INPUT-OUTPUT-ANALYSIS OF THE CUMULATIVE SOYBEAN RESPONSE TO PHOSPHORUS ON AN ULTISOL

Although biological N2 fixation (BNF) by legumes can provide significant N inputs to crop systems on highly weathered tropical soils, potential inputs from BNF largely depend on soil P supply. We compared the cumulative effects of P input regimes on yield, N and P budgets, and soil P availability in four consecutive soybean crops in a 2-year period on a Humoxic tropohumult. In each crop cycle, nodulating (nod) and nonnodulating (nonnod) isolines were subplots in P-regime mainplots (kg P ha-1 by crop cycle): P0= control without P inputs; LP= 50, 35, 35, 35; MP= 100, 70, 70, 70; and HP= 300, 210, 210, 210. Seed yields of the nod isoline in the HP regime were 3700 kg ha-1 in the two summer seasons and 2400 to 2500 kg ha-1 in the fall seasons, with a mean increase of 85% compared to yields of the nod P0 control. Nonnod seed yields and N accumulation were unaffected by the P regime, averaging 870 and 48 kg ha-1, respectively. The contribution of BNF to nod soybean N assimilation was linearly related to P uptake, and mean P uptake by nod plants was 60% greater than by nonnod soybean, despite 35% greater root length of nonnod plants at 0-50 cm depth. For the four crop cycles, total BNF input to the system ranged from 330 kg N ha-1 (P0). or 65% of total aboveground N, to 710 kg N ha-1 (HP), which was 78% of total aboveground N. After accounting for P removal at harvest, a net P input of just 99 kg P ha-1 after four crops increased cumulative seed yield by 3600 kg ha-1 and BNF by 227 kg N ha-1 in the LP treatment. A positive net P balance also resulted in (1) an increase in extractable P in the 0 to 25 cm topsoil, (2) a reduction in the proportion of P that would be fixed from subsequent additions as indicated by a shifted P sorption isotherm, and (3) greater apparent P uptake efficiency from applied fertilizer in subsequent crop cycles. As a result, the yield response of nod soybean per unit P input increased from 13 kg seed kg-1 P applied in the first crop to 44 kg seed kg-1 P applied in the fourth crop of the LP treatment. Cumulative effects of this magnitude emphasize the need to consider the longer-term nutrient balance of the crop system in developing cost-effective P management strategies on highly weathered soils. The potential for greater P input use efficiency with time when inputs exceed outputs means that farmers' average and marginal return from investment in P fertilizer will also increase with time.