The effect of inter‐genotypic competition on yield andother attributes was evaluated in both hill and row plots utilizing four diverse adapted varieties. The experimentaldesign permitted the measurement of competition dosage effects as well as compensatory effects on adjoining genotypes. There was a drastic effect of competition in hills and rows on seed yield, seed number, and efficiency (expressed as a ratio of seed yield to straw yield). Certain genotypic combinations exhibited overcompensatory effects, a result which suggests that genotypic behavior in a competitive situation may be a reliable criterion for predictlng superior varietal blends. The data further suggest a nine‐hill field plot which should be effective in evaluating large populations for superior yield genotypes. The ninehill plot removes about 70% of the competition bias exhibited by an unbordered hill and would require much less land area and seed for testing than standard row plots.