Since the first quantification 10 years ago of a large regional depth anomaly in French Polynesia, numerous studies have explored the origin and the geophysical and geochemical consequences of this ''South Pacific Superswell.'' These efforts would be widely viewed as a waste of time if this superswell were proven not to exist. Recently, Levitt and Sandwell (1996) have proposed just that, based on a modal analysis of multibeam bathymetric data collected on sea floor aged 15 to 35 Ma in eastern French Polynesia. Citing a discrepancy between the ETOPO5 gridded depths and ship data from 4 expeditions in their study area, they suggest that the entire Superswell could be an artifact of poor ETOPO5 sampling and gridding. Here we present a more comprehensive analysis of original ship soundings from 82 oceanographic expeditions with satellite navigation collected since 1967. The soundings span sea floor from less than 30 to more than 110 Ma. We confirm the conclusion of Levitt and Sandwell that depth anomalies calculated from ETOPO5 slightly overestimate the true values from original ship soundings on sea floor aged 30 to 35 million years, but a positive depth anomaly of 250 m exists nevertheless. On older sea floor, we find that ETOPO5 slightly underestimates the magnitude of the depth anomaly based on a modal analysis, but overall the agreement in the mode of the depth/age data between ETOPO5 and the original ship soundings is quite striking, especially given the fact that less than 20% of the ship data used in our analysis had been collected at the time that the global bathymetric grid was prepared. There indeed is a South Pacific Superswell, and the magnitude and age-dependence of the depth anomaly is as originally calculated from the mode of gridded bathymetry. ¿ American Geophysical Union 1996