Prediction of phosphate requirements of Swedish turnips from soil phosphate values

Yields from 199 field experiments, extending over about 15 seasons and covering six soil groups in northeast Scotland, were used to calculate correlation coefficients between the responses of Swedish turnips to 120 1b P 2 O 5 /acre and readily soluble phosphate values for the soils, determined by six methods. There were marked effects of soil parent material and drainage on the correlations, which ranged from −0.05 to −0.74. the Morgan and bicarbonate methods were inferior to the NH 4 F‐HCl, Truog, acetic and lactate methods; the last method was usually the best. By fitting a Mitscherlich equation to the yields the optimum economic phosphate dressings were calculated for four combinations of fertiliser and crop prices. the average profit from these was about 30 shillings/acre greater than from using the mean optimum rate for the soil group or from using a standard rate of 120 1b P 2 O 5 . Calculations and adjustments tested so far have shown no convincing benefit from using the soil values to recommend phosphate dressings for individual soils. Neither inclusion of other soil properties nor use of the logarithms of the soil phosphate values in multiple regressions produced any appreciable improvement in response predictions. The main reason for the generally rather poor correlations seems to be variations in site characteristics and agricultural factors, rather than soil chemical and textural properties, coupled with the fact that the optimum rates are not well defined. the present indications are that the main practical usefulness of the soil P values is in monitoring of the phosphate status to guide rotational manuring and ensure that adequate dressings can be given for sensitive crops, rather than in the precise prediction of optimum dressings for individual fields.