Tamaki, K., Suyehiro, K., Allan, J., McWilliams, M., et al., 1992Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 127/128, Pt. 2

53. LOW-PRESSURE MELTING RELATIONS OF A BASALT FROM HOLE 797CIN THE YAMATO BASIN OF THE JAPAN SEA1

P. Thy2

ABSTRACT

One-atmosphere melting experiments, controlled to approximately the fayalite-magnetite-quartz oxygen buffer, performedon a basalt from Hole 797C crystallized olivine and plagioclase nearly simultaneously from about 1235°C and augite from about1175°C. The liquid compositions indicate systematic trends of increasing FeO and TiO2 and decreasing A12O3 with decreasingMgO. Experimental olivine compositions vary from Fo90 to Fo78, plagioclase from An79 to An67, and augite from En49 toEn46. The KD value for the Fe

2+ and Mg distribution between olivine and liquid is 0.31. The KD value for the distribution ofpetotai an (j fyjg between augite and liquid averages 0.24. These KD values suggest experimental equilibrium. The KD values forNa and Ca distribution between plagioclase and liquid range between 0.55 and 0.99 and are dependent on crystallizationtemperature. Projected on pseudoternary basaltic phase diagrams, the liquid line of descent moves toward increasing quartznormative compositions, revealing a typical tholeiitic crystallization trend with marked Fe and Ti enrichments. Such enrichmentsare a reflection of the dominance of plagioclase in the crystallizing assemblage. The experimental results can explain the markedFe- and Ti-enrichment trends observed for the sills of the lower part of Hole 797C, but have no direct bearing on the origin ofthe relatively evolved high-Al basalts of Hole 794C.

INTRODUCTION

Basement rocks composed of thick basaltic sill complexes in-truded into soft sediments were drilled at two sites during OceanDrilling Program Legs 127 and 128 in the Yamato Basin of the JapanSea. The geochemical and petrogenetic implications of these sillcomplexes have important bearings on tectonic models for the open-ing and evolution of backarc spreading in the Japan Sea. At Site 794in the Yamato Basin, relatively primitive basaltic to evolved high-Albasaltic sills were drilled (Tamaki et al., 1990). At Site 797, also inthe Yamato Basin, a sill complex composed of an upper suite ofbasaltic sills and a lower suite of Ti-enriched ferrobasaltic sills weredrilled. Accompanying chapters discuss the groundmass mineralogyand the rock chemistry of the Japan Sea basalts (Thy, this volume).The presence of two major suites of relatively evolved basaltic rocks(high-Al basalts and ferrobasalts) suggests complex petrogenetic andtectonic regimes during the opening of the Japan Sea. The presenceof ferrobasalts suggests low pressure, tholeiitic fractionation trends,while high-Al basalts point toward high pressure, calc-alkalic frac-tionation trends (Grove and Baker, 1984).

To evaluate these possible petrogenetic models, one-atmospheremelting experiments were conducted on a relatively primitive samplefrom the upper suite of Hole 797C. The selected sample (127-797C-12R-2,74-79 cm; Table 1) was taken from the center of what appearsto be a massive sill and was relatively unaffected by secondaryalteration and hydration. This sill contains granular groundmassolivine (Fo87_89) and subophitic to granular plagioclase (An72_78). Afew plagioclase phenocrysts (An78) are present in the sample. Augiteis the third phase to have crystallized (Mg/(Mg + Fe) = 0.65-0.72)and occurs as skeletal grains growing on the plagioclase. Titanomag-netite appears after augite crystallization as a fine-grained skeletalphase growing on the augite. This chapter presents the experimentalresults and briefly addresses the implications. A subsequent chapterdiscusses in more details the implications of the experimental resultsfor the evolution of the basalts drilled during Legs 127 and 128 (Thy,this volume).

Tamaki, K., Suyehiro, K., Allan, J., McWilliams, M , et al., 1992. Proc. ODP, Sci.Results, 127/128, Pt. 2: College Station, TX (Ocean Drilling Program).

2 NASA, Johnson Space Center, SN2, Houston, TX 77058, U.S.A. (Present address:Department of Geology, University of Botswana, Private Bag 0022, Gaborone, Botswana).

EXPERIMENTAL PROCEDURES

All experiments were performed at atmospheric pressure in avertical quench furnace (Williams and Mullins, 1981) followingprocedures described by Lofgren (1983). Oxygen fugacity was con-trolled by a constant CO2 and CO gas mixture calibrated at about1100°C to the fayalite-magnetite-quartz (FMQ) buffer and measuredby a ZrO2 ceramic oxygen electrolyte cell. Temperature was moni-tored by Pt/90Pt-10Rh thermocouples calibrated to the melting pointof gold and diopside, according to the 1968 International PracticalTemperature Scale. About 100 mg of rock powder was pressed intopellets and sintered to 0.008-in. Pt-wire loops (Donaldson et al., 1975)and inserted into the furnace directly at the experimental temperature.The experiments lasted from 24 hr near the liquidus to about 190 hrcloser to the solidus (Table 2).

Phases were analyzed with the electron microprobe using anacceleration voltage of 15 kV and a beam current of approximately25 nA. A scanning-mode counting over an area of approximately200 µm2 was used for analyzing glasses to minimize the loss of Na.Counting times were 30-70 s, except for Na, which in glasses wascounted first and for 10 s. An evaluation of the accuracy and precisionof glass analyses using the VG-2 glass standard (Jarosewich et al.,1979) is given in Table 3.

EXPERIMENTAL RESULTS

Details of the 1 arm runs are presented in Table 2 and are summarizedin Figure 1. The selected sample crystallized olivine and plagioclasealmost simultaneously from 1235°±6°C and augite at 1175°±5°C. Thephase proportions and the amounts of liquid remaining were estimatedby least squares, linear regression calculations using the analytical datapresented in Tables 4—7. The calculations weighted all oxides by 1.0,except SiO2 and A12O3, which were weighted 0.4 and 0.5, respectively.The result of these calculations (Table 2) reflects a systematic variationin the liquid remaining, temperature, phase proportions (Fig. 1), andMg/(Mg + Fe2+) of the liquid (Fig. 2). The liquid remaining reaches 27wt%, below which the liquid cannot be analyzed with the broad-beamtechnique because of the restricted size of the analyzable areas.

For each experimental charge, 4-6 points were analyzed for eachphase, and average analyses are given in Tables 4-7. Attempts weremade to analyze only equilibrium phases. The standard deviations ofreplicate analyses of glass range between 0.12 and 0.25 for SiO2,0.02

861

P. THY

Table 1. Chemical composition of starting material(127-797C-12R-2, 74-79 cm).

SiO2TiO2A12O3FeO*MnOMgOCaONa2OK2OP2O5Total

Mg/(Mg + Fe)

1

49.311.12

17.967.560.169.59

11.302.820.070.11

100.00

0.693

2

49.031.06

17.987.820.159.79

11.023.030.11

100.00

0.691

Analyses are normalized to 100%. FeO* - Total iron calculatedas FeO. Mg/(Mg + Fe) calculated with all Fe as Fe +.

1. Shipboard analysis as modified by S. Yamashita (pers.comm., 1991).

2. XRF analysis by S. Grundvig, University of Aarhus.

Table 2. Experimental conditions and results.

Runno.

237215216217218219220221222223

Temperature(°C)

1240123112211212120011901180117111601149

Time(hr)

422424687072

119117186189

Run products8

glgl, ol, pigl, ol, pigl, ol, pigl, ol, pigl, ol, pigl, ol, pigl, ol, pi, auggl, ol, pi, auggl, ol, pi, aug

gl

938781726959553727

Phaseproportionsb

ol

35589

11131516

pi

48

14202230324449

aug

058

Losses (wt%)

FeO

0.010.220.500.210.380.190.020.000.00

Na2O

0.010.000.000.000.000.000.080.080.08

agl = glass; ol = olivine; pi = plagioclase; aug = augite.bPhase proportions (wt%) estimated from least-squares linear regressions using the analytical datain Tables 4-7. All oxides were weighted by 1.00, except SiO2 and A12O3, which were weighted 0.40and 0.50, respectively. The Na2O and FeO losses are based on the results of the mass balancecalculations and the original sample composition (Table 1). Most calculated losses are insignificant,except for FeO, runs 216-220.

Table 3. Precision of microprobe analyses, Juan deFuca ridge basalt glass VG-2.

SiO2TiO2A12O3FeOMnOMgOCaONa2OK2OP2O5Total

1

50.341.85

14.0311.660.216.81

11.252.590.200.23

98.85

S.D.

0.150.030.180.070.010.100.180.070.010.02

2

50.811.85

14.0611.840.226.71

11.122.620.190.20

99.49

1. Average and one standard deviation (S.D.) of 68 analyses.2. Accepted analysis of VG-2 (Jarosewich et al., 1979).

and 0.06 for TiO2, 0.05 and 0.46 for A12O3, 0.04 and 0.44 for FeO,0.01 and 0.04 for MnO, 0.05 and 0.38 for MgO, 0.05 and 0.3 for CaO,0.04 and 0.10 for Na20,0.01 and 0.03 for K2O, and 0.01 and 0.03 forP2O5. The high values were found for the low-temperature experi-ments and result from the absence of perfect equilibrium, the finegrain size, and limitations of the analytical procedure. The lowerrange of these values is reasonably within or below the analytical

40 60% Liquid

80 100

Figure 1. Experimental phase proportions vs. percentage of liquid. Phaseproportions and % liquid have been estimated by least squares, linear regres-sion, as described in the text.

precision of the electron microprobe analyses (Table 3). Elementallosses during the experiments were evaluated by mass-balancing thecompositions of the constituting phases and comparing with theoriginal composition (Table 1). The calculated Fe loss to the Pt wireis 3%-6% of the FeO of the original sample (Table 2), but may bebelow the significance for low-temperature experiments. The Na2Olosses are below significance for all experiments.

MINERAL-LIQUID EQUILIBRIA

In the experimental glasses, FeO, TiO2, K2O, and P2O5 all increase,while A12O3 decreases with decreasing MgO (Table 4; Fig. 3). SiO2,Na2O, and CaO show relatively little variations, except for a slightdecrease in CaO after the start of augite crystallization. These vari-ations reflect a marked Fe-enrichment trend on the AFM diagram(Fig. 4). The Fe of the glasses has been redistributed between Fe2O3and FeO, according to the algorithm proposed by Kilinc et al. (1983)and the FMQ oxygen buffer. These calculated Fe2O3 and FeO values,if not otherwise stated, are used throughout this paper.

Equilibrium olivine ranges in composition from Fo90 to Fo78(Table 5), showing correlation with temperature (Fig. 5). The KDvalues for the Fe2+ and Mg distribution between olivine and liquid(Fig. 6) (KDFe/Mg (ol/liq) = [X

Fe0 (ol) ^ ° (liq)]/[XMgO (ol) X**0

(liq)]) average 0.31 (range, 0.29-0.32).Equilibrium plagioclase ranges in composition from An79 to An67

(Table 6) and shows a dependence on temperature (Fig. 7 A). Non-equilibrium plagioclase grains are present in most of the experiments,and are reflected in the compositional variation of individual micro-probe analyses and the scatter observed in Figure 7. The KDNa/Ca(pl/liq) = [X^° (pi) X 0 0 (liq)]/[XCa0 (pl)XNa2° (liq)] ranges between0.55 and 0.99 and is strongly dependent on crystallization tempera-tures (Fig. 7B). The compositions of coexisting olivine and plagio-clase show a systematic covariance (Fig. 8).

The Ti content of equilibrium augite is negatively correlated withMg/(Mg + Fe), while Cr is positively correlated (Fig. 9). Total Alcontent (and Aliv and Alvi) shows little correlation with Mg/(Mg + Fe),and the Ti/Al ratios are systematically low (1/2-1/5), reflecting therelatively low TiO2 content of the glasses. The Fe

3+ content of theaugites, estimated from balancing charge deficiencies vs. chargeexcesses (Papike et al., 1974), suggests the systematic presence ofFe3+ in the augites. However, the amount is uncertain because of therelatively few augites analyzed. Because of this, evaluation of the Fe2+

862

LOW-PRESSURE MELTING RELATIONS IN BASALT

Table 4. Chemical composition of experimental glasses.

SiO2TiO2A12O3FeO*MnOMgOCaONa2O

P 2 O 5Cr2O3Total

Mg/(Mg + Fe2+)

SiO2TiO2A12O3FeOMnOMgOCaONa2OK,OP 2 O 5Cr2O3Total

Mg/(Mg + Fe2+)

Run 237(4)a

49.040.99

18.637.450.119.61

11.202.570.080.100.05

99.83

0.724

0.160.020.080.090.020.050.030.040.010.010.02

Run 219(4)

51.711.68

15.859.270.197.93

11.473.160.120.180.06

101.62

0.638

0.160.060.510.190.030.210.140.050.010.020.04

Run 215(9)

50.111.02

18.197.510.158.95

11.322.800.080.140.03

100.30

0.709

0.220.030.520.150.030.380.080.060.010.020.01

Run 220(5)

51.361.97

14.998.920.177.18

11.692.970.140.200.04

99.63

0.641

0.720.030.520.440.040.190.200.100.010.030.00

Run 216(10)

51.071.21

17.737.510.148.53

11.432.980.090.140.05

100.88

0.699

0.180.040.460.130.030.210.090.060.010.010.01

Run 221(5)

51.461.98

14.459.310.177.16

11.523.000.150.210.04

99.45

0.614

0.250.040.050.170.010.060.060.060.020.010.01

Run 217(6)

51.461.36

16.827.840.188.64

11.233.050.100.150.05

100.88

0.693

0.180.040.410.140.020.050.070.050.010.020.01

Run 222(5)

50.652.58

13.7210.950.196.61

11.172.930.180.240.04

99.26

0.556

0.140.040.020.050.030.070.060.030.030.020.02

Run 218(6)

50.421.45

16.159.350.187.91

11.153.060.110.170.05

100.00

0.635

0.610.040.130.080.010.030.110.060.010.010.02

Run 223(4)

50.593.04

13.4711.780.166.27

10.833.020.190.320.02

99.69

0.525

0.120.060.250.080.020.160.110.080.010.020.02

a Number in parenthesis refer to total number of analyses used for calculating each average (first column) and one standarddeviation (second column)!

FeO* = all Fe as FeO. Mg/(Mg + Fe2+) is calculated according to Kilinc et al. (1983) and the experimental conditions.

and Mg distribution assumes that all Fe occurs as Fe2+ for both liquidand augite. The KDFe/Mg (cpx/liq) values (defined in a similarmanner as for olivine) average 0.24 (Fig. 10).

EXPERIMENTAL LIQUID LINE OF DESCENT

The experimental glasses and augites have been projected into thenormative basalt tetrahedron of Yoder and Tilley (1962), reducing themulticomponent system to the components plagioclase (pi), olivine(ol), diopside (di), and quartz (q) by using a CIPW molecular normand by following the procedure outlined by Presnall et al. (1979). Ironhas been distributed between Fe2+ and Fe3+, according to Kilinc et al.(1983), and assumes the FMQ oxygen buffer. For augite, this assump-tion of the redox state of Fe is an approximation in the absence ofbetter estimates.

Figure 11 depicts a perspective drawing of the pseudoquaternarysystem pl-di-ol-q, and Figure 12 presents two pseudoternary projec-tions ol-di-q and ol-pl-q. As expected, initially the liquid compositionsmove away from a point on the line pl-ol, reflecting the proportionsof the crystallizing phases. The starting composition is hypersthene-normative. After augite begins to crystallize, the liquids are quartz-normative and move away from a point on the plane pl-ol-di inaccordance with the equilibrium phase assemblage.

The relative proportions of the crystallizing mineral phases for themultisaturated liquids reaches 22 wt% olivine, 67% plagioclase, and11% augite. These proportions are significantly more felsic thancommonly observed experimentally for transitional and tholeiiticbasalts of the ocean floor, which typically average 29% olivine, 53%plagioclase, and 18% augite (Thy, 1991; Grove and Bryan, 1983;Tormey et al., 1987). Mildly alkalic basalts will crystallize an evenmore mafic assemblage (Mahood and Baker, 1986; Thy et al., 1991).

0.75

0.5020 50 60 70

% Liquid100

Figure 2. Mg/(Mg + Fe2+) of the experimental glasses vs. percentage of liquid

remaining. Fe2+/Fe3+ was calculated according to Kilinc et al. (1983). The appear-

ance of liquidus phases is indicated; ol = olivine; pi = plagioclase; aug = augite.

DISCUSSION

Experimental equilibrium is suggested by the exchange of Fe and Mgbetween olivine or augite and coexisting liquid. The obtained KD forolivine and liquid (0.31) is consistent with that commonly observed(Roeder and Emslie, 1970; Mahood and Baker, 1986; Baker and Eggler,

863

P. THY

20

1 8

1 β

1 4

1 2

1 0

8

6

• AI2O3DCaO• FβO

D

6

3.2

2.8

2.4

2.0

1 .6

1 .2

0.8

0.4

0.0

θ

ONa2O• K2OQFe2θ3

TiO2

1 0

55

53

51

49

47

45

0.4

0.3

0.2

0.1

Siθ2

6 1 0

aug

t>

• P2O5DMnO

α pl + ol

πD

8MgO (wt%)

1 0 8MgO (wt%)

1 0

Figure 3. Composition of the experimental glasses plotted vs. MgO (wt%). Oxides calculated to 100% with Fe distributed between FeO and Fe2O3 after Kilinc et

al. (1983), assuming the FMQ oxygen buffer. The start of crystallization of augite (aug) and plagioclase and olivine (pi + ol) is indicated.

1987; Ulmer, 1989; Ussier and Glazner, 1989) and is independent oftemperature. This contrasts with the observation of Bender et al.(1978) that KD varies as a function of temperature or degree of crystal-lization. The KD for augite and liquid found here is 0.24, which appearsto be independent of temperature. This finding is consistent with theexperimental observations of Grove and Bryan (1983) and Thy et al.(1991) and for natural lavas by Perfit and Fornari (1983).

The exchange coefficients for Na and Ca between plagioclase andliquid are strongly dependent on crystallization temperature anddegree of crystallization. The KD values range from 0.60 at 1230°-1220°C to 0.99 at about 1150°C. A quantitatively similar temperaturedependence was observed by Ussier and Glazner (1989). Modelingcrystal fractionation without taking into consideration the tempera-ture dependence of KD will lead to relatively higher Na contents ofresidual liquids. Another complicating effect is that KD for plagioclaseand liquid will be dependent on composition (Drake, 1976; Glazner,1984; Ussier and Glazner, 1989).

The augites are hypersthene-normative (Fig. 12). This places thethermal divide defined by coexisting augite, olivine, and plagioclasewithin the hypersthene-normative volume and constrains the possibleliquid evolution path toward increasing normative quartz (Fig. 12).The location of the multisaturated liquids is consistent with the liquidline of descent defined by similar experiments on tholeiitic basalts byWalker et al. (1979) and Grove and Bryan (1983), among manyothers. Because the starting composition is hypersthene-normative(Fig. 12), the liquid path during equilibrium crystallization can be

predicted to reach low-Ca pyroxene-saturated, pseudoinvariant relationsand there to be fully consumed. However, because of the relatively low,but unknown, amount of liquid remaining at this point, direct determina-tion was not possible. Also, the effect of a possible low-temperaturecrystallization of Fe-Ti oxide minerals was not determined. The liquidline of descent, nevertheless, is typical for olivine tholeiitic basalts.

The similarities between the produced crystallization order (pla-gioclase, olivine; augite) with that petrographically observed (pla-gioclase; olivine; augite; titanomagnetite) suggest that the Hole797C sills crystallized at relatively low-pressure and at a low wateractivity. The strong FeO and TiO2-enrichment and Al2O3-depletionexperimentally observed for the Hole 797C basalt is typical fortholeiitic evolution trends and is an effect of the dominance ofplagioclase in the crystallizing assemblage. The low-temperatureexperimental liquids are ferrobasalts that are similar to those seen atslow-spreading segments of active rift zones. In this respect, theexperimental liquids show some general similarities with the Fe- andTi-enriched basalts drilled in the lower part of Hole 797C (Tamakiet al., 1990). On the other hand, it is clear that the evolved high-Albasalts of Hole 794C cannot be produced by low-pressure, anhy-drous crystallization. Instead, these compositions are more likely tohave evolved during crystallization under high pressure or highwater activity, either of which can suppress plagioclase crystal-lization and consequently produce A12O3 enrichments (Grove andBaker, 1984; Michael and Chase, 1987). The experiments, therefore,have no direct bearing on the origin of the high-Al basalts.

864

LOW-PRESSURE MELTING RELATIONS IN BASALT

FθO

Na2O+K2O MgO

Figure 4. AMF (Na2O + K2O, FeO*, MgO) diagram for the experimental

glasses. Fe is calculated as total FeO (FeO*). Arrow indicates direction of

falling temperature.

CONCLUSIONS

One-atmosphere, anhydrous melting experiments performed ona primitive basalt from the upper part of Hole 797C producedresidual liquids, which show a typical tholeiitic evolution trend withincreasing FeO and TiO2 as a function of decreasing temperature andincreasing degree of crystallization. The similarities between theexperimental crystallization order and that observed in the sills sup-port the conclusion that the sills crystallized at very low-pressures andlow water activities. From a major-element point of view, the experi-mental data can explain the Fe- and Ti-enriched sills of the lower partof Hole 797C, but have no direct bearing on the origin of the relativelyevolved high-Al basalts of Hole 794C.

ACKNOWLEDGMENTS

Part of this work was done in G.E. Lofgren's experimental petrol-ogy laboratory at NASA's Johnson Space Center. S.R. Yang helpedwith the microprobe and A.B. Lanier and W. Carter with the experi-mental procedures and equipment. S. Grundvig, J. Allan, and S.Yamashita helped with obtaining reliable XRF analyses of the startingmaterial. The Danish Natural Science Research Council supported theauthor during Leg 127. The shore-based study was initiated during aU.S. National Research CouncilNASA Research Associateship. Criti-cal comments from J. Allan, M. Fisk, and an anonymous reviewerimproved the manuscript.

REFERENCES

Baker, D. R., and Eggler, D. H., 1987. Compositions of anhydrous and hydrousmelts coexisting with plagioclase, augite, and olivine or low Ca pyroxenefrom 1 atm to 8 kbar: application to the Aleutian volcanic center of Atka.Am. Mineral., 72:12-28.

Bender, J. R, Hodges, F. N., and Bence, A. E., 1978. Petrogenesis of basaltsfrom the project FAMOUS area: experimental study from 0 to 15 kbars.Earth Planet. Sci. Lett., 41:277-302.

Donaldson, C. H., Williams, R. J., and Lofgren, G., 1975. A sample holdingtechnique for study of crystal growth in silicate melts. Am. Mineral, 60:324-326.

Drake, M. J., 1976. Plagioclase-melt equilibria. Geochim. Cosmochim. Ada,40:457-465.

11400.70 0.95

Figure 5. Crystallization temperature (°C) vs. Fo mol% for olivine. The bend

in the liquidus slope coincides with the appearance of augite (aug).

Glazner, A. E, 1984. Activities of olivine and plagioclase components insilicate melts and their application to geothermometry. Contrib. Mineral.Petrol, 88:260-268.

Grove, T. L., and Baker, M. B., 1984. Phase equilibrium controls on thetholeiitic versus calc-alkaline differentiation trends. J. Geophys. Res.,89:3253-3274.

Grove, T. L., and Bryan, W. B., 1983. Fractionation of pyroxene-phyricMORB at low pressure: an experimental study. Contrib. Mineral Petrol,84:293-309.

Jarosewich, E., Parkes, A. S., and Wiggins, L. B., 1979. Microprobe analysesof four natural glasses and one mineral: an interlaboratory study ofprecision and accuracy. Smithsonian Contrib. Earth Set, 22:53-67.

Kilinc, A., Carmichael, I.S.E., Rivers, M. L., and Sack, R. O., 1983. Theferric-ferrous ratio of natural silicate liquids equilibrated in air. Contrib.Mineral. Petrol, 83:136-140.

Lofgren, G. E., 1983. Effect of heterogeneous nucleation on basaltic textures:a dynamic crystallization study. J. Petrol, 24:229-255.

Mahood, G. A., and Baker, D. R., 1986. Experimental constraints on depth offractionation of mildly alkalic basalts and associated felsic rocks: Pantel-leria, Strait of Sicily. Contrib. Mineral Petrol, 93:251-264.

Michael, P. J., and Chase, R. L., 1987. The influence of primary magmacomposition, H2O and pressure on mid-ocean ridge basalt differentiation.Contrib. Mineral. Petrol, 96:245-263.

Papike, J. J., Cameron, K. L., and Baldwin, K., 1974. Amphiboles andpyroxenes: characterization of other than quadrilateral components andestimates of ferric iron from microprobe data. Geol. Soc. Am. Abstr.Programs, 6:1053-1054. (Abstract)

Perfit, M. R., and Fornari, D. J., 1983. Geochemical studies of abyssal lavasrecovered by DSRV Alvin from eastern Galapagos rift, Inca transform, andEcuador rift. 2. Phase chemistry and crystallization history. J. Geophys.Res., 88:10530-10550.

Presnall, D. C , Dixon, J. R., O'Donnell, T. H., and Dixon, S. A., 1979.Generation of mid-ocean tholeiites. /. Petrol, 20:3-35.

Roeder, P. L., and Emslie, R. E, 1970. Olivine-liquid equilibria. Contrib.Mineral. Petrol, 29:275-289.

Tamaki, K., Pisciotto, K., Allan, J., et al., 1990. Proc. ODP, Init. Repts, 127:College Station, TX (Ocean Drilling Program).

Thy, P., 1991. High and low pressure phase equilibria of a mildly alkalic lavafrom the 1965 Surtsey eruption: experimental results. Lithos, 26:223-243.

Thy, P., Lofgren, G. E., and Imsland, P., 1991. Melting relations and theevolution of the Jan Mayen magma system. /. Petrol, 32:303-332.

Tormey, D. R., Grove, T. L., and Bryan, W. B., 1987. Experimental petrologyof normal MORB near the Kane fracture zone: 22°-25°N, mid-Atlanticridge. Contrib. Mineral. Petrol, 96:121-139.

Ulmer, P., 1989. The dependence of the Fe2+-Mg cations-partitioning betweenolivine and basaltic liquid on pressure, temperature and composition. Anexperimental study to 30 kbars. Contrib. Mineral. Petrol, 101:261-273.

865

P. THY

Ussier, W., and Glazner, A. F., 1989. Phase equilibria along a basalt-rhyolitemixing line: implications for the origin of calc-alkaline intermediatemagmas. Contrib. Mineral. Petrol, 101:232-244.

Walker, D., Shibata, T., and DeLong, S. E., 1979. Abyssal tholeiites from theOceanographer fracture zone. II. Phase equilibria and mixing. Contrib.Mineral. Petrol., 70:111-125.

Williams, R. J., and Mullins, O., 1981. JSC systems using solid ceramicoxygen electrolyte cells to measure oxygen fugacities in gas-mixingsystems. NASA Tech. Mem., 58234.

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

Date of initial receipt: 27 November 1990Date of acceptance: 15 November 1991Ms 127/128B-203

0.70 0.75 0.80An (mol%)

0.85

0.51140 1160 1180 1200 1220

Temperature (°C)1240

0.4 0.6 0.8Fe/Mg liquid

1.0

Figure 7. A. Crystallization temperature (°C) vs. An mol% for plagioclase. B.KD

Na/Ca (pl/liq) vs. crystallization temperature (°C) for the experimental pla-gioclases.

Figure 6. Fe/Mg for olivine vs. Fe/Mg for coexisting liquid calculated, asdescribed in text. KD is estimated from the slope of the linear regression to 0.305.

Table 5. Chemical composition of experimental olivine.

SiO2FeOMnOMgOCaOTotal

Fo mol%K D

SiO2FeOMnOMgOCaOTotal

Fo mol%K D

Run215(6)a

41.6710.380.17

49.550.29

102.06

89.50.29

0.350.520.021.010.02

Run 220(4)

39.4114.740.32

44.220.51

99.20

84.20.31

0.230.200.050.510.06

Run 216(6)

41.1611.390.22

48.520.40

101.69

88.40.31

0.270.330.020.460.03

Run 221(4)

39.6515.420.25

44.770.49

100.58

83.80.31

0.200.080.020.170.06

Run 217(4)

40.0511.860.20

47.440.43

99.98

87.70.32

0.470.100.020.500.03

Run 222(5)

39.0618.080.29

41.910.49

99.83

80.50.30

0.340.150.030.780.06

Run 218(4)

39.4414.040.24

46.160.49

100.37

85.40.30

0.190.050.020.660.08

Run 223(3)

38.9820.19

0.2640.550.46

100.44

78.20.31

0.160.120.030.420.02

Run 219(3)

39.2314.470.26

44.660.45

99.07

84.60.31

0.100.170.020.130.01

aNumber in parenthesis refers to total number of analyses used for calculating each average (first column) and onestandard deviation (second column).

bKq for the distribution of Fe and Mg between olivine and liquid calculated assuming the FMQ oxygen buffer andFe /Fe3+ according to Kilinc et al. (1983) (see text).

866

LOW-PRESSURE MELTING RELATIONS IN BASALT

Table 6. Chemical composition of experimental plagioclase.

SiO2TiO2A12O3FeOMgOCaONapK2OTotal

An mol%K D

SiO2TiO2A12O3FeOMgOCaONa2OK2OTotal

An mol%

Run215(5)a

50.150.02

32.300.600.28

16.312.400.02

102.08

0.7890.59

0.470.011.450.050.060.360.190.01

Run 220(5)

50.940.07

30.410.710.26

14.333.290.04

100.05

0.7050.90

1.070.021.360.120.080.790.390.01

Run 216(5)

49.680.04

32.040.690.25

16.422.350.02

101.49

0.7930.55

0.310.020.560.100.050.240.120.01

Run 221(8)

50.770.05

31.970.610.20

14.853.030.03

101.51

0.7290.78

0.740.010.610.060.030.500.290.01

Run 217(6)

49.240.04

31.940.620.34

15.592.630.02

100.42

0.7650.62

1.160.011.190.090.090.840.460.01

Run 222(4)

52.230.05

30.380.730.25

13.673.520.05

100.88

0.6800.97

1.100.010.730.060.090.710.370.01

Run 218(4)

50.050.05

31.630.670.27

14.913.010.03

100.62

0.7310.74

1.050.020.430.040.030.670.310.01

Run 223(5)

52.580.06

30.340.790.19

13.373.690.03

101.05

0.6660.99

1.200.020.620.090.020.850.380.01

Run 219(5)

50.270.06

31.840.740.32

15.102.950.02

101.30

0.7380.71

0.440.020.780.130.110.280.180.01

aNumber in parenthesis refers to total number of analyses used for calculating each average (first column) and one standarddeviation (second column).

bKD for the distribution of Na and Ca between plagioclase and liquid (see text).

0.06

0.04

0.02

0.5 0.6 0.7 0.8 0.9Fo mol %

1.0

Figure 8. An mol% of equilibrium plagioclase vs. Fo mol% for coexistingolivine. Arrow indicates direction of falling temperature.

0.00

0.024

0.020

0.016

0.012

0.008

0.004

0.000

Ti

\ • •75 80 85 90

Cr

75 80 85Mg/(Mg+Fe)

90

Figure 9. Minor elements Cr and Ti (calculated to 6 oxygens) vs. Mg/(Mg +Fe) of augite. All analyses are shown, and all Fe was assumed to occur as FeO.A few analyses plotting with Mg/(Mg + Fe) above 0.85 are considereddisequilibrium phases.

867

P. THY

Table 7. Chemical composition of experimental augite.

SiO2TiO2A12O3FeOMnOMgOCaON^OCr2O3Total

En (mol%)Fs (mol%)Wo (mol%)bK

Run 221(4)a

52.660.742.305.320.15

16.9120.05

0.300.51

98.94

49.38.7

42.00.24

0.440.160.690.490.020.970.780.010.12

Run 222(4)

52.240.792.336.460.16

16.6420.470.280.45

99.82

47.610.442.00.24

0.480.080.160.530.010.340.320.020.03

Run 223(5)

51.481.072.567.510.17

16.0720.460.360.31

99.99

45.912.042.10.25

0.700.150.520.970.021.310.660.030.12

aNumber in parenthesis refers to total number of analyses used for cal-culating each average (first column) and one standard deviation (secondcolumn).

bKD for the distribution of Fe and Mg between augite and liquid calculatedassuming all iron as FeO (see text).

Figure 11. Pseudoquaternary normative projection of experimental liquids andaugites. Molecular CIPW-norm calculation and projection methods are thoseof Presnall et al. (1979), except that Fe was distributed between FeO and Fe2O3,according to Kilinc et al. (1983). The range in tie-lines between coexistingaugite and liquid is indicated. The temperature 1231° is the first appearance ofolivine and plagioclase. Arrows indicate falling temperature.

0.6 0.8 1.0Fe/Mg liquid

1.2

Figure 10. Fe/Mg for augite vs. Fe/Mg for coexisting liquid, calculated asdescribed in text. KD is estimated from the slope of the linear regression to 0.243.

ol

ol q

Figure 12. Pseudoternary normative projections of experimental liquids andaugites. Data, calculations, and projection methods as for Figure 11. The range intie-lines between coexisting augite and liquid is indicated on the ol-di-q projection,while the income of augite is indicated for the ol-pl-q projection. Arrows indicatedirection of falling temperature, ol = olivine; pi = plagioclase; di = diopside; q =quartz; aug = augite. Open diamond is the starting composition (Table 1).