CONCLUSIONS

  1. The basaltic clasts in the cobble conglomerate that overlay the SDRS at Site 990 are probably derived from the Precambrian basement. They are unlikely to be derived from the Site 917 lower series, and there are, therefore, no positive indications that this series was exposed and being eroded at an early stage. The reason why the lower series is missing at Site 989 may rather be the depositional offlap of the volcanic pile toward the embryonic spreading center, although erosion or faulting cannot be excluded.
  2. The lava flows drilled at Sites 989 and 990 all have chemical compositions similar to the previously drilled MORB-like (oceanic) postbreakup basalts of Sites 915 and 918. A dikelet from Site 990 and a previously drilled dike from Site 917 are similarly oceanic. We believe that the flows from Site 989 were emplaced after breakup despite their setting on the inner part of the continental margin.
  3. The lava succession at Site 990 shows a slight development upsection, from less to more fractionated compositions (Mg# 62.3-49). However, the variation is within that displayed by the oceanic rocks as a whole, and the establishment of magma chambers in the new-formed oceanic crust, characterizing the transition from the variable synbreakup conditions seen in the Site 917 upper series and to the uniform postbreakup conditions seen at Site 990, must be achieved during the deposition of the short undrilled lava succession between the two sites. The transition was accompanied by a decrease of about 3-5 kbar in the mean pressure of melting.
  4. All the lava flows from Site 990 and one from Site 989 are contaminated with continental crustal material. Only flow Unit 989-1 and two dikes appear to be uncontaminated. The contamination probably took place in the magma chambers within the young oceanic rift when this still contained fragments of continental material. The lavas would be erupted within the rift and flow subaerially away from it and onto the continent. We predict that the contamination signal will rapidly decrease in the lava succession overlying that at Site 990.
  5. The dikes show that lateral injection of oceanic magma into the continental crust also took place, even without ensuing contamination. The flows at Site 989 could have been erupted from such a dike. They constitute a rare example of oceanic basalts deposited on a continent.
  6. The primitive magma that gave rise to the oceanic basalts had an estimated 18% MgO. After fractionation of 30 mol% olivine, it shifted to gabbro fractionation (olivine + plagioclase + clinopyroxene), and the melts started to erupt. The chemical variation within the erupted lavas (8.23%-6.27% MgO) can be explained by up to 24 mol% gabbro fractionation at 3 kbar (31%-54% in the fractionation scheme for the primitive magma), but the average amount of gabbro fractionation is only 14 mol% (giving the average MgO content of 7.3%). The figures 100% primitive melt = 30% olivine cumulates + 14% gabbro cumulates + 56% melts (dikes and lava flows) compare well with the thicknesses for the three igneous layers of the thick oceanic crust modeled from published and new geophysical data over the area and give an independent indication of the character of the layers in the oceanic crust.

NEXT