Fig 1.
Geographic distribution of the Crotalus mitchellii group in western North America, including currently recognized subspecies of C. mitchellii (based after [15]).
Inset shows islands in the Los Angeles Bay region of the central Gulf of California that are inhabited by C. mitchellii, sensu lato.
Fig 2.
Ordination (PCA) and clusters of mtDNA sequence (A), nDNA SNP (B), phenotypic (C), and climatic (D) datasets in geographic space, (E,F,G,H, respectively).
Clusters are based on PC axes that explained > 5% of total variation for each analysis (excluding axis 1 for nDNA, which was correlated with amount of missing data). Choice of PC axes presented on scatterplots is arbitrary (e.g., PC3 and PC4 are presented in A because the “grey” cluster, which represents specimens from Cabeza de Caballo Island, is overlaid with other clusters and is not visible using other axes); see S2 Fig for scatterplots of additional PC axes.
Fig 3.
Genealogy of mitochondrial ATPase 8 and 6 genes for 108 individuals of the Crotalus mitchellii group (including all recognized taxa).
Internodes are labeled with posterior probabalities/ML bootstrap support values. PI = Piojo Island, CCI = Cabeza de Caballo Island, SI = Smith Island, EMI = El Muerto Island.
Fig 4.
nDNA SNP genealogy for the Crotalus mitchellii species complex, with posterior assignment of each individual to clusters using STRUCTURE.
The SNP genealogy is based on maximum likelihood analysis of 2318 concatenated SNPs; numbers indicate boostrap support for major clades. The STRUCTURE analysis is based on 1826 biallelic SNPs. PI = Piojo Island; SI = Smith Island; CCI = Cabeza de Caballo Island.
Fig 5.
Allometric scaling of head size as a function of rattle size among the species of the Crotalus mitchellii complex.
Each species demonstrates distinct differences in developmental trajectories (ANCOVA homogeneity of slopes test; df = 2,321, F = 536.73, P < 0.0001).
Fig 6.
Species tree for species recognized within the Crotalus mitchellii group.
The species tree was assembled using a full coalescent model applied to 2409 biallelic SNPs; support values represent proportion of trees sampled with a particular topology. Dashed lines with arrows indicate taxa for which evidence of introgression was found using Patterson’s D-statistics. Photograph of C. mitchellii by M. Ingrasci.