• Welcome to BirdForum, the internet's largest birding community with thousands of members from all over the world. The forums are dedicated to wild birds, birding, binoculars and equipment and all that goes with it.

    Please register for an account to take part in the discussions in the forum, post your pictures in the gallery and more.
ZEISS DTI thermal imaging cameras. For more discoveries at night, and during the day.

Species recognition between geographically isolated populations (1 Viewer)

This may be correct, but if you read the header "Song Experiments Reveal 21 Possible New Tropical Bird Species" you may think otherwise :eek!: (see here)...

A statement like "Song experiments reveal SACC taxonomy to be seriously outdated" would have been equally correct...(it is in fact almost hilarous that exactly JV Remsen is cited in this note to support the findings...sorry, I couldn't resist |:$| )

The other thing I found quite odd from a SACC bias perspective was to look at the list of examples of good species listed in the paper and to see how they performed under the playback tests. I've grouped these into splits proposed by "SACC members, advisors and their mates" (SACC), splits proposed in my papers with coauthors from back when I was with ProAves ("ProAves") and others.

Dull-mantled Antbird–Magdalena Antbird Myrmeciza laemosticta–M. palliata Suboscine 0.71 (14) SACC
Zeledon’s Antbird–Blue-lored Antbird Hafferia zeledoni–H. immaculata Suboscine 0.91 (11) Proaves
Streak-headed Antbird–East Andean
Antbird Drymophila striaticeps–D. caudata Suboscine 0.3 (10) SACC
Blackish Antbird–Riparian Antbird Cercamacroides nigrescens–C. fuscicauda Suboscine 0.3 (10) SACC
Long-tailed Tapaculo–Rufous-vented
Tapaculo Scytalopus micropterus–S. femoralis Suboscine 0.2 (10) SACC
Paltry Tyrannulet (Central America)–Paltry
Tyrannulet (Venezuelan Andes) Zimmerius vilissimus–Z. improbus Suboscine 0.73 (15) ProAves
Northern Schiffornis–Russet-winged
Schiffornis Schiffornis veraepacis–S. stenorhyncha Suboscine 1 (5) ProAves
Costa Rican Warbler–Three-striped Warbler Basileuterus melanotis–B. tristriatus Oscine 0.56 (9) ProAves
Blue-black Grosbeak (west of the Andes)–
Blue-black Grosbeak (east of the Andes) Cyanocompsa cyanoides–C. rothschildii Oscine 0.71 (14) SACC

On their scale used here, 0 is bad, 1 is good and c.0.6 is a surrogate for species rank. ProAves-related splits scored an average of 0.75 (range: 0.56-1). Of those, 2 were accepted by SACC, one was rejected and one has not been considered yet, one of those which did pass was subject to discussions and adverse comments. SACC-related splits scored an average of 0.44 (0.2-0.91) and despite some of them being very marginal were generally subject to a sycophantic chorus of opprobrium and unanimous positive votes (subject to one comment and vote on Cyanocompsa) . What are not included here are scores for the countless splits proposed by other authors which were rejected by SACC. It would be interesting to see how those measure up. The odd pattern of citation in this paper, which does not include a single one of the ProAves-related papers relating to the splits above, is also noteworthy especially given the prominence given to SACC-authored papers.

[Also, this statement next to the table is an incorrect summary: "We include in this list 2 taxon pairs (Zimmerius and Cyanocompsa) for which proposals to define populations as distinct biological species are currently under consideration by the South American Classification Committee of the American Ornithological Society." As above, the Zimmerius split was rejected and is up for reconsideration. The Schiffornis proposal passed only on a second time of asking. And the Basileuterus proposal has been affirmatively considered by NACC but is yet to be considered by SACC.]

I've been banging on for a long time on Birdforum about the corruption in a system whereby "mates' papers" on dubious or borderline situations are lauded and accepted whilst those dealing with more clear-cut cases but published by persons outside of a narrow clique are inventively trashed for elusive reasons (which surely can't include much of a regard for rationality in taxonomy). Reading this paper, at least in the above respect, is something which caused me a sense of validation!

Saludos, Thomas

PS I should say, odd patterns of citation aside, this is one of the best and most useful and interesting papers Auk has published in years, so well done to the authors.
 
Last edited:
Had several, e.g. Firecrest coming to Goldcrest recording.

Playback is generally very dependent from the time of breeding season, territory density and 'personality' of individual males.

I am especially missing the effect of distance per se. Do birds react much worse to a song of 'different' allopatric species from 400 km away than 'same' species from continous population from 400 km away in another direction?

When I was doing research on Cerulean Warblers, one oddity was that on several occasions we would get responses to playback from Louisiana Waterthrushes, which by all accounts are very different both taxonomically and ecologically. Strange that the waterthrushes would find this worth the energy, while the other local warblers - Kentucky, Hooded, Ovenbirds, and especially the similar (vocally, ecologically, and taxonomically) Northern Parula had no responses to my knowledge.

I don't know if this is addressed in the paper at hand, but I wonder if playback experiments are more effective for relatively sedentary tropical species as opposed to those that experience seasonality. Here in eastern North America, it seems that there are times of year when playbacks will cause vigorous attention and others when the birds couldn't seem to care less. As others have mentioned, there are more factors to playback response than just taxonomy.

That said, from what little time I've been able to spend with this article I agree that it is a good contribution to a topic that needs more attention.
 
An important factor is what you are trying to show. If you are trying to show to show that two populations are different species, then it is the non-response to the songs of the other population that is important. Individuals that respond, due to conspecifity, similar songs, interspecific competition, etc., are not really that important.

Andy
 
Re: individual differences to playback.

Response to playback is very variable. As anybody can check trying to repeatedly playback some common bird in the area it is legal and not otherwise disturbed, for example Chaffinch. The response depends from weather, day, male etc. Two very common reasons of non-response are: that the male is in a distant part of its territory and does not hear the playback or has no time to come. Or the male responds, but by coming quietly and listening from the cover, so is overlooked by the human observer.

I also witnessed several times that completely unrelated and differently singing bird species are interested in playback. Especially if they were themselves singing intensively before. It is my feeling that male songbirds often learn songbirds of other species sharing their territory, and new sound in the neighborhood attracts curiosity.

Re: Studies in playback as evidence of species difference face two other big challenges:

First, they are normally not blind. The observer knows which sound it plays and this can affect his behavior and evaluation of response. But basically it is possible, if difficult, to blind such studies. E.g. by setting the speaker to randomly play one sound or another, and automatically taping the sounds. These would be evaluated by a second observer in the lab, unaware to which sound came the response.

Second, defending territory from other males and attracting a female are two different functions. Playback measures the first, while lack of hybridization depends from the second. It is well known for many bird songs that different characteristics of the song are responsible for one and another.
 
Proposal (754) to SACC

Benjamin G. Freeman and Graham A. Montgomery (2017) Using song playback experiments to measure species recognition between geographically isolated populations: A comparison with acoustic trait analyses. The Auk: October 2017, Vol. 134, No. 4, pp. 857-870.

Proposal (754) to SACC

Elevate 13 taxa to species rank based on playback experiments

A. Elevate Pseudocolaptes johnsoni to species rank
B. Elevate Automolus virgatus to species rank
C. Elevate Grallaria andicola to species rank
D. Elevate Scytalopus intermedius to species rank
E. Elevate Ochthoeca thoracica to species rank
F. Elevate Myadestes venezuelensis to species rank
G. Elevate Pheugopedius schulenbergi to species rank
H. Elevate Amazonian populations of Tunchiornis ochraceiceps to species rank
I. Elevate South American populations of Basileuterus culicivorus to species rank
J. Elevate Myiothlypis chlorophrys to species rank
K. Elevate Myiothlypis striaticeps to species rank
L. Elevate Atlapetes tricolor crassus to species rank
M. Elevate Amazonian populations of Arremon aurantiirostris to species rank
 
Hello all --

Was just pointed to this thread. Lots of interesting comments. I'm happy to address any questions or concerns.

A couple thoughts:

- All taxon pairs are allopatric (geographically isolated). Playback experiments simulate secondary contact - do the birds recognize each others' song? In some cases we conducted experiments in allopatry but the two taxa meet somewhere. In these cases, further work is needed at that place where they meet.

- We mention all this in the article, but there is some confusion here. The paper deals with less than half of the taxon pairs we have studied with playback experiments. The main project is to study how song evolves in geographic isolation in Neotropical birds; we have good data (good = lots of territories tested) for ~ 140 taxon pairs. The paper provides matched data for acoustic trait divergence & behavioral response to playback for 72 taxon pair comparisons. Quantifying acoustic trait divergence is very time consuming, thus the smaller sample. Of the 21 taxon pairs we highlight for species limits only ~ 8 come from the sample of 72.

- One key finding is that sometimes populations happily respond to song from an isolated population that is (to my ears) pretty different. And sometimes they consistently ignore song from an isolated population that is (to my ears) really similar. Of course song recordings are terrifically important. But we show that some birds care when differences are statistically minor and others do not. Remember that the job here is a tough one -- are two geographically isolated populations sufficiently different that they would not interbreed? So really we only care about differences that are relevant to reproductive isolation, and we argue that responses to playback are pretty good data in that regard.

- Another key thing is that we did not conduct these experiments to "prove" that a population deserves species status. We were instead interested in evolutionary patterns across many comparisons. It is noteworthy that we almost always had some responses even when songs are wildly different - responses could have been incidental, curious etc, but they are still a bird approaching the speaker.

- Someone mentioned that some birds like Henicorhina will respond to anything. Actually our previous paper showed that Henicorhina leucosticta in Costa Rica ignore song from NW Ecuador populations! And we had plenty of other wrens that ignored allopatric song. So playback experiments are even useful for wrens.

- Regarding the idea that someone could set up an online database where people could enter their own playback data. Yes yes yes! That would be amazing. I can imagine many folks could do an experiment here or there, and this data could be really useful! Please write to me or continue this thread if that is of interest to anyone.
 
With SACC voting starting to come in on this proposal, the authors may consider that they sold themselves short (and sold previous authors short) by not consulting and citing other literature supporting some of these proposals.

Most notably this one on Atlapetes crassus:
https://www.researchgate.net/public...tion_in_the_Andes_The_Atlapetes_brush-finches

And, for example, this short note on how Woodhaunters (then Hyloctistes, now Automolus) vocalizations differ, citing various other authors who split them already:
http://www.proaves.org/wp-content/uploads/2012/11/Con_Col_17_1-14_Actualizacion_Listado.pdf

Moreover, IOC and other authorities split a number of these "species", but that was not mentioned in the proposal. The proposal and related paper somewhat give the impression of original findings in instances when this is dubiously the case. Moreover, by omitting to reinforce their position on these splits with reference to other consistent works, their proposals on several of these taxa come across as weak, for the reasons discussed by Gary Stiles in the SACC proposal online...

My impression is that many of these splits are good ones, but prospects of progress seem a bit bleak, in part due to the above issues.
 
Last edited:
Hello Thomas --

Thanks for your message.

As you say, I did not submit comprehensive and detailed proposals to SACC. If that is cause for many of the proposals to be rejected, so be it. At least Graham and I have put out worthwhile data into the public sphere. I do expect that all of these proposals will eventually, one way or another, be adopted. Reasonable people can disagree with whether the playback evidence that I have presented is sufficient to split taxa at the moment, at least for some cases (other cases I think are beyond obvious, but of course I would say that). I do think it is worth pointing out that:

1) Because all taxa considered in my proposals are allopatric, genetic distances really don't mean much with respect to reproductive isolation. We expect populations that live in different places and do not interbreed to differ genetically. But what we want to know is whether allopatric populations are reproductively isolated (= different species) or not (= same species). The degree of genetic differentiation in allopatry doesn't help us make this assessment, unless it is absolutely huge (which would suggest that hybrid offspring would be likely to have defects).

2) So far all comparisons that we tested reciprocally (how does pop A respond to song from pop B? and also: how does pop B respond to song from pop A?) show symmetric responses. E.g. if pop A ignored song from pop B, pop B also ignored song from pop A, and vice versa. Thus I am very comfortable making recommendations based on comparisons that were only tested one direction.

3) We used many recordings in our playback experiments. Thus our results can be generalized to tell us how birds from one population respond to song from another population (if we had used only one recording we would be technically limited in our inference to how birds from one population respond to one particular song from another population).

4) Our metric of "discrimination" is rather severe -- we only call it discrimination when a territorial bird completely ignores song from a related population. Not that it shows a weak response, or a slow response, but completely ignores. I think this is the form of discrimination that is most relevant to trying to estimate premating reproductive isolation. (And note too that the territorial bird in question always responded strongly to local song).

5) I think playback data are a big step up from acoustic analysis of songs. Again, we expect populations that live in different places to sing a bit differently from one another. If these differences are huge and completely obvious, then it is pretty good inference to think that these two populations would ignore each other if they came into contact. But if the differences are more minor, it is really important to assess biological significance (behavioral response) rather than statistical significance. Acoustic analyses are extremely helpful and playback data are not always feasible to collect. But I do think playback data presented alone are more meaningful than acoustic data presented alone, and acoustic data presented alone are typically convincing enough for SACC to make decisions.

6) Last, I disagree with the viewpoint that "these species complexes are complex, let's wait until a thorough analysis deals with everything al at once." A "whole-complex" perspective is of course necessary to make taxonomic decisions in some cases, but this criticism is not fatal when we are only considering allopatric taxa. For example the chat-tyrants. Perhaps Venezuelan populations (which differ in plumage) ought to be considered distinct from other populations N of the Maranon. But whether that is true or not has no bearing on whether populations immediately N and S of the Maranon Gap are reproductively isolated from one another.

Whatever SACC decides, I think it is worthwhile to re-emphasize (again and again) that what we are trying to do is figure out which populations are reproductively isolated from one another and which are not. For allopatric taxa, this is, in the strict sense, impossible. But for taxa in which we think song plays an important role in mate choice, playback experiments are likely the best available method for helping us assess premating reproductive isolation.

Best,

Ben
 
Last edited:
Interesting that in the OP, the species will be recognised on the basis of song, under the BSC. Surprising to me, does this mean that they're moving away from the emphasis on DNA?


A
 
At least for the AOU/AOS Committees, results for DNA analyses on species level questions are often ambiguous,especially when trying to show that allopatric populations are reproductively isolated. They can be very useful in sympatry to show lack of gene flow (e.g, two populations are reproductively isolated and thus two species). They are also useful when they show a lack of a assortative mating, thus indicating extensive gene flow and the lack of reproductive isolation (the populations should be lumped as one species). There will never be an absolute metric for defining species (at least for those following the Biological Species Concept) using DNA.

Andy
 
At least for the AOU/AOS Committees, results for DNA analyses on species level questions are often ambiguous,especially when trying to show that allopatric populations are reproductively isolated. They can be very useful in sympatry to show lack of gene flow (e.g, two populations are reproductively isolated and thus two species). They are also useful when they show a lack of a assortative mating, thus indicating extensive gene flow and the lack of reproductive isolation (the populations should be lumped as one species). There will never be an absolute metric for defining species (at least for those following the Biological Species Concept) using DNA.

Andy


I thought that the BSC had a benchmark differential for speciation, something like 2.7%?


A
 
I thought that the BSC had a benchmark differential for speciation, something like 2.7%?


A

Nope. Reproductive isolation could arise by theoretically a single base pair (e.g., changing the color of the throat from red to violet). Lack of assortative mating could occur between populations much greater than 5% different. For example, Common Ravens in California vs the rest of North America show no phenotypic/behavioral/vocal differences but are quite different in DNA (I think around 5%). The populations were isolated for a very long time, but never accrued differences that would lead to reproductive isolation.

Andy

Andy
 
Nope. Reproductive isolation could arise by theoretically a single base pair (e.g., changing the color of the throat from red to violet). Lack of assortative mating could occur between populations much greater than 5% different. For example, Common Ravens in California vs the rest of North America show no phenotypic/behavioral/vocal differences but are quite different in DNA (I think around 5%). The populations were isolated for a very long time, but never accrued differences that would lead to reproductive isolation.

Andy

Andy

I know even less now than I did before!



A
 
Andy Kratter said it perfectly --

Genetic data is of utmost importance when trying to determine whether species that are sympatric are reproductively isolated. Want to know how many species of Scytalopus there are in an Andean cloud forest? Then go out, sample thoroughly, sequence and see how many genetic groups are present. If they all interbreed (no reproductive isolation), then there will just be one genetic group. If two groups are reproductively isolated from one another and do not interbeed, then there will be two genetic groups.

But genetics is not especially useful when we are looking at allopatric populations. In the Andes, it is very common for related but allopatric populations to differ dramatically in mitochondrial DNA from one another (e.g. >5% sequence divergence, which is ~ 2.5 million years of evolving in isolation). It is *so* tempting to want to assign species limits based on sequence divergence. But what we really want to know is whether two groups are reproductively isolated or not. And we know from the ravens that Andy mentioned and other examples (e.g. Xanthomixis in Madagascar - (open access) http://onlinelibrary.wiley.com/doi/10.1002/ece3.1639/full) that lineages can live in allopatry for millions of years but not evolve reproductive isolation (we know this because we infer that when they came back together they freely interbred).

Defining species status of allopatric populations is, has always been, and will always be a difficult task. But I think it is important to assemble data to try to define species limits as best as possible -- my personal bugaboo right now is that species are more narrowly defined in the temperate zone than the tropics, and this taxonomic bias may skew the results of comparative analyses that assume that taxonomy is applied consistently across Earth.

Best,

Ben
 
Last edited:
Andy Kratter said it perfectly --

Genetic data is of utmost importance when trying to determine whether species that are sympatric are reproductively isolated. Want to know how many species of Scytalopus there are in an Andean cloud forest? Then go out, sample thoroughly, sequence and see how many genetic groups are present. If they all interbreed (no reproductive isolation), then there will just be one genetic group. If two groups are reproductively isolated from one another and do not interbeed, then there will be two genetic groups.

But genetics is not especially useful when we are looking at allopatric populations. In the Andes, it is very common for related but allopatric populations to differ dramatically in mitochondrial DNA from one another (e.g. >5% sequence divergence, which is ~ 2.5 million years of evolving in isolation). It is *so* tempting to want to assign species limits based on sequence divergence. But what we really want to know is whether two groups are reproductively isolated or not. And we know from the ravens that Andy mentioned and other examples (e.g. Xanthomixis in Madagascar - (open access) http://onlinelibrary.wiley.com/doi/10.1002/ece3.1639/full) that lineages can live in allopatry for millions of years but not evolve reproductive isolation (we know this because we infer that when they came back together they freely interbred).

Defining species status of allopatric populations is, has always been, and will always be a difficult task. But I think it is important to do our best in all cases -- my personal bugaboo right now is that species are more narrowly defined in the temperate zone than the tropics, and this taxonomic bias may skew the results of comparative analyses that assume that taxonomy is applied consistently across Earth.

Best,

Ben

Thanks for this guys,
I can't pretend to be a scientist, I just always assumed that if two species had DNA that was different by a minimum of whatever, then under the BSC, they would be different.

Obviously I know even less than I thought I did and was surprised that the BSC would split a species based on vocalisations alone, that seems, to my very limited understanding, to be more like a PSC thing to do.

As Ben says though, uniformity is important and even with my scant knowledge, I can see that it's probably not being achieved.


A
 
From bitter experience, Ben, I'm sorry to suggest that AOU/SACC has little to no interest in promoting rational taxonomies based on the latest research. Their main aim seems to be to promote "conservatism" which is a by-word for promoting irrationality over anything but detailed studies of a particular species group in the right kind of journals by the right kinds of authors using museum specimens, DNA and voice. That your study is more informative and based on more data or more relevant data than the efforts of past taxonomists is something I have tried before to little avail with SACC.

I would agree with all your proposed splits. It's a new approach using field data and provides a more rational and consistent basis for species limits for all groups treated compared to present treatments based on the hunches of 1910-1960s museum-based taxonomists. Many of them are a statement of the blindingly obvious. For those in "complexes", to me it is pretty obvious where other named subspecies fall. But SACC are not much interested in rationality in taxonomy. They may be interested in doing something with groups like Atlapetes given the other paper on this situation. They might even be interested in addressing cases where they are calling the earth flat in the face of all evidence and all other taxonomic authorities, like Automolus. I think you might struggle with the others, but surely there are other situations here where there is some "broadly consistent" molecular data or vocal study or field guide split or IOC treatment to point to, in terms of showing this study to be consistent with others and increasing the perceived cogency of your suggestions?
 
Obviously, the argument that these vocal analyses were not peer-reviewed can still be given but that's something else...

I might point out that, beyond not being peer-reviewed, the vast majority of these reports themselves don't cite previous publications. Furthermore, most do not provide the actual analyses or data other than a representative sonogram and the author's conclusion. Thus, as "gray literature," one mustn't be too upset if they are not cited by others.
 
Thanks Peter and Dan for your messages --

Peter - Thanks for the links to the HBW vocal analyses you compiled. They are interesting to read. They do fall short of what would be necessary to convince a skeptic. I agree that ideally, playback experiments (to examine behavioral response) would be paired with detailed acoustic analyses (to examine acoustic geographic variation in song within and between populations).

Dan -- you are an active tour leader so are in a good position for me to ask the following question: Is it feasible to conduct occasional playback experiments while guiding tours? It's easy for me to get excited and imagine that it would indeed be feasible -- that you are often already using playback to lure in an interesting bird for the group to admire, and you are probably already telling your clients/friends about geographic variation in song that might be pertinent to species limits, perhaps even briefly playing allopatric song so everyone can hear how different songs are. But it may well be that my imagination is too rosy as I sit here at my desk -- in the field with a big group things may be way too busy and you have your hands full.

The broader idea is that I think it would be great if guides, ornithologists and birders could team up on playback experiments (each person doing a couple as they are able to) to compile data relevant to assessing species limits. Not sure if this would work in practice, but I do think it is a good idea. Playback experiments are easier to conduct then ever before (wireless speakers, lots of good quality recordings publicly available). One of the main goals of publishing this paper was simply to motivate others to try doing playback experiments themselves.

Best,

Ben
 
Thanks Peter and Dan for your messages --

Peter - Thanks for the links to the HBW vocal analyses you compiled. They are interesting to read. They do fall short of what would be necessary to convince a skeptic. I agree that ideally, playback experiments (to examine behavioral response) would be paired with detailed acoustic analyses (to examine acoustic geographic variation in song within and between populations).

Dan -- you are an active tour leader so are in a good position for me to ask the following question: Is it feasible to conduct occasional playback experiments while guiding tours? It's easy for me to get excited and imagine that it would indeed be feasible -- that you are often already using playback to lure in an interesting bird for the group to admire, and you are probably already telling your clients/friends about geographic variation in song that might be pertinent to species limits, perhaps even briefly playing allopatric song so everyone can hear how different songs are. But it may well be that my imagination is too rosy as I sit here at my desk -- in the field with a big group things may be way too busy and you have your hands full.

The broader idea is that I think it would be great if guides, ornithologists and birders could team up on playback experiments (each person doing a couple as they are able to) to compile data relevant to assessing species limits. Not sure if this would work in practice, but I do think it is a good idea. Playback experiments are easier to conduct then ever before (wireless speakers, lots of good quality recordings publicly available). One of the main goals of publishing this paper was simply to motivate others to try doing playback experiments themselves.

What an excellent idea! A way for we birders to make up for some of the environmental harm caused by our frivolous hobby.
 
Last edited:
Warning! This thread is more than 6 years ago old.
It's likely that no further discussion is required, in which case we recommend starting a new thread. If however you feel your response is required you can still do so.

Users who are viewing this thread

Back
Top