Availability of P from 32P-labelled endogenous soil P and 33P-labelled fertilizer in an alkaline soil producing cotton in Australia

Yield responses of irrigated, field‐grown cotton to phosphorus fertilizer application in Australia have been variable. In an attempt to understand better this variability, the distribution of fertilizer P within soil P fractions was identified using ³²P and ³³P radioisotopes. The soil chosen, an alkaline, grey, cracking clay (Vertosol), was representative of those used for growing cotton in Australia. Chang and Jackson fractionation of soil P from samples collected within 1 h of application indicated that 49, 7 and 13% of the P fertilizer was present as 0.5 m NH₄F, 0.1 m NaOH and 1 m H₂SO₄ extractable P, respectively. Over 89% of the P fertilizer was recovered as Colwell extractable P in these samples, suggesting that the majority of these reaction products was in a highly plant‐available form. Fertilizer‐P remained in an available form within the band 51 days after application, and 68% of the applied fertilizer‐P was recovered as Colwell‐P (1071 mg kg^?1). The Colwell‐P concentration in the band was 35 times that in the unfertilized soil. Thus, the variability in crop response to P fertilizer application in these soils is not a consequence of fertilizer‐P becoming unavailable to plants. These results confirm the suitability of the Colwell (1963) sodium bicarbonate extraction method for measuring available P in these soils.