Phosplate solutions were added to samples of a soil and the samples were then incubated at a range of temperatures between 4° and 42°C for periods of up to 320 days. After various intervals, the proportion of the added phosphate which exchanged with P32was measured by shaking the soil with solutions which were 0.01 M with respect to calcium chloride and which contained appropriate concentrations of labeled phosphate. The proportion of the added phosphate which exchanged was found to be independent of the quantity of phosphate added. It decreased with increasing periods of incubation at a rate which depended on the temperature. The effects of time and temperature of incubation were comparable with previously observed effects on the concentration of phosphate in the soil solution and on the decrease in effectiveness of phosphate for plant growth. For a given period and temperature of prior incubation, the proportion of the added phosphate which exchanged was related to a fractional power of the period of equilibration with P32. The effects of level of application, period and temperature of incubation, and period of equilibration could be described by a single equation. It is suggested that the effects observed in this and in preceding papers can be explained if it is assumed that the initial adsorption reaction of phosphate is followed by a further reaction in which a second hydroxyl group of the phosphate reacts with the surface. It is postulated that there is a range of reaction velocities at individual sites. A distribution of reaction velocities which is compatible with the results is presented.