Using migrated common depth point reflection profiles, we find the structural differences along the ultrafast spreading (>150 mm/yr) East Pacific Rise south of the Garrett fracture zone are second-order, suggesting a remarkably uniform process of crustal accretion. The rise axis south of the Garrett transform is underlain by a narrow (<1.0 km) melt lens which shows great along-strike continuity. The depth of the axial melt sill is approximately 1200 m beneath the seafloor which is about 400 m shallower than along the slower spreading East Pacific Rise at 9¿30'N. This observation strengthens the argument that the depth to the top of the crustal velocity inversion is spreading rate dependent. Melt sill width, however, shows little variation along the East Pacific Rise, suggesting no dependence of magma chamber size on spreading rate. The melt reservoir decreases in width toward/across the 14¿27'S ridge axis discontinuity by a modest 250--300 m and appears to be continuous across this feature. Given the small aspect ratio (~1.0 km by ~50 m by ten's of kilometers) of the axial melt lens, the previously recorded jump in MgO content across the 14¿27'S offset is likely the result of a mixing boundary which is sustained through an along-strike impedance in convection. Wide-angle reflections originating at the base of seismic layer 2A, assumed to coincide with the extrusive layer, reveal a twofold to threefold increase (200--250 to 500--600m ) in thickness within 1--2 km of the rise axis. The pattern of extrusive thickening imaged south of the Garrett transform is similar to that observed along the slower spreading (110--120 mm/yr) East Pacific Rise at 9¿N.

Outside of the neovolcanic zone mean extrusive thickness is relatively invariant along a profile and from profile to profile. This implies a degree of temporal stability of the along-strike magma supply when integrated over the 10 kyr that corresponds to the width of the neovolcanic zone. The inferred uniformity of off-axis mean extrusive thickness is inconsistent with the conjecture that decreases in axial volume toward the 14¿27'S discontinuity are caused by long-term reductions in magma supply. Second-order differences in the style of extrusive thickening may be related to structural differences within the low-velocity zone underlying the rise axis and/or changes within the stress field in the overlying carapace which results in the diffuse emplacement of lavas near the rise axis. Images of Moho on cross-axis profiles may be traced to within ~1.0 km of the melt sill edge; this observations is in agreement with rise crest models which generate the lower crustal section through the advection of material down and outward from the axial melt lens rather than through cumulate deposition at the base of a large magma chamber. ¿ American Geophysical Union 1994