marraskuu 1, 2023

The current state of Heterotheca on iNaturalist (and how it got this way)

Edit: Updated considerably on 2024/05/21

I've been working on the genus for my doctorate research for the better part of two years or so at Univ. of Oklahoma, trying to resolve discrepancies between Nesom's and Semple's taxonomy through a mixture of observations (in the field, herbarium, and online mostly through iNat), molecular phylogenetics (DNA work using HybSeq and RAD-seq), hopefully some greenhouse experiments (am currently growing ~20 species this year mostly from seed collected from the last two years of field work), and likely some morphometrics (in a similar way that Semple used when making taxonomic decisions) or biogeographic comparisons (ENM/SDM and e-space tests among potentially other analyses). There is a lot of work still to do, but the greenhouse ('common garden') work is underway, and most of my first chapter on the group's general phylogeny with Hyb-Seq is being processed (lots of paralogs in the thousand genes or so sequenced for many/most samples). I'll talk a bit about the latter here, but keep in mind that not all of the paralogous sequences have been dealt with yet so a few things may be subject to change.

Taking a step back, Heterotheca is an exceptionally complicated group with an enormous amount of variation and intergradation of vegetation and floral features (often revolving around pubescence and glandularity patterns). Hybridization has been extensively speculated - in many instances with good reason either because of clear phenotypic intermediacy or populations' polyploid status (or both). While much of the hybridization talk likely bears merit, I've run across very few instances (outside of H. (stenophylla var.) angustifolia) of things which definitively look like hybrids where both parents are locally present. I think the hybridization matter may have been overblown and used to explain natural variation that makes certain populations look different from what most people think that a given species/variety should look like. Because of these complexes or swarms of morphological intermediacy, things have been listed as comprising complexes (particularly in Semple's work).

Several species complexes for the perennial species (H. fulcrata, H. oregona, H. sessiliflora, H. stenophylla, and H. villosa) do not make any sense at first glances; it has taken me this long to vaguely understand each author's treatment for these complexes while also coming up with what previously unrecognized concepts/taxa might warrant some formal recognition. In some cases, there seem to be geographically-based groups (such as the H. sessiliflora complex mostly occurring in southern California outside of the Mojave Desert and Sierra Nevada and the H. stenophylla complex occuring predominantly in the southern/central Great Plains). The biggest example here is probably the "H. fulcrata complex" if you look at the southwestern representatives (foothills or mountains Chihuahuan Desert, Sonoran Desert, Interior Hills & Plains, or Arizona-New Mexico Mountains) of Semple's four H. fulcrata varieties + H. marginata + H. rutteri + H. viscida (excluding the CO/WY populations of H. fulcrata vars. amplifolia/fulcrata + H. pumila for sake of arguement).

These "H. fulcrata complex" taxa do share some features - namely the presence of subcapitular "bracts" that protect/shield the head during development and perhaps during seed production post-flowering. This group is likely polyphyletic and an artificial grouping, so the presence of these "bracts" likely represents convergent evolution rather than shared homology. The "H. villosa complex" is far more ambiguous, and I suspect that Semple lumped everything else (bar the "H. fulcrata complex", "H. sessiliflora complex", "H. stenophylla complex", or "H. oregona complex" for the perennial and non-heterocarpic species) more so out of wanting to wrap up his work (as his research into Solidago picked up) rather than thinking they represent a coherent lineage. The data within the main group (comprising most of section Phyllotheca) are pretty messy because of rampant paralogy by-and-large, hence why there are a lot of "maybes" to be worked out still.

I've largely sided with Semple's treatment - despite its overly lumping nature - historically because there was at least some data (albeit most unpublished) supposedly backing up a lot of these decisions. Semple's work also seemed to go through more peer-review, and there were a lot more people (mostly undergrads) tending to many of the measurements used in making decisions (I'll note here that I think it was very wrong to never credit the "army of undergrads" more formally in publications given their contributions). Nesom's work I've always been very skeptical of, in large part because the journal he's used in the last decade very extensively - Phytoneuron - is regularly not peer-reviewed, especially not when Nesom is the sole author and editor of the journal (and particularly when peer reviewers, who he selects personally, are not listed in his own work). The journal has its place for some taxonomic purposes, but its lack of modern peer-review standards (especially when he is the sole author) is not really debatable. There aren't rules strictly preventing Nesom from self-publishing whatever he wants in Phytoneuron, but the findings published in such should be viewed with great skepticism. No morphometrics, statistical comparisons, sequence data, or any other form was used in his treatment of Heterotheca. Morphometrics (as Nesom's mentioned to me) need not necessarily be correct in all instances, but I think the disregarding of any data in a group this complicated is a poor judgement call.

In several instances, the significance of these morphological analyses were discarded. His treatments come much more from his personal glances at many specimens (often virtually so that he can't actually see many of the pubescence and glandularity features so critical to understanding many/most of the taxa involved). Not looking into things in this way I consider problematic, just as his maps don't usually have lists of specimens or observations backing up each county listed (which he's noted is "bad practice"). This further makes the maps unreproducible and thus unreliable since it seems counties were listed based perhaps on suspicions of where things went rather than having justifications for each plotted area. Further, Nesom has also criticized people for relying upon molecular phylogenetic data (see his article of Toiyabea and much of Urbatsch's work on Astereae) and does not appear to fundamentally understand how they work (see his concepts of Doellingeria/Eucephalus). Skepticism of trees with few loci is understandable, but the arguments made aren't typically those from someone who really understands the caveats and limits of molecular phylogenetics. He is a very good 19th century botanist as Semple has acknowledged, with a keen eye for small details - but his publications should never been blindly accepted as if they were subject to the same standards as other journals require (either in terms of data used or peer-review).

That said, this is not to say that Semple's treatment should be taken as bible either. I disagree with both's treatments of the H. fulcrata and H. villosa complexes. Semple's conception of the H. stenophylla complex (H. canescens and H. stenophylla var. angustifolia and var. stenophylla) makes a lot more sense to me than Nesom's does, especially after field work this year and last year, but neither appears wholly correct. There is a lot of population-level variation that was probably beyond the scope of Semple's really extensive work which I think is important to address for something so comprehensive. Semple, like Nesom, lacked any robust molecular phylogenetic to inform his decisions better (though he did include a mixed morphology/chloroplast tree in his '96 monograph). He also lacked access to digitized specimens the way that Nesom had, making it more challenging to know what is where (though he did annotate >10,000 specimens by the mid-2000's). The lumping decisions made likely reflect the times in decades past, but I'm growing increasingly convinced that some groups really ought to be split up. This is particularly true for "H. villosa var. scabra"; I think that Nesom's conception is closer to what's actually going on in breaking up H. cinerascens, H. excelsior, H. joshuana, and H. polothrix. I think H. cinerascens is the problem child here because it sort of blends into the others morphologically, but Semple certainly could have broken things up far more.

Semple's lumping likely has had some unforeseen consequences. The really broad treatments of "H. fulcrata", "H. sessiliflora", and "H. villosa" have led to many (or perhaps most) naturalists and botanists calling the plants they see one of those three things, not delving further into subspecies or variety. This - especially for H. villosa - meant that some people call everything "H. villosa", regardless of whether that's the right "group" to place it in. This isn't as bad for the "H. sessiliflora" or "H. stenophylla" complexes, but it's more prevalent in the "H. fulcrata complex" because the two overlap more geographically. While Semple was generous enough to provide a key to the genus, the broad lumping made some species or groups so highly variable that one person wisely told me that "the variation within species is greater than those among species." I didn't fully understand what it meant at the time, but the further I've delved into the genus, the more obvious the flaws in both treatments (especially Semple's in this regard) became more glaring. They're the best we have so far, until I or someone else can make something more coherent, so it's not going to be clear what to call many things because the two treatments are so radically different.

The reason that I've been going along with the names on iNaturalist (which follow Nesom's treatment) is because another user had switched things over and had identified >3,000 observations on here already. The amount of work, as someone new to the genus during spring '2022, seemed pretty insurmountable. It would be going against the tide to follow Semple's treatment at the time (and even more so now) when people are adopting the new treatment, leading to a greater proportion of taxa left at "Heterotheca" (the subgroups of "Phyllotheca", "Heterotheca villosa complex", and so on per my request had not yet been generated) and thus less useful to countless people (myself principally included since I've been going around looking for these taxa/populations). The amount of work to go along with things was substantial, given that virtually everyone at the time followed Semple's treatment (since all of the floras followed it essentially), but trying to halt it and revert things back would be of similar resistance and difficulty (or at least as much as I predicted).

The problem now is that almost no one knows what to call anything anymore since the name changes from the two treatments was colossal in some regions (not so much in states like Illinois, drastically so in places like Colorado). This, in combination with the seemingly endless combination of names used in local floras, has made it really difficult to know what names to apply to what populations. One authority says one name, another refers to said first name as a species or something wholly different, and then various floras use many different combinations and ranks for various names. For instance, Semple calls some plants H. stenophylla var. angustifolia, but Nesom calls them H. angustifolia while the Flora of Colorado refers to them as H. villosa var. angustifolia. What on earth are most people supposed to do with that conflicting information?

The name changes have overall been so complex that it's not just a variety raised to species but now one variety is now like five species all falling under the umbrella of 'Phyllotheca' or their respective 'complex' taxon. This isn't helped by conflicting boundaries between treatments to where one species/variety occurs in one region but doesn't in another treatment, and floristic works may have separate maps altogether. Or it's things lumped in a way that's hard to backtrack, and the effort required to learn all of these changes and recognize these new taxa relative to the old concepts is deeply prohibitive. It just isn't worth it for 99% of people to try figuring out what all changed, so most stuff is now just 'Heterotheca' or 'Heterotheca villosa' because it just isn't worth it for anyone not specifically working on the group. I bear no ill will towards anyone who leaves things as "Heterotheca" or "Heterotheca villosa complex" - it's taken me a couple years to get up to speed on what most of these things are, and it's unrealistic to ask people to invest that much time in something when it could just be "Heterotheca sp." in a spreadsheet or observation.

E.g. some people are now calling all 'Heterotheca villosa' they see 'Heterotheca hirsutissima' (including those from Michigan or Canada as I recently saw), but most people just leave things at H. villosa (leaving Patrick or myself to try fixing the ID's sent our way). The other problem is that there is a "real" Heterotheca villosa, but now that only encompasses H. villosa var. villosa (and var. foliosa) from Semple's treatment, so the actual number of "correct" H. villosa's is ridiculously small compared to the amount of "H. villosa" sensu Semple that encompassed most of the western United States (and parts of Canada). Finding those is now virtually impossible unless multiple people (to ensure the ID's reach a majority consensus) go back and identify the thousands of observations uploaded from all over North America. The amount of effort taken to switch everything over is going to take literal months, and there will always be a lag as new observations are made according to the old treatment (leading to mixed iNat observations and herbarium collections). Some specimens with duplicates now go by multiple names (ex. H. excelsior, H. villosa, H. villosa var. scabra), and it just isn't realistic to have someone spend weeks trying to sort out all of the names and correct the tens of thousands of existing Heterotheca herbarium specimens out. So basically, unless Nesom, Patrick, or myself personally visit an herbarium to correct things, there's going to be pretty substantial confusion as to what's going on. This has been particularly frustrating for someone trying to study the group because I have to convert all of the names to one system to understand what occurs where and hope/pray that things are identified in the right species complex at least.

For what it's worth until things are sorted out better, pick either Nesom or Semple's treatments. The latter is probably easier to do because things can be left at species rather than variety but comes at the cost of potentially over-lumping variable populations as the same taxon. For our U.S. naturalists/biologists, Semple published a key to the U.S. species in the Flora of North America ( if the enormous 1996 work is too extensive ( If you want to use Nesom's treatments, someone made keys to various states if you want to follow that treatment ( They follow Nesom's treatment better than most local floras which may follow either Semple's, Nesom's, or some admixture of the two. It's taking quite some time to work out what all of the species should be, so I won't have any definitive works on the groups out for some time.

For those of you who have read thus far, thank you for your patience and interest in such a complicated group. A great deal of work remains to be done, but I'm hoping to generate some maps and keys this year to try and make things more understandable for those with a vested interest. Feel free to message me either on iNaturalist or at either or for specific questions, and I'll try to address them as quickly as possible. I also have a website where some additional information can be found: Thank you for your time today.

Julkaistu marraskuu 1, 2023 12:36 IP. käyttäjältä marisaszubryt marisaszubryt | 11 havaintoa | 0 kommenttia | Jätä kommentti