OMPETITION stress is exerted on a plant by the spatial arrangement and phenotype of surrounding plants. The extent to which competition stress influences a plant character is an important consideration in selecting for that character in a breeding program. If a predetermined plant spacing can be identified at which competition stress has a negligible effect on genotypes, relative to their performance in drilled rows, effective selection may be practiced on an individual plant basis with considerable economy in time and field space. Plant spacings much wider than commercial planting rates would be desirable to facilitate plant measurements and identification and to provide adequate planting seed for progeny testing. Heritability estimates summarized by Johnson (9) indicate that spaced F~ plants are frequently poor indicators of their genetic potential for yield, are better indicators of their genetic potential for chemical composition, and are very reliable indicators of some agronomic characters. Differences in competition stress are undoubtedly one reason for the failure of spaced plants to indicate the relative performance of their less heritable characters at commercial planting rates. A differential response of genotypes to spacings was obtained for yield (~0, 11) and number branches (10)at within-row spacings likely to be encountered in commercial production. Border-row competition resulted in significant yield differences (5, 6) but had appreci.able effect on chemical composition of seed. Differences m response to photoperiod are likely to be very important since plant development is controlled to a considerable extent by the response of a plant to photoperiod (i, 2, 4). Limitations imposed by the necessity of small plot size for individual plants are obvious but have no satisfactory solution. This paper presents the results of an experiment conducted at Gainesville, Florida, in 1955 and 1956 to study the effect of competition on yield, chemical composition of seed, and agronomic characters of soybeans and to determine if a within-row spacing could be identified at which the effects were negligible.