% crystallization (F =0.5), except where indicated in the figure caption. Thedistribution coefficients (D) used to calculate the vectorsare taken from the compilations of Pearce and Norry(1979), Ottonello et al. (1984), Anderson and Greenland

(1969), and Philpotts and Schnetzler (1970). Providedthat neither distribution coefficients nor phase propor-tions vary markedly during basalt fractionation, linearfractionation trends should be obtained for each crystal-lizing assemblage.

Figures 13A and 13B, based on Y and P, show that allsamples plot on a single linear trend, this trend beingbetter defined using the more precise Y analyses. SinceZr, Y, and P are all incompatible with respect to the fourcrystallizing phases, the vectors and the observed trendall have slopes close to unity. The presence of a singletrend also confirms the conclusion, reached earlier, thatall the units plotted were derived from mantle sources ofsimilar compositions. Since the lengths of the vectorscorrespond to 50% crystallization, it can be seen thatsome 75% crystallization is necessary to explain the der-ivation of the ferrobasalts from magma having the com-position of Unit III of Hole 597C.

Figures 13C and 13D, based on Ti and V, were cho-sen to investigate the role of titanomagnetite as a crys-tallizing phase. As the crystallization vectors indicate,titanomagnetite crystallization causes both Ti and V tobe depleted in residual melt. That the ferrobasalts ofHole 601B plot off the main Hole 597C fractionationtrend thus indicates that a small proportion of titano-magnetite was involved in the later stages of crystalliza-tion. The relatively large scatter on the V-Zr plot can beattributed to curved mixing trends caused by plagioclase

444

GEOCHEMISTRY AND PETROGENESIS OF BASALTS, LEG 92

Table 6. Linear programming model for derivation of Hole 601B "ferrobasalts" from ba-salts of Unit II, Hole 597C by closed-system fractional crystallization (see text for dis-cussion).

Starting composition: 597C-10-5, 55 cm

Phaseproportions (%)

Phasecompositions (%)

Residual composition: 601B-2-1, 105 cm

Startingcomposition

Phaseproportions (%)

Phasecompositions (%)

Ol Cpx PI Fo Wo En Fs An

Starting composition: 597C-10-5, 55 cmResidual composition: 601B-2-1, 105 cm

Model 1: Fe/Mg(cpx):Fe/Mg(ol) = 0.8; statistical weighting of elementsModel 2: Fe/Mg(cpx):Fe/Mg(ol) = 1.0; statistical weighting of elementsModel 3: Fe/Mg(cpx):Fe/Mg(ol) = 1.2; statistical weighting of elementsModel 4: Fe/Mg(cpx):Fe/Mg(ol) = 1.0; no weighting of elements

Sum ofsquares of

Residue

601B-2-1, 16 cm601B-2-1, 43 cm601B-2-l,69cm601B-2-1, 105 cm

F

0.160.180.160.17

Ol

8.07.77.77.8

Cpx

41.240.341.641.2

PI

50.852.050.751.0

Fo

70.672.970.471.1

Wo

36.837.636.737.0

En

46.247.046.146.4

Fs

17.015.417.216.7

An

72.973.273.274.0

residuals

0.0010.0010.0010.001

Sum ofsquares ofresiduals

597C-10-4, 94 cm597C-10-5, 55 cm597C-11-3, 128 cm597C-11-4, 73 cm

0.210.170.240.24

10.47.89.29.6

40.641.241.038.4

48.951.049.852.0

70.171.172.671.3

36.637.037.537.0

46.046.446.946.4

17.416.715.616.5

72.474.079.176.1

0.0470.0010.4130.098

Params.

Model 1Model 2Model 3Model 4

F

0.180.170.160.16

Phaseproportions (%)

Ol Cpx

8.5 40.27.8 41.27.5 41.77.8 41.2

PI

51.351.050.850.9

Fo

68.671.172.470.8

Phasecompositions (%)

Wo

37.737.036.636.8

En

47.046.446.146.3

Fs

15.316.717.316.9

An

74.674.073.773.5

Sum ofsquares ofresiduals

0.0010.0010.0020.001

cumulation. This process has a much smaller effect onthe Ti-Zr plot because, in this case, the plagioclase com-position lies on a linear backward extrapolation of thefractionation trend, and fractionation and cumulationtrends are thus subparallel.

Figures 13E and 13F, based on Sr and Ni, were cho-sen to investigate the roles of plagioclase and olivine, re-spectively, during fractional crystallization. On Figure13E, the main fractionation trend appears to be subhor-izontal, although there is also a secondary trend to high-er Sr values, which reflects the effect of plagioclase cu-mulation. Assuming a plagioclase-liquid distribution co-efficient of 1.7 (Philpotts and Schnetzler, 1970), it canbe calculated that 55% plagioclase is needed to explainthe observed trend for Hole 597C, a value consistentwith the least-squares calculations (Table 5). The trendto high concentrations of Sr in the ferrobasalts could re-flect plagioclase cumulation. On Figure 13F there is ageneral trend toward lower Ni with increasing Zr. Theproportion of olivine crystallization needed to producethis trend can be calculated, but is very sensitive to thevalue chosen for the olivine-liquid distribution coeffi-cient for Ni. The coefficient of 10 used in the figuregives a proportion of about 15%; this is nearly doublethat calculated by the least-squares method, suggestingthat a coefficient of 20 may be more appropriate. Figure13F also reveals trends toward high Ni values for the

units containing glomerocrysts (Unit I of Hole 597C andHole 60IB), possible evidence for olivine cumulation.

Trace elements can be further employed to estimatedegrees of partial melting and early fractionation histo-ries, using diagrams based on a compatible element (e.g.,Cr, Ni) and an incompatible element (e.g., Y, Yb) (Pearce,1982). On such a diagram, the composition of the man-tle source can be estimated, since neither element is greatlyaffected by source heterogeneities. Moreover, partial melt-ing trends and olivine fractionation trends are almostorthogonal on the relevant part of a diagram of thistype, enabling degrees of melting to be determined byback-projection. Figure 14 shows the Ni-Y diagram. Thesource composition is estimated from nodule composi-tions. The partial melting trend is based on the meltingof a spinel lherzolite composition (60% olivine + spi-nel, 20% clinopyroxene, 20% orthopyroxene) in whichclinopyroxene disappears after 40% melting and ortho-pyroxene disappears after 60% melting. The olivine +spinel fractionation trend is drawn from the Rayleigh frac-tionation equation, and fractionation vectors are drawnto enable subsequent multiphase crystallization trendsto be interpreted. Distribution coefficients were chosenaccording to the likely melt temperature, as proposed byOttonello et al. (1984). It is apparent from this diagramthat the various basalt units all plot along fractionationtrends that are consistent with the results of the major-

445

J. A. PEARCE, N. ROGERS, A. J. TINDLE, J. S. WATSON

B1 5 r

14

13

ocβ

O 11

10

10

597C (III)

8

18

17

16

s? 15

OCM

< 14

13

12

1110

597C (III)

597C(ID

MgO (wt.%)8 7

MgO (wt.%)

25

20

15

10

ono*

pi 2

50 40 30 20MgO (wt.%)

10

50

40

cp2•ft*

//

/

//

—xi Φ

/s/

/s

/

• •?*"

+ X2

TwVßx

sf

OCM

<

30

20

10

-à• 50

p l >

40 30 20MgO (wt.%)

10

Figure 11. (A) CaO-MgO, (B) Al2O3-MgO, and (C) FeO-MgO covariation diagrams for Hole 597C basalts and Hole 601B ferrobasalts. The upperdiagrams show data points and 90% probability ellipses, the lower diagrams the ellipses and the compositions of the crystallizing phases. Thepoint XI on the olivine (oll)-plagioclase (pll) tie-line represents the likely crystallizing assemblage on the olivine-plagioclase cotectic; the pointX2 within the olivine (ol2)-plagioclase (pl2)-clinopyroxene (cp2) triangle represents the bulk crystallizing assemblage for fractionation from UnitIII of Hole 597C to Hole 601B, as calculated by least-squares analysis. Samples from Unit I of Hole 597C are shown by crosses, samples fromUnit II of Hole 597C by circles, samples from Unit III of Hole 597C by triangles, and samples from Hole 601B by diamonds.

element linear programming. This trend has been con-tinued backward until the estimated point at which theolivine-plagioclase cotectic was reached. Subsequent pro-jection along an olivine vector intersects the partial melt-ing line at between 25 and 30% melting. Of course, thismodel does depend on the parameters chosen, but no re-alistic alternative model deviated greatly from that shown:

the general effect of varying the petrogenetic parametersis demonstrated for the Cr-Y diagram by Pearce (1982).

GENESIS OF THOLEIITES FROM HOLE 597C

The geochemical data on Hole 597C have revealed asurprisingly large downhole variation, all of which can

446

16

15

14

13

12

11

10

601B

597C(III)

GEOCHEMISTRY AND PETROGENESIS OF BASALTS, LEG 92

95

45. 50 60Mg number

10 8 7 6 5MgO (wt.%)

Figure 12. Plot of plagioclase composition against Mg number of thehost rock for tholeiitic basalts from Leg 92.

on

25 r

20

15

10

* .

5

ol2.

cp2

X 2 \

×r50 40 30 20

MgO (wt.%)10

^ - * p l

Figure 11 (continued).

be explained in terms of differences in the degree of frac-tional crystallization. Nonetheless, any attempt to pro-duce a detailed model for the magma-chamber process-es operating is greatly constrained by the lack of "fieldevidence." In particular, we have insufficient informa-tion for deciding whether the three units are all lavaflows or whether any are sills; we also need more infor-mation for deciding whether or not all the units were de-rived from the same magma chamber. Despite these con-straints, however, some interpretations are possible.

The lowermost unit (Unit III) has been shown to bethe most basic; the major-element data indicate that on-

ly olivine and plagioclase were liquidus phases when themagma was erupted. It is overlain by the most evolvedunit (Unit II), for which olivine, plagioclase, and clino-pyroxene were liquidus phases. The geochemical stratig-raphy (Fig. 2) shows no sharp boundary, but rather achemical gradation over some 1.5 m. This latter obser-vation makes it difficult to interpret these units as suc-cessive flows. Moreover, extensive fractionation within asingle flow by in situ fractionation or olivine cumula-tion is not supported by the major- or trace-element var-iations. It is possible, therefore, that the transition zonerepresents a zone of mixing between two molten, or par-tially molten, magmas. The ophitic textures that domi-nate both units, the fine-grained zones that may repre-sent internal cooling zones, and the absence of vesicles,can be cited as further tentative evidence in favor of thishypothesis. If the hypothesis is correct, at least one unitcan be thought of as a sheet of magma intruded withina lava lake or pre-existing sill that had not fully crys-tallized. In addition, the petrogenetic modeling demon-strated that Unit II could be derived from Unit III by 48to 69% crystallization. If the two magmas were moltenat the same time, then their compositional differencescould indicate that they were erupted from separate cham-bers in different stages of evolution. More realistically,however, they could have been derived from a single cham-ber in which a more evolved magma was injected by amore basic magma, thus forcing out the evolved mag-ma (Unit II), which was followed by the more primitivemagma (Unit III) (Wright and Fiske, 1971; Natland, pers.comm., 1984).

By contrast, the upper unit is distinctive in its contentof xenocrysts of calcic plagioclase and (pseudomorphosed)olivine. Thus, although the magma had olivine, plagio-clase, and clinopyroxene as liquidus phases, the xeno-crysts may provide evidence for an earlier period of mix-ing with more primitive magma which lay on the oliv-

447

J. A. PEARCE, N. ROGERS, A. J. TINDLE, J. S. WATSON

100

10

o, p1000

30 100

Zr (ppm)

300100

o, p

2000 r

1000

200

o, p, c, m

B

O

30

400 r

100Zr (ppm)

300

o, c, m

100

i i i J I30

30,000 r

100Zr (ppm)

30040

30 100Zr (ppm)

300

10,000

3,000

o, p200

100

P, m

V V

30 100Zr (ppm)

30020

O

30 100Zr (ppm)

300

Figure 13. Trace element-Zr covariation diagrams from Hole 597C and Hole 601B basalts. A. Y-Zr. B. P-Zr. C. Ti-Zr. D. V-Zr. E. Sr-Zr. F. Ni-Zr.Theoretical fractionation vectors are for 50% crystallization of single phases, except for magnetite on Ti-Zr and V-Zr diagrams (5% crystalliza-tion) and for olivine on the Ni-Zr diagram (10% crystallization), o = olivine, p = plagioclase, c = clinopyroxene, m = magnetite. Symbols asfor Figure 11.

ine-plagioclase cotectic. In accordance with current ideason the dynamics of magma chambers (e.g., Sparks etal., 1980), it can be proposed that the xenocrysts wereintroduced into the superheated upper levels of the mag-ma chamber by upwelling plumes of hotter, more basic

magma. When rapid mixing of the two magmas tookplace, these xenocrysts would have remained suspendedin the resultant magma to act as nuclei for further crys-tallization once the temperature again fell to that of theliquidus.

448

GEOCHEMISTRY AND PETROGENESIS OF BASALTS, LEG 92

3000

S

\

1000

100

20

p, m

i i i I i i t i i I10

Y (ppm)100

Figure 14. Petrogenetic pathways for Hole 597C and Hole 60IB ba-salts, as illustrated by a Ni-Y diagram. The partial melting trend,annotated for 10, 30, and 50% melting, is given by line P, and theolivine ( + spinel) fractionation trend is shown by the dashed line,Ol. Fractionation vectors appropriate to subsequent crystallizationare also shown (cf. Fig. 13F).

On a regional scale, it is interesting that the lavas allclosely resemble tholeiites analyzed from other parts ofthe eastern Pacific, and from the ocean crust in general.This is particularly evident in the shapes of their trace-element patterns (Fig. 15), which show a good corre-spondence to the pattern (compiled by the authors) foraverage N-type MORB and thus provide further proofof the uniformity of the MORB reservoir.

100i-

ε 10

N-MORB

Rb Th K Ta Nb La Ce Sr Nd P Hf Zr Sm Ti Tb Y Yb

Figure 15. Primordial-mantle-normalized trace-element patterns for themost and least fractionated, "fresh" samples from Hole 597C (solidlines), compared with the pattern of average N-type MORB (dashedline).

GENESIS OF FERROBASALTS FROM HOLE 601B

The ferrobasalts sampled from Hole 601B representonly a tiny proportion of the rock recovered during Leg92, but do emphasize the importance of this rock typein the eastern Pacific. One of the most significant petro-genetic features of these lavas is the presence of xeno-crysts of calcic plagioclase and pseudomorphosed oliv-ine, which produce cumulation trends on the Sr-Zr andNi-Zr diagrams in Figure 13. It is not, however, clearfrom such limited data whether these were (1) scavengedfrom solid rock during magma injection, (2) scavengedfrom the margins of a magma chamber prior to erup-tion, or (3) introduced into the magma chamber duringan episode of mixing with more primitive magma.

Compared with other evolved MORB suites from theeastern Pacific, the ferrobasalts contain less SiO2 thanthose from the Galapagos Rift, and they are slightlymore evolved than those from the Siqueiros FractureZone (e.g., Byerly et al., 1976; Natland, 1980). There isinsufficient "field" evidence to enable the rocks to beplaced into a tectonic framework, or to evaluate modelsrelating ferrobasalts to rift propagation (e.g., Sinton etal., 1983).

GENESIS OF THE ALKALI BASALT FROMHOLE 602B

The basalt fragment from Hole 602B again consti-tutes a tiny proportion of the basalts recovered duringthe leg, but nevertheless emphasizes the importance ofthe rock type in the region. The genesis of an alkali ba-salt such as this, although once the subject of debate,can be simply attributed to the melting of incompatible-element-enriched mantle. The cause of such enrichmentis variously explained by enrichment of suboceanic litho-sphere resulting from (1) metasomatism by CO2-rich flu-ids and small-degree melts, (2) melting of enriched con-vecting upper mantle (mantle plumes), or (3) some formof dynamic melting. Clearly, no distinction between thesealternatives can be drawn on the basis of such a smallsample. It is worth noting, however, the contrast in en-richment pattern between the alkali basalt sampled andother alkali basalts from the eastern Pacific, as exempli-fied by the basalt from the Society Islands (Dostal et al.,1982) (Fig. 16). The enrichment in most elements iscomparable, but there is a marked difference in the nor-malized abundances of Y and Yb. This feature may re-flect a genetic difference between E-type MORB andwithin-plate oceanic alkali basalts, perhaps involvingthe role of garnet (which retains Y and Yb) in the en-richment or melting process.

SUMMARY

1. Basement recovery on DSDP Leg 92 consisted pri-marily of 48.5 m of core from the 91-m Hole 597C and5.4 m from the 26.4-m "pilot" Hole 597B, both of whichwere drilled into 28-Ma crust originally formed at thenow-extinct Galapagos Rise. Small quantities of corewere also recovered from Holes 597 and 597A (also onGalapagos Rise crust) and Holes 599, 599B, 601B, and602B (on crust formed at the East Pacific Rise).

449

J. A. PEARCE, N. ROGERS, A. J. TINDLE, J. S. WATSON

100

o 10E

I I I I I I I I I IRbTh K TaNbLa Ce Sr Nd P Hf Zr Sm Ti Tb Y Yb

Figure 16. Primordial-mantle-normalized trace-element pattern for thealkali basalt from Hole 601B (solid line), compared with that of analkali basalt from the Society Islands (SOC—dashed line).

2. "Conventional" petrologic classification diagramsrevealed that all except the basalts from Hole 602B, whichproved to be alkalic (Ne-normative), exhibited the iron-enrichment trends and hyp-normative characteristics ofabyssal tholeiitic basalts; the basalts from Hole 60IBadditionally had the high TiO2 and FeO values of ferro-basalts.

3. The "deep" hole, 597C, could be divided petro-graphically and geochemically into three lithologic units.The lowest unit (Unit III), at least 13 m thick, is themost basic (containing >8% MgO) and contains modalolivine. The middle unit (Unit II), about 20 m thick, isthe most evolved and is olivine free. The upper unit,about 48 m thick, is intermediate in composition andcontains both modal olivine and plagioclase glomero-crysts. The upper unit is slightly vesicular, exhibits arange of textures, and was probably a massive flow; thelower units are coarser and nonvesicular and may bemassive flows or sills. There is some evidence of a tran-sitional boundary (magma mixing?) between Units II andIII.

4. Alteration intensity varies from moderate (in mostcores) to very weak (in Units II and III of Hole 597C).The earliest alteration episode (deuteric?) is marked byblue smectite; the two later alteration episodes can bedivided into "relatively reducing" (dark green smectite,talc, chlorite, and rare sulfides and native copper) and"relatively oxidizing" (brown smectite, calcite, low-tem-perature zeolites, and iron oxyhydroxides). Except dur-ing replacement of olivine in Unit III of Hole 597, themain chemical effect of alteration is the increase in LOIand in potassium and related elements.

5. Petrogenetic studies show that all basalts except thealkali basalts from Hole 602B were derived from sourcesof similar composition. In overall terms, the commonsource was the incompatible-element-depleted reservoirof the convecting upper mantle that is considered thesource for all N-type MORB. Other variations can beascribed primarily to magma-chamber processes. Least-squares modeling of major elements and graphic model-ing of major and trace elements give consistent results in

showing that Unit II of Hole 597C can be derived fromUnit III of Hole 597C by about 60% closed-system crys-tallization of a bulk assemblage of olivine (about 10%),plagioclase (about 54%), and clinopyroxene (about 36%),and that the ferrobasalts can be derived from Unit III ofHole 597C by about 80% closed-system crystallizationof a bulk assemblage of olivine (about 9%), plagioclase(51%), and clinopyroxene (40%). They also indicate thatclinopyroxene was not a liquidus phase during crystalli-zation of Unit III of Hole 597C. The presence of some-times partially resorbed or rimmed calcic xenocrysts inboth Unit I of Hole 597 and the ferrobasalts providespossible evidence of magma mixing, but the extent ofthis mixing is difficult to quantify.

6. The three rock types recovered (primarily MORBtholeiites with rare ferrobasalts and alkali basalts) are alltypical of rocks previously recovered by dredging anddrilling in the eastern Pacific, although the range of tho-leiite compositions in the 91-m hole is surprisingly large.

ACKNOWLEDGMENTS

J.A.P. would like to thank all participants of Leg 92 for a stimulat-ing experience and for discussions that helped in the preparation ofthis chapter. We also thank Jim Natland for a constructive review ofthe manuscript.

REFERENCES

Anderson, A. T., and Greenland, L. P., 1969. Phosphorus fractiona-tion diagram as a quantitative indicator of crystallization differen-tiation of basaltic liquid. Geochim. Cosmochim. Acta, 33:493-505.

Basaltic Volcanism Study Project, 1981. Basaltic Volcanism on theTerrestrial Planets: New York (Pergamon Press).

Bass, M. N., 1976. Secondary minerals in oceanic basalt, with specialreference to Leg 34, Deep Sea Drilling Project. In Yeats, R. S.,Hart, S. R., et al., Init. Repts. DSDP, 34: Washington (U.S. Govt.Printing Office), 393-432.

Byerly, G. R., Melson, W. G., and Vogt, P. R., 1976. Rhyodacites,andesites, ferro-basalts and ocean tholeiites from the Galapagosspreading centre. Earth Planet. Sci. Lett., 30:215-221.

Clague, D. A., Frey, F. A., Thompson, G., and Rindge, S., 1981. Mi-nor and trace element geochemistry of volcanic rocks dredged fromthe Galapagos spreading center: role of crystal fractionation andmantle heterogeneity. J. Geophys. Res., 86:9469-9482.

Deer, W. A., Howie, R. A., and Zussman, J., 1966. An introductionto the rock-forming minerals: Harlow, England (Longman).

Dostal, J., Dupuy, C , and Liotard, J. M., 1982. Geochemistry and ori-gin of basaltic lavas from Society Islands, French Polynesia (southcentral Pacific Ocean). Bull. VolcanoL, 45:51-61.

Dungan, M. A., and Rhodes, J. M., 1978. Residual glasses and meltinclusions in basalts from DSDP Legs 45 and 46: evidence formagma mixing. Contrib. Mineral. Petrol., 67:417-431.

Le Bas, M. J., 1962. The role of aluminium in igneous clinopyroxeneswith relation to their parentage. Am. J. Sci., 260:267-288.

Mallet, J.-L., 1978. Reconstitution des concentrations minéralogiquesà partir des concentrations géochimiques. Sci. Terre, 12:35-64.

Mattey, D. P., and Muir, I. D., 1980. Geochemistry and mineralogy ofbasalts from the Galapagos spreading center, Deep Sea DrillingProject Leg 54. In Rosendahl, B. R., Hekinian, R., et al., Init.Repts. DSDP, 54: Washington (U.S. Govt. Printing Office), 755-771.

Mazzullo, L. J., and Bence, A. E., 1976. Abyssal tholeiites from DSDPLeg 34: the Nazca Plate. J. Geophys. Res., 81:4327-4351.

Muir, I. D., and Mattey, D. A., 1982. Pyroxene fractionation in fer-robasalts from the Galapagos spreading centre. Min. Mag., 45:193-200.

Nakamura, N., 1974. Determination of REE, Ba, Fe, Mg, Na and Kin carbonaceous and ordinary meteorites. Geochim. Cosmochim.Acta, 38:757-775.

Natland, J. H., 1980. Effect of axial magma chambers beneath spread-ing centers on the compositions of basaltic rocks. In Rosendahl, B.

450

GEOCHEMISTRY AND PETROGENESIS OF BASALTS, LEG 92

R., Hekinian, R., et al., Init. Repts. DSDP, 54: Washington (U.S.Govt. Printing Office), 833-850.

Natland, J. H., Adamson, A. C , Laverne, C , Melson, W. G., andO'Hearn, T., 1983. A compositionally nearly steady-state magmachamber at the Costa Rica Rift: evidence from basalt glass andmineral data, Deep Sea Drilling Project Sites 501, 504, and 505. InCann, J. R., Langseth, M. G., Honnorez, J., Von Herzen, R. P.,White, S. M., et al., Init. Repts. DSDP, 69: Washington (U.S.Govt. Printing Office), 811-858.

O'Hara, M. J., 1977. Geochemical evolution during fractional crystal-lization of a periodically refilled magma chamber. Nature, 266:503-507.

Osborn, E. R, and Tait, D. B., 1952. The system diopside-forsterite-anorthite. Am. J. Sci. (Bowen Volume), 413-433.

Ottonello, G., Ernst, W. G., and Joron, J. L., 1984. Rare earth and 3dtransition element geochemistry of peridotitic rocks: I. Periodo-tites from the western Alps. J. Petrol., 25:343-372.

Paul, D. K., Potts, P. J., Gibson, I. L., and Harris, P. G., 1975. Rareearth abundances in Indian kimberlite. Earth Planet. Sci. Lett.,25:151-158.

Pearce, J. A., 1982. Trace element characteristics of lavas from de-structive plate boundaries. In Thorpe, R. S. (Ed.), Andesites: NewYork (John Wiley and Sons), pp. 525-548.

Pearce, J. A., and Norry, M. J., 1979. Petrogenetic implications of Ti,Zr, Y and Nb variations in volcanic rocks. Contrib. Mineral. Pet-rol., 69:33-47.

Perfit, M. R., Fornari, D. J., Malahoff, A., and Embley, R. W, 1983.Geochemical studies of abyssal lavas recovered by DSRV Alvin fromEastern Galapagos Rift, Inca Transform, and Ecuador Rift. 3.Trace element abundances and petrogenesis. J. Geophys. Res., 88:10,551-10,572.

Philpotts, J. A., and Schnetzler, C. C , 1970. Phenocryst-matrix par-tition coefficients for K, Rb, Sr and Ba with application to anor-thosite and basalt genesis. Geochim. Cosmochim. Ada, 34:331-340.

Potts, P. J., Thorpe, O. W, and Watson, J. S., 1981. Determinationof rare-earth element abundances in 29 international rock stan-dards by instrumental neutron activation analysis: a critical ap-praisal of calibration errors. Chem. Geol., 34:331-352.

Potts, P. J., Webb, P. C , and Watson, J. S., 1984. Energy-dispersiveX-ray fluorescence analysis of silicate rocks for major and trace el-ements. X-Ray Spectrom., 13:2-15.

Rea, D.K., 1981. Tectonics of the Nazca-Pacific divergent plate bound-ary. Mem. Geol. Soc. Am., 154:27-62.

Scheidegger, K. F., and Corliss, J. B., 1981. Petrogenesis and second-ary alteration of upper Layer 2 basalts of the Nazca Plate. Mem.Geol. Soc. Am., 154:77-107.

Shido, F., Miyashiro, A., and Ewing, ML, 1971. Crystallization of abys-sal tholeiites. Contrib. Mineral. Petrol, 31:251-266.

Sinton, J. M., Wilson, D. S., Christie, D. ML, Hey, R. N., and Dela-ney, J. R., 1983. Petrologic consequences of rift propagation onoceanic spreading ridges. Earth Planet. Sci. Lett., 62:193-207.

Sparks, R. S. J., Meyer, P., and Sigurdsson, H., 1980. Density varia-tion amongst mid-ocean ridge basalts: implications for magmamixing and the scarcity of primitive lavas. Earth Planet. Sci. Lett.,46:419-430.

Stakes, D. S., Shervais, J. W, and Hopson, C. A., 1984. The volcano-tectonic cycle of the FAMOUS and AMAR valleys, Mid-AtlanticRidge (36°47'N): evidence from basalt glass and phenocryst com-positional variations for a steady-state magma chamber beneaththe valley midsections. J. Geophys. Res., 89:6995-7028.

Tarney, J., Wood, D. A., Saunders, A. D., Cann, J. R., and Varet, J.,1980. Nature of mantle heterogeneity in the North Atlantic: evi-dence from deep sea drilling. Phil. Trans. R. Soc. London, A297:179-202.

Wood, D. A., 1979. A variably veined suboceanic upper mantle—ge-netic significance for mid-ocean basalts from geochemical evidence.Geology, 7:499-503.

Wright, T. L., 1974. Presentation and interpretation of chemical datafor igneous rocks. Contrib. Mineral. Petrol., 48:233-248.

Wright, T. L., and Doherty, P. C , 1970. A linear programming andleast squares computer method for solving petrologic mixing prob-lems. Geol. Soc. Am. Bull., 81:1995-2008.

Wright, T. L., and Fiske, R. S., 1971. Origin of the differentiated andhybrid lavas of Kilauea Volcano, Hawaii. J. Petrol., 12:1-65.

Yeats, R. S., Hart, S. R., et al., 1976. Init. Repts. DSDP, 34: Wash-ington (U.S. Govt. Printing Office).

Yoder, H. S., and Tilley, C. E., 1962. Origin of basalt magmas: an ex-perimental study of natural and synthetic rock systems. /. Petrol.,3:342-532.

Date of Initial Receipt: 11 October 1984Date of Acceptance: 15 May 1985

APPENDIX AGeochemical Data on Some USGS Standard Rocks (BCR1, AGV1, GSP1, and a DSDP

Leg 92 Standard, ILB1)

Std.

SiO2TiO2A12O3F e2°3MnOMgOCaONa2OK 2 O

P2O5

ZrYNbRbSrCrNiVCuZn

BCR1Observed

Mean (s.d.)

54.84 (0.38)2.24 (0.02)

13.59(0.18)13.31 (0.09)0.19(0.01)3.56 (0.12)7.05 (0.06)3.36 (0.30)1.66(0.04)0.36 (0.02)

184(1.9)37 (0.4)

11.8(0.2)49 (0.4)

322 (0.4)12(3)14(1.1)

387 (29)19(1.5)

121 (2.2)

BCR1Recommended

Mean (s.d.)

54.35 (0.51)2.22(0.10)

13.63 (0.25)13.41 (0.31)0.18 (0.01)3.45 (0.17)6.95 (0.15)3.27(0.11)1.69(0.08)0.37 (0.02)

191 (5)39(7)

14.0 (3)47(1)

330 (5)16(4)13(4)

404(40)19(4)

129 (1)

AGV1Observed

Mean (s.d.)

59.69(0.11)1.07 (0.01)

17.11 (0.14)7.03 (0.01)0.10(0.00)1.57 (0.12)5.09 (0.01)4.19(0.24)2.94 (0.01)0.51 (0.03)

238 (0)21(0)

14.6 (0)70(0)0(0)

18(0)123 (0)64(0)88(0)20(0)

AGV1Recommended

Mean (s.d.)

59.25 (0.58)1.06(0.05)

17.15 (0.34)6.77(0.19)0.10(0.01)1.53 (0.10)4.94(0.14)4.25 (0.12)2.90(0.10)0.44 (0.03)

225 (18)21(6)

15.0 (3)67(1)

662 (9)12(3)17(4)

123 (12)60(6)88(2)

GSP1Observed

Mean (s.d.)

67.54 (0.09)0.67 (0.00)

15.24(0.10)4.47 (0.01)0.04 (0.00)0.97 (0.08)2.02 (0.00)2.63 (0.17)5.61 (0.02)0.28 (0.01)

498 (2.4)29 (0.5)

14.4 (0.5)151 (2.1)233 (1.5)

11(3)9 (2.6)

62(9)34 (1.7)

103 (2.4)

GSP1Recommended

Mean (s.d.)

67.37 (0.49)0.66 (0.04)

15.16(0.28)4.30(0.13)0.04 (0.01)0.99 (0.08)2.04(0.10)2.80 (0.09)5.51 (0.14)0.28 (0.02)

530 (70)29(6)

26.0 (4)254 (2)234 (3)

13(3)10(3)53(7)34(5)

103 (9)

ILB1Observed

Mean (s.d.)

49.77 (0.26)1.32(0.01)

14.60(0.19)11.62(0.05)0.16(0.00)7.40(0.14)

11.78(0.06)2.39 (0.30)0.09 (0.01)0.13 (0.03)

79 (0.7)31 (0.4)

4.5 (0.3)3 (0.3)

87 (0.8)215 (10)

78 (9.0)346 (16)121 (4.2)

89 (0.6)

Note: Number of analyses: BCR1 (majors = 12, traces = 3); AGV1 (majors = 37, traces = 1); GSP1 (majors = 6, traces =19); ILB1 (majors = 7, traces = 5). Zero signifies no data available. Total iron expressed as Fe 2θ3 s.d. = standard devi-ation.

451

J. A. PEARCE, N. ROGERS, A. J. TINDLE, J. S. WATSON

APPENDIX BMajor-Element Data, XRF Trace-Element Data, and CIPW Norms for DSDP Leg 92 Basalts

HoleCore-Sec.Position, cm

5978.CC

597B2-17

597B2-136

597B2-155

597B2-166

597B2-2138

597B2-325

597B2-363

597B2-3110

597B3-116

597B3-175

597B3-1141

597B3-227

597B3-251

597B3-256

597C3-13

597C3-152

597C3-220

597C3-2117

597C3-39

SiO2TiO 2A12O3Fe2O3MnOMgOCaONa2OK2OP 2 O 5LOI

Total

ZrYNbRbSrCrNiVCuZn

QOrAbAnNcDiHyOlMl11Ap

49.341.04

15.1811.140.187.54

12.992.540.320.150.51

100.93

6429

2.77.0

84363

94321131107

0.001.89

21.4929.08

0.0028.10

4.0310.132.421.980.35

50.171.00

14.4710.260.158.85

12.362.210.150.100.73

100.45

6025

3.63.1

7935117128412982

0.001.89

18.7029.13

0.0025.5817.013.172.231.900.23

48.611.06

15.479.400.147.75

13.712.240.290.112.47

101.25

6727

3.55.6

9427313029813082

0.001.71

18.9531.31

0.0029.21

4.058.432.042.010.25

49.741.08

14.6910.630.168.12

12.732.640.220.120.52

100.65

67274.14.7

8533412733212381

0.001.30

22.3427.59

0.0028.31

5.679.382.312.050.28

49.971.11

14.8310.200.167.86

12.512.280.220.100.83

100.07

7128

4.25.5

8632414731513284

0.001.30

19.2929.59

0.0025.8915.742.002.222.110.23

50.001.23

15.0710.240.157.44

12 262.230.100.140.40

99.26

7829

4.63.0

9025215033312982

0.790.59

18.8730.82

0.0023.6918.340.002.232.340.32

50.441.28

15.219.830.157.48

12.272.520.220.100.50

100.00

8332

3.99.3

9926114935214191

0.001.30

21.3229.55

0.0024.9315.44

1.312.142.430.23

50.111.33

15.2910.230.167.12

12.662.200.290.130.52

100.04

8533

5.012.3

106267

7234519898

0.421.71

18.6231.000.00

25.2416.620.002.222.530.30

49.371.27

13.7213.180.166.37

11.292.391.210.140.99

100.09

8431

4.037.694

22365

30813091

0.007.15

20.2223.14

0.0026.39

8.736.732.872.410.32

SO.081.35

14.8111.290.157.44

11.972.420.100.120.15

99.88

8331

4.33.0

9219180

36113089

0.000.59

20.4829.26

0.0023.9718.800.372.462.560.28

48.721.24

14.0412.890.156.19

11.522.261.120.140.72

98.99

8032

4.527.989

18854

3518671

0.006.62

19.1224.86

0.0025.92

9.146.032.802.360.32

49.741.24

15.0910.980.156.97

12.362.590.560.120.64

100.44

7933

5.118.6

103262

9031811788

0.003.31

21.9227.90

0.0026.70

7.106.922.392.360.28

50.101.22

14.6010.520.157.20

12.411.970.110.160.32

98.76

7631

4.12.5

9227815134212489

2.450.65

16.6730.680.00

24.3617.770.002.292.320.37

50.141.20

15.2111.270.177.80

12.022.290.090.100.61

100.90

7127

3.92.6

82253

9432412076

0.000.53

19.3830.960.00

22.8619.58

1.052.452.280.23

51.101.20

15.069.970.176.84

12.722.780.230.150.23

100.45

7633

2.08.5

92261

7128410780

0.001.36

23.5227.94

0.0027.8912.10

1.762.172.280.35

48.530.95

15.7011.360.196.54

13.161.870.310.090.55

99.25

5927

4.13.9

83395

9434112591

0.001.83

15.8233.54

0.0025.6114.88

1.562.471.800.21

49.681.13

14.4410.190.168.05

12.242.020.150.130.69

98.88

7026

3.74.3

8331211327712474

0.740.89

17.0929.90

0.0024.3619.680.002.222.150.30

49.861.20

15.2610.410.157.81

12.402.130.100.160.47

99.95

7328

4.73.4

88241

8832211278

0.230.59

18.0231.790.00

23.3419.710.002.262.280.37

49.401.20

14.4012.690.156.75

11.772.231.020.120.52

100.25

79314.7

24.882

22087

32312174

0.006.03

18.8726.28

0.0025.8010.086.282.762.280.28

50.171.32

15.0110.670.177.12

12.382.780.480.160.60

100.86

84324.1

13.4105234

8031315588

0.002.84

23.5227.07

0.0027.196,247.302.322.510,37

Note: Total iron has beeof 2 cm diameter.

xpressed as Fe2C>3 in the oxide columns. Norm calculations assume Fe /(Fe + Fe ) = 0.85. Nondetermined Cr and V are given as 0. "Position" row gives upper sampling position of minicore

APPENDIX CReconnaissance Electron Microprobe Analyses (wt.%) of Pyroxenes from DSDP Leg 92 Basalts

HoleCore-Sec.Position, cmPetrology3

SiO2TiO 2A12O3FeOMnOMgOCaONa2OCr2O3

Total

CaMgFe

597B2-2138V

51.070.933.668.980.20

15.4220.140.250.26

100.91

41.444.114.4

597B2-2138V

51.200.753.587.970.24

16.3319.150.260.21

99.69

39.847.212.9

597B3-116P

50.930.973.629.050.20

15.3719.290.260.12

99.81

40.444.814.8

597B3-116P

51.750.772.818.810.23

15.9319.380.260.09

100.03

40.045.814.2

597B3-116P

52.670.632.219.060.25

17.5317.700.200.03

100.28

36.650.413.0

597B3-116P

52.250.412.505.390.16

16.8621.020.250.73

99.57

43.248.2

8.6

597B3-116P

49.671.403.25

15.170.39

12.8017.430.320.00

100.43

37.037.825.2

597B3-116

50.541.012.52

16.210.40

14.6614.750.320.00

100.41

30.842.724.5

597B3-227

50.570.982.75

11.960.30

13.9318.940.310.02

99.76

39.740.719.6

597B3-227

50.710.911.83

19.250.52

14.4911.760.220.00

99.69

25.043.032.0

597B3-227

48.780.871.44

23.030.638.53

15.950.250.00

99.48

34.825.939.3

597B3-227

52.470.602.757.380.20

16.5420.260.260.15

100.61

41.346.911.8

597B3-227

50.910.812.41

13.150.31

14.4217.500.290.00

99.80

36.641.921.5

597B3-227

50.460.822.26

16.090.42

14.9614.080.320.00

99.41

29.743.916.5

597B3-256I

50.161.133.81

12.610.31

13.8018.290.410.05

100.57

38.640.620.8

597B3-256I

50.780.814.356.910.20

16.0819.710.310.28

99.43

41.547.111.4

597C3-22"P

50.740.843.338.670.20

15.2020.040.260.27

99.55

41.744.114.1

597C3-220P

50.820.974.068.530.22

14.6720.00

0.300.29

99.86

42.543.414.1

597C3-220P

51.130.773.976.680.17

15.5520.950.310.24

99.77

43.845.310.9

597C3-2117

I

51.710.753.259.470.25

15.6019.400.250.20

100.88

40.044.815.2

597C3-2117

I

51.350.803.049.340.22

15.3619.800.260.19

100.36

40.944.115.2

597C3-2117P

53.360.582.088.600.25

17.8117.960.180.09

100.91

36.350.113.6

Note: For multiple analyses of a single sample, the analysis numbers (referred to in the section on least-squares modeling, but not given here) begin with 1 for the first analysis of any sample and progress to the right.a I = intersertal; V = variolitic; P = poikilitic; p = microporphyritic.

APPENDIX DReconnaissance Electron Microprobe Analyses (wt.%) of Feldspars from DSDP Leg 92 Basalts

HoleCore-Sec.Position, cmPetrology3

SiO2A12O3FeOCaONa2OK2O

597B2-155P

47.9933.240.52

17.491.730.01

597B2-155P

47.6432.790.52

17.081.930.02

597B2-155P

47.9932.770.56

17.001.930.01

597B2-155P

51.3329.240.86

14.683.250.01

597B2-155P

51.2029.750.80

14.433.280.02

597B2-155

50.5030.37

0.6914.892.990.03

597B2-2138

47.8833.360.55

17.391.780.01

597B2-2138

48.0333.280.49

17.371.790.01

597B2-2138

46.6733.950.37

18.071.340.00

597B2-2138

51.6829.290.83

13.893.580.03

597B2-2138

51.5329.74

0.8214.123.460.01

597B2-2138P

52.4029.71

0.7713.793.760.02

597B3-116

53.4028.86

0.7613.084.050.02

597B3-116

52.8628.94

0.8313.004.150.02

597B3-116

51.1730.250.65

14.493.280.01

597B3-227

54.6127.740.87

11.755.130.05

597B3-227

54.5427.79

0.9311.754.980.05

597B3-227

54.3928.35

0.9112.184.730.04

597B3-256

52.4729.64

0.6313.673.840.01

597B3-256

52.3729.51

0.6013.663.800.02

597B3-256

51.8130.09

0.5514.043.730.02

Total 100.98 99.98 100.26 99.37 99.48 99.47 100.97 100.97 100.40 99.30 99.68 100.45 100.17 99.80 99.85 100.15 100.04 100.60 100.26 99.96 100.24

An 84.7 82.9 82.9 71.4 70.8 73.2 84.3 84.2 88.2 68.1 69.2 66.9 64.0 63.3 70.1 55.7 56.4 58.6 66.3 66.5 67.4

Note: For multiple analyses of a single sample, the analysis numbers (referred to in the section on least-squares modeling, but not given here) begin with 1 for the first analysis of any sample and progress to the right,a P = phenocryst/glomerocryst.

452

GEOCHEMISTRY AND PETROGENESIS OF BASALTS, LEG 92

Appendix B (continued).

597C3-358

597C4-192

597C4-1126

597C4-229

597C4-336

597C4-374

597C4-3107

597C4-411

597C4-4127

597C4-576

597C4-5120

597C4-6

8

597C4-649

597C4-665

597C5-132

597C5-1133

597C5-212

597C5-224

597C5-255

597C6-169

597C6-1128

597C6-244

49.951.35

14.9710.740.167.09

12.222.380.360.120.46

99.80

8733

4.413.5

104239

8935616194

0.002.13

20.1429.11

0.0025.1015.611.172.342.560.28

48.741.32

13.4412.020.177.12

12.732.030.610.122.36

100.66

8631

5.016.194

20360

34416586