AOU Committee on Classification and Nomenclature (North & Middle America)

Comments on Proposals 2017-B

2017-B-1: Recognize additional species in the Aulacorhynchus "prasinus" toucanet complex
2017-B-2: Treat the subspecies (a) spectabilis and (b) viridiceps as separate species from Eugenes fulgens
2017-B-3: Elevate Turdus rufopalliatus graysoni to species rank
2017-B-4: Recognize newly described species Arremon kuehnerii
2017-B-5: Revise the classification of the Icteridae: (a) Add seven subfamilies; (b) Split Leistes from Sturnella; (c) Resurrect Ptiloxena for Dives atroviolaceus; (d) Modify the linear sequence of genera
2017-B-6: Revise familial limits and the linear sequence of families within the nine-primaried oscines
2017-B-7: Lump Acanthis flammea and Acanthis hornemanni into a single species
2017-B-8: Split Lanius excubitor into two or more species
2017-B-9: Add Mangrove Rail Rallus longirostris to the main list
2017-B-10: Revise the generic classification and linear sequence of Anas

Recognize additional species in the Aulacorhynchus "prasinus" toucanet complex

YES on A, B, E, F, G.

YES on A, B, E, F, G for reasons outlined in the proposal. In particular, I found the correlation between morphometric differentiation and genetic distance a helpful analysis when thinking about this case.

YES. I have gone back and forth on this. I am sympathetic to the argument that the lack of clear vocal differences is an argument against species status. But there is (by toucan allotaxa standards) pretty substantial morphological variation in this complex. In terms of voice, my read is that we are not consistent within toucans on how we treat vocally identical (or at least very similar) allotaxa. We lump the big Amazonian toucans, but we split the various Amazonian Selenidera, which to my ear sound pretty identical (at least across southern Amazonia), and we split Andigena hyoglauca and cucullata, as well as Pteroglossus viridis and inscriptus.  Even within Aulacorhynchus whitelyi and derbianus  are described as having essentially identical voices. Given that, plus the complex geographic variation in plumage and bill morphology, as well as consistent genetic results with that variation, I vote to split. On English names I am okay with the suggested names except I think we need something better than Northern and Southern Emerald Toucanet. Without the hyphen, we are not implying sister-status. I don’t have a good option for prasinus at present.

YES. As noted in the proposal, "the treatment relied more on the likelihood of postzygotic RIMs (increasing evidence of morphological divergence making successful hybrids and reticulation less likely) than on prezygotic ones." The fact that hybridization is only occurring regularly across one of the closest-approach zones (and that between the two most similar morphologically and genetically), coupled with its absence or low frequency at the other closest-approach zones (and the remarkably tight correlation between morphological and genetic differences), this looks like a reasonable solution for a good species complex. I think the exclusive focus on prezygotic isolating mechanisms is a mistake, particularly given the increasing number of morphometric differences with genetic distance in the group and (in my view) the telling absence of hybrids at other close-approach zones. In other words, I don't think it is an accident that the most frequent hybridization in this group occurs between the two that are most closely related. Flocking species especially are prone to having one-offs among them. The argument is that these lineages have undergone adaptation morphometrically (highly correlated with genetic distance; see Fig. 5), and that as such adaptations/difference accrue it gets increasingly unlikely that hybrids will have equal or greater fitness. Individuals do move around quite a bit during the nonbreeding season, and I don't think it's an accident that hybrids are not evident at most zones of closest approach (only where the two forms most similar genetically and morphometrically come together). While I don't generally like genetic distance, some of these are old lineages within the genus. From the proposal (with a tree there): "The South American radiation of the species haematopygus, whitelianus, derbianus, and sulcatus likely began after that of the prasinus clade (~4.5 Mya vs. ~5.2 Mya; Bonaccorso et al. 2013, figure inserted here)." In sum, I think these things have effectively made it through the speciation process without prezygotic RIMs (like voice) playing a strong role. As we know, there are different routes to speciation.

NO on A, B, F, and G. YES on E. A separate proposal on English names would be desirable, should any of these pass. As for G, we should ensure that the sequence follows Winker (2016) as closely as possible after votes are in. It does seem that the dispersive abilities of some of these taxa should lead to hybridization if they recognize each other as conspecifics in cases for which there is no such evidence. I don't think we can necessarily assume that because hybridization occurs in one or more groups of Ramphastidae that all the others are blind to phenotypic differences when it comes to mate choice. And I'm struck by the congruence between genetic and morphological divergence (except in the striking case of cyanolaemus-atrogularis). That said, a multi-way split is borderline and will be contentious, but the relatively deep divergence between the two major clades, prasinus and albivitta, indicates that at least these should be considered separate species. I would also agree with Northern and Southern Emerald-Toucanets for a two-way split. Not ideal, but probably better than anything else.

NO on A and B, YES on E, F, G. This is a complex situation and I appreciate all of the comments, especially since I'm not familiar with this group. I think that there remain too many questions to make a definitive decision to split the Middle American forms into multiple species. However, for reasons given in the proposal, it does make sense to me to split northern and southern populations into two species. However, I am also uneasy with the NACC and SACC coming up with a different decision based on the same proposal, and it appears that the SACC is waiting for the NACC to vote. I am fine with Northern and Southern Emerald-Toucanet.

NO (tentative). I almost abstained on this one as it is a group that I’m not overly familiar with. I've seen several of the taxa involved, but don't feel in the least that I really know them. I'm also a little uneasy with different Committees coming up with different votes and possibly a different decision based on basically the same set of issues. The issue of voice being a key RIM in this group (well-discussed in the motion) is not easily explained away, and if plumage (e.g. throat color) isn't an RIM, than what is left? The one ssp. that is phenotypically distinctive, isolated, and may have different vocalizations (as noted in the proposal) is wagleri from southwest Mexico. They certainly looked different in the field (from northern Oaxaca birds) when I saw them in Guerrero. Southwest (or western) Mexico is a known region of endemism so A. p. wagleri may be another species warranting separation. But splitting only that taxon is not currently a choice from the options. I learned a great deal about this group from the well-outlined motion.

NO on A and E for now. YES on B if 2017-B-1A passes. YES on F for those English names if 2017-B-1E passes. YES on G if  2017-B-1A or -1E passes. I read through the proposal and kept looking back at the plate, and was struck on how the two of the most dissimilar taxa (cyanolaemus and atrogularis) were the two that actually hybridized. These taxa differed in bill markings, eye-skin, and throat color), indicating that none of these is very useful in reproductive isolation.

NO to all. This is a tough one, I've gone back and forth, and I really appreciate the effort and analysis on this proposal. Although some of the rationale presented is reasonably convincing concerning the accumulated phenotypic differences among these taxa as likely being RIMs, I can't get past the following points:

1) In the one "test case" in which we have two parapatric taxa, cyanolaemus and atrogularis hybridize freely … or do they? Hybridization per se among BSC species is acceptable, but whether this is the case with these two is either not known or not made specific in the proposal. Lacking time and energy to investigate this myself thoroughly, I am going to assume that the signal from the contact zone is that there are no pure parentals, i.e. non-assortative mating. If that is the case, then empirically we conclude that strong differences in color of the bill, throat, and facial skin are not RIMs in the "test case." That’s a lot of "ifs" but that’s all I have to go on. By perilous extrapolation to the allotaxa, none of those phenotypic differences can be used as reasons for recognition of taxa that differ in all three of those characters, much less just one or two. I emphasize that that this string of conditional statements rests on what is likely an imperfect knowledge of the contact zone and thus easily reversed with more data.

2) Outside the prasinus group but within Aulacorhynchus, the other test case of parapatric taxa is calorhynchus vs. sulcatus, which was studied by Schwartz (1972). They differ dramatically in bill color but are similar in throat and face color, and there are no vocal differences. The contact zone sends a clear signal: non-assortative mating (nothing but intermediate birds). Therefore, by extrapolation, major differences in bill color among allotaxa are not sufficient evidence for species rank (with all the obvious caveats concerning such extrapolations … but it’s all we have to go on.). (By the way, HBW treats these two as separate species despite free interbreeding, in part because in their scoring scheme, hybridization counts as a "+1" point towards considering the two as separate species – go figure …. ).

3) Outside Aulacorhynchus, in Pteroglossus and Ramphastos, Haffer has shown that major differences in plumage and bill color do not serve as RIMs. Although none of the contact zones has been sampled as thoroughly as we would like, the signal sent from the contact zones is existence of hybrid swarms and lack of RIMs. Therefore, in Ramphastidae as a whole, divergence in plumage and bill color does not insure that these populations are evolving independently. In contrast, Haffer showed that sympatric species from different lineages within Ramphastos are more similar to each other in plumage and bill color than they are to their own closest relatives, i.e. plumage and bill color make no difference in their treatment as separate species. However, as Haffer also showed, vocal differences were clear predictors of genetic isolation.

4) Concerning the deep divergence between the groups as rationale for a split, this is often an indicator, yes, but it is a continuous scale with no objective way to be used in delimiting taxa. Also, Ben Marks showed that Bleda syndactylus bulbuls are 8-11% divergent in DNA sequence across minor African rivers with NO phenotypic divergence detectable among populations. Degree of genetic difference, IMO, marks time and divergence in explicitly neutral characters, not necessarily in any characters associated with cessation of gene flow.

The loud signal sent by toucan contact zones in general is that all those differences in bill and plumage color are insufficient for cessation of free gene flow and are not RIMs. In contrast, vocal differences are associated with absence of gene flow despite plumage and bill similarities.

In summary, although I appreciate the point that voice should not be regarded a priori as the only character that indicates species status, in this particular group, the limited empirical data indicate that vocal differences predict absence of free gene flow, whereas coloration patterns are irrelevant. So, if our species concept focuses on free gene flow, or lack of it, then vocal differences or lack of them should indeed be the criterion by which we assign rank to allopatric taxa. Recognizing that we all know the dangers of such extrapolations due to the serendipitous nature of speciation, I nonetheless see no alternatives other than whimsy. As noted by Donegan et al. (2015), all members of the prasinus group are, as far as is known, vocally extremely similar if not indistinguishable, and therefore, within the comparative framework of what we know about gene flow between toucan populations, this lack of difference indicates lack of divergence to the level associated with known cases of assortatively mating toucans, i.e., we should treat all taxa as subspecies pending further data.

IF the final vote was just for a 2-species split, I'm fine with Northern Emerald-Toucanet and Southern Emerald-Toucanet.  Keeping "Emerald" in the name emphasizes their sister species relationship, and follows the "new names for daughter species" principle. On the other hand, prasinus is "the" original Emerald Toucanet, as in Cory-Hellmayr, with other names applied to wagleri ("Wagler's") and various other names for the albivitta group. Unfortunately, Cory-Hellmayr used "Southern Emerald Toucanet" for A. prasinus virescens (Honduras-Nicaragua), but I doubt anyone would object an expanded use a century later.

ABSTAIN.  I find the toucanets confusing, and hardly know them in the field. Only in Costa Rica and Panama. I find the morphological differences relatively minor. I think that vocalizations can be important, esp. in suboscines, but these birds are diurnal and surely use visual cues also. I'll be happy with the committee's consensus.


Treat the subspecies (a) spectabilis and (b) viridiceps as separate species from Eugenes fulgens

(a) YES. (b) NO - 4 without comment.

(a) YES. The differences in plumage and gorget color of spectabilis are clear, the genetic data recover spectabilis as a distinct group, there is anecdotal evidence for vocal differences (see Pam's comments), and no justification for conspecific treatment was given to begin with. I like "Rivioli's Hummingbird" and am ok with "Admirable Hummingbird," although I am also open to other names. (b) NO. There is no evidence that justifies elevating viridiceps to species.

(a) YES. The apparent lack of data to lump them in the first place, combined with the fact that there is more evidence for separating them than maintaining a single species. I am fine with using Admirable Hummingbird for fulgens. (b) NO. There is some genetic structure here but not what we expect of species. Although Bayesian species delimitation found three groups, I’m not a big fan of this approach for diagnosing species (see Sukamaran and Knowles 2016).

(b) YES. (b) NO. Having spent considerable time in the high mountains of Costa Rica, where spectabilis is a common feeder bird, I've always been struck by how different spectabilis is—really, it’s unrecognizable as Magnificent, if you're only familiar with fulgens. The genetic data reinforce this impression. Re song, see and XC 106649, vs. the many recordings of perched male fulgens on XC, which show them to be strikingly different. Of course, a full analysis of vocalizations is needed, but I don’t see any fulgens song examples that are at all similar to those spectabilis songs. I think we should maintain viridiceps as a subspecies, but certainly not a species, at least until someone examines a decent series. I think "Rivoli's" is the best name for fulgens, but I don't have a great suggestion for spectabilis. I don’t mind "Admirable" although it does sound a bit odd. We should not continue to use "Magnificent" for fulgens, as spectabilis is common and rather widely distributed.

(a) YES. (b) NO. I see no reason to give deference to taxonomic changes made long ago by authorities with not a single sentence of justification. The phenotypic differences seem more different than some other Middle American species of hummingbirds I know, namely Calothorax Lucifer and Callothorax pulcher or Selasphorus helosia and Selasphorus ellioti. As for English names I agree that we should restore the old Rivoli's Hummingbird for the northern taxon and that leaving "Magnificent" for spectabilis is confusing. Why not go back to Ridgway's English name? He was treated rudely by having his English name and species rank pulled without any explanation, and, surely if the two taxa together (ignoring viridiceps) were thought to be "magnificent" before, surely either of them could be thought of at least "admirable" now.

(a) YES. This seems rather marginal. The degree of plumage differences would be definitely by of subspecies caliber, were it not for these being hummingbirds. But gorget color and other iridescent portions are often thought significant for species identification. We often harken back to Ridgway’s treatments, especially when they align with a proposed split. But how often was Ridgway wrong? (b) NO. May not even be a good subspecies.

(a) NO. (b) NO. Disjunct ranges are not species limits. Allopatric populations will through neutral processes develop genetic differences that will look like "species" using Bayesian species delimitation. To me that is not a useful approach to delimiting biological species. Based on the photographs and descriptions, I agree that a case could be made, relative to other hummingbird species, that spectabilis  is a good biological species. With that comparison not published, however, I would have to vote no at this time.


Elevate Turdus rufopalliatus graysoni to species rank

NO - 2 without comment.

NO, as per rationale outlined in proposal.

NO. The proposal gives good reasons for not splitting graysoni at this time.

NO. I agree that until more data are obtained, this is a subspecies. The genetics reflect island isolation, as expected. Same with the morphological traits. The mainland occurrences are intriguing, but seasonal movements with subspecies in sympatry part of the year without interbreeding are common.

NO. Lack of vocal information and very weak evidence for sympatry during the breeding season lead me to think that the status quo is the appropriate treatment.

NO. This case represents the classic problem of isolated taxa that are slightly differentiated and that apparently don’t actually come into contact. It’s hard to evaluate from a BSC perspective, but available evidence right now indicates a "no" vote.

NO. No vocal data and no other full species endemics from the Tres Marias. Phillips' assertion that they breed on the mainland is surely in error. There are lots of birders and birding tours that cover this part of Mexico, and I have to believe that territorial singing birds of this taxon would have been found by now in the late winter or early spring.

NO. Considering that island forms are often different in plumage from mainland forms (e.g., several taxa in the Queen Charlotte Islands), I see no reason to split graysoni. The sample sizes given in the proposal appear off (mainland n=32, Tres Marias n=12). DNA from mainland samples of graysoni would have been nice, but not sure what they would have added. It appears the only reason to suspect syntopy is the one June specimen. If we knew more about what was happening on the mainland during the breeding season, it would be easier.

NO, at least for now. Recordings and vocal analyses are needed, at least. Data from the mainland presumed zone of contact could be confirmatory, but if the June specimen was an accidental, no amount of effort would help here.


Recognize newly described species Arremon kuehnerii

NO - 3 without comment.

NO, as per rationale outlined in proposal.

NO. I'd like to see more data.

NO. I don’t really know what to make of this situation, but think it is certainly premature to split this off, at very least.

NO. Interesting data, but I agree with the proposal's rationale and recommendation for not recognizing it as a separate species.

NO. If we don’t recognize ssp. based on only molecular characters (e.g. Juniper Titmice) than surely we shouldn't recognize species on this basis.

NO. I agree that it is an intriguing situation, but if we start naming every diagnosable population based on putatively neutral genetic criteria we will have a name for every isolated population. Species do not have to be monophyletic at all loci (we humans aren't), and this description ignores the monophyly of the polygenic phenotypic traits of its obviously close relative suttoni. It will be a population genetics issue between two close relatives; recognizing a new species for it is not a good solution.

NO. The results are interesting but do not argue that kuehnerii is a separate species. It looks like either: 1) a northern population of suttoni had some past hybridization with virenticeps, and that some of that latter’s taxon’s DNA has become fixed with no plumage evolution. It would be surprising that nuclear DNA markers showed the pattern as well as mtDNA, but not inconceivable. Or 2) that the ancestor of virenticeps, which looked like suttoni, split into two populations: one, leading to virinticeps, had lots of plumage evolution; the other population, leading to kuehnerii, did not have any plumage evolution. In either case, reproductive isolation would not exist between kuehnerii and suttoni.


Revise the classification of the Icteridae: (a) Add seven subfamilies; (b) Split Leistes from Sturnella; (c) Resurrect Ptiloxena for Dives atroviolaceus; (d) Modify the linear sequence of genera

YES - 5 without comment.

YES. Rationale outlined in proposal.

YES. All recommendations are consistent with the phylogeny, and subfamily recognition is consistent with what we did with the tanagers.

YES. SACC will need to think about whether to maintain Meadowlark for the "Pezites."

YES to six subfamiles (a). I think keeping Amblycercus in Cacicinae makes more sense given the morphological similarities. (b) YES. This is not necessary for monophyly, but indications from a very large data set indicate that the age of this split was genus-level split, and some morphological/behavioral also back this up. (c) YES. Necessitated because of non-monophyly. (d) YES.

YES. Regarding (b), although I prefer morphological criteria in addition to just split depth for genera, I presume such existed for Leistes.


Revise familial limits and the linear sequence of families within the nine-primaried oscines

YES - 3 without comment.

YES. Reasons outlined in the proposal.

YES. This solution is a marked improvement over the current treatment - recognizing that some, especially the Nesospingidae and Spindalidae, may see future changes at the family level..

YES. For me, the big question is whether we 1) accept some tentative relationships and fix what is clearly wrong now in the list, or 2) wait until we have better resolution but accept that what we currently have is wrong. The first solution may necessitate further rearrangements, which I like to avoid. But the second solution may be waiting on resolution that may never come. This is not a huge data set compared to what is possible today, so better resolution may be coming. Nonetheless, it is difficult to let sit a taxonomy we know is incorrect.

YES to all proposed changes. The "Nine-primaried Oscine Coalition" has demonstrated convincingly the monophyly of these groups and the rationale for treating each in a separate higher-level taxonomic category. The question is … what level of higher category? At one point, I leaned towards a collapse of many of the traditional 9-p families into a single family, with subfamily designations for each lineage, i.e., each of the current groups would be treated at the subfamily level, along with Parulinae, Icterinae, etc. However, using the time-calibrated tree in Barker et al. (2013; Fig. 9), all but one of the lineages in the current proposal are estimated to be at least 10 million years old, which is consistent with the lower end of the estimates of family-level divergences in other groups (barely). The slight exception is Nesospingidae and Spindalidae, with an estimated divergence time of 9 MYA; however, given the inherent assumptions in such calculations combined with the current lack of formalized use of lineage age for assigning ranks in higher-level classification, it is pointless to argue about an arbitrary threshold at this time.

YES. One thing I would like to see the NACC take responsibility for is when we tinker with families and create new ones (multiples with this motion) why not give the families new English names? This was annoying in NGS 7th. when in a few cases we had to come up with English names. For instance, in the case of Muscicapidae, calling it simply "Old World Flycatchers" is misleading as the family now includes many genera (e.g. Luscinia, Tarsiger, Oenanthe, and Saxicola) that are referred to as "chats" in the Old World, so we called it Old World Flycatchers and Chats. What do we now call the monotypic family where Icteria virens has been placed. Surely, not "Chat," which would invite endless confusion, and what other name than "Yellow-breasted Chat?" Some like the monotypic Rhodoinocichlidae seems easy (Thursh-Tanager), but for some of the other newly formed families where there are a few West Indian species in each (or just one), the English names aren’t so easy, but are still necessary. We are a Committee given the task of nomenclature. Some of our most vigorous debates have involved English names for species. If we give English names for species, why not for families? Keeping it short, yet informative, and clear to our readers should be the goal.

NO (reluctantly). I would like to have seen more discussion of why subfamilies would not be appropriate for some of these taxa. The number of species involved among them all is in my view not a cogent argument. I realize that I will likely be in the minority here, but the only voice I have on this issue is through my vote, and I think we’re starting to look ridiculous in finding seat-of-the-pants family and order limits that are inconsistent among the class Aves. Compare node ages here with those in the same genus in the Anas proposal later in this group. If we consider clade age to be a useful yardstick for higher taxonomy (and I think we are headed strongly that way for families and above), we are wildly out synch: at present, the genus Anas would contain family-level taxa if we extended Parulidae reasoning to Anatidae (yes, that would look ridiculous – but we seem to be comfortable with the unevenness of these treatments? Please let me know if I am looking at this wrong.). I predict eventually we will decide what a family-level entity is across Aves and revise some of these down a notch (I think this will likely pass). Other than that, this is an acceptable step forward and based on a good body of work.


Lump Acanthis flammea and Acanthis hornemanni into a single species

YES - 1 without comment.

YES. Although questions remain, at what point does the preponderance of data overshadow all doubts? If we decide to exclude cabaret (fine by me), are there any studies that show or suggest that hornemanni and flammea are reproductively isolated?

YES. After reading the paper again and reading all the comments, I still think the M & T study indicate there is greater evidence for treating them as one species.

YES. The burden of proof should now be on showing that there is more than one species. Inclusion of the missing taxa in phylogenetic analyses and field work in key sites is needed, but may not happen for a long time. Even when/if they do, it seems likely that Salomonsen's findings of hybridization in Greenland would be further substantiated, and anything other than low or no levels of genetic divergence would be surprising. And, given that there are intermediate individuals, the explanation that environmental factors lead to differential expression of genes, resulting in large, whiter, smaller-billed individuals the farther north one goes seems plausible. Phenotypic non-differentiation between pewees and Empidonax alnorum/traillii does not seem particularly relevant here, since they differ conspicuously in song. As for keeping cabaret separate, I think it is appropriate that we leave responsibility for that decision to other taxonomic bodies.

YES. To me the data seem to overwhelmingly indicate that these are not two species in North America. It seems clear that these are broadly integrating forms, and this genomic dataset is actually quite good at telling us that these phenotypes are one gene pool effectively in panmixia. The question is not genetic diagnosability. In this case we have phenotypic diagnosis between two populations (Mason & Taylor Supplementary Table S1) that have an awful lot of gene flow between them from a popgen perspective. Even under strong divergent selection, it only takes a few 'migrants' per generation (genetics term for effective gene flow) to bind lineages together evolutionarily (and the effects are nonlinear). M&T's evidence is that these forms are far above that. If we decide to tie ourselves to the idea that genomic data are not strongly informative if we do not pin down loci associated with key phenotypic differences, we're making a huge mistake. Adaptation often causes some loci to segregate with high frequencies between populations; that can be very different from speciation. I don't think there is any other interpretation from those genomic data than that gene flow among neutral markers (which will be the great majority of those data) is very high. And we prefer to infer levels of gene flow using neutral loci. While those data suggest panmixia, we can also infer from phenotype that it is not that extreme because the two forms are extant. In a species-limits context it is not a question of whether there are differences between the genomes of the two forms -- I would be surprised if we could not find differences. What matters in my mind is what we can infer from population genetics, and that's a lot of gene flow. That is a strong dataset for that question (and I hope I'm clear on why genomic similarity/dissimilarity is a different question).

The concept of a normal contact zone is only applicable in western Greenland, where tundra and taiga habitats are not highly interdigitated, unlike most of the rest of the Holarctic range. I'd say that given the widespread evidence from Alaska + genomics that the two are not reproductively isolated, it doesn’t matter to me if they appear to be at one place, e.g., if they mated assortatively in Greenland. That’s ring-species-like. Genomically not being able to discern two populations in North America from that much data is telling us a whale of a lot. You can have a ton of assortative mating and still have enough non-assortative mating to produce high levels of gene flow while retaining phenotypically pure forms. And there is a lot of assortative mating. Assortative mating would have to be high at most if not all contact zones, and despite mate guarding and males feeding females at nests, forced copulation attempts add a dimension of breeding behavior that would also have to be accounted for. In the BNA account for flammea I see that 20% of nests surveyed showed extra-pair fertilization (1 of 18 young).

There may be a fairly simple solution for now: Move exilipes into flammea, continue to recognize it as a subspecies for now, and leave nominate hornemanni for future work to determine. It seems that there is evidence of assortative mating in some places and not in others (anecdotal accounts of mixed pairs in Alaska are common).

NO. I don’t find the additional data provided in the proposal compelling. I am frustrated by how hard it is to tell exilpes and flammea apart. The genetic data are suggestive that these at least should be lumped, but I agree that we would really like to see fieldwork that really shows us how these forms interact. We still remain with no genetic data on hornemanni (s. s.) relevant to this taxonomic question. 

NO. Detailed comments fromn others, including those regarding the genomic sampling and the need for field studies of breeding pairs where they contact, have convinced me that its best to hold off on lumping these taxa.

NO. I have no field experience in Baffin Is or Greenland, but I still worry about equating redpolls from there as being conspecific with Alaska-northern Canadian birds.  I would have no trouble accepting that the latter blend into Common -- based on my experience and that of Declan Troy and (via Clayton White) of Springer. But I think that equating all "Hoary" Redpolls as one thing is dangerously typological.

NO. Until there is a detailed study of birds on the breeding grounds to determine if there is assortative mating between these two taxa (Sensu lato), I will have little interest in making any taxonomic changes between these species, subspecies or "color morphs." When I first saw this proposal last year, I voted yes, mainly I saw it as a relief that I didn't have to separate a large number of vexing birds (not just mine, but many that I was sent photos of – why I was thought to have some special insight I haven't a clue). I had become to believe what had become the word on the street, that redpolls were a hybrid swarm worthy of little, if any, taxonomic recognition between the various types. This was sort of the attitude expressed in Kessel (1989): "Recently, however, because of an uninterrupted cline in morphologic characteristics, a geographic and ecologic overlap in breeding ranges, a lack of karotypic differences, and extensive hybridization, Troy (1980, 1985) has recommended merging these two forms into a single taxon, C. fl flammea. The 'Hoary' and 'Common' forms appear to represent population extremes of a single variable species, so although both extremes are present on the Seward Peninsula, I have followed Troy in treating them as a single taxon and have not attempted to distinguish the 'forms'." This period of relief was followed by a gradual descent into worries about what have we done? If they aren't species, than what are they? We didn’t vote on that, we just voted to lump, the old "repeal" but no "replace." Did we continue to recognize them as separate ssp? And, how could we do that if the ranges of some of the taxa from the two groups were broadly sympatric? And, if they are just morphs than why the mensural differences? So, in the end we stepped back from the brink and agreed to revisit the issue this year. Now, it’s back along with "new information" that I don't find particularly any more compelling.

This time I took a few days to carefully review Knox (1988) and also Stoddart's (2013) long overview in British Birds. I would strongly recommend that NACC members carefully review these papers. I do apologize if you have and I still have a different view. Knox's dissertation (1988) is particularly compelling for maintaining redpolls as two separate species (Lesser Redpoll was maintained as ssp. of A. flammea but Knox foretold the situation by stating that if it overlapped with A. flammea flammea during the breeding season and showed assortative mating than it would need to be recognized as a separate species. Shortly afterwards this happened in southern Norway and the BOU and others recognize Lesser Redpoll, A. cabaret, as a separate species. Some of the key points to emphasize from Knox’s (1988) overview:

Both Knox (1988) and especially Stoddart (2013) give a detailed overview of the complex situation in Iceland. Here, both dark types and pale types exist. Salomonsen (1951) indicated that the breeding population contained both dark and pale birds, but he regarded them as a 'hybrid swarm' where the isolating mechanisms had broken down. This might be the 'smoking gun' but more recent work conducted by Herremans (1990) showed biometric differences between these two "forms," the pale birds being smaller billed and longer tailed than the dark birds. The measurements of the "pale birds" are distinct from all types of dark redpolls and while similar to exilipes and hornemanni they are different from both. The dark birds are basically indistinguishable from rostrata. Given this, Herremans (1990) concludes this cannot represent color morphs of a single form, and they do not support the notion of a "hybrid swarm." I can only conclude that this represents yet another location where these two types (Hoary and Common) are sympatric and may well be interacting as separate species, but again more study needed, especially on whether these two types show assortative mating.

The authors of the proposal believe that the 'burden of proof' has shifted to those who would treat flammea and hornemanni as separate species, this despite no genetic testing of the two most different taxa and most importantly no definitive studies on the breeding grounds examining known nesting pairs. The majority of the more anecdotal evidence (some fairly strong in my view) does show assortative mating. At least 'the burden of proof' should fall on those who believe the conclusions are in error. Remarkably, the authors then go on to suggest maintaining cabaret as a separate species, this despite no distinctive genetic separation, but rather on the one breeding study that showed assortative mating (based on eleven nests). They base this on the following: "evidence of assortative mating and apparent reproductive isolation in certain populations and the phenotypic distinctiveness of A. cabaret." Huh? The very same arguments can be made for the situation with the various taxa in the Common and Hoary groups. Additionally they add, that A. cabaret is an extralimital taxon (well, a Greenland record) and we should defer to "local authorities." I’m not sure who the "local authorities" are, but Knox is, to the best of my knowledge, the world’s authority on redpolls and others such as Lars Svensson has opined on this matter and is internationally respected on all European taxonomic issues. More-over our problems are their problems. All taxa, except islandica occur in the both the Nearctic and the Palearctic. I think Old World authorities have been trying to solve this problem a good deal longer than we have.

Joseph R. Jehl, Jr. who did a great deal of work at Churchill, particularly on shorebirds, but studied other species too, including passerines, discussed  the interesting situation there on redpolls and detailed how the number of exiliples varied greatly from year to year, in some being equally as  numerous as Common Redpolls, while in others being nearly absent. He (2004) writes a very useful essay on redpolls (p. 137, "The kinds of Redpolls") that states: "Species limits among North American redpolls have been a long-standing question. Jehl reported "assortative mating between white-rumped (C. h. exilipes) and dark-rumped (C. f. flammea) indicating that the two forms are behaving as distinct species. There are also differences in plumage (Seutin et al. 1992, 1995), habitat (Hoary more abundant in open country) and song (Hoary’s higher and tinklier). From studies of the skins of breeding adults Seutin et al. (1992, 1994) concluded that the plumage differences were associated with mensural differences and that intermediates were rare, which further supports the idea of different species. What remains to be studied in detail is how the birds respond to each other with respect to mate choice. Such data should be easy to amass, because males feed females at the nest. This would allow both members of the pair to be captured in a mist net for examination."

Concise and well-stated. The sympatric overlap between these two species is huge, well really circumpolar. There are lots of easily accessible places where a breeding study with the protocols outlined by Jehl (2004) could be done. Any taxonomic decisions made now before such a study or studies are undertaken and published strikes me as imprudent. I have been well-educated about the genetic similarities between these two groups, but as long as they behave as separate species and show assortative mating, aren't they still good species, or have we redefined what really is a species, or isn't?

References cited in vote:

Further comments: I would agree that lumping nominate hornemanni in with flammea at the species level is worse than lumping exilipes in with flammea at the species level as in the former case the phenotypic differences are greater (especially in comparison to rostrate) and there is no molecular data to support such a move. But why not await an actual nesting study to see what the birds actually do? Jehl  (2004) has outlined the methodology and says this would be easy to accomplish. Locations where both taxa are nesting (exhilipes and flammea) are easy to access. There is enough out there (both Old World and New World) to indicate that even with exhilipes and flammea there is assortative mating taking place (or not). Seems to me that lumping now would be like saying we don’t care what the birds actually do, the only thing that matters is the molecular evidence. At least recognizing the various taxa as ssp. is an improvement on calling them color morphs, but if that’s the case I can’t think of any other comparable case in nature (well, birds) where there is such a huge overlap zone between two ssp. 

NO. If Lifjeld and Bjerke (1996) documented total assortative mating between flammea and cabaret, yet Mason and Taylor (2015) found no evidence for differentiation between them, then why should lack of consistent differentiation between flammea and hornemanni be treated as sufficient evidence for their treatment as conspecific? As I see it, the reason for the difference in treatment is the lack of phenotypic discreteness between the latter two plus some evidence for lack of reproductive isolation. However, we treat pewees and Alder/Willow flycatchers as separate species despite lack of clean phenotypic differentiation between them, so this concerns me. For treatment as conspecific, I would be soooooo much more comfortable with a good field study, i.e. transects through the contact zone with analysis of paired individuals, genetic samples, and vocalizations. I realize that logistics of this might be difficult at those latitudes. However, the tight mate-guarding of carduelines would offset some of the disadvantages of remote terrain. A nagging problem for me is that the circumpolar contact zone between the two types of redpolls is the longest of any on the planet, by far, yet the two types seem to be maintained to some degree (vs. smooth cline). If I understand the situation correctly, this would require somewhat abrupt environmental induction of the two phenotypes throughout their ranges in response to some universal factor(s). Thus, the system would be extremely unique in birds, and so I think the standard for acceptance of this interpretation should also be extremely high. Also, the incomplete taxon-sampling makes it a true dilemma as to what to do, taxonomically, with rostrata, and, most critically, nominate hornemanni.  Sampling the latter is obviously necessary for taxonomic assignment of the name hornemanni (and is a design flaw in this study if it is to have a taxonomic interpretation). Finally, as for the lack of overall differences between the genomes of the two, without a comparative context, this is difficult, for me anyway, to interpret. Are there comparative data? Galapagos finches? In fact, comparisons to the latter system would seem relevant in that those birds are also extremely recent in origin, with occasional to regular hybridization, bill size differences, and subtle plumage differences.

Further comments: I’ve always been suspicious of the Hoary Redpoll as a species. However, after reading other comments and the citations provided, eagerness to lump these taxa has definitely been tempered. Then, I started thinking about the genomic data that at first seemed so convincing and wondering whether our confidence in their interpretation might be premature — without comparative data from recently diverged taxa for perspective, do those data really mean anything? Without comparative data, we really don’t know how to interpret this degree of similarity. Although the genomic data show an impressive lack of differences between the two forms, until we have similar data from very recently diverged species, I think we should be cautious in their interpretation. What would comparable data look like from, say, Nelson’s-Saltmarsh sparrows, Western-Clark’s grebes, Eastern-Spotted towhees, and other species that are ranked as such because of assortative mating but that are also exchanging genes? In general, I worry about over-interpretation of relatively new types of data that have yet to have time to "mature." For example, some of us may remember the very early days of tree-building from genetic data in which we did not appreciate the need to establish support values for nodes, or the early days of DNA sequencing when we did not understand the perils of single locus studies. Therefore, rather than potentially repeat that initial stage of over-enthusiasm for new data, I will remain cautious until the field matures.

I would further argue until there is an actual formal field study of frequency of mixed pairs vs. pure pairs, we don’t really know. In fact, if you CAN tell the two phenotypes apart in the field to the point of telling that a pair is mixed, that alone says that mating is assortative — otherwise where they are in contact no pure birds would be evident after a mere 10 generations of non-assortative mating. That we have two distinct phenotypes may indeed be an example of environmental induction, but the genomic evidence is not sufficient for that. In general, I would argue that all of the Nearctic data is bassackwards concerning species limits — let’s see what the birds do in terms of pairing and mating in the field. Serious mate-guarding in carduelines should facilitate such a study. In contrast, the Palearctic data evidently generally lean towards some degree assortative mating.

Bottomline is that the Mason-Taylor data are fascinating and provocative, but not sufficient for concluding that the genomes are indistinguishable. Personally, all I would need is one actual field study in North American where the two types occur to convince me one way or another, i.e. same standards of evidence required for any and all species-ranking decisions between parapatric/sympatric taxa, and one that would include genetic sampling to detect EPCs. If we lump these two, this would be the first time (at least since the egregious Lumperama era of the 40s-60s) in which two putative species have been lumped without a formal study of their contact zone.


Split Lanius excubitor into two or more species

YES - 5 without comment.

YES. I found myself wondering what made the River Ob important biogeographically as a species separating barrier on shrikes (eastern limit of excubitor with sibiricus farther east). The range, as mapped in the proposal looks continuous so one wonders about interbreeding, etc. But, if my interpretation of Harris and Franklin (2000) is right than the more easterly taxa, including N.A. borealis retain a juvenal plumage well into the winter, and at least for our birds we know this to be true, while in excubitor the juvenal plumage is much more adult like with the more adult plumage being acquired by fall. This seems important and was enough to tip the balance for me along with the molecular evidence.  Please correct, if I’m wrong on these assumptions. Reference: Harris, T., and K. Franklin. 2000. Shrikes and Bush-Shrikes. Christopher Helm, A & C Black, London.

YES. Although given that shrikes are known to hybridize rather frequently, and that mtDNA may be a particularly poor marker to elucidate relationships in this group, there is paraptry (or narrow sympatry) between the westernmost borealis (mollis) and the easternmost race of excubitor (pallidirostris; Panov, 2011, cited from del Hoyo and Collar 2016, HBW Illustrated Checklist). The species differ consistently in the light barred and very pale ocher underparts of borealis, aside from the race of the latter in the Kuril Islands, which is like excubitor, but genetically the most distinct race of borealis (del Hoyo and Collar 2016).

YES. I am concerned that a new taxonomy would be based largely on mtDNA gene trees, which are vulnerable to incomplete lineage-sorting and hybridization, as clearly articulated and suspected in these shrikes by Olsson et al. (2010):"Since the mitochondrial gene tree deviates substantially from the (non-cladistic) interpretation of relationships based on morphological and ecological characteristics, and there are indications that the gene tree might not fully conform with the organismal phylogeny, any proposed taxonomy is uncertain." For example, two borealis samples are nested within ludovicianus (although I would like to have seen a sentence confirming ID of the vouchers). However, unless comments from others convince me that holding off until genomic data are available for this group is the wisest decision, the available data require a taxonomic change, namely from our perspective, recognize Northern Shrike (L. borealis) as a separate species from Great Grey Shrike (L. excubitor) and relatives, which are members of opposite clades in the complex.

Out of curiosity, I looked at online recordings of borealis vs. excubitor, and although the differences aren’t great, borealis seems to have a song composed of simpler notes, e.g.,

NO. I am having trouble with this one. The Johnsen et al. (2010) tree is weak for determining species limits. However, with Olsson et al. (2010) and Peer et al. (2011) supporting that general relationship it looks like species limits are involved. Non-monophyly of excubitor sensu lato does not bother me much. The location of the contact zone seems odd. Phenotypic diagnosability? From Tajkova and Red’kin (2014, esp. Fig. 7), it looks like L. borealis sibiricus is no more distinctive than the two subspecies e. excubitor and e. leucopterus (i.e., subspecies-level difference). How much hybridization in the contact zone? On their p. 96 they state that "The wide individual color variability of European populations may be easiest explained as a consequence of the wide interspecific hybridization between L. excubitor and L. borealis." This looks to me like a case where we are likely to see substantial levels of gene flow across a hybrid zone, and I am more comfortable considering them as subspecies. It reminds me of the situation in Green-winged Teal, except that the hybrid zone is in a weird location.


Add Mangrove Rail Rallus longirostris to the main list

YES - 4 without comment.

YES. Good work on obtaining the specimens and elucidating the taxonomic relationships.

YES. Genetics confirm that these rails beling with Mangrove Rail and not R. crepitans 

YES. Specimens verified by DNA and plumage by expert.

YES. This seems pretty straightforward.

YES. The evidence looks rock solid.


Revise the generic classification and linear sequence of Anas

YES - 3 without comment.

YES. Although most of the odd South American genera (except to my knowledge Tachyeres) have at some time or other been placed within Anas, doing so results in an excessively broad genus incorporating several ancient lineages. The solution proposed is far better, in my view.

YES. It has been a trashcan genus and this looks like a good solution. See also my comment on the Parulidae revision.

YES. Although a phylogeny that includes genes other than mtDNA would be helpful, the tree in Gonzalez et al. (2009) is quite well-resolved to deal with questions as to placement of species in genera. Splitting of Sibirionetta and Spatula are a must given the non-monophyly of Anas, and no desire to incorporate Tachyeres et al. in Anas. Given the depth of the splits of formosa from Spatula and Mareca from the other Anas, it seems a good idea to recognize these genera as well.

YES. This is entirely about maintaining TachyeresSpeculanas and Lophonetta have often been treated as Anas, and Amazonetta to my eye could go in easily. But I agree that lumping Tachyeres in Anas would be a difficult pill to swallow. Thus, we need to split Spatula from Anas at least. I would rather not split Baikal Teal off into a separate genus, but find it hard to accept lumping it into Spatula, which otherwise seems like a pretty clear group. I accept the return to Mareca because of the strength of the split, and the relative age.

YES. I like this arrangement. I'm not sure about wigeon and Gadwall being that close but female Falcated Duck looks very much like a female wigeon. I like having Blue-winged and Cinnamon Teal clustered with Northern Shoveler (all with a similar wing pattern including by sex). Garganey is more uncertain to me. As different as Northern Pintail and Green-winged Teal look the males give calls that to my ear are not separable. Seems odd to have Baikal Teal start the sequence. I'm wondering about the old motion to split A. p. diazi, the Mexican Duck. If I remember correctly, Mallard might not even have been the closest relative (Mottled Duck was, I think). Was it included in the sampling?

YES. As noted in the proposal, Dickinson & Remsen 2013 had arrived at the same classification as the best solution to the existing data. Genomic data will soon be available for this group (and everything else), but I predict that it will support the mtDNA tree because the groupings are consistent with morphological data that was used in pre-Delacour et al. classifications.