2018-C-1: Adopt (a) a revised linear sequence and (b) a subfamily classification for the Accipitridae

YES. 2 without comment.

YES. Neither study would stand up to today’s genomic level analyses. There probably is much of the resulting taxonomy that is correct in the proposal, but I feel a new much more thorough study will undoubtedly necessitate further changes. But leaving things as they are is not viable.

YES. I agree that the proposed changes are better than what we have now.

YES. Matches SACC and H&M.

YES. Three subfamilies is much better than 14. I do look forward to someday having a set of guidelines for what depths families and subfamilies should be across the whole class.

YES. The work supporting this change is a little older than ideal. There may be changes required down the line, but I think this is enough of an improvement over the status quo that it is worth making these changes.

YES. Newer data are needed but this proposal is better than the current classification.

YES. One thing I wondered about was the linear sequence within Buteogallus, in particular having Savanna Hawk (B. meridionalis) between Common Black Hawk (B. anthracinus) and Great Black Hawk (B. urubitinga). This was done after AOU 7th (1998), but I don’t recall it. At some point a split of Cuban Kite (Chondrohierax uncinatus wilsonii) from mainland populations of Hook-billed Kite should probably be considered. Also, Gundlach’s Hawk (Accipiter gundlachi) for me might better be considered a subspecies of Cooper’s (A. cooperi) as it closely resembles that species in plumage and especially structure. Most importantly, the vocalizations are nearly identical. I have had better luck in getting Gundlach’s Hawks in Cuba to respond to play-back of Cooper’s than of Gundlach’s!

YES. There is also a more recent paper by Lerner et al. (2017) that is restricted to the “booted eagle” group, which is primarily Old World; this paper further supports both the existing and proposed sequence of Aquila and Spizaetus.

Lerner, H., L. Christidis, A. Gamauf, C. Griffiths, E. Haring, C. J. Huddleston, S. Kabra, A. Kocum, M. Krosby, K. Kvaløy, D. Mindell, P. C. Rasmussen, N. Røv, R. Wadleigh, M. Wink, & J. O. Gjershaug. 2017. Phylogeny and new taxonomy of the booted eagles (Accipitriformes: Aquilinae). Zootaxa 4216: 301–320.


2018-C-2: Split Yellow Warbler (Setophaga petechia) into two species

NO. Thanks to Ralph to putting together this detailed proposal, which has bits and pieces that suggest possible species status of aestiva. However, I am not convinced by the data to date. A more comprehensive study that involves modern genomic methods, quantitative morphological and plumage color analyses, and quantitative song analysis with playbacks is warranted.

NO. Nothing really new is presented in this detailed proposal, except that the Galapagos taxon may (or may not) be genetically distinct. The analyses of differences between aestiva and “tropical” petechia/eritharoides in genetics, behavior, morphology, are not very rigorous. None of these differences, either separately or collectively, would lead me to say that these groups are reproductively isolated.

NO. More modern genetic approaches and comprehensive geographic sampling are needed. Brown and Klein is based on restriction site mtDNA, and the sequencing studies that have been done have limited taxonomic sampling. Of course, more study of potential reproductive isolating mechanisms, such as a more quantitative analysis of song, would be helpful in applying the BSC. It’s too bad we lost Nedra, she would probably have the answer by now.

NO. I’m rather strongly against this split, as my reading of the evidence is that there is not strong support for a biological species division anywhere in the range of this highly variable and widely distributed species. The proposal is admirably comprehensive but at the same time it does not do a very good job of highlighting the information that is most relevant to determining species limits within this complex; it is more like a review of variation across the entire range. It also places too much emphasis on the mtDNA results from decades ago, and compares mtDNA divergence values that derive from different markers. To me the evidence here seems much less strong than for some other potential splits that have not passed recent Committee review (for example, the Yellow-rumped Warbler split proposals).

NO to split at this time; NO also to changing English name to American Yellow Warbler because I anticipate that with increased work on the group we will eventually make this split. This is one of the most thorough proposals I’ve read and does an excellent job of summarizing the situation. We’re basically assessing allopatric lineages to determine whether they are different enough to be full biological species. MtDNA is not a good species limits marker, in my view. So divergence levels are not compelling. And while there are many differences between aestiva and the tropical taxa and the preponderance of evidence suggests species limits could very well be present, nothing for me pushes it over the edge from being good subspecies to being clearly biological species.

NO. As others have mentioned, there was a lot of detailed information presented in this proposal, but much of it not relevant to taxonomy, or if it is relevant it was difficult to parse the wheat from the chaff. There may very well be multiple species here, but the proposal itself did not get me there. This system is begging for a detailed and thoroughly-sampled of genomic analysis of geographic variation.

NO. I think that the northern migratory group is a distinct species from the tropical group, but this proposal provides little evidence that supports that change. I do agree that we need to add a modifier for Yellow Warbler. I am hopeful that adding American as a modifier will not cause much angst. We’ve done similar things with Black Duck and Tree Sparrow. If the northern aestiva group is split, I would favor Northern Yellow Warbler for it, leaving American Yellow Warbler for the entire complex.

NO. I greatly appreciate Ralph taking the time to write all of this up into a formal motion. He has certainly spent a near lifetime working on this group, and carefully worked me through the nearly 50 sspp. at the USNM (1990’s), that he details in his epic paper on geographic variation in the species (Browning 1994). I can well appreciate the comments in defense of multiple species. What causes me to ultimately vote no is that I just don’t hear distinctive differences in either the song or the calls in comparison of North American sspp. to the more tropical subspecies. And, in play-back, the tropical sspp. always respond vigorously to North American songs and calls. And, I’ve conducted this play-back on dozens of occasions. Yes, plumage differences, wing shape (including shorter primary projection) are distinctive, and the immature plumages within the tropical sspp. group are just plain dull, mainly gray, unlike most of the North American sspp.

It is interesting to me that the northern taxa within the “Golden” group (gundlachi of Cuba and extreme southern Florida and flaviceps of the Bahamas) usually completely lack any chestnut on the crown, the classic mark for the subspecies group, at least for most of the recognized subspecies, while aestiva, the widespread and migratory subspecies from eastern continental North America show chestnut on the crown somewhat frequently. I’ve studied a small number of such birds in the field.

The gap between the breeding ranges of the two groups is interesting, but then I wondered about other species, for instance in Blue-gray Gnatcatcher there is a large gap in Florida between the breeding range of the migratory nominate subspecies and the morphologically distinctive and  resident Bahama birds. I wouldn’t be surprised if Ralph’s intuitions are right, but I just can’t split, yet.

Assuming that the split does not go through, I do not favor any English name change. I’m not worried about the various African taxa that are less “yellow” causing confusion –  certainly not enough to change this so well and perhaps the most familiar (other than Yellow-rumped) North American wood warbler.

Literature cited: Browning, M.R.  1994.  A taxonomic review of Dendroica petechia (Yellow Warbler).  Proceedings of the of the Biological Society of Washington 107 (1):27-51.

YES to splitting aestiva vs. all other taxa (hence two species). However, I am confused by the English names recommendations, which seem first to approve of the use of Mangrove Warbler for the petechia group (all but aestiva), and then to steer away from it on the grounds it has been used historically for just one taxon group. In the case of American Yellow Warbler for aestiva, they are all of course broadly American, so perhaps Northern Yellow Warbler would be preferable (though bland)? Mangrove Warbler is appropriate for the petechia group although it will be familiar to some people for a more restrictive taxon subset; still, it reflects the generally distinctive habitat well.

YES. Songs of only one resident taxon have been analyzed and compared to aestiva: Mennill’s results on quantifying bryanti (Central America) song differences vs. aestiva are suggestive but are based on just 7 male bryanti. His metrics revealed no overlap between a range of aestiva recordings and those of bryanti. I predict that additional data on song and call will show that the two major groups have diverged to the point in vocalizations associated with species rank in other for-certain good species in Setophaga spp.  However, as far as I know, we don’t have any playback trials, and so the pro-split evidence is admittedly weak.

Nonetheless, I vote YES for the following reasons:

1. Song: Song: Mennill’s quantitative analysis corresponds with the qualitative assessments of several of my esteemed field ornithologist colleagues with broader geographic experience than just with bryanti. For example, Bob Ridgely has been advocating a split for decades based on his experience. My experience (N=1) from two days ago with a bird here in Louisiana makes it clear that what my field colleagues say is correct: the song of the Louisiana bird was not recognizable as aestiva Yellow Warbler. Useless anecdote, yes, but if you roam online recording on xeno-canto, as I have done for all 11 subspecies represented there, you can hear consistent differences between them and aestiva songs, including dugesi from Mexican plateau. I like to think that I can tell that the tropical yellow warbler songs are closer to those of aestiva than to those of any other parulid, and in fact I suspect that some of the less emphatic songs of aestiva sound very similar to some of the tropical ones. Obviously, there is no substitute for analyses like Mennill’s, so I do have considerable anxiety.

Differences in song to do not keep GWWA-BWWA or TOWA-HEWA from hybridizing extensively, but these are dynamic, “aggressive” hybrid zones for which, as far as I can tell, there is no real hybrid swarm, which implies assortative mating to some degree, and in both cases, we continue to treat them as separate species. In contrast, no one has documented song differences between Audubon’s and Myrtle warbler, and this is likely related to their non-assortative mating and stable hybrid swarm. (Those two don’t seem to treat each other as separate species when it comes to mating … so why should we humans?). Empirically, therefore (as in the case of all non-controversial North American parulid species), song differences are associated with barriers to free gene flow.

2. Call note: “Everyone” says that the call notes are “not close.”  This could be dismissed as unquantified birder hearsay, but again, perusal of the considerable number of call notes from many tropical taxa online confirms this qualitative assessment. Again, for what it’s worth (= virtually nothing) my experience from two days ago with a bird here in Louisiana was consistent with the hearsay and the qualitative assessment of online recordings. There is no way I could have recognized that Louisiana bird’s call as an aestiva, a call note with which I am so familiar (e.g., nested in my childhood Colorado backyard) that I can shamelessly boast that I identified Bolivia’s first record by call before I saw the bird.

3. Breeding habitat-niche. Our aestiva breeds in a variety of second-growth and riparian habitats from Alaska, including the Aleutians, to Mexico, from Arctic scrub to hot desert. It is especially associated with willows. However, even though the lowland southeastern USA has extensive willow stands right to the Gulf Coast, Yellow Warbler does not breed there. The lower Mississippi has vast stands of willow-cottonwoods that farther north would be seething with breeding aestiva.  Something is preventing their southward spread into what looks like optimum habitat, structurally. Although aestiva Yellow Warblers can tolerate temperatures much higher than those of summertime southeastern USA, evidently there is something about the heat-humidity combination that directly or indirectly limits their breeding range despite wintering in hot-humid habitats. In stark contrast, Mangrove Warbler tolerates heat-humidity combinations that are as bad or worse as the southeastern USA throughout most of its range. Although Mangrove Warbler is not completely restricted to mangroves, expanding locally into dry scrub and humid forest, mainly in West Indies, something keeps it from expanding out of mangroves into early successional habitats in, for example, south Florida, where gundlachi is restricted to mangroves. The point of all this is that these two lineages have, in my opinion, diverged past the “point of no return.” e.g. at which free interbreeding could ever occur; for whatever reason, their genomes cannot handle each other’s range for breeding, despite considerable flexibility in terms of habitat and (for aestiva) climate within their respective ranges.

4. Lack of evidence for treating as conspecific. Why are they considered conspecific?  Ridgway (1902) considered them as separate species.  So did Hellmayr (1935), who regularly lumped taxa considered separate species by Ridgway. I can’t yet pinpoint the AOUCL Supplement in which they were lumped, but it was before the 1954 (29th Supplement). If typical of the era, then the only justification likely provided was along the lines of “it was decided.” I’m sure that the argument was that the tropical populations with reduced chestnut on head bridge aestiva and chestnut-headed ones in terms of plumage. Try presenting that as evidence nowadays. I really think burden-of-proof falls on treating them all as conspecific.

I am a little concerned that we are overly influenced by compilations of differences between two taxa that are not directly relevant to species rank.  We even include a laundry list of these differences in the Supplement statements.  Browning’s in-depth proposal represents a lot of work with a very useful and interesting synopsis; however, many of those characters summarized are not directly relevant to species limits.  For example, we know of no two species ranked as such just because one is migratory and the other not, or just because they differ in molt schedules, etc.  Obviously, assuming all these differences are under genetic control, then, yes, they reflect differentiation.  However, I worry that we lapse into some sort of numerical taxonomy philosophy here (as in the Tobias-Collar-BLI phylogeny-free numerical taxonomy scheme) without focusing on what really determines (or is associated with) species limits in a group.  I’m not saying those differences aren’t important biologically, only that they may be irrelevant to assigning taxon rank in a group.  This is also analogous to comparisons of genetic distance at neutral loci, which really cannot be used to determine species rank, as in bar-coding.  Even when used in comparison to other members of the genus or family, I would argue that degree of differentiation at neutral loci tell us lots about degree of isolation and history, but nothing about differences in the loci that might be related to cessation of gene flow.  After hearing Scott Taylor’s terrific seminar a few days ago on the genomic approach to searching for those critical genes, it seems to me that it is only a matter of time before we hit gold on this.  (I also worry that proposals as in-depth as Browning’s may intimidate others from doing proposals that focus only on the essential data for assigning species rank.  At the same time, many thanks to Ralph for getting the ball rolling on this difficult group!)

Regardless of outcome of the vote, a separate proposal on English names is needed. This one is complicated. Losing the traditional name simple “Yellow Warbler” for aestiva would be painful — it’s never been known by any other name AND it is a remarkably apt name (in contrast to many wood-warbler names). This species is so thoroughly yellow that it shows up even in feather margins and shafts. In a family with an unusually high % of misnomers for E names (e.g. Connecticut, Magnolia, Worm-eating, Hermit, Nashville, etc.), it would be nice to preserve a good one, without the ugly compound modifier that most people dislike. Also, I think Mangrove Warbler is just fine for the tropical group despite the deep internal split. Although some populations occur outside mangroves, this is the primary habitat for both groups. If we can live with Prairie Warbler and Palm Warbler, then we can deal with local deviations in Mangrove Warbler. That two species of African Iduna warblers are called African Yellow Warbler and Mountain Yellow Warbler and two species of African Calamonastides warblers are called Papyrus Yellow Warbler and Zambian Yellow Warbler is a problem for international naming committees to fix in some other way than forcing a novel, compound name on one of the most widespread North American birds; other Iduna species are not called Something Yellow Warbler (for good reason). Certainly, all that creative energy in those committees could be focused to solve that problem, particularly since the African “Yellow Warblers” are no more yellow than some Hippolais warblers in same family. Hyphens or not, that they are all called “Something Yellow Warbler” is misleading concerning their relationships.


2018-C-3: Revise the classification and linear sequence of the Tyrannoidea (with amendment to original proposal)

YES. I very much trust Terry on this one — my votes mirror his preferred recommendations in the amendment.

(1) YES to 1b. (2) YES. (3) YES. (4) YES. (5) NO. (6) YES. (7) YES to a, NO to b. (8) YES to a, NO to b. (9) YES to move Tyranninae, NO to change in sequence. (10) YES  to movie Piprites, no to new subfamily.

(1) YES to 1b. (2) YES. (3) YES. (4) YES. (5) NO, a more complete phylogeny is imminent, so I’m told… (6) YES, though Doug makes a good case to wait. (7) YES to a, NO to b. (8) YES to a, NO to b. (9) NO to change in species sequence. YES to moving Tyranninae. (10). NO to raising family; YES to moving Pipritinae.

(1) YES to 1a. I think it makes sense to merge these genera into Tityridae. (2) YES. (3) YES. (4) YES. (5) NO – follow recommendation in proposal amendment. (6) YES. (7) YES to a, NO to b. – follow recommendation in proposal amendment. (8) YES to a, NO to b. – follow recommendation in proposal amendment. (9) NO, but YES to change in placement of Tryanninae following the recommendation in the proposal amendment. (10) NO – follow recommendation in proposal amendment.

YES. (1a) Tityridae already is pretty morphologically and behaviorally diverse; adding Tyranneutes, Myiobius, Oxyruncus, and Onychorhynchus is okay. (2) YES. (3) YES. (4) YES. (5) NO. Keep in Tyrannidae. (6) YES. Odd island endemic, definitely deserves species status. (7) YES to a, NO to b. (8) YES to a, NO to b. (9) NO change to sequence; YES to moving Tyranninae to precede Fluvicolinae. (10) NO to raising family; YES to moving subfamily for Pipritinae.

(1) YES to Option 1b, placing OnychorhynchusTerenotriccus and Myiobius in Onychorhynchidae, and for Option 1c, moving Oxyruncidae to follow Onychorhynchidae. (2) YES to this move but not to the sequence change. (3) YES to new placement of Platyrinchinae. (4) YES to restriction of Platyrinchinae. (5) NO. (6) YES, this is clearly necessary. (7) YES to a, NO to b. (8) YES to a, NO to b. (9) NO to change in sequence, YES to moving Tyranninae. (10) NO to adoption of Pipritidae, YES to moving subfamily.

I spent hours on this, seemingly just to understand the numerous issues and motions presented! And, it doesn’t help that I have little or no field experience with the species detailed. Seldom have I felt so inadequate on a motion. I do feel that elevating species to a family rank requires pretty solid evidence, so agree with all of the cautionary statements on that. I thank others for carefully detailing the issues, and will cheerfully follow exactly Terry’s suggestions on how to vote.

(1) YES to 1a. I prefer a subfamily Onychorhyncinae and Oyyruncinae as part of Tityridae. (2) YES. (3) YES. (4) YES. (5) NO. I’d prefer to see some objective criteria for what seems to me to be a very vague and unevenly applied “old” to justify family over subfamily status. Ohlson et al.’s Fig. 4 suggests that a much more even-handed definition for that evolutionary age a family should encompass would substantially alter their family-level conclusions. (6) YES. Zucker et al. (2016) clearly show that Nesotriccus needs to be merged with  Phaeomyias. (7) YES to a, NO to b. (8) YES to a, NO to b. (9) NO, but YES to change in sequence proposed. (10) NO to family status, for reasons given in my earlier comments on 5, but YES to moving Piprites.

First, many thanks to Dale Dyer for tackling this complex situation. (1) Clearly, Onychorhynchus etc. cannot remain in Tyrannidae.  I favor 1b (and thus 1c), i.e. new family Onychorhynchidae and retaining Oxyruncidae.  (By the way, a synapomorphy for Onychorhynchidae that bird skinners know painfully well would be highlighted if only we could just change the name to Ohnorippedtheneckidae.) (2) YES. Essentially mandatory based on independent data sets. (3) YES, ditto. (4) YES, I think, assuming that what was meant was also including Neopipo and Calyptura in that subfamily, not in Platyrinchus, and retaining Piprites in its own subfamily. (5) NO.  I recommend waiting until we have a fully time-calibrated phylogeny of these groups to assign family rank; meanwhile, no harm done retaining them in Tyrannidae. (6) YES, assuming that what is really meant is to merge Phaeomyias into Nesotriccus, which has priority. (7) YES, but do not change the sequence until UCE phylogeny published. (8) YES, but do not change the sequence until UCE phylogeny published. (9) YES, but do not change the sequence until UCE phylogeny published. (10) NO.  Although I submitted and strongly supported this same proposal myself to SACC, it was rejected, mainly for the same reasons I voted against 5. No harm in waiting until we have a time-calibrated phylogeny that will allow careful assessment of absolute and relative lineage ages so that we can do this carefully. Piprites needs to be designated as a subfamily for now.

In general it seems clear that there are some genera that are misplaced in the current big subunits of Tyrannidae (Rhynchocyclidae, Elaenidae, Fluvicolinae, and Tyrannidae). These cover subproposals 2,3,7,8,9 and appear uncontroversial to me. I think subproposal 6 does not belong within this proposal. It is not a consequence of the Tello or Ohlson work and is at a different scale than these other proposals. Ideally a larger more complete proposal would be provided for it. I have, however, voted on this subproposal below.

Subproposal 1 is voted on below. Subproposals 4, 5 and 10 are all related. The first issue is whether Platyrinchus and relatives (where is Calyptura shown to be related to Platyrinchus and Neopipo?) and Piprites should be kept in separate taxa or folded into Rhynchocyclinae. I feel most comfortable with folding these taxa into Rhynchocyclinae, so I voted against subproposal 4, and on subproposal 10, while I favor moving Piprites out of incertae sedis, I don’t want to create a new family or subfamily for it, thus my partial vote for that one. Given that the topology of Piprites and Platyrinchus relative to Rhynchocyclinae varies between Tello et al and Ohlson et al, I am most confortable recognizing all of these within the bigger Rhynchocyclinae. Finally as you might begin to suspect, given that I am not creating new units for Piprites andPlatyrinchus, I think that we should maintain Rhynchocyclinae as a subfamily of Tyrannidae, rather than raising all of these units to family level. So I vote no on subproposal 5.

(1) YES I vote for option 1a, moving Terenotriccus, Myiobius, Oxyruncus, and Onychorhynchus. This is complicated. Oxyruncus has long been recognized as weird. The genetic data seems clear that these genera are not “flycatchers.” I can’t see creating another small family for Terenotriccus, Myiobius, and Onychorhynchus. There doesn’t seem to be much gained by placing them in Oxyruncidae, which exists in recognition of the weirdness of Oxyruncus. These other genera do not share distinctive morphological, behavioral or ecological characteristics with Oxyruncus. As a result I think moving all of thee genera into the already morphologically variable family Tityridae is the way to go.

(2) YES. (3) YES. (4) NO. See comments at top. (5) NO. See comments at top. (6) NO. I don’t think we should do this until we address the species level issues that are raised by doing this. The paper by Zucker et al give a pretty clear argument for at least a minimal split of Phaeomyias murinus. Otherwise it seems to me we are only acknowledging a piece of the story here. (7) YES to a, NO to b. (8) YES to a, NO to b. (9) NO to change in species sequence. YES to moving Tyranninae.

(10) NO or only partially. I think that Piprites has been clearly shown in these two studies to belong with Platyrinchinae, and we should remove it from incertae sedis and place it at or near the beginning with Platyrinchinae. The issue of what rank to give it is dependent in my view on your treat of Platyrinchus and the rank you assign these taxa. My full thoughts on this are at the top of my comments on this proposal.


2018-C-4: Split Cory’s Shearwater (Calonectris diomedea) into two species

YES, based on the published information (genetics, vocalizations, morphology, and chemical differences). However, the evidence for intermediates would make me reconsider.

YES. Good evidence for assortative mating in sympatry and concordant differences in ecology, morphology, and genetic markers. I had not looked at the comments or other votes before coming to this conclusion and now I see that I am an outlier on this vote – it could easily be that I overlooked something important.

YES. Although the proposal presents a large number of studies each pointing to reproductive isolation, the comments about intermediates suggest the situation is less straightforward. I would like to see resolution of the degree of intermediacy. However, after reading the Zidat et al. (2017) paper on Calonectris, I am changing my vote from NO to YES on the split. Given that they breed largely assortatively on the same islands evidently driven by different vocalizations and chemical signatures, they are surely now sympatric species. (I don’t think there will ever be complete uniformity among committee members as to exactly how borderline cases will be interpreted, nor should there be, otherwise what is the point of having a committee rather than a single arbiter who checks to see whether any given species candidate ticks all the boxes?)

That said, I need to bring two further papers to the attention of the committee. Avian paleontology papers, even when highly relevant, tend to be overlooked, as they have been here. I agree that whether we can identify cryptic species on their wintering grounds is irrelevant to their species status—it’s whether they can identify each other that counts.

(1) Olson and Rasmussen (2001): Fossils show that borealis and diomedea have been separated for a very long time (c. 5 my), and hence should be treated as separate species. Of course, these assignments to taxa were qualified (as many others in the same paper) as Calonectris aff. borealis and C. aff. diomedea, meaning that they are indistinguishable from those taxa but as fossils cannot be conclusively demonstrated to actually be those species. And, there was only 1 partial humerus of aff. borealis and 3 humeri and 1 femur of aff. diomedea in the sample.

Olson, S.L. & P.C. Rasmussen. 2001.  Miocene and Pliocene birds from the Lee Creek Mine, North Carolina. In: C.E. Ray & D.J. Bohaska (eds.). Geology and Paleontology of the Lee Creek Mine, North Carolina,  III. Smithsonian Contributions to Paleobiology 90:233–365.

(2) Subsequently, Olson (2008) named a new species, Calonectris wingatei, from Bermuda. In the interim since Olson and Rasmussen, of course, further work had come out on Calonectris that suggested a lack of nuclear divergence, and Olson therefore revised his opinion on the species status of borealis (but without referencing Olson and Rasmussen (2001) for having split it). His change of opinion reflects the data from and conclusions of studies that had appeared by that time. However, in my opinion the more recent papers such as Zidat et al. (2017) should overturn the consensus in favor of species status.

Olson, S. L. 2008. A new species of shearwater of the genus Calonectris (Aves:
Procellariidae) from a middle Pleistocene deposit on Bermuda. Proceedings of the Biological Society of Washington 121:398-409.

NO. I have been working on these taxa at FLMNH from the many beached specimens we get (and a visit to the NMNH). About 125 specimens total. Using criteria to differentiate in Pyle, we have a good number (48%) of large birds, without any white on the undersides of the primaries (=borealis). We have a smaller number (15%) of small birds with lots of white on the undersides of  p9 and p10 (= diomedea). But we have a whopping 36% that are intermediate. The intermediates are dispersed throughout the range between diomedea and borealis. Howell pointed out there are a large number of intermediate specimens. Two measured out as edwardsii.  We have started doing some DNA analyses: one of the edwardsii did not amplify, the other came out as borealis. The intermediates fell both into diomedea and borealis.  Even some of the ones we thought we were sure about did not align with their phenotype. I feel that there must be more mixing than the research on the breeding grounds has found. We are working on a paper, but waiting for the full genetic analyses.

NO. I was ready to vote YES on this proposal based on what was presented, but I think it makes sense to hold off based on the evidence for intermediates. Let’s wait until that paper is published and then we can revisit this issue.

NO, but extremely reluctantly, based only on comments about intermediates. The evidence in the proposal seems to support overwhelmingly a split. If Zidat et al. found that 13 of 14 pairs were pure, then these taxa should be ranked as species under any concept of species other than a non-BSC one that treats any sign of successful interbreeding as KOD for species rank, and I don’t recall species defined that strictly in ornithology in 50+ years. So, something is amiss comparing published data to the unpublished results. I suppose it’s best to wait for good genetic sampling to sort this out. By the way, even in our series at LSUMZ, we also have several intermediates not assignable to taxon based on currently used characters.

NO (reluctantly), based on the anecdotal report of conflicting data.

NO for now (or even better, pend). The comments about intermediates has certainly raised a red flag and better to wait, but once that is published I think we should revisit the issue. If it really is as muddled as those comments suggest, something must be amiss about the multiple studies on the breeding grounds which give solid BSC reasons for a split.

NO. The Zidat et al. (2017) study shows what appear to me to be rates of introgression that are too high for these to be full biological species (1 of 14 pairs mixed; other evidence of hybridization). They are differentiated to a small degree, but they have not achieved sufficient levels of reproductive isolation (disassortative mating still too high). This, with the other comments, leads me to consider that these are good subspecies.

NO, for now. Concerns regarding a high level of intermediacy in specimens does not jive with the research on breeding birds. I am uncertain what to make of that. I am hesitant to make this change given that this may indicate there is something going on elsewhere in the breeding range of these birds. I could be talked into changing my vote, as I have gone back and forth. But I would hate to make this change and then change it back, if we can avoid that.


2018-C-5: Split Puffinus boydi from Audubon’s Shearwater P. lherminieri

YES (option 3). I realize I’m alone on this, but I’m not sure how accepted the “default” position is to place this with Audubon’s Shearwater. Storrs Olson’s theory is certainly interesting. I’ve checked Petrels night and day (Robb et al. 2008) and Howell (2012) and relate the following, possibly interesting tidbits. The former source says that Audubon’s has pink feet. Howell says legs and feet pink overall in adults, “but on some birds (probably juveniles) legs and toes can be bluish , with webs pinkish.” On Boyd’s, legs and feet are grayish-blue. Boyd’s is smaller (Robb et al. 2008, Howell 2012) and has more white around the eye (Howell 2012). As for calls, Robb et al.  (2008) were supplied with recordings of Audubon’s by William Mackin, but comparisons of the Audubons’s from burrows was difficult for Robb et al. (2008) to compare to calls of Boyd’s taken in flight. They (ibid) do make an anecdotal observation that in comparison too Boyd’s, Audubon’s has the same number of notes per phrase as Boyd’s but with shorter phrases and a faster delivery.

When one looks at the distribution of Audubon’s it is basically a Caribbean species with two described subspecies (nominate and loyemilleri, a slightly smaller subspecies). In the non-breeding season the species ranges widely through the Gulf of Mexico and in the western Atlantic north to off southern Massachusetts. Thus, they move significant distances, unlike boydi, which is as far as known is completely sedentary around the Cape Verde Islands. This along with the fact that it is well isolated from the Caribbean taxon, and along with slight plumage differences, and possible vocal differences, tips the balance for me to recognizing it as a separate species. In any event the status quo for me was considering for decades that baroli and boydi were subspecies of Little Shearwater (P. assimilis). Obviously, I don’t favor returning to that treatment and Robb et al.  (2008) offer a host of reasons why not to merge boydi with baroli as the Macaronesian Shearwater.

I’m left with the thought that we not long ago considered Newell’s (P. newelli) and Townsend’s (P. auricularis) as the same species. In any event getting more rigorous comparisons of the vocalizations of both taxa would be desirable.

Howell, S.N.G. 2012. Petrels, Albatrosses & Storm-Petrels of North America, a photographic guide.  Princeton University Press.
Robb, M, K. Mullarney, & The Sound Approach. 2008. Petrels night and day. The Sound Approach.

NO (option 1).

NO, for reasons stated in proposal.

NO. I agree with the proposal’s recommendation to keep boydi as a subspecies of lherminieri for now.

NO. I vote for option 1 (= do not split), until/unless future aDNA data suggests otherwise.

NO. Continue to treat as described in Option 1.

NO. I agree with the recommendation and support Option 1.

NO. I agree with the proposal that boydi is best kept as part of lherminieri. The description of boydi, with dark undertail coverts, a long tail, and dark face, sounds a lot like lherminieri. Olson may be correct, but it is quite speculative, and does not make much sense biogeographically.

NO (option 1). More data are needed. Paraphyletic species are predicted outcome of unequal rates of differentiation at the population level and peripatric speciation.

NO. I vote for Option 1, retaining boydi as a subspecies of lherminieri. I am not concerned about this making lherminieriparaphyletic with respect to baroli. In the absence of compelling morphological or vocal characters, I can’t vote to split boydi, and it seems like it fits better with lherminieri than with baroli.


2018-C-6: (a) Split extralimital Gracula indica from Hill Myna G. religiosa; (b) Move Gracula religiosa from the main list to Appendix 1

YES to both (a) and (b) – 5 without comment.

YES to both (a) and (b). Reasons outlined in proposal.

YES to both (a) and (b). Seems uncontroversial.

YES to both (a) and (b). I agree to both parts of this proposal.

YES to both (a) and (b). I agree with using Common Hill-Myna for the English name. Note: in eBird there is a sight report from July 2015 in Puerto Rico, so there may be a few birds hanging on, but in no way “established.”

YES to both (a) and (b). indica is a strong case for species status, despite being allopatric. YES to common names in general usage. The recent PR bird could of course be a recent escape.

Note that HBW Alive also considers Tenggara Hill Myna (G. venerata) specifically distinct; I mention this because, given its (at least formerly) wide distribution over accessible islands, it seems quite possible that this could have been the source population, more likely I would think than either enganensis or robusta (both the latter are restricted to small remote islands and though they have been in high demand for regional markets would have been less likely to be exported). If the introduced PR population was determined as belonging to the religiosa group (narrowly defined) rather than the broader religiosa group (that would include venerata), then that would not be an issue. Do we know what specific subspecies the introduced population has been identified as?


2018-C-7: Split Melozone occipitalis from White-eared Ground-Sparrow M. leucotis

YES (weakly). There are three allopatric taxa, but the clumping of nigrior with leucotis seems appropriate given the morphology and apparently the vocalizations. Given the vocal differences between the two groups, I’m marginally persuaded to go along with this well-written proposal. 

YES. Plumage, morphological, and importantly vocal differences separate the two forms. Genetic data include both mitochondrial and nuclear sequences. The proposal includes a combined analysis of these data; however, Sandoval 2017b also includes separate analyses of the mitochondrial data and a separate analysis of the nuclear genes alone. Both of these trees clearly separate the two forms into clades, with branch lengths deeper, or similar to that seen among related taxa.

YES. I agree that the phenotypic (morphometrics, plumage) differences are along the lines of species-level in the clade that includes Melozone. The vocal and genetic data also lean that way.

YES. The degree of divergence in plumage, structure, vocalizations, and genetics is consistent with species status in congeners.

NO. I am borderline on this one. There are clearly phenotypic, vocal, and genetic differences between the allopatric northern vs southern populations. The phenotypic (especially plumage) differences are notable but could also be consistent with subspecies. The genetic distances are comparable to those between other species in the genus. The male ‘solo’ song is the most distinctive, while the duets and chip calls are not all that different from the sonograms. Sandoval 2017(a) states: “Future research that involves playback to explore the reactions of the northern versus southern forms to each other’s vocalizations would help to clarify the importance of vocal differences to the animals themselves.” Given the allopatric distributions, I would like to see the results of such playback experiments to see whether the differences are sufficient to act in reproductive isolation. I am comfortable keeping them as subspecies for now.

NO. Seems right in the grey zone, however, such that either outcome would be reasonable.

NO. The differences in the vocalizations don’t seem very strong, and plumage differences also seem slight. I am not convinced that the differences would reproductively isolate these populations

NO. These look like perfectly good allopatric subspecies to me. While occipitalis is significantly different from leucotis (+nigrior), the case here seems to rest mostly on genetic distances being similar to congeners we recently elevated to full species.

NO. I’m on the fence on this one. Degree of plumage divergence seems slight relative to most allopatric taxa with species status. However, a better photograph of a specimen series might have pushed me in the other direction.

NO. The plumage differences are interesting and important but they aren’t directly relevant to species limits. Here, for example, that one of the three subspecies is more divergent than the other two in plumage in morphology is interesting and is consistent with the genetic data in reflecting an older split, but this pattern does not in itself mean that occipitalis should be treated as a species. As for the comparative genetic data, as noted under the Yellow Warbler proposal, interpretations of genetic distance using neutral loci for taxonomic purposes seems irresistibly tempting as an objective metric but is not valid, in my opinion. Genetic distance is a good indicator of relative lineage age but not necessarily of the genetic differences associated with cessation of free gene flow. The vocal data, however, look tempting if you look only at the sonograms in the Zootaxa paper, which suggest that the song of occipitalis is qualitatively different from that of the other two (Fig. 4) in addition to the quantitative data mentioned in the proposal. However, if you look at those metrics that differ significantly among subspecies (Table 4 in the paper), the ranges still overlap among the three taxa in every measure, and in one case (number of elements), occipitalis is more similar to nominate than either is to nigrior. Further, although the N for nominate is outstanding, N for nigrior and occipitalis is 3 to 9 depending on vocalization type, and thus mostly substandard in my opinion.

All in all, Sandoval has shown important divergence among the three taxa in several characters, and these are important biological data. However, they do not convince me that a change in species limits is adequately supported. What would count, in my opinion, would be some careful playback experiments to see if those suggestive differences in vocalizations are at the level of species recognition.

As for English names, given the multiple problems pointed out by Terry, especially our guidelines concerning a daughter species not retaining the parental species name, we need a well-reasoned, separate proposal on this to make sure we make the best decision if the proposal passes.


2018-C-8: Split White-collared Seedeater (Sporophila torqueola) into two species

YES. 1 without comment.

YES to split, and YES to Morelet’s Seedeater.

YES. This is straightforward.

YES. Both Mason et al. (2018) and the proposal are well presented and convincing. Given that torqueola and moreletti are not very closely related, and differ markedly in plumage and vocalizations, makes this an easy decision.

YES. The data presented in the proposal are convincing for making this split. I am fine with the proposed English names.

YES. From field experience I have long thought these subspecies were likely good biological species. The Mason et al. (2018) study provides genetic and phenotypic evidence that species-level differences are present. I am not so troubled by the mtDNA results showing that they are not each others’ closest relatives. As nuclear datasets accumulate we’re finding to no surprise that mtDNA often does not accurately track the species tree (e.g., recent Motacilla work). Yes to English names, too.

YES. This the split of the west Mexican torqueola  is straightforward and I like Terry’s suggested English names, Cinnamon-rumped and Morelet’s  Seedeaters for the English names. I am left wondering about the taxon, sharpei of northeast Mexico that looks pretty different (adult males) from birds farther south. Howell and Webb (1995) call the alternate male plumage “distinctive but variable” (see facing plate notes on plate 62). They (ibid) list sharpei. I can appreciate that genetically it clumps with morelleti but there is no spectrogram of the song of sharpei in the proposal. The subspecies is largely confined to Tamaulipas and south (Zapata to Del Rio) Texas. Northeast Mexico is a region of some endemism.  I’m just curious……

Literature cited: Howell, S.N.G., and S. Webb. 1995. A guide to the birds of Mexico and Central America.  Oxford University Press.

YES. However, repeating themes from my comments above concerning potential lapses into numerical taxonomy, the plumage differences, although important on their own, are essentially irrelevant to species limits in this group. We already knew that torqueola was the most divergent phenotypically of the taxa involved, a distinctive west Mexican endemic (thus, there is no “hidden endemism” as in the paper’s title). As for using the deep split as a criterion, how deep does the split have to be before two groups are considered separate species, or alternatively, how shallow would this split have to be before it is no longer considered “deep?” This is the problem with any continuous scale: any break is arbitrary unless it has a conceptual threshold justification. At the other extreme, Campagna et al. and Mason & Burns (2013) have shown that there is little if any genetic differentiation among the capuchino group of Sporophila. Should they all be treated as one species, therefore? Also, that Mason et al. (2018) found no evidence for gene flow between the two groups is also hardly surprising given that there is no contact zone (as stated in the paper), and is basically irrelevant to taxon rank in allotaxa. (However, what about the mismatched individual in Fig. 4?)

Regardless, the phylogeny shows that current S. torqueola consists of taxa from different branches of Sporophila, so the split is mandatory. Further, just on song differences alone, the two taxa warrant separate species treatment (although as Nick noted, the variation among song samples within each group is impressive).

YES. Given pretty significant differences in plumage and distinctive voices, along with the fact that they are not members of the same clade within Sporophila make this a straightforward split. I will vote for the English name Morelet’s Seedeater. I have not been able to confirm for certain that Sporophila moreletti was named for the Morelet that Terry found. But he traveled through the Yucatan and Peten between 1846 and 47 collecting specimens for the French Academy of Science “collecting specimens from as many bird species as possible.” The timing and location of his collecting make it seem plausible that he is who the Sprophila is named for. His specialty was malacology and he named a good number of mollusks. He also has multiple mollusks named for him, as well as a crocodile. All the other name I’ve seen use moreleti or moreletii, so not sure why Bonaparte spelled this differently.

ABSTAIN. Conflict of interest. If passed, I vote yes on the English name amendment.


2018-C-9: Lump Taiga Bean-Goose Anser fabalis and Tundra Bean-Goose A. serrirostri

YES. This split appears to have been a bad call when it was made.

YES. I do not like to undo recent decisions, but the 2007 decision to split looks hasty and base on too little data from too small an area. Given that other more-split happy authorities have not followed suit, I think we should revert back to one species until the situation is more thoroughly examined.

YES. A well-written proposal by Tom and I think his proposal makes a good deal of sense. Most on the Committee agree that in retrospect we should not have passed the proposal. This looks especially awkward because these aren’t our taxa and Old World taxonomic groups seem divided on the issue. The Dutch and I believe the Swedes do split, the Brits do not. And, they live with them, and in general are more inclined to split than we are. Perhaps the Swedes are the most balanced. I agree with Tom’s summary of many of the pertinent differences, or supposed differences, in habitat usage and vocalizations. On genetic grounds, the split, if valid, isn’t right under the present arrangement. More-over, there are many birds seen that simply can’t be assigned in the field. The famous (infamous?) bird at the south end of the Salton Sea, was small, not much bigger than A. frontalis but had a pale head and bill shape more like Taiga. It was widely reviewed and ultimately the CBRC simply accepted the bird as a Bean Goose sp. Photos of one from the Philippines, the first record, were variously labeled both Tundra and Taiga Bean-Goose. I have to admit that it didn’t look different from the Salton Sea bird.

In regards to the consideration of the Salton Sea (California) record, correspondence between Brian Daniels and Leo Ohtsuki with Masayuki Kurechi in Japan, a goose authority, was most informative, but left more questions than answers. He indicated that in eastern Honshu, moderate numbers of the classic huge middendorffii wintered. He also indicated that a smaller population of Taiga types winter on the north side of Honshu north of Tokyo (Tohoku); these birds breed farther west than classic middendorffii, as I remember west of the River Kolyma. He thought that perhaps the Salton Sea bird might be that unnamed taxon. But, upon reviewing the photos (in comparison to the small sponsa Greater White-fronted Goose), he reconsidered and said too small for those north Honshu birds, and opined that it could well be a Taiga type, but thought from an undescribed and unknown breeding population breeding even farther west of the River Kolyma in Russia.

A small group of us (Brian Daniels, Leo Ohtsuki and Larry Sansone) visited Izunuma (Sendai) area in eastern Honshu on 10-11 February 2015, where middendorffii and serrirostris wintered. We found a mixed flock of both together late one afternoon and the next morning. They separated out from the more numerous Greater White-fronted Geese (nominate albifrons I suppose) which were present by the thousands, but did not separate out from each other. I think these birds were roosting here. How and if they separate out in the fields where they graze, I do not know, but the Japanese present and at the headquarters did not mention any differences. They did indicate that the small Lesser White-fronted Goose (Anser erythropus) present in this area in very small numbers, did feed in a different and specialized habitat. In direct comparison of both taxa (and photos), the middendorfii were indeed distinctly larger and had a longer more attenuated bill. They also had a paler head giving a uniform color to the head and body, unlike the darker headed serrirostris. The Ornithological Society of Japan continues to maintain the Bean Goose complex as one species (Ornithological Society of Japan 2012), though admittedly they are a conservative taxonomic committee. As expected they list serrirostrois and fabilis as the occurring subspecies. But intriguingly they list another subspecies as a casual or accidental visitor to Honshu (Niigata, Mar. 2003; and Ogawawara Is., Mukojima I.). They name this subspecies (Lönnberg, 1923) and state that it is “rare and breeding range little known; breeds probably from Novaya Zemlya to the Taimyr Peninsula. Winter range also poorly known.”  Are these the small “Taiga” Geese?

I suppose part of my doubt about this split goes back to the Salton Sea bird, but that bird is not alone. Many, perhaps most, of the records from continental North America are debated as to which side of the Taiga and Tundra divide they fall. The same may well apply (Thede Tobish pers. comm.) in Alaska where small numbers of Bean Geese appear annually. I am reminded of Andy’s ongoing studies with Cory’s Shearwaters. The split all looks good on paper, but many (or most) of the specimens he has from Florida can’t be assigned. What gives?! The difference is that given the ambiguity, we continue with the status quo which means two species for Bean Geese, one species for Cory’s Shearwater. But for me, the lynchpin is a lack of taxonomic agreement with Old World authorities, plus comments (Tom and Pam) that our split doesn’t reflect the genetic realities. Let’s go back to the way we were, until the issue can be properly resolved. The answer, I suppose, lies in the tundra and taiga regions of northern Mother Russia, but when will those studies be undertaken probably not in the near future, or even perhaps not for decades)? Until then, we will continue to maintain a decision, that most of us regret, and agree wasn’t even done correctly.

Within Anser I am more intrigued by our own elgasi Greater White-fronted Goose which have a small confined breeding area (between Cook Inlet and Mt. Denali) in a distinctive habitat and winter in small confined area in the Sacramento Valley. There they primarily feed on tubers in the ponds rather than forage with the thousands of sponsa in the rice fields also present. They have a distinctive plumage and differ substantially on structure. And from the published information, and my own limited experience the calls differ substantially.  And thankfully, I see few, if any, ‘tweeners.’

Literature cited:

Dickinson, E. C., and J. V. Remsen, Jr. (Eds.). 2013.  The Howard & Moore Complete Checklist of the Birds of the World.  4th Edition, Vol. 1, Aves Pres

NO. There are more recent molecular phylogenies of bean geese (Ottenburghs et al. 2016, Honka et al. 2017) than Ruokonen et al. (2007), and although their results differ somewhat and no one is claiming resolution as yet, there are strong patterns. From Ruokonen et al. and Honka et al., it looks like middendorffi and fabalis may both also be better treated as separate species from serrirostris. Ottenburghs et al. found a sister relationship between Pink-footed Goose A. brachyrhynchus and Tundra Bean Goose A. serrirostris, rather than between the latter and Taiga Bean Goose A. fabalis (middendorffi evidently was not sampled).

Given that Tundra and western Taiga are well-known to segregate strongly on breeding habitat, that they have different morphologies associated with different feeding styles, that they have largely different wintering ranges and habitats, that they show strong genetic geographic structuring, AND that we currently consider them to be two separate species, I don’t think reverting to a more agnostic position is warranted. What is needed is better resolution of the whole complex, especially focusing on fabalis vs. middendorffi. If we go back to a single species of Bean Goose, then we have a further problem with Pink-footed as well, and then we might well have to switch yet again when the next study comes out showing support for four species.

I briefly checked out recordings on xeno-canto; as with geese generally there is confusing intrataxon variation, and consistent vocal differences between any of the bean goose taxa are not obvious. Thorough study of confidently identified recordings of homologous context is clearly needed.

Additional refs:

Ottenburghs, J. et al., 2016. A tree of geese: A phylogenomic perspective on the evolutionary history of true geese. MPE 101: 303-313

Honka, J. et al., 2017. Determining the subspecies composition of bean goose harvests in Finland using genetic methods. Eur J Wildl Res 63: 19.

The Ornithological Society of Japan.  2012.  Check-list of Japanese Birds, 7th revised edition. The Ornithological Society of Japan.

NO, based on other comments.

NO. I was initially going to vote Yes on this, but after reading the comments I concur.

NO. Like others, I was going to vote Yes until seeing other comments. Given this was something the committee recently split, best to make sure before we reverse the decision.

NO. I originally voted yes with the view that we should follow the majority in this case until better data are available to support two species. However, I agree that we should hold off given other comments.

NO. I was going to vote YES because the proposal had me convinced that we had made the wrong decision. However, the additional papers suggest that the split might be warranted after all. With the BOU abandoning their taxonomic committee to a group of non-systematists who clearly do not understand the importance of phylogeny and who not understand that the BSC treats free interbreeding as evidence for conspecificity, not the opposite (!), we can’t even say that we are waiting for a legitimate European committee decision.

NO. While I admit that perhaps we were premature in making this split, the papers added to the discussion suggest that it may be premature to lump them again. I am inclined to leave these birds alone until the dust settles on the taxonomy of this group.


2018-C-10: Recognize Mexican Duck Anas diazi as a species

YES. The current treatment is notably inconsistent, with fulvigularubripes, and wyvilliana being considered separate species but diazi a subspecies. I am not in favor of lumping all these together, and cannot see any real difference between the diazi case and those of these other New World drab Mallard-complex members.

YES. Thanks Tom for doing this proposal. Other committee members indicate that we should treat all the hen-plumaged mallard types the same, taxonomically, whether as species or subspecies. In fact, Tom’s proposal is framed this way. Indeed, strictly in terms of genetic differentiation at neutral loci, a symmetrical approach could be justified. This symmetrical approach has an intuitive appeal in terms of consistency, especially if we had no data on contact zones. However, we DO have contact zones, and I see no a priori reason why we should expect each contact zone to behave in the same way. Furthermore, the genetic data are largely from neutral loci and do not necessarily reflect genes that might be involved in selection against hybridization, such as those association with salt and heat tolerance or sexual dichromatism. It is clear from other committee members’ comments that we have no clear consensus on how to interpret genetic introgression with respect to neutral loci in terms of species limits. My personal view is that introgression at neutral loci is inconsequential in the speciation process and an expected outcome of almost any level of hybridization.

Fortunately, we do have contact zones in three cases: American Black, Mottled, and Mexican. What would we expect if each were conspecific with Mallard? As in all cases that I can think of in North American allotaxa, the one expected by the BSC is a hybrid swarm at the contact zone, with the hybrid zone occupied by almost nothing but mixed birds. The consistency that should be sought is not treating all three cases the same way inter se, but treating all three of them in the same way we treat other parapatric taxa. The problem with these ducks is that pair bonds are formed in the winter, when Mallard likely winters throughout much of the range of the hen-plumaged taxa, and so the “contact zone” is really the area where the two overlap during pair formation, not necessarily the breeding range as in the usual contact zone situation. Therefore, comparisons with non-duck contact zones are compromised. Because geographic overlap during pair formation is greater than during the nesting season, this difference is biased towards increased opportunities for hybridization and gene flow. Although my summaries below are no substitute for the thorough literature review for which I lack time, I think they provide a quick view of the situation of each. See also the discussion in the proposal.

American Black Duck:  As is widely known, hybridization with Mallards is frequent. This has been well-documented for decades and is usually attributed to the range expansion of Mallard into the breeding range of AMBD. Therefore, whatever the data show, they are not likely to be in any sort of equilibrium, likely distorted by deliberate introduction of Mallards, and should be interpreted with caution as a potentially artificial situation.  Even so, mating is evidently assortative. The best paper I can find is Brodsky and Weatherhead (1984), although as Tom noted, we really need something like this for northeast USA. More recently, a likely post-mating isolating mechanism has been discovered (see Kirby et al. 2004). On the other hand, the rate of hybridization seems to be influencing the degree of genetic differentiation between the two (Mank et al. 2004). Nonetheless, current data seem to support continued treatment of the two as separate species, i.e. our current NACC treatment. If non-assortative mating were widespread, then I suspect that there would be no phenotypically pure AMBD at this point, after at least a half-century of hybridization — all we would have is one big hybrid swarm. If AMBD and Mallard largely treat each other as separate species when it comes time to mating, then so should we? (P.S.: The Hybridization section of the 2000 BNA account is in bad need of updating.  It doesn’t even cite Brodsky et al.).

Mottled Duck: The situation between the two subspecies may differ; in fact, one gene tree suggest that the two subspecies may not be sisters (McCracken paper). As noted in the proposal, in the nominate subspecies of Florida, hybridization with Mallard is frequent, e.g. Williams et al. 2005. I haven’t gotten access to the paper beyond the abstract, so I hesitate to try to fit these results into a BSC framework. Clearly, hybridization is frequent, but perhaps artificially boosted by habitat alteration and introductions, and so doubly difficult to interpret. As for maculosa of TX-LA, see Peters et al. 2014. Again, plenty of hybridization and gene flow. Also, phenotype is not 100% accurate for ID of birds as hybrids; this is highly expectable but an important reminder that inspection of plumage alone under-represents hybridization. Nonetheless, the winter distribution of Mallards nearly engulfs the entire range of TX-LA Mottled Ducks, and so opportunities for hybridization are enormous; further, domestic Mallard populations are present in suburban enclaves throughout the North American range of MODU. Therefore, my interpretation of the current situation is that levels of hybridization do not match “freely interbreeding” and nonassortative mating, and that, for now, Mottled Duck deserves species rank under BSC.

Anecdotally, I’ve been looking at MODU here in south Louisiana for 40 years and have yet to see a bird that I thought was a phenotypic hybrid. For the last 3 years, I have photographed every MODU I’ve seen close enough to assess plumage features (and uploaded to Macaulay); N is small, but still no hybrid. Furthermore, MODU is actively expanding its range here NE-ward out of southwestern Louisiana into areas that have much denser populations of both feral Mallards and wintering wild Mallards. Yet these isolated individuals are phenotypically pure, including a tiny colonizing population right here in Baton Rouge area. If mating were nonassortative, then the chances of this small, expanding population being represented by pure MODU pairs would be near zero.

Further, and parallel to my comments on Mangrove Warbler, the heart Mottled Duck habitat is fundamentally different from that of Mallard. In the USA, Mottled Duck is a bird of coastal marshes, especially brackish. Adaptations to this habitat undoubtedly include tolerance of high salt concentrations in its diet, severe heat-humidity combinations, and ability to nest in a habitat that has what I would perceive to be an exceptionally high density of predators such as alligators.  To the best of my knowledge, nowhere have any of the vast populations of Holarctic Mallards evolved adaptations to this habitat. In fact, Mottled Duck is the only duck that has done this in southeastern North America. This all suggests that MODU is a highly specialized genome and that hybridization with anything else would likely lower fitness of progeny throughout most of its breeding habitat.

Mexican Duck: As outlined in Tom’s proposal, the potential artificiality of this situation is high given anthropogenic changes in the arid southwest, particularly at the margins of its distribution in the southwestern USA. Hubbard’s obscure paper is the basis for our treatment of this taxon as a subspecies of Mallard. Formerly, I strongly supported regarding Hubbard as the de facto standard that needed to be superseded before any changes should be made to species limits. I had been predisposed towards treatment of Mexican as conspecific in part because Mexican does not seem as specialized in terms of habitat (e.g. salt tolerance) as in MODU.

As noted in Schulenberg’s proposal, even Hubbard himself was conflicted on interpretation of his data. The cline in variation that he documented potentially could be geographic variation intrinsic to diazi and have nothing to do with hybridization. As summarized on the proposal, there is no current assessment that points to anything but low-levels of hybridization between this and Mallard. I am perplexed at any alternative interpretation. Pure pairs of Mexican Ducks are common in Arizona, New Mexico, and west Texas, which would not be the case if there was a complete breakdown of isolating mechanism and free gene flow. Consistent with Webster (2006), Gary Rosenberg wrote me: “Obviously, this is just anecdotal – but I see (and have seen) many pairs of “Mexican” Ducks in s.e. AZ – mainly breeding pairs along the San Pedro River and Santa Cruz River – and wintering birds at Willcox or places like Whitewater Draw – and I have not knowingly seen birds that showed anything other than pure characteristics – despite actively looking for hybrids.” My recent experience in Arizona was much the same – all I could find were phenotypically pure birds despite scrutiny of every mallard-like bird I could find. Steve Mlodinow (in litt.) and Tony Leukering (in litt.) have been scrutinizing Mexican Ducks in Colorado and elsewhere in the Southwest for decades; the find birds that are clearly hybrids, phenotypically, but the overwhelming majority are pure in terms of phenotype. The point is that if there was unrestricted gene flow between the two, then finding phenotypically pure Mexican Ducks at the northern edge of their range would be a low-probability event barring extensive immigration from pure population.  Perhaps the situation has changed since Hubbard’s data? In contrast to Hubbard’s data, the current situation suggests that hybridization is less frequent than between AMBD and Mallard or nominate Mottled and Mallard. It goes without saying that we are a long way from having definitive field studies to corroborate all this.

Overall: Indeed, there is symmetry in the empirical outcomes of all four contact “zones,” i.e. despite frequent hybridization, nearly universal presence of released and domesticated Mallards, and widespread overlap of wintering Mallards with the sedentary taxa during pair-bonding, the integrity of all four populations remains surprisingly intact. Some neutral loci show strong evidence of introgression, but in general all four populations show a surprising degree of phenotypic purity and assortative mating given the vast potential for genetic swamping. Although all this may change in the future, for now the populations of all four resident taxa show no signs of nonassortative mating. So, YES, as recommended in the proposal.

YES. Tom Schulenberg’s motion is entirely convincing for me. It looks like we made a mistake back in 1983 and adopted a position that not even Hubbard (1977) was necessarily advocating. The present situation with little evidence of introgression carries more weight for me than perceived ratios in past. Richard Webster (his on-line comments are cited in his motion) is an excellent observer, one of the best really, and he lives in Portal and conducts regular surveys in southwest New Mexico east of Douglas where there are some wetlands. His analysis and Van’s supportive comments as relayed to him by Gary Rosenberg are compelling. Mallards are largely absent from the Southwest, especially in the breeding season. Issues like releasing large number of birds by a wildlife agency, or analysis of birds near more domestic locations (e.g. town parks with lakes) is another factor. In a zoo-like situation, just about any puddle duck will interbreed.

Tom is right to suggest that if we don’t re-split Mexican Duck, we should lump all of the others in this closely related complex, but really, why lump any?  Van’s comments carefully detail why a massive lump isn’t justified. They all seem to more or less behave as BSC species as we have defined them. The fact that they don’t fit the “genetic” definition of a species doesn’t trouble me that much. There are obviously factors that work that don’t express themselves genetically but seem to separate the species themselves.

I would suggest that all give a careful read of Baldassarre (2014). He has a separate subspecies account for Mexican Duck and devotes a lot of space (pages!!) to the taxonomic status of this taxon, including a careful analysis of Hubbard (1977), the AOU’s decision to lump, and their follow-up decision not to re-split in 2011. I recall voting against re-splitting then, so I am pleased to correct that error (my opinion) now. Baldasarre (2014) concludes this discussion with “more study is needed for Mexican Ducks, especially those from the northern parts of their range, but I believe the existing evidence and the above perspective on hybridization leans strongly toward species status.” As for further studies, there are a number of morphological differences between Mallards and Mexican Ducks. Alternate males are the easiest to diagnose, of course, but in addition to the overall darker body color of Mexican Ducks, the internal markings are more rufous, and the tail is darker, as noted by many. Lauren Harter and David van der Plum have studied birds extensively along the lower Colorado River (yes, Mexican Duck types and hybrids are farther west than we thought) and have plenty of thoughts on the identification issues.

Literature cited:
Baldassarre, G. 2014. Ducks, Geese, and Swans of North America, volume one. John Hopkins University Press.

Hubbard,  J.P. 1977. The biological and taxonomic status of the Mexican Duck. New Mexico Department of Game and Fish Bulletin number 16.

NO. I don’t think any of the studies provide new data genetic or assortiative mating that would require splitting.

NO. Strongly against, as there is apparently no good published evidence for assortative mating and lots of evidence for extensive intermixing, with a clinal pattern of weak differentiation. In fact, I think this proposal introduces a strong case for lumping more mallard-complex taxa at the species level.

NO. Although the proposal makes the point that we shouldn’t treat diazi differently than other taxa in the Mallard complex, I really would want some good new data that shows that treating them as separate species is the best course, especially since we just voted on this. I don’t think any of the new material mentioned in the proposal is very strong to treat diazi as separate.

NO. I understand the rationale for not treating diazi differently from other taxa in this complex, but I don’t think that argument is sufficient by itself to warrant species status at this time. It would be great to see some definitive data that addressed this problem.

NO. This looks like a subspecies case to me, in the early stages of speciation. Earlier work with quantifications of hybrid pairs indicated substantial levels of gene flow. Hubbard’s observation (“Far more diazi-like birds were paired with similar birds (and vice versa in the case of platyrhynchos), and this is a strong indication that pairing is not random in this complex”.) brings us back to the issue of assortative mating. Because of the strong, nonlinear effects of gene flow at low levels, it cannot be treated as a simple presence-absence phenomenon. The levels of disassortative mating are what is important in these situations. And while there is evidence of selection among some Z-linked loci, this is not equivalent to species limits (or we’d quickly have to recognize population-level adaptive traits all across the genome as indicative of species, when gene flow clearly enables us to distinguish between adaptation and speciation).

NO. This was an interesting proposal and thought experiment, but nothing solid enough for me to feel good about changing status quo. In regards to the comment that introgression at “neutral” loci is not relevant to species limits: This is an important idea, but unless I missed something about one of the duck hybrid papers (which is very possible, since I didn’t read them all), I don’t think we can apply that criteria to this particular situation. I appreciate the point that there is more assortative mating than one might imagine a priori given overall levels of introgression, but how we would know what is a “neutral” locus in a hybrid zone whose RI isolating mechanisms are unknown? I think that binning loci into “neutral” or “consequential for reproductive isolation” requires more knowledge about the system (i.e., what we now have in BCCH/CACH, GWWA/BWWA). Until that point, “lots of introgression” vs. “not a lot of introgression” is all we have to work with, and some of the introgression in the ducks may well involve a lack of RI mechanisms or a breakdown of those mechanisms. In any case, figuring out what are the neutral loci versus RI loci underlying assortative mating in a bunch of brown ducks that like to mate during the winter, sometimes through forced copulations (I suspect), is going to be a formidable challenge…particularly since the main donor of genes is one of North America’s most widespread and ecologically plastic birds.

NO. I recognize that we treat Mexican Duck differently than we treat the other hen-plumaged forms in the Mallard complex. This proposal is interesting and makes a fairly persuasive case that the situation for Mexican Duck is really not different from the other taxa recognized as species. However, rather than making me think that we should recognize Mexican Duck, it makes me think that perhaps we should bite the bullet and seriously consider lumping Mottled, American Black and Hawaiian Duck into Mallard. One complication in all of this is that non-native Mallard populations are being widely introduced into areas where they were not historically present, while habitats for wetland species are being simplified and homogenized. This combination of introduction and loss of habitat distinctiveness are both playing a role in the introgression that is occurring. I am not sure exactly how to take this factors into account. There are other ducks where there are hen-plumaged populations (for example pintails) in which I wonder if they would similarly exhibit extensive hybridization if in contact. This proposal definitely points out a problem of inconsistency in these ducks.  I am not currently convinced that the answer is the one the proposal suggests.


2018-C-11: Transfer Loxigilla portoricensis and L. violacea to Melopyrrha

YES. 3 without comment.

YES. This seems like the best course of action.

YES. Straightforward, as long as Melopyrrha has priority for the clade.

YES. Seems straightforward.

YES. It is a shame that we didn’t split the Cayman Bullfinch (Melopyrrha nigra taylori) from the nominate Cuban subspecies when we considered that issue. The vocalization differences seemed striking, as I recall.

YES. Required for generic monophyly.

YES. Straightforward as based on the newest phylogenetic evidence.

YES. Required by Burns et al.’s (2014) phylogeny. The only real difference I can see between Melopyrrha and Loxigilla is that absence of chestnut in the former. Thus, congeneric treatment does not seem to violate any subjective notions on generic limits. The Supplement will have to explain clearly why Melopyrrha has priority over Pyrrhulagra, contra Burns et al. (2016) and HBW.


2018-C-12: Split Gray Nightjar Caprimulgus indicus into three species, recognizing (a) C. jotaka and (b) C. phalaena

YES to (a) and (b) – 1 without comment.

YES to (a) and (b). Reasons are given in the proposal.

YES to (a) and (b). Good evidence for this split and this change would follow other lists that cover the geographic area of these taxa.

YES to (a) and (b). We should follow our international colleagues (and the evidence) for this essentially extralimital issue.

YES to (a) and (b). Definitely needed.

YES to (a) and (b). Songs are radically different between jotakaindicus, and phalaena. The latter only has one recording on xeno-canto, but along with difference in size, it is extremely disjunct from the breeding range of jotaka.

YES to (a) and (b). I can understand wanting better evidence before granting species status to phalena, but having a highly migratory (at least nominate jotaka if one recognizes Himalaya hazarae) be the same species as a resident subspecies (breeding around mangroves, at least in part) on Palau, makes little bio geographical sense. Whether hazarae is a valid taxon is debatable. Vaurie (1965) recognized it, but Holyoak (2001) did not indicating that “numerous specimens from southern China and a few from lower River Amur appear to be intermediate, suggesting extensive clinal intergradation between jotaka and hazarae as admitted by Vaurie 1965:638-639). If there is a geographic break in the breeding range of these two subspecies in southern China, it must be small.

Literature cited:
Holyoak, D.T. 2001. Nightjars and their Allies. Oxford University Press.

Vaurie, C.  1965. The Birds of the Palearctic Fauna, Non Passeriformes. H.F.  & G. Witherby Ltd.

YES to (a). NO to (b).

YES to (a). NO to (b). (a) The published sonogram in Rasmussen & Anderton (2005) barely meets our minimal standards for published data, especially for nightjars. (b) This split is based mainly on qualitative field guide descriptions of voice, which do not meet our minimum standards for published data, and reference to online sources. I think the latter is fine for supplementary information (as in my Mangrove Warbler comments), but not as the primary basis. Although I’m sure that those who have promoted the split are correct, our policies require at least some published documentation. We’re not asking for much — a simple note with some sonograms is all.

YES to (a). NO to (b).The evidence for splitting indicus and jokata is strong. But as to b, I have a couple of issues: (1) We are not required to act, because phalaena is not in our geographic scope. (2) It looks like there is no peer-reviewed literature to back the proposed split. I agree with the likelihood that phalaena is a good species, but prefer to stick to our bailiwick and rules. If I am wrong on this (e.g., if we should consider guides and checklists to meet our peer-reviewed criteria, and if we should extend our deliberations beyond the checklist area), I’d be happy to change my vote.


2018-C-13: Split Barn Owl (Tyto alba) into three species

YES. Based on genetic and morphological differences, three species seems like the best approach at this time.

YES. I am sad to ‘lose’ one of the world’s most widely distributed terrestrial bird species.

NO. It’s great to see this proposal because we really need to evaluate all this given the new data, and IOC and HBW splits. However, this split is based on degree of genetic divergence of (presumably) neutral loci. It is homologous to bar-coding rationale. Because genetic divergence is measured on a continuous scale, there is no conceptual basis for calling a split “deep” or “shallow” or whatever. Yes, there are three lineages, not surprising for such a widespread species, but are there any consequences relevant to BSC species limits in the divergence at neutral loci? Otherwise, all we would be doing is instituting subjective thresholds on genetic differentiation of loci that by definition are assumed to be irrelevant to the biology of the birds involved. The morphological traits said to be associated with each lineage are interesting but as far as I can tell just geographic variation largely irrelevant to species limits.

I cannot think of a single modern analysis of species limits in any owl (or caprimulgid) that isn’t anchored in vocal differences. Barriers to free gene flow are exceptionally strongly correlated with vocal differences in nocturnal birds. It is widely known that plumage differences are basically irrelevant in these nocturnal birds (for obvious reasons, including that some species have strikingly different color morphs). So, how can a proposal in 2018 NOT concentrate on voice? Until vocal differences, if any, are elucidated, we shouldn’t even be considering this split in my opinion, for the reputation of our Committee.

Not one YES voter has produced a coherent rationale for this split; some have written that the in press paper (with first and last names of authors reversed, and at least two run-on sentences … typical MPE) that tipped the scales, but did not explain why. Vera et al. present from great data on relationships among Tyto, with much broader taxon-sampling than the previous paper. However, their taxonomic conclusions are weakly supported and probably would not have survived a bird journal review. It’s basically a set of gene trees, combined, and as stated clearly by the authors, the paraphyly is driven by mtDNA: “our tree topology is largely dominated by the signal of mitochondrial markers.” The authors themselves noted: “All these pitfalls have to be kept in mind for any taxonomic conclusions drawn from newly established phylogenetic hypotheses in the following, because gene-tree topologies depend on the combined effects of introgression, incomplete lineage sorting and faulty taxonomy.” Further, the paraphyly itself is the consequence of peripheral speciation: three insular taxa currently ranked as species (our T. glaucopsT. nigrobrunnea from the Sula Islands, T. rosenbergii from Sulawesi). If these insular spinoff taxa did not exist, then there would be no paraphyly. For this and other reasons, use of the monophyly criterion at the population level is a misapplication of Hennigian principles; Hennig himself did not use the term monophyly at the species level because, using an early schematic diagram of incomplete lineage sorting, he showed why species are not necessarily monophyletic.

The most interesting finding is that the distinctive T. a. punctatissima from the Galapagos is the likely sister of Hispaniolan T. glaucops (yet another Galapagos-Gr. Antilles connection), and in my opinion, requires a proposal to split punctatissima. Otherwise, all we have is three lineages corresponding to three major areas.

The authors’ rationale for the split of Barn Owl into three species is based (in addition to the paraphyly that they themselves were cautious about) is grounded in the existence of three monophyletic lineages corresponding to three major regions. This is a typical pattern that provides no basis for assigning taxon rank. Although I can’t seem to find where in the text the authors present the actual genetic distances among them (presented in an unavailable table in online supplementary material), it is clear that the reasoning is basically bar-coder rationale. For example, they regard the split of Osprey into four species as a given because they are genetically “well-differentiated species,” when the actual % sequence divergences (1.5-2.6% in Monti et al. 2015) are small; these would be at the low end of % sequence divergences among populations of Andean or Amazonian birds separated by rivers.

The plumage differences among the populations are minor, and perhaps less than those among subspecies within the three major groups. Regardless, plumages in owls are essentially meaningless in terms of species limits. Owls are notorious for color phases, geographic variation, and individual variation that are not considered relevant to species limits. I think there might be more individual variation in Glaucidium brasilianum than among these Barn Owl populations.

What about voice? A cruise through what I can find on xeno-canto indicates that the eerie hissing screech that we are all familiar with in our Barn Owl is also the primary vocalization of European birds (“Western Barn-Owl”):

England: https://www.xeno-canto.org/186611
Belgium: https://www.xeno-canto.org/235525
Netherlands: https://www.xeno-canto.org/384166

African birds (also “Western Barn-Owl”) sound similar:

Senegal: https://www.xeno-canto.org/237647

So do Indian birds (“Eastern Barn-Owl”), at least superficially:

India: https://www.xeno-canto.org/species/Tyto-alba?pg=3

And just so you know, here’s a recording from distant Argentina, which sounds to me just like those here in the USA:

Perhaps an actual formal analysis would find some consistent, diagnostic differences among these complex screeches, but the burden-of-proof in my opinion is showing the differences. Then, do the differences make a difference to the owls themselves? At the other potential extreme, maybe all Tyto sound alike. If that’s the case, then this needs to be addressed explicitly.

Finally, if this proposal somehow passes (despite not a single committee member explaining why it should), a separate proposal on English names should be required: is “American Barn-Owl” really the way we want to go for a species found throughout the Western Hemisphere? Sure, Europeans regard all that as “the Americas,” but it makes me queasy. Eastern and Western are out of our purview, but are to be lackluster in the extreme.

NO. Comments from others have convinced me that we should wait for additional supporting evidence before accepting this split.

NO. I generally think that the levels of mtDNA divergence reported among the clades (~6-9% divergence) are, in the vast majority of cases, strongly indicative of species status. This is also what I consider to be the ‘safe zone’ or ‘sweet spot’ as far as mtDNA dating estimates — enough isolation and sorting to avoid most stochastic problems that plague time estimates in the early stages of divergence, but probably not enough divergence to get repeated mutations at the same sites and underestimate divergence. This is just to say that much of the problems associated with mtDNA dating are probably non-issues here. However, I am voting no because the proposal and the two recent papers do not address phenotype (voice OR plumage). There may be phenotypic differences among the populations that, in conjunction with the genetic divergence, would make a strong case for species status. However, those differences are not discussed in the proposal or the two most recent papers. This places too much onus on the NACC members to go back to the earlier primary literature, or do our own research in museum drawers, to suss out what’s going on. This is fun, but proposals should be thorough evaluations, not just suggestions for NACC members to do their own thorough evaluations. In other words, if the author of the proposal feels a taxonomic change needs to be made, than they should summarize all of the available evidence and make a stronger case. In my opinion there is not really even enough information in this proposal to split based on PSC criteria, because no indication is given that the three clades are at all diagnosable. As far as whether voice is important for Tyto species limits – that’s an interesting question that I don’t know the answer to, and the kind of thing I’d like to see researched and addressed by proposal authors. 

NO. Someone needs to do a study on vocal variation among Tyto taxa to demonstrate whether vocalizations are even significant to species recognition in this group. It is my sense that barn owls all give some type of screech, and that unlike typical owls vocalizations may not have much to do with species limits, but I lack time at the moment to check the literature on this. However, if they ARE significant, then that is a strong argument against the split.

NO. I originally had penciled in a tentative accept to the split. Although I don’t think that vocalizations are that important for species boundaries in this genus (e,g., glaucops and pratinciola in Hispaniola), I think that more similar taxa (e.g., OW alba and NW pratincola) that are more similar in plumage and haunts, may lack reproductive isolation if they came in contact. It is up to the those who would want to split them to show otherwise, be it by vocalizations, morphology, behavior. etc.). But it should be more than just mtDNA. 

NO. The proposal demonstrates well defined clades that make biogeographic sense, but does not provide any information on characteristics that might act as reproductive isolating mechanisms in particular, voice. Following the other comments on vocal variation, I have looked at literature descriptions of Barn Owl calls from all over it Old World range. At least at a gross descriptive level, they seem to be the same as New World Barn Owls, so I think I have to vote NO on splitting the Barn Owl.

NO. The proposal is based too much on genetic distance as a species delimiter. With the addition of Vera et al. 2018 MPE, I leaned toward a yes given its inclusion of comparative work within the core Tyto, but they did not rigorously pursue phenotypic correlates to species limits. Given the time/genetic depths of the divergences among the alba group relative to core Tyto, I suspect these are good biological species, but the conclusion that they are so still rests very heavily on mtDNA genetic distances. These are not reliable as species delimiters. So, ultimately, I vote no.

NO. I find it depressing that vocal differences are often not even considered any more in the consideration of motions, and for night birds?? I can appreciate the apparent strong genetic breaks, but we are asked to believe that one such break occurs in western Pakistan. What is so dramatic about this region that separates the populations from eastern Iran (eastern Iran not mapped by König et al. 1999) from India? König et al. (1999) mention no vocal differences between the groups. I checked various sources for vocal differences found nothing helpful initially, but then checked Robb and The Sound Approach (2015) and found the text under Barn Owl to be most informative. For starters they split three eastern Atlantic island species they had field experience with: T. gracilrostris (Slender-billed Barn Owl) from the Canary Islands, T. schmitzi (Madeira Barn Owl), and T. detorta (Cape Verde Barn Owl). No rationale is given, but it might involve conservation concerns. Robb admits that they all basically sound alike. Morphological differences are slight. In fact Robb indicates that vocalizations within various Old World populations are very much alike, although there is no specific mention of Asian subspecies. Back to the motion for a moment, the Australian delicatula and related subspecies structurally differ from all other subspecies by having a very short tail, but this apparently does not apply to javanica (König et al. 1999).

Much more interesting than Robb and The Sound Approach (2015) discussion of the various Old World subspecies, was the extensive commentary about how the sounds of New World birds differed dramatically from Old World birds. Rather than try to summarize what is stated, I’ll just repeat verbatim the text:

“By contrast, there are dramatic differences between Common and American Barn Owls. American has much shorter perennial screeches, typically less than a second long, and in a wide range of recordings, I have never heard American giving anything remotely like a courtship screech. Gerrit Vyn is the author of an excellent CD on North American Owls (2006). When I sent him an example of Common Barn Owl courtship screeches, he confirmed knew nothing similar from American. This not only supports separating the two species but also the two kinds of screeches.

At the same time, American Barn Owl has a prominent flight call that is completely absent in Common Barn Owl. It was Gerrit who recorded the metallic clicking sound in CD1-06, which he calls the ‘kleak-kleak’’ call (Vyn 2006). Unpaired males use it most often (Gerrit Vyn pers comm), so it must have an important role in mate attraction. Marti et al (2005) reported that males kleak in the vicinity of the nest, soon after leaving the daytime roost, and when approaching with food deliveries. Several other Tyto have similar calls (eg, African Grass Owl T. capensis, Eastern Grass Owl T. longimembris and Australian Masked Owl T. novaehollandiae). So rather than being an American invention it seems that Common stopped using this call and replaced it with courtship screeching.”

Robb and The Sound Approach (2015) include a spectrogram of the kleak-kleak calls of a bird recorded from South Dakota. I have not personally listened to the recordings, but most of know the kleak-kleak…..” series of calls from our birds. I have not listened to the screeches and the qualitative differences between New and Old World birds. But, assuming that Robb and the Sound Approach are correct, there would be a good case for splitting Old World (Tyto alba) and New World (Tyto furcata). As for the split within the Old World, the key for me would be the vocalizations, but in any event we wouldn’t even have to address that issue and we could just split the New World birds as one species. So, hopefully the motion can be reconfigured in the near future with the two way split in mind.

A casual reading of Robb and The Sound Approach (2015) reveals promising material for splits in the future, all based on sounds. These include a split of New and Old World Long-eared Owls (Asio asio), New (Boreal Owl) and Old World (Tengmalm’s) Owls (Aegolius funerus) and South American (including the Galapagos Islands) and West Indian Short-eared Owls (Asio flammeus) from North American and Old World subspecies. The morphological differences between New and Old World Long-eared Owls are also pretty striking.

Literature cited:

König, C., F. Weick, and J.-H. Becking. 1999. Owls, A guide to the Owls of the World. Yale University Press.

Marti, C.D., A.F. Poole, and L. R. Bevier. 2005. Barn Owl (Tyto alba). The Birds of North America online (bna.birds.cornell.edu/bna/species/001).

Robb, M. & The Sound Approach. 2015. Undiscovered owls, A Sound Approach guide. The Sound Approach.

Vyn, G. 2006. Voices of North American owls. 2 CD’s and booklet. Ithaca.


2018-C-14: Split LeConte’s Thrasher (Toxostoma lecontei) into two species

YES, but borderline. Genetic differences in both mtDNA and nuclear DNA indicate separate evolutionary histories. Also, new data indicates the two forms do not come into contact. Vocalizations were not studied, but the proposal argues that these are unreliable evidence of reproductive isolation in thrashers.

NO. 1 without comment.

NO. Despite significant genetic differentiation, there is little in the way of morphological or vocal differentiation that might act as reproductive isolating mechanisms.

NO. Although the Vázquez-Miranda et al. (2017) have done an admirable job of addressing some of the questions that arose from the earlier proposal, the similarities in morphology and ecology would, I feel, not result in reproductive isolation. The genetics work showed that these taxa have been long separated (as long as some other thrasher species) and that current gene flow is not occurring, but all this can be explained by these subspecies being allopatric during this process. They may have diverged genetically and to a small degree in phenotype, but these differences have not led to reproductive isolation, as they have in other thrasher species (e.g., Gray versus Bendire’s, which look very different).

NO. Lack of any evidence for BSC-relevant criteria, and only shallow-level divergence in genetic, morphometric, and ecological traits. Probably an incipient species under the PSC but not the BSC.

NO. Although this is a borderline case I don’t think it is a strong one. The apparent absence of morphological, vocal, and ecological divergence suggests that the genetic divergence is a product of time in allopatry. It wouldn’t take much for the populations to be reunited and then gene flow would seem more than likely.

NO. The recent study by Vázquez-Miranda et al. (2017) provides new data that are consistent with prior work, but I am not convinced that these are – or would behave as – biological species if they did contact. The overlap in morphology and color, apparent lack of song discrimination (anecdotal), and lack of ecological niche differentiation argue for maintaining conspecific status despite the genetic differences.

NO. I do not see much new here since we considered this issue the last time. I recall wanting some sort of vocal analysis, particularly of potential call note differences. The loud ascending wheeep is very familiar to any who have field experience with this species. Does the call of arenicola differ? The calls of nominate lecontei and macmillanorum not surprisingly sound identical to my ear.Yes, songs are long and wander all over the place. Deducing differences would be difficult. Play-back of calls would be more informative than songs. All of the currently recognized species within Toxostoma have distinct calls, although I have precious little field experience with Bendire’s (T. bendirei). Even the two Curve-billed subspecies groups (curvirostre and palmeri) give somewhat different calls. I can appreciate the similarity in songs. For years I attracted LeConte’s Thrashers by playing Crissal tapes, not because I was curious, but LeConte’s wasn’t available on the Peterson tapes! I can understand why Crissal and LeConte’s are sister species. It worked nearly every time! But that toit-toit-toit loud series of calls of Crissal is nothing like LeConte’s. So, please let’s have some recordings, play-backs, and analysis of call notes.

NO. These look like subspecies to me. Relative genetic divergences are not reliable species limits indicators. The lack of consistent phenotypic differences also suggest a lack of divergent selection to cause speciation to go to completion. The lineage sorting through neutral genetic processes present in what seem to be largely allopatric taxa is an expected process coupled more with population sizes than with the divergent selection usually considered to drive speciation.

NO. The proposal indicates that the taxa in question are allopatric. Therefore, any discussion of absence of ongoing gene flow is hardly surprising unless there is evidence for long-distance dispersal of these sedentary taxa. That there is no evidence on ongoing gene flow is consistent with these being two allopatric populations regardless of taxonomic rank and can hardly be presented as evidence for species rank.

The papers and the proposal trivializes playback response. As most field people know, almost any bird will respond to any unusual noise in its territory. (For example, I regularly bump my iPod in the field, which immediately generates the “Albatross, Black-footed” vocalization, first in my alphabetic menu; birds often come in to see what is making the strange noise, but I make no taxonomic assessments based on those responses). True playback trials measure the degree of response on a behavioral scale against a control conspecific tape. Without that protocol, field anecdotes are nearly irrelevant.

Genetic distance in neutral loci is once again used as a metric for comparative taxonomic rank. This is conceptually and empirically flawed. Neutral loci roughly measure time-since-divergence between allopatric populations, with a dose of population size (small populations potentially diverge faster than large ones.). They are explicitly irrelevant to the loci associated with any biological process that might be associated with active cessation of gene flow. The relationship between mtDNA divergence rates is weakly correlated with species rank, and the range is dramatic, with essentially no divergence in many taxa treated as species at one end of the continuum (e.g. capuchino Sporophila mentioned above) to strong divergence between populations that are phenotypically indistinguishable (e.g. 8-11% divergence between populations of Bleda syndactylus separated only by rivers and phenotypically indistinguishable by plumage or voice; Ben Marks’ dissertation).

The point in the Vasquez-Miranda et al. paper concerning competitive exclusion is basically irrelevant. Taxa ranked as subspecies replace each other yet are not ranked as species. In fact, populations that show any degree of discrete genetic differences, taxonomically recognized or otherwise, replace each other allopatrically or parapatrically. As long as the allopatric speciation model is in operation, it can’t be any other way. In fact, the opposite would provide better evidence for species rank, e.g. two populations have diverged to the point that they have specialized on different habitats and are thus likely to be able to coexist. Therefore, my interpretation of the niche models falls on the side of “same species.”


2018-C-15: Revise generic assignments of New World “grassland” sparrows: (a) leave some species in Ammodrammus; (b) move some species to Ammospiza; (c) move bairdii and henslowi to (1) Passerculus sensu stricto, (2) Centronyx or (3) Passerculus sensu latol.

YES to (a) and (b). These are clear cut. (c) I prefer option 2 rather than lumping bairdii and henslowii with Passerculus and Melospiza in a single genus. However, I will go along with that if it is the consensus.

YES to (a) and (b). (c) Option 2 is a better way to go than what was originally proposed.

YES to (a) and (b). For (c), I vote for option 2 (Centronyx) because this option is the least disruptive. Fewer species would have to be reassigned to existing genus. In addition, the genera Melospiza and Passerculus are strongly supported; therefore, I don’t think the argument for changing their name in the course of creating a broader genus is compelling. Although the proposal is correct that there is not consistent strong support for a bairdi and henslowi clade, there is historical precedent for keeping them together as well as 0.97 PP support from mtDNA for a bairdi and henslowii clade. If we go with option 2, the worst-case scenario would be that one of these two species is found to be sister to another existing genus in the future. Given the support values, I don’t think it’s likely that either of these two species would later be found embedded into Melospiza or Passerculus. Thus, if these two species are found not to be a clade, I would guess bairdii would stay with Centronyx and henslowii would be renamed at that point anyway, not lumped into an existing genus. So option 2 would only require potentially one change in the future.

YES to (a) and (b). (c) Option 2. If there is one group of passerines where behavioral considerations are important in assignment to genus, it is our New World sparrows. Most genera have their own distinctive behavioral actions. Some are secretive, some are not, some are gregarious, others are not. I can certainly support having a more restricted Ammodramus genus and the species composition seems reasonable. The split of the “marsh sparrows” has been discussed for some time. Fine. But for part C the only option I can support is option 2. The thought of placing Melospiza  with Passerculus along with Baird’s and Henslow’s makes me cringe. Surely not! The three species of Melospiza all share important behavioral similarities. All are somewhat skulking, but readily perch up and raise their crests when “spished” at. All are somewhat solitary. All have similar seep calls (not the chip calls which are species specific) and most importantly all cock and pump their long rounded tails somewhat when flushed. I can tell a Melospiza at a glance just by how it handles its tail when it flies. The genus Passerculus is a short, notched, tailed species that is somewhat skulking when not singing, but when flushed often perches up and remains for good studies. And they are often gregarious. They do nothing with their tails and their seep calls are dissimilar from the three species of Melospiza. I have only an hour (at most) field experience with Xenospiza but they were surprisingly cooperative and fed in the open and in groups. I can certainly accept placing Baird’s into its own genus, Centronyx and for the time being place Henslow’s there too. There are a number of things they share (grassland habitats, very skulking, and solitary behavior). The songs are sure very different from one another.  Don’t know about contact notes.

YES to (a) and (b). (c) Option 2, Continuing to recognize Melospiza and Xenospiza, restricting Passerculus to Savannah Sparrow, and using Centronyx for Baird’s and Henslow’s Sparrow. I am not convinced that Baird’s and Henslow’s are sisters in which case we would have to revisit this issue. However, I have problems putting Melospiza together with this other grassland sparrows. May be a reflection of the fact I haven’t quite come to terms with the fact that Melospiza is not close to Zonotrichia and Passerella. I have to admit that the variable topologies in this part of the tree makes treating that whole group as a single genus a reasonable option.

YES to (a) and (b). Required by the phylogeny of Klicka and colleagues. (c) Option 2 (Centronyx). I strongly dispute that failure to see that PasserculusMelospiza, etc., are congeners is a result of pre-conditioning to current taxonomy. On the contrary, I see no similarities among, say, Savannah, Henslow’s, and Swamp sparrows other than being confamilial. The ecological and morphological differences among the proposed new congeners would indeed be inconsistent with generic limits in any other sparrows, by far. Furthermore, the node that unites them would be the deepest among any that I can find in Klicka et al. (2014; Fig. 1), by far, and is roughly at the same depth in the tree as the one that unites Fox and American Tree sparrows (treatment of which as congeners would likely result in tar-and-feathering of all NACC members). I would rather live with the potential paraphyly of Centronyx than gag on the thought of, say, Lincoln’s Sparrow and Baird’s Sparrow in the same genus. The depth of the split between Baird’s and Henslow’s is comparable to that between many widely recognized sparrow genera, and so a fourth option, not mentioned, would be recognizing each in a monotypic genus. Even that would be better, in my opinion, than a broad Passerculus.

YES to (a) and (b). Needed because of paraphyly. (c) Option 2, put both bairdii and henslowii in Centronyx. I don’t think we should mess with the long-standing genera of Passerculus and Melospiza. Swamp, Song, and Lincoln’s Sparrows are very different than Baird’s/Henslow’s in plumage, ecology, behavior, and especially vocalizations. Savannah Sparrow has always been an outlier to other North American sparrows. If bairdii and henslowii end up being not sisters, then we can change that without creating more instability in Passerculus and Melospiza.

Now, can we address the issue of the issue of having Five-striped and Black-throated in the same genus (Amphispiza). On structure, behavior, and vocals they are very different, and there isn’t even good genetic support for having them in the same genus. There is apparently a genus available for Five-striped, Amphispizopsis

YES to (a) and (b). (c) Option 3, lump MelospizaXenospizabairdii, and henslowii into Passerculus. This would still not be a broader genus than e.g. Peucaea, Rhynchospiza, or Arremonops.

YES to (a) and (b). This solves the paraphyly problem. While I agree that the unintended reversal of publication order in Swainson’s original pair of papers is unfortunate, I consider that AOU resolved this in 1909. (c) I agree with option 3 to have a broader Passerculus.

YES to (a) and (b). (c) Option 3 as recommended. Excellent proposal on a really challenging situation.