Reorganize the generic classification of the “core tanagers”

Proposal (437) to South American Classification Committee

Reorganize the generic classification of the “core tanagers”

Several genetic studies have consistently identified a large clade of the Thraupidae (including nearly 100 species, at least a third of the Thraupidae, depending upon where family limits are eventually fixed), including the type genus Thraupis, the very large genus Tangara, and a diverse group of more-or-less montane genera (Yuri & Mindell 2002, Burns et al. 2002, Burns & Naoki 2004). For convenience, Burns & Naoki called this clade the “core tanagers”. Sedano & Burns (2010) have now produced a nearly complete phylogeny of this clade (including all current genera and 93 of 99 current species). This phylogeny is based upon large (1000+base pairs) samples of two mitochondrial genes and is rooted by the “Tholospiza” group (Coereba, the geospizine finches and several West Indian finches that have been identified as embedded in the Thraupidae and the closest relatives of the “core tanagers” in some studies. The Sedano & Burns study produced a consistent, mostly well-supported tree with Bayesian and maximum-likelihood analyses that, if accepted, would require considerable modifications in the current generic classification of the “core tanagers”. Sedano & Burns made a number of recommendations based on this tree, but these tended towards lumping of many distinctive groups into a few, highly heterogeneous and virtually non-diagnosable genera. The object of this proposal therefore is to seek a compromise solution that maintains genera as monophyletic groups while at the same time maintaining diagnosability with the least possible disruption of the current nomenclature. Even with these guidelines, it is evident that a considerable number of generic changes will be required. For the recommendations I propose, I have relied principally on the synonymies in Hellmayr (1936) and Ridgway (1911). I will divide the proposal into several parts, following the phylogeny of Sedano & Burns: 1) the Tangara group; 2) Chlorochrysa and the Paroaria clade; and 3) the mountain tanager clades.

1) The Tangara group: among the surprising results of the Burns & Sedano study is that Tangara, considered a monophyletic group by Burns and Naoki (2004), is actually polyphyletic with respect to Thraupis. Although the overall species groupings in Tangara itself found by Burns & Naoki were supported, Thraupis (itself polyphyletic) was found to be embedded in Tangara by Sedano & Burns. This would require either lumping the “true” Thraupis into Tangara to form a huge, heterogeneous genus, or splitting up Tangara into several smaller, monophyletic genera while maintaining a monophyletic “true” Thraupis that morphologically seems decidedly out of place in Tangara. Here I pursue this alternative and recommend the following generic arrangement.

a) Tangara Brisson 1760: the “true” Tangara comprise a large group of 26 species, the type of which is brasiliensis (now considered a subspecies of T. mexicana). The species included are those from vassorii through seledon in the phylogeny. This clade includes several subclades that could be split off if one wishes to maintain relatively homogeneous branch lengths throughout. This would require splitting Tangara into at least five smaller genera: Procnopis Cabanis 1844 for vassorii through fucosa in the phylogeny; a new genus for cyanotis and labradorides; Gyrola Reichenbach 1850 for gyrola and lavinia; Chrysothraupis Bonaparte 1851 for chrysotis through johannae; and Tangara Brisson 1860 for inornata through seledon. Several of these could be split further, but given that branch lengths are often short and support for many of the nodes is not terribly good, I see little point in doing so at this point. For the present, I prefer to retain a broad Tangara for all as they do form a fairly homogeneous group.

b) Ixothraupis Bonaparte 1850: forming a separate clade from the preceding, this genus includes the “spotted” group of 5 species, varia through xanthogastra in the phylogeny; the type species is T. punctata.

c) Chalcothraupis Bonaparte 1851: includes only T. ruficervix, for which the genus was described (although there might be some confusion, as Bonaparte apparently also included labradorides here; but according to Hellmayr, ruficervix is the type of the genus.) This species is one of a two-part basal polytomy of the clade that otherwise includes only the “true” Thraupis but it differs considerably in plumage and morphology from these, and is also rather different in plumage and in its fatter bill from the remainder of Tangara. An alternative would be to include it in Thraupis, which I prefer not to do given the above differences.

d) Thraupis Boie 1826: the seven species included here (episcopus, sayaca, glaucocolpa, cyanoptera, abbas, ornata (the type) and palmarum) form a coherent group in size and morphology; the species bonariensis and cyanocephala were excluded and will be treated below.

e) Euschemon Sclater 1851): includes the remaining 13 species of Tangara, palmeri through cucullata in the phylogeny; the type is flava (now considered a subspecies of cayana). This group includes the most sexually dimorphic members of Tangara as well as several more-or-less “hooded” species.

2) Chlorochrysa and the Paroaria clade: Chlorochrysa forms a well-separated small clade of its own and clearly merits recognition as a separate genus. The Paroaria clade includes a number of small, morphologically distinctive genera showing few resemblances among themselves: Stephanophorus, Diuca, Neothraupis, Lophospingus, Cissopis, and Schistochlamys as well as Paroaria itself. Given the striking degree of divergence among these mostly small genera, I favor maintaining all of them as any lumping would produce virtually undiagnosable salads. The levels of divergence in the phylogeny are high for most as well; the two most closely related, Cissopis and Schistochlamys, are perhaps the mist divergent of the lot.

3) The mountain-tanager clades: this group includes three major clades, each of which includes two or more monophyletic subclades that require evaluation:

a) a clade containing two current genera, Wetmorethraupis and Bangsia. The former is sister to the several Bangsia species, which form a monophyletic group. The differences in plumage and size are not that great: Wetmorethraupis looks a bit like a very fancy big Bangsia. However, all species of Bangsia are trans-Andean, with the group centered in the Chocó region, whereas Wetmorethraupis is cis-Andean, occurring to the south of any Bangsia (as well as on the other side of the Andes, which suggests a long-standing divergence). I tentatively favor recognition of both genera. I should also note that this phylogeny provides no support whatever for one of the most frequent lumping in the past, Bangsia into Buthraupis: the two are not even closely related, let alone sisters.

b) clade that includes several distinctive, monophyletic groups:

i) Thraupis bonariensis and Pipraeidea melanonota: These sister species differ somewhat in size and plumage (the latter suggests a smaller, plainer version of the former), but no more so than do members of several other genera, such that I see no formidable obstacles to including both in Pipraeidea; the alternative would be to name a new, monotypic genus for T. bonariensis (for which no specific previous generic name appears to exist).

ii) Iridosornis: the 5 species of this current genus form a monophyletic group for which generic status can be retained.

iii) Calochaetes, Delothraupis, and Dubusia: Calochaetes is sister to the other two and differs strikingly in plumage, ecology (basically a foothill to cloud-forest species) and somewhat in morphology (more chunky and shorter-tailed) from both of them; I would maintain it as a separate genus. Delothraupis and Dubusia, on the other hand, are similar in morphology and in being high Andean species; they differ mainly in the color of the underparts and somewhat in size. My recommendation would be to lump Delothraupis into Dubusia, as some have done (e.g., Meyer de Schauensee 1970): a number of other genera include species with this degree of differences, and it seems best to recognize their status as sister species. I also note that “Saltator” rufiventris probably falls in this clade; if so, the alternatives would presumably be to include it in Pipraeidea or name a new genus for it, but its generic status can only be resolved by further studies.

iv) The Anisognathus clade: this clade includes a basal polytomy of four groups: (A) Thraupis cyanocephala; (B) Buthraupis wetmorei; (C) Anisognathus somptuosus and notabilis; and (D) A. igniventris, lachrymosus, and melanogenys. Here, two options are available: lump all species into Anisognathus Reichenbach 1850, the oldest name for the entire group; or recognize each group as a separate genus. More work will be required to define the structure of this clade, and if all these are lumped the result would be a very heterogeneous group in size, plumage color, and at least bill morphology; hence, I propose the second alternative of four genera, each of which is well characterized. These would be:

A) Sporathraupis Bonaparte 1850 for T. cyanocephala; the genus name was coined for this species, which certainly has no place in Thraupis;

B) Tephrophilus Moore 1934 for B. wetmorei; again, coined for this species and so used by many authors;

C) Compsocoma Cabanis 1851 for A. somptuosus and notabilis, as used by Hellmayr and many others; and

D) Anisognathus Reichenbach 1850 for A. igniventris, lachrymosus and melanogenys.

Each of these groups is distinctive and easily diagnosed; Hellmayr used the same division of Anisognathus (although he used Poecilothraupis, a synonym of Anisognathus, for group D. Although further research may well reveal more structure in this clade leading to lumping of some of these groups, for the present I think it is best to be consistent with the evidence in hand and, given the clear phenotypic differences among them, recognize all four as genera.

v) The Buthraupis clade: this clade breaks basally into two groups: B. montana on the one hand, and Chlorornis and B. eximia and aureodorsalis on the other, C. riefferii being sister to B. eximia and aureodorsalis. One could justify one, two or three genera here, the oddball being C. riefferii, whose bright green plumage and red bill contrast strikingly with the blue-black-yellow plumages of the others. All are moderately to very large, heavy-bodied, rather short-billed high Andean forest tanagers such that if one were willing to overlook the jarring color clash, one could include all in Buthraupis Cabanis 1851. Recognizing two genera would separate B. montana and restrict the color clash to Chlorornis Reichenbach 1850, which would now include eximia and aureodorsalis as well as riefferii. The three-genus alternative would separate eximia and aureodorsalis from riefferii in the genus Cnemathraupis Penard 1919 (type eximia). My inclination would be to recognize three genera, to retain relatively similar branch lengths for all, but given the sometimes rather low support values of several nodes, one could perhaps justify including all in Buthraupis.

In summary, this proposal breaks into several subproposals:

A. Divide the Tangara group of Sedano & Burns into five genera, as proposed above: Tangara, Ixothraupis, Chalcothraupis, Thraupis and Euschemon. I recommend a YES. A NO vote would be for maintaining all of these (including Thraupis) in a very broad Tangara (as recommended by Sedano & Burns).

B. Maintain a moderately broad genus Tangara, but as restricted above. I tentatively recommend a YES. A NO vote would favor subdividing the restricted Tangara further; the five-way split I suggested above would seem the most reasonable alternative but others are possible, such that a new proposal would be required specifying two or more alternatives.

C. Recognize the genera Stephanophorus, Diuca, Neothraupis, Lophospingus, Cissopis and Schistochlamys and Paroaria. While this might seem like oversplitting, most of the nodes dividing this group are fairly basal and all are very distinctive morphologically. I recommend YES; a NO vote would favor lumping of some of them, presumably starting with Schistochlamys and Cissopis (and if the NO wins, a set of new proposals would be needed to determine which and how many lumpings we favor).

D. Continue to recognize Wetmorethraupis and Bangsia as separate genera (YES) vs. lumping Wetmorethraupis into Bangsia (NO).

E. Lump Thraupis bonariensis into Pipraeidea ; I recommend YES (a NO vote would require devising a new genus for the former). For the moment, I leave open the question of “Saltator” rufiventris for want of sufficient data.

F. Lump Delothraupis into Dubusia. I recommend a YES; a NO would be for maintaining them as separate, monotypic genera.

G. Recognize the genera Sporathraupis for Thraupis cyanocephala, Tephrophilus for Buthraupis wetmorei , Compsocoma for Anisognathus somptuosus and notabilis, and Anisognathus for igniventris, lachrymosus and melanogenys, since they all represent segments of a basal polytomy and are therefore equivalent (at least with current evidence); I recommend a YES. The alternative (NO) would be to lump all four groups into Anisognathus

H. Recognize Buthraupis for montana, Chlorornis for riefferii and Cnemathraupis for eximia and aureodorsalis. I tentatively recommend a YES. A NO would favor either two or three genera, as detailed above, and would require a new proposal.

Perhaps fortunately, this set of proposals, as it stands, would not require erecting any new generic names, although a number of older generic names would now be resurrected; any further splitting (as in the still-broad Tangara) would require naming at least one new genus. I have not presented separate proposals in which the phylogeny is concordant with the current classification, as in the recognition of Chlorochrysa and Calochaetes; I assume that these would be noncontroversial. I also did not treat specifically the case of “Saltator” rufiventris, which current evidence indicates is a tanager, probably related to Pipraeidea. This will merit a separate proposal when more evidence accrues. To summarize, I recommend YES votes on all eight subproposals.

Literature Cited

Burns & Naoki 2004 (see proposal 291)

Hellmayr 1936: Catalogue of Birds of the Americas, Part 9.

Ridgway 1902: Birds of North and Middle America, part 2.

SEDANO, R. E., AND K. J. BURNS. 2010. Are the Northern Andes a species pump for Neotropical birds? Phylogenetics and biogeography of a clade of Neotropical tanagers (Aves: Thraupini). Journal of Biogeography 37: 325–343.

Note: (for visualizing most of these beasties, Isler & Isler 1999 is very helpful; see Narosky & Yzurieta 1987 for a drawing of “Saltator” rufiventris).

Gary Stiles, May 2010

Comments solicited from Kevin Burns: “I thank Gary for proposing generic-level classification changes based on the paper I wrote with Raul Sedano on the "core tanagers". As the committee might guess from reading our paper, I don't agree with most of the recommendations. However, many of them I do find acceptable. I have asked Raul Sedano to provide comments separately, as his opinions might differ from mine.

When considering potential taxonomic changes as a result of our new phylogeny, we tried to follow these guidelines:

1) keep things as simple as possible and make as few changes to the taxonomy (i.e., avoid proposing new generic names or resurrecting old ones that would be unfamiliar).

2) keep the taxonomy consistent with strongly supported nodes in the tree (i.e., no paraphyletic genera and no genera that might easily be rendered polyphyletic with additional data).

3) avoid naming new monotypic genera. Monotypic genera don't tell you anything about relationships to other taxa. All you learn from having a monotypic genus is that whoever recognizes the genus thinks that particular species is morphologically divergent from everything else. To me, this is often a subjective call and that is why I prefer classifications that recognize cladogenesis (nodes) over anagenesis (apomorphies along a branch that aren't shared).

We basically only recommended taxonomic changes when the structure of the tree required us to do so. Our recommendations for taxonomic changes in the group are pretty well spelled out in our paper. Rather than repeat them all here, I would ask that the committee see the discussion in our paper, in particular page 336.

Below I will give my opinion on each of the proposals

A. Divide the Tangara group of Sedano & Burns into five genera, as proposed above: Tangara, Ixothraupis, Chalcothraupis, Thraupis and Euschemon. I recommend a YES. A NO vote would be for maintaining all of these (including Thraupis) in a very broad Tangara (as recommended by Sedano & Burns).

I would vote "no" to this proposal. I think the suggested change represents a pretty radical departure. The name Tangara is an incredibly useful and a familiar word to many Neotropical ornithologists and birders in general. If this taxon were to be split up into all these subparts, we would loose the ability to conveniently talk about this taxon as a group. Yes, the Thraupis that are embedded within Tangara are different from the other members of Tangara, but not so different as to warrant sacrificing Tangara itself. In addition, I am very concerned about Euschemon (the genus proposed for palmeri through cucullata). The support for this node is only 0.85 Bayesian posterior probability and 51% bootstrap (basically no support). Further analyses and additional data could easily render this group paraphyletic.

B. Maintain a moderately broad genus Tangara, but as restricted above. I tentatively recommend a YES. A NO vote would favor subdividing the restricted Tangara further; the five-way split I suggested above would seem the most reasonable alternative but others are possible, such that a new proposal would be required specifying two or more alternatives.

I don't think Tangara should be subdivided for the reasons outlined above.

C. Recognize the genera Stephanophorus, Diuca, Neothraupis, Lophospingus, Cissopis and Schistochlamys and Paroaria. While this might seem like oversplitting, most of the nodes dividing this group are fairly basal and all are very distinctive morphologically. I recommend YES; a NO vote would favor lumping of some of them, presumably starting with Schistochlamys and Cissopis (and if the NO wins, a set of new proposals would be needed to determine which and how many lumpings we favor).

I agree with this proposal. This is basically sticking with the status quo for these genera and our phylogeny is consistent with all of these genera. For that reason, we did not recommend any changes to classification within this clade.

D. Continue to recognize Wetmorethraupis and Bangsia as separate genera (YES) vs. lumping Wetmorethraupis into Bangsia (NO).

I agree with this proposal. Bangsia is monophyletic, and thus we see no reason to change the existing taxonomy here.

The remaining proposals, E - H, relate to the clade containing Pipraeidea to Buthraupis eximia (Fig. 2). In our paper, we recommended that all of these be placed in a single genus, Iridosornis (which is the earliest name). One reason we did this was that species in Buthraupis and Thraupis were spread across the group, and we wanted to avoid using a bunch of new or resurrected generic names. Plus, using a single genus name for all these species provides an opportunity to highlight their shared distributions (mostly Andean) and evolutionary history. I think having a single scientific name would facilitate and promote their study as a single group of "mountain-tanagers".

For the reasons outlined in the paragraph above, I would prefer the committee vote no to proposals E-H and instead merge all these species into Iridosornis.

That said, I realize this opinion might not be popular with the committee, so I did think hard about each of these individual proposals. I do think Gary's proposals for this clade offer a way to add only a few names, while retaining many of the traditional genera. For that reason, I might be ok with E, F, some parts of G, and H. For proposal G, I do not think there is enough evidence to split Anisognathus at this point. As we mention in our paper, although we don't have evidence for a monophyletic Anisognathus, we also don't have evidence against a monophyletic Anisognathus. The two clades of Anisognathus may very well connect together with additional data, so it's probably better to stick with the status quo at this point. I would be ok with other aspects of G (Sporathraupis and Tephrophilus).

To summarize, for the clade containing Pipraeidea to Buthraupis eximia, I would prefer a single genus Iridosornis, but if the committee is really opposed to this, I would be ok with partitioning these species into these genera:

Pipraeidea

Iridosornis

Calochaetes

Dubusia

Tephrophilus

Sporathraupis

Anisognathus

Buthraupis

Chlorornis

Cnemathraupis

Regarding Saltator rufiventris, we didn't include it in Sedano and Burns, but this species is sister to Delothraupis/Dubusia when included in an analysis with all the core tanagers. So, the committee could safely merge Saltator rufiventris into Dubusia at this point.

Again, thanks for the opportunity to comment. I will be very interested to see how the committee votes on this proposal. What we found in this group is pretty representative of tanagers as a whole (i.e., very few traditional genera that are monophyletic).

Comments from Thomas Donegan: “This is a comment on the treatment for the Anisognathus group (part of Proposal G). In Donegan & Avendaño (2010) we looked into the Compsocoma / Anisognathus issue in some detail, but regrettably various paragraphs ended up being excised from the manuscript following peer reviewer comments concerning the scope of the paper.

“As things currently stand, Poecilothraupis looks more likely to be the senior name for this group than Anisognathus. Three names (Compsocoma, Anisognathus, and Poecilothraupis) were all described within a maximum period of 2 years. It would seem that the name Anisognathus was first published only on the text below a plate, on a date before the date on which Compsocoma and Poecilothraupis were published. However, the text describing the genus Anisognathus came out after Compsocoma and Poecilothraupis were published. The date of publication of each of these names is in any event very difficult to pin down. One suspects that a messy Commission case may be needed to sort this out, e.g. by fixing dates or suppressing Poecilothraupis and/or Compsocoma. A discussion on the priority of genus names for this group is included in a separate document, in case this is of interest to committee members or others who want to pursue this matter further.

“Whatever one’s interpretation of the molecular data, the genus Compsocoma is well-defined morphologically and behaviourally. We noted in Donegan & Avendaño (2010) that: “Blue-winged Mountain Tanager A. somptuosus and Black-chinned Mountain Tanager A. notabilis are more robust birds with stronger flight … but were lumped (with little justification) into Anisognathus by Meyer de Schauensee (1966). We treat Compsocoma as a subgenus of Anisognathus herein.” Where the two genera occur together, Compsocoma is often in higher forest strata than Anisognathus and is more mobile. Several taxa within both a narrow Compsocoma and Anisognathus may require species rank and some of them have been split by modern authors (e.g. C. somptuosus), so the genera produced will probably not be as small long-term as one might initially think. If one were to ignore history and act only rationally, then recognition of Compsocoma would be sensible based on comparative morphological differences between other tanager genera and now poor support for a monophyletic Anisognathus. However, having seen what happened to Pipromorpha, I appreciate that historical treatments are considered important by many persons.

“On balance, I like the idea of recognising Compsocoma. However, the main point of this comment is this: moving a whole load of additional species into Anisognathus (as possible under Proposal G but not recommended by either Gary or Kevin) makes for a potentially unstable and unwelcome scenario. If unifying these species were to be thought sensible as a concept (thankfully, it would seem not), it might be worth waiting to sort out name priority issues first.

“Reference:

Donegan, T.M. & Avendaño, J.E. 2010. A new subspecies of mountain tanager in the Anisognathus lacrymosus complex from the Yariguíes Mountains of Colombia. Bull BOC 130(1): 13-32.

Comments solicited from Raul Sedano: “Thanks to Gary for bringing an interesting proposal and the opportunity to provide comments on it. Overall, I do not agree with most the recommendations; however, I find many of them practical to implement. Kevin covered many notions from the paper but I also would like to add few observations from P#437 and the emphasis that the sequence from Sedano and Burns might imply for this group.

“Stiles' Proposal (P#437) for delimiting Core Tanagers in roughly 23 to 27 genera is substantially different from the 11 genera proposed as modification of current sequence in Sedano and Burns (2010). They are also different in the level at which subclades are designated as genera. This conflicting taxonomic view between P#437 and Sedano & Burns (2010) could imply that genera designation in this case will easily end up being a matter of taste and consistency (for example see the interesting discussion in P#417).

“To add on Kevin's comments on P#437, in our paper, subclade designation for few genera also emphasized the homoplastic nature of external phenotypes in the core tanager radiation. One example of this emphasis is the notion of a species rich genus "Iridosornis" with remarkable species variation in i.e. color and body size. This clade "Iridosornis" could also highlight a shared distributional pattern largely in the Andes that resulted from evolutionary processes in and out of the Northern Andes or likely due to restricted gene flow along the Andes itself. Thus, Iridosornis somehow reflects patterns that are the result of long-term evolutionary processes. Designating this as a genus would have the potential of being also an informative key name as the well-played role of the genus name Tangara, for communication among ornithologist and birdwatchers' communities.

“In contrast, Gary's proposal P#437 urges for more fine subclade recognition, emphasizing an eclectic collection of anagenetic diagnostics and with no emphasis whatsoever in share evolutionary patterns among the resulting multiple genera.

“Thus, I rather fewer genera as in Sedano & Burns that coincide with the following three overall diagnostic themes on the evolutionary patterns:

“1. We learned something about the broad relationships of nine monotypic genera and we were able to reduce that taxonomical uncertainty in some extend to four monotypic genera. Sure, far from perfect solution but P#437 will end up with a taxonomical sequence with 7 monotypic genera (some of them resuscitated names) and nothing about their broad phylogenetics is told with that genera sequence.

“2. Taxa whose cladogenesis is concentrated early in their histories partition more of their morphological disparity among, rather than within subclades. Thus, the suggested sequence in our paper reflects this expected pattern. In fact, Tangara and Iridosornis are quite different form each other as you might think for each of the five monotypic genera and Paroaria, Chlorochrysa, Lophospingus and Schistochlamys (This is a generalization under current research). Here, I disagree with P#437 because it may be overemphasizing among subclades differences based on a disparate collection of traits.

“3. The larger two genera (Tangara and Iridosornis) somewhat hold a geographical context among clades and their distinctive evolutionary patterns, in and out of the Andes as a whole. (Iridosornis would show a more restricted pattern to the Andes than Tangara or than any other genera in the core tanager clade)

“That said, this is my opinion on proposal A-H in P#437.

“Subproposal 437 A: I would vote "no", I agree with Kevin comments

“Subproposal 437 B: I do not think Tangara should be subdivided BUT if P437 ends up subdividing its sister clade then why not to subdivide Tangara as well.

“Subproposal 437 C: I would vote "no" (waiting for the new set of proposals)

“Subproposal 437 D: I would vote “yes.”

“Subproposal 437 E.: Lump Thraupis bonariensis into Pipraeidea ; I recommend YES (a NO vote would require devising a new genus for the former). For the moment, I leave open the question of "Saltator" rufiventris for want of sufficient data.

“Subproposal 437 F: Lump Delothraupis into Dubusia. I recommend a YES; a NO would be for maintaining them as separate, monotypic genera.

“Subproposal 437 G: I agree with K. Burns comments on this subproposal. To summarize, for the clade with origin on the most recent common ancestor between Pipraeidea to Buthraupis eximia, I would rather a single genus Iridosornis, but if the committee is really opposed to this, I would be ok with partitioning these species into these genera:

Pipraeidea

Iridosornis

Calochaetes

Dubusia

Tephrophilus

Sporathraupis

Anisognathus

Buthraupis

Chlorornis

Cnemathraupis

“Thank you all for the opportunity of commenting in this taxonomical sequence.”

Comments from Robbins: “I presume we are voting on each subproposal. In general, I support Kevin Burns and Raul Sedano’s guidelines for their proposed taxonomic changes. Hence, I vote as follows:

a. “No” to dividing up Tangara into five genera.

b. “No”; continue to recognize Tangara in the broadest sense.

c. “Yes”. Burns and Sedano’s data support the status quo treatment of these genera.

d. “Yes” for maintaining Wetmorethraupis and Bangsia.

e. “Yes” for placing Thraupis bonariensis into Pipraeidea.

f. “Yes” for including Delothraupis in Dubusia.

g. “No”, given Kevin’s cautionary comments on prematurely splitting up Anisognathus, but I would support recognizing Sporathraupis for Thraupis cyanocephala and Tephrophilus for Buthraupis wetmorei.

h. “Yes”.

Comments from Pacheco: “

a. “No” em dividir Tangara em cinco gêneros, considerando que K. Burns colocou.

b. “No”; eu não acho que Tangara deve ser subdividido.

c. “Yes”. Há suporte para manter tais gêneros.

d. “Yes” para continuidade de Wetmorethraupis e Bangsia em nível genérico.

e. “Yes” para subordinar Thraupis bonariensis ao gênero Pipraeidea.

f. “Yes” para inserir Delothraupis em Dubusia.

g. “No”, considerando que K. Burns sugeriu em comentário.

h. “Yes”. Tentativamente eu apoio esta proposição.

Comments from Remsen: “I am not going to vote at the moment because I have solicited comments on these potential changes from a number of field people who know these birds as well as anyone. Because much of this involves subjective decisions on where to draw genera, I want to see those subjective opinions from key people.

“Meanwhile, I will speak on behalf of monotypic genera. Although widely disliked for the reason Kevin pointed out above, namely a monotypic genus conveys no information on relationships, I would counter by saying (1) we can’t expect all relationships to be reflected in binomial nomenclature, and (2) a monotypic genus DOES convey the information that the morphology is so different that to merge that species into an existing genus would render that genus undiagnosable in terms of phenotypic characters … for me, that is important information. Also, monotypy in many cases is likely a product of extinctions that pruned out other branches. Or it might be an artifact of taxonomy – what if the various subspecies in Eucometis were elevated to species rank? Nothing changes, really, in terms of biology – only taxon-ranking. Would Lanio somehow become a “bad” monotypic genus if its components were re-ranked as subspecies? Finally, I think everyone agrees that monotypic genera are required if the relationships are uncertain, but by implication, this would mean that once sister relationships are determined, and then the genera should be merged. By extrapolation, this would mean that once all the nodes in the Tree of Life are worked out, all living organisms could be placed in a single genus, with genera all collapsing into each other every time a node is solidified. Obviously, such hyperbole is only meant to point out that there is indeed a problem with how far one takes the process of merging monotypic genera once relationships are determined. Just in case you’re worried, let me make it clear that I’m not a champion of large numbers of monotypic genera, and favor them only in cases in which their merger into another genus creates a unit so heterogeneous that there is no way to provide a non-genetic diagnosis for the genus. Once we have calibrated trees for families of birds, I am optimistic that we can also add another criterion in determining generic limits, namely relative lineage age.”

Additional comments from Stiles: “Just a note on my reasoning through some recent proposals. I feel that there may be a certain lack of clarity regarding the possible roles of taxonomy vs. systematics. For what it's worth, here is my take on this. Systematics deals with evolutionary relationships, in particular the establishment of monophyletic groups at different levels of the taxonomy, and I agree that taxonomy should reflect this - i.e., not propose groups that are not monophyletic. However, given that the decisions on when to lump monophyletic clades into larger monophyletic groupings remain somewhat subjective, I feel that the best decision is that which reflects the most ancillary information regarding other attributes of the clades/taxa in question. I agree with Mayr on this one - given the limits on subjectivity set by monophyly, I would decide in favor of the groupings that reflect the most information on morphology, behavior, ecology, etc. provided that this does not create great discordance in the equivalence of groups at the same level of the hierarchy. In this sense, the taxonomy becomes more heuristic and predictive. In the final analysis, taxonomy is systematics’ window on the world, what the rest of the world sees and uses. Moreover, this window becomes especially important at the generic level, since the name of a species is the names of genus + species. Thus, when the question arises as to whether to split a monophyletic group into two or more monophyletic genera or lump all into a single genus, I try to choose the alternative that provides the most information on other aspects of the biology of the birds concerned. Hence, I proposed to split Diglossa and Diglossopis because this split carries with it information on aspects of morphology, behavior and ecology that differ between the two groups; the fact that some branch lengths are short says to me that these differences were acquired relatively soon after the groups split. Similarly, on the tanager proposal, I proposed to split Tangara rather than maintain Thraupis within it because I feel that Thraupis evolved a series of features setting it apart - larger size, plainer plumage pattern with even less sexual dichromatism than occurs in Tangara, loud squealy vocalizations freely given rather than the more discreet utterances in Tangara, adaptation to drier, more open or secondary habitats, more in the lowlands rather than the subtropics where Tangara is most diverse. Given this shared evolutionary history (sensu Burns & Sedano), I consider Thraupis worthy of generic recognition. In turn, in order to avoid a paraphyletic Tangara, splitting this genus becomes mandatory. Perhaps fortunately, there are names available for all of the smaller but still monophyletic genera that this would create, as outlined in my proposal, and there are some features of plumage and/or morphology that fit the more restricted genera. I have tried to apply this logic in the rest of the proposal as well to avoid larger, poorly defined and heterogeneous genera that carry little additional information than that they are, by current information, monophyletic. I have no objection to monotypic genera if the species concerned are very distinctive compared to their nearest relatives - for instance, I would maintain Cissopis and Schistochlamys as genera even though they are sisters because they are such different birds in plumage, ecology, vocalizations etc., and I prefer to place "Thraupis" cyanocephala in its own genus because it differs rather strongly from what are now determined to be its closest relatives (again, an appropriate name is available). On the other hand, I would be inclined to lump Thryophilus and Cantorchilus because as far as I can tell, the biology and morphology of these two is so similar, aside from the somewhat subjective difference in songs. Thus, I will run this rationale up the proverbial flagpole and see whether other members of the committee salute it or shoot it!”

Comments solicited from Dan Lane:

“A) If a gun were pointed at my head, my own inclination would be to maintain Tangara as much as possible, and I would tend to agree with Burns and Sedano that there is not much reason to maintain a large and reasonably heterogeneous Tangara to the exclusion of the other (sub)genera that Gary proposes splitting off just to prevent losing 'Thraupis'. To me, Thraupis could fit comfortably into Tangara (if inornata is in there, coloration is not an issue... it's really just size that differentiates Thraupis). As a sidebar, I would love to know where T. rufigenis fits into this clade!

B) See A.

C) I agree with maintaining these genera apart. They are quite distinctive, and the branch lengths seem to support old relationships.

D) I still don't know Wetmorethraupis in life, and only know Bangsia poorly. I could go either way on this.

E) My inclination is not to lump these two into Pipraeidea. They really don't strike me as being very similar (in behavior, voice, coloration). They both show sexual dimorphism and both can be found in more arid environments than most core tanagers can endure, but to me the similarities end there. I'd prefer a monotypic genus for 'Thraupis' bonariensis.

F) This seems good to me. Although I still don't really know Dubusia taeniata well. Vocally, and habitat-wise there are similarities; behaviorally I'm not so clear, but in my experience they aren't so different.

G) I agree with this suggestion, with the caveat that perhaps further study may result in 'Compsocoma' being returned to 'Anisognathus'.

H) I agree with this suggestion.”

Comments from Zimmer:

“A) YES. I’m with Gary on this, particularly as regards his “Additional comments” on the subjectivity of when to lump monophyletic clades into larger monophyletic groups. I also share his vision for what constitutes a genus. Accordingly, I just can’t see merging Thraupis into Tangara. The species of Thraupis embedded within Tangara form a morphologically, vocally, and ecologically cohesive group that differs dramatically from everything else in Tangara. Kevin Burns is correct in his assertion that the name Tangara is a very useful and familiar word to Neotropical ornithologists and birders, but I would argue that its usefulness to both camps would be drastically compromised if the 7 Thraupis were included under its banner. At the same time, we would be losing Thraupis, which is also one of the most familiar, recognizable tanager groupings to ornithologists and birders. Both camps already refer to the “spotted group” as a natural subgroup of Tangara, so I think the proposed Ixothraupis would work. The proposed “Chalcothraupis” is divergent, so I don’t have a problem with it being a monotypic genus. My only hesitation would be with Gary’s proposed Euschemon, because I’m not as confident that they would constitute a natural grouping. But, Gary’s rationale for this grouping sounds sensible, and I am opposed to the alternative, which would be to retain all of the current Tangara and then fold in the seven species of Thraupis.

B) YES.

C) YES. Each of these genera, as currently defined, makes perfect sense to me. I really can’t see throwing any of them together, especially when you look at the morphological and vocal cohesion between species of Schistochlamys or species of Paroaria.

D) YES.

E) NO. Pipraeidea and T. bonariensis are so different vocally that I can’t come to grips with the idea that they would be in the same genus.

F) YES.

G) YES, although I don’t have strong feelings about it either way.

H) YES, but ditto the previous comment.”

Comments from Steve Hilty: “Gary Stiles’ interesting proposal seeks an alternative that would bring some resolution to the recent molecular genetic findings of Sedano and Burns (2010) on Thraupidae. Before commenting I need to try and cleanse myself of about forty years worth of prejudices (I’m guessing I am not alone here) and examine the facts (such as they are) through virgin eyes. Having almost completed HBW’s tanager tome, I have some insight into the issues, but I bring my own baggage to this discussion. Also, for brevity, I confine my remarks to Tangara and the approximately eight or so current Andean mountain-tanagers (Bangsia, Iridosornis, Anisognathus, Buthraupis, Chlorornis, Calochaetes, Dubusia, Pipraeidea, and Wetmorethraupis), hereafter all just called “mountain-tanagers,” that were the focus of part of the Sedano and Burns manuscript and Gary’s proposals. There are plenty more genera for future discussion.

“Initially I read Raul Sedano and Kevin Burns’ manuscript (2010) with something akin to horror (all my cherished tanager genera disappearing?). Then I read Gary’s’ insightful proposals and realized that I needed to step back and see if I could objectively reconcile these proposals and counter-proposals with the birds in life (but not in a museum tray). Raul Sedano and Kevin Burns found that Tangara is monophyletic assuming one inserts six former Thraupis in their midst; that is a little bothersome, but I’m getting used to it. They also find that a large clade (actually two clades) of mountain-tanagers (above) is best viewed as a single monophyletic group placed, by rules of priority, in Iridosornis. At this point the molecular genetics team has played its tanager hand, and their cards are on the table.

“Gary, on the other hand, has proposed a compromise, part of which is almost precisely the opposite of the suggestions made by Sedano and Burns. His proposal splits Tangara into a half dozen genera and preserves a collection of mountain-tanager genera. His proposals make sense, however, because he seeks to preserve similar node lengths for both groups on the phylogenetic trees and, in the process, do less damage to the existing taxonomy, but on this latter point I am not so sure. In any case, his proposals are consistent and well reasoned. The question now, for those of us judging from the sidelines (and in my case without the benefit of a strong molecular background), is what do biologists/taxonomists do now? In other words, where does one draw the taxonomic lines? And, this is where it becomes difficult to shed old prejudices.

“Tangara have long been presumed to comprise a relatively uniform, easily recognized genus. Or, to borrow a cliché, at least, we all thought we knew one when we saw it—or so it seemed until Thraupis came along. But, as the Islers long ago, and Gary more recently noted, the genus comprises up to 13 discrete groups that separate rather well by plumage, as well as by foraging behavior and, to some extent, also by habitat. Remarkably, most of these same species-groups also were revealed by the molecular work of Sedano and Burns. So splitting Tangara isn’t that far-fetched—except for the fact that molecular genetic evidence also clearly supports the notion that they are monophyletic when the six Thraupis are added.

“Looking at the mountain-tanager genera, we also all thought we knew a Bangsia, an Iridosornis, an Anisognathus and most of the others when we saw them, but molecular work suggests that maybe we don’t. The puzzling aspect of Gary’s proposal is this: why would one chose to split one genus (Tangara) which is monophyletic, in order to preserve, as distinct, a collection of other genera (various mountain tanagers) which, on face value, differ little more in size, plumage, and behavior, than do the various members of Tangara. Damage to present nomenclature may be part of the answer but, setting that aside, I think it is here that we have difficulty overcoming past experience and struggle with objectively (I know that I do). Nevertheless, to briefly illustrate a couple points: some Tangara differ markedly in size (e.g. T. palmeri is almost 3x the weight of T. varia). In fact, sizes and weights differ almost as much in Tangara as they do the eight mountain-tanager genera that Sedano and Burns would magically cause to disappear. Secondly, the “clash” of colors within Tangara (between the various species groups) is at least as dramatic as the color “clash” within the various species-groups of mountain-tanagers—maybe more so, but that may be in the beholders eyes and I am not sure colorfulness can be easily quantified. Additionally, I suggest that the various foraging techniques employed by Tangara are probably more varied than those employed by the eight current genera of mountain-tanagers under discussion, as is the range of habitats they occupy. Interestingly, one piece of evidence that may not be helpful here is voice. We are still trying to discover if some of these birds even sing (or, at least, discover what constitutes a song).

“I conclude that the original proposal of Sedano and Burns (2010) may not be as radical as it seems at first read. At least, from a field observers’ viewpoint, I rather like the idea that all of these little loose ends are related and could be tied up together in a single biogeographical /ecological bundle. From a biogeographical standpoint a single Andean super-genus (Iridosornis) actually makes some sense, and the molecular evidence seems to provide support—at least for now. [I let go of Chlorophonia and Euphonia as they migrated elsewhere, so this genetic two-step doesn’t seem like such a leap of faith anymore.]

“Rebuttal to these comments will almost certainly center on the issue of lumping numerous “distinctive” genera into larger genera that become too large, too heterogeneous and non-diagnosable. Diagnosability, of course, is a function of what criteria we choose in the first place. Traditional taxonomy has long relied on morphology—e.g. bill shapes, plumage color and pattern and so on. Molecular studies reveal that these older criteria have often worked well but are not consistently reliable, and especially not with New World nine-primaried oscines. Ultimately the decisions that we make are (whether we care to admit it or not) influenced by criteria that are not easily quantified and more a matter of taste. We want them (the genera, etc.) to look like they belong together. Another frequent criticism of binary branching trees is the difficulty of defining clades (which, as far as I can tell, can represent just about any taxonomic unit you wish from species to family), an accepted process for naming these units is not clear, and the system depends completely on inherited traits so incomplete sampling can become as issue. Nevertheless, molecular science may to be telling us something else, urging us to probe new boundaries. It is worth considering alternatives, hopefully while also looking through a lens that is not clouded by the baggage of history that we all carry.

What would I do here? I’d preserve the genus Tangara and take on board those ex-Thraupis as well. Regarding the mountain-tanagers (all eight or so genera), I might adopt a wait-and-see approach here before we get busy and start completely rewriting history. There is still a lot of work to be done and there will surely be opportunities to make changes in the future once we are absolutely bludgeoned with irrefutable evidence and dragged against our wills into the taxonomic 21^st century.”

Comments from Cadena: “I apologize to committee members and other interested parties (especially Kevin and Raúl!) for my long delay in getting to vote and comment on this proposal; it is so difficult and important that I needed quite a bit time to digest it thoroughly and it was hard to find such time. OK, so here are my opinions, starting with a general remark.

“In contrast to many committee members and other contributors to this most interesting discussion, I am not familiar with many of the species in terms of their voices, behavior, etc., so my votes and comments are largely based on this premise: I think we need make the fewest possible changes to classification, and only change when well supported nodes in the phylogeny indicate that the current classification is inconsistent with phylogenetic relationships. Thus, while I appreciate Gary's effort (and found his discussion of the roles of taxonomy vs. systematics most interesting and instructive!), I generally tend to favor more the "conservative" positions of Sedano and Burns. So, here we go:

“A-B: NO. I think I can live with having a large-bodied clade (i.e. "Thraupis") nested within Tangara. I would argue this is a minor cost relative to the cost of recognizing multiple genera for birds that we have always recognized as "typical" tanagers. If you know what I mean, a Tangara is a Tangara, except (now) when it is a Thraupis. I realize, as Gary and Kevin point out, that there are traits that would allow diagnosis of different genera if they were to be recognized, but this would require a change in mindset by ornithologists and birders that would be difficult to achieve and get used to. That said, there would be far less changes in names if we simply lump Thraupis into Tangara, so I would suggest this is a simpler course of action causing the least possible disruption to classification. Therefore, I agree with Sedano and Burns on this issue.

“C. YES. Because Paroaria is monophyletic, no changes are necessary here. Maybe one could say there are too many monotypic genera here and thinking of groupings could be good if one were starting to construct a classification from scratch. But considering historical momentum (and the distinctiveness of these lineages, as Van pointed out), I think it is best to make no changes.

“D. YES. Again, because Bangsia is monophyletic, there is no need to change anything.

“E. YES. The only other alternative would be to erect a monotypic genus. Having the two species in a single genus highlights they are sister to each other, information that would not be evident from classification if we establish a monotypic genus.

“F. NO. In rigor, there is no need to change here because one could continue to recognize the two monotypic genera and this would still result in classification being consistent with phylogeny. More importantly, this proposal is absorbed by my take on proposals G and H.

“G-H. NO. These are the hardest ones. Sedano and Burns have suggested lumping a large group of species in a single, heterogeneous genus, and committee members have generally taken issue with this. Gary has proposed an alternative in which multiple (perhaps too many) genera are recognized. Gary also mentioned in passing the possibility of lumping T. cyanocephala, B. wetmorei and Anisognathus in a single genus, but I note this could be risky because support for the monophyly of such a clade is not strong (i.e., 0.74 posterior probability, 43% bootstrap). Steve Hilty's comments highlighting that the heterogeneity of the large genus proposed by Sedano and Burns is not as dramatic, especially when one considers the heterogeneity of the broadly defined Tangara. I also like the idea of having a single genus of mountain-tanagers, which has the benefit of being quite strongly supported by the molecular data. In sum, I would go with Sedano and Burns and recognize a single genus for the clade defined by the most recent common ancestor of Calochaetes coccineus and Buthraupis eximia.”

Comments from Stotz:

A YES
B YES
C YES
D YES
E YES
F YES
G NO
H NO

“The question of whether to split up Tangara or accept Thraupis as part of Tangara is a difficult one. I just find it hard to think of Thraupis as being Tangara. I think I am much more comfortable with recognizing the smaller units that would have to be split out of the narrow Tangara than bringing the loud and obnoxious Thraupis into Tangara.

“On the Mountain-Tanagers (G and H), I was voting against the further splitting of those groups and favoring the retention of a broad Anisognathus and a broad Buthraupis. But the idea of a broad Mountain-Tanager genus as suggested by Burns and Sedano is very intriguing to me. It would be somewhat heterogeneous, but in a different way than the grande Tangara would be. It would be composed a number of currently small genera that don’t just don’t cry out for maintaining. It makes biogeographical sense, and I think also ecological sense.”

Comments from Pérez-Emán: “This is a difficult proposal involving a large and heterogeneous group and my decisions are based on the need to have an informative classification while following guidelines coming from Kevin & Raúl’s phylogeny.

“A: A reluctant NO. I don’t like the idea of lumping Thraupis into Tangara and I feel Gary’s five genera option provides with a good alternative. However, resurrection of Chalcothraupis and Euschemon might render classification unstable based on low node support. If we focus just on well-supported nodes, we run again into the problem of including Thraupis into a second group of “Tangara”. An expanded molecular data set, including nuclear genes, might be helpful to improve node support or provide alternative scenarios more congruent with our conflicting views of tanager relationships.

B: NO as a consequence of previous comments.

C: YES. It keeps classification as it is while retaining several divergent species in their monotypic genera.

D: YES. No need to include these species into one genus, especially when support for the node leading to both Wetmorethraupis and Bangsia is low.

E: A tentative YES. Support for lumping these two taxa into one genus is strong. However, to consider each taxon as different species of one genus or two genera depends on the knowledge we have on these taxa and on the weight we give to such information. As I don’t know Thraupis bonariensis and my experience is limited with Pipraeidea, it could go either way, and my decision is based only on the molecular phylogenetic hypothesis.

F: NO. I would keep both species in their own separate genera, which would be a decision congruent with C.

G and H: YES. Though these proposals increase the number of genera in the group by resurrecting many names in order to comply with the phylogenetic hypothesis, I think it does a good job to preserve an important level of information. The option of a large Iridosornis, including option from E to H, might be unstable as support for the monophyly of such clade is ambiguous (high posterior probabilities but very poor bootstrap support). If this proposal is accepted, further research on priority between Anisognathus and Poecilothraupis should be pursued (as Donegan pointed out).”

Comments from Perez-Emán: “This is a difficult proposal involving a large and heterogeneous group and my decisions are based on the need to have an informative classification while following guidelines coming from Kevin & Raúl’s phylogeny.

B: NO as a consequence of previous comments.

C: YES. It keeps classification as it is while retaining several divergent species in their monotypic genera.

D: YES. No need to include these species into one genus, especially when support for the node leading to both Wetmorethraupis and Bangsia is low.

E: A tentative YES. Support for lumping these two taxa into one genus is strong. However, to consider each taxon as different species of one genus or two genera depends on the knowledge we have on these taxa and the weight we give to such information. As I don’t know Thraupis bonariensis and my experience is limited with Pipraeidea it could go either way and my decision is based only on the molecular phylogenetic hypothesis.

F: NO. I would keep both species in their own separate genera, which would be a decision congruent to C.

Comments from Jaramillo:

“A. YES – divide Tangara into five genera. I think that this is preferable to a hodge-podge Tangara that is much more difficult to define.

B. Confusing here, but assume that this is a NO if I voted for a YES for “A”.

C. YES. Retain these smaller genera.

D. YES – Maintain Wetmorethraupis.

E. YES – Hard to get used to, but the two are similar in many ways, although not vocally.”

F. YES – they are quite similar really.

G. YES – This can be re-arranged later if there is new information, but it seems like the way to go give the available data.

H. YES – Again, this seems like the best way to go given available evidence. Perhaps some of these species may contain more than one species level taxa, and will not be single species genus for long (thinking of montana?).”

Comments from Remsen: “Thanks to Gary for this very useful synopsis and working this all into a sensible set of proposals. Apologies to Gary for taking so long to confront the issues.

A. YES. The lesser of “evils” in my opinion is to resurrect the several genera required to make true Tangara monophyletic, to recognize that current Tangara is actually a heterogeneous group, and to avoid forcing the truly different Thraupis into a broad Tangara.

C. YES. Any mergers here would violate subjective standards of within-genus homogeneity.

D. YES – no reason to change from status quo.

E. YES, if only to avoid another monotypic genus.

F. YES, and this one has plenty of historical precedent.

G. YES.

H. YES.