What is a species? (1 Viewer)

[Jon.Bryant]

Yeah, WGAC are continuing to make progress, they just don't necessarily share that progress on their webpage. Just

I am not trying to criticise the WGAC. My understanding was however that everything should have been reviewed by now. More importantly each agreed review can apparently be returned (once) to the WGAC for reassessment. Finally BirdLife have not committed to adopt all WGAC findings - hopefully they will.

This suggests to me that these authorities have strong opinions and that agreement is not so easily reached. This seems slightly contrary to the fact that nobody on this tread has yet been able to define a species.

If there was not a strong opinion and each party was willing to compromise, a simple majority view, would suffice and the WGAC taxonomy could be produced without the large amount of current work (which is all voluntary).

 

[Jon.Bryant]

Yeah, looks like it's possible to speak BSC with a PSC accent. It's grey zones all the way down.

Not sure. An article by BirdLife in the BTO magazine suggested that the PSC system may result in 16k species. If BirdLife are BSC with a PSC accent, it is a very slight one. Their overall total is not much different to IOC - just quite a few pluses and minuses. I think it is just A different implementation under BSC (and the BTO article stated Birdlife use BSC!).

 

[Jon.Bryant]

There has been, as already stated, many definitions made.

Perhaps I am being an idiot, but I think we have plenty of principles but no standards (definitions). As an example I am an Engineer - principles may state what loads should be considered in a design (say wind load), but I would expect a standard to explain how this load should be derived for any particular location. If the principle states how a species shall be assessed, then we seem to be missing any guidance on assessment criteria - how larger difference in DNA, how different are vocalisations etc. Without a definition it boils down to opinion - rather like suggested before, the system is then only a judgement (but not necessarily a fact).

 

[Welsh Peregrine]

The Tobias criteria are (in your terminology) standards, but even they have difficulties in application; a plumage difference may score 1 or 2 points, but AFAIK there is no rigid definition of where to draw the boundary between 1 and 2. Furthermore, who is to say that our perception of visual differences equates to that of the birds concerned with greater UV sensitivity?

 

[Farnboro John]

I am not trying to criticise the WGAC. My understanding was however that everything should have been reviewed by now. More importantly each agreed review can apparently be returned (once) to the WGAC for reassessment. Finally BirdLife have not committed to adopt all WGAC findings - hopefully they will.

This suggests to me that these authorities have strong opinions and that agreement is not so easily reached. This seems slightly contrary to the fact that nobody on this tread has yet been able to define a species.

If there was not a strong opinion and each party was willing to compromise, a simple majority view, would suffice and the WGAC taxonomy could be produced without the large amount of current work (which is all voluntary).

Critical thinking is an important part of refining both the approach and the individual cases and is considerably aided by having people with different opinions on board. A neutral process agreed beforehand by all involved parties for resolving differences of opinion is ultimately required though.

John

 

[jurek]

You are intentionally implying that taxonomic researchers are intentionally lying so they can get some lower impact papers out, the same papers that major universities won't give a crap about as far as tenure goes nor which will help there grant writing. You are effectively accusing the folks whose papers often drive this forum of fraud.

And taxonomic inflation certainly has been a term that has been argued, but I would say by a minority of individuals whose arguments usually come down to conservatism for conservatism sake, the idea itself.

As I said before: since there is no one agreed definition of species, and all definitions come without detailed criteria (except the most permissive, Tobias), the whole debate is mostly built on sand.

As I said before: there is little outright lying. Rather more subtle: changing between definitions of species towards ones which allow more species. Selectively adopting papers and evidence pro-splitting and ignoring pro-lumping (how old is the first paper which showed that Common and Arctic Redpoll are genetically same?). Mode of career which eliminates conservative scientists and promotes ones producing catchy papers (without scientists themselves changing opinions).

 

[Mysticete]

I am not trying to criticise the WGAC. My understanding was however that everything should have been reviewed by now. More importantly each agreed review can apparently be returned (once) to the WGAC for reassessment. Finally BirdLife have not committed to adopt all WGAC findings - hopefully they will.

This suggests to me that these authorities have strong opinions and that agreement is not so easily reached. This seems slightly contrary to the fact that nobody on this tread has yet been able to define a species.

If there was not a strong opinion and each party was willing to compromise, a simple majority view, would suffice and the WGAC taxonomy could be produced without the large amount of current work (which is all voluntary).

I have no idea where you got the idea that WGAC should be finished by now; There certainly was nothing from WGAC suggesting this would take a year or less.

This isn't simply a matter of someone creating a big list of discrepancies and the group asking whether something should be recognized or not. My understanding is that they have a proposal system, that requires each taxonomic discrepancy to be explained with arguments made for and against a split. these proposals are then shared with voting members of the committee and also sent to other relevant taxonomic authorities such as the NACC and SACC to get there opinions. Not to mention that something like this also has to be done for higher level taxonomic discrepancies, such as genus. And all of this has to be balanced with all of the work committee members have OUTSIDE the WGAC, since the folks in WGAC almost certainly are not being paid for their work and have day jobs.

It's not like a member is just emailing out "Tropical Pewee...one or two species?" and calling it a day. They have to summarize the evidence for one or two species, and then weigh in. And the fact there is debate over how many species something is means these decisions are not clear-cut, and might require deliberation.

 

[Mysticete]

Perhaps I am being an idiot, but I think we have plenty of principles but no standards (definitions). As an example I am an Engineer - principles may state what loads should be considered in a design (say wind load), but I would expect a standard to explain how this load should be derived for any particular location. If the principle states how a species shall be assessed, then we seem to be missing any guidance on assessment criteria - how larger difference in DNA, how different are vocalisations etc. Without a definition it boils down to opinion - rather like suggested before, the system is then only a judgement (but not necessarily a fact).

Any sort of standard/threshold is going to be very difficult to apply across all bird groups.

As an example, say you decide to use a % genetic divergence, and define the threshold as a specific number. The problem there is that number will be effected by generation time: faster breeding species will accrue more genetic divergence in a shorter time than slower and more long-lived species. It also will vary based on the gene dataset used, and is also dependent upon whether there is even data out there to calculate this. Finally, it assumes that the threshold value will represent the same point in all species within the speciation process. We know however that there are birds with fairly low genetic divergences that behave as good species.

 

[Andy Adcock]

Any sort of standard/threshold is going to be very difficult to apply across all bird groups.

As an example, say you decide to use a % genetic divergence, and define the threshold as a specific number. The problem there is that number will be effected by generation time: faster breeding species will accrue more genetic divergence in a shorter time than slower and more long-lived species. It also will vary based on the gene dataset used, and is also dependent upon whether there is even data out there to calculate this. Finally, it assumes that the threshold value will represent the same point in all species within the speciation process. We know however that there are birds with fairly low genetic divergences that behave as good species.

I'm no scientist but, I thought BSC had been using c2% as the baseline?

 

[Mysticete]

I'm no scientist but, I thought BSC had been using c2% as the baseline?

2% is sometimes used as a rough approximation, although a more accurate way to look at thing is to compare the % divergence between populations with other divergences in that same family, to determine if the %divergence is greater than one would expect for that group.

With the huge caveat of course that I don't work in bird systematics, and I am not an expert on this!

 

[njlarsen]

2% is sometimes used as a rough approximation, although a more accurate way to look at thing is to compare the % divergence between populations with other divergences in that same family, to determine if the %divergence is greater than one would expect for that group.

With the huge caveat of course that I don't work in bird systematics, and I am not an expert on this!

I am not working in this field either. The understanding I have of the 2% is that it was used as a rule of thumb for difference level in one particular sector of the mitochondrial genome. There has been a development so that these days, a result only including the mitochondrial genome will not be seen as sufficient evidence to result in making a change. Genetics from the nuclear genome, vocalizations or other data are needed on top of the mitochondrial genome data.
Niels

 

[Jon.Bryant]

I have no idea where you got the idea that WGAC should be finished by now

I recall that this was from an email communication with a participant in the process.

If you look at their website the WGAC started formal voting in Feb 2021. By June 2023 they had completed 165 families (although some of these were however easy ones with no differences of opinion between participants). The assessment process had increased in pace to around 33 issues per month. By the end of July they were hoping to have completed 180 families (including two very specious ones), meaning they would have dealt with 55% of taxa.

From the website data I don’t think that a 12 month period for completion would be that daft. It is a little odd however, that the did updates on progress in January and June 2021 and not a peep since.

 

[Jon.Bryant]

Any sort of standard/threshold is going to be very difficult to apply across all bird groups.

Yes probably. I think we have had the debate before on the fact that Large White Headed Gulls apparently show relatively little genetic divergence. If this is the case then the taxonomist would appear to be working to some underlying guidance - lump on weak genetic divergence in some families but not on others, which are presumably assessed on a different basis.

 

[Paul Clapham]

Interesting is that you find merely species in the paleontological account but rarely subspecies. So a bird bone from the Cretaceous has only a little information whether it is a species or a subspecies.

Like Niels said, it's hard to identify fossils to the subspecies level. However paleontologists are using current bird checklists (IOC, I believe) to assign names to recent fossil specimens. And when splits and lumps occur in the checklist it makes then wonder about whether they need to reconsider specimens which are affected. Or so I'm told.

 

[Mysticete]

Yes probably. I think we have had the debate before on the fact that Large White Headed Gulls apparently show relatively little genetic divergence. If this is the case then the taxonomist would appear to be working to some underlying guidance - lump on weak genetic divergence in some families but not on others, which are presumably assessed on a different basis.

BSC in the matter of gulls has always trumped genetic divergence. It's possible we will see more lumps in the group, just as we recently saw Thayer's and Iceland Gulls lumped.

 

[Mysticete]

I recall that this was from an email communication with a participant in the process.

If you look at their website the WGAC started formal voting in Feb 2021. By June 2023 they had completed 165 families (although some of these were however easy ones with no differences of opinion between participants). The assessment process had increased in pace to around 33 issues per month. By the end of July they were hoping to have completed 180 families (including two very specious ones), meaning they would have deal with 55% of taxa.

From the website data I don’t think that a 12 month period for completion would be that daft. It is a little odd however, that the did updates on progress in January and June 2021 and not a peep since.

They very irregularly post updates, I don't think anyone should read anything more into that. We know that the process is ongoing from comments made by IOC and references in the SACC checklist proposals and it seems like all of the world bird checklists are still onboard. That they are moving slower shouldn't be a surprise, since they are also trying to get AOS on board with there changes, so probably waiting for feedback for some issues from those individuals.

It's also very possible that as things have gone back to normal from COVID times, the folks who vote on these things simply may have less time. I would imagine a lot of the proposals for the early rounds of voting were created when the world was still shut down or partially shut down, and the committee members simply had more available time to work on these things.

 

[Myornis]

Perhaps I am being an idiot, but I think we have plenty of principles but no standards (definitions). As an example I am an Engineer - principles may state what loads should be considered in a design (say wind load), but I would expect a standard to explain how this load should be derived for any particular location. If the principle states how a species shall be assessed, then we seem to be missing any guidance on assessment criteria - how larger difference in DNA, how different are vocalisations etc. Without a definition it boils down to opinion - rather like suggested before, the system is then only a judgement (but not necessarily a fact).

Hello Jon. Please see this attempt at an objective, stats-based system for species.

Donegan TM 2018. What is a species? A new universal method to measuredifferentiation and assess the taxonomic rank ofallopatric populations, using continuous variables. Zookeys. PDF

Note that for sympatrics (species occurring in the breeding season in the same space and time) we really need to study whether they interbreed to produce fertile offspring in the field or lab; I don't think stats will help much. The difficult cases tend to be allopatrics (differentiated populations that do not come into contact) and also populations which hybridise "a bit", "a lot" etc.

For those populations which hybridise, some kind of assessment of "a bit" versus "a lot" needs to be made and whether the hybrids that exist are fertile or not - which is fraught with controversy in some edge cases like Carrion/Hooded Crows and Green-winged/Eurasian Teals. I am afraid you will have to deal with opinion-based subjectivity there for the time being.

For allopatrics, it is possible to come up with objective measures by comparing differences between allopatrics versus those between sympatrics. The differences between for example woodpeckers or toucans (which have lots of plumage variation within species) and tapaculos (virtually no plumage variation between many species) shows that a single measure across all birds will not work. Instead, you have to look at each group (family for sure, possibly each genus) separately to determine a yardstick for species rank. Helbig et al 2002 first came up with a practical set of guidelines for birds based on this concept. The Tobias model is a simplistic and easy to understand way of putting some numbers into such a system, but is fraught with basic errors of statistics in the way it was devised and still relies upon considerable subjectivity (as discussed in the above paper, and a recent series of papers in 2021 in "the journal formerly known as Auk" (Ornithology). Their system is also based upon all birds of all families having a similar differentiation for species, so they oversplit toucans and overlump "little brown jobs" like Basileuterus warblers in South America. The good thing about it is that the HBW series of splits and lumps based on the system had the same authors doing the assessments, and so the subjectivity is controlled and, in my view and as published on separately, I think most (but not all) of the splits are "good". The above paper of mine aims to deal with all those problems in a single mathematical model (whose output varies by bird family).

Molecular data helps inform taxonomy but is no use as a benchmark - some species (e.g. Wedge-billed Woodcreeper) have up to 10% intraspecific mtDNA variation in the same region as other species (African flowerpiercers?) show 0% difference between good, sympatric species. So the only benchmarks must be made on real-world data like voice, measurements and plumage; and since you can't really "measure" plumage easily to do statistical analysis, really an objective system can only be done with biometrics and vocal characters, then also looking at plumages and molecular as a supporting crutch.

Since you are an engineer you may understand maths well and might find the above paper useful.

TD

 

[Jon.Bryant]

Note that for sympatrics (species occurring in the breeding season in the same space and time) we really need to study whether they interbreed to produce fertile offspring in the field or lab; I don't think stats will help much. The difficult cases tend to be allopatrics (differentiated populations that do not come into contact) and also populations which hybridise "a bit", "a lot" etc.

Thank you very much for the paper. Much appreciated. I will try and read and digest. I was actually coming to the conclusion that any model would need to be family or genus specific, as you suggest.

The issue of hybridization is quite interesting. Reading in Dutch Birding (not on taxonomy but in the monthly update section), up to 9% of Pallid Swifts in colonies in France have Common Swift DNA, and there is evidence of ongoing hybridization (presumably undetected previously). On the other hand the Common Swift race pekinensis has now been suggested to inhabit different areas and habitat in the wintering grounds and not mix with the nominate race. The subspecies is therefore isolated on breeding and wintering grounds and morphologically different - I have just come back from Beijing where birds are present around the airport, and to me they don't sound the same either - but perhaps that was just the acoustics.

I always assumed that the original concept of hybridization in sympatric species came from our own experimentation with animals - the case that mules are infertile - or is this urban legend. I have read that this definition creates problems with many ducks, but also with Collared & Pied Flycatcher and Blyth's Reed & Marsh Warbler where hybrids have apparently bred.

I am guessing here, but I presume that we cannot determine whether species will have fertile young apart from through experimentation or 'mechanics' - i.e. I have read that falconers experimenting with birds have crossed Merlin and Peregrine, but only with male Merlins, as the eggs for a hybrid are two large for a female Merlin to carry and lay - again not sure if this is urban myth, but pretty sick if it is true. Through mechanics it is pretty obvious that some birds cannot hybridize successfully, but I think theses pairings are obviously different species in any case. I presume that we do not understand genetic information well enough to determine if a hybrid will be infertile in principle, through a mixing of genes (could we predict the gene mix for a start? I think that hybrid ducks quite often have types, so hybridization does not seem to produce infinite variation).

Assessing the possibility to hybridize and produce hybrid young, would be interesting for certain species which presumably have evolved from a common species and then become isolated and fragmented, perhaps through geological processes. As you say many Tapaculos very similar, and hybridisation cannot be assessed due to isolation - vocals here would seem to be the prime differentiator.

As a aide issue, it is interesting that I would expect certain species not to hybridize in the first place as the species are vocally distinct and therefore attraction should not occur in the first place. This does not seem to necessarily be true, with some interesting apparent hybrids recently found in Britain (Redstart x Winchat & assumed Whinchat x Stegner's Stonechat) - are these explained by forced opportunistic couplings with a properly paired female? In any case, a lack of vocal attraction, would seem to make assessment as to whether birds will hybridize and whether young will be fertile even harder to assess - just because species don't generally 'get it together in nature', may not mean they are not closely related or that they cannot produce fertile offspring.

 

[Jon.Bryant]

They very irregularly post updates, I don't think anyone should read anything more into that.

As a bit of fun, I looked at my mapping of the three global taxaonomies, and using IOC as the base, calculated possible not a species - one other party disagrees, possibly an additional species - an additional species recognised by another party, probably not a species - the other two parties disagree, and probably an additional species - an additional species recognised by both the other parties.

Using this we get 10,650 definite species (agreed by all three parties), 11,041 probably species (including those recognised by two parties) and 11,550 possible species (including those recognised by one party only) . The possible and probably species total 900, so still a fair chunk or work to complete if the WGAC assess all these - even at 33 per month this is 27 months! although as we have previously discussed some decisions will take time to trickle through to each global taxonomy, and may have already been agreed by the WGAC. But, if things have slowed down post Covid, then perhaps I shouldn't hold my breath!

I am not sure how to read my unscientific calculation - 900/11,550 = 7.8%. Does this 'grey area' seem large or small? Does it imply the same principles are being applied relatively consistently by all parties or not? Is my question on how species are defined important, of do the stats indicate that the 'system' seems to be working?

A hiccup with the WGAC process will be that in correspondence with BirdLife, I was informed that the participants are not putting a moratorium on new decisions. This means that quite a lot of the updates in each taxonomy are 'work as normal' assessments based on new evidence, rather than acts of unification. Each new decision is potentially a new area of disagreement to be resolved later. I thought it would have been wiser to have parked future decisions, until at least all the parties had agreed that they would be making the same change in future. Otherwise, it may be extremely difficult to finalize the project and sync all lists. The optimum time to try and do this would seem to be to target completion in early July, the end of the longest interval between the global taxonomy updates (IOC end of Jan and July, Clements Aug and BirdLife Dec).

I will not work through the update notices, but await the formal issue of IOC 13.2. It will however be interesting to see how IOC 13.2, Clements 23 and BirdLife 8 change the above calcs. How about a convergence clock with 1,000 differences representing one hour to midnight? (we would then currently be at 11:06 - 54 minutes to midnight!).

 

[Andy Adcock]

As a bit of fun, I looked at my mapping of the three global taxaonomies, and using IOC as the base, calculated possible not a species - one other party disagrees, possibly an additional species - an additional species recognised by another party, probably not a species - the other two parties disagree, and probably an additional species - an additional species recognised by both the other parties.

Using this we get 10,650 definite species (agreed by all three parties), 11,041 probably species (including those recognised by two parties) and 11,550 possible species (including those recognised by one party only) . The possible and probably species total 900, so still a fair chunk or work to complete if the WGAC assess all these - even at 33 per month this is 27 months! although as we have previously discussed some decisions will take time to trickle through to each global taxonomy, and may have already been agreed by the WGAC. But, if things have slowed down post Covid, then perhaps I shouldn't hold my breath!

I am not sure how to read my unscientific calculation - 900/11,550 = 7.8%. Does this 'grey area' seem large or small? Does it imply the same principles are being applied relatively consistently by all parties or not? Is my question on how species are defined important, of do the stats indicate that the 'system' seems to be working?

A hiccup with the WGAC process will be that in correspondence with BirdLife, I was informed that the participants are not putting a moratorium on new decisions. This means that quite a lot of the updates in each taxonomy are 'work as normal' assessments based on new evidence, rather than acts of unification. Each new decision is potentially a new area of disagreement to be resolved later. I thought it would have been wiser to have parked future decisions, until at least all the parties had agreed that they would be making the same change in future. Otherwise, it may be extremely difficult to finalize the project and sync all lists. The optimum time to try and do this would seem to be to target completion in early July, the end of the longest interval between the global taxonomy updates (IOC end of Jan and July, Clements Aug and BirdLife Dec).

I will not work through the update notices, but await the formal issue of IOC 13.2. It will however be interesting to see how IOC 13.2, Clements 23 and BirdLife 8 change the above calcs. How about a convergence clock with 1,000 differences representing one hour to midnight? (we would then currently be at 11:06 - 54 minutes to midnight!).

How on earth did you crunch those stats so quickly!!!!