Proposal (730) to South American Classification Committee

Revise generic limits in the Thraupidae

Burns et al. (2016) proposed major revisions to generic limits in the Thraupidae based on DNA sequence data published primarily by the Kevin Burns lab and collaborators, especially Keith Barker. Burns et al. (2014) and Barker et al. (2015) produced comprehensive phylogenies that together sampled about 95% of the 372 species of true tanagers (Thraupidae). Their genetic sampling included several mitochondrial and nuclear makers. Burns et al. (2016) presented portions of the overall tree relevant to the proposed, and so please refer to that in evaluating the proposal. The major overhaul of generic limits in the Thraupidae to conform to phylogenetic data is an essential first step for studying character evolution and biogeography of this family and requires immediate attention.

Their results revealed numerous problems with traditional genus-level classification in the family. To bring Linnaean classification in line with the phylogenetic results, Burns et al. (2016) proposed 30 changes relevant to taxa in the SACC area, grouped as subproposals as follows:

730.01. Resurrect Rhopospina Cabanis 1851 for Phrygilus fruticeti. Phrygilus (type species = gayi) is highly polyphyletic, with fruticeti sister to Porphyrospiza (“Phrygilus” alaudinus + (“Phrygilus” carbonarius)). These three “Phrygilus” are distantly related to true Phrygilus, and so clearly a change is required. The alternative to the one proposed by Burns et al. would be to expand Rhopospina to include Porphyrospiza and Corydospiza (as in Dickinson & Christidis 2014). An expanded Rhopospina would unite three species with black and gray plumage patterns, but the fourth, Porphyrospiza caerulescens, is all blue, but nonetheless not dissimilar in size and shape. However, the time-calibrated phylogeny of Barker et al. (2015) predicts that fruticeti diverged from the others ca. 7.5 mya, i.e. significantly older than most taxa ranked as genera in the Thraupidae. For example, in this same clade, the split of Stephanophorus from Cissopis + Schistochlamys is the same age as that estimated for the split of fruticeti from the others in this group.

A YES vote on 730.01 would be to follow Burns et al. in resurrecting Rhopospina. A NO vote would indicate a preference for an expanded Rhopospina. I recommend a Yes because of the relatively old age of the split of fruticeti from the rest. Note that Rhopospina is feminine (Dickinson & Christidis 2014) and thus the species names with variable endings (alaudinus and carbonarius) would be changed.

Here are specimens of the four species:

730.02. Resurrect Corydospiza Sundevall 1872 for Phrygilus alaudinus and Phrygilus carbonarius. See discussion in 730.01. The split of Porphyrospiza from Corydospiza is estimated at ca. 4 mya, i.e. borderline but not totally out of line with the age of many thraupid genera.

A YES vote on 730.02 would be to follow Burns et al. in resurrecting Corydospiza. A NO vote would indicate a preference for an expanded Porphyrospiza. I have no strong recommendation. An expanded Porphyrospiza would eliminate a monotypic genus whose placement is certain but would also create a genus that is somewhat heterogeneous in color. However, blue and black plumage is found within other genera, e.g. Diglossa, true Phrygilus (well, not really blue blue), and possibly others, so is not of major phylogenetic significance. Note that Porphyrospiza is presumably feminine and thus the species names with variable endings (alaudinus and carbonarius) would be changed. Note that if we reject 730.01, then Porphyrospiza would be included in Rhopospina.

730.03. Merge Saltatricula into Saltator. Saltatricula multicolor is sister to Saltator atricollis, and this pair is sister to all other Saltator. This issue has come up before with SACC and basically left unresolved (see SACC 344 and 593 for discussion).

A YES vote on 730.03 would be for the merger, and a NO vote would be for retaining Saltatricula and including atricollis in Saltatricula (as in Dickinson & Christidis 2014). The similarities between multicolor and atricollis have been discussed in previous proposals. I recommend a NO vote on this because in the time-calibrated tree of Barker et al. (2015) of multicolor and atricollis from other saltators is estimated at ca. 9.5 mya, i.e. substantially older than that between most taxa treated as genera in the tanagers. I think the combination of this estimate of lineage age combined with the characters previously discussed provides sufficient evidence for recognizing Saltatricula (and including atricollis in it as required by the phylogeny). (By the way, within remaining Saltator, there are two primary lineages that diverged ca. 7 mya, and so if we move towards genera defined by rough lineage age, Saltator is a strong candidate for a split.)

730.04. Merge Tiaris bicolor into currently monotypic Melanospiza and recognize newly named Asemospiza Burns, Unitt, & Mason 2016 for Tiaris obscurus and Tiaris fuliginosus. Tiaris is polyphyletic, and the type species (olivaceus) is sister to a group of largely Caribbean and Galapagos genera, including all Galapagos finches. Tiaris bicolor is sister to Melanospiza richardsoni, a relationship that Burns et al. noted was consistent with phenotypic features (pink legs and dark plumage without facial markings).

Tiaris bicolor:

Melanospiza richardsoni:

A polytomy in this area of the tree complicates alternative solutions. Tiaris obscurus and Tiaris fuliginosus are sisters, but together they are sister to the Galapagos finches, and thus a new genus must be named for them unless they and all genera of Galapagos finches are merged into a single genus; the latter radical treatment would, however, be consistent with the estimated age of the group, i.e. only 2 MY.

A YES vote on 730.04 would be for the merger, i.e. Tiaris bicolor becomes Melanospiza bicolor (as in Dickinson & Christidis 2014) and for recognition of Asemospiza for obscurus and fuliginosus. Note that Asemospiza is feminine and so the species names with variable endings must change in agreement. I strongly recommend a YES on this one because there really is no other viable solution. The only other options, both untenable in my opinion, would be (a) to merge them with all genera of Galapagos finches plus Loxipasser and Loxigilla, or (b) ignore the phylogenetic data to maintain status quo.

730.05. Recognize new genus Islerothraupis for Tachyphonus cristatus, T. luctuosus, and T. rufiventer. The genus Tachyphonus as traditionally constituted is highly polyphyletic. The black and tawny plumage that currently unites them is found in three different lineages in the Burns et al.’s Tachyphoninae. True Tachyphonus (type species T. rufus) consists of three species (phoenicius, rufus, and coronatus) that comprise the sister lineage to Ramphocelus; they are species of relatively dry, open habitats, and two of the three forage fairly near the ground (as in sister genus Ramphocelus) and in contrast to the species in Islerothraupis. The sister group to Islerothraupis is Eucometis + Trichothraupis. Barker et al. (2015) estimated the split of Islerothraupis from Eucometis + Trichothraupis at ca. 8 MYA, i.e. older than most taxa ranked as genera in the tanagers.

A YES vote on 730.05 would be to recognize Islerothraupis and to include in it Islerothraupis cristata, I. luctuosa, and I. rufiventer (note changes in variable endings). A NO vote would indicate a preference for an alternative to recognizing Islerothraupis, i.e. expand Trichothraupis Cabanis 1851 to include also Eucometis and Islerothraupis, as in Burns et al.’s Table 1, Alternative 2. I favor a YES on this to maintain long-standing and distinctive Eucometis and Trichothraupis and to recognize the relatively old divergence of these lineages. However, looking at specimens of Trichothraupis, I can see more similarities between it and Islerothraupis than between the latter and other Tachyphonus.

730.06. Recognize new monotypic genus Maschalethraupis for Tachyphonus surinamus and recognize new monotypic genus Chrysocorypha for Tachyphonus delatrii. See 730.5 for problems with traditional Tachyphonus. These two species are part of a group in which the topology is poorly resolved; they have no obvious sister species, and the only strongly reported node in this section of the tree (see Fig. 2) unites seven genera, including Coryphospingus and Rhodospingus. That surinamus and delatrii are not sister species cannot be refuted from the data, but there is no support for that relationship either.

A YES vote on 730.06 would be to recognize these two monotypic genera, i.e. Maschalethraupis surinamus and Chrysocorypha delatrii. A NO vote would indicate a preference for an alternative but would require explaining what that alternative would be. I am not thrilled about recognizing two new monotypic genera for species, but I do not see an alternative. Additional genetic data may resolve the topology in Fig. 2, but if the current branching pattern is close to reality, but I don’t see any outcomes that would produce a sensible merger of these two into existing genera; the “best” outcome might be that they are each other’s sisters, thus combining the two into one of the new genera. Furthermore, and critical to my views, the ages estimated for the divergence of surinamus and delatrii are ancient by tanager standards, i.e. 7-8 MYA (Barker et al. 2015). So, I recommend a YES on this for lack of viable alternatives.

730.07. Resurrect Pseudospingus Berlepsch & Stolzmann 1896 for Hemispingus xanthophthalmus and H. verticalis. Hemispingus as traditionally defined is polyphyletic … for those who know these birds, this was expected. Hemispingus (type species superciliaris) currently consists of a collection of Andean tanagers that have rather dull plumage and insectivore bills, but they really differ in voice and foraging behavior. They are scattered in 6 different parts of Burns et al.’s Poospizinae tree. Within that tree, these two species are on a branch all by themselves with no certain close relatives. These two species have long been recognized as sister taxa (forming a classic superspecies).

A YES vote on 730.07 would be to resurrect Pseudospingus for these two species (as in Dickinson & Christidis 2014). I strongly recommend a Yes on this one because no viable solution presents itself other than merging multiple phenotypically divergent genera into one huge genus. Further, Barker et al. (2015) estimated the divergence time between this lineage and Cnemoscopus at ca. 8 MYA.

730.08. Merge two species of Hemispingus (H. rufosuperciliaris and H. goeringi) and the two species of Compsospiza (C. garleppi and C. baeri) into a more restricted Poospiza (type = P. nigrorufa) than currently recognized. The other species retained in Poospiza would be boliviana, ornata, hispaniolensis, and rubecula. This unites a group of species defined by a node with high support (Fig. 3). The two ex-Hemispingus are rare and local in Andean undergrowth. The two Compsospiza are also rare and local Andean species that have previously been treated in Poospiza, so that part is not controversial. The four Poospiza are also relatively rare to uncommon and local residents of Andean regions. You can see the ochraceous plumage theme that all members share except hispaniolensis, which is also the one found in the driest habitats.

Hemispingus goeringi:

A YES vote on 730.08 would be to treat H. rufosuperciliaris and H. goeringi as Poospiza rufosuperciliaris and P. goeringi, and Compsospiza garleppi and C. baeri as Poospiza garleppi and P. baeri. The two Hemispingus are deeply embedded in Poospiza, and so trying to maintain them in a separate genus would require resurrecting Orospingus Riley for the two Hemispingus (as in Dickinson & Christidis 2014), retaining Compsospiza, and naming new genera for hispaniolensis and rubecula. Burns et al.’s revised Poospiza encompasses a group of mostly Andean species with some shared plumage themes; the difference in bill shape between the two Hemispingus and the conical Poospiza and Compsospiza bills is irrelevant in my opinion because this feature is highly plastic in birds, particularly tanagers. Therefore, I recommend a Yes on this one.

730.09. Recognize newly named Kleinothraupis for four species of Hemispingus (atropileus, calophrys, reyi, and parodii). See 730.07 for why Hemispingus cannot be maintained as is. Kleinothraupis unites a strongly supported group whose sister relationships are uncertain other than membership in a group of as many as nine other genera, including taxa as different as Cypsnagra, Nephelornis, Thlypopsis, and former members of Poospiza and Hemispingus. The close relationship of these four species has been recognized in traditional classifications. All four are found in relatively high-elevation Andean cloud-forest.

Hemispingus reyi:

A YES vote on 730.09 would be to treat Hemispingus atropileus, H. calophrys, H. reyi, and H. parodii as Kleinothraupis atropileus, K. calophrys, K. reyi, and K. parodii. I see no other viable alternatives and strongly recommend a Yes on this one. Further, Barker et al. (2015) estimated that this lineage diverged from their sister group ca. 7 MYA.

730.10. Resurrect Sphenopsis Sclater for Hemispingus melanotis and H. frontalis. See 730.07 for why Hemispingus cannot be maintained as is. These two species, traditionally recognized as sisters, constitute a group that is sister to Thlypopsis. The only alternative would be to merge these into Thlypopsis.

A YES vote on 730.10 would be to treat Hemispingus melanotis and H. frontalis as Sphenopsis melanotis and S. frontalis (as in Dickinson & Christidis 2014). A No vote would presumably be for expanding Thlypopsis to include them. I recommend a Yes on this because the two species of ex-Hemispingus differ substantially in my subjective view in morphology and plumage (but I am open to alternative views on this). Barker et al. (2015) estimated the divergence time from Thlypopsis at ca. 6 MYA, so the lineage is well within range of groups ranked as genera in tanagers.

730.11. Merge Pyrrhocoma ruficeps and Hemispingus superciliaris into Thlypopsis. These two species are embedded in a strongly supported group that includes all Thlypopsis. To retain all three genera would require naming new genera for various Thlypopsis. I don’t know Pyrrhocoma other than that it skulks in undergrowth, that its size and shape are consistent with Thlypopsis, and that the female really looks like a Thlypopsis. I used to be very familiar with H. superciliaris, and I am surprised that this canopy-dweller would be embedded in a group that includes Thlypopsis ruficeps, T. ornata, and T. pectoralis.

A YES vote on 730.11 would be to treat Pyrrhocoma ruficeps and Hemispingus superciliaris as Thlypopsis pyrrhocoma and T. superciliaris. Note that the epithet ruficeps is “preoccupied” in Thlypopsis, and thus Burns et al. introduced the new species name pyrrhocoma for this species to preserve the connection to Pyrrhocoma (a logical choice). I recommend a Yes on this one because the alternatives would involve naming two new genera, which would presumably be the alternative indicated by a No vote.

730.12. Resurrect Poospizopsis Berlepsch for Poospiza caesar and P. hypochondria. Poospiza is highly polyphyletic (see subproposals above and Fig. 3), and these two species form a strongly supported group with Donacospiza and Cypsnagra.

When I first wrote the proposal, I thought that any merger of these genera would be unacceptable by any subjective standard of morphological continuity. However, looking at the specimens, this foursome doesn’t look that much more heterogeneous than, say, reformed Poospiza.

A YES vote on 730.12 would be to treat Poospiza caesar and P. hypochondria as Poospizopsis caesar and Poospizopsis hypochondria (as in Dickinson & Christidis 2014). Although this is a relatively young lineage (est. ca. 3 MY in Barker et al. 2015), merging them and Donacospiza into Cypsnagra (which has priority) would seem unpalatable, but maybe there are themes that they share that I don’t see. A NO vote would presumably endorse broad Cypsnagra.

730.13. Recognize newly named, monotypic Castanozoster for Poospiza thoracica. As you can see from Fig. 3, P. thoracica is on a lonely branch within the group that includes the species in 730.12 and also a group that includes Nephelornis, Urothraupis, and a bunch of phylogenetic refugees from Poospiza (see next subproposal), and its relationships within this group are unresolved. Burns et al. were left with no choice but to name a new genus for the species.

Phenotypically, I don’t see anything special about this species; certainly, its traditional placement in broad Poospiza cannot be criticized from the standpoint of external plumage and morphology:

A YES vote on 730.13 would be to treat Poospiza thoracica as Castanozoster thoracicus (note change in variable endings because Castanozoster is masculine). I see no alternatives to this treatment given the phylogenetic data so far and thus strongly recommend a Yes on this one. Barker et al. (2015) estimated the divergence time at ca. 5 MYA.

730.14. Recognize Microspingus Taczanowski, 1874 (type species = trifasciatus), for Hemispingus trifasciatus and merge Poospiza cabanisi, P. lateralis, P. erythrophrys, P. alticola, P. torquata, P. cinerea, and P. melanoleuca into Microspingus. In Fig. 3, you can see that this group is strongly supported, as is its sister relationship to Nephelornis and Urothraupis. Hemispingus trifasciatus is deeply embedded in this group.

A YES vote on 730.14 would be to treat all these species in Microspingus, i.e. Microspingus trifasciatus, M. cabanisi, M. lateralis, M. erythrophrys, M. alticola, M. torquatus, M. cinereus, and M. melanoleucus, as in Dickinson & Christidis (2014). (Note changes in variable endings because Microspingus is masculine.) The only alternative to this treatment would be to also include Nephelornis and Urothraupis in Microspingus; thus, I strongly recommend a Yes on this one.

730.15. Merge Oreomanes into Conirostrum. As you can see in Fig. 3, Oreomanes fraseri is deeply embedded in Conirostrum, and to retain monotypic Oreomanes would require recognizing 4 additional genera for species currently placed in Conirostrum. Although at first the placement of the distinctive, bark-foraging specialist Oreomanes might seem a surprise, it shouldn’t be. Not only is its plumage similar to Conirostrum but it has also hybridized with the most similar species in plumage, C. ferrugineiventre (Schulenberg 1985).

Unfortunately, fraseri is “preoccupied” in Conirostrum by C. cinereum fraseri; the next oldest available name is Oreomanes binghami Chapman, and so the species becomes Conirostrum binghami.

A YES vote on 730.15 would be to treat Oreomanes fraseri as Conirostrum binghami. I see no alternatives to this treatment given the phylogenetic data so far and thus strongly recommend a Yes on this one.

730.16. Recognize newly named genus Ephippiospingus for Phrygilus dorsalis and P. erythronotus and recognize newly named genus Chionodacryon for Diuca speculifera. The genus Phrygilus is perhaps the most highly polyphyletic genus in the tanagers. Most of us suspected this already, but the degree to which Phrygilus species are scattered between and within subfamilies is unprecedented (see Figs. 1 and 4, and 730.01). Diuca is also not monophyletic (see Fig. 4; type species = D. diuca). As can be seen in Fig. 4, these three species plus Idiopsar brachyurus form a strongly supported clade. Burns et al. reasoned that naming two new genera was better than merging all four species into Idiopsar on the basis of the unique bill and foraging behavior of the latter.

A YES vote on 730.16 would be to treat Phrygilus dorsalis and P. erythronotus as Ephippiospingus dorsalis and E. erythronotus, and to treat Diuca speculifera as Chionodacryon speculiferum (note the required change in a variable ending). A NO vote would endorse the treatment rejected by Burns et al., i.e. all four species placed in Idiopsar, as in Burns et al.’s Table 1, Alternative 2. I strongly prefer the latter treatment and thus I recommend a NO on this subproposal. Not only would inclusion of all species in Idiopsar avoid recognizing two new genera, one of which is monotypic, but it would also group together four high-elevation species of the southern Andes that share overall gray plumage and similar size. As for the unique bill shape and foraging behavior of Idiopsar, the same could be said for Oreomanes (see 730.15), which Burns et al. merged in Conirostrum (which I support). Finally, Idiopsar brachyurus and Diuca speculifera are extremely close genetically (closer than almost any two species in the phylogeny), to the point that recognition of Chionodacryon as monotypic is not supported by genetic distance. Another potential solution would be to include speculifera in Idiopsar while recognizing Ephippiospingus. Barker et al. (2015) estimated the age of the node that unites all four species at ca. 4 MYA, and so they are all well within the range of groups included in a single genus.

730.17. Resurrect Geospizopsis Bonaparte 1856 for Phrygilus unicolor and P. plebejus. As can be seen in Fig. 4, these three species plus Haplospiza unicolor and extralimital Acanthidops bairdi form a strongly supported group. Of great surprise (to me anyway) is that the two traditional Haplospiza are not sister species, with rustica sister to Acanthidops. Although the support for this is strong, it is not 100%, and the position of Haplospiza unicolor within the group is uncertain.

A YES vote on 730.17 would be to treat Phrygilus unicolor and P. plebejus as Geospizopsis unicolor and G. plebejus. This would leave our current Haplospiza as likely paraphyletic, but that is preferable, temporarily, to retaining a polyphyletic Phrygilus until the problem with homonymy (see below) is solved. A NO vote would be for an alternative treatment, e.g. to treat all 5 species in a single genus (Haplospiza Cabanis 1851 has priority over Acanthidops Ridgway 1882), as in Burns et al.’s Table 1, Alternative 2. I actually prefer the latter treatment. I have a difficult time with the two species of current Haplospiza being in different genera, which would not be diagnosable on morphological grounds. These two plus Acanthidops are bamboo specialists, and there seems to be a remote chance that additional genetic data might actually group the three species. The two former Phrygilus also share with the three bamboo specialists an all-gray plumage, and the five species together differ primarily in bill shape. Barker et al. (2015) estimated the age of the node that unites H. rustica and Acanthidops at ca. 4 MYA, the age of the divergence between them and Geospizopsis at ca. 5.5 MYA, and the age of the node that unites all 5 taxa, including Haplospiza unicolor, at ca. 6.5 MYA. Thus, lineage age are does not provide any clear guideline here; certainly, an expanded Haplospiza would fit with the ages estimated for most genera in tanagers. However, the main problem for a single genus treatment is one of nomenclature: in an expanded Haplospiza, Phrygilus unicolor was described (as Emberiza unicolor Lafresnaye & d’Orbigny 1837) before Haplospiza unicolor (1851), and so that mess would have to be sorted out in a separate publication. A new name would be required for Haplospiza unicolor, a name that has been in use for nearly 170 years. Therefore, I recommend a YES vote (resurrect Geospizopsis, retain Haplospiza as currently constituted) just as a temporary solution while nomenclature is sorted out.

730.18. Recognize a monotypic Tephrophilus Moore 1934 for Buthraupis wetmorei; recognize monotypic Sporathraupis Ridgway 1898 for Thraupis cyanocephala; and continue to recognize Anisognathus as monophyletic despite lack of support. In Fig. 5, you can see there is a group of montane tanagers, including all species in our current Buthraupis and Anisognathus plus Chlorornis and “Thraupis” cyanocephala, that has mixed support (Bayesian PP 1.0 but Maximum Likelihood bootstrap only 69); within this group the topology is largely unresolved. This was the subject of now-tabled SACC proposal 569 (see for valuable discussion). There is a 6-way polytomy: (1) Buthraupis/Tephrophilus wetmorei; (2) Buthraupis montana (the type species of Buthraupis); (3) Thraupis/Sporathraupis cyanocephala; (4) Chlorornis + Cnemathraupis eximia and C. aureodorsalis (note that we already approved recognition of Cnemathraupis); (5) Anisognathus somptuosus + A. notabilis; and (6) Anisognathus melanogenys, A. lacrymosus, and A. igniventris. Burns et al. recommend the solution as stated in 730.18; concerning the problem with lack of support for monophyletic Anisognathus, they stated: “In addition, support for a few genera (Anisognathus, Certhidea, and Sicalis) was equivocal. Because there was neither support for or against monophyly in these cases, we retained the current genus assignment, pending further data.” Hellmayr (1936) recognized Poecilothraupis Cabanis 1851 for these three species and Compsocoma Cabanis 1851 for somptuosus and notabilis.

A YES vote on 730.18 would support treatment of Buthraupis wetmorei as Tephrophilus wetmorei, Thraupis cyanocephala as Sporathraupis cyanocephala, and Buthraupis eximia and B. aureodorsalis as Cnemathraupis eximia and C. aureodorsalis, and maintain Anisognathus as currently defined (all as in in Dickinson & Christidis 2014). In Table 1, Burns et al. provided an alternative treatment, namely combining all of these genera with Dubusia, Pseudosaltator, and Pipraeidea (with Pipraeidea having priority). The unstated rationale for this was to expand the boundary of the genus to encompass the first strongly supported node (BPP = 1.0, MLB = 96). This alternative would produce an exceptionally heterogeneous genus and would likely be unpalatable. However, I like an intermediate solution, namely combining the six branches that form a polytomy into a single genus: Chlorornis Reichenbach 1850 has priority. Given the uncertain relationships within this group, combining them all into a single genus, at least until relationships within the group are better resolved, has some appeal. Although heterogeneous in plumage, they are roughly similar in size and bill shape, and they share an Andean cloud-forest distribution. Also, to recognize an expanded Chlorornis would (1) avoid maintaining Anisognathus, for which there is no support for monophyly (and was not considered monophyletic by Hellmayr), and (2) avoid resurrecting 3 genera (Cnemathraupis, Sporathraupis, Tephrophilus), two of which would be monophyletic. Barker et al. (2015) estimated that the age of the node that unites this group at about 8 mya, so this fits nicely the age of taxa ranked as genera in tanagers; however, most of the divergence among lineages that Burns et al. proposed as genera are also fairly old, ca. 6 MYA (in other words, relatively little recent divergence), so either recommendation fits my scheme of using lineage age as a guide in delimiting genera. A disadvantage of a single genus is that it would create nine novel combinations with Chlorornis, which is also the only almost-all-green species in the group. I lean towards a NO vote on this subproposal but will decide based on comments.

730.19. Resurrect Ixothraupis Bonaparte 1851 (type species = Tangara punctata), for Tangara punctata, T. varia, T. rufigula, T. xanthogastra, and T. guttata. In Fig. 5, one can see that traditional Tangara (type species = chilensis) is not monophyletic because Thraupis is embedded within it. As Burns et al. (2016) noted, their previous suggestion to merge Thraupis and Tangara was not adopted by anyone, and so Burns et al.’s (2016) solution was to maintain Thraupis by dismembering Tangara. Although many of us will mourn the loss of broadly defined Tangara, that classification is really a Meyer de Schauensee (1966) construct, and previous classifications (e.g., Hellmayr 1936) divided the genus into Tanagra and Calospiza. The 5 species in Ixothraupis form a strongly supported group, as would be predicted from shared plumage patterns.

A YES vote on 730.19 would support treatment of Tangara punctata, T. varia, T. rufigula, T. xanthogastra, and T. guttata as Ixothraupis punctata, I. varia, I. rufigula, I. xanthogastra, and I. guttata. This is consistent with the data and required to keep a pruned Tangara and Thraupis monophyletic; I recommend a Yes on this one. Barker et al. (2015) estimated the divergence time between this lineage and the rest at ca. 6 MYA, so this fits within the range of the age of genera in tanagers.

730.20. Recognize newly named Poecilostreptus for Tangara palmeri (and extralimital T. cabanisi); resurrect Chalcothraupis Bonaparte 1851 for Tangara ruficervix; and recognize newly named Stilpnia for Tangara cyanoptera, T. larvata, T. nigrocincta, T. cyanicollis, T. preciosa, T. peruviana, T, meyerdeschauenseei, T. vitriolina, T. cucullata, T. cayana, T. viridicollis, T, phillipsi, T. argyrofenges, and T. heinei. See 730.19 for the need to break up Tangara to maintain Thraupis. As one can see in Fig. 5, these changes are required within the section of the tree that includes Thraupis to maintain that genus. Stilpnia is a heterogeneous group for which the node has substandard MLB support (72) but strong BPP support (.97); Burns et al. outlined some plumage features that distinguish members of this group from true Tangara.

A YES vote on 730.20 would support recognition of Poecilostreptus, Chalcothraupis, and Stilpnia as outlined above as well as traditional Thraupis (minus cyanocephala; see 730.18). I recommend a Yes on this one. Barker et al. (2015) estimated the divergence time among these lineages at ca. 5 MYA, so this fits within the range of the age of genera in tanagers.

Literature Cited

BARKER, F. K., K. J. BURNS, J. KLICKA, S. M. LANYON, AND I. J. LOVETTE. 2015. New insights into New World biogeography: An integrated view from the phylogeny of blackbirds, cardinals, sparrows, tanagers, warblers, and allies. Auk 132: 333–348.

BURNS, K. J., A. J. SCHULTZ, P. O. TITLE, N. A. MASON, F. K. 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 Phylogenetics and Evolution 75: 41–77.

BURNS, K. J., P. UNITT, AND N. A. MASON. 2016. A genus-level classification of the family Thraupidae (Class Aves: Order Passeriformes). Zootaxa 4088: 329–354.

SCHULENBERG, T. S. 1985. An intergeneric hybrid conebill (Conirostrum X Oreomanes) from Peru. Pp. 390-395 in "Neotropical Ornithology" (P. A. Buckley et al., eds.). Ornithological Monographs No. 36.

Van Remsen, October 2016

P.S. Once the outcomes of these proposals are determined, I will work on a proposal on revising the linear sequence of genera in Thraupidae. Fortunately, Burns et al. (2016) already addressed this and provided a phylogenetic sequence.

___________________________________________________________________________________

Comments from Areta: “This is the longest proposal ever seen in SACC and one that shall not be easily solved, given the many nuances of each subproposal. After spending some hours reading through the papers and the proposal, I realized that many of the most conflicting situations for me did arise from newly described or old monotypic genera that are part of polytomies. Also, I feel that trying to accommodate a classification recurring almost exclusively to sequence data and without including natural history (and other) data is very difficult. I've tried to contribute with other lines of evidence when I felt that the sequence (and sometimes plumage) data were insufficient to gain a good understanding of the birds. I am pleased to see the authors of phylogenetic papers taking the time to propose detailed taxonomic treatments derived from their trees.

“730.01. NO. I prefer an expanded Rhopospina. Fruticeti, caerulescens, alaudinus and carbonarius form a coherent group in terms of plumage (especially when in new plumage, which changes greatly once worn), shape and bill coloration. Alaudinus is sometimes strikingly bluish-gray, approaching the unique color of caerulescens, and bibs are more or less apparent in different individuals or at different times in fruticeti and alaudinus. In terms of habitat and behavior, they also share important features. All species sing from exposed dry perches, and at least fruticeti and carbonarius also make parachuting flights while singing. Structurally, the songs differ either within Corydospiza, in an expanded Porphyrospiza or in my preferred choice of Rhopospina. I find it very easy and intuitive to accommodate variation in these four taxa as occurring within a single genus. This also facilitates phylogenetic allocation of all four species.

“One final note: Porphyrospiza Sclater & Salvin 1873 is younger than Rhopospina Cabanis 1851, so I think that the Alternative 2 in Burns et al. 2016 (Table 1, page 345) including all four species in Porphyrospiza is an error.”

“730.02. NO. For reasons stated in 730.01.

“730.03. NO to their merger in Saltator. However, I do not like both under Saltatricula either. Multicolor and atricollis are very different in vocalizations, plumage, behavior and sociality. It looks like a case in which much differentiation has occurred in comparison to other similarly distant species in the tanager clade.”

“730.04. YES and perhaps NO. Clearly, all three species need to be removed from Tiaris. YES to the placement of fuliginosa and obscura in Asemospiza. However, I have trouble in supporting the relationship between richardsoni and bicolor to the exclusion of other possibilities. To me, bicolor sounds and looks like an Asemospiza. Given the polytomy in which these taxa are placed, may future genetic studies show different phylogenetic relationships? I am open to suggestions/interpretations of this by other SACC members.”

“730.05. YES. I like the idea of keeping Eucometis and Trichothraupis separate while recognizing Islerothraupis for this set of ex-Tachyphonus. I've always had trouble in accommodating all the variation found in Tachyphonus. Having more experience with the southern birds, my encounters with other species always left me scratching my head. It is good to see order brought to this.”

“730.06. NO. I am not in favor of recognizing monotypic genera unless there is a good set of reasons to do so. Given the uncertainties regarding their placement, I imagine they could be accommodated in Rhodospingus or perhaps even in Lanio. If these were long-described genera, I might be slightly more convinced of recognizing them. But in this era, I would like to see better arguments to erect monotypic genera.”

“730.07. YES. Phylogenetic relationships, biogeographic patterns and phenotypic data support the treatment of xanthophthalmus and verticalis in Pseudospingus.”

“730.08. YES? Although an alternative treatment that will also please me is to see ornata, boliviana and nigrorufa/whitii in Poospiza and hispaniolensis, rubecula, baeri, garleppi, superciliaris and goeringi in Compsospiza. If this later alternative is preferred by others, I would vote for it. If not, I can go with this "larger-but-smaller" new Poospiza conception.”

“730.09. YES.”

“730.10. A hesitant YES. See comments on 730.11 below.”

“730.11. YES. Pyrrhocoma females are very similar to other Thlypopsis, and vocally, the males also fit nicely in this genus. Thlypopsis pyrrhocoma is an elegant solution to the taxonomic homonymy between both ruficeps.

“The inclusion of superciliaris in Thlypopsis adds diversity in shape and vocalizations to an otherwise reasonably tight group and also makes me wonder on the need to separate frontalis and melanotis in Sphenopsis. If we accept superciliaris in Thlypopsis, why do not accept those two too? I look forward to receiving input from those who know these birds in life.”

“730.12. YES. Two neat Andean birds that share the same pectoral pattern, inverting the colors of the flanks and breast-band.”

“730.13. A mild YES. Again, I would appreciate input on the life of thoracica to inform a better decision.”

“730.14. YES. Here is the other lot of Poospiza-like birds. It makes sense vocally and morphologically to keep them all under Microspingus.”

“730.15. YES, but YUCK! I do not see any way out of this. It is sad to see Oreomanes fraseri becoming Conirostrum binghami.”

“730.16. NO on several grounds. First, all species in this clade are exclusively high-Andean taxa occurring in a relatively narrow zone, frequently replacing each other geographically or in terms of habitat. Second, as argued by Van repeatedly, bill features are extremely plastic and in this case, brachyurus can be understood as just a version of the other birds with a larger (particularly longer) bill, whereas speculifera is a more boldly patterned bird than the other species. Third, facial patterns (white crescent below the eye, faintly mottled or striped ocular region, red eyes of variable intensity), white throats and gray bellies are very similar among all four species. Fourth, except for the reddish back of dorsalis (and juvenile erythronotus) these are mostly gray birds. Fifth, all species seem to sing during a short period of time and remain vocally little known and all have various similar high-pitched and squeaky calls. Sixth, the glacier-nesting speculifera has taken an extreme evolutionary path in this, as much as brachyurus took the road to a humongous bill and dorsalis is restricted to above 4000 m asl (at least in Argentina, but possibly also elsewhere?). All in all, despite their extremes, I find many similarities among these four birds that make it unjustified to have three genera for them. Lastly (and more subjectively), Ephippiospingus and Chionodacryon are unnecessarily convoluted names that may cause more trouble than clarity. In sum, I do not endorse unnecessary over-splitting of a genus that seems to me coherent: I support placing all the species in Idiopsar.”

“730.17. YES. I would like to see more solid data showing that the two Haplospiza are not sister, since like Van, I believe that they are. I may be wrong, but I want to a solid rejection of my favorite hypothesis. So, retaining a possibly paraphyletic Haplospiza does not bother me.”

“730.18. Difficult proposal in the face of an enduring polytomy. However, I will follow the treatment that I endorsed in Proposal 569, which implies resurrecting several old genera.

--Sporathraupis: given the long branch and distinctive morphology, Sporathraupis cyanocephala seems the way to go.

--Tephrophilus: in the recent paper by Barker et al. (2015) Tephrophilus wetmorei was found to be sister to Buthraupis montana, making it possible for it to be retained in Buthraupis. Nevertheless, even if part of a polytomy, long branches and very different external appearance support the inclusion of wetmorei in a different genus.

--Compsocoma: Compsocoma (with somptuosus and notabilis) is separated from the remainder of Anisognathus by deep branching and is morphologically coherent (yellow nape and plain-black face without facial markings), thus I favor its recognition, which is also consistent with treatment of other genera in the group.

--Cnemathraupis: recognize this genus for aureodorsalis and melanogenys.”

“730.19. YES. Plumage and natural history data support their collective placement in Ixothraupis.”

“730.20. YES. Seems a reasonable way to sort the distinctive "black collared" Poecilostreptus palmeri and cabanisi, accommodate the oddly patterned Chalcothraupis ruficervix and place the remaining ex-Tangara.”

Comments from Robbins:

“730.01. NO. I agree with Nacho that an expanded Rhopospina, with the caveat that perhaps Rhopospina may have priority over Porphyrospiza, is the better choice with dealing with this group of birds.

“730.02. NO, for the reason pointed out by Nacho.

“730.03. NO, to the merger of multicolor and atriceps into Saltator for reasons mentioned in earlier proposals coupled with the new genetic data and timing of the split from the rest of Saltator.

“730.06. NO. I totally agree with Nacho’s philosophy about recognizing monotypic genera. Moreover, the relationships are clearly unresolved, so I prefer we wait until more data arise.

“730.07. YES. This makes total sense in placing xanthophthalmus and verticalis in Pseudospingus, as they clearly are quite distinct in all facets from other “Hemispingus”.

“730.10. A tentative YES, as I can easily be convinced with including these two species in an expanded Thlypopsis.

“730.11. YES, for merging Pyrrhocoma and superciliaris into Thlypopsis. Like Van, I’m surprised that superciliaris is closely related to ruficeps, ornate, and pectoralis.

“730.13. YES. There does not seem to be an alternative, so concur with the new genus Castanozoster.

“730.15. YES. Unfortunately by having to also change the specific name makes the English name the only stability of this wonderful taxon! Jeez……

“730.16. NO. I agree with the comments of Van and Nacho concerning why these taxa should all be included in Idiopsar.

“730.17. YES. Indeed, this is a messy proposal with no easy solution. For now, I support Van's temporary suggestion until some of the details can be resolved that might clarify what is the best course.”

“730.18. Clearly whatever we do here is tentative. I support placing cyanocephala in Sporathraupis. Given that Barker et al. (2015) subsequently found that wetmorei is sister to montana, I do not support resurrecting a new genus for it, i.e., for now, maintain wetmorei in Buthraupis. We continue to place aureodorsalis and eximia in Cnemathraupis and recognize Anisognathus. This provides the most stability until we have the final word on relationships within this clade.

Comments from Stiles:

“730.01: resurrect Rhopospina for ”Phrygilus” fruticeti: YES. It is clearly distinct genetically and phenotypically from the other species of this clade.

“730.02: Resurrect Corydospiza for “P.” alaudinus and carbonarius: YES. The proposal is not clear on whether caerulescens should be included: however, similar color differences also occur in Diglossa and in this case, the genus would have to be Porphyrospiza by priority.

“730.03: NO. Include both species in Saltatricula, which clearly merits generic status.

“730.04: YES: Merge the two ”Tiaris” into Asemospiza and merge bicolor into Melanospiza. This is definitely preferable to sweeping under the rug the great genetic and phenotypic variation in the rest of this clade in a virtually undiagnosable Tiaris. Here, I prefer being flexible with regard to branch lengths and estimated ages, and the status quo is clearly not acceptable either.

“730.05: YES. Clearly these cannot be maintained in Tachyphonus s.s., and the “typical” black-below Tachyphonus plumage is subject to much homoplasy (such homoplasy is also evident among the genera of hummingbirds).

“730.06: YES. Genetic data favor monotypic genera for Maschalethraupis surinamus and Chrysocorypha delattrei, and no alternative treatment is reasonable with the data available.

“730.07: YES. The close affinity of these two canopy species has long been suspected and they cannot be included in Hemispingus, so resurrecting Pseudospingus is the best choice.

“730.08: YES. The difference in bill sizes does not disturb me as it is clearly a flexible character related to diet shifts, which have occurred in several other tanager genera; Poospiza as reconstituted thus seems preferable to more generic splitting in this clade.

“730.09: YES. These four species form a well-supported genus genetically.

“730.10: YES to resurrecting Sphenopsis for these two species; they represent an appropriately old lineage distinct from Thlypopsis and are ecologically similar, although in my experience, melanotis seems to like areas with more bamboo while frontalis seems more indifferent. I disagree with Mark on including these two larger and heavier-bodied species in Thlypopsis, which increasingly would verge on undiagnosability.

“730.11: YES, but with one reservation - I hope that the phenotypically discordant (in Thlypopsis) genetic sample of “Hemispingus” superciliaris is supported by a suitably vouchered and identified specimen – this should be checked and if necessary, repeat this analysis with a duly vouchered specimen. (Such a case with a mislabeled genetic sample did occur with a hummingbird, so best to be sure).

“730.12: YES, given the strong support for placing both in Poospizopsis and the impossibility of including these two species in Poospiza, as redefined above.

“730.13: YES to a monotypic Castanozoster. Cypsnagra and Donacospiza are sufficiently distinct genetically, ecologically and phenotypically that a merger would produce a totally undiagnosable genus.

“730.14: YES to including these seven species in a resurrected Pseudospingus, which is well supported and both its nearest relations, Nephelornis and Urothraupis, are sufficiently distinct to continue as monotypic genera. The “cascade” of branch lengths in this group would make any further splitting within it arbitrary.

“730.15: YES; O. fraseri is best considered an ecologically and dietarily specialized Conirostrum, and the change in the species name is thus mandatory.

“730.16: NO. Including all four species in Idiopsar seems the best choice, as all four form a fairly coherent group, phenotypically and biogeographically; as in the previous case, the unusual bill of I. brachyurus presumably represents a relatively recent dietary shift (cf. Conirostrum). The most feasible alternative to me would be a two-genus split: Idiopsar for brachyurus and speculifera, Ephippiospingus for dorsalis and erythronotus.

“730.18: Yes, at least in part. Another study places wetmorei and montana as sisters, but the relationship is not close; given the very different plumages, I favor recognizing Tephrophilus as distinct from Buthraupis. Sporathraupis should also be recognized as a separate genus: it certainly is not a Thraupis, and its plumage and bill also differ from the remainder of this polytomy. However, I also strongly favor recognizing Compsocoma as distinct from Anisognathus; it is well characterized by plumage, ecology and morphology and genetically, it seems quite distinct from the “true” Anisognathus. The split of Chlorornis from the two Cnemathraupis appears more recent and is therefore somewhat more debatable. Given its completely different coloration including the bill, I lean towards maintaining the two genera to facilitate diagnoses; this also is better for maintaining stability. However, all three are seemingly similar in morphologically, so I could also live with an enlarged Chlorornis, despite the color clash.

“730.19: YES to recognizing Ixothraupis; its species share a distinctive plumage pattern; this split also makes the partial dismemberment of Tangara and maintenance of the vocally and morphologically distinctive Thraupis consistent with the genetic data.

“730.20: YES to recognizing Poecilostreptus, Chalcothraupis and Stilpnia to complete the generic rearrangement of the species of Tangara. Both Poecilostreptus and Stilpnia are well characterized genetically and by plumage; a monotypic Chalcothraupis for ruficervix is justified by its lack of close relatives as well as its distinctive plumage and bill shape. For those that bemoan the “loss” of Tangara, consolation remains: it is still the second-largest largest genus in the Thraupidae, only exceeded by Sporophila!

Comments from Pacheco:

“730.01. YES. Based on the phenotypic and genetic divergences, and especially the age of separation in the clade.

730.02. YES. However, it is necessary - as alerted by Gary - to decide how to subordinate the Blue Finch. Corydospiza Sundevall, 1872, takes precedence over Porphyrospiza Sclater & Salvin, 1873.

730.03. NO. Preferring once again to keep both species apart, in Saltatricula

730.04. YES. It seems to me <<merge the two "Tiaris" into Asemospiza and merge into bicolor Melanospiza >> is preferable. Both genera with only two species.

730.05. YES. From direct field experience, it seems to me quite natural to gather these 3 taxa in a separate genus Islerothraupis.

730.06. YES. Unless a broad genus is maintained, I consider the evidence satisfactory at present for the treatment of these taxa in monospecific genera.

730.07. YES. Recognizing Pseudospingus seems well supported decision.

730.08. YES. In this case, I vote by “simpler approach that does not require two additional new generic names”.

730.09. YES. The treatment of these four taxa under this new genus is well supported.

730.10. YES. Based on Gary's comment.

730.11. YES. In my experience with "Pyrrhocoma" ruficeps, I find quite plausible to combine it with Thlypopsis sordida, the only member of the genus that I know in field but type of this genus

730.12. YES. Well-founded support for placing both taxa in Poospizopsis.

730.13. YES for a monotypic Castanozoster. Put in perspective, thoracica is even fairly distinct from the traditional "Poospiza", at least of the Brazilian species that I know.

730.14. YES. The set of taxa under Microspingus is basically a well-interrelated group.

730.15. YES. A double change (name of genus, name of species) but clearly necessary.

730.16. NO. For the reasons put forward by Van: Idiopsar brachyurus and Diuca speculifera being extremely close genetically, my vote is for the maintenance of all 4 taxa in Idiopsar – the oldest available name for this set.

730.17. YES. On the basis of Alvaro's comment [below], I am also in favor of keeping the two former Phrygilus in a separate genus

730.18. An almost YES. I give my vote for the recognition of Tephrophilus, Sporathraupis, Cnemathraupis, but also of Compsocoma.

730.19. YES. A set very well correlated from among "Tangara".

730.20. YES. The maintenance of the distinctive Thraupis outside of Tangara requires all these arrangements.”

Comments from Claramunt:

“730.01. NO. I agree with Nacho and Mark. Merging these four species into a single genus makes better sense and would result in a fairly coherent genus (habitat, bluish plumage, yellow bill), avoids the resurrection of a monotypic genus and eliminates an additional one. Not impressed by the estimated crown age of the clade; not much older than Incaspiza, for example. Rhopospina Cabanis 1851 has priority over Porphyrospiza Sclater & Salvin, 1873, for the expanded genus.

730.02. NO. See above.

730.03.  NO to merge Saltatricula into Saltator. Although Saltatricula is sister to S. atricollis, atricollis itself is an “aberrant” Saltator and given the diversity already contained in Saltator and the position and distance of the Saltatricula/atricollis clade, I think it is better to maintain Saltatricula, containing multicolor and atricollis, as a separate genus. Although different in size, S. atricollis and S. multicolor share some similarities, including bill shape and color and a mostly-black face that makes them a coherent group, if not homogeneous.

730.04. YES to both changes regarding the disintegration of former Tiaris. Genetic data seems robust, taxonomic changes unavoidable.

730.05. NO to recognize the new genus Islerothraupis. Instead, I see the opportunity to get rid of two questionable monotypic genera based mostly on plumage and taxonomic uncertainty (Trichothraupis and Eucometis) by merging them with the “Islerothraupis” clade to form an expanded Trichothraupis. The result would be a somewhat diverse but coherent genus in terms of behavior and morphology. I’m emphasizing the similarities across species (bill shape, elongated crown feathers, plumage patterns) over the differences. I think it would be a better use of the “genus” category in this highly diverse and complex tanager taxonomy.

730.06. YES to recognize new monotypic genera Maschalethraupis but NO to recognize Chrysocorypha. Relationships are still unresolved but I’m willing to accept the new genus for surinamus given the phylogenetic evidence so far. However, delatrii may be sister to Rhodospingus (Burns et al. 2014, albeit with no statistical support) and thus the most conservative solution is to include delatrii in Rhodospingus instead of erecting a new monotypic genus, at least until relationships are resolved with more confidence.

730.07. Reluctant YES to resurrect Pseudospingus for Hemispingus xanthophthalmus and H. verticalis. There is some signal of them being sister to Cnemoscopus (Burns et al. 2014) thus placing them in that genus may have been a more conservative solution. Phenotypic, behavioral and ecological similarities are evident. But given the uncertainty and that a name is already available, resurrecting Pseudospingus is a sensible solution right now.

730.08. YES to the redefinition of Poospiza. Clade support is not super high but the group makes sense phenotypically, although still heterogeneous in size and color. Splitting the group into two genera, as Nacho suggested, is another possibility. If P. hispaniolensis grouped with the nigrorufa clade, I will be inclined to favor the two-genera solution but given the current tree, I prefer the more widely defined Poospiza suggested by Burns et al.

730.09.  YES to recognize Kleinothraupis.

730.10. Reluctant YES to resurrecting Sphenopsis. The homogeneity of Thlypopsis is already disturbed by the inclusion of Pyrrhocoma and H. superciliaris. If we accept the new more-heterogeneous Thlypopsis, the inclusion of melanotis and frontalis in it would not add much heterogeneity to the group.

730.11. YES to merging Pyrrhocoma ruficeps and Hemispingus superciliaris into Thlypopsis.

730.12. YES to resurrecting Poospizopsis for Poospiza caesar and P. hypochondria. The alternative of merging all four taxa into a single genus would result in a genus that makes no sense phenotypically or ecologically.

730.13. YES. Let’s concede a new genus to this phylogenetic refugee.

730.14. YES to recognize Microspingus.

730.15. YES to merge Oreomanes into Conirostrum. Fascinating result.

730.16. NO. I fully agree with the arguments presented by Van and Nacho. I prefer to include these taxa in Idiopsar.

730.17. YES to resurrect Geospizopsis. I would also merge Acanthidops into Haplospiza, as they seem to form a clade together with H. unicolor and H. rustica (see fig. 5 in Burns et al. 2015 Molecular Phylogenetics and Evolution 75: 41–77). An expanded Haplospiza would not be desirable because of the homonymy involving the two “unicolor" and splitting the group into more than two genera seems unnecessary.

730.18. NO. This is a case where a “radical” solution such as treating all taxa under Chlorornis makes more sense than trying to save some traditional generic names at the cost of erecting some new ones leading to multiple monotypic genera and small genera with no consistent diagnosable traits. After seeing the phenotypic heterogeneity of other genera in the new classification, the heterogeneity within the new expanded Chlorornis is about normal.

730.19. YES to resurrect Ixothraupis. It helps solving the “Tangara/Thraupis” issue and the resultant genus is cohesive.

730.20. NO to the recognize Poecilostreptus and Chalcothraupis but yes to split the clade between Thraupis and Stilpnia (although I don’t like the name because it is difficult to pronounce). I don’t see ruficervix and palmeri distinct enough to deserve their own genera so the argument rests solely on phylogenetic uncertainty, which is not a sufficient reason by itself, in my opinion. A conservative solution would be to include ruficervix in Thraupis and palmeri in Stilpnia, as there is some phylogenetic signal for that (see fig. 6 in Burns et al. 2014).”

Comments from Jaramillo: “I should first say that this was a beautifully well done proposal!

730.01: “YES, fruticeti is quite an unusual bird, quite unlike alaudinus in the other group. Not only in the very large size and overall bulk, but also the voice is almost icterid like. It does stand alone, at least knowing the live bird.”

730.02: “NO, I prefer an expanded Porphyrospiza.”

730.06: “NO. Maybe I am misunderstanding, but doesn’t the data suggest that delatrii is sister to Rhodospingus? Why not merge it with Rhodospingus, and leave surinamus in the new genus?

730.10: “YES resurrect, Sphenopsis. Although I am not averse to expanding Thlypopsis.”

730.12: “NO. Donacospiza always seemed like a “Poospiza” to me, of course that means little now that we have sorted out Poospiza in a different format. However, the oddball in this group is Cypsnagra, yet it is clearly in the group. I don’t know what to do exactly, but it seems to me that uniting these four, although unpalatable, may be the best thing to do. By the way, that specimen of Donacospiza may be a juvenile, the adult would look a lot more like two of the four of this group with an obvious supercilium etc.

730.16: “YES/NO. YES, recognize Ephippiospingus for dorsalis and erythronotus. I am unclear if they are not actually the same species! Juvenile erythronotus have reddish backs, similar to dorsalis. But that will need more research. HOWEVER, on the second part of this I am very happy that the data bear out something that I have been seeing in the field for years, i.e. that speculifera is no Diuca! But rather than give it its own genus, it is clearly fine in Idiopsar. Now, Idiopsar has an odd bill, but this is an incredibly plastic trait. In the field, and even if you look at photos, the look of Idiopsar is similar to that of “Diuca” speculifera, particularly so in the face pattern with the pale area below the eye, and the overall gray color. Including speculifera in Idiopsar is actually preferable.

730.17: “YES. I am in favor of maintaining the two ex Phrygilus as separate from the rest in this group. While the males of unicolor might be outwardly similar, these two are found in open Puna or Paramo habitats, quite different from the rest of the birds in the group. As well, females of unicolor and plebejus are brown and streaked, again different from the rest. I am not comfortable in lumping all of these under one genus. By the way, unicolor is perhaps two different species, but that is another story. Interestingly in these possible two species, the male plumage is conserved, but females are different, and voice is different.

730.18: “YES. Not a strong vote on my part, the additional genera are more palatable to me that lumping all in Dubusia for example.”

730.19: “YES, resurrect Ixothraupis. Morphologically not that far off, and fits the molecular data.”

Comments from Zimmer:

"730.01: YES, because of the predicted age of divergence of P. fruticeti, and, because it seems like a fairly different beast from the others.

“730.02: YES. I’m really on the fence with this one. Porphyrospiza is such a distinctive little bird of the campos rupestres habitats, and also seemingly somewhat nomadic in seeking out areas of this habitat that have recently burned. In addition, it would be the clear plumage outlier within the proposed expanded genus. On the other hand, female plumages of the various species are much more similar to one another, adult males are yellow-billed, and habitat, at least between P. alaudinus and Porphyrospiza is at least structurally similar. This one is a coin toss for me, but I think I’ll cast my lot with a YES vote and keeping caerulescens on its own.

“730.03: “NO, for reasons articulated in previous Proposals dealing with this specific issue.

“730.04: YES. As Van states, this looks like the only palatable solution, given the polyphyly of Tiaris.

“730.05: YES. On ecological, morphological and vocal grounds, the divisions between “true Tachyphonus” (rufus, phoenicius, coronatus) and proposed Islerothraupis (cristatus, luctuosus and rufiventer) makes perfect sense. I would be very opposed to rolling Eucometis and Islerothraupis species into an expanded Trichothraupis. Although Trichothraupis has some plumage similarities to Islerothraupis (black-and-tawny plumage; semi-concealed contrasting coronal patch), it is, to my thinking, a very different beast. For starters, Trichothraupis is not nearly as sexually dimorphic in plumage as is any of the species proposed for Islerothraupis. All of the Islerothraupis tend to forage high in the vegetation (to the canopy), coming lower at forest edges, but typically feeding amongst terminal, leafy branches, usually by direct gleaning insects or small fruits. Trichothraupis (at least in Brazil) forages more in the upper portion of the understory and much more often from open branches and hanging vines, from which it sallies to air or foliage to take insects (behaving almost like a Thamnomanes antshrike). Its song is also much more robust than the very high frequency, thin (almost hummingbird-like) songs of the Islerothraupis species. Aside from the obvious plumage differences, I can see more parallels between Eucometis and Trichothraupis, but I don’t either of those fitting into the same genus as the trio of species proposed for Islerothraupis.

“730.06: There are no truly palatable choices here. YES, for lack of any better idea. The two species seem pretty different from one another, and yet, do not match up well with anything else.

“730.07: YES. The data pretty strongly indicate they are sisters, so I see no other viable alternative.

“730.08: YES. On plumage considerations, this grouping makes sense (with all species having an obvious supercilium above a darker mask, and most having a significant amount of gray-and-rufous in the overall plumage).

“730.09: YES for reasons stated by Van in the proposal.

“730.10: YES. Folding H. melanotis (in particular) into Thlypopsis would introduce some major morphological and plumage variation into what is otherwise a pretty cohesive genus.

“730.11: YES, although the position of H. superciliaris within the Thlypopsis-group seems like a real stretch to me. On the other hand, I know Pyrrhocoma well, and it skulks in the understory, preferring shrubby forest borders, and has somewhat similar vocalizations to those of several of the Thlypopsis species. As Van points out, the females (and immature males) of Pyrrhocoma really resemble Thlypopsis, and even the mature males have a general pattern (rufous head, contrasting with the remainder of the body) that fits in with Thlypopsis.

“730.12: YES. I can’t see an expanded Cypsnagra that includes any of these other three species. It’s not about the plumage as much as it is about behavior (including the raucous duets of Cypsnagra), ecology, and voice.

“730.13: YES. I’m not a huge fan of monotypic genera either, but in this case, I think it clearly beats the other options.

“730.14: YES. I don’t like the alternative much (including Nephelornis and Urothraupis).

“730.15: YES. Sad to see Oreomanes go, but the phylogenetic data are pretty clear, the plumage pattern is not a big departure, and, there is that hybridization issue.

“730.16: NO, for reasons articulated by Van in the proposal.

“730.17: Am I remembering seeing a lot of email discussion about how to handle this one? I notice that no one else seems to have voted on this. It was all happening while I was on tour, so I wasn’t able to pay much attention to what was resolved.

“730.18: May I elect to punt? YES, but not for any truly coherent reason, other than I can’t fathom throwing all of these things in Chlorornis.

“730.19: YES. – consistent with the data, and also forms a morphologically cohesive group.

“730.20: YES.”

Additional comments from Stiles: “To begin with, the objective of this classification is to produce a classification based on phylogeny rather than the traditional classification based largely on colors and patterns and to a lesser extent, on differences in bill shape. We have all been raised and weaned on this classification, but it conflicts with the phylogeny of Burns et al. (2016) at many points. As this is the most solid and complete estimate of tanager phylogeny available, this basis for classification is not without its birth pangs! However, it represents a huge advance over our currently recognized classification with respect to the above objective. I think Van has done a great job of presenting it, including alternatives at several points – so let’s get with it! Here I present a somewhat revised version of my previous votes, trying to include information present in several comments so far received. I am assuming that the subproposals not mentioned here have already been passed (or rejected).”

“730.01: Here, I think Nacho made a good point for keeping all four species in a single genus, for which Rhopospina Cabanis, 1851 would be correct. However, the age of the split between fruticeti and the rest is definitely consistent with generic status for this species. Hence, I think that the most inclusive course would be to separate fruticeti in Rhopospina and the other three in Porphyrospiza; I am not disturbed by the blue-vs-black/gray color clash, which also occurs in several other genera (e.g., Diglossa). My second choice would be to place all four in Rhopospina, as recommended by Nacho.

“730.02: NO; see above.

“730.04: “YES; this seems the best way to resolve the polyphyly of Tiaris while recognizing relationship of bicolor to Melanospiza. A wholesale merger including the Galápagos finches and other Antillean genera would seem to produce an undiagnosable soup, despite the young age of the group as a whole, and I see no justification for simply ignoring the phylogeny, especially given the manifest polyphyly of Tiaris.

“730.05: NO: of the alternatives given, I rather prefer an expanded Islerothraupis, including Trichothraupis and Eucometis. However, I note that a while back, based on extensive studies of ecology and behavior, Willis advocated merging Eucometis into Trichothraupis, so despite the color clash, my alternative choice here would be this two-genus arrangement; the age of the split between the two groups also conforms to our level for generic groupings. Clearly, the remaining three species cannot stay in Tachyphonus, hence Islerothraupis should stand for them.

“730.06: “YES. Even given the problem of the polytomy, the ages of these two are best expressed as two monotypic genera; if they do prove to be congeneric this can be easily arranged by lumping them (whoever might do this would qualify as first reviser and choose which name to use).

“730.12: “The question here is whether or not to include all of these species in Cypsnagra; if we recognize Poospizopsis, we should certainly also recognize a monotypic Donacospiza as well as a monotypic Cypsnagra, however the depth of the polytomy in this group is not especially deep, hence I would not be averse to including both the two “Poospizopsis” plus Donacospiza in Cypsnagra.

“730.16: NO; I agree with Van that the evidence for splitting this group (with the addition of two new monospecific genera) is less than convincing, given that the only justification for the splits is the rather unusual bill of Idiopsar, but the age of the splits is no greater than those unifying Poospiza or Thlypopsis (among others) and the enlarged Idiopsar seems to be reasonably diagnosable.

“730.17: YES to the “temporary solution” recommended by Van. There are several ways to resolve the “nomenclatural mess”: a) resolve the polytomy, which would certainly require additional work with more genes sequenced (this might or might not produce a monophyletic Haplospiza, thus resolving the synonymy of unicolor; b) dig into the literature to find a junior synonym of unicolor – if one exists, which would also resolve this problem; or c) respecting the great depth of the polytomy, name a new monospecific genus for H. unicolor.

“730.18: Clearly more work is needed to fully resolve this polytomy, which is quite deep, implying that however it comes out, multiple genera will surely be required. At this point, I prefer to recognize coherent, diagnosable generic groupings insofar as possible; if some are later to be lumped, so be it. I certainly do not like the idea of recognizing a broad (!) Anisognathus – i.e., sweeping the whole problem under the rug! So, my recommendations are:

1) recognize monotypic Tephrophilus for wetmorei;

2) recognize monotypic Buthraupis for montana;

3) recognize monotypic Sporathraupis for cyanocephala (we certainly cannot leave it in Thraupis, and its plumage, bill and ecology don’t fit Anisognathus either);

4) recognize Compsocoma for somptuosus and notabilis – again, plumage, ecology and vocalizations don’t fit with Anisognathus, and they are not sisters to Anisognathus unless we also sink into it Chlorornis and Cnemathraupis, which would produce a very heterogeneous bunch for diagnosis;

5) either continue to recognize Chlorornis and Cnemathraupis, as we have already done, or lump all three into Chlorornis – I could go either way on this (having become somewhat immured to color clashes);

6) Restrict Anisognathus to lachrymosus, melanogenys and igniventris, which form a clearly diagnosable group on morphology, plumage and ecology.

“730.19: YES to recognize Ixothraupis for five species of “spotty” Tangara; necessary to at least justify preservation of Thraupis, upon which we all agree; the sticky part comes with further division of Tangara.

“730.20: “Really three proposals in one: the first two involve two deep outliers in the first big Tangara clade, ruficervix and palmeri. Given their position and probable time of divergence, both should be recognized as separate genera, respectively, Chalcothraupis and Poecilostreptus. The plumages of both are distinctive, as are the large size of palmeri and the unusual stubby, broad bill of ruficervix. Whether cabanisi should go with palmeri is a reasonable suggestion in the absence of genetic data, but when such data become available, it could be reallocated without much fuss.

“The second is recognition of Stilpnia, which at first sight seems quite heterogeneous. However, there are some common threads in plumages. Three clades exist within Stilpnia: the first includes several species mostly marked with silvery plumage and black caps or hoods, with all showing at least some sexual dichromatism; the second includes three “bright-hooded” species with reduced sexual differences (but not invisible upon close examination, even in the field); the third, another group with varied but often rather more dull-colored plumage, again with more or less marked sexual dichromatism. The probable time of divergence of Stilpnia is similar to that of Thraupis, hence I see no alternative to recognizing it if we wish to conserve Thraupis as a genus (which I do).

“Note: the third clade includes ca. 27 species, which Burns et al. (generously) consider to be the genus Tangara, although the age of the splits therein could easily justify dividing it into three to five smaller genera. However, this one could be worried about in a later proposal for those who feel less stressed at “losing” Tangara as a big, pretty genus! In a nutshell, the phylogeny is the best we have, and we can’t have our cake and eat it too!”

Comments from Stotz:

“730.01: NO. I think we are best served by recognizing this clade of 4 species as a single genus rather than creating a monotypic genus for fruticeti. In the Andes, I tend to think of alaudinus and fruticeti as fairly similar to each other and similarly carbonarius and fruticeti in Patagonia, despite the greater bulk of fruticeti. I think that a monotypic genus for fruticeti does not serve us well, despite the relatively old split of fruticeti.

“730.02: NO. Follows from my vote on 730.01

“730.03: NO

“730.04: YES. Doesn’t seem to be another realistic option.

“730.05: YES. This seems pretty straightforward These 3 Amazonian species seem like a natural unit, and I personally find it difficult to be comfortable merging them in with the very different Eucometis and Trichothraupis. I would not oppose putting Eucometis and Trichothraupis together in Trichothraupis as Gary discusses, although I guess that change is not on the table.

“730.06: NO. I don’t have an answer here that really convinces me. I think the only options are Lanio for surinamus, delatrii and Rhodospingus, or those 3 taxa as monotypic genera. I recognize that as currently defined, Lanio is a pretty tight group ecologically and morphologically, and is distinct from the other 3 taxa. However, the fact that we have no resolution regarding surinamus, delatrii and Rhodospingus relationships, beyond being part of the clade with Lanio leads me to conclude that putting everything in Lanio is unfortunately the best answer we have. I think we provide more information by emphasizing this clade, rather than creating monotypic genera to emphasize that we don’t quite understand these relationships.

“730.07: YES. In the grand scheme of Proposal 730, this seems pretty easy to accept. It is no surprise that Hemispingus is wildly polyphyletic, nor that these two species are closely related to each other and not to anything else.

“730.08: YES. I have a hard time thinking of Hemispingus rufosuperciliaris as a Poospiza, but in part that is due to the fact that when I think of Poospiza, I am really thinking of a different subsection of the genus (now or soon to be in Microspingus and Castanozoster). However, the Burns et al 2016 result is clear.

“730.09: YES

“730.10: NO. Given that these two species have little in common morphologically, and clearly are part of a larger clade with Thlypopsis (which has already had 2 species not traditionally placed in Thlypopsis brought into it), I think we get more information by placing melanotis and frontalis in Thlypopsis.

“730.11: YES. Given that I voted to bring melanotis and frontalis into Thlypopsis, it is clear that I agree these two taxa should be treated as Thlypopsis.

“730.12: YES. Not terribly surprising that, with Poospiza being scattered, Poospizopsis would come back. Not exactly intuitive that hypochondrius would go into it, but doesn’t concern me. Although Donacospiza has always seemed “Poospiza” and I could imagine it being congeneric with caesar and hypochondria, Cypsnagra seems like too much of an outlier morphologically, ecologically and behaviorally to bring in with these other species I guess that currently means maintaining Cypsnagra and Donacospiza.

“730.13: YES. Seems weird that thoracica would stand out so much phylogenetically It and lateralis always felt to me like virtually the same bird, but lateralis (or at least cabanisi; doesn’t look like true lateralis was included in study?) is in heart of Microspingus.

“730.14: YES.

“730.15: YES. In some ways sorry to see Oreomanes go, but it has always seemed like a giant Conirostrum as its English name suggests.

“730.16: NO. I think placing all 4 taxa in Idiopsar is clearly the way to go.

“730.17: YES. I should just say that I am not completely horrified by placing unicolor, rustica and bairdii in separate genera. I prefer not to have monotypic genera, but a lot of older literature used Spodiornis for H. rustica and of course Acanthidops is what we currently call bairdii. Ecologically the 2 ex-Phrygilus are distinctive compared to other three, so I would favor, if it comes to it, keeping plebejus and ex-Phrygilus unicolor separate, no matter what the ultimate topology looks like.

“730.18: YES, sort of. I have gone back and forth on this, but can’t bring myself to vote in favor of everybody in Chlorornis. So I favor recognizing Sporathraupis for cyanocephalus, recognizing Compsocoma for somptuosus and notabilis, and Cnemathraupis for eximia and aureodorsalis. I am inclined to put leave wetmorei in Buthraupis with montana.

“730.19: YES. This seems like a well-defined phenotypically clade.

“730.20: YES. I am fine with Stilpnia, less fine with Chalcothraupis and Poecilostreptus. However, I can’t see a way around recognizing them based on what we currently know.”