But I believe the term variety is also applied to artificially-produced cultivars in which case it may not have to follow the botanical nomenclatural rules.
Yes, for me as soon as a taxon lower than the species is described the nominate taxon is created since the new descriprtion is made in comparison with the nominate taxon.
Notwithstanding it could be that some electronic databases for some species show the subordinate taxa but not the nominate one.
Botanical varieties work the same as subspecies. There are three primary infraspecific ranks, subspecies, variety, and forma. (In that order, if used in hierarchy) https://www.iapt-taxon.org/nomen/pages/main/art_4.html
Most plant taxonomists try to stick to a single one of those ranks, but it is permissible to create an infraspecific classification using a hierarchy of those ranks. Such classifications do not create quadrinomial or multi-nomial names however. Any and all infraspecific names are trinomials combined directly with a species. https://www.iapt-taxon.org/nomen/pages/main/art_24.html
For sure, although I have been noticing a lot more than there used to be. Seems like that is something that POWO could automate with a script, once they have decided to recognize heterotypic infrataxa.
Here is an example, just for fun: Tony Orchard in 2004 undertook a revision of Bedfordia linearis (Labill.) DC. in Tasmania. He described Bedfordia linearis subsp. oblongifolia Orchard, and this automatically created the autonym Bedfordia linearis (Labill.) DC. subsp. linearis.
He then further described Bedfordia linearis subsp. oblongifolia var. curvifolia Orchard (explicitly as a variety of Bedfordia linearis subsp. oblongifolia Orchard). This automatically created the autonym Bedfordia linearis subsp. oblongifolia Orchard var. oblongifolia.
The proper name of Bedfordia linearis subsp. oblongifolia var. curvifolia Orchard acording to the Code of Botanical Nomenclature is Bedfordia linearis var. curvifolia Orchard, and Bedfordia linearis subsp. oblongifolia var. oblongifolia is referred to as Bedfordia linearis var. oblongifolia Orchard (which, confusingly enough, is actually an autonym, but of the subspecies rather than the species).
This way, Bedfordia linearis var. oblongifolia Orchard and Bedfordia linearis subsp. oblongifolia Orchard are both simultaneously valid (in two different ranks).
Shenzen code 26.1/26.2 - An infraspecific autonym can refer only to the species epithet and not a (different) infraspecific epithet. Bedfordia linearis subsp. oblongifolia Orchard var. oblongifolia is not an automatically created autonym as far as I can judge?
They are often referred to as varieties, but are technically called cultivars (with the abbreviation cv.). This distinguishes them from true varieties. Varieties typically apply to extreme populational forms of a species. Itās not a completely consistent usage as I have come across many cases where the plants that varieties represent are actually fully reproductively isolated from other varieties and should thus be treated as species. But, establishing reproductive isolation isnāt always easy.
Back to cultivars. Most of the time, Iāve noticed the cv. is skipped and is implied by the form. E.g., Euphorbia graminea āDiamond Frostā. Note that the cultivar name isnāt italicized and is within single quotation marks. Iām not completely sure if the example represents āproperā form or not, but it seems quite consistent.
I donāt see that wording in the Code, but I interpret this to mean that any autonyms are referrable only to the type (hence āspecies epithetā, which is the name tied to the type). It has nothing to say about ranks.
Botanical nomenclature is much more complicated than zoological. Zoologists have subspecies (and even those are slowly fading from use in many groups) with a nominotypical (nominal) subspecies. Never heard of an autonym, but itās basically the same thing.
26.3. The first instance of valid publication of a name of an infraspecific taxon under a legitimate species name automatically establishes the corresponding autonym.
[my emphasis] Bedfordia linearis subsp. oblongifolia var. curvifolia automatically created Bedfordia linearis var. linearis, not var. oblongifolia. The latter would have needed to be named explicitly as var. oblongifolia Orchard, which I expect it probably was given the classification he employed in this example.
I can imagine @jnstuart rolling his eyes right now