Molecular Phylogenetics and Evolution 65 (2012) 287–293 Contents lists available at SciVerse ScienceDirect Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev DNA sequence data reveal a subfamily-level divergence within Thamnophilidae (Aves: Passeriformes) ⇑ Gustavo A. Bravo , J.V. Remsen Jr., Bret M. Whitney, Robb T. Brumfield Museum of Natural Science and Department of Biological Sciences, Louisiana State University, Baton Rouge, LA 70803, USA article info abstract Article history: The Thamnophilidae is a diverse radiation of insectivorous passerine birds that comprises nearly 220 spe- Received 12 March 2012 cies and is mostly restricted to the lowlands and lower montane forests of the Neotropics. Current clas- Accepted 18 June 2012 sification within Thamnophilidae relies primarily on morphological variation, but recent incorporation of Available online 1 July 2012 molecular and vocal data has promoted changes at various taxonomic levels. Here we demonstrate that the genus Terenura is polyphyletic because Terenura callinota, T. humeralis, T. spodioptila, and T. sharpei are Keywords: phylogenetically distant from the type species of the genus, Terenura maculata. More importantly, the New subfamily former four species are not particularly closely related to any other thamnophilids and represent a clade New genus that is sister to all other members of the family. Because no genus name is available for this previously Avian systematics Suboscines undetected lineage in the Thamnophilidae, we describe the genus Euchrepomis for callinota, humeralis, Antwrens spodioptila, and sharpei, and erect the subfamily Euchrepomidinae. We discuss the taxonomic and evolu- Antbirds tionary significance of this divergent lineage. This study highlights the importance of taxonomic coverage and the inclusion of type taxa to redefine classifications to reflect accurately evolutionary relationships. Ó 2012 Elsevier Inc. All rights reserved. 1. Introduction lutionary diversity of the Thamnophilidae with numerous taxonomic consequences. Most such changes are descriptions of The Thamnophilidae is a diverse radiation of insectivorous pas- new species or reassessments of species limits (e.g. Cháves et al., serine birds that comprises approximately 220 species and is 2010; Isler and Whitney, 2011; and references therein). However, mostly restricted to the lowlands and lower montane forests of few studies have evaluated traditional classifications at deeper tax- the Neotropics (Zimmer and Isler, 2003). Traditional classification onomic levels (Aleixo et al., 2009; Bravo et al., 2012; Isler et al., within the family is based primarily on comparisons of plumage 2006; Moyle et al., 2009) due primarily to incomplete taxonomic and morphometric proportions of external features such as bill, sampling. Moyle et al. (2009) recognized two subfamilies, Myrmor- tail, wings, and tarsi (Cory and Hellmayr, 1924; Peters, 1951; Ridg- nithinae and Thamnophilinae, with the former consisting of the way, 1911; Sclater, 1890). Examination of internal morphological monotypic genera Myrmornis, Pygiptila, and Thamnistes, and features (e.g. the sound producing organ in birds, the syrinx) sug- the latter comprising all remaining genera in the family except gested that members of the family Thamnophilidae are diagnos- the genus Terenura, which they left unassigned to subfamily. Tradi- able anatomically from other passerine families (Ames, 1971). tional linear classifications place Terenura near Epinecrophylla, Isle- Furthermore, molecular phylogenetic studies have shown that ria, Myrmotherula, Microrhopias, Herpsilochmus, Formicivora, the family Thamnophilidae is a monophyletic group, that their Drymophila, and Hypocnemis, with which Terenura shares small closest relatives are other lineages of Neotropical suboscine passe- body size, small thin bills, and in some species, black and white rines in the furnariid radiation, and that phylogenetic reconstruc- streaking on the head and neck. As far as we can determine, the tions are not entirely congruent with traditional taxonomic monophyly of the genus Terenura and its relationships to the adja- classification (Bravo et al., 2012; Brumfield et al., 2007; Irestedt cent genera in linear sequences have never been formally ques- et al., 2004; Moyle et al., 2009; Sibley and Ahlquist, 1990). tioned. Recently, however, Terenura sharpei and T. humeralis were Molecular phylogenetic analyses coupled with analyses of vari- shown to be the sister group to all other Thamnophilidae (Bravo ation of their innate songs have promoted a reappraisal of the evo- et al., 2012; Brumfield and Edwards, 2007; Irestedt et al., 2004; Moyle et al., 2009), but lack of samples of the type species of the ⇑ genus, T. maculata, impeded certainty about the phylogenetic Corresponding author. Fax: +1 225 578 3075. placement of Terenura. E-mail addresses: [email protected] (G.A. Bravo), [email protected] (J.V. Remsen Jr.), [email protected] (B.M. Whitney), brumfl[email protected] Cabanis and Heine (1859–1860) named the genus Terenura for (R.T. Brumfield). the species Myiothera maculata (Wied, 1831) of southeastern 1055-7903/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ympev.2012.06.016 288 G.A. Bravo et al. / Molecular Phylogenetics and Evolution 65 (2012) 287–293 Brazil, which had been placed in Formicivora by Sclater (1858). gene 2 – RAG2, 1152 bp) following the methods described in Groth Cabanis and Heine did not provide a diagnosis or description of and Barrowclough (1999) and Barker et al. (2002). Additionally, their new genus, but Terenura translates as ‘‘soft tail’’ in Greek (Jo- some sequences were obtained from previous publications of our bling, 2010). The species Formicivora callinota (Sclater, 1855) was own work (Bravo et al., 2012; Brumfield and Edwards, 2007; Brum- subsequently included in Terenura by Taczanowski and von Ber- field et al., 2007; Derryberry et al., 2011; Gómez et al., 2010; Moyle lepsch (1885) without comment, but perhaps because Sclater et al., 2009). Analyses were conducted using a concatenated six- noted in his original description of callinota that ‘‘it [callinota] must gene alignment containing 7025 bp. be placed next to the Brazilian Formicivora maculata ...with which We edited sequences using Sequencher 4.7 (Gene Codes Corpo- it agrees in form and style of plumage.’’ This rationale apparently ration, Ann Arbor, MI) and checked that protein-coding sequences led to the association of callinota with maculata. Subsequent classi- did not include stop codons or anomalous residues. We aligned se- fications continued to place both in Terenura, along with three quences using the program MAFFT v. 6 (Katoh et al., 2002), and ob- additional species described later, all noted as close relatives of cal- tained a concatenated dataset using Geneious Pro v5.5 (Drummond linota and all described in Terenura: T. humeralis (Sclater and Salvin, et al., 2011). Newly obtained sequences were deposited in Gen- 1880), T. spodioptila (Sclater and Salvin, 1881), and T. sharpei (von Bank (Accession numbers JX213474–JX213578). Berlepsch, 1901). The classification of Terenura as containing five species remained stable until the addition of a sixth, newly discov- 2.2. Phylogenetic analyses ered species: T. sicki (Texeira and Gonzaga, 1983). At least two phe- notypic groups have been recognized within the genus based on We conducted ML analyses for six partition schemes under the plumage differences: the ‘‘streaked-headed’’ group consisting of GTR + C model of nucleotide substitution using RAxML 7.2.7 (Sta- T. maculata and T. sicki, and the ‘‘standard’’ Terenura consisting of matakis, 2006) on the Cipres Science Gateway V 3.1 (Miller et al., the remaining four species (Ridgely and Tudor, 1994). The 2010). We then calculated the Akaike Information Criterion (AIC) ‘‘streaked-headed’’ Terenura are restricted to the Atlantic Forest, for each partition and established that the most informative whereas the ‘‘standard’’ Terenura are found through much of scheme is the fully partitioned dataset (16 partitions; each codon Amazonia, the Guianan shield, and mid-elevations in the Andes position for each coding gene, and the nuclear intron for separate and southern Central American mountains (Zimmer and Isler, partitions). To evaluate nodal support of the fully partitioned data- 2003). set, we conducted a rapid bootstrap analysis in RAxML using 1000 Here, we present results of morphological and DNA-based phy- bootstrap replicates under the GTR + C model of nucleotide substi- logenetic analyses to test the monophyly of the genus Terenura and tution, following recommendations by the author in RAxML assess its phylogenetic position within the family. We demonstrate manual. that Terenura is polyphyletic and that a subset of its members, not Using the same partition strategy followed in the likelihood including the type species T. maculata, represents the sister clade analysis (16 partitions), we also performed a Bayesian analysis as to the rest of the Thamnophilidae. This subset of species must be implemented in Mr. Bayes 3.1.2 (Huelsenbeck and Ronquist, placed in a separate genus for which no name is available, and they 2001) on the Cipres Science Portal (Miller et al., 2010). To deter- deserve to be treated as a separate subfamily. We describe a new mine the best nucleotide substitution model for each partition, genus for these species, place them in a new subfamily, and discuss we used PAUP (Swofford, 2003) to obtain likelihood