Holotype

Study skin, Centro de Ornitología y Biodiversidad (CORBIDI) AV-12381; tissue, Louisiana State University Museum of Natural Science (LSUMZ) B-93477; adult male; 2.5 km west-northwest of Flor de Café, Cordillera Azul, San Martín Region, Peru (7.390°S, 76.320°W; Figure 1); elevation 1,570 m a.s.l.; collected July 12, 2016; prepared by A.E.M., original catalog number 193; audio-recorded by A.E.M., D.F.L., F.A., and J.F. (Macaulay Library of Natural Sounds [ML], Cornell Laboratory of Ornithology, Ithaca, New York, USA; audio: ML 46435931, ML 224879, ML 224880); GenBank sequences KY849953 (ND2) and KY849952 (ND3).

FIGURE 1.

Map of the Cordillera Azul showing the type locality, which is also the only known locality, of Myrmoderus eowilsoni (yellow star) in relation to other surveyed sites on ridges of the Cordillera Azul (orange circles; O'Neill et al. 2000, Alverson et al. 2001, Merkord et al. 2009, J. Bates personal communication, T. Mark personal communication, LSUMZ field expeditions). The white line represents the boundary of the Cordillera Azul National Park. Areas east of the Huallaga River between 1,300 and 1,700 m a.s.l. are shaded to emphasize potential areas of occurrence of M. eowilsoni. Inset map shows the location of the Cordillera Azul in Peru.

Diagnosis

We assign Myrmoderus eowilsoni to the genus Myrmoderus on the basis of its combination of black ear coverts, extensively rufous brown plumage, and blackish wing coverts with broad white or buff tips (see Isler et al. 2013). Additional characters shared with M. ferrugineus include bluish bare orbital skin, gray feet and tarsi, lack of white interscapular patch, and terrestrial walking behavior. Despite these similarities, it is readily distinguishable from M. ferrugineus by (1) crown and nape color, (2) supercilium color, (3) extent of bare orbital skin, (4) belly color, (5) breast color (females), and (6) song (Figures 2 and 3). Compared to M. ferrugineus, the new species has a colder brown crown and nape, a gray rather than white supercilium, and less extensive bluish bare orbital skin. The belly of M. eowilsoni is black (males) or dark brown (females), whereas both sexes of M. ferrugineus show substantial white in this area. Additionally, females of M. eowilsoni have a rufous breast, whereas females of M. ferrugineus have a black breast (Figure 2A). The song differs substantially from M. ferrugineus by having fewer notes and a slower pace.

FIGURE 2.

Ventral (A) and lateral (B) views of both species of Myrmoderus. Left to right: male M. eowilsoni (CORBIDI 12381; holotype), female M. eowilsoni (CORBIDI 12380; mate of holotype), male M. ferrugineus (LSUMZ 178458), and female M. ferrugineus (LSUMZ 178456). Scale bars = 5 cm.

FIGURE 3.

Sonograms of representative vocalizations of Myrmoderus species. Recordings are available at Macaulay Library (ML), Xeno-canto.org (XC), or Isler and Whitney (2002), with the exception of Macedo recordings, which are in the collection of the recordist and used here with permission. Specimen number, recordist, date and location recorded, and archival catalog number follow in parentheses, as available. (A) Song of holotype male M. eowilsoni (CORBIDI 12381; D.F.L.; July 12, 2017; ML224880). (B) Song of female M. eowilsoni (CORBIDI 12380; mate of holotype; D.F.L.; July 12, 2017; ML224880). (C, D) Variation among songs of female M. eowilsoni (D.F.L.; July 13, 2017; ML224882). (E) Song of female M. eowilsoni (LSUMZ 190882; D.F.L.; July 12, 2017; ML224881). (F) Chatter call of M. eowilsoni with insect noise between 3 and 5 kHz (CORBIDI 12380 or 12381; D.F.L.; July 12, 2017; ML224880). (G) Song of male M. ferrugineus (B. Whitney; Isler and Whitney 2002). (H) Song of female M. ferrugineus (B. Whitney; Isler and Whitney 2002). (I) Chatter call of M. ferrugineus (D.F.L.; March 13, 2011; Suriname, Palumeu; XC74997). (J) Song of male M. ruficauda (T. Parker III; Isler and Whitney 2002). (K) Song of male M. loricatus (G. Macedo field recording no. 0327; November 30, 2014; Brazil, Rio de Janeiro, Estrada do Contorno). (L) Song of female M. loricatus (G. Macedo field recording no. 0328; November 30, 2014; Brazil, Rio de Janeiro, Estrada do Contorno). (M) Song of male M. squamosus (G. Macedo field recording no. 0469; September 3, 2015; Brazil, São Paulo, Paranapiacaba). (N) Song of female M. squamosus (G. Macedo field recording no. 0469; September 3, 2015; Brazil, São Paulo, Paranapiacaba).

Description of Holotype

Capitalized color names are based on Ridgway (1912), and alphanumeric color codes in parentheses refer to Munsell (no date). Crown and nape Warm Sepia (10YR3/4) with feather tips edged slightly darker, imparting a faintly scaled appearance. Forehead and lores black. Back Brussels Brown (5YR3/6) tending toward Amber Brown (5YR4/8) on upper scapulars and upper mantle. Two mantle feathers have a small amount of white fringing toward the vane bases, suggesting a rudimentary interscapular patch, but otherwise mantle feather bases are variably dark gray to blackish. Rump and uppertail coverts Brussels Brown (5YR3/6). Dorsal side of tail Bone Brown (7.5YR2/2) tending toward Fuscous-Black (slightly darker than 10YR2/2) at tip with outer vanes fringed Brussels Brown (5YR3/6). Ventral side of tail uniformly Fuscous (10YR2/2). Undertail coverts Brussels Brown (5YR3/6). Chin, throat, auriculars, breast, sides, upper flanks, and belly black. Lower belly color blending toward Raw Umber (10YR3/4) on vent and lower flanks. Supercilium Light Gull Gray (N6.75), beginning above anterior edge of eye and extending to and broadening at nape. Lesser coverts black with narrow white tips, becoming entirely white on leading edge of wing. Greater and median secondary coverts black, broadly tipped Warm Buff (7.5YR8/8) to Ochraceous-Buff (7.5YR6/8), forming 2 obvious wing bars. Primary coverts black, narrowly tipped Warm Buff (7.5YR8/8). Underwing coverts blackish, greater coverts tipped white. Primaries and secondaries Fuscous-Black (slightly darker than 10YR2/2) fringed with Brussels Brown (5YR3/6). Tertials Fuscous (10YR2/2) with diffuse black subterminal spot on secondaries 7 and 8 and diffusely tipped with Cinnamon (5YR5/8) on outer vane. Soft-part colors recorded at the time of collection: irides dark brown, skin of orbit blue-gray, feet and tarsi dark gray, mandible blue-gray, and maxilla blackish with blue-gray tomium. Measurements are provided in Table 1; no molt or fat, stomach contained insect parts, left testis 4.5 × 2 mm.

TABLE 1.

Measurements from Myrmoderus eowilsoni specimens (n = 8).

Paratypes

There are 7 paratypes of M. eowilsoni. Six were collected near Flor de Café between July 12 and 19, 2016 (CORBIDI 12380, 12382, 12383; LSUMZ 190882, 190883, 190884), and one was collected 6 km southeast of Flor de Café on September 30, 2016 (LSUMZ 190885). Data from field labels and morphometric data for all specimens of M. eowilsoni are shown in Table 1. Female plumage differs strikingly from that of males in the following characters: throat and malar white with fine black tips, more extensively black toward chin; breast and sides between Xanthine Orange (5YR6/10) and Amber Brown (5YR4/8); center of belly Dark Gull Gray (N4.25) blending to Raw Umber (10YR3/4) on the surrounding belly and through the flanks and vent. All individuals of both sexes show an area of paler Amber Brown (5YR4/8) on the upper mantle and scapulars, but to varying degrees of intensity and extent. The tips of the lesser coverts are slightly variable in color, with 2 females (LSUMZ 190882 and 190884) showing buffy tips and all other individuals showing exclusively white tips. The wing bars formed by the tips of the median coverts are also slightly variable in color, with some individuals tending toward white. The amount of Raw Umber (10YR3/4) on the belly is variable among females, being largely replaced by gray on one individual (CORBIDI 12380). A male and a female in formative plumage (LSUMZ 190883 and 190884), aged on the basis of molt limits in the rectrices and between the greater secondary and greater primary coverts and by the presence of bursa; like adults in definitive plumage, except that the young female has a nearly pure white throat and chin with greatly reduced scaling.

Etymology

We name Myrmoderus eowilsoni in honor of Dr. Edward Osborne Wilson to recognize his tremendous devotion to conservation and his patronage of the Rainforest Trust, which strives to protect the most imperiled species and habitats in the Neotropics and across the globe. We select the English name to draw attention to the little known but biogeographically important and biodiverse mountain range that contains the type locality of the species.

Phylogenetic Relationships

To assess the evolutionary relationships within Myrmoderus, we estimated a phylogeny based on sequence data from all taxa in the genus, using 2 or 3 individuals of each taxon except M. r. ruficauda, for which we had genetic data from only one individual (Appendix Table 3). We obtained sequences of 2 mitochondrial genes, NADH dehydrogenase subunit 2 (ND2; 1,041 base pairs) and NADH dehydrogenase subunit 3 (ND3; 351 base pairs) from 11 individuals sequenced previously by Bravo (2012). To augment this dataset, we generated new sequences from 2 individuals of M. eowilsoni and 2 individuals of M. ferrugineus elutus Todd, 1927. For these 4 individuals, we extracted total DNA from 25 mg of pectoral muscle using a Qiagen DNEasy kit using the manufacturer's protocol and performed polymerase chain reactions (PCR) for ND2 and ND3 following the protocol of Bravo (2012). PCR product was purified and sequenced at Eton Biosciences (Durham, North Carolina, USA) or the LSU Genomics Facility (Baton Rouge, Louisiana, USA). Four outgroup taxa from the Bravo (2012) dataset were used to root the tree: Barred Antshrike (Thamnophilus doliatus Linnaeus, 1764), Spotted Antbird (Hylophylax naevioides Lafresnaye, 1847), White-bellied Antbird (Myrmeciza longipes Swainson, 1825), and White-plumed Antbird (Pithys albifrons Linnaeus, 1766). All sequence data are deposited in GenBank (Appendix Table 3).

Each new sequence was cleaned manually and assembled into contigs in Geneious 10.1.3 (Kearse et al. 2012). The 2 loci were aligned separately in MUSCLE (Edgar 2004) from within Geneious using default values and then concatenated. Using PartitionFinder 2 (Lanfear et al. 2017) to compare DNA substitution models available in RAxML 8 (Stamatakis 2014) in a corrected Akaike's Information Criterion (AIC_c) framework gave a 3-partition scheme that pertained to the first, second, and third codon positions and an overall GTR+I+G model of DNA substitution. We used these settings to conduct a maximum-likelihood analysis in RAxML 8.2.10 (Stamatakis 2014) and selected the tree with the best likelihood score from 100 independent tree searches. Statistical support for the topology of this tree was then computed with the automatic majority-rule convergence criterion in RAxML 8.2.10 (Stamatakis 2014), resulting in 800 bootstrap replicates (Felsenstein 1985) and a final majority-rule consensus tree (Figure 4). When selecting among the substitution models available in MrBayes 3.2.6 (Ronquist et al. 2012), PartitionFinder 2 (Lanfear et al. 2017) gave the same partition scheme as for the maximum-likelihood analysis, but with substitution models of GTR+G, HKY+I, and GTR+I for the first, second, and third codon positions, respectively. We used these partitions and substitution models to perform the Bayesian analysis, comprising 4 independent runs of 4 MCMC chains, a chain temperature of 1.75, sampling for 20 million generations, a sampling frequency of 1,000, and a burn-in of 20%. We set a variable substitution-rate prior, an exponential alpha of 0.05, and a flat Dirichlet prior on the exchangeability parameters and base frequencies. Results from our Bayesian analyses were visualized in Tracer 1.6.0 (Rambaut et al. 2014) to assess convergence and stationarity. All runs reached convergence, and the average standard deviation of split frequencies from MrBayes 3.2.6 (Ronquist et al. 2012) was 0.002. Topologies of the maximum-likelihood and Bayesian trees are identical, so we report the maximum-likelihood tree with nodal support values from both methods (Figure 4).

FIGURE 4.

Phylogeny of the genus Myrmoderus estimated from a concatenated ND2–ND3 dataset (1,392 bp) using a maximum-likelihood analysis. The Bayesian analysis produced a tree that was topologically identical. Nodal support values are the likelihood bootstrap values and Bayesian posterior probabilities, respectively.

We recovered M. eowilsoni as sister to M. ferrugineus with high support in all analyses (Figure 4). Myrmoderus eowilsoni has an uncorrected pairwise genetic distance of 8.4% in ND2 and 8.7% in ND3 from M. f. ferrugineus (found north of the Amazon River), and 9.2% in ND2 and 7.6% in ND3 from M. f. elutus (found south of the Amazon River). These distances are high when compared to those between other antbird sister-species, even within the same genus (M. loricatus and M. squamosus: ND2 uncorrected pairwise distance of 4.5%). Myrmoderus f. elutus has an uncorrected pairwise distance of 3.6% in ND2 and 4.3% in ND3 from M. f. ferrugineus, although it differs only slightly in plumage (Todd 1927) and voice (average 5.7 notes second⁻¹ in M. f. elutus, n = 5 individuals vs. average 7.0 notes second⁻¹ in M. f. ferrugineus, n = 20 individuals).

Isler et al. (2013) presented a taxonomic revision of the highly polyphyletic Myrmeciza antbirds in which they resurrected the genus Myrmoderus (Ridgway 1909) for M. loricatus, M. squamosus, M. ruficauda, and M. ferrugineus. Although Myrmoderus (sensu Isler et al. 2013) formed a clade supported by molecular, morphological, and some behavioral data, taxonomic relationships within this clade were unclear except for the sister relationship of M. loricatus and M. squamosus. With the discovery of M. eowilsoni, the sister relationship of this new species and M. ferrugineus becomes clear. The placement of M. ruficauda remains uncertain in our analyses, because of low support values on the subtending node. It was placed, however, as sister to M. loricatus–M. squamosus in all analyses (Figure 4). Adding nuclear loci to the mitochondrial DNA dataset would likely increase confidence in our placement of M. ruficauda (Maddison 1997), but we note that plumage, behavior, and vocalizations are all consistent with the genetically inferred relationships in this study.

Taxonomy

Myrmoderus eowilsoni and M. ferrugineus share several distinctive morphological traits, including bluish bare orbital skin, gray feet and tarsi, and white leading edges to the inner bend of the wing (proximal to the wrist), and lack the white interscapular patch shown by other species of Myrmoderus. Conversely, the other members of Myrmoderus, including the type of the genus (M. loricatus), have pink legs and feet and obvious white interscapular patches, the latter of which are regularly used in visual signaling in the forest understory (D. F. Lane personal observation). Myrmoderus eowilsoni and M. ferrugineus differ behaviorally from the other Myrmoderus by walking rather than hopping as a means of locomotion over the ground (Zimmer and Isler 2003; B. Whitney personal communication). Furthermore, M. eowilsoni and M. ferrugineus exhibit strong sexual dimorphism in song, whereas the 3 other Myrmoderus do not (B. Whitney personal communication; Figure 3). Todd (1927) noted morphometrics, extensively bare orbital skin, plumage coloration, and dark feet to justify the creation of a monotypic genus, Myrmedestes, for Myrmoderus ferrugineus. However, the name Myrmedestes Todd, 1927 fell into disuse when this genus was subsumed into a broadly conceived Myrmeciza Gray, 1841 by multiple authorities (Zimmer 1932, Peters 1951, Meyer de Schauensee 1970), prior to separation of Myrmoderus by Isler et al. (2013). We consider it appropriate, given the sister relationship of Myrmoderus ferrugineus and M. eowilsoni and the presence of numerous morphological and behavioral synapomorphies, to recognize Myrmedestes as a subgenus in Myrmoderus to include these 2 species.

Morphological Diagnosability

We took standard morphological measurements from 105 specimens representing all recognized taxa in the genus Myrmoderus (Appendix Table 4). For consistency, O.J. took all measurements. Measurement protocols followed Baldwin et al. (1931; their figure numbers follow the measurements) for tail length (fig. 120), wing chord (fig. 100), tarsus (fig. 136), exposed culmen (fig. 3), bill tip to nares (fig. 8), bill width at base (fig. 13), and bill depth at base (fig. 10; Table 2).

TABLE 2.

Morphometric data from all Myrmoderus taxa (n = 105). Mean values are given, followed by 2 standard deviations, with sample sizes in parentheses.

To evaluate the morphological data, we performed a principal component analysis with the PCA function in Scikit-learn (Pedregosa et al. 2011) using all specimens for which we had complete morphological data (n = 101). The first 2 principal components explained 91.4% of the variation in our dataset and clustered the species into 2 non-overlapping groups: one containing M. loricatus and M. squamosus and the other containing M. eowilsoni, both subspecies of M. ferrugineus (M. f. ferrugineus and M. f. elutus), and both subspecies of M. ruficauda (M. r. ruficauda and M. r. soror Pinto, 1940) (Figure 5). Despite this overlap between M. ferrugineus and M. ruficauda in morphological measurements, the genetic data suggest that M. ruficauda is allied with M. loricatus and M. squamosus (Figure 4). In addition, M. ruficauda has pink legs and feet, a white interscapular patch, and hopping locomotory behavior, which all suggest a closer relationship to M. loricatus and M. squamosus.

FIGURE 5.

The first 2 principal components (PC1 and PC2) of 7 morphometric variables from all species of Myrmoderus. The loadings for PC1 and PC2 accounted for 73.5% and 17.9% of the variation in the dataset, respectively.

Vocalizations

Songs of both male and female M. eowilsoni are distinguished from those of other members of the Thamnophilidae by a combination of a clear whistled quality, few notes, and a simple pattern. Compared to songs of M. ferrugineus, those of M. eowilsoni consist of fewer notes (2–6 vs. 8–14) that are delivered more slowly (average 2 notes second⁻¹ vs. average 7 notes second⁻¹; Figure 3). The male song seems quite stereotyped, but with some minor variation, particularly in pitch. Males sing 4 whistled notes given as 2 couplets (songs of 2 or 3 notes heard rarely), usually with the first note highest in pitch, the second lower in pitch, the third of similar or slightly higher pitch, and the final note lowest in pitch (Figure 3A). Although it is distinctive to the ear, the song of male M. eowilsoni can be likened to the slowed-down terminal 2 couplets of the song of male M. ferrugineus. The song of female M. eowilsoni shows wider variation within and among individuals in pitch and number of notes (3–6) than in the male, and the female song is delivered more slowly and often with a raspier quality (Figure 3B–3E). The call (Figure 3F) is generally a sputtering series of notes similar to that of M. ferrugineus (Figure 3I), but it is delivered at a slower pace and occasionally consists of only single or doubled notes. The songs of M. ferrugineus (Figure 3G, 3H), M. loricatus (Figure 3K, 3L), and M. squamosus (Figure 3M, 3N) are similar to one another in also being composed of couplets that generally descend, but differ from that of M. eowilsoni in having more notes. By contrast, the dry, rattled song of M. ruficauda (Figure 3J) is immediately distinguishable from that of the other members of the genus. The songs of M. eowilsoni and M. ferrugineus are sexually dimorphic (i.e. female songs differ consistently from those of males in pattern, note number, and note shape), whereas the songs of the other 3 Myrmoderus are not (B. Whitney personal communication; Figure 3).

Distribution

Myrmoderus eowilsoni is known currently only from the Flor de Café ridge around the type locality, between 1,340 and 1,670 m a.s.l. (Figure 1). This distribution encompasses just 15 km² (minimum convex polygon of known localities), although extrapolating the range to include habitat from 1,300 to 1,700 m a.s.l. on the same ridge increases the distribution to ∼78 km². We believe it is very likely that this species is found on additional ridges within the Cordillera Azul, and it should be looked for farther afield, perhaps as far as the Cordillera El Sira. Multiple previous ornithological surveys in the Cordillera Azul since the 1990s did not detect M. eowilsoni, yet there are ∼2,480 km² of habitat between 1,300 and 1,700 m a.s.l. in the Cordillera Azul that may be suitable for this species, most of which are within the Parque Nacional Cordillera Azul (Figure 1). With knowledge of the elevational distribution and habitat requirements of the new species, as well as song playback, we believe that efforts to revisit previously surveyed localities within the Cordillera Azul are warranted in addition to exploring new localities.

Habitat, Ecology, and Behavior

Myrmoderus eowilsoni appears to be restricted to tall, montane, evergreen forest with intact understory characterized by small trees (diameter at breast height about 5–10 cm) and ferns, many moss-covered logs, and extensive dead leaf matter on the ground. We did not observe any individuals in second growth, edge, or tree-fall gap habitats, although in a few cases we could hear individuals singing from within 10 m of large plantations of sun-grown coffee. Most sightings occurred on or near level ridge tops, although this may have been biased by a lack of trails in steeper terrain. We did encounter a few individuals up to ∼200 m downslope of ridgeline trails. Despite conducting extensive playback (at intervals of <100 m) along trails through forested habitats around Flor de Café from 900 to 1,850 m a.s.l., we did not detect the species below 1,340 or above 1,670 m a.s.l. Both members of the pair defend their territory, and most of our observations were of pairs responding to playback of the song together, typically walking or occasionally flying toward the source of playback, and singing for as much as an hour after playback ceased. Only one female (CORBIDI 93483) had an enlarged ovary suggestive of breeding activity. Aside from the initial detection by J.R.B., we detected the species only after playback, and it seemed largely silent under natural conditions. We note, however, that we were generally not near known territories at dawn or dusk, when unprovoked singing would be most likely. Alternatively, unsolicited singing may be seasonal; other species of Myrmoderus can become quieter in the dry season, and their period of maximum singing is typically early in the morning between about 0630 and 0700 hours (B. Whitney personal communication). We observed M. eowilsoni only within 1 m of the ground, and individuals typically moved by short, low flights and by walking slowly along the ground. Walking involved deliberate steps accompanied by a jerking, dove-like head bob, with the tail held straight out behind. As with many terrestrial walking antbirds, antthrushes, and tapaculos, M. eowilsoni often hopped onto fallen logs or low branches in response to playback. We noted few instances of tail movements, mostly when an agitated individual was perched off the ground and responding to playback, when it would give occasional tail-raising flicks, bringing the tail slightly above horizontal from a position well below horizontal. Brief foraging observations consisted of birds picking through leaf litter to search for arthropod prey, and preserved stomach contents of specimens (CORBIDI 12380, 12381; LSUMZ 190882, 190883, 190884) contain arthropod parts.

When surveying for the species with playback, we located several adjacent territories along a ridgetop trail, which helped us estimate territory lengths of roughly 300–500 m. We revisited territories on consecutive days, and playback did not always elicit a response; this suggests that the pair may have been far enough away as to be out of earshot and that territories are quite large. We tentatively estimate a maximum territory size of 0.25 km², by squaring our estimate of the longest territory length.

Conservation

Myrmoderus eowilsoni is presently known only from the ridge of Flor de Café, yet little fieldwork has been conducted within the nearby Parque Nacional Cordillera Azul, one of the largest national parks in the country with an area of >13,500 km². Future surveys should target localities within this national park, which contains ∼1,940 km² of area between 1,300 and 1,700 m a.s.l. that may be appropriate habitat for M. eowilsoni (Figure 1). Habitat alteration, primarily clear-cutting, around Flor de Café is extensive and ongoing. Sun coffee plantations, the primary agricultural land use there, do not provide habitat for M. eowilsoni or other forest species. Nevertheless, we found M. eowilsoni to be fairly common in forest fragments around the type locality, and these fragments seem to maintain a remarkably intact avifauna, at least for the time being. Human colonists from other parts of Peru are rapidly clearing forest in the area and may soon eliminate much of the remaining forest around Flor de Café. Therefore, confirming the presence of M. eowilsoni in the national park is a high priority. Considering the proximity of extensive protected habitat within the national park, we are optimistic that the species will be found to be a common resident in other parts of the mountain range, and we hope that this discovery spurs further conservation and research in this unique region.

We here make a preliminary estimate of the population size of the species. Based on a territory size ranging from 0.066 km² (territory size reported for M. ferrugineus by Johnson et al. 2011) to 0.25 km² (our maximum estimate), 2 birds territory⁻¹, territories in only 45.2% of appropriate habitat (taken from mean estimate for occupancy of M. ferrugineus by Stouffer 2007), and an area of 1,940 km² between 1,300 and 1,700 m a.s.l., the predicted population size within the national park is 7,000–27,000 individuals. If we include ridges of the Cordillera Azul found outside the park (including the Flor de Café ridge) but within the latitudinal range of the park, the area between 1,300 and 1,700 m a.s.l. increases to ∼2,480 km², and the predicted population size is 9,000–34,000 individuals. These estimates of population size are crude and highly preliminary because of a large number of assumptions, but they nonetheless serve as a starting point for assessment of the actual population size. Flor de Café lies on the leeward side of the Cordillera Azul, and we suspect that the lower elevational limit of M. eowilsoni is locally maintained by drier habitats as one descends toward the rain shadow of the Huallaga valley. In more windward and humid regions of the Cordillera Azul, appropriate habitat may extend considerably lower in elevation, perhaps as low as ∼900 m. Survey work is urgently needed to determine the full distribution of M. eowilsoni, to assess territory size and occupancy, and to improve estimates of population size. Because most of the estimated distribution of M. eowilsoni lies within the Parque Nacional Cordillera Azul, we do not expect this species to face extensive habitat loss. Despite the restricted distribution and habitat loss in the immediate vicinity of the type locality, we consider an IUCN Red List status of “Data Deficient” as most appropriate until the distribution and population trends can be further evaluated.