How much expertise should we have when identifying?

To add a little bit of nuance that at first may seem completely contradictory to my previous post, it CAN sometimes be good to ID things well below the level you actually know anything about. I can think of two scenarios in which this is the case:

  1. at higher levels, it’s often extremely useful to give ANY guess at all, so that the observation can be seen by people who may be familiar with things that just look like that. For example, let’s say you find a green filamentous thing in the water. Even if you know absolutely nothing about ANY kind of algae, IDing it as a “green alga” (chlorophyta) will help it be identified much faster than if you just leave it at “life”. The people who are familiar with green algae will know that it is not a green alga, and will likely be able to at least ID it as a red alga (or whatever).

  2. Sometimes, with some taxa, and at certain times in this platform, it is much better to give what is probably a wrong ID simply to make it easier for a correct ID to be provided eventually. Let me give you an example- Let’s say genus X has 15 species, and you see an observation that could be one of three species, but you have no clue which. Unfortunately, there are very few people identifying that group of organisms in iNat right now, and none of them are particularly knowledgeable about it… but you think it IS possible to ID it to the species level, you just don’t have the information to do so. In addition, there are already hundreds or even thousands of observations of that genus that need ID right now. In that case, giving an ID that is more likely to be wrong than to be right (2/3rds chance of being wrong) is actually the best choice. That’s because if you leave it at the genus level, it’s likely that nobody will ever see that observation again. It will probably just get lost in a sea of observations needing ID, and nobody will ever go through them. An ID, even if it’s wrong, will help narrow down the options, and it does actually help get you closer to a true ID.

Ok… I think I’m done for now. I like “overanalyzing” things :P


@mnharris you’ll find many more thoughts on this here:


let’s address the specific example of Eschscholzia californica. i think what a botanist might do if they’re not super familiar with E. californica is consult a dichotomous key to aid in identification to the species level. in this case, it looks like all species of Eschscholzia occur in the US. so if you can confidently get to the genus level, then a US dichotomous key for Eschscholzia should work for you worldwide.

looking at a couple of respected keys available online, it seems like identifying E. californica is fairly clear-cut, if you can get a look at the receptacle.

Jepson ( says that if the outer receptacle is rim 0.5–5 mm, then it’s E. californica. other species have receptacle rims 0-0.3mm.

FNA ( says if the receptacular cup has a spreading free rim, then it’s E.californica. other species don’t have a spreading free rim.

they’re basically saying the same thing, just in different ways. so in this case, it seems fairly straightforward to identify to species if you can see the receptacle. otherwise, you can use the other details in the species descriptions from these sources to help you narrow things down.

you can also use ranges from Jepson, iNat, BONAP (, and other sources to help you identify, as well.

i would try the keys out on some previously identified observations, and once you’re comfortable, i don’t see why in this particular case, you couldn’t confidently get to species in a lot of cases, or at least do what some others suggest and identify to genus but suggest a species in the comments.

EDIT: if you’re more of a visual learner, this page ( has this graphic (image), which shows what the receptacle and the receptacle rim look like.


I take a more relaxed view, particularly for the Neotropical, African and Oriental regions. Please help with IDing, even if you cannot be 100% sure! There are not enough IDers around, and the few have to deal with 10- or 100-thousands of observations. Therefore, I think it is very helpful if there is a reasonalble headline taxon. It could be one subfamily off, or one genus wrong, but still it will be searchable under that taxon.

Let me make an example: if you see a brown butterfly from Central- or South America, it’s likely to be a Satyrinae. Now, if people are very strict and cautions, it will be labelled as “Lepidoptera” (could possibly be a moth?) and the observation can be found within 270,000 other observations! If the observer is sure it is really a butterfly and not a day flying moth: “Papilionoidea” reduces it to 138,000 other observations. This is obviously not very motivationg to scan through those numbers of pictures to find what you want! What I am saying is, I would be courageous and help ID it the most likely subfamily or genus, even if it may not be 100% sure. “Satyrinae” would make it 18,000 observations, and the popular genera “Taygetis” or “Pedaliodes” would reduce it further to 1,100 or 640 observation, respectively. A specialist is much more likely to sort out the few wrong identifications in Taygetis or Pedaliodes, as compared to look through hundred-.thousands of 100% secure Lepidpoptera. - What would really help the community, would be a search with a photo as input, not a text! I have suggested that elsewhere. That means, I would search with a photo of a Taygetis and I would get all the observations with similar photos, regardles of what the headline taxon says. Think about it. Thanks.


It seems like this problem might be important enough to not allow people to do that?

1 Like

I am fairly new to iNat and not an expert in any particular field, but I recently discovered that even I can help get Observations IDed.

It started with the City Nature Challenge. I participated in a web conference after the weekend to help get more Observations in NYC to Research Grade. The organizer suggested that those of us who don’t have a specialty could help steer Unknowns to a better ID. It turns out (most of you probably know this) that experts search for Observations in their area of expertise–and that new users often don’t think “Plant” is enough of an ID to be worth using, or maybe they haven’t figured out ID at all.

So I IDed a bunch of Unknowns. Some of them got refined to better identifications.

This weekend I tackled a bunch more Unknowns. I found about half a dozen dandelions that were not IDed by the observer! You don’t need to be an expert to ID common local weeds, and you’d be surprised how many of them need identifying. (I have also come to suspect that this is not due to the user’s lack of knowledge about nature but rather their unfamiliarity with using iNaturalist. There is a collection of scripts you can use or adapt to educate users about improving their Observations:

Imagine my glee this morning when I had 15 notifications, almost all because something I’d dragged out of Unknown limbo had been IDed.

Also, if you don’t feel confident working at the species level, check out this recent discussion:

Do you know the difference between alive and dead? Flowering or not? Can you recognize an egg? You might be more “expert” than you think you are.

Just observe the caveats mentioned elsewhere in this discussion and don’t identify anything to a greater degree of accuracy than you are sure of. I’ve come to think of it not as a one-time ID, but as steering unidentified Observations toward a better ID down the road.


More detail than my last reply: When a person posts an observation, and then agrees with somebody else’s species level ID, we could at least reject it unless the ID has a comment. That rejection could include text explaining that an explanation for your ID is required.

Or agreeing with a species level ID on your own observation could be entirely rejected.

This could be an issue for folks who are more conservative on their identifications. For instance I will often do a genus level or other appropriate level on my records and write what I think it is as a comment for instance (not that I cant ID this species just an example) do an ID of Larus and write I think it is a Ring-billed Gull.

This would block people from doing this or force them back to id’s they are trying to be cautious on.


No doubt it would have disadvantages, I’m asking if people agreeing with other people’s species level IDs without any reason is enough of a problem for this to be worth doing.

I think you’ll appreciate this forum thread. And we appreciate your help!


The latter is the thing that kind of drives me crazy. I’ve come across a lot of things where the observer didn’t put anything, and reflexively agreed with the first person to give an ID even if it was clearly wrong. Correcting it means getting two observers who are specialists or else it gets bumped down to the consensus (typically order or phylum), unless they do the same automatic agreement again.

That said I’m not sure blocking it is a solution. For things like insects it would mean very few research grade IDs.

Sometimes new people see ‘agreeing with an ID’ as a polite way of acknowledging help and saying thank you (equivalent to a FB Like).


When I’m not confident I’ll often just ID to genus so it doesn’t go to Research Grade. Because of that I sometimes get more confident IDers identifying it further and saying something like “this is the only species in that genus in the region…”, but I’d rather that than inaccurate RG observations.



I learned that there are actually two kinds of dandelions here: the common dandelion (which is apparently more common, appropriately); and the red-seeded dandelion. But it isn’t easy to tell them apart, so I identify almost everything as genus dandelion.

I’ve seen a couple of instances where people who obviously knew their stuff debated in IDs and comments over slight details. In one somebody pointed out that one species was well known in the area and so people tended to ID things as such, but that another species in the same genus is possible, though rare in that area, and defaulting to the most common species might actually be skewing the data.

So many reasons to be careful, but I try not to let it scare me off. I’m learning as I go.


You may be interested in this thread. :)


I study Carex, a large and often confusing genus of sedges. My approach is simple. If I’m 99% confident of the ID, I provide it. If not, I leave a comment rather than an ID such as, “This is probably C. complanata or C. hirsutella (or perhaps something closely related).” The parenthetical statement may not be needed in many cases if I can narrow it down to 2-3 species. With sedges, it often takes a specimen in hand to make a good ID so I have no problems just sending people in the right direction or to the right group. Sometimes, I can easily provide an ID in my home range (SE USA), but often even there I just leave comments and let others sort it out.


As someone who benefits from suggestions, I appreciate this sort of approach. As I learn more about the organisms that fascinate me, I realize some are fairly easy to identify to species - i.e. Northern Cardinal. Others are less so: spiders, midges, etc.

So, for my own purposes, I know that I can work hard at determining which kind of sparrow that is and I can get better at that. But I also know (at my level of interest and expertise), knowing that is a ‘non-biting midge’ or a ‘type of wolf spider’ is about as good as it’s going to get for my observations.

I love sedges! They seem to be blooming right now. Thanks for being a sedgehead! My sedge observations may not make Research Grade but I am still personally learning about them when you (and/or others) take the time to make comments and share your knowledge. Thanks for that. :-)


Definitely this! At higher taxonomic orders there is no risk of the tentative ID you give being falsely accepted, and being confirmed as research grade, and indeed identifying things to class, order, even family, can lead to the right people looking at the observation.


I can add a reassurance about big-looking numbers: Many id’ers start in their local area, so to them a pile of just Papilionoidea out of Lepidoptera may still be a useful narrowing down, because they will just see their local sliver of the 138K. So, any extra distinction can help, even a still pretty coarse one.

(Relatedly I happened to send a set of years-old Lep > Pap nudges recently and saw it made a difference in some getting to species after all this time- amzamz I think you helped on a number of those, thanks! :) )


Yep, this is a good approach that I also use a lot. With my own observations of insects, a lot of the time I know the species because I collected it afterward and identified it based on characters that you can’t see in the photo. Hopefully that doesn’t lead people to identify other things based on just looking like them! Sometimes when identifying others’ you can see a particular small characteristic that allows species level ID for certain, and I’ll mention what it is. And also if it’s not visible, so people know what to look for next time.