A general model relating relative hydrothermal activity to the rate of plate creation requires data from the full spectrum of mid-ocean ridge spreading rates. To obtain data from a superfast spreading environment, the Japanese/U.S. Ridge Flux Project used continuous hydrographic/optical tow-yos in 1993 to map the distribution of hydrothermal plumes along the southern East Pacific Rise from 13¿30' to 18¿40'S. Plume incidence, the fraction of the spreading axis length overlain by a significant plume, was 0.6, including a virtually continuous vent field stretching 150 km from 17¿20' to 18¿40'S. Hydrothermal venting was most concentrated along portions of the ridge crest with an inflated cross-sectional area and an observable axial magma chamber reflector. This pattern agrees with previous results from the northern East Pacific Rise. Such consistency implies that hydrothermal circulation on fast spreading ridges is vigorous where the relative volume of partial melt is high and meager where melt volume is low or undetectable; ''hot rock'' alone is generally insufficient to drive significant hydrothermal circulation. Using a mean temperature anomaly of 0.014¿0.010 ¿C, we estimated the hydrothermal heat flux from the study area as (1.5¿1.1¿107 MW)Ux, where Ux is the cross-axis flow at the ridge crest. No direct measurements of Ux are yet available. Combining the plume distribution found here with prior data from slow, intermediate, and fast spreading ridges yields a significant linear correlation between plume incidence and spreading rate that extends across the full range of plate motion. This correlation, which mirrors that for near-axis heat flux calculated by recent models, implies that magma supply rate is the principal control on the large-scale distribution of axial hydrothermal venting. ¿1996 American Geophysical Union.