PATH-ANALYSIS OF RICE YIELD AND YIELD COMPONENTS AS AFFECTED BY SEEDING RATE

Understanding relationships among rice (Oryza sativa L.) yield and yield components (panicle density, filled grain per panicle, unfilled grain per panicle, and grain weight) is critical to utilizing these relationships effectively. Interrelationships among rice yield and yield components for direct-seeded rice cultural systems using path-coefficient analysis that describes a priori cause-and-effect relationships have not been shown. Our objective was to study direct-seeded rice yield and yield component relationships as affected by seeding rates using path-coefficient analysis. 'Lemont', a semidwarf rice, was seeded at 43, 85, 128, 170, and 213 kg ha-1 in four environments between 1986 and 1988. The seeding rate by environment interaction was nonsignificant for all traits, indicating stable performance across environments. As seeding rates increased, panicle density increased significantly, and filled grain per panicle decreased significantly, an indication of the compensatory nature between these two traits. Higher panicle densities at the lower seeding rates were due in part to increased tillering of the rice plants. Path-coefficient analysis revealed that panicle density had the largest positive direct effect on rice yield, even at low seeding rates where filled grain per panicle increased to compensate for decreased panicle density. Direct effects for filled grain per panicle and grain weight were of secondary and/or tertiary importance in determining rice yield. Unfilled grain per panicle had negligible effects on rice yield. To achieve optimum rice yields and grain quality in a direct-seeded cultural system, adequate panicle density per unit area of uniform maturity must be achieved.