rediscovery of a cryptic species and description of a new ... - BioOne

The Auk 119(2):362–378, 2002

REDISCOVERY OF A CRYPTIC SPECIES AND DESCRIPTION OF A NEW SUBSPECIES IN THE MYRMECIZA HEMIMELAENA COMPLEX (THAMNOPHILIDAE) OF THE NEOTROPICS MORTON L. ISLER,1,5 JOSE´ ALVAREZ ALONSO,2 PHYLLIS R. ISLER,1 THOMAS VALQUI,3 ALFREDO BEGAZO,4 AND BRET M. WHITNEY3 Division of Birds, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560, USA; 2 Instituto de Investigaciones de la Amazonıá Peruana; Av. Quinõnes km 2.5, Apdo. 784, Iquitos, Peru; 3 Museum of Natural Science, 119 Foster Hall, Louisiana State University, Baton Rouge, Louisiana 70803, USA; and 4 3920 SE 14th Terrace, Gainesville, Florida 32641, USA 1

ABSTRACT.—The recording in 1992 of an unknown antbird loudsong initiated a journey of discovery that led to the resolution of the conflict between John Zimmer (1932) and Melvin Carriker (1934) regarding the taxonomic standing of the population of Myrmeciza hemimelaena, a widespread southern Amazonian antbird, that occurs in lower Andean elevations in San Martıń, Peru. Zimmer had described a new subspecies, M. h. castanea, from the Moyobamba Valley, but Carriker, on the basis of specimens he collected at the same elevation in an adjoining valley, could not find morphological differences between his specimens and the widespread nominate form of M. hemimelaena. Both authors were correct. Two cryptic sister taxa coexist in the foothills of San Martıń. Diagnostic vocal and morphological characters and syntopy confirm their status as distinct species. Myrmeciza hemimelaena castanea Zimmer is revived and raised to species status. Concurrently, analysis of the vocalizations and morphology of the lowland population north of the Rıó Maranõń in Peru, the loudsong recording of which initiated the project, revealed that this population was closely related to M. castanea of which it is described as a subspecies, M. c. centunculorum, under the biological species concept. The discovery that there are two cryptic species in the Myrmeciza hemimelaena complex exemplifies the continuing need to pursue field knowledge and analysis of avian species limits in the Neotropics required for the development of conservation strategies as well as phylogenetic understanding. Received 17 April 2001, accepted 7 December 2001. RESUMEN.—La grabacioń de un canto fuerte (i.e., ‘‘loudsong’’) desconocido de una especie de Thamnophilidae en 1992 inicio´ un camino de descubrimientos que llevarıá a la resolucioń del conflicto entre John Zimmer (1932) y Melvin Carriker (1934) con respecto al estatus taxono´mico de la poblacioń de Myrmeciza hemimelaena (una especie de amplia distribucioń en el sur de la Amazonıá) que se encuentra a bajas elevaciones en los Andes de San Martıń, Peru´. Zimmer habıá descrito una nueva subespecie (M. h. castanea) del Valle del Moyobamba, pero Carriker no encontro´ diferencias morfolo´gicas entre especı´menes que e´l colecto´ a la misma elevacioń en un valle adyacente y la ampliamente distribuida forma nominal de M. hemimelaena. Los dos autores tenıán razoń, pues dos taxa hermanos crı´pticos coexisten en el piedemonte de San Martıń. Caracteres vocales y morfolo´gicos diagno´sticos y sus distribuciones sinto´picas confirman su estatus como dos especies distintas. Ası´, el taxon Myrmeciza hemimelaena castanea Zimmer es revivido y elevado al estatus de especie. Ana´lisis de las vocalizaciones y la morfologıá de la poblacioń de tierras bajas que se encuentra al norte del Rıó Maranõń en Peru´ (en donde se hizo la grabacioń que inicio´ este proyecto) indican que e´sta esta´ estrechamente relacionada con la especie M. castanea, por lo que bajo el concepto biolo´gico de especie, se describe como una nueva subespecie (M. c. centunculorum). El descubrimiento de que existen dos especies crı´pticas en el complejo de Myrmeciza hemimelaena ejemplifica la necesidad de continuar acumulando conocimientos en el campo y analizando los lı´mites especı´ficos entre las aves neotropicales. Esto es necesario para el desarrollo de estrategias de conservacioń ası´ como para mejorar el entendimiento de la filogenia.

RAPIDLY EXPANDING KNOWLEDGE of Neotropical birds is providing surprising evidence 5

E-mail: [email protected]

(behavioral, vocal, molecular, distributional) that a number of previously described subspecies, although resembling sibling taxa closely in appearance, should be recognized as distinct

362

April 2002]

A Cryptic Antbird Species

363

FIG. 1. Locations in which taxa in the Myrmeciza hemimelaena complex have been found in northern Peru and adjoining regions. Symbols represent locations for which there are documented records of taxa in the Myrmeciza hemimelaena complex. Open circle 5 M. h. hemimelaena; open star 5 both M. h. hemimelaena and M. c. castanea; solid diamond 5 M. c. castanea; solid square 5 M. castanea ssp. nov.; U in a circle 5 taxon uncertain. Locations mentioned in the text follow. (1) Rıó Negro (type locality of M. c. castanea) and Rıó Seco, San Martıń, Peru (;068129S, 778059W). (2) Saposoa, San Martıń, Peru (068569S, 768489W). (3) Intuto, Loreto, Peru (038329S, 748449W). (4) About 50 km by road west of Rioja, San Martıń, Peru (;058499S, 778189W). (5) Jesu´s del Monte, San Martıń, Peru (068039S, 768449W). (6) San Martıń, Loreto, Peru (038539S, 738279W; type locality of M. castanea ssp. nov). (7) Sarayacu, Pastaza, Ecuador (018449S, 778299W). (8) Kapawi Lodge, Pastaza, Ecuador (028329S, 768519W). (9) Chinapinza (Cordillera del Cońdor), Zamora-Chinchipe, Ecuador (048009S, 788349W). (10) Miazi (Cordillera del Cońdor), Zamora-Chinchipe, Ecuador (048179S, 788389W). (11) Jeberos; Loreto Peru (058179S, 768139W). (12) Huambo, Amazonas, Peru (068229S, 778289W)

species. Within suboscine avian families, examples may be found in Bierregaard et al. (1997), Isler et al. (1997), Pierpont and Fitzpatrick (1983), Whitney et al. (2000), Willis (1992), and Zimmer (1997). Such discoveries, although vital for their scientific application to studies of biogeography and phylogeny, also have important implications for defining regions of endemism and land conservation strategies (Whitney et al. 1995). The Myrmeciza hemimelaena complex provides another case in point.

J. T. Zimmer (1932) described Myrmeciza hemimelaena castanea from specimens obtained by H. Watkins from the Rıó Negro, Moyobamba Valley, San Martıń, Peru at 2,600 ft (793 m; locations cited are mapped on Figs. 1 and 2). Zimmer described the plumage of castanea as ‘‘darker than other subspecies of M. hemimelaena known at present’’ and listed a number of plumage characters that he considered diagnostic. Later, M. A. Carriker, Jr. (1934) reviewed the status of castanea after collecting specimens

364

ISLER

ET AL.

[Auk, Vol. 119

FIG. 2. Geographic ranges of taxa in the Myrmeciza hemimelaena complex and geographic distribution of sample of vocalizations analyzed. Narrowly spaced dots are cordilleran divides. Symbols represent the occurrence of taxa within small geographic sectors (see Isler 1997). Open circle 5 M. h. hemimelaena; open triangle 5 M. h. pallens; open star 5 both M. h. hemimelaena and M. c. castanea; open diamond 5 M. c. castanea; open square 5 M. castanea ssp. nov.; ‘‘U’’ in a circle 5 taxon uncertain. Asterisk 5 location for which vocalization recordings were examined visually; double asterisk 5 location for which a full set of vocal measurements were obtained; asterisk and a plus sign 5 location for which a partial set of vocal measurements were obtained. Heavy dashes outline the region detailed in Figure 1, and the following numbers of localities mentioned in the text succeed the numbered locations of that figure. (13) Mapiri, La Paz, Bolivia (158159S, 688109W; type locality of M. h. hemimelaena). (14) Borgonã, Junıń, Peru (118059S, 758209W; type locality of M. h. ‘‘spodiogastra’’). (15) Vila Bela da Santı´ssima Trindade (5Villa Bella de Mato Grosso or Mato Grosso), Mato Grosso, Brazil (158009S, 598579W; type locality of M. h. pallens). (16) Puerto Umbria, Putumayo, Colombia (008529N, 768339W). (17) Cordillera de Galeras, Napo, Ecuador (008509S, 778359W). (18) East bank Rio Madeira opposite Humaita´, Amazonas, Brazil (;078319S, 638029W). (19) Manicore´, Amazonas, Brazil (058499S, 618179W).

near Saposoa, a locality in the same department and at a similar elevation, 2,500 ft (762 m), but located in the Saposoa river valley immediately to the south (see Fig. 1). When Carriker compared the Saposoa specimens to specimens of nominate hemimelaena further south in Peru, he did not find the plumage differences described by Zimmer. Assuming (implicitly) that the Saposoa and the castanea specimens represented the same population, Carriker concluded that castanea was not a valid taxon. Meyer de Schauensee (1948) concurred. Subsequently, castanea was synonymized with nominate hemimelaena by Peters (1951), although without reference to Carriker’s paper.

In 1992, along the Rıó Tigre, a northern affluent of the Rıó Maranõń in north-central Peru, near Intuto (;125 m elevation) Alvarez tape recorded a loudsong of an unseen thamnophilid antbird that could not be identified to species by any of us. Shortly thereafter, Alvarez saw the bird, which he identified as a form of Myrmeciza hemimelaena, and he collected a specimen for the Museu de Historia Natural de la Universidad Nacional Mayor de San Marcos, Lima, Peru (MUSM). Since then, Alvarez obtained additional specimens and along with B.M.W. made over 40 vocal recordings of that population whose range was found to extend east to the Rıó Nanay and to the urbanization

April 2002]

365


limits of the city of Iquitos, and northwest to the Rıó Corrientes and (by voice) to Kapawi Lodge, Pastaza, Ecuador, (;200 m elevation) near the border with Peru. Subsequently, that population is referred to as the lowland population north of the Maranõń. In the process of comparing vocalizations of the lowland population north of the Maranõń to available recordings of Myrmeciza hemimelaena, the Islers discovered a previously unexamined recording of a loudsong that resembled recordings of the north of the Maranõń lowland population and was dissimilar from the vocalizations of nominate hemimelaena. The recording had been made by T. A. Parker III in 1977 in the Moyobamba Valley at a site 50 km (by road) west of Rioja, San Martıń, at an elevation of ;2,700 ft (822 m). Simultaneous with the recording, Parker and his colleagues collected four specimens from that site that were deposited in the Louisiana State University Museum of Natural Science (LSUMZ). We examined the three males and one female and found that two of the males and the female were a ‘‘dark form’’ of hemimelaena and appeared to agree with the original description of castanea; the third male was a juvenile. Subsequently, the Islers were given copies of recordings of an antbird that was identified visually as hemimelaena obtained by a field party (Valqui, C. Bushell, R. Innes, D. Parsons, B. Walker et al.) in 1998 in the Moyobamba Valley at Jesu´s del Monte (also known as ‘‘15 km northeast of Jirillo’’), San Martıń, at an elevation of ;1,350 m. The observers had recognized that the loudsong delivered by that individual differed from those of lowland hemimelaena with which they were familiar. That individual was netted, photographed, and identified as a ‘‘dark form’’ of hemimelaena before its release. In 1999, Valqui and Begazo returned to Jesu´s del Monte and once again found the dark form of hemimelaena. Equally significantly, they discovered a male of a ‘‘light form’’ vocalizing within hearing distance of the dark form. Unfortunately, they did not have a collecting permit, but they tape recorded vocalizations of both forms (the light form was not recorded at the site, but another individual was recorded ;3 km distant), and noted habitat and behavior of each. At last, almost 70 years later, field data were available that permitted the differences in

the conclusions of Zimmer and Carriker to be reconciled. The goal of this study was to establish appropriate taxonomic treatments of populations in the Myrmeciza hemimelaena complex from the Moyobamba Valley and from north of the Rıó Maranõń and Rıó Amazonas by comparing them with one another and with populations from the remaining geographic range of M. hemimelaena. A secondary purpose was an initial exploration of the vocal and morphological characteristics of the latter populations, setting the stage for future, more detailed study. METHODS We developed a locality based map of geographic distribution in the Myrmeciza hemimelaena complex on the basis of sites listed in museum inventories, sites referenced in the literature, and sites of vocal recordings (Fig. 2). Specimens were examined and compared at the Academy of Natural Sciences of Philadelphia (ANSP), the American Museum of Natural History (AMNH), LSUMZ, and the National Museum of Natural History, Smithsonian Institution (USNM). We obtained plumage color measurements of selected specimens with a calibrated colorimeter (CR-221 Chroma Meter, Minolta Corporation, Ramsey, New York) with a 3.0 mm aperture following the protocol of Graves (1999) which derives the mean of nine measurements in each plumage location (e.g. center of breast). Resultant opponent-color coordinates (L, a, b) describe color in terms of dark-light (L 5 0 to 100, low values are dark), red-green (a 5 1100 to 2100, more positive the redder the color, negative if green, near zero if neutral), and yellow-blue (b 5 1100 to 2100, positive if yellow, negative if blue, near zero if neutral), always in that order. Plumage color measurements of specimens of both sexes were taken at the following plumage locations: center of crown, center of nape, center of lower mantle, center of rump, upper side of rectrices, center of breast, center of belly, and posterior flanks. Additional plumage locations measured for females included the center of forehead, center of throat, and malar region. Color measurements were taken for all available specimens from the Moyobamba Valley and from north of the Rıó Maranõń (total of nine males and seven females). A full set of color measurements was initially taken from a sample (four males, six females) in the region south of the Rıó Maranõń excluding the Moyobamba Valley; that size sample was deemed appropriate given that we did not have the objective of studying plumage differences within that region. To provide more comparative data, however, an additional six males and five females from that region were measured for colors of the flanks of both sexes and the

366

ISLER

breasts and bellies of females after those body locations were found to be representative of overall plumage darkness and color intensity. English color names used in verbal descriptions were adapted from Munsellt Soil Color Charts (New Windsor, New York). Measurements of bill width, depth, and length; tarsus; tail; and wing chord were taken with MAX-CAL electronic digital calipers of specimens selected for color measurements. We reviewed 225 recordings of populations in the Myrmeciza hemimelaena complex and made a spectrogram of each type of vocalization given by every individual on each recording. Recording locations are identified in Figure 2, and the inventory of vocalizations is listed in Appendix 1. Following visual inspection of the spectrograms, we selected individuals from 30 recordings for quantitative measurement of loudsongs. They included 12 individuals from north of the Rıó Maranõń, 5 individuals from the Rıó Moyobamba Valley (four of the dark form, one of the light form), and a geographically well-distributed sample of recordings of 21 individuals from the remaining region (locations identified in Fig. 2). In addition to seeking geographic representation and despite the possibility of introducing bias, recordings used in this step were selected on the basis of recording clarity, but even with that precaution, some elements of loudsongs, especially note length, were difficult to measure with precision because of intense sound reverberation in the dense understory where members of the complex typically are found. Another problem was that recordists were often unable to see vocalizing individuals well enough to identify them by sex, and in a few cases we deduced whether the recording had been made of a male or a female. We took 25 measurements of each loudsong that reflected eight independent vocal characters, and three additional characters were established by visual inspection of spectrograms. We also measured the lengths of all raspy calls (termed ’’churrs’’). We omitted from the analysis an example of an apparently prematurely terminated call and another of an abnormally extended call given by an individual whose other calls were normal. Measurements taken, analytic procedures, and tests of significance of possible differences follow earlier papers (Isler et al. 1998, 1999). Employing results of those analyses, we determined morphological and vocal characteristics of populations in the following three regions: (1) the Moyobamba Valley in the drainage of the Rıó Mayo, San Martıń, Peru, the region of the type locality of castanea; (2) the remainder of the region south of the Rıó Maranõń and Rıó Amazonas/Solimo˜es; and (3) the region north of the Rıó Maranõń and Rıó Amazonas/Solimo˜es. We then compared characteristics from each region to other regions in order to identify diagnostic differences, if any, between them. Finally, we assessed the taxonomic status of each population,

ET AL.

[Auk, Vol. 119

employing guidelines developed in previous papers (Isler et al. 1997, 1998, 1999). Those guidelines provide support to recommendations as species or subspecies on the basis of the extent of diagnostic differences in morphology and vocalizations. As described in more detail in the earlier papers, such differences must be unambiguously distinct character states, or, in the case of continuous variables, ranges could not overlap and the means (x) and standard deviations (SD) of the population with the smaller set of measurements (a) and the population with the larger set of measurements (b) had to meet the requirement: xa 1 taSDa # xb 2 tbSDb where ti is the t-score at the 97.5 percentile of the t distribution for n 2 1 degrees of freedom (except for ratios where this statistical test is inappropriate).

RESULTS CHARACTERISTICS OF POPULATIONS Moyobamba Valley of San Martıń, Peru, and nearby locations west of the Rıó Huallaga. We compared plumage coloration and body measurements of two male and four female AMNH specimens of castanea from the Rıó Negro (type locality) and Rıó Seco with the two LSUMZ adult males and one female collected 50 km (by road) west of Rioja noted above (henceforth ‘‘Rioja specimens’’). Posterior flanks were slightly darker in the AMNH males (L 5 27.1 6 1.2, range 26.2 to 27.9) than the Rioja males (L 5 30.9 6 0.7, range 30.4 to 31.4), and the center of the belly was yellower in AMNH than Rioja males (b 5 10.8 6 1.9, range 9.6 to 12.3, and 5.3 6 1.3, range 4.4 to 6.2, respectively) and females (b 5 18.5 6 2.6, range 16.3 to 22.2, and 7.7, respectively). Otherwise, all 26 other color and body measurements obtained for the Rioja specimens fell (or in a few instances came close to falling) within the range of color and body measurements of the AMNH castanea specimens. We judged the few differences to be within the range of individual variation and concluded that the AMNH specimens of castanea and the Rioja specimens came from the same population. That result allowed us to identify the Rioja recording of a loudsong (Fig. 3A) as that of castanea, important because Rioja is the only location from which we had both specimens and a recording of the dark form. As noted above, both dark and light forms were tape recorded at Jesu´s del Monte, al-

April 2002]


367

FIG. 3. Loudsongs of Myrmeciza castanea. (A) Fifty kilometers by road west of Rioja, San Martıń, Peru (Parker MLNS 12891). (B) Jesu´s del Monte, San Martıń, Peru (Valqui ISL-TV.1:03). (C) Varillal Expetroleros, Loreto, Peru (Alvarez ISL-JA.7:12). (D) Miazi, Zamora-Chinchipe, Ecuador (Parker ISL-MLNS.143:05). Archive codes of this and other spectrogram figures explained in Appendix 1.

though only loudsong recordings of males were obtained. When we compared spectrograms (Fig. 3) and measurements of recordings of eight individuals (three measured for quantitative data) of the dark form at Jesu´s del Monte with the Rioja recording of castanea, we found no differences either visually or in measurements. Notes of loudsongs from both locations rise in frequency, although the final few notes drop off, whereas the series accelerates in pace. Regarding plumage, although no specimens were available from Jesu´s del Monte, we had a photograph of a mist-netted male of the dark form, and we could detect no difference between it and the AMNH or Rioja specimens of castanea. On the basis of vocal and morphological evidence, we concluded that the AMNH castanea specimens, the Rioja specimens and re-

cording, and the Jesu´s del Monte recordings of the dark form all represented the same population which henceforth is referred to as ’’castanea’’. When we compared the two loudsong recordings of a male of the light form obtained at Jesu´s del Monte (Fig. 4A) with recordings of castanea, we found substantial differences. The light form loudsongs were, however, indistinguishable from those of male hemimelaena (Figs. 4B and C) which are described below. Notes of those loudsongs are initially flat in frequency but then descend while their pace accelerates. Henceforth, the light form recorded in the Moyobamba Valley is referred to as ’’hemimelaena’’. Figure 1 identifies three other Peruvian locations for the Myrmeciza hemimelaena complex west of the Rıó Huallaga and south of the Rıó

368

ISLER

ET AL.

[Auk, Vol. 119

FIG. 4. Loudsongs of Myrmeciza hemimelaena. Recordings (B) and (C) were the closest available to the type localities of M. h. hemimelaena and M. h. pallens, respectively. (A) Jesu´s del Monte, San Martıń, Peru (Valqui ISL-TV.1:01). (B) Rıó Tuichi, La Paz, Bolivia (Whitney ISL-BMW.107:13 [male] and ISL-BMW.112:18 [female]). (C) Flor de Oro, Santa Cruz, Bolivia (Whitney ISL-BMW.88:23). (D) Manicore´, Amazonas, Brazil (Whitney ISL-BMW.187:20).

Maranõń. All are represented solely by specimens. Those from Saposoa, San Martıń (Carriker 1934), and Huambo, Amazonas (Berlepsch and Stolzmann 1894) were identified a light forms of hemimelaena by the authors. The specimen from Jeberos, Amazonas, was simply listed as M. hemimelaena by Sclater and Salvin (1873). We considered it as an unknown taxon pending further examination of the specimen or additional field work. Apparent differences in habitat preference between castanea and hemimelaena were found in the vicinity of Jesu´s del Monte by Valqui and Begazo. Lower elevations in the Moyobamba Valley are in a rain shadow and moderately xeric (Davis 1986, Davis and O’Neill 1986), but at higher elevations a tall cloud forest emerges

with up to 40 m canopy height in areas of good soil, but that in turn gives way to a semistunted forest of about 20–30 m canopy height on poorquality sandy soil, and finally to a forest of stunted and dispersed trees (height ;4 m) that occurs on ridges and other outcroppings. Valqui and Begazo found that semistunted forest surrounding outcroppings of stunted forest appeared to be the sole habitat of castanea, whereas they only found hemimelaena in the tallest forest. More observations are needed to confirm that possible difference in habitat preference. However, even though castanea and hemimelaena may occupy different habitats, they were in hearing distance of each other’s vocalizations, and should be considered syntopic at this locality.

April 2002]


Myrmeciza hemimelaena south of the Rıó Maranõń and Rıó Amazonas/Solimo˜es. In examining museum specimens, we were faced with the same problem as previous authors who reviewed hemimelaena (Zimmer 1932; Carriker 1934; Gyldenstolpe 1945a, b), namely the variability of darkness and color intensity of specimens obtained in different locations. Those authors characterized males from some locations as ‘‘dark’’ and as having a gray belly that was sometimes mixed with black; some as ‘‘light,’’ having a substantial portion of the center of the belly white; and some as intermediate. Female plumages were found to vary in darkness and color intensity consistent with that of males at the same localities. Dark females had the belly pale reddish-yellow-brown, whereas light females had a white belly. Berlepsch and Stolzmann (1894) described a dark form from Borgonã, Junıń, Peru, ;800 m, as Myrmeciza spodiogastra on the basis of comparisons to lighter specimens from the south in Bolivia (location unclear) and from the north at Huambo, Amazonas, Peru, a lower montane location in the Rıó Huambo valley immediately to the south of the type locality of castanea. Later, Berlepsch and Hellmayr (1905) described a light form from Villa Bella de Matto Grosso (5Vila Bela da Santı´ssima Trindade), Mato Grosso, Brazil, as M. hemimelaena pallens. When Hellmayr (in Cory and Hellmayr 1924) reviewed the species, he recognized three subspecies: M. h. spodiogaster (spodiogastra), whose range he limited to Junıń; M. h. pallens, which he placed in the eastern portion of the range in Brazil west to the Rio Jurua´ and south to the Rio Guapore´; and nominate hemimelaena, occupying the intervening area extending from Colombia through Ecuador, Peru (except Junıń), southwestern Brazil, and Bolivia. The clarity of that pattern was put into question by Zimmer (1932) who in reviewing specimens of M. hemimelaena noted that specimens from the region immediately south of the Rıó Amazonas in Peru and western Brazil were also dark and had the gray belly of spodiogastra. He also found substantial individual variation within regions. Later, Gyldenstolpe (1945a) found dark and light forms in the middle and upper Rio Jurua´ in Amazonas, Brazil, which he assigned to spodiogastra and h. hemimelaena, respectively, and he confirmed the variability of Bolivian specimens (1945b). Finally, Peters

369

(1951) synonymized spodiogastra with hemimelaena without explanation, although he maintained pallens as a subspecies. Results of our specimen examinations and color measurements of selected specimens were consistent with the confused nature of earlier findings. The lightest specimens we examined were from the southeast corner of the distribution, near the type locality of pallens. The darkest specimens came from Andean foothills in Cuzco and Junıń, Peru, near the type locality of spodiogastra, but, as noted by Zimmer, also from the geographically distant region immediately south of the Amazon between the mouths of the Rıó Urubamba and Rio Jurua´. The geographic pattern of specimens intermediate in color was unclear. For example, the darkness and intensity of belly coloration was about the same in specimens from near the mouth of the Rio Tapajo´s in Para´, Brazil, near the base of the Andes in Cochabamba, Bolivia, and along the middle Rıó Ucayali, in Ucayali, Peru. Although specimens from the west of the Rio Madeira in Brazil were generally darker than from east of the river, with the specimens in hand, we could not distinguish distinct geographically defined populations. Turning to vocalizations, visual examination of loudsong spectrograms showed a quite similar pattern throughout the range of hemimelaena, including recordings from near the type localities of hemimelaena (Fig. 4B) and pallens (Fig. 4C). One difference among regions was that male loudsongs from west of the Rio Madeira in Brazil and the Rıó Mamore´ (and its extension, the Rıó Grande; see Fig. 2) in Bolivia typically accelerated more rapidly than those east of the Madeira/Mamore´. The difference was quantified by dividing the loudsongs into three sections, obtaining the pace of each section, and then the ratios between those values (Table 1). Given the overlap in range of values of the ratios, the measure was not diagnostic but suggested genetic divergence in those populations. Such divergence was also reflected in an allozymic and mitochondrial DNA analysis of Myrmeciza hemimelaena samples obtained at eight locations in Bolivia and Peru (Bates 1993). Those included six locations in the vicinity of the Serranıá de Huanchaca, Santa Cruz, Bolivia (near the type locality of pallens) east of the Madeira/Mamore´; and two locations west of the Madeira/Mamore´, one in the vicinity of Cobija,

370

ISLER

ET AL.

[Auk, Vol. 119

TABLE 1. Selected measurements of male loudsongs of taxa in the Myrmeciza hemimelaena complex. Mean 6 SD (Range). (A) M. hemimelaena east of Madeira n59

(B) M. hemimelaena west of Madeira n 5 17

(C) M. castanea upland n59

(1) Number of notesa 9.6 6 1.1 (8–11) 8.5 6 1.1 (7–10) 8.8 6 0.7 (8–10) (2) Length second noteb 87 6 9 (75–102) 92 6 9 (83–111) 188 6 16 (165–206) (3) Length first note/ next-to-last notec 1.0 6 0.2 (0.8–1.2) 1.4 6 0.4 (0.9–2.1) 5.9 6 0.9 (4.5–7.6) (4) Pace section 1d 4.7 6 0.4 (3.7–5.1) 4.0 6 0.4 (3.4–4.8) 2.9 6 0.2 (2.6–3.1) (5) Pace section 2/ e section 3 1.0 6 0.1 (0.9–1.2) 0.8 6 0.1 (0.7–1.0) 0.5 6 0.1 (0.4–0.6) (6) Peak frequency second notef 4.1 6 0.1 (3.8–4.2) 4.1 6 0.2 (3.8–4.4) 3.5 6 0.02 (3.4–3.5) (7) Frequency first note/frequency 1.05 6 0.02 1.04 6 0.01 0.82 6 0.03 middle noteg (1.03–1.09) (1.02–1.06) (0.77–0.86)

(D) M. castanea lowland n55 6.4 6 0.5 (6–7) 205 6 8 (197–213) 9.0 6 1.9 (6.7–11.6) 2.1 6 0.2 (1.7–2.4) 0.4 6 0.04 (0.3–0.4) 3.9 6 0.1 (3.7–4.0) 0.77 6 0.07 (0.65–0.85)

No overlap of (C) with (D) but difference not significant. In ms. No overlap of (A) and (B) with (C) and (D) and differences significant. No overlap of (A) and (B) with (C) and (D). Significance test not applicable. d In notes per second. No overlap of (A) and (B) with (C) and (D) and differences significant. No overlap between (C) and (D) but differences not significant. e No overlap of (A) and (B) with (C) and (D). Significance test not applicable. f In kilohertz. No overlap of (C) with (A), (B), and (D) and differences significant. g No overlap of (A) and (B) with (C) and (D). Significance test not applicable. a

b c

Pando, Bolivia and one 65 km east-northeast of Pucallpa, Ucayali, Peru. Bates found that genetic distance between the Huanchaca samples and those from Cobija and Peru was high, at a level similar to that distancing closely related species of other Neotropical birds. Loudsong recordings from the same sites exhibited the differences in acceleration described above. The foregoing comparisons exclude two recordings of male loudsongs from near the Rio Madeira’s east bank that are unique for hemimelaena in having the frequency of the initial notes lower than those that follow, so that note frequencies (measured at the peak) rise and fall as seen in spectrograms (Fig. 4D). Given that the east bank recordings were obtained at Manicore´ (by B.M.W.) and east of Humaita´ (by M. Cohn-Haft), both in Amazonas, Brazil, but ;275 km apart, they could reflect the presence of a population distinct from other populations both east and west of the Madeira/Mamore´, but the sample size is small, and consideration of differences is beyond the scope of this article. Female loudsongs were not analyzed quantitatively, but visually they were longer overall, slower in pace, and had longer and more upward inflected notes than male songs (Fig. 4). Note shapes of female songs are somewhat variable, and difference between Figure 4B and C

represent individual, not geographic, variation. For example, female loudsongs with note shapes similar to those of Figure 4B were recorded at the same location as the female loudsong depicted in Figure 4C. Two calls were recorded for hemimelaena, and examples are provided from east and west of the Rio Madeira. One was short, downward inflected, and frequency modulated (Fig. 5A and B) that we have termed a ’’churr’’ (corresponding to the ’’chirr’’ of Willis 1985). The other was an abrupt note spanning a large frequency range that we have termed a ’’chak’’ (Fig. 5E and 5F). The latter was repeated two (corresponding to the ’’dit-it’’ of Willis 1985) or three times or strung into a series resembling a rattle; spacing between the notes was variable and not stereotyped. North of Rıó Maranõń and Amazonas. Prior to the last decade, Myrmeciza hemimelaena was known from only two published specimens in this region: Puerto Umbrıá, Putumayo, Colombia (ANSP) and Sarayacu, Pastaza, Ecuador (Natural History Museum [BMNH]). Additional unpublished specimens had been deposited in the Los Angeles County Museum (LACM), collected in 1934 at Cordillera de Galeras (Cerro Galeras), Napo, Ecuador, and the Pontifica Universidad Cato´lica del Ecuador Museo

April 2002]


371

FIG. 5. Calls of taxa in the Myrmeciza hemimelaena complex. (A) hemimelaena west of Madeira: ’’churr’’; Explorer’s Inn, Madre de Dios, Peru (van den Berg MLNS 29015). (B) hemimelaena east of Madeira: ’’churr’’; Caima, Para´, Brazil (Whitney ISL-BMW.161:02). (C) castanea lowland: ’’churr’’; Varillal Expetroleros, Loreto, Peru (Alvarez ISL-JA.7:12). (D) castanea upland: ’’churr’’; Miazi, Zamora-Chinchipe, Ecuador (Parker MLNS 80067). (E) hemimelaena west of Madeira: ’’chak’’; Parque Nacional Madidi, La Paz, Bolivia (Whitney ISLBMW.150:40). (F) hemimelaena east of Madeira: ’’chak’’ Ruro´polis, Para´, Brazil (M. Isler MLNS 47876). (G) castanea lowland: ’’chak’’; Varillal Expetroleros, Loreto, Peru (Alvarez ISL-JA.7:12).

(QCAZ), collected at Sarayacu in 1963. As noted above, Alvarez began to collect and tape record individuals of a population in that complex in the early 1990s in the region northwest of Iquitos, Loreto, Peru, and consequently found the form at a number of localities in the drainages of the Rıó Tigre and Rıó Corrientes (localities available on request). Additional individuals were found in the same decade in Ecuador at Miazi (048179S, 788389W; Parker tape recordings and Schulenberg and Awbrey 1997) and Chinapinza (048009S, 788349W; ANSP), both in the Cordillera del Cońdor, ZamoraChinchipe, and at Kapawi Lodge, Pastaza (028329S, 768519W; Whitney tape recordings). Characteristics of specimens and vocalizations from north of the Rıó Maranõń and Amazonas are discussed below in comparison to castanea. COMPARISONS OF CHARACTERISTICS OF POPULATIONS Comparison of castanea in the Moyobamba Valley and hemimelaena. In his description of castanea, Zimmer (1932) identified six diagnostic morphological differences between females of castanea and other populations of hemimelaena.

Of those, three were supported by our analysis. First, regarding bill coloration of females, both the mandible and maxillary of castanea are black; in hemimelaena the mandible is gray, contrasting with the black maxillary. Second, in both sexes, spots on the tips of the inner remiges of castanea are slight, vague and indistinct, whereas they are clearly demarcated in all populations of hemimelaena. Third, in females of castanea the dark breast contrasts sharply with the whitish center of the belly (mean of ratio of the darkness [L value] of belly divided by breast 5 1.6; range 1.4–1.8; n 5 5), whereas in hemimelaena the contrast between the breast and belly is reduced because populations of hemimelaena with whitish bellies have correspondingly pale breasts and populations with dark breasts had pale reddish yellow brown rather than whitish bellies (mean of ratio of L value of belly divided by breast 5 1.3; range 1.1–1.4; n 5 11). Of the three characters, female bill color appears to be the most useful. One specimen (AMNH 177448 from Saposoa, Peru) of hemimelaena had somewhat indistinct spots on the inner remiges approaching those of castanea, but that individual had a leaden blue (color noted by collector) lower mandible and low level of contrast be-

372

ISLER

ET AL.

[Auk, Vol. 119

TABLE 2. Means of color measurements of specimens of Myrmeciza catanea taken at central points of selected plumage regions. Measurements in L, a, b notation (see text). Puerto Umbria, Colombia Male Crown Mantle Breast Belly Flanks Female Crown Mantle Throat Breast Belly Flanks

16.18 21.64 11.03 74.79 39.54

n51 1 2.12 1 1 5.08 1 1 1.90 1 2 0.23 1 1 5.62 1

2.48 10.98 1.83 9.31 21.16

NW of Iquitos, Peru n53 1 0.90 1 1.93 1 5.37 1 14.17 1 0.90 1 1.00 1 0.18 1 4.72 1 6.50 1 24.18 n52 23.10 1 0.60 1 4.09 24.44 1 1.76 1 8.39 60.44 1 9.80 1 39.58 48.18 1 13.04 1 43.43 75.49 1 2.97 1 24.32 34.92 1 5.09 1 19.09 18.29 23.57 12.21 80.54 43.05

tween breast and belly color; another individual from the same location had distinct wing spots. Differences between loudsong vocalizations of male castanea and hemimelaena were more clearly diagnosable than plumage differences. First, the frequency of initial notes of the castanea loudsong ascended whereas those of hemimelaena were flat or descended (except in the two loudsong recordings from the east bank of the Rio Madeira described above). The difference could be readily seen in visual comparison of spectrograms (Figs. 3 and 4) and quantified (Table 1). Second, the pace of the castanea loudsong accelerated at a substantially faster rate than that of hemimelaena (Table 1). Third, length of the initial two notes of castanea were diagnosably longer than those of hemimelaena (Table 1). Fourth, notes of castanea diminish substantially in length throughout the loudsong, whereas those of hemimelaena shortened imperceptibly, and the difference, as reflected in the ratio between the length of the first and next-to-last notes, was diagnostic (Table 1). Comparison of populations north of the Rıó Maranõń and Amazonas with castanea in the Moyobamba Valley. Plumages of male specimens from north of the Rıó Maranõń and Rıó Amazonas were consistent with castanea from the Moyobamba Valley in having poorly defined pale terminal spots on the inner remiges, but they varied substantially in color intensity. A male from Chinapinza, Cordillera del Cońdor, Ecuador, was the darkest; its color measurements, except for the rump, were darker than in

Chinapinza, Ecuador 13.94 18.28 9.81 50.14 26.26

n51 1 0.96 1 1 4.94 1 1 0.81 1 1 0.41 1 1 4.59 1

1.25 9.08 0.48 2.87 15.21

Moyobamba Valley, Peru n54 1 1.18 1 5.13 1 3.90 1 10.54 1 1.52 1 1.69 1 0.37 1 8.04 1 8.98 1 21.14 n55 21.08 1 2.73 1 7.87 20.19 1 5.47 1 12.60 49.72 1 13.03 1 38.76 43.53 1 15.58 1 39.34 69.81 1 2.34 1 19.09 33.29 1 6.15 1 19.67 18.06 20.65 12.86 70.49 28.99

castanea from the Moyobamba Valley (Table 2) and its belly lacked the broad white patch found in Moyobamba males. A male from Puerto Umbrıá, Colombia, and males from the lowlands north of the Maranõń in Peru were similar in color intensity to the Moyobamba males except they were paler on the belly and especially the flanks (Table 2). Female specimens from north of the Rıó Maranõń were available only from lowlands in Peru and were consistent with females of castanea in having black lower mandibles, vague terminal spots on the inner remiges, and strong contrast in darkness of the breast and belly (mean of ratio of L value of belly divided by breast 5 1.6; range 1.5–1.6; n 5 2). Female specimens from north of the Maranõń lowlands in Peru were paler than Moyobamba specimens, and diagnostic differences are described in Appendix 2. Loudsongs recorded in the lowlands of Peru north of the Rıó Maranõń (Fig. 3C) and in southern Ecuador resembled loudsongs of castanea of the Moyobamba Valley (Fig. 3A and B), although with some differences that are described in Appendix 2. Loudsongs of males from Miazi, Cordillera del Cońdor, Ecuador, 900 m elevation (Fig. 3D), could not be distinguished from Moyobamba Valley loudsongs recorded at similar elevations. Female loudsongs, available only for lowlands north of the Maranõń and Cońdor sites, had the same pattern of frequency shift as males and similar note shape but tended to consist of more notes and longer notes (Fig. 3); sample sizes were too small for quantitative comparison. A similar long

April 2002]


’’churr’’ was recorded at both the lowland and in the Cońdor sites (Fig. 5C and D). Comparison of populations north of the Rıó Maranõń and Amazonas and hemimelaena. Specimens and male loudsongs obtained from north of the Rıó Maranõń differed from hemimelaena in the same characters as did castanea of the Moyobamba Valley. Additionally, female loudsongs and calls, unavailable for the Moyobamba Valley, were available for the northern populations. On the basis of those, it appeared, pending larger samples, that male and female loudsongs of northern populations were similar in note shape, whereas male and female loudsongs of hemimelaena differed in note shape (compare male with female songs in Figs. 3 and 4). Also, the raspy ’’churr’’ differed significantly in length between 20 individuals of hemimelaena (0.22 s 6 0.04; range 0.16 to 0.28) and 14 individuals of northern populations (0.54 s 6 0.10; range 0.41 to 0.79); a difference that was obvious in spectrograms (Fig. 5). The shorter ’’chak’’ call, however, could not be distinguished between populations. DISCUSSION

AND

RECOMMENDATIONS

Consideration of species limits. In addition to the finding of syntopy in the vicinity of Jesu´s del Monte in the Moyobamba Valley of Peru, the analysis identified both vocal and morphological characters that distinguished Myrmeciza castanea from M. hemimelaena. Under guidelines developed in earlier papers (Isler et al. 1997, 1998, 1999), castanea differed sufficiently from hemimelaena to be considered specifically distinct under the Biological and Phylogenetic Species Concepts. We therefore recommend that Myrmeciza castanea Zimmer be revived and raised to the species level. We suggest English names of Northern Chestnut-tailed Antbird for M. castanea and Southern Chestnut-tailed Antbird for M. hemimelaena, reflecting their distributions. Geographically, Myrmeciza castanea ranges from extreme southern Colombia in Putumayo south through eastern Ecuador to lowlands of northeastern Peru west of the Rıó Napo and north of the Rıó Maranõń and Rıó Amazonas, extending south of the Rıó Maranõń at elevations of 750–1,050 m to the Moyobamba Valley of San Martıń. Recordings of vocalizations and specimens of females are unavailable north of

373

extreme southern Ecuador, and more northern records are assigned to castanea primarily on geographic grounds, although the indistinct spots on the inner remiges of a male specimen from Colombia are consistent with that determination. The known records of castanea indicate a ‘‘patchy’’ distribution apparently reflecting confinement to patchy habitats growing over very poor clay or sandy soils. The geographic range of M. hemimelaena is confined to south of the Rıó Maranõń and Rio Amazonas along the eastern base of the Andes to Santa Cruz, Bolivia, (ascending to ;1,300 m in Peru and Bolivia), and in the lowlands eastward to the drainage of the Rio Xingu in Para´, Brazil, the drainage of the Rio Teles Pires in northern Mato Grosso, Brazil, and the headwaters of the Rio Guapore´ in the southwestern corner of Mato Grosso, Brazil. Consideration of subspecies. With regard to Myrmeciza castanea, we found vocal and morphological differences between the upland population in the Moyobamba Valley and the Cordillera del Cońdor and the lowland population in the region northwest of Iquitos, Peru. Although, following the guidelines cited earlier and given current information, we could not accept those differences as reflective of specifically distinct populations under the Biological Species Concept, the two populations differ sufficiently to describe the lowland population as a new subspecies. The subspecies description including the diagnosis of vocal and morphological differences between the two populations is presented in Appendix 2. Those populations would be considered distinct species under the phylogenetic species concept. Although it was not a goal of the study to establish differences among populations of Myrmeciza hemimelaena, we were aware of such differences in describing characters that differentiate M. hemimelaena from M. castanea. Regarding morphology, we were unable to improve on the geographic confusion in plumage intensity described by earlier authors. Comparing recordings from east and west of the Rio Madeira and its central Bolivian tributaries (Rıó Mamore´, Rıó Beni), we did not find diagnosable vocal characters, although one character overlapped only slightly, and two recordings from along the east bank of the Madeira were suggestive of a possibly distinct, geographically intermediate population. The

374

ISLER

strongest evidence supporting the possibility that M. hemimelaena populations are specifically distinct came from Bates (1993) who found allozymic and mitochondrial DNA differentiation between populations east and west of the Rio Madeira. Pending further analysis, we recommend that Peters’ (1951) treatment be maintained and that populations east of the Rio Madeira in Brazil and its central Bolivian tributaries (not the Rio Guapore´ as described by Peters) be considered as M. h. pallens with the nominate form of M. hemimelaena extending west of the range of pallens to the eastern slope of the Andes. Conservation. The extremely limited known range of the nominate form of Myrmeciza castanea lies within a region that is under intense human pressure for cultivation of coca, coffee, and other crops. Even if the poor quality, sandy soil habitat of castanea is not destroyed for agriculture, the increasing human occupation of the region and the consequent gathering of firewood puts heavy pressure on any woody habitat. Another endemic, Herpsilochmus parkeri, that appears to be restricted to San Martıń has been considered ‘‘Endangered, (a taxon) in danger of extinction and whose survival is unlikely if the causal factors continue operating’’ (Collar et al. 1992). Herpsilochmus parkeri was found by Valqui and Begazo in San Martıń on stunted ridge tops as well as in the surrounding semistunted forest where castanea was recorded, and the same factors that threaten H. parkeri in that region would affect Myrmeciza c. castanea. There is hope, however, that the governments of Ecuador and Peru will reach and sustain agreement on the preservation of the forests of the Cordillera del Cońdor (Schulenberg and Awbrey 1997) which could provide an important refuge for M. c. castanea. The extremely few records of (presumed) M. castanea in Ecuador and Colombia suggest that this species distribution in those countries is even more constrained than it is in northern Peru. The lowland population of Myrmeciza castanea, described in Appendix 2, is only known with certainty from the north of the Maranõń region of northern Peru and adjacent Ecuador where it occurs in patches of poor clay or sandy soil forests known as ‘‘irapayal’’ and ‘‘varillal’’ (see Whitney and Alvarez 1998). Human activities threaten those habitats, and urgent measures are needed to save them as sanctuaries

ET AL.

[Auk, Vol. 119

for a number of specialized avian inhabitants that they house. In March 1999, the Peruvian government took an important step in preserving the rare white sand habitats of northern Amazonian Peru by establishing the Zona Reservada Allpahuayo-Mishana of 57,600 ha. Despite the formal creation of that reserve, a burgeoning human population continues to threaten the habitats of M. castanea, and additional financial resources are needed to support measures to enforce the sanctity of the refuge (Alvarez and Whitney 2001). In summary, both highland and lowland forms of M. castanea have limited ranges in patchy habitats that are threatened by human activity to an extent that both populations are candidates for endangered status. Future work. More field work and analysis is needed to complete our understanding of the Myrmeciza hemimelaena complex. Further genetic analysis seems vital in order to clarify geographic variation found between and within M. h. hemimelaena and M. h. pallens, and it would be useful to accompany molecular studies with detailed measurement of vocalizations and the assembly and review of a much larger series of specimens than employed in this study. With regard to M. castanea, the rarity of records north of Peru needs to be explained, and more natural history information and vocal recordings are needed in general, including female songs and calls from near the type locality in the Moyobamba valley. ACKNOWLEDGMENTS This paper is dedicated to the memory of Theodore A. Parker III who left behind vocal recordings, specimens, and volumes of field data that remain an extraordinarily valuable scientific resource. In the present study, Ted’s recordings and specimens of M. castanea obtained in San Martıń, Peru, almost 25 years ago were the key to establishing relationships between specimens and vocalizations necessary to the analysis. We are grateful to J. V. Remsen and S. Cardiff, Louisiana State University Museum of Natural Science, for extended loan of specimens while the puzzle of the San Martıń populations of ’’hemimelaena’’ was being unraveled. L. Joseph and N. Rice kindly afforded access to the collection at Academy of Natural Sciences of Philadelphia and provided timely loans of specimens. M. Isler’s work at American Museum of Natural History was supported by a Chapman Collections Study Grant and the assistance of G. Barrowclough and P. Sweet. Alvarez is

April 2002]


grateful to Fauna and Flora International (Cambridge, United Kingdom), Bergstrom Awards (Association of Field Ornithologists), and Idea Wild (USA) for financial and material support for field work; to the Direccioń General de Areas Naturales Protegidas y Fauna Silvestre (INRENA) of Peru for study and collection permits; and to I. Franke of the Museo de Historia Natural in Lima for especially helpful assistance. T.V. and A.B. thank M. and L. Sokol for help provided toward their San Martıń expedition. We are indebted to C. Bushell who provided the Islers with the first recording of castanea from Jesu´s del Monte and D. Parsons who contributed photographs of that individual. J. Freile provided data from the Pontifica Universidad Cato´lica del Ecuador Museo collection. S. Olson helped with nomenclature. J. Bates, M. Braun, G. Graves, J. V. Remsen, T. Schulenberg, and K. Zimmer afforded useful discussions and encouragement on the project to us over the years. B. Loiselle, E. Willis, and J. Bates reviewed the final submission and provided a number of helpful comments. We continue to be indebted to the curators and collections managers at the museums cited in earlier papers who provide vital geographic range data and other information for specimens in their care. Finally, we are most deeply appreciative of all those recordists whose contributions were essential to this paper and whose names are acknowledged in Appendix 1 and to the Macauley Library of Natural Sounds, Cornell Laboratory of Ornithology, the archive for a number of these recordings. LITERATURE CITED ALVAREZ, A. J., AND B. M. WHITNEY. 2001. A new Zimmerius tyrannulet (Aves: Tyrannidae) from white-sand forests of northern Amazonian Peru Wilson Bulletin 113:1–9. BATES, J. M. 1993. The genetic effects of forest fragmentation on Amazonian forest birds. Ph.D. dissertation, Louisiana State University, Baton Rouge. BERLEPSCH, H. VON, AND C. E. HELLMAYR. 1905. Studien u¨ber wenig bakannte Typen neotopischer Vo¨gel. Journal fu¨r Ornithologie 53:1–33. BERLEPSCH, H. VON, AND J. STOLZMANN. 1894. Descriptions of quelques espe`ces nouvelles d’oiseaux du Pe´rou central. Ibis 6:385–405. BIERREGAARD, R. O., JR., M. COHN-HAFT, AND D. F. STOTZ. 1997. Cryptic biodiversity: An overlooked species and new subspecies of antbird (Aves: Formicariidae) with a revision of Cercomacra tyrannina in northeastern South America. Ornithological Monographs 48:111–128. CARRIKER, M. A., JR. 1934. Descriptions of new birds from Peru, with notes on the nomenclature and status of other little-known species. Proceedings

375

of Academy of Natural Sciences of Philadelphia 86:317–334. COLLAR, N. J., L. P GONZAGA, N. KRABBE, A. MA˜ N IETO, L. J. N ARANJO , T. A. P ARKER III, DRON AND D. C. W EGE . 1992. Threatened Birds of the Americas. ICBP/IUCN Red Data Book, 3rd ed., part 2. Smithsonian Institution Press, Washington D.C. CORY, C. B., AND C. E. HELLMAYR. 1924. Catalogue of birds of the Americas. Pteroptochidae–Conopophagidae–Formicariidae. Field Museum of Natural History (Zoological Series) 13, Pt. 3:1– 369. DAVIS , T. J. 1986. Distribution and natural history of some birds from the Departments of San Martıń and Amazonas, northern Peru. Condor 88:50–56. DAVIS , T. J., AND J. P. O’NEILL. 1986. A new species of antwren (Formicariidae: Herpsilochmus) from Peru, with comments on the systematics of other members of the genus. Wilson Bulletin 98:337– 352. GRAVES, G. R. 1999. Diagnoses of hybrid hummingbirds (Aves: Trochilidae). 8. A provisional hypothesis for the hybrid origin of Zodalia glyceria (Gould, 1858). Proceedings of the Biological Society of Washington 112:491–502. GYLDENSTOLPE, N. 1945a. The bird fauna of Rio Jurua´ in western Brazil. Kungliga Svenska Vetenskapsakademiens Handlingar 22:1–338. GYLDENSTOLPE, N. 1945b. A contribution to the ornithology of northern Bolivia. Kungliga Svenska Vetenskapsakademiens Handlingar 23:1–300. ISLER, M. L. 1997. A sector-based ornithological geographic information system for the Neotropics. Ornithological Monographs 48:345–354. ISLER, M. L., P. R. ISLER, AND B. M. WHITNEY. 1997. Biogeography and systematics of the Thamnophilus punctatus (Thamnophilidae) complex. Ornithological Monographs 48:355–381. ISLER, M. L., P. R. ISLER, AND B. M. WHITNEY. 1998. Use of vocalizations to establish species limits in antbirds (Passeriformes; Thamnophilidae). Auk 115:577–590. ISLER, M. L., P. R. ISLER, AND B. M. WHITNEY. 1999. Species limits in antbirds (Passeriformes; Thamnophilidae): The Myrmotherula surinamensis complex. Auk 116:83–96. MEYER DE SCHAUENSEE, R. 1948. The birds of the Republic of Colombia, part 3. Caldasia 5:645–871. PETERS, J. L. 1951. Check-list of Birds of the World, vol. 7. Museum of Comparative Zoology, Cambridge, Massachusetts. PIERPONT, N., AND J. W. FITZPATRICK. 1983. Specific status and behavior of Cymbilaimus sanctaemariae, the Bamboo Antshrike, from southwestern Amazonia. Auk 100:645–652. SCHULENBERG, T. S., AND K. AWBREY. 1997. The Cordillera del Cońdor region of Ecuador and Peru: A biological assessment. Rapid Assessment Pro-

376

ISLER

gram Working Paper no. 7, Conservation International, Washington, D.C. SCLATER, P. L., AND O. SALVIN. 1873. On the birds of eastern Peru. Proceedings of the Zoological Society of London 43:252–311. WHITNEY, B. M., AND J. ALVAREZ A. 1998. A new Herpsilochmus antwren (Aves: Thamnophilidae) from northern Amazonian Peru and adjacent Ecuador: The role of edaphic heterogeneity of terra firme forest. Auk 115:559–576. WHITNEY, B. M., J. F. PACHECO, D. R. C. BUZZETTI, AND R. PARRINI . 2000. Systematic revision and biogeography of the Herpsilochmus pileatus complex, with description of a new species from northeastern Brazil. Auk 117:869–891. WHITNEY, B. M., J. F. PACHECO, AND R. PARRINI. 1995. Two species of Neopelma in southeastern Brazil and diversification within the Neopelma/Tyranneutes complex: Implications of the subspecies concept for conservation (Passeriformes: Tyrannidae). Ararajuba 3:43–53. WILLIS, E. O. 1985. Myrmeciza and related antbirds (Aves, Formicariidae) as army ant followers. Revista Brasiliera de Zoologia 2:433–442. WILLIS, E. O. 1992. Three Chamaeza antthrushes in eastern Brazil (Formicariidae). Condor 94:110– 116. ZIMMER, J. T. 1932. Studies of Peruvian birds. VI. The formicarian genera Myrmoborus and Myrmeciza in Peru. American Museum Novitates 545:1–24. ZIMMER, K. J. 1997. Species limits in Cranioleuca vulpina. Ornithological Monographs 48:849–864. Associate Editor: B. A. Loiselle APPENDIX 1. Inventory of vocalization recordings examined. The following list identifies recordings used in the study by taxon, country, state or department, and recording location. Recordist names aid in identifying the recording as well as providing acknowledgments. Numbers following the recordist name identify the number of cuts per recordist per location. Acronym for recording archive: MLNS 5 Macauley Library of Natural Sounds, Cornell Laboratory of Ornithology, Ithaca New York ISL 5 recordings not yet archived in an institutional collection but that have been copied into the inventory maintained by the Islers. Myrmeciza c. castanea. 15 Recordings. ECUADOR: Zamora-Chinchipe: Miazi (Parker 7 MLNS). PERU: San Martıń: Jesu´s del Monte (Begazo 4 ISL; Bushell 1 ISL; Valqui 2 ISL); Rioja, 50 km by road west of (Parker 1 MLNS). Myrmeciza castanea ssp. nov. 42 recordings. ECUADOR: Pastaza: Kapawi Lodge (Krabbe 2 ISL; Whitney 8 ISL). PERU: Loreto: Allpahuayo (Alvarez 1 ISL); Corrientes, mouth of Rıó (Alvarez 1 ISL); Corrientes, upper Rıó (Alvarez 2 ISL); El Dorado (Alvarez 6 ISL; Whitney 10 ISL); El Milagro (Al-

ET AL.

[Auk, Vol. 119

varez 1 ISL); Varillal Expetroleros (Alvarez 3 ISL); Fundo Piura (Alvarez 1 ISL; Whitney 2 ISL); Tigre, middle Rıó (Alvarez 2 ISL); Yuto (Alvarez 3 ISL); Myrmeciza h. hemimelaena. 108 recordings. BOLIVIA: Cochabamba: Sajta (Mayer 1 ISL); La Paz: Alto Madidi (Parker 1 ISL); Araona Palmasola, Campamento (Tello 1 MLNS); Caranavi (Whitney 1 ISL); 33–48 km north of Caranavi (Parker 2 MLNS; Whitney 1 ISL); Madidi, P. N. (Whitney 13 ISL). Pando: Camino Mucden (Parker 2 MLNS). Santa Cruz: Amboro, P. N. (Parker 2 ISL); Buena Vista (Lane 2 ISL). BRAZIL: Acre: Branco, Rio (Whitney 1 ISL); Catuaba, Reserva (Whitney 4 ISL); Chico Mendes, P. A. (Whitney 2 ISL); Humaita´, Reserva (Whitney 1 ISL); Mancio Lima (Whitney 1 ISL); Minas, Riozinho das (Whitney 1 ISL); Paratari, Rio (Whitney 1 ISL); Seringal Ocidente (Whittaker 2 ISL); Serra do Divisor, P. N. (Whitney 3 ISL). Amazonas: Cruzeiro do Sul/Santa Ba´rbara road, kilometers 38–45 (Whitney 1 ISL); Palmari, R. N. (Whitney 2 ISL); Tefe´ (Pacheco 2 ISL). PERU: Cuzco: Amazonia Lodge (Arvin 1 ISL; Behrstock 2 ISL; Innes 3 ISL; Whitney 1 ISL). Loreto: Cayari, Rıó (Whitney, ISL); Contamana, 79 km NW (Lane 4 ISL); Jesu´s del Monte (Begazo 1 ISL; Valqui 1 ISL); Juanjui, about 86–90 km SE (Lane 2 ISL; Valqui 2 ISL); Yarapa Reserve (Michael 1 ISL). Madre de Dios: Explorer’s Inn (Clayton 1 ISL; Fjeldsa˚ 1 ISL; M. Isler 2 MLNS; P. Isler 1 MLNS; Jakobsen 1 ISL; Kibler 1 MLNS; Palmer 1 MLNS; Parker 11 MLNS; van den Berg 1 MLNS; Zimmer 5 ISL); Manu Wildlife Centre (Innes 1 ISL). Pasco: Villa Rica (Widdowson 3 ISL). Puno: Ta´vara, Cerros del (Parker 1 ISL); Topo Dos (Schulenberg 1 ISL); Topos Tres (Schulenberg 2 ISL). Ucayali: Abujao (Meyer 1 ISL); Balta (O’Neill 1 MLNS); Tahuayo, Cerro (Meyer 1 ISL); Tingo Maria (Coffey 1 ISL; Parker 1 MLNS); Yarinacocha (Coffey 4 ISL). Myrmeciza hemimelaena ssp.? E bank Rio Madeira. Two recordings. BRAZIL: Amazonas: Humaita´, E bank opposite (Cohn-Haft 1 ISL); Manicore´ (Whitney 1 ISL). Myrmeciza h. pallens. 58 recordings. BOLIVIA: Santa Cruz: Flor de Oro (Parker 1 ISL; Whitney 4 ISL); Huanchaca, Serranıá de (Bates 1 ISL); Los Fierros (Whitney 3 ISL); Perseverancia (Fisher 2 ISL; Parker 1 ISL). BRAZIL: Mato Grosso: Alta Floresta (Cohn-Haft 1 ISL; M. Isler 1 MLNS; P. Isler 1 MLNS; Parker 1 ISL; Whitney 4 ISL; Zimmer 11 ISL); Cristalino, Rio (Whitney 1 ISL). Para´: Altamira, 15 km south (Whitney 2 ISL); Altamira, 25 km east-southeast (Whitney 1 ISL); Aveiro (Whitney 5 ISL); Boim, 36 km west (Whitney 1 ISL); Caima (Whitney 1 ISL); Itapoama (Whitney 2 ISL); Kayapo´, R. I. (Whitney 1 ISL); Porto do Meio (Whitney 4 ISL); Ruro´polis (M. Isler 1 MLNS); Saõ Benedito, mouth of Rio (Whitney 2 ISL). Rondoˆnia: Nazare´, L bank Cachoeira (Schulenberg 1 MLNS); Pacaa´s Novos, Serra dos (Whittaker 5 ISL).

April 2002] APPENDIX 2. NEW

A Cryptic Antbird Species TAXON DESCRIPTION .

Myrmeciza castanea centunculorum subspecies novum Holotype. Museo de Historia Natural de la Universidad Nacional Mayor de San Marcos, Lima, Peru (hereafter ‘‘MUSM’’), no. 21109, adult female from San Martıń, on the Rıó Nanay, Departamento de Loreto, Peru; 038549S, 738289W; ;120 m elevation; collected 8 August 1999 by Jose´ Alvarez Alonso; voice tape recorded previous day of same individual as judged by location. Voice specimen to be archived at the Macauley Library of Natural Sounds (MLNS), Cornell Laboratory of Ornithology, Ithaca, New York; currently in Isler inventory as JA.7:16. Diagnosis. Plumage.—A medium-sized, sexually dichromatic typical antbird (Thamnophilidae) most similar to Myrmeciza castanea. Females of the new subspecies differ from females of nominate castanea from the Moyobamba Valley by their paler and less intense reddish-yellow brown coloration (Table 2), although differences between populations do not meet our statistical test given the small sample size of centunculorum. Males of centunculorum can be distinguished from males of nominate castanea of the Moyobamba Valley by a more extensively white belly (dimensions about 45 3 20 mm in centunculorum; about 30 3 15 mm in c. castanea); paler posterior flanks and crissum (Table 2); and differences in crown feathers (blackish centers with paler gray edges in centunculorum, more uniform gray in c. castanea). Diagnosis. Voice.—Comparisons were limited to loudsongs of males because female loudsongs and calls of nominate castanea were unavailable. One vocal character differed diagnostically between populations. Loudsongs of centunculorum were higher in pitch as exemplified by frequencies of second notes (Table 1). Other vocal characters may be found to be diagnostic with larger samples. Loudsongs of centunculorum had fewer notes (Table 1), and were slower paced in the initial section (Table 1), but accelerated more rapidly in the final two-thirds of the song (Table 1), than those of c. castanea. Although ranges of the latter three measures did not overlap, differences were not significant statistically given small sample sizes. Finally, with finer recordings than were

377

available, we suspect a difference may be found between the two populations in the pattern of final loudsong notes, with those of centunculorum shifting in frequency and amplitude more abruptly as suggested by a comparison of spectrograms (Fig. 3). Distribution. Known only from the drainages of the Rıó Nanay and the Rıó Tigre (including the Rıó Corrientes), Loreto, Peru, and from the Kapawi Ecotourism Lodge (multiple tape recordings but no specimens) on the Rıó Pastaza, Pastaza, Ecuador (Fig. 1). Subspecific identification of specimens further north in Ecuador to Putumayo, Colombia, remains to be determined. Description of holotype. Forehead, crown, and nape gray tinged olive with forehead feathers blackish along rachis and at edges (25.3 10.6 13.2), crown feathers with broad blackish band along rachis (22.6 10.5 14.3), and nape slightly more olive (25.5 11.1 14.7). Mantle dark-yellowish-brown (25.4 11.3 18.1) with feathers dusky in midsection and white at the base forming a white interscapular patch extending to the scapulars and ;25 mm long. Rump strong-reddish-yellow-brown (24.0 17.3 116.5) becoming lighter on uppertail coverts (30.0 17.5 120.0). Upper side of rectrices darkreddish-brown (21.1 19.2 111.9). Gray of the crown extends down to lores and to ;2.5 mm below eye socket. Outer vanes of remiges brown with paler brown margins (25.2 16.6 113.2 measured on folded wing, 22.4 15.7 112.6 measured on center of inner remiges), inner remiges with indistinct light-reddish-yellow-brown tips. Wing coverts dark brown (14.6 12.0 13.2) with pale-buff spots (;1.7 mm deep and 2.5 mm wide) on greater and median secondary coverts; spots smaller (;1 3 2 mm) and whiter on lesser coverts. Primary coverts and alular quills with thin white outer margins. Underwing coverts brownish gray with broad white edges. Chin and center of throat lightreddish-yellow-brown (61.0 17.8 136.4). Malar darker and redder (45.0 115.2 138.4). Breast reddish-yellow-brown (48.2 113.5 144.7) becoming tinged with olive on the sides (36.5 17.2 128.0). Center of belly white faintly tinged yellowish-brown (77.3 11.4 119.4) in a patch ;16 mm wide by 45 mm long separated from breast color by a buffy-yellowish-brown (69.2 11.7 125.9) transition. Flanks brown becoming strong-reddish-yellow-brown (34.4

378

ISLER

15.7 120.4) posteriorly. Undertail-coverts reddish-yellow-brown (40.3 18.3 130.0). Underside of rectrices dull-reddish-yellow-brown (27.0 16.1 112.8). Tibial feathers buffy-brown. Soft parts (original tag notation in Spanish followed by translation): Iris ‘‘marroń’’ (brown). Bill ‘‘negro’’ (black). Tarsi, feet, and soles ‘‘gris claro rosaćeo’’ (light rosaceous gray). Body mass 14.6 g. Skull ‘‘100% ossif.’’ Tissue saved in buffer (#9-8-98) at Field Museum of Natural History. Measurements of dried specimen: wing (unflattened chord) 52 mm; tail 34 mm; tarsus 23 mm; bill from anterior edge of nares to tip 9.5 mm; bill depth at anterior edge nares 4.1 mm; bill width at anterior edge nares 3.9 mm. Description of allotype. MUSM no. 21110, adult male from Allpahuayo, on the Carretera Iquitos-Nauta, Departamento de Loreto, Peru; 038589S, 738259W; ;120 m elevation; collected 1 July 1997 by Jose´ Alvarez Alonso. Forehead, crown, and nape black; forecrown and crown narrowly (;1 mm) edged gray (19.2 11.7 11.9); nape tinged brownish (19.9 11.6 16.2). Mantle dark-reddish-yellow-brown (23.9 14.1 113.6) with feathers blackish in midsection and white at the base forming a white interscapular patch extending to the scapulars and ;25 mm long. Rump dark-reddish-yellowbrown (25.0 17.8 117.5) becoming slightly paler and brighter on uppertail coverts (27.7 110.7 122.7). Upper side of rectrices yellowish-redbrown (25.6 110.8 115.6). Black of the crown extends down to lores and to ;2.0 mm below eye socket. Outer vanes of remiges brown with light-reddish-yellow-brown margins (25.8 15.8 114.4 measured on folded wing, 22.6 15.0 110.3 measured on center of inner remiges), inner remiges with thin (,0.5 mm) indistinct light-reddish-yellow-brown tips. Wing coverts black with dark-reddish-yellow-brown margins and pale-reddish-yellow-brown spots (;2.5 mm deep and 5.0 mm wide) on greater and median secondary coverts; spots smaller

ET AL.

[Auk, Vol. 119

(;1.5 by 3 mm) and whiter on lesser coverts. Primary coverts and alular quills with thin white outer margins. Underwing coverts dark gray with broad white edges. Chin, throat, and breast black (9.5 11.1 11.0). Outer feathers of sides of breast edged gray, similar to gray edges of crown feathers. Center of belly white in a patch ;20 mm wide by 45 mm long. Flanks yellowish-brown (42.0 12.1 111.4) becoming more reddish posteriorly (41.2 17.2 127.4). Undertail-coverts light-reddish-yellow-brown (46.4 16.8 130.4). Underside of rectrices dullreddish-yellow-brown (30.1 17.1 114.5). Tibial feathers light-brownish-gray. Soft parts (original tag notation in Spanish followed by translation): Iris ‘‘marroń’’ (brown). Bill ‘‘negro’’ (black). Tarsi, feet, and soles ‘‘rosa claro’’ (light rose). Body mass 15.1 g, length 11.1 cm. Skull ‘‘100% ossif.’’ Measurements of dried specimen: wing (unflattened chord) 53 mm; tail 34 mm; tarsus 23 mm; bill from anterior edge of nares to tip 10.4 mm; bill depth at anterior edge nares 4.0 mm; bill width at anterior edge nares 3.7 mm. Plumage variation. Two additional males and one additional female were collected by Alvarez near the type locality. The female (MUSM 21108, skull ‘‘100% ossif.’’), in comparison to the type specimen, has crown feathers without darker coloration along the rachis although they are dark basally, center of belly more yellowish, edges of remiges paler, and outer rectrices with indistinct light-reddishyellow-brown tips. No obvious differences were found between the allotype and the two additional male specimens (MUSM 21106 and 21107). Body mass (all specimens, male and female combined): 14.7 6 0.8 (range 13.4 to 15.4; n 5 5). Etymology. The subspecific name is taken from the male diminutive centunculus and the suffix orum, meaning ‘‘of the little patches,’’ reflecting the patchy distribution of habitats in which the new subspecies is found.