Introducing Taxon Frameworks

We're in the process of rolling out new functionality called Taxon Frameworks. Please read about them here. But in a nutshell, taxon frameworks allow us to track and communicate what we mean by a particular branch of the tree of life and thus what we're all agreeing to reference and curate towards. Taxon Frameworks can be sourced to external secondary authorities like the Reptile Database and deviations from the authorities can be explicitly mapped. We've also folded in the existing 'complete taxa' and 'taxon curators' functionality into taxon frameworks.

Some background: the core activity on iNaturalist is observing and identifying observations by hanging them on the Tree of Life. The tree of life on iNaturalist is a single global taxonomy that we all share. A big thank you to the few dedicated iNaturalist Curators who are responsible for maintaining this taxonomy.

The policy on iNaturalist has generally been to adhere to secondary taxonomic authorities (like the Reptile Database) for a number of reasons. However, as iNaturalist has grown, its been increasingly difficult to manage information about these secondary taxonomic authorities including what they are, when we defer to them, and when we deviate from them.

Over the last year, I've been experimenting with ways to better manage iNaturalist taxonomy with a few informal taxonomy working groups and some pilot features like 'complete taxa' and 'taxon curators'. These features allow branches of the tree of life to be curated only by a smaller subset of curators assigned to a particular branch and to make it clear that these branches have all taxa added.

While this pilot has been productive, using journal posts, projects, external spreadsheets, and flags hasn't been the best way to maintain, share, and communicate all of this information. Keeping track of 'deviations' - places where we wish to depart from whatever secondary taxonomic authority we're otherwise following - has been particularly difficult with the existing system.

We're still in the process of rolling out the Taxon Framework features including migrating over content managed by the old system and updating documentation. But I wanted to give everyone a heads up about this functionality incase you stumble across it. Also for the moment, only iNat staff can create/edit taxon frameworks and taxon framework relationships while we continue testing things and getting the bugs out.

As we continue experimenting with ways of leverage the Tree of Life on iNaturalist - from making computer vision suggestions and consensus identifications more sophisticated to tracking range extensions and other discoveries - its becoming increasingly important that we have a well formed and robust tree of life to build these systems on. While we recognize that the tree of life will always be constantly changing, we're striving towards systems where it changes in a more structured way. Taxon frameworks are a step in this direction by letting us explicitly lay out what taxonomy we're all agreeing to reference by stitching together secondary sources and the judicious use of deviations.

At the moment, we have about 25 taxonomic frameworks covering much of the tree of life:

When we scale the tree by the number of observations, about 75% of observations are covered by taxon frameworks. The major 'holes' in this coverage are Lepidoptera and several other insect orders (Diptera, Hymenoptera, Coleoptera, and Hemiptera), Fungi, and Bryophytes, and various marine invertebrates such as Malacostraca, Cnidarians, Echinoderms etc.

I believe the World Registry of Marine Species can be used as a basis of frameworks for many of these marine invertebrate groups. For Bryophytes, Goffinet, B., W.R Buck and A.J. Shaw (2008), Morphology and classification of the Bryophyta looks promising.

For Fungi, it would be great if Fungi leaders in the iNaturalist community could come together and provide some guidance. In the past we used Index Fungorum but there were too many instances where the community wanted to deviate from this reference so Fungi are currently a bit of a free for all. If we could use taxon frameworks and deviations to rein Fungus taxonomy in to something that would have broad buy in by the community that would be fantastic.

This leaves Lepidoptera and the other Insect orders. The Orthoptera Species Files and a few other "Species Files" exist for smaller orders but guidance on what to do about the large orders like Lepidoptera would be very helpful. Remember, while ideally taxon frameworks would extend all the way down to species, they can rather only extend down to coarser nodes. For example, even constraining Lepidoptera taxonomy down globally to tribe would be hugely helpful in terms of reining in taxonomic free-for-all.

People often get very passionate about taxonomy - which I guess is good. But as we work towards a consensus taxonomy that we all share, I'll close with a plea for a spirit of compromise and flexibility. The iNat tree of life exists in this weird place between phylogeny (a tree that exactly matches evolutionary history) and taxonomic nomenclature (a set of rules for categorizing things). The taxonomies behind Catalogue of Life, GBIF, the IUCN Redlist, EOL and others occupy similar spaces. This means that while most of the changes come from trying to make the trees reflect evolutionary history, they still have many useful but non-'natural' (non-monophyletic) groups like 'Reptiles'. iNat curators coming from more of the phylogenetic philosophy often want to totally restructure the tree with something that better reflects evolutionary history. While this is a critical component of guiding taxonomic decisions, other factors like stability and simplicity must also be considered too. At its core, the iNaturalist tree is still a Linnaean taxonomy and unnamed ranks such as 'clade' can't be accommodated by the system at the moment. Please try to be open minded and flexible as we compromise towards a consensus taxonomy even if it doesn't result in a perfect phylogeny.

Similarly, iNaturalist isn't built to accommodate all of the rules and metadata that have governed taxonomic nomenclature back to the days of Linnaeus. Some things like quadrinomials are allowed under the rules of nomenclature but can't be accommodated by the iNaturalist system at the moment. Please be similarly flexible regarding some of the strict rules of nomenclature and the omission of certain nomenclatural metadata like types and authors. Once taxon frameworks become editable by iNat curators, we hope they will be places for curators to add some of this metadata if they wish.

Posted on November 21, 2018 12:08 AM by loarie loarie


Excellent -- good news all round. We will need to pull together though, and not be too rigid about our own opinions and preferences.

Posted by susanhewitt over 5 years ago

P.S. When you rein something in, it is spelled rein, because you are pulling back on the reins, like when you are riding a horse. A reign is when a monarch holds sway over a country -- Queen Elizabeth II is the longest reigning Monarch of Britain.

Posted by susanhewitt over 5 years ago

thanks Susan - fixed!

Posted by loarie over 5 years ago

I am no curator, but I appreciate your rules that "force" compromise and lead to consensus. I feel lucky to be witnessing the assembly of an immediate, global, interactive taxonomic system. I appreciate the utility of this system.

Posted by edwardrooks over 5 years ago

What great news! Also, I love the visuals in this post.

Posted by durling over 5 years ago

good news indeed.
Lepidoptera.... ho hum.... where to start. There is no published concencus :(
probably the most reputable resource (and online with no paywall) is van Nieukerken et al., 2011, which has a clear format.

This seminal paper has only recently been updated for all Lepidoptera (Regier et al., 2013 -, and I would suggest iNat follows figures 2 & 3 in this more recent publication as it's foundation for Lepidoptera herein. The key problem (and one I have avoided during my seven years of Lepidoptera curation on iNat) is the use of clades at the ranks of sub and infra order, where there remains a high degree of uncertainty regarding taxonomic positions of more recently evolved superfamilies, never mind the primitive non Ditrysia groups :( [compare figs. 1A and 1C in Regier et al]
Hope this helps !

Thereafter matters get messy. There are quite a few projects online that claim to have global coverage of all Lepidoptera, but not one is truely representative (imho). There are so many people trying to use molecular data now that it's confusing trying to keep up with who is publishing what. There is very little that actually gives an holistic overview. Most research seems to be at the family rank, or lower, or on a geographical basis at the regional or local scale.
For some families there is progress - Noctuoidea, Geometroidea, Pyraloidea (the big 3 groups, accounting for upwards of 60% of all Lepidoptera) are now fairly robust in their systematics and generally accepted globally, but most of the microleps (Pterophoroidea and below in fig 1A) are still poorly understood at most ranks from superfamily down. I don't know how long it will be until there is general concencus about the phylogeny of Lepidoptera - a while yet. Until then we will likely have to muddle through the mire of publications to try and put a coherent framework together for iNat users. The WorMS link is woefully inadequate and outdated for Lepidoptera.

Posted by hkmoths over 5 years ago

@loarie This may be a silly question, but if there are no taxon curators for a specific family, say Acanaloniidae, who handles the flags for curation? The taxon details for the Family Acanaloniidae says "Not covered by any taxon frameworks," so my understanding is that it doesn't have a taxon curator. Earlier this month, I placed a flag for curation for the genus Parathiscia, but it was never resolved. Being that there are no taxon curators for the Family Acanaloniidae, is it just going to stay listed under Existing Flags? There were a few observations incorrectly listed under Parathiscia, but I placed them in the correct taxon. If there are existing flags, would be possible for people to know about them without actually having to go Flag for Curation. With it's current location, Existing Flags are kind of hidden. If someone goes to the page of a specific genus, they may not know that there is an issue with it.

Posted by feistyone over 5 years ago

There are no ordinary Palaearctic species, apparently especially from the territory of Russia (and what about the tropics ?! And not too ancient taxonomic monographs - 10 years). I have a few observations, but there are already a few. Yes, I checked by species names if they appeared to be in continuously changing genus names (good function, but a short selection list is shown, and in vain from other kingdoms). It is necessary every time to call an unknown curator so that he manually adds something that is absent in external databases?

Also Im unhappy about common names of low ranks that obscure true latin names in taxonomy chain string. Need option to swithch off or choose "latin only")) - it can be user-allowed. And ban someone's amateur names rusification! - every case it partially and crooked and outside the box (in entomology, imho).

@hkmoth Molecular data is good if taken from a type specimen. (and just a few genes are the evolutionary tree of these genes, not of the species as a whole. How many genes are used to establish paternity in the USA? -I know only a few works, where about 30 sequences were used to build the macrosystem, whereas usually 1-2-3 - as only for distinguish, and maybe polymorphism.) But it is still unknown who determines how, what then becomes a sequence reference. ...looking at the general level of "non-classical systematics", that also will choose a tree from proposed at the level of any "70 percent"..

Posted by entomokot over 5 years ago

Nice reading the article & kudos for behemoth task you have taken up. Agree fully with your comment on simplicity. Look forward to be an active member of community....

Posted by anil_kumar_verma over 5 years ago

I see that some plant hybrids are sourced to POWO - but POWO does not list any hybrids (e.g. I presume that this is just temporary and such discrepancies will be cleaned up in due course? Or is that the curator's task?

Posted by tonyrebelo over 5 years ago

Where would the discussion about fungal taxonomy take place?

Posted by daviswj over 5 years ago

Very promising and interesting addition. One thing i was unclear on after going through it a couple of times was how exactly this will be used to manage deviations. Maybe the info is not there, or it is and just did not find an unaffiliated brain cell to attach to.

Not the philosophy behind deviations, I get that, but the actual mechanics of managing them. i know you are still porting certain informatin over, but can you comment on that, or show an example if any have been loaded yet ?


Posted by cmcheatle over 5 years ago

I was going to flag @susanhewitt for being too reasonable, but decided not to!

Being reasonable is key to this discussion about taxonomic problems. At it's core, iNaturalist is a platform for people - whatever their level of knowledge - to post photos of wild life. And hope for an identification. Higher level taxonomic concerns (based on genetic analysis & etc) are not a major concern for most users, and without them the site is not going to work. This kind of problem even gets down to the 'lower levels' where most of us live. I work mainly with NA Noctuidae, and am encountering more instances of species level observations that are impossible to identify. Some require genital dissection, or DNA analysis to identify, both of which are beyond the scope of someone looking at a picture. The default is to Genus.
I appreciate the importance of higher level taxonomy, but for most folks it is not an issue. I believe that most users want to take pictures of the life around them, and get an answer to what it is. Researchers who use these data should be the ones to sort out the higher taxonomic identities, as they are the experts.
I guess what I'm trying to say is yes, go ahead with higher taxonomic revisions, but don't lose sight of the main purpose of iNat.


Posted by mamestraconfigurata over 5 years ago

Great, @loarie, this sounds like great progress. Just to clarify: does this mean that we will be able to deviate from databases when their taxonomy does not reflect the consensus in the literature?
Looking forward to seeing this fully implemented.

Posted by markscherz over 5 years ago


Shannon, thank you for the heads-up about the ancient state of the Tropiduchidae and Acanaloniidae on iNaturalist. And Flatidae isn't much better. Unfortunately I have to disagree with @mamestraconfigurata [above] about the importance of higher ranks, especially those between genus and family: they are indeed "a major concern" of ordinary users observing invertebrates in the global south. We desperately need our subfamilies, tribes and subtribes (and sometimes intercalated ranks) and we use them all the time, since a genus-level ID is most often a luxury we cannot attain. I'll write more about this matter if you are unfortunate enough to get more from me to read.
Regarding this subthread in general, I agree that outstanding flags are hidden beyond plain sight and that this may be improved upon. Shannon, I think that your Parathiscia flag has gone unattended simply because it wasn't seen by any curator who felt up to the task, despite these being easy pickings from the FLOW database and the Uni Delaware project. I've previously stumbled upon some flags, which I could resolve, a number of times, some of them being very old at the time. If there is a better way for me [curator] to check on flags of which I am unaware, I shall appreciate advice.
Regarding you specifically, Shannon, I undertake to within days fix up the whole of the Tropiduchidae after Gnezdilov (2013) and Gnezdilov, Bartlett & Bourgoin (2016). There are two reasons for this: (1) The Fulgoroidea are my non-focal interest group of the week, and (2) I need you to do a thorough sweep of the flatid-like hoppers of southern Africa on iNaturalist. [I know that you have begun with that as I am typing here]. You're welcome to @ me with comments, advice and admonitions, and I may just pepper you with questions. A forthcoming scientific paper partly based on iNat observations hangs in the balance.

Thanks, and insert favourite emoji here: ___

Posted by beetledude over 5 years ago

Taxon Frameworks seem incredibly helpful. I thank you @loarie and anyone else who worked to put this together. It provides insight with what those "Complete Taxon" concepts were suppose to achieve.
I might have something helpful for insects. Over the course of several months, I have been looking for potential authorities for our insect orders and have been documenting my findings in this journal post. Due to class obligations, I have not had the time to personally investigate each of these sources yet, and - to be honest - some of them seem better curated and up-to-date than others. However, I think this is a good starting point. Maybe other insect curators could help me review these at some point.

I have a few questions:

At present, you are the sole taxon curator for a number of complete vertebrate taxa. Is this subject to future change? What criteria are you looking for in taxon curators?
I recently committed a taxon change in Cephalopoda that deviates from WoRMS/MolluscaBase. Does this get automatically reflected in the Taxon Framework Relationships or do iNat staff have to manually implement the deviation on the Relationships page themselves? Is it possible for taxon curators to add notes for individual deviations on the Relationships page to indicate that the discrepancy is intentional and not an oversight?
Could Taxonomy Frameworks be used to compare the iNaturalist tree with an authority we are considering switching to? You probably know this, but I was asked by @carrieseltzer to facilitate a comparison between iNat/IUCN to the Mammal Diversity Database (MDD). It might be more helpful to identify differences between iNat to the MDD using the Framework system than the Google spreadsheet we've been using...

Posted by bobby23 over 5 years ago

@feistyone Just a question - are generic level assessments a luxury because the organisms have not been defined well enough, or that there are no resources to identify them (or both)? Is there any way for us (meaning the general users of iNat) to help? With NA nocuidae I have the luxury of a fairly defined taxonomy, although it always seems to be in flux. I don't have much knowledge of the situation in the Global south. I'm not against higher level identification or taxonomy specifically, but I don't want iNat to lose the focus on the general user. If you want to contact me via email rather than cluttering up this thread, please do so. Sometimes we get so isolated we can't see the larger picture.

Posted by mamestraconfigurata over 5 years ago

Just sorted out that my comment above was intended for @beetledude. Sorry

Posted by mamestraconfigurata over 5 years ago

@mamestraconfigurata Oh, good. I was wondering there, but I was gonna try to answer you, lol. :)


Riaan, you're welcome and are correct about Flatidae as well. Yes, I do believe we need the subfamilies, tribes, and subtribes, even the genus levels if a species is able to be applied to that. I'm aware that it's difficult to get species down to genus levels in the area that you're located in.

Oh okay, thanks for explaining about the flag that I put in. I wasn't quite sure how it worked. It would definitely help if there were a section available to curators just for flags that need to be checked.

Oh, that would great! My primary interest is in the Fulgoroidea, especially those outside of the U.S. I've taken a look at the Fulgoroidea in South Africa a number times. I've been trying to find as much information available that I can, but as you're aware it's very limited. Some species look like they could belong to one family, when they actually belong to another. I will definitely do what I can, especially if someone is able to put forth a scientific paper about the species there! I'm very fascinated with many of the species in your area, so that would actually be awesome!

Thanks :)

Posted by feistyone over 5 years ago

In my opinion, most study groups have the more up-to-date databases and knowledge than IPNI, POWO, etc.

Posted by wolfgangb over 5 years ago

Up to date is not necessarily acceptable. I think the advantage of having a single reference far outweighs the disadvantages.

A better rational is to approach these databases and provide the references with the revisions, and get them to update them. Certainly POWO has been excellent, updating changes with days, and providing valuable insights in cases were they were correct in the first place and our "more-up-to-date" databases had missed issues with the revision, or subsequent papers, or previous papers that had addressed the issue that the latest revisions had missed.

Deviations are after all allowed, but should be documented and justified. That is fair!

Posted by tonyrebelo over 5 years ago


It would be very helpful if the existing and old flags can be more easily accessed, not only for a particular taxon but its children.

Currently one can use:
via the taxon page:
via the flags page:
But it would be nice to have this as a clickable link in the "Curation" or "Taxonomy Details" tabs.
But far more useful, would be if this could show all the children's flags as well please. Both current "?new" and resolved flags (sometimes issues have already been resolved).
Especially as "Deviations" might have been made at lower ranks than the one expected.

Posted by tonyrebelo over 5 years ago

This is great, and very promising! I am glad you are moving towards something like that! (and I promise flexibility in understanding as well! :-) )


Posted by amarzee over 5 years ago

@tonyrebelo That is an excellent idea about the flags! You made some very valid points. I think that would actually be the best way to show flags, new and old alike, with all the children flags, including the deviations.

Posted by feistyone over 5 years ago

@bobby23, yes, as far as I understand this new system should be able to replace the spreadsheet system for reconciling differences with iNat and MDD because we'll be able to document deviations. @loarie can hopefully address your other questions. (Just a heads up for those outside the US, it's Thanksgiving here so please forgive staff delays as we celebrate with family and friends).

Posted by carrieseltzer over 5 years ago

Thanks everyone for the feedback.

@feistyone, if a brach isn't covered by a framework that has taxon curators, then any curator can edit it.

@tonyrebelo, correct, what needed to happen there was for a 'Taxon Framework Relationship' to be added to describe the relationship between that taxon and the Taxon Framework's source (POWO). I added one, in this case the relationship is 'not external': At the moment, only staff can create/edit Taxon Framework Relationships while we finish testing, but as soon as we open that up that will be a big new job for curators, to make/edit taxon framework relationships to help articulate how things differ from whatever external source we're otherwise using.

@daviswj I'd think a journal post, the google group, or a flag would be a good place to start a conversation about fungi taxonomy. There might already be an existing thread that could be picked up.

@cmcheatle every taxon automatically gets a 'taxonomy details' page eg. but the functionality to create/edit a taxon framework to a taxon, or to create/edit a 'taxon framework relationship '(which could be a match or a deviation) for a taxon in the context of some upstream taxon framework is staff-only at the moment. Once we open that up, I'll add a second tutorial, but that should explain how curators can manage this content by creating/editing taxon frameworks and taxon framework relationships.

@mamestraconfigurata I agree with you that higher taxonomy usually isn't the barrier to identification. But this functionality also allows for frameworks that extend down to species which is the goal when possible (and is already the case for chordates) but to do this we need good global species level taxonomies which I don't think exist yet for Noctuidae - do they?

@markscherz yes, if we are following a source for some branch (e.g. Reptile Database for Reptiles) but the iNat community really wants to deviate (e.g. by going with some new paper that hasn't been yet included in RD) then it can be accommodated with a deviation (which is a 'taxon framework relationship' other than type 'match'). This doesn't help sort out decision making for when we should deviate and when we shouldn't. We still need to hash out some decision making policies for when curators disagree.

@bobby23 very cool insect post - thanks for adding. Re: number of taxon curators - my vision is for there to be more like 10 taxon curators associated with most of these frameworks and for us to come up with a nice federated system to divide up curation responsibilities. At the moment, only staff can make/edit taxon framework relationships (ie deviations). But once we open that up, creating/editing those will be a big new curation responsibility. Re: switching sources, you could do this with notes (e.g. this is a match under IUCN but would be a split under GMD) but I'm not sure if that would be any more helpful than the spreadsheet since it wouldn't be searchable.

@tonyrebelo check the 'flagged covered taxa' section on the taxonomy details page

Posted by loarie over 5 years ago

@loarie Awesome, thanks for that info. I appreciate it!!

Posted by feistyone over 5 years ago

'flagged covered taxa' - cool thanks.
Can one see resolved flags as well? - or will that be unnecessary given the documentation on the page?

Posted by tonyrebelo over 5 years ago

@loarie Thanks!

I couldn't find a recent Google group; so, I started a new thread. Once it clears I'll post the link. :)

Here is the link:

Posted by daviswj over 5 years ago

Thanks Scott, I had not gone into the reptiles area, the ones I had looked at did not yet have any imported or created.

I guess that covers how it is shown that a deviation is in place. I was more thinking about the process to initiate one, the documentation of the discussion etc, the why as opposed to simply the end result. Will tracking and documenting this be a part of the framework process ?

Posted by cmcheatle over 5 years ago

That's great news @loarie—we should talk about microhylids.
New question: Is there any prospect for incorporating species complexes under this scheme? As you know, in Madagascar we have a lot of reptile and amphibian species that are complexes, and identifying individuals to species level is pseudo-confidence. Right now we have a mix, where some complexes are just assigned a more-or-less arbitrary name from the complex, while others are given to genus level, with a comment indicating which complex they belong to. Our case is a special one, I acknowledge, but it would be helpful to have a framework (or at least a guideline) to incorporate this taxonomic uncertainty.

Posted by markscherz over 5 years ago

It is not just Madagascar - the same applies in many other cases.

Why can we not just allocate "species complexes" and "species aggregates" as "superspecies"? Although not a formal taxonomic rank, they are well accepted in the literature informally.

Most taxonomical accounts do circumscribe these "superspecies", without creating formal names. Could we just not give them the rank of superspecies, choose the earliest name, and annotate them with "group" (rather than, or alternatively, "aggregate" or "complex").

Some obvious restrictions would be:

not to add superspecies which are synonymous with their genus, subspecies, section or series: merely use those instead.
use the earliest published species name for the superspecies name.
dont use compound names (such as Hypsiboas calcaratus–fasciatus group - especially when there are six species in the group).
the complex must be recognized in the literature - dont create complexes for observations where IDs can be done, but diagnostic features are often missing from some photographs - rather inform and train users to provide the additional evidence.

dont go overboard with adding "complexes" to the dictionary unless there are observations present for them.

There appears to be a demand and need for this. Should we not consider motivating for it?

Posted by tonyrebelo over 5 years ago

@tonyrebelo I think this is an excellent idea to solve the problem! It would have my vote!

Posted by markscherz over 5 years ago

What tony lays out could certainly be done, providing 'complexes' have a specified 'rank level' (e.g. 11 between subsection and species):
genus:20, subgenus:15, section:13, subsection:12, species:10
keep in mind this would mean you couldn't have complexes of complexes. It would also mean that species couldn't belong to more than one complex. But aside from the 'can we' and on to the 'should we' I'd like to see specific examples of where this is needed. Maybe someone could write a journal post describing a particular genus, its species and how complexes would really help move forward observations sitting at the genus level

Posted by loarie over 5 years ago

I think that those are two good additional rules:

complexes should be exclusive: any species can only belong to a single complex - in the event of complications: merge.
complexes are complex: subcomplexes are not an option - rationalize.

@alexanderr - please provide some herp examples where complexes will be useful: see last four entries above. (e.g. Amietia; Tomopterna; etc.)

Posted by tonyrebelo over 5 years ago

@loarie @tonyrebelo
Scott, Tony - many moth complexes would benefit. Off the top of my head, just for Hong Kong, the following genera have distinct complexes with undescribed species that fit into either named or un-named groups :- Eoophyla (Crambidae, Acentropinae), Dichomeris (Gelechiidae, Dichomeridinae), Arctornis (Erebidae, Lymantriinae), Cyana (Erebidae, Arctiinae, Lithosiini), Glyphodes (Crambidae, Spilomelinae), Bastilla (Erebidae, Erebinae, Poaphilini). One should take a look at Terry Whitaker's Pyraloidea of Borneo to see how vast the issue is. Many of the genera there have undescribed species, and a fair number of these are in distinct complexes.

Posted by hkmoths over 5 years ago

Yes, this is something that could help some frustrating ID conundrums, especially as more cryptic species are recognised.
Some herp examples (to list a few):

Tomopterna tandyi and T. cryptotis
Amietia fuscigula and A. poyntoni
Leptopelis bocagei and L. parbocagei

Trachylepis varia complex (often easy to recognise T. varia senso stricto)
Hyperolius nasutus complex (often easy to recognise H. nasutus senso stricto)

Posted by alexanderr about 5 years ago

OK: have started a post here:
@alexanderr - I need information on Trachylepis varia complex and Hyperolius nasutus complex (?: acuticeps, nasutus, parkeri) is not an issue in southern Africa where the species are geographically distinct (and the only way to ID them), but I dont have information for central Africa.

Posted by tonyrebelo about 5 years ago

See these publications:
Weinell, J.L. and Bauer, A.M., 2018. Systematics and phylogeography of the widely distributed African skink Trachylepis varia species complex. Molecular phylogenetics and evolution, 120, pp.103-117.

Channing, A., Hillers, A., Lötters, S., Rödel, M., Schick, S., Conradie, W., Rödder, D., Mercurio, V., Wagner, P., Dehling, J.M. and Du Preez, L.H., 2013. Taxonomy of the super-cryptic Hyperolius nasutus group of long reed frogs of Africa (Anura: Hyperoliidae), with descriptions of six new species. Zootaxa, 3620(3), pp.301-350.

Posted by alexanderr about 5 years ago

@alan_rockefeller any thoughts on how this conversation could include species complexes in fungi? Perhaps making the taxonomy more flexible in this way would entice more MO folks to migrate

Posted by dgreenberger about 5 years ago

Might fungi need any deviations from the "rules"?
Otherwise why not?

Posted by tonyrebelo about 5 years ago

thanks alexanderr - anyone want to wade through the 82 Trachylepis species on iNat and sort out which ones are in the Trachylepis varia species complex according to that paper? not it.

Posted by loarie about 5 years ago

I have read the Trachylepis varia species complex paper. it deals only with southern Africa in any detail and three species: T. varia , T. damarana & T. laevigata. While very similar there are several morphological features usually visible in a dorso-latero view of the body and head that should allow identification. The varia complex is more a taxonomical issue than an identification issue, and would probably not merit adding to the dictionary unless unexpected issues arise. The 5 species in the complex to the north will be dealt with in a future paper.

Posted by tonyrebelo about 5 years ago

No time to go into a lot of detail, but the mantellid frog genus Mantidactylus would benefit enormously:

30 undescribed species
six subgenera
Mantidactylus (Ochthomantis) femoralis: at least 13 species-level lineages with two available names and no way to tell them apart
Mantidactylus (Hylobatrachus) cowanii: identity of cowanii and lugubris unclarified, plus six more undescribed species-level lineages
Mantidactylus (Chonomantis) charlottae: at least seven lineages, identity of type lineage is quite clear, but most examples on iNat are not the 'true' charlottae, but instead an undescribed member of the complex
Mantidactylus (Brygoomantis) betsileanus: often more than one species per location, total taxonomic mess that hasn't really even been touched, and true species identity is unclear.
et cetera.
Actually, a massive proportion of Madagascar's reptiles and amphibians would benefit from this.

Posted by markscherz about 5 years ago

markscherz - iNat already supports subgenera, so what would complexes get you beyond that?
In all cases what we need is someone to do the work to group the actual iNat taxa, e,g,:

Subgenus Ochthomantis:

Mantidactylus aerumnalis
Mantidactylus albofrenatus
Mantidactylus alutus
Subgenus Chonomantis:
Mantidactylus ambohimitombi

Or even further into complexes, e.g.

Subgenus Ochthomantis:

Complex Mantidactylus-aerumnali-complex

Mantidactylus aerumnalis
Mantidactylus albofrenatus
Mantidactylus alutus
Subgenus Chonomantis:
Mantidactylus ambohimitombi
I'd like to see a real world example laid out like that that we can all look into and discuss exactly how general internodes like subgenera or a new rank level like complexes could or could not help with identification

Posted by loarie about 5 years ago

@markscherz Um. I think we need to be very clear.
Species complexes are groups of known species that cannot be easily identified apart. They are not taxonomical messes requiring attention. For instance:

Mantidactylus femoralis: - why not just identify this to the species? What is the difference between calling this a species requiring taxonomical attention and a "complex". This is a taxonomical issue, and not an identification issue. When the taxonomy is resolved, there may well be a species complex here, but at present it is just a "described species", that requiring urgent attention that might result in 13 species. Why not just identify them as M. femoralis until the taxonomy is done?
Mantidactylus cowanii: - the six undescribed species should just be identified as Mantidactylus (Hylobatrachus). No need for a complex.
Mantidactylus charlottae: are the lineages described? can they be not be distinguished? Perhaps then they are a complex. But at present it sounds like a taxonomical issue, best dealt with by using M. charlottae and the rest of the lineages as (Chonomantis). If you can identify the non-charlottae members then this is not an identification "complex" - why not motivate for MS name (with a brief diagnosis, and ideally a reference specimen) to be used for the taxon until it is described.
Mantidactylus betsileanus: this sounds like a mess, but it does not sound like a complex. Identify the various species as either Mantidactylus (Brygoomantis) or Mantidactylus betsileanus (or one of the other 12 species) until the taxonomy is resolved.

Your examples conveniently have subgenera to help collate the undescribed species.
But I think things will rapidly go out of hand if we create complexes for taxonomical issues. Their purpose on iNaturalist is to resolve identification of cryptic groups that cannot be identified further based on morphological features. For taxonomical issues park the undescribed species in the nearest subgenus or genus.
In southern Africa we have created an observation field to assist us with these undescribed species: see (which we manage via the project
However, we are fortunate in the Cape that we have a published conspectus that lists all undescribed species, and to which we can add new ones. (e.g. see - page 552-558 where some 19 undescribed species are listed for the genus Indigofera in the Cape)

Posted by tonyrebelo about 5 years ago

@tonyrebelo in almost all of these cases, there are several different species described, and several species that are undescribed.

I disagree with you that undescribed taxa within a complex do not make it a complex. In any case of a 'species' that we know to contain multiple different species-level lineages, it is a species complex, irrespective of whether or not names are available for those species.

I should have emphasised that because the majority of undescribed diversity in Madagascar is characterised genetically, we understand the extent of the taxonomic mess involved, and all (or almost all) undescribed taxa carry candidate species numbers. Because of this, I am always hesitant to identify something as a species that we know it is not, simply because the species that it actually is, is not yet described. For example, identifying all members of the Mantidactylus femoralis complex as 'Mantidactylus femoralis' gives the impression of false confidence that these areally are 'M. femoralis'. There should be some indiication that the term 'M. femoralis' refers to a complex and you cannot be sure which species is concerned.

For the sake of argument, even if we ignore the undescribed species, identification to species level by dorsolateral imagery is almost impossible in many cases. Identification to genus or subgenus though is an understatement of the level to which it is possible to identify the species though, and is therefore less valuable than being able to identify to complex. Identification to species level is an overstatement of the reliability of our ID.

I should note, by the way, that under the current system, iNat does not implement the subgenera for Mantidactylus, because for some reason the ASOTW database does not implement them. So while in theory, you are right, there are these convenient subgenera that would help in a lot of these cases, they currently do not. This, by the way, is a great example of a place where deviation from a certain database would make a lot of sense!

@Loarie Let me give an alternative example:
Gephyromantis plicifer complex
Gephyromantis luteus, Gephyromantis sculturatus, and Gephyromantis plicifer.
• These three frogs differ subtly in size, in the extent of webbing between the toes when they are fully stretched, in call, and in the shape of the femoral glands. These frogs are comparatively abundant in eastern Madagascar, and overlap in range.
• Generally, photos on iNaturalist are unidentifiable to species level, because we do not have a sense of scale, we do not have calls, and we do not have ventral shots that would show the femoral glands.
• Identification to subgenus would put these in Duboimantis, a subgenus of >15 species, all of which they differ from in obvious characters (strong scapular ridges, broad feet, strong foot webbing, spines above eyes…). Identification to complex level would be a considerable improvement in the refinement of identity.
• If taxonomic rearrangements are made in higher taxa, species complexes will almost always be moved as blocks of species, which means they are quite stable—even more stable than genera, for example.

To apply your scheme to this subgenus:


Gephyromantis (Duboimantis)

G. (D.) plicifer complex

Gephyromantis (Duboimantis) luteus (Methuen & Hewitt, 1913)
Gephyromantis (Duboimantis) sculpturatus (Ahl, 1929)
Gephyromantis (Duboimantis) plicifer (Boulenger, 1882)
G: (D.) granualtus complex
Gephyromantis (Duboimantis) salegy (Andreone, Aprea, Vences & Odierna, 2003)
Gephyromantis (Duboimantis) tandroka (Glaw & Vences, 2001)
Gephyromantis (Duboimantis) granulatus (Boettger, 1881)
Gephyromantis (Duboimantis) zavona (Vences, Andreone, Glaw & Randrianirina, 2003)
Gephyromantis (Duboimantis) grosjeani Scherz, Rakotoarison, Ratsoavina, Hawlitschek, Vences & Glaw, 2018
Gephyromantis (Duboimantis) saturnini Scherz, Rakotoarison, Ratsoavina, Hawlitschek, Vences & Glaw, 2018
Gephyromantis (Duboimantis) leucomaculatus (Guibé, 1975)
Gephyromantis (Duboimantis) redimitus (Boulenger, 1889)
G. (D.) cornutus complex
Gephyromantis (Duboimantis) moseri (Glaw & Vences, 2002)
Gephyromantis (Duboimantis) cornutus (Glaw & Vences, 1992)
Gephyromantis (Duboimantis) tschenki (Glaw & Vences, 2001)
G. (D.) schilfi complex—Note: arguably (phylogenetically) within tandroka complex
Gephyromantis (Duboimantis) schilfi (Glaw & Vences, 2000)
Gephyromantis (Duboimantis) tohatra Scherz, Razafindraibe, Rakotoarison, Dixit, Bletz, Glaw & Vences, 2017
… (all other Geyphromantis subgenera)

Posted by markscherz about 5 years ago

I agree with the usefulness of allowing species complexes in iNat.

@tonyrebelo species complexes are groups of species that cannot be easily identified apart. They may or may not be taxonomical messes requiring attention.

@loarie in some cases there may be named subgenus that could be used as an equivalent of a species complex. If this is the case, then I agree that is probably a better option. However, there is often not an existing subgenera for us to use. In that case, something like a species complex would be very useful.

Current DNA work is showing that many taxa that were previously thought to be valid species are actually species complexes that include several cryptic species. The iNaturalist community is trying to accommodate this research, and is including many of these new species. The problem is that these are often not a "complete" taxon, there may be several unnamed cryptic species which look very similar, or may even be morphologically identical. I think that this could be solved if iNaturalist allowed a taxa at the rank of "species complex". This is a strategy that is used by BugGuide, and I think it is quite useful there.

Two similar examples that I have recently encountered on iNaturalist that highlight this issue, for a fungus and a liverwort:

Fomitopsis pinicola is a species complex of at least 4 cryptic species. Fomitopsis pinicola sensu stricto is restricted to Europe, while North America has the other 3 cryptic species, one of which has been described as Fomitopsis ochracea, while the other two have yet to be named.

Conocephalum conicum is a species complex of at least 5 cryptic species. Conocephalum conicum sensu stricto is restricted Europe, while North America has 3 of the other cryptic species, one of which has been described as Conocephalum salebrosum, while the other two have yet to be named.

The question is what to do with the undescribed cryptic species.

In the case of the Conocephalum conicum species complex, the prevalent strategy on iNat has been to name the European specimens C. conicum and the North America specimens C. salebrosum. This is incorrect, as the majority of the North American specimens are not C. salebrosum, they are undescribed cryptic species in the C. conicum species complex. A more accurate strategy would be to name all the North American specimens "C. conicum species complex" unless they can be positively identified as C. salebrosum.

In the case of the Fomitopsis pinicola species complex, the prevalent strategy on iNat has been to apply the name F. pinicola to North American specimens that are not identifiable as F. ochracea. This is incorrect, because F. pinicola sensu stricto does not occur in North America. A more accurate strategy would be to name all the North American specimens "F. pinicola species complex" unless they can be positively identified as F. ochracea.

The other solution would be to just leave the cryptic species identified to the genus level. However, this seems like a less desirable option, if they can be identified to a finer level. It does not differentiate the specimen from other species in the genera, when we could otherwise do so.

I would also suggest allowing a species complex ID to have a Research Grade designation. This would remove the incentive to try to identify it to a finer level, if there is no way to do so with confidence.

Posted by stu_crawford about 5 years ago

@stu_crawford - correct! "species complexes are (closely-related) groups of species that cannot be easily identified apart (by external morphology). They may or may not be taxonomical messes requiring attention."

And: the concept should not be used for taxonomical messes in which the species can be told apart, but are currently merely unnamed.

Posted by tonyrebelo about 5 years ago

I guess that I'm just not familiar with examples of where the species can be told apart by external morphology, but have yet to be described. If they haven't been described, how do we know how to tell them apart?

But either way, where I see the value of a species complex designation is for taxa that are a taxonomic mess AND are very difficult (or perhaps impossible) to tell apart from external morphology (at least based on our current knowledge).

Posted by stu_crawford about 5 years ago

@stu_crawford Perhaps a simple example; Moths (noctuidae) are phenotypically variable. Two common species of Xestia can only be told apart by genital dissection. That is not possible with a photo. They are not a taxonomic mess, just cannot be identified by sight alone, so are classed as a species complex. The default setting on iNat has to be Genus.

Posted by mamestraconfigurata about 5 years ago

@stu_crawford in Madagascan herpetology, species description lags far behind discovery, even if new species are clearly identifiable, simply because we discover more species per year than we can describe.

Posted by markscherz about 5 years ago

Getting back to the question of a consensus higher taxonomy - I have a global taxon database that I maintain ( which also needs to address this question, and I have decided that for extant taxa I will attempt to follow Ruggiero et al.'s higher classification (down to order) as given at (updated version also available, with some minor errors corrected, here: ), except where already superseded or where I decide to deviate for my own reasons. (Extinct taxa are a somewhat different matter). For families, there is/was an extended 2014 version (i.e. a few discrepancies vs. the 2015 version) called "FALO" (Families of Living Organisms) which is downloadable (as "GGI Family Data") from . Just mentioning these in case you have not seen them - they represent a "consensus taxonomy" acccording to the usual players (with whom you may agree or disagree) but are at least one standard that can potentially be followed if desired, down to order then family.

For extinct taxa, I try to follow major sources but they sometimes need a bit of modification (e.g. for fossil plants) to make them play nicely with the Ruggiero et al. classification. I can supply more info on the latter if interested... Regards - Tony Rees

Posted by tonyrees about 5 years ago

'complexes' are up - see for a pilot use case

Posted by loarie about 5 years ago

Kool. Thanks!
Most of these taxa will be in "curated taxon frameworks" Do we just flag these for the taxon curators?

Posted by tonyrebelo about 5 years ago

that sounds like a good approach. FWIW I'm in favor of using sparingly and reactively to emerging issues (rather than proactively)

Posted by loarie about 5 years ago

in case folks are interested, we rolled out some changes so that Taxon Framework Relationships can be edited by Taxon Curators. But we're purposefully just piloting them with a small group of mammals at the moment until we get the kinks worked out. You can read more here:
if you want to get involved with the mammal pilot please chime in there. But also, we hope to roll these out more broadly to other taxon frameworks as soon as we finish this test with mammals so you can just hang tight.

I realize that I unintentionally made myself the bottle neck for lots of taxon curation things by locking down a lot of these branches before these tools to get more curators involved (via taxon frameworks) were ready for show time. I apologies for that and am looking forward to removing myself as the bottle neck ASAP by expanding the set of taxon curators with permissions and training to curate each of these Taxon Frameworks

Posted by loarie about 5 years ago

I'm outraged! What about moths! Joking of course. It really will be good to have some guidance surrounding the whole issue of complexes. Both because it's difficult to verify some moths below genus, and how this will work with folks doing research. I read some of the comments on the other thread, and some were pretty heated. Thanks for all the hard work, and I hope this works out.

Posted by mamestraconfigurata about 5 years ago

But if it is difficult to verify below genus, then just use the genus as the ID. Complexes are for small sets of species within large genera that cannot be identified by photographs. If the bulk of species in a genus cannot be identified, then generic level IDs should be suitable in the majority of cases. Dont create complexes if "synonyms" in genera, subgenera, sections or subsections already exist.

Posted by tonyrebelo about 5 years ago

@tonyrebelo I suspect it will be a while before this all gets down to the moth level. There is a North American 'complex' of Xestia c-nigrum/dolosa, which are different species but cannot be distinguished by pattern alone. Range can make a difference, but for the most part these common moths cannot be told apart by photos alone. There are no synonyms in the current taxonomy as far as I know. All I want is to know what to do with photos which are identified to species. It's not really a big problem.

Posted by mamestraconfigurata about 5 years ago

It will be a while before taxon frameworks gets to moths, but you can create a complex right now:
Xestia c-nigrum/dolosa - sounds like a good candidate - lots of species in the genus but this species pair indistinguishable and overlapping in part of range. (if there is total segregation then I would argue ID it by location and dont create a complex: but be upfront about how you are doing it: after all if in the future someone finds a feature, they can just be reviewed; if they are found to overlap, then the problem is really only in the overlap area - outside this they can tentatively by ID'd to species).

Create it and see if it works for you!

Posted by tonyrebelo about 5 years ago

This is a great step forwards for many things! Thank you for all the work!

Posted by amarzee about 5 years ago

Excellent ! Yet " for a spirit of compromise and flexibility" you may leave away the Clement's check-list for birds.

Posted by jujurenoult over 4 years ago

On another topic, it would also be good to recommend uploading more photos of bacteria, Glaucophyte archaea, acanthocephalans and other underestimated minor taxa.

Posted by srsvivos 5 months ago

Add a Comment

Sign In or Sign Up to add comments