Assessment of fertilizer P residues in a calcareous vertisol

The amounts of available P in a Typic Pellustert containing different levels of residual fertilizer P were assessed in a field experiment and chemically with cation-anion-exchange resin, using: (i) P desorbed after 1 extraction, (ii) cumulative P from 8 extractions [Resin8-P] and (iii) the asymptote of the extraction curve. P exchangeable to 32p, Olsen-P and P extracted by dilute salt solution (0.03 M KC1) were also measured. P available to sorghum in 1989 was strongly influenced by fertilizer P applied in 1988, but not by P applied in 1987. Olsen-P and Resin8-P both reflected well the effects of residual P on yield and P uptake and could therefore be used to predict available residual P. The amounts of Resin8-P were much larger than Olsen-P, so it was easier to observe the differences between soil treatments, but Olsen-P gave the better correlations with grain yield and P uptake. Multiple regressions of crop yield or P uptake with amounts of P applied in 1987 and 1988 gave the best correlations, without the need for practical work. Thus, the development of an appropriate model based on records of P fertilization is seen as the most effective way to predict availability of residual P. Assessments related to the P intensity and desorption rate using resin were not able to improve predictions of available P, indicating that within one soil the quantity of residual P is most important, but for comparisons between soils, account will need to be taken of differences in soil characteristic