Proposal (168) to South American Classification Committee

Lump Amazilia cupreicauda with A. viridigaster

This proposal lump Amazilia cupreicauda (Coppery-tailed Hummingbird) (with three subspecies) with A. viridigaster (Green-bellied Hummingbird) (with two races); they are treated as separate species in our baseline list, following Schuchmann (1999) and Weller (2000).

Amazilia viridigaster occupies the eastern Andean foothills and lower slopes, from extreme W Venezuela (Táchira) to E-C Colombia, with A. cupreicauda occurring at similar elevations in the tepuis of Venezuela, the Guianas and extreme N Brazil.

Amazilia viridigaster was described in 1843 by Bourcier from the E Andes of Colombia (supposedly from Fusagasugá, on the W. slope, but this is almost certainly an error as the species has never been recorded reliably from this slope of the Eastern Andes; Weller accepts the locality in spite of this, stating that "there is no indication that the type locality was wrongly given by Bourcier, nor that the habitat has significantly changed". The latter statement is inexplicable as the area was heavy forest through Chapman's time -- most inappropriate for the species, which Chapman did not find there -- and subsequent deforestation has been very extensive, making the area more suitable for the species, which is still unrecorded any nearer than Choachí on the E slope. Also, the type was presumably collected by Goudot or a native collector employed by him, and at that date the practice of labeling individual specimens was by no means universal, especially by native collectors -- hence, the probability of a faulty type locality seems overwhelming).

The subspecies iodura was described from Táchira, Venezuela, by Reichenbach in 1854 based upon the more purple, less bluish color of the tail. In 1884 Salvin & Godman described cupreicauda from Mt. Roraima based on its differently colored tail, upper tail-coverts and rump. Cory (1918) recognized viridigaster and cupreicauda as species, with iodura as a subspecies of the former. Subsequently Chapman (1929) described duidae as a race of cupreicauda from Mt. Duida of S Venezuela. Peters (1945) considered cupreicauda and duidae subspecies of viridigaster and lumped iodura into nominate viridigaster without comment, and this was followed by most subsequent authors, including Meyer de Schauensee (1966), Meyer de Schauensee & Phelps (1979) and Sibley & Monroe (1990). The most recent taxa to be described in the group were laireti (as a subspecies of viridigaster) by Phelps & Aveledo (1988) from extreme S Venezuela near the Brazilian border, and pacaraimae by Weller (2000) from the Sierra de Pacaraima on the Brazil-Venezuela border. Schuchmann (1999) split cupreicauda (including duidae, pacaraimae and laireti) from viridigaster, while reinstating iodura as a valid race of the latter species, although the rationale for this was published later by Weller (2000). This was followed by Hilty (2003).

In order to evaluate the differences among these forms, and thus the relative merits of placing them in one species or two, it is essential to compare the "morphological differences, especially the colors of the rump, upper tail-coverts and tail" advanced by Schuchmann (1999) and Weller (2000) to support the split. A brief summary follows, based on Chapman (1929), Meyer de Schauensee & Phelps (1979), Phelps & Aveledo (1988), Schuchmann (1999) and Weller (2000), as well as examination of specimens (of viridigaster, here in ICN). The characters are (arranging the races roughly from W to E and N to S):

• viridigaster: rump extensively buffy-brown (to olive-green, fide Schuchmann, although I have not seen a green one), upper tail-coverts violet (near "true violet" of Smithe), tail violet-blue to deep blue (actually the rectrices have more or less of dark steel-blue centrally, violet (like coverts) on borders and tips, but much variation: in most the violet borders are broad and the blue occupies a limited area along the shaft (even reduced to a faint blue gloss in a few), in some the rectrices are mostly blue with only a narrow violet edge. In males, the crissum feathers are purplish centrally, rufous distally with narrow grey fringes. In females the lateral rectrices are brownish to bronzy towards the edges or tip, the crissum feathers mostly dark grey with whitish fringes (Weller 2000). 

• iodura: differs from the former in its copperier to purple, less violet tail (presumably no blue). Weller gives more details: rectrices purplish with bronze lateral webs, "inner rectrices often bicolored", upper tail-coverts reflect purplish to bronze (especially in females), the crissum feathers pales, less glittering (?), grey-green or bronze-green to purplish with broad rufous margins apically. Weller describes the backs of both of these forms as "golden-green", not mentioning specifically the rump.

• duidae: lower back and rump copperier, rectrices bronzy to purplish with no rufous basally (Schuchmann); more strongly contrasted (sic) on the hindneck, back and tail. bronzish coloration and lustrous copper to purple coloration towards the rump and upper tail-coverts, especially in males. The contrast in the rectrices varies from bronze reddish to purple red, with bronze-green to dark purplish tips. Females are less intensively colored with more bronze-green in the tail; ..color transitions in the tail more variable, less abrupt than in cupreicauda (Weller).

• laireti: darker green above and below than other forms, upper tail-coverts golden-bronze to coppery (Schuchmann), or: rump and upper tail-coverts darker (rather than paler) brown (cf. cupreicauda), rectrices blackish-brown as opposed to violaceous-brown (cf. viridigaster), according to Phelps & Aveledo; or more uniform, lacking coppery back feathers, the rectrices have chrome bronze (??) instead of purplish tinge (Weller).

• pacaraimae (from Weller): differs from cupreicauda in its darker green color, with the back more coppery, the rump with purple reflections without rufous traces at the base (?), the tail rather uniform purplish to copper red; generally darker green than duidae or laireti; the tail lacks the variable coloration with bronze-green parts of the former and the chrome bronze tinge (whatever that means) of the latter. Head and neck bronze green, back and rump bronze green to copperish, rump with purplish reflections; upper tail coverts purplish fringed with rufous "possibly due to molt", feathers of chin with inconspicuous whitish bases, underparts and flanks glittering golden green, crissum light rufous (male). Female with white in chin more conspicuous, back and rump less copperish, more bronze laterally on inner rectrices. Weller considers this form to "connect cupreicauda with the southern races, especially duidae)".

• cupreicauda (Schuchmann): back and rump golden-green to bronze or coppery, sometimes with purplish reflections. Rectrices with bases and centers rufous to chestnut, broadly edged coppery to bronze. Weller describes it as having less of a copperish tinge on the back and rump than duidae, laireti; more rufous in upper tail-coverts and rectrices than other races; inner rectrices bicolored (rufous and bronze-green) but rarely with purplish. Hilty (2003) describes cupreicauda more briefly as "foreparts green, hind parts (rump, tail) mostly rufous", which coincides with the rather garish figure of this form in Schuchmann (1999) but not the description in the text.

The overall impression I get from all this is that there is considerable variation in color characters of most forms. The situation is further complicated by the fact that the color descriptions of the characters of the various forms by Phelps & Aveledo, Schuchmann, and Weller often do not coincide; I find Weller's descriptions in particular to be confusing and subjective (use of a color standard would have helped greatly). However, as I read it, there do appear to be some fairly consistent trends in coloration from W to E in the complex as a whole, moving from viridigaster to cupreicauda, the two forms most divergent both in color and geographically (when I figure out what "chrome-bronze" is, I will be better able to evaluate laireti). The rump progresses from more brownish to more bronzy to coppery or golden-green, the upper tail-coverts from violet through purplish to bronzy to partly rufous, the tails from more blue to violet through more coppery to bronze, and the rufous in the rectrices appears to be an exclusive feature of cupreicauda itself. The difference in depth of tail fork cited by Weller between the Pantepui and Andean groups is not great (less than that between some races of Chlorostilbon mellisugus in the restricted sense). All of this could be taken to indicate either that some gene flow might be occurring, or a series of successive colonizations from W to E, such that the number of species into which one divides the complex might depend upon which characters one chooses to emphasize. An extreme splitter might go with as many as three Pantepui species. Schuchmann and his collaborators often place much emphasis upon distribution and disjunct ranges in elaborating their speciation models, and the fact that an Andes-Tepui disjunction occurs in a number of taxa probably influenced them to split these forms from their Andean relatives. However, given the situation as described, I am by no means assured that the four races of cupreicauda are more closely related to each other than any is to viridigaster-iodura and the possibility of multiple independent invasions of the tepuis should not be excluded (especially for laireti). Hence, I feel that the evidence for the two-species split is not convincing as it stands, and that genetic evidence will probably be required to sort out the true relationships in this complex. For the present, I feel that the best and most conservative course is to recognize all as races of viridigaster with the proviso that two (or more) species might be involved, and recommend a YES vote on this proposal. 

(Incidentally, "Green-bellied Hummingbird" is really a most inappropriate name for this species as only the throat, breast and sides are green, the belly itself being a contrasting dark brownish-grey. Would it be worth a proposal to change the name to "Purplish-tailed Hummingbird" or something similar?? Although such "name-tinkering" is in principle opprobrious, this is one case where the traditional name is actively misleading. Or should we simply let tradition and stability stand in spite of inaccuracy?).

LITERATURE CITED

Chapman, F. M. 1929.

Cory (1918)

Hilty 2003

Meyer de Schauensee 1966.

Meyer de Schauensee & Phelps 1979

Peters 1945.

Phelps, W. H. Jr. & R. Aveledo H. 1988. Una nueva subespecie (Trochilidae) de la Serranía de Tapirupecó, Territorio de Amazonas, Venezuela. Boletín de la Sociedad Venezolana de Ciencias Naturales 145:7-10. 

Schuchmann 1999. HBW vol. 5.

Sibley & Monroe 1990.

Weller, A.-A. 2000. A new hummingbird subspecies from southern Bolívar, Venezuela, with notes on the biogeography and taxonomy of the Saucerottia viridigaster-cupreicauda species group. Ornitologia Neotropical 11:143-155.

Gary Stiles, March 2005

Comments from Remsen: "YES. I agree with Gary that the conservative course is to treat all as one species until better data dictate otherwise."

Comments from Zimmer: "YES. It looks to me as if there is either 1 species or several, but not two. Morphological differences do seem to follow a west to east trend, with the most different taxa (morphologically) also being the ones most geographically distant from one another. Allocation of some of the Pantepui populations to either cupreicauda or viridigaster appears arbitrary on available data. As an aside, I can't see that Schuchmann addresses pacaraimae at all, although his range maps for cupreicauda imply that it is included therein. Although I suspect that more than one species will ultimately emerge from this mess (based as much on biogeographical considerations as anything), I think the conservative course at present is to recognize a single species."

Comments from Nores: "SI, pero no demasiado convencido. Si uno mira las figuras que aparecen en el HBW (5:600) parece evidente que son especies diferentes, y aunque el color del dorso podría ser subespecífico, no lo es el de la cola. Además, hay una enorme distancia en el rango de las dos especies. Sin embargo, después de ver los comentarios de Stiles y la gran variación que tiene la especie, parece que las diferencias morfológicas, como señala Zimmer en su comentario, siguen una tendencia de oeste a este con los dos extremos diferentes como los que se ven en las figuras del HBW."

Comments from Robbins: "YES. Having observed birds, that are quite different, at the two extremes of this complex's distribution (Andean & the Acari Mts. on the Guyana/Suriname/Brazil border), I suspect Schuchmann and Weller are at least partly correct in their species division of Andean vs. Guianan Shield birds. However, as Gary's review indicates, there are a number of details that need to be clarified before we start recognizing additional species. Thus, I vote "yes" on lumping these."

Comments from Pacheco: "YES. É forçoso reconhecer que, em vista das considerações de Stiles e a depender de evidência genética, o tratamento mais conveniente, por ora, seja manter reunidos todos os táxons envolvidos sob o nome de Amazilia viridigaster."

Comments from Jaramillo: "YES - for the record, as this has passed already. I do think that there is likely more than one species level taxon here, and that makes me uncomfortable lumping the lot, but it certainly is not clear where a good, provisional division could be made now as we await other data sets."