Jacobaea vulgaris (Asteraceae), also known as Senecio jacobaea, Germany, 2005. This species is a widely introduced weed, and among the places it has been introduced to is Australia.
I had known for some time that it and several relatives had been officially segregated out into Jacobaea. Having met it first as a student under a Senecio name, not seeing any obvious morphological differences to species that remained under that genus, and, perhaps most importantly, without any pressing need to research the matter, I was unsure what to make of it.
Perhaps this was just some random splitting based on a whim? Or there would have been a splitting and a lumping option to make something monophyletic, and the splitting option that creates two indistinguishable genera was preferred over a lumping option that would create a heterogeneous genus? For all I knew this may well have been one of those scenarios that 'evolutionary' systematists always complain about.
Recently, however, a colleague discussed the matter with me, and that prompted me to properly study the literature. The phylogeny of the Senecioneae was thoroughly resolved in a series of papers published around ten years ago by Pieter Pelser and collaborators. Perhaps the best overview for present purposes is provided by figure 1 of Pelser et al. (2007, Taxon 56: 1077-1104). It is a single phylogeny broken up over ten pages. The part that concerns Jacobaea is figure 1F, and it can conveniently be summarised as follows:
So this is not just paraphyletic. Even 'evolutionary' systematists should be unable to accept a genus that includes Jacobaea and Senecio in the strict sense. But surely then two clades separated by so many other clades, among them illustrious genera like Crassocephalum and Dendrosenecio, must have some diagnostic characters? Sadly, no:
Although Jacobaea is distinguished from Senecio s.str. and other genera in Senecioneae by DNA sequence characters (Pelser & al. 2002), clear morphological synapomorphies for Jacobaea have not been identified to date.That may be very unsatisfying, but life isn't always fair. There is no rule that says that we will always find diagnostic characters to tell all distantly related groups apart.