Proposal (607.1) to South American Classification Committee

Results of the voting on 607 (see below) were that 5 of 8 votes favored option 1, i.e. no change in species limits but a transfer of one subspecies to another species.

Therefore 607.1 becomes a YES/NO proposal as follows:

1. Retain three species of trumpeters, transferring ochroptera from P. leucoptera to P. crepitans to avoid a paraphyletic P. leucoptera.

(Van Remsen, 23 January 2015)

Comments from Remsen: “YES. This is the safest course given the lack of direct studies of interactions at contact zones and lack of study of isolating mechanisms. I am even a little reluctant to make the subspecies transfer because this is based on an mtDNA gene tree. I think Oppenheimer and Silveira, and Ribas et al., have done a terrific job in setting the stage for more focused and thorough sampling of populations in this group. The preliminary data from the headwaters all point towards multiple species, and I look forward to one transect through one of the potential contact zones for solid data on which to base a decision. Because Psophia are of conservation concern, I think that this is a case in which a series of photographs might be sufficient establish the pattern of plumage variation across a transect. Pegan and Hruska did a great job pulling these data together to make it easy for us or anyone else to get a clear and quick grasp of the problem.”

Comments from Stiles: “YES. I agree with Van on this one ... more intensive genetic analyses and attention to vocalizations seem warranted. Incidentally, I find Bret’s observation of singing by a trumpeter in captivity interesting – a source of vocal data? Has anyone tried mirror-image stimulation with these birds? If males, it might work!”

Comments from Jaramillo: “NO – I am going to go contrarian on this one, largely to state that there is something here in this paper that would be good to acknowledge. I am not sure that the splitting up of Psophia will occur in this vote, but keeping things as they are seems less palatable today to me, than beginning to incorporate some of these findings.”

Comments from Pacheco: “YES. In view of insufficient sampling present in Ribas et al. and Oppenheimer/Silveira, it is not appropriate changing boldly the arrangement of trumpeters.”

Comments from Zimmer: ““YES. I think we have to do this given the analysis of Ribas et al (2012), and given that the original Proposal 607 on changing species limits in Psophia failed to pass. However, like Alvaro, I think there is more that we should have done with the original proposal regarding species limits.”

Comments from Areta: “YES. This is the minimum necessary change to make our classification consistent with phylogenetic data. Given insufficient sampling and meager biological data, this is clearly a conservative change. As stated by others, the data presented by Oppenheimer & Silveira (2009) and by Ribas et al. (2012) suggest that there are more species-level entities in Psophia, but critical information is needed before this move can be made with confidence.”

=====================================================

Proposal (607) to South American Classification Committee

Recognize a new species-level taxonomy of trumpeters (Psophiidae)

Effect on SACC: This proposal reviews the taxonomy of the trumpeter family (Psophiidae). If passed, multiple new species would be recognized: the revision we recommend would split Psophia viridis into three species, P. viridis, P. dextralis, and P. obscura. Psophia crepitans napensis would be elevated to species status as P. napensis, and P. leucoptera ochroptera would be elevated to species status as P. ochroptera.

Background:

Psophiidae – the trumpeters -- is a family in the Gruiformes currently consisting of three species: P. leucoptera, P. crepitans, and P. viridis. New genetic and morphological evidence suggests that this arrangement underestimates trumpeter diversity, as many (or all) of the eight trumpeter taxa could be elevated from subspecific to species status. Here we consider the evidence for how these eight taxa should be classified.

Current taxonomy recognizes the following taxa as species. The Gray-winged Trumpeter (P. crepitans) occurs north of the Amazon River and is currently composed of two subspecies, Psophia crepitans crepitans and P. crepitans napensis (hereafter referred to as P. crepitans and P. napensis). The Pale-winged Trumpeter (P. leucoptera) occurs south of the Amazon River and west of the Madeira River and includes two subspecies, P. leucoptera leucoptera and P. leucoptera ochroptera (hereafter referred to as P. leucoptera and P. ochroptera). The Dark-winged Trumpeter (P. viridis) is endemic to Brazil and has three widely recognized subspecies, P. v. viridis, P. v. dextralis, and P. v obscura. A fourth subspecies, P. v. interjecta, has generally not been considered valid. Hereafter these taxa will be referred to as P. viridis, P. dextralis, P. obscura, and P. interjecta. The taxa within the P. viridis complex were recently reviewed by Oppenheimer and Silveira (2009).

Trumpeters are ground-dwelling birds that live in family groups that defend territories together and breed cooperatively (Sherman 1996). They are poor fliers and prefer to walk around obstacles, including water, rather than fly over them. They have been reported flying over small streams and, in rare cases, swimming to escape danger, but appear to seldom attempt to cross rivers. Hence, rivers are thought to be an important barrier to trumpeter movement and therefore gene flow between trumpeter populations. Trumpeter taxa are morphologically similar, predominately distinguished by the color of their hind-wing patch (or mantle); thus their English names, Gray-winged, Pale-winged and Dark-winged. This hind-wing patch appears to be important in maintaining visual contact between flock members and may play a role in displays (Sherman 1996).

Genetic Data:

Ribas et al. (2012) sampled and sequenced genetic mitochondrial genes (cyt b and ND2) from 62 individuals, including representatives of all species and subspecies of Psophia.

Phylogenetic Analysis:

Ribas et al. used a Bayesian analysis to build a phylogenetic tree for all Psophia taxa (Figure 1). This tree identified the eight trumpeter subspecies as monophyletic entities, and found P. crepitans and P. leucoptera, as currently defined, to be paraphyletic. Support values were high for branches separating taxa, with the exception of weaker support for their node 5, which separates P. crepitans and P. ochroptera. The Ribas phylogeny contains two main clades, one corresponding to taxa north of the Amazon River and one corresponding to taxa found south of the Amazon River. In the Northern clade, P. napensis was found to be sister to P. crepitans and P. ochroptera, while in the Southern Clade, P. leucoptera was sister to the four subspecies currently subsumed within P. viridis, with P. v. viridis sister to the remaining three taxa. In addition, P. obscura was found to be sister to a clade comprised of P. dextralis and P. interjecta (Ribas et al. 2012). This phylogeny demonstrates clear differences between taxa but little to no structure within them, indicating high gene flow within populations and lack of gene flow between them. Psophia viridis (i.e., the population formerly called P. viridis viridis) is somewhat of an exception, showing higher nucleotide diversity and signs of structure in its haplotype network; still, no evidence of hybridization with any other taxa was found in this case.

Vocal Data:

There are no vocal data studies that can inform this proposal. The three current trumpeter species seem to produce similar vocalizations, but this has not been well studied.

Morphological Data:

As previously described, all trumpeters are somewhat similar in appearance, differing mainly in the color of their wings (mantle) and lower throat (Sherman 1996).

Oppenheimer and Silveira (2009) analyzed the P. viridis complex and found three morphologically distinct groups defined by plumage color (based on a set color key) and geography: the taxa are each unique to regions separated by large rivers. Their proposed taxa are defined below:

1) P. viridis: Found from the region between the Madeira and Tapajos rivers, individuals have the distal and intermediate parts of the mantle Parrot Green, with the proximal portion Dark Green. These individuals have a conspicuous purple iridescence on the lower neck and wing.

2) P. dextralis: Found between the Tapajos and Tocantins rivers, individuals have a mantle whose distal portion is Olive Green, and whose intermediate and proximal portions are Very Dark Brown. The lower neck and wing have either discrete or absent purple iridescence.

3) P. obscura: Found from eastern Tocantins to Maranhão, individuals had the distal and intermediate portion of the mantle consistently Dark Green, with the proximal portion Dark Brown. Their lower necks and wings had very much reduced or no purple iridescence.

These morphological differences, in conjunction with the absence of known contact and hybridization zones, support the hypothesis that these taxa are genetically isolated from one another and thus likely to retain their genetic and phenotypic uniqueness in the future. In addition, this fact is supported by the lack of clinal variation found in these morphometric parameters, both within and across subspecies of Psophia (Oppenheimer and Silveira 2009). These criteria coincide with those utilized by the British Ornithologists’ Union when recognizing species based on morphology alone (Helbig et al. 2002). Oppenheimer and Silveira found that P. interjecta was not unambiguously diagnosable from P. dextralis based on plumage morphology so they recommend that it be considered a subspecies of P. dextralis. However, genetic data (see Genetic Data above; Ribas et al. 2012) shows evidence of structure between these two populations, perhaps indicating an absence of gene flow between them.

Contact zones:

These eight taxa are allopatrically distributed and separated by large rivers, which likely form isolating barriers to members of this ground-dwelling family. Nonetheless, the headwaters of these rivers, where rivers are relatively narrow, may allow contact between different populations.

Oppenheimer and Silveira studied specimens from the headwaters of rivers that run between the populations (where the rivers would serve as less effective barriers) and found no evidence of clinal variation or intergradation in their mantle and neck coloration. This observation suggests there is no gene flow between the populations even where the rivers may serve as less effective barriers and the populations may be in contact. Similarly, Sherman (1996) reported that P. ochroptera may be in contact with P. napensis at the western part of its range, but that there is no evidence that interbreeding occurs there. Finally, Ribas et al. also studied specimens collected from headwaters areas and found no mtDNA evidence for inter-taxon hybridization in these areas. They also found little geographic structure within populations, and that in some instances specimens from headwaters had the same haplotypes as those of the same populations from near the river mouths. This indicates that there is little isolation within these taxa, but strong isolation between them (see also Genetic Data section above).

Conclusions:

This proposal is largely based on the results of two studies. Oppenheimer and Silveira (2009) reviewed morphological evidence surrounding species limits in Psophia viridis; Ribas et al. (2012) used mitochondrial DNA to examine all three of the current trumpeter species.

Taken in conjunction with these birds’ natural history, the two studies complement each other: Ribas et al. (2012) provides genetic support for the eight trumpeter taxa and Oppeneheimer and Silveira (2009) shows that in the case of the P. viridis complex, genetic differentiation is accompanied by morphological divergence that suggests reproductive isolation and lack of hybridization. We suggest that the morphological evidence presented in Oppenheimer and Silveira (2009) can inform our inferences based on the Ribas mtDNA phylogeny. Specifically, this bears on the question of whether distinct taxa with rather shallow branch lengths ought to each be considered species or whether subspecies affinities merely need to be rearranged. The fact that Oppenheimer and Silveira’s morphological isolation data largely agrees with the genetic conclusions of Ribas et al. suggests that P. viridis should be split into multiple species; moreover, in this case they report no evidence of clinal variation at headwaters, indicating that color, especially mantle color, may be an important indicator of isolation. This idea is further supported by the fact that the colored mantle appears to be important in the social behavior of the species (Sherman 1996).

Given this conclusion, it does not seem advisable to consider P. crepitans and P. ochroptera as the same species given their readily diagnosable differences in mantle coloration. Recognizing these taxa as species also supports the recognition of P. napensis as a species despite the fact that it is very similar in coloration to P. crepitans; to do otherwise would create a paraphyletic species.

The one area where morphological and genetic taxonomic recommendations differ is their treatment of P. interjecta. This proposed subspecies is usually not recognized and has been thought to be an intergradation between P. dextralis and P. obscura (Sherman 1996). While Oppenheimer and Silveira find it too similar to P. dextralis in plumage morphology to be diagnosable as a species, Ribas et al. find genetic support for its distinction as a taxon. This further strengthens the idea that no clinal variation exists between trumpeter populations separated by large rivers; if P. interjecta is a valid taxon and not an intergradation, no examples of this sort of clinal variation exist. Ribas et al. found little to no mtDNA structure within each of the eight taxa but strong structure between them, and use P. interjecta and P. dextralis as particular examples of this result. However, given the lack of morphological evidence, the relatively few samples of this taxon used in the genetic study, and the shallow branch lengths of these taxa’s node on the tree shown in Figure 1, it is less clear whether this taxon ought to be elevated to a species or whether it should be recognized as a subspecies of P. dextralis.

English names: Should this proposal be passed, we leave it up to the SACC to determine English names of the resulting species.

Taxonomic possibilities:

There are quite a few permutations of taxonomic possibilities because of the number of taxa involved, but we have identified three possibilities that appear most reasonable given the new genetic and morphological data:

1. Retain three species of trumpeters, transferring ochroptera from P. leucoptera to P. crepitans to avoid a paraphyletic P. leucoptera.

2. Recognize seven species of trumpeters; recognize P. napensis and P. crepitans and elevate all P. viridis subspecies to the species level except for P. interjecta, based on lack of morphological evidence for this latter taxon’s distinctness from P. dextralis; classify interjecta as a subspecies of P. dextralis.

3. Recognize eight species of trumpeters; all those described in the above option, and additionally recognize P. interjecta based on lack of evidence of structure within the taxon (contrasting with its genetic separation from P. dextralis) and its geographic separation from P. dextralis by a large river.

Recommendation:

We suggest that recent morphological and genetic evidence is sufficient to warrant a novel species-level phylogeny for trumpeters. As described above, we consider the evidence sufficiently strong to support breaking trumpeters into multiple species instead of rearranging subspecies into the currently recognized species. We therefore recommend a “No” vote on Option 1. Instead, available evidence supports recognizing at least seven species of trumpeters, and we therefore recommend a “Yes” vote on Option 2. We are agnostic on the question of whether P. interjecta ought to be elevated to full species status, but suggest that this approach (Option 3) also a valid possibility.

Literature Cited:

Helbig, A.J., Knox, A.G., Parkin, D.T., Sangster, G., & Collinson, M. 2002. Guidelines for assigning species rank. Ibis, 144(3), 518-525.

Oppenheimer, M., & Silveira, L. F. 2009. A taxonomic review of the dark-winged trumpeter Psophia viridis (Aves: Gruiformes: Psophiidae). Papeis Avulsos De Zoologia, 49(41), 547-555.

Ribas, C.C., Aleixo, A., Nogueira, A.C.R., Miyaki, C.Y., & Cracraft, J. 2012. A palaeobiogeographic model for biotic diversification within Amazonia over the past three million years. Proc. R. Soc. B 279, 681-689.

Sherman, P.T. 1996. Family Psophiidae (trumpeters). Pp. 96-107 in: del Hoyo, J., Elliott, A., & Sargatal, J. eds. (1996). Handbook of Birds of the World. Vol. 3. Hoatzins to Auks. Lynx Edicions, Barcelona.

Teresa Pegan and Jack Hruska, December 2013

========================================================

Comments from Stiles: “YES to proposal 1. The splitting of P. viridis into multiple species seems a bit premature, as the genetic distances are not all that great. Moreover, trumpeters are very vocal birds that spend most of their time in dark forest understory; hence the importance of mantle color as an isolating mechanism (vs. vocalizations) may be uncertain; here I think that voice might be more important, and that some sonograms (hopefully with at least some playback experiments) could help to shift the balance toward splitting up the viridis complex, given that all forms are apparently allopatric.”

Comments from Nores: “YES, to proposal 2. However, I agree with Gary that voice might be very important, and that some sonograms and some playback experiments could help especially in the viridis complex.”

Comments from Zimmer: “YES. I’m conflicted on the specifics. I think that Proposal 1 is the minimum that needs to be done – the genetic data are clear for removing ochroptera from leucoptera. I agree with Gary that vocalizations seem likely to be of more importance than slight differences in mantle color. Unfortunately, I’m guessing that it is going to be a long time before anyone comes up with meaningful sample sizes of auditory recordings of each of these taxa. As we have already seen with attempts to evaluate geographic differences in vocalizations in Ortalis, it can be very difficult to assess vocal differences in Psophia due to variation from one recording to the next in the number of individuals vocalizing (and this usually being unknown). Also, I can’t help but think that the plumage differences between ochroptera and crepitans/napensis have to be important when it comes to signal recognition, mate choice, etc. – unlike the relatively conservative plumage distinctions between the various taxa in the viridis complex, ochroptera versus anything else is pretty striking. If we treat ochroptera as a different species from crepitans, then we also have to split crepitans and napensis to avoid a paraphyletic situation within crepitans. If we split those two taxa, then that opens the door for recognizing the various populations of viridis as specifically distinct, an outcome that may be the correct way to go, but one for which I am less enthusiastic (lacking vocal data) than I am about splitting ochroptera and crepitans. Despite the short branch lengths, I do think that all of these taxa are likely on independent evolutionary trajectories, especially given the seeming effectiveness of the river barriers in restricting any potential for gene flow. I guess I would weakly favor Option 2, to recognize 7 species (leucoptera, ochroptera, crepitans, napensis, viridis, obscura and dextralis). I can’t make myself go the extra mile in splitting interjecta, which, from what I can tell, isn’t even necessarily morphologically diagnosable.”

Comments from Robbins: “After reading the proposal and Kevin’s comments, I support proposal 2, in recognizing seven species of Psophia. Certainly, at a minimum, proposal one should pass as the genetic data demonstrate that ochroptera must be transferred to crepitans if one is going to take the narrow view that only three species should be recognized. Given the relatively little genetic differentiation I appreciate how others might want to be conservative and recognize only three species. Kevin is correct that it may be a long time before pertinent vocal and display data are obtained for these taxa. From what we currently know, it does seem that the rivers, including the headwaters, are acting as barriers to gene flow. Until there are data to indicate otherwise, I lean more toward recognizing the seven taxa outlined in proposal 2 as species.

Comments from Bret Whitney: “I see Ribas et al, like most phylogeographic studies to date, as suffering from poor sampling, especially for the two taxa east of the Xingu (= 2 for interjecta and 1 for obscura), and uneven application of “results”.  The proposal says that Ribas et al. recovered “high support values for branches separating taxa” except for “weaker" support value for crepitans and ochroptera.  I’d say several others are no or little better, especially where sampling was poor (as it was for ochroptera = 2).  Ribas et al (and, implicitly, the proposal) don’t seem to have any reservations about calling viridis, dextralis, and obscura monophyletic species despite poor support, but do not mention similar, possibly geographically linked structure within napensis (perhaps across the Napo, a river known to separate a number of species-level taxa).  Had someone given a name to birds up in nw. Amazonas, Brazil, I imagine the proposal would be recommending species-status for that, too.  The proposal, however, echoes a statement from Ribas et al to the effect of “the phylogeny demonstrates clear differences between taxa butt little to no structure within them…” The napensis example negates this assertion and I don’t think they have enough of a sample of any taxa to make this statement.  Headwaters regions were woefully unsampled. Sampling was a little better in the morphological study of Oppenheimer and Silveira, but it was also not robust.

“Another observation I have is two groups of ochroptera of 4-10 birds each that I have seen in Jaú National Park and near Barcelos that had some individuals decidedly more orangish than others.  This made me wonder if the orange tint in the backs and remiges of these birds is real (= genetically determined), or possibly an artifact of bathing constantly in blackwater… something to at least think about, I suggest, in the same vein as White-faced Whistling-Ducks showing variably orangish faces (but rarely “white”) due to feeding habits. I can’t say I see orangish herons or egrets, for example, more frequently in the Negro basin, but it does seem plausible that pale-backed trumpeters bathing in dark-blackwater streams inside forest could become heavily stained — and perhaps that is inevitable and to be expected.  Specimens of ochroptera are rare, and there are not many available for analysis (Blake did have four males and a female and didn’t mention any variation, but I would not expect him to unless it was dramatic).

“I would vote to accept proposal #1 and part of #2, with napensis and ochroptera as subspecies of crepitans (with a caveat on ochroptera as a valid taxon [tentatively maintain it for now]); a monotypic leucoptera; and with the split of viridis, dextralis, and obscura as species because the analyses of Oppenheimer and Silveira, and Ribas et al., despite very weak sampling, are congruent and mantle colors are quite different (in my opinion); I think the split of the Dark-winged group represents an advancement from where we are at present.  I also have two recordings of adult, captive obscura that are notably different from viridis and dextralis, which are much more similar to each other — indicative of differentiation, I would guess, but not enough on which to base judgments.  I would wait for more analysis of better sampling — morphological, vocal, and genetic — of all taxa and putative populations before moving for further revision; I do not see compelling reason(s) for doing more than this at the moment.  It certainly appears that interjecta is individual variation, and we do not know whether overlooked cantatrix Boeck 1884 from Beni represents an additional taxon or not; no sample from that region was included.

“Some final thoughts — Would it not be possible to include most or all of the specimens sampled by Oppenheimer and Silveira in molecular analysis?  It seems to me that most of them would have sufficient material in their feet to permit extraction of tissue.  I think it would be especially wise to analyze all of the type specimens and a few selected others, at least, to try to clear up some lingering doubts (that I have, at least) about their provenance. Specifically, Spix’s two leucoptera were labeled from the Rio Negro, but Hellmayr decided they had to have come from south of the Amazon and west of the Madeira, so fixed the type locality “left bank of the Rio Madeira”.  He may be correct in imagining that Spix obtained the specimens from villagers who had them as pets, and they may well have originated west of the Madeira, but this should be checked.  Trumpeters are exceptionally fine pets and are traded and kept by people all over the Amazon.  I think it probable that the type of P. viridis, from Parintins in the extreme lower part of the Madeira-Tapajós interfluvium on the largest island of the Tupinambarana region, was a pet that had been taken elsewhere; I doubt that trumpeters occur naturally around Parintins.  This should also be investigated genetically, along with interviews of hunters and other locals on the island, who will definitely know if trumpeters are in the forests there, or not.  Oppenheimer and Silveira cited a single specimen of viridis (MNRJ 9645) from “Machado, Mato Grosso”, an indeterminate locality.  It was the only example among their sample of 24 viridis that had aberrant mantle coloration (two characters), and it was the only specimen from Mato Grosso.  They attributed the differences to individual variation because another specimen “collected 70 km of this specimen (MZUSP 76728) is typical…” Oppenheimer and Silveira did not indicate how they determined the location of “Machado, MT” and did not map any Mato Grosso localities for viridis (their figure 4).  It would be desirable to analyze this interesting specimen genetically in an attempt to understand something more about it.”

Comments from Jaramillo: “YES on 1, and part of 2. Hmm, a problematic one, but I find Bret’s comments valuable and convincing. Keep napensis and ochroptera as part of crepitans. This moves us along, taking to consideration the new data we have, but conservative on the parts that are still missing from this analysis.”

Comments from Stotz: “YES to part 1. Currently I am unwilling to split up viridis based on current data.”

Comments from Remsen: “YES on 1, following from Bret’s comments and waiting for additional data and a proposal that would allow accepting of the part of 2 for which data are solid.”

Comments from Pérez-Emán: “YES to Part 1. Nodal support grouping ochroptera and crepitans has the weakest support indicating that these three taxa are included in a polytomy. Besides, in the supplementary material you can find the nuclear tree and there is a haplotype shared by both crepitans and napensis. This might be ancestral polymorphism but it could also be gene flow. I think it is better to be conservative until more data are available.”