The aim of this research was to optimize experimental planning for soybean crop in experiments with restrictions for grain yield, in order to reduce experimental plot size. The experiment was carried out at Santa Maria Federal University, in the agricultural year of 2000/2001, using a randomized block design, with three replicates, and two soybean cultivars (BRS137 and Fepagro RS10) used as treatments. Each replication was constituted by 18 rows of 8-m long and 44 cm of spacing between rows. Harvest was performed in basic units (ub) of 1 m of row (0.44 m²), totaling 144 ub. Plot size and true difference between treatment means were estimated by Hatheway method, with three different numbers of replications (4, 8 and 16) to verify, in the grain yield variable, which combination would minimize experimental area for soybean crop. The same plot size can be used for different soybean cultivars, keeping constant precision degree and number of replicates. Experiments can be optimized using larger number of replications and reducing plot size.
Glycine max L; limited experimental area; plot size and experimental precision
