4 June 1986 PROC. BIOL. SOC. WASH. 99(2), 1986, pp. 218-224 SYSTEMATICS OF THE GORGETED WOODSTARS (AVES: TROCHILIDAE: ACESTRURA)

Gary R, Graves

Abstract. —Acestrura heliodor (sensu Peters 1945) forms a superspecies com- posed of two allospecies: A. astreans Bangs of the Sierra Nevada de Santa Marta; and A. heliodor (Bourcier) of the Andes of Venezuela, Colombia, and northeastern Ecuador. Acestrura h. meridae Zimmer and Phelps is synony- mized with A. h. heliodor. Acestrura h. cleavesi Moore of northeastern Ecuador is a valid subspecies. Wing and tail lengths are positively correlated with latitude in the heliodor superspecies. Gorget color in males is subject to postmortem change.

The diminutive woodstars of the genus Acestrura heliodor heliodor (Bourcier) Acestrura of the Andean region in western Ornismya heliodor Bourcier, 1840:275. South America are poorly known, under- "Bogota." represented in museum collections, and are Acestrura heliodor meridae Zimmer and frequently misidentified. Acestrura ber- Phelps, 1950:1, Paramo Conejos (4000 lepschi, A. bombus, and A. mulsant are cur- m), Merida, Venezuela. rently considered to be monotypic. A fourth , A. heliodor, exhibits considerable Characters. — Male: Upperparts and Hanks geographic variation that is recognized at are green and the gorget purple. The formula the subspecihe level (Peters 1945, Zimmer of rectrix length in closed tail is: 3 > 4 > 1953): A, astreans of the Sierra Nevada de 5 > 2 > 1 (retrices numbered from the in- Santa Marta, Colombia; A. h. heliodor of side outward). The width of rectrix 4 is the Eastern, Central and Western Cordil- greater than '/> the width of rectrix 3 and leras of the Colombian Andes: A. h. meri- nearly intermediate in width between rec- dae of the Venezuelan Andes; and A. h. trices 3 and 5 (Fig. 2). Females: The chin, cleavesi of northeastern Ecuador, throat, breast, and flanks are rich buffy cin- The purpose of this paper is to re-evaluate namon. In adult females, the tail is rufous the of woodstars in the A. helio- with a broad black band across the center, dor complex, incorporating data from pre- A trace of green occurs just proximal to the viously unreported series of specimens col- band on the central rectrices. The lower lected from 1946-1952 in Colombia by M. rump and upper tail coverts are rufous; a A. Carrikcr, Jr. and deposited in the Na- few feathers have a green central spot. tional Museum of Natural History, Smith- Distribution.— Scattered localities in the sonian Institution (USNM). Differences Andes of Venezuela, and the Eastern, Cen- among populations in plumage color in both tral, and Western Cordilleras of the Colom- sexes and tail configuration in males suggest bian Andes (Fig, I). The reported occur- that the Acestrura heliodor (Fig. 1) is ac- rence of A. heliodor at Cana, Cerro Pirre, tually composed of two allospecies: A. as- Panama (Wetmore 1968, Ridgely 1976) is treans (Santa Marta Woodslar) of the Sierra based on a misidentified specimen of Cal- Nevada dc Santa Marta, and A. heliodor liphlox mitchelli (Robbins et al. 1985). () (A. h. heliodor and A. Specimens examined. —VENEZUELA: h. cleavesi) of the main Andes. Pinos (USNM 1 <3); "Merida" (USNM 1 3, VOLUME 99, NUMBER 2 219

Fig. 1. Distribution of the Acestruru helivdor superspecies based on specimens examined in this study. Some symbols represent two closely spaced localities. Squares = Acestmra astresw, circles — A. h, heliodor, dia- monds = A. h. cleavesi. Hatching indicates areas above 1000 feet elevation. 220 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

(FMNH 1 2).-"New Grenada" (USNM I 3, 1 2; AMNH 2 66, 2 29). None de Santan- heliodor der: Buenos Aires (USNM 7 66, 3 22).— Ocana (USNM 1 3).-Ramirez (CM 3 SS, 1 2; ANSPl2).-Cachiri(CM 1 2).-Las Yen- tanas (CM 2 52). Caldas: Laguneta (ANSP 1 3, 3 22). Huila: San Agustin (USNM 4 dd, 2 22; AMNH 1 3; ANSP 3 66, 2 22).- Belen (USNM I 3). Cauca: El Tambo (FMNH 1 3; WFVZ 1 3).-Tijeras (FMNH 1 3).— Moscopan (USNM 1 2).

Acestrura heliodor cleavesi (Moore) cleaves: Moore. 1934:1. Cu- c leaves! yuja, Ecuador. Characters. — Male: The formula for the Length of rectrices in A. h. cleavesi is: 3 > 4 > 2 > 5 > 1, instead of: 3 > 4 > 5 > 2 > 1, as in A. h. heliodor and A, astreans. Rectrices 4 and 5 arc similar in width, and much narrower than rectrix 3. Female: Sim- ilar to A. h. heliodor but more richly colored on chin and throat and with rump more astreans extensively rufous. Distribution. —Known only from the Amazonian slope of the Andes in north- eastern Ecuador. Specimens examined. —ECUADOR; Baeza (MLOC 3 66, 6 22; USNM 1 6, 1 2; AMNH 1 3).-Cuyuja (MLOC 1 3. 1 2).- Pallatanga (MLOC 1 2).-Rio Hollin Fig. 2. Tail patterns of Acestrura h. heliodor, A. h. clearest and A. astreans. Males on the left, females on (MLOC 1 9).- Rio Oyacachi Abajo (AMNH the right. Color of rectrices: stippling = green: un- 1 3, 3 92).-Rio Tigre (MLOC 1 3, 3 22).- marked = rufous or rich hutf; shaded — black. 1.8 x Tumbaco (MLOC 1 2). natural size. Acestrura astreans Bangs Acestrura astreans Bangs, 1899:76. San Se- 2 22; AMNH 19 33.9 22).-Conejos (USNM bastian (6600 feet), Sierra Nevada de San- 1 2; AMNH I 3 type "meridae," 1 £).- ta Marta, Colombia. Tambor (USNM I 3, 4 22; AMNH 9 33, 2 22).—Escorial (AMNH 4 33).—Tierra Characters. —Mate: In adults, the back, (USNM 1 3). COLOMBIA: (unspecified lo- rump, upper tail coverts, and flanks are me- calities) "Bogota" (USNM 4 33,222; AMNH tallic blutsh-green, instead of green as in A. 24 66 including type of A. h. heliodor, 9 h. heliodor and A. h. cleavesi. The gorget is 22).-"Lower Magdalena" (USNM 1 3, 1 red or reddish-purple, depending on the an- 2).-"Colombia" (MLOC 3 66; USNM 4 66, gle of reflection, not purple as in A. h. he- 2 22; AMNH 2 33). -"Santiago" near Pasto? liodor. The formula of rectrix length is the VOLUME 99, NUMBER 2 221

same as in heliodor. Rectrices 4 and 5 arc gets than more recently collected specimens very narrow, less than '/z the width of rectrix from the Eastern Cordillera. For example, 3. Female: Adult females differ from those specimens from Buenos Aires, Norte de of A. heliodor in having the upper tail co- Santander, taken in 1946, have slightly more verts green instead of rufous. The central purplish, less pinkish gorgets than speci- rectrices of A. astreans are green, instead of mens collected in 1916 at Ramirez, Norte rufous with a black band like the other rec- de Santander. In turn, the "1916" speci- trices, as in A. heliodor (Fig. 2). The venter mens have gorgets that are more purplish, of A. a.s//*ra/w is lighter, not as richly colored less pinkish, than those of pre-1900 "Bo- as in A. heliodor. gota" specimens. Specimens collected at ap- Distribution. — Restricted to the Sierra proximately the same time in the Eastern Nevada de Santa Marta. Specimens have Cordillera and the Venezuelan Andes are been taken on the western, eastern, and indistinguishable in gorget coloration. Aces- southern slopes at elevations between 2700 trura h. meridae Zimmer and Phelps, 1950, and 6600 feet (ca. 825-2010 m). should thus be considered a synonym of A. Specimens examined. — Sierra Nevada de h. heliodor Boucier, 1840. In another ap- Santa Marta: San Sebastian (USNM 2 22).— parent example of postmortem change from Chinchicua (USNM 4 22).-Vista Nieve shorter to longer wave lengths, the gorgets (USNM 1 a. 2 22).-Cincinati (ANSP 1 6, 1 of specimens (USNM 333521; MLOC 7015. 9).-El Mamon (AMNH 1 6). — No further 7016, 7023, 10367, 10377) included in the locality data (USNM 2 % AMNH 1 un- type series of A, h. cleavesi) arc presently sexed). matched closely by Rose Color (capitalized names from Ridgway 1912), instead of Variation in Male Gorget Color Rhodamine Purple (Moore 1934). Gorget color in the Acestrura heliodor su- The magnitude of postmortem change pers pecies exhibits significant geographic observed in specimens collected from a sin- variation. In contemporaneously collected gle locality equals thai of the contempora- specimens from the main Andes, gorget col- neous geographic variation found among the or varies clinally from pinkish-purple in Colombian populations of A. h. heliodor. northeastern Ecuador to purple in the East- The gorgets of from the Central (Be- ern Cordillera of Colombia and the Vene- len, Tijeras, Laguneta) and Western Cor- zuelan Andes; specimens from the Western dilleras (El Tarn bo) are pinker, less purplish and Central Cordilleras are intermediate. than those of contemporaneously collected The Santa Marta population is character- (1942-1957) specimens from the Eastern ized by a dark red gorget. Cordillera, but match those of older speci- Zimmer and Phelps (1950) separated the mens from the same region. Venezuelan (A. h. meridae) populations These observations suggest that there are from those of Colombia on the basis of "a some discrete differences among heliodor darker, more purplish, less reddish throat" populations. The lack of contemporaneous- in males. Unfortunately, assessment of this ly collected series from key populations, character is hampered by previously unrec- however, prevents the subspecifie partition ognized postmortem change in gorget color. of A. heliodor on the basis of gorget color They compared their Venezuelan series, alone. collected mostly between 1903 and 1921, Size and Shape Variation with a large series of pre-1900 "Bogota" specimens, most of which were probably Wing length is positively correlated with obtained from the Eastern Cordillera. These latitude in males (Fig. 3; n = 31, r2 = 0.586, have consistently pinker, less purplish gor- P < 0,0001) and females (n = 39, r2 = 222 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

36 26 :: J 34- * L i : . U r * 1 «32 9. i! : 2* Z * * * i:

i ^30- : 14 » i D sDS o 5' 10° 15°N .; * M >.•*•* LATITUDE cr j. -J ' 2G 5°s 0° 5° \V^ 15°N 32 • LATITUDE Fig, 3. Relationship of wing length with latitude in o30 "" the Acestrura keliodor superspeeies. Squares = A. as- Z treans; triangles = A. h. keliodor, diamonds = A. h. &28 • cleavesi. »v • 26 26 0.522, P < 0.0001). Tail length in males, 16 18 20 22 24 TAIL an important taxonomic character (Moore [934), is also positively correlated with lat- Fig. 4. Relationships ol'tail length wiih latitude and 2 wing length in males of the Acestntra heliodor super- itude (Fig. 4; T = 0.761, P < 0.0001) and species. Squares = A, astreans; solid triangles = A. h. wing length (H = 0.658, P < 0.0001). These heliodor; empty triangles = pre-1 900 "Bogota" speci- correlations conform to Bergman n's Rule men identified as A. h. keliodor on basis of plumage (cf.Handfbrd 1983, Rent sen 1984) and sug- but lacking definite locality data; diamonds = A. h. gest that tail length is more closely linked cleavesi. with allomctry and aerodynamic function, than to sexual display. Culmen length is un- of tail shape, but are easily distinguished co rrel a ted with either latitude or wing length from A. h. cleavesi and A. astreans. As ex- (P > 0.05) in either sex. Tail and wing length pected, "Bogota" specimens have wing/tail in males of A. astreans and A. keliodor do ratios similar to specimens from known lo- not overlap. 1 performed a Principal Com- calities in the Eastern Cordillera (Fig. 4). ponents Analysis of culmen, wing, and tail lengths of males. Not surprisingly, all three Taxonomic Conclusions taxa (A. astreans, A. h. keliodor, A, h. cleavesi) had nonovcrlapping distributions Acestrura astreans differs from A. hello- along the axis of the first principal com- dor in several important characteristics that ponent (which explained 60.8% of the vari- are not the terminal stales of observed clinal ance). variation, and that support the recognition These taxa are also readily identified by of A. astreans as a full species. The most the proportional width of the outermost rec- important of these are: (1) the distinctive trices (numbers 3-5) in males. The propor- shape of the rectrices in the males; (2) the tional width of rectrices is not clinal. Pop- coloration of body plumage in both sexes: ulations of,4. h. heliodor from the Western, (3) pattern and color of the central rectrices Central, and Eastern Cordilleras of Colom- of females; (4) the gorget color in males. bia and the Venezuelan Andes arc not dis- Because oftheir allopatric distributions, the tinguishable from one another on the basis existence of reproductive isolating mecha- VOLUME 99, NUMBER 2 223 nisms between astreans and heiiodor can 1949, Paynter and Traylor 1981), Santiago only be surmised. However, me morpho- (1°08'N) is ESE of Pasto, midway between logical differences between ihese taxa are of populations of A. h. cleavesi at Cuyuja the same scale as those observed among the (0°24'N) on the Amazonian slope of the Ec- other species of Acestrura and Chaetocercus uadorian Andes and those of A. h. heiiodor jourdanii. The high degree of endemism in at San Agustin (1D53'N) in the Upper Mag- the Santa Martas has long been recognized dalena Valley. Apparently, no specimens of (Chapman 1917. Todd and Carrikcr 1922). Acestrura heiiodor have been collected be- At least 12 other avian taxa endemic to the tween Cuyuja and San Agustin during the Santa Marta massif, with affinities in the past century. The Santiago specimen is in- northern Andes, are recognized as specifi- distinguishable from typical females of A. cally distinct from their most closely related h. heiiodor, and differs from all specimens congeners {Pyrrhura viridicata, Campylop- of A. h. cleavesi in the intensity and distri- terus phalnopeplus. Coeligena phalerata, bution of rufous and cinnamon buff on the Ramphomicron dorsale, Synallaxis fusco- underparts and rump. rufa. Cranioleuca hellmayri, Grallaria Appropriate habitat is available in the bangsi, Myiotheretes pernix, Myioborusfla- distributional hiatus, and the two taxa al- vivertex, Basiieuterus basilicus, Anisogna- most certainly intergrade or come into con- thus melanogenys, At/apetes melanocepha- tact somewhere between Santiago and Cu- lus). yuja. I recommend that cleavesi should best The systematic status of Acestrura h. be regarded as a subspecies, rather than a cleavesi is uncertain. Except for rectrix shape full species, until specimens from the ap- in males, plumage differences between A. h. propriate regions are obtained. cleavesi and A. h. heiiodor appear to be pri- marily quantitative and clinal. The report Acknowledgments of A. h. heiiodor in Ecuador (Moore 1934) is based on three specimens with definite For the loan of specimens I thank the locality data: Two adult females (MLOC curators and staff of the American Museum 3083, Pallatanga, lD59'S; MLOC 3084, of Natural History (AMNH), Academy of Tumbaco, 0°13'S); and an immature male Natural Sciences, Philadelphia (ANSP). (MLOC 3086, Pallatanga). Both localities Carnegie Museum of Natural History (CM), are on the Pacific slope of the Andes. Be- Field Museum of Natural History (FMNH), cause the plumage of the immature male of Moore Laboratory of Zoology, Occidental A. h. cleavesi is unknown, the male speci- College (MLOC), and the Western Foun- men cannot be identified to subspecies. The dation of Vertebrate Zoology (WFVZ). I females resemble the more heavily pig- thank Storrs L, Olson, Kenneth C. Parkes. mented individuals of the nominate race J. V. Pern sen, and Richard L. Zusi for many from Colombia but also match the palest helpful comments. This work was support- examples of A. h. cleavesi from Baeza, near ed by a Smithsonian Postdoctoral Fellow- the type locality. Adult male specimens will ship, be needed to determine the racial affinity of the Pacific slope population in Ecuador. On Literature Cited geographical grounds alone, 1 tentatively consider these specimens as A. h. cleavesi. Bangs, O. 1899. On a small collection ol" birds from A female (FMNH 45429, received from San Sebastian, Colombia. — Proceedings oT the New England Zoological Club 1:75-80. the "Museum Boucard") was collected by Bourcier. J. 1840. Oiseau-mouche nouveau, —Revue Delattre before 1850 at "Santiago. Colom- Zoologique, p, 275- bia." If correctly located (de Schauensee Chapman, F. M. 1917. The distribution of -life 224 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

in Colombia. —Bulletin American Museum The avifauna of Cerro Pirre, Darien, eastern Natural History 36:1-729. Panama. In P. A. Buckley, M. S. Foster, E, S. de Schauensee, R. M. 1949. Birdsof the Republic of Morton, R. S. Ridgely, and F. C Buckley, eds.. Colombia. Part 2.-Caldasia 23:381-644. Neotropical ornithology. Ornithological Mono- Handford, P. 1983. Continental patterns of morpho- graphs No. 36, pp. 198-232, logical variation in a South American spar- Todd, W. E. C, and M. A. Carriker, Jr. 1922. The row.—Evolution 37:920-930. birds of the Santa Marta region of Colombia: A Moore, R. T. 1934. A new species of , study in altitudinal distribution.—Annals of the genus Chaetucercus from eastern Ecuador.— Carnegie Museum 14:1-611. Condor 36:1-6. Wetmore, A. 1968, The birds of the Republic of Pan- Paynler, R. A., Jr., and M. A. Traylor. Jr. 1981. Or- ama. Part 2 —Columbidae (pigeons) to Picidae nithological gazetteer of Colombia. Harvard (woodpeckers). —Smithsonian Miscellaneous University, Cambridge, Massachusetts, 31 I pp. Collections 1:1-605. Peters, J. 1945. Check-list of birds of the world, vol. Zimmer, J, T. 1953. Studies of Peruvian birds. No. 5, Cambridge. Massachusetts: Museum ofCora- 63. The hummingbird genera Orconympha, parativc Zoology. Schisles, fletiothryx, Loddigesia, lleliotnaxier, Remsen. .1. V.. Jr. 1984. Geographic variation, zoo- Rhodopis, Thaumasrura, Calliphtox, Mynis, geography, and possible rapid evolution in some Myrmia. and Accstrura. — American Museum Cranioleuca spine tails (Furnariidae) of the An- Novitates 1604:1-26. des.-Wilson Bulletin 96:515-523. , and W. H. Phelps. 1950. Three new Vene- Ridgely.R.S. 1976, A guide to the birds of Panama. zuelan birds. —American Museum Novitates Princeton University Press, Princeton. New Jer- 1455:1-7. sey, 354 pp. Ridgway. R 1912. Color standards and color no- Department of Vertebrate Zoology, Na- menclature. Washington, D.C. [Published by the author] tional Museum of Natural History, Smith- Robbins, M.B..T. A. Parker III. and S.E.Allen, 1985. sonian Institution, Washington, D.C. 20560.