Proposal (541) to South American Classification Committee



Elevate Myrmeciza immaculata zeledoni to species rank


Proposal:  This proposal would result in an additional Myrmeciza being recognised based on a recent publication.


Donegan (2012) published a detailed study of available type specimens, plumages, voice, and biometrics of M. immaculata.  Species limits were not the central point or objective when work on the paper started – that instead being subspecies limits and the undescribed, vocally divergent Colombian Central Andes population.  However, data showing present species limits to be misled came up and, unfortunately, could not be ignored in a serious taxonomic review like this. We included this split in the Colombian field guide, Spanish language version (McMullan et al. 2011) after sharing the MS with co-authors, among only 3-5 considered deviations from SACC.

The two groups for which species rank is proposed here meet the Isler tests for vocal differentiation (including when subspecies within the groups are subject to cross-pairwise comparisons, not just the species groups).  That is to say, there are multiple fully diagnosable vocal differences, which in number exceed the differences observed between sympatric Thamnophilidae. 

M. zeledoni and M. immaculata are also diagnosable through a range of both male and female plumage characters, with significant but non-diagnosable differences in biometrics and distributional patterns suggesting ecological differentiation.  The plumage differences (in carpal patch extent in both sexes, the shade of brown in upperparts and underparts of females and lores feathering) greatly exceed those between some known-to-be-good antbird species occurring in the same forests of Colombia (e.g. C. parkeri/tyrannina) and other recently proposed splits accepted by this committee, such as the Xingu Scale-backed Antbird.  Ridgway (1909) made this split (and also split zeledoni from macrorhyncha) on this basis.

As summarised in the abstract: "Available data … support splitting Immaculate Antbird into two species, under any modern version of the Biological Species Concept. Western Immaculate (or Zeledon’s) Antbird M. zeledoni inhabits foothills and mountains from Costa Rica southwards, and includes M. z. macrorhyncha of southern Panama to Ecuador. (Andean) Immaculate Antbird M. immaculata occurs in the Central, East, Perijá, and Mérida Andes of Colombia and Venezuela (and includes the new subspecies). Vocal differences exceed those between parapatric Goeldi’s Antbird M. goeldii and White-shouldered Antbird M. melanoceps and those between sympatric thamnophilids in other genera."

This proposal is not novel in that Ridgway (1909) also split them (and he split zeledoni from macrorhyncha/berlepschi too).  The proposal in Donegan (2012) is a minimalist treatment in that macrorhyncha is also diagnosable vocally from zeledoni based on its more equal-length first note of male songs.  New taxon concepcion shows small subjective note shape differences from immaculata in both song and call and is near-diagnosable (over 90% by actual data but not 97.5% using t-distributions) by a quantitative vocal measure (song speed).  As a result, some of the other taxa proposed for subspecies rank by Donegan (2012) present similar situations to the borderline split of M. palliata, which was recently approved by this committee.  Both concepcion and macrorhyncha appear to be good phylogenetic species and more ardent splitters would go further than just separating [zeledoni + macrorhyncha].  This proposal, therefore, reflects a conservative approach.  PSC / evolutionary species concept advocates might recognise four species.  The two split groups recognised in this proposal are vocally and morphologically cohesive whilst mutually highly differentiated.

Because zeledoni and berlepschi are contemporaneously described, and because the latter name may resurface with generic revisions, zeledoni was selected as the name that should apply by explicit first reviser action.


I have produced a "feature" on xeno-canto drawing attention to the vocal differences, available at this link, which may be of interest and allows those interested to "see" and hear additional recordings to those in the paper:



In past proposals, I have sought balance in highlighting possible dubious contrary rationales for rejecting proposals, e.g. small sampling gaps or lack of statistical analysis of unconfusable voices.  (This has often been intended as the equivalent of the BBC attempting balanced coverage by giving airtime to a "creation scientist" following a piece on some discovery relevant to human evolution.  But some committee members have found these contrary arguments compelling.)  This one is different because I cannot think of a contrary rationale to the present proposal that has legitimacy (even based on a very liberal test).  Any such approach would be based in rejection of the Isler model for species limits in antbirds (i.e. allowing parapatric species to have greater vocal and plumage differences than sympatric related species) or a novel requirement to base decisions only on molecular data.  By way of pre-emption, there is a typo (missing word "Todd" middle of line 5 on page 10) – which, if the Zimmerius proposal is anything to go by, could present an "opportunity" to one or two committee members... but please do consider the paper as a whole.  This conclusion of "no good contrary rationale" is based on the assumption that the paper is OK – and no-one has written to me with any critique to date.  If anyone has a technical comment on the paper, I would of course be glad to discuss it, privately (tdonegan A T proaves DO T org), here or on birdforum.

Finally, an appeal to your collective good nature and the good of rational taxonomic treatments:  This paper was over 10 years in the making.  It is a long one and long-gestating attempt at a taxonomic review, with time taken due to difficult nomenclatural issues (including over types of various names) and only recently-filled sound recording gaps.  Some of the data was collated, and types of new subspecies concepcion collected, at significant personal risk in difficult parts of Colombia.  A lot of people made their recordings available.  Analysing and collating data on 8 vocal variables each for 1202 recordings, alone and without any third party support, takes a very, very long time.  It is a labour of love probably never to be repeated for this group; and will be one of my last bird papers too.


Vernacular names: Western Immaculate Antbird and Andean Immaculate Antbird should be adopted on passing of the proposal, being preferred for reasons stated in the paper.  A separate proposal on vernacular names will be raised in the event that this proposal passes, to change the former's name to Zeledon's Antbird (retaining Immaculate for a more restricted immaculata).  In the event that this proposal fails, a series of proposals for various antbirds currently ranked as species to be lumped or re-lumped for consistency would have to be presented: including just in this genus M. goeldii into M. melanoceps and both M. palliata and M. berlepschi back into M. laemosticta.


Donegan, T. M. 2012. Geographical variation in Immaculate Antbird Myrmeciza immaculata, with a new subspecies from the Central Andes of Colombia. Bull. Brit. Orn. Club 132: 3–40.


Other references are cited in the above paper.


Thomas Donegan, August 2012




Comments solicited from Mort Isler: “I should start by revealing that I was a reviewer for the BBOC on what became the Donegan 2012 paper.  The paper has a wealth of valuable information regarding the Myrmeciza immaculata group, but I will confine my remarks to the taxonomic recommendations.  After reexamining the paper, I continue to support the author’s recommendation that M. immaculata and M. zeledoni be considered distinct species based primarily on the multiple vocal characters that distinguish their loudsongs.  Although I did not attempt to replicate the analysis, I spot-checked the vocal character measurements of the author, and I found them to be altogether consistent.  I might also add for future consideration that the data supporting subspecies within each of the species is also appears well founded.


“One problem.  The published version of the paper states that vocalizations of the two species differ diagnostically  “in the note shape of single-note calls”.  This statement was not in the manuscript that I reviewed, which only stated, “there are small differences in the note shape of single-note calls”.  After reviewing spectrograms, I conclude that the single-note calls of the two species are so similar that “blind tests” by non-participants (not mentioned in the methodology) would be required to come to the determination that they differ diagnostically in note shape.  Moreover, the difference in shape is not made explicit in the paper except to state that in zeledoni “the up-down stroke is thicker and longer at peak in most recordings”, which I believe, are not descriptions of note shape but are measurable characters.  (The Xeno-canto commentary mentioned in the proposal also speaks of a more rounded note shape.) This weakness in the paper is unfortunate and should not detract from the otherwise excellent analysis and the taxonomic conclusions.”


I also mention that when I reviewed the proposal, I recommended to Thomas that he use ”Zeledon’s” and “Immaculate” as English names for the two species, at least partially because zeledoni occupies portions of the Western Andes of Colombia and Ecuador so “Andean Immaculate Antbird” for immaculata isn’t completely appropriate.  It also seems like a good idea to recognize the contributions of Zeledon.”


Comments from Thomas Donegan: “I'd like to thank Mort Isler for taking the time to look through a long paper again, for his thoughtful comments above and also in the peer review process, as well as in making various otherwise un-archived sound recordings available.  These sorts of taxonomic reviews can only proceed with collaboration; and he and others made available many sound recordings without which sample sizes would have been lower and this paper's conclusions would not have had the same statistical support.  I also agree with his comments substantively.


“Whilst I am uneasy with making aspects of the peer review process a matter of public record, the following further observations should probably be made.  In Appendix 3D, below the table on page 34 of the published paper, it states: "macrorhyncha and zeledoni: up-down stroke, thicker and longer at peak in most recordings".  In Appendix 3D on page 32 of the submitted MS, it stated: "macrorhyncha and zeledoni: up-down stroke, thicker at peak in most recordings".  In Appendix 4 (of both the published and submitted MS), it states "_NS(4)" under each cell comparing members of the zeledoni group and immaculata group.  This denotes subjective diagnosability of note shapes of single note songs based on available recordings but without statistical analyses.  These cell entries did not change from submission to publication.  The only material change is that in the species limits section of the published paper, subjective differences in note shape - previously noted only in appendices - were mentioned within a list of differentiating characters.


“Methods for analysing note shape involved writing down a description for each vocalisation analysed, alongside the mensural data, as part of the process when each vocalisation was analysed.  This gave raw data for the basis of textual note shape descriptions.  When writing up the MS, I then lined up all available sonograms of all taxa for single notes in order to verify the descriptions and differences.  This verification process was repeated at least thrice more for single note shapes: once when selecting sonograms, when I used the opportunity to double-check descriptions; at least once when writing the diagnosis section of the new subspecies (this was itself checked a few times); and a third time after having more recently commented on Chaves et al. (2010)'s approach to note shape variables, to double-check for consistency of approach with my comments and other studies.  There is no blind test involved, but there were at least 4 self-verifications after initial data collection.  None of this checking and double-checking is mentioned in the methods simply because anyone publishing a major revision should be expected to check, double-check and double-double-check their results.


“I agree that more research could be done into differences in single note calls using statistics and blind tests.  I decided against spending more time on this because there were only 2 recordings available of this sort of vocalisation for four key populations (Merida, East Andes, Central Andes and the nominate population of zeledoni).  It is doubtful that statistical tests of diagnosability or even t-tests of average difference could be "passed" with samples of this size even if some truly diagnostic variable could be identified given (i) that data were compared not between putative species at a macro-level but at subspecies and population level, and (ii) the high numerical value that t-distribution tables give with such low degrees of freedom.  I therefore took only basic frequency and length data for such calls.  There are some small apparent non-diagnosable differences in acoustic frequency between these calls evident when one looks at the raw data, but little by way of statically backed differentiation was found for these sorts of vocalisations.  Of course, small sample sizes also mean that conclusions about note shape variation are tentative and provisional.  But rather than a weakness or shortcoming in the paper, this is more a function of a non-ideal data set for this particular sort of rare vocalisation and an author focusing analysis on the types of vocalisations and variables with greater samples that are more likely to produce useful results.


“Based on samples available, the zeledoni group (for which there is a much greater sample of subspecies macrorhyncha) have a more "n", less "chevron" shaped single note, with a thicker piece of noise of more or less stable frequency at the peak of the note, compared to a narrower / more pointed / faster and less stable peak in immaculata group as a whole.  The wording in the paper and XC summary sought to capture this difference more concisely, and did so accurately.  The small differences can be seen from Figure 5 (compare Figs. 5A-C of zeledoni group with Figs. 5D-H of immaculata group).  Committee members who are interested in this issue should look at the published sonograms and can make their own mind up as to what they think of the observed differences.  Time will tell whether these observed differences can also be supported statistically or remain true when greater samples of some taxa are considered.


“As suggested by Mort Isler in his comments, I would like to really emphasise that no "hat" should be "hung" on single call note shape differences in the species-level taxonomic proposals set out in the paper.  Sample sizes for male loudsongs (and indeed other songs and calls and some other populations' single-note calls) were considerably larger than n=2.  See the data in Appendix 3 for sample sizes and in Appendix 4 for the full details of statistical tests passed for different variables.  The two proposed species groups subject to this proposal differ from one another diagnosably in several more variables than is the standard species benchmark for antbirds, even disregarding any putative differences in the note shape of single note calls.


“As for vernacular names, the preference for "Western Immaculate Antbird" (zeledoni) and "Andean Immaculate Antbird" (immaculata) is for two reasons.  First, there are probably issues with restricting "Immaculate" to a species whose range does not coincide with the region where probably most birders have seen these (western Ecuador and Costa Rica, where zeledoni in the species sense occurs).  The zeledoni group occurs to the west of immaculata.  Whilst zeledoni does itself also occur in the western cordillera and slope of the Andes and achieves similar elevations to immaculata, a split immaculata is restricted to Andean slopes mostly at 800-2000 m, which is an unusual distribution for a thamnophilid antbird.  (In Colombia, Parker's Antbird, Rufous-winged Antwren, various Dysithamnus and Unicolored Antshrike are others that spring to mind as truly Andean in distribution; this compares to many tens of lowland antbirds.)  Secondly, this suggestion is based on a personal bias against using patronyms generally where possible.


“Sorry for boring you all with the length of this comment.  As they say over here: "I'll get my coat."


Comments from Pacheco: “YES.  Bem corroborado por evidźncias vocais bem manejadas na análise.”


Comments from Zimmer: “YES.  The vocal differences are, I believe, of a number and degree consistent with a ranking of separate species.  I would echo Mort Isler’s comments regarding the wealth of information provided in Donegan (2012) supporting this split –– the paper does not suffer from any lack of detail.  I, for one, would have preferred that the author of the proposal/paper had let his detailed treatment speak for itself, without the preemptive scolding/lobbying directed at the committee in his closing remarks.”


Comments from Stiles: “YES. The vocal differences are sufficiently solid and morphological differences, though rather subtle, are also consistent with treating zeledoni as a separate species from immaculata; the split makes biogeographical sense as well.  However, I strongly endorse Isler’s suggestions regarding English names: Immaculate for immaculata and Zeledón’s for zeledoni.  Unlike the hapless Schiff, Zeledón was a pioneering Costa Rican ornithologist (and friend of Ridgway) who made important collections that formed the nucleus of the Museo Nacional de Costa Rica’s collection, which I have used on numerous occasions - and I can testify to the quality of Zeledon’s skins!”