Fig. 1.

Hesperia colorado type specimen traced to Lake County, Colorado. Different methods consistently partition specimens into four populations (named subspecies): sublima (red), ochracea (orange), colorado (blue), and idaho (green). (a) Principal component analysis (PCA) of covariance between the SNPs in samples from Colorado using Eigensoft (Patterson et al. 2006). The 1st and 2nd components are shown. (b) Truncated map of Colorado showing specimen localities. Arrow points to the locality in Lake County we deduced for the colorado lectotype. (c) t-SNE (parameter: perplexity = 10) reduced the first ten PCA dimensions to two, revealing populations as clusters. (d) TREEMIX results showing the clustering and evolutionary history of specimens. (e) Population structure inferred by STRUCTURE, showing the proportion of each population’s features in each specimen. Names of taxa (colorado, oroplata, ochracea) are given for type specimens (HT holotype, LT lectotype, PLT paralectotype), and primary types are highlighted in magenta; other historical specimens are highlighted in gray; county names are given for other specimens (see other data in supplementary table S1, Supplementary Material online, referred to by the numbers from 1 to 86 given as leaves in the TREEMIX tree); a question mark after the name of a type indicates a previously uncertain collection locality that we identify here using genomic comparisons.