The Contribution of Variance in Lifetime Reproduction to Effective Population Size in Tigers

Abstract

Abstract: We studied tiger (Panthera tigris) reproduction in Royal Chitwan National Park to examine variance in lifetime reproduction and its contribution to estimates of inbreeding‐effective population size in this species. A total of 36 breeding animals (14 males, 22 females) and 144 young were monitored between 1973 and 1989. Mean litter size was 2.98 and young were born throughout the year. Mortality during the first year was 34%. Loss of entire litters accounted for 73% of first‐year mortality, suggesting that chance events were more important in determining first‐year survival than the quality of a female's territory. Second‐year mortality was 17%, and most (71%) was within‐litter mortality. The number of young that survived to dispersal or 2 years was used as a measure of lifetime reproduction. To calculate the effect of variance in lifetime reproduction on inbreeding‐effective population size, we estimated the distribution of number of offspring that reach breeding age. High variance in lifetime reproduction resulted in an N e ‐to‐N ratio of 0.41 and an N e , of 26 Because the rate of inbreeding in the Chitwan population is approximately 2% per generation and Chitwan is one of the largest tiger populations on the Indian subcontinent, it is possible that inbreeding depression occurs in many tiger populations. Inbreeding depression may be reflected in many measures of fitness: litter size, birth weight, infant and early survival, and sperm viability. The field data presented here demonstrate the length of time and effort needed to obtain even a single moderate‐sized data set. Were it possible to obtain two such data sets, comparisons between populations or time periods would‐be problematic Because of these difficulties, conservation decision‐makers will need to rely on population viability analysis modeling. Long‐term field data on individual populations and information on the dynamics of metapopulations are needed to build and validate these models.