ALBERT

Description: The basalt stratigraphy of the Deccan Trap between Mahabaleshwar Ghat and Belgaum over-steps the basement from north to south. Sr-isotope and Zr/Nb ratios, and Sr, Rb, and Ba concentrations correlate portions of the post-Poladpur stratigraphy over 250 km along the Western Ghats, thereby confirming a southerly component of dip of 0·06°. At the southwestern margin, the stratigraphy extends upwards from the compositionally uniform Ambenali Formation (Cox & Hawkesworth, 1984) into a sequence of grossly heterogeneous flow units which have been allocated to the Mahabaleshwar and Panhala Formations (Lightfoot & Hawkesworth, 1988). The Mahabaleshwar Formation is represented only by a sequence of highly fractionated flows (termed the Kolhapur unit) with similar 87Sr/86Sr0 to the Mahabaleshwar (0·7045), but with Sr〈240 ppm and TiO2〉2·25%. Succeeding the Kolhapur unit are a series of flows with high 87Sr/86Sr0 (0·7045-0·705), Zr/Nb 〉 13, and low Sr (〈 200 ppm), which have been allocated to the Panhala Formation, and a group of flows with high 87Sr/86Sr0 (0·707–0·708) and Sr (〉230), but trace element concentrations similar to the Mahabaleshwar Formation; these have been allocated to the Desur unit of the Panhala. Geochemical variations in flows overlying the Ambenali define two distinct trends: one is attributed to gabbro fractionation, and the other to variations in the compositions of the parental magmas, and arguably their source regions. There is little evidence for significant crustal contamination in these flows, and the degree of fractionation and the composition of the phase extract are shown to vary along strike within the Mahabaleshwar Formation. The high TiO2 content of Kolhapur unit flows is shown to be the result of shallow-level gabbro fractionation, rather than the presence of a primitive high-Ti magma. Mahabaleshwar Formation basalts exhibit a broad negative correlation between the degree of fractionation and Sr-isotopic composition. The endmember with lower 87Sr/86Sr0 has different Zr/Y from the Ambenali basalts, and would appear to have been generated by lower degrees of melting of a similar source. The other endmember has more radiogenic Sr, lower Zr/Nb, similar Zr/Y, but higher mg-number. The simplest interpretation is that these magmas were more primitive and hence hotter and more able to interact with the lithosphere en route to the surface, and that they then mixed to produce the Mahabaleshwar array. The Panhala Formation basalts plot on the Sr-Nd array defined by the Mahabaleshwar Formation, and the Desur unit basalts plot on an extension of this array; this suggests that the source characteristics are also lithospheric. The absolute elemental abundances may then be a function of melting and fractionation. We are impressed by the apparent switch from crustal lithospheric contributions to mantle lithospheric contributions through the stratigraphy, and suggest that this, together with the more protracted fractionation of the magma, reflects a change in the availability of the lithospheric components accompanying the southerly migration of the volcanic edifice.