Genes Suggest Ancestral Colour Polymorphisms Are Shared across Morphologically Cryptic Species in Arctic Bumblebees.

Paul H Williams,Mikhail V Berezin,Frode Ødegaard,Claus Rasmussen,Alexandr M Byvaltsev,Björn Cederberg,Cory S Sheffield,Jiaxing Huang,Suzanne T Williams,Leif L Richardson

doi:10.1371/journal.pone.0144544

Abstract

Our grasp of biodiversity is fine-tuned through the process of revisionary taxonomy. If species do exist in nature and can be discovered with available techniques, then we expect these revisions to converge on broadly shared interpretations of species. But for the primarily arctic bumblebees of the subgenus Alpinobombus of the genus Bombus, revisions by some of the most experienced specialists are unusual for bumblebees in that they have all reached different conclusions on the number of species present. Recent revisions based on skeletal morphology have concluded that there are from four to six species, while variation in colour pattern of the hair raised questions as to whether at least seven species might be present. Even more species are supported if we accept the recent move away from viewing species as morphotypes to viewing them instead as evolutionarily independent lineages (EILs) using data from genes. EILs are recognised here in practice from the gene coalescents that provide direct evidence for their evolutionary independence. We show from fitting both general mixed Yule/coalescent (GMYC) models and Poisson-tree-process (PTP) models to data for the mitochondrial COI gene that there is support for nine species in the subgenus Alpinobombus. Examination of the more slowly evolving nuclear PEPCK gene shows further support for a previously unrecognised taxon as a new species in northwestern North America. The three pairs of the most morphologically similar sister species are separated allopatrically and prevented from interbreeding by oceans. We also find that most of the species show multiple shared colour patterns, giving the appearance of mimicry among parts of the different species. However, reconstructing ancestral colour-pattern states shows that speciation is likely to have cut across widespread ancestral polymorphisms, without or largely without convergence. In the particular case of Alpinobombus, morphological, colour-pattern, and genetic groups show little agreement, which may help to explain the lack of agreement among previous taxonomic revisions.

Highlights

Authors have sought to reduce subjectivity in the practice of species discovery or ‘delimitation’ [1] by comparing the results from applying multiple criteria in an explicitly integrative taxonomy [2]
For some arctic bumblebees, taxonomic revisions by experienced specialists have shown unusually poor agreement on the number of species, we suggest in part because of poor agreement among results from applying different criteria
More than a century ago, it was realised that the groups supported by skeletal morphology are more clearly circumscribed, and that within these morphological species, bumblebees are highly variable in colour patterns [7,8,9,10]

Summary

Introduction

Authors have sought to reduce subjectivity in the practice of species discovery or ‘delimitation’ [1] by comparing the results from applying multiple criteria in an explicitly integrative taxonomy [2]. More than a century ago, it was realised that the groups supported by skeletal morphology are more clearly circumscribed, and that within these morphological species, bumblebees are highly variable in colour patterns [7,8,9,10] This is supported by recent genetic studies [11,12,13,14], reaffirming the long-held view that bumblebee colour patterns do not necessarily diagnose species [15]. Detailed studies of bumblebee variation have led to the realisation that species can be cryptic in both morphology and colour pattern [16, 17], confirmed by genetic studies [18, 19]

Methods

Results

Discussion

Conclusion