lov7tattes/AMERICANt MUSEUM PUBLISHED BY THE AMERICAN MUSEUM 0OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NIEW YORK, N.Y. 10024 Number 2777, pp. 1-17, figs. 1-3, tables 1-6 January 30, 1984

Patchy Distribution and Systematics of Oreomanes fraseri (Aves, ?Coerebidae) of Andean Woodlands

FRANqOIS VUILLEUMIER1

ABSTRACT The Giant Conebill, Oreomanesfraseri Sclater, Range disjunctions and geographical isolation of 1860, is an endemic, monotypic ofcoerebid populations do not appear conducive to incipient or thraupid-like distributed in the high speciation. The pattems of geographical distri- of South America from southern Colombia to bution and variation suggest either: (a) that the southwestern Bolivia. This well-marked taxon ap- origin of the disjunctions is recent (i.e., there has pears to be restricted ecologically to "islands" of not been enough time for the isolates to become woodland vegetation above timberline that are morphologically and presumably also genetically dominated by trees of the endemic Andean genus differentiated); (b) the geographical isolation has Polylepis (). The geographical distribu- been insufficient to promote differentiation (i.e., tion of Oreomanes fraseri is quite patchy. Oreo- gene flow exists between isolated populations and manesfraseri shows geographical variation in bill counters the disrupting effects ofgeographical iso- size and plumage color. Geographical variation, lation); or (c) geographical isolation is effective and however, is very minor, and so it does not seem is relatively old in origin but the rate of morpho- useful to describe it in terms of . The logical differentiation in the isolates is slow. Data pattern of this variation is partly clinal and partly are lacking at present to permit one to check which checkerboard, as one would expect from the nature of these hypotheses is the most likely to explain of the discontinuous distribution of the species. the observed patterns. RESUMEN El Ptajaro de los queniuales, Oreomanes fraseri suroeste de Bolivia. Esta especie parece restringida Sclater, 1860, es un ge'nero endemico y monotipico ecol6gicamente a islas de vegetaci6n lefnosa enci- de ave relacionado a los Coerebidae o a los Th- ma del limite superior del bosque andino nublado, raupidae, distribuido en los altos Andes de Ame- dominadas por arboles del genero andino ende- rica del Sur, desde el sur de Colombia hasta el mico Polylepis (Rosaceae). La distribuci6n geo-

ICurator, Department of Ornithology, American Museum of Natural History.

T4V4zTi C((-rf5l9) / PTric V.9 05 AMERICAN MUSEUM NOVITATES NO. 2777

grafica discontinua de Oreomanes esta relacionada geografica sugieren tres hipotesis explicativas: (a) con la distribucion insular de los bosques de Poly- el origen de las discontinuidades de distribuci6n lepis. Oreomanesfraseri tiene variaci6n geografica es muy reciente y el tiempo no ha sido en suficiente algunos caracteres morfologicos externos, es- para una mayor diferenciaci6n morfologica (y ge- pecialmente el tamafio del pico y el color de su netica?) de las poblaciones el plumaje. Esta aisladas; (b) aisla- variaci6n es irregular, como se hu- miento geogrifico no es suficiente para resultar en biera podido predecir segun la distribucion dis- variaci6n marcada, pues el intercambio continua de genes de las poblaciones de la especie. Sin em- entre poblaciones aparentemente aisladas se opo- bargo, esta variabilidad geografica es menor y no ne a los efectos del aislamiento parcial; (c) la di- parece biol6gicamente significativa describirla con se ha una ferenciaci6n morfol6gica producido muy nomenclatura subespecifica. Afunque distri- despacio, auinque podria haber sido muy antiguo buida en poblaciones aisladas, el patr6n de dife- el origen de las discontinuidades distribucionales. renciaci6n en Oreomanes fraseri no parece apro- No tenemos datos todavia piado a para averiguar cual se- especiaci6n. Tal distribuci6n y variaci6n ria la mas probable de esas tres hipotesis.

INTRODUCTION The Giant Conebill, Oreomanes fraseri dean in distribution, are quite uncertain. For Sclater, 1860, is a remarkable bird ofthe high convenience I shall Andes of South America. It retain in this paper the is about the size term Coerebidae to include them all. of a large Sitta (Sittidae), and it Two also interesting problems are posed by roughly has the color, aspect and branch Oreomanes. The first is of the or tree-climbing behavior (H.-W. Koepcke, taxonomic af- 1961, p. 176; finities of this genus. Its describer, Sclater George, 1964; Pearson and (1860, p. believed Pearson Ralph, 1978, p. 29; personal ob- 76), Oreomanes to be "more nearly allied to Diglossa ... than to serv.). Bull (personal commun.) was struck other by the resemblance in color any genus of Coerebidae." A hundred and general be- years later, havior between Oreomanes and Sitta casta- George (1964, p. 26) suggested nea. Oreomanes is again "a near relationship" between the two dark slate gray above and genera. chestnut brown below, with a conspicuous Specifically, George (1964) stated that white cheek patch. Without this Oreomanesfraseri was perhaps close to Di- white cheek glossa carbonaria taxon patch, the color pattern ofOreomanes resem- (a belonging to a se- bles that of ries of allopatric and parapatric semispecies ferrugineiventre, a and allospecies. See high Andean coerebid that is occasionally in Vuilleumier, 1 969a, found sympatrically press; Graves, 1982). Storer (1970, p. with Oreomanes (Parker remarked that "This 38) and O'Neill, 1980; Schulenberg, in press). The nuthatch-like genus may family be related to Diglossa and to the emberizine Coerebidae is now considered by some finches. Its workers to be a polyphyletic assemblage of systematic position remains to be unrelated genera. determined" (he placed it between Xenodac- Paynter and Storer (1970, nis and p. ix) placed Xenodacnis, Oreomanes, and Diglossa in the Thraupinae of the Diglossa in the subfamily Thraupinae of the Emberizidae). However, bill shape in Oreo- family Emberizidae, and kept Conirostrum manes is unlike that in any species of Di- out. Lowery and Monroe (1968, p. 3) con- glossa but more like that in species of Con- sidered Conirostrum (and Coereba) to be gen- irostrum, whereas the color pattern of era incertae sedis, which they Oreomanes resembles as much that of some placed "at the of end of the family" Parulidae. Paynter and species Diglossa (e.g., D. sittoides) as that Storer (1970, p. ix) stated that "Oreomanes of some species of Conirostrum (e.g., C. fer- may be close to Diglossa or to Conirostrum rugineiventre). Clearly, the affinities of Or- or both." Recently, Bock (in press) argued on eomanes are in need of further study. Ana- anatomical grounds that Diglossa may be a tomical evidence and biochemical data (from polyphyletic genus. Thus the interrelation.- electrophoresis ofproteins and from DNA x ships of the genera DNA hybridization) would be Xenodacnis, Oreomanes, ful. especially use- Conirostrum, and Diglossa, which are all An- The recent description of a hybrid spec- imen 0. fraseri x Conirostrum ferrugineiv- I1984 VUILLEUMIER: OREOMANES FRASERI entre (Schulenberg, in press) adds a substantial bird's patchy distribution. If the distribution element strongly supporting the hypothesis of Oreomanes is island-like, one might expect of close taxonomic relatedness of these two this species to be made up of geographically genera. and ecologically isolated populations varying The second problem stems from the fact from one another either in checkerboard that the geographical distribution of Oreo- fashion, thus showing a population structure manes fraseri is patchy (figs. 1 and 2), and (sensu Mayr, 1959, p. 294) conducive to spe- seems to coincide ecologically with the dis- ciation, or in clinal fashion, thus showing the tribution of its preferred habitat, the rather influence of environmental conditions along open woodlands dominated by trees of the the Andes, such as moisture gradients. genus Polylepis (Rosaceae). These woodlands are ofexceptional interest in Andean ecology and biogeography because they usually occur ACKNOWLEDGMENTS as small islands ofarboreal vegetation above timberline in the open, scrubby or grassy ex- I am grateful to the following people for panses of the alpine-like paramo or puna their help while I was studying specimens in vegetation formations (Simpson, 1979, in collections in their care, and who sent me press). All specimen labels ofOreomanes that specimens on loan, or provided information include an indication of habitat mention about specimens or other matters: Dr. Ernst either "Polylepis" or "queflual" (the Quechua Sutter, Naturhistorisches Museum, Basel name for these woodlands), and all authors (detailed information about the specimens who have written about Oreomanes have collected by the late Dr. W. Markl, formerly mentioned this preference (e.g., M. Koepcke, ofPiura, ); Dr. Raymond A. Paynter, Jr., 1970, p. 111; H.-W. Koepcke, 1961, p. 176; Museum of Comparative Zoology, Harvard; George, 1964, p. 26; Vuilleumier, 1969b, p. Mr. James Bond and Mr. Rodolphe Meyer 606; Parker and O'Neill, 1980, p. 173). Fur- de Schauensee, Academy of Natural Sci- thermore, H.-W. Koepcke ( 1961, p. 176) no- ences, Philadelphia; Dr. Kenneth C. Parkes, ticed that the color pattern ofOreomanes cor- Carnegie Museum of Natural History, Pitts- responds very closely ("hervorragend burgh; Mr. Ian C. G. Galbraith and Dr. Da- uibereinstimmend" are his words!) to the gray vid Snow, British Museum (Natural History), and reddish color of the flaky bark of Polv- Tring; Dr. Jean Dorst, Museum National lepis trees. George (1964, p. 26) went so far d'Histoire Naturelle, Paris; Dr. William as to state that "the bird is so well camou- George, Southern Illinois University, Car- flaged that it ranks high among species that bondale; Dr. John W. Hardy, Florida State provide support for the theory of protective Museum, Gainesville (formerly at Moore coloration in ." Whether he is right Collection, Occidental College); Dr. Richard or not, the striking resemblance between bird L. Zusi and Dr. Paul Slud, Smithsonian In- and bark color can easily be gauged by look- stitution, Washington, D.C.; the late Dr. ing at Arthur Singer's attractive color plate George H. Lowery, Dr. John P. O'Neill, Dr. illustrating George's article. James Van Remsen and Mr. Thomas S. In the present paper I do not discuss the Schulenberg, Louisiana State University Mu- systematic position of Oreomanes within the seum ofZoology, Baton Rouge; Mr. John Bull, Coerebidae, the Thraupidae, the Emberizi- American Museum of Natural History, New dae, or the nine-primaried oscines, leaving York; Dr. Stephen Hilty, Tucson, Arizona; this problem to anatomists or biochemists Dr. Ram6on Ferreyra, the late Dr. Maria (see also Schulenberg, in press). Instead, I Koepcke and Dr. Hernando de Macedo, "Ja- investigate two interrelated aspects ofthe sec- vier Prado" Museum, Lima; Mr. Manuel ond problem mentioned above. First, I will Plenge, Lima; and Dr. G. Diesselhorst, Mu- verify whether the patchy distribution of nich Museum. This study, as part of a re- Oreomanes corresponds to the patchy distri- search program on speciation and zoogeog- bution of Polylepis woodlands. Second, if raphy of Andean , was financially there is geographical variation in Oreomanes, supported by grants from the F. M. Chapman I will check whether it is correlated with the Fund and the Sanford Fund of the American 4 AMERICAN MUSEUM NOVITATES NO. 2777

TABLE 1 Localities of Oreomanes fraseri (see fig. 1) Number of Specimens Exam- ined in Re- 1rhis References or Specimensa Locality ported Sttudy [specimens examined in brackets] 1. "Pasto," Narifio, Colombia (no altitude 2 2 Meyerde Schauensee (1951, p. 958) [BM given) 85.4.1.80; USNM 90465] 2a. Papallacta, Napo, Ecuador (no altitude 1 [Moore Collection Occidental College Ec- given) L 320] 2b. Papallacta Road, Napo, about 3700 m. - John Bull (personal commun.); sight record 3. Cerro Huamani arriba, Pichincha, 2 2 Chapman (1926, p. 641) [MCZ 233512; Ecuador (no altitude given) AMNH 180788] 4. Guamani Pass, Pichincha, Ecuador 1 [BM 1916.8.24.87] (16,000 ft. = 4850 m.) 5. Cerro Puntas, Pichincha, Ecuador (no 1 2 [PM CG 1931 No. 830, PM CG 1931 No. altitude given) 831] 6. Chaupi, Cotopaxi/Pichincha, foot of Mts. - Salvadori and Festa (1899, p. 12) Iliniza and Coraz6n, Ecuador (3200 m.) 7. Panza, s slope of Mt. Chimborazo, 2 - Sclater (1860, p. 75) Ecuador (14,000 ft. = 4200 m.) (type locality offraseri) 8. Mt. Chimborazo, Ecuador (14,000 ft. = 2 - Taczanowski and von Berlepsch (1885, p. 76) 4200 m.) 9. Sical, eastern Ecuador (no altitude given) 1 Sclater (1886, p. 13); locality "not located" (Paynter and Traylor, 1977, p. 117) [BM 85.4.1.79] 10. Mocha Canon, Tungurahua, Ecuador 2 2 Chapman (1926, p. 641) [AMNH 145898, (1 1,000 ft. = 3300 m.) AMNH 145899] 11. El Parotillo, S. Jose, Loja, Ecuador (no 1 [Moore Collection Ec-R 263] altitude given) 12. Laguna Querecocha, Ancash, Peru (4200 1 [Basel Museum, Markl No. 166] m.) 13. Llanganuco, Yungay, Ancash, Peru (4000 2 2 [Basel Museum, Markl No. 864; Museo m.) "Javier Prado" Lima No. 4116] 14. Yanac, Ancash, Peru (13,000-1 5,000 9 7 See Carriker (1933, pp. 9-10 and 28-29 about ft. = 3900-4550 m.); Bosque Quipis Yanac); LSUMZ 82300, LSUMZ 82301 Munte, above (south) YAnac, ca. 13,000 ft. [ANSP 109336, ANSP 109337, ANSP (=3900 m.); Quebrado Tutapac, 13,200 ft. 109338, ANSP 109339, LSUMZ 80873, (=4000 m.) LSUMZ 80874, MCZ 179385] 15. Above Huariaca, km. 335, Carretera 2 2 [LSUMZ 75364, LSUMZ 79125] Central, Cerro de Pasco, Peru (no altitude given) 16. Cerro Pumaj, upper valley of Rio Santa [Museo "Javier Prado" Lima No. 3233] Eulalia, Lima, Peru (4100 m.) 17. Ca. 13 km. W Milloc, Lima, Peru 8 4 See M. Koepcke (1970, p. 111) for description (12,600-13,200 ft. = 3800-4000 m.) of range in Department of Lima; LSUMZ 79987, LSUMZ 97929, LSUMZ 106753, LSUMZ 106940 [LSUMZ 79126, LSUMZ 79127, LSUMZ 79128, LSUMZ 79630] 1 984 VUILLEUMIER: OREOMANES FRASERI 5

TABLE 1-(Continued) Number of Specimens Exam- ined in Re- This References or Specimensa Locality ported Study [specimens examined in brackets]

18. Lachocc, Huancavelica, Peru (13,000 ft. = 3 3 Morrison (1939, p. 486) [BM 1946.49.757, 3900 m.) BM 1946.49.758, BM 1946.49.759]

19. Pampa Galeras, Ayacucho, Peru (about - - Brokaw (1976, p. 29), Venero G. and Brokaw 4000 m.) (1980); sight records 20. Cedrobamba Ruins, near , 1 Chapman (1919, p. 331) [USNM 273004] Cuzco, Peru (12,000 ft. = 3600 m.) (type locality of binghami)

21. Anta, Cuzco, Peru (3500 m.) 1 - von Berlepsch (1900-1901, p. 197)

22. 14 km. NE Abra Malaga on 2 2 Parker and O'Neill (1980, p. 173) [LSUMZ Ollantaitambo-Quillabamba Road, Cuzco, 79130, LSUMZ 79131] Peru (13,000 ft. = 3900 m.) 23. 6 km e Nunioa, via road to Macusani, 2 2 George (1964, p. 26) [W. George No. 1630, Puno, Peru (13,000 ft. = 3900 m.) AMNH 786678] 24. Ollachea, 20 mi. (=32 km.) W (N?) of 3 3 Hellmayr (1912, p. 159) [Munich Museum Macusani, Carabaya, Puno, Peru (11,500 11.251, Munich Museum 11.252, Munich ft. = 3500 m.) Museum 11.253] 25. Rochanga (road from Chuquibamba to 1 1 [Museo "Javier Prado" No. 4115] Pampacolca), Arequipa, Peru (4000 m.)

26. 20 km. e Chiguata, Arequipa, Peru (12,800 I 1 [LSUMZ 79129] ft. = 3900 m.)

27. 20 km. ne Tarata, Tacna, Peru (no altitude I 1 George (1964, p. 26) [W. George No. 1993] given)

28. 6 km. ne Tarata, Tacna, Peru (3900 m.) - - Pearson and Pearson Ralph (1978, p. 29); sight record 29. 10 mi. (=16 km.) n of Viloca [=Viloco], 1 1 Bond and Meyer de Schauensee (1942, p. 370); La Paz, Bolivia (11,500-13,500 ft. = altitudes of "Viloco" given as 13,800- 3500-4100 m.) 17,000 ft. in Paynter, Traylor, and Winter (1975, p. 68) [ANSP 134212] 30. 43 km. from Cochabamba City along road 1 1 Vuilleumier (1969b, p. 606) [AMNH 793132] to Morochata, Cordillera Tunari, Prov. Quillacollo, Cochabamba, Bolivia (3630 m.) 31. Pocona, Cochabamba, Bolivia (2700 m.) I Vuilleumier (1969b, p. 606) [CM 120242] 32a.Finca Salo, Oploca, Potosi, Bolivia (1 1,500 2 1 Bond and Meyer de Schauensee (1939, p. 1) ft. = 3500 m.) [ANSP 13421 1] 32b.Mina Isca-Isca, above La Torre, about 25 - Vuilleumier (1969b, p. 606); sight record km. NNE Tupiza, Sud Chichas, Potosi, Bolivia (3900 m.) Totals 60 47 a Museum abbreviations: AMNH, American Museum of Natural History; ANSP, Academy of Natural Sciences of Philadelphia; BM, British Museum (Natural History); CM, Carnegie Museum ofNatural History; LSUMZ, Louisiana State University Museum of Zoology; MCZ, Museum of Comparative Zoology; PM, Museum National d'Histoire Naturelle, Paris; USNM, United States National Museum, Smithsonian Institution. 6 AMERICAN MUSEUM NOVITATES NO. 2777

Museum of Natural History, the National cording to Bond and Meyer de Schauensee Science Foundation (Grant nos. G19729 and the taxon sturninus corresponds to the pop- GB3 167 to Harvard University's Committee ulations from Bolivia, ofwhich they saw sev- on Evolutionary Biology), and the National eral specimens from the Departments of La Geographic Society (Grant no. 1443). The Paz, Cochabamba, and Potosi. manuscript was read critically by Mr. John It appeared at one time that the wide rang- Bull and typed by Mrs. Mae Lackner and Mr. ing 0.fraseri was geographically variable (two James Drobnick. The final illustrations were subspecies, sturninus and nominate fraseri) prepared by Mr. Juan C. Barberis of the and that the localized 0. binghami occurred Graphic Art Department of the American in Cuzco, perhaps sympatrically withfraseri. Museum of Natural History. The problem of whether one or two species exist in the genus Oreomanes was solved by MATERIALS AND METHODS Hellmayr (1 919, pp. 11-1 2) who showed that the type specimen of 0. binghami is an im- The present study is based on 47 of the 60 mature bird referable to 0. fraseri. Chapman ;pecimens of the species known to me (table (1921, p. 1 15) agreed with [), and deposited in various museums in the this view and rel- United States, England, West Germany, egated his own taxon to the synonymy of 0. France, Switzerland, and Peru. The speci- fraseri. Zimmer (1942, p. 16), however, mens are cited (table 1) by museum number thought that "There is reason to believe that or collector number, or by references to the Chapman's proposed name should be re- literature cited. I have had personal field ex- vived for Peruvian examples of Oreomanes perience with Oreomanes in the Andes ofBo- fraseri." On the basis of his examination of livia (see Vuilleumier, 1969b, p. 606), and I 10 specimens (three from Ecuador, four from of in Peru, and three from Bolivia) and his study have visited stands Polylepis Venezuela, of the literature (Hellmayr, 1912, 1919; and Colombia, Ecuador, Peru, Bolivia, and Hellmayr in Morrison, 1939), Zimmer con- northwestern Argentina. cluded that 0. fraseri showed geographical variation in back color, crown color, white THE GENUS OREOMANES line above the rufous superciliary stripe, and Oreomanesfraseri was described by Sclater bill length. He recognized three subspecies: (1860, p. 75) on the basis of two specimens, 0. f fraseri (Ecuador; and presumably also one prepared as a skin and one in spirits, southern Colombia), 0. f binghami (Peru), obtained at 14,000 ft. (=4200 m. locality 7 and 0.f sturninus (Bolivia). Bond and Meyer oftable 1) on Mt. Chimborazo, Ecuador. Ad- de Schauensee (1939) also mentioned geo- ditional specimens collected in Ecuador and graphical variation in several color characters in Peru at the end of the nineteenth century, and in bill length. Finally, Storer (1970) rec- and during the first years of the twentieth ognized the three subspecies admitted by century, were referred to the same species. In Zimmer (1942). 1919 Chapman described a second species, Chapman (1919, 1921) seems to be the only 0. binghami, on the basis of one specimen author who had noticed at an early stage in from the Cedrobamba ruins near Machu Pic- the study of Oreomanes the problem of dis- chu, Cuzco, southern Peru (locality 20, table continuity in its range. "Such gaps in distri- 1, fig. 1). In that paper Chapman mentioned bution indicate lack ofcontinuity in the Tem- that the record of 0. binghami represented a perate Zone itself, and when the species southward range extension of the genus of formed in these Temperate Zone islands show approximately a thousand miles (about 1600 no appreciable differentiation, it seems evi- km.) but in fact one specimen of 0. fraseri dent that they have become isolated at a com- had already been collected near Cuzco in 1899 paratively recent date" (Chapman, 1921, p. at Anta (locality 21, table 1, fig. 1, von Ber- 115). Chapman did not know, apparently, lepsch, 1900-1901). The third and last taxon that 0. fraseri was ecologically specialized to be described in Oreomanes was the sub- and restricted in its distribution to islands of species 0. fraseri sturninus (Bond and Meyer Polylepis woodland. de Schauensee, 1939; see also 1942). Ac- The distributional records of 0. fraseri 1984 VUILLEUMIER: OREOMANES FRASERI 7

TABLE 2 The "Populations" of Oreomanes fraseri (see fig. 2) No. Popu- Altitudes Specimens lation Localities Region (m.) ("sample") A I Nariiio, Colombia 2 B 2a, 2b, 3, Eastern Andes, Ecuador 3200-4850 6 4, 5, 6 C 7, 8, 10 Mt. Chimborazo, Ecuador 3300-4200 2 D 11 Loja, Ecuador E 12, 13, 14 Ancash, Peru 3900-4550 10 F 15, 16, 17 Cerro de Pasco and Lima, Peru 3800-4100 7 G 18 Huancavelica, Peru 3900 3 H 19 Ayacucho, Peru about 4000 I 20, 21, 22 Cuzco, Peru 3500-3900 3 23, 24 Puno, Peru 3500-3900 5 K 25, 26 Arequipa, Peru 3900-4000 2 L 27, 28 Tacna, Peru 3900 1 M 29 La Paz, Bolivia 3500-4100 1 N 30, 31 Cochabamba, Bolivia 2700-3630 2 0 32a, 32b Potosi, Bolivia 3500-3900 1 Total included in sample 46

known to me are set out in table 1 and figure Andes is the monotypic coerebid Xenodacnis 1, and the "populations" in table 2 and figure parina, which interestingly also lives only in 2. These illustrations provide an updated view Polylepis woodlands (often associated with of the geographical patchiness in the distri- , Compositae) in Ecuador (see, e.g., bution of 0. fraseri, and of some of the gaps Ridgely, 1980, pp. 247-248) and in Peru (see, that might be correlated with present-day e.g., Short and Morony, 1969, p. 115). (All Andean barriers. or most other species of birds living in Polylepis woodlands seem to inhabit one or whether GEOGRAPHICAL ECOLOGY more other vegetation types, para- mo, puna, dry montane woodlands, or cloud OF OREOMANES forests; Vuilleumier, in press.) Oreomanesfraseri occurs along the Andes Polylepis woodlands are a very distinctive from southern Colombia (Pasto) to Ecuador, ecological feature of the high Andes. The ge- Peru, and Bolivia as far south as the De- nus Polylepis (family Rosaceae) occurs from partment of Potosi (figs. 1 and 2). All locality Venezuela (Merida) to northern Chile and records accompanied by a description of the northwestern Argentina (Tucuman) (Bitter, habitat stated that the species was collected 191 1; Herzog, 1923; Hueck, 1966; Simpson, in Polvlepis woodlands or quefiual scrub, as 1979, in press). Trees of the genus Polvlepis already mentioned in the Introduction. One form distinct stands or patches above tim- may therefore assume, insofar as Oreomanes berline right within the various grassland and has not been collected or observed outside scrub associations of the paramo and puna these woodlands, that this genus is indeed (e.g., Troll, 1959, p. 31). restricted to them. It is one of the few in- Because potential habitat for Oreomanes, stances in birds of a well-marked taxon at the that is, islands of woodland made up mostly genus level found only in a specific kind of of Polylepis trees, occurs in all Andean coun- vegetation type. Another example in the high tries from Venezuela to Argentina, it is sur- NO. 2777 8 AMERICAN MUSEUM NOVITATES

ARGENTINA 80 75 65

FIG. 1. Localities of Oreomanesfraseri. Black dots = localities from which specimens were examined; ircles = literature records or unpublished specimens not examined personally; triangles = sight records. se table 1 for description of localities. 1 984 VUILLEUMIER: OREOMANES FRASERI

COLOMBIA

FIG. 2. The "populations" of Oreomanesfraseri. The populations are identified in table 2. 10 AMERICAN MUSEUM NOVITATES NO. 2777

TABLE 3 Sexual Variation in Mensural Characters and Weight in Oreomanes fraseri Mean Size of Character (n = sample size)a Wing Tail Tarsus Culmen Bill Length Length Length Length Depth Weight Population Sex (mm.) (mm.) (mm.) (mm.) (mm.) (g) B (Eastern Andes, M 86.5 (1) 62.1 (2) 22.3 (2) 12.1 (2) 6.4 (2) Ecuador) F 85.5 (2) 59.3 (3) 22.0 (4) 13.1 (4) 6.2 (3) - E (Ancash, Peru) M 89.0 (2) 60.0 (2) 23.0 (1) 13.0 (1) 5.8 (2) 25.1 (2) F 88.6 (8) 62.2 (8) 22.9 (7) 13.4 (5) 6.4 (6) 24.5 (1) F (Pasco and Lima. M 92.2 (3) 66.0 (3) - - 5.5 (2) 23.75 (3) Peru) F 89.8 (3) 64.0 (2) - - 5.7 (3) 25.0 (2) All available birds M - - - 24.8 (9) (weights only) F - - -_ 24.7 (4) a Measurements were taken as follows: Wing length = chord; Tail length = from base of innermost two rectrices to tip oftail; Tarsus length = diagonally from articulation to last scute prior to digits; Culmen length = from anterior edge of nostril to tip of bill; Bill depth = at anterior edge of nostril. prising that Oreomanes should not have been Chiles and Volcan Cumbal certainly have suf- discovered farther north than Pasto in south- ficient elevation and fairly extensive Polyle- ern Colombia or farther south than Potosi in pis woodland, and they are remote enough to southern Bolivia. As far as is known at pres- have escaped extensive habitat alteration. I ent the genus Oreomanes is absent from Ven- think this is where the birds likely were taken." ezuela, Colombia except the extreme south According to Hilty (in litt.), no recent records (Narinio), Chile, and Argentina. In view of exist for Colombia. the rather thorough ornithological work car- In the remainder of the Andes, especially ried out in the Andes in recent years, it seems in Ecuador, Peru, and Bolivia, the distribu- unlikely that a bird as striking as Oreomanes tion of Polylepis woodlands is patchy, as is would have been overlooked there. One may the distribution of Oreonmanes. Large gaps in therefore surmise that its absence in Vene- distribution, illustrated on figure 2, exist be- zuela, most of Colombia, Chile, and Argen- tween some of the 34 localities and some of tina is genuine and not due to the artifact of the 15 "populations" where Oreomanes has "collecting gaps," although to be sure many been collected or observed. Examples ofsuch Polylepis woodlands remain to be explored gaps are the area between the southernmost ornithologically in all four countries. Oreo- Ecuadorian record in Loja (sample D) and manes may be absent from some areas with the northernmost Peruvian ones in Ancash apparently suitable Polylepis habitat. For ex- (E) (Northern Peruvian Low); and the region ample, I visited extensive stands of Polylepis between populations N (Cochabamba) and 0 in western Bolivia near Volcan Sajama in (Potosi) in Bolivia. In both gap areas Polv- 1968 but failed to observe Oreomanes there, lepis occurs (personal observ.) but Oreo- although I was constantly on the lookout for manes has not been observed or collected so it. Although few ornithologists have visited far. At present it is not possible to say whether the Polvliepis woodlands in the western Andes such distribution gaps are due to lack of col- of Bolivia, nevertheless it is possible that in lecting efforts and hence are artifacts or not. the future Oreo'nanes could be discovered there. GEOGRAPHICAL VARIATI9N The distribution of Oreomanes in Colom- Figure 1 illustrates the geographical range bia is based on two old specimens from "Pas- of localities from which specimens of Oreo- to," both ofwhich I examined (table 1). Hilty manes have been collected (see table 1). Fig- (in litt.) believes that these two birds were ure 2 and table 2 group these localities into indeed taken in Colombia, and that "'Volcan "populations" in the same way I pooled col- 1 984 VUILLEUMIER: OREOMANES FRASERI I I

TABLE 4 Geographical Variation of Mensural Characters in Oreomanes fraseri (adult males and females pooled) PoP- Wing Length (mm.)b Tail Length (mm.)b Tarsus Length (mm.)b Culmen Length (mm.)" ula- tiona Range (Mean) (n) Range (Mean) (n) Range (Mean) (n) Range (Mean) (n) A 87.0-91.0 (89.0) (2) 61.0-62.0 (61.5) (2) 21.5-23.5 (22.5) (2) - (12.0) (2) B 84.0-87.0 (85.5) (4) 59.0-63.2 (61.0) (5) 20.5-23.0 (21.9) (6) 12.4-14.0 (12.8) (6) C _ (88.5) (2) 62.5-65.0 (63.8) (2) 22.0-22.5 (22.3) (2) 12.5-13.0 (12.8) (2) D - - E 86.0-93.0 (89.3) (9) 58.5-66.5 (61.8) (10) 21.5-24.0 (22.9) (8) 13.0-14.0 (13.3) (6) F 86.5-94.0 (91.0) (6) 60.5-67.5 (65.2) (5) - 24.5 (1) 14.0-14.5 (14.3) (2) G 92.0-93.0 (92.5) (2) 62.5-65.0 (63.8) (2) 24.0-24.5 (24.3) (2) 14.0-14.5 (14.3) (2) H I 89.0-90.5 (89.8) (2) 63.0-63.5 (64.0) (2) - 22.9 (1) - - - J 83.5-89.0 (86.8) (4) 59.0-60.5 (59.5) (4) 23.0-24.0 (23.5) (3) 13.0-14.5 (13.8) (4) K - 91.0 (2) 65.0-65.5 (65.3) (2) - - - - L - 89.0 (1) - 61.0 (1) - 23.5 (1) - 15.0 (1) M - 87.0 (1) - 63.0 (1) - 20.5 (1) - - - N 86.5-89.0 (87.8) (2) 61.0-63.0 (62.0) (2) 23.0-23.5 (23.3) (2) 14.5-16.0 (15.3) (2) 0 86.0 (1) - 61.5 (1) - 24.5 (1) - 15.5 (1) a Populations lettered as in table 2. h Wing, tail, tarsus, and culmen measurements: see footnote to table 3. lecting localities when studying geographical be no important geographical variation in variation in the Asthenesflammulata super- wing length, tail length, or tarsus length. species (Vuilleumier, 1968). The reason for However, there is a trend of clinal variation the geographical grouping is that the number in culmen length (table 4, fig. 3), the south- of specimens available from any single lo- ernmost birds from Bolivia being larger than cality is so small that it is impossible to assess the northernmost ones from Colombia and intra-populational variation. I hoped that Ecuador, whereas the geographically inter- pooling localities into populations would en- mediate ones from Peru are morphologically able me to obtain samples with a sufficient intermediate also. But because of the small number of specimens to be able to analyze number of individuals in each sample area trends ofvariation statistically. In fact, as can one cannot be entirely certain that geograph- be seen from table 2, the numbers of speci- ical variation in culmen length is truly clinal, mens in each pooled sample are so small even or is in fact slightly discontinuous. So further after such grouping that statistical analysis discussion of what appears to be a disconti- appears meaningless. nuity (marked "-" in table 5) must be post- Table 3 summarizes sexual variation for poned until more specimens are available. three sample areas for which both males and Geographical variation in some color char- females were available in small numbers. acters (e.g., forehead, dorsum) appears to be Males appear on the average to be slightly clinal (table 6). Thus the darkest birds are larger than females in wing, tail, and tarsus from the northern part of the range of 0. length, whereas females may be slightly larger fraseri (Ecuador), whereas the palest ones than males in culmen length and bill depth. come from the southern part of the range Because these sexual differences are very (Bolivia), with intermediate birds in the geo- small, and because the samples are small, graphical center (Peru). This decrease in the males and females have been pooled in the intensity ofpigmentation from north to south analysis of geographical variation. Only the along the high Andes is found in a number immature specimens have been excluded of other bird species, including Nothoprocta from consideration. ornata (Tinamidae), Attagis gayi (Thinocor- Tables 4 and 5 show that there appears to idae), Metriopelia melanoptera (Columbi- I12 AMERICAN MUSEUM NOVITATES NO. 2777

TABLE 5 parts), so that the overall trend ofgeographical Trends of Geographical Variation in Mensural variation in color and color pattern appears Characters in Oreomanesfraseri to be more checkerboard than gradual in na- (adult males and females pooled) ture. Trend of Increase or Decrease Beginning with In summary, the data presented in tables Northernmost Populationb 4, 5, and 6 and in figure 3 show that geo- variation in Popula- Wing Tail Tarsus Culmen graphical mensural and color tiona Length Length Length Length characters in Oreomanes fraseri is partially clinal, and partially checkerboard.minor, A It appears to be clinal for some characters B - - - + taken but checkerboard when + + + -iniiulyindividually, D (?) (?) (9) (?) all characters are considered together. If this E + + + is so, then geographical variation in Oreo- F + + ?) + manesfraseri may reflect both the major eco- G + - + = climatic influences prevailing from north to H ...... south in the high Andes and the geographical I - c (?) (?) disruptions in the range ofthe species. From J a taxonomical point ofview, I conclude from K + + (?) (?) this analysis of geographical variation that L (?) (?) (?) (?) the formal recognition of subspecies is un- M ) (- ( warranted in view of the lack of clearcut dif- N ( - ( + 0 (?M?? (? ferentiation within the species as a whole. a Populations lettered as in table 2. GEOGRAPHICAL BARRIERS ,'Only samples of two or more specimens (see table 4) have been included; the "(?)" indicates that no trend In other papers on speciation in Andean can be detected because of lack of data (no, or too few birds I was able to correlate patterns of dis- specimens); a "+" indicates an increase in size; a "-" a continuous geographical variation and dis- decrease; an "=" no change (trend from north to south); tribution in several species with geographical and a ". . ." no specimen. and ecological barriers along the Andes. This was the case, for instance, in the Asthenes flammulata superspecies (Vuilleumier, 1968), dae), Bolborhynchus aurifrons (Psittacidae), in several superspecies of the genus Diglossa Phrygilus unicolor (Emberizidae) (Vuilleu- (Vuilleumier, 1969a), in the Andean cara- mier, unpubl. data), the Muscisaxicola alpina caras (Vuilleumier, 1970), and in several superspecies (Tyrannidae) (Vuilleumier, groups ofground tyrants (Vuilleumier, 1971). 1971), and Diglossa caerulescens (Coerebi- The geographical distribution of Oreomanes dae) (Vuilleumier, 1 969a). This trend of rel- is in many ways parallel in its patchiness to atively gradual geographical variation, in Or- that of these taxa, but there is no evidence, eomanes and in the other species cited above, either that the patchiness is directly corre- can be correlated with the amount and sea- lated to geographical variation, or that the sonality ofrainfall along the Andes from Co- variation is correlated to obvious ecological lombia southward to Bolivia. Thus rainfall or geographical barriers. Hence, unlike sev- is both more abundant and less seasonal in eral other high Andean taxa, Oreomanes does Colombia and Ecuador than in Peru and Bo- not show incipient speciation correlated with livia, and the Peruvian and Bolivian Andes eco-geographical barriers (Vuilleumier, 1977, are progressively drier from east to west 1980). (Lauer, 1952). The trends ofvariation in birds In other words, whereas species such as living under such climatic regime can be Attagis gayi (Thinocoridae) or Iridosornis ru- ascribed to Gloger's Rule. In Oreomanes, fivertex (Thraupidae) have morphologically however, discontinuities in the clines of de- well-differentiated populations on either side creasing color saturation can be found in sev- of the Northern Peruvian Low that can be eral characters (supraciliary stripe, under- considered isolates of potential speciational 1984 VUILLEUMIER: OREOMANES FRASERI 13

12 2.5 13 13.5 14 14.5 15 15.5 16 mm 1 1 1 -I 1 1. II -j A (2) + B (6) C (2) D E (6) I F (2) I G (2) I H

J (4) K L (I) + M

I N (2) I

0 (I) + FIG. 3. Geographical variation in culmen length (adults) in Oreomanesfraseri. See table 4 for mea- surements. significance (Vuilleumier, 1977, p. 37), the phologically (table 6; see also fig. 3). Similar populations of Oreomanes fraseri living . lack ofmarked morphological differentiation across this gap are poorly differentiated mor- across other eco-geographical barriers also 14 AMERICAN MUSEUM NOVITATES NO. 2777

0 0 0 .o E .0 0i n 0 0 CK) VDao U C)

>. C/ ,0(C) (-_ C) 0 P 11 ; - L.. OSa

- 0 .0 0 UO C.) "? - 0 C. C. V) 0 u

S2Q ._0n U .> u? , 0 t3 0 Q V) 0 6 () 0 _O r-0 )- 0

C1. 03 0 1* 0 =1 * 3 ._ 0 Ho o C) QCA Q.0 C.) tDE A0U U 0-

c) 0I _

0 0. 0w

C) *C) Q U 0 C)0 - =0 ._ C) C)0 -o- A 0). a °

X; ._. -C) .._ 0C_)SC I0 C) 0T ;). 0 CZ =C, C/

U CL_ 0 o C.) C/) LL U0Q 1984 VUILLEUMIER: OREOMANES FR4SERI 15

exists elsewhere in the range of Oreomanes aging behavior of this species, including its fraseri (e.g., Rio Apurimac and Rio Pampas precise food preferences. Ornithologists who in Peru, or Rio La Paz in Bolivia). Oreo- have the opportunity to study Oreomanes in manes fraseri is thus a well-marked genus detail will be able to answer some ofthe ques- with a single species having only minor geo- tions about the ecology and distribution of graphical variation. Its patterns of variation this unique bird. and distribution, by themselves, give no clues (2) Distributional patchiness and incipi- as to the history of the taxon in the Andes. ent speciation: The patchy distribution of Hybridization between Oreomanes and Oreomanes can be correlated with geograph- Conirostrum (Schulenberg, in press), how- ical variation in color and size, but the minor, ever, strongly suggests that Oreomanes orig- clinal and checkerboard nature of this vari- inated in the Andes from ancestors living in ation does not appear conducive to incipient the high, wet, and cool environments of the speciation, unlike what has been found in cloud forest type. several other Andean taxa having similar dis- tribution patterns. It is noteworthy that the DISCUSSION other Polylepis specialist, Xenodacnis parina, shows more geographical variation than The two problems investigated in this pa- Oreomanes fraseri. The correlation between per can now be addressed. patchiness and gene flow (or its absence) in (1) Oreomanes and Polylepis: The patchy Oreomanes is thus unclear. George (1964) distribution of Oreomanes does indeed cor- thought that the resemblance between the respond to that of Polylepis woodlands. So color of Oreomanes and that ofPolylepis bark far as I know, Oreomanes has not been ob- was an example of protective coloration. If served or collected in another habitat, al- this were the case, then some selection pres- though it might be expected to occur in mon-< sure arising from predation could be held re- tane forests at or just below timberline. sponsible for the "mimicry" between bird and Observers should look for Oreomanes in bark. One could therefore expect that, as a habitats other than Polylepis woodlands. They result ofsuch selection, the geographical vari- are also urged to search for this taxon in Poly- ation in color of Oreomanes would parallel lepis woodlands in other areas than those cit- the geographical variation in bark color of ed in this paper, both within and outside its the different species of Polylepis. I attempted known range. For example, why does Oreo- to match the color ofpieces of Polylepis spp. manes not live in Polvlepis woodlands ofmost bark from Peru and Bolivia with bird skins of Colombia or of Chile and Argentina? Are from these areas, but failed to obtain good the ecological conditions of the Polylepis correlations, perhaps because I had too few woodlands found in these countries unsuit- bark samples for a precise analysis. A thor- able for Oreomanes firaseri? Wet Polylepis ough comparison with adequate samples woodlands of Colombia, where Oreomanes should be undertaken and future investi- is absent, do not seem very different from gators should also look for potential preda- those of Ecuador, where it occurs (personal tors of Oreomanes in order to test the hy- observ.). Both kinds of woodlands have avi- pothesis of protective coloration. faunas including such typical montane forest species as Margarornis squamiger (Furnari- LITERATURE CITED idae), Grallaria quitensis (Formicariidae), Myioborus melanocephalus (Parulidae), Di- Berlepsch, Comte H. von glossa carbonaria, and D. cyanea (Coerebi- 1900-1901. Sur quelques esp6ces nouvelles ou dae). A detailed of the peu connues recueillies dans le departe- investigation ecolog- ment de Cuzco (Perou central) par M. ical requirements of Oreomanes should be Otto Garlepp. Ornis, vol. 1 1, pp. 197- undertaken at several localities along its range, 198. including both moist and dry Polylepis wood- Bitter, G. lands. Besides its nuthatch-like feeding be- 191 1. Revision der Gattung Polylepis. Bot. havior, very little else is known of the for- Jahrb. Syst., vol. 45, pp. 564-656. 16 AMERICAN MUSEUM NOVITATES NO. 2777

Bock, Walter J. Koepcke, H.-W. [In press] Is Diglossa (?Thraupinae) monophy- 1961. Synokologische Studien an der West- letic? In, Neotropical ornithology. seite der peruanischen Anden. Bonn Bond, James, and Rodolphe Meyer de Schauensee Geogr. Abhandl., Heft 29, pp. 1-320. 1939. Descriptions ofnew birds from Bolivia. Koepcke, Maria Part I.-Oscines. Notulae Naturae, no. 1970. The birds of the Department of Lima, 12, pp. 1-5. Peru. Wynnewood, Pennsylvania, Liv- 1942. The birds ofBolivia. Part I. Proc. Acad. ingston Publ. Co., 144 pp. Nat. Sci. Philadelphia, vol. 94, pp. 307- Lauer, Wilhelm 391. 1952. Humide und aride Jahreszeiten in Suid- Brokaw, Howard P. amerika und Afrika und ihre Beziehun- 1976. Birds of Pampa Galeras, Peru. Delmar- gen zu den Vegetationsgiurteln. Bonn. va Omithologist, vol. 11, no. 1, pp. 26- Geogr. Abhandl., Heft 9, pp. 15-98. 30. Lowery, George H., Jr., and Burt L. Monroe, Jr. Carriker, Melbourne A., Jr. 1968. Family Parulidae. Pp. 3-93, In, R. A. 1933. Descriptions of new birds from Peru, Paynter, Jr. (ed.), Check-list of birds of with notes on other little-known species. the world, vol. XIV. Cambridge, Mas- Proc. Acad. Nat. Sci. Philadelphia, vol. sachusetts, Mus. Comp. Zool., Harvard 85, pp. 1-38. Univ. Chapman, Frank M. Mayr, Ernst 1919. Descriptions ofproposed new birds from 1959. Trends in avian systematics. Ibis, vol. Peru, Bolivia, Argentina and Chile. Bull. 101, pp. 293-302. Amer. Mus. Nat. Hist., vol. 41, pp. 323- Meyer de Schauensee, Rodolphe 333. 1951. The birds of the Republic of Colombia. 1921. The distribution ofbird life in the Uru- Fourth part: Alaudidae-Fringillidae. bamba Valley of Peru. U.S. Natl. Mus. Caldasia, vol. 5, no. 25, pp. 873-1112. Bull., vol. 117, pp. 1-138. 1966. The species of birds of South America 1926. The distribution ofbird-life in Ecuador. and their distribution. Narberth, Penn- A contribution to the study ofthe origin sylvania, Livingston Publishing Co., of Andean bird-life. Bull. Amer. Mus. xvii + 577 pp. Nat. Hist., vol. 55, xiii + 784 pp. Morrison, A. George, William G. 1939. The birds of the Department of Huan- 1964. Rarely seen songbirds of Peru's High cavelica, Peru. Ibis, vol. 81, pp. 453- Andes. Nat. Hist., vol. 78, no. 8, pp. 486. 26-29. Parker, Theodore A., III, and John P. O'Neill Graves, Gary R. 1980. Notes on little known birds ofthe Upper 1982. Speciation in the carbonated flower- Urubamba Valley, southern Peru. Auk, piercer (Diglossa carbonaria) complex vol. 97, pp. 167-176. of the Andes. Condor, vol. 84, pp. 1- Paynter, Raymond A., Jr., and Robert W. Storer 14. 1970. Check-list of birds of the world, vol. Hellmayr, C. E. XIII. Cambridge, Massachusetts, Mus. 1912. Uber neue und seltene Vogel aus Sild- Comp. Zool., Harvard Univ., xiv + 443 peru. Verhandlungen Ornithol. Gesell- pp. schaft in Bayern, vol. 1 1, pp. 159-163. Paynter, Raymond A., Jr., and Melvin A. Traylor, 1919. Ein Beitrag zur Ornithologie von Sud- Jr. ost-Peru. Archiv. fur Naturgeschichte, 1977. Ornithological gazetteer of Ecuador. vol. 10, pp. 1-131. Cambridge, Massachusetts, Mus. Comp. Herzog, Theodor Zool., Harvard Univ., 151 pp. 1923. Die Pflanzenwelt der bolivischen An- Paynter, Raymond A., Jr., Melvin A. Traylor, Jr., den und ihres ostlichen Vorlandes. No. and Blair Winter XV. In, A. Engler and 0. Drude (eds.), 1975. Ornithological gazetteer of Bolivia. Die Vegetation der Erde, Sammlung Cambridge, Massachusetts, Mus. Comp. pflanzengeographischer Monographien. Zool., Harvard Univ., 80 pp. Leipzig, Wilhelm Engelmann Verlag, Pearson, Oliver P., and Carol Pearson Ralph 258 pp. 1978. The diversity and abundance of verte- Hueck, Kurt brates 1966. along an altitudinal gradient in Die Walder Sudamerikas. Stuttgart, Peru. Memorias Mus. Hist. Nat. "Javier Gustav Fischer Verlag, 422 pp. Prado," no. 18, pp. 1-97. 1984 VUILLEUMIER: OREOMANES FRASERI 17

Ridgely, Robert S. Venero G., Jose L., and Howard P. Brokaw 1980. Notes on some rare or previously un- 1980. Ornitofauna de Pampa Galeras, Aya- recorded birds in Ecuador. Amer. Birds, cucho, Peru. Publ. Mus. Hist. Nat. "Ja- vol. 34, pp. 242-248. vier Prado," Zool. Ser. A, no. 26, pp. Salvadori, T., and Enrico Festa 1-32. 1899. Viaggio del Dottore Enrico Festa nell' Vuilleumier, Fran9ois Ecuador. Ucelli. Parte prima, Passeres 1968. Population structure of the Asthenes oscines. Bolletino dei Musei di Zoologia flammulata superspecies (Aves: Fur- ed Anatomia Comparata della R. Uni- nariidae). Breviora, Mus. Comp. Zool., versita di Torino, vol. 15, no. 357, pp. no. 297, pp. 1-21. 1-31. 1969a. Systematics and evolution in Diglossa Schulenberg, Thomas S. (Aves, Coerebidae). Amer. Mus. Novi- [In press] An intergeneric hybrid conebill tates, no. 2381, pp. 1-44. (Conirostrum X Oreomanes) and its 1 969b. Field notes on some birds from the Bo- taxonomic implications. In, Neotropi- livian Andes. Ibis, vol. I 1 1, pp. 599- cal ornithology. 608. Sclater, Philip Lutley 1970. Generic relations and speciation pat- 1860. Lists ofbirds collected by Mr. Fraser in terns in the caracaras (Aves: Falconi- the vicinity of Quito and during excur- dae). Breviora, Mus. Comp. Zool., no. sions to Pichincha and Chimborazo. 355, pp. 1-29. Notes and descriptions of new species. 1971. Generic relationships and speciation Proc. Zool. Soc. London, part 28, pp. patterns in Ochthoeca, Myiotheretes, 73-83. Xolmis, Neoxolmis, Agriornis, and 1886. Catalogue ofthe birds in the British Mu- Muscisaxicola. Bull. Mus. Comp. Zool., seum, vol. XI. London, 431 pp. vol. 141, no. 5, pp. 181-232. Short, Lester L., and J. J. Morony 1977. Barrieres ecogeographiques permettant 1969. Notes on some birds of central Peru. la speciation des oiseaux des hautes An- Bull. Brit. Ornithol. Club, vol. 89, pp. des. Pp. 29-51, In, H. Descimon (ed.), 112-115. Paris, Biogeographie et evolution en Simpson, Beryl B. Amerique tropicale. Paris, Publ. Labs. 1979. A revision of the genus Polylepis (Ro- Zool. Ecole Normale Superieure, no. 9. saceae, Sanguisorbae). Smithson. Contr. 1980. Speciation in birds of the high Andes. Bot., vol. 43, pp. 1-62. Acta XVIII Congressus Internationalis [In press] Speciation and specialization in Poly- Ornithologici, Berlin, 1978, vol. II, pp. lepis (Rosaceae, Rosoideae, Sanguisor- 1256-1261. bae). In, M. Monasterio and F. Vuilleu- [In press] Zoogeography of Andean birds: two mier (eds.), Adaptations and evolution major barriers; and speciation and tax- in biota of high tropical montane eco- onomy of Diglossa carbonaria subspe- systems. New York, Springer-Verlag. cies. National Geographic Research Re- Storer, Robert W. ports. Vol. 16, pp. 713-731. 1970. Subfamily Thraupinae, . Pp. [In press] Origins of the tropical high Andean 246-408, In, R. A. Paynter, Jr. (ed.), avifaunas. In, M. Monasterio and F. Check-list of birds of the world, vol. Vuilleumier (eds.), Adaptations and XIII. Cambridge, Massachusetts, Mus. evolution in biota of high tropical mon- Comp. Zool., Harvard Univ. tane ecosystems. New York, Springer- Taczanowski, L., and H. von Berlepsch Verlag. 1885. Troisieme liste des oiseaux recueillis par Zimmer, John T. Monsieur Stolzman dans l'Ecuador. 1942. Studies of Peruvian birds. No. XLIII. Proc. Zool. Soc. London, 1885, pp. 67- Notes on the genera Dacnis, Xenodac- 124. nis, Coereba, Conirostrum, and Oreo- Troll, Carl manes. Amer. Mus. Novitates, no. 1193, 1959. Die tropischen Gebirge. Bonn Geogr. pp. 1-16. Abhandl., Heft 25, pp. 1-93.