Proposal (623) to South American Classification Committee


Recognize newly described Campylorhamphus gyldenstolpei and Campylorhamphus cardosoi and split Campylorhamphus procurvoides into four species



Effect on South American CL: If adopted, this proposal would add two newly described species of Campylorhamphus to the list and recognize three current subspecies of C. procurvoides (multostriatus, probatus and sanus) as species level taxa.


Background: The Curve-billed Scythebill (Campylorhamphus procurvoides) has been treated as a polytypic species for 79 years (Zimmer 1934), with five recognized subspecies distributed in two groups according to vocal and morphological differences: multostriatus (multostriatus and probatus) and procurvoides (procurvoides, sanus, and successor), with the former group being more similar to C. trochilirostris than to the remaining subspecies of Campylorhamphus procurvoides (Marantz et al. 2003). More recently, Portes and Aleixo (2009) showed that the type series of C. p. successor consists of a mix of individuals belonging to C. trochilirostris (including the holotype for C. p. successor) and specimens whose plumage characters closely matched those of an undescribed taxon probably closely related to the nominate subspecies of C. procurvoides.


New information: DNA sequence data for the mitochondrial genes cytochrome b (cyt b, 1048 bases pairs) and NADH subunit 2 (ND2, 1041 bases pairs) were obtained for 41 individuals of all species in the genus Campylorhamphus, including all taxa currently grouped under the polytypic C. procurvoides (Aleixo et al. 2013, Portes et al. 2013). The phylogeny obtained with Bayesian inference strongly supports that C. procurvoides, as traditionally defined, represents a polyphyletic species, with taxa currently classified as subspecies of procurvoides found in three separate clades with disparate phylogenetic affinities (Aleixo et al. 2013, Portes et al. 2013). The three highly supported and reciprocally monophyletic clades that group taxa currently classified as subspecies of the polyphyletic C. procurvoides are as follows: (1) birds occurring south of the Amazon and east of the Xingu River in the Xingu center of endemism (Silva et al. 2005; corresponding to the taxon multostriatus, whose type locality lies in this area of endemism); (2) birds found south of the Amazon and west of the Madeira rivers (in the Inambari center of endemism, corresponding to the other new Campylorhamphus taxon described by Aleixo et al. [2013], i.e. gyldenstolpei) and birds distributed north of the Amazon, which correspond to the taxa sanus (found in the Napo and Imeri areas of endemism) and procurvoides (corresponding to the Guiana area of endemism); and (3) birds found south of the Amazon and between the Madeira and Xingu rivers, corresponding to the taxa probatus (distributed in the Madeira center of endemism) and the other new Campylorhamphus taxon described by Portes et al. (2013), i.e. cardosoi, endemic to the Tapajós area of endemism. 


Because C. multostriatus appears to be the sister group to all remaining taxa grouped until now in C. procurvoides and C. trochilirostris, and the clade C. probatus / cardosoi was recovered with high statistical support as sister to C. trochilirostris, the C. procurvoides complex cannot include C. multostriatus, C. probatus, and C. cardosoi.


Analysis/Recommendation: The fact that the Bayesian phylogeny recovered cardosoi, gyldenstolpei, multostriatus, probatus, procurvoides, and sanus as reciprocally monophyletic taxa with strong statistical support in conjunction with morphological and vocal diagnoses of these taxa, support their recognition as valid species level taxa as follows: Snethlage's Scythebill (C. multostriatus; distributed in the northern portion of the Xingu-Tocantins interfluve); Rondônia Scythebill (C. probatus; distributed in the Madeira - Tapajós interfluve); Tapajós Scythebill (C. cardosoi; distributed in the Tapajós - Xingu interfluve); Curve-billed Scythebill (C. procurvoides; distributed on the Guianan shield north of the Amazon and east of the Negro - Branco rivers in Venezuela, Brazil, Guyana, Surinam, and French Guiana); Zimmer's Scythebill (C. sanus; distributed west of the Branco - Negro rivers in Amazonian Brazil and Venezuela westward towards the base of the Andes in Colombia, Ecuador, and Peru north of the Amazon/Solimões rivers); and Tupana Scythebill (C. gyldenstolpei; distributed west of the Madeira and south of the Solimões rivers in Amazonian Brazil).


Literature Cited


Aleixo, A., Portes, C. E. B., Whittaker, A., Weckstein, J. D., Gonzaga, L. P., Zimmer, K. J., Ribas, C. C. and Bates, J. M. (2013).  Molecular systematics and taxonomic revision of the Curve-billed Scythebill complex (Campylorhamphus procurvoides: Dendrocolaptidae), with description of a new species from western Amazonian Brazil.  In J. del Hoyo, A. Elliott, J. Sargatal & D. Christie (Eds), Handbook of the Birds of the World.  Special Volume: New Species and Global Index, pp.253-257.  Lynx Edicions, Barcelona.

Marantz, C., A. Aleixo, L. R. Bevier and M. A. Patten (2003). Family Dendrocolaptidae (Woodcreepers). Pp. 358-447 in: del Hoyo, J., A. Elliott, and D. A. Christie (eds.) (2003). Handbook of the Birds of the World. Volume 8: Broadbills to Tapaculos. Lynx Edicions, Barcelona.

Portes, C. E. B. and A. Aleixo, A. 2009. Campylorhamphus procurvoides successor  (Aves: Dendrocolaptidae) is a junior synonym of Campylorhamphus trochilirostris notabilis. Zoologia 26: 547 – 552.

Portes, C. E. B., Aleixo, A., Zimmer, K. J., Whittaker, A., Weckstein, J. D., Gonzaga, L. P., Ribas, C. C., Bates, J. M. and Lees, A. C. (2013).  A new species of Campylorhamphus (Aves: Dendrocolaptidae) from the Tapajós-Xingu interfluve in Amazonian Brazil.  In J. del Hoyo, A. Elliott, J. Sargatal & D. Christie (Eds), Handbook of the Birds of the World.  Special Volume: New Species and Global Index, pp.258-262.  Lynx Edicions, Barcelona.

Silva, J. M. C., S. B. Rylands, and G. A. B. Fonseca (2005). The fate of the Amazonian areas of endemism. Conservation Biology 19: 689-694.


Carlos Eduardo B. Portes and Alexandre Aleixo, December 2013





Comment from Thomas Donegan: “We considered these and other HBW splits relating to Colombian taxa for purposes of the Colombian checklist update this year. We accepted some of the proposed changes to Campylorhamphus taxonomy set out in the paper referred to in this proposal, but not all of them.  We noted as follows: "Populations of this species in Colombia (subspecies sanus) are closely related to C. (p.) gyldenstolpei, which was described by Aleixo et al. (2013). Vocal differences between these two taxa and nominate procurvoides are only in the note shape of part of the song; and in note shape of a call comprising a single note similar to the variable note of the song. Typically for antbirds, multiple (>3) diagnostic vocal differences are recommended to treat populations as separate species (Isler et al. 1998). In the absence of studies into the extent of vocal variation between sympatric scythebills (e.g. Brown-billed and Red-billed Scythebills in Colombia), we treat sanus, procurvoides and gyldenstolpei as allopatrically distributed subspecies of the same species. Molecular (<1.2% mtDNA), biometric and plumage differences between the three taxa initially appear relatively minor. We instead adopt Aleixo et al. (2013)’s alternative three-way split of procurvoides (including sanus and gyldenstolpei) from extralimital species probatus (including recently-described cardosoi of Portes et al. 2013) and multostriatus. Such a treatment separates out vocally highly divergent populations with deep (4.3%-6.7%) molecular differentiation and avoids paraphyly with eastern populations of Red-billed Scythebill C. trochilirostris.  However, further splitting of procurvoides (or other taxa in this complex) lack support of a detailed vocal study. Further consideration needs to be given to vernacular names for this group and to the inclusion of vocally distinctive Colombian populations of C. trochilirostris (cf. Donegan 2012a) in molecular studies. This proposal would be better unbundled into various separate parts as follows: (A) recognising procurvoides, probatus and multostriatus as separate species; (B) splitting sanus from procurvoides, (C) splitting gyldenstolpei from procurvoides, (D) splitting cardosoi from probatus.  We only adopted part A; the other proposed splits here are much more borderline and may be better left as subspecies in the absence of a detailed vocal study.


Reference: Donegan, T.M., McMullan, W.M, Quevedo, A. & Salaman, P. 2013. Revision of the status of bird species occurring or reported in Colombia 2013. Conservacion Colombiana 19: 3-10.http://www.proaves.or/wp-content/uploads/2013/12/Checklist-Update-2013-Conservacion-Colombiana-19-3-10.pdf.”


Comments from Stiles: “NO.  And back to HBW. As in several other cases, data presentation leaves much to be desired.  Starting from the genetic data, three clades appear to be definitely recognizable within the procurvoides complex: multostriatus; procurviodes+sanus+gyldenstopei; probatus+cardosoi. Vocalizations also identify these three clades well enough.  However, hereafter, things get messy, and good statistical analyses of large samples of vocalizations and perhaps, morphometrics would have helped.  The fact that the third clade groups with C. trochilirostris leaves open the possibility that this clade might actually belong in the latter species.  To address this, it would have been useful, to say the least, to have data from that species – distribution and vocalizations – available for comparison.  Given the mixed nature of “successor”, it is clear that at least gyldenstolpei apparently is sympatric with trochilirostris over at least part of its distribution: is the probatus+cardosoi clade also sympatric with trochilirostris?  If so, this would clearly mark this clade as representing a separate species.  Given that Todd evidently confused the two in his description of “successor”, the morphological differences between the two species appear to be subtle.  However, given the minuscule genetic differentiation between cardosoi and probatus and the rather subtle difference in vocalizations, I see little justification for considering each of these as different species without at the least, a more thorough analysis of a better series of vocalizations (and perhaps, playbacks) – evidently morphometrics were not examined at all).  In the northern-western clade, things are perhaps a little clearer: at the least, it represents a separate species, but within this clade, the decision to split it up into three species is more questionable (especially with respect to separating gyldenstolpei and sanus, which are very similar genetically).  Again, more thorough analyses of vocalizations, perhaps morphometrics and hopefully playbacks could settle the question.  In sum, I think that the authors have provided good evidence that procurvoides in the broad sense is indeed polyphyletic and that at the least, multostriatus and the procurvoides+sanus+gyldenstolpei clade represent separate species; the probatus+ cardosoi clade represents a separate taxon from these two, but its relation to trochilirostris was not resolved: it could represent a separate species or be lumped into trochilirostris.  Existing data for deciding this may exist, but were not presented.  The evidence for splitting the procurvoides+sanus+gyldenstopei clade into three (or two) species is less convincing as presented, likewise that for splitting probatus and cardosoi as species.  I think that this proposal should be broken into parts: A-split multostriatus from procurvoides; B1: recognize procurvoides as a species with three subspecies; B2: split procurvoides from sanus+gyldenstolpei; B3: split gyldenstolpei from sanus; C: remove the probatus+cardosoi clade from species procurvoides; this seems clear enough, bur hereafter we are left with the question of what to do with it – either lump it with trochilirostris (C1?) or consider it a separate species with two subspecies (C2), and then decide on whether or not to split probatus and cardosoi as separate species (C3).  Given that the proposal as it stands is all-or-nothing, I would have to vote NO at this time.  I´m not saying that the authors are wrong, just that there are too many loose ends that need tying.”


Comments from Remsen:  “NO.  Clearly the current taxonomy is incorrect, and clearly more than two species are involved.  However, by packaging all of the proposed splits into a single proposal, those with weak support bring down those with strong support.  As recommended by Thomas and Gary, this one needs to be re-packaged and subdivided accordingly, stratified according to strength of evidence in each case.”


Comments from Nores: “NO. After analyzing the seven proposals made by Aleixo et al. (#617-623), it is evident that these authors are enthusiastic to create new species, and with scant foundations. What it took nature millions of years to create, they try to do in just a few weeks.”


Comments from Zimmer: “As the proposal is currently written, with several separate decisions bundled into one proposal, NO.  As Gary and Van both note, this proposal needs to be broken down into several specific parts.  In the interest of full disclosure, I must point out that I was a junior author on each of the two papers upon which this proposal is based.  My authorship was based on contributions of data (particularly tape recordings) and the fact that I was among the first to point out some of the variation in vocal characters within the procurvoides group, particularly as it pertains to multostriatus, cardosoi and probatus.  When we first got into this project (long before the HBW special volume was even on the radar), I was slated to do a quantitative vocal analysis, and toward that end, I stockpiled recordings of the various taxa and did some initial spectrographic analysis to identify potential vocal characters for analysis.  The press of other commitments bogged me down, and my analysis never got beyond this initial character-identification stage.  Meanwhile, the HBW new species chapter concept was hatched, and the more senior authors committed the Campylorhamphus papers to publication in that venue.  Once that commitment was made, fairly rigid production deadlines (to adhere to publication deadlines) created external pressures to maintain a certain momentum.  By this time, I had already turned over all of my accumulated audio samples to other authors.  I came back from an extended tour and found that both scythebill papers were already written and submitted.  I still had opportunity to comment prior to the galley stage, but there was little opportunity to make substantive changes at that point in the game.  All of this is just to explain why the opinions that I am about to express are not in lockstep with the taxonomic recommendations made in two papers for which I am listed as an author.


I think that the current evidence strongly supports splitting “Curve-billed Scythebill” into three species:  1) nominate procurvoides + sanus + gyldenstolpei; 2) multostriatus; and 3) probatus + cardosoi (probatus has priority).  This three-way split is supported by fairly deep genetic differences and by some solid vocal distinctions (although our papers did not do a particularly thorough job of making these clear).  I also feel that cardosoi is a valid taxon, worthy of formal recognition, but I think that it should be considered a subspecies of probatus.  It is vocally distinct from multostriatus (and also differs in plumage characters), which replaces cardosoi east of the Xingu, and I have no doubt that multostriatus should be considered specifically distinct from cardosoi + probatus.  Vocal distinctions between cardosoi and probatus are tenuous.  As far as I can tell, each of these taxa has two different song types, a long trill that gains in amplitude through the middle of the song and then fades at the end, and which lacks a strongly differentiated initial note, and a shorter song with an emphatic, highly differentiated introductory note that is followed by a short trill of even pitch and amplitude.  The difference between cardosoi and probatus is in the relative frequency with which the two song types are given.  In cardosoi, the shorter song with the emphatic, highly differentiated introductory note is the vocalization that I hear most regularly, whereas the homologous vocalization seems to be rarely delivered in probatus.  Conversely, probatus most often gives the long trill without the differentiated intro note, and that vocalization, although given by cardosoi, is much less frequently heard.  Similarly, I am not convinced that there are diagnosable vocal differences between gyldenstolpei, sanus and nominate procurvoides, and even if there are, I don’t think our paper made a strong enough case, given the lack of a quantitative analysis.


“There is one additional fly-in-the-ointment in this whole mess.  I would have preferred that we had dealt with the entire trochilirostris and procurvoides complexes simultaneously.  This approach was deemed too unwieldy (and, to be fair, we were lacking adequate vocal and tissue samples for some of the key taxa in the trochilirostris group), and was abandoned altogether once we were committed to publishing in HBW.  However, I suspect that some taxa that are more properly treated as part of the procurvoides complex are currently treated as part of trochilirostris, and I think that it is going to prove impossible to get either complex correctly restructured without simultaneously looking at the other.


“So, to summarize:  I would vote NO to the proposal as currently structured, and strongly favor splitting this proposal up into several component proposals that could be dealt with separately.  With that in mind, I would strongly favor splitting procurvoides into three species:  procurvoides (including sanus and gyldenstolpei), probatus (including cardosoi) and multostriatus, with the two new taxa (= cardosoi and gyldenstolpei) recognized at the subspecific (but not specific) level.”


Comments from Robbins: “NO, as has been pointed out by those who have already evaluated this proposal, clearly multiple species need to be recognized, but the proposal does indeed need to be repackaged and presented in separate proposals.”


Comments from Pacheco: “NO. É evidente que há mais de um táxon no presente complexo. Aguardarei – como sugerido pelos demais membros – que a proposta seja reestruturada, sob a ótica do BSC, para contemplar a divisão em três espécies.”