In 1982 we undertook a seismic refraction experiment, known as the MAGMA expedition, to examine the detailed structure of the East Pacific Rise near 12¿50'N. This segment of the rise, where the full spreading rate is about 110 mm/yr. is near the projected trace of the O'Gorman fracture zone and is the siet of an overlapping spreading center as well as ''black smoker'' hydrothermal activity. In this paper we describe the analysis of a subset of the travel time data collected during the MAGMA expedition, namely the data from profiles which were oriented normal to the rise axis. These profiles provide a data set roughly equivalent to those collected other experiments and sample the crosssectional structure of the rise. We have modeled these data using a two-dimensional ray-tracing program. We have found that the seismic velocities in the young oceanic crust are rather high, with velocity gradients of 4.0--5.5 s-1 in the uppermost crust. The highest velocities at the seafloor occur beneath the rise axis itself and seem to decrease as the crust ages to 0.1 Ma. This decrease in velocity must result from an increase in porosity in the upper crust and may coincide with the development of abundant surface fissures as the crust spreads. The decrease in velocity does not appear to penetrate deeper than about 0.5 km and may reverse itself as hydrothermal alteration fills the pores and cracks. The fact that the highest velocities occur under the rise axis suggests that the hydrothermal circulation responsible for the black smokers is confined to seismically unresolved conduits, a result consistent with the high temperatures and discrete nature of the vents. Our best model includes a magma chamber some 4 km wide and extending from the Moho to within about 1.1 km of the seafloor. This magma chamber is far smaller than many models for the rise axis have predicted but larger than those inferred from seismic refraction experiments at other sites on the East Pacific Rise. These discrepancies probably arise because the magma chamber under the East Pacific Rise is not a steady state feature but changes with time because of hydrothermal cooling and perhaps because of an episodic supply of magma from the mantle or along the rise axis.