Tailoring release protocols to individual species and sites: one size does not fit all.

Reintroduction programs for threatened species often include elaborate release strategies designed to improve success, but their advantages are rarely tested scientifically. We used a set of four experiments to demonstrate that the influence of release strategies on short-term reintroduction outcomes is related to both intrinsic and extrinsic factors. We compared different reintroduction strategies for three mammal species in an arid environment where exotic mammalian predators were removed. Wild greater stick-nest rats selected vegetation shelter sites with greater structural density than captive-bred rats, travelled further from the release site and experienced lower rates of mortality. In comparison, there was no difference in mortality or movement between wild and captive-bred greater bilbies. Burrowing bettongs and greater bilbies were also subjected to immediate and delayed release strategies and whilst no difference was detected in bilbies, bettongs that were subjected to delayed releases lost less weight and took less time to establish burrows than those that were immediately released. Interspecific differences in treatment response were attributed to predation risk, the nature of the release site, and behavioural traits such as shelter investment and sociality. Our varied results highlight the inadequacies of review articles focusing on optimum release protocols due to their attempt to generalise across species and release sites. We provide an example of a predictive model to guide future release strategy experimentation that recognises the range of intrinsic and extrinsic factors influencing reintroduction outcomes. We encourage researchers to treat programs experimentally, identify individual site and species characters that may influence release strategies and record data on movements, mortality, weight dynamics, and settling times and distances. The inherent issues of small sample size and low statistical power that plague most reintroduction experiments suggests there is also a need for increased standardisation and publication of data sets to enable appropriate meta-analyses to occur.