About those dandelions

To some extent, I think this is a question of whether you’re trying to address a research question vs. get into a high-profile journal—which is to say, I think there are a lot of research questions that could be answered just fine with datasets and techniques that would be considered passé.

Whether that particular study is an example, I do not happen to know. Sounds likely, though.

1 Like

Oh, if only dandelion sex were as simple as “if they produce pollen, they reproduce sexually” vs. apomictic. Some species produce seeds by apomixis (without sex) but produce pollen that can fertilize ovules of other species of dandelion.

Dandelions have no respect for our human need for neat categories.


I agree. And yet if I agree to an observer’s ID at Taraxacum and mark “as good as can be,” I’m likely to get pushback. Not much pushback, admittedly, but enough that I’ve stopped doing it. And that means the Needs ID Taraxacum observations will mount up till there’s so many no one will ever want to tackle them.


Definitely. For a problem like dandelions, I think we need as many lines of evidence as possible. However, many taxonomic questions can be answered just with morphology. My colleague and I believe we have correctly delineated a Bentgrass species by morphology alone. And I remember a Reedgrass question for which my major professor got a small grant to determine if what was called one species was really one or two. Well, it was clearly two, based on morphology once we actually looked carefully at all the specimens. However, we still had to do the genetic study because that was what the contract required. Surprise, surprise, the genetics indicated that we had two species.


If North Carolina has any similar number, it would be worthwhile to make these determinations. Although it would be important to know if different exotic species occur in proximity, and how many could be expected in one town, for instance.

Incidentally, about a year ago, I came upon an observation which I wondered if it might be one of the native species. Look how different it looks from the usual ones we see – for one thing, it has only a single row of rays: https://www.inaturalist.org/observations/55220704

Unfortunately there are not enought photos to say. It doesn’t look like T. officinale, but it would be nice to see other angles, the inflorescence on the side and the shapes of the seed.

Do you mean within dandelions specifically? Technically yes; the original description of each species will probably say if it’s sexual or asexual, and other sources may as well, it’s been a while since I’ve read through identification guides or taxonomy papers on it. My understanding is that generally the asexual populations are weedier than the sexual ones, so the sexual ones will tend to be more restricted to certain areas of Europe whereas asexual ones can be much more widespread and will be the ones that make it to North America. So I think you can assume that weedy non-native dandelions are apomictic.

There are complexes containing apomictic populations in a variety of other plant taxa as well, including raspberries, buttercups, hawkweeds, hawthorns… They have similar taxonomical issues, in most cases observations in those complexes are not identified past complex (many of my observations of those genera are stuck at genus).

This is the most thorough collection of resources for identifying them past section, for British species: Dandelions – Botanical Society of Britain & Ireland (bsbi.org)
In this video the leading British authority goes through around 30 sample dandelions and explains their identification.

If you haven’t read the BC paper yet that might clarify, but absolutely you’d expect many species in a state and quite a few in one town, depending on the variability of the habitat. For example in BC you get more sections represented near the coast and further north, whereas in inland urban areas it’s mostly just sect. Taraxacum. So I expect here in Ontario we’d have significantly less diversity because we don’t have coastal or alpine habitats. A lot of dandelion diversity comes from coastal and alpine habitats in Scandinavia, northern Europe, the Mediterranean etc.

1 Like

The small flower size immediately tells me it’s not section Taraxacum. Fortunately it’s in British Columbia so we can try the key on it. Unfortunately we don’t even know what sections occur in other states or provinces, and there’s no key for all the sections in Europe, so that’s a significant obstacle for the rest of North America.

The BC key is linked here or in the paper here. I should note that this observation is from July, so it won’t be the spring leaves. Apparently later “generations” of leaves are more variable and consequently less reliable, which is frustrating… Worth a shot anyway though to see where we get.

1: It’s at a low latitude in BC but in the Rocky Mountains so it could be a high elevation. From the map/background it looks like it’s growing in a gravelly area near a road (disturbed habitat). We can’t see the bracts, but the leaves have many lobes, with a couple complex lobes. So probably non-native. 1b → 2
2: Again we can’t see the bracts, but the gravel and leaves look dry… Also if you look at representatives of sect. Palustria the flower and especially leaf shape is very different. 2b → 3
3: Stuck here because we can’t see the bracts. Always get a photo of the back of flowers of Asteraceae! (and the underside of mushrooms! :P)
However step 4 is quick, we can check out the 2 sections there. The key descriptions aren’t useful since they only discuss the seeds, but we can look at the descriptions and images of those sections later in the paper. They both seem to be small dandelions with complex leaves, which matches the one in the observation. However the T. fulvicarpum group is exclusively coastal, which eliminates it. Let’s put a bookmark on sect. Erythrosperma and try the rest of the key.
5: Issue with the bracts again here, but I think details about the leaf shape, leaf colouration, and range eliminate the sections in steps 6 and 7.
8: We can’t see the bracts but given the tiny size of the flower I’m sure the bracts are also small. The leaves aren’t large enough to be “crisped”/rugose (indented where the veins are). 8b → 9
9: The petioles (leaf stems) have no wings (thin area of leaf beside the stem). The capitula (entire flower-head of an aster) is small. The leaves aren’t egg-shaped or lance-shaped, but then neither are the leaves of the example sect. Borea in figure 1… At least it is not sect. Boreigena.

So after all that I would say that the dandelion in the observation is probably not native, and it’s probably either section Erythrosperma or Borea, leaning slightly towards the former due to the leaf shape. Let me know if you disagree with my logic anywhere there… We could have more confidence with photos of the back of the flower (to show the bracts and the stripes on the underside of the petals), and having spring leaves and knowing the shape and colour of the seeds wouldn’t hurt either.
If I had more confidence about the section I might try cross-checking the 6 sect. Erythrosperma species that Bjork lists with European identification resources, to see if they’re different enough that I might be able to decide between them and if this observation is a good match for one of them.
Another potential complication is that this plant could be smaller and have fewer leaves because it’s stressed for one reason or another. I think that’s the case with this one I found growing on a rock. It’s hard to tell without more context and other examples of similar plants in the area.

I have a couple observations of dandelions that I don’t think are sect. Taraxacum, but have had similar challenges to this with identifying them to section. It’s especially challenging without having someone more experienced to confer with most of the time.
At least you can often tell when plants are sect. Taraxacum (larger plants with large flower heads, strongly recurved bracts, complex lobed leaves with rugose centres and 3D edges, usually with pollen on the stigmas). After that you can try to get lower with the British key (sect. Ruderalia = sect. Taraxacum), with the caveat that there are definitely species in North America from elsewhere in Europe as well.

1 Like

Don’t disagree entirely, but I think it depends on how good the a priori concepts are to begin with and what evidence you think will fill in your understanding. In general, though, I think the actual evidence that phylogenetic studies provide in supporting a species is often oversold and the importance of coming up with good morphospecies to test a priori, undersold (i.e., garbage in, garbage out). I often here people say, “the problem needs DNA” as if that will magically solve the problem. The problem of determining species is so much more detailed than that and if someone think it’s just matter of naming monophyletic groups, they’re missing out on so many important aspects of understanding what a species is. Honestly, I’m also a bit skeptical when researchers use limited datasets like this to address what appears to be infraspecific variation before they know that these are species level processes (truly infraspecific processes should be governed by population genetics where assuming that one or a few gene trees equates to species trees doesn’t really make sense). You also have incomplete lineage sorting and hybridization that complicates things, but if you get a simple tree, maybe you can assume it’s rare enough in your group to ignore.

My post was mostly aimed towards what I think you would need to provide enough evidence in the abstract to support a species quantitatively (i.e., articles that provide compelling evidence to support species delimitation in general).

Just to be clear, I don’t disagree with this at all. More that I don’t want people to think that, just because you have DNA, it will easily solve all the problems. It can be super effective at testing species, but you really have to be critical if there isn’t some well understood taxonomy based on some other rational (discrete morphological units or otherwise) to fall back on when you get the phylogeny.



1 Like

Best summary of this whole discussion so far!


I know but what is the point of species complex then ?There is a lot of species complex in iNaturalist, which are easier to determine than dandelions.

Part of the reason for that is that the iNaturalist taxonomy structure doesn’t have an option for “species group”, which is a categorization at a similar level but without the implied complications about separating species. So a lot of groups on iNat are called complexes when it would be more accurate to call them species groups.

All the taxonomic levels and how we decide where to categorize organisms into them can be pretty arbitrary, and as @aspidoscelis said earlier it’s done a bit differently between different taxa. However generally a complex covers just a handful of complicated and closely related species. A genus can cover a handful or hundreds or thousands of species depending on how closely related they are and which taxon it’s in; my impression is that plants tend to have larger genera than animals. The genus Taraxacum has hundreds of species, which can be subdivided into like 50 smaller groupings (a lot of Asian ones in addition to the European ones in discussion here) which are called sections. It happens to be that a bunch of the European sections do apomixis and so there’s a proliferation of microspecies in those sections. If it only affected a smaller number of species then it would make more sense to call them complexes. And again maybe when raspberries do apomixis, different taxonomists have decided on a different solution for categorizing them, I’m not sure.

1 Like


Personally, I’m on the cladist side. ‘Monophyletic’ is not a descriptor of species. Attempting to delimit species as monophyletic groups, then, indicates that one does not understand the relationship between the data and the question at hand.

(To some extent, this is a difference in emphasis from how the situation is more often presented; we could think of lineage sorting as the process by which phylogenetic relationships are created.)

It’s for cases when there is some group of organisms that people would like to recognize under a single designation, but that designation is not considered assignable to a particular taxonomic rank. They aren’t scientific names, but sometimes it’s appropriate to use them to fill in one of the gaps in scientific names. If you can just use a legitimate scientific name, though, that’s the better option.

1 Like

This paper gives an overview of all the apomictic plant groups and how taxonomists have classified them: http://www.ibot.cas.cz/hieracium/studygroup/pdfs/1017_1040_Majesky.pdf


The genus is traditionally classified into several subgenera, but only one – subg. Rubus – represents a group with predominantly asexually reproducing species (Alice & Campbell, 1999). … Taxonomic treatment of the apomictic species involved several approaches (for a review, see Weber, 1996).
(page 1026)


The section is the key infrageneric category for sorting the large diversity of Taraxacum species (). … A section contains one or a few diploid sexual taxa and a number of morphologically distinct apomictic polyploids (for which more than 3000 names have been published). However, there are several sections which contain only sexual or only asexual taxa (Kirschner & Štěpánek, 1996). Sexual taxa are described as species with a wide distribution (but restricted in comparison to apomicts) and with large phenotypic and genotypic variation (). Polyploid apomictic taxa are recognized based on their distinctive morphology and are traditionally classified as microspecies. … Some botanists, however, prefer to refer to broad species/species aggregates … This practice, however, is not correct [as it’s been done in the past, at least], as used taxon names do not always represent the types, based on which sections were described and defined, and thus should not be used instead of a sectional name ().
(page 1030, I took out most of the citations here for legibility)

Here’s the Kirschner & Štěpánek, 1996 paper cited in there: https://www.jstor.org/stable/4181475

In large areas agamospermy is known to predominate absolutely (Siberia, N Europe, N
America) but sexuality is by no means rare in the genus (only sexuals are known among
native temperate taxa in the southern hemisphere). …
We should mention that the absolute majority of sections or subsections without any known
sexuality (about 25 groups) are arctic or other northern groups, or are limited to smaller areas
in high mountains. On the other hand the most widely distributed derived or precursor European
and Asian sections usually exhibit a certain, often considerable degree of sexuality (Tab. 1).

(pages 416-417)


With insects, people often use species complexes when there are a few species that are each others’ closest relatives and the individual species can be hard to determine without very specific information (for example, dissection of the genitalia for insects) or the taxonomists are still figuring out exactly how many species there are in the group/how to tell them apart. So it’s a way of being more specific than just “genus X” without claiming that you know exactly what species it is.

As opposed to sections, which (at least to my understanding) are intended to recognize more similar groups/clades of (usually identifiable) species within a genus. For example, red and white oaks are different sections of Quercus

1 Like

Last night, after I finished reading the forums, I went to iNaturalist and looked at Taraxacum in North Africa, just to get an idea of the range of variation in another part of the world – I picked North Africa because I expected that the dandelions I know wouldn’t succeed there. My goodness, there are a lot of dandelion species which look nothing at all like Section Taraxacum.

I appreciate all the links. I will go through those resources carefully and see if they can lead somewhere.


Okay, we may not all accept the microspecies, but the consensus remains that the Sections are distinct, meaningful entities and that we should at least try to identify our dandelions to that level. This thread has given me a new “plantasy” – I want to find those Celtic dandelions.

In my exploration of Greenville, I am finding distinct types. By far the most common one in suburban lawns (which I think is the same kind as the observation I originally linked, but I have yet to go back and check that specimen) consistently has acute leaf tips and a very raggedy appearance:

However, there appears to be another kind, less common, with blunt leaf tips and is not at all raggedy:

I believe that they are both in Section Taraxacum, but I will definitely be checking to see if they key out sensibly.