24 June 1998 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON I I l{2>;420-424. IWH Taxonomic notes on (Aves: Trochilidae). 1. Eriocnemis dyselius Elliot, 1872 is a melanistic specimen of Eriocnemis cupreoventris (Fraser, 1840)

Gary R. Graves Department of Vertebrate Zoology. National Museum of Natural History, Smithsonian institution, Washington. DC. 20560. U.S.A.

Abstract.—Eriocnemis dyselius Elliot, 1872 is hypothesized to be a mehi- nistte specimen of Eriocnemis cupreoventris (Fraser. 1840). a put'Heg hum- mingbird restricted to the Mountains of and .

Among the families of , the system- mount with glass eyes, is adult as judged by atic of the Trochilidae are the most con- the lack of striations on the maxillary ram- fused, in absolute numbers, by hybrids, ge- phothecum (see Ortiz-Crespo 1972). Previ- netic variants, and the problems associated ously housed in both the Bourcier and Elliot with tax a described from unique specimens collections (see Greenway 1978), the speci- (e.g.. Berlioz & Jouanin 1944: Banks & men is now catalogued in the American Mu- Johnson 1961; Greenway 1978; Bleiweiss seum of Natural History (AMNH 38452). I 1988; Graves I WO. 1993. 1996, 1997a. studied the specimen taking the approach out- 1997b; Hinkelmann el al. 1991). One such lined in Graves (1990) and Graves & Zusi questionable tax on is Eriocnemis dyselius (1990). In determining the scope of the spe- BUiot. 1872 a of in- cies pool to be investigated, Elliot's (1872: determinate origin. Salvin (1892:369) sug- 294) description offers little guidance: gested thai the hlack-plumaged specimen "Four specimens, precisely alike, were, was "perhaps a melanism of E. cupreiven- as I was informed, contained in the small tris." an inhabitant of forest borders and collection of birds from which my ex- shrubby slopes (1950-3000 m) of the Ven- ample was taken: and. although no local- ezuelan Andes and the Eastern Cordillera of ity was given, it is supposed that the Colombian Andes (Hilty & Brown is the habitat of the species." 1986. Fjeldsa & Krabbe 1990). Salvin did The existence of four similar specimens of not elaborate on this proposal. Subsequent unknown origin should not be interpreted as authors (Cory 1918, Berlioz & Jouanin evidence of a differentiated population. Mil- 1944. Peters 1945. Greenway 1978. Fjeldsa linery dealers in the 19th century sorted and & Krabbe 1990) agreed with Salvin but high-graded shipments of hummingbird skins likewise provided no further support for for unusual specimens to otter to collectors this hypothesis. Consequently, the taxono- of natural history specimens. Although the my of E. dyselius is still uncertain. Here 1 circumstances of Elliot's acquisition of E. dy- present evidence that supports Salvin's selius are unknown, the four black-plumaged (1892) conjecture that the holotypc of E. specimens could have been gleaned from dyselius is a melanistic specimen of Erioc- commercial lots consisting of tens ol nemis cupreoventris (Fraser, 1840). thousands of specimens (Doughty 1975). For Materials and Methods the purpose of analysis, the holotype of E. The unsexed holotypc of Eriocnemis dy- dyselius may have originated anywhere in selius (Fig. 1), a partially relaxed taxidermy northwestern South America (see Berlioz & VOLUME- 111. NUMBI-R 2 421

Fig I. Hototype of Eriocnemis dyssiitu Ulliot. 1872 (AMNH ,1X452).

Jouanin 1944). a region inhabited by more member of Eriocnemis. I took standard than 150 species of hummingbirds (Hilty & measurements (rounded to the nearest 0.1 Brown 1986, Graves 1990). mm) of adult male and female specimens I considered Salvin's hypothesis as the of Eriocnemis cupreoventris. E. nigrivestis, most likely, a priori. The possibility of hy- and E. vestiius (the three species most sim- bridization was judged to be negligible be- ilar in size and shape to the holotype of E. cause the plumage ol" E. dyselius is sub- dyselius) with digital calipers: wing chord; stantially darker and less reflective than that lengths of rectrices (from point of insertion of any potential parental species in north- of central rectrices); and bill length (from western South America (i.e., Coettgeaa anterior extension of feathers) (Table 1). prunellei, Eriocnemis nigrivestis, Helian- Color comparisons were made under natu- gelus zusii, H. regalis). ral light; the plumage was examined under The downy tibia) puffs and body propor- 10x magnification (Appendix). tions of E. dyselius clearly mark it as a I used principal components analysis 422 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON

Table I,—Ranges and means (± standard deviation) of measurements (nun) of Eriocntmis cuprepvtntris, E. irififn.v. and the type specimen oi Briocngmii dy$*Bia Elliot, l%7J-

fptm Hrt vffstUm S4 (« "si .,:' ',., In = II) AMN1I 1SJS:

Wing chord 59.6-^,3.8 57.2-60.7 57.8-61.3 54.7-59.4 58.8 61,6 * 1.3 58.8 ± LI 59.4 ± 1.1 57.7 t 1.4 Bill lenglh 16.6-18.6 17.0-19,6 15.9-18.8 17.4-19.2 17.5 17.8 * 0.5 18.5 ± 1.0 17.5 ± 0.8 18.5 ± 0.5 Reclrix J 24.5-28.2 24.6-26.6 26.2-29.7 29.7-32.8 26.9 26.4 ± LI 25.8 ± 0.7 28.4 i 1.1 31.1 ± 0.9 Reclrix 2 26,9-3 1.2 26.1-28.2 28.0-31.5 31.3-35.1 29.8 28,6 ± 1.3 27.4 * 0.8 29.9 ± 1.1 32.9 ± 1.1 Reclrix 3 3 1.9-36.6 30.7-33,3 31.1-35.3 33.5-37.7 34.6 33.8 ± 1.5 32.1 t 0.9 33.1 t 1-4 35.9 ± 1.3 Rcetrix 4 36.9-42.4 34.8-38.5 35.3-39.7 35.9-41.0 37.9 39.5 ± 1,8 37.0 ± 1.3 37.9 ± 1.5 38.9 ± 1.5 Reetrix 5 40.4-46.3 36.0-40.1 40,6-45.3 37,6-42,2 39.5 42.9 ± 1.7 (H.h - i (> 42.7 ! 1.4 39.7 ± 1.5

' All measurements on left side.

(PCA) on log,,, transformed measurements lacks glittering iridescence, an observation to reduce the dimensionality of data. Un- consistent with the hypothesis of melanism. related principal components were extract- Although melanism is thought to occur at a ed from covariance matrices (Wilkinson very low frequency in the Trochilidae (e.g., 1989). Factor scores were projected on a Salvin 1892. Greenway 1978), the line bivariate plot to illustrate the relationship of structure of melanism in hummingbirds has reciricial measurements in Eriocnemis (Ta- not been formally investigated, and I will ble 2. Fig. 2). For brevity, the holotype of only briefly address the topic here. Irides- E. dyselius will be referred to as dyselius in cence in hummingbirds is caused by the in- the remainder of the paper. terference of light reflected from the upper and lower surfaces of air-filled vacuoles in Results and Discussion melanin granules, which are closely stacked Currently recognized species of Erioc- in 7-15 layers in the outer keratin of the nemis (Sibley & Monroe 1990, Graves expanded dorsal flanges of feather barbules 1996) exhibit areas of glittering or brilliant (Dorst 1951, Greenewalt et al. 1960, Lucas plumage which probably serve as signaling & Stettenheim 1972). Perceived colors vary badges during agonistic and sexual dis- according to the size of the vacuoles. the plays. The dull black plumage of dyseHits thickness of melanin granules, and the an-

Tablc 2.—-Factor loadings from a principal components analysis (PCA) of reciricial measurements of Elioc- nemis iupreovenths. E. vtstittu. and ihe holotype of Eiincm'inh dyselba Elliot. 1872.

Vjdatiles FT A I

Reclrix 1 (innermost) (103 14 0.0054 Rcetrix 2 0.0298 -0.0035 0.0005 Rcetrix 3 0,0195 0.OO55 -0.0102 Rcetrix 4 0.0093 0.0150 -0.0071 Reetrix 5 0.0012 0.0228 0.0090 Percent variance explained 66.2% 23.1% S V. VOLUME 111. NUMBER 2 423

004 —A thus eliminating that species as a possibili- 003 ty A' i ^A Feather shape of dyselius provides addi- A A 002 tional clues as to its identity. The outermost AA 0.01 A rectrices and longest uppertail coverts of E. o A A cupreoventris are slightly narrower and < 000 * • A A A A more attenuate than those of E. vestitus and -001 0<> A E. nigrivestis, although some overlap oc- • X ii A -003 curs among the species. The shape of these • feathers in dyselius is most similar to those -003 o • of E. cupreoventris. -004 1 i When viewed head-on under direct light, -O10 -005 000 006 010 015 the throat of dyselius emits a dull plum- PCA I beous iridescence but exhibits no evidence Fig. 2, Bivariatc relationship of factor si tires I IT A of a centrally demarcated area correspond- I & PCA 1. see Table 2) from a principal components ing to the gorget found in both sexes of E. analysis ,,| rectricial measurement! ol Eriocnemis ett- nigrivestis and E. vestitus. Eriocnemis cu- prroventris (diamonds: females = tilled, males = emp- ty p, E. vr\ttiti.\ i(riaitgle\: Iemale filled; males = preoventris lacks a defined gorget. Instead, empty): and I he hi i loly pc o f Erim item is tiyxe /i"u.v Elliot. the entire throat and upper breast exhibits 1872 (filled circle). brilliant iridescence in both sexes. The gra- dation of feather size, shape, and reflectivity across the throat of dyselius resembles that of E. cupreoventris. Moreover, the pattern gle of observation. The intensity of color is of mekmizaf ion in dyselius corresponds pre- enhanced by reflectance from multiple lay- cisely with the distribution of iridescent ers of granules. An overabundance and ran- plumage in E. cupreoventris. dom placement of melanin granules in the In summary, the holotype of Eriocnemis keratin would lead to a disarrangement of dyselius corresponds well in size to male E. the reflective layers, the absorption of light, cupreoventris. Subtleties of rectrix shape, and a damping of iridescent brilliance. the lack of a well-developed gorget, and the Eriocnemis cupreoventris and E. vestiuts general pattern of melanization of dyselius are remarkably similar in size and shape also are consistent with Salvin's (1892) hy- and melanistic specimens would he difficult pothesis that dyselius is a melanistic ex- to distinguish. The mean bill and wing ample of E. cupreoventris, and provide no lengths of the respective sexes of the two reason to believe that dyselius represents ei- species differ by 0 to 3.7%. The difference ther a hybrid or a valid species. Thus, the in mean rectrix length varies from 0.4% to name Eriocnemis dyselius Elliot. 1872 cor- 7.5% in males and 5.1% to 20.9% in fe- rectly is placed in the synonymy of Erioc- males. Comparison of raw measurements nemis cupreoventris (Eraser, 1840). and inspection of bivariate plots of PCA variables extracted from rectricial measure- Acknowledgments ments show that dyselius is most similar in size and shape to male E. cupreoventris The critiques of Richard Banks. Tom (Table I, Fig. 2). The bill length of dyselius Schulenberg. Michael Walters, and Richard (17.5 mm) falls outside the range of mea- Zusi significantly improved the manuscript, surements for E. nigrivestis (males: n — 15; 1 am grateful to the curators and staff of the 14.4-15.8 mm. X = 15.1 ± 0.5. and fe- American Museum of Natural History, New males: n = 6: 15.6-16.5 ram. X = 16.0 ± York, for permitting me to study the spec- 0.3: see measurements in Graves 1996), imen. Photographic prints were prepared by 424 PROCEEDINGS OF THE BIOLOGICAL SOCIETY OF WASHINGTON ihe Smithsonian photographic services. hummingbird [HeOodosta leadbeateri x Heliun- Travel was supported by the Alexander geltu anwihysthi'tlis) from northern Colombia. Condor 92:754-760. Wetmore Fund and the Department of Ver- Greene wait. C, H.. W. Brandt. & D D, Iriel. 196(1. Ir- tebrate Zoology, Smithsonian Institution. idescent colors of hummingbird leathers. -Journal of the American Optical Society 50:1005-1016, Literature Cited Green way. J. C. Jr. 1978. Type specimens of birds in the American Museum of Natural History Pan Banks. R. C. & N. K. Johnson. 1961. A review of Ninth 2.—Bulletin of Ihe American Museum of Natural American hybrid hummingbird;,.—Condor 63:3- History 161:1-305. 28. Hilly, S. L., & W. L. Brown. 19*6. A guide to the birds Berlioz. J., & C. Jouanin. 1944. Lisle de Trochilide.s of Colombia. Princeton University Press. Prince- trouves dan-, tes collections commerciales de Bo ton. New Jersey. 836 pp. giila.—Oiscau 14:126-155, ilinkehnqnii, C. B. Nicolai. & R. W. Die ken nan. 1991 Bleiweiss. K 1988. Plumage ontogen) and (axonomic Notes on a hitherto unknown specimen of Afec Maw*, nt die Dusky Starfrontlel CoeOgena oiina h:\bin nchrktnm IBcrlcpseh, 1887; Trochilidae) Wetmore.—Bulletin of ihe British Ornithologists with a discussion of the hybrid origin of I his 'spe- Cluh 108:127-131. cies.'—Bulletin of the British Ornithologists Club Cory. C. B. 1918. Catalogue of * of the A merit, as. 111:190-199. Part 2. No. 1.—Field Museum of Natural History Lucas. A. M.&RR, Steiienheim. 1972. Avian anatomy, Zoological Scries 13:1-315. Integument. Pan 2. United States Department of Dorsi, J. I9M. Rccbeiches sur la structure des plumes Agriculture. Washington DC. Agricultural Hand- des TrochilidCs.—Mc moires du Museum National book 362:341-7^0. D'Hisloire Nalurellc (Scrie A. Zoologic) 1:125- Ortiz-Crespo. H I. 1972. A new method to separate im- 260 mature and adult hummingbirds.—Auk 89:851- Doughty. R. W. 1475. Feather fashions and bird pres- 857. ervation: A study in nature protection. University Peters. J. 1945. Check-list of birds of Ihc world. Vol. 5. of California IVess. Berkeley, California. 1X4 pp Museum of Comparative Zoology. Cambridge. Elliot. D. G 1872. Description of a supposed new spe- Massachusetts, 306 pp. cies of humming bird of tlie genus Eriin-ntttib.— Salvin. O, 1892. Catalogue of the birds in (be British Ibis (Second Series I : 29? 295. Museum. Vol. 16 London. 703 pp. Kjeldsa, J.. & N. Kruhbe. 19%). Birds of the high Andes. Sihley. C. G.. & B L. Monroe. Jr. 1990. Distribution Zoological Museum. University of Copenhagen. and of birds of the world Yale 1 ntser Denmark, 876 pp. sity Press, New Haven. Connecticut. 1111 pp. Eraser. 1L.| 1840. | diameters of several species of hum- Wilkinson. L. 1989, SYSTAT: the system for statistics. ming birds, which had been placed in his hands SYSTAT. Inc.. F.vanston. Illinois. 822 pp. by the tiarl of Derby].—Proceedings of the Zoo- logical .Society of London (pan H):14—19. Appendix Graves. G R. 199(1, Systematics of the "green-throated sunangels" (Aves: Trochilidae): valid laxa or hy- Description of Eriimifmis dvseliiu Elliot. 1872. The brids?—Proceedings of Ihe Biological Society of plumage of tfystlius is entirely black (with the exception Washington [03:6-25. of tibia I plumes), glossier on the crown (bluish sheen I. . 1993. Relic of a lost world: a new species of with faint greenish reflections on the uppcriail covens and sunaugcl fTrochilidae: Heli/mgettlS) from "Bogo- pronounced bronzy -green re 11 eel ions on the innermost ta."—Auk 110:1-8. secondaries. Sides of the head, lores, anil auricular,, are . 1996. Diagnoses of hybnd hummingbirds about same color as the hmdncck and crown but lack the (Aves: Trochilidue). 2 Hybrid origin of Eriimietn- bluish sheen. Dorsal body plumage is subtly darker than i.t sixirrxtromi Buder.—Proceedings of (he Biolog- ventral plumage; feather bases are grayish-bult. pales: ical Society of Washington 109:764-769. neat ihc rachis. Hie throat lacks a structurally demarcated . 1997a. Diagnoses of hybrid hummingbirds gorget; however. Ihe terminal discs reflect a faint plum- (Aves: Trochilidue). 3. Parentage of Le.shia ortahi beous iridescence in direct light (dull black in diffuse Lawrence.—Proceedings of ihe Biological Society light i. The basal margins of some throat leathers arc huf- 110:314-319. fy-while, imparting a somewhat molded or scaled ap- , 1997b. Diagnoses of hybrid hummingbirds pearance lo the throat. Undertail coverts are black with a (Aves: Trochilidae |. 4. Hybrid origin of CuUnhor- bluish sheen. Primaries are dull black hut paler than the ax dt'roiuhi.\ Gould.—Proceedings of the Biolog- dorsal body plumage, Kcctrices are glossy bluish-black ical Society 110:320-325. on the dorsal and ventral surfaces. The well-developed , & R. L. Zusi. 1990. An inlergenerie hybrid tibial "puffs" are white.