Body size evolution simultaneously creates and collapses species boundaries in a clade of scincid lizards

Abstract

Speciation is generally viewed as an irreversible process, although habitat alterations can erase reproductive barriers if divergence between ecologically differentiated species is recent. Reversed speciation might also occur if geographical contact is established between species that have evolved the same reproductive isolating barrier in parallel. Here, we demonstrate a loss of intrinsic reproductive isolation in a clade of scincid lizards as a result of parallel body size evolution, which has allowed for gene flow where large-bodied lineages are in secondary contact. An mtDNA phylogeny confirms the monophyly of the Plestiodon skiltonianus species complex, but rejects that of two size-differentiated ecomorphs. Mate compatibility experiments show that the high degree of body size divergence imposes a strong reproductive barrier between the two morphs; however, the strength of the barrier is greatly diminished between parallel-evolved forms. Since two large-bodied lineages are in geographical contact in the Sierra Nevada Mountains of California, we were also able to test for postzygotic isolation under natural conditions. Analyses of amplified fragment length polymorphisms show that extensive gene exchange is occurring across the contact zone, resulting in an overall pattern consistent with isolation by distance. These results provide evidence of reversed speciation between clades that diverged from a common ancestor more than 12Myr ago.

Figure 1

Relationships and distribution of clades of the P. skiltonianus complex in California. Numbers beside the branches are Bayesian posterior probabilities (asterisks indicate values ⩾0.98). For clarity, the Bayesian phylogram has been pruned to show only the relationships of the major geographical clades (methods for inferring the phylogeny are presented in the electronic supplementary material). Hatched lines indicate the limits of skiltonianus clades 1 and 3 where they occur in sympatry with P. gilberti. Plestiodon lagunensis occurs in disjunct populations in lower Baja California. (a) Juvenile of the southwestern gilberti ecomorph (Photo by Brad Alexander). (b) Juvenile of the Sierran gilberti ecomorph (Photo by Robert Hansen).

Figure 2

Body size divergence and copulation success. (a) Histogram showing frequency of copulation success as a function of size difference (binned into 10.0 mm categories) in no-choice mate trials. Size difference is calculated as male SVL−female SVL. (b) Histogram of SVL measurements for 399 P. skiltonianus and 536 P. gilberti taken on reproductively mature museum specimens, showing sexual dimorphism and divergence between ecomorphs. Males are shown above the central horizontal axis and females are below.

Figure 3

AFLP markers support a model of genetic isolation by distance across the mtDNA contact zone. Ovals delimit populations spanning from the northern Sierra Nevada south to the Mexico border . Brackets indicate paired populations used to estimate θ^B (means, standard errors and 95% CIs are reported) between (a) nearest neighbour populations and (b) populations at increasing distances from the contact zone. The dumbbell between populations 4 and 5 delimits the mtDNA contact zone. In the absence of gene flow, distance from the contact zone will not affect θ^B and the estimate of θ^B across the contact zone will be greater than between other neighbouring pairs of populations. Under isolation by distance, estimates of θ^B increase as the distance between paired populations increases. Our data are not consistent with an absence of gene flow, but are consistent with isolation by distance.