5. TRACE ELEMENT AND ISOTOPIC EVIDENCE FOR MAGMA MIXING IN ALKALIC ANDTRANSITIONAL BASALTS NEAR THE EAST PACIFIC RISE AT 8°N

Rodey Batiza, Department of Earth and Planetary Sciences and McDonnell Center for the Space Sciences,Washington University, St. Louis, Missouri

andJeffrey R. Johnson, Geological Research Division A-015, Scripps Institution of Oceanography, University of

California at San Diego, La Jolla, California

ABSTRACT

Trace element abundances and 87Sr/86Sr ratios of Mg-rich tholei-ites, alkalic basalts, and transitional basalts from very young oceaniccrust near the East Pacific Rise at 8°N show wide variability. Thetholeiites are depleted in large-ion-lithophile (LIL) elements and have8?Sr/86Sr = 0.70270. The alkalic basalts are LIL-enriched, and freshsamples have 87Sr/86Sr = 0.70300. Chemical and isotopic effects ofalteration, fractional crystallization, "dynamic melting," and man-tle heterogeneity are observed in each group. Fresh transitionalbasalts have intermediate LIL-enrichment and 87Sr/86Sr = 0.70287.These basalts are hybrid and originate by mixing of tholeiitic andalkalic basalt near the East Pacific Rise.

INTRODUCTION

Numerous petrologic and geophysical studies haverecently been carried out in the Siqueiros area of theEast Pacific Rise (EPR) (Figure 1) at about 8°N (Batizaet al., 1977; Johnson, 1979; Scientific Staff, 1977;Rosendahl et al., 1976; Orcutt et al., 1976; Lonsdale,1979; Rosendahl, 1976; Crane, 1976). The geophysicalstudies have established that the crest of the EPR isunderlain by a narrow crustal low-velocity zone which isprobably a magma chamber. This magma chamber pro-vides a locus for production of a wide range of crystalfractionation products from an Mg- and Ni-rich, andlarge-ion-lithophile (LIL) element depleted parentaltholeiite liquid. Abundant Fe- and Ti-rich tholeiites ofthis type have been recovered from the Siqueiros area(Batiza et al., 1977; Johnson, 1979; Scientific Staff,1977). In addition, Ne-normative, less LIL elementdepleted alkalic basalts have been dredged in the Si-queiros area (Batiza et al., 1977). Basalts that arechemically and isotopically intermediate between Mg-and Ni-rich, LIL element depleted tholeiite and thesealkalic basalts have also been recovered (Johnson,1979). In this report we present additional trace-elementand isotopic data supporting the hypothesis that thesetransitional basalts are hybrid magmas produced by themixing of tholeiitic and alkalic basalt melts. These dataalso provide additional constraints on the origin of theSiqueiros alkalic basalts and on the nature of chemical

and isotopic mantle heterogeneities below very youngcrust at the EPR.

RESULTS

Major oxide, trace element, and isotopic data for thetholeiites, alkalic basalts, and transitional basalts aregiven in Table 1. Rare earth element (REE) abundancesare shown normalized to chondrites (Haskin et al.,1968) in Figure 2. The tholeiites are typical of light rareearth element (LREE) depleted mid-ocean ridge basalt(Engel et al., 1965; Tatsumoto et al., 1965; Frey et al.,1974 and references therein; and Kay and Hubbard,1978 and references therein). They are Ol-Hy norma-tive, have very low concentrations of LREE, Rb, Sr, Zr,Hf, and U, and have 87Sr/86Sr = 0.7027. The Sm/Ndand 143Nd/144Nd of these samples (Carlson et al., 1977,1978) is also within the range of mid-ocean ridgebasalts, which are thought to be produced from "de-pleted" portions of the sub-oceanic upper mantle (Gast,1968; DePaolo and Wasserburg, 1977).

The alkalic basalts, on the other hand, are enriched inLREE, other LIL elements, and 87Sr relative to the tho-leiites. The enrichments of LIL elements in the alkalicbasalts are not as large as those observed for oceanic is-land alkalic basalts (Sun and Hanson, 1975; Gast,1968), nor do these basalts have as much normative Neas typical alkalic island basalts have (Batiza et al.,1977). They have higher abundances of heavy rare earthelements (HREE) and less Sr than is typical of island

63

R. BATIZA, J. R. JOHNSON

106 104° 103° 102°

Figure 1. Bathymetric map of the Siqueiros area showing locations of dredge hauls. Rocks from SD-8, DS-3, andDS-4 are described in this paper.

alkalic basalts, but their Sm/Nd and 143Nd/144Nd ratiosare comparable to ocean island values (Carlson et al.,1977, 1978).

The major oxide and trace element abundances of thetransitional basalts are intermediate between those ofthe tholeiites and the alkalic basalts (Table 1 and Figure2). The 87Sr/86Sr ratios of the transitional basalts have arange that overlaps the values obtained for the alkalicbasalts. The Siqueiros transitional basalts have 143Nd/144Nd ratios intermediate between alkalic and tholeiiticvalues (Carlson et al., 1977, 1978). The chemical com-position of the transitional basalts in the Siqueiros areais very similar to that of samples drilled on DSDP Leg 2in the Atlantic (Frey et al., 1974), indicating that suchbasalts may not be altogether rare. Similar basalts havealso been sampled in the North Atlantic (Wood et al.,1979 and references therein).

DISCUSSION

The chemical effects of several individual processescan be identified within the basalt from the Siqueiros

area. These processes include: alteration, fractionalcrystallization, mantle heterogeneity, dynamic melting,and mixing. Evidence for each of these will be brieflydiscussed.

Alteration

Effects of sea-water alteration on the major oxidechemistry of the Siqueiros basalts have been discussedby Batiza et al. (1977) and Johnson (1979). There is ad-ditional evidence that the 87Sr/86Sr ratio of some of thesamples has been elevated by sea-water interaction.Even without petrographic evidence of extensive altera-tion, it has been clearly demonstrated that such interac-tion can significantly raise the 87Sr/86Sr ratio of submar-ine volcanic rocks (DePaolo and Wasserburg, 1977;Dasch et al., 1973; Hart et al., 1974). Figure 3 showsthere is good correlation between the Fe + 3/Fe+2 and87Sr/86Sr ratios of some of the basalt samples from theSiqueiros area, although there is no corresponding ex-tensive visible alteration in the rocks. If only the lowestvalues of the 87Sr/86Sr ratio obtained for each rock

64

TRACE ELEMENT AND ISOTOPIC EVIDENCE

TABLE 1Major Oxide and Trace Element and Isotopic Dataa

Oxide orElement

wt. %SiO2

TiO 2

A12O3

F e 2 O 3

FeOMnOMgOCaONa2OK 2 Op 2 θ 5

H 2 O

Total

PpmLaCeSmEuTbYbLuRbSrZrHfUCoCrNiCuZnSc

87Sr/86sr

SD8-3

50.30.93

15.61.277.450.149.80

12.772.150.050.09-

100.55

1.595.412.070.800.572.230.370.40

125-1.40.027

46478130

828044.6

0.70270

DS3-3

49.10.91

17.62.36.10.169.74

12.452.50.080.091.0

102.03

1.384.222.090.830.482.000.30-

9942

1.50.014

60.2427176

726929.9

-

D S b

4-1 a

49.21.90

16.52.896.40.158.14

10.752.70.600.291.5

99.23

9.7025.95

4.731.630.813.180.49—

169206

4.00.224

59.740614348

11033.1

-

D S b

4-6

48.81.87

16.42.896.50.178.51

10.753.20.560.281.4

99.64

10.1325.19

4.951.510.853.370.49—

180229

3.90.237

48307133489332.1

-

D S b

4-14

50.11.90

16.5-9.320.168.38

10.833.10.430.29-

101.01

10.1225.2

4.781.620.993.190.485.51

199224

4.00.224

47.4310129

609033

0.70312

D S b

4-15

48.21.86

16.2-9.200.169.08

10.573.10.380.28-

99.03

9.4624.90

4.791.460.863.120.476.90

223—4.00.228

70349142

607832.3

0.70287

SD8-5

47.62.45

16.12.327.700.136.809.383.581.230.12-

97.41

23.5451.15

6.772.121.213.120.43

13.54380344

5.50.660

39.5157

—_

8025.7

0.70307

DS8-6

48.22.52

15.22.318.560.177.159.693.790.900.44-

98.93

18.4841.19

6.672.021.173.680.54

13.09342289

5.20.359

43196113

449728.8

0.70300

Major oxide data are from Batiza et al.(1979),and Johnson (1979), where the determinative procedures and theiraccuracy are discussed. REE were determined by instrumental neutron activation (Jacobs et al., 1977), as wereHf, Zn, and Sc. Rb was determined by isotope dilution mass spectrometry. Sr, Zr, Co, Cr,and Ni were deter-mined by both INAA and atomic absorption spectrophotometry. Agreement was found to be within 2 per centof the atomic absorption method (Batiza et al., 1977), and values obtained by both methods were averaged. Cuwas determined by AA. U was determined by induced fission track methods. 87s r/86s r w a s determined by massspectrometry; 8 7 s r /86sr values were normalized to an 86§r/88sr value of 0.1194, and values are ±0.00006.From Johnson, 1979.

group are used, then the transitional basalts have87Sr/86Sr ratios intermediate between the alkalic basaltand the tholeiite values. Samples with lowest 87Sr/86Srratios have Fe + 3 /Fe + 2 = -0.2. Isotopes of Nd are lesssubject to alteration by sea-water interaction; indeed,Carlson et al. (1977, 1978) have shown that the Si-queiros transitional basalts have 143Nd/144Nd ratios in-termediate between alkalic and tholeiitic values.

Fractional Crystallization

The importance of shallow-level crystal fractionationin the production of Fe- and Ti-rich basalts in the Sique-iros area has been confirmed by the work of Batiza et al.(1977) and is further supported by trace element andisotopic data (Batiza, unpublished data). However,shallow fractional crystallization of a single parental li-quid cannot be responsible for the production of tholei-

ites, alkalic basalts, and transitional basalts (Johnson,1979). First, the isotopic data preclude such an originfor the three basalt types, and secondly no quantitativefractionation models using low-pressure phase assem-blages can be made to fit the observed major oxide andtrace element abundances in the lavas. The alkalic andtransitional basalts cannot be related to either an Mg-rich tholeiitic liquid or fractionated tholeiitic liquid byfractional crystallization unless other processes are in-voked to account for discrepancies in the isotopic andchemical data.

Mantle Heterogeneities

The Sr and Nd isotopic data for alkalic and tholeiiticbasalts in the Siqueiros area indicate that the mantle be-low is isotopically heterogeneous. Also, chemical het-erogeneity is indicated by large differences in La/Sm,

65

R. BATIZA, J. R. JOHNSON

ALKALIC BASALTS

Figure 2. REE abundance patterns of the samples fromTable 1 normalized to chondrites (Haskin et al,1968). Also shown is the pattern resulting from themixing model described in the text. (Transitionalbasalt patterns are dashed.) Note similarity betweenthe mixing model and the observed REE abundancesof the transitional basalts.

0.7029/ THOLEIITIC

Fe+3/Fe+2

Figure 3. Plot of Fe+3/Fe+2 and 87Sr/86Sr ratios foralkalic and tholeiitic basalts. Most of the tholeiitedata are unpublished but are available from theauthors.

K/Rb, and Zr/U ratios between alkalic and tholeiiticbasalts (Table 1). Smaller differences in such ratios areobserved between individual samples within the tholei-itic and alkalic groups (Table 1) and indicate smaller-scale heterogeneity in the sources of both tholeiites andalkalic basalts (Church and Tatsumoto, 1975; Langmuiretal., 1977).

Dynamic MeltingEvidence for "dynamic melting" (Langmuir et al.,

1977) in the upper mantle in the Siqueiros area comesfrom the crossing chondrite-normalized REE patternsof the alkalic basalts (Figure 2). Langmuir et al. (1977)have shown that such patterns cannot result simplyfrom one-step mantle or crustal processes. Instead, theypropose that simultaneous action of several simple pro-cesses produces a wide range of trace-element abun-dance patterns from a single mantle source. Thus, as-suming that a single source was melted to produce thealkalic basalts from the Siqueiros area, it is necessary toadduce a process such as dynamic melting for the vari-able LREE and HREE abundances in samples SD 8-5and SD 8-6. Other mechanisms, such as batch process-ing, could produce similar patterns (O'Hara, 1977). Theisotopic differences between these two samples are notsignificant (Table 1 and Figure 3).

MixingFigure 4 and Table 2 provide strong evidence that the

transitional basalts originated by mixing, whose effectsare most clearly seen in the transitional basalts. Johnson(1979) proposed that mixing of little-fractionated tholei-itic and alkalic basalt could produce the transitionalbasalts of the Siqueiros area. Mixing did not occur be-tween strongly fractionated East Pacific Rise ferroba-salt and alkali basalts. Furthermore, it is possible toplace quantitative constraints on the composition of themixing end members using the graphical methods out-lined by Langmuir et al. (1978). Not only does Figure 4show that mixing occurred, but also that probably truemagma mixing occurred rather than mixing of sourcesprior to melting. This hypothesis was put to a quantita-tive test for all oxides and trace elements (see Table 2),the results showing remarkably good agreement withobserved major oxide and trace element abundancesand isotope ratios, in the transitional basalts.

Averages of little-fractionated tholeiitic basalts, alka-lic basalts, and transitional basalts were employed inthis test in hopes of eliminating the chemical effects ofother processes discussed previously. This averaging haslittle influence on the validity of the test, since indi-vidual samples within each group do not vary greatly intheir chemistry. In addition to the elements used inFigure 4, many other element and isotope ratio-ratioand element-ratio pairs are consistent with the mixinghypothesis. The results of these graphic methods indi-cate that the end-member compositions are very close tothose of the averaged alkalic and tholeiitic basalts ofTable 2. The large geographic distances between dredgesSD-8, DS-3, and DS-4, however, preclude mixing ofthese particular samples. The modeling, though, pro-vides strong evidence that hybridization of alkalic andtholeiitic basalts did actually occur. Further, the closechemical similarity between the various transitionalbasalt samples, as well as the small inferred volumes of

66

TRACE ELEMENT AND ISOTOPIC EVIDENCE

TABLE 2Mixing Model Dataa

10-

Ce/Yb

~i 1 r

8-5

. -4 -154 .6. .* .4-1 a

,. 4-14

8-3 3:3 ×I ×-i L

10Ce/Yb

Figure 4. Plots of Hf/U versus Ce/Yb (a) and U/Ybversus Ce/Yb (b) for the basalts of this study. Asshown by Langmuir et al. (1978), such data arrayscan be used to interpret the role of mixing in mag-matic processes. The linearity of the U/Yb -Ce/Ybdata array (b) (which is the "companion plot" of theHf/U versus Ce/Yb plot [a]) indicates that magmamixing rather than source mixing prior to melting oc-curred in these samples. Note also that samples SD8-5 andDS 3-3 are very close to the asymptotes of thehyperbolic mixing curve, and are thus close to the ac-tual mixing end members.

such rocks (Johnson, 1979), suggests that the dredginghas sampled the products of only one mixing event.Since the model shows some significant discrepancies(e.g., Zr; Table 2), it is improbable that the actual mix-ing end members have been recovered and analyzed.The general good fit of the model, however, suggeststhat the analyzed samples are very similar to actual liq-uids involved.

Implications

If the mixing hypothesis for the origin of the transi-tional basalts in the Siqueiros is correct, then it holdsseveral implications for magmatic processes at the EPRand perhaps other oceanic ridges and also for the size of

Oxide orElement

SiO2

TiO2AI2O3FeO*MnOMgOCaONa2θK2OP2O5

Total

LaCeSmEuTbYbLuRbSrZrHf

uVCoCrNiCuZnSc

8VSr/86Sr

AverageTholeiite

49.70.92

16.68.390.159.77

12.622.330.070.09

100.64

1.486.072.080.810.522.110.340.40

11042

1.40.020

21053

441152

777437.3

0.70270

AverageAlkaliBasalt

47.92.49

15.610.210.156.989.543.691.070.28

97.77

21.0146.17

6.722.071.193.400.49

13.32354316

5.30.509

29741

183113

4488.527.2

0.70303

AverageTransi-tionalBasalt

49.11.89

16.49.160.168.53

10.733.030.490.28

99.77

9.8525.68

4.811.550.873.210.486.21

208219

4.00.228

26655

319135

549232.5

0.70287

50:50MixingModel

48.81.71

16.19.300.158.38

11.083.010.540.19

99.26

11.2526.124.401.430.852.760.416.86

232176

3.350.264

25347

312132

6081.532.2

0.70295

R 2

0.0900.0320.0900.0200.0010.0230.1230.0020.0020.008

0.39

% Difference12

2-9-8-2

-14-15

910

-18-16

14-5

-15-2-210

-11-1

i The tholeiite column is an average of analyses DS 3-3 and SD 8-3. The alkalicbasalt column is an average of SD 8-5 and SD 8-6. The transitional basaltcolumn is an average of the four transitional basalts from Table 1. The 50:50mixing model column represents a mix of equal weights of tholeiitic and alkalicbasalt. These proportions were determined by obtaining the mean of the indi-vidual proportions required for fitting each element. Iterative computer meth-ods yield essentially the same result. Note the excellent agreement betweenobserved and calculated chemical and isotopic composition of the transitional(hybrid) basalts. Residuals for the major oxides are given in the last column(R2); the sum of R2 = 0.39. Also given are the percentage differences for eachtrace element. The lava compositions were averaged in order to minimize theeffects of differences in degree of partial melting, small-scale mantle heterogene-ity, and crystal fractionation. Complications arising from these processes arediscussed in the text.

mantle heterogeneities in the sub-oceanic mantle. Thetransitional basalts from Siqueiros occur on oceaniccrust dated with sea-floor magnetic anomalies to be 1.6m.y. old (Johnson, 1979), but they could be muchyounger. The alkalic basalts occur on crust <0.5 m.y.old (Batiza et al., 1977). This indicates that productionof alkalic basalts, their eruption, and mixing with tho-leiitic basalt liquids occur very close to the ridge crest.The low Sr and high HREE abundances in these alkalicbasalts compared with other oceanic alkali basalts sug-gest that plagioclase rather than garnet is the dominantAl-rich phase in the zone of melt production. If so, thedepth of production of these melts is <9-10 Kbar(Yoder, 1976, p. 27). This implies that isotopic andchemical heterogeneities of relatively great magnitudeare present in the upper mantle very close to the zone ofproduction of mid-ocean ridge tholeiite. The exactnature, size, geometric arrangement, and cause of theseheterogeneous mantle sources are subjects of consid-erable debate (Tatsumoto, 1978). The results of this

67

R. BATIZA, J. R. JOHNSON

study indicate that many of these heterogeneous do-mains may be small, as shown by the low volumes ofmelt produced, and that these small-size heterogeneitiesare spatially and temporally closely related to the sourceof normal ridge crest tholeiitic volcanism.

CONCLUSIONS

The major conclusions of this study are:1) Isotopic and chemical heterogeneities of variable

magnitude occur very close to the EPR in the vicinity ofthe Siqueiros transform fault.

2) The effects of such heterogeneity are clearly evi-dent in the chemistry of the basalts, as are the effectsof fractional crystallization, sea-water alteration, andcomplex melting phenomena.

3) The origin of the transitional basalts of the Si-queiros area is almost certainly due to mixing of tholei-itic and alkalic basalt; the transitional basalts are hybridliquids.

ACKNOWLEDGMENTS

We thank J. D. Macdougall for providing the fission trackU data and R. Carlson for supplying a preprint. We are espe-cially indebted to C. Hedge and K. Futa of the USGS IsotopeGeology Branch and to L. Haskin, J. Jacobs, W. Day, M.Lindstrom, and D. Lindstrorn for their generosity in allowinguse of their facilities and providing assistance in collecting themass spectrometric and INAA data, respectively. The workwas supported by the McDonnell Center for Space Sciences,St. Louis, Missouri, and NSF Grant OCE 77-27001 to R. B.An earlier draft of this paper was improved by the review com-ments of W. B. Bryan and J. H. Natland.

REFERENCES

Batiza, R., Rosendahl, B. R., and Fisher, R. L., 1977. Evolu-tion of oceanic crust 3. Petrology and chemistry of basaltsfrom the East Pacific rise and the Siqueiros transformfault, Journal of Geophysical Research, v. 82, p. 265-276.

Carlson, R. W., Macdougall, and Langmuir, G. W., 1977.Sm-Nd constraints on the evolution of oceanic rocks,Transactions of the American Geophysical Union, v. 58,no. 6, p. 533-534.

, 1978. Differential Sm/Nd evolution in oceanicbasalts, Geophysical Research Letters, v. 5, p. 229-232.

Church, S. E. and Tatsumoto, M., 1975. Lead isotope rela-tions in oceanic ridge basalts from the Juan de Fuca-GordaRidge area, N.E. Pacific Ocean, Contributions to Mineral-ogy and Petrology, v. 53, p. 253-279.

Crane, K., 1976. The intersection of the Siqueiros transformfault and the East Pacific rise, Marine Geology, v. 21, p.25-46.

Dasch, E. J., Hedge, L. E., and Dymond, J., 1973. Effect ofsea-water interaction on strontium isotope composition ofdeep-sea basalts, Earth and Planetary Science Letters, v.19, p. 177-189.

DePaolo, D. J. and Wasserburg, G. J., 1977. The sources ofisland arcs as indicated by Nd and Sr isotopic studies,Geophysical Research Letters, v. 4, p. 465-468.

Engel, A. E. J., Engel, C. G., and Havens, R. G., 1965.Chemical characteristics of oceanic basalts and the uppermantle, Geological Society of America Bulletin, v. 76, p.719-734.

Frey, F. A., Bryan, W. B., and Thompson, G., 1974. Atlanticocean floor: Geochemistry and petrology of basalts fromLegs 2 and 3 of the Deep Sea Drilling Project, Journal ofGeophysical Research, v. 79, p. 5507-5527.

Gast, P. W., 1968. Trace element fractionation and the originof tholeiitic and alkaline magma types, Geochimica et Cos-mochimica Acta, v. 32, p. 1057-1086.

Hart, S. R., Erlank, A. J., and Kable, E. J. D., 1974. Seafloor basalt alteration: some chemical and strontium iso-topic effects, Contributions to Mineralogy and Petrology,v. 44, p. 219-230.

Haskin, L. A., Haskin, M. A., Frey, F. A., and Wildeman,T. R., 1968. Relative and absolute terrestrial abundances ofthe rare earths. In Ahrens, L. H. (Ed.), Origin and Distri-bution of the Elements: New York (Pergamon Press), p.889-912.

Jacobs, J. W., Korotev, R. L., Blanchard, D. P., and Haskin,L. A., 1977. A well-tested procedure for instrumental neu-tron activation analysis of silicate rocks and minerals, Jour-nal of Radioanalytical Chemistry, v. 40, p. 93-114.

Johnson, J. R., 1979. Transitional basalts and tholeiitesfrom the East Pacific Rise, 9°N, Journal of GeophysicalResearch, v. 84, p. 1635-1651.

Kay, R. W. and Hubbard, N. J., 1978. Trace elements inocean ridge basalts, Earth and Planetary Science Letters, v.38, p. 95-116.

Langmuir, C. H., Bender, J. F., Bence, A. E., Hanson, G. N.,and Taylor, S. R., 1977. Petrogenesis of basalts from theFAMOUS area, Mid-Atlantic Ridge, Earth and PlanetaryScience Letters, v. 36, p. 133-156.

Langmuir, C. H., Vocke, R. D., Hanson, G. N., and Hart, S.R., 1978. A general mixing equation with applications toIcelandic basalts, Earth and Planetary Science Letters, v.37, p. 380-392.

Lonsdale, P., 1979. A pair of young cratered volcanoes onthe East Pacific rise, Journal of Geology, v. 87, p. 157-173.

O'Hara, M. J., 1977. Geochemical evolution during fraction-al crystallization of periodically refilled magma chambers,Nature, v. 266, p. 503-507.

Orcutt, J. A., Kennett, B. L. N., and Dorman, L. M., 1976.Structure of the East Pacific Rise from an ocean-bottomseismometric survey, Geophysical Journal of the RoyalAstronomical Society, v. 45, p. 305-320.

Rosendahl, B. R., 1976. Evolution of oceanic crust, 2, con-straints, implications and inferences, Journal of Geophys-ical Research, v. 81, p. 5305-5311.

Rosendahl, B. R., Raitt, R. W., Dorman, L. M., Bibee, L. D.,Hussong, D. M., and Sutton, G. H., 1976. Evolution ofoceanic crust, 1, A geophysical model of the East PacificRise crest derived from seismic refraction data, Journal ofGeophysical Research, v. 81, p. 5297-5302.

Scientific Staff, 1977. Glomar Challenger completes 54thcruise, Geotimes, v. 22, no. 11, p. 19-23.

Sun, S. S. and Hanson, G. N., 1975. Evolution of the mantle:geochemical evidence from alkali basalts, Geology, v. 3, p.297-302.

Tatsumoto, M., 1978. Isotopic composition of lead in oceanicbasalt and its implication for mantle evolution, Earth andPlanetary Science Letters, v. 38, p. 63-87.

Tatsumoto, M., Hedge, C. E., and Engel, A. E. J., 1965. Po-tassium, rubidium, strontium, thorium, uranium and the

68

TRACE ELEMENT AND ISOTOPIC EVIDENCE

ratio of strontium 87 to strontium 86 in oceanic tholeiitic IPOD Leg 49: Implications for mantle heterogeneity: Earthbasalts, Science, v. 150, p. 886-889. and Planetary Science Letters, v. 42, p. 77-97.

Wood, D. A., Tarney, J., Varet, J., Saunders, A. D., Boug-ault, H., Joron, J. L., Treuil, V., and Cann, J. R., 1979. Yoder, H. S., 1976. Generation of Basaltic Magma: Washing-Geochemistry of basalts drilled in the North Atlantic by ton (National Academy of Sciences Press).

69