A field experiment was used to investigate the effects, if any, of changing group size and racial composition on the attitudes, behaviors, and effectiveness of black and white leaders. Subjects were 288 naval recruits, half black and half white, performing two tasks which were watched by a pair of racially mixed observers through a one-way mirror. This study found that white leaders achieved better performance from their groups than black leaders did; white-led groups of all types performed quicker. This contrasts sharply with laboratory study which showed no differences in the performance of all white, all black, and mixed dyads led by black and white student leaders. Some explanations are offered for the differences, but they remain to be tested and validated. (NTIS/AG)