Using five Argentinian soils, we studied the ability of the Colwell, the Olsen and the Bray-1 tests to reflect four factors that affect the availability of phosphate. The four factors were: the amount and solubility of native P present; the amount of P added; the period of contact between soil and phosphate; and the capacity of the soil to retain P. Phosphate was added to the soils, and they were incubated at a range of temperatures for periods of up to 60 d. Appreciable native P was extracted from some of the soils, especially by the Colwell method. At zero added P there was, for one soil, a slight decrease in P extracted with increasing incubation. For the others there was little change. This contrasted with previous studies, which showed that, on prolonged incubation, these soils released phosphate to the soil solution. In most cases, the proportion of the added P extracted increased as the level of addition increased. This occurred even when there was no prior incubation. It was therefore due to a nonlinear sorption reaction despite the presence of the reagents rather than to a faster reaction at low levels of addition of P. The amount of P extracted from the soil decreased as the period and temperature of incubation increased. These effects of level of addition and of time and temperature were closely described by functions that described the increase with time of both the retention of added P and the amount of native P that was potentially extractable. The recovery of added P decreased as the capacity of the soil to sorb P increased. The effects of both time and P sorption were reflected differently by the different soil tests. The vigorous conditions of the Colwell test resulted in a smaller effect of both the period of incubation and of P sorption.