Proposal (1000) to South American Classification Committee

 

 

Treat Thamnophilus shumbae as a separate species from T. bernardi

 

 

Note:: This is one of several situations that the IOU Working Group on Avian Checklists has asked us to review because published treatments of the species differ, and WGAC has to pick one.  BirdLife International (del Hoyo & Collar 2014) now treat these two as separate species.  I have no first-hand knowledge of this situation, so my goal here is to lay out as many facts as possible that might be of potential relevance just to get the discussion started.

 

Background: As currently defined (e.g. Dickinson & Christidis 2014), Thamnophilus (ex-Sakesphorus; see SACC 278) bernardi consists of two subspecies that show a common pattern of taxon pairs (Tumbesian + Marañon):

 

T. b. bernardi Lesson, 1844: sw. Ecuador (Manabí) to nw. Peru (S to Ancash), including crossing Andes into upper Marañon Valley in Cajamarca (Schulenberg et al. 2007)

 

T. b. shumbae (Carriker, 1934): Marañon Valley

 

Here is the terrific plate by Hilary Burn from Zimmer & Isler (2003), here somewhat desecrated by my green line to show T. bernardi broadly defined.  Two extremes of male plumage variation are shown for bernardi.  The female shown for bernardi has a solid rufous crown, but Zimmer & Isler noted that some have varying degrees of black in the crown – see also specimen photos below.

 

 

Zimmer & Isler synonymized two previously described subspecies under nominate bernardi because their examination of specimens indicated that they were just points on a cline: piurae Chapman, 1923 (type locality = “Piura”), and cajamarcae (Hellmayr, 1917; type loc = Tembladera, Cajamarca, 1200 ft.”).  The type locality for bernardi is near the north end of its distribution: Guayaquil.  Another taxon, baroni Hartert and Goodson, 1917 (TL = “Yonan River, northeast of Trujillo”), was treated as a synonym of cajamarcae by Cory & Hellmayr (1924) and Zimmer (1933), almost certainly because the type locality of cajamarcae, Tembladera, is on the Yonán River, and cajamarcae has priority (barely).  Cory & Hellmayr (1924) and Zimmer (1933) both recognized piurae and cajamarcae.  Note that shumbae was described after publication of Zimmer (1933), who had very few specimens from the Marañon region.

 

Here are a set of photographs of specimens of males and females from LSUMNS:

 

Males (no bernardi available)

 

 

Females

 

 

New information:

Schulenberg et al. (2007, Bird of Peru) noted the song of shumbae differed from bernardi in being “much faster with distinct introductory and terminal notes” but were evidently reluctant to make any taxonomic recommendation.  Kevin Zimmer was evidently the first person to notice this (fide Dan Lane), so I am looking forward to Kevin’s comments especially.

 

Del Hoyo & Collar (2016) treated shumbae as a separate species (“Maranon Antshrike”) from bernardi and other 4 subspecies.  Del Hoyo & Collar’s assessment using the Tobias et al. point scheme is as follows (courtesy M. Iliff):

 

“Maranon: https://birdsoftheworld.org/bow/historic/hbw/colant3/1.0/introduction

Hitherto treated as conspecific with T. bernardi and placed in Sakesphorus (see T. bernardi), but differs in its smaller size, notably shorter tail (effect size −3.3; score 2); buff-tinged whitish vs pale tan-buff underparts in female (2); much reduced black breast patch in male, so that malar and ear-coverts are black-speckled whitish rather than black or blackish (2); darker, less rufous dorsal coloration in male (1); and faster song, with greater pace (2) and shorter note length (2) (1). Monotypic.”

 

The vocal differences are based on Boesman (2016), who quantified substantial differences in pace and note length.  Boesman wrote:

 

We can thus conclude that shumbae has a loudsong which is delivered much faster (score 2-3) and which consists of much shorter notes (score 2-3). If we include the somewhat faster songs of bernardi when singing in duet, effect size drops somewhat. Nevertheless, difference still allows for a score of 2+2 = 4.”

 

As previously noted, I’m a big fan of Boesman’s notes as catalysts to guide more thorough research, but he pumped out 457 of these in a short time and just did not have the time to get them into the shape that would be required for peer review.  In this case, the geographic distribution of the recordings would have been of interest.  No attempt was made to employ directly the Isler-Whitney scoring system for antbird songs in terms of whether the different best fit the species or subspecies category.

 

Some sample recordings:

 

bernardi from Guayas, Ecuador, by John V. Moore: https://xeno-canto.org/258338

bernardi from Lambayeque, Peru, by Andrew Spencer: https://xeno-canto.org/45729

 

shumbae from Jaén, Cajamarca, ca. 30 km from type locality of shumbae, by Fabrice Schmitt: https://xeno-canto.org/543063

shumbae from Bagua Chica, Amazonas: https://xeno-canto.org/45701

 

The differences between the two song types as described by Schulenberg et al. and Boesman et al. are fairly obvious (to me).

 

Oswald et al. (2017) included bernardi and shumbae in their analysis of genetic divergence between Tumbesian and Marañon taxa in six passerines.  Using DNA sequencing and SNPs, they found evidence for substantial gene flow from nominate bernardi into Marañon shumbae.  Although this was the only one of their focal taxa known to be in direct contact, the degree of gene flow was similar to or greater than that detected between Tumbesian and Marañon subspecies of Mimus longicaudus and Campylorhynchus fasciatus; in contrast, gene flow detected between one taxon pair (Melanopareia elegans/M. maranonica) ranked as species was zero The other two subspecies pairs showed zero (Saltator striatipectus peruvianus/S. m. immaculatus) or negligible (Arremon a. abeillei/A. a. nigriceps) evidence of gene flow (and those cases suggest a closer look is needed at their taxonomy.)

 

 

Harvey et al.(2020), with its comprehensive taxon-sampling, included bernardi but not shumbae.  The locality of their sample (ca. 50 km SE Abra de Porculla; see voucher specimen by Dan Lane in specimen photographs) places it in that part of the Marañon drainage within the range of bernardi, in fact close to a putative contact zone.  Here is the relevant portion of the Harvey et al. tree.  The green arrow points to Thamnophilus bernardi and its sister species, a surprisingly closely related T. atrinucha from Darién.  The yellow vertical line marks the depth in the tree of nodes at the species/subspecies boundary according to current taxonomy.  Note that the branch lengths uniting bernardi and atrinucha are actually shorter than those uniting almost any other traditional subspecies.  That’s a shocker.  Boldly assuming that shumbae is more closely related to bernardi than to atrinucha, then that branch length would be even shorter.  Although I don’t think relative genetic distance data are useful in determining species vs. subspecies rank, for those that do, note that shumbae would fall at the extreme short end of branch lengths that unite subspecies.

 

­­­

 

Discussion and recommendation: All we really have to go on, in my opinion, is the difference between the songs.  I’m going to try to get an antshrike song expert to take my vote on this.  If only there were some playback trials!  Obviously, it would be best if we could get a bird’s ear view of whether the differences we can hear are sufficient to be a barrier to gene flow.

 

English names:  Del Hoyo and Collar (2016) used “Maranon Antshrike” for shumbae and retained “Collared Antshrike” for bernardi.  Because their ranges differ so much in size, and because shumbae is more or less a peripheral isolate, retaining the parental names is justified for bernardi.  Marañon is not a good choice for shumbae because Collared also occurs into the western Marañon drainage, in at least one valley: the Río Chamaya (where the LSU sample is from).  Also, we have a plethora of Marañon Somethings.  Do we really need yet another one?  I also foresee a problem with birders naturally calling their first individuals as they enter the Marañon drainage from the west “Marañon Antshrikes.”  With these negatives, I think it’s worth a separate proposal on English names.  Something like “Freckle-faced” would call attention to a conspicuous plumage difference between the two and make people actually look at the bird.  Dan Lane (pers. comm.) suggested Chinchipe Antshrike or Pallid Antshrike.  Jaen Antshrike?  Shumba Antshrike?  Apparently, there is no current place called Shumba but there is a Shumba Alto, a Shumba Baja, and a Cruz de Shumba all north of Jaén.  A local name might boost conservation efforts in the region.  Further discussion postponed until if/when the current proposal.

 

 

References (see SACC Bibliography for standard references)

 

Boesman, P.  2016.  Notes on the vocalizations of the Collared Antshrike (Sakesphorus bernardi).  HBW Alive Ornithological Note 54: In Handbook of the Birds of the World Alive.  Lynx Edicions.

 

OSWALD, J. A., I. OVERCAST, W. M. MAUCK III, M. J. ANDERSEN, AND B. T. SMITH.  2017.  Isolation with asymmetric gene flow during the nonsynchronous divergence of dry forest birds.  Molecular Ecology 26: 1386–1400.

 

 

Van Remsen, June 2024

 

 

_____________________________________________________________________________________________

 

Comments from Lane: “I'd be interested to hear what Kevin has to say, as I had heard that he and the Islers had investigated the shumbae/bernardi situation nearly 20 years ago and the distinctiveness of the two was disappointingly less obvious than they had hoped (or so I was led to believe), and so the project didn't progress further. My gut feeling is: given that both taxon groups are found on the same side of the Porculla Pass (I have localities for the two that are about 28 km apart, see map, straight-line-distance, along the Chamaya River with no obvious biogeographic barriers between them, other than perhaps a constriction of the valley walls along one stretch), and the distinctive female plumage and voice, they probably would react to one another as different species. But that's just a hunch and I've not been able either to document actual overlap or do much playback experimenting. But if Kevin can elaborate on what he and the Islers were able to ascertain of the two forms, that would help me decide how to vote on this split.”

 

 

Comments from Stiles: “A tentative NO for recognizing  T. shumbae, but if information on different vocalizations can be substantiated, this would tip the balance to a YES for me.”

 

Comments from Jaramillo: “NO – but a more substantial vocal analysis would be nice to have. It may tip things the other direction into convincing all a split is the way forward…. Maybe?

 

Comments from Jessica Oswald (voting for Remsen): “NO, based on the evidence presented herein.

 

“Oswald et al. (2017) found that the best fitting demographic model across all focal species was isolation with asymmetric migration from Tumbes into the Maranon Valley. Estimates of divergence between Thamnophilus bernardi subspecies was during the Pleistocene (which could be as recent as ~12,000 years ago). STRUCTURE recovered K=3 as the best K value (number of clusters/populations represented in their dataset) for T. bernardi. “The poorer fit of Thamnophilus [bernardi] and Campylorhynchus [fasciatus] to a two-population (Tumbes – Maranon Valley) model may be due to the shallow divergence of these taxon pairs, potentially gene flow from an unsampled population in Tumbes, or genetic drift (Falush et al. 2016).” Indeed, the geographic sampling of Oswald et al. (2017) of T. bernardi is restricted to a few sites in Peru so additional data from Ecuador would be potentially informative.

 

“Further, I’m curious if there are any plumage or song differences in individuals in Ecuador (coastal individuals or near Loja) as compared to Peruvian birds.

 

“Playback data would be fantastic from across this species range.”

 

Comments from Areta: “I vote NO to the recognition of T. shumbae. In our WGAC exercise I have followed Harvey et al. in their interpretation that the specimen sampled pertained to shumbae (and this is what their supplementary table indicates), thus it was a surprise to see that the specimen actually pertains to bernardi. This does not change the interpretations substantially, as it is clear that a study comparing shumbae, bernardi and atrinucha is needed. Oswald et al (2017) indicated that their STRUCTURE analyses found K=3 to be the best. This is not shown in the main STRUCTURE figure in the paper, but rather in the Figure S3:

 

Pajarografo Sólido:Users:javierareta:Desktop:Screen Shot 2024-07-14 at 2.10.58 PM.png

 

“Incidentally, here is where a map would have been very useful to understand how geography and genetics map together.

 

“The vocalizations sound indeed quite different, as described in Schulenberg et al (2007), and Dan´s recordings of birds only 28km apart provide a very suggestive piece of evidence. I see merit in the hypothesis of shumbae being a different species, but I cannot discard that the current subspecific status is inadequate. This is the type of case in which a focus study can help clear any doubts. It makes me uneasy to split taxa without a more clear understanding of what seems to be, superficially, a good split, and the data on asymmetric gene flow in Oswald et al. (2017) indicates that the situation is not as clear-cut as it may seem in principle.”

 

Comments from Gustavo Bravo (voting for Elisa Bonaccorso): “NO. As in the case of S. canadensis (SACC 999), when you look at the extremes of the range, there seems to be diagnosable phenotypic differences to potentially elevate population to species status. However, as you move closer to contact zones, things get messy and clinal variation associated with environmental variation seems to be the case. What is missing is solid data. In T. shumbae’s case, we at least have Jess’s genomic data, which seems to support my interpretation.”

 

Comments from Zimmer: “YES.  Of all of the many unfinished projects that time conflicts have relegated to my backburner over the years, this one is one of my bigger regrets.  I made the first audio recordings of shumbae on my first visit to the Marañon Valley back in 2002 (I think).  Finding it and recording it was actually high on my radar, because Mort Isler and I were hard at work on the antbird chapter for HBW Volume 8, and shumbae represented something of a mystery, because there were no audio recordings of it in the Islers’ extensive antbird vocal archives, and because the subspecies was described (if I’m remembering correctly) from a female specimen, so we weren’t even entirely sure about the parameters of male plumage characters.  On this initial trip, I was actually scouting out an itinerary for a fairly comprehensive Northern Peru tour, which had me spending a lot of time in the Tumbes Region to the northwest, where other recognized subspecies of S. bernardi were quite common.  Accordingly, by the time I reached the Marañon Valley, I had already seen, videotaped, and tape-recorded a number of bernardi.  While searching for shumbae, I periodically trolled with recordings of nominate bernardi I had just made in the preceding week, and was frustrated to get zero response.  Eventually, I stumbled across a spontaneously singing/calling male shumbae and recorded it.  With the bird in sight, I tried playback of nominate bernardi songs and male-female duets (both of which I had broadcast all over the area off-and-on over the preceding couple of hours), neither of which elicited any detectable response from the male shumbae.  After giving it a rest for a couple of minutes, during which the bird continued to forage quietly, I then played back the bird’s own spontaneously given vocalizations that I had recorded several minutes earlier.  This elicited an immediate response, as the male approached me, raising and lowering its crest, and resuming calling, before starting to sing.  I immediately began taping the song, and before long, a female appeared and joined the male in singing and calling.  I was blown away by how much faster the song was compared to all of the nominate bernardi that I had been hearing and recording for days.  Over the next few days, and again, when I returned in 2003, I made efforts to locate and tape record as many shumbae as possible, and each time I found a new territory, I made a point of performing similar playback trials, presenting the birds first with playback of songs, calls and duets of nominate bernardi, followed by a break, and then presenting them with songs, calls & duets of shumbae.  In each instance, there was a complete absence of response to playback of bernardi vocalizations, followed by a strong and immediate response to playback of shumbae vocalizations.  In the process of my investigations, one of the things I was curious about, was establishing whether or not there was a contact zone between bernardi and shumbae, so I worked my way from the valley floor back upslope toward Abra Porculla, surveying for antshrikes of either taxon as I went.  I ended up finding several more pairs of each, always assortatively mated, no sign of intermediate phenotypes (in voice or plumage), and no actual contact zone.  I conducted reciprocal playback experiments with all of the bernardi that I found, presenting them with songs & duets of shumbae, followed by a break, and then, songs and duets of bernardi.  The results were similar – no bernardi ever responded to songs/duets of shumbae, but did respond strongly to songs of other bernardi.  I did find territories of bernardi and shumbae that were separated from one another by as little as 30-35 road km, with no obvious barrier to dispersal (e.g. large river, steep canyon, major habitat-shift other than gradually increasing aridity the lower downslope you went toward the Marañon Valley floor).  It seemed to me that the two taxa were essentially parapatrically distributed in this area, and that they were behaving as would be expected from two distinct species under the BSC.

 

“Needless to say, I was pretty pumped about this, and so was Mort Isler when I returned home from the first trip and told him about it.  We immediately got to work on it, and Mort began a thorough vocal analysis comparing of all of my shumbae recordings, with what was already a pretty extensive set of archived bernardi recordings from throughout its range.

 

“Somewhere during this process (I can’t remember if it was after my first trip or after my second, the following year.), Dan Lane returned from a Peru trip, wherein he had independently found and recorded shumbae for the first time.  Dan also recognized the distinctiveness of shumbae, and contacted Mort, who told him we were already actively working on it.  My recollection is that Dan contributed additional recordings and data, and, at Mort’s request, made detailed descriptions of specimens of shumbae from the LSUMNS collection, and it was our intention to add Dan as a coauthor on the paper.

 

“For the vocal analysis, Mort divided the overall sample of S. bernardi (sensu lato) recordings into 5 clusters, 4 of which corresponded to the named subspecies at that time (bernardi, piurae, cajamarcae, and shumbae) along with a 5th cluster of recordings that fell midway between the type localities of piurae and cajamarcae, and which Mort labeled as piurae/cajamarcae for convenience, with no taxonomic implications.  This provided a basis for examining the possibility of clinality in vocalizations.  Mort analyzed recordings of 105 individuals from 32 different localities.  Although all vocalizations were examined spectrographically (e.g. for note shape differences), only loudsongs were measured.  Whenever possible, Mort measured 3 songs per individual.  He obtained 63 measures and ratios reflecting 12 vocal characters, including the number of notes; duration, pace (notes per second), change in pace, note shape, change in note shape, note length, change in note length, interval (space) length, change in interval length, frequencies of note peaks, and change of note peak frequencies.

 

“Recording measurements were analyzed in three steps.  First, loudsong measurements for males and females were compared within each population.  Second, data for both sexes were aggregated by population, and characteristics of loudsongs of the five populations were compared.  Third, data for shumbae were compared to an aggregate of the remaining populations.  Spectrographic examination allowed us to identify 6 basic types of vocalizations:  loudsongs, softsongs, bark, bark-roll, growl, and caw (caws were highly variable in length and pitch).  Except for loudsongs, no diagnostic differences could be distinguished between populations for any of these types of vocalizations, although it should be noted that the inventory of calls was very small for some populations.  The remainder of the analysis, and all measurements, was confined to loudsongs.

 

‘The first step, a comparison of male & female loudsongs within each population, found no diagnostic sexual differences, although our sample contained no recordings of female loudsongs for cajamarcae, and only 2 for nominate bernardi.  Consequently, male and female recordings were combined for the remainder of the analysis.

 

“In the second step, the combined date for both sexes were compared for the five populations.  Diagnostic differences were found between shumbae and the remaining 4 populations in pace and change of pace (degree of acceleration), with the acceleration in shumbae songs being most pronounced in the second half of the song.  There were no diagnostic vocal characters distinguishing any of the other 4 populations from one another.

 

“In the final step, data from shumbae were compared to the combined values for all of the (aggregated) remaining 4 populations.  Again, loudsong differences between shumbae and the aggregated populations were found to differ diagnostically in both pace and acceleration of pace.”

 

“Because no vocal characters were found to distinguish loudsongs of populations described as nominate bernardi, piurae, and cajamarcae, and because our examination of specimens revealed that alleged plumage distinctions between the named subspecies were inconsistent, and showed evidence of clinality, we ended up merging cajamarcae and piurae with nominate bernardi, in our HBW chapter (2003), wherein we recognized only two subspecies, bernardi and shumbae.  On the other hand, our vocal analysis revealed that loudsongs of shumbae were distinguished from the remaining populations by two independent vocal characters.  Although fewer in number than typically found in syntopic pairs of closely related thamnophilid species (Isler et al 1998), the two independent diagnostic vocal characters, combined with the finding of parapatry, strongly indicated to us, that shumbae and bernardi should be considered specifically distinct.

 

“We then asked Robb Brumfield to analyze tissue samples of bernardi and shumbae (I don’t remember if this was a bird that Dan collected – he could probably fill that part in.), to see if molecular distance coincided with the vocal differences.  In piecing together old email threads, I see one from Robb to Mort, dated 6/24/2004 that said “The two Sakesphorus bernardi bernardi and S. b. shumbae are divergent by only 0.6%.  That’s pretty minimal for species.  That was followed 2 days later, by an email from Mort to me, in which Mort relayed to me the following:  “I talked to Robb after I spoke with you yesterday.  He is quite open to the possibility that molecular distances found under current procedures might not correspond to species level differences under the BSC, and he agrees that the paper should go ahead.” 

 

“I remember that we were a little discouraged that the molecular data were not strongly supportive of our taxonomic conclusions, but also thinking that perhaps the divergence was just really recent in evolutionary terms, and the fact that the two forms seemed to be existing in parapatry, with no phenotypic signs of intermediacy in plumage or vocalizations, meant that they were behaving as distinct biological species.  And, while it was true that the vocal analysis only revealed 2 independent, diagnostic vocal characters (as opposed to the 3-character yardstick of Isler et al, 1998), the ranges of those two characters were not only non-overlapping between shumbae and bernardi, but mean values were not even close statistically.

 

“Somehow, although it was our intent to proceed with the paper and make the case for separate species status for shumbae, we lost momentum.  Mort was concurrently involved in a big paper that was nearing completion, and I got consumed with work related stuff.  I think I was also hoping to make another trip back to northern Peru to try to further investigate the contact zone before we submitted the paper, and when that trip didn’t come off, we completely lost all momentum, and we were both absorbed in a series of other projects that always seemed to take precedence.  Now, so many of the details seem like ancient history, and it would take some effort to recreate everything.

 

“It should be kept in mind that although we did not identify any diagnostic differences in calls between shumbae and the other four populations of bernardi, that sample sizes of calls were limited, and inspection of calls was limited to visual inspection of spectrograms, without any measurements being made.  There are also other groups of closely related thamnophilid antbirds which are known to differ in their loudsongs, but which do not differ appreciably in their calls.  Although we identified only two diagnosable vocal characters, the extent of the differences in those two characters were striking, and the potential significance of them as isolating mechanisms is reinforced by my series of reciprocal playback trials.  Also, I think it bears pointing out that our vocal analysis did not find any diagnostic differences in vocal characters over the much broader range of bernardi, and then, over the space of < 30 km, with no intervening geographic barrier, bernardi is replaced by a taxon that differs in morphometrics, plumage, and at least 2 vocal characters.  Regardless of what the genetics are suggesting, to me, that’s a pretty strong case for treating these two taxa as separate species.”

 

Comments from Remsen: “I switch my vote to YES because Kevin’s comments above actually fill in the blanks perfectly in the information gaps I identified in the proposal.  I am especially impressed by the first-hand account of the playback trials, and I am highly influenced by the documented abrupt turnover from bernardi to shumbae.  Parapatry without evidence of gene flow is THE strongest evidence one can get for species rank in non-sympatric taxa.  As for genetic distances … well, as you can anticipate, I don’t put much stock in them – they measure time-since-separation divergence at neutral loci, which is correlated broadly with speciation, but the correlation is so weak that I never think it should be use as primary evidence except at its extremes.  If one could measure just the “genetic distance” between the gene complexes that control song in this case, I suspect that the degree of divergence at those loci would be substantial, comparatively speaking.  These taxa may not have been separated for very long but obviously long enough for songs to diverge to the point of not recognizing each other as conspecifics.

 

“The only reason I can see now, in my opinion, for rejecting the proposal is that the critical information is not published in a journal.  I can see that side of the argument, also.”

 

Additional comments from Bravo: “I read Kevin’s vote – some bits of which I already knew – and I think it is very telling. However, I stick to my NO vote. 

 

“It is still necessary to see all those data in published form and explicitly test for clinal variation. I spent the Summer of 2017 following shumbae around in the core of its range and the Summer of 2019 following bernardi in Tumbes. As I said, those two extremes are decisively different vocally (loudsongs), morphometrically, and in plumage. Still, I would like to see how those traits vary across the entire range of the species, particularly given the low genetic divergence and the strong signal of admixed individuals presented by Oswald et al. (2017). What I think is missing from Kevin’s described vocal analyses is an explicit test of gradual variation across space.”

 

Comments from Lane: “YES. Wow, well Kevin really delivered there! I should add my own addendum to his tale. He and I overlapped at the start of a tour he ran in 2002, in which Kevin kindly allowed me to join him and his group on a day trip to Lomas de Lachay. During that day out, he told me about his discovery of the different voices of the two T. bernardi, and that he had found coastal type birds east of Abra Porculla, which was what inspired me to try to collect some when I returned to the area in 2003, and I successfully got one female (which, apparently, is the one that was used in the Harvey et al tree, but was erroneously identified as shumbae). I had collected a series of T. b. shumbae in Sept 2002, with recordings, and those tissues should be available at LSU.

 

“Anyway, given the work Kevin has done in playback experiments, the fact that we both were able to document the two taxa along the same river valley within ~35 road km of one another without evidence of interbreeding, and the fact that the coastal (nominate) group of T. bernardi shows no clinal or other shift in song characters over a huge region, I think the evidence is strong, if unpublished, in support of species status for T. shumbae. YES.”

 

Additional comments from Areta: “NO. The genetic data is suggesting that something complex can be happening here, and like Gustavo, I would like to see a proper paper analyzing this. If there is a mismatch between genomics and phenotype, then it also should be explained. I am impressed by the amount of work that Kevin has done, and taking a formal look at the vocal and plumage data in association to genetics should clarify what´s happening here. Sometime ago I tried to lure Peruvian students into this complex, but apparently failed to create enough enthusiasm. This proposal is now making the case very appealing for further study.”

 

Additional comments from Robbins: “Reading all those great comments/data by Kevin leaves little doubt that shumbae should be considered a species. So, I change my vote to YES.”

 

Comments from Claramunt: “NO. Kevin’s observations are very suggestive of reproductive isolation, but the genetic data suggest free gene flow between the two taxa. Evidently there is a lot of geographic variation in bernardi (plumage and voice) that needs to be analyzed and in general, we need more published information to make an informed decision.

 

Additional comments from Stiles: “YES. Although the data are clearly not  all there yet, I think that the detailed field experience of Kevin, especially including the playback experiments, may have tipped the balance toward considering shumbae a distinct species – so, a tentative YES here.”