Travis Ingram reported on a new method he devised to test whether the anole radiations on the Greater Antilles are more similar than might be expected to occur by chance. We all know that each island has experienced its own radiation, producing more or less the same set of ecomorphs. However, some islands have more ecomorphs than others (Jamaica: 4; Cuba, Hispaniola: 6). In addition, there are non-ecomorph species on the larger islands. It is always possible that it is just a coincidence that the same types have evolved on multiple islands. After all, given large enough evolutionary radiations, one would expect the same morphology to evolve by chance on multiple islands. Travis developed a method to test this hypothesis, and found that, indeed, the Greater Antillean radiations are more similar in morphology than would be expected by random evolutionary change. Read all about it in the abstract:
Ingram, Travis (Harvard University); Mahler, D. Luke (University of California: Davis, Canada); Revell, Liam J. (University of Massachusetts – Boston, Canada); Losos, Jonathan B. (Harvard University, Canada)
A new comparative method that does not require pre-assigned ecomorph categories confirms exceptional morphological convergence in Caribbean anoles.
Convergent evolution is a powerful line of evidence for the efficacy of natural selection, and the replicated evolution of habitat specialist ecomorphs in Anolis lizards is among our best-known examples of convergence. Most statistical tests for convergence suffer from a potential circularity: species are assigned to categories (such as ecomorphs) based of their appearance or ecology, then methods are used to determine whether ecomorphs cluster together in phenotype space. Here, we introduce a new comparative method that identifies cases of convergence using a phylogenetic tree and multivariate trait dataset, without any a priori classification of species to ecomorphs. The method, called SURFACE, first uses stepwise AIC to fit increasingly complex models assigning regions of the tree to distinct stabilizing selection regimes. The second stage of the method assigns independent regions of the tree to shared selective regimes until the model fit ceases to improve further. We apply this method to ecomorphological data from 100 Greater Antillean Anolis species. SURFACE settles on a model with a large number of evolutionary regimes, close to half of which are convergent with regimes in other parts of the tree. Comparison with data simulated under a variety of null models indicates that the method is unlikely to recover this degree of convergence by chance. Many, but not all, of the cases of convergence identified by SURFACE correspond to classically recognized ecomorph designations. This finding supports the long-held view that natural selection has driven exceptional evolutionary convergence in the adaptive radiation of anoles.