Proposal (1031) to South American Classification Committee

 

 

Revise species limits in the Schiffornis turdina complex

 

 

Background:

 

The Schiffornis turdina complex currently comprises thirteen taxa sorted into five species:

 

·      S. turdina (5 ssp.: turdina, intermedia, wallacii, amazonum, steinbachi)

·      S. stenorhyncha (2 ssp.: stenorhyncha and panamensis)

·      S. veraepacis (4 ssp.: veraepacis, dumicola, acrolophites, rosenbergi)

·      S. aenea (monotypic)

·      S. olivacea (monotypic)

·       

They were all formerly treated as a single species (Snow 2004, Kirwan and Green 2012), although many authors have noted — based on marked vocal differences observed in the field — that more than one species was involved (e.g., Ridgely and Tudor 1994).

 

The complex has been broadly treated as consisting of five species since ~2012 after the works of Nyári (2007) and Donegan et al. (2011). Briefly, Nyári (2007) examined mtDNA and vocal variation in the complex, and proposed that it should be split into five species. The proposal, however, was not widely accepted for multiple different reasons (see Proposal 327 for details). Donegan et al. (2011) conducted an additional analysis of vocal variation, and made a new proposal with more evidence for splits (see Proposal 505 for details). Together, the findings of Nyári (2007) and Donegan et al. (2011) led to the current treatment of the S. turdina complex into five species.

 

As several noted in their comments on previous proposals on the S. turdina complex, those splits were a significant step forward in recognizing species-level diversity within the complex. However, several gaps remained, and recognition of additional species was likely with more detailed information. Notably, several taxa remained poorly sampled and vocal differences had never been quantified.

 

New information:

Lima et al. (2024) recently examined geographic variation in song, plumage color, and morphology and population genetic structure across the genus Schiffornis. They analyzed the songs of 314 individuals, 598 study skins, 56 UCE sequences, and hundreds of mtDNA sequences, including all or almost all taxa in each data set.

 

Lima et al. (2024) found eleven phenotypically and genetically diagnosable taxa in the S. turdina complex, including a new taxon and one taxon that has been treated a synonym of S. t. amazonum. They corroborated previous splits and suggested that as many as nine biological species should be recognized in the complex based on a “yardstick” approach, using patterns of divergence between sympatric or parapatric taxa within the genus to assess the rank of the remaining, allopatric taxa.

 

Patterns of phenotypic and genotypic differentiation vary widely among taxa in the S. turdina complex. Some taxa have markedly distinct songs but shallowly diverged genotypes; others have similar songs but deeply diverged genotypes; others have deep mitochondrial but negligible nuclear differentiation. Several taxa are indistinguishable visually. See Lima et al. (2024) for details. Because variation in plumage and morphology is probably of little (if any) relevance in defining biological species limits within this group, I will focus below on the vocal and genetic results.

 

Here’s a map from Lima et al. (2024) that will be useful in understanding the geographic range of each taxon. The map shows sampling locations of sound recordings (circles), study skins (squares), and genetic samples (triangles). Colored stars indicate the type locality of valid taxa based on the revised classification proposed by Lima et al. (2024), and numbered black stars indicate the type locality of taxa synonymized therein. For more details and other relevant figures, see Lima et al. (2024).

 

 

The proposal is subdivided as follows:

 

A. Treat rosenbergi as a separate species from S. veraepacis

Schiffornis veraepacis currently comprises two disjunct populations: one in Central America (including ssp. veraepacis, dumicola, acrolophites) and the other ranging from western Colombia to northern Peru west of the Andes (rosenbergi). The range limits of the four putative taxa within S. veraepacis, based on plumage color variation, have never been fully understood. Previous work on the S. turdina complex showed that these two disjunct populations of S. veraepacis differ diagnosably in song and mtDNA (Nyári 2007, Donegan et al. 2011). Based on these differences, some considered that these populations could be treated as separate species. However, some considered that the differences between the two could reflect sampling gaps (see Proposal 505).

 

Using larger sample sizes, Lima et al. (2024) showed that the Central American and South American populations of what is currently recognized as S. veraepacis do have markedly different songs, with differences equivalent in magnitude to those between the sympatric S. veraepacis and S. stenorhyncha. Additionally, Lima et al. (2024) also confirmed, with greater sampling, that the Central American and South American populations of S. veraepacis are diagnosable on the basis of mtDNA. There are no consistent differences in either phenotype or genotype within these two vocally and genetically distinct populations. Taking all these results together, Lima et al. (2024) suggested that the Central American and South American populations of S. veraepacis should be recognized as two separate biological species, both monotypic: S. veraepacis and S. rosenbergi. I recommend a YES vote on this subproposal, primarily due to significant vocal differences distinguishing rosenbergi from all other taxa in the complex (Lima et al. 2024).

 

B. Treat steinbachi as a separate species from S. turdina

Lima et al. (2024) discovered that two vocally and genetically distinct populations currently considered conspecific within S. turdina are locally sympatric in northwestern Bolivia and southeastern Peru. These populations are steinbachi and intercedens, the latter currently considered a synonym of amazonum (more details about the name intercedens below in subproposal “D”). These two populations differ diagnostically in multiple song characters, and there is no phenotypic evidence of hybridization between the two despite the close geographic proximity and even overlap of sound recordings. Although there is some genetic evidence of hybridization between the two, a genome-wide FST = 0.3 suggests that steinbachi and intercedens remain considerably differentiated despite some gene flow where they come into contact (Lima et al. 2024).

 

Although additional research of the contact zone between steinbachi and intercedens is clearly needed to determine levels of reproductive isolation, current evidence suggests that reproductive isolation between the two is moderate at least (Lima et al. 2024). Based on the vocal differences between the two and the sizeable genotypic differentiation in the face of some gene flow, I believe that steinbachi and intercedens are best considered separate biological species now.

 

C. Treat amazonum as a separate species from S. turdina

What is currently understood as S. t. amazonum actually includes two vocally and genetically distinct populations separated by the Amazon River (more details below in subproposal “D”). Here in subproposal “C” we are dealing with the northern population. This population exhibits significant vocal and genome-wide differentiation in relation to all other taxa of the S. turdina complex except the newly described cracrafti (Lima et al. 2024).

 

See subproposals “F” and “I” before voting on this one.

 

D. Treat intercedens as a separate species from S. turdina

Schiffornis turdina intercedens Todd, 1928 was synonymized with S. t. amazonum (Sclater, 1860) shortly after its description (Hellmayr 1929) and has since been universally treated as a synonym (Pinto 1944, Snow 2004, Kirwan and Green 2012). Using the largest sample of specimens examined to date, Lima et al. (2024) corroborated Hellmayr’s (1929) assessment that the two taxa are indeed visually indistinguishable. However, Lima et al. (2024) found that intercedens and amazonum have diagnostically distinct songs and genotypes, and therefore reinstated intercedens to taxonomically recognize these differences.

 

This population, intercedens, exhibits significant vocal and genome-wide differentiation in relation to all other taxa in the S. turdina complex except wallacii and cracrafti. It differs diagnostically from both in song and mtDNA, but vocal differences between intercedens and wallacii are subtle and — based on a “yardstick” approach — unlikely to generate premating isolation (Lima et al. 2024).

 

See subproposals “E” and “H” before voting on this one.

 

E. Treat wallacii as a separate species from S. turdina

This taxon caused some confusion in previous proposals on the S. turdina complex owing to uncertainties regarding its type locality. Many reference works incorrectly mention that the range of wallacii reaches the Guiana Shield (e.g., Hellmayr 1929, Pinto 1944, Kirwan and Green 2012), where only olivacea occurs (Lima et al. 2024). Given that wallacii and olivacea are visually indistinguishable (Lima et al. 2024) and their ranges were traditionally based on plumage color variation alone, this confusion is not surprising. Donegan et al. (2011) correctly assigned the type locality of wallacii to the city of Belém, in the Brazilian state of Pará. Lima et al. (2024) further restricted the type locality to the vicinities of the Guamá River near Belém, based on the collector’s itinerary.

 

Lima et al. (2024) found sizeable vocal and genome-wide differentiation between wallacii and all other taxa in the S. turdina complex except cracrafti and intercedens. Although wallacii differs diagnostically from cracrafti and intercedens in song and mtDNA, vocal differences between wallacii and intercedens are subtle and — based on a “yardstick” approach — unlikely to generate significant reproductive isolation (Lima et al. 2024). Song differences between wallacii and cracrafti are equivalent in magnitude to those between sympatric Schiffornis species, but overall nuclear genetic differentiation is negligible (genome-wide FST ~ 0).

 

See subproposals “D” and “H” before voting on this one.

 

F. Recognize the newly described Schiffornis cracrafti as a species

Lima et al. (2024) discovered one vocally and genetically diagnosable population largely restricted to the Madeira/Tapajós interfluvium in southwestern Amazonia. They described a new taxon, named cracrafti, to taxonomically recognize this population. Using a “yardstick” approach, Lima et al. (2024) inferred that cracrafti may have evolved significant reproductive barriers from all other taxa in the complex except amazonum.

 

See subproposals “C” and “I” before voting on this one.

 

G. Treat nominate turdina as a separate species from all the rest

Schiffornis t. turdina of the Atlantic rainforest of eastern Brazil was poorly sampled in previous studies on the S. turdina complex (Hellmayr 1929, Nyári 2007, Donegan et al. 2011). Using a geographically comprehensive samples of this taxon, Lima et al. (2024) found that it differs diagnostically from all other taxa in the complex by multiple song characters and exhibits substantial genome-wide differentiation. It is also the most differentiated in morphology (Lima et al. 2024), though this is likely of little relevance here. Because the levels of vocal and genetic divergence between turdina and the remaining taxa in the complex are equivalent to or greater than those between sympatric Schiffornis species, this split must be straightforward.

 

If subproposals A–G pass, we will not have to make combinations of taxa into polytypic species. But if some of those subproposals don’t pass, then we will have to deal with combinations. Here are subproposals that deal with the latter possibility:

 

H. Treat wallacii and intercedens together as a polytypic species separate from all the rest

There is no evidence for reproductive isolation between wallacii and intercedens (Lima et al. 2024). Although they differ diagnostically in song, differences are smaller than those between sympatric Schiffornis species. Therefore, we can’t assume that they would mate assortatively based on song differences if they were to come into contact. Additionally, they have shallow nuclear differentiation (genome-wide FST = 0.12). Therefore, we can’t assume that they would exhibit significant postmating reproductive barriers such as genetic incompatibilities. They have sizeable mtDNA divergence (uncorrected pairwise distance for cyt-b, ND2, and COI: 3.7–4.4%). Is that meaningful in defining biological species limits? Probably not.

 

Note that wallacii and intercedens are not sister taxa. A NO vote on this subproposal may therefore result in a nonmonophyletic species taxon, and with a strange geographic distribution. If this is considered problematic, you should vote NO on this subproposal and YES on subproposals “D” and “E.” Otherwise, vote YES on this subproposal and NO on subproposals “D” and “E.”

 

I. Treat cracrafti and amazonum together as a polytypic species separate from all the rest

There is no evidence for reproductive isolation between cracrafti and amazonum (Lima et al. 2024). Although they differ diagnostically in song, differences are smaller than those between sympatric Schiffornis species. Therefore, we can’t assume that they would mate assortatively based on song differences if they were to come into contact. Additionally, they have shallow nuclear differentiation (genome-wide FST = 0.16). Therefore, we can’t assume that they would exhibit significant postmating reproductive barriers such as genetic incompatibilities. Their mtDNA divergence, although sizeable (3.5–5%), is probably not meaningful in defining biological species limits.

 

Note that cracrafti and amazonum are not sister taxa. A NO vote on this subproposal may therefore result in a nonmonophyletic species taxon, and with a strange geographic distribution. If this is considered problematic, you should vote NO on this subproposal and YES on subproposals “C” and “F.” Otherwise, vote YES on this subproposal and NO on subproposals “C” and “F.”

 

References

Donegan, T. M., A. Quevedo, M. McMullan, and P. Salaman (2011). Revision of the status of bird species occurring or reported in Colombia 2011. Conservación Colombiana 15:4–21.

 

Hellmayr, C. E. (1929). Catalogue of birds of the Americas and the adjacent islands in Field Museum of Natural History. Volume XIII, Part VI. Field Museum Press.

 

Kirwan, G. M., and G. Green (2012). Cotingas and manakins. Christopher Helm.

 

Lima, R. D., F. Bocalini, and L. F. Silveira (2024). Integrative revision of species limits in the genus Schiffornis (Aves: Tityridae) reveals cryptic diversity in the Neotropics. Zoological Journal of the Linnean Society 200:1048–1079.

 

Nyári, Á. S. (2007). Phylogeographic patterns, molecular and vocal differentiation, and species limits in Schiffornis turdina (Aves). Molecular Phylogenetics and Evolution 44:154–164.

 

Pinto, O. M. O. (1944). Catálogo das aves do Brasil e lista dos exemplares na coleção do Departamento de Zoologia. 2° parte, ordem Passeriformes (continuação): superfamília Tyrannoidea e subordem Passeres. Departamento de Zoologia.

 

Ridgely, R. S., and G. Tudor (1994). The birds of South America. Volume 2 (The Suboscine Passerines). University of Texas Press.

 

Snow, D. W. (2004). Family Pipridae (Manakins). In Handbook of the Birds of the World (J. del Hoyo, A. Elliot and D. A. Christie, Editors). Lynx Edicions, pp. 110–169.

 

 

Rafael D. Lima, August 2024

 

 

Note from Remsen: If the proposal passes, we will need a separate proposal on English names.

 

_________________________________________________________________________________________________________

 

Vote tracking chart: https://www.museum.lsu.edu/~Remsen/SACCPropChart968-1043.htm

 

 

Comments from Remsen: “This is an excellent compilation and synthesis of everything available on the turdina complex.  What a fascinating, complex system, and a good example of conservative evolution in plumage.

 

“A. YES. Vocal differences between the two warrant treatment as separate species.  Differences in mtDNA, however, are irrelevant to species limits in allotaxa, and I think all genetic results that conflict with other data in this analysis should be treated with caution – they are mtDNA gene trees that do not necessarily reflect historical branching patterns.

“B. YES. Vocal differences combined with no evidence of free gene flow despite parapatry or sympatry essentially require these two to be treated as separate species.

“C. YES to be consistent in using voice as an indicator of species limits.

“D. NO based on lack of strong vocal differences.

“E. YES based on vocal differences.

“F. YES based on vocal differences.

“G. YES.  Easiest decision of all.

“H. YES. Tough one!  It looks like the intervening taxon cracrafti has diverged from the ancestral form more rapidly than intercedens and wallacii have, leaving them with a bunch of shared ancestral vocal characters, which poses problems for taxonomy.  So, the BSC classification here is uncomfortable if intercedens and wallacii (which has priority) are treated as conspecific.  I think that’s better than treating the two as separate species when they don’t differ in voice, the character most important for speciation in this group.

“I. NO. Difficult case, and a good example of how our typological classification scheme has difficulty with mosaic evolution.  The conclusion that they are not sister taxa based on mt DNA gene trees is premature in my opinion.”

 

Comments from Lane:

“A) A weak YES considering S. rosenbergi separate from S. veraepacis. To my ear, the voices are not strongly differentiated, but I will accept Lima et al.’s conclusions on this.

“B) YES to treating S. steinbachi as a separate species. I think the vocal differences are quite clear to me and worthy of species recognition.

“C) YES to considering S. amazonum separate from S. turdina. Again, biogeographically and vocally, this is a strong case.

“D) I agree to recognizing the taxon intercedens as separate from amazonum, but not as a separate species, so NO, I guess.

“E) This is weirdly worded. I support separating wallacii as separate from turdina, but not from amazonum. So YES to separating it from turdina, but NO to recognizing it as a species on its own.

“F) Similarly to the last, I say YES to recognizing cracrafti as a taxon and as a species apart from S. turdina, but NO to recognizing it as a species on its own. The evidence suggests to me that cracrafti is best considered part of a polytypic S. amazonum.

“G) YES. Vocally, molecularly, and phenotypically, I think the case is good for Atlantic Forest S. turdina being separated from the rest of the complex as a biological species, even if doing so results in a non-monophyletic S. amazonum.

“H) NO. In Lima et al (2024) the branching phylogeny published is a cytB-based tree, and I am reluctant to accept that branching scheme at face value (as Van seems reluctant to do). I would prefer using audio characters as my best guess to taxonomic and species-level affiliation, and to my ear these Amazonian forms are all quite similar (and in my experience, motivation, such as response to playback, can affect song structure enough that I would be reluctant to use anything short of a large sample of unsolicited song recordings to determine vocal characters). Thus, I would accept a polytypic S. amazonum that encompasses cracrafti, wallacii, intercedens, but apart from S. turdina (even at the risk of a non-monophyletic S. amazonum), which is not an option in the present set of votes.

“I) NO. See H.”

 

Additional comments from Rafael Lima: “"I personally have no objection to nonmonophyletic species taxa and am comfortable to any lumping of Amazonian populations, given that evidence for reproductive isolation is equivocal or weak for several of them. However, I want to clarify that the phylogenetic relationships mentioned in both the paper and this proposal are not based solely on mitochondrial markers. Lima et al. (2024) conducted several different phylogenetic analyses using both mitochondrial and genome-wide nuclear markers (specifically UCEs), and these analyses were largely congruent. When the proposal mentions that certain lumps may result in nonmonophyletic species taxa, this includes nonmonophyly across all phylogenetic analyses performed. For example, the conclusion that cracrafti and amazonum are not sister taxa is based on concatenated and coalescent analyses of UCEs as well as mtDNA data. Also, a polytypic S. amazonum encompassing cracrafti, wallacii, and intercedens, but excluding S. turdina, as Lane suggests, would form a nonmonophyletic taxon based on genome-wide markers (see Supplementary Figures S13–15 in Lima et al. 2024). I am comfortable with such a taxon, though."

 

Comments from Zimmer: “

“A) YES.  Based upon vocal differences.  Also, it strikes me as an odd distribution that rosenbergi would be conspecific with polytypic populations in Central America (ssp. veraepacis, dumicola, acrolophites) that are separated from an isolated, monotypic rosenbergi in South America by the range of a different species (S. stenorhyncha), that is sympatric with veraepacis.

“B) YES to treating steinbachi as distinct both from S. turdina, and, as distinct from intercedens, based on differences in multiple song characters and lack of evidence of free gene flow between steinbachi and intercedens despite the two existing either in parapatry, or possibly sympatry.

“C) YES.  Based on vocal differences and biogeographic considerations, this seems straightforward.

“D) This is where I’m starting to get confused as to what I am actually voting on, and, in looking at Van’s and Dan’s comments, it looks like we are all interpreting the question differently.  This subproposal reads:  “Treat intercedens as a separate species from S. turdina.”  Taken at face value, I would vote YES on recognizing intercedens as being distinct from nominate turdina (See Subproposal G – which is to treat turdina as distinct from everything else.), based upon “significant vocal and genome-wide differentiation in relation to all other taxa in the complex except wallacii and cracrafti.” (Lima, in the Proposal).  So, my YES vote is only to treat intercedens as separate from nominate turdina, not as separate from wallacii or cracrafti.  Implicit to me was recognition/reinstatement of intercedens as a valid taxon at some level (Lima et al. 2024), rather than retaining it as a junior synonym of amazonum.  However, I see that Van has voted “No”, citing “lack of strong vocal differences”, which would be the case if we were considering the question of recognizing intercedens as a species distinct from wallacii and cracrafti, but not if the question is one of whether to treat intercedens as distinct from turdina, which is how Subproposal D is worded.  Meanwhile, Dan agreed with recognizing intercedens as a taxon (presumably as a subspecies) distinct from amazonum, but not “as a species on its own”  -- in other words, I am interpreting Dan’s comments to mean that he agrees that intercedens is a valid taxon, but that he doesn’t consider it specifically distinct from amazonum.  But, again, the wording of Subproposal D is asking whether we should consider intercedens as distinct from turdina, not from amazonum.  So, Van and Dan seem to be basing their NO votes on the latter question, whereas my YES vote is based upon the former question.  Dan says as much in his responses to both Part E & F.

“E) As in my comments on Part D, I vote YES to treating wallacii as specifically separate from S. turdina, but NO to treating it as distinct from intercedens, cracrafti, and amazonum, based upon minimal vocal differences.  I would vote YES to treating these 4 Amazonian taxa as part of a polytypic S. amazonum, that is distinct from all other members of the complex, but that doesn’t seem to be one of the voting options in this Proposal.

“F) As in my comments on Parts D & E, I vote YES to treating cracrafti as specifically separate from S. turdina, but NO to treating it as distinct from intercedens, wallacii, and amazonum, based upon minimal vocal differences.  I would vote YES to treating these 4 Amazonian taxa as part of a polytypic S. amazonum, that is distinct from all other members of the complex, but that doesn’t seem to be one of the voting options in this Proposal.

“G) YES.  This one seems particularly straightforward, on the basis of differences in multiple song characters, genetic differences, morphological differences (although, I agree, that morphology is probably not overly relevant in this complex), and from the standpoint of biogeography.

“H) NO.  According to the Proposal’s summary of Lima et al. 2024, intercedens “exhibits significant vocal and genome-wide differentiation in relation to all other taxa in the S. turdina complex except wallacii and cracrafti.”  And “Lima et al. (2024) found sizeable vocal and genome-wide differentiation between wallacii and all other taxa in the S. turdina complex except cracrafti and intercedens.”  So, there are no sizeable vocal differences between these 3 taxa, and therefore, I don’t see the basis, other than the genetic data, for treating wallacii and intercedens together as a species that does not include cracrafti and amazonum.

“I) NO. According to the Proposal’s summary of Lima et al. 2024, “There is no evidence for reproductive isolation between cracrafti and amazonum(Lima et al. 2024). Although they differ diagnostically in song, differences are smaller than those between sympatric Schiffornis species.”  So, if the vocal differences between cracrafti and wallacii & intercedens are unremarkable, and there is no evidence for reproductive isolation between cracrafti and amazonum, which also do not differ strongly from one another vocally, then I see no strong reason for separating the latter two taxa from wallacii + intercedens.  Again, my YES vote would be for a polytypic S. amazonum (amazonum + cracrafti + intercedens + wallacii) as separate from everything else and based upon lack of yardstick-level vocal differentiation between the 4 taxa involved.”

 

Comments from Areta: “I´ve been struggling to integrate the vocal and genetic (mtDNA and UCE datasets) in order to have a firmer grasp of the taxonomy of these Schiffornises. My final take is like that of Kevin and Dan. Although I am conflicted by the non-monophyly of the polytypic S. amazonus (including intercedens, cracrafti, and wallaci), and by the "leapfrog" pattern of taxa with 3 (amazonus and cracrafti) and 4 (intercedens and wallaci) notes, there are enough uncertainties in the phylogenetic relationships suggested by the different DNA sources to cast doubt on what the true relationships are here. Keeping a polytypic amazonus seems the conservative stance, and while the vocal differences seem consistent, it is not exactly clear what do they mean from the perspective of interbreeding, given also how similar these vocalizations sound.

 

“I am also somewhat conflicted by the split of the sister, little diverged, and vocally similar rosenbergi from veraepacis, mostly because of the apparently fairly widespread vocal evidence of hybridization in Panama between the phylogenetically distant and vocally divergent veraepacis and stenorhyncha. I wonder whether these vocally intermediate individuals are really hybrids.

 

“It seems to me that things are more fluid in Schiffornis than what discrete categories might be able to describe, and further genomic work will surely come up with cool and taxonomically useful information.”

 

“A. YES, but quite conflicted on this one: the more I think on this, the less convinced I am of splitting veraepacis from rosenbergi (see comments above)

B. YES

C. YES, but see D, E, F, H, and I to understand what should be included in S. amazonum (and see comments above; no solution pleases me fully).

D. YES to treating intercedens as separate from turdina, but no to S. intercedens as a different species level taxon. Treat intercedens as part of a polytypic amazonum.

E. YES to treating wallacii as separate from turdina, but no to S. wallacii as a different species level taxon. Treat wallacii as part of a polytypic amazonum.

F. YES to treating cracrafti as separate from turdina, but no to S. cracrafti as a different species level taxon. Treat cracrafti as part of a polytypic amazonum.

G. YES to a monotypic S. turdina.

H. NO, treat wallacii, intercedens, cracrafti, and amazonum as part of a polytypic amazonum.

I. NO, treat wallacii, intercedens, cracrafti, and amazonum as part of a polytypic amazonum.”

 

Comments from Stiles: “Schiffornis splits: the problem here is that the wording is confusing or too vague in some parts.

A: split rosenbergi from veraepacis- YES

B. split steinbachi as a species from turdinus - YES (interesting that this species takes off as a foothill species south of where (after a gap) aenea left off.

C. split amazonum from turdinus as a species (?):YES. Here starts the paring off of various taxa from turdinus.

D. recognize taxon intercedens and split from turdinus.

E. ditto for cracrafti

F. ditto for wallacei (although suggested that all should be called species, this seems to have confused some because all were later demoted to subspecies; perhaps here simply name them all as taxa.

G. Split turdina from everyone else (as it has been left monotypic).

H: combine wallacii and intercedens as a species, and

I: ditto cracrafti and amazonum - both based on similarity of vocalizations but both conflicting with (mtDNA) phylogeny and (at least in part) with distributions. More doubts - but the beginning of a possible consensus? Dan, Kevin and Nacho have  suggested that all taxa in H and I be combined in a single polytypic species (amazonum has priority). This also seems to me the best solution, This was not presented as an option in the original proposal, but it could easily be accommodated as option J (which also appears be acceptable to Rafael).

 

Comments from Robbins: “I’ve gone through this proposal twice, about a month apart.  Based on the vocal and limited genetic data (we know the issues associated with mitochondrial data), it isn’t clear what is the best course of action for some of these subproposals.  However, in the end, I too agree with what Kevin and Dan have concluded. Thus, I vote as follows, primarily based on what has already been stated.

 

“A. A “lukewarm” endorsement of voting YES, as the vocalizations are quite similar.

B. YES, the vocal differences are clearly different.

C. YES, this is one of the more clearcut cases.

D. Kevin makes a good point of what are we actually voting on in this proposal. I agree with him, if the question is treating intercedens separate from turdina, then I vote YES. But, look how I vote in E & F.

E. YES. Kevin has clearly laid out my thoughts on this.

F. YES. Ditto to what Kevin states.

G. YES, song and genetic data makes this, like B & C, a straightforward decision.

H. NO for reasons that have been stated by other members.

I. NO for reasons that have been stated by other members.”

 

Comments from Naka: “This case is ​indeed quite complex. I have spent ​over a week thinking how to best solve the taxonomic mess. ​I have gone from many yeas to many nays in my votes. Unfortunately, voice, DNA, and distribution do not seem to tell a clear story, ​and morphology isn’t offering many tips. Therefore, I am having a tough time having to decide whether DNA or voice or distribution makes more sense in this case. I commend the authors ​of the proposal for presenting ​such a rich data​set in their article (+​ Supp. ​material), but the data presented in ​this proposal is limited to fully understand the case.​ I wish trees, structure plots, and spectrograms would have been included here, as I had to go back and forth many times. Unfortunately, I think the wording of the proposals and subproposals is not ideal and somewhat confusing at times.

“Despite the wording of the proposal, I think that for the main problem (S. amazonum and turdina) we have four alternative paths. I have included all four below, which I ordered by my own increasing preference. I think path 1 is worse than 2, 2 worse than 3, and 3 worse than 4.

“I will offer my votes, which are consistent with my view that the path 4 (see below) is the most consistent and less problematic.

“Path 1. Virtually every vote cast seems to be happy creating a paraphyletic clade, where a polytypic S. amazonum includes cracrafti + intercedens + wallacii, excluding S. turdina, as a monotypic taxon. The problem I see with this proposal is that not only champions a paraphyletic taxon, but by doing so, it creates an Amazonian Frankenstein in terms of vocalizations. I would have no problem in having a non-monophyletic lineage excluding turdina, IF turdina would have been an outlier of some sort. But it happens that not only wallacii seems to be its closest relative (both in mt and nDNA), and share high admixture in the nuclear genome, but also has the same vocal type with four notes and relatively similar shaped notes (see types C and E in Lima's Figure 3).

 

“Path 2. Another option is to accept an Amazonian/Atlantic Forest monophyletic clade with considerable vocal variation. This taxon would include turdina + ​wallac​ii + cracrafti + amazonum + intercedens + steinbachi within an expanded polytypic S. turdina. I don’t think this is ideal but looks better to me than path 1.

 

“Path 3. We could also accept two monophyletic clades, where S. turdina includes wallac​ii + cracrafti, and S. amazonum includes intercedens + steinbachi. The problem is that turdina's vocal type with four notes is shared by turdina, wallacii, and intercedens, but not cracrafti. Therefore, the leapfrog pattern mentioned by some in their votes.

 

“Path 4. Finally, I see an option that does not require crossing any red lines. This includes a) having a polytypic S. turdina + wallacii (which by the way are sister taxa, highly admixed, and similar sounding); b) a new species cracrafti, which is a bit problematic because it sounds a lot like amazonum; and c) a monophyletic S. amazonum with intercedens + steinbachi.

 

“This last treatment may solve the taxonomy without creating major disruptions. 

 

​“According to this treatment, I provide my votes below:

 

“A. Treat rosenbergi as a separate species from S. veraepacis: NO. The two taxa (rose​nbergi and veraepacis​) are ​each other's closest relatives and, ​according to Lima's Figure 3​ do not seem to show such a strong vocal variation​, I am happy to go with ​two allopatric populations, shared by the Chocó and C. America. This would not be the first such distribution, and I think is quite interesting and educative in terms of understanding these two regions’ biogeography. I admit that ​the presence of steinbachi​ makes this story quite challenging​, though. However, I don’t know any of these taxa, so I am happy to change this vote under pressure.

 

“B. Treat steinbachi as a separate species from S. turdina: YES, I am happy to treat steinbachi as a separate species from S. turdina​, but I don’t think this is the point here. I agree that steinbachi​ should be included within polytypic S. amazonum.​

 

“C. Treat amazonum as a separate species from S. turdina: YES, I am happy to consider them distinct species, as long as wallacii and potentially cracrafti fall within its closest relative: S. turdina.

 

“D. Treat intercedens as a separate species from S. turdina: YES, despite its vocal similarity, I think intercedens should be split form S. turdina but included within S. amazonum.

 

“E. Treat wallacii as a separate species from S. turdina: NO. In Lima’s mtDNA tree (Figure 8), wallac​ii forms a well-supported clade with turdina and cracrafti, and this close relationship remains close in the SNPs PCA (Fig. 9). Vocally speaking, the songs of ​wallac​ii and turdina seem quite close, with four notes and similar shapes. Biogeographically, other groups show the same close relationship between Easternmost Amazonian populations (Belem area) and the Atlantic Forest taxa, which seems to be the case here. Quite interestingly, Limas’s figure 9, shows a very high admixed populations between the Atlantic Forest samples and those from Belem.

 

“F. Recognize the newly described Schiffornis cracrafti as a species: A tentative YES. I don’t like the idea of keeping this taxon with closely related, yet different sounding turdina + wallacii, but at the same time, its voice looks surprisingly similar to that of S. amazonum.

 

“G. Treat nominate turdina as a separate species from all the rest: NO. As I mentioned before, I think this taxon should include at least wallacii, from the Belem Area of endemism.

 

“H. Treat wallacii and intercedens together as a polytypic species separate from all the rest: NO. I don’t like the idea of creating this paraphyletic taxon, solely based on song and excluding the genetic data available.

 

“I. Treat cracrafti and amazonum together as a polytypic species separate from all the rest: NO. I don’t like the idea of creating this paraphyletic taxon, solely based on song and excluding the genetic data available.”