Proposal (1018) to South American Classification Committee

Treat Xenodacnis petersi as a separate species from X. parina

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

Background: Our SACC note on this is as follows:

51a. The northern petersi subspecies group was formerly (e.g., Bond & Meyer de Schauensee 1939, Zimmer 1942) considered a separate species from Xenodacnis parina. All subsequent classifications since Meyer de Schauensee (1970) and Storer (1970a) have treated them as conspecific. Del Hoyo and Collar (2016) treated it as a separate species, with support from vocal differences found by Boesman (2016m). Aguilar (2021) found that the Ecuadorean population of X. p. bella was significantly larger in body size and had a different habitat preference (Aguilar 2019) than all other taxa.

To set the stage for this one, here is the map from Aguilar (2021):

Long treated as a monotypic species (e. g. Hellmayr 1935), a paucity of specimens of this patchily distributed Polylepis-zone specialist likely prevented earlier authors from noting the fairly pronounced geographic variation that was noted by Bond & Meyer de Schauensee (1939), who described petersi as a separate species from X. parina, and in the same publication described bella as a subspecies of petersi. Zimmer (1942) also treated petersi as a separate species. Zimmer only had 1 petersi to examine but he noted:

“The differences from X. parina are not very pronounced, except for the considerably greater size of petersi. The bright streaks on a large part of the blue feathering are suggested in parina though inconspicuously so. The distinctions between the two forms are such that intergradation could easily exist somewhere in the broad area intervening between the respective ranges, but it has yet to be demonstrated. Until some closer approach in range and taxonomic characters can be demonstrated, the two groups are best left specifically distinct.”

Concerning bella, Zimmer wrote:

“Not in the collection before me. In a sense this form is intermediate between typical petersi and parina, being smaller than petersi and with somewhat less strongly marked bright streaks. The more greenish tinge of the male plumage ought to be substantiated by more material since it is within the possible bounds of individual (and "ageal") variation, as demonstrated by the specimens of parina discussed earlier.”

Here are photos of LSUMZ specimens of the three taxa, males and females. I should have checked petersi vs. bella carefully when I took the photos to see if the plumage difference I see in the photos holds up, especially because they conflict with some statements herein. There is a real size difference, obvious to the eye, evident in both sexes. (Many of you will recognize some of the preparators on the labels, all officially Useful Humans. John O’Neill, and even Mark Robbins, was skinning birds before some of you were born. Note also the cleverness with which I have deftly offset petersi in the photos slightly to the left to highlight that it is primarily a bird of the Western cordillera):

Bond (1955) maintained the two as separate species. (He also provided a tidbit on one of M. A. Carriker’s [see previous proposal] above-average field days “It is of interest that Mr. Carriker obtained a specimen of this interesting bird on the same day [March 26, 1932] that he collected his types of Yanacea alpina and Leptasthenura yanacensis and an example of the recently described Zaratornis stresemanni Koepcke. His field notes for that day state that he ‘went up to Puno’, and ‘worked up to about 15500 feet, but most of the shooting was done between 14000 and 15000 feet.’")

Meyer de Schauensee (1966) was the one who first lump then. Shaun Peters (in litt.) graciously dug out this quote from that work (of which I no longer have a copy). From Meyer de Schauensee (1966) under Xenodacnis parina:

“Includes Xenodacnis petersi (Bond & de Schauensee, Not. Naturae, 1939, no. 40, p.1, Yanac, Ancash, Peru) and X. petersi bella (Bond & de Schauensee, l.c., p.2, Atuen, Amazonas, Peru)

“Recently, W. G. George collected specimens of this rare bird in southwestern Arequipa, Peru and found that birds from this region combined characters of both X. parina and X. petersi, showing them to be conspecific.

“George thinks that this aberrant honeycreeper is probably related to Dacnis. The pale glistening blue stripes on the feathers of the back and chest are found in both Dacnis berlepschi and X. parina."

This seems to have become the accepted subsequent classification (e.g. Storer 1970a). Now I really want to see those George specimens at AMNH to assess the “combined characters” assessment. Shaun Peters again came to the rescue in providing a jpg copy of George (1964), using Archives.org, and there is nothing I can find in that article relevant to species limits.

Isler & Isler (1987, The Tanagers) described the vocalizations in detail, but this was based on Ted Parker recordings, which at that time might have been based on petersi only; no differences among taxa were mentioned.

Fjeldså & Krabbe (1990; Birds of the High Andes) wrote: “Vocalizations of ssp parina considerably weaker than in other sspp.”, although they considered bella “doubtfully distinct” from petersi.

Here is the helpful account from Schulenberg et al. (2007; Birds of Peru guide) with Dale Dyer’s plate:

New information:

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

"Streaked: https://birdsoftheworld.org/bow/historic/hbw/tildac3/1.0/introduction

Hitherto treated as conspecific with X. parina, although given species rank when first described, with good reason: differs in its (in male) bright blue shaft streaks (present but less obvious on race bella) vs none (1); (in female) mid-crown to nape ashy brown vs dull blue (3); blue vs ferruginous on malar area at base of lower mandible (ns[1]); blue vs whitish subocular line (1); much larger size (five male wings comprising nominate 82.5, 79, 78 mm and bella 76.5, 75.5 vs two X. parina 66.5, 61.5 (1); allow 3); moreover, commonest song type very different (lower speed and lower number of notes: 2+2=4) (2). Race bella marginally smaller and duller. Names petersi and bella described simultaneously, but latter as a race of former, so petersi claims priority. Two subspecies recognized (but see below)."

The vocal data are from Boesman (2016m), which presents sonograms from what I assume are three different individuals each from the Ecuador population, petersi, bella, and nominate parina, plus one from the somewhat anomalous Arequipa population:

“Song of petersi and bella is a sequence of some 5-10 clear whistles, typically one or more series of repeated identical notes, which are either sharply upslurred or downslurred. Most notes reach max. frequencies of 8-10kHz. 

“All but one recording of nominate are very different from northern races, song having mainly peculiar subdued lower-pitched bubbling notes at a much higher trilled pace. This would seem a major vocal difference, but then there is a single recording from Arequipa (ML33844 from Ted Parker), which is about identical to examples from Ecuador (!). This greatly reduces the marked vocal difference, but nevertheless, it would seem that the most commonly uttered song-type of nominate is very different from any vocalization known from northern races. Based on high pace and high number of notes, a 2+2 score could be given (albeit with some reservation, as there may be two song-types involved).”

Note that Boesman was placing the Arequipa population with nominate parina based on its southerly distribution, but photos by Tom Schulenberg of George’s specimens at AMNH clearly indicate that it belongs with petersi, thus matching the vocal diagnosis.

Discussion and Recommendation: Here we have the Classic Allotaxa Conundrum, with only fragments of information and data with which to work. There are no contact zones to guide us – in fact, even within these taxa, the tiny populations are all isolated from each other in habitat islands at high elevations. Genetic data would be fascinating, particularly since the effective population sizes are likely tiny.

Reasons to vote YES: (1) No rationale has ever been published for the lump, or at least I can’t find it; so, we would be returning to the original classification. (2) Vocal differences as described by Boesman seem substantial. (3) The plumage differences are substantial, and at the risk of getting touchy-feely, they feel like different species to the touch, to me anyway; it’s not just color but something also to do with texture, perhaps an artifact of the shaft streaking in petersi/bella. Even the females seem to have a different plumage quality.

Reasons to vote NO: Are Boesman’s data sufficient for a split? He sampled all the populations, but the total N is unknown, and I would certainly feel more comfortable with a formal analysis that also specified no doubt on homology of vocalizations. How soft are we going to get in terms of requiring a formal analysis in a potentially complicated issue like this one? I hope someone here has the energy to explore recordings on xeno-canto and Macaulay, which are likely many more now than were available to Boesman. By the way, for those of you not on eBird, you can do this efficiently by going to eBird explore page, then hit “Search Photos and Sounds” in lower right; note that you may have to hit refresh to get anything to appear at all; then, enter a taxon name and under “More filters” check “song” or “call”, and where it says Recently Uploaded on far right, toggle that to Best Quality (although beware that eBird recordists, like eBird photographers, chronically over-rate the quality; also that not all recordings are rated, so it might also be worth toggling to Least Rated and skimming through sonograms for good ones).

I have no strong recommendation, although I lean slightly towards a YES. I have conscripted Niels Krabbe to take my vote on this one.

English names: BirdLife International used “Tit-like Dacnis” for nominate parina, and “Streaked Dacnis” for the petersi group. I have problems with this. If there was only one species, then Tit-like Dacnis works pretty well. The “Tit-like” combined with the genus name Xenodacnis (“foreign Dacnis”, with xeno being one of the Greek roots that is more familiar in meaning through “xenophobia”) adequately signaled, to me anyway, that this was not a “real dacnis”. The Burns Lab genetic data show that it is distant from true Dacnis, and clusters with other high-elevation neo-tanagers like Diglossa and Catamenia. BLI, as per their usual policy, kept the parental name with the nominate form despite the fact that it has the much smaller distribution. By our SACC guidelines on English names, either petersi would retain the parental name because of its wider distribution, or new names should be created for both. Then, there is the potential issue of whether to retain “Dacnis” at all. The English name Dacnis would have a helpful 1-to-1 match with the genus name, which is currently hard to do in Thraupidae, if we changed Xenodacnis to something like “Xenodacnis” as the “last” name, which would retain the former connection but emphasize the “foreign” angle. BLI’s “Streaked Dacnis” implies to most that it is just another one of the real Dacnis and is unacceptable in my opinion; even an awkward “Streaked Tit-like Dacnis” with a comparable modifier for parina s.s. would be preferable in my opinion. Dan Lane suggested to me the potential group name “Tit-dacnis”. Anyway, let all that incubate while we deliberate on species limits. I’m all for stability, but if we destabilize species limits, then that would also be the propitious time to consider a new group name. By the way, Dacnis itself is nothing more than an alphabetical barcode in terms of a meaningful name – from Jobling: “daknis type of bird from Egypt, otherwise unidentified, mentioned by Hesychius and Pompeius Festus.

References: (see SACC Bibliography for standard references)

AGUILAR, J. M. 2019. Análisis de la distribución geográfica del género Xenodacnis (Aves: Thraupidae) utilizando el modelado de nicho ecológico. Revista Peruana de Biología 26: 317-324.

AGUILAR, J. M. 2021. Morphological variation in the Tit-like Dacnis (Xenodacnis parina): a call to revise the taxonomic status of Ecuador’s population. Ornitologia Neotropical 32: 51-55.

BOESMAN, P. 2016m. Notes on the vocalizations of the Tit-like Dacnis (Xenodacnis parina). HBW Alive Ornithological Note 393. In: Handbook of the Birds of the World Alive. Lynx Edicions, Barcelona. https://doi.org/10.2173/bow-on.100393

BOND, J. 1955. Notes on Peruvian Coerebidae and Thraupidae. Proceedings Natural Academy Sciences Philadelphia 107: 35–55.

GEORGE, W. G. 1964. Rarely seen songbirds of Peru’s high Andes. Natural History 78: 25-29.

Van Remsen, July 2024

Comments from Niels Krabbe (voting for Remsen): “NO. I mustered enough energy to compare all the 105 different Xenodacnis recordings in XC and ML (25 "cajaensis" from Azuay, Ecuador, 4 bella from Amazonas and Cajamarca, 54 petersi from W Cajamarca, Ancash, and Lima, 9 "arequipae" from Arequipa, and 13 nominate parina from Junín and Cuzco), 38 of which were added to libraries since Boesman's 2016 analysis.

“I also sorted the measurements given by Aguilar (2021) and compared them with body masses of birds I collected in Ancash and Apurímac, and concluded that size is extremely variable (as I had also noted for Cranioleuca baroni in Ancash and Lima) and not a very useful character, when defining taxa of Xenodacnis. This leaves bella characterized only by the lesser extent of blue on the crown in two females when compared with three females of petersi (Bond 1955), in my opinion rather too small a sample size to accept bella as a valid taxon. That bella should be intermediate in size between parina and petersi is not corroborated by the measurements at hand, rather on the contrary. Similarly, the supposed characters defining "cajaensis" do not seem to stand up to scrutiny. The pale brown vent of the female ("ferruginous vent in all X. parina subspecies from Peru") looks much like vents of females of bella and petersi posted in this proposal. As for size difference of "cajaensis", a male I collected in Ancash (MHNJP, tissue ZMUC 116093) was as heavy (23 g) as in the heaviest male of "cajaensis". So as to morphology and plumage, only two taxa seem to be valid: the rather small nominate parina and the streaked petersi.

As to the vocal analysis, Boesman's reason for placing the Arequipa birds with nominate parina was based on vocalizations (not distribution as stated in the proposal). I concur with Boesman, that the majority of recordings of "arequipae" are most similar to those of nominate parina, the only exception being one recording by Ted Parker (ML33844), which is of the raspy notes so typical of petersi, "bella" and "cajaensis". So, although "arequipae" has plumage like petersi, "bella" and "cajaensis", it is vocally intermediate with nominate parina.

“Another critical recording is XC330954, uploaded 12 August 2016 and evidently not analyzed by Boesman. It was recorded by Dan Lane above Chilifruta, Junín, i.e. nominate parina by range. It includes clear whistled song typical of petersi, "bella" and "cajaensis".

“These two cases of intermediacy lead me to believe it highly likely that any two forms would interbreed, if they were to come into contact. Whether they are likely genetically incompatible (i.e. diverged more than 2.5 my ago), remains to be shown by future analyses.

“On present knowledge, I must vote NO to the proposal.”

Comments from Areta: “NO, but narrowly so, until someone amasses more carefully vetted data (and hopefully gets more recordings and genetic data). This seems like a mess with several holes to me. The Ecuador, petersi-bella, Arequipa, and parina populations seem to have different songs. Overall, the main plumage and song types can be grouped in two clusters Ecuador, petersi-bella, Arequipa on one side and parina on the other. However, the complex vocalizations have not been well characterized by Boesman (2016) or Aguilar (2016). Note for example that some recordings from Ancash (petersi) show long and fast songs, while some recordings from Arequipa also show vocalizations that do not fit nicely with the characterization by Boesman (2016) (see for example the spectrogram F in Aguilar 2016]). Aguilar (2016) also indicated that the Ecuadorian populations deserve subspecific recognition (he coined the name cajaensis, but to my knowledge this has not been adequately published) and reports on their plumage and vocal distinctions. The taxon bella has been considered as somewhat intermediate between parina and petersi (in size and male plumage) by Zimmer (1942) despite its geographic position and it remains vocally little recorded, although how little is known indicates affinity to petersi (Van´s photographs of females seem to provide some support on this intermediacy). The Arequipa population seems also widely allopatric to the petersi populations to which it is supposedly allied, and also seems to differ from them vocally: do they also differ in plumage? Aguilar (2016) erroneously included this population in parina as well.

“I think that there is room for a rigorous study here (those by Aguilar 2016, 2019, 2021 are a beginning). If we were to start from scratch, I would certainly endorse a 2-species split based on plumage, and this was also Zimmer´s (1942) preference.

“In our WGAC discussions Tom Schulenberg clarified that the distributions are not as easy as often assumed. In his own words:

"I just want to point out that the 2016 vocal analysis by Boesman mischaracterizes the distributions of the two groups. (this isn't Boesman's fault, he's simply following the original Hilty account in HBW, which also got this wrong). Boesman assumes that the pattern is one of simple north/south replacement: northern petersi/bella and southern nominate parina, but this being Peru, that's not quite the way it works. Petersi and bella generally occur farther north, on the west slope of the Andes and on both sides of the interior Marañón Valley. but the population on the west slope of the Andes in southern Peru in Arequipa also is of the northern type (presumably petersi), even though it appears to be highly disjunct from the closest (known) population of petersi in northern Lima. the Arequipa population is very poorly known, and it may be that the only specimens in the US are one or two that William George collected in 1962. I've seen this material (American Museum of Natural History) and should have photos somewhere, which I'll try to locate. there may well be additional specimens in the collection in Arequipa, and perhaps we can request photos of these if need be. Anyway, the upshot is that nominate parina of course occurs in southern Peru, but only in the interior valleys, and very locally on the east slope as well; and while petersi and bella primarily occur in northern Peru, there is an outpost on the west slope in southern Peru in Arequipa.’

“Also, regarding the Arequipa populations, Tom provided the following evidence. Three females. from left to right, these are from Ayacucho (parina), Lima (presumed petersi), and Arequipa (Photo courtesy of T.S. Schulenberg):