At 5 to 8 months of age the fighting and tail-rattling of male Marsh mice during paired confrontations was significantly greater for isolated than for mice prehoused in groups of males or in groups of one male with females. The differences were not significant for treatment effects tested in a similar manner with mice 11 to 12 months of age. At 5 months of age, mounting behavior of pregrouped males on paired confrontation was significantly less than that of isolated males. Paper towels instead of pine shavings for bedding abolished test-fighting for isolated mice. In any model with randomly sampled groups, an inherently highly aggressive mouse (killer) required to be eliminated from the grouped series, is not detected in the isolated series. These considerations challenge the use of isolated male mice as a pharmacologic model for evaluating behavior-modifying drugs. A four-fold difference in the area of prehousing cages of isolated males did not influence aggressive behavior. Fights correlated with tail-rattling; r equals 0.93 over the 14 series studied.