We tested the hypothesis that small, isolated populations would show less depression in fitness when inbred than would large, central populations. Laboratory stocks of Peromyscus leucopus and P. polionotus were established from insular, peninsular, and central populations. The isolated populations had one-third to one-half the genic diversity of central populations. Responses to inbreeding were highly varied: some populations had smaller litters, others experienced higher mortality, some showed slower growth rates, and one displayed no measurable effects when inbred. These results suggest that inbreeding depression is controlled by a small number of genes and that the size of the genetic load depends on which alleles are present in the founders of a population. The severity of fitness depression in inbred litters did not correlate with initial genic diversity of the stocks nor, therefore, with the size of the wild populations. Fitness measures appeared linearly related to the inbreeding coefficient of the liters, with no diminution of deleterious effects through subsequent generations of inbreeding. Thus overdominance of fitness traits probably contributed as much to the genetic load as did deleterious recessive alleles. The inbreeding level of the dam negatively affected the size, growth, and survival of litters only in genetically diverse populations, indicating that the load of recessive alleles negatively impacting maternal care may have been reduced by selection in the more peripheral populations during past bottlenecks.