ID of photo vs ID of specimen after collection

I am new to the forums as well as iNaturalist. I searched the forums for similar or comparable topic and I couldn’t find any, so I’m asking.
I have various pictures of insects on plants, either pollinating, eating, mating, etc. At times is hard to determine the species from the picture. In many cases I have collected the specimens and was able to get the species id, or an expert in the group later identified the specimen. Those specimens are deposited in the collection at Saint Louis University, where I am curator.
Is it appropriate to add the species name even though from the picture is hard, if not impossible to determine?

yes. do it. you can add photos of the collected specimen to your existing photos of the specimens in the field, as long as you know the collected specimen is the same individual as the one you photographed in the field.


Thanks! Keeping track should be easy given the catalog number.

Certainly, though if you are adding your ID based on the authority of someone else, please make sure you write who identified it. If the identifier has an iNat account, their account should make the ID and you should only confirm if you personally have the knowledge to do so.
As @pisum hinted at, best practice is to add additional photographs if you can get them and have permission to post them. Keep the metadata of original sighting/collection.
You probably want to denote the catalog number either in the description, or as an Observation Field.

Welcome to the Forum!



1 Like

One potential issue with adding observations from individuals that have also been accessioned into collections is if the collection exports their observations info to GBIF (not sure if St. Louis U does or not). If so, these observations would appear twice in GBIF (once from the collection, once from iNat, assuming they reach RG grade), which could be a problem for some scientific analyses. If it’s just a few observations, this probably isn’t a huge issue, but if it’s a lot, it might be wise to avoid.

You can also cite any specific characters that were used to make the ID that aren’t visible/discernable in the photo which may help others ID.

1 Like

It’s also very useful to add the observation fields “Collection institution” ( and “Voucher number(s)” (

If your photos include interactions with other species, it’s also helpful to add our “interaction->” observation fields (✓&q=interaction->&commit=Search).

(As someone who uses GBIF data, I’d much rather see duplicates than have less data. Duplicate observations of the same taxa with the same date and time and location by the same observer are easy enough to filter out.)


Afraid that ship has sailed. GBIF has multiple instances of the same specimen as they consolidate from numerous sources. So, one of my moth specimens appears from the iNat, BugGuide, Lepidopterists’ Society Season Summary and BOLD (barcoded) contributions to GBIF.

I work on mapping data points for the North American Moth Photographers’ Group which also consolidates records from various sources. Sorting out multiple “copies” of the same specimen to make one record can be an aggravation but it’s always preferable to no record.


Agreed that multiple is preferable to no record.

However, I don’t see that that’s any reason to try to not duplicate data if you’re aware that it’s going to happen and it’s easy to avoid.

Jon, that’s great advice. Thanks a lot. Hopefully, we should be able to keep duplicates to a minimum.

It’s always fine to add an ID that requires information that you observed but is not captured in the photo, but it likely means it won’t end up being confirmed by another user.

You can also take photos after you preserve the specimen to corroborate your ID (also can have a side benefit of teaching the computer vision what a museum specimen of the organism looks like).

Something else I’ve seen more and more folks doing is putting the observation ID from iNat on their herbarium label and the catalog number in their description on iNat.


I do this a lot, but would suggest adding a comment to that effect (something like “specimen was collected after photographing and identified with a microscope”) in order to be clear that you identified it using characters not visible in the photo. This will (hopefully) prevent people from downgrading the identification based on insufficient characters being visible.

1 Like

Nearly uniformly if you’re using GBIF data you’ll remove colocated points (and do a variety of other data cleaning) so it’s probably a low risk error. And since GBIF images (as far as I know) don’t feed into the iNat computer vision training data, it might be worthwhile in rarely collected or rarely identified taxa


For most straight up spatial analyses, this is likely true. And data cleaning will definitely help. However, some data cleaning might not catch duplicate observations. For instance, if entry/observation time or locations end up being slightly different, GPS pts end up being to different decimal places, etc., some data cleaning processes might still treat them as independent points. And there are other analyses like time series/phenology where you would not want to remove observations with the same locality data.

I doubt this is an issue if done on a small scale basis, but for lots of observations double-posted to different databases it could be.

It’s easy enough for iNat observations that you know are already added to GBIF via another route/database to change the individual observation license to “No license (all rights reserved)” which will leave it on iNat but not go out to GBIF.