Use of the subspecies as the basic unit in the conservation of endangered caribou (Rangifer tarandus) would produce a “melting pot” end-product that would mask important genotypic, phenotypic, ecological, and behavioral variations found below the level of the subspecies. Therefore, we examined options for establishing the basic conservation unit for an endangered caribou population: use of subspecies based on taxonomy, subspecies based solely on mtDNA, Evolutionarily Significant Units, and the geographic population. We reject the first three and conclude that the only feasible basic unit for biologically and ecologically sound conservation of endangered caribou in North America is the geographic population. Conservation of endangered caribou at the level of the geographic population is necessary to identify and maintain current biodiversity. As deliberations about endangered caribou conservation often involve consideration of population augmentation, we also discuss the appropriate augmentation protocol for conserving biodiversity. Management of a critically endangered caribou population by augmentation should only be initiated after adequate study and evaluation of the genotype, phenotype, ecology, and behavior for both the endangered caribou and the potential‘donor’ caribou to prevent the possible ‘contamination’ of the endangered caribou. Translocation of caribou into an endangered population will have failed, even if the restocking efforts succeed, if the donor animals functionally alter the population’s gene pool or phenotype, or alter the ecological and behavioral adaptations of individuals in the endangered population. Most importantly, a seriously flawed restocking would risk irreversibly altering those functional characteristics of caribou in an endangered population that make them distinct and possibly unique. It might even result in the loss of the endangered population, thus eliminating a uniquely evolved line from among the caribou species.