A ROLE FOR LOWER CONTINENTAL-CRUST IN FLOOD-BASALT GENESIS - ISOTOPICAND INCOMPATIBLE ELEMENT STUDY OF THE LOWER 6 FORMATIONS OF THE WESTERN DECCAN TRAPS

Flows of the lower six formations of the western Deccan Traps (Jawhar through Khandala) cover a range epsilon(Nd)(T) from 0 to -20, (Sr-87/S r-86)T from 0.7062 to 0.7128, and Pb-206/Pb-204 from 16.72 to 22.43. O xygen isotopic data for fresh clinopyroxene and plagioclase separates indicate magmatic deltaO-18 values between +6.6 and +7.4 parts per tho usand. Previous isotopic studies of the upper four formations (Bushe t hrough Mahabaleshwar) have revealed two major trends that, to a first approximation, correspond to variable contamination of epsilon(Nd)(T) greater-than-or-equal-to +7 source magmas by two very different negati ve epsilon(Nd) lithospheric endmembers. Isotopic data for the lower si x formations describe completely different arrays from those of the up per formations, and the fields for the individual lower formations are also distinct from one another. Significantly, the lower formation ar rays overlap at or converge toward a common range of isotopic signatur es, with epsilon(Nd)(T) almost-equal-to 0.0 to -5.5, (Sr-87/Sr-86)T al most-equal-to 0.7067 to 0.7085, and Pb-206/Pb-204 almost-equal-to 19.2 to 20.9. These values are unlike those of oceanic mantle, and interme diate between the extremes defined by the little-contaminated Ambenali and highly crustally contaminated Bushe formations in the upper part of the stratigraphic sequence. One explanation for this common signatu re is that it represents a mantle source located in the continental li thosphere, and quite distinct from the Ambenali-like source dominating the upper formations. However, the incompatible element patterns of t he common-signature (and other lower formation) samples do not resembl e those of typical Proterozoic or Phanerozoic continental mantle xenol iths and, unlike the Ambenali basalts, all of the lower formation samp les analyzed to date have significantly higher delta O-18 values than oceanic lavas or the great majority of continental lithospheric mantle xenoliths. An alternative possibility is that the common signature ma gmas could be the products of a large-scale, open-system, lower crusta l contamination process similar to that postulated for the thick mafic complex in the Ivrea Zone of northern Italy. If so, then the isotopic arrays emanating from the common signature would represent secondary contamination episodes involving at least three different crustal endm embers. In the upper formations, a two-stage mixing process also appea rs necessary to account for the Pb-Nd and Pb-Sr isotopic relationships displayed by data for the Bushe and Poladpur formations. Model calcul ations indicate that incompatible element patterns and isotopic ratios similar to those of the common-signature samples can be produced, whi le still maintaining a basaltic major and compatible trace element com position, by mixing a large-degree partial melt (approximately 40%) of Indian Archean basic amphibolite into Ambenali-type or Reunion-type p rimitive magma. With the particular amphibolite composition used, the proportion of contamination required is large: roughly 10-30%, compara ble to the amounts proposed for the mafic complex in the Ivrea Zone. M ore siliceous contaminants permit smaller amounts of contamination but generally yield poorer trace element fits.