Do iNat users who are able to identify Chironomidae to genus/species level actually need genital photos on iNat? Don’t these expert users rather use scientific papers with figures?
2 Likes
It’s often more convenient and easier to compare two photos of live specimens (one as a voucher) than it is to compare a line drawing with said fly. Scientific figures may be misleading based on the angle the drawing was made from, which cannot be inferred as well as images can. Not a huge problem, but it does save on clicks.
I think this is only a small piece of the discussion, though.
2 Likes
I understand the desire for an easy reference for identification, I really do, but the very first photo really needs to be something that an average person can get some meaning out of, surely?
That doesn’t mean it has to show the entire organism, but it is possible to show enough to give identifiers a clue without leaving everyone else completely clueless?
(I realize “identifiers” and “everyone else” are not two totally separate categories in all cases, but to the average person, identifications based on close-ups of insect genitalia are…pretty niche.)
7 Likes
My opinion was and always has been that the taxon feature image (used as a thumbnail) is for a good representative image of the organism. That is, not specific details or useful traits, but the organism overall, preferably in-situ and alive.
It should not be the job of this feature photo to convey identification features. That’s what the additional photo ‘slots’ on the taxon pages are best for. I think it’s fine if the taxon photos show something of the field marks, but if your taxon photo of an insect is just the male genitalia zoomed in, I think this fails to accomplish the purpose of the taxon image.
It’s not a “how to identify this organism” image, it’s a “this is what it looks like generally”. I can’t imagine some of the birds having feature images of their butt or a pattern in their wing. While those features are crucial, they are again not part of the general representation of the organism.
The reality is many organisms cannot be readily distinguished from an overall view. That’s okay.
11 Likes
I don’t really know much about Chironomids. @zoology123 is the expert here (if that wasn’t already obvious). Now that you’ve said this though, I do need to be entirely honest about something here.
I was really only aware of this situation because zoology123 messaged me because I had changed some taxon photos. I had deleted them because not only did I not recognize them as fly genitalia, I didn’t recognize them as fly body parts at all! I thought they were just more examples of the recent taxon photo vandalism.
This, to me at least, shows why these fundamentally fail as default taxon photos - I, and an experienced identified (though not of Diptera) didn’t even know what I was looking at - so how could anyone less experienced?
6 Likes
One of the main points I argue is that you will not get much meaning at all besides it’s a green fly if you use the adult images for certain taxa, though most have not even been Identified yet on iNat. There is many genera that can have identical looking males and females. This is compounded further that you can easily confuse other genera with green individuals in the Harnischia Group.
“identifications based on close-ups of insect genitalia are…pretty niche.”
This is partly why i question this is being brought up like it’s an issue. I went through and only counted 10 species with genitalia as the default taxon image. I suspect a number of people here have never heard of Chironomids also. That’s not necessarily a bad thing, but more to illustrate how niche this family of flies is. Most of those species also have only a handful of observations.
I do think these images also have real effects. I have to correct less often taxa that uses genitalia as the main picture. While others like Glyptotendipes, Procladius bellus, Goeldichironomus carus, frequently get misidentified. I truly think that using genitalia for the ones that require it helps to reduce the amount of miss identifications and hopefully gives people second thought.
I don’t think currently with the systems in place it is possible to have a perfect solution. You can either have tons of identical images for a number of taxon for the one’s only IDable by genitalia. Or have them be genitalia which shows the differences.
1 Like
I mean, I’m not surprised. This is a niche group. This group is so niche they really only started getting Identified and corrected on iNaturalist about 8 months ago. Before that, these were extremely inaccurately IDed overall across the entire site. Before I started IDing, the CV only knew about 7 taxa. Now it knows 55 taxa of Chironomids.
5 Likes
This is a startling accomplishment given how many of a species need to be identified for each to be recognized within the CV. (In case no Observer has said so, thank you.)
Do any identifications stand out? Maybe you could detail one or more in this thread.
My inclination is to leave taxa photos to the Identification experts within that taxa. (If they choose poorly, they may see more uploads at species they may not have intended.)
PD: I had no idea there was taxon photos vandalism afoot. Thank you to everyone, including the volunteers, Curators and non-Curators, who are helping clean it up.
7 Likes
I am in favor of having a first image of the whole animal, then select diagnostic images of the genitalia or other diagnostic features secondary. I concur with someone above who mentioned that the visible clue that two Chironomids are essentially identical at the macro scale is useful information. It can also be very important and useful to include subsequent detail images, but the genitalia-first image effort is a slippery slope with possible unintended consequences.
In particular, a great many insect groups primarily rely on genitalia to be certain of an ID (I’m thinking of Orthoptera, Hemiptera, Hymenoptera, Diptera, moths, etc., etc.). And in many, many groups that has become the primary (sometimes only) published detail put forth to separate species.
In the moth world where I live and study, a great many small non-descript moths might currently be indistinguishable but previous authors have published species accounts separating them on genitalia–such examples are very numerous. However, I’m finding that in certain cases that I’ve delved into, too little effort was placed in distinguishing adults by size, pattern, form, behavior, etc. There is a need in moth identification for a huge expansion of our knowledge base on identifying groups of similar species based on patterns that previously published studies (e.g. from museum-based taxonomists) have dismissed or ignored. I will never argue that genitalic examination is not important. But my point is that if you moved iNat towards genitalia-first images of all insect taxa which were defined on genitalic examination, the photo galleries of all those groups would become an uninterpretable wall of jibberish to the larger population of iNaturalists.
10 Likes
I have spent multiple hundreds of hours on this taxa group over the span of 8 months. I have made about 31k identifications. Most are not to species level.
What do you mean stand out identifications?
I do not wish to derail this thread but if you read the first post of the link, the OP has detailed several reasons a particular identification might be one for which an Identifier might feel special pride. Yours might be training the CV. (I think you should be proud.)
4 Likes
I see this continuing to be brought up. “The adults should be the taxa image even if they are identical because then you can see how many identical species there are”. Something like that.
Please understand that certain groups of Chrionomid are so difficult to ID. They literally aren’t getting IDed. You can’t have taxon images of taxa with 0 observations. For Chironomus, only 10 species have been observed in the US. But there are atleast like 90 species with 30-40 undescribed. It is not even possible to really have the other taxa show identical adults to even compare. One of the species IDed in the US was from DNA.
For Orthocladiinae, a very large subfamily. There are nearly 200 genera without any observations. Most species do not have images on the internet.
The species of Chironomus getting IDed in the US also are really only the distinct ones. The bland gray identical looking ones aren’t getting IDed.
If Chironomus crassicaudatus does not use a genitalia image. It will look pretty much like 40-60 other Chironomus species in the US that have not been IDed so have no taxon images to compare. If it is the only one of like 3 gray Chironomus species IDed, because the others are on the verge of impossible to ID. People will just choose that species in the CV because it’s gray and is close enough.
3 Likes
I like Chironomids :)
Back in the day I did quite a bit of sorting out the taxonomy because it was a total mess - hopefully it’s considerably better now than it was. I’ve had a go at identifying a few on slides myself, but much more practice required - and when to find the time?
Anyway. My little view on this is that I don’t really think it matters in a group such as this. It’s probably a good thing that there’s a variety of types of taxon picture - some with genitalia first up, some with the whole body. Anybody looking at these is probably doing so for a reason, and will learn more from seeing both types. Neither will cause significant issues for a meaningful number of people. If anything I’d probably prefer to use the genital picture if it’s a high quality one, just because there aren’t so many good pictures like that.
4 Likes
I believe enough people are and the ones who aren’t won’t be stopped by a taxon image in my experience (I’ve seen enough plants with a normal leaf spot identified as some beetle or spider)
Identifying ladybeetles, I’ve seen quite a few observations left at genus or tribe despite the CV suggestion. For Platypezidae, a species level ID is the exception (though I believe most species aren’t yet covered by CV, so I don’t know whether that is much of an argument).
I agree with this. I really appreciate these kinds of comments by IDers and find them more helpful than any one photo. I learned how to take decent (for identification purposes) photos of mushrooms and of rust fungi that way. And with the iNat Metadata Tool by Megachile it is very quick and easy to add IDs with copypasta comments and annotations with just one click or a keyboard shortcut (if you use the identify page).
Personally, I have not noticed this. It is a problem, but apart from some bigger issues with observations uploaded via the seek app and potentially some that will be caused by the CV implementation in iNat Next, I see it being more of a constant thing perhaps slightly getting better because of a more accurate CV.
Wow, that’s quite some work you’ve put into that! And training the CV to that extent is a huge accomplishment!
I started the thread in question as an identifier equivalent to all the “favourite observation/photo/wildlife encounter” topics. If you want to, feel free to share anything you’re proud of, or any ID that you made which you are particularly fond of for any reason. It doesn’t necessarily have to be an individual ID. :)
Very much this. People who are going to just pick whatever is the first CV suggestion are not going to stop doing that just because you pick a different taxon image, so those people really shouldn’t be the ones you pick the image for.
Personally, if I see a list of CV suggestions and from the thumbnails they all look identical to me, that’s a very good clue that I’ve stumbled across something I’m not equipped to be identifying yet. So that image is conveying useful information to me, even if it is not terribly helpful to experts.
Really though I wonder if this is more a philosophical question about what (and who) the taxon images are for, and that might be more of a matter of general policy that should be added to the general guidelines (e.g. https://help.inaturalist.org/en/support/solutions/articles/151000184018-what-guidelines-should-i-follow-when-choosing-taxon-photos-)
5 Likes
I don’t know chironomids, and if I found one labeled “unknown” I’d ID it as Diptera or even Pterygota. I have, however, been identifying macaws to species, especially the blue-and-gold, of which I’ve known one. Ara militaris and A. ambiguus are hard to tell apart (I suspect that’s how A. ambiguus got its name), and the differences boil down to shades of green, size, and something else.
I think that one of the few photos that show up when you hit Compare should be an adult standing next to a ruler, and another should be the back of the bird in the same light as the other species.
1 Like
You can change the taxon photos to show that useful info?
1 from the front, 2 with something for scale, 3 from the back ?
No, I think that someone should take pictures of these macaws with a ruler and in the same light. AFAIK there isn’t already a picture of each with a ruler.
1 Like
I think it would be a cool feature to have a dedicated section for this on the species page for this reason. Perhaps a “Species Journal” or “Species Wiki” specifically for identification.
I don’t think it is feasible to convey all that data in taxon photos for every species.
(I view the chances of me propping up a ruler next to a live bird, setting up studio lights calibrated to an exact temperature, blocking out enough environmental light, and then taking a bunch of photos from every side, all without it flying away as pretty minimal)
In a wiki you could even write “Species x can only be distinguished from species y through a small difference in height and colour”.
It will not do much to prevent misidentifications, but it might significantly increase the number of identifiers. And that in turn world solve the misidentification problem because there are more IDers to check and correct.
TL;DR: I think CV-based misidentifications aren’t actually the problem, so taxon images shouldn’t need to account for them. Having too few IDers is. And solving that requires better infrastructure for IDers (most notably: better ways to find and share identification information and resources). (This feature request covers a lot of this and more in full detail)
1 Like
I hear you, but is it really a problem to have the genitalia then as a second photo.
3 Likes