Proposal (884x) to South American Classification Committee

Expand Percnostola to include Hafferia and Akletos

With the publication of a massive new data set (Harvey et al. 2020), expansion of Percnostola to include Hafferia and Akletos would make that genus paraphyletic with respect to Gymnocichla. Here is a screen shot of the relevant portion of their tree, which shows that Akletos is sister to Gymnocichla (node in upper left corner) and that Hafferia is sister to Percnostola.

Therefore, to merge Hafferia and Akletos into Percnostola would require merger of all three into Gymnocichla (which has priority), and that would produce a highly heterogenous genus with a node deeper in the tree than many or most antbird genera. So, I strongly recommend a NO vote on this quick revision of the proposal. See also comments from Gustavo Bravo in the Comments on the previous version.

Voting options:

YES = Hafferia and Akletos merged into Percnostola

NO = Hafferia and Akletos retained.

Expansion to include all genera in Gymnocichla would require a new, separate proposal that would explore in detail the problem with respect to node depth, over all tree topology, and phenotypic diversity.

Van Remsen, May 2021

Comments from Jaramillo: “NO. I am hesitant to create these large genera that are heterogenous. There is likely a better intermediate option, but until that idea is brought forward, the above is my vote. Also I am taking into account the notes given by Bravo in the associated proposal.”

Comments from Stiles: “NO, in anticipation of future events/proposal(s).”

Comments from Lane: “NO. Given Gustavo Bravo's comments and the need for a new proposal if we include Gymnocichla, I will change my vote to NO. In any event, the branch nodes may be deep enough that these clades are too old to consider congeneric?”

Comments from Claramunt: NO. Hafferia and Akletos should not be merged into Percnostola given the position of Gymnocichla.”

Comments from Pacheco: “NO. In view of Gustavo's comments, in the previous proposal, I understand that a highly heterogeneous genus is not the best arrangement for now. It is opportune to change my vote to no.”

Comments from Zimmer: “NO. Based upon the publication of the new data set by Harvey et al. 2020, and the additional comments by Gustavo Bravo, I change my vote on this to NO. I was willing to live with this change before we knew that the resulting grouping would be paraphyletic with respect to Gymnocichla, and that we would have to fold all three genera into Gymnocichla, which has priority. I’ve never been a fan of overly heterogeneous genera, and this proposed one (expanded Percnostola) was already verging on borderline for my taste without rolling them all into Gymnocichla. I would consider that move a bridge too far. I would prefer to keep all 4 genera separate until such time as conflicting new data comes along.”

Comments from Robbins: “NO, for reasons stated in the proposal.”

Comments from Areta: I vote YES to a broad Percnostola. I understand Gustavo´s comments, but data are pretty much in agreement in showing that Gymnocichla, Percnostola, Hafferia, and Akletos are part of the same clade (we also don´t know how the different analyses differed). In looking at the divergence times of numerous other antbird genera in the Harvey et al. 2020 tree, I think that even a broad Pyriglena is within reason. The Percnostola, Akletos, and Myrmeciza immaculata that I am familiar with, have always felt much like the Pyriglena in terms of behavior and vocalizations to me. Take a look at genera such as Myrmoborus, Dysithamnus, Myrmelastes, Hylophylax, Drymophila, Cercomacra, and many others, and you might also start feeling that the "Pyriglena" clade is being oversplit at the generic level. So, either we need to multiply, say, by 4, the number of genera in the Thamnophilidae or we should give some serious consideration to merging some of the genera in this clade. To me, a broad Percnostola is fine, and I could also accept a broad Pyriglena. But a proliferation of genera would be my last option.”

Additional comments from Stiles: I will support a YES for the broad Percnostola.

“

Proposal (884) to South American Classification Committee

Expand Percnostola to include Hafferia and Akletos

Effect on SACC: This would subsume two genera of antbirds into the genus Percnostola.

Background. This is a “do-over” of SACC 628G. That proposal favored recognition of Hafferia (for fortis, zeledoni, immaculata) and Akletos (for melanoceps and goeldii), but as Terry Chesser pointed out to me, the vote tally did not match what we implemented, and even so, none of the options attained a 2/3 majority. This error was my fault, and so I hereby re-open the question with a new proposal.

Please consult SACC 628G for all the background material. Below are the trees from the Isler et al. publications.

Figure 1. Maximum-likelihood tree of a subset of the Thamnophilinae (Myrmeciza species are emboldened). The color of the circles at nodes indicates bootstrap support values, > 70% (black), 50-70% (gray), < 50% (white). This figure is Figure 1 from Isler et al. (2013):

Figure 2. Time-calibrated tree showing relative ages of former members of Myrmeciza. Bars at nodes indicate the 95% highest posterior density for the inferred divergence time estimates. This is Figure 3 from Isler et al. (2013) with the Akletos revision.

Below is the rationale by Mort Isler, Gustavo Bravo, and Robb Brumfield in favor of three genera.

“Proposal 628G (option 1). Maintain Percnostola and recognize the genera Akletos and Hafferia. Traits distinguishing genera other than Myrmeciza in the final clade are fewer than the preceding clades and almost entirely confined to morphology, yet the clade includes well-established genera such as Myrmoborus and Pyriglena. The principal challenge involves the Percnostola, immaculata, and melanoceps subclades ("G" and "H" in Figure 1). A sister relationship between any of them is not strongly supported so we currently have a polytomy. Future analysis with additional data can either support: (a) a sister relationship between Hafferia and Akletos or (b) a sister relationship between Percnostola and either Hafferia or Akletos. Given the uncertainties, we wanted to avoid lumping Akletos and Hafferia into one genus that could become paraphyletic. Our preferred options are either to treat them as separate genera (628G option 1) or to merge them into Percnostola (628G option 3). With either of these, no matter what the final topology will be, no taxon can become non-monophyletic. On the other hand, if Percnostola and Hafferia + Akletos are placed into two genera (628 option 2), scenario (b) would render a paraphyletic genus. We recognize that, despite this possibility, option 2 might be considered a more conservative approach. Given the choice of placing them in the same genus as the morphologically dissimilar Percnostola or recommending three genera, the latter was deemed consistent with the morphological distinctions currently distinguishing other genera in the Pyriglenini. Note that subsequent to publication, it was brought to our attention that the proposed name for one of the recommended genera, Inundicola, was a junior synonym of Akletos, and a correction has been made (Isler et al. 2014).”

And here is their rationale opposing a broad Percnostola:

“Proposal 628G (option 3). Recognize the genus Percnostola for P. rufifrons, P. arenarum, Myrmeciza melanoceps, M. goeldii, M. fortis, M. zeledoni, and M. immaculata. Not recommended. This option creates a genus that includes species whose traits are inconsistent as compared to other thamnophilid genera.”

In the Comments section, Thomas Donegan provided an extensive review of the characters that do or do not distinguish these genera and presented rationale for Option 3. Several committee members favored Option 3 or even Option 5, which would have included all three of these genera in a broadly defined Pyriglena. Most voters did not strongly support any of the options.

Analysis and recommendation: I now personally favor Option 3. I think Donegan had some good points on the phenotypic characters that are shared among a broader Percnostola. I also think that the depth of the node that unites Percnostola, Hafferia, and Akletos is more consistent with recognizing the taxa that it unites as belonging to a single genus – they are the two youngest genera in the tree.

References.

Isler, M. L., G. A. Bravo, and R. T. Brumfield. 2013. Taxonomic revision of Myrmeciza (Aves: Passeriformes: Thamnophilidae) into 12 genera based on phylogenetic, morphological, behavioral, and ecological data. Zootaxa 3717 (4): 469–497.

Isler, M. L., G. A. Bravo, and R. T. Brumfield. 2014. Inundicola Bravo, Isler, and Brumfield 2013 is a junior synonym of Akletos Dunajewski 1948 (Aves: Passeriformes: Thamnophilidae). Zootaxa 3779 (3): 399–400.

Van Remsen, September 2020

Comments from Stiles: “Especially after rereading Donegan's piece and noting the depth of the nodes in question, I agree that option 3 (uniting Akletos and Hafferia under Percnostola) is the best option.”

Comments from Areta: “YES. I think that merging all in Percnostola is a good option. In my original vote, I also mentioned that putting all in Pyriglena is also appealing to me, based on plumage and vocalizations. Both options satisfy me.”

Comments from Claramunt: “YES. I think it makes sense to merge these three genera. In particular, Hafferia and Akletos species are very similar and would fit in the same genus comfortably. The two Percnostola (sensu stricto) species are somewhat different but overall, I favor an expanded Percnostola over an atomized classification with 3 genera and very few species in each.”

Comments from Robbins: “YES to merging Hafferia and Akletos into Percnostola, and I agree with Nacho, I would be fine with placing all three of these in Pyriglena.”

Comments from Zimmer: “YES. As Isler et al. made clear in Proposal #682G, this option would at least ensure that paraphyly is avoided. Donegan’s points regarding the lack of diagnosable characters distinguishing Percnostola, Hafferia, and Akletos are well taken – eye color goes out the window when you consider the gray-eyed minor subspecies of P. rufifrons, and most of the other characters are even less consistent or impressive – the three genera, as currently constituted, are probably better regarded as representing three different superspecies. And, if we could live with the Schistocichla group being considered part of Percnostola (as we did for a long time), we should certainly be able to welcome Hafferia and Akletos into the fold. However, unlike some others on the committee, I would balk at extending this logic toward the end of folding these three genera + Gymnocichla into an expanded Pyriglena. It would be defensible on genetic grounds, but looking for the most inclusive and most heterogeneous monophyletic group possible is antithetical to my concept of what a genus should be. None of the species currently included in Percnostola, Hafferia or Akletos approaches the complexity of the vocal repertoire found in any of the species of Pyriglena, all of which have equally complex social structures centered related to their habitual ant-following habits. And, although an approach to Pyriglena can be seen as regards group congregation over ants, vocal complexity, and stereotypical tail movements within both Gymnocichla and immaculata + zeledoni, each of these three taxa (not to mention fortis, melanoceps and goeldii) has obvious morphological distinctions that would make them outliers with respect to Pyriglena. Pyriglena, as currently constituted, is very homogeneous with respect to morphometrics, eye color, presence and prominence of interscapular patches, male plumage, voice (across multiple different types of calls as well as loudsongs), social structure, and overall ecology. The various species differ most obviously in the female plumage, and even there, all of the various taxa are more similar to one another than any of them are to females of nudiceps, fortis, melanoceps, goeldii, immaculata, zeledoni, rufifrons, or arenarum.”

Comments from Lane: “YES. Since we are still not yet ‘settled’ in a relatively final taxonomy of the Thamnophilidae after what has been nearly Earth-shaking change, why not make last tweaks before it is set and dried? Just by going over vocalizations, I hear enough similarity between members of Percnostola (sensu stricto) and Akletos to see that these two are quite similar. Hafferia is a bit more distinct, but the bulk and shape of its members are similar enough to Akletos that it doesn't result in a particularly heterogeneous assemblage... and I appreciate having to memorize fewer generic names!”

Comments from Jaramillo: “YES – The differences seem marginal between these three taxa, it makes sense to lump them under Percnostola. But to retain Pyriglena as separate.”

Comments from Pacheco: “YES. It seems that lumping Percnostola, Hafferia or Akletos in a single genus as a good solution.”

Comments from Remsen: “NO. The just-published phylogeny by Harvey et al. (2020 Science) indicates that Gymnocichla is part of this group thus forcing us to reconsider.”

Additional comments from Robbins: “NO. I agree that the proposal should be modified (or create a new one) to include the new molecular data.”

Comments from Gustavo Bravo: "I read carefully this proposal to expand Percnostola to include Hafferia and Akletos and I decided to contribute my two cents. I'd like to start by highlighting that I still favor the treatment that Mort Isler, Robb, and I put forward in 2013, whereby those genera are best treated separately. The problem with the proposed treatment – Percnostola + Hafferia + Akletos in an expanded Percnostola – is that different types of loci and different types of analyses yield different topologies, likely suggesting that our available data and phylogenetic tools are not capturing fully the historical complexity at the base of that clade. In Harvey, Bravo et al (2020), we basically conducted two kinds of analyses – concatenation vs coalescence – and they produced conflicting results using the same data in this part of the phylogeny. These two methodological frameworks have fundamental differences in their assumptions and how they treat data (not dwelling into those details here), and such differences must not be taken lightly if we are aiming at having a stable classification that reflects evolutionary history to the best of our capacities. I personally think that coalescent-based analyses are more robust – which by the way do not conflict with this proposal – but given the incongruence, I think that we need further analyses with a more restricted sampling aiming at resolving the base of this clade. I know that this requires extra efforts causing delays in taxonomic decisions, but that is simply how things go: The more data, the more heterogeneous and conflicting signals.

“Finally, I'd like to send a cautionary note regarding similar cases in our suboscine analyses. In Harvey, Bravo et al (2020) we chose to use the concatenation-based tree for downstream analyses and to go in Figure 1 due to its more stable structure toward the base of the tree and the way it behaved with samples with high proportions of missing data, such as toe-pads. However, it doesn't necessarily mean that it is the "true" phylogeny– whatever that means – or even better than the coalescent-based tree. Hence, whenever conflicts appear, taxonomic decisions must not be made solely based on a single tree. I am aware of many more cases like Percnostola's across suboscines. and my advice is to move forward with caution when dealing with them. As I mentioned above, ideally, I'd like to see analyses targeted at those specific taxa before making taxonomic decisions, and that's the way I am moving forward in producing a taxonomic classification of the Thamnophilidae."

Additional comment from Claramunt: “I totally agree with Nacho in that a broad genus make more sense in this group, but we need to deal with Gymnocichla somehow. I think Gustavo is right: we should wait for phylogenetic analyses that focus on this particular question.”