Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Jul 1.
Published in final edited form as: Syst Entomol. 2015 Jan 28;40(3):532–546. doi: 10.1111/syen.12120

A tale of two haplotype groups: Evaluating the New World Junonia ring species hypothesis using the distribution of divergent COI haplotypes

Amber P Gemmell 1, Jeffrey M Marcus 1,*
PMCID: PMC4532355  NIHMSID: NIHMS650555  PMID: 26279602

Abstract

The New World Junonia butterflies are a possible ring species with a circum-Caribbean distribution. Previous reports suggest a steady transition between North and South American forms in Mesoamerica, but in Cuba the forms were thought to co-exist without interbreeding representing the overlapping ends of the ring. Three criteria establish the existence of a ring species: a ring-shaped geographic distribution, gene flow among intervening forms, and genetic isolation in the region of range overlap. We evaluated mitochondrial cytochrome oxidase I haplotypes in Junonia from 9 species in the Western Hemisphere to test the Junonia ring species hypothesis. Junonia species are generally not monophyletic with respect to COI haplotypes, which are shared across species. However, two major COI haplotype groups exist. Group A predominates in South America, and Group B predominates in North and Central America. Therefore, COI haplotypes can be used to assess the degree of genetic influence a population receives from each continent. Junonia shows a ring-shaped distribution around the Caribbean, and evidence is consistent with gene flow among forms of Junonia, including those from Mesoamerica. However, we detected no discontinuity in gene flow in Cuba or elsewhere in the Caribbean consistent with genetic isolation in the region of overlap. Though sampling is still very limited in the critical region, the only remaining possiblity for a circum-Caribbean discontinuity in gene flow is at the Isthmus of Panama, where there may be a transition from 98% Group B haplotypes in Costa Rica to 85–100% Group A haplotypes in South America.

Keywords: Junonia, ring species, adaptive radiation, cytochrome oxidase I, phylogeography

Introduction

Ring species are groups of species or subspecies with ring-like geographic distributions around regions of unsuitable habitat such that the extreme ends of the range overlap (Mayr 1942). Gene flow occurs through intermediate forms around the ring circumference, but in the overlapping region forms do not interbreed. Ring species have been of particular interest to evolutionary biologists because they may represent a transitional state in species formation by geographic vicariance (Irwin et al. 2001). Many organisms have been identified as possible ring species, some of which have been studied intensively. Among these, some species have interruptions in gene flow at various points around the supposed ring (e.g. herring gulls, genus Larus (Liebers et al. 2004)), and are thus not ring species. Other species (e.g. the Asian Phylloscopus warblers (Irwin et al. 2005; Irwin et al. 2008) and the Californian Ensatina salamanders (Wake 2001; Pereira & Wake 2009)) appear to show gene flow around the ring, but genetic isolation where the ends of the range overlap, thereby fulfilling the criteria of a ring species. Other proposed ring species have not been studied sufficiently to determine whether they meet the criteria of a ring species: ring-shaped geographic distribution, genetic isolation in the region of range overlap, and gene flow among the intervening forms (Irwin et al. 2001).

The New World Junonia butterflies are one example of an understudied possible ring species (Table 1). The ring-shaped Caribbean distribution of Junonia was identified by Forbes (1928) who noted that there was a steady phenotypic transition in Mexico and Central America from North American color pattern phenotypes (similar to J. coenia), to phenotypes typical of South American forms (similar to J. evarete and J. genoveva). The two most prominent of these phenotypes that increase from North to South are the degree of orange suffusion within the white apical patch on the dorsal forewings and the appearance of blue iridescent scales as a component of the ground coloration on both the dorsal forewings and hindwings (Forbes 1928). However, two apparently separate colonization events of Cuba and the nearby Isle of Pines (one originating from North America and one from South America) and the descendants of these two Junonia colonizations were thought to co-exist without interbreeding. New World Junonia was popularized as a possible ring species by Mayr (1942, 1963), but no additional information was added beyond what was reported by Forbes (1928). Brown and Heineman (1972) suggested that North American Junonia penetration into the Caribbean may not be limited to Cuba, with specimens with North American-like color patterns known from Jamaica and the Cayman Islands. Remington (1985) revisited the Junonia ring species hypothesis and provided a map, supplemented with photographs of specimens, showing geographic variation in the genus. Finally, Irwin et al. (2001) examined the Caribbean Junonia as part of a review of many proposed ring species and speculated that gene flow between forms around the Caribbean was unlikely.

Table 1.

Described Junonia species from the Western Hemisphere with selected subspecies. Also included are some forms with ambiguous species assignments.

Species
Junonia coenia Hübner, 1822 J. coenia grisea Austin & Emmel, 1998
Junonia divaricata C. & R. Felder, 1867
Junonia evarete (Cramer, 1779)
Junonia sp. flirtea (Fabricius, 1793)
Junonia genoveva (Cramer, 1780)
J. sp. hilaris C. & R. Felder, 1867
J. sp. affin. hilaris Borchers and Marcus, 2014
Junonia litoralis Brévignon, 2009
Junonia neildi Brévignon, 2004
J. sp. nigrosuffusa Barnes & McDunnough, 1916
J. sp. pallens C. & R. Felder, 1867
Junonia vestina C. & R. Felder, 1867
Junonia wahlbergi Brévignon, 2008
Junonia zonalis C. & R. Felder, 1867
Open in a new tab

Other studies have shown that forms of New World Junonia share identical karyotypes (N=31)(Maeki & Remington 1960; Turner & Parnell 1985), can hybridize and produce fertile offspring in the lab (Hafernik 1982; Paulsen 1994; Paulsen 1996), and engage in interspecific courtship (Minno & Emmel 1993) and possible hybridization (Scott 1986) in the field. Mitochondrial markers show historic gene flow between Junonia populations in Mexico (Pfeiler et al. 2012a). More recently, molecular genetic work using both mitochondrial and nuclear markers shows hybridization between Junonia species appears to take place in the wild and individuals that show genetic evidence of hybrid ancestry are common in some populations (Borchers & Marcus 2014; Gemmell et al. 2014).

Barcoding studies of the mitochondrial cytochrome oxidase I (COI) gene in Junonia show that there are 2 New World mitochondrial haplotype groups (A and B), with 3.9% sequence divergence between them (Pfeiler et al. 2012b). However, most mitochondrial genotypes are shared across several Junonia species and most New World Junonia species contain individuals with mitochondria of both haplotype groups (Brévignon & Brévignon 2012; Borchers & Marcus 2014; Gemmell et al. 2014). Regardless of species, COI haplotype group A is most abundant in South America (85% frequency or higher) while haplotype group B predominates in North America (nearly 100% frequency) (Pfeiler et al. 2012b; Gemmell et al. 2014). Haplotype group frequencies on Martinique and Guadeloupe are intermediate between those in North and South America (Gemmell et al. 2014). This suggests that while COI barcode haplotypes are not diagnostic for Junonia species, they are good markers for measuring the genetic influence from continental North and South America on Junonia populations.

The Junonia ring species hypothesis (Forbes 1928), predicts that the island of Cuba will show strong genetic influences from both North and South America, but the rest of the Caribbean will show primarily South American influences. It is also expected that there will be a continuous transition in allele frequencies between COI haplotype groups A and B in Mexico and Central America to parallel the transition in color pattern phenotypes reported from this region. Finally, the ring species hypothesis predicts a discontinuity in COI haplotype group frequencies in the Caribbean in association with the overlapping region where supposedly non-interbreeding forms of Junonia co-occur in Cuba at the termini of the ring. We use a combination of phylogenetic analysis, haplotype networks, and biogeographic approaches to test these predictions of the Junonia ring species hypothesis. To place the possible New World Junonia ring species in a phylogenetic context we also consider prospective Old World sister taxa and the potential origins of haplotype groups A and B.

Materials & Methods

Specimens and DNA Preparation

Junonia specimens from all 9 currently recognized New World species (Borchers & Marcus 2014) were collected from the wild using hand-held butterfly nets and frozen at −20°C or obtained from other collectors (Supporting Information Table S1). Specimens were identified on the basis of morphological characters (Turner & Parnell 1985; Neild 2008; Brévignon & Brévignon 2012), however, it should be noted that Junonia taxonomy is in flux due to recent genetic comparisons (Pfeiler et al. 2012b; Borchers & Marcus 2014; Gemmell et al. 2014). Thorough molecular comparisons between Florida, Texas, and Caribbean Junonia are incomplete making it unclear which forms are conspecific (Gemmell et al. 2014). We provisionally refer to the Junonia from Florida as J. coenia (the common buckeye), J. “evarete” (the mangrove buckeye), and J. “genoveva1” (the tropical buckeye, all sensu Turner and Parnell, 1985) because this is the current common usage of scientific names in Florida (Glassberg et al. 2000; Cech & Tudor 2005; Mitter 2013) and is the taxonomy used for our collection and importation permits. For the same reasons, we refer to Junonia from Texas as J. coenia (the common buckeye), J. sp. nigrosuffusa (the dark buckeye), and J. “genoveva2” (the mangrove buckeye)(Pfeiler 2011; Pfeiler et al. 2012b). We recognize that alternate taxonomic hypotheses for the taxa exist (Neild 2008; Brévignon & Brévignon 2012), but prefer to avoid causing further confusion by withholding taxonomic judgment until appropriate tests using nuclear markers are conducted. Junonia “genoveva1” from Florida and Junonia “genoveva2” from Texas are not conspecific with each other, and it is yet to be determined whether either of these forms is conspecific with J. genoveva from South America.

Individual legs were removed from each specimen and DNA was isolated using the DNEasy Blood and Tissue kit (Qiagen, Düsseldorf, Germany). Legs were ground with a ceramic mortar and pestle in 180 µl of tissue lysis buffer ATL. Once homogenized, 20 µl of proteinase K (Qiagen, 600 mU/mL) was added and the mixture was incubated for one hour in a 55°C water bath until tissue was lysed. Samples were then processed manually as previously described (Borchers & Marcus 2014) or loaded into a Qiagen QIAcube extraction robot and processed using the standard instrument protocol for purification of total DNA from animal tissue (Gemmell et al. 2014). Samples were evaluated for DNA concentration using a Nanodrop 2000 spectrophotometer (Nanodrop, Wilmington, Delaware, USA) and stored at −20°C until needed.

PCR and Sequencing

Cytochrome oxidase I (COI) PCR products were generated by amplification with LCO1490 and HCO2198 primers (Table 2) (Folmer et al. 1994). Quick-Load Taq 2× Mastermix (New England Biolabs, Ipswich, MA, USA) was used in PCR reactions with total volumes of 25 µl. Amplification protocols were run on MyCycler or S1000 Thermal Cyclers (BioRad, Hercules, California, USA) for these and all other PCR amplifications. PCR reaction conditions were: 95°C for 5 minutes; 35 cycles of 94°C for 1 minute, 46°C for 1 minute, and 72°C for 1.5 minutes; and a final 5 minute extension at 72°C before being placed on a 4°C hold. PCR products were loaded into a 1% agarose gel in TAE buffer, run for 1 hour at 78 V, and visualized using ethidium bromide under ultraviolet light.

Table 2.

Primer sequences used for cytochrome oxidase I (COI) PCR reactions.

Primer Sequence Source
LCO1490 GGTCAACAAATCATAAAGATATTGG (Folmer et al. 1994)
HCO2198 TAAACTTCAGGGTGACCAAAAAATCA (Folmer et al. 1994)
miniCOIF2 ATACTATTGTTACAGCCTCATGC (Gemmell et al. 2014)
miniCOIR2 TGTTGTAATAAAATTAATAGCTCC (Gemmell et al. 2014)
miniCOIF3 CCCCACTTTCATCTAATATTGC (Gemmell et al. 2014)
miniCOIR3 TATTTCGATCTGTTAAAAGTATAG This study
Open in a new tab

PCR products were sequenced using the Sanger dideoxy method as previously described (Borchers & Marcus 2014). PCR products were sequenced in both directions with the primers used to generate the products. Sequencing reactions were analyzed on ABI 3130 or ABI 3730×l automated sequencers (Applied Biosystems, Carlsbad, California, USA) and edited using Sequencher 4.6 software (Sequencher 2005). Primer sequences were removed leaving sequenced amplified products of 658 bp, which were then aligned in CLUSTAL W version 2.1 (Thompson et al. 1994; Larkin et al. 2007).

Phylogenetic Analysis

In addition to the sequences generated by this study, additional sequences from our earlier work (Borchers & Marcus 2014; Gemmell et al. 2014) and previously published Old World Junonia COI sequences (Kodandaramaiah & Wahlberg 2007; Kodandaramaiah 2009; Hebert et al. 2013; Shi et al. 2013) were included in a phylogenetic analysis. COI sequences from each New World Junonia specimen were compared with each other and duplicate sequences were removed from the analysis (identical sequences are listed in Supporting Information Table S2). Therefore, only one exemplar of each genotype was included in the phylogenetic analysis, minimizing computation time. CLUSTAL W COI sequence alignments were converted to NEXUS format and analyzed phylogenetically using several reconstruction methods (distance, parsimony, likelihood) that rely on vastly different assumptions about sequence evolution, each of which recovered nearly identical trees. For the sake of brevity, we only present the maximum likelihood analysis (HKY model, 10 replicate heuristic searches with random number seeds, tree bisection and reconnection branch swapping algorithm) of PAUP version 4.0b10 (Swofford 1998). In addition, a maximum likelihood bootstrap analysis of the dataset (500 fast addition replicates, retaining nodes with frequency less than 50%) was conducted to assess the degree of support for each node.

Haplotype Network Construction

The aligned COI sequences generated by this study, along with sequences from previously published studies from our laboratory (Borchers & Marcus 2014; Gemmell et al. 2014) and exemplar COI sequences for Old World outgroup species J. orithya and J. villida (Kodandaramaiah & Wahlberg 2007; Hebert et al. 2013; Shi et al. 2013) (each previously identified as possible sister taxa to the New World Junonia based on morphology and wing color patterns (Seitz 1914; Corbet 1948; Vane-Wright & Tennent 2011)), were formatted for input into Arlequin 3.5 (Schneider et al. 2000). No J. villida sequences in Genbank contain the complete COI barcode, so we used the consensus sequence between a published sequence (NW99-1) (Kodandaramaiah & Wahlberg 2007) and a short J. villida COI sequence (AUS1, collected Sydney, Australia, 15 Dec.1999) that we generated ourselves (DNA Database of Japan (DDBJ) Accession Number AB981649) for network analysis. An AMOVA analysis was run using the following settings: 1000 permutations, determining the minimum spanning network (MSN) among haplotypes, computing distance matrix, and pair-wise difference with a gamma value of 0. The minimum spanning tree output from the AMOVA analysis was used as input for HapSTAR-0.7 (Teacher & Griffiths 2011), which displays the haplotype network in graphical form. This analysis, as implemented in Arlequin, requires all sequences to be the same length. Therefore, only full-length COI barcode fragments were used. Additional graphical adjustments to this network were made using Canvas X (ACD Systems, Seattle, Washington, USA) which included scaling the size of population circles to reflect the number of individuals with a specific COI haplotype, and the addition of divisions and colors to reflect geographical regions of a specific haplotype.

Restriction Fragment Length Polymorphism Genotyping

As a supplement to Junonia COI sequences, we evaluated COI PCR products from additional New World Junonia specimens by restriction digest to detect cut sites diagnostic for haplotype groups A and B. The quantity and quality of the PCR products amplified from the DNA of some specimens was insufficient for successful DNA sequencing, but was suitable for genotyping by restriction digest. PCR products were generated using a two-step amplification with LCO1490 and HCO2198 primers (as described above) followed by re-amplification using miniCOIF2 and HCO2198 primers (Table 1) (re-amplification conditions were: 95°C for 2 minutes; 5 cycles of 95°C for 1 minute, 46°C for 1 minute, 72°C for 30 seconds; 35 cycles of 95°C for 1 minute, 53°C for 1 minute, and 72°C for 30 seconds; and a final 5 minute extension at 72°C before being placed on a 4°C hold). PCR reactions were prepared using reaction volumes of 25 µl and Qiagen TopTaq Mastermix for these and all other PCR re-amplifications described below.

PCR products of the correct size were evaluated by a triple digest using BseYI, AflIII, and BamHI restriction endonucleases (New England Biolabs, Ipswich, MA, USA) (Fig. 1). A 10 µL aliquot of the PCR product was incubated with 10 µL of the triple digest cocktail (2 µL of NEB Buffer3, 2 µL of BSA (10×, 1 mg/mL), 4 µL of ddH2O, 0.5 µL of BseYI, 0.5 µL of AflIII and 1 µL of BamHI) for 1 hour in a 37°C water bath. Samples were then heated in a 70°C water bath for 10 minutes to deactivate the enzymes. Samples were stored on ice until they were run on a QIAxcel Advanced instrument (Qiagen) fitted with a DNA Screening Cartridge with QX Size Markers (25–500 bp v. 2.0) and QX Alignment Markers (15 bp-1 kb) using the AL320 electrophoresis method. Most genetic variants (~85%) within haplotype group A contain a BamHI cut site (GGATCC) at 256 bp producing 2 bands of sizes 256 bp and 313 bp. Haplotype group B genotypes lack this BamHI cut site, but instead, contain an AflIII cut site (ACR(G/A)Y(T/C)GT) at 351 bp and a BseYI cut site (CCCAGC) at 501 bp, thereby producing 3 bands of sizes 351 bp, 150 bp and 68 bp. A small minority of haplotype group B genotypes (~5%) lack the BseYI cut site, resulting in 2 bands of sizes 218 bp and 351 bp, which are still easily distinguished from digests of group A haplotypes. Assessing a New World Junonia COI haplotype at these 3 restriction sites allows for a completely unambiguous assignment to one of the 2 major haplotype groups (Fig. 1).

Fig. 1.

Fig. 1

Open in a new tab

Map of PCR primer binding sites and restriction enzyme cut sizes for the barcode fragment of COI. BamHI restriction sites are only found in haplotype group A alleles while AflIII and BseYI restriction sites are only found in haplotype group B alleles.

Samples that failed to amplify with miniCOIF2 and HCO2198 primers were re-amplified from the original LCO1490/HCO2198 PCR products with miniCOIF2 and miniCOIR2 primers (Table 1) using the same re-amplification thermocycler protocol followed by a restriction digest using BamHI. The other two restriction cut sites lie outside the 339 bp PCR amplification product produced by these primers (Fig. 1). Samples were also re-amplified from the LCO1490/HCO2198 PCR products with miniCOIF3 and HCO2198 or miniCOIF3 and miniCOIR3 primers (Table 1) using the same re-amplification protocol and digested with AflIII and BseYI before being assigned to a COI haplotype group based on bands produced by all digests (Fig. 1).

Phylogeographic Mapping of COI Haplotype Groups

To maximize the geographic representation of Junonia beyond our own sampling, we evaluated additional New World Junonia COI sequences from previously published studies (Kodandaramaiah & Wahlberg 2007; Janzen & Hajibabaei 2009; Marcus et al. 2009; Hebert et al. 2010; Escobedo 2011; Brévignon & Brévignon 2012; Janzen 2012; Pfeiler et al. 2012b; Mitter 2013; Borchers & Marcus 2014; Gemmell et al. 2014) and the Junonia haplotype group assignments determined by restriction digest in our laboratory, to calculate the proportion of A and B haplotype groups for each species in various geographic regions. This was done to evaluate differences in haplotype group allele frequencies between species, using samples of the same species from different countries (or in some cases in sub-regions within a country (e.g. Texas and Florida, USA)) as replicates to estimate within-region variation. Since in most instances, COI haplotypes are not confined to particular species in Junonia and occur at very similar frequencies in all species found in a given country (or sub-region), we combined specimens from all species together to calculate and map the haplotype group frequencies geographically. For comparison, species and selected subspecies ranges were determined from collection localities of specimens in our research collection and from published reports, and were plotted onto the same map (Forbes 1928; Turner & Parnell 1985; Brown et al. 1992; Elster et al. 1999; Neild 2008; Menezes & Peixoto 2009; Pfeiler 2011; Brévignon & Brévignon 2012; Gernaat et al. 2012; Pfeiler et al. 2012a; Pfeiler et al. 2012b).

To check for discontinuities in the distribution of group A and B haplotypes across the Caribbean we plotted the proportion of group A alleles (as determined above) in each Caribbean population versus the distance from the Southern tip of mainland Florida, USA. We used the “As the Crow Flies” distance calculator (http://tjpeiffer.com/crowflies.html) to determine the straight-line distance from Everglades National Park in South Florida to the closest landfall in Cuba, Jamaica, Hispaniola, Puerto Rico, Antigua, Guadeloupe, Dominica, Martinique, and French Guiana. Plotted haplotype group frequency and distance data were then analyzed by logarithmic regression because the expected genetic influence of North American populations on West Indian populations is expected to be inversely proportional to geographic distance.

Results

Phylogenetic Analysis

New COI sequences generated by this project were deposited in Genbank, accession numbers KM287911 - KM288274 (364 accessions). Maximum likelihood analysis of these COI sequences, plus those from our prior studies (Borchers & Marcus 2014; Gemmell et al. 2014), produced 3 phylogenetic trees of likelihood score 3281.34255 (Fig. 2). As previously reported (Pfeiler et al. 2012b; Borchers & Marcus 2014; Gemmell et al. 2014), there are two distinct monophyletic lineages of mitochondrial haplotypes found in the New World Junonia. Haplotype group A is primarily found in South American and Caribbean specimens, but is also present in a few North and Central American Junonia. Haplotype group B includes many North American, Central American, and Caribbean specimens, but also includes some South American Junonia. Only two named forms of New World Junonia can be identified unambiguously on the basis of COI barcode haplotype alone: J. vestina from the South American Andes (Haplotype A1 of Pfeiler et al. (2012b)) and J. coenia grisea from California. Of the two major sublineages identified by Pfeiler et al. (2012b) within Haplotype group A, Haplotype A1 (which includes only J. vestina) is monophyletic, while Haplotype A2 (which includes individuals from 7 Junonia species) is unresolved in our analysis (Fig. 2).

Fig. 2.

Fig. 2

Open in a new tab

Consensus of 3 maximum likelihood trees depicting the two distinct mitochondrial COI haplotype groups found among the New World Junonia. The single node that differs between these 3 trees is indicated by an asterisk. Numbers adjacent to nodes indicate bootstrap support. Bootstrap values of less than 30% are omitted. Group A haplotypes occur primarily in South America and the Caribbean. Group B haplotypes occur throughout the Western Hemisphere, but are most common in North and Central America. Most Junonia species include individuals with both haplotypes. Junonia vestina is unusual in that it is characterized by the unique haplotype group A1. Only one exemplar of each haplotype was included in the phylogenetic analysis, but the relative abundance of each haplotype is indicated by the vertical width of each triangle. Taxonomic abbreviations: J.c. (J. coneia), J.d. (J. divaricata), J.e. (J. evarete), J.”e.” (J. “evarete”), J.f.(J. sp. flirtea), J.g. (J. genoveva), J.”g.” (J. “genoveva”), J.h. (J. sp. hilaris), J.n. (J. nigrosuffusa), J.ne. (J. neildi), J.sp.a.h. (J. sp. affin hilaris), J.w. (J. wahlbergi) J.z. (J. zonalis). Geographic abbreviations: AR (Argentina), BO (Bolivia), CR (Costa Rica), CU (Cuba), FG (French Guiana), FL (Florida, USA), GD (Guadeloupe), JA (Jamaica), KY (Kentucky, USA), MA (Martinique), TX (Texas, USA).

The maximum likelihood trees with the best scores place a clade of Old World taxa (bootstrap support 75%) that includes African, Asian, and Indo-Australian species as the sister group to the New World Junonia (bootstrap support of 87%). Substantially similar trees were recovered by phylogenetic analysis using Parsimony and Distance reconstruction methods (data not shown). All of these trees also place haplotype groups A and B as sister to each other, making the New World Junonia monophyletic, but in the maximum likelihood analysis, this relationship is only supported by a bootstrap value of 36%. A minority of alternate maximum likelihood bootstrap tree topologies maintain the monophyly of the New World Junonia as a whole and the monophyly of haplotype group B, but haplotype group A is paraphyletic with respect to haplotype group B (not shown). These alternative bootstrap trees differ from each other with respect to which lineages within haplotype group A are most closely related to haplotype group B. Finally, a small final group of maximum likelihood bootstrap tree topologies do not support the monophyly of the New World Junonia (not shown), which is based on a very small number of informative sites (Borchers & Marcus 2014).

COI Haplotype Network

The minimum spanning haplotype network of New World Junonia COI sequences is in concordance with the results of the phylogenetic analysis, reinforcing the conclusion that haplotype groups A and B are distinct lineages (Fig. 3). The two most similar sequences between the two haplotype groups (LCB320 and LCB164) are 11 mutational steps apart. Haplotype group A shows substantially greater within-group genetic structural diversity and geographic structuring of its component genotypes. Group A is 17 mutational steps away from the COI sequence of the nearest outgroup species (J. villida, which has an Indo-Pacific distribution). Haplotype group B has lower internal genetic diversity and less geographic structuring with the majority of sequences falling into a single component genotype (LCB166) and is 18 mutational steps away from the J. villida outgroup sequence. As in the phylogenetic analysis, the COI sequences from J. vestina and J. coenia grisea are represented by distinct branches within the haplotype network.

Fig. 3.

Fig. 3

Open in a new tab

Haplotype networks generated using complete barcode fragment mitochondrial COI haplotypes. Circle areas are scaled to represent the number of individuals that contain a specific COI haplotype with the exception of the specimens carrying the LCB166 allele (40% of full scale). Divisions and colors of circles reflect geographic locality. COI haplotypes are generally shared across Junonia species with the exception of those associated with J. coenia grisea and J. vestina.

Phylogeography of Junonia COI Sequences

We determined the mitochondrial haplotype group assignments of 718 New World Junonia specimens for phylogeographic analysis. In most cases, species that co-occur in the same region show statistically indistinguishable haplotype group frequencies (Table 3). However, in different regions, even the same Junonia species can show dramatically different haplotype group frequencies. Thus, while there is only a weak species-specific signal in the COI data, there is a strong geographic signal in the distribution of group A and B COI haplotypes in the New World (Fig. 4). North and Central American Junonia carry COI alleles that are almost exclusively from haplotype group B. Of the 434 specimens evaluated from this region, only 8 specimens with group A haplotypes were identified (1 each from Panama, Costa Rica, Belize, and Mexico, plus 4 from southern Florida). In South America haplotype group A predominates with the frequency of haplotype group B never exceeding about 15% frequency and is often much lower. The continental transition between predominantly group A haplotypes and predominantly group B haplotypes occurs approximately at the isthmus of Panama, but this transition does not coincide with the range limit of any Junonia species (Fig. 4).

Table 3.

Haplotype group assignments for taxa in different regions of the Western Hemisphere. # Areas sampled refers to the number of countries or major regions within countries for each species. Haplotype group proportions are reported as % abundance in the region +/− 1 standard error. Within each region, taxa generally show statistically indistinguishable haplotype group frequencies and are clustered in the table. Exceptional taxa with distinct haplotype group frequencies from other taxa in the same region are listed separately in the table.

Taxon # Areas
Sampled		 Total #
Specimens		 % Haplotype
Group A		 % Haplotype
Group B
North America
J. coenia 5 193 0 +/− 0% 100 +/− 0%
J. "evarete" 1 60 3.3% 96.7%
J. "genoveva1" 1 22 9.1% 90.9%
J. “genoveva2” 2 20 0 +/− 0% 100 +/− 0%
J. sp. nigrosuffusa 2 47 2.1 +/− 5.9% 97.9 +/− 5.9%
Junonia coenia grisea 1 22 0% 100% +/− 0%‡
South America
J. sp. hilaris* 2 21 90.5 +/− 8.8% 9.5 +/− 8.8%
J. sp. flirtea 4 23 95.7 +/− 2.9% 4.3% +/− 2.9%
J. evarete 4 52 98.1 +/− 1.7% 1.9 +/− 1.7%
J. genoveva 5 35 94.3 +/− 3.6% 5.7 +/− 3.6%
J. litoralis 2 7 71.42 +/− 23.6% 28.6 +/− 23.6%
J. wahlbergi 1 10 80.0% 20.0%
J. divaricata 1 5 80.0% 20.0%
J. sp affin. hilaris* 1 7 57.1% 42.8%
J. vestina 1 7 100%† 0%
Central America
J. evarete 2 2 50.0 +/− 70.7% 50.0 +/− 70.7%
J. genoveva 2 56 3.6 +/− 69.4% 96.4 +/− 69.4%
Caribbean
J. neildi 3 13 53.8 +/− 38.1% 46.2 +/− 38.2%
J. zonalis 7 103 67.0 +/− 29.9% 33.0 +/− 29.9%
Open in a new tab
*

J. sp. hilaris and J. sp affin. hilaris are extremely difficult to distinguish based on morphological characters without supplemental genotyping with nuclear markers. Some specimens that have not undergone nuclear genotyping have been tentatively assigned to J. sp. hilaris.

†

Exclusively Junonia vestina alleles in haplotype Group A1.

‡

Exclusively California-specific alleles in haplotype Group B.

Fig. 4.

Fig. 4

Open in a new tab

Map of the Western Hemisphere showing Junonia species ranges and the frequency of group A and group B COI haplotypes in different regions. For each geographic region all species were considered together to calculate haplotype group frequencies except for J. vestina in Peru, which occurs in higher elevation habitats and is genetically distinct from other forms of Junonia. Ranges of species and selected subspecies are indicated by shaded and patterned areas on the map.

In the Caribbean, the frequency of the 2 haplotype groups varies considerably with haplotype group A ranging from 35% (Cuba) to 87% (Dominican Republic). Plotting the abundance of group A haplotypes in each locality versus the linear distance from the North American mainland and fitting a trend line yields a highly significant logarithmic regression (R2=0.640, p=0.005) across the entire Caribbean, with no evidence of discontinuity in gene flow in Cuba or elsewhere in the West Indies (Fig. 5).

Fig. 5.

Fig. 5

Open in a new tab

Plot of the proportion of haplotype group A genotypes found in Junonia populations on the Caribbean Islands and in French Guiana versus the distance of each island from the Southernmost point in mainland Florida, USA. Logarithmic regression analysis of the data yields a highly significant result, with an apparent asymptote at a Group A haplotype frequency of approximately 85%, consistent with the frequency of that haplotype group in much of mainland South America.

Discussion

Evaluating the Ring Species Hypothesis

In order to recognize the New World Junonia as a ring species, it must meet three criteria: a ring-shaped distribution, gene flow among intervening forms, and reduced or absent gene flow at the extreme ends of the distribution where forms overlap (Irwin et al. 2001). The ring-shaped distribution of Junonia around the circumference of the Caribbean and on the islands of the Antilles has long been known and was likely a key factor leading to the proposal of the Junonia ring species hypothesis (Forbes 1928).

Gene flow among intervening forms in Junonia is well-supported by laboratory studies of hybridization (Hafernik 1982; Paulsen 1994; Paulsen 1996), gradual geographic transitions in color pattern phenotypes around the circumference of the Caribbean (Forbes 1928; Remington 1985), and studies of nuclear and mitochondrial markers showing evidence of ongoing genetic exchange in Argentina, Mexico, and among populations in French Guiana and the Antilles (Pfeiler et al. 2012a; Borchers & Marcus 2014; Gemmell et al. 2014). In this study, it is clear that mitochondrial haplotypes are shared between nominal Junonia species and that the frequency of different COI haplotype groups in the Caribbean is a function of the distance of a population from the North or South American mainland, which is also consistent with gene flow.

The final prediction of the Junonia ring species hypothesis, as formulated previously in the literature (Forbes 1928; Remington 1985; Irwin et al. 2001) would be a discontinuity in gene flow in association with the supposed divergent, overlapping, and non-interbreeding ends of the Junonia distribution in Cuba. We would expect to see evidence of this discontinuity with abundant North American mitochondrial haplotypes (group B) present in Cuba, but rare elsewhere in the Caribbean. A logarithmic regression fits the COI haplotype group frequency data very well, and no discontinuity is apparent anywhere in the Caribbean (Fig. 4). Instead there appears to be a gradual change in haplotype group frequencies across the archipelago, suggesting genetic admixture (Fig. 5). We therefore reject the ring species hypothesis as originally articulated by Forbes (1928) because gene flow in COI is apparently not interrupted in Cuba or elsewhere in the Caribbean. However, it would be highly desirable to evaluate the population genetic structure of the New World ring species hypothesis using nuclear markers to see if the patterns of gene flow observed using mitochondrial COI are also evident in other parts of the Junonia genome.

While no discontinuity has been detected in Cuba, there may be a discontinuity in the distribution of mitochondrial haplotype groups elsewhere in the range of New World Junonia. In particular, the transition between North and South American haplotypes in the region around the Isthmus of Panama might be seen as abrupt (Fig. 4) with few group A alleles penetrating northwards into Central America, and an unknown number of group B alleles found in the northern parts of South America. However, our sampling in this region so far is relatively weak. If the well-sampled populations of Junonia from French Guiana are representative of populations from elsewhere on the Guiana Shield, the allele frequency of group B haplotypes would be expected to reach about 15% (Brévignon & Brévignon 2012; Gemmell et al. 2014).

An alternative to the ring species hypothesis of Forbes (1928) still requires evaluation: if a circum-Caribbean buckeye ring species exists, it may be that the terminal ends of the ring are not in Cuba as originally hypothesized, but at the transition between Central and South America. The formation of the land bridge making up the Isthmus of Panama was completed approximately 3 million years ago (Knowlton & Weigt 1998). In butterflies, COI evolves at a rate of about 1.8% per million years (Brower & DeSalle 1998), putting the divergence of haplotype groups A and B, which differ from each other by 3.9% (Borchers & Marcus 2014), at approximately 2.2 million years ago. The rough congruency of these two estimates is curious, but may be completely coincidental because Junonia have been documented to fly across substantial bodies of water during migration events (Williams 1930; Harris 1988). Unfortunately, our sampling in Panama, Columbia, Venezuela, and Ecuador is currently very limited, so we cannot fully evaluate whether this apparent steep discontinuity in mitochondrial genotypes is real or whether it is an artifact of our limited sampling. Junonia is abundant in this region (Neild 2008; Gernaat et al. 2012) and it should be possible to evaluate this part of the circum-Caribbean distribution of Junonia in more detail and test this alternative ring species hypothesis in a future study.

Other Biogeographic Points of Interest in Junonia

There are several additional geographic regions within the New World in which the study of population genetic structure of Junonia will be very interesting. First, it appears that California populations of J, coenia grisea (Austin & Emmel 1998) have COI haplotypes that are distinct genetically from populations of J. coenia coenia from elsewhere in North America (Pfeiler et al. 2012b), whereas most species of Junonia cannot be distinguished from one another on the basis of COI haplotypes (Brévignon & Brévignon 2012; Borchers & Marcus 2014; Gemmell et al. 2014). The ranges of the two subspecies of J. coenia are largely separated by the Rocky Mountains, but apparently meet in Arizona and New Mexico, USA and in Northern Mexico (Brock & Kaufman 2003). Different populations of Junonia coenia have markedly different host plant preferences (Camara 1997; Gemmell et al. 2014), which could contribute to temporal or habitat partitioning, but it is not known to what extent hybridization between subspecies might occur in this region. Given the distinctive J. coenia grisea COI haplotype (Pfeiler et al. 2012b) and the phenotypic differences between this form and nominate J. coenia (Austin & Emmel 1998), if gene flow between them is limited or absent, it may be appropriate to elevate J. coenia grisea to full species status.

Junonia vestina also has a distinctive COI haplotype (A1) that is not shared with other Junonia species (Pfeiler et al. 2012b). This species occurs at mid-elevations of the Andes Mountains in Ecuador, Peru, and Bolivia (Neild 2008), but thus far only material from Peru has been studied with molecular tools. Whether the range of J. vestina abuts or overlaps with the ranges of lower elevation forms of Junonia (both to the East and West of the Andes, which almost exclusively carry COI alleles from haplotype group A2) is unknown, as is whether there is any genetic exchange between them. Based on the limited sampling so far with mitochondrial markers (Pfeiler et al., 2012a; this study), there is no apparent gene flow. Additional fieldwork, sampling, and molecular analysis of Junonia specimens of all species from this region would clarify the degree to which J. vestina is unique among New World Junonia in its lack of hybridization with other forms.

Finally, Florida, USA also represents a region of particular interest with respect to Junonia population genetics. Unlike the regions described above, Junonia populations in Florida carry almost exclusively mitochondrial haplotype B. The only exceptions to this are 2 mangrove buckeye specimens (J. “evarete” sensu (Turner & Parnell 1985)) in the lower Florida Keys in Monroe County and 2 tropical buckeye specimens (J. “genoveva” sensu (Turner & Parnell 1985)) from the Everglades Greenway in Dade County. However, the colonization history of Junonia in Florida is better known than in any other geographic region. Junonia coenia and J. “evarete” are resident in Florida, but J. “genoveva1” was first recorded from Key Largo in 1961 and became abundant both there and in locations on the Florida mainland by 1978 (Minno & Emmel 1993). J. “genoveva1” is phenotypically similar to J. zonalis from the Caribbean, but the two forms should be compared using nuclear genetic markers before they are determined to be conspecific. Due to variations in climate (the larval host plant is very sensitive to frost) and land development by humans, there have been large fluctuations in the abundance of J. “genoveva1” (Glassberg et al. 2000). Currently, J. “genoveva1” is difficult to find in the Florida Keys, and all but one of the specimens included here come from a single locality on the Florida mainland. Museum specimens exist covering the entire history and geographic range of J. “genoveva1” in Florida. Examining the genetic changes that occurred in J. “genoveva1” as it invaded Florida, dispersed, and possibly hybridized with the resident Junonia species would be an informative case study in invasion biology.

Origins of the New World Buckeyes and COI Haplotype Groups

Phylogenetic approaches have yet to establish the sister taxon to the New World Junonia or even determine the continent from which the New World was colonized (Kodandaramaiah & Wahlberg 2007). In our minimum spanning haplotype network analysis we included sequences from specimens from the two species that are most morphologically similar to the New World forms: J. villida and J. orithya (Seitz 1914; Vane-Wright & Tennent 2011). Junonia villida has a range extending from the Chagos Islands in the West Indian Ocean to Tonga and the Tuamotu Archipelago in the South Pacific, while J. orithya has a range that extends from sub-Saharan Africa to New Guinea. Minimum spanning haplotype networks are a useful supplement to phylogenetic analysis, particularly in cases where complete coalescence of haplotypes in the various daughter lineages has not yet gone to completion and ancestral genotypes are still extant and present among the sequences being analyzed (Posada & Crandall 2001), as is often the case in determining intraspecific phylogenies, or ascertaining relationships among very closely related taxa (Bandelt et al. 1999; Jeratthitikul et al. 2013).

Our network analysis suggests that the COI haplotypes present in the New World Junonia are much more similar to that found in J. villida than to the haplotypes from J. orithya (Fig. 3). It also shows that while haplotypes from J. orithya originating in Asia and Australia are similar to one another, J. orithya COI from Africa is highly divergent, perhaps suggesting that the two forms are not conspecific. The association between J. villida (which has an Indo-Pacific distribution) and the New World Junonia COI haplotypes, supports an old hypothesis that J. villida is the sister taxon to New World Junonia species (Seitz 1914; Vane-Wright & Tennent 2011).

Junonia villida is marginally closer (17 vs. 18 mutational steps) to genotypes in haplotype group A, but the genetic diversity and the geographical structure of haplotype group A is also much greater than that of haplotype group B. Typically, in the absence of selection, a lack of substantial genetic structure is attributed to high rates of gene flow among populations or to recent colonization (Sotka & Palumbi 2006; Guzmán et al. 2011). Since haplotype groups A and B co-occur in the same Junonia species, we do not anticipate large differences in gene flow between the 2 haplotype groups, suggesting that the diversification of haplotype group B may be more recent than that of haplotype group A. This could be interpreted as evidence against a hypothesis proposed by Borchers and Marcus (2014) that the co-existence of the two haplotype groups may be the result of an ancient mitochondrial polymorphism present in the founding population of Junonia in the New World. Two remaining explanations for the origins of haplotype groups A and B are that mitochondrial haplotype group B evolved within North America from group A genotypes or that there were 2 separate Junonia invasions of the New World followed by hybridization events between the descendants of the 2 invasions. Junonia villida is extremely widespread, inhabiting islands from the Indian Ocean to the mid-Pacific Ocean (Vane-Wright & Tennent 2011), and it is possible that two invasions of the Western Hemisphere from different source populations of this taxon could contribute to the patterns of haplotype diversity that we observe among the New World Junonia. While there has been a recent review of the morphological diversity found in J. villida (Vane-Wright & Tennent 2011), no comparable survey of molecular diversity in this species has been conducted to date, representing an opportunity for further study. It is also possible that selection is acting on Junonia mitochondrial haplotypes (perhaps on COI sequences, but possibly on other mitochondrial sequences linked to COI), but this has not been demonstrated and is another opportunity for additional research. If selection is demonstrated, possible scenarios for New World Junonia COI haplotype diversification should be revisited (Irwin 2012).

Future Directions

The New World Junonia represent an exciting system in which to study both morphological and molecular diversification. The rapid radiation of the New World Junonia to occupy diverse habitats, utilize different larval host plants, and display variation in color pattern phenotypes is reminiscent of some other well-known radiations such as the haplochromine cichlids of Lake Victoria (Seehausen 2002), the North American sunflowers (Rieseberg et al. 2007), and the swordtails and platyfish of Central America (Marcus & McCune 1999). These radiations share with Junonia the propensity for interspecific hybridization in the wild (Cui et al. 2013; Keller et al. 2013; Bock et al. 2014) making them particularly amenable for studying the roles of reticulate evolution in the generation of novel phenotypes. These biological characteristics, the ease of Junonia laboratory culture (Marcus 2005), and the diverse tools available for experimentally manipulating Junonia butterflies (Marcus 2005; Nijhout 2010; Dhungel et al. 2013; Beaudette et al. 2014) provide much potential for identifying and assigning function to genetic variation associated with ecologically relevant traits, linking genotype and phenotype.

Supplementary Material

Supp TableS1
Supp TableS2

Acknowledgements

We thank Fred Nijhout for introducing us to the Junonia ring species puzzle. Thanks to Andy Anderson, Christian Brévignon, Bill Dempwolf, Bob Dowell, Loran Gibson, Nick Gillham, Victor Hitchings, Jorge Monteiro-Moreno, Fred Nijhout, Charlie Sassine, and Chad Stinson for generously contributing specimens for this project. Christian Brévignon and Andrew Neild provided invaluable insights concerning the ranges of Junonia species. Thanks to Jason Depew, Phil Lounibos, Marc Minno, Bill Perry, Sue Perry, Mark Salvato, and Don Stillwaugh for their logistical support in collecting butterflies in Florida. Travis Evans, Amber Harper, Sarah House, Tia Hughes, Brooke Polen Jackson, Joey Marquardt, Amanda Maupin, Tara Powell Cox, Timothy Shehan, and Joey Simmons collected butterflies and processed butterfly DNA samples. Thanks to Roohollah Abbasi, Kahlia Beaudette, Tanja Borchers, Ashley Haverstick, Moiz Kapasi, Melanie Lalonde, Bonnie McCullagh, Jacob Miller, Melissa Peters, Stephanie Rozbacher, and Colin Rumbolt for advice and assistance in the laboratory. The manuscript benefitted greatly from the insights of Christian Brévignon, Andrew Neild, Darren Irwin, Thomas Simonsen, and one anonymous reviewer. Butterflies were collected under permits from the Pinellas County (Florida) Department of Environmental Management, the Texas State Parks, the Florida Department of Environmental Protection Division of Recreation and Parks (Permits: 5-04-31, 03090510, 06120610a), Everglades National Park (Permit: EVER-2007-SCI-0018), National Key Deer Refuge (Permit: 41580-2006-19), and Aransas National Wildlife Refuge (Permit: FY07026). Junonia were imported into Canada under permits P-2009-04040, P-2010-01130, and P-2010-01464 from the Canada Food Inspection Agency. This project was part of a research program supported by NSERC Discovery Grant RGPIN386337-2011. Support for this project was provided by a University of Manitoba Undergraduate Research Award and an NSERC-USRA (to A.P.G.). This work was also supported by grants from the National Science Foundation and the Commonwealth of Kentucky (EPS-0132295 and 0447479), the US Environmental Protection Agency (X796463906-0), the National Institutes of Health and the National Center for Research Resources (P20 RR16481), the Canada Foundation for Innovation (212382), and the Canada Research Chair program (950-212382). This paper is subject to the NIH Public Access Policy.

References

  1. Austin GT, Emmel JF. New subspecies of butterflies (Lepidoptera) from Nevada and California. In: Emmel TC, editor. Systematics of western North American butterflies. Gainesville, FL: Marioposa Press; 1998. pp. 501–522. [Google Scholar]
  2. Bandelt HJ, Forster P, Röhl A. Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution. 1999;16:37–48. doi: 10.1093/oxfordjournals.molbev.a026036. [DOI] [PubMed] [Google Scholar]
  3. Beaudette K, Hughes TM, Marcus JM. Improved injection needles facilitate germ-line transformation of the buckeye butterfly Junonia coenia. Biotechniques. 2014;56:142–144. doi: 10.2144/000114147. [DOI] [PubMed] [Google Scholar]
  4. Bock DG, Kane NC, Ebert DP, Rieseberg LH. Genome skimming reveals the origin of the Jerusalem Artichoke tuber crop species: neither from Jerusalem nor an artichoke. New Phytologist. 2014;201:1021–1030. doi: 10.1111/nph.12560. [DOI] [PubMed] [Google Scholar]
  5. Borchers TE, Marcus JM. Genetic population structure of buckeye butterflies (Junonia) from Argentina. Systematic Entomology. 2014;39:242–255. [Google Scholar]
  6. Brévignon L, Brévignon C. Nouvelles observations sur le genre Junonia en Guyane Française. (Lepidoptera: Nymphalidae) (Seconde partie) Lépidoptères de Guyane. 2012;7:8–35. [Google Scholar]
  7. Brock JP, Kaufman K. Butterflies of North America. Houghton Mifflin New York: 2003. [Google Scholar]
  8. Brower AVZ, DeSalle R. Patterns of mitochondrial versus nuclear DNA sequence divergence among nymphalid butterflies: The utility of wingless as a source of characters for phylogenetic inference. Insect Molecular Biology. 1998;7:73–82. doi: 10.1046/j.1365-2583.1998.71052.x. [DOI] [PubMed] [Google Scholar]
  9. Brown FM, Heineman B. Jamaica and its Butterflies. London: E. W. Classey, Ltd.; 1972. [Google Scholar]
  10. Brown JW, Real HG, Faulkner DK. Butterflies of Baja California: Faunal Servey, Natural History, Conservation Biology. Lepidoptera Research Foundation; 1992. [Google Scholar]
  11. Camara MD. A recent host range expansion in Junonia coenia Huebner (Nymphalidae): Oviposition preference, survival, growth, and chemical defense. Evolution. 1997;51:873–884. doi: 10.1111/j.1558-5646.1997.tb03669.x. [DOI] [PubMed] [Google Scholar]
  12. Cech R, Tudor G. Butterflies of the East Coast. Princeton, NJ: Princeton Univ. Press; 2005. [Google Scholar]
  13. Corbet AS. Papers on Malaysian Rhopalocera. V. The conspecificity of the American Precis lavinia (Cramer) with the Oriental Precis orithya (Cramer) Entomologist. 1948;81:54–56. [Google Scholar]
  14. Cui R, Schumer M, Kruesi K, Walter R, Andolfatto P, Rosenthal GG. Phylogenomics reveals extensive reticulate evolution in Xiphophorus fishes. Evolution. 2013;67:2166–2179. doi: 10.1111/evo.12099. [DOI] [PubMed] [Google Scholar]
  15. Dhungel B, Ohno Y, Matayoshi R, Otaki JM. Baculovirus-mediated gene transfer in butterfly wings in vivo: an efficient expression system with an anti-gp64 antibody. BMC Biotechnology. 2013;13:27. doi: 10.1186/1472-6750-13-27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Elster C, Perdomo L, Polania J, Schnetter M-L. Control of Avicennia germinans recruitment and survival by Junonia evarete larvae in a disturbed mangrove forest in Colombia. 1999;15:791–805. [Google Scholar]
  17. Escobedo E. Junonia evarete cytochrome oxidase subunit I isolates from Quintana Roo, Mexico.Genbank Accessions GU659425-GU659427, GU659429-GU659432, GU659435-GU659436, HQ990188. 2011 http://www.ncbi.nlm.nih.gov/nuccore/321485383.
  18. Folmer O, Black MB, Hoch W, Lutz RA, Vrijehock RC. DNA primers for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotechnology. 1994;3:294–299. [PubMed] [Google Scholar]
  19. Forbes WTM. Variation in Junonia lavinia (Lepidoptera, Nymphalidae) Journal of the New York Entomological Society. 1928;36:306–321. [Google Scholar]
  20. Gemmell AP, Borchers TE, Marcus JM. Genetic Population Structure of Buckeye Butterflies (Junonia) from French Guiana, Martinique, and Guadeloupe. Psyche. 2014;2014:1–21. [Google Scholar]
  21. Gernaat HBPE, Beckles BG, van Andrel T. Butterflies of Suriname: A Natural History. Netherlands: KIT Publishers, Berkel en Rodenrijs; 2012. [Google Scholar]
  22. Glassberg J, Minno MC, Calhoun JV. Butterflies through binoculars: A field, finding, and gardening guide to butterflies in Florida. Oxford, New York: 2000. [Google Scholar]
  23. Guzmán BE, Nuñez JJ, Vejar A, Barringa EH, Gallardo CS. Genetic diversity and population structure of two South American marine gastropods, Crepipatella dilatata and C. fecunda (Gastropoda: Calyptraeidae): distinct patterns based on developmental mode. Italian Journal of Zoology. 2011;78:444–454. [Google Scholar]
  24. Hafernik JE. Phenetics and ecology of hybridization in buckeye butterflies (Lepidoptera: Nymphalidae) University of California Publications in Entomology. 1982;96:1–109. [Google Scholar]
  25. Harris AC. A large migration of the Australian meadow argus butterfly Junonia villida calybe (Lepidopera: Nymphalidae) to southern New Zealand. New Zealand Entomologist. 1988;11:67–68. [Google Scholar]
  26. Hebert PDN, deWaard JR, Landry JF. DNA barcodes for 1/1000 of the animal kingdom. Biology Letters. 2010;6:359–362. doi: 10.1098/rsbl.2009.0848. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hebert PDN, deWaard JR, Zakharov EV, Prosser SWJ, Sones JE, McKeown JTA, Mantle B, La Salle J. A DNA ‘Barcode Blitz’: Rapid digitization and sequencing of a natural history collection. PLOS One. 2013;8 doi: 10.1371/journal.pone.0068535. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Irwin DE, Irwin JH, Price TD. Ring species as bridges between microevolution and speciation. Genetica. 2001;112–113:223–243. [PubMed] [Google Scholar]
  29. Irwin DE, Bensch S, Irwin JH, Price TD. Speciation by distiance in a ring species. 2005;307:414–416. doi: 10.1126/science.1105201. [DOI] [PubMed] [Google Scholar]
  30. Irwin DE, Thimgan MP, Irwin JH. Call divergence is correlated with geographic and genetic distance in greenish warblers (Phylloscopus trochiloides): a strong role for stochasticity in signal evolution? Journal of Evolutionary Biology. 2008;21:435–448. doi: 10.1111/j.1420-9101.2007.01499.x. [DOI] [PubMed] [Google Scholar]
  31. Irwin DE. Local adaptation along smooth ecological gradients causes phylogeographic breaks and phenotypic clustering. 2012;180:35–49. doi: 10.1086/666002. [DOI] [PubMed] [Google Scholar]
  32. Janzen DH, Hajibabaei M. Junonia evarete cytochrome oxidase subunit I isolates from Area de Conservacion, Guanacaste, Costa Rica. Genbank Accessions GU1572880-157300, GU334033-334037. 2009 http://www.ncbi.nlm.nih.gov/nuccore/GU157280.1.
  33. Janzen DH. Junonia evarete cytochrome oxidase subunit I isolates from Area de Conservacion, Guanacaste, Costa Rica. Genbank Accessions JQ529754, JQ535680-JQ535697, JQ536229, JQ539189. 2012 http://www.ncbi.nlm.nih.gov/nuccore/JQ529754.1.
  34. Jeratthitikul E, Hara T, Yago M, Itoh T, Wang M, Usami SI, Hikida T. Phylogeography of Fischer’s blue, Tongeia fischeri in Japan: Evidence for introgressive hybridization. Molecular Phylogenetics and Evolution. 2013;66:316–326. doi: 10.1016/j.ympev.2012.10.009. [DOI] [PubMed] [Google Scholar]
  35. Keller I, Wagner CE, Greuter L, Mwaiko S, Selz OM, Sivasundar A, Wittwer S, Seehausen O. Population genomic signatures of divergent adaptation, gene flow and hybrid speciation in the rapid radiation of Lake Victoria cichlid fishes. Molecular Ecology. 2013;22:2848–2863. doi: 10.1111/mec.12083. [DOI] [PubMed] [Google Scholar]
  36. Knowlton N, Weigt LA. New dates and new rates for divergence across the Isthmus of Panama. Proceedings of the Royal Society of London Series B-Biological Sciences. 1998;265:2257–2263. [Google Scholar]
  37. Kodandaramaiah U, Wahlberg N. Out-of-Africa origin and dispersal-mediated diversification of the butterfly genus Junonia (Nymphalidae: Nymphalinae) Journal of Evolutionary Biology. 2007;20:2181–2191. doi: 10.1111/j.1420-9101.2007.01425.x. [DOI] [PubMed] [Google Scholar]
  38. Kodandaramaiah U. Eyespot evolution: Phylogenetic insights from Junonia and related butterfly genera (Nymphalidae: Junoniini) Evolution & Development. 2009;11:489–497. doi: 10.1111/j.1525-142X.2009.00357.x. [DOI] [PubMed] [Google Scholar]
  39. Larkin MA, Blackshields G, Brown NP, et al. Clustal W and Clustal X version 2.0. Bioinformatics. 2007;23:2947–2948. doi: 10.1093/bioinformatics/btm404. [DOI] [PubMed] [Google Scholar]
  40. Liebers D, de Knijff P, Helbig AJ. The herring gull complex is not a ring species. Proceedings of the Royal Society of London Series B-Biological Sciences. 2004;271:893–901. doi: 10.1098/rspb.2004.2679. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Maeki K, Remington CL. Studies of the chromosomes of North American Rhopalocera. 4. Nymphalidae, Charaxidae, Libytheinae. Journal of the Lepidopterists' Society. 1960;14:179–201. [Google Scholar]
  42. Marcus JM, McCune AR. Ontogeny and phylogeny in the northern swordtail clade of Xiphophorus. Systematic Biology. 1999;48:491–522. [Google Scholar]
  43. Marcus JM. Jumping genes and AFLP maps: Transforming Lepidopteran color pattern genetics. Evolution & Development. 2005;7:108–114. doi: 10.1111/j.1525-142X.2005.05012.x. [DOI] [PubMed] [Google Scholar]
  44. Marcus JM, Bell DB, Bryant AN, et al. The Upper Green River Barcode of Life Project. Journal of the Kentucky Academy of Science. 2009;70:75–83. [Google Scholar]
  45. Mayr E. Systematics and the Origin of Species. New York: Dover Publications; 1942. [Google Scholar]
  46. Mayr E. Animal species and evolution. Cambridge, MA: Harvard University Press; 1963. [Google Scholar]
  47. Menezes LFTd, Peixoto AL. Leaf damage in a mangrove swamp at Sepetiba Bay. Vol. 32. Brazil: Rio de Janeiro; 2009. pp. 715–724. [Google Scholar]
  48. Minno MC, Emmel TC. Butterflies of the Florida Keys. Gainesville, Florida: Mariposa Press, Scientific Publishers; 1993. [Google Scholar]
  49. Mitter KT. Junonia genoveva isolate AV-91-0109 cytochrome oxidase subunit I. Genbank Accession KF491814. 2013 http://www.ncbi.nlm.nih.gov/nuccore/558477287.
  50. Neild AFE. The Butterflies of Venezuela, Part 2: Nymphalidae II (Acraeinae, Libytheinae, Nymphalinae, Ithomiinae, Morphinae) Meridian, London: 2008. [Google Scholar]
  51. Nijhout HF. Molecular and Physiological basis of color pattern formation. Advances in Insect Physiology. 2010;38:219–265. [Google Scholar]
  52. Paulsen SM. Quantitative genetics of butterfly wing color patterns. Developmental Genetics. 1994;15:79–91. [Google Scholar]
  53. Paulsen SM. Quantitative genetics of the wing color pattern in the buckeye butterfly (Precis coenia and Precis evarete): evidence against the constancy of G. Evolution. 1996;50:2360–2372. doi: 10.1111/j.1558-5646.1996.tb03931.x. [DOI] [PubMed] [Google Scholar]
  54. Pereira RJ, Wake DB. Genetic leakage after adaptive and nonadaptive divergence in the Ensatina escholtzii ring species. Evolution. 2009;63:2288–2301. doi: 10.1111/j.1558-5646.2009.00722.x. [DOI] [PubMed] [Google Scholar]
  55. Pfeiler E. Confirmation of black mangrove Avicennia germinans (L.) L. as a larval host for Junonia genoveva (Cramer) (Nymphalidae: Nymphalinae) from Sonora, Mexico. Journal of the Lepidopterists Society. 2011;65:187–190. [Google Scholar]
  56. Pfeiler E, Johnson S, Markow TA. Insights into population origins of Neotropical Junonia (Lepidoptera: Nymphalidae: Nymphalinae) based on mitochondrial DNA. Psyche. 2012a;2012:1–6. [Google Scholar]
  57. Pfeiler E, Johnson S, Markow TA. DNA barcodes and insights into the relationships and systematics of buckeye butterflies (Nymphalidae: Nymphalinae: Junonia) from the Americas. Journal of the Lepidopterists' Society. 2012b;66:185–198. [Google Scholar]
  58. Posada D, Crandall KA. Intraspecific gene genealogies: trees grafting into networks. Trends in Ecology & Evolution. 2001;16:37–45. doi: 10.1016/s0169-5347(00)02026-7. [DOI] [PubMed] [Google Scholar]
  59. Remington CL. Genetical differences in solutions of the crises of hybridization and competition in early sympatry. Bolletino di Zoologia. 1985;52:21–43. [Google Scholar]
  60. Rieseberg LH, Kim S-C, Randell RA, Whitney KD, Gross BL, Lexer C, Clay K. Hybridization and the colonization of novel habitats by annual sunflowers. Genetica. 2007;129:149–165. doi: 10.1007/s10709-006-9011-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Schneider S, Roessli D, Excoffier L. Arlequin ver. 2000: A software for population genetic data analysis. University of Geneva, Geneva, Switzerland: Genetics and Biometry Laboratory; 2000. [Google Scholar]
  62. Scott JA. The butterflies of North America : a natural history and field guide. Stanford, CA: Stanford University Press; 1986. [Google Scholar]
  63. Seehausen O. Patterns in fish radiation are compatible with Pleistocene desiccation of Lake Victoria and 14,600 year history for its cichlid species flock. Proceedings of the Royal Society of London Series B-Biological Sciences. 2002;269:491–497. doi: 10.1098/rspb.2001.1906. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Seitz A. Die Gross Schmetterlinge der Erde. Stuttgart: Alfred Kernen; 1914. [Google Scholar]
  65. Sequencher . Version 4.5. Ann Arbor, MI: Gene Codes Corporation; 2005. [Google Scholar]
  66. Shi Q, Huang D, Wang Y, Hao J. The complete mitochondrial genome of Blue Pansy, Junonia orithya (Lepidoptera: Nymphalidae: Nymphalinae) Mitochondrial DNA Early Online: 1–2: DOI: 10.3109/19401736.19402013.19823182. 2013 doi: 10.3109/19401736.2013.823182. Early Online 1–2. [DOI] [PubMed] [Google Scholar]
  67. Sotka EE, Palumbi SR. The use of genetic clines to estimate dispersal distances of marine larvae. Ecology. 2006;87:1094–1103. doi: 10.1890/0012-9658(2006)87[1094:tuogct]2.0.co;2. [DOI] [PubMed] [Google Scholar]
  68. Swofford DL. PAUP*, Phylogenetic analysis using parsimony (*and other methods) Sunderland, Massachusetts: Sinauer Associates; 1998. [Google Scholar]
  69. Teacher AGF, Griffiths DJ. HapStar: automated haplotype network layout and visualization. Molecular Ecology Resources. 2011;11:151–153. doi: 10.1111/j.1755-0998.2010.02890.x. [DOI] [PubMed] [Google Scholar]
  70. Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research. 1994;22:4673–4680. doi: 10.1093/nar/22.22.4673. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Turner TW, Parnell JR. The identification of two species of Junonia Hübner (Lepidoptera: Nymphalidae): J. evarete and J. genoveva in Jamaica. Journal of Research on the Lepidoptera. 1985;24:142–153. [Google Scholar]
  72. Vane-Wright RI, Tennent WJ. Colour and size variation in Junonia villida (Lepidoptera, Nymphalidae): subspecies or phenotypic plasticity? Systematics and Biodiversity. 2011;9:289–305. [Google Scholar]
  73. Wake DB. Speciation in the round. Nature. 2001;409:299–300. doi: 10.1038/35053264. [DOI] [PubMed] [Google Scholar]
  74. Williams CB. The Migration of Butterflies. Edinburgh: Oliver and Boyd; 1930. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp TableS1
NIHMS650555-supplement-Supp_TableS1.xlsx (113.1KB, xlsx)
Supp TableS2
NIHMS650555-supplement-Supp_TableS2.xlsx (44.7KB, xlsx)

RESOURCES