Proposal (635) to South American Classification Committee


Recognize newly described Cyanocorax hafferi



Effect on South American CL:  This proposal would add a recently described species to our main list.


Background:  Cohn-Haft, Santos, Fernandes and Ribas (2013) recently described a new species of jay, Cyanocorax hafferi (Campina Jay) from 8 specimens collected south of the rio Solimões and west of the rio Madeira, with all but one known locality lying within the Madeira-Purús interfluve, in the Brazilian state of Amazonas.


The new species is seemingly restricted to the interface between Amazonian savannas (= campinas) and forest-edge, particularly the islands of low-stature forest within campinas, and, as such, could be considered a habitat specialist.


As stated by Cohn-Haft et al. (2013) in their diagnosis, the new species “clearly belongs to the genus Cyanocorax on the basis of its stiff frontal crest covering the nostrils (Ridgeway 1904) and overall size and plumage similar to other members of the genus.  The new species is the only member of the genus with the following combination of characteristics: pale iris, pale-tipped tail, three blue facial marks, and pale blue breast.


It is most similar to C. heilprini, from which it differs in having a pale azure-blue versus dark purplish-blue wash to underparts and three blue facial marks (supraocular, subocular, malar) versus one (malar).  Dorsally, C. heilprini and C. hafferi are more similar, differing in the latter’s paler and duller-blue tone of the back, wings, and tail.  All other congeners with pale eyes and white-tipped tail have a white breast (or yellowish-white with no hint of blue) below the black bib.


All eight specimens were very similar to one another genetically (< 0.3% divergence), and differed from their nearest relatives, C. heilprini and C. affinis by approximately 0.7% and 1.8% sequence divergence respectively, in the mitochondrial genes ND2 and cytochrome b.


Although they were able to document a varied repertoire of vocalizations for the new species, the authors were unable to confirm any vocal differences between it and C. heilprini.  Furthermore, the authors state that C. hafferi responds with equal vigor to audio playback of C. heilprini vocalizations as they do to playback of their own vocalizations.  I assume that there has been no opportunity by anyone to perform reciprocal playback experiments with the only marginally better known and similarly range-restricted C. heilprini.


Analysis & Recommendation:  There is no question that C. hafferi represents a previously unknown taxon.  The only other jay known to occur within the region occupied by C. hafferi is C. violaceus, which differs substantially in plumage and vocalizations, and which is restricted to várzea forest along the rio Purús.  Genetic data show that the new species is only slightly divergent from its closest relative, C. heilprini, which is further indicated by the vocal data. [As an aside, I find it interesting that C. hafferi is as close genetically as it is to the geographically distant (trans-Andean) and ecologically very different C. affinis.]  I have no field experience with C. hafferi, but I have listened to audio recordings, and their vocalizations were indistinguishable to my ears from vocalizations of C. heilprini, a species that I have recorded in southern Venezuela.  This squares with the conclusions of the authors that they were unable to identify any vocal differences between the two species, as well as with the results of playback experiments.  The genetic data, taken together with the vocal data and playback results, do suggest that C. hafferi and C. heilprini are somewhat tenuously divergent.  However, the degree of differentiation in plumage pattern and color saturation is consistent with that seen between some other pairs of taxa within the genus that are considered to represent separate species.  That, combined with a pretty substantial range disjunction that is greater than just the distance between the closest seemingly appropriate enclaves of campina habitat on opposite banks of the rio Solimões/Amazon, would seem to point to two taxa that are diverging and on independent evolutionary trajectories.


In several respects, this case reminds me of that involving the recently described Cinclodes espinhacensis and whether or not it merited recognition as a species distinct from Cinclodes pabsti:  minimal genetic divergence; few (if any) diagnostic vocal characters separating the two taxa; strong response of one taxon to audio playback of the other in one-way trials; similar ecological requirements; highly disjunct distributions.  In the case of the two cinclodes, I favored treatment as subspecies.  Something about the jay situation feels a little different to me.  The extent of plumage differences between the two forms is greater, involving pattern (presence/absence of supraocular and subocular marks) and color (dorsally and ventrally), rather than just degree of color saturation (as in the two cinclodes).  Also, the nature of the range disjunction is different.  The ranges of C. hafferi and C. heilprini are separated by two rivers (the Solimões/Amazon and the Negro) that are both major biogeographic barriers, and the range of C. heilprini does not even extend to seemingly suitable campina habitats close to the Amazon.  It seems pretty clear that the two are sister taxa that descended from a common ancestor relatively recently.  It also seems pretty clear that the prospects for secondary contact are beyond poor.


In summation, I would recommend a YES vote to recognize Cyanocorax hafferi as a new species to our list.  If adopted, it should, following our conventions, occupy a position immediately following C. heilprini in the linear sequence.  The English name of “Campina Jay” chosen by the authors strikes me as entirely appropriate, highlighting, as it does, the unique and threatened habitat to which the species is restricted.


Literature Cited:


Cohn-Haft, M., M. A. Santos, Junior, A. M. Fernandes, and C. C. Ribas.  2013.  A new species of Cyanocorax jay from savannas of the central Amazon.  Pp. 306-310 in Handbook of the Birds of the World. Special Volume: New Species and Global Index (del Hoyo et al., eds.).  Lynx Edicions, Barcelona.


Ridgway, R.  1904.  The birds of North and Middle America, part 3.  Bulletin U.S. National Museum, no. 50, pt. 3.


Kevin J. Zimmer, July 2014




Comments from Pacheco: “YES.  In view of the arguments of Kevin, my opinion that this relictual taxon should be treated at the species level.”


Comments from Stiles: “A very tentative YES… this is clearly a borderline case, but given the geographical framework and the two major rivers intervening, it seems reasonable to conclude that hafferi and heilprini are on separate evolutionary trajectories and that contact is impossible in the foreseeable future.  The small genetic distance between hafferi and the very different affinis does suggest that this is a case where genetics (at least the markers used) don´t tell us much regarding speciation.”


Comments from Remsen: “NO, with great pain.  The differences between hafferi and other members of the group are, in my opinion, minor geographic variation.  The absence of any vocal differences is highly suspicious.  Despite the strong geographic isolation, the degree of genetic divergence is small and comparable to that typical between taxa ranked as subspecies (although as you know I am opposed to using genetic distance as a deal-maker or deal-breaker on taxon rank).  Anyone who voted against Capito fitzpatricki or Cinclodes espinhacensis as a species-level taxon should not vote for ranking hafferi as a species.”


Comments from Nores: “NO.  The differences between hafferi and heilprini are for me at the subspecies level.  As Van commented, despite the strong geographic isolation, the degree of genetic divergence is small. Is also noticeable, the absence of any vocal differences.”


Comments from Stotz: “NO This is a tough one.  The lack of vocal differences and limited genetic difference make it hard to vote for this one, despite the big disjunction and plumage characters.  A lot of campinas birds have some big range disjunctions accompanied by limited morphological differentiation.  What about north and south Amazonian Elaenia ruficeps, or Tachyphonus phoeniceus, Neopelma chrysocephalus, Euphonia plumbea and Cotinga cotinga in Peru.  Not sure what this means, but something about either recent dispersal or connection within a subunit of these white sand species.”


Comments from Jaramillo: “YES – Borderline of course. I am going with Kevin’s suggestion; the issue of recognition of song is a difficult one. Recognition of voices of closely related taxa is expected, by default, and the informative situation is when they do not respond to each other’s voices. The fact that they do respond to the voice of a related species is not very informative at all particularly in oscines.”


Comments from Areta: “A very painful NO. As stated by others, this is clearly a borderline case and one that makes us think on what a species is and how meaningful such distinction is to understand biodiversity. Given this, I think a much more thorough work should have been done to justify species status for hafferi. It is a difficult decision for anyone, and I am not completely clear as to what would be the best choice. However, some key elements inclined me toward a no vote.


“a) The authors mention the use, at least occasionally, of terra firme forest by hafferi, which together with the very small genetic divergence indicate that birds tend to move around more than what one would initially suspect. Given this fact, the 600km gap between hafferi and heilprini doesn't seem an insurmountable barrier to this jay (indeed, 600km is about the SW-NE length of the geographic range of hafferi).


“b) The 0.3% genetic distance within hafferi in comparison to the 0.7% distance between hafferi and heilprini seems to be easily accounted for by the present day geographic distance between them. I imagine the same genetic difference could be found between extremes in a population of jays spanning the joint and continuous geographic range of hafferi+intervening gap+heilprini.


“c) Vocal differences between some Cyanocorax jays are not easy to notice, let alone characterize. However, hafferi and heilprini seem to be among the least different in the genus, further complicating the recognition of hafferi as a valid species.


“d) The single qualitative morphological difference that caught my attention were the three (hafferi) vs. one (heilprini) facial markings. Yet, its importance as a mating or recognition cue is not clear and I am reluctant to recognize a different Cyanocorax species based on minor color differences and slight differences in facial pattern.”


Comments from Cadena: “NO, though this is a hard one. Others have noted that genetic distances are rather minor, but I actually find them to be quite remarkable given that these birds are relatively large-bodied, highly dispersive species from open areas, in which population genetic differentiation is often rather shallow. That said, although this is clearly a new taxon (phylospecies) and a spectacular finding, I am unconvinced that the evidence is sufficient to treat it as a distinct biological species, especially given the apparent lack of vocal differences. In other groups we treat plumage variation of this sort as indicative of subspecific differentiation. Kevin's point about range disjunction is well taken (I have advocated for disjunctions of this sort as relevant in other cases, e.g. Arremon, Anthocephala), but given the minor phenotypic differences, no vocal differences, and results of playback trials, I think that hafferi is best considered a subspecies of heilprini.”


Response from Mario Cohn-Haft:


Please know, everyone, that I have no personal or even strong professional reaction to this decision.  I write now, not to defend any particular conclusion (either of which -- species/subspecies -- seems reasonable to me), but rather to call attention to what seem to me to be the key arguments in the decision and to try to understand what sort of evidence would be needed to make a stronger conclusion.  The overall impression I got from reading the discussion of the case was of a severely arbitrary process (which I suspect is unavoidable), in which the criteria for decision-making remain vague or inexplicit (which IS avoidable).  I sort of wish I'd treated the question of taxonomic status directly in the description (one of the reviewers even suggested subspecies status), but I suspect it wouldn't make much difference, and it certainly was easier to finish the manuscript on time without going into that.


“So, here's my take on what I just read on the SACC site:


“Four members voted YES for species, 5 voted NO, and one member did not vote.  Everyone who expressed an opinion about the validity of the taxon agreed that it deserves a name, regardless of level. 


“I won't make much out of the order of the votes, since I'm not sure it's shown accurately on the site, except to point out that the proposal was a YES and the first votes (in the order listed on the site) agreed with the proposal; the first NO vote seemed to release a flurry of NOs, suggesting (also quite understandably) an aspect of peer pressure or self consciousness in the process. 


“Relevant criteria:


“Genetic divergence

Some members mentioned that they don't like hard genetic distance cutoffs to serve as taxonomic level criteria.  However, it seems to me fairly obvious that a larger difference (more like that found between other Cyanocorax spp. pairs) would have made a YES vote more comfortable.  One member felt the differences to be remarkably large considering other ecological aspects (!), but did not follow up on the significance of that. 


“The question I pose is "What does a relatively small genetic difference mean?"  One possible answer, and the one I expect is correct, is that the two phenotypically and geographically distinct populations (taxa) have been genetically isolated for a relatively short time.  Does that make them more or less likely to be distinct species?  That's a valid question that was not tackled directly by the members' arguments, but would seem to be crucial to the debate.  Perhaps it would be useful for SACC to have an explicit stance or policy with respect to this subject.


“I note here that, although the phylogenetic relationships between C. hafferi, heilprini, and affinis were unresolved in the tree presented, nevertheless hafferi differed in DNA sequence of the genes analyzed much less from heilprini than from affinis.  What does that mean?  What sort of situations lead to unresolved polytomies in the face of strong differences in genetic distance?  Is this biological or methodological?  What sort of new evidence could help resolve the question?  Answers to these questions might point to a clear solution to the dilemma or at least make explicit what new analyses might shed light.


“Lack of vocal distinction.

“This clearly was very important to almost everyone.  Yet one member pointed out that lack of vocal distinction seems to be widespread in the genus and reciprocal response to playback is common in birds with similar voices, regardless of species status; thus, lack of response is stronger evidence of species distinctiveness than positive response is evidence of conspecificity.


“I'd like to add a couple of observations here:


“1) absence of difference is much harder to prove than presence of difference, so stating no differences were found is not the same as stating there are none (especially in birds with big intraspecific repertoires that are hard to describe);


“2) the vocal differences between hafferi and the other close spp. besides heilprini were not carefully addressed in the original description, so it's possible that there are few if any vocal differences among ANY of these species in the larger clade;


“3) having a "yardstick" for the relevance of this trait (voice) in the genus would make the decision less arbitrary; for example, do cyanopogon and chrysops differ vocally?


“4) it's not unusual for sister or closely related species not to differ vocally; for example, apparent lack of vocal distinction among Selenidera spp. (except piperivora) or Rhegmatorhina or Gymnopithys spp. doesn't seem to bother anyone.


“I recognize that by addressing 1-3 above directly in the published description I might have made the decision easier for you.  However, point 4 to me is the clincher that lack of vocal distinction is not an important character in this discussion, in the case of jays specifically, and perhaps with most South American birds in general.  This was also stated more or less explicitly by one of the members.


"Minor" morphological differences.

“This is an important argument, I think, that was really not tackled head on.  It's entirely arbitrary to call differences minor or major, without some kind of measure, be it a comparative one (inter- vs. intraspecific yardstick) or a functional one (e.g., face patterns matter in mate choice, but overall color doesn't) or a combination of those.  Interestingly, the SACC proposal suggests that the plumage differences are equivalent to or greater than those found in other species in the genus.  However, other members imply that the differences are too small for species distinction and equivalent to subspecies differences.  But nobody is explicit about which taxa are informing these personal yardsticks.


“I suspect both the plumage differences (overall color and face pattern) between hafferi and heilprini are important.  But less subjective than that, I believe that their intrapopulational uniformity points to good gene flow within taxon, lack of geographic distance effects within, and lack of gene flow between taxa.  So, without dwelling on whether the differences are minor or major, the pattern they show jibes with the DNA story:  recently split populations without evidence of gene flow between them.  Does that make them biological species?  I sure don't know.


“Geographic distance and isolation.

“Hmm, this is a real juicy one that nobody seems to know what to make of!  The apparent long distance and presence of well-known important biogeographic barriers between the two taxa help people to consider treating them as distinct species.  Why?  That would appear to be an argument about the chances of their coming into contact any time soon, suggesting they'll continue to have time to accumulate fixed differences between them that will prevent them from interbreeding, i.e., that they'll keep speciating until such a time when they might come into contact eventually.  So, I think that's a guess about the future.


“If they were separated only by one big river and found immediately on opposite banks, as is so often the case with distinct Amazonian sister taxa, the logical extension of the argument in the paragraph above would be that these only-river separated taxa have a greater chance of eventual future contact and that their only subtle differences are likely to be tested with regard to their efficacy as reproductive isolating mechanisms in the near future.  Thus, presumably, committee members would have less confidence in the taxons' specific distinction.  On the other hand, if they're found that close to one another now and don't show signs of gene flow, then it could be argued that they've already passed the test--they could interbreed, but don't.  In other words, the current spatial pattern doesn't make predictions about the future, but rather tells the story of their past.


“Seems a little wishy-washy to have it both ways.  Does some sort of affirmation about dispersal ability make a difference in this story?  If the two were on oceanic islands separated by open sea of the same distance, what would committee members think?  How does having a permeable matrix between them (terra firme forest) and even patches of suitable habitat (campinas here and there), apparently uninhabited, affect the decision?  This would appear to be another area where a clear expression on the part of the committee would help everyone understand what the expected relationship is between pattern and process and taxonomic rank. 


“So, to wrap up, I think perhaps some more explicit arguments in the description, if they'd been made, might have helped members decide.  However, I also think that a clear expression by SACC members of what new evidence would help them decide and how each kind of evidence would influence the decision would be helpful, not only to me (or anyone else tempted to try to gather more information relevant to taxonomic status), but also to fellow committee members in an attempt to generate a consensus about what's a species and what's a subspecies.


“I'm inclined to describe the situation as one in which a widespread, western Amazonian, campina-specialist jay fairly recently split into 2 allopatric populations that show consistent plumage differences, signs of good dispersal ability, but a lack of occupation of all currently available habitat, and little or no vocal differences (in birds that don't show much vocal differentiation anyway).  I suggest that the lack of vocal difference means nothing (except lack of time for drift and lack of selective pressure for differentiation in allopatry or even strong stabilizing selection on a big repertoire with lots of communicable information content), but that the morphological and genetic differences do mean something.  But what?  Are they distinct species?  How to tell?


“If this is a timeless and solutionless issue and nobody's up for answering, then I'll be happy to throw up my hands too, but also inclined to conclude that every committee is just as arbitrary as ever and the conclusions don't really mean much.  But if there are some consistent answers, either with respect to what more info committee members want to know about the case and how this would affect the answer, or with respect to a new consensus based on further contemplation of the subject by the committee, then I'll be real eager to hear.


“I understand that everybody has tons of other stuff to do and that, even within the context of SACC activities, this is just one case in zillions.  I also don’t mean to be smug or ungrateful for all the work you do and for making it public.  On the contrary, I found myself stimulated by this to try to understand the process better and to try to stimulate you all to be a little more explicit.  Maybe it's already written down somewhere and I'm just being dense.  But if by chance anybody finds this to be a worthy nudge to discuss the big issues more, then I'll be very pleased.  Thanks!”