Proposal (548) to South American Classification Committee

Recognize newly described Cinclodes espinhacensis

Effect on South American CL: This proposal would add a newly described species to the list.

Background: Freitas et al. (2012) discovered a new population of Cinclodes, most closely related to C. pabsti, in Minas Gerais, Brazil, where it inhabits campo rupestre habitat in the Serra do Cipó area in the southern portion Espinhaço Range, over 1100 km N of the range of C. pabsti. It is thus by far the most isolated population of any Cinclodes.

Freitas et al. compared their 10 specimens of the new population to 6 of C. pabsti. Their published photos and description make it clear that this new population differs in being much darker overall and is thus represents a diagnosable unit worthy of taxonomic recognition.

The authors designated their new taxon as a species under BSC criteria because songs were diagnosably different in terms of frequency and duration. They analyzed 26 recordings of songs of 4 individual espinhacensis and 30 of 4 individual pabsti, and found statistically significant differences in 8 of 17 parameters. They also analyzed 475 recorded call notes of 12 individual espinhacensis and 225 of 8 individual pabsti, and found statistically significant differences in 3 of 6 parameters measured. (I am concerned, however, about independence – using multiple recordings of the same individuals as independent units of analysis is questionable.) The sonograms show the longer duration of the call notes and the lower frequency of the song.

The authors also analyzed mtDNA sequence data (2528 bp, 4 loci) for 11 espinhacensis and found that it is reciprocally monophyletic with respect to its sister taxon, pabsti. Genetic distance between the two is small, although slightly greater than that between C. taczanowskii and C. nigrofumosus and only slightly less than that between C. oustaleti and C. olrogi.

Analysis and Discussion: This one is right on the border of the subspecies/species boundary, and the authors have done almost everything they could to produce data relevant to the basic question under the BSC for allopatric taxa, namely, has this new form diverged to the degree associated with known species limits in related birds? The only missing piece is playback trials.

I lean towards treating espinhacensis at the species rank because there are more data to support ranking it as a separate species than there are for treating C. taczanowskii as a separate species from C. nigrofumosus, or C. aricomae from C. excelsior. Further, we know from Sanín et al. (2009) that morphological differentiation among good species in the C. fuscus is minimal. Overall, species differences are slight among species in the various species groups in Cinclodes.

Recommendation: I recommend a "yes" vote to add this newly described species to the South American list.

References:

FREITAS, G. H. S., A. V. CHAVES, L. M. COSTA, F. R. SANTOS, AND M. RODRIGUES. 2012. A new species of Cinclodes from the Espinhaço Range, southeastern Brazil: insights into the biogeographical history of the South American highlands. Ibis 154: 738-755.

SANÍN, C., C. D. CADENA, J. M. MALEY, D. A. LIJTMAER, P. L. TUBARO P.L., AND R. T. CHESSER. 2009. Paraphyly of Cinclodes fuscus (Aves: Passeriformes: Furnariidae): Implications for taxonomy and biogeography. Molecular Phylogenetics and Evolution 53: 547-555.

Van Remsen, September 2012

Comments from Zimmer: “YES. I agree with Van that this one seems borderline in terms of being divergent enough from pabsti to be recognized as a species rather than a subspecies. Then again, as Van also points out, plumage (and voice) is evolutionarily conservative in this group. The Espinhaço Range is a mini-center of endemism (e.g. Asthenes luizae, Augastes scutatus), and the range disjunction from pabsti is large, so treating the two as separate species makes sense biogeographically, and, they would appear to be safely on independent evolutionary trajectories.”

Comments from Stiles: “YES, for reasons summarized by Kevin and Van.”

Comments from Nores: “YES. It is evidently a new bird, although that is not definite that is a species and not a well-differentiated subspecies. It is surprising that the authors have not marked anything that is unique in the habitat of Cinclodes: water independence.”

Comments from Nacho Areta and Mark Pearman: “The description of a new species of Cinclodes by Freitas et al. (2012) is an interesting case. We believe that the authors of the paper describing Cinclodes espinhacensis have missed the opportunity of making a more useful comparison of vocalizations in the genus and have based their conclusion mostly on our previous lack of knowledge to justify their species instead of putting forward a better frame to understand what a species should be in Cinclodes. There is no thorough comparative study of other species vocalizations or morphology and no evaluation of behavioral responses to playback, and thus no possibility of evaluating whether the differences described should be interpreted as indicating the existence of a new species or of a new subspecies. Weighing up the evidence mentally in comparison to Cinclodes with which we are familiar (including pabsti), does not reveal any meaningful biological difference between espinhacensis and pabsti to justify their treatment as different species. More objective reasons to support this view are summarized below.

“Vocalizations: even though Freitas et al. (2012) have found differences in some acoustic parameters, these are highly variable, and there is no mention on whether the cuts were unsolicited or in response to playback, which may influence parameters such as pitch and number of notes per vocalization (let alone differences due to different recording gear and recording quality), several acoustic parameters also show no difference (9 out of 17 parameters show no difference, while 8 out of 17 do; how should we interpret this?), the shape of notes in the songs is identical (the call is also extremely similar in every respect with 3 out of 6 parameters showing no differences; without a clear ‘yardstick’ we cannot evaluate just how conservative songs are in Cinclodes), and finally there is a whole lot of pseudoreplication in their measurements as it is not clear how many calls and songs from each individual they've measured (i.e., individual consistency and species level differences are conflated).

“

Morphology: there is no diagnosable morphometric difference except for tarsus (with a large p, though...) and weight (which is awkward, given that morphometrics usually correlated with weight such as wing size are identical; we are in the presence of two birds with very similar shapes which have dramatic weight differences, perhaps an artifact of weighing birds in different seasons and/or with different tools?)

“Comparison to Upucerthia: given the lack of comparative data within Cinclodes, looking at its sister genus Upucerthia is warranted. The authors didn't mention the existence of ecogeographic variation in Upucerthia (e.g., U. dumetaria and U. jelskii), which may easily explain the difference in coloration between espinhacensis/pabsti without necessarily meaning that a different species is involved (indeed, all undisputedly valid species in Cinclodes [and Upucerthia] differ in pattern and coloration, not just coloration). Tail pattern is particularly important in Cinclodes yet pabsti and espinhacensis show identical tail patterns. The level of differentiation found between espinhacensis and pabsti seems equivalent to that between Upucerthia jelskii and U. validirostris two ‘species’ that answer readily to playback of each other (whose vocalizations are so variable that they defy any meaningful quantitative characterization despite being structurally undistinguishable as is the case of note shape [and pattern?] in espinhacensis/pabsti) and which should be considered as part of a single geographically variable species (Areta and Pearman in press). If the darker saturation of dorsal plumage in espinhacensis is the only difference between it and pabsti, then similar distinctions at subspecies rank or indeed at color morph level can be found in numerous furnariid species (Remsen 2003).

“Genetic differences: the meager genetic differentiation, with values falling between those of the conflictive cases oustaleti-olrogi and nigrofumosus-taczanowskii, adds little to answer the question of the species-level of espinhacensis but if anything, it should raise suspicion.

“Geographical gap: finally, we would like to stress that the gap between espinhacensis and pabsti should not be used as a tool for species-level assignment. A large gap (for the poor flying Scytalopus) of 450 km is known to exist (or was known to exist) between records of S. iraiensis in Paraná and Minas Gerais, which may well be a sampling artifact (Vasconcelos et al. 2008). How much of the gap between espinhacensis and pabsti is an artifact? Let us also consider that species such as Cinclodes fuscus may easily fly 3000 km from Patagonia to Buenos Aires and another 3000 km back every season during their migration (and perhaps much more, when reaching south-east Brazil). To be clear: the gap between espinhacensis and pabsti is within flight distance of at least one migrant Cinclodes, and all inland Cinclodes have powerful flights.

“In sum, we do not see evidence for considering the Espinhaço birds as a different (biological) species from southern birds, and hope these arguments will generate an interesting discussion for SACC members.

“References

Areta, J.I. and M. Pearman. In press. Species limits and clinal variation in a widespread high Andean Furnariid: the Buff-breasted Earthcreeper (Upucerthia validirostris). Condor.

Remsen, J.V. Jr. 2003. Family Furnariidae (Ovenbirds). Pp. 162-357 in: del Hoyo, J., Elliot, A. & Christie, D.A. eds. (2003) Handbook of the Birds of the World. Vol. 8. Broadbills to Tapaculos. Lynx Editions, Barcelona.

Vasconcelos, M. F., Maurício, G. N., Silveira, L. F. and G. M. Kirwan. 2008. Range extension for Marsh Tapaculo Scytalopus iraiensis to the highlands of Minas Gerais, Brazil, with an overview of the species’ distribution. Bull. B. O. C. 128:101-106.”

Comments from Thomas Donegan: “This is a comment on some of the analysis of data on continuous variables (biometrics and voice) in this paper, following up on Nacho Areta's comments.  The authors include means, standard deviation, and range data for all the variables measured. They cite various variables as differing in the "diagnosis" section, making it clear which variables overlap and which do not.   If one looks at table 2 (page 7), there are no vocal variables that are diagnosable based on recorded values (range data overlapping) - even if t and p shows they are statistically mildly differentiated. No attempt in the paper is made to use statistical tests for diagnosability.  Following the nomenclature for stats in my recent series of papers, the authors of this new taxon therefore looked only at "Level 1" (statistical significance of differences between means using t; the weakest form of differentiation) and "Level 4" (diagnosability based on recorded values; which is the form of differentiation most subject to biases related to sample size).  The Isler et al. 97.5% test of diagnosability was not applied, nor were any tests of 50% or 75% diagnosability.  Out of the continuous vocal and biometric variables studied, only body mass appears to have met the "Level 4" test.  However, here, the means are less than four average standard deviations apart, so would not meet statistical tests of diagnosability (Level 5), even before controlling for sample size.  I do not know these birds and do not want these comments to detract in any way from appreciation of the authors' efforts in discovering and publishing many interesting details of this previously undescribed new taxon.  Other data, notably molecular and plumage data, will be relevant to the question of species rank too."

Additional comments from Zimmer: “One of the highlights of the tour was our couple of days in the Espinhaço range, where we were able to locate a couple of pairs of the recently described Cinclodes espinhacensis.  I had voted on the relevant proposal (#548) to the SACC just prior to the trip, rendering my vote based solely on available information as presented in Freitas et al. (2012), and upon extensive field experience with the closely related Cinclodes pabsti.  With the additional benefit of field experience with the newly described taxon, I would like to add some additional perspective to my earlier comments.

“I found the morphological distinctions of espinhacensis relative to pabsti to be just as described by Freitas et al. (2012),  i.e. espinhacensis is markedly darker dorsally, with a slightly, but noticeably more clouded breast (lending a slightly "bibbed" appearance), and appears more slender and less bulky in build.  Collectively, I still feel that these morphological distinctions meet the yardstick test of other species-pairs of Cinclodes in what is clearly a group in which morphology has proven evolutionarily conservative.

“My field experience did however provide revelations concerning the vocalizations of espinhacensis relative to those of pabsti.  We already knew from Freitas et al. (2012) that the two taxa had broadly similar loudsongs and single-note calls.  I had the opportunity to evaluate the vocalizations (both songs and calls) from 2 pairs of espinhacensis, and to perform unidirectional tape playback experiments by presenting both pairs of espinhacensis with playback of songs and calls of pabsti.  When we encountered the first pair of espinhacensis, they were foraging on the ground ca. 100 m away from us.  I barely initiated playback of songs (no calls) of pabsti, before the pair of espinhacensis came rocketing in toward us like heat-seeking missiles.  They perched on the nearest fence posts and sang back repeatedly and without letup, all the while, vigorously flapping their wings in a display nearly identical to what I have seen pairs of pabsti do on numerous occasions.  I might add that the wind was blowing in the range of 20-30 mph at the time, and passerines in general were not exposing themselves any more than they had to.  The pair of cinclodes had trouble staying on their perches, such was the strength of the wind.  On 3 occasions the birds flew off and perched more than 100 m away from us, where they resumed singing in agitation. Each time I attempted playback of the song of pabsti (using pre-recorded songs of 4 different individual pabsti from Rio Grande do Sul), the pair of espinhacensis came roaring back in, singing and displaying wildly.

“On our hike out of the area, we encountered another individual of espinhacensis foraging along the margins of a large pond. This was clearly a different bird, located 500 m+ from the first pair, which were still vocalizing in the distance behind us.  This time, we presented the bird with playback of the single-note calls (but no songs) of pabsti (also recorded in Rio Grande do Sul). Once again, the reaction was immediate, with the bird we were watching flying straight at us and landing on the ground nearby and calling (single-note calls) over and over in apparent distress. Meanwhile, its presumed mate, which we had not seen, flew in from the far side of the pond, circled us, and landed nearby and gave several songs.

“So, we had 2 different pairs of espinhacensis respond repeatedly and vigorously to playback of audio recordings of pabsti.  The first pair responded to loudsongs of 4 different individuals (These were presented separately (i.e. The birds came in the first time to playback of pabsti #1, then flew off, then returned in response to pabsti #2 and so on.).  These recordings of pabsti varied noticeably in the length and number of notes of the songs, and also in the pace and change of pace in the songs.  I've noticed this variation in songs of pabsti before, and the variation between songs within a single individual is at least as great as the variation between individuals, often reflecting differences in degrees of agitation in response to playback.  We witnessed the same phenomena with the two pairs of espinhacensis, i.e. individual birds changed the length and pace of their songs from one song to the next.  We tape recorded this as best as we could, although the persistent high winds were hellish for recording.  The second pair of espinhacensis reacted equally strongly to pre-recorded single-note calls (in the absence of songs) of pabsti.  In all instances, the vocalizations of espinhacensis (both songs and calls) were indistinguishable (to my ears at least) from those of pabsti.

“My personal opinion is that negative results from playback trials are more significant than positive results.  In other words, many taxa that are universally recognized as being good biological species will still respond positively to playback of other vocally similar species in the same genus (for example, playback of almost any Thamnophilus antshrike can be used to stimulate almost any other Thamnophilus species), perhaps due to interspecific competition. On the other hand, a complete absence of response to playback is usually more informative, especially if the recipient of the playback responds to its own voice without responding to the other taxon. And, I do think that the voices of the various cinclodes are largely similar enough that a positive response from one species to playback of another is not necessarily a deal-breaker. However, in the current case, the response from 2 pairs of espinhacensis to both single-note calls and loudsongs of pabsti was immediate, dramatic, and vigorous, and accompanied by the same sorts of wing-flapping displays that are used by other cinclodes in defending their territories against conspecifics.  I can think of plenty of closely related species-pairs with parapatric/allopatric distributions in which the loudsongs are nearly identical but the calls are very different (e.g. some pairs in the Hypocnemis cantator complex), or, where the calls are nearly identical but the loudsongs are very different (e.g. Schistocichla antbirds), but seldom are both the calls and the songs as well as the associated territorial displays identical.

“Also of note, is that my sense from reading Freitas et al. (2012) was that espinhacensis differed further from pabsti in its ecological associations, being restricted to higher elevations with a highly endemic and specialized flora. While that may be the case for the bulk of the population, the pairs of espinhacensis that we encountered were at the lower elevational limit, and were frequenting highly human-modified pasture and margins of seasonal ponds, all subject to heavy grazing by cattle and horses -- in other words, the same kinds of habitats where one finds pabsti in Rio Grande do Sul.

“In summation, the morphological distinctions of espinhacensis do meet the yardstick test for species recognition (but could also be merely an example of Gloger's Rule), but I do not believe the calls or the songs of these birds are diagnosable from those of pabsti, and limited playback trials suggest that these two populations, were they to come in contact, would still treat one another as conspecific.  There is no doubt in my mind that espinhacensis is a valid taxon, and I do believe that they are on an independent evolutionary trajectory from pabsti but at this point, I don't feel that they meet our criteria for being recognized at the species level.  Given the advantage of hindsight, I would like to change my vote to NO, and recognize espinhacensis as a distinct subspecies of Cinclodes pabsti.”

Cipo Cinclodes (espinhacensis) by Kevin Zimmer:

Cipo Cinclodes (espinhacensis) by Kevin Zimmer :

Long-tailed Cinclodes (C. pabsti) by Kevin Zimmer :

Additional comments from Remsen: “In view of Kevin’s findings and Nacho’s comments, I change my vote to NO. This is a valid new taxon but best ranked as a subspecies.”

Comments from Robbins: “NO. Based on Kevin’s field experience, especially the limited playback experiments, it seems best to treat this new taxon as a subspecies.”

Comments from Pérez-Emán: “NO. This is another borderline case as the one in Proposal 545. Species level definition in Cinclodes is not clearly associated to particular levels of divergence either in morphological or vocal characters. Similarly, level of molecular divergence among sister “species” is characterized for a wide range of values, a pattern that might be the result of different historical and/or ecological processes associated with their divergence or problems with current taxonomy in the genus. In the particular case of this proposal, espinhacensis differs from pabsti in plumage coloration and size but in general are very similar birds with similar vocalizations. Moreover, after reading the article by Freitas et al. (2012), vocal analyses might suffer from lack of independence and lack of diagnosability tests, as others have indicated before. Kevin’s data on playback further incline the balance toward recognition of two subspecies instead of two valid species under the BSC. Final outcome of this proposal might lead to a reevaluation of Cinclodes species limits.”