The results of phosphate (P) extraction (Pe) from soil as a function of shaking time (t) are presented. Three different extractant solutions and four soils from the Argentinean wheat belt were used. Shaking times were up to 80 hours with soil samples of one gram each. The solutions used were: 20 mL 0.5M sodium bicarbonate at pH 8.5; 7 mL 0.025M hydrochloric acid plus 0.03M ammonium fluoride; and 10 mL 0.01M calcium chloride. The effect of shaking time on Pe is explained in terms of competition effects. A model is presented which assumes that phosphate is first displaced from the soil, then partially readsorbed from the solution. Thus, the net effect on the amount of P present in the extracting solution at a given time, depends on the algebraic sum of the two separate terms involved. In general, Pe increases with shaking time for the sodium bicarbonate, decreases for the ammonium fluoride, and changes only slightly for the calcium chloride solution. Both terms, the amount of P initially displaced and the amount of P readsorbed, increased with shaking time. For each soil, the net effect on Pe is the result of both the shaking time and the other particular conditions of the extractant method. The results show that the rate of change in Pe is high enough to warrant an accurate shaking time control if the usual methods are used (Bray 1, 1–5 min; and Olsen, 30 min).