The natural and human-mediated expansion of a human-commensal lizard into the fringes of Southeast Asia

Abstract

Background: Human-commensal species often display deep ancestral genetic structure within their native range and founder-effects and/or evidence of multiple introductions and admixture in newly established areas. We investigated the phylogeography of Eutropis multifasciata, an abundant human-commensal scincid lizard that occurs across Southeast Asia, to determine the extent of its native range and to assess the sources and signatures of human introduction outside of the native range. We sequenced over 350 samples of E. multifasciata for the mitochondrial ND2 gene and reanalyzed a previous RADseq population genetic dataset in a phylogenetic framework.

Results: Nuclear and mitochondrial trees are concordant and show that E. multifasciata has retained high levels of genetic structure across Southeast Asia despite being frequently moved by humans. Lineage boundaries in the native range roughly correspond to several major biogeographic barriers, including Wallace's Line and the Isthmus of Kra. Islands at the outer fringe of the range show evidence of founder-effects and multiple introductions.

Conclusions: Most of enormous range of E. multifasciata across Southeast Asia is native and it only displays signs of human-introduction or recent expansion along the eastern and northern fringe of its range. There were at least three events of human-introductions to Taiwan and offshore islands, and several oceanic islands in eastern Indonesia show a similar pattern. In Myanmar and Hainan, there is a founder-effect consistent with post-warming expansion after the last glacial maxima or human introduction.

Keywords: Biogeography; Eutropis multifasciata; Invasive species; Phylogeography; Scincidae; Southeast Asia.

Fig. 1

A Map displaying phylogeographic structure within Eutropis multifasciata, with the inset showing the mitochondrial BEAST timetree. Triangular tips in the tree have been collapsed and indicate more than one sample within a clade. The colors of geographic sampling points correspond to the tip colors in the tree and are nested according to the larger colored shapes corresponding to the encircled colored clades in the tree. Subclades are labeled corresponding to the tree tip labels. Nodes in the timetree with posterior probability greater than 0.95 are shown with a black dot, and timescale is in units of million years. Relevant biogeographic features discussed in the text are labeled. The dotted black lines indicate the geographic extent of the major numbered clades. Wallace’s Line and Huxley’s modification are shown in light blue for reference. The underlying map is shaded according to a digital elevation model (SRTM). The sea level bathymetry displays the 40 m (darker grey) and 120 m (lighter grey) depth contours (GEBCO), and indicate the areas exposed during periods of reduced sea levels. Location of inset maps shown by boxes. B Inset map of Sulawesi. White-rimmed points within the green T3 clade display additional phylogenetic substructure not indicated on the collapsed timetree. C Inset map of Taiwan, Ludao, and Lanyu

Fig. 2

Comparison between the simplified ND2 BEAST tree (left) and the concatenated RADseq BEAST tree (right). Clades recovered in both datasets are indicated by colored boxes. Solid lines between the boxes represent samples that are present in both trees, whereas dashed lines represent clades that are presumed to be equivalent based on geographic location and tree position. The two magenta lines refer to instances of mismatch between the two trees and possible introgression. Black dots at nodes represent posterior probabilities greater than 0.95, or are labeled otherwise, and bars at nodes represent the 95% confidence interval on node height

Fig. 3

Simplified species distribution model for Eutropis multifasciata under the present climate (orange) and projected on the climate of the last glacial maximum (LGM; Purple). Much of its northern range is likely to have been uninhabitable during the LGM. Back points indicate all georeferenced localities of E. multifasciata, and blue points indicate the rarefied localities used to train and test the species distribution model to avoid spatial autocorrelation. Note that all of Hainan and most of Taiwan are predicted to have been uninhabitable during the LGM