A greenhouse experiment was conducted to evaluate varietal differences in the uptake and availability of P from Gafsa phosphate rock (PR) to five cowpea cultivars grown in a low-P Paleustult soil from Ghana, using the A value technique. The P-32 radioisotope used as a tracer was P-32-labelled triple superphosphate (TSP). Each cultivar received sole or a combined application of the two fertilizers (TSP, PR). From these treatments it was possible to estimate for each cultivar, A(R) + A(S), A(S) and by difference A(R) (A(R) and A(S) stand for A values for phosphate rock and soil, respectively). Using this approach we measured significant genotypic differences in P uptake from PR and A(R) values. The ranking of the cultivars in P uptake from PR was the same as for A(R), i.e. Asontem > Vallenga > Soronko > IT81D-1137 > Amantin. Similarly, ranking for uptake from soil P was the same as A(S), i.e. Vallenga, Soronko, IT81D-1137, Asontem and Amantin. Thus although PR could not be labelled directly, using the A value approach it was possible to distinguish between P availability from PR and soil to the plant. The recoveries of applied TSP ranged from 8.0% to 9.4% and those of PR from 2.3% to 3%. The other advantage of the A value method is that it made it possible for the different genotypes tested to be compared directly in units of a standard fertilizer, TSP in this case. Thus for Vallenga in this soil 3.01 kg of P in Gafsa PR was capable of supplying the same amount of P that could be supplied by 1 kg P of TSP; whereas for Amantin a higher amount i.e. 3.5 kg P as Gafsa PR was needed. This information is useful for adjusting application rates to be recommended for different P fertilizer sources in field trials so as to achieve similar effects.