Comparison of three soil tests for phosphate on lateritic soil fertilized with superphosphate, crandallite rock phosphates and apatite rock phosphates

Abstract

Different amounts of superphosphate, crandallite-millisite rock phosphate (three different samples of C-ore from Christmas Island), 500�C calcined C-ore (three different samples of Calciphos) and apatite rock phosphate (two different samples from the Duchess deposit, Queensland) were applied once only in May 1977 on lateritic soil in Western Australia. Samples of the top 10 cm of soil were collected in the summer (January-February) of 1979 and 1980 for measurement of extractable soil phosphate by the Colwell, Olsen and Bray soil tests, and the data were compared with dry matter production of subterranean clover pasture in the following spring (August). For each fertilizer in each year, the Olsen and Brayl soil test values (as dependent variables) were closely correlated with the Colwell soil test values (independent variable), and the Bray 1 soil test (dependent variable) was closely correlated with the Olsen soil test (independent variable). For the relationship between yield and soil test values, separate calibrations were required in most cases for each fertilizer and each soil test, and these calibrations showed large differences in two successive years. None of the soil tests was significantly better for predicting clover yields for the different fertilizers applied two and three years previously.