Supplementary Information
The Shimodaira-Hasegawa test was used to ask whether the molecular phylogenetic data statistically reject topologies in which ecological similarity and phylogenetic proximity are congruent. We examined a series of phylogenetic topologies including those in which pairs of the most ecologically similar species were constrained to be sister taxa, as well as a topology maximizing ecological similarity of all 11 species; the latter tree was determined by constructing a UPGMA phenogram from a matrix of pairwise ecological distances to obtain the topology expected if species were grouped phylogenetically according to ecological similarity1.
Table S1. Statistical Tests of Phylogenetic Topologies.
Alternative Topology p*
1. Topology of all 11 species
based on ecological similarity < 0.001
2. A. allogus & A. mestrei 0.002
3. A. homolechis & A. vermiculatus < 0.001
4. A. angusticeps & A. barbatus 0.007
5. A. loysiana & A. luteogularis < 0.001
6. Two major groupings < 0.001
*Probability that the maximum-likelihood topology (Figure 1) and the alternative topology are equally compatible with the molecular phylogenetic data using the Shimodaira-Hasegawa test.
Comparison of the phylogenetic topology based on DNA data to topologies constrained in various ways to group ecologically similar species: (1) Topology based on ecological similarity as determined by UPGMA; (2-5) maximum-likelihood topologies constrained to group particular pairs of ecologically-similar species; pairs chosen were 4 of the 5 most ecologically similar pairs of species; we did not include the pair of A. mestrei and A. homolechis because A. mestrei is more ecologically similar to A. allogus; (6) maximum-likelihood tree constrained to the two major groupings of species as determined by UPGMA (group 1 contains A. allogus, A. homolechis, A. mestrei and A. vermiculatus; group 2 contains all other species except A. alutaceus).
1. Swofford, D. PAUP*: Phylogenetic analysis using parsimony (*and other methods). (Sinauer, Sunderland, MA., 2002).