Proposal (838) to South American Classification Committee

Treat Cinclodes olrogi as conspecific with Cinclodes oustaleti

Effect on South American CL: This proposal would place Cinclodes olrogi within the Cinclodes oustaleti species complex that includes various allopatric taxa:-

C. o. oustaleti: C Andes of Argentina and Chile south to N Patagonia

C. o. hornensis: S Patagonia and Tierra del Fuego

C. o. baeckstroemii: Isla Alexander Selkirk formerly Masafuera, Chile

C. [o.] olrogi: Sierras of Córdoba and San Luis, CN Argentina

Background: Our current SACC note reads as follows:

14a. Nores (1986) considered Cinclodes olrogi to be a subspecies of C. fuscus (broadly defined); others (Olrog 1979, Navas & Bó 1987, Vuilleumier & Mayr 1987, Mazar Barnett & Pearman 2001) considered it more likely to be closely related to C. oustaleti. Genetic data support the latter relationship (Chesser 2004a, Derryberry et al. 2011). See also Note 14b. SACC proposal needed.

Cinclodes olrogi has long been known as a controversial "endemic" species in Argentina ever since its description in 1979. Different authors have classified it as either a distinct species, as a subspecies of C. oustaleti or as a subspecies of C. fuscus. The brief type description (Nores & Yzurieta 1979) treats olrogi as a monotypic species although in the same year C. C. Olrog (1979) immediately treated olrogi as a subspecies of oustaleti. It is also worth noting that M. Nores subsequently changed his stance and considered olrogi to be a subspecies of Cinclodes fuscus (Nores 1986, 1996). However, and in the only exhaustive and critical morphological study of the oustaleti complex, J.R. Navas and N. Bó (1987) convincingly demonstrated that olrogi should be regarded as a subspecies of oustaleti based upon plumage and morphometric data.

As invariably noted, virtually all of the plumage features found in olrogi are present in either nominate oustaleti or hornensis or both, and the white center of the abdomen is a shared feature of all of these taxa (Fjeldså & Krabbe 1990, Navas & Bó 1987, pers. obs.).

The concluding remarks by Navas and Bó (1987) are also pertinent when they suggested that “the most important feature [distinguishing olrogi from oustaleti] was a shorter mean difference of 1-3 mm in the culmen of olrogi compared to oustaleti and hornensis.”; Jorge R. Navas spent much of his life making meticulous specimen comparisons as the MACN curator (Buenos Aires) of 60,000 specimens.

New information: A phylogenetic analysis (Chesser 2004) based on mitochondrial DNA (complete COII and ND3 genes) showed a divergence of 0.5% between olrogi and oustaleti. Chesser (2004) concluded that “it seems likely that they are potentially interbreeding taxa, and they are perhaps best considered biologically conspecific, pending further study.”

Using the same genes and partly the same dataset, Sanín et al. (2009) recovered, without much support, a sister relationship of olrogi and oustaleti, but interestingly a haplotype of what is currently known as C. albiventris tucumanus was more similar to oustaleti than to olrogi (Figure 2). They also mentioned that “Although the status of the forms [of the then C. fuscus] occurring in the central Andes and Argentinean highlands is not entirely clear with respect to C. olrogi and C. oustaleti, we suggest they should be maintained separate and elevated to the species-level, taking the name C. albiventris. This species would provisionally comprise five subspecies: C. a. albiventris, C. a. tucumanus, C. a. yzurietae, C. a. riojanus, and C. a. rufus. Additional work will be necessary to determine whether these taxa are maintained as specifically distinct from C. olrogi and C. oustaleti. (Sanín et al. 2009:554).”

Fig. 1. Chesser (2004): Phylogram of the ML tree constructed using combined sequences of COII and ND3.

Fig. 2 . (Sanín et al. 2009). Maximum clade credibility tree for ND3 and COII haplotypes of C. fuscus and the other Cinclodes species, inferred using Bayesian analysis as implemented in BEAST.

“Most haplotypes of the central subspecies C. f. albiventris/ tucumanus formed a weakly supported clade (0.59, 28%, 24%), which grouped with C. olrogi and C. oustaleti (0.93, 40%, 42%). However, one haplotype, observed in two individuals from Tucumán, Argentina, was identical to one of the two haplotypes of C. oustaleti reported by Chesser (2004). In the Bayesian analysis, C. f. albiventris/tucumanus was sister to a clade formed by C. olrogi and C. oustaleti (posterior probability: 0.98), whereas in the MP and ML trees C. oustaleti was weakly supported (<20%) as sister to a clade formed by C. olrogi and C. f. albiventris/tucumanus. Divergence of some haploypes within C. fuscus albiventris/tucumanus (range 0.09–1.06%) was greater than the divergence of haplotypes between C. f. albiventris/tucumanus and C. oustaleti (range 0. 38–0.96%, without considering the shared haplotype between them) or C. olrogi (range 0.29–0.87%). The haplotype network (Fig. 3) provided a different perspective on relationships among these taxa. Although most C. f. albiventris/tucumanus haplotypes clustered together, two haplotypes from central Peru were distant from the rest. The two haplotypes of C. olrogi formed a cluster, as did those of C. oustaleti (including the fuscus individuals that share the haplotype), but these two clusters did not form a distinct unit. These clusters were separated from the rest of the network by only two mutations each, whereas the separate C. f. albiventris/tucumanus haplotype cluster was four steps removed.” (Sanín et al. 2009:551)

Fig. 3 (Sanín et al. 2009). Median-joining haplotype network showing relationships among C. fuscus albiventris, C. oustaleti, and C. olrogi haplotypes. Sizes of the circles are proportional to the number of individuals sampled with that haplotype. Connection lengths are proportional to the number of mutations between two haplotypes. Shadings and patterns correspond to different regions or species. Small black vertices represent unsampled or extinct haplotypes.

Analysis: Since Chesser (2004) and Sanin et al. (2009), there has been a long period of silence, yet published morphological and genetic data both support the notion that olrogi is conspecific with oustaleti. There seem to be no lines of evidence at all that would support recognition of olrogi as a valid species. This in turn questions the recognition of albiventris as separate from oustaleti, but more comparative material is needed to resolve this problem.

The taxonomic status of hornensis and baeckstroemii in relation to oustaleti seemingly require further study but are unrelated to this proposal.

All of these taxa are largely sedentary with some north-south migration in hornensis, and also slight altitudinal migration in oustaleti, hornensis and olrogi.

Recommendation: We recommend a YES vote to lumping olrogi with Cinclodes oustaleti for the reasons mentioned above.

Literature Cited:

Chesser, R.T. 2004. Systematics, evolution, and biogeography of the South American ovenbird genus Cinclodes. The Auk 121 (3), 752-766.

Fjeldså, J. & Krabbe, N. 1990. Birds of the High Andes: A Manual to the Birds of the Temperate Zone of the Andes and Patagonia, South America. Zoological Museum, University of Copenhagen,

Navas, J.R. & Bó, 1987. Notas sobre Furnariidae argentinos (Aves, Passeriformes). Museo Arg. Ciencias. Naturales 14(4): 55-86.

Nores, M. 1986. Diez nuevas subespecies de aves provenientes de islas ecológicas argentinas. Hornero 12(4): 262-273.

Nores, M. 1996. Avifauna de la Provincia de Córdoba. Pp 255-337 in Di Tada, I.E. & Bucher, E. eds. Biodiversidad de la Provincia de Córdoba, 1. Univ. Nac. Rio Cuarto.

Nores, M. & Yzurieta, D. 1979. Una nueva especie y dos nuevas subespecies de aves (Passeriformes). Academia Nacional Ciencias Córdoba Misc. 61: 4-8.

Olrog, C.C. 1979. Nueva lista de la avifauna argentina. Opera Lilloana 27.

Sanín, C., Cadena, C.D., Maley, J.M., Lijtmaer, D.A., Tubaro, P.L. & Chesser, R.T. 2009. Paraphyly of Cinclodes fuscus (Aves: Passeriformes: Furnariidae): Implications for taxonomy and biogeography. Molecular Phyl. and Evol. 53: 547-555.

Mark Pearman and Nacho Areta, July 2019

Comments from Claramunt: “NO. The arguments for conspecificity presented are very weak and information on potential reproductive compatibility is completely lacking:

“Plumage: Statements by Navas & Bo, and Nores are based on “degree of similarity;” they concluded that the differences are “at the subspecific level.” I admit that these species are very similar (there is an obvious general plumage conservatism in Cinclodes and other open country furnariids). Nothing strange there, especially given the recent divergence times. However, overall similarity should not be interpreted as evidence of conspecificity (or pheneticism is back?). olrogi is fully diagnosable, at least by the shape and color of the alar band, which is pure white and broad distally in olrogi (conspicuous even in the closed wing), versus ochraceous and narrower in oustaleti. This difference should not be downplayed, as it would be very conspicuous during the singing display typical of this genus. Other plumage differences such as the browner tones of olrogi may be also diagnostic.

“Morphometrics: Differences in bill length are not tremendous but actually they show no overlap: olrogi 14 to 16 mm (mean of 15 mm, 13 individuals), oustaleti 17 mm (9 individuals, no detectable variation). So, these species seem diagnosable by bill length (at least no evidence of continuous variation).

“Genetics: The genetic evidence only suggests that there may be some issue of incomplete lineage sorting or past hybridization but between albiventris and oustaleti. C. olrogi is not involved in this problem (it has its own, non-shared, haplotypes). So, no evidence of conspecificity there.

“Genetic distances between olrogi and oustaleti are similar to those between oustaleti and albiventris, or between fuscus and antarcticus. So, no evidence of anything there.

“Phylogeny: olrogi appears as sister to oustaleti in the mtDNA tree, but the node that unites them does not have statistical support, so we cannot rule out the possibility that oustaleti is closer to albiventris, with olrogi as a more distant relative.

“Therefore, I don’t see any evidence of conspecificity between olrogi and oustaleti. Moreover, evaluating the status of olrogi without considering the status of albiventris seems problematic.”

Comments from Remsen: “NO. I agree with all of Santiago’s points above. I would also add that I wonder if the phenotypic assessments of Navas, Bó, and Nores would have been different if they had not been in the framework of C. fuscus sensu lato (prior to Sanín et al.).

“Even if this proposal passed, it’s a good example of why any proposal that focuses the issues towards what additional data are needed is productive (and more valuable than the immediate outcome of the voting). Specifically, what data are needed to assess species limits in this genus? Are there any contact zones to provide insight on barriers to gene flow that could be applied as a yardstick to allotaxa (with all appropriate caveats)? Are the vocalizations perceived by the birds as being as similar as they do to us? Given that these are among the most vagile furnariids, it seems likely that there are intrinsic barriers to free gene flow.”

Additional comments from Pearman and Areta:

“We would like to add some more information to expand our proposal, because we feel that some of the arguments put forward by Santiago demand clarification on our part.

“Plumage

“Santiago stated that "olrogi is fully diagnosable, at least by the shape and color of the alar band, which is pure white and broad distally in olrogi (conspicuous even in the closed wing), versus ochraceous and narrower in oustaleti. This difference should not be downplayed, as it would be very conspicuous during the singing display typical of this genus. Other plumage differences such as the browner tones of olrogi may be also diagnostic.”

“However, although we agree that olrogi is diagnosable, his characterization may create a false impression of great distinctiveness that is not borne in the field or in specimens. Diagnosability does not equate to species status, and the critical point lies in understanding what does diagnosability entails from the point of view of speciation. In our proposal we did not enter into the finer detail of plumage distinctions because these taxa are extremely similar and the proposal focused on the genetic data.

“Wing-stripes in Cinclodes can be one color on the upperwing and a different color on the underwing, and upper and underwing seem to differ on their importance during display. The wing-raising display in all Cinclodes involves a high, vertical wing spread with fluttering and rotation on an almost vertical axis, meaning that the bird is always exposing the wing-flash or ulnar bar of the underwing to a potential mate from an elevated perch; not the upperwing.

“The underwing pattern of oustaleti and olrogi is 100% identical in coloration; a more recently collected olrogi specimen was very slightly cleaner white than older oustaleti specimens. MP measured the maximum width of this ulnar bar and found it to be 22 mm in both olrogi and oustaleti which at least demonstrates that from a single comparison they were identical, and not broader in olrogi as Santiago stated.

MACN specimens: C. oustaleti oustaleti (left) and C. olrogi (right)

“Moving on to the upperwing, the wing-stripe of both taxa is largely a fairly bright cinnamon, with the difference that in olrogi the most distal portion becomes whitish; just barely visible in the photograph below. This white portion can sometimes be seen on the closed wing, but it is not used in display. Thus, both the upper and underwings of these taxa are extremely similar. Importantly, Cinclodes oustaleti hornensis has an identical upper and underwing pattern to olrogi.

MACN specimens: C. oustaleti oustaleti (left) and C. olrogi (right)

“We were surprised by Santiago´s suggestion that we are endorsing a “pheneticist” point of view, because we are attempting to use whatever is known of these taxa; and morphology seems to be the main source of characters at stake. We submit that a small difference in bill length (not controlling for the possible existence of latitudinal clines in oustaleti-hornensis) is meager evidence to grant species status to olrogi.

“Intraspecific clinal variation leading to mildly or wildly different looking populations is known to occur along the Andes and neighbouring mountain chains (e.g., Upucerthia dumetaria and U. validirostris for which see Areta & Pearman (2009, 2010), and the recent merger of Cranioleuca baroni and C. antisiensis for which see Seeholzer & Brumfield (2017). We see that available evidence for considering olrogi as a different species from oustaleti is exiguous, whereas considering them as conspecific is more consistent with available data on morphology and genetics. On a similar vein, Cinclodes atacamensis schocolatinus (from the Sierras de Córdoba and San Luis) is as different from the Andean Cinclodes atacamensis as olrogi is from oustaleti. Yet, nobody seems to be arguing for a split here. Interestingly, in both cases the Córdoba populations are darker than the Andean populations.

Genetics

“We do not advocate for a genetic-threshold for species recognition, and are aware that good biological species can differ minimally in this regard. However, we fail to see how having two slightly different haplotypes provides solid evidence of the existence of two species per se. What we argued is that the genetic data is weak, and thus it should be interpreted with care. The fact that “genetic distances between olrogi and oustaleti are similar to those between oustaleti and albiventris” is put forward as an argument, but it can instead be used to undermine the recognition of albiventris as distinct from olrogi and oustaleti. The issue of whether albiventris is conspecific or not with olrogi-oustaleti is a lingering one. Interestingly, the taxa yzurietae-riojanus-famatinae are but reddish versions of oustaleti-olrogi, while albiventris is a browner version sharing the same structural plan.

“The lack of support for the “albiventris” group (see Figure 2) and the sharing of haplotypes of albiventris with oustaleti suggest that the story is not simple, yet, the split of albiventris was accepted without serious conflict on this topic by SACC (http://www.museum.lsu.edu/~Remsen/SACCprop415.htm). Recognizing the extremely similar and genetically weakly differentiated olrogi and oustaleti as a single species can be considered as a necessary first step to solve the taxonomy of this group.

“At any rate, genetic evidence can be interpreted in many difference ways. From our perspective, however, the most difficult interpretation is that the scant genetic differentiation provides evidence in favou\r of the recognition of two species. It can be consistent with two species, given that some species exhibit almost null genetic differences, but it is difficult to propose it as a solid argument in favor of species status for olrogi.

Vocalizations

“Vocal differences, if any, remain to be shown. In the few recordings available, vocalizations of all taxa in the albiventris complex are very similar. More recordings are needed, because individual birds can change the pattern of song within the same bout (pers. obs.). So caution is advised when jumping quickly to conclusions regarding differences based on a few recordings.

Final remarks

“Although Santiago routinely mentioned “no evidence” indicating conspecificity of olrogi and oustaleti, we argue that evidence supporting their recognition as different species does not fit well with current species limits in Cinclodes. To this end, it might be relevant to bring back the case of Cinclodes espinhacensis (which was considered to be subspecies of C. pabsti by this committee: http://www.museum.lsu.edu/~Remsen/SACCprop548.htm), and the problematic C. nigrofumosus/taczanowskii duo, which seem to be more a sign of historical momentum than of solidly sustained biological species (http://www.museum.lsu.edu/~Remsen/SACCprop33.htm).

“In sum, this is not an easy case, but C. olrogi is a little-differentiated taxon both in morphology and plumage, vocalizations resemble those of C. oustaleti/albiventris and it shares a distributional pattern with subspecies schocolatinus of C. atacamensis in the Sierras de Córdoba. This, added to the nature of geographic variation in Cinclodes (and other Andean furnariids, including the closely related Upucerthia) should put the burden of proof in those advocating full species status for olrogi.”

Additional comments from Remsen: “I am swayed by the new comments from Mark and Nacho, and I change my vote to YES. More broadly, the issue of species limits in this genus really needs re-evaluation, especially from the standpoint of voice and playback trials.”

Comments from Zimmer: “YES”. I must confess to being influenced in this case by my field experience with Cinclodes espinhacensis and C. pabsti (which seem to differ at least as much in plumage characters as olrogi and oustaleti), which showed 2 pairs of espinhacensis to respond instantly, vigorously and repeatedly to playback of both single-note calls and songs from multiple pabsti (see my detailed comments to Proposal #548). This experience not only caused me to change my vote on Proposal #548, but it also left me feeling less than confident about the morphological “yardsticks” used for evaluating species-limits within this evolutionarily conservative genus.”

Additional comments from Claramunt: “This case is not similar to the espinhacensis case and the discussion of the wing patch by Nacho and Mark is misleading: 1) the dorsal side of the wing is clearly visible during displays in Cinclodes, 2) specimens in the pictures are inadequate to show anything useful as the difference between oustaleti and olrogi is in the distal section of the wing, in the primary feathers that are concealed in the specimens (not in the “ulnar” = secondary feathers). Regarding morphometrics, I have measured specimens of both species and, although sample sizes are mall (n=3, 4) I can confirm the bill size differences using bill length from nares (olrogi 10.6-11.4 vs. oustaleti 12.2- 13.3mm) and I realized that they also differ in hallux length with no overlap (olrogi 16.2-17.2 vs. oustaleti 18.5- 19.4mm). So, this is different from the case of pabsti/espinhacensis, which show wide overlap in all measurements. Instead, olrogi shows diagnostic differences in at least one plumage trait and two morphometric traits. Only a phenetic point of view would disregard these differences on the ground of being of small magnitude. Instead, I think that these differences demonstrate a history of reproductive isolation that explains the accumulation of diagnostic differences between these two taxa.”

Comments from Bonaccorso: “NO. Driving away from the “pheneticist” discussion, lack of genetic divergence is just one line of evidence. Both plumage and morphological differences seem subtle, but are discrete. Also, it is important to evaluate plumage differences in the context of subtle plumage differences found among other species of Cinclodes (e.g. C. fuscus and C. albidiventris). I have not seen C. olrogi displaying, but if the white wing patch is conspicuous during display, it is definitively a key feature that could point to reproductive isolation. Also, even if future analyses confirm that Mark and Nacho are right, it will be easier to just lump these two taxa. However, if future analyses show that they are not conspecific, we will have to reverse the lumping, creating more chaos.”

Comments from Robbins: “YES. The minor level of differentiation in both morphological and genetic data of olrogi suggest that it should be treated as conspecific with oustaleti. Yes, there needs to be a re-evaluation of species limits within this entire genus. I vote for subsuming olrogi into oustaleti.”

Comments from Stiles: “In view of the most interesting (and confusing, at least to me) discussions on the olrogi-oustaleti question, I think it best to abstain (at least for now). It seems that the same genetic and plumage data can produce divergently opposite opinions regarding the question of the proposal itself ... and looking at the Sanin genetic data, it appears that the olrogi-oustaleti question is embedded within the wider question of species limits in albiventris vs. fuscus. It looks to me like a question of back to the drawing board - i.e., the birds in the field. Given that plumage differences appear subtle, at least to my uneducated eye, vocalizations and displays, coupled with extensive reciprocal playback experiments, may be the way to go (would make a nice doctoral thesis?).”

Comments from Pacheco: “In view of the difficulty in deciding from the information available at this moment, I choose to abstain.”

Comments from Jaramillo: “NO. I am most swayed by Santiago’s viewpoints on this problem. But also I will note something pretty remarkable about C. oustaleti, at least the nominate population. They are migratory; now, the situation might be a tad different on the Argentine side, but in Chile the majority move to the coast in winter. What is striking is that they do not just move “coastward” but they setup immediately in the splash zone of the coast and share the habitat with C. nigrofumosus (Seaside Cinclodes). This strikes me as a particularly amazing shift in habitat for the majority of the population, going from a High Andean breeding habitat to a marine habitat. This is possible through physiological adaptations that are similar to that of C. nigrofumosus. Apart from a move to marine habitats, there is a slight northward movement of oustaleti compared to the breeding range. Here is one of the papers on the physiology of oustaleti.

https://link.springer.com/article/10.1007/s00360-004-0428-2

“I think that apart from the missing details mentioned by Santiago in the discussion. It is important to note this quite different ecology, at least of the major proportion of the nominate subspecies of oustaleti. This situation of some taxa of Cinclodes being marine or partly marine is interesting, and it has occurred in different branches of the Cinclodes tree. Additionally, something unusual may be happening in the far south, with oustaleti. It appears that there are Andean breeders and perhaps marine breeders, within oustaleti. Which one of these, or both, refers to hornensis might have bearing on oustaleti (sensu stricto). But indeed this, and the Juan Fernandez population do not have direct bearing on the question at hand.

“I am open to the idea that these taxa may be conspecific, but I think we are not quite there in determining that.”

Comments from Stotz: “NO. I am uncertain what the best treatment here is; given my uncertainty, I am inclined to continue with the status quo.”

Comments from Pacheco: “NO. I took a closer look at the arguments on both sides: Pearman / Areta versus Santiago and chose to consider the same cautious position as Elisa. Despite genetic similarity, the existence of an objective diagnosis (without overlapping) and evidence of a potential reproductive isolation make me decide to keep both taxa at the species level.”