Proposals (408-411) to South
American Classification Committee
408. Change
linear order of current Thryothorus wrens
(II)
409. Recognize
genus Pheugopedius (II)
410. Recognize
genus Thryophilus (II)
411. Recognize
genus Cantorchilus (II)
Summary: This proposal would result in a
new genus name or names for the current SACC Thryothorus wrens. In
Proposal 219, all
committee members were in favour of taking such a step, but there was no
agreement on using Mann et al. (2006)'s three-genus approach for South American
Thryothorus. This decision should be reconsidered in light
of vocal analysis by Mann et al. (2009).
However, if the three genus approach is again to be rejected, then SACC
should either revert to Hellmayr's two genus treatment for S American Thryothorus or move them to other genera. The
current baseline remains as Thryothorus for
all these species, which cannot be
supported.
Discussion: In Proposal 219, in a close vote,
SACC decided not to split S American Thryothorus
into three new genera. However, there
was considerable support in the comments of dissenters for moving all South
American Thryothorus to Pheugopedius. The proposition that Thryothorus is not monophyletic has strong support from multiple
studies, as discussed in Proposal 219 and further shown by Mann et al. (2006)
and papers cited therein, with true Thryothorus
occurring only outside the SACC region.
There was some scepticism over
Proposal 219 given that Mann et al. (2006)'s diagnosis of the new genus Cantorchilus involved only molecular
characters: "Currently, no known
uniquely derived morphological characteristics diagnose the genus Cantorchilus,
as defined here. Given current taxonomic and character sampling, the genus is
diagnosable by 9 unreversed synapomorphies in cytochrome b (including: A150C,
A156G, C297T, A876C, C903A, C924A, A948G, C960A and C1116A, where aNb refers to
the ancestral state a and derived state b at position N), all at third codon
positions, including 6 transitions and 3 transversions, one of which results in
an amino acid replacement (I372M)."
Daniel Cadena (who voted in favour of
adopting the three genera) in his comments on Proposal 219 noted "it appears to me that the songs of
Thryophilus (rufalbus and nicefori, which I am familiar with) stand out as
rather unique when compared to those of other "Thryothorus" I know." This has now been borne out with
analysis. Mann et al. (2009), based on
sampling of almost all relevant species, showed the three proposed genera to be
diagnosable on the basis of duetting behaviour and also noted some behavioural
differences (separation by forest strata) between members of different proposed
new genera where they occur together.
For example, Thryophilus sensu
stricto don't duet or duet in a very temporally uncoordinated fashion, whereas Cantorchilus show the highest degree of
temporal coordination. Van Remsen (who
voted in favour of three genera) noted previously (in summarising the comments
of dissenters) that the "three
separate genera … are … morphologically "below" the traditional level
of generic distinctiveness".
However, Mann et al. (2009) suggest that the three groups are
diagnosable by ecological characters in addition to molecular characters. As a result, the proposed genera are at least
as supportable as some of the N American wrens related to this group.
Given the committee member comments
on Proposal 219, changing the genus for S American Thryothorus would appear to be unfinished business. For those that consider the three genus
approach still to be weakly supported, there are two options: (i) reverting to
Hellmayr's two genus treatment, which could be regarded as a status quo ante when Thryothorus is properly restricted to N
America; or (ii) place all S American Thryothorus
in Pheugopedius. In either of these treatments, the
non-recognized genera of Mann et al. (2006) could be treated as
sub-genera. If the SACC later comes to
the conclusion that a greater Pheugopedius
is not monophyletic, as suggested by Mann et al. (2006, 2009), then a further
rearrangement could be considered in future, but this would affect less than
half of the species currently recognised in this group if Pheugopedius is used, or just two species if SACC reverts to
Hellmayr's treatment.
An unfortunate consequence of Proposal
219 being rejected was that the proposed new linear order was also
rejected. Proposal 408 would re-arrange
the genus (whatever its name) so as to reflect the relationships suggested by
Mann et al. (2006) and recommended in Proposal 261, which maintain the existing
order generally but rearrange to show the suggested three (sub-) genera
(together with minor tweaks to the order of the Cantorchilus group), with "incertae sedis" (within Cantorchilus) griseus at the end of the order:
Old linear order |
Proposed new linear order |
spadix |
(sub-)genus Pheugopedius: spadix (sub-)genus Thryophilus: rufalbus (sub-)genus Cantorchilus: leucopogon |
A positive vote on Proposal 408
would result in this new linear order being adopted.
As a result of Proposal 219 not
passing, all Thryothorus were moved to Pheugopedius
in the Colombian checklist three years ago (Salaman et al. 2007, 2008,
2009) assuming that SACC would also do so in due course based on comments of
committee members. One of the authors of
this proposal has sought to adopt such an approach in some journal papers
including site/regional checklists and range extensions, but has become
somewhat fed up with editors and peer reviewers commenting that these birds
should be Thryothorus due to the
SACC's treatment (particularly for the "core" Pheugopedius that would be recognised in this genus under Mann et
al.'s suggested treatment where 10/10 SACC members supported recognising this
genus!). SACC action on this matter
would be welcomed as other persons working with Neotropical birds may have had
similar experiences.
(Another approach would be to merge Henicorhina, Uropsila, Cinnycerthia,
Cyphorhinus and S American Thryothorus
into one genus, but that would cause more nomenclatural instability than it
is worth and rob us of some well-defined genera.)
The following votes would have the
following results, assuming that proposal 408 (on linear order) passes:
409 passes, 410 rejected (or 409 and 411 pass
but 410 does not): |
409 and 410 pass, 411 does not: |
409, 410 and 411 pass: |
Pheugopedius: spadix |
Pheugopedius: spadix Thryophilus: rufalbus |
Pheugopedius: spadix Thryophilus: rufalbus Cantorchilus: leucopogon |
References:
Mann, N.I., et
al. 2006. Molecular data delineate four genera of "Thryothorus"
wrens. Mol. Phylogenet. Evol. 40:
750-759. http://www.tc.umn.edu/~barke042/Publications.php
Mann, N.I.,
Dingess, K.A., Barker, F.K., Graves, J.A., Slater, P.J.B. 2009. A comparative
study of song form and duetting in Neotropical Thryothorus wrens. Behaviour 146, 1-43
Recommendations: (From Thomas Donegan): I would strongly recommend a
YES vote to 408 and 409; for votes 410-411, I can see reasons why committee
members might vote either way. My own
view is that a yes vote is more sensible in both instances. Mann et al. (2006, 2009) taken together
provide a solid rationale for their recommendations and have clearly considered
the issues at length. Their diagnosis of
the three new genera now has real world as well as molecular support. Given that all the genus names of these
species will be changed anyway, a priori,
their treatment should probably best be followed rather than being second
guessed.
(From Keith
Barker): I would strongly recommend a YES vote for all four of these proposals,
for the reasons set out in Mann et al. (2006, 2009). However, if Cantorchilus is not to be recognized, I would strongly recommend
voting yes on 408-110, taking us to the status
quo ante Hellmayr’s lumping of all within Thryothorus, and avoiding placement of taxa into Pheugopedius that have never been there.
Thomas Donegan & F. Keith Barker, August 2009
=====================================================
Comments
from Stiles:
“408. YES. The new linear order presented clearly reflects phylogeny
better than our current arrangement.
“409. YES, to recognize Pheugopedius;
the Neotropical species clearly do not belong in Thryothorus; their inclusion would render this genus polyphyletic.
“From here on, however, things get a bit dicier. The
proposal contains some confusion regarding whether the current practice of
lumping all into Thryothorus vs. the “two-genus” approach of recognizing
Thryophilus and Pheugopedius is due to Hellmayr or is “pre-Hellmayr”. Looking up what Hellmayr actually wrote in
“birds of the Americas” (1934), I find that he indeed lumped all the Neotropical
forms (as well as ludovicianus, the
only “true” Thryothorus) into a
single genus. However, the idea was not
original with him; he was following van Rossem (Trans. San Diego Soc. Nat.
Hist. 6, p. 208, date not given) in stating:
“…the differences between Thryophilus,
with open nostrils, and Pheugopedius,
with partly operculate nasal groove, is so completely bridged by intermediate
species that no dividing line can be drawn.
Moreover, the two types of nostrils, used as criteria for generic
distinction, even occur within the same species, the case of T. modestus being very appropriately
cited by van Rossem as a striking example of such variation. If Pheugopedius
and Thryophilus be merged, there
is no valid ground for the retention (sic) of Thryothorus, since a good many species of “Thryophilus” agree with the Carolina Wren in the lesser graduation
of the tail.”
“If these were the only features defining Pheugopedius and Thryophilus, their lumping would seem to be in order, but Hellmayr
did not give diagnoses for genera. I therefore went to Ridgway (Birds of N. and
Middle America, USNM Bull. 50, vol. 3, 1904) for these. The following table summarizes the points of
difference that I was able to extract from Ridgway’s diagnoses:
Pheugopedius |
feature |
Thryophilus |
Longitudinal to fusiform |
Nostril |
Small, round or oval |
Decumbent operculum overhangs nostril, posterior end contacting with
or overhung by feathers of nasofrontal antiae |
Operculum |
No operculum, but a naked membrane may extend along upper edge of nostril,
separated from nasofrontal antiae |
Graduated for 1/3 or more of its length |
Tail |
Graduated for 1/3 or less of its length |
Straight for ≥ ½ of its length, abruptly decurved terminally |
Exposed culmen |
Straight to slightly decurved, gradually decurved terminally |
Slightly concave, deflexed basally, slight indication of subterminal
notch |
Maxillary tomium |
Straight to slightly decurved terminally with distinct subterminal
notch |
Short but distinct, occasionally 1-2 well developed |
Rictal bristles |
Obvious, 1-3 well developed |
Short, rounded; primaries 3-7 or 4-6 longest, 8<<3 |
Wing shape |
Moderate, rounded; primaries 4-7 longest, 8 variable |
D(igit) 4 w/o claw reaches to middle of subterminal phalanx of d3,
claw reaches base of claw of d3 |
Length of toes |
D4 w/o claw reaches subterminal joint of d4, claw does not reach base
of claw of d3 |
As a basis for separating genera, this is cutting it pretty fine, but is
fairly typical of generic diagnoses of Ridgway’s time. I then went to our collection to check out
these characters. Surprisingly, I found
that in specimens in which the nasal area was intact, the difference cited by
Ridgway does seem to hold in most cases.
All species show a naked membrane partly covering the nostrils but in
all of the Pheugopedius, this extends
from the dorsal edge of the nasal fossa and does partly cover the nostril –
although in some where the membrane has been pulled back, the nostril appears
more rounded. In Thryophilus, this membrane extends from the proximal end of the
fossa distally to the nostril, which does indeed appear more rounded – but in a
few the dorsal extension is more pronounced, though never so much as in the Pheugopedius. (We do not have T. modestus here, so I could not check directly van Rossem’s
statement.) The tail is indeed more
graduated in Pheugopedius, but in
some Thryophilus it approaches pretty
closely this degree of graduation.
Regarding the curvature of the culmen, there is much variation within
species: the tendency for a more strongly decurved tip in Pheugopedius holds, but some (young?) individuals show a more
gradual curvature; a few individuals of several species in Thryophilus have more abruptly curved bill tips, though not so
pronounced as in most Pheugopedius. The same goes for the subterminal notch:
though on average it is more pronounced in Thryophilus,
one can find the full range of variation in virtually all species. I did not go into detail regarding wing
shape, but again, primary 3 seems to average relatively longer in Pheugopedius but the difference is
pretty slight. I could see no
consistent difference in the relative length of the toes, and the rictal
bristles seemed to vary more according to preparation than anything else. In conclusion, there is some morphological
basis for recognition of Thryophilus as
distinct from Pheugopedius but the
differences are often slight and bridged by individual variation in many cases
(perhaps related to age for bill shape?).
Given that these differences, such as they are, do correlate with the
genetic data, I could be induced to vote a lukewarm YES on proposal 410 to
recognize Thryophilus (although I
would not be surprised – or heartbroken - if many SACC members disagreed with
me).
“However, I could find
no morphological character(s) that would convincingly separate Thryophilus (sensu stricto) from Cantorchilus.
I am reluctant to base a genus on
zero morphological differences (but if Keith or Thomas wish to conduct a more
thorough morphological analysis and do find something, I could go along with
it). The song differences are
interesting, but I am also reluctant to use such differences to separate genera
in oscines where song learning is pronounced and local song dialects are frequent. In species like rufalbus, sometimes one hears perfect coordination between males
and females (the female adding her notes to the end of the male’s with no
discernable break, as in modestus, for
example), at other times (especially in minlosi
of the Llanos) coordination seems lacking or less developed. Hence, I will vote NO on 411, at least for
the present, though I am willing to recognize Cantorchilus as a subgenus of Thryophilus
(or Pheugopedius) based on the
genetics. In passing, I note that the recognition
of species in Scytalopus is based to
a great extent upon song differences that line up with genetic differences, but
here we are dealing at the species, not genus level in a group in which song
learning is presumed to be absent.
“I also note that no
genetic data are available for T. griseus,
which is morphologically much more distinctive than anything in Pheugopedius or Thryophilus with its much shorter and nearly non-graduated tail and
grey coloration (and canopy habitat?).
Not having the song paper I don’t know if its song was included in the
pertinent paper of Mann et al., but on the face of it, it might well be worth
looking at (we have a recent specimen here, Daniel!)”
Comments from Robbins: “YES.
I’m on the fence on this one, as it really doesn’t bother me
that Cantorchilus is defined
primarily by genetics (vocal data are suggestive). The morphological differences between the Pheugopedius and Thryophilus clades are so slight that if those were the only data
supporting recognition of those two genera then I would vote no on that
proposal. Thus, in an attempt to be
consistent, I vote for the recognizing Cantorchilus,
although I’m not totally convinced that this is the right course of action.”
Comments from Zimmer: “
“Proposal #408: Change the linear order of current Thryothorus wrens. YES, our current arrangement does not square
with the available evidence.
“Proposal #409: Recognize Pheugopedius. Again, YES, as the evidence is pretty clear
that Thryothorus applies only to ludovicianus/albinucha, and that all of
the South American species belong somewhere else.
“Proposal #410: Recognize Thryophilus. YES, based on the combination of molecular evidence,
vocal evidence (as outlined in Mann et
al. 2009), and Gary’s assessment of Ridgway’s diagnosis of Pheugopedius versus Thryophilus.
“Proposal #411: Recognize Cantorchilus. Gary’s points on this one are well
taken. Although there is definite morphological
cohesion between most of the species that would be placed in Cantorchilus (superciliaris, leucotis, longirostris, guarayanus), there is little
if anything in the way of morphological characters that would allow the genus
to be diagnosed relative to Thryophilus. I also share Gary’s hesitance regarding the
use of vocal differences to define a genus of oscine passerines, which can
learn their songs. However, I think that
Mann et al are really on to something
as regards vocal differences in duetting behaviors in these wrens. When you read through the list of different
singing styles (Table 4 in Mann et al. 2009)
it is somewhat confusing because of the sheer complexity of the vocalizations
and duetting behaviors involved. But, if
you have field experience with the species involved, you can see the patterns
that Mann et al refer to, and many of
the species in each proposed “new” genus do sort out as having similar singing
“styles”. It seems to me that the
underpinnings that allow males and females of one species to frequently engage
in highly coordinated, complex, antiphonal duets, as opposed to those species
in which duetting is relatively rare and/or less coordinated are more likely to
be genetically based than are simple differences in song dialects, which, in
this family, can obviously be learned.
We still don’t really have a handle on what constitutes genus-level
vocal distinctions, although I am reminded of that old quote regarding
obscenity – “we can’t really define it, but we know it when we see it”. Clearly, there is more work to be done in
refining some of Mann et al’s song
style categories, and, as the authors themselves point out, not every species
in each proposed genus displays the same song styles. However, the authors have provided a good
platform for investigating the extent of the genetic basis for the differences
in vocal behavior in these wrens, and I do think that where there is smoke
there is fire.
“Gary raised the question of T. griseus, which would seem to be a
potential fly-in-the-ointment for any proposed phylogenetic arrangement,
because without DNA we really don’t know where it belongs. Morphologically, it is more different from
all other “Thryothorus” than any of
them are from one another. And no, Mann et al 2009 did not have vocal data on griseus in their paper. However, I find it interesting that under
their proposed arrangement, griseus
would be in Cantorchilus, as would leucopogon and thoracicus (of Central America).
I recently published a paper detailing observations on vocalizations
(including duetting behavior), ecology, and nesting of T. griseus (Zimmer & Whittaker 2009, Cotinga 31:80-85), and off the top of my head, I would say that the
singing style of griseus would be
closest to Style 9 as defined by Mann et
al. (which was published while our Gray Wren paper was in press), with
males having different categories of song (series of single tonal notes as well
as more complex phrases, both of which are repeated and increase in amplitude
through the course of the song), and duets formed by females joining in with
complex phrases. Mann et al. classified both leucopogon and thoracicus as Style 9, and I would strongly agree with that
classification. It is also interesting
that both leucopogon and thoracicus show some ecological
similarities to griseus, in that all
three species tend to forage mostly in low or mid-level vine tangles. Molecular analysis may ultimately show that griseus isn’t particularly close to any
of the other South American “Thryothorus”,
but I do find it interesting that ecologically and vocally, it is closer to thoracicus & leucopogon, which
molecular data would place in Cantorchilus.
“When all is said and done, I have to
vote YES on recognizing Cantorchilus
along with Pheugopedius and Thryophilus, with the recognition that
we may end up moving some of the constituent species between genera as more
molecular and vocal data become available (especially as regards griseus).”
Comments from Cadena:
“409-411.
YES to all three proposals. I think we made a mistake when considering proposal
219. Back then, the discussion led to rejection of the proposal because
committee members disagreed on whether one should recognize one or three genera
for the South American "Thryothorus".
This implied that we have maintained a genus that has been convincingly shown
not to be monophyletic (owing to the position of ludovicianus).
“When
discussing proposal 219, I tried to emphasize that the three genera proposed by
Mann et al., and now endorsed by Donegan and Barker in these proposals, are
strongly supported clades. This suggests that one cannot go wrong by giving
names to these groups. In contrast, support for the monophyly of a clade formed
by all three proposed genera to the exclusion of other wren genera is
nonexistent. Therefore, I contend that recognizing a broad Pheugopedius is untenable with the evidence at hand. I would also
like to emphasize that even if the three clades turn out to form a monophyletic
group, a classification that ranks each of them at the genus level would be
stable (i.e. monophyly of the group would not require making any further
changes). In contrast, if we adopt a broad Pheugopedius
as suggested by committee members before, and this genus turns out not to be
monophyletic (note that the molecular data point in this direction), we will
need to revisit the classification of the group once again, leading to
instability.
“Finally,
it is nice to see that there is some evidence from vocalizations, etc. that
aids with the diagnoses of the new proposed genera. However, I think that this
should not be a requirement to justify changes in situations like this: solid
phylogenetic analyses tell us that our current taxonomy is inconsistent with
evolutionary history, so we need to change it. It would be nice if all named
groups had phenotypic diagnoses, but I believe that it is more important that
they represent monophyletic groups. As Kevin said, we might end up needing to
move a few species between genera as data accrue in the future, but this is fine.
Let's act based on what we already know.”
Comments from
Jaramillo: “
“Proposal 408. Change linear order of current Thryothorus wrens
(II)
Yes – this appears to better fit molecular, as well
as behavioural data sets.
Proposal 409. Recognize genus Pheugopedius (II)
Yes – this one is clear based on the data.
Proposal 410. Recognize genus Thryophilus (II)
Yes – Again, this seems to be the right
course of action based on molecular data, morphology and voice.
Proposal 411. Recognize genus Cantorchilus (II)
Yes – This was the most problematic one
for me. What I found most helpful were the comments by Gary Stiles and then
Kevin Zimmer in the argument for whether we should recognize a genus based
entirely on molecular characters. The clarification came in how the vocal data
is viewed, and I think that Kevin really hit the nail on the head. The vocal
data is there and they duet and vocalize in a different way, consistently
within each genus, but it is very difficult to describe properly. It is
described awkwardly but in print in the Mann paper, so on the whole I am
cautious but believe that a yes vote is the way to go on this. Voice is
incredibly important in wrens, and in a rather complex manner.”