All remaining subspecies of tigers (Panthera tigris) are endangered in the wild and space for captive breeding in western zoos is limited How should this space be allocated to enhance survival and retention of genetic variation in the face of uncertainty about genetic relationships among tiger subspecies and about the future of tigers in the wild? We solicited expert judgments on subspecies population trends and on the ratio of effective to census size in captivity and analyzed different scenarios for dividing the 1000 captive spaces among four tiger subspecies To maximize survival and retention of gene dimity in individual subspecies, it is best to divide the space equally among the four subspecies To maximize retention of gene diversity for the species as whole, it is better to allocate more space to the subspecies that appears most variable on the basis of limited electrophoretic data (p. t. tigris). Allocating half of the captive space to tigris and dividing the remainder equally among the other three subspecies is a satisfactory compromise between species welfare and subspecies welfare that ensures survival of all four subspecies (at least in captivity) and retains about 80 percent of existing gene diversity within subspecies and about 93 percent of diversity for the species as a whole. Sensitivity analyses showed that our recommendations were robust to uncertainty about the demographic and genetic status of tiger subspecies. For other species with several endangered subspecies; such as rhinos, resource limitations may be too severe to permit an easy compromise between species and subspecies welfare and acquiring better genetic data becomes critical. Copyright © 1990, Wiley Blackwell. All rights reserved