Analyzing the past to understand the future: Natural mating yields better reproductive rates than artificial insemination in the giant panda

Conservation breeding programs require active management and thus selection among various management alternatives is a common practice. As in the practice of adaptive management used in ecology, it is important to reduce uncertainty about the outcomes of various management actions. Ideally this evaluation will be done using a priori hypothesis testing, but retrospective analyses can provide important insights as to which methods work better than others. The giant panda (Ailuropoda melanoleuca) conservation breeding program has a long history of active management and therefore is rich in potential lessons learned for panda and other endangered species breeding programs. Now self-sustaining and experiencing exponential growth, the panda population also provides sample sizes large enough to support rigorous statistical evaluation. A fundamental decision for any breeding program is whether to invest in development and application of assisted reproduction techniques or to promote natural mating. Here we analyze 21 years (1996-2016) of giant panda reproductive data from 304 insemination events to determine relative success rates of insemination methods and evaluate management strategies. The birth rate after natural mating was 60.7%, 50.6% for combined natural mating and artificial insemination techniques, and 18.5% for artificial insemination (AI). Within the combined insemination technique group, 81.8% of births could be attributed to the natural mating event with only 18.2% attributed to AI. These results suggest that while techniques for improving AI should continue to be explored and will play an important role for some conservation applications, behavioral and biological management to encourage natural mating should be the most important goal for conservation breeding of this species.