Abstract. The behaviour of potassium (K) in a range of arable soils was examined by plotting the change in exchangeable K of the topsoil (Δ Kex) at the end of a 3–5 year period against the K balance over the same period (fertilizer K applied minus offtake in crops, estimated from farmers' records of yield and straw removal). Based on the assumption that values for offtake per tonne of crop yield used for UK arable crops MAFF 2000) are valid averages, 10–50% of Δ Kex was explained by the balance, relationships being stronger on shallow/stony soils. Excess fertilizer tended to increase Kex and reduced fertilization decreased it, requiring between 1.2 and 5.4 kg K ha−1 for each mg L−1Δ Kex. However, merely to prevent Kex falling required an extra 20 kg K ha−1 yr−1 fertilizer on Chalk soils and soils formed in the overlying Tertiary and Quaternary deposits, despite clay contents >18%. Whereas, on older geological materials, medium soils needed no extra K and clays gained 17 kg K ha−1 yr−1. It is unlikely that the apparent losses on some soil types are anomalies due to greater crop K contents. Theory and the literature suggest leaching from the topsoil as a major factor; accumulation in the subsoil was not measured. Recommendations for K fertilization of UK soils might be improved by including loss or gain corrections for certain soil types.

Abstract. Ten chalk topsoils (0‐25 cm) were repacked into columns in the laboratory. After leaching similar to one year's throughflow in the field, loss of K was equivalent to between 9 and 74kg K/ha. This represented between 3 and 30% of the initial exchangeable K with which loss was poorly correlated. Loss was dependant on the soil solution concentration and was inversely proportional to potassium buffer power. The loss of magnesium in the same columns was between 10 and 22 kg Mg/ha (6‐21% of the initial exchangeable Mg). Magnesium loss was poorly correlated with exchangeable Mg. When KCl fertilizer was incorporated into the soils, the increase in leaching of potassium was 1–35% of the K addition. Application to the top of the column resulted in less leaching than when the K was incorporated. Leaching of magnesium was increased by up to 5 kg Mg/ha. Potassium leaching may be delayed by the underlying A/C horizon but pure chalk, with an extremely low buffer power for K, has little ability to retain K. Extremely calcareous topsoils were the most leaky although in practice it is the organic chalk soils on which it is most difficult to attain adequate K levels. On all chalk soils, maintenance of a high K level with K fertilizer is likely to cause unnecessary long‐term leaching losses. Annual, rather than biennial, fertilizer applications are to be preferred.

Potassium release from saturated spherical aggregates of three clay soils to Ca-resin was measured and compared with release in suspensions (Talibudeen et al., 1978). The diffusion coefficient for K was calculated from measurements of the tortuosity factor using 36Cl, volumetric water content and buffer power. Release of K is in the order: suspensions > 2-3 mm > &8 mm aggregates. A computer model incorporating diffusion analysis and release of K shows that in the aggregates the rapidly diffusing exchangeable K, K,,, is best measured by the extrapolated intercept of a desorption isotherm (24 h equilibration), and is 8-S5% less than that extracted by NH,OAc leaching. Release of K within the aggregates (K,,,, a fraction approximating to Talibudeen's peripheral K) is delayed until after about 95% of K,, has been depleted which takes 1-2 days. The effective release rate of K,,, inside the aggregates is -17% less than in soil suspensions.Diffusion imposes only small limitations on the release from aggregates of K,,,. This rate of release depends on the frequency of resin changes.

The behaviour of P in a range of English arable soils was examined by plotting the change in resin P in the topsoil (ΔPres) at the end of a 3‐ to 5‐year period, against the P balance over the same period (fertilizer P applied minus offtake in crops, estimated from farmers’ reported yields and straw removal). Based on the assumption that values for offtake per tonne of crop yield used for UK arable crops are valid averages, 20–60% of ΔPres was explained by the balance. Applying excess P fertilizer increased Pres, and reducing P fertilizer use decreased it; typically 3–4 kg P ha−1 was required for each mg L−1ΔPres (6–8 kg ha−1 for each mg L−1 of Olsen P). About half the P balance seems to be resin extractable and this differed little between soil groups, except in cases of very low P (index 0) in which the P buffering was stronger, and on very high P soils (index 4/5) when buffering was less. However, on calcareous soils and red soils, when fertilizer was applied in accord with offtake, Pres fell by up to 4 mg L−1 year−1 (2 mg L−1 yr−1 olsen P) and to prevent this an extra 3–10 kg P ha−1 year−1 fertilizer was required. But on most non‐calcareous soils, replacing offtake maintained Pres, with perhaps slight rises on soils of low clay content or greater organic matter content. In soils under arable rotations, the apparent recovery of P from fertilizer was often around 100%, falling to 85% on Chalk soils and 75% on medium–heavy soils on limestone or Lower Chalk. The fate of the ‘missing’ P needs clarification. The case for corrections to current P fertilizer recommendations in the UK on certain soil types is discussed.

The behaviour of potassium (K) in a range of arable soils was examined by plotting the change in exchangeable K of the topsoil (D K ex ) at the end of a 3-5 year period against the K balance over the same period (fertilizer K applied minus offtake in crops, estimated from farmers' records of yield and straw removal). Based on the assumption that values for offtake per tonne of crop yield used for UK arable crops (MAFF 2000) are valid averages, 10 -50% of D K ex was explained by the balance, relationships being stronger on shallow/stony soils. Excess fertilizer tended to increase K ex and reduced fertilization decreased it, requiring between 1.2 and 5.4 kg K ha 21 for each mg L 21 D K ex . However, merely to prevent K ex falling required an extra 20 kg K ha 21 yr 21 fertilizer on Chalk soils and soils formed in the overlying Tertiary and Quaternary deposits, despite clay contents . 18%. Whereas, on older geological materials, medium soils needed no extra K and clays gained 17 kg K ha 21 yr 21 . It is unlikely that the apparent losses on some soil types are anomalies due to greater crop K contents. Theory and the literature suggest leaching from the topsoil as a major factor; accumulation in the subsoil was not measured. Recommendations for K fertilization of UK soils might be improved by including loss or gain corrections for certain soil types.