Proposal (881) to South American Classification Committee



Treat Lepidocolaptes falcinellus as a subspecies of L. squamatus


The current SACC notes reads:


"138. Lepidocolaptes falcinellus was formerly (e.g., Pinto 1937, Peters 1951, Meyer de Schauensee 1970, Sibley & Monroe 1990, Ridgely & Tudor 1994) considered a subspecies of L. squamatus, but Silva & Straube (1996) provided evidence for why it should be treated as a species, and this was followed by Marantz et al. (2003)."


However, the information to justify the treatment of falcinellus as a separate species from squamatus is weak, and a return to the previous classification seems warranted.


Silva & Straube (1996) analyzed 217 specimens of wagleri, squamatus, and falcinellus. In doing so, they (1) obtained measurements (which were not very informative, and are not discussed any further here) and (2) defined discrete character types for three body parts which form the bulk of their arguments.


We synthesize the data relevant to this proposal in the following summary table.


Summary table






Voice (call)





Río São Francisco





Río Paraíba/ 22°S






HEAD: 1 = brown feathers with pale shaft lines; 2 = brown feathers with a small buff or whitish spot, frequently edged with black or dusky at the tip; 3 = black feathers with broad buff stripes (see Fig. 1)

BACK: 1 = Sanford’s Brown × Burnt Sienna; 2 = Tawny × Ochraceous Tawny; 3 = Antique Brown

TAIL: 1 = Burnt Sienna; 2 = Chestnut





The main arguments by Silva & Straube (1996) are summarized in these paragraphs (see also our Summary table):


"The three allopatric populations of Scaled Woodcreeper (Fig. 4) may be separated from one another with 100% confidence by using a combination of two discrete plumage characters (HEAD and BACK). The third character (TAIL COLOR) separates the three populations into only two groups: wagleri plus squamatus and falcinellus (Fig. 2). Body measurements show little congruence among them in their pattern of variation."


"Based on the spatial congruence of the location of zones of morphological changes in the three plumage characters evaluated, Scaled Woodcreepers may be divided into three diagnosable population clusters Fig. 4): (a) one restricted to the left bank of the São Francisco River, which from now on will be referred as wagleri (Spix 1824); (b) one from the right bank of the São Francisco River to around 22°S, which will be referred to squamatus (Lichtenstein 1822), and (c) one from around 22°S to 30°S, which will be referred to falcinellus (Cabanis & Heine 1859)."


"The changes between the different types of each character are sharp and well defined. The only exception is a single specimen (MPEG 45202, female) collected by one of us (JMCS) in a locality on the left bank of the São Francisco River (Bahia, Coribe, 13°45'S, 44°28'W); the specimen has HEAD of type 2, but BACK of type 1 (Fig. 3). "


"With the exception of one single specimen (out of 11 specimens collected in the same locality) that is within the range of wagleri but has HEAD similar to that of squamatus, no other evidence of hybridization or intergradation among the three population clusters was found. "


Silva & Straube (1996) proposed considering wagleri, squamatus, and falcinellus as three separate species applying the Phylogenetic Species Concept and the Evolutionary Species Concept.  Marantz et al. (2003) provided further plumage features of all taxa and only partially followed this proposition, splitting the three taxa in two species: wagleri + squamatus on one side and falcinellus on the other.”


Genetic data

García-Moreno & Silva 1997 studied the phylogenetic relationships of several Lepidocolaptes and wrote that "The average uncorrected sequence divergence between the Lepidocolaptes species in this study is 6.3% ± 2.6 (excluding L. lacrymiger, for which we lack the full dataset, and the comparison between L. wagleri and L. squamatus, which clearly belongs to a different time frame; see below and Table 2). Assuming a substitution rate of 2% per million years (e.g., Klicka & Zink, 1997), they appear to be the product of a radiation that occurred between 1.9 and 4.5 million years ago (our estimates differ little whether we use data from ND2, cyt b, or both fragments; see Tables 1 and 2). Using the same clock, but only the cyt b data, we estimate L. lacrymiger to have split from L. souleyetii and L. albolineatus about 2.3 mya (4.6% divergence), which is well within the radiation of Lepidocolaptes. L. wagleri and L. squamatus, with a difference of only 1.7% (although diagnosable by plumage characters, see Silva & Sträube 1996), appear to have split from each other more recently (850 000 years ago). " More recently, Arbeláez-Cortés et al. (2012) used a concatenated dataset and found that squamatus-wagleri (samples from Bahía) were sister to falcinellus (samples from Uruguay and Argentina), but distant geographic samples, and partial genetic datasets precluded more detailed analyses; no genetic distance was calculated for any gene between the three taxa.

Intermediate specimens (squamatus/wagleri and squamatus/falcinellus)


In three other studies, Vasconcelos & D'Angelo Neto (2009, 2018) and Vasconcelos et al. (2012), collected specimens in areas where squamatus could meet with either falcinellus or wagleri. In the first study (Vasconcelos & D'Angelo Neto 2009), at the Serra do Juncal, on the border of Minas Gerais and São Paulo, two specimens were collected. The authors expected these to be either falcinellus or squamatus since this was an area where neither taxon had been recorded. One specimen appeared to be typical of falcinellus while the other an intermediate falcinellus-squamatus head pattern and ventral pattern of squamatus. However, the photographs of the specimens show more variation than expected, indicating that this variation needs further assessment. In two other studies (Vasconcelos et al. 2012, Vasconcelos & D'Angelo Neto 2018) report numerous intermediate specimens between wagleri and squamatus from 8 wide ranging localities (1 in Bahia and 7 in Minas Gerais, see Appendix 1 in Vasconcelos & D'Angelo Neto 2018).


Figure 3 from Vasconcelos & D'Angelo Neto (2009)



 Pajarografo Sólido:Users:javierareta:Desktop:Screen Shot 2020-08-22 at 11.53.32 AM.png



Figure 6 from Vasconcelos & D'Angelo Neto (2018)



Pajarografo Sólido:Users:javierareta:Desktop:Vasconcelos & Neto 2018 Figure 6.jpg



Additionally, Vasconcelos & D'Angelo Neto (2009) elaborated on the hybridization between squamatus and falcinellus:


"This is the first record of intergradation between these two species (see Silva & Straube, 1996; Marantz et al., 2003). Furthermore, L. falcinellus was known only from southern bank of Paraíba do Sul river (see Marantz et al., 2003) and these records show that at least phenotypes of this species occurs north of this river. The occurrence of phenotypes of L. falcinellus in the Araucaria forests of the north slopes of Serra da Mantiqueira can be related to palaeoecological connections to the nucleus of Araucaria forests from southern Brazil (where L. falcinellus is the only species of this complex – see Marantz et al., 2003). In the Serra da Mantiqueira and adjacent regions, expansion of Araucaria forests is hypothesized to have occurred between 9,7008,200 years before present and, later, after c. 3,5003,000 years before present, when climate became cooler and moister than today (Behling, 1997, 1998, 2002; Garcia et al., 2004). Curiously, L. squamatus is the taxa recorded at Itatiaia, in the Atlantic (south) slope of this mountain range (Pinto, 1951, 1954). This could be explained by the occurrence of Araucaria forests in the north slopes of Serra da Mantiqueira, whilst the Atlantic slope is mostly covered by montane and cloud forests (see Hueck, 1972) and should has different zoogeographical affinities."


A different interpretation

We interpret the available data in a different manner to Silva and Straube (1996) and consider that the evidence indicates that there is a single, widespread, biological species at stake. First of all, recently collected specimens have shown that the presumed distributional limits of squamatus, falcinellus and wagleri are different from those proposed by Silva and Straube (1996), and that the presumed barriers do not exist. Second, the fact that the plumage features exhibit stepwise leaps and that squamatus appears to be intermediate between wagleri and falcinellus in crown and back colour suggests that they are linked through a cline. It might be a steep or a stepped cline, we do not know, but the very nature of the typological analyses carried out by Silva & Straube (1996) does not seem particularly suited to detect such a variation, while it will clearly be able to detect the main geographical leaps in variation. Third, intermediate individuals have been collected where each form meets the nearest taxon (with numerous intermediates between squamatus and wagleri, and few known intermediates between squamatus and falcinellus), adding weight to the argument that falcinellus and squamatus are conspecific.

         Additionally, although no thorough vocal analyses are available, the call (the most frequently heard and recorded vocalization in this complex) is surprisingly similar across the geographic range, with much variation within each taxon but nothing consistent that we could detect among them (see the table above for some examples). We refrain from analyzing songs here, as there appear to be few recordings and variation is not well understood. Many recordings labelled as "song" in Xeno-Canto and the MLNS pertain to calls, while few "songs" are available for all taxa. Some voices seem different (perhaps not all homologues:,,,, other seem identical (,, and various other recordings at the Macaulay Library seem misidentified (descending vocalizations; sound like Xiphorhynchus fuscus tenuirostris). With so few samples to compare and the generally variable nature of long songs in Lepidocolaptes, we cannot confidently assess what is going on. However, any such analysis would need to include samples from across the presumed hybrid zones between taxa if it is to rigorously test the species limits between them.

         In sum, the lack of complete genetic data, minor plumage differences linked through intermediates where their ranges abut, and vocal similarities, we propose a return to the historical treatment of falcinellus (and wagleri) as subspecies of squamatus.

         We thank Vitor Piacentini for questioning a previous, less detailed proposal, which prompted us to look for further evidence on the taxonomic status of these woodcreepers, resulting in a much more interesting and richer perspective.


Additional references

ArbeláezCortés, E., NavarroSigüenza, A. G. & GarcíaMoreno J. (2012). Phylogeny of woodcreepers of the genus Lepidocolaptes (Aves, Furnariidae), a widespread Neotropical taxon. Zoologica Scripta 41: 363–373.

García-Moreno, J. & Silva, J. M. C. (1997). An interplay between forest and non-forest South American avifaunas suggested by a phylogeny of Lepidocolaptes woodcreepers (Dendrocolaptinae). Studies on Neotropical Fauna and Environment 32: 164–173.

Vasconcelos, M.F. & D'Angelo Neto, S. (2009) First assessment of the avifauna of Araucaria forests and other habitats from extreme southern Minas Gerais, Serra da Mantiqueira, Brazil, with notes on biogeography and conservation. Papeis Avulsos de Zoologia. 49: 49-71.

Vasconcelos, M.F., Souza, L.N., Duca, C., Pacheco, J.F., Parrini, R., Serpa, G.A., Albano, C., Abreu, C.R.M., Santos, S.S. & Fonseca-Neto, F.P. 2012. The avifauna of Brejinho das Ametistas, Bahia, Brazil: birds in a caatinga-cerrado transitional zone, with comments on taxonomy and biogeography. Revista Brasileira de Ornitologia 20: 246-267.

Vasconcelos, M.F. & D'Angelo Neto, S. (2018) First avifaunal survey of a Cerrado dry forest enclave on the right bank of the São Francisco River, Minas Gerais, Brazil, with insights on geographic variation of some species. Papeis Avulsos de Zoologia 58: e20185815



Juan I. Areta and Mark Pearman, August 2020




Comments from Stiles: “YES to placing falcinellus as a subspecies of squamatus. Clearly, more data on distribution and a genetic analysis of haplotype sharing are needed, and retaining all three subspecies as part of a single species seems the best course with the data available.”


Comments from Pérez-Éman: “This proposal aims to treat Lepidocolaptes falcinellus as a subspecies of L. squamatus. The basis for this lump is summarized at the end of the proposal and it includes four main points: the lack of complete genetic data, minor plumage differences, intermediate forms in the contact zones, and vocal similarities. I think decisions could go either way depending on the importance given to particular criteria and/or the presence/lack of information. My take on each of the four main points follows.


“1. Molecular sequence data available for these taxa is few and widely scattered throughout the literature. Arbeláez-Cortés et al. (2012) used concatenated sequences of both mitochondrial cytochrome oxidase subunit 1 (CO1) and NADH subunit 2 (ND2) genes (with sequences of each gene coming from different individuals, known as chimeric sequences) to estimate the phylogenetic relationships among species in the genus Lepidocolaptes. For falcinellus, sequences came from individuals of both Misiones (Argentina) and Cerro Largo (Uruguay), whereas, for squamatus, both of them came from Bahia, but from either side of the San Francisco River (Coribe and Palmas Monte, Bahia, Brazil), potentially including individuals of both wagleri and nominal squamatus. This strategy might have not influenced the phylogenetic goals of the study but certainly might mask some of the genetic variation between these two taxa. In fact, preliminary and rather short sequences (cytochrome b and ND2 genes) from the same squamatus individuals were produced by García-Moreno & Silva (1997) but were identified as both wagleri and squamatus (and showed a moderate genetic divergence; see below). It is worth to notice that the only individual with intermediate plumage characters between these two taxa, reported by Silva & Straube (1996), was found in Coribe, Bahia (Brazil). Other CO1 gene sequences available are from Chaves et al. (2015, Molecular Ecology Resources) and Klippel et al. (2015, PLoS ONE), both coming from work on DNA barcoding. Chaves et al. (2015) sequence is from an individual collected in Brejinho das Ametistas, Caetité (Bahia, Brazil) and the Klippel et al. (2015) is from northern Espiritu Santo. Unfortunately, it seems to be unclear how to label both Coribe and Caetité individuals, as intermediate individuals have been found in both localities at either side of the San Francisco River in Bahia (Silva & Straube 1996, Vasconcelos & D´Angelo Neto 2018). Information on phylogenetic relationships and genetic divergence can be extracted from these sequences. Arbeláez-Cortés et al. (2012) phylogenetic hypothesis of the genus Lepidocolaptes, based on the concatenated dataset (CO1+ND2), showed a strongly supported sister-taxon relationship between squamatus and falcinellus, a result that is congruent with results based on just ND2 sequences or genomic data (UCEs) (posterior probabilities of 1.0 in Bayesian analysis and Maximum Likelihood bootstrap support of 100; see my comments on Proposal 868 that include both phylogenetic hypothesis). An uncorrected genetic divergence of 1.7% between wagleri and squamatus, based on both cytochrome b and ND2 genes, was reported by García-Moreno & Silva (1997); however, these data came from very short sequences a bit longer than 200 base pairs (a normal sequence length obtained at that time). Unfortunately, these are the only two samples sequenced for these genes. For CO1, uncorrected genetic divergence ranges between 0.5 and 1.5%, with the northern Espiritu Santo sample the most divergent, but keep in mind the Bahia samples came from a contact zone with reports of intermediate plumage individuals. ND2 uncorrected genetic divergence between falcinellus and squamatus averaged 4.35% (only two individuals for each taxon, with the additional falcinellus coming potentially from Brazil (Rodrigues et al. (2013) but collection locality requires the missing Supplementary Information on this article). Divergence between both falcinellus samples was 0.4%. How does this genetic divergence between falcinellus and squamatus compare to other species pairs? These are five examples of average uncorrected genetic divergence: affinis/leucogaster = 4.2%, angustirostris/souleyetti = 4.6%, albolineatus/angustirostris = 4.5%, fatimalimae/angustirostris = 4.7%, and angustirostris/squamatus = 4.0%. Average uncorrected genetic distance within the genus is approximately 7%, a bit higher than the 6.3% previously reported by García-Moreno & Silva (1997) given that L. lacrymiger, the most divergent species (approx. 7%) was not included in such calculations. In summary, molecular data support a sister-taxon relationship between falcinellus and squamatus (including here both wagleri and nominal squamatus), and uncorrected genetic divergence between both taxa is at the level of several species pairs within Lepidocolaptes; a lower divergence is found between potential individuals of both wagleri and squamatus.


“2. Plumage differences among the three taxa (wagleri, squamatus, and falcinellus) seem to be minor. However, based on Silva & Straube (1996), they are consistent at the broad spatial scale they were evaluated. L. falcinellus differed from the first two taxa in the complete set of three characters, while wagleri and squamatus just in head and back plumage pattern and coloration (similar in tail coloration). However, more recently, Vasconcelos & D´Angelo Neto (2009, 2018) found individuals with intermediate plumage patterns in two of the potential contact zones. Fourteen (14) out of 29 individuals of squamatus (and wagleri) collected mostly at the right bank of the San Francisco River showed intermediate plumage patterns (at both sides of the river). A similar pattern, though with less number of individuals, was found in Serra da Mantiqueira where individuals of the falcinellus phenotype were collected, known previously only to the south of the Paraíba do Sul river. Reading the description of these intermediate forms, you can come to the conclusion that plumage comparative studies need to include more characters than the originally considered by Silva & Straube (1996); for example, underpart colors (breast and belly), have shown to be very informative when comparing variation among these taxa and were not considered by these authors. Besides, such descriptions should be thorough and emphasize contact zones and the geographical distribution of such intermediate phenotypes.


“3. I have already commented on the intermediate forms issue. These findings suggest hybridization between these taxa (currently or in the past) and, if so, clearly shows that the potential riverine barriers associated to geographical limits of these taxa are not unsurmountable or physical/climatic conditions have been dynamics through the history of the region. Environmental conditions have changed and have affected many taxa in the Atlantic Forest Region, a pattern that is clearly shown in the literature. Studies cited by Vasconcelos & D´Angelo Neto (2009) indicate that, at least, the area with potential gene flow between falcinellus and squamatus has been clearly dynamic in the recent history. This information, however, does not help to understand the dynamics of this contact/hybrid zone. Is there current gene flow? Is symmetric or asymmetric? is the hybrid zone expanding or forces such as natural selection are counteracting a potential expansion, such as the findings recently found for the Baltimore and Bullock´s oriole hybrid zone by Walsh et al. (2020, Auk)? I think we have an interesting setting to expand on these findings and conduct a morphological/molecular/behavioral study that provide with some answers to these questions.


“4. Vocal similarities/differences are the weakest point here because there is no much information and the available one is complex to evaluate because of its variable nature. In fact, these were the reasons the proposal did not provide any vocal analyses. However, Vitor Piacentini´s comments on the differences among songs of these taxa and the distinct use of their repertoire, even when not evaluated quantitatively, clearly suggest that there is a story beyond what is currently available that needs to be evaluated before taking any taxonomic decision.”


“In summary, falcinellus and squamatus are sister taxa with a level of divergence comparable to many species pairs within the genus Lepidocolaptes. Even when genetic distance does not represent a yardstick to diagnose speciation events, such a level of divergence suggest a long evolutionary history in isolation. At this point, one still could go either way and look for information in other characters. Plumage pattern though minor are consistent (apparently) in large ranges of their distribution and intermediate plumages are concentrated (so far) in contact zones (more extensive for wagleri/squamatus with intermediate forms found at both sides of the San Francisco River in Bahia and Minas Gerais, Brazil). If these intermediate forms are evidence of hybridization, the only existence of an hybrid zone does not invalidate the recognition of species level for both taxa. We need to study the dynamics of such hybrid zone. A pattern that requires further study is the potential altitudinally replacement pattern, when in parapatry in Rio de Janeiro, with falcinellus occupying the upper portion of the gradient (Piacentini´s comments). Finally, a vocal study is missing and the lack of evidence is not evidence. The pattern found for these two taxa is very similar to the one found for Myrmoderus squamosus/loricatus (study case mentioned by Vitor Piacentini): sister taxa, a potential origin associated to vicariant geological events previous to the Pleistocene and a dynamic demographic history due to a recent history of climate changes (Raposo do Amaral et al. 2013, Molecular Ecology). These Myrmoderus species are somehow similar in plumage pattern and vocalizations with a replacement zone associated to the Paraíba do Sul river. Some (few) hybrids have been found but apparently minor differences in songs have a strong effect in species discrimination, competition, and possibly mate recognition (Macedo et al. 2019, Behavioral Ecology). In conclusion, my position is similar to the one expressed by Vitor Piacentini and, for the time being, is to keep falcinellus and squamatus as different species with this last taxon including both nominal squamatus and wagleri (the Silva & Straube (1996) proposal considering the Biological Species Concept). There is no clear available evidence so far that justify a treatment in which these two taxa are better considered as a single species. There are important gaps that need to be filled if we want to make informed decisions. It is true that we also lack the evidence to treat them, confidently, as different species but a return to a historical one species treatment without evidence is to risk the need for further adjustments in the near future.”


Comments from Claramunt: “NO. I think that the interpretation in the proposal is reasonable, but we need to see some actual evidence. Is variation clinal? We need to see plots and maps. Is there a hybrid zone between squamatus and falcinellus?  It seems so, but is the hybrid zone wide and with lots of intergrades or narrow and composed of sterile F1s?  Narrow hybrid zones are not evidence of conspecificity in modern species concepts. I think that a published paper is needed to flesh out the required information.”


Comments from Robbins: “I’m on the fence on this one.  If we hadn’t already recognized falcinellus as a species I would lean towards treating it as a subspecies of squamatus given the incompleteness of some of the data sets.  However, instead of flipping back and forth, like Jorge and Vitor, I’m for continuing our current treatment until there are additional data to suggest otherwise.”


Comments from Zimmer: “NO. I’m in complete agreement with Vítor’s comments on this topic.  His observations regarding differences in the songs (as opposed to calls) of falcinellus relative to those of squamatus, and, especially, to the behavioral differences between the two in the relative frequency with which the two taxa use songs versus calls, squares very well with my experience.  I have little doubt that a quantitative vocal analysis of songs of falcinellus and squamatus would reveal diagnostic differences in more than one vocal character, but Areta & Pearman admit in their Proposal, that they excluded songs altogether from their analysis due to small sample sizes of recordings of songs, and variation in those small samples that was “not well understood.”  Even the comparison of calls across taxa cited in the Proposal appears to have been a qualitative comparison rather than a quantified, thorough analysis.  I agree that the calls of the three taxa (including wagleri) in this complex are quite similar to the human ear (as are the calls of some other Lepidocolaptes, such as angustirostris), but I would not be at all surprised to find that spectrographic analysis might reveal diagnostic between-taxon differences in the note shapes, degree of frequency modulation, peak frequencies and note width (duration) of those calls, even if those differences are not readily apparent to the human ear.  Differential usage frequency within vocal repertoires is also important, as Vítor suggests.  [As an aside, I would liken the difference in usage frequency of songs versus calls between falcinellus versus squamatus to that seen between N bank (of Amazon) versus S bank populations of Lophotriccus galeatus.  N bank galeatus seem to deliver lengthy or intermittent series of “tic” notes as their primary vocalization, only occasionally reverting to the trilled, gravelly song.  S bank populations on the other hand, routinely deliver the trilled song, with only occasional “tic” calls as punctuation.  Similar geographic differences in usage rates of analogous/homologous vocalizations have also been noted in populations of  the Zimmerius gracilipes/acer complex, the members of which are now treated as distinct species.]   The point is, no thorough analysis of vocalizations in this complex has been undertaken, and given consistent plumage differences and the genetic distance between falcinellus and squamatus (which, as Jorge notes, reveals comparable levels of divergence to that seen in other species-pairs in Lepidocolaptes, and suggests a long evolutionary history in isolation), I would argue that the burden of proof for producing such an analysis would fall on those arguing for a one-species treatment, particularly in light of the anecdotal observations presented by Vítor and myself regarding differences in songs (available for qualitative comparison in publicly accessible sound archives, such as Xeno-canto), and differential song versus call frequency within repertoires.  The number of Lepidocolaptes specimens from the contact zone in Bahia and Minas Gerais exhibiting plumage characters seemingly intermediate between wagleri and squamatus, adds further weight to the treatment advocated by Marantz et al. (2003), who considered wagleri and squamatus as conspecific (while treating falcinellus as a separate species).  The scarcity of such intermediates between falcinellus and squamatus from the contact zone, along with the elevational parapatry of the two taxa in southern Rio de Janeiro, bolsters the case for continuing to recognize two species, and suggests that any “hybrid zone” is very narrow.  To me, falcinellus is very much a bird of the upland Araucaria-Podocarpus forests that typify higher elevations of Paraná, Santa Catarina and Rio Grande do Sul, and, as such, is an ecologically different beast from squamatus/wagleri, an impression that is only reinforced by the elevational parapatry of the two species in the Serra da Bocaina as described by Vítor.”


Comments from Jaramillo: “NO – I will admit I am most influenced by Kevin’s analysis of the situation here, thanks Kevin for your thoughts. On the other hand, the proposal is a reasonable proposal. There is clearly a need for more data, in particular song as noted. I am usually not the person who asks for more data, I tend to recoil at the thought because often this is kicking down the road a problem that we already can act on to a later date due to a bit missing details on only part of the problem. In this case, I think the available data can be interpreted in two ways, essentially it is on the fence, rather than there being an obvious choice.”


Comments from Lane: “NO. I am swayed by Kevin's and Vitor's comments, and believe the prudent stance is to maintain the current status quo until additional fieldwork addressing the remaining issues is conducted and results presented.”


Comments from Pacheco: “NO. In this case, I have a position similar to Kevin and Dan. It is preferable to wait for additional field data to resolve the treatment given here.”


Comments from Bonaccorso: “NO. More evidence is needed to decide, so we don´t have to revert this decision in a few years. A 4% difference in mitochondrial DNA suggests that there might be some genetic differentiation in fast introns or SNPs; looking into this type of evidence may help solve the issue. Also, more quantitative data on vocalizations is needed. I am not worried about intermediates if they are confined to a narrow hybrid zone. I think this is the most conservative approach to the issue, at least until more evidence is available.”