This has been a recurring problem with the taxa I work heavily on, but one curator, who is an expert/scientist I might add, has added a long list of complex taxa that do not show in our taxonomic source. Keep in mind that our taxonomic source does list plenty of “superspecies” and we have been using complex rank as an equivalent. It also creates the slightly bothersome issue that since iNaturalist does not have the superspecies rank, and we cannot have a “matching” relationship in the taxonomy framework, so no matter what I do, the complex is going to be listed as a “deviation” and the species within the complex are alternate position. I do have a spreadsheet that is keeping track of “true” deviations, and taxa that don’t transition well into iNat. So far, we only have 2 true deviations; a hybrid taxon and an undescribed species that has over 500 observations (we were also promised a name for this species like 4 years, which is why we kept it).
There were many more deviations in my local for this group of organisms a couple years ago, but we’ve pretty much rolled back on all of them in order to be in agreement with the taxonomic source, even if we disagree with them. Now we have this long list of complexes with no framework relationships, simply because the complex is not listed in our taxonomic source. The curator who added them is hard-set on the idea that these deviations are necessary, despite most of them being monotypic (only containing one species).
Anyway, this is the advice I’m looking for:
It was my understanding through the curator guide that complexes are to be used sparingly, so monotypic complexes are not ok. Is this true? (btw, I have never a monotypic complex, though they do occur in our taxonomic source)
On a totally different debate earlier this summer, it was affirmed that we identify according to the taxonomic source. If we are creating complexes purely because you don’t want to use the species rank as a sensu lato, does this fall under creating your taxonomy?
When are deviations “acceptable”? Like previously stated, in my area of the world, we had a lot of deviations, but have gotten rid of all but two in the past year, simply to follow the taxonomic source. Though the deviation was almost certainly the correct application.
This is why deviations exist. Blindly following an external authority purely for the sake of following it is not good practice. So you already answered your own question re
Yes there will be scope for interpretation and some subjectivity as to what the ‘correct application’ is. But if the consensus is that a deviation is useful and allows for a better representation of a taxonomic relationship or standpoint, then there is no reason not to implement it.
I think monotypic complexes are very strange. They don’t really seem in line with the curator guide, but if someone wants to put in the curatorial effort then I won’t argue too strongly. Are there purported undescribed species in the complex?
But staff told us earlier this summer that we should be identifying to the taxonomic source (this is over an entirely different topic), so why is changing the taxonomy to fit your ideal of identification acceptable?
Once again, I state that here in North America, we as a collective group of Orthopteran experts decided to roll back on our deviations because the purpose of the taxonomic source is to follow it. Curator guide even states that taxonomy is subjective, that’s why we have a framework, so wouldn’t deviating be the exact opposite of that policy?
That’s what I’ve been told, but I don’t see why undescribed species justify the complex. For example, there’s one genus I’m helping to revise, and we know about 45 undescribed species, with about 4 superspecies/complexes, of which 3 of them already have a named taxon within them. But I identify everything to genus level unless it is the OG species in the complex, because I’m identifying in accord to the taxonomic source, and I’m not making placeholder taxa, which is basically what a lot of these are. I don’t see why these “deviations” can’t be identified via observation fields.
I did not make any of the complexes you’re referring to, and I have nothing to do with Orthoptera curation, so I haven’t changed anything to fit any ideal, mine or otherwise.
ok, that’s great for you that you all agree on that, but I’m unsure why that is relevant to the situation you described in your original post, which involves Australian taxa, not North American taxa.
the curator guide in no way mandates that each external authority must be followed. Indeed, it explicitly states that there are many cases in which we do deviate:
All that being said, there are also many situations in which we do not (or cannot) follow authorities
The guide then provides direct hyperlinks to explicit deviations for many of our major external authorities. If it was dictated that we must always exactly follow every name and decision made by our authorities, then why does the functionality to deviate exist?
Deviations are perfectly acceptable provided they make sense, are applied judiciously, offer utility, and are clearly documented.
I have no direct, strong opinion either way on the specific complexes you’re referring to in your original post, and am largely indifferent about whether you get rid of them or not, but from a broader curatory perspective, there is nothing inherently ‘wrong’ or ‘bad’ about deviations on iNaturalist. So any decisions to remove or retain these complexes should be based on legitimate reasons rather than just a blanket principle of ‘we should get rid of all deviations because they are deviations’
I think we could reflect on why researchers use complexes in their works. It is just because complexes are convenient in order to pinpoint a group of species that are, at least morphologically, related with each other in those cases sections or subsections are not available of considered not suitable.
In this light, I think that convenience could be a good rule for the use of complexes here in iNat. As regards, the introduction of a given complex can be convenient here if, for example, the species x, y and z are hard to be identified with photos and/or often revised in literature thus requiring extensive re-identifications.
Of course, what can be convenient here in iNat is not necessarily convenient also for researchers. Thus, it is not granted that researchers have already addressed in literature to a certain group of species as, for example, “the group/complex of Species x”.
All summed up, I would humbly suggest to consider the possibility to slightly loosen the rules for the adoption of complexes and to try to evaluate the possibility to take into account the usefulness/convenience here in iNat of a possible complex, even in the cases it has not already been cited in literature.
With this I do not intend to legitimize a sort of taxonomic anarchy. Of course, it would be important to validate such complexes by expert users and to a void any misuse.
This is the main reason I add complexes, and I think it’s a very good reason. I will only add complexes if IDers can’t ID to species from most/all iNat observations, IDers will use the complex and manually mark observations as RG where applicable, there is a mention of the complex in the literature, and the species form a clade (as best as we can tell). That’s more strict than the curator guide, but I think it fits both the community needs and the taxonomic situation.
This actually touches on a much bigger issue, which is the fact that taxonomists are moving more and more to unranked (non-Linnaean) clades. A huge amount of the current spider taxonomy cannot be reproduced on iNaturalist because of this. Do we need to make a feature request for an “unranked” rank? Or is the current limitation a blessing in disguise (preventing us from creating an infinitely complex unmaintainable taxonomy)?
Not exactly to the point of your three questions, but regarding using complexes in general I think (at least with herps, my area of interest) complex taxa are best applicable on iNat only for when an observation cannot be IDed and brought to research grade as a species because of visual similarity with other related species (i.e. Gray Treefrogs, Hyla versicolor and H. chrysoscelis). Complexes should not be used for populations that are visually identifiable but are genetically in an area of uncertainty because of the poor methods of iNat’s current sources (i.e. Agkistrodon sp., Lampropeltis sp., Pantherophis sp.). Rather (at least in my opinion) iNat should attempt to use more scientifically supported taxa, even if that means deviating from sources. I do understand the need for consistency, but iNat’s current ‘up-to-date’ herp taxonomy sources are often very poor done with limited sampling and data, hence the ‘areas of uncertainty’. iNat is currently using hybrid or complex taxons in these situation, which I think is incorrect usage.
Yeah, what is going on with spiders? " Zoosubsection Oval calamistrum clade" “RTA Clade”
Personally I’m in favor of more taxa ranks being added. So that there doesn’t need to be more inelegant, technically wrong taxa groups because INat only supports certain taxa levels.
Issue with West Indian orchids: several complexs have been introduced recently, which allows to go ahead the genus and give closer information for IDs, especially when flowers are not available. One example is the complex Epidendrum difforme with many very closed taxons difficult to identify. But strangely, others have appeared like the very variable Epidendrum ciliare, may be due to some proposals of splits mostly based on ecological induced variability. in that case, it may be barely acceptable but of limited interest. Or, others that add confusion to confusion. In the case of a group of endemic Epidendrum, complex E. mutelianum : E. boricuo-mutelianum, E. mutelianum, Epidendrum pallidiflorum, E. patens. One endemic to Montserrat has been forgotten, and a quite different species from Brazil has been added to this complex ! There is no issue of identification when flowers are present but historically, based on dry specimen analysis, these taxons have frequently been mixed in floras and herbariums (today corrected). When only vegetative plant is present, it is less easy to make an ID except for the endemic (to 2 small isands) Epidendrum mutelianum. But the name of this group has been based on that one with narrow distribution, easy to identify and not one of the more largely distributed and common E. patens, less easy to identify when not in flower.
It would be helpful to complete the complex with all but only the related species (adding E. montserratense and removing Epidendrum kautskyi) and to choose a better name like Epidendrum patens. In some other cases, identification of some West Indian populations of relative species is really difficult if not impossible bases on pictures with the available literature and would beneficiate of the creation of a complex (i.e. Epidendrum anceps including E. amphistomosum, and other taxons).
In conclusion, I think I understand only part of the rationale of the creation of some of these complexs that in some specific groups, may request a field knowledge in addition to taxonomic skills.
I may give the useful information to the curator(s) if requested.
A lot of comments have been made while I was gone, hopefully I can round back on all the things I’d like to cover.
I will touch on this later on, but it seems like accepting these would muddy the waters between accepted complexes (because our source does accept complexes) and complexes that are simply best guesses. Like I said before, I have been told on more than one account that most of these complexes are being used as placeholders for suspected undescribed taxa, so it just seems these Australian curators are trying to find another loophole to add these undescribed taxa (which they already tried a few years ago when a couple Genera novums were added to the taxonomy with the justification of “Rentz 2001 listed it”)
For the complexes in question, many of them are either monotypic (only one species listed) or 80-90% of the observations in the complex are already RG at one of the species listed. Most complexes I have looked at tend to have a much lower RG rate. Melanoplus sanguinipes has a RG rate of 35%, and that’s probably a little generous because identifiers are agreeing with species level ids on female subjects.
I think a good clarification between herps (and all vertebrae for that matter) and insects, is that our taxonomic authority lists complexes. To date, we have approximately 230 complexes in Orthopterans that are recognized and listed in our taxonomic authority. So in a way, we aren’t truly “deviating” by adding these taxa, which is why I bring up the complexes that are specifically not recognized by our source. In herps and mammals, it’s different because your source doesn’t not list these superspecies, and it’s useful to add them because of the overzealous attempt to split everything phylogenetically.
I probably wouldn’t create these complexes, but just because I don’t think it would be worth my curatorial effort. Others are welcome to disagree.
In the case of monotypic complexes with undescribed species, I would just ID everything as the one species and then split it once the other species are named. But that’s just my opinion.
In some cases where species are found to be complexes I have seen people edit the rank and change it to complex (with no children). This is generally a bad idea for a variety of reasons and I believe staff have discouraged this.