I discovered a taxon change the other day that seemed ill-advised and thought I’d make a discussion in the forum for people to throw their two cents in for best practices given the situation. I intend for this to be a general discussion about the topic since it keeps coming up, but I don’t want to “call out” anyone involved. Obviously the particulars relevant to the change influence how it should be carried out, and these shouldn’t be discussed on the forum, but I wanted to give an overview and see what other curators thought. Please don’t seek out the change/curator.
The gist of the situation is that a global, multi-continental “Species 1” is thought for a long time to probably be many species, but solid evidence actually confirming this presumption is lacking. A paper comes out in current year showing with genetic evidence that “Species 1” is actually a dozen species in Europe, but the paper only covers Europe, and a mystery still hangs over everywhere else. The paper correlates morphology to phylogeny, so the species are identifiable, and its a good paper, but it only covers Europe. This is significant because less than half of the observations of “Species 1” are actually in Europe.
As I see it, there are three viable options:
Don’t commit the split. Wait for descriptions of species outside of Europe or a complete treatment of the group. European observations don’t have access to their split taxa for a while, even if they want them.
Commit a restricted version of the split. Bump “Species 1” up to complex only in Europe by fabricating atlases that cover all of Europe for the new segregate taxa and delete the atlases after the split. Moves “Species 1” to sensu stricto in Europe, where the segregate taxa are known and available, but leaves the sensu lato everywhere else intact until further study can occur.
Commit an unrestricted version of the split. Bump every observation everywhere up to complex, regardless of whether an appropriate name exists to replace it. More than half of the observations of “Species 1” sensu lato now cannot be assigned to a species, as no verification exists that any of the new taxa, which were found strictly through genetic work (but have been correlated to diagnosable morphological traits), exist outside of Europe.
A curator on iNat wants to draft and commit a split intended to bump all observations of “Species 1” up to a newly created complex that covers all of the new species, effectively killing the concept of “Species 1” on iNat everywhere outside of Europe, which now have no species to split into (Option 3). My recommendation was to postpone the split for a while for discussion, as the paper literally just came out, then follow Option 2, but the other curator insists that Option 3 is the only possible option given that “Species 1” is now confirmed to be a number of cryptic species.
Just from this description I see multiple reasons to think that this paper probably isn’t relevant to iNat curation and shouldn’t be acted on.
For many taxa there are secondary taxonomic authorities which have priority over primary literature precisely because of dilemmas like this.
If the paper only demonstrates evidence for multiple species and doesn’t actually state that they’re multiple species, then it would seem that the authors aren’t confident enough to formally revise their status. This is often the case with phylogenetic papers where they will propose revisions to a clade but not actually formally revise it. This suggests that they think more work should be done before modifying the formal taxonomy. Species should only be added when they’ve been formally described, or if this is a case where they were described in the past, and then synonymized, then it should be clear in the paper that they’re being formally reinstated.
In cases where there are conflicting versions of a clade in literature, I think ones with a globally consistent understanding should be prioritized even if they’re older. This is another reason why secondary authorities are helpful.
No secondary taxonomic authority for this group. There’s been discussions about acquiring one, but they haven’t borne fruit yet. Thanks for bringing this up.
The paper states they’re multiple species, provides formal names/descriptions, and provides a morphology-based key (informed by the results of their phylogenetic analyses) for the new species. The curator I mention earlier has already added them all to iNat.
I wouldn’t say there are conflicting versions of a clade per se, more that the European members of the sensu lato are now the most well-studied members of the group. The opposing view is one big species, which is known to be impossible because there is a species that breaks the monophyly of “Species 1” that is extremely distinct, trivially easy to identify, enigmatic, of conservation concern, etc., etc.
Ah, considering all of that then it makes sense to me to add the new species. Observations outside of Europe will have to be tried in the key and identified to the next rank up (complex/subgenus/genus) if there’s anything ambiguous (I think if the split is organized properly on iNat this should happen automatically?).
Unfortunately it’s not known if there are species outside of Europe that key incorrectly in the new paper’s key, as nothing outside of Europe was studied in the paper. I’m certainly curious how my local things key (and will be keying them), but I wouldn’t assign them based on the key with any confidence. The phylum is infamous for species that look identical between continents.
but, based on your characterization of the situation, I would lean toward Option 2 (split in Europe, and leave IDs at Species 1 sensu lato elsewhere). If there’s a good reason to expect the species outside Europe to receive a similar treatment relatively soon (1 year?), then I might follow Option 1 and hold out for a more complete taxonomic understanding. If this were for a plant species, I might use the excuse of waiting for POWO! But I don’t think iNat can defer accepting a sound taxon split indefinitely.
As to Option 3, it makes no sense to reassign IDs for non-European observations if the authors of the recently published paper did not examine specimens outside Europe. I guess from one perspective, moving all new observations to “Species 1 complex” is equivalent to stating that we define “Species 1” narrowly in Europe and more broadly outside. But I have several practical concerns with bumping up non-European observations to complex:
A large number of observations will likely go from RG to Needs ID and be removed from GBIF.
Many users will be unsatisfied with a complex-level ID and try to apply the ID Species 1 outside Europe, leading to messy data and unproductive exchanges with identifiers.
CV will try to suggest species-level IDs for non-European observations and will have to be frequently corrected. (Against that, in the Option 2 scenario, I think you’re going to have problems with CV also.)
It sounds to me like this is, at most, a non-vascular plant;
In that case, POWO will not cover the situation. Of course, if this instead were a vascular plant, there’s also no particular reason POWO would adopt a sensible global solution, and might attempt to incorporate the revision without regard to its consequences for the rest of the world. Kew has so few people working the desk (so to speak) and they intend it clearly enough to be a tertiary source that POWO decisions are not really considered treatments in the manner of flora.
As for the case in this thread, obviously Option 3 is the most conservative and probably the least likely to be “wrong” in the sense of being likely to contravene subsequent evidence. However, I would go with Option 1, unless there is some absolutely pressing need (conservation, gathering data for an imminent followup publication) to see the distributions of the new segregates that have yet to be discussed in any responding source. Option 2 is, I’ll warrant, the most likely to happen in a place like iNaturalist, but leads to a (deliberately) inconsistent and incoherent usage of the “Species 1” name, even if it’s “practical”.
I do not see the issue with “a large number of observations likely go[ing] from RG to Needs ID and be[ing] removed from GBIF” if the consensus solution points to their current determination probably being inconsistent with the evidence. (If the split is performed a certain way, though not necessarily the way being advocated, the complex can also be the one in the Computer Vision suggestions, which means that users accepting CV suggestions will not be using the species name.) In addition, I can’t picture “users [being] unsatisfied with a complex-level ID and try[ing] to apply the ID Species 1” being such a widespread problem that it would account for a plurality of new identifications.
insists that Option 3 is the only possible option given that “Species 1” is now confirmed to be a number of cryptic species.
But is it *not* confirmed to be a number of cryptic species outside Europe. Outside of Europe we wait for a paper that actually says these are separate species (and I don’t mean a paper that says “should probably be split in the future”, but an actual paper doing the split outside Europe)\
*When I say 2 is the only option, I am assuming that one of the following is true (using letters instead of numbers to avoid confusion with OP’s scenarios)
A: The findings of this paper are accepted by iNat’s taxonomic authorities.
B: The findings of this paper are not accepted by iNat’s taxonomic authorities, but there is a consensus to deviate and treat this paper as authoritative.
C: There is no authority for the taxon in question, and consensus is to treat this paper as authoritative.
If none of the above are true, we are talking about a primary source that there is no active consensus to treat as authoritative, then I see no reason why any taxon change is being considered regardless of the details of the change, iNat just does not use primary sources outside of specific circumstances.
I think the most important thing to do here is to wait for a bit and see what the general reaction to the paper is among other specialists. Sometimes a big split happens and everyone is in agreement and happy that it finally got done, and sometimes when a big split happens everyone thinks it’s rubbish and will be reverted. There’s no point in doing a bunch of taxon swaps to then find that you have to revert them all when someone publishes a rebuttal. As said above a few times, we should wait until there is consensus that this is authoritative.
In terms of which scenario to follow, I think it’s simply down to what is most useful. If this were a group I was studying, I would tend to vote for option 3. It’s been shown that the previous concept of Species 1 was actually multiple species, and now there is a much more restricted concept of Species 1 in use. The individuals outside Europe have been identified on iNat based on the old concept of Species 1, and if we are then changing that, then these records need to be re-evaluated based on our new concept. This is especially true if there is a new Complex taxon for the old species complex, because it’s essentially a one-to-one relationship between them, rather than having any records get bumped up to an even higher taxonomic level and lost. It’s also been a good incentive to get some work done on these groups, so that species names can actually be assigned to everything.
At the end of the day though, all that is irrelevant if it’s not useful. I think that if we are recognising these new species as valid, then option 1 is not particularly useful (unless, as others have said, there is reason to believe that outside Europe will be treated soon). But it’s often a tossup which of options 2 or 3 are more useful. Is it useful to treat the restricted taxon in Europe the same as the broader taxon outside of Europe? Or conversely, is it useful to bump everything up to complex level? I think it is more confusing to have two different concepts for the same taxon, but it will depend on the specific taxon, the number of observations, the geographic range, and whatever else may be relevant.
Are yout alking abpout Aneura pinguis? (if so, the articles are focused on europe, but not exclusive to it, so at eleast for some species, ther is already evidence the occur elsewhere)
If yes, I made a flag and suggested we keep this as is, for the exact resons you listed, but somebody ignored (or missed it) and commited the change anyway.
IMO, the best case solution is that we make it a complex Aneura pinguis anyway (so option 3), as it does reflect reality and our current knowledge - even for experts, it is impossible to really assign it to species, even in europe not all of the taxa were described.
The split will also be likely accepted, this has been in the works for 3 decades and there is a broad agreement on it already, and there are at least 3 groups in Europe alone working on it, and the recent publications will likey spur a bit of activity elsewhere too.
I would say a version of option 2 is best. I would atlas all species, and the sensu lato should be atlased as sensu lato (including sensu stricto). I don’t know why you would need to delete atlased after the fact though.
This is actually fairly common in understudied groups like invertebrates (and I’d guess plants), where work on diverse groups is done in chunks. We’ll be waiting til the cows come home for a comprehensive revision to be done in every group, so we’re almost always working with a partial or outdated revision. There’s always a reason to wait, which is sort of a reason not to wait.
As others have said, though, this is clearly dependent on the situation and what IDers and observers of this group prefer.
Yes, I was also confused about the idea of deleting the atlases after the split. It would seem that we would generally want to maintain the same distribution guidance for the post-split taxa, at least until there’s new research clarifying the non-European species and their distributions.
Yes, but I made it fairly explicit that I wanted to discuss this general recurring situation and not the specifics involved in the change for Aneura pinguis. I saw the flag but didn’t comment because the paper literally just came out and I wanted to understand their results better. The split was pushed through almost immediately but it failed because it was drafted wrong, so we’re still in the same place as before the split.
Making a complex is not the same as option 3 (the complex exists in option 2 as well). I don’t doubt that the split will be accepted and that the European taxa exist elsewhere, this discussion is more of a question of whether the split is currentlyfollowable at this time. We can accept both that the results of the paper are true and also that immediately following the paper could bring unfortunate and unfollowable results (such as with Conocephalum, for which we are approaching 20 years with no good name for the species in North America so everyone in North America is still using conicum or salebrosum anyway). iNat isn’t well-equipped to handle taxa that don’t have names/descriptions; the recent developments in fungi would be a step towards that, but such a robust system does not exist for plants.
When I think about option 3, what comes to mind are other splits informed by phylogeny that lack the necessary combinations to actually have anything split out yet. For a hypothetical, if a paper split the genus Polygala but only made necessary combinations available for European taxa, I would not expect sources out of Europe to immediately stop using the incorrect binomials until new combinations are provided for non-European taxa. In that same vein, I would not expect anyone outside of Europe to start using the new European Aneura until it is known what restricted species concepts occur regionally.
In my example the atlas is just a temporary means to split out Europe and so the range would not be accurate. I know my vague example leaves a lot to the imagination, but my intention was to describe a situation where the ranges of these new segregates are not fully understood so an atlas of all of Europe and just Europe would be almost certainly inaccurate.
Well, unfortunately, I think to resolve such matters sensibly, there can be no general rule over all taxa, and it is kind of imoprtant to know what we are dealing with here.
It is, by making Species 1 (wither Aneura, Conocephalum or anything else in a similar position) a species complex, knowing the species list in it would not be exhaustive. With the current knowledge, we anyway know that what we call Species1 in America is probably not really species1, since studies form other continents have already shown it is a very speciose species complex, or there is already evidence that it does not align with the type of species 1.
It is also the only “correct*” solution for many underreseached taxa. This is also a matter of reality, some taxa are much harder to ID to species than others and I think there should be more messaging to iNat users that complex is quite often the most you can get from a picture.
*correct in the sense it is following evidence and current scientific knowledge. It might not be correct in the sense of making things easier/clearer in the field or for observers, and I appreciate there are some tradeoffs between the two
I am a retired Medic and have some doubts about the current trend in splitting although I apply the example of H. sapiens that may be unfair. Re DNA sequencing, many Human polymorphisms eg. ABO blood groups have large and varied DNA coding sequences and yet we are all H. sapiens. I could be wrong but wonder if some of the sequences used to split non human species are shorter sequences and not relevant to reproduction? Also splitting on minutiae of phenotypic detail would infer that H. sapiens is multiple split species;
For a long time, phylogenetics used single genes (often several at a time), and a good amount of research went into deciding which were diagnostic. This is still common, especially for understudied and diverse groups, but new genomic tools are also available now to look at millions of sites across the genome rather than just a few hundred bases of individual genes. It’s still important to select diagnostic sites, but this is generally a much more robust way of constructing phylogenies.
I’d favor option one. Doing a half-way change is wasted effort. I’ve also heard negative feedback from other iNat users about things getting changed “all the time”. Where new species are concerned, people got away with unbelievable things in the past, for example one based on a single preserved adult male. The author was right after all but I’d think at a minimum an author should have found a breeding population, good life photos of both sexes as adults, and some coverage of juvenile stages. Unfortunately its not easy to fight the rush to publish.
Option 3 is arguably more phylogenetically correct, but I lean towards Option 2 as a little more pragmatic for us. (The Conocephalum situation came to mind.) I agree that people outside the monographed range are likely to continue using the old, broader circumscription of the species without qualification. I understand not wanting to use a paraphyletic concept of a species, but being strict about this also implies that we should be changing rank to species complex for species that have been shown to be multiple entities but which no one has yet attempted to divide.