Search for a command to run...
Abstract Reptiles, and turtles in particular, are disproportionately represented in the illegal wildlife trade. Chemical analysis of animal tissues is a powerful tool to combat wildlife laundering, which is the act of disguising animals poached from the wild as legally raised in captivity by means of counterfeit import or export documentation. We used stable isotope ratios and trace element concentrations derived from the claw tips of 449 wood ( Glyptemys insculpta ), spotted ( Clemmys guttata ), and Blanding's turtles ( Emydoidea blandingii ) from the eastern United States to develop a multispecies statistical model for determining the probability that a confiscated turtle was poached from the wild. During model development, our simple model that used 4 stable isotope ratios (δ 13 C, δ 15 N, δ 18 O, δ 2 H) and 4 trace elements (Ti‐47, Fe‐57, Cu‐63, Ba‐137) misclassified only 2 turtles out of 449, yielding a predictive accuracy rate of 99.55%. We further validated the predictive accuracy of our multispecies model by calculating the probability of being wild for 9 eastern box turtles ( Terrapene carolina ) and a replicate sample of 5 wood turtles, all with known classifications of being wild or captive. We now aim to promote our model to aid conservation law enforcement in combating the illegal turtle trade by helping to close the wildlife laundering loophole. We also hope to provide a forensic framework for developing a conservation tool for other taxa of conservation interest.