Get ready for a whole bunch of changes to taxon photos

I realize many people like taxon photos to be unchanging. However, with the implementation of suggestions in the CV using the best matching taxon photo, I predict there will be a wave of changes to taxon photos as people try to tweak the existing set to better
represent a species. In response, those who like the existing set will change them back and we’ll enter a period of back-and-forth struggle over which photos to include.

5 Likes

I do value taxon photo stability, but really only for the top photo since it is used as a visual shortcut by IDers and Observers alike. I generally think the top photo shouldn’t be changed unless there’s a compelling reason to do so, and that top taxon photos shouldn’t be shared by multiple taxa to avoid confusing one taxon for another.

That doesn’t mean that the other taxon photos shouldn’t be changed though! I think changing these to add views of different sexes, life stages, fruits, etc. will be beneficial overall. The new system also means that having the same photo as a taxon photo for multiple taxa will be a bit of a waste as it would reduce the options for the CV to work with. Get as much diversity as we can in these!

30 Likes

I think people will just add more photos to better represent characteristic features or life stages of an organism. The CV will select whichever is closest to the observation being identified.

I don’t delve into the world of taxon photo curation but I don’t see a reason to be upset that there are more photos. I can, however, say that as an end user I have looked at a species’ taxon photos in the past and have been frustrated that while there are many pretty photos and those of aberrant individuals, few or none were useful in demonstrating characteristics key in identifying the species. Not saying what the frequency of that happening is, but if this change encourages more taxon photos that would aid as reference materials for identifying an organism, I am in full support.

12 Likes

In the new system, when someone posts a blurry picture of an iguana, the CV says, oh here’s a blurry taxon picture of an iguana that looks most peculiarly similar. Not the iconic first photo from before, but for whichever matches best. Another example I’ve already noticed, a partially obscured photo in an observation yields the most similarly obscured taxon partial photo. Granted in both cases removing the taxon photos of blurry or partially obscured taxon photos is the best bet, but for as long as they remain in the taxon pool, bad observation evidence will yield the most similarly bad photo of the taxon pool.

2 Likes

I agree, maybe the change will encourage better cleaning of taxon photos of this nature than in the past?

3 Likes

The thing is, currently, the vast majority of taxon photos don’t really work well as taxon photos. Taxon photos need to be clearly identifiable and show well as a tiny thumbnail image. Foe this change, it didn’t really matter what most of the taxon photos were, as long as the first one was good. But now we have to curate all of these taxon images? Honestly I’m not a fan of this update.

4 Likes

One can only hope. I’ve already started that process for some taxa. The real challenge is with those taxa for which there only exist the minimum number of observation to be suggested. Often there are not many good options and so sub-optimal photos are included as examples.

2 Likes

As someone who IDs a lot of caterpillars and frequently has to rearrange the Taxon Photos after someone has decided to put a bunch of pretty (but not diagnostic) adult photos in front of the caterpillar photos, I can promise you, people are already having to curate all these taxon images. At least this way, the caterpillar (or pupa) photos won’t have to be in the first 2 in order to be visible in the Identify preview.

22 Likes

One benefit is that it is bringing a lot of misidentified taxon photos to my attention. Back in the early iNat days, before even BugGuide had good information on identifying Misumenini crab spiders, a bunch of misidentified Flickr photos made it into the taxon photos of many taxa such as Mecaphesa asperata. I don’t think I ever would have noticed that if not for this new change.

11 Likes

I’ve noticed the same thing in the lizards as I search for, say, a genus to represent in a subfamily collection…why is THAT coming up in the search for THIS genus?

1 Like

Earlier, I used to use this project to roll back the photos in some taxon pages:
https://www.inaturalist.org/projects/taxon-page

Then came this new feature, 2 years ago:
Taxon photo changes are now recorded in taxon history
https://forum.inaturalist.org/t/taxon-photo-changes-are-now-recorded-in-taxon-history/39364

But I haven’t rolled back photos for a long time. I gave up.

For genus Senna, in almost all cases, it is useless to have a closeup photo of the flower as the cover photo.

For instance, for Senna occidentalis, this photo was to my opinion the best one for the cover photo as one can instantly identify it as this species:
https://www.inaturalist.org/photos/3030253?size=large
Even reduced to a thumbnail, the photo was still identifiable and harmonious.

Another example is this closeup photo of a Senna phyllodinea phyllode, showing it is curved and with only one edge toothed, which is enough to identify this species without any doubt. I think this photo must remain the cover photo:
https://www.inaturalist.org/photos/71081623?size=large

The observations I selected in the project Taxon Page could be added to a project ID Tips, after adding identification tips in comment.
https://www.inaturalist.org/projects/observations-with-helpful-id-tips-umbrella-project

3 Likes

Have restored your chosen taxon default picture

2 Likes

I think one of the reasons that the main taxon photos get swapped out is that we’d all like to see photos that showcase the whole organism in the wild and are identifiable, and those two things are mutually exclusive for many taxa. For lots of plants, fungi, and insects, a photo showing the whole organism is inherently unidentifiable (because identification requires close-ups of specific features), and photos showing the identifying features are inherently bad at showing what the organism actually looks like (because they’re magnified images of specific anatomical features). Personally, I don’t see the point in having a beautiful hi-res in situ photo of the whole organism as the main taxon photo for 10 similar species if those photos don’t give any indication as to how the identity was reached (cut to people using the “compare” feature and reading all sorts of non-distinguishing features into pictures that show none of the necessary marks to make an ID). But I also see there’s an argument to be made that showing a scoped image of the setae on the underside of a beetle’s fourth abdominal segment as the main taxon photo will leave people confused as to what type of organism they’re even looking at. I like that this update allows both those types of photos to have a chance to show up as a thumbnail, but I suspect the disagreements about what makes a good “cover photo” will continue as they always have. Questions like “which matters more- consistency in showing the same life stage as the top photo for all species in a group, or the ID benefits of showing whichever life stage is most identifiable on a case by case basis?” will continue to cause debate, but maybe the answers to those questions won’t matter as much now.

10 Likes

They are often quite useless for plants. I often need fruit or leaves to verify what I am keying in the local, or regional key, but instead all I get are pretty pretty flowers for days. One should be able to use only a key, sure, but photos are nice. I like the idea of showing the most crucial part of the plant first, even if it is an “ugly” seed or fruit. Also, they are still less trustworthy than Calflora in California, in my opinion. I’ve had to flag one that was not the correct species.

3 Likes

The answer is not to flag anything, but to open the observation the photo comes from and correct its ID, then remove that photo from the taxon photos (I’d love removing an observation from a species to automatically remove it from taxon photos, but it doesn’t work that way).

I’ve found and fixed a number of observations based on seeing incorrect taxon photos. I’ve also found various wrong IDs while selecting new taxon photos. In fact, it was finding misIDs in this way that started me checking through RG observations of species I’m familiar with as quality control.

Bonus: we can help make them more trustworthy!

9 Likes

I did just that today.

3 Likes

I am just a middle school teacher who fell down a rabbit hole doing some research for my class on newly discovered species as an intro to our unit on Classification and Taxonomy. It started with a Smithsonian article about Alto Mayo, Peru and ended up here looking for more details about Daptomys - where I found out that my new favorite mouse, and a squirrel friend Microsciurus ( who just generated their own genus ) have very little info and no picture. So wanted to do my part to help out by adding the links where I started from, and let all of you take over with gathering and posting the info and photos for these two and the other species from that trip ( 27 total ).
https://www.conservation.org/stories/new-discoveries-in-a-changing-landscape-uncovering-alto-mayos-hidden-biodiversity
Expedition Discovers 27 New Species in Peru, From an ‘Exceedingly Rare’ Amphibious Mouse to a Blob-Headed Fish | Smithsonian

The Smithsonian article contained a link to a 47 page report on all these new species. I honestly did not open it ( I should be home making dinner instead of at school researching ) but hopefully it has lots of useful data for you all.

1 Like

Yes, please don’t use flags for issues that non-curators can deal with.

2 Likes

Welcome to the iNat forum Michelle!

Wow, a middle school class covering taxonomy! That sounds quite amazing.

Thanks for the link to the Smithsonian article about the 2022 expedition to the Alto Mayo region. If I understand your comment correctly I think there probably a couple of connected issues…

How do the 27 new species discovered by the expedition get added to iNaturalist’s taxonomy? iNaturalist bases its taxonomy mostly on a series of authoritative “external taxonomic authorities”. These authorities review articles published in the scientific literature formally describing new species (or recategorizing existing ones) and in most cases will add them as new species. If there’s real dispute about the validity of particular species, the authority will decide whether or not to accept it. iNat mostly follows the guidance of these authorities, but curators can deviate where there’s a good reason with community support.

So, when the Daptomys and Microsciurus mentioned in the article are formally described, then they should get added to our taxonomic authority and make it into iNat’s taxonomy.

How do photos get added for new species (or for any species that is missing photos)? Once a species is listed in iNat, people who observe it (or who think another iNat user has observed it) can start using that species name for identifications. Of course, it really helps to have some pictures of the new species to compare with other related species. iNat does some of that automatically, but the concept of “taxon photos” allows us to prioritize up to 12 photos that are good representations of a particular species. A subset of these is then shown in a bunch of contexts when people are considering possible IDs (which is the focus of this thread).

Taxon photos have to be selected from existing iNat observations, or a few sources with free licenses such as Wikimedia Commons, Flickr and the Encyclopedia of Life. All the photos in the Smithsonian article are copyright reserved, but if someone added an observation of the new Daptomys species to iNat we could then use those photos to illustrate the species.

Now that you’ve been mesmerized by the whiskery Daptomys, we’re all looking forward to seeing the photos you’re going to post from your class trip to the Alto Mayo (or to the local park, which is pretty valuable, too).

3 Likes

I welcome disruption to taxon photos. Anytime I look at a species that’s new to me I check the photos and it’s common that the species are incorrect, or using unverifiable flicker pics, or just plain unhelpful.

As mentioned about the sennas, there are so many plants on inat with only a variety of photos of their flowers. Ok, sure, flowers are often important for identification, but most plants only flower part of the year. What about the leaves or fruits or form? Is it a herb? A tree? What about when it’s young? or mature? You get no idea about any of that, just flowers. And for genus+ level taxon is very common to see just a couple of species represented even when it’s a huge varied group.

This seems like a good idea so those who know what to look for and those who don’t can both get suggested taxon photos that are relevant and helpful for them.

4 Likes