A study was initiated in 1980 in Germany on a soil with 4% humus and 18.2 mg P 20 5/100 g soil and 35.7 mg K 20/100 g soil. Crops included winter wheat, barley, rye, oats, rape, and maize. Annual soil analyses showed an unexpectedly large variation following inputs of up to 60 kg P/ha and 80 kg K/ha. The difference in P and K supply between fertilized and unfertilized plots was around 20%. However, the effect on yields was less than might have been expected. Non-application of P did not influence yield, while non-application of K resulted in 5% decrease (despite high K availability). No direct relationship was found between soil analysis and yields.