Soil properties as predictors of yield response of clover (Trifolium subterraneum L.) to added P in soils of varying P sorption capacity

Abstract

Thirty-five unfertilised soils collected in south-western Australia were used to measure the effect of soil properties on (i) shoot yield responses of 50-day-old clover (Trifolium subterraneum L. cv. Nungarin) plants to applied phosphorus (P), and (ii) extractability of bicarbonate soil test P (slope of the linear relationship between Colwell P and the amount of P applied). Data for the relationship between shoot yield and the amount of P applied were fitted to a rescaled Mitscherlich equation to calculate the amount of P required to produce 50% and 90% of the maximum yield (P50% and P90%) and determine the curvature (c) and n coefficients of the equation. When the value of n is 1.00, the response curve is exponential, and as the value of n increases above 1.00 the response curve becomes more sigmoidal. The c, n, P50%, P90%, and extractability values were related to properties of the 35 soils.There was a significant (P < 0.05) trend for the values of c and extractability to decrease as the capacity of the soil to sorb P increased. Consequently, as the soil sorbed more P, the trend was that (1) more P needed to be applied to produce the same yield, so both P50% and P90% tended to significantly (P < 0.05) increase; (2) shoot yield responses to applied P became more sigmoidal so the value of the n coefficient tended to significantly (P < 0.05) increase; (3) more P needed to be applied to a soil to produce the same soil test P value; and (4) larger soil test P values were needed to produce the same yield. No single soil property adequately predicted P50%, P90%, extractability, c, or n. Stepwise multiple regression indicated that (1) clay content and P buffer capacity (PBC) of soil together accounted for 48% of the variation in P50%, 56% of the variation in P90%, and 52% of the variation in c; (2) PBC and soil pH together accounted for 17% of the variation in n; and (3) PBC, percentage clay and percentage organic carbon content of soil together accounted for 68% of the variation in extractability.