Last June, I discussed a study with big ambitions: to experimentally compare the effects that competition and predators have on island populations of brown anoles, Anolis sagrei. Now the current issue of the journal that carried that study, Nature has a brief communication from the godfather of anole evolutionary ecology himself, Jonathan Losos. Losos and his coauthor Robert Pringle raise some serious questions [$a] about the results of that experiment.
The authors of the original study [$a], Ryan Calsbeek and Robert Cox, concluded that competition was more important than predation because natural selection acting on anoles was stronger on experimental islands with higher anole population density, while the presence or absence of predators on those islands made no difference in the strength of selection. Losos and Pringle object that anole population density is entangled with other factors that may make Calsbeek and Cox’s results uninterpretable.
This experimental design is confounded in three fundamental ways. First, density is confounded with island area. All analyses treat lizard density as a surrogate for intraspecific competition. However, an inverse correlation with island area explains 95% of the variation in density, such that it is impossible to disentangle the two factors statistically. This is a crucial problem, because multiple factors related to both predation and competition are known to vary with island area. For example, as island area increases, so too do the number of bird species (which increases the number of potential predators) and mean vegetation height (which might increase lizards’ susceptibility to avian predation). Likewise, because larger islands have lower perimeter/area ratios, they receive relatively lower input of marine-resource subsidies and have lower arthropod densities; a study of A. sagrei in this system showed that lizard densities vary significantly with the amount of seaweed deposition, and that experimental seaweed deposition increased lizard densities by more than 60%. [In-text citations removed for clarity.]
That point alone is a pretty big problem with Calsbeek and Cox’s result. Then Losos and Pringle re-analyze the data presented in the original study, and discover the very odd result that anoles in the experimental populations had higher rates of survivorship on the high-density islands—which is exactly the opposite of what you’d expect if competition for important resources were more intense in high-density populations. At the very least, this indicates that there could be more going on than Calsbeek and Cox originally supposed, in which case their data don’t support their conclusions.
Losos and Pringle raise other objections, including the issue of small sample size I noted in my original post. You should read the whole thing [$a] for the details, as well as the response [$a] from Calsbeek and Cox.
Calsbeek, R., & Cox, R. (2010). Experimentally assessing the relative importance of predation and competition as agents of selection. Nature, 465 (7298), 613-616 DOI: 10.1038/nature09020
Calsbeek, R., & Cox, R. (2011). Calsbeek & Cox reply. Nature, 475 (7355) DOI: 10.1038/nature10141
Losos, J., & Pringle, R. (2011). Competition, predation and natural selection in island lizards. Nature, 475 (7355) DOI: 10.1038/nature10140