Proposal (1019) to South American Classification Committee

Treat Granatellus paraensis as a separate species from G. pelzelni

Note: This is a high-priority issue for WGAC.

Background: Our new SACC note on this is as follows:

34a. Del Hoyo & Collar (2016) treated the subspecies paraensis of se. Amazonia as a separate species (Rose-bellied Chat), based in part on vocal differences from nominate pelzelni described by Boesman (2016n).

Here is the basic set-up. Granatellus pelzelni is considered to consist of two subspecies:

• nominate pelzelni: largely at the periphery of the Amazon Basin from e. Colombia, c. Venezuela, the Guianas, ne. Brazil and then S of the Amazon River in central Brazil, W of the Rio Tocantins, to e. Bolivia

• paraensis: E of the Rio Tocantins in extreme e. Brazil.

They differ in the extent of white on the flanks and the amount of black in the face and crown. Here’s the plate by David Quinn from HBW:

The peculiar and patchy geographic distribution is best understood visually – here is the eBird map, onto which I placed a blue line to roughly represent the Tocantins.

Where it gets fuzzy for me because I am not familiar with the area is which localities go with which taxon in the upper Tocantins, Also, the possibility seems strong to me, again keeping in mind I am unfamiliar with the region, that there could well be a contact zone in the headwaters of the Tocantins-Araguaia river basin in southern Tocantins and northern Goias that would be extremely informative.

New information:

Del Hoyo & Collar (2014) treated as a separate species based on the Tobias et al. point scheme as follows:

"Rose-breasted Chat (Rose-bellied)

“Hitherto treated as conspecific with G. pelzelni, but differs in its grey vs black mid-crown and ear-coverts (2); lack of white flanks (pink extending further onto them or perhaps more grey extending out from inner flanks) (2); distinctive song, a variable, rather rhythmic phrase alternating one or more mellow low-pitched down-slurred whistles and harsher short high-pitched notes, vs a series of repeated rather similar notes, sometimes ending with a different trilled part or two series of repeated notes, to be scored as fewer repeated notes (2), shorter song phrases (ns[2]), different note shapes (ns[1]) and mellow notes having a much narrower frequency range (2); and possible absence (not yet recorded) of the typical “jrt” call of G. pelzelni (ns[2]) (2). Monotypic.”

The vocal data come from Boesman (2016n), who stated:

“A comparison of song (illustrated with multiple sonograms in the pdf version of this note): nominate (in the Guianas, N of Amazon), nominate (S of Amazon) and paraensis (n=4).

“There is a striking difference in song between paraensis and nominate! Nominate in all regions (N and S of Amazon) has a song consisting of a series of repeated notes which are all rather similar-shaped (sometimes ending with a different trilled part or 2 series of repeated notes). Typically 6-15 repeated notes. Phrase length 1.8 - 3s. Notes have a freq. range of c. 0.7 - 3kHz. Paraensis seemingly has a song which is a variable, rather rhythmic phrase with alternating a mellow low-pitched down-slurred whistle (or a short series of these) and harsher short high-pitched notes. Repeats 1 - 6. Phrase length 0.5 - 1.5s. Mellow notes have a narrow frequency range of c. 0.7 - 1.1kHz.

The vocal difference can be scored based on paraensis having less repeated notes (score 2-3), shorter song phrases (score 2-3) and different note shapes (1-2) , with mellow notes having a much narrower freq. range (2), which leads to a total vocal score of about 5. This obviously is based on only 4 examples of paraensis, and thus needs further confirmation, but nevertheless seems quite convincing.

With just N=4 in Boesman’s analysis of paraensis, I would object in principle to making any taxonomic conclusion on these putative song differences. In my opinion, Boesman’s “needs further confirmation” statement is correct. Just from checking recordings in Macaulay, I noticed this one:

https://macaulaylibrary.org/asset/220845741 (by Ciro Albano Tocantins, from the Rio Javaés, a right bank tributary of the R. Araguaia, which itself is a left bank tributary of the Tocantins. This is classic pelzelni song with its evenly spaced set of upward-inflected notes. It is west of the Tocantins itself, so if the Tocantins is the barrier, then it fits that pattern, but the Araguaia is actually bigger than the Tocantins, so again I want to know (1) is it the Tocantins itself that is the barrier or the Tocantins drainage? And (2) what did this individual look like. It is identified as paraensis in Macaulay, but the basis of this uncertain.

For reference, here’s one that fits Boesman’s characterization of paraensis (by Alex Aleixo, Pará at Municipio de Tomé-Açu) from Pará, east of the Tocantins:

https://macaulaylibrary.org/asset/219437

New genetic data: I had originally overlooked the published genetic data on Granatellus in this region. Alex Aleixo alerted me to his paper (Dornas et al. 2022), which included the two taxa of Granatellus in their analysis of genetic variation in the Tocantins drainage. See Areta comments below for full citation and description of the data There is a strong genetic break between the two taxa, and the Tocantins is the barrier, with a 5.3% sequence divergence. But before anyone gets excited about that, here goes more than my usual diatribe on genetic distances as a metric for taxon rank.

Below are comparisons of genetic distance (ND2) between 62 taxon pairs that are in contact or approach each other in the Rio Branco region studied by Luciano Naka for his dissertation here at LSU. This remains as far as I know the largest comparative data set of its kind, and note that because they were all done by the same methodology using the same genetic markers on “sedentary” tropical birds, mostly forest-dwellers, this controls for several potential sources of variation. I am going to keep using this table every time genetic distance is used in our assessments of species rank. Note also the within-family variation – I see no evidence for a phylogenetic effect. To no one’s surprise, taxa we rank as species on average show greater between-taxon genetic distance. That is not in dispute. But note that the ranges overlap widely. Just within this sample, if forced to use genetic distance alone as a metric, any GD measure between 2.0 and 13.0% cannot be used as a diagnosis for species rank. What if the sample were expanded beyond sedentary Amazonian birds? Certainly, those ranges would expand in both directions, but especially downwards, below 2.0%. Some degree of taxonomic circularity may be biasing this analysis, and Naka et al. rightfully advocated for closer examination of some of the taxon pairs currently ranked as subspecies. Nevertheless, at this point, the general signal from genetic distance data is that it cannot be used as a means of assigning species vs. subspecies rank, as emphasized by Naka et al.

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Data from Naka et al., 2012, Am. Nat. 179: E115-E132, Table1)
Taxon A	Taxon B	rank	Gdist	species	subspecies	species + SU/SP	subspecies - SU/SP
Hylophilus ochraceiceps luteifrons	H. o. ferrugineifrons	SU/SP	13.3		13.3	13.3
Zimmerius acer	Zimmerius gracilipes	SP	12.9	12.9		12.9
Pheugopedius c. coraya	P. c. griseipectus	SU	12.9		12.9		12.9
Tyranneutes virescens	Tyranneutes stolzmanni	SP	11.3	11.3		11.3
Hylophilus muscicapinus	Hylophilus hypoxanthus	SP	11.3	11.3		11.3
Phaethornis s. superciliosus	P.s. insolitus	SU	10.8		10.8		10.8
Willisornis p. poecilinotus	W. p. duidae	SU	10.8		10.8		10.8
Microcerculus b. bambla	M. b. albigularis-caurensis	SU	10.1		10.1		10.1
Schiffornis t. turdinus	S. t. amazonus	SU/SP	9.7		9.7	9.7
Myrmotherula guttata	Myrmotherula hauxwelli	SP	11.3	11.3		11.3
Hypocnemis cantator	Hypocnemis flavescens	SP	9.0	9.0		9.0
Iodopleura fusca	Iodopleura isabellae	SP	8.8	8.8		8.8
Glyphorynchus s. spirurus	G. s. rufigularis	SU	8.5		8.5		8.5
Veniliornis cassini	Veniliornis affinis	SP	8.5	8.5		8.5
Monasa atra	M. morphoeus	SP	8.5	8.5		8.5
Onychorhynchus c. coronatus	O. c. castelnaui	SU/SP	8.2		8.2	8.2
Galbula d. dea	G. d. brunniceps	SU	8.2		8.2		8.2
Terenotriccus e. erythrurus	T. e. venezuelensis	SU	8.2		8.2		8.2
Pyrilia caica	Pyrilia barrabandi	SP	7.8	7.8		7.8
Galbula a, albirostris	G. a. chalcocephala	SU	7.7		7.7		7.7
Capito niger	Capito auratus	SP	7.7	7.7		7.7
Topaza pella	Topaza pyra	SP	7.6	7.6		7.6
Epinecrophylla gutturalis	E. haematonota	SP	7.4	7.4		7.4
Euphonia cayennensis	Euphonia rufiventris	SP	6.9	6.9		6.9
Lepidocolaptes a. albolineatus	L. a. duidae	SU	6.9		6.9		6.9
Dendrocincla m. merula	D. m. bartletti	SU	6.4		6.4		6.4
Xiphorhynchus guttatus polystictus	X. g. guttatoides	SU	6.4		6.4		6.4
Xenops minutus ruficaudus	X. m. remoratus	SU	6.3		6.3		6.3
Dendrocincla f. fuliginosa	D. f. phaeochroa	SU	6.1		6.1		6.1
Thalurania f. furcata	T. f. nigrofasciata	SU	5.7		5.7		5.7
Percnostola r. rufifrons	P. r. minor	SU	5.4		5.4		5.4
Tolmomyias assimilis neglectus	T. a. examinatus	SU	5.3		5.3		5.3
Xiphorhynchus pardalotus	Xiphorhynchus ocellatus	SP	5.3	5.3		5.3
Hemithraupis f. flavicollis	H. f. aurigularis	SU	5.0		5.0		5.0
Thamnophilus amazonicus divaricatus	T. a. cinereiceps	SU	4.3		4.3		4.3
Automolus infuscatus cervicalis	A. i. badius	SU	3.7		3.7		3.7
Microbates collaris torquatus	M. c. collaris	SU	3.0		3.0		3.0
Tolmomyias poliocephalus sclateri-klagesi	T. p. poliocephalus	SU	3.0		3.0		3.0
Myrmothera c. campanisona	M. c. dissors	SU	2.8		2.8		2.8
Pteroglossus aracari	Pteroglossus pluricinctus	SP	2.8	2.8		2.8
Cymbilaimus l. lineatus	C. I. intermedius	SU	2.7		2.7		2.7
Myrmoborus myotherinus ardesiacus	M. m. elegans	SU	2.7		2.7		2.7
Momotus m, momota	M. m. microstephanus	SU	2.3		2.3		2.3
Gymnopithys rufigula	Gymnopithys leucaspis	SP	2.3	2.3		2.3
Dendrocolaptes c. certhia	D. c. radiolatus	SU	2.1		2.1		2.1
Brotogeris chrysoptera	Brotogeris cyanoptera	SP	2.0	2.0		2.0
Piculus f. flavigula	P. f. magnus	SU	2.0		2.0		2.0
Celeus u. undatus	Celeus u. grammicus	SU	2.0		2.0		2.0
Thamnophilus murinus cayennensis	T. m. murinus	SU	2.0		2.0		2.0
Cercomacra cinerascens immaculata	C. c. cinerascens	SU	2.0		2.0		2.0
Schistocichla l. leucostigma	S. 1. infuscata	SU	1.9		1.9		1.9
Trogon r. rufus	T. r. sulphureus	SU	1.9		1.9		1.9
Cercomacra tyrannina saturatior	C. t. tyrannina	SU	1.9		1.9		1.9
Synallaxis rutilans dissors	S. r. confinis	SU	1.7		1.7		1.7
Piprites c. chloris	P. c. tschudii	SU	1.7		1.7		1.7
Sittasomus griseicapillus axillaris	S. g. amazonus	SU	1.6		1.6		1.6
Tachyphonus s. surinamus	T. s. brevipes	SU	1.5		1.5		1.5
Pithys a. albifrons	P. a. peruvianus	SU	1.3		1.3		1.3
Hemitriccus zosterops rothschildi	H. z. zosterops	SU	1.2		1.2		1.2
Myrmotherula a. axillaris	M. a. melaena	SU	1.2		1.2		1.2
Celeus t. torquatus	C. t. occidentalis	SU	1.1		1.1		1.1
Celeus e. elegans	C. e. jumanus	SU	0.8		0.8		0.8

mean corrected genetic distance

5.6

7.7

4.8

8.1

4.4

	range	N
species (SP)	2.0-12.9	17
subspecies (SU)	0.8-13.3	45
species incl. possible splits (SU/SP)	2.0-13.3	20
subspecies excl. possible splits (SU/SP)	0.8-12.9	42
SU/SP = taxa currently treated as subspecies but under active investigation

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

In the case of these two Granatellus taxa, a genetic distance of 5.3% falls very close the mean genetic distance of all taxa in Naka et al.’s analysis and cannot be considered, in my opinion, relevant to taxon rank.

Discussion and Recommendation:

Here is an expanded view of the David Quinn plate so that you have a broader context for comparisons of plumage among species in Granatellus:

The basic plumage pattern in the genus is fairly conservative, with the differences between paraensis and pelzelni (extent of white on flanks, cheek color, and superciliary strength) also reflected in differences among the other two species. As you can see, paraensis is the extreme in terms of reduction of white flanks but otherwise does not stand out.

I don’t have a strong recommendation on this. I will vote NO for now because of the small N of songs of paraensis and Boesman’s cautionary statement. The genetic data certainly suggest a barrier to gene flow. Terra firme taxa ranked as subspecies frequently show that level of genetic distance across major rivers, although I did not think G. pelzelni was a species for which a river would be a barrier, likely because of my ignorance of habitat preference. Regardless, this seems to me to be a case in which detailed characterization of phenotype distribution in the region of potential contact is critical, especially towards the upper Tocantins. I suspect that more detailed mapping of vocal and plumage characters will confirm species rank for these two, but I feel uneasy about endorsing this without those data.

English names: BirdLife International used Rose-bellied Chat for paraensis and retained parental Rose-breasted Chat for pelzelni s.s. Nominate pelzelni is considerably more widely distributed, and so retention of the parental name with that taxon is in line with our SACC guidelines on English names. The similarity between the names might be confusing to some but helpful in others in linking the two. The problem is that BOTH taxa have rose breasts and rose bellies, and so I suspect many people would have a problem remembering which is which. I suggest that “Tocantins Chat” has some merit for paraensis because it refers to its center of distribution in terms of the river and to a lesser extent to the Brazilian state; also, this part of Amazonia has undergone major deforestation and so the toponym might have some merit in terms of promoting conservation. On the other hand, if pelzelni replaces paraensis on the west bank of the Tocantins, the attractiveness of that name is diminished. In my opinion, we should do a separate proposal on the English name if the proposal passes.

References: (see SACC Bibliography for standard references)

Boesman, P. (2016). Notes on the vocalizations of Rose-breasted Chat (Granatellus pelzelni). HBW Alive Ornithological Note 384. In: Handbook of the Birds of the World Alive. Lynx Edicions, Barcelona. https://doi.org/10.2173/bow-on.100384

Van Remsen, July 2024

Comments from Areta: “I vote NO to the split. Although there is a reasonable case for the split based on minor plumage distinctions and deep mtDNA differentiation, there is conflicting vocal evidence that I would like to see sorted before changing our taxonomy.

“Dornas et al. (2022) sequenced 22 pelzelni (including the Guianan population) and 2 paraensis. Terry Chesser downloaded 4 and 2 sequences and found that mtDNA distance was 5.3%.

“Here is Dornas et al´s figure 2D:

“And their supplementary Figure S12: ‘Supplementary Figure 12. Phylogenetic tree (A) obtained by Bayesian Inference (BI) for Granatellus pelzelni accompanied by the respective haplotype network (B). Dots represent sampling localities within our target area, with different colors denoting phylogenetic/haplotype relationships among sampled individuals (C). From 24 sequenced individuals (Table S2), a total of 14 haplotypes for the mitochondrial marker ND2 in 1033 bp fragments with the following base frequencies were retrieved: A = 0.3035, C = 0.3643, G = 0.0991, T = 0.2332. As outgroup, sequences from Granatellus sallaei (GB: EF529922.1), G. venustus (GB: EF529921.1) and Pheucticus ludovicianus (GB: AF290108.1) were used. A total of 81 sites with nucleotide substitutions were identified. The evolutionary model selected for BI by JMODELTEST was TRN+I (nst=6, rates=equal, pinvar=0.6100). CEB = Belém Area of Endemism; CEX = Xingu Area of Endemism; and ITA = Tocantins Araguaia Interfluve. Populations of Granatellus pelzelni presented strong phylogeographic structure, with the occurrence of three independent lineages: G. p. paraensis (BAE), G. p. pelzelni (TAI and XAE and westward), and a third and apparently unnamed lineage on the Guiana Area of Endemism (GAE). An asterisk (*) denote the photo ́s authorship along its accession number in the WikiAves (wikiaves.com; WA) citizen science portal.’ “

Now, I am confused by some recordings from Maranhão, where supposedly only paraensis should be present. A recordist attached sex to two birds that were singing together (and based on his comments, he seems quite familiar with the taxon paraensis, to the point that he indicates that the females often give this song which males never do). According to this, then the male paraensis would have a song that is overall similar in structure to that of pelzelni, and not radically different as Boesman (2016) indicated. However, I cannot rule out that the recordings might have been misidentified, as most other recordings of paraensis do follow Boesman´s description (certainly, all the ones in XC do so, as well as those by Aleixo in the ML):

Male
https://macaulaylibrary.org/asset/194395401

[sounds quite similar to this one, from Pará, west of the Tocantins river https://macaulaylibrary.org/asset/375366741]

Female
https://macaulaylibrary.org/asset/194393221

Other recordings from eastern Pará:

Female?
https://macaulaylibrary.org/asset/214418071

Male?
https://macaulaylibrary.org/asset/194197101

“If this interpretation is correct (I have seen Granatellus pelzelni a few times, and maybe recorded it in SE Venezuela ages ago, but I do not claim having a deep knowledge of this taxon), then the vocal differences need to be reassessed, as clearly it is a weak basis to compare voices of males of one taxon to females of the other one (although maybe only females of paraensis give this multi-noted song?).

“The genetic break seems quite deep and abrupt, but I would like to have more clarity on the situation with vocalizations. There is a possibility that (vocally) Boesman might have compared pears to apples. This does not explain the deep genetic break, but (if correct) seriously undermines the vocal arguments. paraensis and pelzelni are presumably allo/parapatric, but if one takes Boesman´s analyses as such, then the recordings I point out above would indicate sympatry (alternatively, one can interpret that the vocal differences are not as described and that both taxa have structurally similar songs)..

“Dornas T, Dantas SM, Araújo-Silva LE, Morais F and Aleixo A (2022) Comparative Phylogeography of Birds Across the Tocantins–Araguaia Interfluve Reveals a New Biogeographic Suture in the Amazon Far East. Front. Ecol. Evol. 10:826394. doi: 10.3389/fevo.2022.826394”

Comments from Robbins: “NO. Although the mitochondrial data suggest two species, as Nacho points out there seems to be confusion on vocal comparisons. The plumage differences may not be significant, note the differences between nominate Granatellus venustus and G. v. francescae, where the differences between those two are greater than those between the two pelzelni taxa.

“Although I'm on the fence on this one, I would like to see clarification on the vocalizations before voting Yes for recognizing paraensis as a species.”

Comments from Stiles: “NO for now, for about the same reasons as for the Xenodacnis split, we need a more thorough vocal analysis with definite data for the locations of the birds recorded for a final decision. The plumage differences, to my eye, are of the magnitude that better characterizes subspecies rather than species.”

Comments from Del-Rio: “YES. 5.3% mitochondrial divergence has to mean something... I do not oppose "NO" votes... I agree that more data are needed, but it's quite hard to find this level of mitochondrial divergence in populations with similar geographic ranges...”

Comments from Lane: “NO. From the information provided in others’ comments above, it seems that separating paraensis from pelzelni is poorly supported by existing evidence. I would much rather wait for better evidence than act on what little is available at present.”

Comments from Claramunt: YES. The birds look similar, but all Granatellus species look similar. The fact is that paraensis differs from pelzelni in multiple plumage traits. Songs seem to differ too, and mtDNA, although the sample size is small, suggests substantial divergence.”

Comments from Zimmer: “NO. I suspect something interesting might be going on here, given the mitochondrial divergence and the plumage differences, which, although relatively slight, are about on par with plumage divergence in this genus, the members of which are mostly just small variations on a common theme. However, I don’t think we have enough to go on at this point to draw any definite conclusions. Reading this Proposal, and listening to the vocal sample (ML219437, from E of the Tocantins, by Alex Aleixo) that Van included in the Proposal, stirred something from the distant past in my brain. In August of 2007, Andy Whittaker and I spent a week surveying birds based out of a logging camp in the Rio Capím region (E of the Tocantins). I remembered encountering Rose-breasted Chat there, and was pretty sure that I had tape-recorded it. I just went back through my tape logs from that trip, and a quick perusal turned up only a single recording of Granatellus, which came at the end of a long sequence that began with Formicivora grisea, and then switched to various species calling overhead from a mixed-species, mid-level flock. At some point, with the tape still running, I switched onto some repeated single-note calls, that may have been from a Granatellus, but if so, sounded somewhat different from what I am used to from nominate pelzelni. In any case, I then locked onto several songs that are almost a perfect match for the Aleixo recording of paraensis, and very similar to the ones labeled as “Female” and “Female?” that Nacho shared from eastern Pará. I would characterize these short, but variable (compared to typical repetitive series given by nominate pelzelni) songs as suggesting a sweeter sounding Dusky-capped Greenlet or the phrases of some vireo. Anyway, after recording several of these songs, there is an obvious break in my recording, without a voice annotation, followed by a long series of very excited sounding Granetellus songs that are pretty typical to my ears, of songs given by nominate pelzelni. I recorded the Granatellus right down to the end of Side B on the cassette, and barely squeezed in an annotation of “Songs of Granatellus pelzelni, preceded by Formicivora grisea” before the tape abruptly ran out. In listening to the recording, it’s pretty obvious to me that I switched off of the Formicivora and onto the flock passing overhead, when I heard the initial short, variable, greenlet-like songs, which would have struck me as unfamiliar. After recording several of those, there was the quick, but obvious break in the recording, without an annotation, which suggests that after taping several of the mystery voices, I tried playback, which resulted in an aggressive and persistent bout of amped-up songs from the male Granatellus, and those continued to the end of the tape. I seem to remember having a pair of Granatellus there, but can’t say for certain. But if the recordist that Nacho quoted is correct, and that females of paraensis often give the short song, but that males never do, then I probably recorded spontaneous songs of the female of the pair, and when I played those back, it elicited the more typical sounding pelzelni-like songs from the male. All of which means nothing, other than calling into question Boesman’s (2016) assertion of striking differences in song between the two taxa. Now, if no such female song exists in nominate pelzelni (and I certainly can’t remember ever having heard or taped this song anywhere else), then that does suggest that there are some vocal differences. But the fact that the male songs are so similar, at least in my limited sample, does suggest that we simply don’t know enough about this particular situation to justify a split.”