In Praise of Subspecies

The definition of what's a species or subspecies starts getting fun when you consider ring species[0].

[0] https://en.wikipedia.org/wiki/Ring_species?useskin=vector

As someone who has recently been dealing with taxonomic classification, this world is a complete mess. For one, there isn't even a singular authority on classification. Secondly, the "data model" we have for classification isn't even capable of actually representing proper lineages.

Linnaean classification, i.e. Domain > Kingdom ... Genus > Species is a tree data structure, wherein each node has one parent. Now what happens if you have something like a Leyland Cypress, a hybrid?

A Leyland Cypress is a very popular tree that is the product of hybridizing an Alaskan Cypress and a Monterey Cypress. Let's begin analyzing all of the problems with this:

1.) As mentioned, we can only choose one parent, even though all species have 2. Which lineage to choose?

2.) We know the exact parents of this new species, and yet the fact that Linnaean taxonomy uses a fixed number of ancestor node ranks means we can't classify this under a parent, but rather go to the parent's genus

3.) The idea of "species" being the only thing that can produce viable offspring is completely incorrect, so being stuck with a parent's genus as a foolproof way of reparenting a node is not reliable. While an Alaskan Cypress and Monterey Cypress used to both be classified as Cypresses (Cupressus nootkatensis and Cupressus macrocarpa, respectively), they've now each been reclassified multiple times into multiple different genuses, and are now Callitropsis nootkatensis and Hesperocyparis macrocarpa respectively. So we've created a viable offspring not just from two different species, but two different genuses.

4.) Many of these classifications were come up with purely phenotypically and have nothing to do with genetic lineages.

So is a Leyland Cypress classified as Callitropsis or Hesperocyparis, or the former Chamaecyparis or Cupressus? Just choose an arbitrary grandparent node, and throw away the information of half of its lineage, as well as the information of its species-level parentage.

And then within the species level, as mentioned by the original article, we have subspecies (or infraspecific ranks). Subspecies, varieties, subvarieties, forms, subforms, proles, cultivars, etc. If you think there's sanity found in that world, you'd be mistaken.

Fortunately scientists dealing with actual genetic lineages have begun switching to a different system which is not only founded in genetics, but also has a proper data model to represent it, which is a graph wherein any node can be associated with a particular ancestor. So a Leyland Cypress could be parented to an Alaskan cypress and a Monterey cypress. It can also be parented to ancestors of any rank, and this parentage can be any number of levels deep. Each of these assertions of parentage puts a particular entity in what is referred to as a "clade," making this form of classification referred to as "cladistic" or "phylogenetic."

[flagged]