Proposal (1087) to South American Classification Committee

Treat Willisornis poecilinotus as consisting of six species

Background: The current SACC footnote is as follows

60. Pinto (1978) treated lepidonota (with duidae) as a separate species from W. poecilinotus, and Hilty (2003) and Zimmer & Isler (2003) suggested that more than one species was involved; however, see Zimmer (1934d), Willis (1982), and Ridgely & Tudor (1994). Isler & Whitney (2011) presented evidence that the subspecies vidua merits species rank (including also nigrigula). SACC proposal passed to treat vidua as separate species. Pulido-Santacruz et al. (2018) found strong evidence for post-zygotic reproductive isolation between W. poecilinotus and W. vidua in their headwaters contact zone and found strong evidence for barriers to gene flow. Naka and Brumfield (2018) and Silva et al. (2019) found that W. poecilinotus is a paraphyletic species with respect to W. vidua. Quaresma et al. (2022) found that broadly defined W. poecilinotus was paraphyletic with respect to W. vidua; they recommended recognizing six species of Willisornis. SACC proposal needed.

The subspecies are defined primarily on the basis of major differences in female plumage, and subspecies distributions are mostly river-delimited. We have already separated W. vidua from Willisornis poecilinotus, and there have been several suggestions that multiple species are involved. Isler and Whitney (2011) analyzed songs from throughout the Amazon Basin and showed that those of vidua (with nigrigula) differ from those of all other taxa to the degree associated with species limits in antbirds. However, they emphasized that gaps in their sampling, especially in headwater contact zones, left the status of some taxa not fully resolved.

The distribution of the subspecies is as follows (from Quaresma et al. 2022; note that five of the seven subspecies were subdivided as separate sampling units).

New information:

GENETICS

Pulido-Santacruz et al. (2018) did a densely sampled transect (n=144) through the headwaters region of the Teles Pires and Xingu rivers (along with another 70 individuals from elsewhere in Amazonia) and found (using SNPs) that although W. p. griseiventris and W. vidua showed no sign of assortative mating that they also showed no signs of gene flow between them, i.e., evidence of almost complete post-zygotic reproductive isolation. This is the sort of result that those of us focused on mate choice and assortative mating need to process in our thinking about the BSC. One interpretation is that natural selection has not yet produced a prezygotic isolating mechanism despite vocal and female plumage differences.

Naka and Brumfield (2018; Fig. S39) used mtDNA sequence data in their broad study of differentiation across the Rio Branco/Rio Negro region, and found the following relationships, which showed that vidua (see “S Amazonia” section of the tree) is embedded within poecilinota, although note the weak support for the sister relationship of vidua and nominate poecilinotus.

Silva et al.’s (2019; Fig. S2 below) analysis of mtDNA and nDNA sequence data also showed that W. poecilinotus as currently defined is paraphyletic, with the southwestern Amazonian subspecies griseiventris closer to W. vidua than either is to nominate poecilinotus from the Guianas.

Harvey et al. (2020) included W. vidua and four samples of W. poecilinotus in their genomic analysis. The taxa in their tree are as follows: LSU16954 is from Pará (vidua); INPA “2032” (=2031) is from the Jaú, Amazonas, Brazil (duidae); L55367 is from Suriname (nominate poecilinotus); LSU 8861 is from Pando (griseiventris), Bolivia; MCP 2362 is from Rôndonia (griseiventris).

Note that this conflicts with the results from Naka and Brumfield and Silva et al. in that vidua is sister to all other Willisornis, most importantly including duidae. Therefore, although individual and taxon sampling is much weaker in Harvey et al., the much broader genetic sampling in Harvey et al. suggests that the results using mtDNA in the other studies are driven by incomplete lineage sorting or hybridization, and the case for using paraphyly as evidence for further species splits is weakened.

Quaresma et al. (2022) performed the most comprehensive genetic study of Willisornis with 182 samples from throughout the range of the genus – see their Fig. 1 above. Two mitochondrial (cytb, ND2) and two nuclear (BF5, MUSK) genes were sequenced. They identified 13 reciprocally monophyletic lineages, and the relationships among them were as follows:

PLUMAGE

Quaresma et al. assessed several aspects of male and female plumage of over 1000 specimens, and found that each taxon is diagnosable based on plumage, at least for females, although whether lepidonota is diagnosable from duidae is an open question. I will add photos of female specimens here soon. The most important point is that in some respects females of W. vidua and W. p. griseiventris are more similar to each other, at least superficially, than either is to females of the other subspecies in W. poecilinotus; they have plain wings and gray bellies, whereas all the others have heavily marked wings and tawny bellies, as in W. p. lepidonota shown below, or if gray bellies, then heavily marked wings, as in nominate poecilinotus below; however, note that vidua and griseiventris differ strongly in face pattern and back color, so I consider the resemblance superficial, and in a touchy-feely way, vidua to me appears to be the outlier.

Left: griseiventris female by Joe Tobias. Right: vidua female by Beto Guido Méndez

Left: female lepidonota by Matthew Bruce. Right nominate poecilinotus by Hugo Foxonet

Here is Hilary Burn’s plate as reproduced in Isler and Whitney (2011):

VOCALIZATIONS

Isler and Whitney (2011) showed that vidua was the outlier in terms of voice, and that was the rationale for its treatment as a separate species, and that variation among all the other taxa was minimal. See SACC 495 for full details – no need to repeat here.

TAXONOMY

Quaresma et al.’s proposed new taxonomy is as follows, with 5 monotypic species and 1 polytypic species:

Willisornis poecilinotus (Cabanis, 1847) – Guianan shield region

Willisornis gutturalis (Snethlage, 1914) – limited area in n. Peru and w. Brazil – see map above

Willisornis vidua (Hellmayr, 1905) – southeastern Amazonia east of the Xingu

Willisornis nigrigula (Snethlage, 1914) – southern Amazonia – see map above

Willisornis griseiventris (van Pelzeln, 1869) – southwestern Amazonia – see map above

Willisornis lepidonota (Sclater & Salvin, 1880 – western Amazonia – see map above

W. l. lepidonota

W. l. duidae

Discussion and recommendation: Quaresma et al.’s paper is excellent, with a massive amount of data and many analyses. This is a very complex system with 13 distinct lineages and 7 named taxa, with varying amounts of gene flow between some of them. Plumage differences among the seven taxa range from dramatic (within an antbird context) to subtle. This is another Amazonian “pie” (as in Cercomacra cinerea - -see my general comments in SACC 1085) that is difficult to force into Linnaean nomenclature. I can state at the outset that there is no single taxonomic solution that I can see that will not be disputed. It’s complicated even trying to devise a voting scheme that will produce a logical outcome. I apologize in advance to the authors if I have misconstrued anything or botched their interpretations.

In trying to sort all this out, here are the key points in my view. The material in quotes is from Quaresma et al. unless indicated otherwise. Please feel free to add, subtract, and correct:

• Isler and Whitney established that vidua should be treated as a species based on the Isler-Whitney framework of antbird vocal differences and species limits. Also, subjectively, the female plumage is the most distinctive of the lot, in my opinion.

• The issue of paraphyly of our current polytypic W. poecilinotus is unresolved in my opinion because the paraphyly is probably due primarily to mitochondrial gene trees, although I am uncertain of this without seeing separate trees for nDNA alone. Also, Harvey et al.’s much more comprehensive genetic sampling shows that among the taxa that they sampled, poecilinotus was indeed monophyletic. To this, Quaresma et al.’ response was “even though [Harvey et al.] was based on the largest molecular dataset available for Willisornis and several other suboscine lineages, it lacked samples of the genetically divergent W. p. duidae, W. p. gutturalis, W. p. lepidonota and W. v. nigrigula, and, therefore, a sparser taxon sampling could have influenced the recovered topologies.” But unless I’ve botched the localities, duidae was sampled by Harvey et al. (Jaú, Amazonas), and whether inclusion of the missing taxa would affect the internal topology I leave to the tree-builders.

• poecilinotus and duidae have two contact zones, but there is no evidence for widespread intergradation and only a couple of putative hybrids. If indeed there is no hybrid zone despite putative direct contact, then these two have to be treated as separate species.

• gutturalis and griseiventris meet somewhere along the R. Juruá, “where at least some limited hybridization occurs (Fig. 5; Supporting Information, Fig. S4).”

• vidua vs. nigrigula: “Our data did not support the existence of a contact zone between W. vidua and W. nigrigula in the upper Xingu River (Fig. 1; Supporting Information, Fig. S1), so it remains speculative whether these taxa meet in the sector of the Xingu–Tocantins interfluve (Isler & Whitney, 2011)”.

• There might be an unsampled contact zone between vidua and griseiventris in the upper Xingu.

• nigrigula vs. griseiventris: “Both of these populations are in direct contact with W. griseiventris in the middle sector of the Tapajós–Madeira interfluve and the upper reaches of the Tapajós and Xingu rivers, where they hybridize (Supporting Information, Fig. S5), although selection against highly introgressed hybrids indicates the existence of post-mating reproductive barriers between them, causing their hybrid zone to be extremely narrow, at least in upper Tapajós drainage (Pulido-Santacruz et al., 2018).”

• on a personal note, I continue to object to the use of “reciprocal monophyly” as an absolute term. It’s contextual. It depends on the number of individuals sampled and how close they are geographically to known or potential contact zones. Any statement about two populations being reciprocally monophyletic is one just one additional sample from being falsified.

• The work of Quaresma et al. strengthens the statement made by Isler and Whitney 15 years ago, namely: “Six pairs of subspecies are apparently parapatric and lack a known barrier to intergradation in at least a portion of their contact zone; yet confirmed hybrids are known only for one pair in one location.” See Isler and Whitney (2011) for their analysis of each contact zone.

So, what is the classification that best reflects all this? Despite the lack of strong vocal differences, to me the absence of any hybrid swarms is any of the contact zones indicates that there is no “free interbreeding” between any of the proposed species-level taxa in Quaresma et al. As with Cercomacra cinerascens and the Rhegmatorhina berlepschi-hoffmannsi case, something besides voice seems to be acting as an isolating mechanism. Personally, I regard parapatry without evidence of free gene flow to be better evidence for species rank than any other criterion. The most detailed sampling of any of the contact zones (vidua vs. griseiventris; Pulido-Santacruz et al. 2018) suggests strong post-zygotic isolation between the two despite no evidence for voice being an isolating mechanism. Although one can’t extrapolate cleanly from that case to the other contact zones because this is the only one in which there are vocal differences, the phenotypic signal from all the other contact zones is similar, i.e. no hybrid swarm. Some of that might be an artifact of limited sampling, but we have to go on what we know at the moment. So, and it wasn’t until after I had assembled all this work that I had reached this conclusion, I recommend that we follow Quaresma et al.’s six species classification until additional data say otherwise. Congrats to the Aleixo Lab for producing such a volume of critical data.

As for voting, let’s try this scheme. Vote YES/NO on each one – you only get one YES!

Option A. Treat all taxa as subspecies of one big polytypic species, thus retreating to the classification used by Meyer de Schauensee (1970) and others. This would eliminate the apparent problem of paraphyly.

Option B. Maintain status quo, i.e. two species, with vidua retained as a separate species because the vocal data and the most detailed sampling of the contact zone indicate that at least two species are involved, but wait for additional, more detailed sampling of the other contact zones.

Option C. Adopt the six species classification of Quaresma et al. as the best classification given current available information, while acknowledging that additional data from the contact zones might cause future revisions in some cases.

Option D. Some other combination. Pick and choose from the various remaining combinations (and provide rationale).

English names: Obviously, there is no point in acting on this until we see if and how species limits are changed, although retaining Scale-backed Antbird as the group name seems a wise course.

References:

HARVEY, M. G., G. A. BRAVO, S. CLARAMUNT, A. M CUERVO, G. E. DERRYBERRY, J. BATTILANA, G. F. SEEHOLZER, J. S. MCKAY, B. C. O’MEARA, B. G. FAIRCLOTH, S. V. EDWARDS, J. PÉREZ-EMÁN, R. G. MOYLE, F. H. SHEDLON, A. ALEIXO, B. T. SMITH, R. T. CHESSER, L. F. SILVEIRA, J. CRACRAFT, R. T. BRUMFIELD, AND E. P. DERRYBERRY. 2020. The evolution of a tropical biodiversity hotspot. Science 370: 1343-1348.

ISLER, M. L., AND B. M. WHITNEY. 2011. Species limits in antbirds (Thamnophilidae): the Scale-backed Antbird (Willisornis poecilinotus) complex. Wilson J. Ornithology 123: 1–14.

NAKA, L. N. AND R. T. BRUMFIELD. 2018. The dual role of Amazonian rivers in the generation and maintenance of avian diversity. Science Advances 2018 (4):eaar8575.

PULIDO-SANTACRUZ, P., A. ALEIXO, AND J. T. WEIR. 2018. Morphologically cryptic Amazonian bird species pairs exhibit strong postzygotic reproductive isolation. Proceedings of the Royal Society B 285: 20172081.

QUARESMA, T. F., A. A. CRONEMBERGER, R. BATISTA, AND A. ALEIXO. 2022. Diversification and species limits in scale-backed antbirds (Willisornis: Thamnophilidae), an Amazonian endemic lineage. Zoological J. Linnean Society 196: 1408–1430.

SILVA, S. M. A. T. PETERSON, L. CARNEIRO, T. C. T. BURLAMAQUI, C. C. RIBAS, T SOUSA-NEVES, L. S. MIRANDA, A. M. FERNANDES, F. M. D'HORTA, L. E. ARAÚJO-SILVA, R. BATISTA2, C. H. M. M. BANDEIRA, S. M. DANTAS, M. FERREIRA, D. M. MARTINS, J. OLIVEIRA, T. C. ROCHA, C H. SARDELLI, G THOM, P. SENA RÊGO, M. P. SANTOS, F. SEQUEIRA, M. VALLINOTO, AND A. ALEIXO. 2018. A dynamic continental moisture gradient drove Amazonian bird diversification. Science Advances 2019 (5)eaat5752.

Van Remsen, July 2026

________________________________________________________________________________________________________________

Vote tracking chart:

https://www.museum.lsu.edu/~Remsen/SACCPropChart1044+.htm