Revise species limits in Habia rubica: A. Treat Middle American H. rubicoides as a separate species; and B. Also treat Amazonian rhodinolaema group as a separate species.

Proposal (1048) to South American Classification Committee

Revise species limits in Habia rubica: A. Treat Middle American H. rubicoides as a separate species; and B. Also treat Amazonian rhodinolaema group as a separate species.

Note from Remsen: This is a proposal submitted to NACC (as “Treat Red-crowned Ant-Tanager Habia rubica as six or seven species“) that I here modify for SACC. First, is the text of the original proposal, left largely intact, including the emphasis on Middle American taxa not directly relevant to SACC; you can skip big chunks of this. Then, I included comments from NACC members. Finally, I restructure the proposal and voting along lines consistent with the NACC results and relevance to SACC.

Background

The Red-crowned Ant Tanager Habia rubica (Cardinalidae) is a highly polytypic taxon with marked geographical variation; up to 14 subspecies have been recognized, most of which were described based on variation in the hue and intensity of plumage color (Hilty, 2011). Its current distribution ranges from central Mexico to northeastern Argentina and southeastern Brazil and encompasses regions with very different ecological conditions or separated by recognized biogeographical barriers (Hilty, 2011).

Ridgway (1902) placed Habia rubica in the tanagers, which included 21 genera. One of them was Phoenicothraupis (Cabanis 1850). According to Ridgway (1902), the genus Phoenicothraupis (described by Cabanis 1850 with the type Saltator rubicus Vieillot) had three species: Phoenicothraupis rubica (with five subspecies: P. r. rubicoides, P. r. nelsoni, P. r. vinacea, P. r. affinis, P. r. rosea), Phoenicothraupis salvini (with five subspecies), and Phoenicothraupis fuscicauda (monotypic). Ridgway (1902) described the genus Phoenicothraupis as follows:

Medium-sized Tanagers superficially resembling the more uniformly colored species of Piranga, but outermost (ninth) primary shorter than second (instead of decidedly longer than third); adult males with a scarlet crown-patch and with more or less red on under parts (sometimes confined to the throat); females and young brown or olive above, paler below. Bill as in the more slender-billed species of Piranga but narrower (width at base scarcely if at all exceeding basal depth), the gonys relatively shorter, and distinctly, though slightly, convex, and maxillary tomium without any indication of a tooth-like projection. Nostrils narrower. Rictal bristles strong, conspicuous, and frontal bristles (over nostrils) well developed. Wing about three and three-fourths to a little more than four times as long as tarsus, much rounded (seventh to fourth primaries longest, ninth shorter than second); primaries exceeding secondaries by much less than length of tarsus. Tail shorter than wing by much less than length of tarsus, sometimes nearly as long as wing, more or less rounded, the rectrices rather broad, with rather loose webs and somewhat pointed tips. Tarsus decidedly longer than middle toe with claw; outer claw reaching about to or a little beyond base of middle claw, the inner claw falling short of the latter; hind claw shorter than its digit.

Coloration.—Adult males reddish brown, reddish gray, or dusky, with bright red throat and crown, the feathers of the latter sometimes developed into a more or less obvious crest; females and young usually brownish above, paler beneath, with or without a yellowish-buffy or tawny crown-patch; adult female sometimes similar to the male, but duller.

Range.—Southern Mexico to southern Brazil, Paraguay, Bolivia, and western Ecuador.

Phoenicothraupis was placed in Habia by Hellmayr (1936) and subsequent classifications. The genus Habia was until recently placed in the Thraupidae, but we now treat it in the Cardinalidae (Burns et al. 2014).

The AOS currently treats Habia rubica as single species (AOU 1983, AOU 1998). Hilty (2020) recognized 17 subspecies as follows:

· H. r. rubica: southeastern Brazil (southern Minas Gerais) to eastern Paraguay and northeastern Argentina

· H. r. bahiae: tropical eastern Brazil (Bahia)

· H. r. rubicoides: southern Mexico (Puebla and eastern Veracruz) to northern Nicaragua

· H. r. holobrunnea: subtropical eastern Mexico (southern Tamaulipas to Veracruz and northern Oaxaca)

· H. r. nelsoni: southeastern Mexico (Yucatán Peninsula north of southern Campeche)

· H. r. alfaroana: northwestern Costa Rica (Guanacaste Peninsula)

· H. r. vinacea: Pacific slope of southwestern Costa Rica (Nicoya Peninsula) to eastern Panama

· H. r. affinis: Pacific slope of southern Mexico (Oaxaca)

· H. r. rosea: Pacific slope of southwestern Mexico (Nayarit and Jalisco to Guerrero)

· H. r. rubra: Trinidad

· H. r. crissalis: coastal mountains of northeastern Venezuela (Anzoátegui to Sucre)

· H. r. mesopotamia: Venezuela (Río Yuruán region of eastern Bolívar)

· H. r. coccinea: eastern base of eastern Andes of north-central Colombia and western Venezuela

· H. r. rhodinolaema: southeastern Colombia east of the Andes to northeastern Peru and far northwestern Brazil

· H. r. peruviana: tropical eastern Peru to central Bolivia and adjacent western Brazil

· H. r. hesterna River (eastward to Rio Xingu), southward to northern Mato Grosso

· H. r. perijana: Sierra de Perijá (Colombia/Venezuela border)

New information: Three recent studies were published on Habia rubica. One (Lavinia et al. 2015) was a phylogeographic analysis that suggested that this species originated in South America, where there are at least two clearly differentiated clades: one in the Atlantic Forest of Brazil and another in the Amazon basin. However, limited taxon-sampling precluded detailed investigation of diversification within Central America and southern Mexico. The other two studies (Ramirez-Barrera 2018, 2019) provided new evidence, one using multilocus data and the other, plumage color.

Lavinia et al. (2015) sequenced mtDNA and nDNA from 100 individuals from Mexico to Argentina. Their results are shown below (Bayesian and maximum parsimony, Fig. 1). Their geographic sampling did not include Panama and much of western Mexico. The most important result was the identification of two lineages in South American that split ca. 3.5 MA and two lineages in Middle America. Their analyses of vocalizations and plumage coloration found differences between the four main groups consistent with the molecular data.

Ramirez-Barrera et al. (2018) amplified mitochondrial and nuclear markers from 125 individuals of H. rubica covering the species’ distribution, the other three species of Habia (fuscicauda, atrimaxillaris, gutturalis), and 16 samples from Chlorothraupis (C. olivacea, C. carmioli and C. stolzmanni). They found that H. rubica can be divided into three main clades: (1) western Mexico, (2) eastern Mexico to Panama, and (3) South America. Within these main clades they recognized seven main phylogroups, as shown in the next figure:

Figure 1. Geographical distribution, phylogenetic consensus tree, and haplotype network. (A) Geographical distribution (indicated by pink shading) and sampling points of H. rubica; the mitochondrial DNA sampling is represented by the color of the dots and the nuclear DNA sampling is highlighted with a black dot on the dot’s color. ArcGIS (ArcMAP 10.2.2; Esri, Redlands, CA, USA). (B) Phylogenetic consensus tree representing the relationship among populations of H. rubica, based on Bayesian inference from a multilocus dataset. Values above branches denote posterior probabilities (PP). (C) Haplotype network, where the phylogroup ”D'' corresponds to individuals from the Chiapas-Yucatan peninsula to Costa Rica and “d'' corresponds to individuals from Guatemala and El Salvador (the numbers inside of circles indicate the number of individuals that shared each haplotype).

Ramirez-Barrera et al.’s (2018) species delimitation analysis (BP&P) is shown on the following page; note the high speciation probabilities (0.97 to 1.0).

Ramírez-Barrera et al. (2019) analyzed genetic, coloration, and morphometric data from specimens from collections in Mexico and the United States and used the Multiple Matrix Regression with Randomization (MMRR) approach to evaluate the influence of geographic and environmental distances on genetic and phenotypic differentiation at both the phylogroup and population levels. They found that geographic isolation was the main factor structuring genetic variation within populations of Habia rubica; this suggests that climate did not playing a major role in within-species genetic differentiation.

Recommendation:

We present two options:

(1) Separate H. rubica into seven species: This is based on phylogenetic evidence and some differences on plumage color.

1. Habia rosea (Nelson, 1898): Pacific coast of western Mexico (Jalisco, Nayarit, and Colima; lineage A in Figure 2)

2. Habia affinis (Nelson, 1897): Pacific coast of southwestern Mexico (Michoacan, Guerrero, and Oaxaca; lineage B in Figure 2)

3. Habia holobrunnea (Griscom, 1930): E Mexico from S Tamaulipas S to N Oaxaca (lineage C in Figure 2

4. Habia rubicoides (Lafresnaye, 1844): S Mexico (from Puebla, E Oaxaca, Tabasco, and Chiapas), Guatemala and Belize S to Honduras, El Salvador, and Nicaragua (lineage D in Figure 2).

5. Habia vinacea (Lawrence, 1867): Panama (lineage E in Figure 2)

6. Habia rhodinolaema (Salvin and Godman, 1883): Amazon basin (lineage F in Figure 2)

7. Habia rubica (Vieillot, 1817): southeastern Brazil, Argentina, and Paraguay (lineage G in Figure 2)

(2) As in Option 1 but treat lineages A and B as the same species (Habia rosea – Habia affinis). This option is to separate Habia rubica into six species based on two main arguments, the first being the low speciation probability value (<0.95) presented by the clade comprising the subspecies H. r. rosea and H. r. affinis in the BP&P analyses performed with multilocus data (Ramírez-Barrera et al. 2018). The second argument is the consistency of this grouping through two independent tests of the same analysis and with adjustment of specific parameters. At the morphological level, both subspecies are described as a polytypic group characterized by having paler plumage than the other subspecies of H. rubica (del Hoyo 2020). At the geographical level, the distribution of these two subspecies from western Mexico is clearly delimited by three large mountain ranges: the Eje Neovolcánico Transversal, the Sierra Madre del Sur, and the Sierra Madre Oriental. These geographical formations seem to have great influence on the genetic structure of the populations, acting as barriers to gene flow, which has possibly promoted the differentiation between the populations of eastern and western Mexico.

Species	Proposal 1	Species	Proposal 2
1	Habia rosea (Nelson, 1898)	1	Habia affinis (Nelson, 1897)
2	Habia affinis (Nelson, 1897)	1	Habia affinis (Nelson, 1897)
3	Habia holobrunnea Griscom, 1930	2	Habia holobrunnea Griscom, 1930
4	Habia rubicoides (Lafresnaye, 1844)	3	Habia rubicoides (Lafresnaye, 1844)
5	Habia vinacea (Lawrence, 1867)	4	Habia vinacea (Lawrence, 1867)
6	Habia rhodinolaema (Salvin and Godman, 1883)	5	Habia rhodinolaema (Salvin and Godman, 1883)
7	Habia rubica (Vieillot, 1817)	6	Habia rubica (Vieillot, 1817)

We recommend Option 1 for the following reasons:

1. The phylogenetic evidence supporting the differentiation between H. r. rhodinolaema and H. r. rubica in South America is consistent across two independent studies (Lavinia et al. 2015, Ramírez-Barrera et al. 2018), which present broad sampling and cover most of the distribution of both phylogroups. In addition, significant differences in traits such as song and coloration have also been reported (Lavinia et al. 2015), which support the probable differentiation between the two phylogroups. Furthermore, there is a geographic correspondence between the distribution ranges of the identified phylogroups (del Hoyo 2020), which could also indicate that geographic barriers in South America have influenced their genetic differentiation.

2. The deep genetic structure reported between the phylogroups of Central America and eastern Mexico (Ramírez-Barrera et al. 2018) shows high support for the phylogenetic hypothesis obtained with multilocus data. This genetic structure corresponds to barriers to dispersal and gene flow such as the Talamanca Mountain Range in Costa Rica, the Motagua-Polochic-Jocotán fault system, and the Isthmus of Tehuantepec in southern Mexico, which could be promoting the reported genetic isolation and genetic differentiation. Analyses of differences between these groups for traits such as song and coloration could not be tested in detail due to the low sampling available; however, a clear difference is reported between these phylogroups and the phylogroups distributed in South America (Lavinia et al. 2015).

3. The five phylogroups distributed from central Mexico to South America show evidence of genetic and phenotypic differentiation, as well as geographic correspondence in their distributions. We consider this evidence sufficient to identify at least five clearly differentiated species. This would allow us to better explain the relatively weak patterns of variation among the subspecies described for this geographic range of H. rubica.

4. The phylogroups distributed along the western slope of Mexico also show robust molecular differentiation supported by a phylogenetic hypothesis based on multilocus data that shows a profound divergence between those distributed in the north and south of this region (Ramírez- Barrera et al. 2018). This evidence is supported by very complete sampling, which covers the entire distributional range of H. rubica and leaves no room for doubt about the genetic structure of the species in this region. In addition, there is evidence of geographic correspondence with barriers such as the Transversal Neovolcanic Belt, the Sierra Madre del Sur, and the Sierra Madre Oriental, which limit the distribution of populations and have kept them isolated for long periods of time, which has favored the deep genetic differentiation (Ramírez- Barrera et al. 2018). There is no evidence of song differentiation as in the previous cases; however, both phylogroups are recognized as a polytypic group whose coloration is markedly paler than that reported in the rest of the species (del Hoyo 2020), which could add evidence on the degree of differentiation that these groups present.

5. Habia rubica is widely described as a polytypic species with a very wide distribution; however, it has also been described as a species in which “differences between the racial groups are not always clear-cut or pronounced” (del Hoyo 2020). “Several of the numerous races are weakly differentiated and seem barely worth recognizing” (del Hoyo 2020). This apparently weak phenotypic differentiation, coupled with the complicated field identification of the different subspecies (especially those with sympatric distribution), makes it necessary to consider all those genetic, geographical, and phenotypic elements (e.g., pink coloration of phylogroups from western Mexico) as sufficient arguments to be able to recognize these groups as independent species.

English names: If the proposal passes, a follow-up proposal on English names is needed. We should also coordinate with SACC on this proposal.

Literature Cited:

Burns, K. J., A. J. Shultz, P. O. Title, N. A. Mason, F. Keith Barker, J. Klicka, S. M. Lanyon, and I. J. Lovette. 2014. Phylogenetics and diversification of tanagers (Passeriformes: Thraupidae), the largest radiation of Neotropical songbirds. Molecular Phylogenetic and Evolution 75: 41-77.

Hilty, S. L. 2011. Family Thraupidae (Tanagers). Handbook of the birds of the world. Tanagers to New World blackbirds, vol. 16. Barcelona: Lynx Editions, 46 329.

Lavinia, P. D., P. Escalante, N. C. García, A. S. Barreira, N. Trujillo-Arias, P. L. Tubaro, K. Naoki , C. Y. Miyaki, F. R. Santos, and D. A. Lijtmaer. 2015. Continental-scale analysis reveals deep diversification within the polytypic Red-crowned Ant Tanager (Habia rubica, Cardinalidae). Molecular Phylogenetics and Evolution 89:182 193. DOI 10.1016/j.ympev.2015.04.018

Ramírez-Barrera, S. M., B. E. Hernández-Baños, J. P. Jaramillo-Correa, and J. Klicka. 2018. Deep divergence of Red-crowned Ant Tanager (Habia rubica: Cardinalidae), a multilocus phylogenetic analysis with emphasis in Mesoamerica. PeerJ. 1-19 DOI: 10.7717/peerj.5496

Ramírez-Barrera, S. M., J. A. Velasco, T. Orozco-Téllez, A. M. Vázquez-López and B. E. Hernández-Baños. 2019. What drives genetic and phenotypic divergence in the Red-crowned Ant Tanager (Habia rubica, Aves: Cardinalidae), a polytypic species? Ecology and Evolution. DOI: 10.1002/ece3.5742

Ridgway, R. 1902. The birds of North and Middle America: a descriptive catalogue of the higher groups, genera, species, and subspecies of birds known to occur in North America, from the Arctic lands to the Isthmus of Panama, the West Indies and other islands of the Caribbean Sea, and the Galapagos Archipelago. Bulletin of the United States National Museum No. 50, Part 2.

Blanca E. Hernández-Baños & Sandra M. Ramírez-Barrera, May 2025

Here are the comments from NACC members (names removed to protect anonymity as per NACC policy):

Voter 1: “YES to A. I find the combination of very old (mid-Pliocene!) genetic lineages and qualitative differences in song and plumage to be indicative of multiple species in the group. However, until someone quantifies the differences in song and plumage, I am only comfortable with recognizing three (and at most four) species in this group: the Atlantic Forest, Amazonian, and Middle American lineages. There seem to be some consistent differences with the two West Mexican lineages, and if other committee members move towards recognizing that clade as a species, I am open to the idea.

“So, for now, I will vote to recognize as species the 3 clades listed above. The Lavinia et al. (2015) study cited in the proposal does have some data on song and plumage, but they had data mostly just for the Amazonian and Atlantic Forest clades, but they showed that the latter has a darker / more saturated red plumage and a distinctly slower song. The part that concerns me is that they mention that there was more plumage variation within each clade than between. This agrees with my experience with the species, which like many Cardinalids shows considerable variation in the carotenoid-based colors, even within one flock. A thorough study of plumage variation would be helpful, but even with those caveats in mind, there are consistent differences between the clades in my experience. The Atlantic Forest birds, in addition to being redder, are a distinctly intense and evenly colored red that is quite different than that of the Amazonian and Mesoamerican birds. The females are also distinctly brown (vs olive). There also seem to be some qualitative structural differences with the Atlantic Forest birds, especially in the head and bill shape. I was unfamiliar with the song of this group, but in listening to recordings, they do have very widely spaced notes, typically repeated, sound ‘clipped’, and are quite different than the other clades. The calls also seem shorter. In contrast, the Amazonian birds are somewhat paler red, more often have gray on the flanks, have a more contrasting red throat, have clear whistled songs that are run together and rise to a kind of crescendo at the end, and have longer harsher calls. The Mesoamerican birds look to me more like the Amazonian group, but there is so much plumage variation that I’m not sure if I can pick out a consistent difference versus the Amazonian birds (a bit concerning). The Mesoamerican birds do have distinctly rising notes that are given in pairs, at least in the north. The southern Central American birds tend to have a kind of bouncy staccato song, which likely warrants investigation. The ones in West Mexico are clearly pale and small-billed, but again, there is much variation. The songs are more slurred, too. These do seem to be allopatric from the populations in eastern Mexico and are a distinct lineage on the tree.

“We will likely need another proposal on English names if this passes, but I’ll just toss some ideas out there. First, I don’t think that we should further modify “Red-crowned”, as that will create unnecessarily long names. eBird/Clements has subspecies group names, which are a good starting point, but do not perfectly match the clades if we adopt fewer than five species. Given the brighter red plumage of nominate H. rubica, I suggest following eBird/Clements’ and adopt Red Ant-Tanger for this clade. For the other two, I suggest geographic or plumage-based names. eBird/Clements uses Scarlet-throated Ant-Tanager for rubra of the Amazon, which fits for the bird given its brighter throat, but could I suppose cause a perceived association with Red-throated Ant-Tanager. Amazonian Ant-Tanager could also work, but this species would also include the perijanus of the Perijá mountain range (outside the Amazon). Regardless, both are SACC issues. For our birds, eBird/Clements uses “Northern” for rubicoides of Panama to east Mexico and “West Mexican” for the west Mexican clades. I suggest Mesoamerican or Middle American Ant-Tanager for the combined group.”

Voter 2: “YES to A. As with other committee members, I am most comfortable now with recognizing three species: Atlantic Forest, Amazonia, and Middle America. The Lavinia et al. (2015) study had only one sampling locality for western Mexico (n = 2), and that clade was also not included in the vocal analysis. Furthermore, I am concerned by the lack of study of any potential contact zone in Middle America; the only mention of a contact zone there is the statement in the paper that “the possibility of a contact zone in southern Oaxaca and Chiapas remains to be explored and therefore the presence of gene flow cannot be fully discarded.” Until such a study is done, I think it’s best to wait on splitting western Mexico populations from those in eastern Mexico and Middle America.

Voter 3: “(A) YES; split the species; (B) Three species. I’ve enjoyed reading the series of detailed papers on this complex; there are definitely multiple species involved. There is good evidence for 5 or 6 species under many concepts, such as the phylogenetic species concepts. However, since we adhere to the biological concept, I think we need more evidence regarding the proposed Mexico and Central American species, in particular vocal data. As others have noted, to date we have song, plumage, and genetic data to justify a three-way split (Atlantic Forest, Yungas & Amazonia, and Mexico/Middle America). Thus, at this point, I would be in favor of voting for a 3-way option. As a side note, one of my MS students has an unpublished MS thesis (Scott 2019) with a UCE phylogeny that includes one sample each of these three, and it shows the same set of relationships as the studies discussed in this proposal. Levels of relative divergence is similar to divergence among currently recognized species of other cardinals/grosbeaks.

Voter 4: “YES to a split. The proposal and the papers it is based on make it clear that at least three BSC species are involved. I also consider that the evidence favors treating the two West Mexican taxa as a separate species (affinis) rather than as a subspecies of rubicoides, based on the genetics, morphology, and song.
Song.---I downloaded and compared songs of the two groups in sonogram composites, and it seems pretty clear that the notes of affinis songs are more modulated and more vertical, while the notes of rubicoides are more prolonged and horizontal, and often with multiple note types per strophe rather than a single note type or a tiny grace note as is typical of affinis. It also seems that affinis may not sing that often, with many more recordings of calls than song, while there are plenty of songs of rubicoides. Calls of these two groups are less tractable, with multiple call types represented, some perhaps misidentified (?), and not differing obviously in quick comparisons. (If anyone wants to see the sonogram composites I’d be happy to share).

Plumage.---Not necessarily a species characteristic, but I was quite impressed with the pallid plumage of birds I saw in Jalisco last year, which seemed very unlike those I’ve seen elsewhere.

Genetics.---The two phylogenies presented in the proposal show a divergence between the west Mexican and east Mexican/Central American taxa more or less on a par with the other clades.

Distribution.---It’s not clear from either the range maps in Howell and Webb, or eBird records, that there even is a zone of overlap between the affinis and rubicoides groups. There seems to be a small break in Oaxaca between the northern Oaxaca rubicoides group and southern Oaxaca affinis group.

“So, I’m voting for four species (first choice) and failing that, three species (second choice).

Voter 5: “YES. As other committee members have already commented, I am only comfortable accepting a 3-species split, recognizing Middle American, Amazonian, and Atlantic Forest birds as separate, although I too would be open to also recognizing the West Mexican subspecies as separate if the rest of the committee feels that is appropriate.”

Voter 6: “YES. I agree with the other committee members that there is good evidence for a 3-way split of Middle American, Amazonian, and Atlantic Forest birds. There may very well be additional splits, but I’d like to see more analyses of phenotypic variation and putative contact zones among the Middle American taxa. BP&P analyses certainly have value but tend to oversplit lineages compared to more conventional species delimitation approaches.”

Explanation of SACC voting procedure from Remsen: As a consequence of these comments and votes, NACC voted to recognize three species, as follows:

1. Habia rubicoides (extralimital to SACC): includes all Middle American taxa listed in the proposal, from Mexico to Panama.

2. Habia rhodinolaema: Amazonia etc. (includes peruviana and hesterna by genetic sampling and perijana, rubra, crissalis, mesopotamia, and coccinea by phenotype; note that rhodinolaema was not genetically sampled)

3. Habia rubica: Atlantic forest (including bahiae)

Note that the NACC proposal really didn’t get into the details of the South American taxa – the focus was on Middle America. So, let’s try this as a voting structure, with two YES/NO choices.

A. TWO SPECIES. YES is to treat Middle American H. rubicoides as a separate species from all South American taxa, i.e. treat the rhodinolaema and rubica groups as a single species, at least for now. NO is for continuing the status quo, i.e., all one species. A NO vote on this automatically translates to a NO on B (I think).

B. THREE SPECIES. YES is to treat all three groups as separate species, as per NACC voting results. NO is for continuing the status quo, i.e., all one species, or for more than two species in South America (an option not considered in the NACC proposal).

Note from Remsen on English names: A separate proposal will be needed on English names if either part of the taxonomic proposal passes, but meanwhile note discussion in NACC comments.

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

Comments from Bonaccorso: “NO. I think the level of discussion for South American-Panamanian forms in the current proposal is not sufficient to make an informed choice. The three proposed species are clearly divergent in mitochondrial DNA, but the level of nuclear divergence is difficult to tell because the concatenated mtDNA-nuclear tree could be driven by mitochondrial divergence. Plumage differences, however, are consistent with the subspecies level.”

Comments from Stiles: “YES to A and B. For A, this definitely shifts the burden of proof onto those recommending the status quo. B is a well-substantiated spilt based genetic and vocal evidence as well as more subtle differences in plumage. The split also makes biogeographic sense because the two populations are rather widely allopatric, following an Amazonia vs. eastern Brazil separation seen in many other groups.”

Comments from Areta: “Well, as routinely happens, the lack of a paper focusing on species-limits issues integrating data and a proliferation of papers lacking complete taxon sampling creates a situation in which some things appear to be obvious, some appear to be likely correct, and some other issues go undetected or at least are not given the foreground place that they deserve. I find it highly intriguing that none of the papers, nor the proposal, have cited Parkes (1969), which is a highly relevant paper discussing important specimens (e.g., the distribution of rhodinolaema and coccinea, and the birds west of the Andes in Colombia) and characterising two new subspecies from Venezuela (crissalis and mesopotamia).”

“Clade III in Ramírez-Barrera et al. 2018 is the South American clade, which is deeply diverged from Clades I and II in Mesoamerica that should be the focus of our deliberations. The deep split between these taxa was also recovered by Lavinia et al. (2015). Thus, it is clear that there are two very deeply diverged main lineages, one in South America and the other one in Mesoamerica. mtDNA and concatenated mt+nucDNA are concordant.

“Lavinia et al. (2015) made the following taxonomic recommendations:

‘ 4.5. Taxonomic implications

We found H. rubica to be monophyletic within our dataset with high support in spite of the deep divergences among its main lineages. The Crested Ant Tanager (H. cristata), a species that has not yet been included in any molecular studies and that based on morphological and behavioral evidence seems to be closely related to H. rubica (Willis, 1972), should be included in future analyses to confirm this result. In addition, and as evidenced by Klicka et al. (2007) and Barker et al. (2015), H. rubica appeared more closely related to species of the genus Chlorothraupis than to its congeneric species, making the genus Habia paraphyletic as currently defined. Since further analyses including H. cristata are needed to definitely establish the phylogenetic affinities among the species of these two genera, we are anyway keeping the genus name Habia for the new species suggested below.

“Based on consistent genetic, morphological and behavioral evidence, we believe that splitting H. rubica into three species is the most reasonable and conservative alternative for the moment. One of the species should be the lineage present in the Atlantic Forest, which should keep the name H. rubica. Despite the fact that our genetic analyses do not distinguish between the subspecies rubica and bahiae from the Atlantic Forest, we believe that further morphological and behavioral objective analyses (like the ones performed here among main lineages) are needed to fully address their validity. Therefore, we propose that this species should keep the subspecies rubica and bahiae for the moment. Although we lack representatives of some races distributed north of the Amazon River, we recommend that all South American subspecies other than the ones from the Atlantic Forest are included in a new species called H. rubra (given its taxonomic priority). Until a more comprehensive sampling of the subspecies from northern South America allows further genotypic and phenotypic analyses of this group, H. rubra would include the subspecies rubra, peruviana, rhodinolaema, hesterna, perijana, coccinea, crissalis, and mesopotamia. Finally, we suggest considering the Mexican and Middle American lineage a single species named H. rubicoides. This species would temporally comprise the subspecies rubicoides and affinis, as treated by Navarro-Sigüenza and Peterson (2004), and alfaroana and vinacea, until further sampling in western Mexico and southern Middle America allows a deeper analysis of these subspecies."

“The deepest splits in Clade III are group F (which we can call the rhodinolaema complex) and group G (which we can call the rubica AND rubra complexes).

“The rhodinolaema complex: taxon rhodinolaema was apparently sampled by Lavinia et al. (2015) but not by Ramírez-Barrera et al. (2018) (or so I understand; it is tough to check all the samples for ssp. assignation, which would have ideally been done by the authors; maybe it was sampled only for mtDNA?), so its recognition would rest on the sampling of rhodinolaema, hesterna, and peruviana (i.e., all its constituent taxa).

“The rubica complex: a tight grouping of AF birds, coupled to vocal distinctions indicate that it deserves to be ranked as a species different from rhodinolaema. This would include ssp. rubica and bahiae, both genetically sampled.

The rubra complex: this group was not recognised as such by any of the published papers or proposals, but I think that it can be reasonable to propose its existence. One sample from the eastern base of the Andes of Venezuela (coccinea) is eye-catching, as it is recovered as part of III-G, but it could as well have been placed on its own "lettered" clade given the depth of the divergence and the biogeographically implausible case of a species having a disjunct distributional pattern in the E Andes of Venezuela and Colombia and in the Atlantic Forest (see map in Ramírez-Barrera et al. 2019, which makes this more evident). This Atlantic Forest/northern South America connection is in need of more attention. Interestingly, the subspecies rubra from Trinidad and crissalis, from the NE mountains of Venezuela (Paria and Turimiquire; type from "Mirasol (3,000 feet), about 15 km. S. of Cumanacoa, Sucre, Venezuela") have very similar "double- or triple-ticking" calls to coccinea and perijana (https://search.macaulaylibrary.org/catalog?taxonCode=rcatan1&mediaType=audio&sort=rating_rank_desc&view=list&regionCode=VE), and very unlike the harsh long rasp of Atlantic Forest birds or the dry, short, staccato calls of the rhodinolaema complex. Taxon rubra, the second oldest name for Habia rubica sensu lato, from Trinidad is also missing genetic data. Populations west of the Andes in Colombia seem to also belong to this group and lack a name (if one is needed; see Parkes 1969). Finally, the taxon mesopotamia from SE Venezuela (E Bolivar) also lacks samples (both, vocal and genetic) and it was not even mapped by Lavinia et al. 2015 or Ramírez-Barrera et al. 2018, 2019). So, we lack sampling of key subspecies level entities that would presumably occur within the Clade III.

“I think it is troublesome to place the northern Andes/Paria-Turimiquire/E Bolivar taxa in the same species with the Atlantic Forest birds, when the admittedly meager genetic and vocal data indicate that they are deeply diverged and vocally distinctive (as in the proposal). I likewise think it is problematic to mix the rubra and rhodinolaema complexes under an enlarged H. rubra as in Lavinia et al. 2015 (i.e., include every South American taxon here, except for H. rubica rubica and H. rubica bahiae).

“It seems that there are AT LEAST 2 species in clade G rubica (with rubica and bahiae) and rubra (including, it seems, rubra, crissalis, coccinea, and perijana, and one could stretch this to include mesopotamia). This is a classic problem: whether we should follow our gut feeling, experience, and sense of likelihood, or whether we should wait for solid data. I tend to follow the later approach, because this is how solid knowledge is built. So far, I have proposed a different taxonomic hypothesis to those presented in the proposal or by Lavinia et al. 2015, but to me this hypothesis has not been rigorously tested. This group is still begging for a taxonomically minded genetic work, and for more sound recordings.

“After writing all this, I´ve found that eBird has a "rubra" group, that includes what I have here termed rhodinolaema and rubra complexes. These two are better treated separately, based on the single genetic sample from Barinas of coccinea in Ramírez-Barrera et al. (2018) and on the vocal differences of perijana, coccinea, crissalis, rubra (and maybe also the unrecorded, at least on XC and ML, mesopotamia) from the rhodinolaema complex.

“What to do? There are hundreds of sound recordings that are crying to be analysed. In a rapid assessment, it seems that they can help diagnose at least 3 main vocal (call) groups in Clade III which agree with genetic data (but many ssp. need more vocal samples, some remain unrecorded, and some are lacking genetic data). So, I think that Clade III might better be treated as at least 3 species: rubica (with ssp. rubica and bahiae), rhodinolaema (with ssp. rhodinolaema, hesterna, and peruviana), and rubra (with ssp. rubra, coccinea, birds W of the Andes in Colombia?, and perijana, and perhaps even mesopotamia, although this one might as well be something different, and the composition of this third species remains speculative). Other scenarios are likely as well, one species could be Andean (coccinea and perijana), one limited to NE Venezuela (rubra) and another one to E Venezuela (and Brazil? Guyana?; mesopotamia). I think that the 3-species solution might eventually be correct, but do we have enough data to decide? What to do? Move forward piecemeal or wait for an integrative approach? I look forward to the thoughts of others, knowing that I am usually on the "wait for thorough data" team.

“Disclaimer: it seems to me that there are numerous situations here that should be properly sorted in a rigorous taxonomic work. I approached this with an inquisitive mind, and I think that I could grasp part of the complexity of the situation. However, revising type (and other) specimens, assessing the distributions of taxa, properly sampling for genetic data, and studying the vocalizations should be done with greater attention to detail than what I could afford to vote here. Yet, I think that I have uncovered a novel pattern that should be tested. Maybe I erred here and there, but I feel that the overall picture is accurate, and so here it is to be subjected to scrutiny and criticism.”