Abstract

The effects of plant density on yield and yield components in upland rice cultivation were examined by conducting a series of field experiments in Central Uganda, using three African and one Japanese improved upland rice varieties. The estimation of plant-density response functions with respect to yield components and yield revealed that an increase in plant density significantly decreased the number of panicles per hill, number of spikelets per panicle, and 1000-grain weight, and significantly increased the number of panicles per square meter. The percentage of filled grain was not affected by plant density. Compared to the Japanese variety, the three African varieties were characterized by more numbers of panicles/hill, less numbers of spikelets/panicle, higher grain-filling ratio and lighter 1000-grain weight, but differences in the degrees of response to plant density were less distinct between them. Rice yield increased in the range of plant density tested, though the marginal increase in yield due to an increase in plant density by 1 hill/m2 diminished from 100 kg/ha at the plant density of 11 hills/m2 to 30 kg/ha at 33 hills/m2. No significant differences were found among the four varieties for the level of yield as well as for its degree of positive response to plant density. The yield components that determined the increase in yield were the number of panicles per square meter and the number of spikelets per panicle, or combined together, the number of spikelets per square meter, which was estimated to reach the maximum at the plant density of 35 hills/m2. When the differences among the treatments in the costs of seeds and weed-control were considered, the optimum plant density was found to be 22 hills/m2 (plant-spacing of 30 cm × 15 cm), lower than the plant density that gives the maximum yield.      

Key words: Economic optimum, maximum yield, NERICA, plant-density response function, yield components.