Identifiers as Volunteer Curators, and Obs with Probs

I know what this is about, and I fully support you.

This is just to comment and voice my support as long as this post is still visible.

The issue you refer to should be addressed. Curators have no tools(*) to deal with this, so it will have to be a public discussion related to site policy.

(*) Well, now curators can hide identifications, but that tool wasn’t available until very recently.

1 Like

In general this is done by museum staff or researchers using the collections. Accessing a collection can be quite a task (need to go to a specific physical location, so travel costs can be serious, need to find the specimens and pull them out of storage, need to handle very carefully in the case of older/fragile specimens). Because of this collections are often only examined if there is a specific need or reason like someone doing a specific project, suspicion that there is an error needing to be corrected, or general updating/moving of a collection (the last tends to be very infrequent)


I ID (and study) African butterflies. You’ll find observations of “mud-puddles” (congregations of butterflies feeding at mud, excrement, etc.) featuring several species in the photograph.

My philosophy is that a photo which contains several species is still a record (an observation) of any one of those species. An observations that remains “Pieridae” is much less scientifically useful than an observation identified as “Belenois solilucis” (even if there are also Belenois aurota, Belenois creona, and various Terias in photo).

My approach is as follows:

  1. Discern what the observer is intending to record with the observation, and identify that.
  2. If it’s unclear which individual butterfly the observer is intending to record, and I’m the first identifier, I’ll identify the most interesting (& usually the least commonly recorded) species in the photograph (or, sometimes, ask which is intended).
  3. If it’s unclear which species the observer is intending to record, and someone else has already offered an identification, I’ll follow their lead.
  4. Once I’ve left an ID, write a comment suggesting the observation be split so that species X, Y, & Z can also be recorded.

I feel like this approach is good because:
a. It aims to honour the observer’s intention, it provides them with value (a specific ID), and it directs them to better practice.
b. It, at the very least, leads to the observation being of practical scientific value. (A record of a single species in a particular location is much more useful than a generic record which covers several species.)

(With all that written, I hope I’ve understood the issue being discussed…)


Your way of handling such a case is fine. However, the type of observation that is creating problems is a bit different – not multiple species in one photo, but multiple unrelated photos showing completely different species (for example, a rose, a giraffe, and a pigeon) in one observation. This is a relatively common mistake with new users who don’t realize that they need to create a new observation for each organism. It also sometimes happens to more experienced users, for example, if there are multiple similar-looking insects or birds flying about and the observer doesn’t realize that not all the photos show the same species.


First of all, I’d just like to say that I’m very sorry you had unpleasant experiences, and I’m sad to see you leave iNaturalist. We haven’t interacted much, but I’ve always appreciated all the work I see you putting in.

I think a major part of the “regional differences” in iNat culture come from certain areas previously using a different website which migrated into iNaturalist, with a pre-formed community that still sticks to the culture and rules of the old site.

From my own experiences, I deliberately exclude the African continent from my identify queue, and the only time I venture in there is if I’m specifically tagged in something. It’s just too stressful to interact with. Fortunately California has enough observations to keep me busy for a lifetime anyway…


Thank you so much, @graysquirrel! I normally would have avoided African observations also, because of prior bad experiences trying, but in this case there was an appeal for people to help reduce a large backlog there and I thought things might be different.


I handle those the situations in the same way.

ID the first image (that way something in the observation provides value, and it is an observation of a specific species after all), and request the user split the remaining photos into separate observations.
It’s not uncommon with the mudpuddling scenario that the user will aim to be helpful and post a 2nd, 3rd, 4th photo of close-ups of the species they’re observing, not realizing that photos 2 and 4 are a different species than in 1 and 3.

Edit: I guess I just don’t see any practical upside/benefit for IDing to the lowest commonly shared taxon.

The problem is that you end up with photos 2 and 4 being associated with an ID for species A, when in fact they show species B.

If photos were viewed only in connection with the observation, this would not be a big deal, because one can see an explanation about which photos are what. But this is not necessarily the case – photos may be viewed or distributed in other contexts (taxon pages, downloaded/shared with other sources) with the ID that is assumed to be correct. If it is not obvious (like a giraffe being labelled a rose), then you end up with mislabelled photos that are not recognized as such.

I ask the observer to edit the photos in such a case, of course, but there is a non-insignificant portion of cases where the observer is no longer on iNat and the observation never gets fixed.


First, I think we should remember that everyone here is trying to do the best they can to improve iNat, even though we have different philosophies and ways of going about it. I think it’s also best to acknowledge that there will always be messiness, gray areas, and edge cases.

With the upcoming mobile app, I suspect this problem will be reduced due to the new observation workflow and guidance in it. It will still happen of course, but I think the occurence will be greatly reduced. :crossed_fingers:

There’s official guidance here, although it’s buried and should have its own FAQ.

If you see an observation that has two or more photos depicting different organisms in each, it’s best to identify to the level that fits all photos and make a comment politely asking the user to separate the photos into different observations. For example, if the first photo is a flower and the second is a bug (without the flower), identify it as “Life”. If the user is unresponsive to requests to separate the photos, mark the observation as “Based on the evidence, can the Community ID still be confirmed or improved? No, it’s as good as it can be” in the Data Quality Assessment section.

I can make a new FAQ for that today.

I totally understand where you’re coming from, but I’d say there are a few issues it causes:

  • observations record encounters with individual organisms, and the photos/sounds attached to the observation are evidence of that encounter. So IDing only the first photo is going against the system and definitions everyone is being asked to follow, and on which iNat is based, and it’s an inaccurate evaluation of the evidence provided.

  • Because observations are defined as such, all photos from an observation are included in the taxon browser for the observation’s taxon, which creates confusing and weird situations. The photos might be used to train the computer vison model as well.


Yeah, OK, I had skimmed the help/FAQ pages to see if there was any official guidance, but advice for observers is not where I would be looking as an IDer trying to figure out what the best practice is for dealing with observations with problems.

If I might quibble about semantics:

it’s best to identify to the level that fits all photos

This is ambiguous and would apply to both “ID as life” and “ID to lowest common taxon”.


Thanks! I did added it to the Identifications section as well and tried to clarify the text a bit.


An observation consisting of multiple photos, each of a different species, is a useless observation in its current form. (This is very sad; often the photos are good quality and sometimes one or more of the species is unusual in some way.) It’s important to ask the observer to split the observation, giving the link to some instructions. Sometimes the observer even splits it! But what should we as identifiers do after the observation has sat for months or years in a useless state?

To me, it doesn’t seem to matter much. (Just IDing the first photo seems the wrong answer because it’s misleading, though I’ve done it at times, by mistake.) We just need to minimize the harm the observation causes. Move past it? Fine. Click “reviewed” and move on? Fine. Better options might be to get the observation out of the way for all of us. Label it “Plants” (if the photos are all plants) and comment about the need to split? Fine. Get it out of the way of people IDing plants by IDing it as “Life” with a hard disagreement? Fine. Label it “casual”? Fine (assuming it’s been sitting there for a while; this is not the best choice right away, when the observer might still split it if encouraged to do so).

What wouldn’t be fine would be my putting other people down for choosing a different approach that I would use. My acting as if my approach is THE approach, the ONLY good choice. There isn’t one.

Those of us who ID a lot can become very terse. That could be understood as insulting or arrogant, but it’s generally not intended to be such. (In a few cases, I wonder, . . . but then I suspect I’m seeing the kind of frustration I feel when I explain for the millionth time that the posted plant is Wild Carrot, NOT Poison Hemlock.)

Should we discuss this and establish a “best practice”? Well, we can discuss it until the cows come home, possibly even come to a conclusion, but I don’t think it will matter. Why not? New IDers won’t know about it. Experienced IDers will forget. And some of us (I’m looking in the mirror here) are stubborn bastards and won’t do what we decided unless it’s what we wanted to do anyway.

What might help is putting in a button we can click to mark a multiple-photo-with-multiple-species observation. Then whatever has been done to its identification, it could go out of the way (to “casual”?) until the observer redeems it.


I see why’d you exclude African observations, I’ve considered it at times. However, I’ve found the issue lies with only a handful of users. Prominent ones, but not representative of all observers in Africa.

Because of this, I don’t see it fair to avoid all African observations as there are plenty of well-intentioned and friendly people there on iNat. I wouldn’t want those gaining an interest in nature to not get attention on their observations because of some bad actors.

I’m not sure if this is exactly what you meant, but I think it’s something to consider.


I was not very involved in the Africa Mission Impossible effort, so I may be missing context. I think the question is how should we treat observations where a user posts multiple photos and the intended subject of the photos appears to be different species.

My process:

  1. If I encounter a new multiple species observation, I usually add a comment without ID: “Looks like you’ve observed a few different species in the photos here. Each species should be on its own separate observation. There are some tips for splitting the observation up here: Please make a comment if you split the photos, so I can update my ID.”

  2. If I encounter a multiple species observation that already has a comment, but no ID I will ID as “Life” and make sure that comment includes the link for help and request to comment if they fix the problem. I think this is the step of concern, so I will elaborate below.

  3. If I encounter a multiple species observation that has a high level ID already (Life or Otherwise), I add an ID that is the most specific common ID I can add. I make sure that the comments include education, the recommendation to make a comment and that I believe this is the best ID that can be provided until the species are split. At this point, I check the DQA for “as good as it can be”. I don’t see the need to bump a multi-species observation to “Life” once it already has another appropriate high-level ID. “Agree” and DQA is less work.

  4. I monitor Casuals for 3 US states. If I catch a multiple species observation being DQAed without notice/education to the observer, I add a comment similar to the one above.

I really focus on the education/communication to the observer because I think most of these are mistakes. I want observers to have the information that will enable them to have a better experience on iNat. I know that not all IDers have the patience for doing this.

I add “Life” an an interim ID so that it will get into the pathway of high volume IDers who are willing to help with these situations. I know that most of these IDers will make a second ID that is more specific when appropriate. If not, I can do that when I get the notification of 2nd ID. I also have a project that is monitoring Needs IDs unknowns for my primary focus area. So if something is at “Life” in this area, I can come back to it in the off season and tend to it.

I hope that making the Identification experience better will lead to more engagement and improved IDs over time. People IDing high level IDs (not Life) are likely are looking to improve Plants, Fungi, Athropods, etc. We really need this. These people may also be less familiar with community conventions for what to do with multi-species observations.

What next?
If we aren’t going to do ID to Life anymore, we need a work around. Perhaps we could add them to the Variable Species and Other Life project. I suspect that will be more work and I don’t think my above approach is devaluing anyone’s observations or IDing efforts. But I’m happy to adjust to evolving community norms.

Edited: I removed an idiomatic expression that I thought might not work well on this multi-cultural forum or in the context of recent events.


Oh, I agree, it’s definitely not fair to them. However, my own mental well-being has to take precedence. Those kinds of very confrontational interactions with a few users can take a big toll on me, so I find it’s better to put my energy into other areas.
Also I don’t know much about African taxa anyway, so I’d only be doing basic IDs there, and there’s more of that work than I can keep up with in my own area anyway.


That’s fair.

I admit, I’m guilty of posting observations showing multiple species. (Case in point: The Bluet Brothers.) I knew starting out, though, that one species per observation is the rule, so I make certain to say which organism is the target of the observation. Since I ID too, I don’t want to make the job of anybody who’s making an ID for me any harder than it has to be.

Any time I run across a new-ish user posting multiple species in one observation, my practice has been to leave a note along the lines of “great photos; you’ll be more likely to get IDs if you break them out into separate observations. Feel free to drop me a message or @ me for assistance or questions”. I don’t know if there’s a greater number of multiple-organism observations being posted through the website or one of the apps, although I suspect it’s the latter. Maybe have the ‘how-to’ posted more prominently in the app interface? I only use the website on my laptop, so I have no idea how the intros are set up for mobile devices.

I do think that a more prominent guide on how to handle the multi-species issue—both for observers and IDers—would be helpful. Thank you, @tiwane, for getting that ball rolling.


I don’t know that the clashing of different communities with different cultures is necessarily only an “Africa” issue, and I suspect framing it as such might limit this conversation in ways that aren’t productive. It is a more general issue: I see it with different taxon specialities, too, or sometimes even with different groups within the same taxon.

(A couple of the international IDers in my taxon of interest have very strong opinions about how things should be done; this may harmonize with the practices of the other taxon IDers in some regions, but it is not always the same as the culture among the local IDers here, and there are definitely a few times I have been left badly hurting by a curt comment or criticism when I was already doing my best and feeling swamped by too many observations and not enough IDers. This is exacerbated in situations where there is both a power imbalance – the user in question being a major expert in the field – and a discursive imbalance – the high-volume IDer who expects others to read the comments they make but doesn’t read their own notifications, thus turning their instructions into a one-way communication.)

I think it’s fairly normal that people who work together closely develop their own ways of dealing with recurring challenges, and likewise normal that these strategies may not always be identical with the practices developed by other groups. There isn’t inherently anything wrong with this – some work flows and solutions may be better suited to certain constellations than others.

It is a problem when these differences lead to an inability to be flexible and accommodate others outside the group, when they lead to unilateral enforcement of norms that are not shared by the iNat community as a whole, and when other users feel bullied or intimidated by the norm enforcers rather than being treated as equals.

It is also a problem when things are managed by unwritten community norms that rely on complex insider knowledge when they should be managed using other, simpler and clearer mechanisms.


Unlike any other section of the DQA, this question “…can the community ID… be improved,” requires the participation of two identifiers. The observation has to have a community ID (2+ individual IDs) order to go to casual; marking the box when you are the first identifier does nothing until a second identifier arrives.

Friction happens when only one of the two identifiers knows about and intends on following the help page procedure. Plenty of people simply don’t know about it. Some people know about it and don’t like it. But truly it is this requirement for two identifiers cooperating that makes the potential for conflict. And again, no other QDA section has this. When you want to mark some captive, etc, you can do that all by yourself.


Certainly not, I’ve also noticed differences in “ID culture” between US and Europe, for example, or even eastern US and and southwest. There are also differences by taxonomic group, e.g. insect ID’ers handle things differently sometimes from plant ID’ers.

One thing to keep in mind when bumping things back to “Life” is that iNat essentially treats that as an unknown. By default, when you check the icon for “unknowns” in identify, you get unknowns + “Life” and that adds stuff to the queue for those who are trying to clear out as many unknowns as possible. There is I think a URL hack to get only the “true” unknowns excluding Life but not everyone is aware of it or uses it.

Another thing to keep in mind when using DQA to “dump” things into casual is that there are a number of identifiers who look through the casual pool as well. This already comes with its own challenges (e.g. no way to automatically exclude casuals that already have a community ID) so any additions to that pool have the potential to make it even harder to work with. Not every region has people dedicated to doing this though, so it might not be an issue for some areas since nobody will notice while it does annoy a bunch of folks in other areas where the casual pool is more actively monitored.