We obtained 1403 base pairs of mtDNA sequences of cyt b and ND2 from four specimens of the Pale-faced Bareeye (Skutchia borbae, representing a monotypic genus of the Thamnophilidae) and evaluated for the first time the species' phylogenetic position among core obligate army-ant-following antbirds. All trees placed Skutchia borbae consistently and with high statistical support as sister to the Black-spotted Bare-eye (Phlegopsis nigromaculata) in a clade with the two known species of Phlegopsis. This result supports the return of S. borbae to the genus Phlegopsis, where it was originally placed. When the plumage characters originally used to diagnose Skutchia are mapped onto the phylogeny, it is evident that they have evolved convergently multiple times in core obligate army-ant-following antbirds. This discovery further supports merging Skutchia with Phlegopsis.
The monotypic genus Skutchia was erected by Willis (1968) for the species originally described as Phlegopsis borbae (Pale-faced Bare-eye) on the basis of a single specimen from Borba, on the east (right) bank of the lower Madeira River (Hellmayr 1907). The species' range is bounded by the Madeira and Tapajós rivers in the Brazilian Amazon. Until Willis (1968), knowledge of P. borbae was restricted to two specimens (including the holotype) obtained from what appeared to be the margins of its distribution. Willis (1968) reported extensive data on behavior and ecology of P. borbae and demonstrated that it belongs to the guild of obligate army-ant-following birds along with Phlegopsis and several other genera. He argued that P. borbae was morphologically so distinct from any other army-ant-following species that it should be treated as separate genus, which he named Skutchia in honor of Alexander Skutch, the famous neotropical ornithologist.
In the original diagnosis, Willis (1968) Skutchia was distinguished from Phlegopsis by (1) the reduced extent and paler color of the bare skin in its orbital area, (2) its patch of black stiffened feathers above the eye, (3) the distinct patch of elongated feathers on the lores, (4) the body coloration in which adult males are predominantly brownish rather than blackish, and (5) its distinct pectoral band. Since the description Skutchia, the placement of P. borbae in this genus has been followed by most authors (Zimmer and Isler 2003, Brumfield et al. 2007, Remsen et al. 2009), even though the consistency of the diagnosis with respect to Phlegopsis has been questioned at least once (Pinto 1978).
Brumfield et al. (2007) proposed the first phylogenetic hypothesis for the army-ant-following antbirds (Thamnophilidae), but for lack of tissues, Skutchia was the only obligate army-ant follower genus they did not analyze (R. Brumfield, pers. comm.). Nonetheless, from its taxonomic history, Brumfield et al. (2007) argued that Skutchia is likely positioned within Phlegopsis.
Here, we analyze 1403 bp of cytochrome b (cyt b) and NADH dehydrogenase 2 (ND2) mtDNA sequences from four specimens of S. borbae in association with the data presented by Brumfield et al. (2007) to provide the first phylogenetic analysis that includes Skutchia. The analysis demonstrates not only the phylogenetic position Skutchia but also evaluates the plumage characters used to diagnose the genus.
In addition to Skutchia, we sampled all core obligate army-ant-following genera of the Thamnophilidae (Pithys, Gymnopithys, Rhegmatorhina, Phlegopsis, and Phaenostictus), shown to be consistently monophyletic (clade A of figure 2 in Brumfield et al. 2007). We obtained sequences for Skutchia from ethanol-preserved muscle tissues vouchered by four study skins deposited at the ornithological collection of the Museu Paraense Emílio Goeldi in Belém, Brazil (MPEG) and collected at two localities separated by about 200 km in the southern part of the Brazilian state of Amazonas (Table 1). To ensure dense taxon sampling of Phlegopsis, the genus purportedly closest to Skutchia, we also sequenced at least one representative of each taxon in this genus recognized by Zimmer and Isler (2003; see Table 1). All remaining sequences used in this study were downloaded from GenBank (Brumfield et al. 2007), representing the White-plumed Antbird (Pithys albifrons: EF640047, EF639980), White-masked Antbird (P. castaneus: EF640060, EF639993), Bicolored Antbird (Gymnopithys leucaspis leucaspis: EF640062, EF639995, and G. l. bicolor: EF640061, EF639994), Rufous-throated Antbird (G. rufigula: EF640064, EF639997), White-throated Antbird (G. salvini; EF640016, EF639949), Lunulated Antbird (G. lunulatus: EF640063, EF639996), Bare-eyed Antbird (Rhegmatorhina gymnops: EF640065, EF639998), White-breasted Antbird (R. hoffmannsi: EF640050, EF639983), Hairy-crested Antbird (R. melanosticta: EF640066, EF639999), Scale-backed Antbird (Willisornis poecilinotus: EF640020, EF639953), Black-spotted Bare-eye (Phlegopsis nigromaculata: EF640046, EF639979), Reddish-winged Bare-eye (P. erythroptera: EF640068, EF640001), and Ocellated Antbird (Phaenostictus mcleannani; EF640045, EF639978).
General information on the samples of Skutchia borbae, Phlegopsis erythroptera, and Phlegopsis nigromaculata sequenced in this study. All samples were ethanol-preserved muscle tissues. Voucher numbers refer to study skins deposited at the ornithological collection of the Museu Paraense Emílio Goeldi in Belém, Brazil (MPEG).
For all samples shown in Table 1, we amplified portions of the cyt b and ND2 mitochondrial protein-coding genes by PCR under standard conditions (available upon request) with the following primers, respectively: L14990 (Helm-Bychowski and Cracraft 1993)/H16065 (Kocher et al. 1989) and L5215 (Hackett 1996)/H6313 (Sorenson et al. 1998). PCR products were cycle-sequenced with a DYEnamic ET Dye Terminator kit (GE Healthcare) and run on a DNA MegaBace (GE Healthcare) sequence analyzer according to the manufacturer's instructions. The following measures were taken to ensure that the DNA fragments sequenced were accurate and of mitochondrial origin (not pseudogenes): (1) both DNA strands were sequenced, (2) sequences were inspected for insertions, deletions, and stop codons that would result in a nonfunctional protein with the aid of the software BioEdit (Hall 1999) and DAMBE (Xia and Xie 2001), and (3) sequences were expected to exhibit high transition-to-transversion substitution ratios characteristic of mitochondrial, not nuclear, substitution patterns.
Phylogenetic analyses were performed on the concatenated cyt b/ND2 dataset, after a partition-homogeneity test as implemented in PAUP* 4.0b10 (Swofford 2002) failed to detect significantly different phylogenetic signals between the two genes analyzed (P > 0.05). Trees were rooted in Willisornis poecilinotus (a regular army-ant follower), shown with strong statistical support to be the sister taxon to the entire core obligate army-ant-following clade (Brumfield et al. 2007).
Maximum-parsimony (MP) and maximum-likelihood (ML) searches were conducted with PAUP* 4.0b10, with MP based on unweighted sequence data and ML following the model of molecular evolution chosen by the Akaike information criterion (AIC) that fit our dataset, as selected by the software Modeltest 3.7 (Posada and Crandall 1998). We used 1000 nonparametric bootstrap replications to evaluate confidence levels of nodes for all phytogenies obtained with MP and ML (Felsenstein 2004). Using the software MrBayes 3.1.2, we also assessed the support of nodes in the phylogeny with a site-partitioned Bayesian analysis (Ronquist and Huelsenbeck 2003) with the following specifications: (1) assuming separate optimal ML models selected for each data partition by Modeltest 3.7; (2) running the Markov chain for 5 000 000 generations and sampling 1 tree every 1000 generations. We discarded the first 499 trees sampled (10% of the sample) as “burn-in”, with the remaining 4501 trees used to estimate posterior probabilities of the trees' topologies.
Distribution of plumage characters used by Willis (1968) to diagnose Skutchia among core obligate army-ant-following antbirds.
We mapped the plumage characters used by Willis (1968) to diagnose Skutchia onto the phylogeny of core obligate army-ant-following antbirds from figure 2 of Brumfield et al. (2007), complemented with our data for Skutchia borbae (Table 2). For each terminal taxon in the phylogeny, we examined adult museum specimens at MPEG and used Zimmer and Isler (2003) as a reference to code characters of taxa missing from the collection (Gymnopithys leucaspis bicolor, G. lunulatus, and Pithys castaneus). All characters were scored as unordered and either binary or multistate. We used MACLADE (version 3.07; Maddison and Maddison 1992) to reconstruct patterns of character evolution and to assess whether characters have significant phylogenetic signal. We calculated the consistency index (CI) and retention index (RI) for each character mapped onto the phylogeny. We used Maddison and Slatkin's (1991) randomization test to assess the phylogenetic signal or inertia of these characters. For each character, we randomized the character states 1000× on the phylogeny and compared the reconstructed number of character-state changes to the random distribution of character-state changes. Characters were reconstructed by both acctran and deltran optimization.
Cyt b and ND2 sequences from 11 individuals of Skutchia borbae, Phlegopsis nigromaculata, and P. erythroptera obtained in this study (see Table 1) were trimmed to 453 bp and 950 bp, respectively, spanning positions 15307 to 15759 and 5286 to 6235 of the cyt b and ND2 sequences of the domestic chicken (Gallus gallus) (Desjardins and Morais 1990). Aligned sequences revealed no signs of saturation (plot available upon request) and were deposited in GenBank (see Table 1).
The newly sequenced dataset of 1403 nucleotides of Skutchia and Phlegopsis was combined with data on Pithys, Phaenostictus, Gymnopithys, Rhegmatorhina, Phlegopsis, and Willisornis downloaded from GenBank (Brumfield et al. 2007). This 25-taxon dataset had 464 variable sites (33%), 347 of which were phylogenetically informative. An equally weighted MP heuristic search of the entire dataset resulted in a single most parsimonious tree (length 949; CI = 0.57; RI = 0.76; not shown). Modeltest analysis selected the TrN + 1 + G model of nucleotide substitution with estimated base frequencies (A = 0.32, C = 0.34, G = 0.09, T = 0.25), substitution model (A-C = 1, A-G = 22.77, A-T = 1, C-G = 1, C-T = 15.04, G-T = 1), proportion of invariable sites (I = 0.60), and rates for variable sites following a gamma distribution (Γ shape parameter = 4.51). The ML tree (-In likelihood = 6409.32293; Fig. 1) had a topology similar to that estimated by MP, except for the position of the Pithys /Phaenostictus clade, which was placed as sister to the Rhegmatorhina/Gymnopithys clade in the latter analysis.
Regardless of the tree-building method employed, all phylogeny estimates placed Skutchia borbae consistently and with high statistical support in a clade with the two known species of the genus Phlegopsis; relationships within this clade were also well resolved and statistically supported and showed S. borbae and P. nigromaculata as sister taxa, to the exclusion of P. erythroptera (Fig. 1). MP and ML estimates differed, respectively, in placing an entire Pithys/Phaenostictus/Rhegmatorhina/Gymnopithys clade or just a Pithys/Phaenostictus clade as sister to the Phlegopsis/Skutchia clade. Irrespective of the tree-building method, a very short branch and poorly supported node joins the Phlegopsis/Skutchia clade to its sister group (Fig. 1), reflecting the same lack of resolution concerning the sister taxon of Phlegopsis found by Brumfield et al. (2007).
When mapped onto molecular phylogeny (cladograms not shown), none of the five plumage characters used by Willis (1968) to distinguish Skutchia from other obligate army-ant-following antbird genera exhibited a significant phylogenetic signal (Table 3). The characters (3) patch of stiffened black feathers above the eye and (4) patch of elongated feathers on the lores represent autapomorphies that have evolved independently in Skutchia and Phaenostictus and in Skutchia and Pithys, respectively. Similarly, character 5 (extensive barring on the chest and belly) is an autapomorphy that has evolved independently in Skutchia, Rhegmatorhina, and Phaenostictus. The lack of black plumage in males of Skutchia (character 1) is also shared with two independent lineages in the genus Gymnopithys, and all have evolved from a common ancestor in which males had black in the plumage. Finally, the light bare orbital skin (character 2) of Skutchia represents either a reversal from the dark red shown by Phlegopsis (in acctran optimizations) or retention of a character inherited from an ancestor with light bare skin (in deltran optimizations).
The phylogenies in Fig. 1 and Brumfield et al. (2007) confirm the long-standing belief based on external morphology, vocalizations, behavior, and ecology that Skutchia belongs in the clade of core obligate army-ant-following antbirds (Willis 1968, Zimmer and Isler 2003). A most unexpected result, however, was the clear sister relationship between Phlegopsis nigromaculata and Skutchia to the exclusion of P. erythroptera, rendering the genus Phlegopsis paraphyletic (Fig. 1). This finding is also supported by comparative levels of pairwise uncorrected sequence divergence, which average lower between Skutchia and P, nigromaculata (5.45%; range 4.78–6.06%) than between P. nigromaculata and P. erythroptera (7.69%; range 7.2–8.48%).
Measures of homoplasy for plumage characters used by Willis (1968) to diagnose the genus Skutchia. Consistency index (CI) and retention index (RI) are for individual characters mapped onto the phylogeny of core obligate army-ant-following antbirds presented by Brumfield et al. (2007), where Skutchia borbae is placed as the sister taxon of Phlegopsis nigromaculata as shown in Fig. 1. Signal calculations were done by the method of Maddison and Slatkin (1991).
When Hellmayr (1907) described and placed S. borbae in the genus Phlegopsis, he was motivated primarily by an overall resemblance between the brown plumages of the holotype of S. borbae (a male) and of female P. erythroptera. In the original description, Hellmayr (1907) never mentioned P. nigromaculata (the actual sister species of S. borbae; Fig. 1) because in this species the male is predominantly black (Zimmer and Isler 2003). Willis (1968) speculated that Hellmayr (1907) tried to reconcile the brown male plumage of S. borbae, when he placed it in Phlegopsis, in which males of both P. erythroptera and P. nigromaculata are mostly black, by considering the holotype to be an immature, despite the lack of any evidence for this conclusion. In fact, Willis (1968) discussed evidence indicating that the holotype of S. borbae is an adult male. This apparent contradiction was the main factor prompting Willis (1968) to propose the genus Skutchia. Willis (1968) noted that he and other ornithologists had been misled by Hellmayr's (1907) original description to expect that the adult male of P. borbae would be black, and it was not until Willis rediscovered the species in 1966 that ornithologists realized the adult male is brown.
The phylogenies presented in this study (Fig. 1) and in Brumfield et al. (2007) provide a novel framework for the study of character evolution in the clade of obligate army-ant-following antbirds, allowing a critical re-evaluation of the taxonomic usefulness of the characters used to distinguish Skutchia from other genera and species in this group. When mapped onto molecular phylogeny, all plumage characters originally used to distinguish Skutchia from other lineages in the clade (especially Phlegopsis) are revealed as also shared by the closely related genera Rhegmatorhina, Gymnopithys, Pithys, and Phaenostictus, likely as a result of convergence and/or retention of ancestral characters, weakening the validity of Skutchia as a monotypic genus. On the other hand, sexual dimorphism in plumage color (not mentioned by Willis 1968) exhibits a strong phylogenetic signal when mapped onto phylogeny (CI = 0.33, RI = 0.67, P = 0.03; cladogram and character-state matrix available upon request), and represents a shared character loss uniting P. nigromaculata and S. borbae to the exclusion of the strongly dimorphic P. erythroptera.
Inasmuch as generic limits are often drawn arbitrarily (see Willis 1968 for an interesting discussion on the subject pertaining to the obligate army-ant-following species) and are often difficult to define objectively, the data we present support the return of S. borbae to the genus Phlegopsis, where it was originally placed by Hellmayr (1907). Since the type species of Phlegopsis is P. nigromaculata (Gray 1855), merging Skutchia and Phlegopsis is a more conservative taxonomic approach to retaining Skutchia and erecting a new genus for P. erythroptera. This alternative is also at odds with levels of comparative genetic variation found in other genera of the core obligate army-ant-following clade, such as Gymnopithys, in which the range of uncorrected sequencedivergence levels (0.71–8.42%) is similar to that of the “inclusive” genus Phlegopsis as defined here (0.57–8.48%).
In conclusion, the genus Skutchia Willis, 1968, should henceforth be considered a junior synonym of Phlegopsis Reichenbach, 1850, which now includes three biological species in the following systematic sequence: P. erythroptera (Gould, 1855), P. nigromaculata (d'Orbigny and Lafresnaye, 1837), and P. borbae Hellmayr, 1907.
This study was funded by the Brazilian Ministry of the Environment (MMA) PROBIO initiative (#01/2003) in association with the Brazilian National Research Council (CNPq #0014-00/04). Field work resulting in the collection of samples of Skutchia was made possible by another PROBIO grant (edition 02/2001; CNPq #2504-00/02) and the Kanindé Foundation from Porto Velho, Rondônia. We are especially grateful to Marcos Pérsio Dantas Santos and several other collectors for contributing with specimens and tissue samples used in this study. Silvanira Barbosa and Soraya Andrade helped with laboratory work. Jason D. Weckstein helped with calculations of the Maddison and Slatkin (1991) test. Fred H. Sheldon and an anonymous reviewer provided insightful comments on this manuscript. During this study TCTB was supported by a master's degree fellowship from CNPq while MFCS held a productivity fellowship from CNPq (#309545/2003-0).