This study attempted to: (1) replicate the density effects found in a previous study; and (2) test the hypothesis that partitioning mediates the effects of high density on children's behavior. Twelve children (5 female, 7 male, ranging in age from 40 to 68 months) were observed under each of four conditions that crossed high and low density with partitioning and no partitioning. These factors were crossed with sex in a 2 x 2 x 2 factorial design. A rating instrument was used to record social interaction, the individual's relationship with the environment, and affect. Results indicate that the higher level of density decreased associative and cooperative behavior. Partitioning was shown to mediate the relationship between density and constructive behavior while independent effects on aggressive and cooperative behavior were also found. Implications of the results for child development are discussed. (Author/SB)