The fundamental problem of the west Iberia/Newfoundland conjugate pair of nonvolcanic rifted margins is one of plate geometry. How can we explain plate tectonic reconstructions that restore the Flemish Cap against Galicia Bank that at the same time further south leave an area hundreds of km wide of unusually thin crust (Fig. 1) which, according to finite element modeling, cannot be explained by any physically reasonable extension of continental crust (Bassi et al., 1993)? The problem cannot be solved by a simple geometrical translation of an independent Galicia Bank or Flemish Cap microplate.
We have presented two hypotheses, neither of which we find to be fully satisfactory, which attempt to solve the problem in the light of all available tectonic, magmatic, and other geophysical information. Hypothesis 1 emphasizes an idealized vertical cross section for oceanic crust formed by ultraslow seafloor spreading and the correspondence of the geological and geophysical evidence with such a cross section. It can explain the existence of gabbro and serpentinized peridotite at the top of acoustic basement, but the basalt that the model predicts should also exist there is found in only minute quantities in the cores. Hypothesis 2 emphasizes the plan view development of rifting along the margin. It can explain the geophysical observations and the existence of gabbro and the peridotite ridge, but has to explain the Site 899 peridotites as atypical of, and even unique to, the southern Iberia Abyssal Plain margin. In the final analysis, both hypotheses, although they approach the problem from different standpoints, tend to converge on a concept of some sort of highly heterogeneous crust that may in fact turn out to be hard to distinguish by remote (geophysical) observations. Any successful model must eventually take all the above-mentioned factors into account as well as appearing to be physically and Theologically feasible.
Some of the evidence we have is ambiguous or neutral. For example, the total tectonic subsidence (Sawyer, 1985) of the Iberia Abyssal Plain crust west of Site 901 and east of the peridotite ridge (Wilson et al. this volume) is consistent with the Sclater et al. (1980) prediction for 130-150 Ma oceanic crust. This is not a strong discriminant between the two hypotheses, however, because the predicted subsidence of very highly extended continental crust is about the same. Evidence of crustal extension on seismic reflection profiles likewise is not a strong discriminant; both hypotheses expect strong tectonic extension. Although it is generally believed that, on the basis of seismic velocity structure, normal oceanic crust can be distinguished from continental crust, which has been stretched by factors of say 3 or less, it becomes increasing clear that the situation may be far more difficult to decipher when comparison is made between crust generated by seafloor-spreading at less than 7.5 mm/yr half-rate and highly evolved disrupted and intruded continental crust. We also find that some geochemical evidence can also be ambiguous and difficult to decipher.
In spite of the apparent differences between the two hypotheses we would like to emphasize that Leg 149 drilling, and particularly the basement cores, have provided some important new and original information about the ocean/continent transition of the southern Iberia Abyssal Plain and presumably about ocean/continent transition zones of nonvolcanic margins in general. Sites 897 and 899 showed that serpentinized peridotite outcrops can exist not only at, or close to, the oceanward edge of the ocean/continent transition but also within the ocean/continent transition. Both peridotites experienced very similar melting and exhumation histories from the deep mantle, underwent important hydrothermal alteration during prolonged exposure on a seafloor elevation, and were uplifted shortly after the onset of sea- floor spreading. Site 900 provided important evidence of a cumulate gabbro, of an almost certainly synrift age, that cooled at about 13 km depth and was exhumed tectonically. The gabbro was emplaced either in the upper mantle beneath oceanic crust formed by ultraslow seafloor spreading or was underplated beneath thinned continental crust.
The differences between the Galicia Bank and southern Iberia Abyssal Plain segments of the west Iberia Margin are also important and deserve further study. It is clear that the ocean/continent transition, however formed, is much narrower off Galicia Bank. Sampling there has revealed widespread granitic rocks and only rare igneous (basaltic or gabbroic) rocks whereas in the southern Iberia Abyssal Plain segment gabbroic and basaltic material, either in situ or as clasts, features at Sites 899 and 900, and granitic material is completely absent (although the latter is probably a function of a lack of granitic outcrops on the adjacent land).
In future, all aspects of the problem deserve further study. There is a need for more extensive and detailed geophysical observations, as well as for further sampling of acoustic basement that extends, and fills gaps in, the transect of drill sites begun during Leg 149. Only by this means will we obtain a clearer view of the temporal and spatial development of the continent-to-ocean transition. It is salutary to realize that, although significant progress has been made from the cores obtained so far, basement has been sampled at only three sites in the southern Iberia Abyssal Plain. Until samples of in situ lava flows are recovered from beneath the southern Iberia Abyssal Plain, the first hypothesis must remain tentative, and, until samples of unequivocal continental basement material are obtained there, the second hypothesis must also remain tentative. We also suggest that coordinated studies of analogous active nonvolcanic rifted margins, such as the northern Red Sea, might help to progress studies of the difficult problem we have presented here.