The Southwest Indian Ridge is a slow spreading end-member of the mid-ocean
ridge system. The deepest borehole penetrating the lower oceanic crust, Oce
an Drilling Program hole 735B, lies on the eastern transverse ridge of the
Atlantis II Fracture Zone at 57 degreesE. A wide-angle seismic survey in th
e vicinity of the borehole reveals a crustal structure that is highly heter
ogeneous. To the east of Atlantis Bank, on which hole 735B is located, the
crust consists of a 2-2.5 km thick high-velocity-gradient oceanic layer 2 a
nd a 1-2 km thick low-velocity-gradient layer 3. The transform valley has a
2.5-3 km thick crust with anomalously low velocities interpreted to consis
t largely of highly serpentinized mantle rocks. The seismically defined cru
st is thickest beneath the borehole, where layer 2 is thinner and the lower
crust is inferred to contain 2-3 km of partially serpentinized mantle. The
seismic velocity models are consistent with gravity data which show weak r
esidual mantle Bouguer anomalies because the regions of thinner crust have
lower crustal densities. Stress variations deduced from mass balances betwe
en the transform valley floor and the adjacent transverse ridges are much l
arger than the likely threshold for lithospheric failure and therefore indi
cate that the relief is supported dynamically. The variation of crustal thi
ckness with spreading rate defined by data from the Southwest Indian Ridge
and elsewhere is consistent with models of melt generation in which the upw
elling mantle is cooled by conductive heat loss at very slow spreading rate
s? resulting in reduced melt generation under the spreading axis. Large seg
ment-scale variations in crustal thickness suggest subcrustal along-axis mi
gration of melt toward segment centers.