Proposal (952) to South American Classification Committee

Treat Metallura primolina (with M. w. atrigularis) as a separate species from M. williami

Background

Andean hummingbirds of the genus Metallura show complex patterns of geographic replacement and are represented by both micro-endemics and more widespread but morphologically highly variable species, such as M. tyrianthina (Loddiges 1832) and M. williami (Delattre & Bourcier 1846). Both of the latter have subspecific variation characterized by dramatically different tail colors, and in the latter case, also throat color. Based in part on this morphological variation, M. w. atrigularis (Salvin 1893) from southern Ecuador and M. w. primolina (Bourcier 1853) from southern Colombia to southern Ecuador have been considered separate species in the past (e.g. Cory 1918), but were subsequently treated as subspecies of M. williami (Peters 1945, Zimmer 1952).

Metallura has been studied from a phylogenetic perspective on two occasions, first by Garcia-Moreno et al. (1999) and subsequently by Benham et al. (2015). The former authors found that M. w. primolina and M. w. atrigularis were not sister taxa, with M. baroni (Salvin 1893) from the Cajas Plateau in southern Ecuador nested between the two. Also, M. odomae (Graves 1980) from southernmost Ecuador and northern Peru, and M. phoebe (Lesson & Delattre 1839) from Peru south of the Marañón Valley, were sister to M. w. primolina. However, their study did not include M. w. williami from the Colombian Central Andes (nor M. w. recisa, Wetmore 1970, from Páramo de Frontino in Colombia) and was solely based three on mitochondrial genes. Benham et al. did include M. w. williami (but not M. w. recisa) and used one mitochondrial and three nuclear genes. Benham et al. found that M. w. williami was sister to M. odomae, rather than to M. w. primolina and M. w. atrigularis, which were sister to M. baroni. The split between the clade including M. w. williami and the clade including M. w. primolina/M. w. atrigularis was well-supported (posterior probability=1.0). Further, this divergence is supported by morphological data, as M. w. primolina and M. w. atrigularis have all green tails in male plumage, whereas M. w. williami has a blue tail.

Fig. 1. Four-locus (ND2, AK1, Bfib7 and MUSK) species tree from Benham et al. (2015).

Paraphyly within the M. williami complex has taxonomic implications. Note that the study suggests other taxonomic implications, particularly involving M. tyrianthina, but more widespread geographic sampling will be necessary to resolve this.

There are two possible courses of action; either M. odomae and M. baroni are treated as subspecies of a broader M. williami, or M. w. primolina and M. w. atrogularis are separated from M. w. williami. The latter option safeguards taxonomic stability for M. baroni and M. odomae, but may also be preferable, because:

1. M. baroni and M. w. atrigularis are known to occur syntopically (Ridgely & Greenfield 2001)

2. M. odomae and M. w. atrigularis are known to occur sympatrically (Ridgely & Greenfield 2001)

3. M. w. williami and M. w. primolina are apparently separated by the R. Caquetá in Colombia, without known introgression (although a possible contact zone is poorly studied, such barriers are known to strongly reduce gene flow in Metallura hummingbirds, Benham & Witt 2016)

I further propose to treat atrigularis as a subspecies of M. primolina for the time being, because of individuals showing signs of introgression (variable amount of black throat feathers) from both Mt. Chimborazo and Papallacta, Ecuador (Moore 1940). No data are available for M. w. recisa, but its potential phylogenetic affinities do not influence the status of other taxa in the complex. I tentatively propose to treat recisa as a subspecies of M. williami, due to their geographic proximity, but recognize that recisa may well be closer to M. primolina based on tail color or it may even deserve species status.

A. A YES to part A is for treating Metallura primolina (with atrigularis as a subspecies of M. primolina)

B. English name: I suggest the English name Ecuadorian Metaltail for M. primolina, as the ranges of both of its subspecies are mainly in Ecuador.

REFERENCES

Benham, P.M., A.M. Cuervo, J.A. McGuire, & C.C. Witt, 2015. Biogeography of the Andean metaltail hummingbirds: contrasting evolutionary histories of tree line and habitat‐generalist clades. Journal of Biogeography 42(4): 763-777.

Benham, P.M., & C.C. Witt, 2016. The dual role of Andean topography in primary divergence: functional and neutral variation among populations of the hummingbird, Metallura tyrianthina. BMC Evolutionary Biology 16(1): 1-16.

Bourcier, J., 1853. Metallura primolina. Revue et Magasin de Zoologie Pure et Appliquée 2(5): 295.

Cory, C.B. (1918). Catalogue of birds of the Americas. Zoological Series; Field Museum of Natural History, Chicago.

Delattre, M.M.A. & J. Bourcier, 1846. Trochilus williami. Revue Zoologique 9: 308.

García-Moreno, J., P. Arctander, & J. Fjeldså, 1999. Strong diversification at the treeline among Metallura hummingbirds. The Auk 116(3): 702-711.

Graves, G.R., 1980. A new species of metaltail hummingbird from northern Peru. The Wilson Bulletin 92: 3.

Lesson, R.P. & M.M.A. Delattre, 1839. Ornysmia phoebe. Revue Zoologique 2: 17.

Loddiges, G., 1832. Trochilus tyrianthinus. Proceedings of the Zoological Society of London 2(15): 6-7.

Moore, R.T., 1940. The nomenclature and habits of the Black-throated Copper-tailed Hummingbird. The Condor, 42(5): 251-254.

Peters, J.L. 1945. Check-list of the birds of the world. Museum of Comparative Zoology, Cambridge, Massachusetts, USA.

Ridgely, R.S., & P.J. Greenfield, 2001. The birds of Ecuador, status distribution and taxonomy. Helm, London.

Salvin, O., 1893. Metallura baroni. Bulletin of the British Ornithologists' Club 1: 49.

Wetmore, A., 1970. Descriptions of additional forms of birds from Panamá and Colombia. Proceedings of the Biological Society of Washington 82: 767-776.

Zimmer, J.T. (1952). Studies of Peruvian birds. No. 62, The hummingbird genera Patagona, Sappho, Polyonymus, Ramphomicron, Metallura, Chalcostigma, Taphrolesbia, and Aglaiocercus. American Museum Novitates, 1595.

Paul van Els, July 2022

Comments from Areta: “I vote NO to the split. The lack of sampling of recisa makes part of the taxonomy a guesswork. I find it difficult to endorse that the green tails of primolina and atrigularis males add support to their sister relationship to baroni, whereas the also green-tailed males of recisa would be placed with the blue-tailed males of williami based on geographic proximity. The ND2 topologies differ from each other and from the multilocus one in Benham et al. (2015), and support for some clades in the multilocus tree is not great, which casts doubts on which phylogenetic hypothesis is closer to the true phylogenetic relationships in this recently diverged clade. I would like to see a more dedicated study exploring geographic variation in greater detail, including genetic samples of recisa and solving whether the break between primolina and atrigularis is abrupt or there is evidence of intermediacy. Do vocalizations differ among taxa in this clade?”

Comments from Stiles: “NO. The phylogenetic relationships of the members of this genus are indeed complicated, especially in view of the fact that several of the named taxa have not been sampled genetically. For this proposal to split primolina from williami at the species level, the problem lies with the subspecies recisa: although its plumage color (specifically, its tail) resembles that of primolina, its distribution is separated from that of the also green-tailed recisa by that of nominate williami, in which the tail is blue. Therefore, this conflict of characters can only be resolved by genetic data of recisa (none is mentioned in the proposal). I therefore vote NO on this proposal as it stands (but I also note that tissue samples of recisa surely exist in Colombia, provided by the recent collections of this race by J. P. López. The specimens are in the collection of the Instituto de Ciencias Naturales of the Universidad Nacional de Colombia; the tissues are in the tissue collections of either the Instituto Alexander von Humboldt or the Universidad de los Andes, which should be consulted).”

Comments from Lane: “NO. For reasons given by others above.”

Comments from Robbins: “NO for treating Metallura williami primolina as a species for the well-articulated reasons stated by Gary.”

Comments from Bonaccorso: “YES. The nodal support from Benham et al.´s tree is not great but decent for the clade M. baroni + M. w. primolina + M. w. atrigularis clade (0.92). I recognize that this tree is only based on a few genes, one sample for each subspecies (and no recisa), and probably driven by mitochondrial genes, BUT judging by the topology and branch length’s, a monophyletic M. williami seems unlikely.”

Also, the relationship between M. baroni + M. w. primolina + M. atrigularis makes a lot of biogeographic sense. Metallura baroni is isolated from the “main” Ecuadorian Andes at the Cajas Massif. According to the phylogeny, it is sister to M. w. primolina + M. w. atrigularis from the main Ecuadorian Andes, with primolina to the north of the Girón-Paute depression (in southern Ecuador), and atrigularis to the south. I think M. w. recisa is not really a problem since it can stay as M. w. recisa until molecularly sampled. Then, if it is sister to or part of M. primolina, we will have to deal with it.”

“As a tangent, the relationship of M. w. williami with M. odomae seems very difficult to believe (and the support is very low), but even if it is not true, it does not affect the possibility that M. w. williami is not monophyletic.”

Comments from Pacheco: “NO. As pointed out by Nacho and Gary, due to the absence of genetic sampling of Metallura williami recisa.”

Comments from Claramunt: “NO. I think it is premature to make taxonomic changes with the evidence at hand. Very few specimens have been sequenced, and there is little statistical support for odomae and nominate williami to be sister. The situation seems complex.”

Comments from Remsen: “NO, for all the reasons outlined by others, especially the gaps in taxon sampling but also weak genetic sampling for such closely related taxa. I do have the feeling that stronger data will support the proposed split, as noted by Elisa, but I think it is best to wait for those data.”