DNA barcodes reveal different speciation scenarios in the four North American Anthocharis Boisduval, Rambur, [Duménil] & Graslin, [1833] (Lepidoptera: Pieridae: Pierinae: Anthocharidini) species groups

The mitochondrial DNA COI barcode segment sequenced from American Anthocharis specimens across their distribution ranges partitions them into four well-separated species groups and reveals different levels of differentiation within these groups. The lanceolata group experienced the deepest divergence. About 2.7% barcode difference separates the two species: A. lanceolata Lucas, 1852 including A. lanceolata australis (F. Grinnell, 1908), from A. desertolimbus J. Emmel, T. Emmel & Mattoon, 1998. The sara group consists of three species distinctly defined by more than 2% sequence divergence: A. sara Lucas, 1852, A. julia W. H. Edwards, 1872, and A. thoosa (Scudder, 1878). Our treatment is fully consistent with morphological evidence largely based on the characters of fifth instar larvae and pupal cone curvature (Stout, 2005, 2018). In barcodes, it is not possible to see evidence of introgression or hybridization between the three species, and identification by morphology of immature stages always agrees with DNA barcode identification. Interestingly, A. thoosa exhibited the largest intraspecific divergence in DNA barcodes, and several of its metapopulations are identifiable by haplotypes. The cethura group is characterized by the smallest divergence and is best considered as a single species variable in expression of yellow coloration: A cethura C. Felder & R. Felder, 1865. Notably, the most sexually dimorphic subspecies A. cethura morrisoni W. H. Edwards, 1881 is the most distinct by the barcodes. Finally, the midea group barcodes do not always separate A. midea (Hübner, [1809]) and A. limonea (A. Butler, 1871) and we observe gradual accumulation of differences from north (northeastern USA) to south (Hidalgo, Mexico). This barcode gradient suggests a recent origin of the two midea group species and provides another example of vicariant sister species well defined by morphology, ecology and geography, but not necessarily by DNA barcodes.