AGRONOMIC COMPARISONS BETWEEN PHOSPHATE ROCK AND SUPERPHOSPHATE IN SOILS OF THE CERRADO OF BRAZIL

The potential utilization of Patos de Minas phosphate rock (PR) for direct application to Cerrado soils was evaluated in field and laboratory experiments. In field studies, the effects of broadcast PR or ordinary superphosphate (OSP), with or without band-applied OSP, on the growth and yield of soybeans were studied on a Typic Acrustox in the Cerrado of Brazil during two consecutive crops. In laboratory studies, the dissolution patterns of North Carolina PR and Patos PR were compared in seven Cerrado soils during a 158-day incubation period.; Soybean yield maxima for two consecutive crops were higher with broadcast OSP than with broadcast PR. The low crop growth rates (CGR) and P accumulation rates (PAR) for plants receiving Patos PR indicated that this low reactivity source was unable to supply adequate P for optimal soybean growth. This observation was confirmed by the small increases in available soil P values. Estimates of critical P levels extracted by the Bray 1, Olsen, and anion resin methods were 11-15, 6-8, and 8-12 ppm P, respectively. The relationships between yield response and soil-test P values were identical for both sources of broadcast P. Although Bray 1 P levels at pH 4.4 were almost double the levels obtained at pH 4.8, increases in grain yield were not observed at this low pH, presumably because of Al toxicity.; Increasing the levels of available soil P increased the CGR and PAR throughout the growing season. Differences in CGR among the broadcast P treatments were evident at an earlier stage of growth with OSP than with PR.; Preplant band-applied OSP combined with either source of broadcast P improved initial growth and maintained CGR and PAR values superior to broadcast P alone during the first 70 days after planting. The band application of 44 kg P/ha combined with 176 kg P/ha of broadcast OSP provided yields equivalent to the maximum achieved with 352 kg P/ha as broadcast OSP, providing a reduction of 88 kg P/ha in the total P fertilizer application for two crops and decreased P costs by 25%. Yield increases due to band-applied P decreased with increasing soil-test P values.; The dissolution rate for Patos PR was approximately linear during the 158 days of incubation, as opposed to an exponential pattern for the North Carolina PR, indicative of an initially rapid release of P during the first 43 days of incubation. Inherent differences in the PR sources, namely the degree of isomorphous substitution of CO(,3) for PO(,4) in the apatite structure, were associated with these differing patterns of dissolution. Changes in Olsen-extractable P were found to be more indicative of the PR dissolution process than P extracted by an anion resin procedure. Maximum dissolution of the PR, when measured by increases in exchangeable Ca, occurred in soils with the highest P sorption capacity. The "Ci" modifier of the Fertility Capability Classification system was an effective criterion for separating soils with high and low levels of PR dissolution. Increases in exchangeable soil Ca during incubation ((DELTA)Ca) was a useful parameter for describing PR dissolution patterns.