In West Africa, agriculture is confronted with many production constraints, of which the most important is soil nutrient deficiency, especially phosphorus deficiency. Cowpea (Vigna ungiculata) is grown on acid, sandy soils that are very low in plant-available phosphate. Commercial phosphate fertilizers are not used by most farmers because they are expensive and/or unavailable. A possible alternative P source, particularly for leguminous species, is rock phosphate, which occurs in many West African deposits. Our objective was to assess genetic variability for P utilization efficiency in a core collection of cowpea genotypes using a sandy, P-deficient soil with (300 mg P/kg) and without added rock phosphate. Six hundred ninety six cowpea accessions were grown under greenhouse conditions for eight weeks in a Betis sand (sandy, silicious,thermic Psammentic Paleustalf) with a pH of 4.7 in water and 3 ppm Melich III P. The P treatment was added as Tahoua rock phosphate with a particle size range of 100 to 30 mesh in-1. All other nutrients were provided at optimal rates through P-free Hoagland’s solution. At harvest, plant height, and shoot and root dry weights were measured. Shoot P-content was also determined, and shoot-to-root ratios were computed. High variability in P utilization efficiency was evident from all measured parameters. Shoot:root ratios exhibited the greatest amount of genotype by P treatment interaction, suggesting that it is an appropriate indicator of genetic variability. P acquisition from added rock phosphate indicated responsive and unresponsive genotypes. Genotypes also varied in their ability to make use of the naturally low Fe-oxide bound P in the 0 P treatment. Study results may provide breeders with the raw material for selection of cowpea germplasm that is (1) adapted to West African soils of low to moderate P fertility, and (2) able to utilize P from poorly soluble rock phosphate.