Several (well, thousands of) insect species induce galls in plants. For many, this is where they reside for most of their life cycle, only emerging briefly to mate and lay eggs.
Should the main photo associated with these insect taxa be:
A photo of the insect itself?
pro: emphasizes that the taxon is, in fact, referring to the insect, not the host plant
con: often not the form of evidence most frequently encountered, especially when seeing the free-flying insect itself is a rare occurrence.
or
A photo of its characteristic gall?
pros: often the form of evidence of the insect most frequently encountered by iNaturalists
cons: seems to confuse many casual users, who click on it thinking it is the scientific name of the plant (which they may or may not realize is galled)
Example: Rhopalomyia solidaginis is a fly that makes galls in the apical buds of goldenrods, common late-summer wildflowers in my region. These galls look like tufts of many narrow leaves at the top of the stalk. Maybe due to overzealous identifications in the past, the CV now aggressively suggests “Rhopalomyia solidaginis” for photos of a great diversity of leaf-clusters on diverse plant species. Casual users of the site frequently select this suggestion, even for un-galled plants, seemingly unaware that they are selecting a species of fly, not a species of plant. I think the photo might be partly to blame, which shows the characteristic gall formed by this fly. The actual adult flies themselves are rarely encountered. You basically have to rear them yourself if you want to photograph them.
I think it’s going to vary a lot by taxon and gall type.
re: Your example
Ordinarily, I’d say that since the suggestion’s thumbnail in the dropdown list is only one pic (unless they view it or use the Identron), to use the gall…after all, if you only have 1 pic, make it the one that matches what people will likely see.
I see your point about it looking like a plant, though, so maybe swapping the first and second photos would help…the one that’s currently second shows there’s something at the center of the plant tissue (helpful for the suggestion’s single thumbnail), while the other shows the gall itself.
As someone who does a lot of IDing and looking-for-galls, I strongly favor making the representative photos ones that will actually help me to ID what I’m most likely to see–and it is way more likely that I will find oak galls than oak gall wasps, and more likely that I will actually be able to identify an oak gall from a photo than a gall wasp. I would be encumbered and frustrated if a policy of “make all the main gall-inducing insect photos into adult forms when possible” because it would greatly reduce the utility of the site for me when I’m trying to identify my and others’ galls.
When I’m adjusting the photos on a taxa, I like to add all life stages/forms when possible, but research grade photos for use are not always available for adult forms of gall-inducing arthropods. Also, sometimes the gall-inducing organism is something like a virus, which is even less likely to be useful to have useful photos available
The AI being overaggressive with suggestions is definitely annoying sometimes, and I understand why this would be confusing for someone who doesn’t realize what they’re looking at when they see a scientific name thrown and a slightly aberrant plant photo thrown at them. I think changing photos to ones that still show the gall clearly but are harder to confuse with the flower itself, for example, might be worthwhile (so, in the Goldenrod Bunch gall Midge, making the 4th photo in the lineup the first one instead).
my take is that the main photo should be whatever is the most charismatic thing. for example, a lot of oak galls are really interesting and beautiful (ex. Andricus coronus), but the wasps that cause them are somewhat nondescript. so the gall should be featured in the main photo. on the other hand, some, say, leaf mining flies are colorful and interesting (ex. Liriomyza commelinae), while their mines arre rather nondescript. so even if the adults are less often noticed by humans, i think a case can be made that the adult fly is more interesting than the mine. so the adult fly should get the main photo, if a good photo exists.
(similarly, i’m much more likely to encounter and make an observation of scat, tracks, and other coyote signs than an actual live coyote. but i don’t think there’s a question that a photo of an actual coyote deserves to be the default photo because the actual animal is more charismatic than its scat.)
It is useful to sort by life stage which can be done in the picture section. It would also be useful to be able to search by host plant. Maybe this is possible but I do not know how.
I’m just some guy who’s bored during quarantine, with no scientific background whatsoever nor any hobbies or particular interests in plants/insects/whatever, so take that into consideration, but here’s my thought about it:
Before a few weeks ago I had never heard of a gall. I found what was clearly a plant, a green ball on the ground, and the Seek app IDd it as some kind of “gall wasp”. What is going on? This is clearly some kind of plant, not a wasp. Uhh, I think. Do I need to be scared that this “wasp” is going to sting me? Is this just a plant with a funny name the way an ant lion isn’t anything like a lion? Is the image recognition AI just confused? I had no idea. I had to sit down and read a bit when I got home that evening to learn what I was looking at.
Now I know what galls are and have seen a few more since then. It all makes sense now. But I think it would be helpful to have the main photo be of the organism causing the gall. I think it’s fairly common knowledge that a bug may look drastically different in its larva, nymph, egg, and adult forms. If I see some kind of carrion beetle larva that my phone tells me is a beetle even though it looks nothing like a beetle, it’s fairly intuitive that it’s a beetle in a different stage of life. I’d like to see a photo of the adult when I look it up in Seek or iNaturalist. For me, that’s what would be the most intuitive.
I am afraid that “charismatic” will not help in IDing. Main photo of an organism should be a quality photo of the character by which the organism is most IDed. If the photo is charismatic, too, that is only an additional bonus. For gall-makers and miners the keys are focused on galls and mines they make. Keys do not even mention adult phase because it is useless for identification. Main photo has to have primary aim - help for ID and only secondary aim of being aesthetic and admirable. I’ve seen many highly admirable photos (main ones, too), which are useless for ID and even misleading, especially for the less experienced users.
That’s an interesting point. Again, I think it’s highly dependent on taxon, and what is going to result in the fewest number of wrong selections from suggestions.
To your point, if someone were to post an observation of a dog’s pawprint, and the suggestions all had various carnivoran footprints (dogs, bobcats, foxes, coyotes, cougars, wolves, etc) as the default pic, it’s very likely they’d pick the wrong one. In that case, the charismatic pic helps reduce wrong selections. Ditto for a leafminer’s eating pattern, if it isn’t distinctive…seeing a bunch of moths, flies, etc in that thumbnail at least gives them pause to know that little winding path could be caused by several things.
In OP’s case, the midge may be the more charismatic photo, but it’s far less likely to be observed than the gall. The CV’s still going to be looking at leafy crowns (because that’s the majority of observations submitted) regardless of which thumnail is in place. OP wonders if putting the insect as the default photo may reduce incorrect gall selections on plant observations.
I see OP’s point, but it may also increase the amount of mis-IDed actual gall observations, because someone submitting a photo of an what they think is just a very leafy goldenrod may not select the suggestion with an insect pic in the thumbnail (“it looks nothing like my photo!”)
Another solution might be if someone makes an observation with a collage.
Eg. This used to be the default taxon photo for Belonocnema treatae:
I can see why it was changed in the case of B. treatae; the galls are more likely to be encountered than the wasp, and while it looks a bit like some like other galls, it doesn’t look like a plant alone, unlike your case with Rhopalomyia solidaginis.
I agree with @ekmes that with galls and leafmines, it is usually most helpful for the first icon photo to show the gall or the leaf mine, and the second (or third) photo to show the adult.
If the common name of the organism includes the word “gall” or “leaf mine”, then is it not a problem if the icon image looks like a plant. However, many galls and leafminers don’t have a common name, so yes, some people will be confused, but we can’t make everything work for everybody.
I also wanted to add here that the first icon pic (of anything) needs to work really well at tiny thumbnail size, so it needs to be simple, and very clear even when tiny.
I find that some people select for the default icon pic the most impressive image of a species, like a mating pair or a close-up of the head of an insect, but it is better to leave that further down the series of icon images.
People use the icon pic as a reminder, or a first idea of what the species looks like. If they want to know more about how it looks, or how it can look, they will learn to go to the taxon page and see a good variety of images.
I don’t think it is reasonable to compare a coyote to a leafminer in terms of what to use as the icon pic; different guidelines apply there for sure.
Well, if they think it’s just a very leafy goldenrod (and in a very real sense, it is!) then they’d be correct to pass over the insect ID and choose the plant ID instead. The observation is also evidence of the plant, after all.
My first question to users making puzzling IDs of this and similar taxa is often: “Are you meaning to indicate the plant or an organism deforming the plant?” The answers are about 50/50.
It looks like these helpful guidelines from @ekmes have a lot of support from the community.
For anyone reading along, I have just changed the photo gallery for the example species in my original post (R. solidaginis) to show a clearer photo of the characteristic gall as the main photo, keeping in mind @susanhewitt 's image choice suggestions. I also selected other photos for its gallery that show the insect itself at various life stages, as well as a representative photo of an infrequently encountered gall type (this species makes two types of galls at different times of the year).
Hopefully, at least some of the users who are confused by an insect taxon being associated with an image of a plant organ will click on the species page and get a quick idea of what’s going on from the image gallery alone.
I also agree with @ekmes and @susanhewitt on this one! I recently got more interested in galls, and by happy circumstance, I’ve recently photographed galls of several species that have very few to no observations on iNat (aka too few observations to be suggested by the computer vision). But since I’m not a gall expert by any means, ID’ing these rare species is tricky for an amateur like myself, since the photos available often aren’t the clearest or most reliable. With the help of a couple iNatters with way more gall experience than me (namely @Megachile, @calconey, and @mileszhang), I’ve gotten better at recognizing some of the local species, and as a result, I’ve tried to help improve the default photos for those species that I recognize. I always prefer to make the first photo a high-quality close-up of the gall, both since it’s the most recognizable stage to find and because this way there’s no mistaking the gall for the host plant. Then, the second photo is preferably a picture of the adult (if available, the VAST majority of galls don’t have pics of the adults on iNat), and then after that, pictures of any particular unique features that make ID’ing easier (such as texture, inside architecture in a cross-section, the shape of the exit, characteristic morphology changes in the host plant, etc). It helps tremendously even just adding more photos so that a species has 4 or 5 main photos showing the phenotypic variety instead of just one picture that may not be representative of what you find in the field!
Side note, I use basically these same principles when helping to clean up the photos for plant species as well: The first photo should be a clear close-up of the most recognizable feature (whether that’s the flower, fruit, leaf shape, etc), and then the next 4 photos should be pics of key identifying features (e.g. characteristic bark, buds, leaf shape, leaf scar, pubescence, full-form of the plant, etc). You can add up to 12 photos per taxon, so I will normally add as many additional photos as I can to include all the important key characteristics and/or show the phenotypic variety if there’s not that many important/consistent features. Highlighting important features in the first five taxon photos is SUPER helpful even just to remind myself of key differences between very similar species (i.e. Elaeagnus umbellata vs Elaeagnus pungens). I know not everyone will take the time to look past the first photo, but I hope that ID’ing becomes a little easier for the dedicated learners by making the main photos more practical/informative for ID’s!
I think non-insect plant (or animal) diseases provide a useful point of reference here. We’re never going to have eg microscope slide photos of leaf spot fungi or mosaic viruses or bacteria. iNat taxon photos are not a representation of the essence of a species but tools to help users recognize it. I agree with what others have already said: the main photo should show the most common, apparent, charismatic view of the organism as it’s encountered by naturalists (ie the gall), and the other photos can be used to add a broader perspective on its life cycle/biology.
