Fig 1.
Examples of color polymorphism in Polythore taxa used in this study.
For each wingform, the typical form is shown on the left, and extrema are shown on the right.
Fig 2.
Collection localities of Polythore used in this study and phylogenetic reconstruction using Cytochrome oxidase I (COI).
(A) map of collection localities of and (B) phylogenetic reconstruction (best ML phylogram) using COI data, values above the branches represent bootstraps support (ML) and posterior probabilities (BI). The magenta dots on the nodes represent the OTUs or species boundaries estimated by the PTP species delimitation model; lighter color represents less-supported probability for that OTU.
Fig 3.
“Default wing” used as a template for landmarking analysis of fore- and hindwings of Polythore wingform.
The positions of six major bands (I–VI) are marked as the cross six longitudinal veins: anterior edge or costa (C), radius anterior (RA), second branch of radius posterior (RP2), third branch of radius posterior (RP3), media posterior (MP), and posterior edge (venation terminology per Riek & Kukalová-Peck (1984)). (A) LMs 1–14 are major morphological points representing the basic venation pattern (~outline) of the wing, and LMs 15–50 represent the proximal edges of the six bands. (C–D) red arrows depict protocol for “collapsing” LMs in (C) proximodistal and (D) anteroposterior directions in wings where bands are missing or do not extend the full width of the wing (see Methods text for full description).
Fig 4.
Mean wing shapes for Polythore wing morphs from landmark analysis.
Fifty landmarks were taken from individuals of the 7 wing morphs, Procrustes superimposed, and averaged together for each wing morph. Outline landmarks (LMs 1–14) are dashed; color bands (LMs 15–50) are solid lines and numbered from I to VI. See Fig 2 and text for detailed description of landmarking protocol. M = male, and F = female.
Fig 5.
Discriminant analysis of morphological results.
Discriminant analysis of principal components (DAPC) plots for (A) landmarking, (B) chromaticity, and (C) GWT analyses. (D–F) areas of the wings corresponding to the 5 most discriminating coefficients for DAPC axes 1 and 2 for each of the analyses (see Results text for further explanation of D–F). Note: here, these locations are superimposed onto an image of a P. aurora male, for presentation, but represent variation among all individuals/wingsforms.
Table 1.
Mahalanobis distances (MDs; calculated from first 6 DAPCs) between wingforms using for different combinations of morphological coefficients.
Fig 6.
Comparison of mtDNA haplotypes among Polythore.
(A) mtDNA haplotype network and (B) FST values across the geographical population. The COI haplotype network shows the relationships among 29 haplotypes. Each circle represent an haplotype, the size of the circles represents the number of individuals sharing the haplotype, colors represent the geographical population, black small circles show missing or unsampled haplotypes, branch lengths are fixed and represent the genetic distance between the haplotypes, parallels lines represent more than 100 missing haplotypes and high genetic distance between the haplotypes. The population matrix is organized from Northern to Southern geographic location, colors underneath the diagonal represent the FST values (i.e. gradient) and above the diagonal shows the significant p-values.
Table 2.
Peruvian geographical population polymorphism statistics.
Fig 7.
Comparison of the differing topologies among phylogenetic reconstructions.
Simplified topologies are shown based on (A) morphological and (B) COI data, showing typical examples of each wing morph (center). Only males were included in these analyses. For full trees, see Fig 5 and S1 Fig. G. Note: (a) P. ornata was recovered as monophyletic (although with low, 9% bootstrap branch support) except for two specimens, which were recovered within P. victoria (see S1 Fig. G); (b) in B, some P. neopicta grouped with P. ornata, while others grouped with P. victoria.