Proposal (780) to South American Classification Committee

Change the generic classification of the Trochilinae (part 1)

Two recent studies of the generic classification of the Trochilini or “Emeralds” detected numerous instances of polyphyly and other incongruences with respect to the DNA-based phylogeny of this group (Stiles et al. 2017a, 2017b), the largest major clade of hummingbirds with over 100 species. The first study addressed problems of generic nomenclature in the Trochilidae, with particular reference to the two of the largest and most problematic genera, Amazilia and Leucippus. The second paper proposed a new generic classification of the Trochilini to bring it into the best possible accord with the phylogeny of McGuire et al. (2014), which treated 275+ species, including most or all species in all of the ca. 30-35 currently recognized genera. Our overall objective was to produce a classification taking as its base the branching pattern of the phylogeny, while preserving stability of existing nomenclature wherever possible. We tried to produce cohesive, diagnosable genera while avoiding producing large, undiagnosable genera on the one hand, and an excessive number of small or monotypic genera on the other; this necessitated a rather more flexible treatment of branch lengths. In the process, we found numerous instances of homoplasy in plumage color and pattern as well as discordance of plumages in other monophyletic groupings. We have all been weaned on a classification whose basis was to a very large extent on plumage, so the new classification resulted in many drastic reallocations of generic circumscriptions: this required the resurrection of nine generic names currently considered synonyms, synonymizing seven currently recognized genera and the creation of one new genus. We here present the new generic classification for review by the SACC. This classification is presented in Figure 1 of Stiles et al. (2017b) below, and as we work through this, we present our reasoning for each change in brief; for further details regarding nomenclatural issues, see Stiles et al. 2017b.

We began by dividing the Trochilini into four large groups (A, B, C and D), within each of which we recognized from two to seven subgroups, and further divided these to produce new generic groupings. We found that many of these new groupings showed strong geographical coherence, sometimes at odds with similarities in plumages. In this proposal we treat groups A, B and C; a subsequent proposal will deal with group D, by far the most difficult, including untangling the chaos associated with the generic names Amazilia and Leucippus.

Group A includes the currently recognized genera Chlorostilbon, Cyanophaia and Cynanthus. Given that Cynanthus as currently constituted includes the species sordidus, one alternative would be to lump all the other genera into Cynanthus. We rejected this option because it would mask considerable genetic and phenotypic diversity. Because sordidus is clearly an outlier (subgroup A1, sister to the rest of group A), we advocate returning it to its status as the monotypic genus Phaeoptila (it had been lumped into Cynanthus without explanation by Peters). Examining the remaining groupings, a broad Cynanthus includes two coherent subgroups of Chlorostilbon (subgroups A2 and A4), separated by Cynanthus itself (subgroup A3): in effect, Chlorostilbon as a genus is polyphyletic. Subgroup A2 includes three species of the Greater Antilles and Cyanophaia bicolor of the Lesser Antilles, the most divergent in plumage; Subgroup A4 includes the Chlorostilbon species of southern Middle America and South America including the type species, mellisugus. We therefore advocate resurrecting the genus Riccordia for subgroup A2, and including within it Cyanophaia. We ascribe the greater divergence in plumage of bicolor to rapid evolution on isolated small islands: a similar case in the Polytminae is the Lesser Antillean genera Eulampis and Sericotes, which the phylogeny found to be nested within Anthracothorax.

Cynanthus forms a compact generic group A3, the surprise being that nested within it are three to five (depending upon how finely one splits these taxa) species nearly always included in Chlorostilbon because of their plumages. Recognizing these as a separate genus Chloanges is not acceptable, as this would make Cynanthus itself paraphyletic; we therefore include these species in Cynanthus. One conclusion is that the “typical” plumage of Chlorostilbon shows homoplasy; however, another conclusion is that Cynanthus represents a coherent biographical radiation in northern Middle America.

Group B includes species in several genera. Subgroup B1 comprises two Mexican species often included in Hylocharis in the past, leucotis and xantusii. This is clearly untenable because the type species of Hylocharis is in Group D in the phylogeny. We therefore follow several recent authors in placing these species in the genus Basilinna. Subgroup B2 includes two Mexican species of Campylopterus, separated from the rest of this genus by subgroup B3. We therefore advocate resurrecting the generic name Pampa, as used and diagnosed by Ridgway, for these species including as well its type species, curvipennis, not included in the phylogeny but close to (and sometimes lumped with) excellens, thus resolving the apparent polyphyly of Campylopterus.

Subgroup B3 includes five small genera (Klais, Abeillia, Orthorhynchus, Anthocephala and Stephanoxis, all on long branches. One alternative would be to lump all five into Orthorhynchus, the oldest name. A second would be to lump Klais into Abeillia, and the remaining three into Orthorhynchus. However, the lack of morphological or biogeographic coherence among this group leads us to continue recognizing all five genera, which also promotes stability. Subgroup B4 includes the bulk of the genus Campylopterus including its type species largipennis, with species ranging from Middle America through much of South America. Although some of the branch lengths are rather long, we see nothing to be gained by splitting a well-diagnosable genus like Campylopterus into three or four small genera, at least one of which would require a new genus name; we therefore recommend continued recognition of a broad Campylopterus, again preserving stability.

Group C includes only two subgroups. Subgroup C1comprises two clades. The first is a tight group of three species in two genera, Microchera and Elvira, all of which inhabit lower middle elevations of the mountains of Costa Rica and western Panamá. Given the short branch lengths joining them, we consider that all are best considered congeneric; Microchera has priority. Microchera has long been considered monotypic due to the very distinctive male plumage of albocoronata, however the female plumage is quite similar to those of Elvira.

The second clade breaks into two groups: the first comprises the monotypic genera Goldmania and Goethalsia of the Darien highlands of eastern Panama and adjacent Colombia; the second includes Thalurania ridgwayi and the several species of the genus Eupherusa. We see no reason for maintaining two monotypic genera in the former group, and lump Goethalsia into Goldmania, which has priority. The two species are similar in morphology and share an unusual type of undertail coverts, and differ only in color patterns; they show a somewhat leapfrog-like pattern of distribution on isolated mountaintops in the Darien. These two species are adjacent in all recent classifications.

The surprise in the second group is Thalurania ridgwayi, which has always been included in this genus since its description, based upon its green throat and chest, dark abdomen and bright blue-violet crown. However, the genetic data preclude inclusion of ridgwayi in Thalurania; moreover, a closer examination of its plumage reveals previously overlooked similarities in plumage with Eupherusa. Furthermore, its Pacific slope distribution accords much better with that of Eupherusa than that of Thalurania, which extends northward in the Caribbean lowlands to Guatemala and only occupies the Pacific slope from southwestern Costa Rica southwards into South America. Hence, we advocate inclusion of ridgwayi in the genus Eupherusa. The only other option would require naming a new genus for ridgwayi, which we deem unnecessary given its close genetic relationship to Eupherusa.

We now present the following proposals for consideration by SACC. Although several of these are most strictly in the domain of the NACC, we present them here because they do affect the classification of some genera of South America as well.

1. [extralimital --- advisory only]

A. Expand the genus Cynanthus to include Chlorostilbon sensu lato.

B. Separate the species sordida in the genus Phaeoptila; doing so then permits further consideration of the circumscription of Chlorostilbon. We strongly favor this option.

2. [extralimital --- advisory only]

A. Restrict Cynanthus to exclude the canivetii group of species of Chlorostilbon, segregating these in the genus Chloanges.

B. Include the aforementioned species in Cynanthus. We favor this option because option A would render Cynanthus paraphyletic.

3. [extralimital --- advisory only]

A. Retain the Antillean species in Chlorostilbon.

B. Split Chlorostilbon into two genera, reviving the generic name Riccordia for the Antillean species including Cyanophaia, with the second genus including the majority of the species of Chlorostilbon including its type species; nearly all of these species are South American. We strongly favor this option, because option A would produce a polyphyletic Chlorostilbon.

4. [extralimital --- advisory only]

A. Retain the species excellens and its close relatives in the genus Campylopterus.

B. Split Campylopterus into two genera, reviving Pampa for the aforementioned northern species excellens and its relatives of Mexico and Guatemala. We strongly favor this option because option A would render Campylopterus polyphyletic.

5. A. Retain the non-Pampa species of Campylopterus in this genus, which includes its type species.

B. Split the restricted Campylopterus into three or four small genera. In this case, we favor option A, especially as at least one new generic name might be required for option B, and Campylopterus as restricted is well diagnosable.

6. A. Retain generic rank for Orthorhynchus, Abeillia, Klais, Anthocephala and Stephanoxis.

B. Lump all of these genera into Orthorhynchus.

C. Lump the first two genera into Abeillia and the last three into Orthorhynchus. Here we favor option A because all of these genera are separated on long branches, and the because of the lack of morphological or biogeographical concordance between them.

7. [extralimital --- advisory only]

A. Continue to recognize Microchera and Elvira as separate genera.

B. Lump Elvira into Microchera. We favor this option because of biogeographical concordance, short branch lengths and previously overlooked similarities in female plumages.

8. A. Continue to recognize Goethalsia and Goldmania as separate genera.

B. Lump Goethalsia into Goldmania. Here also, we favor this option as the two species are similar morphologically and biogeographically, the difference between them being only coloration, especially of the males.

9. [extralimital --- advisory only]

A. Name a new genus for “Thalurania” ridgwayi, because the genetic data preclude its inclusion in Thalurania.

B. Include ridgwayi in the genus Eupherusa reflecting hitherto overlooked similarities in plumage, biogeographical concordance and genetic proximity. We favor this option.

10. A. Continue to recognize the genera Thalurania and Chalybura.

B. Lump Chalybura into Thalurania, which are sister genera in the phylogeny. We favor option A. Although sharing a few similarities in plumage, these genera are separated on long branches and are well diagnosable from each other.

References (downloads available at https://www.researchgate.net/profile/J_Remsen):

McGuire, J. A., C. C. Witt, J. V. Remsen, Jr., A. Corl, D. L. Rabosky, D. L. Altshuler, & R. Dudley. 2014. Molecular phylogenetics and the diversification of hummingbirds. Current Biology 24: 1-7.

Stiles, F. G., V. de Q. Piacentini, & J. V. Remsen, Jr. 2017. A brief history of the generic classification of the Trochilini (Aves: Trochilidae): the chaos of the past and problems to be resolved. Zootaxa 4269: 396–412.

Stiles, F. G., J. V. Remsen, Jr., & J. A. McGuire. 2017. The generic classification of the Trochilini (Aves: Trochilidae): reconciling classification with phylogeny. Zootaxa 4353: 401-424.

Gary Stiles, March 2018

__________________________________________________________

Comments from Jaramillo: “9 A – Yes, B – No. It seems more informative to put ridgwayi in a new genus.”

Comments from Areta: “I vote YES on the following: 1B, 2B, 3B, 4B, 5A, 6A (However, it may well be the case that lumping all these species into a single genus Orthorhynchus will result in finding deeper homologies than hitherto acknowledged) 7B, 8B, 9B (This is striking, as ridgwayi does look remarkably like traditional Thalurania species! If someone erects a new genus for this, please don´t use “Pseudothalurania”!), 10A

Comments from Claramunt: “1-3 “Group A” is very uniform phenotypically and is not excessively species-rich or old, so I don’t see the rationale for a solution that involves more genera instead of fewer. The resultant genera would not be diagnosable. So, I vote YES for an expanded Cynanthus (Yes to 1A and 2B, and NO to all other alternatives).

“6C. guimeti and abeillei seem to form a superspecies rather than two different genera, so lump into Abeillia. Anthocephala is distinct, but the similarity between Orthorhyncus and Stephanoxis is striking. So, two genera instead of five.

“7A. Microchera is strikingly diagnosable, one of the most distinctive hummingbird genera.

“8B. Yes to lump these two together. I also think that they could be merged into Eupherusa.

10A Yes to maintain Chalybura and Thalurania since there is no phylogenetic reason to change things here. But I don’t strongly oppose a potential lumping.

Comments from Zimmer:

“1B. YES. This would be the logical path based upon the genetic data.

“2B. YES. Chlorostilbon, as currently constituted, is polyphyletic, so something has to be done with the canivetii-group. As Gary points out, isolating the canivetii-group in the genus Chloanges would render Cynanthus paraphyletic, so, including these species in an expanded Cynanthus would appear to be the only solution.

“3B. YES. This is mandated by the genetic data, and results in two biogeographically coherent genera.

“4B. YES. This change is clearly mandated by the genetic data. Retaining excellens & rufus in Campylopterus would render the genus polyphyletic.

“5A. YES. Based upon the length of some of the branch lengths within this species-cluster, one could make a case for recognition of multiple genera. But, when considering morphology, the broader Campylopterus, as seen in Subgroup B4, really seems like a pretty cohesive, readily diagnosable group, so my inclination is to keep them together.

“6A. YES. Any course other than retaining the status quo would result in some groupings that would be hard to defend on either morphological or biogeographical grounds, and looking at the branch lengths separating these genera, I don’t think the genetic data points strongly in that direction anyway.

“7B. YES. Based simply upon the genetic data, we could go either way on this, but lumping them into one genus makes perfect sense on distribution, ecology, biometrics, and female plumage, and thus, is more informative of relationships than would be maintaining the current arrangement, based solely upon the rather different male plumage of the Snowcap.

“8B. YES. This option makes perfect sense on biogeographical and morphological grounds.

“9A. I have to go against the grain on this one. It’s clear from the phylogeny that we can’t continue to treat ridgwayi in Thalurania. But other than a close genetic distance to Eupherusa, I really don’t see it belonging there. Minus ridgwayi, Eupherusa is a tight-knit, coherent group (all still more closely related to one another than to ridgwayi), all members of which have a similar tail pattern and show prominent rufous in the wing. As far as I know, all of them, with the possible exception of cyanophrys, are also more montane (middle-elevation) in their distribution and habitat preferences, as opposed to ridgwayi, which occurs mostly below or at the lower end of the elevational distribution of Eupherusa. I think ridgwayi is simply an oddball, closest genetically to Eupherusa, but morphologically, highly suggestive of Thalurania. The proposal doesn’t spell out what the “hitherto overlooked plumage similarities” of ridgwayi to Eupherusa are, but I am curious, since they aren’t at all evident to me. This strikes me as a very different situation from Microchera and Elvira, where genetic distances were similar to those in the present case, but the only real disparity was in the male plumage of one species, whereas biometrics, female plumage, ecology and distribution were all pointed toward inclusion in a single genus.

“10A. YES. Absolutely. I would be strongly opposed to lumping these two morphologically and ecologically distinctive genera, and given the branch lengths, I don’t think the genetic distances point in that direction either.”