Colin Walker wrote: ↑Tue Jun 14, 2022 9:29 am
Agreed Ralph there's no tightly defined definition of a genus. However, with modern molecular evidence, we're getting closer. So, the components of a genus, species or whatever rank you wish to consider, have to have a common evolutionary origin and form a discrete branch (clade ) in the family tree.
I agree in principle with Colin's idea that whatever we stick a label on should be a discrete branch of the tree, but there are still some sticky issues:
(1) Clades, just being bits of the tree with a common ancestor, are a more generally useful idea than species or genera, which are (or should be) just names for clades which satisfy certain additional requirements (unable to reproduce with others outside the clade, for example). The requirements should in principle be fixed according to the level (species, genus, etc) we are considering. But in practice these additional requirements are not strict - applying commonly used ones about being able to reproduce, we not only get intraspecific hybrids, but intrageneric hybrids. If you tweak them to get "common-sense" answers, they become subjective.
A more objective way of doing it would be to say that things within a certain genetic distance of each other should be considered to be a genus, or a species, for example. However, I doubt that they can be made objective - how do you weight the importance of removing or duplicating a gene, with putting it somewhere else on the chromosome, or a mutation, for example? What about aberrant organisms with extra chromosomes? Furthermore, epigenetics can turn genes on or off, which can muddle things further.
(2) It's not really a tree anyway. (It's a directed acyclic graph, DAG, if you want to be precise - directed just says it is clear which is parent and which is offspring, and acyclic just means you can't be your own ancestor). We call it hybridisation if two reproducing organisms' last common ancestor was a long way back up the "tree". Assuming it's a tree is a simplification that is typically used in computer based analysis of DNA, to make the mathematical problem of determining the most plausible explanation easier. Other assumptions are made too, e.g. that the simplest explanation is correct, so for example an explanation in which something is always present is preferred over a less likely answer in which it is lost and regained. Less likely of course does not mean mean "never happens". How often hybridisation is important for cactus evolution, i don't know, although some seem to hybridise pretty readily.
Anyway, the useful side of DNA analysis is that it can show us mistakes, where big differences in DNA between things we though were closely related are actually far apart (because they looked similar, for example). But I can't see how it can really be used to resolve or justify questions of what a species is, or a genus is, in terms of a chunk of the DAG. And it can only give us the most likely answer, given some assumptions, which is not the same as what actually, historically happened. When there are so many "species" in the world, less likely explanations will be the real ones from time to time.