Proposal (759) to South American Classification Committee


Treat Pyriglena (Thamnophilidae) as consisting of five species


Effect on SACC: This proposal would add two species to the list.


Background and Analysis: As currently recognized by SACC, Pyriglena includes three species: the White-backed Fire-eye (P. leuconota), the Fringe-backed Fire-eye (P. atra), and the White-shouldered Fire-eye (P. leucoptera). Ten subspecies of P. leuconota were recognized by Peters (1951). A recent dissertation explored their evolutionary history, and this analysis found that the twelve currently described Pyriglena taxa fell into five well-supported clades (Maldonado-Coelho 2010). Using this phylogeny as a framework, Isler and Maldonado-Coelho (2017) examined 1430 vocal recordings from 186 localities, mapped published locations to review biogeography, and reexamined morphology. Plumage differences among the three species and among the ten subspecies of leuconota described and illustrated by Zimmer & Isler (2003) were confirmed as were recently-published (Isler et al 2013) morphometrics. Regarding vocal comparisons, diagnostic differences in song were confined to differences in pace (notes/sec), but numerous differences were found among populations in their four principal types of calls (see Table 2 in Isler and Maldonado-Coelho 2017).


Recommendation: Diagnosable differences in vocalizations and plumage, supported by a molecular based phylogeny, meet the Isler et al. 1998 yardstick for elevation of maura and similis to species rank and confirm species status for atra. Subspecies included and recommendations for English names follow:


Pyriglena maura Western Fire-eye

         P. m. pacifica Chapman, 1923

         P. m. picea Cabanis, 1847

         P. m. marcapatensis Stolzmann and Domaniewski, 1918

         P. m. hellmayri Stolzmann and Domaniewski, 1918

         P. m. maura (Mntris, 1835)


Pyriglena similis Zimmer, 1931 Tapajos Fire-eye

Pyriglena leuconota East Amazonian Fire-eye


         P. l. interposita Pinto, 1947

         P. l. leuconota (von Spix, 1824)

         P. l. pernambucensis Zimmer, 1931

Pyriglena atra (Swainson, 1825) Fringe-backed Fire-eye


Pyriglena leucoptera (Vieillot, 1818) White-shouldered Fire-eye


Note: P. l. castanoptera is synonymized with picea following recommendations of the paper.



Isler, M. L., Bravo, G. A. & Brumfield, R. T. 2013. Taxonomic revision of Myrmeciza (Aves: Passeriformes: Thamnophilidae) into 12 genera based on phylogenetic, morphological, behavioral, and ecological data. Zootaxa, 3717, 469497.

Isler, M. L., P. R. Isler, and B. M. Whitney. 1998. Use of vocalizations to establish species limits in antbirds (Passeriformes; Thamnophilidae). Auk 115:577590.

Isler, M. L., and M. Maldonado-Coelho. 2017. Calls distinguish species of Antbirds (Aves: Passeriformes: Thamnophilidae) in the genus Pyriglena. Zootaxa 4291 (2): 275294.

Maldonado-Coelho, M. 2010. Evolution and Biogeography of South American Fire-eye Antbirds (Genus Pyriglena): Insights from Molecules and Songs. Ph.D. dissertation, University of Missouri-St. Louis, St. Louis, MO, USA, 196 pp.

Peters, J. L. 1951. Check-list of birds of the world, vol. 7. Museum of Comparative Zology, Cambridge, Massachusetts. 318 pp.

Zimmer, K. J., and M. L. Isler. 2003. Family Thamnophilidae (typical antbirds). Pages 448681 in Handbook of the Birds of the World. Volume 8: Broadbills to Tapaculos (J. del Hoyo, A. Elliot, and D. A. Christie, Editors). Lynx Edicions, Barcelona, Spain.


Mort Isler and Marcos Maldonado-Coelho, October 2017





Comments from Areta: "NO (for now). I have spent many hours moving between the tables, spectrograms and texts of the paper and still feel that the variation therein described does not fit well with this species scheme. The presentation of data is somewhat problematic, as one has to examine spectrograms of some (not all) of the populations being compared, and quite often the vocal similarities do not match the phylogenetic relationships. Similarly, it is not clear when the authors have performed critical diagnosability tests and when diagnosability was assessed by other means, and often the variation is described in little detail and without showing it. The amount of data is impressive, and dealing with so many data points must be exceedingly difficult, but this may also require the usage of other analytical tools to reach sound conclusions. Although current species limits are certainly defective, I am not ready to embrace a five-species treatment as here proposed.

         "Several differences, but also many similarities, are apparent in songs AND calls among presumably different species, and when taken together a fairly unclear pattern emerges. For example, the songs in putative subspecies of "P. leuconota" (interposita, leuconota and pernambucensis) are diagnosably different from each other, while those of atra and leucoptera are indistinguishable, and presumably songs of maura in Brazil and of maura in Bolivia are different or at least inconsistent. The medium calls of "P. maura" are shared with "P. similis", while those from "P. atra" and "P. leucoptera" had 'many variants' and some even may appear to be not safely diagnosed. Also surprising is the sharing of rattle calls between many "species", which apparently do not differ in note shape but rather in their frequency of appearance and perhaps in their pace: Type "E" is shared between pernambucensis, interposita and castanoptera (=picea), Type "C" is shared between members of three clades proposed to be species (similis, pernambucensis and leucoptera), and "Type B" is shared between similis and the "Eye-browed" clade, finally Type "A" is shared between the phylogenetically distant Atlantic Forest clade and the Andean clade. Perhaps more noticeable, is the lack of sharing of rattle calls within "P. maura": the Andean clade possesses a Type "A" note, while the Eye-browed clade possess a Type "B" note. The differences between the long calls are not diagnostic throughout, and indeed again maura is problematic: long calls of maura from Brazil were more similar to those of the East Amazonian clade instead of to the Andean clade with which it would be part of the same "P. maura" species (but no evidence is presented on this). The short calls are the only ones that may appear to be diagnostic based on shape and the ones that would be fully consistent with species limits proposed. However, I am confronted with the problem of having to decide on why would these short calls be more important than other ones for the establishment of species limits (assuming all notes being compared in each category are indeed homologous).

         "I will thus vote NO, until someone else provides an alternative analysis of this paper that may make me change my vote. I assume that a different suite of analyses may show clearer differences that may be consistent with species limits proposed by Mort and Marcos, but at present I am troubled by the sharing of many different calls, notes and the patterns of song sharing between different putative species."


Comments from Stiles: "NO, at least for now, based upon the comments by Areta: Hopefully, the authors of this proposal will respond in detail to help clarify this situation!"


Comments to Areta comments from Isler and Maldonado-Coehlo: We appreciate the time and care spent by Juan Areta in evaluating the taxonomic recommendations of the Pyriglena paper. Our understanding is that Juan's negative response to the taxonomic recommendations results from his perception of a "fairly unclear pattern" of what are considered significantly differentiated vocal characters among recommended species. The examples he uses describe this lack of consistency in the use and characteristics of each type of vocalization in our recommendations. We have no disagreement with his perception of inconsistency (see Table 2 in the paper). We believe, however, that inconsistency of vocal characters is not relevant to the recommendations. To explain our view, we will try in a few sentences to provide our rationale. We have not included citations supporting each sentence (some are in the paper), but we can make these available. Also, we do not discuss the specifics of his comments. We can extend these comments to specifics (e.g., emphasis on calls rather than song, the parapatry of atra and leucoptera) as the committee desires.


The Pyriglena study populations are currently named species or subspecies. We recommend maintenance of a subspecies if we found it is diagnosable by a single morphological or vocal character. In the case of Pyriglena subspecies, their maintenance as independently evolving populations (where recommended) is supported by a molecular based phylogeny. The question is not whether diagnosable allopatric populations are evolving independently but whether they have differentiated sufficiently that they meet the model of the CBSC to be considered species. To this end, we concentrate on vocalizations. Suboscine vocalizations are not learned. An earlier paper found that presence of three diagnostic vocal characters provides a "yardstick" for making species recommendations based on vocalizations of sympatric sister species. But there is no available basis for concluding that any of these vocal distinctions function in mating. The yardstick is simply a measure of the extent vocal differentiation that has been found in sympatric thamnophilid taxa that meet the CBSC model as species. It is possible that one or more of them might be employed as signals in pairing of individuals, but it is also possible that none of them are so employed. Inconsistency in the evolution of the various types of vocalizations (or conversely in the maintenance of ancestral vocal characters) across widely separated populations that diverged hundreds of thousands of years ago is to be expected. Such inconsistency is irrelevant to the recommendations in the Pyriglena paper. Juan suggests that additional analyses would help clarify our conclusions. We refrain in doing so, as other analytical methods (i.e. multivariate statistics) will not tell us anything more about reproductive isolation of the allopatric forms.


Response from Areta: "Thanks Mort and Marcos for taking the time to respond to my points. The main problem I perceive is how the vocalizations are sorted out to subspecies or populations. It seems that the methods used to characterize vocalizations lead to vocalizations being (spuriously?) shared between taxa. This is the inconsistency I refer to, and the one that needs to be fully addressed (I think) before the species limits can be set with more confidence. May the a prior delimitation of vocal types be in part responsible for this pattern? To me, consistency in this regard is important and as you point out, this is possibly our main disagreement. What needs to be addressed is not only whether subspecies differ in some characters, but also whether these subspecies share many other vocal characters! Without a better understanding of which signals are used in mating, I am worried that populations sharing many vocalizations are placed in different species because they do differ in some other vocal aspect, and that populations with differences in some vocalizations (e.g. maura) are placed in the same species because they share other vocalizations. The paper puts emphasis in calls over songs, and that's fine with me. However, I also find inconsistencies in the sorting of calls among subspecies."