We report the analysis of seismograms from two orthogonal refraction lines in the ROSE area of the East Pacific Rise at 11¿20'N recorded by a three-component ocean bottom seismometer (OBS). Converted shear waves are clearly observed on the horizontal seismometers for both lines, including paths crossing the rise axis. Neither a travel time delay nor measurable attenuation is observed for shear or compressional waves which have passed beneath the rise crest at depths down to 4 km. On the basis of these observations we conclude that no large crustal magma chamber is present beneath the rise at this latitude. We invert the travel times of P and converted S waves to infer the crustal structure in the vicinity of the rise, and we use P wave amplitudes to constrain further the range of solutions. The P wave velocity distribution does not differ markedly from that of ''normal'' young oceanic crust, but the shear velocity distribution results in a Poisson's ratio (0.32) that may be more typical of highly fractured crust in rise-axis regions. A thin (~80 m) layer of sediment is inferred to be present beneath the OBS on the basis of the measured particle motions, the observation of an Sp phase, and the reverberatory character of the seismograms. A combined consideration of the results of this experiment and other observations of rise-crossing crustal phases elsewhere in the ROSE area leads us to conclude that any magma within the crust beneath the rise axis of this region must be confined to narrow dikes or pipes or to small isolated bodies with a vertical thickness of less than about 1 km.