Treat Rhegmatorhina berlepschi (Harlequin Antbird) as conspecific with R. hoffmannsi (White-breasted Antbird)

Proposal (1024) to South American Classification Committee

Treat Rhegmatorhina berlepschi (Harlequin Antbird) as conspecific with R. hoffmannsi (White-breasted Antbird)

Background:

The genus Rhegmatorhina currently comprises five mostly allopatric species, historically treated as separate species taxa based on diagnostic differences in plumage (Cory and Hellmayr 1924, Zimmer and Isler 2003). Most are similar in behavior, and some taxa considered separate species, such as R. berlepschi and R. hoffmannsi, are undifferentiated vocally (Willis 1969). Although differences in vocalizations are considered the major phenotypic marker for species differences in antbirds, plumage differences may be important in antbirds with ornamented plumages or conspicuous, colorful bare skin patches, such as Rhegmatorhina (Skutch 1996, Zimmer and Isler 2003). Genetic work documented structured mitochondrial genotypes congruent with taxonomic boundaries, but shallowly diverged nuclear genotypes (Ribas et al. 2018, Harvey et al. 2020). The classification of the genus into five species seemed reasonable given the information we had until now.

New information:

Recent genetic research revealed extensive nuclear introgression between R. berlepschi and R. hoffmannsi across hundreds of kilometers (Del-Rio et al. 2022). The data is mostly consistent with neutral diffusion of their nuclear genomes, although narrower mitochondrial clines imply some selection maintaining mitochondrial differentiation (Del-Rio et al. 2022). That berlepschi and hoffmannsi hybridize extensively despite differences in plumage (Del-Rio et al. 2022) indicates that their plumage differences have little (if any) impact on reproductive isolation. In summary, this data indicates weak overall reproductive isolation between berlepschi and hoffmannsi.

Recommendation:

Given the evidence of free nuclear gene flow between berlepschi and hoffmannsi (Del-Rio et al. 2022), their taxonomic status warrants reconsideration. The hybrid zone between these taxa is significantly wider than those between many bird lineages considered conspecific (Price 2008). The correct taxonomic treatment seems straightforward in light of the new evidence: berlepschi and hoffmannsi are weakly reproductively isolated and, therefore, should be treated as conspecific. The name hoffmannsi (Hellmayr, 1907) has priority over berlepschi Snethlage, 1907, by some months.

References

Cory, C. B., and C. E. Hellmayr (1924). Catalogue of birds of the Americas and the adjacent islands in Field Museum of Natural History. Volume XIII, Part III. Field Museum Press.

Del-Rio, G., M. A. Rego, B. M. Whitney, F. Schunck, L. F. Silveira, B. C. Faircloth, and R. T. Brumfield (2022). Displaced clines in an avian hybrid zone (Thamnophilidae: Rhegmatorhina) within an Amazonian interfluve. Evolution 76:455–475.

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. C. Faircloth, S. V Edwards, et al. (2020). The evolution of a tropical biodiversity hotspot. Science 370:1343–1348.

Price, T. (2008). Speciation in birds. Roberts & Company.

Ribas, C. C., A. Aleixo, C. Gubili, F. M. d’Horta, R. T. Brumfield, and J. Cracraft (2018). Biogeography and diversification of Rhegmatorhina (Aves: Thamnophilidae): Implications for the evolution of Amazonian landscapes during the Quaternary. Journal of Biogeography 45:917–928.

Skutch, A. F. (1996). Antbirds and ovenbirds: their lives and homes. University of Texas Press.

Willis, E. O. (1969). On the behavior of five species of Rhegmatorhina, ant-following antbirds of the Amazon basin. The Wilson Bulletin 81:363–395.

Zimmer, K. J., and M. L. Isler (2003). Family Thamnophilidae (Typical Antbirds). In Handbook of the birds of the world (J. del Hoyo, A. Elliot and D. A. Christie, Editors). Lynx Edicions, pp. 448–681.

Rafael Lima, July 2024

Comments from Stiles: “YES. This is an easy choice, with a solid publication based upon abundant and appropriate data - in contrast to all those recent ones to evaluate splits based on fragmentary, incomplete and often conflicting pieces of data!”

Comments from Lane: “NO. Glaucia’s study is an intriguing one, but even she didn’t propose lumping the two taxa because of the hybrid zone. The hybrid zone seems to be moving, and doesn’t cover a large area (compared to the Myioborus case in Prop 1003, for example), suggesting that it is not a static system, but rather a dynamic one that may end with one form swamping out the other.”

Comments from Robbins: “NO. As we have seen over and over, mitochondrial data has limited utility and can mislead interpretations of taxonomy. Given the nuclear data coupled with the reported lack of differentiation in primary vocalizations, this seems to be rather straightforward in that these two should be considered conspecific.”

Comments from Areta: “NO. I was waiting for Glaucia´s input on this. The genetic situation is not as simple as described in the proposal, and there are several nuances discussed in the meaningful paper by Glaucia and collaborators. Meanwhile, in the face of a proper study measuring vocalizations of Rhegmatorhina, I decided to take a listen. The vocal similarities between hoffmannsi and berlepschi are striking and obvious even without a quantitative study (and in agreement with the broad hybridization area found), but what caught my attention are the also very similar vocalizations of gymnops (east of the Tapajós): songs and the burry/churring calls sound pretty much identical. Glaucia´s work has also shown a shallow divergence between gymnops and hoffmannsi/berlepschi. Even if we know of no intergrades between berlepschi and gymnops (or is there any evidence for this?), the seeming lack of vocal differences in their primary songs, minor plumage differences (akin to those between berlepschi and hoffmannsi), and shallow genetic divergence (see haplotype network in Del-Rio et al. 2022) all would seem to indicate that, if east-Tapajós gymnops is given the chance to interact reproductively with west-Tapajós berlepschi, they would hybridze freely. Then I remembered a possibly relevant work and found it: Weir et al. (2015; Evolution 69-7: 1823–1834) have found hybridization close to the Tapajós headwaters (the black star in the figure shows a hybrid between "hoffmannsi" and gymnops, I use quotation marks as it is not clear to me how these birds looked like and what would happen with phenotypes in that area):

“Weir et al. (2015) also add that "Mitochondrial sequence based estimates of gene coalescence times between parental populations from the Rondonia and Tapajos endemism centers ranged from just 0.7 Ma in L. iris / L. nattereri and R. hoffmannsi / R. gymnops to 4.1 Ma in W. poecilinota / W. vidua (Fig. 4)." and "Ironically, the taxon pairs with the most clearly differentiated plumage patterns, and which have always been recognized as specifically distinct (R. hoffmannsi/R. gymnops; L. nattereri/L. iris), are also the youngest (<1 Ma) in our sample."

The discussion of Willis (1969) provides interesting perspectives and somehow prepared the scenario for the modern findings: "The forms gymnops, berlepschi, and hoffmannsi are practically identical in behavior. Oddly, berlepschi looks like a hybrid between gymnops to the east and hoffmannsi to the west." And more:

“Finally, note that gymnops was described by Ridgway in 1888, so that name would have priority over berlepschi and hoffmannsi. It seems to me that a 3-way lump would be a reasonable treatment, but I would like to hear from those with extensive Amazonian experience (I am eagerly waiting for Bret´s and Kevin´s surely encyclopedic experience on this group). Even if at present the Tapajós is impassable, I would like to see a deeper genetic and vocal study in order to change our current taxonomic treatment. It seems that the current 3-species treatment of the east-Madeiran Rhegmatorhina is wrong, but how many species should be recognised is debatable.”

Comments from Robb Brumfield (who has Remsen vote): “NO. Hybridization between hoffmannsi and berlespschi is clearly extensive. The center of the hybrid zone is composed almost entirely of recombinant individuals, and introgression of alleles extends deep into the distributions of the two hybridizing taxa. These patterns are typical of most avian hybrid zones for which we have genome-scale data. However, Del-Rio et al (2021) found five diagnostic SNPs that fully distinguish the two taxa and, using those, identified two individuals from the hybrid zone as putative F1s. Both individuals were heterozygous at all five diagnostic nuclear SNPs, and both had intermediate plumage. The presence of these F1s is atypical of most avian hybrid zones. The two F1s suggest that strong reproductive barriers exist between the two taxa in some parts of their nuclear genomes. There are of course nuances to hybrid zone models that can change interpretations based on assumptions, but it may be that strong selection on a few nuclear loci is sufficient to maintain nuclear divergence between hoffmannsi and berlepschi. Whether the presence of strong reproductive barriers at a few nuclear loci is sufficient evidence for species status is unclear to me. Weir et al. (2015, Evolution 69:1823) found evidence of hybridization between R. hoffmannsi and R. gymnops where they come into contact. My take is that Rhegmatorhina is reminiscent of Manacus manakins, in that narrow hybrid zones occur wherever species come into contact. Based on the data in Del-Rio et al. (2015) and the fact that Rhegmatorhina hybridization and speciation is the focus of ongoing research, my vote is to not make any taxonomic changes now.”

Additional comments from Rafael Lima: “I agree with Areta that a three-way lump seems like the most reasonable approach, making gymnops the oldest name. That option initially escaped my attention, and I thank Juan for pointing it out.

“I really appreciate Robb’s thoughtful input on this matter and would like to offer some additional reflections based on his comments.

“Robb mentions that Rhegmatorhina hybrid zones are reminiscent of those in Manacus, where narrow hybrid zones occur wherever species come into contact. However, I am not entirely convinced that the hybrid zone between hoffmannsi and berlepschi can be accurately described as "narrow." Of course, the terms "narrow" and "broad" are subjective, but if the Rhegmatorhina hybrid zone is considered narrow, I wonder what a "broad" hybrid zone would look like. The hybrid zone between hoffmannsi and berlepschi is, in fact, wider than many hybrid zones typically referred to as "wide" in the literature, including those between bird lineages considered conspecific (e.g., Price 2008).

“I completely agree with Robb's point regarding whether strong reproductive barriers at a few loci should suffice for species status when there is otherwise free gene flow throughout the genome. I don’t have an answer for that question. However, regardless of what the preferred cutoff should be, I believe this case may set a precedent for future proposals. In the interest of consistency, those voting "no" on this proposal might need to consider whether they would also support splitting the many bird taxa currently classified as subspecies that have similar hybrid zones and reproductive isolation levels. A "no" vote might imply that bird lineages forming hybrid zones several hundred kilometers wide, dominated by late-generation hybrids, should be considered separate species under the BSC because of reproductive barriers at a few loci amid widespread introgression elsewhere across the genome.

“I wonder why two vocally undifferentiated antbird lineages that hybridize extensively over hundreds of kilometers (i.e., berlepschi and hoffmannsi) should be regarded as separate biological species while other similar antbird lineage pairs should be considered conspecific. For instance, I recall the Thamnophilus capistratus proposal (Proposal 890). Why should Thamnophilus doliatus capistratus and the rest of Thamnophilus doliatus be considered conspecific when they may have a hybrid zone less than 100 km wide, while Rhegmatorhina taxa with substantially wider hybrid zones should be considerate separate species? The Thamnophilus capistratus proposal is merely an example that came to mind, but there are numerous similar cases within the Thamnophilidae. While I no longer think that capistratus should necessarily be treated as a separate species, I do believe that consistent application of criteria is crucial. If criteria are to be applied consistently across taxa, we may need to reconsider species limits in many antbird taxa based on the precedent potentially set by Rhegmatorhina.

“As a final note, I personally view species limits as hypotheses formulated within the context of currently available evidence. This is how I understand science to work, or at least how it should ideally function, in my opinion. We should make decisions based on the evidence at hand, not on potentially different evidence that may emerge in the future. Therefore, I am not entirely convinced by the argument that taxonomic changes should be postponed due to ongoing research on Rhegmatorhina hybridization. If we avoid taxonomic changes because of ongoing research, we might never make any changes. I believe the decision here should be as straightforward as the recent NACC decision to treat Corvus caurinus as conspecific with C. brachyrhynchos, for example. Common to both cases is the existence of recent studies providing crystal clear evidence for extremely weak reproductive isolation between the two lineages involved. If Del-Rio et al. (2021) did not provide sufficient evidence of weak reproductive isolation between berlepschi and hoffmannsi, I wonder what future research could possibly demonstrate conspecificity between these two taxa. Should that evidence arise, we can adjust accordingly, as we scientists routinely do.”

Comments from Claramunt: “Tentative NO. I was going to vote Yes, persuaded by Rafael’s proposal and a cursory examination of Glaucia’s superb* paper. R. berlepschi and hoffmannsi are clearly hybridizing and forming a relatively wide hybrid zone with intergradation of traits, and admixture of genomes. Songs are the same and the plumage distinction didn’t impede interbreeding. It seems that there are no barriers to interbreeding. I’m not convinced by the argument about the purported identification of F1 individuals. In a sample of 222 individuals and only five diagnostic loci, it would not be hard to find a handful of specimens that look like F1s by chance, just by being heterozygotes for the five loci. Unless I’m missing something, I don’t see how this method would lead to correct identification of F1s. And it seems that it would be impossible for true F1s to be formed anyway given that there are no parental pure forms in geographic proximity.

“The unavoidable conclusion under the biological species concept is that these two taxa should be lumped. Many recent proposals for species status are being rejected because of the speculation of potential interbreeding in hypothetical sympatry. Well, here there is actual evidence of widespread interbreeding and no signs of reproductive barriers.

“Having said that, it is possible that the degree of genomic admixture is being overestimated. Each SNP has very little information about genealogy, and at least some methods used to analyze SNPs do not take a historical/evolutionary perspective, e.g. allele A suggests ancestry X if A is frequent in X, regardless of whether A is the ancestral state or not. If allele A is ancestral, sharing it does not indicate close relationships in the genealogical network (basic phylogenetics applied to genealogies). I’m going to do some research to see if I can back up my suspicion with some evidence. In the meantime, I feel entitled to suspect that the width of the SNP clines and the degree of admixture is being overestimated. This may be a case of a, maybe not narrow but relatively constrained hybrid zone (200km?) on genomic backgrounds that are differentiated. Then the question would be whether the hybrid zone is a tension zone maintained by some sort of selection against hybrids or assortative mating. Is the hybrid zone stable or moving? Is it shrinking or expanding? There may be room for still considering these two taxa separate species even if hybridizing. For that reason I tentatively vote NO, waiting for more comments and discussion about this case, that may set a precedent for decisions based on this kind of scenario.”

Comments from Del-Rio: “NO. It is a complicated case, primarily because most of the cases we analyze here at SAAC are based on plumage, song, morphology, a few genes, or—in the best-case scenarios—reduced representation genomic data (usually less than 1% of whole genomes). In the case of Rhegmatorhina, however, we have all of the above plus a high-quality whole genome and whole genome resequencing data at a decent coverage for almost one hundred individuals.

“In the era of whole genomes, let’s see what happens. How are we going to define what constitutes strong versus weak levels of reproductive isolation?

“The results and ideas in the Evolution paper (Del-Rio et al., 2022) have evolved over the last couple of years. I won’t go into details here, but my most recent results show: (1) several genomic islands of differentiation between Rhegmatorhina hoffmannsi and R. berlepschi; (2) strong differential selection indicating assortative mating on regions of the genome involved in the definition of plumage coloration (Del-Rio et al in prep.).

“I vote NO because: (1) The argument that the Rhegmatorhina hybrid zone is too wide is flawed, as the hybrid zone width varies depending on which part of the genome is being analyzed. Some regions of the genome and some plumage features show narrow and steep clines. That’s the beauty of the system—a unique opportunity to understand which parts of the genome are preventing species fusion; (2) Additionally, with the available methods, the complex geography of the system leads to an overestimation of cline widths; (3) I am still working to understand how much these islands of genomic differentiation translate into reduced fitness in hybrids; (4) I am still trying to perform experiments to test for assortative mating based on plumage and vocal characters; (5) I am still analyzing vocal and environmental variation between the two forms; (6) In Brazil, “subspecies” status is rarely considered in conservation efforts, and R. berlepschi might become a species of concern with the advance of deforestation in Amazonia.

“Regarding the idea of gene flow with R. gymnops, my data does not corroborate the findings of Weir et al. (2015) so far. Again, this is another aspect I am still investigating.”

“In the era of whole genomes, let’s see what happens. How are we going to define what constitutes strong versus weak levels of reproductive isolation?

“The results and ideas in the Evolutio paper (Del-Rio et al., 2022) have evolved over the last couple of years. I won’t go into details here, but my most recent results show: (1) several genomic islands of differentiation between Rhegmatorhina hoffmannsi and R. berlepschi; (2) strong differential selection indicating assortative mating on regions of the genome involved in the definition of plumage coloration (Del-Rio et al in prep.).

I vote NO because: (1) The argument that the Rhegmatorhina hybrid zone is too wide is flawed, as the hybrid zone width varies depending on which part of the genome is being analyzed. Some regions of the genome and some plumage features show narrow and steep clines. That’s the beauty of the system—a unique opportunity to understand which parts of the genome are preventing species fusion; (2) Additionally, with the available methods, the complex geography of the system leads to an overestimation of cline widths; (3) I am still working to understand how much these islands of genomic differentiation translate into reduced fitness in hybrids; (4) I am still trying to perform experiments to test for assortative mating based on plumage and vocal characters; (5) I am still analyzing vocal and environmental variation between the two forms; (6) In Brazil, “subspecies” status is rarely considered in conservation efforts, and R. berlepschi might become a species of concern with the advance of deforestation in Amazonia.

“Regarding the idea of gene flow with R. gymnops, my data does not corroborate the findings of Weir et al. (2015) so far. Again, this is another aspect I am still investigating.”

Additional comments from Rafael Lima: “Considering the comments above, it seems unlikely that any amount of evidence or argument will alter the trajectory of this proposal. Nevertheless, I would like to share some final thoughts:

“1. My impression from all this is that there is a strong reluctance to change, as most arguments for maintaining the current classification hinge on predictions about ongoing studies or conservation-oriented reasoning, rather than drawing much more straightforward conclusions from the available evidence. While I’ve already expressed my thoughts on this, I feel it’s important to reiterate that taxonomic decisions should be grounded in the evidence we currently have, not on what future research might reveal. I fully understand the desire for taxonomic stability, especially for those working in downstream fields that use checklists like this. However, prioritizing stability over scientific rigor severely undermines taxonomy as a discipline. I therefore urge that decisions be made based on present evidence rather than futurology.

2. The conclusion that these taxa should be considered conspecific appears so evident in the results of Del-Rio et al. (2022) that I initially thought this proposal would be straightforward. However, based on the comments above, it seems I may not have fully elaborated on my interpretation of these results. One key point I thought would be clear to anyone reviewing the results from Del-Rio et al. (2022) is the modality of the hybrid zone. The modality of a hybrid zone is a critical indicator of the strength of reproductive isolation between two interbreeding populations (nicely reviewed in Jiggins & Mallet 2000, Trends Ecol. Evol. 15: 250-255). In Rhegmatorhina, the hybrid zone is clearly unimodal. Most importantly, it is late-generation hybrids that dominate. I can’t think of a more compelling indication that overall reproductive isolation is very weak in this system. If assortative mating existed, the zone would be bimodal. If strong postzygotic isolation occurred, through selection against hybrids or genetic incompatibilities, first-generation hybrids, not backcrosses, would dominate. Neither assortative mating nor postzygotic isolation seems to be in play.

“Glaucia mentions the presence of steep clines, possible assortative mating, and reduced hybrid fitness as potential reasons to maintain the taxa as separate species. Theory predicts that as differentiation accrues between any two populations, some degree of reproductive isolation will almost always be present. Unsurprisingly, it is not unusual to observe steep clines, assortative mating, and reduced hybrid fitness within species. The critical question when applying the BSC is not whether these phenomena are at play, but whether they are strong enough to prevent the fusion of the populations involved. In the case of Rhegmatorhina, there are some narrow clines, but they exist alongside many broader ones. Are these narrow clines causing moderate or strong overall reproductive isolation? The evidence suggests not. Regarding assortative mating and hybrid fitness, the unimodal nature of the hybrid zone makes it implausible that either is occurring at a significant level. It is not entirely clear to me how there could be assortative mating given that there are no parentals in the hybrid zone, or how hybrids could suffer from reduced fitness when backcrosses are so predominant. Even if we assume some degree of non-random mating and reduced hybrid fitness, the current evidence strongly suggests that these putative barriers are insufficient to prevent extensive gene flow. Gene flow remains abundant, despite any potential barriers. The way I see it, the burden of proof is clearly on the current multiple-species treatment, for which the evidence has long been weak (as noted by Willis 1969).

“Finally, it is not true that subspecies taxa are not considered in conservation efforts in Brazil. The Brazilian Red List of Threatened Species does include subspecies-level assessments for birds. However, this point should not even be a factor in this discussion, as the decision here should be independent of conservation concerns. (Taxonomic decisions motivated by conservation have already been addressed in other proposals, so I will refrain from elaborating further.) Otherwise, species limits in Rhegmatorhina will, after this voting, be based more on speculation and conservation-oriented reasoning than on existing evidence.”

Additional comments from Claramunt: “I change my vote to YES. At the end of the day, I’m with Rafael on this. The evidence of introgressive interbreeding is just overwhelming. There are no signs of selection against hybrids: the paper concludes that neutral diffusion cannot be rejected. There are some hints of “islands of differentiation” and reduced hybrid fitness in some part of the genomes, but those are clearly not sufficient as reproductive barriers and genes are profusely flowing in both directions. The conclusion is inescapable under any species concept: reproductive isolation is nonexistent, diagnosability is broken, lineages are intertwined.

“It doesn’t get better than that when it comes to evidence of interbreeding. It is not only the absolute width of the hybrid zone what matters but how broad it is in comparison with the rest of the distribution of the species involved. In this case, the tail of the hybrid zone gets so deep into the berlepschi territory that there are barely any pure berlepchi anymore (see the ancestry analysis in Fig. 3). The paper by Glaucia et al. is superb, and we are in a period in which we need to pay close attention to analyses of genomic data. The genomic revolution would likely change our perspectives on how we delimit species and even on what species are. But today, under current species concepts, the evidence in this case is conclusive. Keeping berlepschi and hoffmannsi as separate species in the face of seemingly free introgressive hybridization would be a radical departure from a century of application of the biological species concept. If our mission is to produce a coherent and evidence-based classification, we can’t maintain berlepschi and hoffmannsi as separate species. There is no better evidence of interbreeding than genomic analyses like this. Plumage and song differences and similarities, and playback experiments, can suggest reproductive compatibility, but never prove it. We can’t ignore the best evidence in this case and at the same time lump other taxa based on much lower quality evidence, in many cases, educated speculation about what birds would do if in sympatry. Is this a breaking point in our approach to species delimitation? Actual evidence of introgressive hybridization does not count anymore? We must surrender to the evidence if we are to produce a scientific classification.”