Understanding patterns of species coexistence is a fundamental topic for community ecology. Ecologists have hypothesized that a competitively superior colonizing species will lead to the displacement of any closely related co-occurring species. Long-term population data indicate that lakes in northern Wisconsin and Michigan can contain either native virile crayfish (Orconectes virilis) or invasive rusty crayfish (O. rusticus) or both. We conducted observations and experiments to determine the ecological factors that allow coexistence of these closely related species. We hypothesized that within-lake habitat heterogeneity and differential survival according to habitat type are important determinants of species coexistence. To test this hypothesis, we analyzed observational habitat association data for both species and conducted experiments to quantify relative predation pressure on the two species in two habitat types.
Results/Conclusions

Regarding habitat associations, we found rusty-only lakes have more littoral cobble habitat (51% vs. 30%, p=0.008) and less vegetated habitat (11% vs. 41%, p=0.005) than coexistence lakes. Moreover, in coexistence lakes, virile crayfish are found more commonly in vegetation and less commonly in cobble than would be expected by random chance alone (p<0.001). Conversely, rusty crayfish are found less commonly in vegetation and more commonly in cobble than expected (p<0.001). In addition, there are fewer rusty crayfish found in vegetated habitats compared to cobble habitats. These findings suggest reduced predator-mediated competition between the two species in vegetation. To investigate the role of predation in determining coexistence, we used a tethering experiment to measure inter-specific differences in survival in the presence and absence of fish predators in both sand and vegetated habitats. Without predation there were no differences in survival between either crayfish species in either habitat (crayfish p=0.74, habitat p=0.07). When predators were present, mortality from predation for both species was lower in vegetated than sand habitats, and rusty crayfish experienced lower predation rates than virile crayfish in both habitats (crayfish p=0.045, habitat p<0.005). Nevertheless, measured against rusty mortality, predation rates on virile crayfish was1.3 times that of rusty crayfish in vegetation and 8 times that of rusty crayfish in sand habitats. Combining observations on habitat association with results of predation experiments, we conclude that while both crayfish species prefer cobble, vegetated habitats are characterized by reduced interspecific difference in predation pressure (i.e., virile are less disadvantaged). Thus the less desired vegetated habitat provides a refuge for the native virile crayfish.