The objective of this work was to establish plant arrangement effect on the optimal plot size estimates of grain sorghum yield experiments. The experimental design was a completely randomized block with factorial combination of two row spacings (0.50 m and 0.80 m), and three plant densities (100 thousand, 160 thousand and 220 thousand plants ha-1), and four replications. The area of each replication was composed by 12 basic units measuring 0.50 m in row length. Models were adjusted in optimum plot size estimates that correlate variance or variation coefficient with four simulated plot sizes. Plot size was 3.2 m² for grain sorghum yield experiments. Increasing number of plants in row did not result in higher seed yields, however it improved sorghum quality experiment. Estimation of the ideal plot size depends on number of plants used in the basic unit. Row spacing did not affect estimation of the ideal plot size.
Sorghum bicolor; basic unit; experiment planning
