抄録
Data on four yield components obtained from a long-term experiment, in which NERICA 4 was continuously planted under upland conditions for 17 seasons, were analyzed together with rainfall data. Results showed that rainfalls gave significant positive effects on the number of spikelets panicle-1 (Spikelets), the rate of grain filling (Gf), and 1000-grain weight (Gwt), and yield, but no significant effect on the number of panicles m-2 (Panicles). Of three equally-divided growth periods, the effect of rainfalls during the second period, including the meiosis-anthesis stage, was more pronounced. An increase in rainfall by 1 mm in this stage increased rice yield by 10 kg ha-1, with Gf as the largest contributor to this yield increase, followed by Spikelets and Gwt. The adverse effect of continuous mono cropping on yield was not large, 0.6%, or 14 kg ha-1, per an additional planting. This observed yield decline due to additional planting every season was brought about wholly by the decline in Gwt, and not by other yield components. The application of N-fertilizer mitigated the adverse effect of continuous cropping for Panicles and Spikelets but aggravated it for Gf and Gwt. The P-deficient soil at the experimental site made the effects of N-application very low if used without P-application. When applied with P-fertilizer, N-application significantly increased Panicles and Spikelets, but significantly decreased Gf and Gwt, resulting in the marginal effect of N on yield of 13 kg ha-1 per 1 kg N.
