An experiment with soybeans, Glycine max (L .) Merrill, was planted at Cidra, Puerto Rico, on May 23, 1973, on a Torres clay, an Ultisol (Orthoxic Palehumults, clayey, mixed, isohyperthermic). The experiment was one of a series designed to study the relative efficiency of banded vs. broadcast P and of its residual effects on various successive crops. This was the fourth in the series after two corn and one rice experiments cropped in the same field in the seasons immediately preceding its installation. The identity of the plots was maintained throughout the whole series. There was no grain response to the broadcast or banded P, whether limed or unlimed, which had been applied to the previous crops, although soil analyses by both the Olsen and the Bray No.2 method indicated deficiency of available P. However, a field-wide grain average of 3698 kg/ha, equivalent to 55 bu/acre, was obtained. The Bray No.2 extractable P content had fallen to less than 20 p/m 2 years after applying 359 kg/ha or less fertilizer P. This yield is almost twice that of the average commercial 1973 U.S. yield and demonstrates the potential for growing soybeans on Ultisols of the humid tropics.