Observing plants without reproductive structures...

If you want to make the world a better place, take a look at yourself and make a change.

😂😂😂
Have fun in Saudi Arabia! Were you practicing in the Davis Mtns?

I have found it useful in my own identification to find other people's observations on inat where the plant isn't in bloom! So my weightless opinion is that if you feel confident in the ID go for it!

Thanks for sharing your thoughts, Sam. I appreciate your candor and willingness to change your mind-- two hallmarks of good science.

I've always believed in, and mostly practice, the hoover approach to documenting biodiversity. Suck up everything and sort the wheat from the chaff later. Most herbaria frown on sterile specimens (those without reproductive parts). Where I worked for thirty years, these were routinely rejected. I won't say where, but I've seen specimens collected by famous botanists in remote locations tossed into the trash because they were sterile. I was shocked and appalled. But then I wasn't the curator responsible for managing a limited budget, staff and space. The discipline to focus on fertile material in the field was productive in the end.

But it's different with iNaturalist, The administrators might take issue, but basically, the constraints of budget, space and staff just aren't there-- at least to a much lesser extent. That leaves the scientific value of the observation as the sole criterion. So what if an observation can't be identified with a high degree of confidence? Even at the genus level, it conveys taxonomic, ecological, temporal and spatial information at a very minimal cost.

In my work on Smartweeds (Persicaria), particularly the Water Smartweeds (Persicaria amphibia and Persicaria coccinea), I'm grateful for every observation. Obviously some are more informative than others, but every one conveys useful information. I've become familiar enough with them that I can tell them apart with confidence even even from sub-optimal evidence. In fact, these sterile observations have helped train me to see more and learn the plants better. And in the process I'm helping train the AI (and real people) to see more and make better predictions. Sterile specimens are particularly useful for this as some plant characters are only evident in the vegetative state.

Are we training the AI to replace us? Well.... that's a topic for another discussion.

Thanks for the opportunity to share my thoughts.

Best wishes,

Daniel

Sam, I remember those early iNat gatherings when I would take you a plant specimen to identify and you would look at it, then look at me, sigh and say "Linda Jo, where are the flowers?
Over the years I have learned that the ability to identify rosettes is very valuable, especially back when I was a vegetable gardener. In fact, I had a Texas A&M professor tell me that a person who is able to ID plant rosettes is a dang good botanist.

Foolish consistency is the hobgoblin of little minds. -Emerson

Thanks for all you do for the rest of us and for being big-minded enough to change.

Saudi Arabia! Safe travels! Excited to see what you find!
I've found that my ID's are stronger when I've observed a species with reproductive parts and without. Everything is always changing!

Sam, this change sounds totally reasonable. I hope you have a great trip to Saudi Arabia and get lots of new-to-you species documented!

I'm going to weigh in heavily on the side of documenting plants without reproductive structures, which I do often and without apology.

First, iNat is not, was not designed to be, and never will be a herbarium or a natural history museum. As such, the rules for what makes an acceptable "specimen" differ greatly. Herbaria (which I will use from now on as a shorthand for "herbaria and other physical natural history collections") can only document a particular occurrence of an organism at one point in time and space. iNat, on the other hand, can document multiple occurrences of the same organism through time (and space for animals). As such, iNat can document when the first bluebonnets emerge in spring in one's favorite "patch, when that first individual sets its first bud, when it first flowers, and when it first fruits, as well as when the last buds, flowers, and fruits emerge each season, which a herbarium collection cannot. Thus, iNat can and does build a useful phenology of plant reproductive activity (and non-activity) that can be tracked from year to year. Ignoring non-reproductive aspects of plant phenology ignores a large and active part of their life cycle.

Secondly, many plants, particularly trees and shrubs, have inconspicuous reproductive structures and are eadily identified on vegetative characters alone. Maples, for instance, are readily identified by their leaves, but only a specialist could identify them from the flowers or samaras photographed without the clues provided by the leaves and geography. Certainly, many groups cannot be identified to species without flowers or fruits, but we must get over the need to identify everything to species and learn to be content to leave things at a generic level.

I struggled long and hard to learn botany, to no avail, and to the frustration of any number of mentors. It was iNat that has really helped me to learn such botany as I have, and much of that has come through the posting of photos of plants without reproductive structures.

If the argument about reproductive structures is to be extrapolated (ad absurdum) to the animal world, we would be discouraged from posting observations of fall warblers or wintering warblers, since they are not in "reproductive condition," yet such observations are important for any number of reasons, including phenology and distribution. Likewise, we would not be encouraged to post pictures of caterpillars, as they lack "reproductive structures," yet they are often as recognizable, if not more so, than their respective imagines, and often are longer-lived. Yet, iNat allows for such posting and even has an attribute for "larva."

Finally, it is dangerous in the extreme to equate iNaturalist with herbaria and other physical natural history collections. No photo will ever be good enough to get DNA from. No "holotype" left to run wild for later "collection" can ever be consulted again. No photograph can ever substitute for physical specimens in stable isotope or environmental contamination studies. The discussions that often accompany the "simplest" identifications in iNat demonstrate just how difficult is can be to identify even the most common of organisms from photographs. Physical collections and responsible collecting remain as vital as they have ever been. To suggest that iNat can substitute in any way for these collections is the sort of thing that narrow-minded administrators jump on to justify the closing and disposition of these valuable repositories of our natural heritage as a "cost saving" measure.

I've been turning this post around in my head for a few days, since coming across it.

First, Sam, your engagement with this platform is always appreciated, and I can personally say that your comments encouraged me to keep observing and enjoying my time out and about! I like reading through your thought process and evolution on this. It has made me think through my own thoughts, which involve the accuracy of identifications.

--

Tonight, I'm working through a set of BRIT herbarium samples, georeferencing specimens collected in 1949, looking at the lovely reproductive structure of Physostegia pulchella collected roadside in a bar ditch near Royse City. This species has fewer than 600 observations today on iNat, and roughly a third are research grade. (I'll swing back to this a bit.) It's not that commonly observed.

The specimen I'm looking was collected by V.L. Cory (1949) and was later reviewed by three trusted identifiers (Shinners, 1951; Mahler; 1978; Cantino; 1980). All identifiers agree on the genus. But there are three different species-level ID, with the last identifiers settling on P. pulchella. This isn't a case of synonyms; it's three different IDs provided by four identifiers.

Was the final ID based on a species that was undescribed when the specimen was collected? Possibly. Did the literature improve over time? Is the final ID provided correct?

... and are the iNat IDs correct?

--

When I first started using iNat, it served both as a learning and engagement tool. I would have never gone down this path without it. It was actually the first app I ever installed on my phone, and it's the only social media I use of any kind. (You'd never know I work as a web developer in my off time.) But, I was planning on going on a group walk at Bison Hill and I didn't want to be the only person who didn't know how to use iNaturalist.

Thus, the start of my amateur botany hobby. I've never looked back once I started using iNat. Within a week I was pretty much documenting everything I could... here I am today. I've located three previously undocumented prairie remnants; I can identify hundreds of hyper-local limestone remnant species; I am hopefully about to begin serving as a volunteer land manager for one of those remnants.

So!

I see such good in the way we use iNaturalist because of how it's impacted me, regardless of the imperfections of the data we collect.

I see such good in the way others benefit from using iNaturalist as well, imperfections and all.

--

These days iNat serves more as a documentation tool for me. I like to document plants throughout the year precisely for the reasons others have said - it's helpful to know how plants look when they aren't flowering. When do they sprout? What do they look like after the first frost? I focus on items I can ID with high confidence when I do this. I also occasionally document cultivated species for this reason, so long as species can also be found in remnants within a mile from my home.

--

And... now going back to my 1949 herbarium specimen, with all its beauty and reproductive structure on display.

Spending time with herbarium specimens has made me less and less confident about IDs (which is not a bad thing). I assume a 3% inaccuracy rate across the board on iNat for plants. The computer vision has gotten increasingly better over the 21 months I've been here.

And yet...

As I study the past, the more I see that species are missing from our documentation on iNat.

Today, there are people on iNat who painstakingly work through thousands of species to get them right for accuracy. They add missing species. They make journals to describe them.

But I wonder... how much are we missing?

How are we doing on with Symphotrichum, for example? Why has nobody documented Symphotrichum eulae?

Why do I have the only observation of (non-native) Erysimum repandum in the county?

What if 3% is too low, what is the inaccuracies are higher?

...

I don't know how to square this.

But I have to balance the imperfections of data we collect and how we collect it with the net positive. iNat changes lives; changed lives advocate for the future.

So, here is to loosening up.