An index of the buffering capacity for phosphate of a group of soils was obtained by measuring adsorption of phosphate from dilute solutions of calcium chloride. The effect of buffering capacity on the amount of phosphate initially displaced by solutions of sodium bicarbonate and on the amount of secondary adsorption from bicarbonate was then studied. These two effects were separated using a regression procedure in which the soil: solution ratio was the independent variable. As buffering capacity increased the amount of phosphorus initially displaced decreased and the amount of secondary adsorption increased. Both these changes resulted in a decrease in the amount of phosphate in the extract. The effect of buffering capacity was greater with the Olsen method (soil : solution ratio 1 : 20; 30 min) than with the Colwell method (soil : solution ratio 1 : 100; 16 h). The relation between phosphorus extracted and buffering capacity was of a similar shape to that between effectiveness of fertilizer and buffering capacity. However, the first relation depends on the conditions of extraction and the second on the kind of plant grown and on the conditions of growth. Hence the two relations do not necessarily coincide. The effect of buffering capacity on the proportion of added phosphate initially displaced from the soil became more marked as the period of incubation prior to extraction was increased. Bicarbonate soil tests would therefore indicate that, on soils of low buffering capacity, the decrease with time in availability of applied phosphate would be smaller than on soils of high buffering capacity. This effect differs from that observed with plants.