A systematic program of genetic improvement was initiated by assessing the phenotypic variation of Steinernema feltiae strains for two traits assumed to limit efficacy: ultraviolet tolerance and host-finding ability. All of the strains assayed showed both low ultraviolet tolerance and poor host-finding ability, indicating that the likelihood of improving these traits through more extensive population sampling is remote. Limited genetic variation was detected among the strains for tolerance to ultraviolet, suggesting that selective breeding for increased tolerance would be inefficient. By contrast, highly significant phenotypic differences were found with regard to host-finding ability, suggesting that this trait would be responsive to selection. A genetically heterogeneous population was constructed by round-robin mating of 10 strains; it will serve as the foundation population for selective breeding.