Proposal (181) to South American Classification Committee

Treat Pyrrhura griseipectus and P. pfrimeri as distinct species from Pyrrhura leucotis

The systematics of the Pyrrhura leucotis-picta complex is one of the major taxonomic troublesome matters within Neotropical psittacines. Recent work (Joseph 2000, 2002, Olmos et al. 1998) has suggested that many taxa formerly considered subspecies are best treated as species, thus reverting to former arrangements, and dramatically increasing the number of recognized species. Of seven taxa formerly lumped within Pyrrhura leucotis, five are now accorded specific status.

The three endemic Brazilian taxa are: White-eared Parakeet Pyrrhura leucotis, a bird of lowland and foothill forests (below 500 m) east of the Brazilian coastal range from southern Bahia (c.14°S), to Sepetiba, just south of the city of Rio de Janeiro; Pfrimer's Parakeet P. pfrimeri, an endemic of dry, deciduous forests on a narrow band of limestone-derived soils west of the Serra Geral massif in Goiás and southern Tocantins (Olmos et al. 1998), and Gray-breasted Parakeet P. griseipectus. All three possess distinctive morphology, habitats and ecology, wholly allopatric ranges and complete lack of intermediates or contact zones, and can be considered full species under any concept available (Joseph 2000, Olmos et al. 1998).

Pyrrhura griseipectus, P. pfrimeri, and P. leucotis have completely allopatric distributions, being effectively isolated from each other by thousands of km. Pyrrhura pfrimeri is both the most disjunct and occupies the most distinctive habitat compared to all other taxa in the group, being the only species restricted to dry forests and to occur outside the Atlantic Forest. P. pfrimeri and P. leucotis are similar in size, with no significant differences in wing length, culmen length and mandible depth, but differ in coloration, the most striking character in pfrimeri being the complete absence of the white to pale buff auricular patch of leucotis and griseipectus, and the contrasting dark red face and pale blue forehead, crown, occiput and nape, this color spreading to the neck- and throat-sides and grading to green on the breast. In contrast, leucotis has blue only on the forehead (sometimes also just above the eyes), the crown and nape being buff or grey. P. pfrimeri is thus one of the most distinctive taxa in the picta-leucotis species group (Joseph 2000).

Although both Atlantic Forest taxa might appear to be closely related, based on their geographical proximity, P. griseipectus seems to be more morphologically divergent from P. leucotis than P. pfrimeri. Both are distinguished by differences in the periophthalmic ring (dark blue in leucotis and pfrimeri, whitish to slaty in griseipectus), auricular patch (cream to yellowish in leucotis, often with a buff tinge; pure white or cream in griseipectus, and notably larger in the latter), head color (front, nape and neck-sides suffused blue in leucotis; no blue in griseipectus) and breast feathers (green with a blue suffusion, more intense near the neck, and with a broad pale grey or buff subterminal band and narrow blackish terminal one in leucotis; dusky grey with a broad cream to pale buff terminal band in griseipectus) (Olmos et al. 1998, Olmos, Cotinga in press). P. griseipectus is the same size as leucotis in wing length, but an important difference is the significantly longer bill of griseipectus and its deeper mandible (Olmos et al. 1998, Olmos Cotinga in press). P. griseipectus is proportionally larger headed than other Brazilian taxa, which should be visible in skull comparisons and make skeletal remains diagnosable, but there is no difference in bill width between griseipectus and leucotis. As in P. pfrimeri, the even more marked morphological differences, the geographical isolation, and very different habitat are sufficient to accord species status to P. griseipectus, as proposed by several authors. The degree of morphological differentiation of the three is at least comparable to that between accepted species taxa such as P. frontalis, P. devillei and P. molinae. It should be pointed out that some recent illustrations of the Brazilian taxa bear little resemblance to the birds in life (Juniper & Parr 1998), another factor that has delayed their acceptance as species.

Recent molecular work based on cytochrome-b sequences (Ribas 2004) has shown leucotis and griseipectus to be closely related, against morphology, whereas pfrimeri has a more basal position, with a divergence of 0.021

Recommendation: I would recommend a YES on this proposal, thus treating both Pyrrhura griseipectus and P. pfrimeri as a distinct species from Pyrrhura leucotis.

Literature cited:

Joseph, L. (2000) Beginning an end to 63 years of uncertainty: the Neotropical parakeets known as Pyrrhura picta and P. leucotis comprise more than two species. Proc. Acad. Nat. Sci. Philadelphia 150: 279-292.

Joseph, L. (2002) Geographical variation, taxonomy and distribution of some Amazonian Pyrrhura parakeets. Orn. Neotrop. 13: 337-364.

Juniper, T. & Parr, M. (1998) Parrots: a guide to the parrots of the world. Yale: Yale University Press.

Olmos, F., Martuscelli, P. & Silva e Silva, R. (1998) Ecology and habitat of Pfrimer's Conure Pyrrhura pfrimeri, with a reappraisal of Brazilian Pyrrhura leucotis. Orn. Neotrop. 8: 121-132.

Ribas, C.C. 2004. Filogenias moleculares e biogeografia histórica em psitacídeos (Aves; Psittacidae): padrões e processos de diversificação no Neotrópico. Unpublished Thesis, USP, São Paulo University.

José Fernando Pacheco, September 2005 (with assistance from Fabio Olmos)

________________________________________________________________________________________

Comments from Robbins: "NO. As we all recognize Pyrrhura taxonomy is a mess. Furthermore, I'm sure the majority of us suspect that the relatively minor differences in color patches, presence or absence of ventral markings, etc. may be quite plastic in this group. Indeed, in the current proposal there is a reference to Ribas (2004; I have not seen this work) in which the molecular data do not support the morphology and the genetic differentiation appears to be relatively small. It is unfortunate that Ribas's work is unpublished, as I think that work would offer insight into this particular proposal as well as a broader perspective on Pyrrhura.

"In my opinion, what is needed is a thorough molecular-based treatment of the Pyrrhura in order to inform us about the evolution of plumage morphology in this group. In sum, I take the conservative approach and wait for more information before supporting this split, i.e., I vote "no" for the present moment."

Comments from Silva: " YES. These two species mentioned by Pacheco are allopatric and completely different in plumage from leucotis. They live in different habitats and have small populations. There is not any possibility to find natural hybrids between these species and leucotis. This is a different situation from the Pyrrhura picta complex in Amazonia."

Comments from Stiles: "YES. The morphological and ecological picture seems to indicate species status clearly. I am less worried about the genetic data in questions of species status for reasons I mentioned in an earlier comment. Mutations that might involve mate choice, especially if sexual selection might occur, may spread rapidly in a population and produce reproductive isolation with very little change in the overall genome (and those few genes sequenced in genetic studies bear no known relation to reproductive isolation).  If allopatric speciation, especially in isolated peripheral populations, is the dominant mode of speciation (as Mayr proposed), then at some point most or all ancestral species will be paraphyletic. In sum: at this level, I go with morphology, behavior, ecology and distribution more than genetic data. Once we get to higher taxa (genera on up) genetic data become much more relevant."

Comments from Jaramillo: "YES - The morphological differences as described appear to be significant, but I am impressed by the ecological/habitat differences shown by these taxa. Given their differences in coloration, presumably related to mate choice, allopatric distributions and differences in habitat, it seems to me that these taxa have been separate for some time, and are unlikely to be able to mix in the future. I am comfortable thinking of them as species."

Comments from Zimmer: "YES". The morphological differences between these taxa are at least as great as those between several congeners that are currently treated as good species (e.g. the devillei, molinae, frontalis group), and it is my field impression that the vocal distinctions between each of them are probably on a par with typical interspecific differences in the genus. While I would echo some of Mark's concerns regarding the P. picta group, I really think this is a completely different case. The various components of the picta mess are largely parapatrically distributed, and based on my own museum work, I think that some errors have been made in where the boundaries between different forms have been drawn. In the present case, we are dealing not only with allopatrically distributed taxa, but taxa with hugely disjunct ranges and that occupy very different habitats (Fernando's nice summary didn't mention that griseipectus is isolated in the Serra de Baturité and Serra de Ibiapaba of Ceará and Serra Negra in Pernambuco, where it occupies "sky islands" of humid evergreen forest above 500 m, in a "sea" of lowland arid caatinga). In other words, these three taxa, no matter how recent their divergence, are now on independent evolutionary trajectories, with no chance that I can see of ever achieving secondary contact. In fact, given the continued erosion of the already tiny amount of humid montane forest in Ceará and Pernambuco, and the seemingly relentless conversion of dry forests in Goiás and Tocantins to soybean production, both griseipectus and pfrimeri will do well just to maintain their present tiny ranges."

"I also think that Gary's comments are right-on regarding the weight that should be given to molecular studies when attempting to resolve species-limits questions. If the genetic distance between populations under discussion is as great or greater than between currently recognized congeneric species pairs, I would see that as more ammunition for species recognition. I do not see the reverse case, in which genetic distances are slight, as a deal-breaker. I really think that the true value of molecular evidence is in sorting out phylogenies and higher-level relationships, not in resolving species limits."

Comments from Nores: "YES. Pienso que las diferencias de plumaje, genéticas y de hábitat que existen son suficientes para separarlas, especialmente a pfrimeri. No obstante, pienso que antes de tomar una decisión sería importante consultarle a Ribas para que opine si griseipectus y leucotis son especies diferentes, ya que él encontró una estrecha relación entre estas dos especies."