Proposal (930) to South American Classification Committee

Revise generic limits in the Lesbiini: A. Expand Oxypogon to include Oreonympha and Chalcostigma, and B. Modify linear sequence

The comprehensive phylogeny of the Trochilidae by McGuire et al. (2014) has generated major revisions of generic limits in the family that have already been addressed by SACC. We (mostly Gary) are in the process of addressing a series of minor revisions with each tribe to bring the SACC classification up to date.

This proposal addresses one case within the Lesbiini: the possible extensive paraphyly of Chalcostigma.

To put the proposed change in context, an overview of the phylogeny with respect to current generic limits is presented below.

When more than one option exists, we usually emphasize branching order over branch lengths, and in most cases prefer the option that least affects stability.

Clade 1. Includes only the genus Sephanoides, which definitely represents a separate genus.

Clade 2. Includes two monophyletic groups of similar age, Lophornis and Discosura, both well characterized morphologically and entitled to generic status.

Clade 3. Includes only the genus Phlogophilus, which definitely retains its generic status.

Clade 4. Includes only the genus Heliangelus. The species regalis is a genetic outlier, with a very distinct adult male plumage (although the female plumage is closer to that of other Heliangelus. Two options exist: A. Separate regalis in its own genus, for which a new name would be required; or B. retain it in Heliangelus, recognizing the similarity of the female plumage to others of the genus. We recommend this option, which also preserves stability.

Clade 5. Includes only the genus Adelomyia, which retains its generic status; although some recent studies suggest that the several subspecies include distinctive lineages, evidence to date appears insufficient to justify recognition of any of these as species.

Clade 6. Includes Heliangelus zusii, Taphrolesbia and Aglaiocercus. Heliangelus zusii is now thought to be a hybrid, and hence may be excluded from further consideration. The branch lengths separating Taphrolesbia and Aglaiocercus are rather short; hence, two options exist: A. lump both into a single genus (in which case, Taphrolesbia (Simon, 1918) has priority over Aglaiocercus (Zimmer, 1930), or B. maintain both as separate genera. The two are quite distinctive in plumage, and this option favors stability; we prefer this option.

Clade 7. Includes the genera Sappho, Ramphomicron, and Lesbia. The three genera separate in this order by two rather small steps. Here also there appear two alternatives: A. lump all three into a single genus for which Lesbia (Lesson, 1833) takes priority over Sappho (Reichenbach, 1849) and Ramphomicron (Bonaparte, 1850). And B. maintain three separate genera. All three are distinctive morphologically, and this option maintains stability; hence, we favor it.

Clade 8. This includes only Oreotrochilus, which clearly should retain its generic status.

Clade 9. This includes only Polyonymus, on a long branch: definitely a monotypic genus.

Clade 10. Another morphologically distinctive species with no close relatives: Opisthoprora also merits generic status.

Clade 11. Includes members of three genera: Chalcostigma, Oreonympha, and Oxypogon. The very short branches between them indicate that their speciation was rapid; a complication is that as it stands, Chalcostigma is likely paraphyletic with respect to the other two genera, but note the low support values for the topology. We therefore consider a novel option: include all members of this clade in a single genus, for which Oxypogon Gould, 1848, takes priority over Chalcostigma Reichenbach, 1854, and Oreonympha Gould, 1869. It is noteworthy that this is the first time that these three genera have been treated together, without other intervening genera (usually Metallura) in the linear sequence. The support for the node that unites them as a group is very strong. An expanded Oxypogon would include all of the species in which the gorget is restricted laterally to form a narrow, multi-colored “beard” (see photos below), which could be considered a synapomorphy; also, Oreonympha and Oxypogon s.s. (including the three species we now recognize) share conspicuous white in the tail. All are species of high elevation cloud-forest, elfin forest, and páramo.

Clade 12. Includes only the genus Metallura. Because the species are mostly separated in a sequence of moderate to small steps, we see no reason to attempt to subdivide the genus, which is morphologically coherent, and recommend continuing to consider all species as congeneric, which also maintains stability.

A. Expand Oxypogon to include Oreonympha and Chalcostigma: As noted above under Clade 11, we think this merger warrants consideration because there really isn’t any evidence that Chalcostigma is monophyletic. A practical problem is that McGuire et al. (2014) did not have access to tissue of the type species (heteropogon) for Chalcostigma Reichenbach, 1854; therefore, even the assignment of species to Chalcostigma s.s. would depend on assumptions that cannot be supported by genetic data. Traditional Chalcostigma is fairly heterogeneous, with C. ruficeps decidedly smaller than the others, and the female is phenotypically more similar to Metallura tyrianthina than to its current congeners; it also occurs at lower elevations. One could make the case for splitting Chalcostigma at least 4 ways based on the topology above, at least 3 of which would be monotypic: (1) herrani, (2) ruficeps (sister relationship to olivaceum has weak support); (3) olivaceum; and (4) stanleyi, and hope that type species heteropogon is sister to one of these. This would allow maintaining Oxypogon as a tight genus of four allospecies (see SACC 609) and Oreonympha as a distinctive monotypic genus. However, the branch lengths among these are more similar to those among many congeners within other genera in the Lesbiini.

Photos of males from LSUMNS specimens:

A YES vote would have the added benefit of highlighting the close relationship among all these rather phenotypically divergent but similar species and suggests that there is something about this lineage that allows rapid plumage diversification, perhaps facilitated by their isolation and likely very small populations during the height of interglacial warm cycles. An expanded Oxypogon would be certainly monophyletic. A NO vote would maintain a likely paraphyletic Chalcostigma but would emphasize the need for better-supported typology and possible inclusion of the type species of Chalcostigma.

B. Modify linear sequence. If we do not change generic limits, a change in linear sequence is required. Our current sequence is: Chalcostigma, Oxypogon, Oreonympha. Maintaining those three genera as is would require an inversion of that sequence under standard sequencing conventions, with most diverse genera listed successively, e.g., Oreonympha, Oxypogon, Chalcostigma.

Gary Stiles and Van Remsen, November 2021

Comments from Areta:

“A. NO. I feel uncomfortable making this move, given the low support of several branches and the lack of samples from the type species of Chalcostigma. In the long run, this might be a reasonable generic arrangement, but I prefer to wait for more solid data before making changes.

“B. YES, modify the sequence to match the richness of each group.”

Comments from Robbins:

“A. YES for subsuming all of the former Chalcostigma taxa and Oreonympha into Oxypogon given the node that unites this group is strongly supported and that priority has been clearly established.

“B. YES, for changing the linear sequence.”

Comments from Lane:

“A. YES. I don't think it's likely that C. heteropogon will be outside this clade, so I think it is fairly safe to go ahead with the merging of genera into Oxypogon now.

“B. YES to reordering the members of the larger clade.”

Comments from Pacheco:

“A – YES. As the node that unites this group shows strong support it is acceptable to combine these 3 genera to expand Oxypogon.

“B – YES. The change in the linear sequence must be made.”

Comments from Bonaccorso:

“A. NO. Many of McGuire et al.´s relationships are driven by mitochondrial DNA and in this case many branches are short. Because this is not a “life or death” situation, I think it is better to wait for more data and inclusion of C. heteropogon in the analysis. In the meantime, we can maintain stability for a while longer.”

“B. YES, but just to reflect the phylogenetic sequence.”

Comments from Claramunt:

“A. NO. The three genera are so distinctive phenotypically that I think we should wait for stronger evidence before making any change in this group.

“B. NO. For the same reason.”

Comments from Zimmer: A.“YES for subsuming all Chalcostigma taxa and Oreonympha into a single, expanded Oxypogon, given that the node uniting this group is strongly supported, and that Oxypogon has priority. Normally, I’m a proponent of smaller, more internally cohesive genera, particularly for such morphologically distinct groups such as Oxypogon sensu stricto. However, the apparent paraphyly of the current arrangement bothers me (even though the support levels for the topology are not particularly impressive), and a merger into an expanded Oxypogon would seem to result in an unquestionably monophyletic group. I’m also intrigued by the potential for uniting all of these proportionately short-billed, long-tailed, high-elevation species with narrow, multicolored “beards” into a single group, as mentioned in the proposal. As Dan has opined, I don’t see C. heteropogon falling outside of the clade, so, although it would be preferable for the type species of Chalcostigma to have been sampled by McGuire et al. (2014), its absence from their phylogeny shouldn’t be an impediment to making the proposed generic merger in my opinion. The alternatives would be A) Maintain a status quo with three seemingly paraphyletic genera; or B) maintain the stability of Oxypogon and Oreonympha, while splitting Chalcostigma into as many as four genera (at least 3 of which would be monotypic), and still not knowing for certain how any of them relate to heteropogon. Despite the phenotypic distinctions between the three genera as currently recognized, I think there are more phenotypic, ecological, and distributional characters that unite Chalcostigma, Oxypogon and Oreonympha than there are differences within currently recognized Chalcostigma. B. YES to changing the linear sequence of the expanded clade.

Comments from Remsen: NO. Based on Elisa and Santiago’s comments on the genetic data, and because the genera as currently defined seem to be mostly cohesive units, I’m backing away from my recommendation in the proposal. I do think it would be OK to change the linear sequence because the node that unites the three genera looks solid, and so rearrangement within that group should follow sequencing conventions; the support for altering generic limits is weak, but the support for them as a group is strong … at least as far as that data-set goes, and those are the best data we have so far.”

Additional comments from Stiles: “I very much doubt that C. heteropogon would fall outside Chalcostigma - but surely there should be tissue samples in Colombia … and the node connecting the 3 genera seems well supported, such that to avoid paraphyly of Chalcostigma and place all 3 in Oxypogon still seems to me to be the best option.”

Comments from Jaramillo: “A. YES – I may be in the minority, but these hummingbirds are not all that phenotypically diverse. They are uniquely “bearded” in their gorget shape. They are also relatively short billed, and also somewhat uniquely long-tailed. The tails are sometimes broad and have more area than the wings, not the norm in most hummingbirds. I think this group looks pretty tight as a single genus. YES – if necessary.”