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