Validation and description of two new north-western Australian Rainbow skinks with multispecies coalescent methods and morphology.

Huw A. Ogilvie,Ana C. Afonso Silva,Natali Santos,Craig Moritz

doi:10.7717/peerj.3724

Abstract

While methods for genetic species delimitation have noticeably improved in the last decade, this remains a work in progress. Ideally, model based approaches should be applied and considered jointly with other lines of evidence, primarily morphology and geography, in an integrative taxonomy framework. Deep phylogeographic divergences have been reported for several species of Carlia skinks, but only for some eastern taxa have species boundaries been formally tested. The present study does this and revises the taxonomy for two species from northern Australia, Carlia johnstonei and C. triacantha. We introduce an approach that is based on the recently published method StarBEAST2, which uses multilocus data to explore the support for alternative species delimitation hypotheses using Bayes Factors (BFD). We apply this method, jointly with two other multispecies coalescent methods, using an extensive (from 2,163 exons) data set along with measures of 11 morphological characters. We use this integrated approach to evaluate two new candidate species previously revealed in phylogeographic analyses of rainbow skinks (genus Carlia) in Western Australia. The results based on BFD StarBEAST2, BFD* SNAPP and BPP genetic delimitation, together with morphology, support each of the four recently identified Carlia lineages as separate species. The BFD StarBEAST2 approach yielded results highly congruent with those from BFD* SNAPP and BPP. This supports use of the robust multilocus multispecies coalescent StarBEAST2 method for species delimitation, which does not require a priori resolved species or gene trees. Compared to the situation in C. triacantha, morphological divergence was greater between the two lineages within Kimberley endemic C. johnstonei, which also had deeper divergent histories. This congruence supports recognition of two species within C. johnstonei. Nevertheless, the combined evidence also supports recognition of two taxa within the more widespread C. triacantha. With this work, we describe two new species, Carlia insularis sp. nov and Carlia isostriacantha sp. nov. in the northwest of Australia. This contributes to increasing recognition that this region of tropical Australia has a rich and unique fauna.

Highlights

Cryptic species—when two or more distinct species are inaccurately classified under one species name (Bickford et al, 2007)—present great challenges for taxonomy and species delimitation due to the desirability of validating candidate species using multiple lines of evidence (Fujita et al, 2012)
We examined five morphometric characters taken to the nearest 0.1 mm with Mitutoyo electronic callipers: snout-vent length (SVL), axilla-groin length (AGL), head length (HL) measured from anterior edge of tympanum to snout, head width (HW) measured at widest point of the head, and head depth (HD) measured at parietal scales
While topology for the original datasets was as expected by 99% of the models (Fig. 2), for each gene set without recombining blocks only 64% and 85% of the models supported the same topology

Summary

Introduction

Cryptic species—when two or more distinct species are inaccurately classified under one species name (Bickford et al, 2007)—present great challenges for taxonomy and species delimitation due to the desirability of validating candidate species using multiple lines of evidence (Fujita et al, 2012). To more robustly infer species boundaries, the use of integrative taxonomy is increasingly common (Bickford et al, 2007; Padial et al, 2010). The objective of this approach is to corroborate taxonomic validity with independent, distinct types of evidence. Given deep genetic divergence, fixed morphological differences are not necessary to diagnose species boundaries since speciation itself does not require phenotypic characters to evolve at the same rate as the genome (Leaché & Fujita, 2010). In taxa with inherently conservative morphology, it may be that the primary evidence for distinct species will come from genetic data

Methods

Results

Discussion

Conclusion