Influence of intensity/quantity characteristics of soil phosphorus tests on their relationships to phosphorus responsiveness of wheat under field conditions

Abstract

Six soil phosphorus tests (lactate, Brayl, Bray2, neutral fluoride, Olsen and Colwell) were regressed against potassium chloride-soluble phosphorus (intensity) and isotopically exchangeable phosphorus (quantity) measured in 59 soils of the northern and central wheat belts of New South Wales. Wheat nutrition experiments on these soils during 1986-89 measured yield responses to phosphate and nitrogen fertilizers. Soil tests varied widely in their correlations with yield responsiveness to phosphate, with the lactate and Bray2 tests accounting for more than twice the variance accounted for by other soil tests. The intensity parameter was also highly correlated but the quantity parameter was not. All soil tests, except Bray1, were very highly correlated with the intensity parameter, so this relationship did not differentiate the relative efficacies of the soil tests. Soil tests were less correlated with the quantity parameter, but those soil tests (neutral fluoride, Olsen and Colwell) that were most highly correlated (r2 > 0.62) with this parameter were most weakly correlated (r2 < 0.29) with yield response. It was concluded therefore that exchangeable phosphorus is not a satisfactory measure of the quantity factor and that an effective soil test for wheat-growing soils will be highly correlated with intensity but not necessarily with exchangeable phosphorus. The critical value of the lactate test was the same (17 mg/kg) as in previous studies with wheat but was lower (14 mg/kg) in 1989 when very low in-crop rainfall occurred. With deeper sampling (15 cm rather than 10 cm) the lactate test was slightly less accurate and the critical value was lower (11 mg/kg).