Proposal (997) to South
American Classification Committee
Treat Deconychura
longicauda (Long-tailed Woodcreeper) as three species
Note:
This is one of several situations that the IOU Working Group on Avian
Checklists has asked us to review because published treatments of the species
differ, and WGAC has to pick one.
In
this case BirdLife International split Deconychura into three species,
whereas Clements/eBird, IOU, and Howard-Moore (as well as SACC) continue to
treat it in the traditional way as one species.
Here
is the BirdLife rationale, as provided by Marshall Iliff:
“Little Long-tailed
Woodcreeper [D. typica, including also subspecies darienensis
and minor; from Honduras through Central America to N. Colombia]: In
past, sometimes considered a separate species from D. longicauda
and D. pallida, and this arrangement restituted here, based on
smaller size (male wing 91–99 mm, n=9, vs respectively 106–110, n=3, and
102–111, n=4; allow 2); spots vs streaks on breast (2); chestnut
undertail-coverts (2); clearer buff-white chin and throat (1); song a long fast
series of short piping notes, starting slowly and slowing at end, hence high
number of notes (4) and very short note lengths (4) (1). Birds from NW Colombia
(Córdoba) may be intergrades between darienensis and minor. Three subspecies
recognized.
“Northern
Long-tailed Woodcreeper [D. longicauda; the Guianas and Brazil N of the
Amazon]: Usually considered conspecific with D. typica and D. pallida
(see both). Monotypic.
“Southern
Long-tailed Woodcreeper [D. pallida, including also subspecies connectens;
W and S Amazonia]: Hitherto considered conspecific with D. longicauda
and then usually also with D. typica, but differs from latter in
characters given under that species and from former in its shorter tail (in
male 92–105 mm, n=4, vs 107–109; at least 1); whitish vs strong buffish chin
and streaks on throat to breast (1); and distinctive song, a series of c. 8
flat whistles vs a series of 6–10 long upslurred whistles (both gradually
descending in pitch), hence having much shorter notes (3) with pitch of first
note much higher (3) and note shape different (ns[2]) (1). Populations in
Andean foothills of E Ecuador apparently belong to connectens, but
adjacent lowland birds may be of race pallida; in these foothills,
either connectens or an undescribed taxon sings a very different song, a
series of c. 10–14 double notes, slightly descending in pitch, suggesting a
species-rank difference; undescribed taxon, perhaps same, reported from
foothills of NE & NC Peru (2). Recent analysis of whole genus using voice,
morphology and genetics (3) proposed promotion to full species status of the
three races currently included herein, as well the undescribed taxon from
Andean foothills; however, differences appear slight, and vocal analysis based
on very small samples, some of which may be unrepresentative. Names pallida
and connectens published simultaneously; former awarded priority by
First Reviser (4). Three subspecies recognized.
Our current SACC note reads as follows:
“109. The subspecies typica was formerly (e.g., Ridgway 1911,
Cory & Hellmayr 1925) treated (with minor and Panamanian dariensis) as a separate species from Deconychura
longicauda, but Zimmer (1934) treated it as a subspecies of D. longicauda
without comment, and this was followed by Peters (1951) and subsequent
classifications. Marantz et al. (2003)
indicated that vocal differences among populations suggest that more than one
species might be involved, with the typica group possibly more closely
related to Certhiasomus stictolaema than to Amazonian longicauda group. Barbosa (2010) found vocal evidence that D.
longicauda consists of three of more species. Boesman’s (2016f) analysis of vocalizations
also supported recognizing at least three species, and this was done by del
Hoyo & Collar (2016): D. typica of Middle American, D. longicauda
of the Guianan Shield, and D. pallida of Amazonia.
”
Here are LSUMNS specimens of the three proposed
species, top to bottom: typica (represented by darienensis), longicauda,
and pallida). Note that HBW/BLI
awarded something like 7 points in the Tobias et al. scheme based on phenotypic
characters such as throat color, undertail coverts color, and size (if I am
interpreting the numbers above correctly); make your own call on whether what
you see below would be sufficient for species rank. See also photos in Barbosa (2010), although
some of those did not reproduce well.
The vocal data in the BLI account come from
Boesman’s (2016) brief synopsis, one of 457 he prepared to evaluate vocal
differences in advance of del Hoyo & Collar (2016) for use in the Tobias et
al.’s point system: once a total of 7 points is achieved using various
morphological and vocal characters, a taxon is ranked as a species. The analysis of the Deconychura
referred to as “(3)” in the accounts is an unpublished MS thesis (Barbosa
2010). The reference to treatment as
separate species in the past likely refers to Cory & Hellmayr (1927), which
treated the complex as two species, D. typica and D. longicauda;
at that time no subspecies were described within longicauda and it thus
including all Amazonian populations. Of
interest is that the linear sequence of species in Deconychura in Cory
& Hellmayr implies that they did not consider them sister species (D.
stictolaema separating them), as is also implied in their discussion of
similar species in the footnotes.
This situation in an increasingly familiar one
in species limits in Neotropical birds: fragmentary and anecdotal information
strongly suggest that more than one species is involved yet no peer-reviewed
paper has been published that evaluates the evidence. The conundrum is whether to anticipate what a
formal analysis would indicate or wait for the formal analysis to appear.
In this case, we do have an unpublished formal
analysis, and so the conundrum also includes whether to treat an unpublished
analysis, albeit available online, as sufficient evidence. (It’s a solid thesis; Alex Aleixo was
Barbosa’s advisor, and John Bates and Jason Weckstein served on the committee.)
As for the Tobias et al. point system used by
BLI/HBW, none of the phenotypic characters mentioned in the quoted accounts
above is any more associated with species-level differences than with subspecies-level
differences. In the dendrocolaptids,
differences in all those features are routinely found between taxa treated as
subspecies because of vocal similarities.
So, it all boils down to vocal differences in my opinion.
Marantz et al. (2003) noted that the typica
group had once been considered a separate species, possibly more closely
related to then=Deconychura, now=Certhiasomus stictolaema. They
also outlined the strong vocal differences between the typica, pallida
and longicauda groups. At that
time the Andean foothill taxon was not well known and was thought to be D.
p. connectens.
Barbosa (2010) presented sonograms for each of
the 8 well-supported clades found in the genetic analysis; those 8 clades
include the 3 recognized as species by BLI/HBW plus the undescribed foothill
taxon plus and additional 3 clades within the Amazonian pallida
group.
They all appear qualitatively different, but as
noted by Boesman (2016), there are problems with taking each sonogram as
representative of the clades because of small N, differences in “motivation”
(as emphasized by Marantz et al. 2003 for interpreting woodcreeper
vocalizations in general). Further,
although longicauda and typica appear distinctive, so to varying
degrees do some of the groups within pallida. I suspect this is in part due to
non-homologous vocalizations, but it’s hard to say. Barbosa’s Discriminant Function Analysis of
standard vocal characters shows that one of the pallida clades (5) is
actually the most distinctive, and that clade 1 (longicauda sensu
stricto) is actually more similar to some of the pallida group than are
some pallida clades to each other, although this may be a consequence of
the way characters are scored.
Barbosa’s (2010) genetic data (mt DNA gene
trees) are presented below:
Note that clade 1 represents longicauda,
clade 2 the undescribed Andean foothill taxon, clades 3-7 pallida, and
clade 8 typica. My personal view
is that these data say nothing about taxon rank per se.
Recommendation: I don’t have a strong
recommendation. When I listen to
cherry-picked sample recordings, my subjective reaction is “have to be
different species”:
• typica from
Puntarenas: https://xeno-canto.org/168203 (by Chris Benesh).
• longicauda
from Sipaliwini: https://xeno-canto.org/519473 (by Rolf A. de By)
• what would become
nominate pallida (type loc. rio Purus) from Rondônia: https://xeno-canto.org/427298 (by Caio Brito).
But there are dangers in that approach, and I
can see the rationale for waiting for more detailed, published analyses,
especially given the tricky nature of dendrocolaptid vocal analysis.
In support of a split, note that typica
was formerly treated as a separate species (e.g., Ridgway 1911, Cory &
Hellmayr 1925) and was subsequently treated as conspecific with D.
longicauda without published rationale (as far as I can find).
English names: A separate proposal would be needed
for English names if this one passes.
Perhaps we can do better than the somewhat misleading and insipid names
used by BLI, but I’m not optimistic.
References:
Barbosa, I. 2010. Revisão
sistemática e filogeografia de Deconychura longicauda (Aves -
Dendrocolaptidae). MSc Thesis. Universidade Federal de Pará, Belem.
Boesman, P. 2016.
Notes on the vocalizations of the Long-tailed Woodcreeper (Deconychura
longicauda). HBW Alive
Ornithological Note 78, In: Handbook of the Birds of the World Alive. Lynx
Edicions, Barcelona.
Marantz, C. A., A. Aleixo, L. R. Bevier, and M. A. Patten. 2003. Family Dendrocolaptidae (woodcreepers).
Pp. 358-447 in "Handbook of the Birds of the World, Vol. 8.
Broadbills to tapaculos." (J. del Hoyo et al., eds.). Lynx Edicions, Barcelona.
Van Remsen, May 2024
_________________________________________________________________________________________________
Comments
from Lane:
“This is a topic that has been on my mind a long
time, and I reviewed Barbosa's dissertation work with many comments and
suggestions while it was still in the works, but it appears he incorporated few
of them, which makes me reluctant to accept his results at face value. Among my
concerns were: 1) he had a few specimens that he misidentified to taxon through
a misunderstanding of localities. 2) he made little to no effort to control for
vocal variation due to agitation to playback versus "calmer"
emotional state, and given the small sample sizes of recordings he used in his
comparisons, this resulted in magnifying differences that are probably not
real, and 3) he completely ignored minor Todd, 1917 (type loc. El
Tambor, Santander; recognized by Peters 1951) as a taxon
worth investigating, simply lumping it in with typica and darienensis,
without much rationale.
“That said, his phylogenetic tree does show that it is necessary
to split Deconychura up into several species, and there is strong voice
distinction that correlates with this. The present 3-way split (plus the
undescribed Andean taxon that Jonas Nilsson and I and others have a manuscript
in the works to describe) seems to be the best move with current phylogenetic
and voice information available to us. I think minor, presently nearly
unknown in life, may be another taxon to keep an eye on, and I believe Andres
Cuervo has re-encountered it recently, so I am interested to see how that
shakes out.
“So at present, between Barbosa's dissertation and my own
experience with several of the forms, I would say YES to splitting Deconychura
up into three named species as listed in the proposal, with the Andean taxon
yet to be described as a fourth:
D. typica (including darienensis,
but with the caveat that minor probably doesn't belong here)
D. longicauda
(monotypic)
D. pallida (including
all other Amazonian lowland forms for now).
“A more in-depth study may require more fine-tuned splitting of
the pallida complex, and minor, but for now, this is the best option.”
Comments
from Claramunt:
“YES. Songs and genetic data suggest the presence of fairly divergent lineages.
In particular, the separation of typica from the Amazonian taxa seems supported
by both. Phenotypic differences are subtle, some claimed to be diagnostic but
with no formal analysis. But subtle differences in the shape of the light spots
or the extension of spotting across the plumage should not be disregarded as
insignificant, as woodcreepers are very conservative in plumage and such
differences are associated with species-level taxa in other genera. At the very
minimum, we have to accept that the historical lumping of typica into longicauda was done without presenting any
evidence, and we don’t have any evidence today. On the contrary, songs and
mtDNA suggest a clear divergence between the two. For the separation of pallida from longicauda, the vocal and
genetic evidence also show similar levels of divergence and I don’t see any
evidence of gene flow or admixture between the nominate form and the other
Amazonian forms. Therefore, having similar evidence, the conclusion should be
the same: separate species. Once the situation of the Andean clade is clarified
we can revisit this problem.”
Comments
from Stiles:
“YES to the 3-way split of Deconychura longirostris. I note that the
species epithet longirostris was proposed in Natterer’s (unpublished?)
catalogue in the genus Dendrocolaptes, but was apparently first
published by Pelzeln (1868) in the genus Dendrocincla. The genus Deconychura
was named by Cherrie (1891) with the species epithet typica as its
type species. Again, I presume that a proposal pending regarding E-names will
be forthcoming. As an aside, I recall seeing a recommendation (but I
unfortunately can’t recall where) to discourage use of the name typica for
proposed type species because with generic transfers, this could result in
messy homonyms.”
Comments
from Robbins:
“NO. Given the widely recognized limitations of using cyt B and ND2 for
discerning species limits coupled with concerns of comparing appropriate
primary vocalizations, I lean for waiting until there is a more complete
assessment of this complex before making changes. Clearly, plumage morphology does not add
insight into species limits within this complex. I look forward to Dan et al.’s
upcoming paper that will address some of the issues that he brings up in his
comments on this proposal. At that
point, it may become clear just how many species should be recognized.”
Comments
from Jaramillo:
“YES – Thanks Dan for your comments, they persuaded me at least. I suggest
Short-tailed longtailed Woodcreeper as a name for one
of the taxa.”
Comments
from Curtis Marantz (voting for Remsen): “YES. Based on my
field experience with at least several of the Amazonian taxa, the work that we
did putting together the HBW account, and the genetic work done subsequently by
Barbosa, my feeling is that the split of this complex into three species is
well supported. I know nothing about the recent Andean taxon, and
therefore cannot comment on it, but it appears that this one can be dealt with
following the publication of a description.
“We knew even before the HBW accounts were published that there
were multiple vocal groups, with the Central American ones being especially
distinct, and both looking and sounding much like what is now recognized as Certhiasomus,
which the genetics appear to show as being quite different. The genetic, morphological, and vocal datasets
all support recognition of this complex as a full species, especially given
that there appears to be no real justification for lumping it with the
Amazonian taxa in the first place.
“The Amazonian populations present a more complex situation. Despite their morphological similarity, D.
longicauda is moderately different vocally from the members with which I am
familiar in the pallida-group. I
will nevertheless note that, in my opinion, the vocal differences between these
groups represent more a case of moderate variation on a theme rather than
totally different themes, as is true for the Amazonian versus Central American
groups. Moreover, within
"Southern" Long-tailed Woodcreeper group, which I might add is poorly
named because it occurs almost as far north as does the "Northern"
Long-tailed Woodcreeper, I find minimal variation in the songs. Looking at the sound spectrograms shown in
Figure 4 of the proposal taken from Barbosa, I would have to argue that the
representative examples were either based on a very small sample of
unrepresentative recordings, or instead, they were "cherry picked" to
show differences that are unlikely to be real. For comparison, check out the songs in the
four recordings at the links below taken from widely separated locations within
the range of the pallida-group, and clearly representing both D. l.
pallida and D. l. connectens as the ranges of each are described. (Ignore the black squares – they come with
the link and can’t be deleted without deleting the link.)
Ucayali, Peru - ML127031131 - Long-tailed Woodcreeper - Macaulay Library
|
Para, Brazil - ML115073 - Long-tailed Woodcreeper (Southern) - Macaulay
Library
|
Napo, Ecuador - ML343373841 - Long-tailed Woodcreeper - Macaulay Library
|
Amazonas, Venezuela - ML65714 - Long-tailed Woodcreeper (Southern) - Macaulay
Library
|
“Having just looked superficially at the spectrograms and listened
to a song or two from each recording, I will not deny that there could be
subtle differences in the songs included in each of these recordings, but to
me, they sound pretty much the same. Moreover, almost more so than any woodcreepers
that I know, Deconychura can get really worked up after playback, and
they can emit a wide variety of sounds, so one must be careful t compare songs given under similar motivational states,
this being most easily done by looking at natural songs that were not recorded
during a territorial dispute.
“This said, I do think the song differences between nominate D.
longicauda and the pallida-group are consistent, and when used in
combination with the genetic data, probably sufficient to treat these two
groups as separate species.
“Given the morphological, vocal, and genetic variation presented
in our HBW accounts, Barbosa's thesis, and the proposal, I do feel that
treating Deconychura as three species is warranted. The decision about whether to split these
populations up now, or instead, wait until more information is available for
the new Andean population and those in the zone between the Amazonian and
Central American forms in Colombia is
another issue. I suppose I fall more in
the camp of waiting until the whole picture is clear before making changes,
which is why we treated these populations as one species for HBW even though we
knew that more than one species was almost certainly involved. Given my experience with at least several of
the Amazonian populations, s exemplified by the recordings above, I do not see
support at this point for further subdividing the pallida-group into
multiple species.
“Given their scientific names, I might
suggest naming the typica-group the Little Woodcreeper, keeping
Long-tailed Woodcreeper for D. longicauda, and maybe Pallid Woodcreeper
for D. pallida, though admittedly, it is not overly "pallid"
as far as woodcreepers go... “
Additional
comments from Remsen:
“Curtis, inspired by Alvaro and clearly trying to audition for a voting
position on SACC for English names, further made the following, logical
suggestion, perhaps implicit in Alvaro’s:
Little Long-tailed Woodcreeper
Long-tailed Long-tailed
Woodcreeper
Short-tailed Long-tailed
Woodcreeper”
Comments from Areta: “YES. I am not enthusiastic about this kind of
splits. I would really like to see the paper describing the new taxon and
assessing the species limits in Deconychura
as a whole, which will surely provide solid nomenclatural and taxonomic
assessments. I admit that it is difficult, at this point in time, to recognize
a single species when the information is screaming that there are multiple
species. I do not see the need for rapid assessments when there is people
working in good, detailed studies that will shed light on this. This being
said, I support the 3-way split that has been around for a long time.”
Comments
from Bonaccorso:
“NO. The song differences are compelling, but the
vocal analysis is based on a handful of recordings. The plumage differences
don´t seem to add much; I think they are far from diagnostic, although that
seems to be the case for many good species in this group of birds. I think more
genetic information (nuclear loci) based on more specimens, especially at
potential contact zones, are needed.”
Comments from Andrew Spencer (voting for Del-Rio): “YES -
There's not much I can say here that hasn't already been said by previous votes
for the split, but to reiterate: the vocalizations of these groups are so
drastically different, especially in typica, that I
think the burden of proof is squarely on those who want to keep them together.
I personally feel that in the age of enormous sound collections that are
continually vetted by experts from all over the world, that requiring the
perfect rigorous analysis in cases of obviously very divergent vocalizations is
overkill. Yes, I'd love to see that analysis. But I don't think we need it to
state the obvious that has already been thoroughly vetted. Also, and this is
purely circumstantial and minor evidence, I have played the songs of southern
birds to individuals in Costa Rica and Panama and was completely ignored. Both
before and after playback of typical typica (ha that
was fun to write). And I have also played
typica to a bird at Cristalino, Mato Grosso,
with the same lack of response.”
Comments from Zimmer: “YES to the
proposed three-way split. Even given
some of the problems with the vocal analyses of Barbosa (as highlighted by Dan
and Curtis), the vocal differences between the typica-group and both longicauda
and the pallida-group are not even in the same ballpark in my opinion,
and the distinctiveness of the typica-group is further supported by
morphometric and (to a lesser extent) plumage differences, as well as by the
genetic data from Barbosa (2010). That
the Central American birds were described as a species separate from longicauda,
and subsequently lumped without justification, is all the more reason to split
them in my opinion. I would also add,
that like Andrew Spencer, I have conducted informal playback trials by
presenting individuals of typica, encountered on the Pacific Slope of
Costa Rica (specifically, in Carara NP, and the
Wilson Botanic Gardens), with playback of songs of various subspecies in the pallida-group
that I had personally recorded in Brazil, and had zero response. I have performed similar trials, using
playback of recordings of typica from Costa Rica to pallida-types
at various spots south of the Solimões/Amazon in Brazil, and had identical zero
response, and this, from birds that were highly responsive to geographically
appropriate, taxon-specific vocalizations.
So, in sum, the split of the typica-group from the others is a
slam-dunk in my opinion.
”I also feel that N bank longicauda
is vocally distinct from everything I’ve encountered on the S bank of the
Solimões/Amazon, and although these vocal distinctions are not as striking as
those of Central American versus Amazonian populations, they are, to my ears,
both significant and consistent, and the genetic data of Barbosa (2010) further
support treating nominate longicauda as distinct from the pallida-group. That’s as far as I’m willing to take it at
this point. I know nothing about the
Andean population that Dan is working on, nor do I know anything of minor,
so I have nothing to add there. As
regards the 3 additional clades that Barbosa recovered within the pallida-group,
I remain unconvinced regarding the purported vocal distinctions. Echoing the concerns of others regarding the
problems of small sample sizes, lack of proof that homologous vocalizations are
being compared, and lack of accounting for the motivational state of the
audio-recorded individuals involved, I’m reluctant to give the qualitative
distinctions evident in Barbosa’s spectrographs too much weight. The importance of the motivational state of
the recorded birds cannot, in my opinion, be emphasized enough. As Curtis states in his comments,
woodcreepers in general, but particularly these pallida-types, can get
really wound-up by playback. One thing
that I’ve noted in particular among pallida, is that when presented with
playback, they have a decided tendency to modulate the frequency of each note,
often dramatically so, which can change the entire quality of the song, in
addition to changes in pace, and total number of notes. This tendency, when coupled with small sample
sizes, could result in spectrographs that seriously inflate perceived
differences in songs. In summation, I would vote YES to the suggested 3-way
split, but hold off on any further action until Dan & co-authors sort out
some of these remaining issues.”