Use of the Term "Species Complex" in iNaturalist Taxonomy

[paul_dennehy]

@matthewvosper

The key thing is not to invent groups with no literature basis just because someone thinks they are confusing.

I suppose this is my main concern about the way the term is being applied. The Curator Guide lists as two of the requirements for a taxon:

• Species complex is monophyletic (i.e. sibling groups of species)
• Complex is recognized in the literature

And neither of these is true of some of the examples listed. Again, I’m all for the inclusion of a group like “Datana drexelii/major”, from a pragmatic point of view (I’ve ID’d 649 of the 787 observations at that “complex” level personally). But it’s not a monophyletic group, and it’s not called a “Complex” in any literature. They’re just the Willow/Alder Flycatcher of the Notodontid moth world, and seeing their larvae is the equivalent of hearing the Empid calling.

@keirmorse

The point of names is to put things into groups so that you can refer to that group. If you can reduce a group from 20 relatively distinct species to 3 relatively cryptic species, that is a very useful tool.

Agreed, completely. Whether there’s any consistency in how those names are “ranked” is my main concern.

There’s also inconsistency in when one of these “complex” levels gets added. Essentially, “how hard to identify is hard enough?” For the Desmia funeralis complex, you just need to flip the moth over and look at its underside. For the Feltia subgothica complex, you just need the antennae to be in-focus and at decently high resolution. Amphipoea americana and interoceanica require genitalia dissection, but there’s no “complex” for them, so they have to go back to a genus containing 15 species (or just be randomly ID’d as one species or the other, which is what usually happens). I’d love to suggest adding a complex level to put them in, but are they even sister species? And is there a literature basis for the complex? According to a quick look at BOLD, interoceanica is sister to senilis, a distinctive southwestern species, that pair is sister to a Palearctic species group of three species, and americana is on an outgroup branch from there. But americana/interoceanica is the pair that presents ID difficulty in photos, so… a complex?

@keirmorse

Maybe a better name would be a species morphocomplex

Yes! I think the more common term for what we’re calling “Complexes” here is “morphospecies”. At bioblitzes, this is always how we divide the little brown moths that require dissection when we’re trying to get a day-of-the-event quick count of species. The term explicitly includes “morpho-” to emphasize that this is a pragmatically-applied non-taxonomic grouping based entirely on appearance. I like this idea.

@sedgequeen

Maybe we can set up a Bombus vosnesenskii clade. (Species almost identical to B. vos. aren’t its sister taxa.)

@taylorse

In birding this pair is routinely reported as “Traill’s Flycatcher”, a combination which has its own banding code.

@jurga_li

even though the complex does not neesarily include taxonomically closely related species

These are pretty much exact parallels for what some of the iNat insect “complexes” are- not sister taxa, just groups that present an ID challenge. From a pragmatic standpoint, it’s great to have a name to put on them besides genus. From a taxonomic standpoint, we’re sometimes inventing polyphyletic groups to organize our data. I’m not inherently opposed to doing that, but I think a term like “complex” at least implies monophyly, no matter how broadly you interpret it. “Morphocomplex” or “Morphospecies” seems like a better descriptor of what we mean.

I don’t want the takeaway of this thread to be “eliminate all the complexes that aren’t monophyletic”, just “we have quite a few polyphyletic/paraphyletic complexes on here so maybe we shouldn’t be calling them complexes anymore”.

[sedgequeen]

Actually, the two Willow and Alder Flycatchers are sister taxa. But I see your point.

One the other hand, you write, “For the Desmia funeralis complex, you just need to flip the moth over and look at its underside” but for the photos of living moths that get posted on iNaturalist, that’s not practical. The taxa are cryptic in the iNaturalist context.

I really think complexes or morphocomplexes (or whatever we call them) help a lot here. The complex names would help researchers find the taxa they want and give observers satisfaction that could help keep us observing.

[upupa-epops]

I see no reason not to have a Traill’s Flycatcher complex on iNat, except that bird taxonomy is more strictly and formally curated than insects or plants. The usefulness of that category here would be pretty obvious, and they have been referred to as a “species complex” occasionally. There is an open flag for adding it.

There used to be a taxon for Rufous/Allen’s Hummingbird but it was removed in 2017(?), before the species complex rank was introduced in 2019. I was just looking into whether we could argue for re-introducing that complex and found this rule in a flag discussion which I’d never heard of before, from @loarie:

There are 17,969 current IDs sitting at Selasphorus
compared to 21,604+18,035 = 39,639 current IDs on S. rufus and S. sasin which is < 1 to 2 (even if we assume that all those ~18k IDs on Selasphorus would be on a complex sitting above S. rufus and S. sasin were it available)

We tend to only make complexes when a strong majority of IDs (e.g. > 2 to 1) are sitting a coarse node which isn’t the case here. Lets check back if that becomes the case

Exceptions are cases where there’s an indeterminate number of species descending from a genus so complexes can be the natural stopping place. Only 9 well defined Selasphorus species so also not the case here.

Also adding nodes gets much more disruptive when there are lots of IDs (there are ~60k below Selasphorus) so there’s a large cost to adding them.even if it did meet the criteria I mentioned.

As I said this is the first time I’ve heard of this, which is a bit frustrating since the rest of us have been discussing for years about the merits of various complexes and now it turns out there’s another rule behind the scenes. And limitations like this based on site resources are also frustrating since a complex like this one would still be really helpful for literally thousands of observations.

[Averixus]

@paul_dennehy

I suppose this is my main concern about the way the term is being applied. The Curator Guide lists as two of the requirements for a taxon:

• Species complex is monophyletic (i.e. sibling groups of species)
• Complex is recognized in the literature

That’s exactly it. The problem is really that there’s a disparity between the purported rules (the curator guide is very clear and explicit that complexes can only be added if they are monophyletic and recognised in literature), and what people actually do in practise (many iNat complexes don’t meet those criteria).

This means that either:

1. A huge number of curators are constantly breaking the rules. All their changes should be rolled back, all those complexes should be removed, and all those curators should be suspended or otherwise penalised?! or:
2. The rules are written wrong.

It would be a lot easier in the short term to change the rules than to undo all of the clearly rule-breaking complexes that already exist. The trouble is that changing the rules would mean admitting that taxa on iNat don’t actually always have to be monophyletic or recognised in literature. And doing that opens the floodgates for all the taxonomic debate that iNat specifically tries to avoid by having strict rules and following established authorities in the first place.

So instead we’re in this strange position where the powers that be are pretending that the rules are correct and still apply, but “everyone” knows that they really don’t. Which is a pretty unpleasant and uncomfortable environment for newcomers, because it means there’s no way to learn what the actual rules are other than seeing which ones other people break without consequence.

Even adding one or more specific levels which were explicit exceptions to the usual taxnomy rules (e.g. the morpho- taxa some people have suggested) would still mean that definitions at those levels would be a free-for-all. How do we decide whether species are similar enough to be grouped? At the moment we decide by, any curator willing to ignore the rules goes along with the most convincing argument and makes them a complex. Which doesn’t seem like a great system, but I can’t think of a great alternative either. Plus it would probably(?) just be impossible to do in practise because iNat’s system is built on every taxon having one parent. What if there’s a “morphocomplex” containing species from two or more genera or subgenera, how would that be added into the taxonomy?

[matthewvosper]

@Averixus

What if there’s a “morphocomplex” containing species from two or more genera or subgenera, how would that be added into the taxonomy?

I have always understood this to be the point of the ‘monophyletic’ rule. i.e. all members need to have the same parent. You can’t create a complex in genus A and then create B’us species’us beneath it, when that truly belongs in genus B.

I agree that that is not strictly what is said, but otherwise we are limiting the use of complex to groups where an actual phylogenetic study has been performed - which rules out the vast majority of all things! I don’t think that’s the intention.

I understand the rules to mean ‘only include species that share the same taxonomic parent’, ‘only add things that have some support in literature’ (which need not be a formal taxonomic act), ‘don’t create something that doesn’t significantly refine the parent’, ‘name it after a single species, and if the literature does not define one species for the name, use the oldest name’.

Generally I’m in favour of the ‘add species group as a taxon’ proposal:

@zygy

Please vote for https://forum.inaturalist.org/t/create-a-new-taxon-rank-for-inaturalist-the-species-group/14543 to fix this!

…as I make clear on that thread.

I do find the current situation tolerable however.

[charlie]

@Averixus

The trouble is that changing the rules would mean admitting that taxa on iNat don’t actually always have to be monophyletic or recognised in literature. And doing that opens the floodgates for all the taxonomic debate that iNat specifically tries to avoid by having strict rules and following established authorities in the first place.

Well i can tell you the debate is prolific already. It’s not always clear cut whether something is monophylletic. I think it makes sense to have a different name and designation for known monophylletic groups, but it completely breaks my brain that someone would want to use that as a criteria for group validity given this is a citizen/community science site. The site simply will not work if you insist on only monophylletic groupigns and still insist on the level of taxonomic change and splitting that occurs here. I won’t get into that further because most people already know how i feel about this… but these polyphylletic groupings are a way to create the dynamic taxonomy some curators really want while also allowing the site to be usable by others, so i don’t see any downside other than semantics.

@Averixus

So instead we’re in this strange position where the powers that be are pretending that the rules are correct and still apply, but “everyone” knows that they really don’t. Which is a pretty unpleasant and uncomfortable environment for newcomers, because it means there’s no way to learn what the actual rules are other than seeing which ones other people break without consequence.

Yes agreed. I’m autistic and in my own autistic profile i can’t even really function in an environment that works that way. I see others breaking rules on the site in a bunch of ways, and not getting asked to stop. Then i break a rule on what i see is a similar magnitude and suddenly i’m some big villain and have tons of flags on my content and am getting scolded. Meanwhile others engage in what i see as awful behavior towards me and when i try to flag it, i get told off but when i try to argue back i myself get flagged. I totally agree i should hold myself to a higher standard than people whos behavior i think is wrong, but this goes beyond that, significantly beyond. This issue of inconsistent rule enforcement is why i stopped being a curator. I can’t function in that environment. And while autistic people are all different there are a whole lot of autistic people on iNat and i know others have this issue as well.

@Averixus

At the moment we decide by, any curator willing to ignore the rules goes along with the most convincing argument and makes them a complex. Which doesn’t seem like a great system, but I can’t think of a great alternative either.

Curators also do this to split or lump species, change names, etc. There have been multiple cases where these changes don’t meet the criteria. The same people are up in arms about ‘polyphylly’ who changed dramatic swaths of iNat and made a bunch of data unusable beased on unpublished research.

To be clear I don’t think the people who run iNat are malicious or intend to create a system like this. I think they are overwhelmed by the vast size of the site relative to staff, a very flawed moderation system (again, the taxonomic curators are also all moderators), and a lot of external factors that pressure them to bend rules this way or that. I don’t know the full answer either, but i do myself feel the site is on a very problematic trajectory right now. I know i’ve been arguing here too much and am tryign to disengage again, but i do appreciate that others on the forum recognize these issues too.

[sandboa]

I know the non-sister species group causes some structural complexities for iNaturalist, but it is a useful tool for encouraging people to make the best identification they can.

If there was a way to remove any taxonomic node implications to a “species group” while retaining those for a “species complex” it seems it would serve all the multiple interests/goals of iNaturalist.

I don’t think creating a new term is the answer. We want non-scientists to be able to use this platform seamlessly. They won’t really care about the taxonomic implications of species complex vs species group. They just want to be able to make the most precise identification they can.

And maybe on those species group/complex pages, a curator or volunteer could write a succinct explanation as to why they are lumped together.

When two species can be separated geographically with certainty, there should be no grouping whatsoever.

[sandboa]

I would like to see us avoid “species level” labeling of complexes though. As an example, if I want to find all the records for Hyla chrysoscelis and its sister Hyla versicolor, I have to search under Hyla versicolor. If I type in Hyla chrysoscelis into the seach box, I don’t get the “complex” identifications. Maybe if we could name it Hyla chrysoscelis/versicolor instead?

[matthewvosper]

That gets pretty unwieldy when there’s more than two species in the complex. I think that’s why the naming rule is as it is.

[john_hepburn]

I’m glad my confusion is not just the result of ignorance, and grateful for the explanation.