Blackwell Science, LtdOxford, UKBIJBiological Journal of the Linnean Society 0024-4066The Linnean Society of London, 2005? 2005 862 227251 Original Article

PHYLOGENY OF NYMPHALINAE N. WAHLBERG ET AL Biological Journal of the Linnean Society, 2005, 86, 227–251. With 5 figures .

Phylogenetic relationships and historical biogeography of tribes and genera in the subfamily Nymphalinae (Lepidoptera: Nymphalidae)

NIKLAS WAHLBERG1*, ANDREW V. Z. BROWER2 and SÖREN NYLIN1

1Department of Zoology, Stockholm University, S-106 91 Stockholm, Sweden 2Department of Zoology, Oregon State University, Corvallis, Oregon 97331–2907, USA

Received 10 January 2004; accepted for publication 12 November 2004

We infer for the first time the phylogenetic relationships of genera and tribes in the ecologically and evolutionarily well-studied subfamily Nymphalinae using DNA sequence data from three genes: 1450 bp of cytochrome oxidase subunit I (COI) (in the mitochondrial genome), 1077 bp of elongation factor 1-alpha (EF1-a) and 400–403 bp of wing- less (both in the nuclear genome). We explore the influence of each gene region on the support given to each node of the most parsimonious tree derived from a combined analysis of all three genes using Partitioned Bremer Support. We also explore the influence of assuming equal weights for all characters in the combined analysis by investigating the stability of clades to different transition/transversion weighting schemes. We find many strongly supported and stable clades in the Nymphalinae. We are also able to identify ‘rogue’ taxa whose positions are weakly supported (the different gene regions are in conflict with each other) and unstable. Our main conclusions are: (1) the tribe Coeini as currently constituted is untenable, and Smyrna, Colobura and Tigridia are part of Nymphalini; (2) ‘Kallimini’ is paraphyletic with regard to Melitaeini and should be split into three tribes: Kallimini s.s., Junoniini and Victorinini; (3) Junoniini, Victorinini, Melitaeini and the newly circumscribed Nymphalini are strongly supported monophyletic groups, and (4) Precis and Junonia are not synonymous or even sister groups. The species Junonia coenia, a model system in developmental biology, clearly belongs in the genus Junonia. A dispersal-vicariance analysis suggests that dispersal has had a major effect on the distributions of extant species, and three biotic regions are identified as being centres of diversification of three major clades: the Palaearctic for the Nymphalis-group, the Afrotropics for Junoniini and the Nearctic for Melitaeini. © 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251.

ADDITIONAL KEYWORDS: DIVA – molecular phylogeny – Partitioned Bremer Support – sensitivity analysis.

INTRODUCTION 2001; Wahlberg, 2001). Despite the long-standing interest in these butterflies, the phylogenetic relation- Butterflies belonging to the currently recognized sub- ships among the various tribes and genera have family Nymphalinae (Wahlberg, Weingartner & Nylin, remained remarkably obscure. Improving our under- 2003b) have contributed extensively to our knowledge standing of the phylogenetic resolution of such scien- of ecological and evolutionary processes, from hybrid tifically popular taxa should be a high priority, so that zones and ring-species (Forbes, 1928; Silberglied, this abundance of knowledge can be placed in an evo- 1984; Dasmahaptra et al., 2002; Austin et al., 2003), lutionary framework. metapopulation dynamics (Hanski, 1999) and evolu- Since Nymphalinae is the type subfamily of the tionary developmental biology (Carroll et al., 1994; diverse family Nymphalidae, its delineation has Brakefield et al., 1996), to insect–plant interactions enjoyed a dynamic history, as various authors have (Singer, 1971; Nylin, 1988; Janz, Nylin & Nyblom, considered diverse subsets of tribes and genera to rep- resent ‘typical nymphalids’. This confusion has led to several competing classification schemes based on dif- *Corresponding author. E-mail: [email protected] ferent data sets (Ackery, 1988; Harvey, 1991; Kuznet-

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 227

228 N. WAHLBERG ET AL. zov & Stekolnikov, 2001; Wahlberg et al., 2003b). The however, several cladistic analyses have been pub- higher systematics of Nymphalidae is still in a state of lished, based on either morphology (DeVries et al., flux and the delineation of Nymphalinae has not yet 1985; de Jong et al., 1996; Penz & Peggie, 2003; Freitas reached stability, though there is a growing consensus & Brown, 2004) or DNA sequences (Weller et al., 1996; that the classification of Harvey (1991) (based on the Brower, 2000; Wahlberg et al., 2003b). Three of these classification of Müller, 1886), with the addition of the studies (Brower, 2000; Wahlberg et al., 2003b; Freitas tribe Coeini (= Coloburini of authors), currently pro- & Brown, 2004) sampled enough representatives of vides the most natural definition of the subfamily Nymphalidae to provide evidence bearing upon the (Brower, 2000; Wahlberg et al., 2003b; Freitas & monophyly and circumscription of the Nymphalinae. Brown, 2004). This concept of Nymphalinae comprises In all three, sampled taxa belonging to Nymphalini, a monophyletic group that includes the supposedly Coeini, Kallimini and Melitaeini form a monophyletic well-defined tribes Nymphalini, Coeini, Melitaeini group, with Coeini as the sister group to Nymphalini. and Kallimini. It is this hypothesis of nymphaline Freitas & Brown (2004) and Wahlberg et al. (2003b) relationships that we take as our point of departure in suggest an association of Kallimini with Melitaeini, our analysis and discussion. while Brower’s (2000) study has a basal, paraphyletic Ehrlich’s (1958) influential paper on the classifica- Kallimini, with regard to Melitaeini and Nymphalini + tion of butterflies included a much broader concept of Coeini. The sister group to the Nymphalinae remains Nymphalinae, which was based mainly upon symple- in doubt, with Freitas & Brown (2004) proposing Hel- siomorphic and homoplastic characters and appeared iconiinae, Brower (2000) suggesting Biblidinae + Apa- to comprise those taxa that do not fall into any of his turinae and Wahlberg et al.’s (2003b) data implying more restricted and homogeneous nymphalid subfam- Apaturinae. ilies (Danainae, Ithomiinae, Satyrinae, Morphinae, Within Nymphalinae, several taxa have received Calinaginae, Charaxinae, Acraeinae). The taxa in Ehr- attention from systematists in recent years. The rela- lich’s Nymphalinae are today considered to represent tionships among genera and species groups have been several different subfamilies, including Heliconiinae, investigated in Melitaeini (Kons, 2000; Wahlberg & Limenitidinae, Biblidinae and Apaturinae. Several Zimmermann, 2000) and Nymphalini (Nylin et al., subsequent authors have used much the same group 2001; Wahlberg & Nylin, 2003) using both molecular of tribes and genera to represent the ‘core nymphalids’ and morphological data. In Melitaeini, it has become (e.g. Ackery, 1984; Scott, 1985; Scott & Wright, 1990), clear that Harvey’s (1991) proposed division into three though often splitting off some components as distinct subtribes (Euphydryina, Melitaeina, Phyciodina) is subfamilies. At the other extreme, some authors have not satisfactory. Wahlberg & Zimmermann (2000), delineated Nymphalinae in a very narrow sense, to based on mtDNA, found that the Euphydryas-, Meli- include only the tribes Nymphalini and Kallimini taea-, Chlosyne- and Phyciodes-groups of species and (Ackery, 1988) or Nymphalini and Melitaeini (Clark, genera were of equal status. Kons (2000), using mor- 1948). phology, also found these four major groups and All of the above classifications have suffered from a additionally described a fifth group containing Gna- lack of clearly described morphological synapomor- thotriche species. However, the relationships of the phies that diagnose the various circumscriptions. Har- four major groups are in conflict between these two vey (1991) proposed a classification of the family studies. Wahlberg and Zimmermann inferred the Mel- Nymphalidae based on a set of larval characters, that itaea-group to be sister to the Chlosyne-group, was accepted by many authors (e.g. Ackery et al., whereas Kons found Phyciodina and Gnathotriche to 1999). He placed the tribes Nymphalini, Kallimini and be sister to the Chlosyne-group. Melitaeini together, based on the arrangement of The two studies on Nymphalini (Nylin et al., 2001; spines on the larvae, but was unable to resolve the Wahlberg & Nylin, 2003) have established the mono- relationships among them due to character conflict phyly of the the Nymphalis-group of genera (i.e. within the subfamily. Harvey (1991) placed the genus Aglais, Nymphalis and Polygonia) and the close rela- Amnosia in the tribe Kallimini, but Wahlberg et al. tionship between Araschnia, Mynes and Symbrenthia, (2003b) showed that Amnosia does not belong in but otherwise the relationships of genera within Nymphalinae, but in Cyrestinae along with other Nymphalini remain unclear. A further recent morpho- members of Pseudergolini (with which Amnosia was logical study of Symbrenthia, Mynes and Araschnia affiliated prior to being moved to Kallimini (Ackery, (Fric, Konvicka & Zrzavy, 2004) suggests that Mynes 1988). is nested within Symbrenthia. Most historical classifications of the Nymphalidae In addition to these tribal-level studies, species- have been intuitive rather than the product of rigor- level phylogenetic studies have been done for the gen- ous phylogenetic analysis of character state distribu- era Euphydryas (Zimmermann, Wahlberg & Desci- tions formalized in a data matrix. More recently, mon, 2000), Chlosyne (Kons, 2000), Hypanartia

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251

PHYLOGENY OF NYMPHALINAE 229

(Willmott, Hall & Lamas, 2001), Anartia (Blum, Ber- in a 20 mL reaction volume. The cycling profile for COI mingham & Dasmahapatra, 2003), Phyciodes (Wahl- and wingless was 95 ∞C for 5 min, 35 cycles of 94 ∞C for berg, Oliveira & Scott, 2003a) and Symbrenthia, 30 s, 47 ∞C for 30 s, 72 ∞C for 1 min 30 s and a final Mynes and Araschnia (Fric et al., 2004). Of these, only extension period of 72 ∞C for 10 min. The cycling pro- Kons (2000), Wahlberg et al. (2003a) and Fric et al. file for EF1-a was 95 ∞C for 7 min, 35 cycles of 95 ∞C (2004) included sufficiently extensive outgroup sam- for 1 min, 55 ∞C for 1 min, 72 ∞C for 2 min and a final pling to test the monophyly of the genus being studied. extension period of 72 ∞C for 10 min. For all three From this brief review of the current state of genes, the PCR primers were also used for sequencing. nymphaline systematics, it is clear that many ques- In addition, we developed two internal primers for tions remain unanswered. The relationships among sequencing: (EFmid 5¢-CAA TAC CRC CRA TTT genera in Kallimini and Coeini have never been inves- TGT-3¢) for EF1-a and (Patty 5¢-ACW GTW GGW tigated, and relationships within Melitaeini and GGA TTA ACW GG-3¢) for COI. Sequencing was done Nymphalini are still contentious. In this study we test with a Beckman-Coulter CEQ2000 or CEQ8000 capil- the monophyly of the most recent definition of Nymph- lary sequencer (Bromma, Sweden). We checked the alinae, comprising Nymphalini, Coeini, Kallimini and resulting chromatograms using BioEdit (Hall, 1999) Melitaeini (Wahlberg et al., 2003b), and endeavour to and aligned the sequences by eye. Clear heterozygous resolve the relationships among the tribes and genera positions in the nuclear genes (chromatogram peaks within the subfamily. In the current circumscription, almost or exactly equal) were coded according to the the subfamily Nymphalinae comprises about 496 spe- IUPAC ambiguity codes, but were treated as missing cies in 56 genera. We have studied the relationships of characters in further analyses. The sequences have representative species belonging to the subfamily with been submitted to GenBank (Accession numbers in DNA sequences from three gene regions. These are Appendix 1). cytochrome oxidase subunit I (COI) from the mito- We searched for the most parsimonious cladograms chondrial genome, and elongation factor-1a (EF1-a) from the equally weighted and unordered data matrix and wingless from the nuclear genome. Based on our consisting of 189 taxa using a heuristic search algo- results, we investigate the biogeography of the group rithm in NONA 2.0 (Goloboff, 1998) via WINCLADA using dispersal-vicariance analysis (Ronquist, 1997). 1.00.08 (Nixon, 2002). The heuristic searches were The investigation is intended to identify broad biogeo- conducted with 1000 random addition replicates using graphical patterns for closer inspection at a later date. TBR branch swapping with ten trees held during each step and a final swapping to completion. We did this for each gene separately and for all three genes com- MATERIAL AND METHODS bined. Trees were rooted with Heliconius for display. We sampled as many species as possible from almost For the separate analyses, we evaluated clade support all of the genera belonging to Nymphalinae, a total of using bootstrap with 100 pseudoreplicates. It is now 161 species in 49 genera. Of the seven missing genera, widely recognized that assessing incongruence among six are in Melitaeini and one is in Coeini. In addition, data partitions is much more complex than measuring we sampled 28 outgroup species, representing all sub- with a simple all-or-nothing significance test (Farris families of the nymphaline clade (sensu Wahlberg et al., 1994; DeSalle & Brower, 1997; Miller et al., et al., 2003b), i.e. Cyrestinae, Biblidinae and Apaturi- 1997; Darlu & Lecointre, 2002). We have thus chosen nae, each of which may be the sister group to Nymph- to analyse the three gene regions as a single data set, alinae (see Wahlberg et al., 2003b). We also included a and have assessed the impact of each gene region on specimen of Heliconiinae and Limenitidinae, which the support values of each node using Partitioned belong to the putative sister clade to the nymphaline Bremer Support (PBS) analyses (Baker & DeSalle, clade. The species sampled and their collection locali- 1997; Baker et al., 1998). ties are listed in Appendix 1. We evaluated the character support for the clades in We extracted DNA mainly from one or two legs of the resulting cladograms using Bremer support (BS) freshly frozen or dried butterflies using QIAgen’s (Bremer, 1988, 1994). We calculated BS and PBS DNEasy extraction kit, although some specimens values using anticonstraints in PAUP* 4.0b10 for were extracted following the protocol of Brower (1994). Windows (Swofford, 2001). As in previous studies The spread voucher specimens can be viewed at http:/ (Wahlberg & Nylin, 2003; Wahlberg et al., 2003b), we /www.zoologi.su.se/research/wahlberg/. For each spec- refer to the support values as giving weak, moderate, imen we sequenced 1450 bp of COI, 1077 bp of EF1-a good or strong support when discussing our results. and 400–403 bp of the wingless gene. Primers for COI We define ‘weak support’ as BS values of 1–2 (gener- were taken from Wahlberg & Zimmermann (2000), for ally corresponding to bootstrap values <50%–63%), EF1-a from Monteiro & Pierce (2001) and for wingless ‘moderate support’ as values between 3 and 5 (boot- from Brower & DeSalle (1998). We performed all PCRs strap values 64%-75%), ‘good support’ as values

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 230 N. WAHLBERG ET AL. between 6 and 10 (bootstrap values 76%-88%) and occurred between positions 91 and 137 in the Colobura ‘strong support’ as values >10 (bootstrap values 89%- dirce sequence (GenBank accession number 100%). We strongly endorse BS values over bootstrap AY090162) and were easily detected when aligning by values because they are a parameter of the data eye (flanking regions were relatively conserved). Gaps rather than an estimate based on manipulated sub- were treated as a ‘fifth base pair’ in all phylogenetic samples of the data, and have no upper bound as sup- analyses. The basic statistics for the three gene port for a given clade increases. regions are given in Table 1. All three gene regions We evaluated the stability (sensu Wheeler, 1995; showed an ample amount of variation, though EF1-a Judd, 1998; Giribet, 2003) of clades inferred for the was less variable than the other two genes. The two equally weighted data set to different character-state nuclear genes show about equal base frequencies, transformation weighting assumptions under parsi- while COI has a high AT bias, in accordance with all mony searches using PAUP*. We weighted transver- previous publications on this gene in insects. sions 2, 3, 5, 7 and 10 times transitions for the combined data set. Such sensitivity analyses may help identify potential instances of long branch attraction GENERAL PHYLOGENETIC PATTERNS (Giribet, 2003), and can provide a valuable heuristic When analysed as separate data partitions, none of tool to guide subsequent sampling strategies for the three gene regions recovered the subfamily refinement of the current hypothesis. We refer to Nymphalinae as a monophyletic entity (data not clades that are recovered under all the tested weight- shown). There are, however, several phylogenetic com- ing schemes as stable. ponents common to the separate hypotheses: Meli- We investigated the historical biogeography of the taeini and Nymphalini (as delineated below) are subfamily using dispersal-vicariance analysis (Ron- monophyletic, the genera Junonia, Precis, Hypolim- quist, 1997) as implemented in DIVA (Ronquist, nas, Yoma, Salamis and Protogoniomorpha together 1996). Species distributions were recorded at the level form a monophyletic group, and the Nymphalis-group of zoological biomes, i.e. Nearctic, Neotropical, of species (see Wahlberg & Nylin, 2003) is monophyl- Afrotropical, Palaearctic, Oriental and Australasian. etic. Clades for which component species had identical dis- Combining the three data sets yields eight trees of tributions were collapsed into a single terminal. The length 17547 steps (CI = 0.13, RI = 0.58), the strict maximum number of ancestral areas was either not consensus of which gives a much clearer picture of constrained or restricted to two. the relationships among the various groups in Nymphalinae (Figs 1–4). The subfamily as currently delimited is inferred to be polyphyletic, with the RESULTS genera Historis and Baeotus branching off close to the base of the larger nymphaline clade (sensu Wahl- CHARACTERISTICS OF THE DATA SETS berg et al., 2003b). The monophyly of Nymphalinae The total combined data set consisted of 2935 nucle- without Historis and Baeotus receives moderate sup- otides, of which 12 positions were coded as gaps in port. Nymphalini and Melitaeini form monophyletic some taxa. All inferred gaps were in the wingless data groups, while Coeini and Kallimini are poly- and set and represent indel events of whole codons. These paraphyletic, respectively. Three genera tradition- include an inferred codon insertion in all Melitaeini ally placed in Coeini (Colobura, Tigridia and sequences (as noted in Brower, 2000), an inferred Smyrna) are associated with Nymphalini with good codon insertion in Tigridia acesta, an inferred codon support, while the two other sampled genera, His- insertion in Rhinopalpa polynice and an inferred toris and Baeotus, are outside Nymphalinae with codon deletion in the three species of Aglais (as noted weak support. Kallimini + Melitaeini has moderate in Nylin et al., 2001). All inferred indel events have support, but the kallimine taxa form a paraphyletic

Table 1. The basic statistics for the three molecular data sets in 174 species of Nymphalidae

Empirical base frequencies (%)

Gene No. of sites No. variable No. informative TCAG

COI 1450 775 614 39.9 14.7 31.9 13.6 EF1-a 1077 469 364 22.6 28.1 25.9 23.4 wingless 412 238 192 21.0 26.2 24.9 27.9

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 PHYLOGENY OF NYMPHALINAE 231

Heliconius Adelpha 30 Marpesia chiron 8 12,10.9,7.1 Marpesia orsilochus Cyrestini 20.7,-9.9,-2.8 10 Chersonesia 16.8,-4.1,-2.7 Cyrestis 1 Historis odius 17 8.4,-2.6,-4.8 Historis acheronta 19 Baeotus japetus Coeini 21.3,2.8,-7.0 Baeotus deucalion 32 28.0,-0.1,-8.9 1 8.7,-3.3,-4.4 4 Baeotus aeilus 13.5, 15.4,-0.2,-11.2 Baeotus beotus 17.1, 1.4 10 Dichorragia 17 28.2,0.8,-19.0 Stibochiona Pseudergolini 32.1,1.2,-16.3 6 Amnosia 1 20.5,1.3,-15.8 Pseudergolis 3 8.5,-2.6,-4.9 31 Apatura 23.9,-6.0,-14.9 17 27.7,11.4,-8.1 Mimathyma 12 Timelaea Apaturinae 7.2,9.1,0.7 Eulaceura 9.6,1.4,1 4 5.5,1.9,-3.4 2 Asterocampa 9.4,-2.2,-5.2 Chitoria 2 Sevenia 7.9,-0.2,-5.7 1 Dynamine 22.1,-6.6,-14.5 Hamadryas 1 13 19 Catonephele 8.8,-2.8,-5 14,3.4,-4.4 19.8,6.5,-7.4 Myscelia 2 2 Panacea Biblidinae -0.6,6.2,-3.6 12 Callicore 0.2,5.1,-3.3 1.5,4.3,6.2 Nica 21 Ariadne Eurytela 18.4,9.4,-6.8 2 -1.0,6.9,-3.9 2 Byblia 0.4,6.0,-4.4 Mesoxantha Colobura 1 9 Tigridia Smyrna 18.2,-5.7,-11.5 22.5,-8.6,-4.9 Antanartia Symbrenthia Araschnia Nymphalini Mynes Hypanartia Vanessa Aglais Nymphalis 4 Polygonia Kallimoides 14.1,-7.4,-2.7 Siproeta Napeocles Victorinini Metamorpha 1 Anartia 22.2,-6.7,-14.5 Rhinopalpa Vanessula Hypolimnas Precis Junonia 5 Salamis Junoniini 18.6,-6.6,-7.0 Yoma Protogoniomorpha 1 Doleschallia Kallima 19.9,-5.4,-13.5 Catacroptera Kallimini 1 Mallika 21.9,-6.5,-14.4 Euphydryas 39 Poladryas 25.6,8.5,4.9 Microtia Chlosyne Melitaea Gnathotriche Higginsius Mazia Melitaeini Tegosa Phyciodes Anthanassa Telenassa Castilia Eresia Janatella

Figure 1. Summary of strict consensus of eight trees found for the combined data set when all changes weighted equally (length = 17547, CI = 0.13, RI = 0.58), pruned to show only genera. For unpruned trees, see Figs 2–4. Numbers above the branches are Bremer support values and numbers below are Partitioned Bremer Support values for the COI, EF1-a and wingless data partitions, respectively. Thickened branches are stable to changes in weighting schemes (transversions weighted 1, 2, 3, 5, 7 and 10 times transitions).

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 232 N. WAHLBERG ET AL.

19 Colobura dirce 33.6,-5.7, Tigridia acesta -8.9 9 Smyrna blomfildia 22.5,-8.6,-4.9 62 Antanartia delius 37, Antanartia schaenia 6 13.4, 11.6 9 Araschnia levana 11.5,4.4,-9.9 3.6,2.7, 10 2.7 Mynes geoffroyi 24.7,-4.4, 11 Symbrenthia hypselis -10.3 15 11 2.5,3.5,5.1 Symbrenthia lilea 2.4,6.6,6.1 12.7,7.2,-8.9 Symbrenthia hypatia 22 Hypanartia bella 17 7,9.9, Hypanartia lethe 5.1 23.9,- 0.6, 21 Hypanartia lindigii -6.3 2 Hypanartia charon 3 21.8,5.3,-6.2 4 2,0.9,-0.9 Hypanartia kefersteini 21.5,-1.5,-16.0 2,1.9, Vanessa annabella -0.9 25 Vanessa gonerilla 8 11,6.9, 7.1 Vanessa itea 2.9,4.3,0.8 1 Antanartia abyssinica 1 -2.8,3.7, 7 Vanessa atalanta 0.1 -2.8,3.7,0.1 2.9,4.6,-0.5 Vanessa indica 7 Vanessa cardui 4 24 7.1,-1.1, 1.1 Vanessa kershawi 19,-1.4,-13.5 9.2,10.2,4.6 5 Vanessa virginiensis 6,- 2.1,1.1 9 Vanessa braziliensis 0.2,5.4,3.4 Vanessa myrinna 16 Aglais io

5,0.9, 26 Aglais milberti 10.1 9,11.4, Aglais urticae 5.6 Polygonia canace 8 1 Nymphalis l album 21.2,-0.3,-12.9 1,-1,1 6 5 Nymphalis polychloros 4.5,-0.6,2.1 Nymphalis antiopa 8 3.2,1.4,0.4 5 Nymphalis californica -4.5,5.4,4.1 2,3.9,2.1 Nymphalis xanthomelas Polygonia c-aureum 2 1 Polygonia egea 3,-2.1,1.1 0.4,0.6, 11 Polygonia c-album 6 0 5.6,4.2,1.2 Polygonia faunus 0.5,4.5,1.1 7 Polygonia interrogationis Polygonia comma 9.3,-2.3,0 4 Polygonia progne 9.9,-5.2,-0.7 7 8,-1.1,0.1 1 Polygonia haroldi Polygonia satyrus 7.3,-4.3,-2.0 1 7.2,-4.3,-1.9 4 Polygonia oreas 5.4,-1.2,-0.2 Polygonia gracilis

Figure 2. Relationships of sampled species from the tribe Nymphalini as delimited in Fig. 1. Tree statistics and branch thickness as in Fig. 1. grade with respect to Melitaeini. The sister group of Results of the PBS analysis show that all three gene Nymphalinae (without Historis and Baeotus) is Bib- regions are congruent at 52 of the 183 resolved nodes lidinae in the most parsimonious trees from the com- (Figs 1–4). These nodes generally have high BS values bined data set, although this clade has weak support (range 3–63, mean 19) and tend to be concentrated in (Fig. 1). the clades describing Nymphalini and Melitaeini. At

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 PHYLOGENY OF NYMPHALINAE 233

Kallimoides rumia

3 16 Siproeta epaphus 19.9,-6.6, 27.5,-1.2, 1 Napeocles jucunda -10.4 -10.3 8 23.3,-8.0,-14.3 Siproeta stelenes 31.4,-6.9,-16.5 1 Metamorpha elissa 16.7,-5.6,-10.1 40 Anartia jatrophae 43.8,-0.7,-3.2 33 Anartia amathea 40.6,0.1,-7.7 Anartia fatima 1 4 Rhinopalpa polynice 22.2,-6.7,-14.5 18.4,-5.3, -9.1 Vanessula milca 7 Hypolimnas anthedon 26.8,-4.9, 17 -14.9 Hypolimnas usambara 9.7,-0.6, Hypolimnas misippus 7.9 11 5.5,4.3,1.2 2 Hypolimnas alimena 7 2,-0.1,0.1 1 Hypolimnas bolina 1 7.5,-0.3, 1,-0.1,0.1 21.8,-6.4,-14.4 -0.2 Hypolimnas pandarus Precis ceryne 25 4 Precis archesia 31.3,1.7,-7.9 23.6,-4.7, 1 -14.9 Precis octavia 23.3,-7.4,-14.9 Precis sinuata 2 2 Precis cuama 1.1,-0.2,1.1 3.0,-2.1, 19 2 1.1 Precis tugela 27.2,1.1,-9.3 3.2,-2.4,1.2 3 Precis andremiaja 4,-1.1,0.1 Precis antilope 62 Salamis anteva 46.6,9.9, 3 5.5 Salamis cacta 6.9,0.5, -4.4 5 Yoma algina

4.2,1.6,-0.8 34 Protogoniomorpha anacardii 5 34,9.9,-9.9 Protogoniomorpha parhassus 21.3,-5.6,-10.7 26 Junonia erigone 20,3.9, 2.1 Junonia iphita 13 3 Protogoniomorpha cytora 14.5,-1.9,0.4 6.0,-2.1, 16 Junonia artaxia -0.9 1 29.2,0.3,-13.5 Junonia touhilimasa 21.9,-6.5,-14.4 Kamilla cymodoce 1 28 Junonia natalica 22,-6.5,-14.5 33.4,-1.5, 1 -3.9 Junonia terea 22.1,-6.6,-14.5 23 Junonia coenia 35.2,-0.5,-11.7 4 Junonia oenone 20.6,-5.6,-11 Junonia sophia

Figure 3. Relationships of sampled species from the tribes Junoniini and Victorinini, as delimited in Fig. 1, as well as species in the genera Kallimoides, Vanessula and Rhinopalpa. Tree statistics and branch thickness as in Fig. 1.

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 234 N. WAHLBERG ET AL.

Doleschallia bisaltide 1 37 Kallima paralekta 19.9,-5.4, 16 37.8,4.3, Kallima inachus -13.5 -5.5 15.7,2.2,-1.9 17 Catacroptera cloanthe 29,-0.6,-11.4 Mallika jacksoni 28 Euphydryas editha 30.6,3.8, 2 Euphydryas chalcedona 36 -6.5 3,-1.1,0.1 Euphydryas phaeton 1 24.7,19.7, Euphydryas gillettii -8.4 9 21.9,-6.5, -14.4 7.3,5.1,-3.3 30 Euphydryas aurinia 20.2,6.7,3.1 Euphydryas desfontainii Poladryas arachne 3 16 Microtia elva 3,7,6 4 Dymasia dymas -0.3,4.2, -0.9 6.9,1.9,-4.8 Texola elada 1 39 Chlosyne janais -2.1,3.9,-0.8 29 25.6,8.5,4.9 10.1,8.6, 22 Chlosyne gaudealis 10.4 17 7.1,14.3,0.6 Chlosyne narva 9.1,8.2,-0.3 6 Chlosyne cyneas Chlosyne theona 10.3,0.1,-4.3 1 Chlosyne nycteis -1.5,3.5,-1 6 8 Chlosyne gorgone 6.2,-0.3,0.1 1 8.7,0.5, Chlosyne lacinia -1.2 22 -1,2.9,-0.9 8 Chlosyne harrisii 21.1,8.8,-7.9 6.6,1.8,-0.4 4 Chlosyne acastus 0,4.9,-0.9 Chlosyne palla 18 Melitaea scotosia 7.6,7.4, Melitaea punica 12 3 Melitaea trivia 3.7,9.1, -0.8 2 13 Melitaea cinxia Melitaea varia -1.3,2.5,0.8 6,5.9, 22 1.1 1 11,8.9,2.1 9 Melitaea britomartis 7,1.9,0.1 1.9,-1,0.1 Melitaea deione Melitaea arduinna 6 2 25 Melitaea persea 9.3,2.1,-5.4 -1.3,2.3,1 6,13.9,5.1 13 Melitaea didymoides 10,0.9,2.1 Melitaea latonigena 16 Gnathotriche exclamationis 3.4,10,2.6 Higginsius fasciata 5 Mazia amazonica Tegosa anieta -1,5.9,0.1 55 31.5,17.7, Tegosa tissoides 33 5.8 Phyciodes graphica 12.9,15.5,4.6 2 Phyciodes cocyta 15 3 -2.9,7, Phyciodes tharos -2.1 5.4,11.7, 6,-2.1, Phyciodes batesii 8 -2.1 -0.9 6 15.2,-0.9,-6.3 28 3,2.9,0.1 Phyciodes pulchella 14.4,11.2,2.4 Phyciodes phaon 2 18 Phyciodes pallescens -1.3,7.9,-4.6 7.9,10.7, Phyciodes picta 2 -0.5 1,4.5,-3.5 Phyciodes orseis 2 11 7.4,5.4,-1.8 4 Phyciodes mylitta 8.3,-0.4,-6.0 4,0.9,-0.9 Phyciodes pallida 4 Anthanassa texana 1.7,1.2, Anthanassa tulcis 3 1.1 0.9,1.1, Anthanassa otanes 1 9 3,3.9,2.1 3 Anthanassa ardys 6 1.2,-0.4,2.2 Anthanassa drusilla 7.7,1.4,-3.1 3 Telenassa trimaculata 2,-0.1, 15 Castilia eranites 1.1 3 4.9,8.6,1.5 Castilia ofella 4 Eresia eunice 4.3,-0.3,-1 2.7,2.2, Janatella leucodesma 5 -0.9 15 Eresia pelonia 5.8,0.6,-1.4 6.9,5, Eresia quintilla 5 3.1 30 Eresia coela 2.5,2.5,0 1 23.2,3.7, Eresia sestia 3.2 3.1,-1.3,-0.8 10 Eresia clio 12,-1.7,-0.3 Eresia letitia

Figure 4. Relationships of sampled species from the tribe Melitaeini and Kallimini as delimited in Fig. 1. Tree statistics and branch thickness as in Fig. 1.

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 PHYLOGENY OF NYMPHALINAE 235

Table 2. Number of nodes in the ingroup at which differ- support, Colobura and Tigridia are sister genera with ent patterns of incongruence were observed. +, PBS score strong support and they are sister to Nymphalini + positive or 0; -, PBS score negative (magnitude not taken Smyrna. The monophyly of Smyrna, Tigridia, into account) Colobura and Nymphalini has good support and is sta- ble, though there is conflict from the nuclear genes Number of nodes COI EF1-a wingless (Fig. 1). Historis and Baeotus form a monophyletic group with strong support and both genera are mono- 47 + + + phyletic (Fig. 1). Historis odius and H. acheronta (the 46 + + –latter sometimes placed in the genus Coea, from which + + 20 – the tribal name is derived) are sister species with only 6–+ + weak support and moderate conflict from the two 56 + –– nuclear genes. Baeotus, on the other hand, is a 0––+ strongly supported monophyletic group with no con- 8–+ – flict from the different data partitions. Constraining the five coeine genera to form a monophyletic group results in trees that are 19 steps longer than the most 131 nodes, one or two gene regions provide signal that parsimonious trees found for the combined data sets. is incongruent with the weight of the combined evi- Constraining Colobura, Tigridia and Smyrna to form dence. The different patterns of incongruence are sum- a monophyletic group results in trees seven steps marized in Table 2. Nodes which show incongruence longer than the most parsimonious trees. tend to have lower BS values (range 1–19, mean 7.7). The association of Colobura, Tigridia and Smyrna Most of the incongruent nodes lie in the basal with Nymphalini is stable under all the different branches of the cladogram (Fig. 1) and in the kal- weighting schemes, as is the clade containing Historis limine grade (Figs 3, 4). and Baeotus species. However, the position of Historis The PBS results show that the data partitions are + Baeotus changes as one increases the weight of contributing unequally to the phylogenetic patterns transversions to transitions. When transversions are inferred in this study. The COI data set conflicts at 17 weighted 2 times transitions, this clade is sister to nodes, of which one node has a PBS score of <-4. The Nymphalinae. At higher tested weighting schemes total PBS for the COI data partition is 2176.1 (mean (transversions weighted 3–10 times transitions), His- 11.9). The EF1-a data partition is incongruent at 70 of toris + Baeotus appears as the sister clade to Apaturi- the 183 resolved nodes, 23 of which have a PBS score nae. of <-4, and has a total PBS of 370.3 (mean 2.0). The wingless data alone have a negative PBS score at 106 of the 183 resolved nodes, 64 of which have a PBS PHYLOGENETIC PATTERNS IN NYMPHALINI score of <-4, and the gene region has a total PBS of - Within the monophyletic Nymphalini, relationships 533.5 (mean -2.9). within the Nymphalis-group of genera (i.e. Nymphalis, Many of the clades are very stable (sensu Giribet, Polygonia and Aglais) are almost identical to those 2003) and are present in a strict consensus of all trees hypothesized in a previous study (Wahlberg & Nylin, found across the different weighting schemes from the 2003). The exception is Polygonia canace, which is sis- equally weighted analysis to the analysis with 10 : 1 ter to Nymphalis in the present study, whereas it was TV/TI weighting (Figs 1–4). These include the larger sister to the rest of Polygonia in the 2003 study. Wahl- nymphaline clade (including Biblidinae, Cyrestinae, berg & Nylin (2003) was based on an additional gene Apaturinae and Nymphalinae; see Wahlberg et al., sequence and a morphological data set and thus we 2003b), Apaturinae and Biblidinae, Pseudergolini, prefer the hypothesis supported in that study. Polygo- Cyrestini, Nymphalini and Melitaeini (Fig. 1), and nia oreas and P. haroldi were not included in previous many of the well-defined genera (Figs 2–4). Weighting phylogenetic studies of the Nymphalis-group, and our transversions 2 times transitions is the only analysis results suggest that the former is related to P. gracilis which gives a monophyletic Nymphalinae, with His- and the latter is sister to a clade containing P. oreas, toris + Baeotus coming out as sister to the rest of P. gracilis and P. satyrus. Nymphalinae. Other genera in Nymphalini are here sampled more broadly than in previous studies and we are now able to identify three additional clades within Nymphalini. PHYLOGENETIC PATTERNS IN COEINI Two species of Antanartia (A. delius and A. schaenia) The tribe Coeini does not form a monophyletic group form the sister group to the rest of Nymphalini. The in any of the analyses (Figs 1, 2). In the combined next most basal clade is formed by the genera Ara- analysis, Smyrna is sister to Nymphalini with good schnia, Mynes and Symbrenthia, which group together

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 236 N. WAHLBERG ET AL. with good support. The third major new clade com- Within the Anartia-clade, Napeocles and Siproeta prises the genera Vanessa and Hypanartia, though form a strongly supported, stable subclade, as do the this clade has only moderate support and is not stable. three species of Anartia. The sister group of Anartia Vanessa + Hypanartia is sister to the Nymphalis- may be either Siproeta/Napeocles or Siproeta/Napeo- group clade. A surprising result is the position of cles + Metamorpha. It is clear that Metamorpha is an Antanartia abyssinica within Vanessa; it is not at all entity distinct from Siproeta (as suggested by Fox & closely related to the other Antanartia species. Within Forbes, 1971), though historically Siproeta stelenes Vanessa, species often placed in the genus Cynthia do has occasionally been placed in Metamorpha. The not form a monophyletic group, as one species, clade containing these neotropical genera is stable in V. annabella (and presumably its putative sister spe- sensitivity analyses. The position of Kallimoides as cies V. carye; Shapiro & Geiger, 1989), appears to be the most basal taxon of this clade has only weak sup- sister to the rest of Vanessa. port and the node is not stable, suggesting that its Sensitivity analyses suggest that the relationships placement requires further investigation. within the Nymphalis-group are very stable; the only The Junonia-clade exhibits some of the greatest sur- node which is unstable is the sister relationship prises of this study. The first is that Junonia and Pre- between Polygonia egea and P. c-album + P. faunus. cis are separate genera that are not even sister Hypanartia is stable, as are the relationships of spe- groups. The genus Kamilla is clearly within Junonia, cies within the genus. Within Vanessa, the sister rela- as was concluded by Shirôzu & Nakanashi (1984) tionship of A. abyssinica to V. atalanta + V. indica is based on morphology (but see Larsen, 1991). In addi- stable, as is the sister species relationship of tion, the species Protogoniomorpha cytora is found V. gonerilla + V. itea, and the V. cardui clade (V. cardui within Junonia, rather than with other speices of Pro- to V. myrinna in Fig. 2). Other stable clades are togoniomorpha or Salamis, with which it has always Antanartia delius + A. schaenia and Symbrenthia. been associated. The sister group relation between Precis and Hypolimnas has good support and is stable, except when transversions are weighted 3 times tran- PHYLOGENETIC PATTERNS IN KALLIMINI sitions, when it is sister to the Salamis + Yoma + Pro- The tribe Kallimini as currently circumscribed does togoniomorpha + Junonia clade. Another surprise is not form a monophyletic group, but rather a paraphyl- that the African Protogoniomorpha, which has been etic grade with regard to Melitaeini. Constraining it to usually considered to be a synonym of Salamis, is the be monophyletic results in trees that are four steps sister group to the Australasian Yoma, and that Sala- longer than the most parsimonious trees found for the mis is the sister group to Protogoniomorpha + Yoma. combined data set. The strict consensus of the most Vári (1979) has argued that species of Protogoniomor- parsimonious trees shows two clades. The basal clade pha should not be considered to be congeneric with contains the following subclades: (1) a largely Neotro- species of Salamis based on genitalic differences, a pical subclade including Anartia, Siproeta, Napeocles position corroborated by our data. The sister group and Metamorpha (termed the Anartia-clade), (2) the relationship of Yoma and Protogoniomorpha has good African Kallimoides, a subclade with Vanessula and support and is stable. To complete the surprise, Sala- Rhinopalpa (both monotypic genera that have not pre- mis, Yoma and Protogoniomorpha, usually considered viously been associated with each other or indeed with to be related to Hypolimnas, are placed in our results any other kallimine genera) and (3) a subclade that as sister group to Junonia, though this clade disap- contains the largely African and Asian genera pears when transversions are weighted 5 or more Junonia, Kamilla, Precis, Hypolimnas, Salamis, Pro- times transitions. togoniomorpha and Yoma (termed the Junonia-clade). The implied position of the Kallima-clade as sister The second clade in the Kallimini grade, which is sis- to the Melitaeini is somewhat surprising given tradi- ter to the tribe Melitaeini, includes the African Catac- tional classifications, but this grouping has weak sup- roptera and Mallika, the Asian Kallima and the port and may be due to long branch attraction. Indeed Australasian Doleschallia (termed the Kallima-clade). the sister group relationship is not stable and disap- Weighting transversions more than transitions pears when transversions are weighted 2 or more causes Kallimoides and Vanessula to become sister times transitions. The relationships between Catac- taxa. These two form the sister clade to Rhinopalpa roptera, Mallika and Kallima are strongly supported and the Anartia- and Junonia-clades when transver- and stable. The strongly supported, stable sister rela- sions are weighted 2 or 3 times transitions. However, tionship of Catacroptera and Mallika (both monotypic) weighting schemes of 5 and above cause Kalliomoides, has been suggested earlier by Shirôzu & Nakanishi Vanessula, Rhinopalpa and the Anartia-clade to (1984). Catacroptera and Mallika are usually associ- become the most basal clades in the entire tree, after ated with Junonia rather than Kallima (Larsen, the outgroups Heliconius and Adelpha. 1991). The recently suggested (Parsons, 1999) associ-

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 PHYLOGENY OF NYMPHALINAE 237 ation of Doleschallia with Kallima is corroborated by DISCUSSION our data, although support is not strong and there is This is only the second comprehensive phylogenetic strong conflict between the nuclear genes and the analysis of the relationships within a nymphalid mitochondrial gene regions. subfamily, and the first to use molecular data. The sole previous attempt to infer relationships within a PHYLOGENETIC PATTERNS IN MELITAEINI nymphalid subfamily is the recently published study by Penz & Peggie (2003) on Heliconiinae. This is not The monophyly of Melitaeini is very strongly sup- altogether surprising, given that the circumscrip- ported and stable, as are Euphydryas, Melitaea, tions of the subfamilies have only recently stabi- Phyciodina and the sister relationship between Gna- lized (Harvey, 1991; Wahlberg et al., 2003b), and thotriche and Higginsius (termed the Gnathotriche- because the high degree of variation in morphologi- group). The Chlosyne-group has moderate support but cal characters among species in Nymphalidae has is unstable, with both the COI and wingless gene confounded previous attempts to delineate natural regions conflicting with the EF1-a data. The genus groups (de Jong et al., 1996). In our study, we have Chlosyne itself is a strongly supported, stable clade. been able to identify clades that are well-supported The position of Euphydryas as the sister group to the by the three gene regions and stable to varied char- rest of the melitaeines is strongly supported and sta- acter state transformation weights, as well as clades ble, in agreement with previous studies (Kons, 2000; that are less robust and therefore likely to change Wahlberg & Zimmermann, 2000). Relationships with the addition of more data. Based on our between the Chlosyne-, Melitaea-, Gnathotriche- results, we are proposing a new classification of the groups and Phyciodina are not well-supported and are subfamily Nymphalinae, which is shown in Appen- unstable. The most parsimonious trees from the dix 2. equally weighted analysis suggests that the Gnathot- riche-group is sister to Phyciodina with moderate sup- port, and that the Melitaea-group is sister to these two with good support (Fig. 4). This arrangement is stable THE CIRCUMSCRIPTION OF NYMPHALINAE when transversions are weighted 2 and 3 times tran- We have found that Nymphalinae as currently sitions, but breaks down at higher weights. The sister circumscribed is not monophyletic, although Nym- relationship of the Gnathotriche-group and Phyciod- phalini, Kallimini and Melitaeini do form a mono- ina is not stable and disappears after weights of 3 phyletic group with the inclusion of three genera times. This study presents a third novel arrangement traditionally placed in Coeini (i.e. Colobura, of these four clades, with Kons (2000) having the Chlo- Tigridia and Smyrna). Two genera also tradition- syne-group sister to Phyciodina + Gnathotriche-group ally placed in Coeini (Historis and Baeotus) are and Wahlberg & Zimmermann (2000) having the clearly not related to these three genera and indeed Chlosyne-group sister to the Melitaea-group. do not appear to be closely related to Nymphali- The positions of two genera (Poladryas and Higgin- nae. Our results are not unprecedented, as Muys- sius) are not in agreement with previous studies. In hondt & Muyshondt (1979) argued that based on contrast to Wahlberg & Zimmermann (2000) but in larval morphology Smyrna should be placed in agreement with Kons (2000), Poladryas is here related Nymphalini and that Colobura is closer to Nympha- to the Chlosyne-group. Morphological evidence that lini than Historis and Baeotus. However, the sug- Poladryas is associated with Higginsius (Kons, 2000) gestion by Muyshondt & Muyshondt (1979) that is quite decisively contradicted by molecular evidence, Coeini is an unnatural group has not been followed which places Higginsius as sister to Gnathotriche with by subsequent authors. On the other hand, Freitas strong support and stability. Within the Chlosyne- & Brown (2004) found that, based on morphologi- group, the genera Microtia, Texola and Dymasia form cal data, Historis, Colobura and Smyrna formed a a strongly supported, stable monophyletic group, as monophyletic group sister to Hypanartia and Van- found by Kons (2000). essa. We were unable to sample the remaining Relationships of the Neotropical Phyciodina (repre- ostensibly coeine genus, Pycina, which has morpho- sented by the genera Mazia, Tegosa, Eresia, Castilia, logical features in larvae and pupae that appear to Telenassa, Janatella and Anthanassa) are generally be intermediate between those in Historis/Baeotus not well-supported or stable at the deeper nodes. Ere- and Colobura (Muyshondt & Muyshondt, 1979). The sia, Castilia, Telenassa, Janatella and Anthanassa do unstable behaviour of the Historis + Baeotus clade form a stable monophyletic group to the exclusion of in our study and its quite different position in the Tegosa and Mazia, even though the wingless data par- study by Freitas & Brown (2004) suggests that titions is in conflict with the COI and EF1-a data at more data are needed to clarify its relationship the node. within Nymphalidae.

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 238 N. WAHLBERG ET AL.

THE CIRCUMSCRIPTION OF NYMPHALINI ter relationship between Mynes and Araschnia is stable up to a weighting of transversions 7 times tran- The well-supported association of Colobura, Tigridia sitions. A recent morphological study of this group and Smyrna with Nymphalini suggests that they found that Mynes is within Symbrenthia (Fric et al., should be incorporated into the tribe, though the tra- 2004). Normally, Symbrenthia has been associated ditional delineation of the tribe is also well-supported with Mynes (Parsons, 1999; Nylin et al., 2001; Wahl- and, in addition, is stable. berg & Nylin, 2003). For our results to be congruent The placement of Antanartia abyssinica within Van- with those of Fric et al. (2004), we would have to find essa is clear. Vanessa (including A. abyssinica) is a that Mynes geoffroyi is sister to Symbrenthia hypselis, well-supported clade and the sister relationship of which is clearly not the case. Obviously, more species A. abyssinica to V. atalanta + V. indica is stable. of all three genera need to be sampled to resolve the Indeed, larval morphology of A. abyssinica corrobo- conflicting results of the two studies. rates our evidence that it is unrelated to other This study contains a much more comprehensive Antanartia species and related to Vanessa (Nakanishi, sampling of species in Nymphalini than our two pre- 1989). Antanartia has been divided into two species- vious studies (Nylin et al., 2001; Wahlberg & Nylin, groups, the delius-group, comprising A. delius, 2003). The Nymphalis-group continues to be a well- A. schaenia and the unsampled A. borbonica, and the supported and stable clade, though the stable position hippomene-group, to which A. abyssinica, and the of Polygonia canace as sister to species in the genus unsampled A. hippomene and A. dimorphica belong Nymphalis is in contrast to earlier results. Polygonia (Howarth, 1966). Whether A. hippomene and canace is usually placed in its own genus Kaniska, and A. dimorphica should also be placed in Vanessa is not the uncertainty of its position may warrant the use of clear; Nakanishi (1989) noted that the larvae of that genus name. However, our study confirms the sta- A. abyssinica differed greatly from those of ble relationship of Aglais io (usually placed in its own A. schaenia and A. hippomene. Clearly, the missing genus Inachis) with other species of Aglais and the species need to be sampled. stable relationship of Nymphalis l-album with other The divergent position of Antanartia with respect to species of Nymphalis. Also stable and well-supported Hypanartia is also surprising, as all species were is the sister relationship of Polygonia and Nymphalis, included in Hypanartia prior to the description of with Aglais as sister to these two. This study includes Antanartia by Rothschild & Jordan (1903). Antanartia two species of Polygonia that have not been included has always been assumed to be the sister group of in previous phylogenetic studies, P. oreas and Hypanartia (e.g. Willmott et al., 2001). However, our P. haroldi. The position of P. oreas as sister to results suggest that the sister group to Hypanartia is P. gracilis is surprising as it is often considered to be a Vanessa and that Antanartia is sister to all the rest of subspecies of P. progne (Scott, 1984). Polygonia the traditionally recognized Nymphalini. These posi- haroldi is a little known species and morphologically it tions are stable when transversions are weighted 2 is somewhat intermediate between P. progne and and 3 times transitions, but they break down at higher P. satyrus. The relationships of species in Polygonia weights. Weighting 5 and 7 times suggests that are being investigated in more detail currently (E. Antanartia is sister to the Symbrenthia/Mynes/Ara- Weingartner, N. Wahlberg & S. Nylin, unpubl. data). schnia clade and that Hypanartia is sister to these. In no analysis does Antanartia appear as the sister to Hypanartia. THE FATE OF KALLIMINI Vanessa is usually thought to be the sister group to The monophyly of Kallimini appears to be doubtful. the Nymphalis-group of genera (Wahlberg & Nylin, The tribe, as previously circumscribed, never formed a 2003). Our study suggests that Vanessa is sister to monophyletic group in our analyses. Species belonging Hypanartia and that these two clades make up the sis- to Kallimini have occasionally been placed in higher ter group to the Nymphalis-group. This arrangement taxa of their own, and our results show that some of is stable when transversions are weighted 2 and 3 these groups are strongly supported and stable. These times transitions, while at higher weights only Van- well supported groups of genera should be recognized essa is sister to the Nymphalis-group. Thus, it is likely at the tribal level, unless one cares to resort to the (in that either Vanessa or Vanessa + Hypanartia is more our opinion, unacceptable) synonymization of Kal- related to the Nymphalis-group than to the other gen- limini with Melitaeini. Names are available for two era in Nymphalini, and further new characters tribes: Victorinini Scudder, 1893 for the Anartia-clade (molecular and morphological) will help in refining the and Junoniini Reuter, 1896 for the Junonia-clade. The relationships further. Kallima-clade constitutes the newly circumscribed The relationships of Symbrenthia, Mynes and Ara- tribe Kallimini. The three remaining genera (Kalli- schnia are rather surprising, especially since the sis- moides, Vanessula and Rhinopalpa) have to remain

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 PHYLOGENY OF NYMPHALINAE 239 incertae sedis until their positions are stabilized by Melitaeini appears to be the newly circumscribed Kal- further investigation. limini and/or Junoniini. Within Melitaeini, the sister There are 11 species in four genera in the newly cir- relationship of Euphydryas to the rest of Melitaeini is cumscribed tribe Victorinini. The close relationship of in agreement with all previous studies and has a clear, Napeocles and Siproeta has never been considered, well-supported, stable position in this study. The rest though Metamorpha has been considered to be a of Melitaeini appears to be divided up into four dis- derived Siproeta (Fox & Forbes, 1971). Perhaps the tinct groups: Phyciodina, Melitaeina (including only superficial similarity of Metamorpha to Siproeta and Melitaea), the Gnathotriche-group and the Chlosyne- the highly distinctive wing patterns of Napeocles has group. Of these, only the Chlosyne-group is not stable, led previous investigators to ignore the latter. How- though Poladryas remains associated with Chlosyne ever, weighting transversions higher than transitions with transversions weighted up to 7 times transitions. causes Napeocles to become sister to S. stelenes, which The Microtia + Texola + Dymasia clade becomes the is also recovered in some of the most parsimonious sister to the rest of Melitaeini (excluding Euphydryas) trees of the equally weighted analysis. All weighting when transversions are weighted 5 and 7 times tran- schemes other than equal weighting also recover a sis- sitions, though it returns to being sister to Chlosyne at ter relationship between Metamorpha and Anartia, the highest weighting scheme. The close relationship suggesting that Metamorpha is not as closely related of Microtia, Texola and Dymasia was also found by to Siproeta as previously thought. The relationships of Kons (2000), who suggested that Texola and Dymasia the three species of Anartia included in this study are should be synonymized with Microtia. in concordance with a previous study on the genus The close relationship of Higginsius to Gnathotriche (Blum et al., 2003). It is clear from our study that Vic- was suggested by Higgins (1981), but Kons (2000) torinini is a well-defined entity that deserves the rank found Higginsius to be associated with Poladryas. Our of tribe in the Nymphalinae. results strongly corroborate Higgins’ (1981) hypothe- The Junoniini, as circumscribed here, is also a well- sis. The two genera are very curious members of the defined and stable group. The relationships of the gen- Andean fauna and appear to be always rare (Higgins, era in Junoniini found in the equally weighted analy- 1981). They comprise a total of six species and, with sis are stable under fairly severe weighting schemes the unsampled Caribbean Atlantea and Antillea (also (transversions weighted up to 5 times transitions). At very species-poor and always rare), create something higher weights, the Yoma + Protogoniomorpha clade of a biogeographical mystery. We find that the sister becomes sister to the Precis + Hypolimnas clade, and relationship of Gnathotriche and Higginsius to Phycio- Salamis is sister to the rest of Junoniini. The close dina is stable with transversions weighted up to 7 relationships of Protogoniomorpha and Yoma and Pre- times transitions. When transversions are weighted cis and Hypolimnas are stable under all weighting 10 times transitions, the Gnathotriche-group becomes schemes. The clean separation of Precis and Junonia sister to Melitaea. Since Phyciodina is largely a Neo- in our study is both surprising and gratifying. These tropical subtribe, its close relationship to the Gnathot- two genera have long been conflated in the literature, riche-group is perhaps not surprising, but the possible despite their biogeographical disjunction and the fact origin of both groups in South America requires fur- that de Lesse (1952) showed clear genitalic differences ther investigation. between them. de Lesse’s (1952) delimitations of the Of the six missing genera, four (Tisona, Dagon, Orti- two genera are corroborated here. For North American lia and Phystis) putatively belong to the subtribe Phy- researchers it is important to emphasize that all New ciodina (Higgins, 1981). The remaining two genera, World species (including the well-studied J. coenia) Antillea and Atlantea, are restricted to the Greater belong to Junonia and that Precis is restricted to Antilles in the Caribbean and have until recently not Africa. been associated with other melitaeine genera. Kons The once common concept of Kallima (i.e. containing (2000) placed both in the Chlosyne-group, with Kallimoides, Kamilla and Mallika) is clearly untena- Atlantea being sister to Higginsius. It is clear that ble, as firmly concluded by Shirôzu & Nakanishi these two genera need to be sampled for molecular (1984). The Kallimini, as delimited here, contains only data. four genera: Kallima, Catacroptera, Mallika and Dole- schallia. A NOTE ON PHYLOGENETIC PATTERNS IN THE OUTGROUPS THE DELINEATION OF MELITAEINI Although our study has concentrated on Nymphali- The monophyly of Melitaeini is beyond doubt. All nae, we have extensively sampled several potential three gene regions support the clade and it is stable in sister groups to the subfamily. In particular, we have a variety of weighting schemes. The sister group to sampled all genera in the tribes Cyrestini and Pseud-

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251

240 N. WAHLBERG ET AL. ergolini, which constitute the Cyrestinae in Wahlberg (excluding Colobura, Tigridia and Smyrna). An alter- et al. (2003b). However, in the present study the two native scenario would have a widespread common tribes do not form a monophyletic group; rather, ancestor in South America, Africa and the Palaearctic Cyrestini appears as the most basal group in the that then speciated through a series of vicariance nymphaline clade and Pseudergolini is sister to Apa- events (Fig. 5). The Palaearctic has, however, been a turinae (Fig. 1). There are four clades in the outgroup very important area for the diversification of the gen- appearing with good to strong support that correspond era Aglais, Nymphalis, Polygonia and possibly Van- to the Apaturinae, Biblidinae, Cyrestini and Pseuder- essa. There have clearly been several independent golini. However, the relationships among these four colonizations of the Nearctic from the Palaearctic by taxa and Nymphalinae are very poorly supported, are species in the the former three genera. unstable and show much conflict among partitions. The clade that includes the tribes Junoniini and Vic- Clearly, the nymphaline clade (sensu Wahlberg et al., torinini, as well as Kallimoides, Rhinopalpa and Van- 2003b) needs to be more extensively sampled and essula, appears to have originated in Africa (Fig. 5). more data need to be added before any robust There is a colonization of South America from Africa conclusions emerge about relationships of the major by the ancestor of Kallimoides and Victorinini. Also, clades. the Oriental region appears to have been colonized independently several times. The patterns in this clade are obscured somewhat by the weak support for BIOGEOGRAPHICAL HISTORY OF NYMPHALINAE the current hypothesis of relationships; in addition, Despite the limitations of our current phylogenetic many missing species in the genera Hypolimnas and hypothesis, some strong patterns emerge from our dis- Junonia are likely to have had a decisive effect on the persal-vicariance analysis (Fig. 5). First of all, many results due to their presence in the African, Oriental dispersal events are required to explain the distribu- and Australasian areas. However, it is clear that the tion patterns seen today: 44 when the maximum num- African region has been instrumental in the diversifi- ber of ancestral areas is not constrained and 46 when cation of taxa in the clade. they are constrained to two. This should not be sur- The tribe Kallimini appears to have originated in prising, since Nymphalinae contains some of the most the Oriental region, with a colonization of Africa by mobile butterflies known, such as Vanessa cardui, the common ancestor of Kallima and Catacroptera/ which is famous for being able to disperse over thou- Mallika (Fig. 5). This hypothesis needs to be tested by sands of kilometers in one generation. Indeed, the including the several species of Doleschallia found genus Vanessa, which is present in all the major zoo- exclusively in the Australasian region. logical biomes of the world, has a highly ambiguous The Nearctic region has been an important area for reconstruction of its ancestral distribution. Perhaps the diversification of the tribe Melitaeini (Fig. 5). One the ancestor of Vanessa was a widespread species, interpretation of the patterns recovered by DIVA is much like V. cardui today, that subsequently speciated that the group originated in the Nearctic and subse- during intervals of isolation due to climatic or other quently colonized the Palaearctic (the ancestor of factors. Melitaea) and the Neotropics (ancestor of the Gnathot- The distribution of the most recent common ances- riche-group and Phyciodina). Interestingly, the cur- tor of Nymphalinae (excluding Historis and Baeotus) rent phylogenetic hypothesis suggests a disjunct appears to be widespread (Fig. 5), though this may be distribution for the ancestor of Melitaea + Gnathot- an artefact of the program used and might be fixed by riche/Higginsius + Phyciodina being found in the including outgroups with known ancestral distribu- Palaearctic and the Neotropics. Sampling the two tions (Ronquist, 1996). We refrained from doing so, as missing Caribbean genera will be important to under- our sampling of the outgroups is not sufficient to standing the historical biogeography of Melitaeini. resolve their ancestral distributions; also, we are not In summary, it is clear that several regions have confident about the current hypothesis of relation- been important areas of diversification of clades in ships among major groups in the nymphaline clade. Nymphalinae, viz. the Palaearctic for the Nymphalis- The tribe Nymphalini may have originated in South group of genera, Africa for Junoniini and related gen- America, one of the seven possible ancestral areas in era and the Nearctic for Melitaeini. Our DIVA analy- the unconstrained analysis (Fig. 5) and one of two in sis indicates that dispersal has had a major effect on the constrained searches. Africa may have been colo- the distributions of extant species in Nymphalinae. nized from South America by the common ancestor of Whether dispersal or vicariance has been the more Smyrna and the rest of Nymphalini. Once the lineage important process in generating the deeper diver- leading to Smyrna split off, it appears that the Palae- gences in Nymphalinae can be tested by dating diver- arctic was colonized from Africa by the common ances- gences using molecular clocks calibrated with fossils. tor of Antanartia and the rest of Nymphalini Such a study is in preparation and the implications of

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 PHYLOGENY OF NYMPHALINAE 241

B BD BCD Colobura/Tigridia B BDF BC Smyrna B BCDF BD Antanartia C BDEF BCD Symbrenthia E F DF ADF BDF ABDF CDF ACDF BDF DE BCDEF DF Araschnia D BCDF ABCDF DEF ADEF BDEF ABDEF BCDF EF CD DEF Mynes F CDEF ACDEF BCDEF ABCDEF BDEF CDF BCDEF BD Hypanartia B CDEF ABD Vanessa annabella A AD BCD D ACD CD Vanessa abyssinica C ABCD DF ADF D Vanessa atalanta AD BDF DEF ACDF D Vanessa indica DE ABDF DF Vanessa gonerilla/itea F BCDF CDF ABCDF B Vanessa virginiensis AB D Vanessa braziliensis/myrinna B Vanessa cardui ACDE AF CF ACF DF Vanessa kershawi F ADF CDF ACDF EF D Aglais io D AEF CEF ACEF DEF AD Aglais milberti A ADEF CDEF ACDEF Aglais urticae D D Polygonia c-aureum D BD ABD BCD ABCD D BDE ABDE BCDE AD Polygonia North American clade A Polygonia egea D ABCDE BF ABF BCF ABCE D D ABCF BDF ABDF BCDF Polygonia faunus A ACDE AD ABCDF BEF ABEF BCEF ABCDE Polygonia c-album D ABCEF BDEF ABDEF ACDEF D Polygonia canace DE BCDEF ABCDEF ABCDEF D Nymphalis l-album AD D Nymphalis polychloros D D Nymphalis xanthomelas D AD Nymphalis californica A Nymphalis antiopa AD BC Kallimoides rumia C Victorinini B CE Rhinopalpa E C Vanessula C Precis C C Hypolimnas anthedon/usambara C C C F Hypolimnas misippus CEF CF CF CEF F Hypolimnas alimena F C F Hypolimnas bolina CEF Hypolimnas pandarus F C Salamis C CE Yoma EF AC C Protogoniomorpha C Junonia erigone F ACE CE EF ACDE CF Junonia iphita E ABCDE Junonia artaxia/touhilimasa/cytora C CEF C Junonia cymodoce C C C Junonia terea/natalica C AC Junonia coenia A Junonia sophia/oenone C E Doleschallia EF CE Catacroptera/Mallika C Kallima E AE A Euphydryas North American clade A ADE AD Euphydryas gilllettii A ABDE Euphydryas aurinia/desfontainii D Melitaea D A BD A = Nearctic Gnathotriche/Higginsius B AD B Mazia B B = Neotropics ABD B Tegosa B C = Afrotropics B Phyciodes A D = Palaearctic AB AB Eresia/Castilia/Telenassa B E = Oriental AD B ABD B Anthanassa rest B F = Australasia AB Anthanassa texana A Anthanassa tulcis B Poladryas A A A Microtia elva AB A Texola/Dymasia A A Chlosyne Central American clade AB A Chlosyne cyneas A A Chlosyne theona A A Chlosyne nycteis A A Chlosyne acastus/palla A A Chlosyne gorgone A Chlosyne lacinia AB

Figure 5. Reconstructed ancestral distributions according to a dispersal-vicariance analysis. Letters after the names of taxa give the current distribution of those taxa.

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 242 N. WAHLBERG ET AL. it will be discussed elsewhere (N. Wahlberg, unpubl. Chichvarkhin, Tim Davenport, Phil DeVries, André data). Freitas, Daniel Janzen, Chris Jiggins, Tjeerd Jonge- ling, Darrell Kemp, Norbert Kondla, Runar Krogen, Gerardo Lamas, Torben B. Larsen, Yi-Hsin Lee, Jorge PBS AND SENSITIVITY ANALYSIS León-Cortés, Mike Leski, Debra Murray, Naomi The ease with which molecular data can be generated Pierce, James Scott, Mike Soukop, Constant’ Stefa- has led to an ever-increasing number of published nescu, Martin Steinbauer, Man-Wah Tan, Bruce phylogenetic studies. Common to almost all these Walsh, Andy Warren, and Keith Willmott for helping studies is the lack of sensitivity analyses and, in the us to obtain specimens. We thank Tiina Berg and Sim- case of multiple independent data sets, evaluation of ina Vintila for expert help in the lab. We also thank congruence at given nodes (but see Reed & Sperling, Andy Warren, Keith Willmott, Carla Penz and Torben 1999). An exception to the former are studies based on B. Larsen for useful comments on a draft version of the direct optimization of ribosomal DNA sequences manuscript. (Wheeler, 1995; Giribet, 2003). Congruence among data partitions is usually evaluated at the level of entire data sets through tests such as the ILD (Farris REFERENCES et al., 1994), rather than through evaluation at every resolved node. Again, a few exceptions are notable (e.g. Ackery PR. 1984. Systematic and faunistic studies on but- Gatesy et al., 1999; Cognato & Vogler, 2001; Lambkin terflies. In: Vane-Wright RI, Ackery PR, eds. The biology of et al., 2002; Damgaard & Cognato, 2003; Wahlberg & butterflies. Princeton, NJ: Princeton University Press, 9– Nylin, 2003). Both the PBS and sensitivity analysis 21. allow detailed evaluation of which nodes are likely to Ackery PR. 1988. Hostplants and classification: a review of nymphalid butterflies. Biological Journal of the Linnean be robust and stable to the addition of new data, as Society 33: 95–203. well as which nodes require further investigation. We Ackery PR, de Jong R, Vane-Wright RI. 1999. The butter- emphasize that these are heuristic tools to assess the flies: Hedyloidea, Hesperoidea and Papilionoidea. In: Kris- quality of the most parsimonious hypothesis and guide tensen NP, ed. Lepidoptera, moths and butterflies. 1. future sampling of characters and taxa, rather than Evolution, systematics and biogeography. Handbook of Zool- alternate analytical approaches that challenge the ogy 4 (35), Lepidoptera. Berlin: de Gruyter, 263–300. philosophical basis of our preferred optimality Austin GT, Murphy DD, Baughman JF, Launer AE, criterion. Fleishman E. 2003. Hybridization of checkerspot butter- flies in the Great Basin. Journal of the Lepidopterists’ Society CONCLUSION 57: 176–192. Baker RH, DeSalle R. 1997. Multiple sources of character The main conclusions of this study are: (1) Smyrna, information and the phylogeny of Hawaiian drosophilids. Colobura and Tigridia are associated with Nympha- Systematic Biology 46: 654–673. lini, (2) ‘Kallimini’ is more closely related to Melitaeini Baker RH, Yu X, DeSalle R. 1998. Assessing the relative con- than to Nymphalini, (3) ‘Kallimini’ is not monophyl- tribution of molecular and morphological characters in etic, but is paraphyletic with regard to Melitaeini, (4) simultaneous analysis trees. Molecular Phylogenetics and Melitaeini and Nymphalini plus the three coeine gen- Evolution 9: 427–436. era are strongly supported monophyletic groups, and Blum MJ, Bermingham E, Dasmahapatra K. 2003. A (5) Precis and Junonia are not synonymous or even molecular phylogeny of the neotropical butterfly genus Anar- sister groups. tia (Lepidoptera: Nymphalidae). Molecular Phylogenetics Major unanswered questions are: (1) the positions of and Evolution 26: 46–55. Historis + Baeotus in the nymphaline clade, and (2) Brakefield PM, Gates J, Keys D, Kesbeke F, Wijngaarden PJ, Monteiro A, French V, Carroll SB. 1996. Develop- the sister group of Nymphalinae. Increased sampling ment, plasticity and evolution of butterfly eyespot patterns. of species in the subfamilies Biblidinae and Apaturi- Nature 384: 236–242. nae, as well as increased character data (molecular Bremer K. 1988. The limits of amino acid sequence data in and morphological), will help to resolve these angiosperm phylogenetic reconstruction. Evolution 42: 795– questions. 803. Bremer K. 1994. Branch support and tree stability. Cladistics ACKNOWLEDGEMENTS 10: 295–304. Brower AVZ. 1994. Phylogeny of Heliconius butterflies As always, this study would not have been possible inferred from mitochondrial DNA sequences. Molecular Phy- without the truly wonderful, altruistic help of both logenetics and Evolution 3: 159–174. amateur and professional lepidopterists. We are eter- Brower AVZ. 2000. Phylogenetic relationships among the nally grateful to Jun-ichi Asahi, Alexei Belik, Anton Nymphalidae (Lepidoptera), inferred from partial sequences

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 PHYLOGENY OF NYMPHALINAE 243

of the wingless gene. Proceedings of the Royal Society of Lon- Giribet G. 2003. Stability in phylogenetic formulations and its don B 267: 1201–1211. relation to nodal support. Systematic Biology 52: 554–564. Brower AVZ, DeSalle R. 1998. Patterns of mitochondrial ver- Goloboff PA. 1998. NONA. 2.0. Published by the author. sus nuclear DNA sequence divergence among nymphalid Hall TA. 1999. BioEdit: a user-friendly biological sequence butterflies: the utility of wingless as a source of characters alignment, editing and analysis program for Windows 95/98/ for phylogenetic inference. Insect Molecular Biology 7: 73– NT. Nucleic Acids Symposium Series 41: 95–98. 82. Hanski I. 1999. Metapopulation ecology. Oxford: Oxford Uni- Carroll SB, Gates J, Keys DN, Paddock SW, Panganiban versity Press. GEF, Selegue JE, Williams JA. 1994. Pattern-formation Harvey DJ. 1991. Higher classification of the Nymphalidae, and eyespot determination in butterfly wings. Science 265: Appendix B. In: Nijhout HF, ed. The development and evo- 109–114. lution of butterfly wing patterns. Washington, DC: Smithso- Clark AH. 1948. Classification of the butterflies with the allo- nian Institution Press, 255–273. cation of the genera occurring in America north of Mexico. Higgins LG. 1981. A revision of Phyciodes Hübner and related Proceedings of the Biological Society of Washington 61: 77– genera, with a review of the classification of the Melitaeinae 84. (Lepidoptera: Nymphalidae). Bulletin of the British Museum Cognato AI, Vogler AP. 2001. Exploring data interaction and of Natural History 43: 77–243. nucleotide alignment in a multiple gene analysis of Ips Howarth TG. 1966. Revisional notes on the genus Antanartia (Coleoptera: Scolytinae). Systematic Biology 50: 758–780. (Lepidoptera: Nymphalidae). Bulletin of the British Museum Damgaard J, Cognato AI. 2003. Sources of character conflict of Natural History 18: 21–43. in a clade of water striders (Heteroptera: Gerridae). Cladis- Janz N, Nylin S, Nyblom K. 2001. Evolutionary dynamics of tics 19: 512–526. host plant specialization: a case study of the tribe Nympha- Darlu P, Lecointre G. 2002. When does the Incongruence lini. Evolution 55: 783–796. Length Difference test fail? Molecular Biology and Evolution de Jong R, Vane-Wright RI, Ackery PR. 1996. The higher 19: 432–437. classification of butterflies (Lepidoptera): problems and pros- Dasmahaptra KK, Blum MJ, Aiello A, Hackwell S, Davies pects. Entomologica scandinavica 27: 65–101. N, Bermingham EP, Mallet J. 2002. Inferences from a Judd DD. 1998. Exploring component stability using life- rapidly moving hybrid zone. Evolution 56: 741–753. stage concordance in sabethine mosquitoes (Diptera: Culi- DeSalle R, Brower AVZ. 1997. Process partitions, congru- cidae). Cladistics 14: 64–94. ence, and the independence of characters: inferring relation- Kons HLJ. 2000. Phylogenetic studies of the Melitaeini (Lepi- ships among closely related Hawaiian Drosophila from doptera: Nymphalidae: Nymphalinae) and a revision of the multiple gene regions. Systematic Biology 46: 751–764. genus Chlosyne. Gainesville: Butler Department of Entomol- DeVries PJ, Kitching IJ, Vane-Wright RI. 1985. The sys- ogy & Nematology, University of Florida, 799. tematic position of Antirrhea and Caerois, with comments on Kuznetzov VI, Stekolnikov AA. 2001. New approaches to the the classification of the Nymphalidae (Lepidoptera). System- system of Lepidoptera of world fauna. St. Petersburg: Nauka. atic Entomology 10: 11–32. Lambkin CL, Lee MSY, Winterton SL, Yeates DK. 2002. Ehrlich PR. 1958. The comparative morphology, phylogeny Partitioned Bremer support and multiple trees. Cladistics and higher classification of the butterflies (Lepidoptera: 18: 436–444. Papilionoidea). University of Kansas Science Bulletin 39: Larsen TB. 1991. The butterflies of Kenya and their natural 305–370. history. Oxford: Oxford University Press. Farris JS, Källersjö M, Kluge AG, Bult C. 1994. Testing the de Lesse H. 1952. Note sur les genres Precis Hb. et Junonia significance of incongruence. Cladistics 10: 315–319. Hb. (Lep. Nymphalidae). Bulletin de la Société Ento- Forbes WTM. 1928. Variation in Junonia lavinia (Lepi- mologique de France 57: 74–77. doptera: Nymphalidae). Journal of the New York Entomolog- Miller JS, Brower AVZ, DeSalle R. 1997. Phylogeny of the ical Society 36: 305–320. neotropical moth tribe Josiini (Notodontidae: Dioptinae): Fox RM, Forbes AC. 1971. The butterflies of the genera comparing and combining evidence from DNA sequences and Siproeta and Metamorpha (Lepidoptera: Nymphalidae). morphology. Biological Journal of the Linnean Society 60: Annals of the Carnegie Museum 43: 223–247. 297–316. Freitas AVL, Brown KSJ. 2004. Phylogeny of the Nympha- Monteiro A, Pierce NE. 2001. Phylogeny of Bicyclus (Lepi- lidae (Lepidoptera: Papilionoidea). Systematic Biology 53: doptera: Nymphalidae) inferred from COI, COII, and EF- 363–383. 1alpha gene sequences. Molecular Phylogenetics and Evolu- Fric Z, Konvicka M, Zrzavy J. 2004. Red & black or black & tion 18: 264–281. white? Phylogeny of the Araschnia butterflies (Lepidoptera: Müller W. 1886. Südamerikanische Nymphalidaeraupen: Ver- Nymphalidae) and evolution of seasonal polyphenism. Jour- such eines natürlichen Systems der Nymphaliden. Zoolo- nal of Evolutionary Biology 17: 265–278. gische Jahrbücher 1: 417–678. Gatesy J, O’Grady P, Baker RH. 1999. Corroboration Muyshondt AJ, Muyshondt A. 1979. Notes on the life cycle among data sets in simultaneous analysis: hidden support and natural history of butterflies of El Salvador. III C. His- for phylogenetic relationships among higher level artiodactyl toris odius and Coea acheronta (Nymphalidae-Coloburinae). taxa. Cladistics 15: 271–313. Journal of the Lepidopterists’ Society 33: 112–123.

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 244 N. WAHLBERG ET AL.

Nakanishi A. 1989. Immature stages of Antanartia abyssin- Silberglied RE. 1984. Visual communication and sexual ica (Felder) (Lepidoptera, Nymphalidae). Japanese Journal selection among butterflies. In: Vane-Wright RI, Ackery PR, of Entomology 57: 712–719. eds. The biology of butterflies. Princeton, NJ: Princeton Uni- Nixon KC. 2002. WINCLADA, 1.00.08. Published by the versity Press, 204–223. author. Singer MC. 1971. Evolution of food-plant preference in the Nylin S. 1988. Host plant specialization and seasonality in a butterfly Euphydryas editha. Evolution 25: 383–389. polyphagous butterfly, Polygonia c-album (Nymphalidae). Swofford DL. 2001. PAUP*: phylogenetic analysis using par- Oikos 53: 381–386. simony (*and other methods), Version 4.0b10. Sunderland, Nylin S, Nyblom K, Ronquist F, Janz N, Belicek J, MA: Sinauer Associates. Källersjö M. 2001. Phylogeny of Polygonia, Nymphalis and Vári L. 1979. The butterflies of southern Africa. Part IV: related butterflies (Lepidoptera: Nymphalidae): a total-evi- Nymphalidae; Nymphalinae. Pretoria, South Africa: Trans- dence analysis. Zoological Journal of the Linnean Society vaal Museum. 132: 441–468. Wahlberg N. 2001. The phylogenetics and biochemistry of Parsons M. 1999. The butterflies of Papua New Guinea: Their host plant specialization in melitaeine butterflies (Lepi- systematics and biology. London: Academic Press. doptera: Nymphalidae). Evolution 55: 522–537. Penz C, Peggie D. 2003. Phylogenetic relationships among Wahlberg N, Nylin S. 2003. Morphology versus molecules: Heliconiinae genera based on morphology (Lepidoptera: resolution of the positions of Nymphalis, Polygonia and Nymphalidae). Systematic Entomology 28: 451–479. related genera (Lepidoptera: Nymphalidae). Cladistics 19: Reed RD, Sperling FAH. 1999. Interaction of process parti- 213–223. tions in phylogenetic analysis: an example from the swallow- Wahlberg N, Oliveira R, Scott JA. 2003a. Phylogenetic rela- tail butterfly genus Papilio. Molecular Biology and Evolution tionships of Phyciodes butterfly species (Lepidoptera: 16: 286–297. Nymphalidae): complex mtDNA variation and species delim- Ronquist F. 1996. DIVA, 1.1. Program and manual available itations. Systematic Entomology 28: 257–273. from Uppsala University (ftp.uu.se or ftp.systbot.uu.se). Wahlberg N, Weingartner E, Nylin S. 2003b. Towards a bet- Ronquist F. 1997. Dispersal-vicariance analysis: a new ter understanding of the higher systematics of Nymphalidae approach to the quantification of historical biogeography. (Lepidoptera: Papilionoidea). Molecular Phylogenetics and Systematic Biology 46: 195–203. Evolution 28: 473–484. Rothschild W, Jordan K. 1903. Lepidoptera collected by Wahlberg N, Zimmermann M. 2000. Pattern of phylogenetic Oscar Neumann in north-east Africa. Novitates Zoologica 10: relationships among members of the tribe Melitaeini (Lepi- 491–542. doptera: Nymphalidae) inferred from mtDNA sequences. Scott JA. 1984. A review of Polygonia progne (oreas) and P. Cladistics 16: 347–363. gracilis (zephyrus) (Nymphalidae), including a new subspe- Weller SJ, Pashley DP, Martin JA. 1996. Reassessment of cies from the southern Rocky Mountains. Journal of butterfly family relationships using independent genes and Research on the Lepidoptera 23: 197–210. morphology. Annals of the Entomological Society of America Scott JA. 1985. The phylogeny of butterflies (Papilionoidea 89: 184–192. and Hesperoidea). Journal of Research on the Lepidoptera Wheeler WC. 1995. Sequence alignment, parameter sensitiv- 23: 241–281. ity, and the phylogenetic analysis of molecular data. System- Scott JA, Wright DM. 1990. Butterfly phylogeny and fossils. atic Biology 44: 321–331. In: Kudrna O, ed. Butterflies of Europe. 2. Wiesbaden, 152– Willmott KR, Hall JPW, Lamas G. 2001. Systematics of 208. Hypanartia (Lepidoptera: Nymphalidae: Nymphalinae), Shapiro AM, Geiger H. 1989. Electrophoretic comparisons of with a test for geographical speciation mechanisms in the vicariant Vanessa: genetic differentiation between V. anna- Andes. Systematic Entomology 26: 369–399. bella and V. carye (Nymphalidae) since the Great American Zimmermann M, Wahlberg N, Descimon H. 2000. Phy- Interchange. Journal of the Lepidopterists’ Society 43: 81–92. logeny of Euphydryas checkerspot butterflies (Lepi- Shirôzu T, NakanashiNakanishi A. 1984. A revision of the doptera: Nymphalidae) based on mitochondrial DNA genus Kallima Doubleday (Lepidoptera, Nymphalidae). Tyô sequence data. Annals of the Entomological Society of to Ga 34: 97–110. America 93: 347–355.

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 PHYLOGENY OF NYMPHALINAE 245 wingless a her specimens, see http:// her specimens, AY788596 AY788698 AY788461 APPENDIX 1 03-srnp-11744 NW62-5RB249NW82-6NW62-3 Butterfly farm supplier RICA: COSTA NW83-5NW109-4Ariquemes Rondonia, BRAZIL: Amani TANZANIA: NW85-11 Butterfly farm supplier RICA: COSTA NW109-8 AY090215 Kibale Forest UGANDA: NW88-15 ECUADOR AY090181Yurimaguas Road to PERU: AY090216 ECUADOR AY788597 AY090148 Harare ZIMBABWE: AY090182 AY788699 AY090149 AF246581 AY218242 AY788598 AY218245 AY218262 AY788700 AY218265 AY218280 AY788462 AY218283 AY218247 AY788599 AY218267 AY788600 AY788701 AY218285 AY788702 AY788463 AY788464 NW70-6NW107-16NW100-11NW111-3 Butterfly farm supplier RICA: COSTA RB250 Oregon USA: RB227 Sylhet Division BANGLADESH: NW101-1 Seko Sulawesi, INDONESIA: NW111-10 AY090202NW118-1Ariquemes Rondonia, BRAZIL: NW100-10 AY090168 Brastagi Sumatra, Ariquemes Rondonia, W. BRAZIL: INDONESIA: NW82-10 Hinunangan Leyte, S. PHILIPPINES: AY218240NW88-14 AY090135 Lao Cai VIETNAM: NW119-3 Sylhet Division BANGLADESH: AY218260 AY788601 AY788602 AY218235 Kibale Forest UGANDA: AY788604 Marondera ZIMBABWE: AY218278 AY788703 AY788603 AC Guanacaste, AY788704 RICA: COSTA AY218254 AY788591 AY788706 AY788465 AY218249 AY788705 AY788466 AY218273 AY788693 AF246532 AY218269 AY788467 AY788457 AY788605 AY218237 AY218287 AY788595NW69-6 AY788707 AY218256NW82-15 AY788697NP95-Y227 AY788468 AY218274 NW84-7 AY788460 Butterfly farm supplierNW97-8 Arizona USA: (MCZ voucher) MALAYSIA NW97-11 Primorye RUSSIA: County Taitung TAIWAN: County Taitung TAIWAN: AY788593 AY090199 AY788695 AY090165 AY218251 AY788592 AF246588 AF187734 AY090132 AY788594 AY218271 AY788694 AY218257 AY788696 AY218289 AY788458 AY218275 AY788459 NW130-16NW130-13NW130-14NW130-15 Lorocachi Pastaza, ECUADOR: NW81-7Yuturi Rio Napo, ECUADOR: PE-18–16 Cerro Lumbaqui Sucumbios, ECUADOR: NW68-11 Rio Cachaco Imbabura, ECUADOR: AY788614 Butterfly farm supplier RICA: COSTA Pres.Tambopata Madre de Dios, PERU: AY788613 AY788719 Butterfly farm supplier RICA: COSTA AY788615 AY788616 AY788631 AY788718 AY788632 AY788480 AY788720 AY788721 AY090228 AY788750 AY788479 AY788751 AY788481 AY788482 AY090196 AY788511 AY788512 AY090162 Catonephele numilia Dynamine maeon Eurytela dryope Hamadryas februa Mesoxantha esothea Myscelia capensis Nica flavilla regina Panacea Sevenia boisduvali Heliconius hecale Adelpha bredowi Cyrestis thyodamas rahria Chersonesia Marpesia orsilochus Marpesia chiron Amnosia decora nesimachus Dichorragia Pseudergolis wedah nicea Stibochiona Ariadne enotrea Byblia anvatara Callicore pacifica Apatura iris Asterocampa leilia Eulaceura osteria Mimathyma schrenckii maculata Timelaea Chitoria chrysolora Baeotus japetus Baeotus deucalion Baeotus aeilus Baeotus beotus Historis odius Historis acheronta Colobura dirce ELICONIINAE YMPHALINAE YRESTINAE IBLIDINAE PATURINAE IMENITIDINAE A L C B N www.zoologi.su.se/research/wahlberg Higher taxon Species code Voucher Locality COI EF1- List of species sampled in this study along with the GenBank accession numbers three genes sequenced. images of vouc For ‘Coeini’ H

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 246 N. WAHLBERG ET AL. wingless a Continued APPENDIX 1 RB349NW85-2Ariquemes Rondonia, BRAZIL: São Paulo BRAZIL: AY788684 AY788822 AY788582 AY788678 AY788816 AY788576 NW86-7NW82-4NW65-5NW39-2 Kitumbeine TANZANIA: NW63-20 Kibale Forest UGANDA: NW65-2 CAMEROONNW97-2 ESTONIANW97-3 Queensland AUSTRALIA: PE-10–7 Selangor MALAYSIA: NW89-11 Kauhsiung County TAIWAN: NW89-6 Kauhsiung County TAIWAN: NW36-6 Quebrada Chapimayo Cuzco, PERU: A-15PE-05 AY788609 Sucumbios ECUADOR: NW74-4 AY788610 Sucumbios ECUADOR: NW63-21 AY788711 Quebrada San Luis Cuzco, PERU: Minas Gerais BRAZIL: NW89-4 AY248778 AY788712 AY788638NW63-3 AY788472 AY788680 Wyoming USA: NW63-4 AY248803 AY788473 AY788679 Stockholm SWEDEN: AY788757NW63-9 AY248779 AY218236 AY788818 Esmeraldas ECUADOR: NW63-14 AF412760 AY248781 AY788817 Missouri AF246590 USA: NW77-3 AY248804 AY248780 AY218255 AY788578 NEW ZEALANDNW36-5 AY788639 AY248806 AY788577 JAPANNW77-16 AF412784 AY248805 AF412780 NEW ZEALAND AY788640NW63-16 AY788758 AF187774 AF412759 Australia South AUSTRALIA: NW77-14 AF412762 AY788759 Minas Gerais BRAZIL: NW63-3 AY788518 Tennessee USA: AY788760NW70-2 Stockholm SWEDEN: AY090221 AY788519 NW74-14 AY788686 Washington USA: AY788520 NW78-1 AY788685 AY090187 Stockholm SWEDEN: NW62-2 AY788824 Stockholm SWEDEN: NW84-1 AY788689 Oregon USA: AY788823EW19-11 AF412772 AY788584 AY248782 AY248784 British Columbia CANADA: NW70-3 AY788827 AY788583 Öland SWEDEN: NW65-8 AY788688 AY788690 AY248807 Yakutia RUSSIA: NW65-6 AY248809 AY788587 JAPANNW77-15 AY788826 AY248785 AY788828 Stockholm SWEDEN: AF412770 NW74-12 AF412782 AY788687 AY248783 JAPAN AY788586 AY248810 Tennessee AY788588 USA: AY248787 AY248786 AY788825 GREECE AY248791 AY248808 AY218246 Oregon USA: AF412766 AY248812 AY248811 AY788585 AY248816 AY248827 AY218266 AY248828 AF412777 AY248789 AY248832 AY218284 AY248788 AY248790 AY248814 AY090222 AY248813 AY248815 AY248830 AY090188 AY248829 AY248831 AY248792 AY248794 AY090154 AY248817 AY248819 AY248799 AY248798 AY248800 AY248833 AF412781 AY248824 AY248823 AY248825 AF412786 AY248837 AY248838 Tigridia acesta Tigridia Smyrna blomfildia Antanartia abyssinica Antanartia delius Antanartia schaenia levana Araschnia Mynes geoffroyi Symbrenthia hypatia Symbrenthia hypselis Symbrenthia lilea Hypanartia bella Hypanartia charon Hypanartia kefersteinii Hypanartia lethe Hypanartia lindigii annabella Vanessa atalanta Vanessa braziliensis Vanessa cardui Vanessa gonerilla Vanessa indica Vanessa itea Vanessa kershawi Vanessa myrinna Vanessa virginiensis Vanessa Aglais io Aglais milberti Aglais urticae Nymphalis antiopa Nymphalis californica Nymphalis l-album Nymphalis polychloros Nymphalis xanthomelas canace Polygonia c-album Polygonia c-aureum Polygonia comma Polygonia egea Polygonia faunus Polygonia Nymphalini Higher taxon Species code Voucher Locality COI EF1-

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 PHYLOGENY OF NYMPHALINAE 247 wingless a NW112-3NW77-12NW74-10NW74-9Yecora Sonora, MEXICO: NW74-6 Tennessee USA: Oregon USA: Oregon USA: Oregon USA: AY788662 AY788800 AY788560 AY248793 AY788663 AY248818 AY248796 AY788801 AY248834 AY248797 AY248821 AY788561 AY248822 AY248835 AY248836 NW68-5NW66-5NW36-6NW64-7 Butterfly farm supplier RICA: COSTA NW69-5 Butterfly farm supplier RICA: COSTA NW85-1 Minas Gerais BRAZIL: PE-10–10 Butterfly farm supplier RICA: COSTA NW81-1 AY788606 Butterfly farm supplier RICA: COSTA NW68-8 AY788607 Mato Grosso BRAZIL: NW62-6 AY788708 Quebrada Chaupimayo Cuzco, PERU: NW68-3 AY788677 AY788709 Queensland AUSTRALIA: NW80-11 AY218248 AY788469 Butterfly farm supplier KENYA: NW66-4 AY788815 AY788470 AY788658 Queensland AUSTRALIA: NW111-6 AY218268 Butterfly farm supplier KENYA: NW88-4 AY788575 Ceram Island INDONESIA: AY788608 AY788784NW88-6 AY218286 Butterfly farm supplier KENYA: NW88-3 Mandraka MADAGASCAR: AY788710NW83-13 AY788544 AY788634 Marondera ZIMBABWE: AY788661NW68-9 AY788633 Marondera AY788471 ZIMBABWE: NW83-1 AY788635 AY788753 Marondera ZIMBABWE: NW114-15 AY788788 Harare ZIMBABWE: AY090224 AY788752NW81-5 AY788637 AY788754 AY788636 AY788514 Butterfly farm supplier KENYA: NW111-9 AY788548 AY090190 AY788513 Kibale Forest UGANDA: NW82-3 Lesombo R N of Mwinilunga, ZAMBIA: AY788756 AY788515 AY788664 AY788755NW80-13 AY090156 MALAYSIA AY788517 Mandraka MADAGASCAR: AY788802 AY788516 AY788671NW73-15 AY788665 Kibale Forest UGANDA: NW82-7 AY788666 AY788669Valley Wau Morobe Prov., PNG: AY788562 NW123-23 AY788667 AY788809 AY788803NW114-11 AY788804 AY788807 Butterfly farm supplier KENYA: AY788668NW85-13 AY788805 AY788569 AY788563 Kibale Forest UGANDA: NW68-17 AY788670 GHANA AY788564 AY788567 NW81-2 S of Mwinilunga ZAMBIA: AY788806 AY788565 AY788675 AY788692NW68-13 AY788808 Tennessee USA: NW68-1 AY788566 Butterfly farm supplier AY788813 AY090223NW83-10 AY788830 AY788676 AY788568 Bulolo Morobe Province, PNG: NW68-15 Butterfly farm supplier KENYA: AY788573 AY090189NW95-15 AY788590 AY788814 AY788674 Butterfly farm supplier KENYA: Kibale Forest UGANDA: AY788673 AY090155 AY788574 Butterfly farm supplier AY788642 KENYA: AY788812 Kalene Hill ZAMBIA: AY788811 AY788762 AY788645 AY788644 AY788572 AY090225 AY788646 AY788571 AY788522 AY788764 AY788763 AY788643 AY788648 AY788672 AY090191 AY788765 AY788524 AY788523 AY788647 AY248801 AY788767 AY788810 AY090157 AY788525 AY788766 AY248826 AY788527 AY788570 AY788649 AY788526 AY788768 AY788528 Polygonia haroldi Polygonia interrogationis Polygonia oreas Polygonia satyrus Polygonia gracilis Polygonia Anartia amathea Anartia fatima Anartia jatrophae Siproeta epaphus Siproeta stelenes Napeocles jucunda Metamorpha elissa Hypolimnas alimena Hypolimnas anthedon Hypolimnas bolina Hypolimnas misippus Hypolimnas pandarus Hypolimnas usambara Precis andremiaja Precis antilope Precis archesia Precis ceryne Precis cuama Precis octavia Precis sinuata Precis tugela Rhinopalpa polynice Salamis anteva Salamis cacta algina Yoma Protogoniomorpha anacardii parhassus Pr. cytora Pr. artaxia Junonia coenia Junonia iphita Junonia erigone Junonia natalica Junonia oenone Junonia sophia Junonia terea Junonia touhilimasa Junonia ‘Kallimini’ Higher taxon Species code Voucher Locality COI EF1-

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 248 N. WAHLBERG ET AL. wingless a Continued APPENDIX 1 NW114-8NW85-15NW62-8NW96-8 Lesombo R N of Mwinilunga, ZAMBIA: NW88-1 Butterfly farm supplierNW122-6 Butterfly farm supplierNW64-5 AY788652 Ashanti GHANA: RegionNW96-5 Marondera ZIMBABWE: AY788771 TANZANIA Butterfly farm supplier AY788531 Ashanti GHANA: Region AY788650 AY090229 AY788651 AY788769 AY788619 AY090197 AY788770 AY788529 AY788724 AY788621 AY090163 AY788691 AY788530 AY788485 AY788735 AY788829 AY788653 AY788496 AY788589 AY788772 AY788532 NW6-4NW14-4NW70-4NW5-8NW24-6 Cervières FRANCE: California USA: NW13-3 El Guix SPAIN: NW35-15NW38-17 California USA: Montana USA: NW37-2 Maryland USA: NW34-4 Colorado USA: NW35-10 Arizona USA: NW62-1 La Selva RICA: COSTA NW62-4 Colorado USA: NW37-3York New USA: NW34-5 AF187746 Butterfly farm supplier RICA: COSTA NW20-4 Butterfly farm supplier RICA: COSTA NW27-6 AF187752 AY788743 La Selva RICA: COSTA NW27-7 AY090226 Colorado USA: NW61-1 AY788620 AY788744 AY788504 AF187765 California USA: NW7-1 AY090227 AY090193 AF187771 Arizona USA: NW27-4 AY788730 AF187797 AF187770 AY788505 AY788745 Arizona USA: NW23-5 AF187735 AY090195 AY090159 AY788746 Chiapas MEXICO: NW69-8 AY788491 AY788747 AY788727 AF187757 AY788506 NW73-14 Texas AY788725 USA: AY090161 AY788507 California USA: NW95-5 AY788508 AF187772 AY788488 AY788726 AF187773 Pissoderi GREECE: NW26-1 AF187786 AY788486 SWEDENNW25-3 Stockholm SWEDEN: AY788728 AY788487 AY788729 AY788731 Aude FRANCE: Buryatia AY788489 RUSSIA: AY788490 AF187788 AY788492 Buryatia RUSSIA: AF187791 AY788732 AF187808 AY788660 AY788733 AF187764 AY788493 AY788734 AY788787 AY788494 AF187742 AY788785 AF187740 AY788656 AY788659 AY788495 AY788547 AY788774 AY788545 AY788799 AY788776 AY788786 AY788534 AF187762 AY788657 AY788655 AY788559 AY788536 AF187780 AY788546 AY788777 AY090194 AY788775 AY788778 AY788537 AY090160 AY788535 AY788538 Kamilla cymodoce Kallima inachus Kallima paralekta Kallimoides rumia Catacroptera cloanthe Mallika jacksoni bisaltide Doleschallia milca Vanessula Euphydryas aurinia Euphydryas chalcedona Euphydryas desfontainii Euphydryas editha Euphydryas gillettii Euphydryas phaeton Chlosyne acastus Chlosyne cyneas Chlosyne gaudealis Chlosyne gorgone Chlosyne harrisii Chlosyne janais Chlosyne lacinia Chlosyne narva Chlosyne nycteis Chlosyne palla Chlosyne theona Dymasia dymas Microtia elva elada Texola arachne Poladryas Melitaea arduinna Melitaea britomartis Melitaea cinxia Melitaea deione Melitaea didymoides Melitaea latonigena Melitaeini Higher taxon Species code Voucher Locality COI EF1-

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 PHYLOGENY OF NYMPHALINAE 249 wingless a NW34-10NW34-11NW27-11NW23-6 Mohafazat Beharré LEBANON: NW24-13 Mohafazat Kesronan LEBANON: Hebei Province CHINA: NW89-9 Pissoderi GREECE: PE-10–20 Laus de Cervières FRANCE: NW76-6NW12-6 AF187796 AF187803 Sucumbios ECUADOR: NW22-4 Quebrada Chaupimayo Cuzco, PERU: NW24-4 AY788779 AY788781 Esmeraldas ECUADOR: NW104-12 Texas USA: NW76-2 AY788539 AY788630 AY788541 AF187804 Monteverde RICA: COSTA NW105-3 AF187812 Monteverde RICA: COSTA NW76-5 Gamboa PANAMA: AY788749NW104-3 AY788780 AY788783 AF187810 Pichincha ECUADOR: NW92-5 Achiote Road PANAMA: NW91-9 AY788510 AY788540 AY788543 AY788629 Esmeraldas ECUADOR: NW76-3 AY788782Alta to Gloria Path PANAMA: NW108-11 AY788611 AY788748 São Paulo BRAZIL: NW76-8 AY788542 AF187743 Sucumbios ECUADOR: NW85-16 AF187790 AY788714 AY788509 Esmeraldas ECUADOR: NW76-6 PERU AY788713NW72-4 AY788715 AY788475 Esmeraldas ECUADOR: NW11-4 AY788612 Gamboa PANAMA: AY788623 AY788474 AF187806 AY788618 AY788617NW11-10 AY788476 ECUADORNW67-3 AY788622 AY788717 AY788737 AY788716 Ontario AY788723 CANADA: NW64-2 AY788722 British Columbia CANADA: NW34-6 AY788736 AY788478 British Columbia CANADA: AY788498 NW35-11 AY788477 AY788484 AY788483 AY788625 AY788624 California USA: NW34-7 AY788627 AY788497 Michoacán MEXICO: NW67-14 AY788739 Colorado USA: NW34-2 AY788738 AY788628 AY788741 Mazatlan MEXICO: NW67-9 AY788500 Colorado USA: NW91-6 AY788499 AY788742 AY788502 Oregon USA: AY788641 AF187755NW91-11 AF187785 Minnesota USA: NW76-4 AY788503 AY788761 AY788626 Mexico State MEXICO: AY090192 AF187747 AY788791 Sucumbios ECUADOR: Sucumbios ECUADOR: AY788521 AY788654 AY788740 AY090158 AY788789 AY788551 Esmeraldas ECUADOR: AY156640 AY788773 AY788501 AY156631 AY788549 AF187798 AY788793 AY788533 AY788792 AF187792 AY788795 AY788553 AY788552 AY156684 AF187800 AY788794 AY788683 AY788555 AY156662 AF187807 AY788681 AY788790 AY788796 AY788554 AY788682 AY788821 AY788797 AY788798 AY788819 AY788550 AY788556 AY788820 AY788581 AY788557 AY788558 AY788579 AY788580

Melitaea persea Melitaea punica Melitaea scotosia Melitaea trivia Melitaea varia Gnathotriche exclamationis Higginsius fasciatus Anthanassa drusilla Anthanassa texana Anthanassa ardys Anthanassa otanes Anthanassa tulcis Castilia eranites Castilia ofella Eresia clio Eresia coela Eresia eunice Eresia letitia Eresia quintilla Eresia pelonia Eresia sestia leucodesma Janatella Mazia amazonica Phyciodes batesii Phyciodes cocyta Phyciodes mylitta Phyciodes orseis Phyciodes pallescens Phyciodes pallida Phyciodes phaon Phyciodes picta Phyciodes pulchella Phyciodes tharos Phyciodes graphica trimaculata Telenassa anieta Tegosa tissoides Tegosa Higher taxon Species code Voucher Locality COI EF1-

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 250 N. WAHLBERG ET AL.

APPENDIX 2 Tribe Victorinini Scudder, 1893 Anartia Hübner, 1819 Proposed higher classification of the subfamily = Celaena Doubleday, 1849 Nymphalinae based on results presented in this paper. = Celoena Boisduval, 1870 Genera marked with asterisks were not sampled for = Anartiella Fruhstorfer, 1907 this study. For a full synonymic list of the species Siproeta Hübner, 1823 please see http://www.zoologi.su.se/research/wahlberg = Victorina Blanchard, 1840 = Amphirene Doubleday, 1845 Subfamily NYMPHALINAE Rafinesque, 1815 = Amphirene Boisduval, 1870 incertae sedis = Aphnaea Capronnier, 1881 Pycina* Doubleday, 1849 Napeocles Bates, 1864 Kallimoides Shirôzu & Nakanishi, 1984 Metamorpha Hübner, 1819 Vanessula Dewitz, 1887 Tribe Junoniini Reuter, 1896 Rhinopalpa Felder & Felder, 1860 Junonia Hübner, 1819 Tribe Coeini Scudder, 1893 = Alcyoneis Hübner, 1819 Historis Hübner, 1819 = Aresta Billberg, 1820 = Coea Hübner, 1819 = Kamilla Collins & Larsen, 1991 = Aganisthos Boisduval & Le Conte, 1835 Salamis Boisduval, 1833 = Megistanis Doubleday, 1845 Yoma Doherty, 1886 = Megistanis Boisduval, 1870 = Yoma de Nicéville, 1886 Baeotus Hemming, 1939 Protogoniomorpha Wallengren, Tribe Nymphalini Rafinesque, 1815 1857 Colobura Billberg, 1820 Precis Hübner, 1819 = Gynoecia Doubleday, 1845 = Coryphaeola Butler, 1878 Tigridia Hübner, 1819 = Kallimula Holland, 1920 = Callizona Doubleday, 1848 Hypolimnas Hübner, 1819 Smyrna Hübner, 1823 = Esoptria Hübner, 1819 Antanartia Rothschild & Jordan, 1903 = Diadema Boisduval, 1832 Araschnia Hübner, 1819 = Euralia Westwood, 1850 Symbrenthia Hübner, 1819 = Eucalia Felder, 1861 = Laogona Boisduval, 1836 Tribe Kallimini Doherty, 1886 = Brensymthia Huang 2001 Kallima Doubleday, 1849 Mynes Boisduval, 1832 Doleschallia Felder & Felder, 1860 Hypanartia Hübner, 1821 = Apatura Hübner, 1819 = Eurema Doubleday, 1845 Catacroptera Karsch, 1894 Vanessa Fabricius, 1807 Mallika Collins & Larsen, 1991 = Nymphalis Latreille, 1804 Tribe Melitaeini Newman, 1870 = Cynthia Fabricius, 1807 incertae sedis = Pyrameis Hübner, 1819 Antillea* Higgins, 1959 = Bassaris Hübner, 1821 Atlantea* Higgins, 1959 = Ammiralis Rennie, 1832 Gnathotriche Felder & Felder, 1862 = Neopyrameis Scudder, 1889 = Gnathotrusia Higgins, 1981 = Fieldia Niculescu, 1979 Higginsius Hemming, 1964 Aglais Dalman, 1816 = Fulvia Higgins, 1959 = Ichnusa Reuss, 1939 Chlosyne Butler, 1870 = Inachis Hübner, 1819 = Morpheis Geyer, 1833 = Hamadryas Hübner, 1806 = Synchloe Doubleday, 1845 Nymphalis Kluk, 1780 = Anemeca Kirby, 1871 = Scudderia Grote, 1873 = Coatlantona Kirby, 1871 = Euvanessa Scudder, 1889 = Limnaecia Scudder, 1872 = Roddia Korshunov, 1996 = Charidryas Scudder, 1872 Polygonia Hübner, 1819 = Thessalia Scudder, 1875 = Kaniska Moore, 1899 Microtia Bates, 1864 = Eugonia Hübner, 1819 Texola Higgins, 1959 = Comma Rennie, 1832 Dymasia Higgins, 1960 = Grapta Kirby, 1837 Poladryas Bauer, 1961

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251 PHYLOGENY OF NYMPHALINAE 251

Subtribe Euphydryina Higgins, 1978 Subtribe Phyciodina Higgins, 1981 Euphydryas Scudder, 1872 Phyciodes Hübner, 1819 = Lemonias Hübner, 1806 Phystis* Higgins, 1981 = Occidryas Higgins, 1978 Anthanassa Scudder, 1875 = Eurodryas Higgins, 1978 = Tritanassa Forbes, 1945 = Hypodryas Higgins, 1978 Dagon* Higgins, 1981 Subtribe Melitaeina Newman, 1870 Telenassa Higgins, 1981 Melitaea Fabricius, 1807 Ortilia* Higgins, 1981 = Lucina Rafinesque, 1815 Tisona* Higgins, 1981 = Schoenis Hübner, 1819 Tegosa Higgins, 1981 = Cinclidia Hübner, 1819 Eresia Boisduval, 1836 = Mellicta Billberg, 1820 Castilia Higgins, 1981 = Didymaeformia Verity, 1950 Janatella Higgins, 1981 = Athaliaeformia Verity, 1950 Mazia Higgins, 1981

© 2005 The Linnean Society of London, Biological Journal of the Linnean Society, 2005, 86, 227–251