#IdentiFriday is the happiest day of the week

General
1741

This really is one of the most rewarding parts of identifying, isn’t it? The satisfaction of having figured out a tricky ID, the successful detective work (or sometimes just recognizing that it was a rare observation and getting it to the attention of relevant experts).

I thought of you this summer when I came across a couple of observations of Ripiphorus subdipterus that had ended up in Hymenoptera. I started by looking at sawflies, but something about these insects just didn’t fit, so I did a bit more hunting, followed up on an intuitive lead – and there it was.

3 Likes
1742

I have been hit by that nasty little gremlin, which scrolls the list up, just enough pixels that I hit - what? - I would NEVER have chosen that! Never even heard of it.
Click F to prompt fungus or fern, and they are in gremlin reach.
I wonder why the list scrolls, just a tiny but significant bit?

2 Likes
1743

In my case, the list scrolls because my internet connection is slow, such that it takes longer to finish showing all of the map than it does for me to type in an ID. That causes me to choose the wrong taxon to click or, sometimes, to miss hitting the Agree button when I’m aiming for it. It’s quite annoying.

2 Likes
1744

The other niggle around slow internet is when you see an ID, withdrawn, then added again. Because it didn’t take while they waited, so they tried again - now iNat shows it, twice. Little sigh.

3 Likes
1745

Yes, very annoying.

1 Like
1746

I have been back at my laptop a few days after only IDing on my phone for a long time… it is interesting how different the IDing experience is.

On my phone I prefer to go to needs ID batches that will either require me to suggest a finer ID or to agree with an excisting ID at species level. The identify pages does not work well on my phone, so I either can do the agree on the overview page (need to be pretty charismatic species then that I can agree on in this small view) or I need to open the observations in new tabs. I usually do the latter, skimming to through order or family level for the observations I can push further until I have about 30-40 tabs open and will then go through them to suggest my ID.
I haaaaate doing e.g. genus level agreements on my phone. It quickly gets super boring to open all these observations, agree, scroll down to the DQA to mark “good as it can be” and go on… not doing this.

On my laptop this is much easier and faster. So I went to one of my long abandoned pet projects - pisaurids of Europe - and ran into 99 new pages of needs ID :tired_face: … I like doing those on the laptop, as they are super quick and will help push many observations out of the ID chain (did 22 pages = ~650 obs in just a bit over an hour of baby sleep yesterday while listening to the news ;-))

Just thought it was interesting how different my ID behaviour is on different devices

5 Likes
1747

Absolutely! Figuring out an ID of something previously lacking on iNat is really satisfying. Even as a so-called expert that never gets old.

1748

Sometimes I just get tired of looking at plants. Been working on inverts in the Pre-Maverick project this weekend. I got started doing isopod vs myriapod vs beetle larvae on Friday and continued Saturday. Today became mostly slug/worm vs caterpillar/other larvae. I really should go to bed now. :sweat_smile:

6 Likes
1749

In case you haven’t seen it, iNat just released an “anomaly detector” that uses the Geomodel to show you potentially erroneous or novel observations. See https://www.inaturalist.org/blog/99727-using-the-geomodel-to-highlight-unusual-observations

13 Likes
1750

@tiwane, let me congratulate the iNat staff on yet another really fantastic tool. Only about 12,000 such anomalous observations to sort through in my region - I guess I better get started.

5 Likes
1751

I’ve been working through these unknowns. May I suggest that the next time we work on a big bioblitz or school project we set up a subproject or a code to paste into the observation meaning, “I think this could be identified, if the right person came along”? Something to put on plants that are in focus and have the flowers and/or fruits as well as leaves? (Or sometimes just really distinctive leaves.) Something to direct relevant identifiers to the observations that are most likely to “pay off” in terms of identification? In among the thousands of poorly focused photos or nice photos of utterly unidentifiable entire leaves (and the soft-focus Acacia leaves) there are some really nice observations.

4 Likes
1752

Day by day, what is still waiting for an ID, gets blurrier and
whimpers off into a corner, harder to get any further than … dicot. (I hate to do that, but it is a leetle bit better than plant)
The ones with field marks to work with, are a reward for plodding on.

2 Likes
1753

This is a game-changer in motivating and focusing identification effort on a fraction of the overall observation pool. Kudos to the iNat team behind it, again and again and again.

2 Likes
1754

Today I’m helping out with correcting the many RG obs of Potentilla indica in east Texas and Louisiana which are actually Potentilla hebiichigo, a species that I only learned of last week.

8 Likes
1755

Which I sincerely appreciate and applaud!
If anyone would be interested in learning the difference: Distinguishing Potentilla Indica vs P hebiichigo · iNaturalist is the information I have compiled from research papers.
I’m going through the US “P. indica” ascendingly and so far am up to 01/01/2019!

6 Likes
1756

This is a case where adding pictures would de-muddy the waters.

1757

I have just about got to the monocot/dicot stage for the blurry ones. But it is frustrating as Sedgequeen said above, there are some great observations of plants that someone will be able to recognise straight away but I have no idea so just have to pass over them.

2 Likes
1758

For the good pictures, with clear field marks - I am inclined to cheat - and use a ‘possible’ CV suggestion. As in - a wrong ID gets attention … then I can withdraw. Or perhaps I learn something new.
Trying this ID for this Unknown.

1 Like
1759

I was inspired by some of the comments on the iNat blog post on their new Anomaly Detector to work up some numbers on the percent of Needs ID plant observations currently above Family level vs. Family level and below.

These numbers are for Verifiable plant observations in New England as of 5 October 2024. Currently, there are 2,779,871 such observations. Of those, 1,624,603 are Research Grade (58%) and 1,155,279 are Needs ID (42%).

Of the Needs ID observations, 111,515 (10%) are at Epifamily level and above; 1,043,756 are at Family level and below (90%).

Having 10% of the Needs ID plant observations in the Kingdom-to-Epifamily range seems quite reasonable to me, (especially given the number of student observations I’ve been forced to leave at Vascular Plants in the past couple of weeks!). But New England has a very active and competent community of plant identifiers, so maybe the region is anomalous compared to the rest of the world.

So, I ran the numbers for ALL plant observations worldwide. That’s 85,890,668 Verifiable observations - 51,592,544 Research Grade (60%) and 34,298,135 Needs ID (40%).

Of the Needs ID observations, 3,484,198 (10%) are currently at Kingdom through Epifamily. 30,813944 (90%) are currently at the Family level and below. Again, that ratio seems quite reasonable to me.

But, of course, the real question is not where the already identified-to-some-point plant observations lie in the classification; it’s what to do with Unknown observations. Unknown observations are a small percentage of all Needs ID observations, but still important, to some degree. I don’t know of a way to estimate what percent of Unknown observations are of plants - @pisum, @jeanphilippeb, can either of you help with this?

My guess is that around 60% of Unknowns are plants, but I could be wildly off. For that matter, even if ALL Unknowns that are plants get identified to somewhere in the Kingdom-to-Epifamily range, would it really make that much of a difference? Currently, there are 684,574 Verifiable Unknowns world-wide; if all of those are plants and all get an initial ID of Epifamily or above, that changes the numbers to 4,168,772 out of 3,4982,709, or 12%.

Assuming I’ve done the numbers correctly, 10% or 12% hardly seems worth all the fuss. Or am I missing something?

1 Like
1760

across all observations, a little over 40% are of plants. is there any reason to believe there would be relatively that many more plants in the unknown pool than across all observations?

the only case that i can think if where people are purposely leaving plants at unknown is in Africa. so i guess whatever that effect is would probably be the main thing that could bump up unknown plants. (i’m not going to estimate that effect though because the thought of people purposely leaving things at unknown when they do know makes me shake my head.)

1 Like