INFLUENCE OF SOIL PROPERTIES ON THE RESPONSE TO PHOSPHORUS IN SOME TROPICAL SOILS .2. INITIAL RESPONSE TO FERTILIZER

The availability of fertilizer P in six P-deficient tropical soils from Brazil, Kenya, Malaysia and Indonesia was assessed by grass in a pot experiment. Grass dry matter yield (D) and fertilizer P (F) were fitted to a Mitscherlich equation: D = a - b exp(-cF), and P uptake (U) and F to the linear equation: U = alpha + betaF. Fitted parameter beta equals the proportion of P recovered in one crop and it varied widely between soils, ranging from 12 to 51%. Quantitative assessments of fertilizer-P availability could also be made using dry-matter data alone if the rates of fertilizer used were well distributed along the response curve, when Mitscherlich parameter c was correlated well with beta. Chemical measurements were made on uncropped soil. Phosphate sorption isotherms were measured, using P-32 to assess exchangeable and non-exchangeable phosphate. The availability parameters c and beta were correlated best with parameter b(e), the phosphate buffer capacity derived from the fitted Freundlich isotherm for exchangeable phosphate, suggesting that the mobility of exchangeable phosphate is a major influence on P availability. Al and Fe were extracted with acid oxalate, citrate-dithionite and pyrophosphate reagents, and parameters c and beta correlated best with Al extracted by acid oxalate. These relationships were inverse, showing that Al in disordered mineral forms lowers the availability of fertilizer P.