Direct and Indirect Mass Selection for Grain Yield in Bulk Oat Populations 1

One cycle of direct mass selection was conducted separately for heading date, plant height, grain yield, width/seed, weight/seed, and spikelets/panicle of oats ( Avena sativa L.). One cycle of indirect mass selection for grain yield was also conducted in each of the remaining five traits and weight/primary panicle. Selection was practiced among 1,500 random, nearly‐homozygous lines grown at two levels of interplant competition, and was bidirectional with selection intensities corresponding to 10 and 30%. Variations of basic hill‐plot technique simulated the two selection environments, noncompetitive (one oat seed sown/hill) and competitive (1 oat seed plus 30 barley seeds sown/hill). Hill plots sown with 30 oat seeds/hill simulated planting arrangements common to commercial practice and provided a standard by which gains from mass selection were determined. Direct mass selection was effective for each trait. Actual gains ranged from 15 to 83% of gains predicted from selection among line means in replicated trials. Selection for heavy primary panicles was the most effective method of indirect mass selection for grain yield. In general, the level of interplant competition had little effect on gains from mass selection. Considerations of technique and monetary costs seem of greater importance than interplant competition in the choice of an environment for mass selection in oats. Environmental effects on genotypic expression of grain yield were not sufficiently large to prevent improvement of this trait by mass selection at either level of interplant competition.