Monday, January 12, 2015

Much ado about absent ancestors

(The following is the seventh part of a series of posts on an Annals of the Missouri Botanical Garden special issue on “Evolutionary Systematics and Paraphyly”. All posts in this series are tagged with “that special issue”.)

Pages 79 to 88 of this special issue meant to promote the acceptance of supraspecific paraphyletic taxa in biological classifications are taken up by a contribution entitled Paraphyly and polyphyly in the worldwide tribe Rubieae (Rubiaceae): Challenges for generic delimitation. Written by Friedrich Ehrendorfer, Michael Barfuss and Vladimir Vladimirov, it is perhaps best described as a concise review article summarising the current state of knowledge about phylogenetic relationships in their group of interest, and as such I found it extremely interesting and rewarding to read.

Even if they are not familiar with the scientific name, most people in the northern hemisphere will know Rubieae. They are characteristic herbs with seemingly whorled leaves; only seemingly because in reality there are only two true leaves, and the additional members of each whorl are derived from stipules, as demonstrated by the fact that only up to two side branches arise from each node. The flowers of these plants are usually small, radiate and white or yellow. The group has an extremely wide distribution; there are even native representatives here in Australia.

The phylogenetic studies discussed in the paper show that the two largest genera that are traditionally accepted in the Rubieae, Galium and Asperula, are horribly non-monophyletic in their current circumscriptions. Faced with this situation and considering the context in which the paper has been published, one would now expect, or at least I would expect, that the authors argue for the continued acceptance of these genera as non-monophyletic. That was, after all, the whole point of the symposium in Vienna that inspired the special issue.

Nonetheless, if one were to expect such an argument one would be disappointed. Maybe it is a failing on my part and I simply do not understand, but nowhere in the text have I seen the desire to classify, and the phrasing here will soon become important, into paraphyletic supraspecific taxa any of the extant species that are actually available for classification. Instead, the authors argue that:
it is obvious that today's taxonomic subdivisions within the tribe are still strongly influenced by tradition and should not be maintained.
Somehow I imagine that this is not a stance that Richard Zander for example would be comfortable with. To the best of my ability to follow their review of phylogenies, none of the taxa in their suggested re-classification will be non-monophyletic. However, that is not how they see it; in various places, the authors say that they will accept paraphyletic taxa even though their classification will not include a single species that will make any of them paraphyletic:
A problem emerges if Kelloggia is not regarded as sister but rather as a surviving relic of Rubieae ancestors. This would make Keloggiinae paraphyletic from the standpoint of cladistics because the subtribe then should include all of its increasingly apomorphic extant descendants (cf. Schmidt- Lebuhn, 2011: fig. 2, X+C vs. A, B). [...]
One could interpret Galium sect. Cymogalia as a relic of an ancestral paraphyletic group that has given rise to the remaining, more apomorphic Rubieae clades VA–D and VI (see Fig. 2). [...]
Thus, contrary to the arguments recently summarized by Schmidt-Lebuhn (2011) and in agreement with Hörandl (2006, 2007), Stuessy (2009), Hörandl and Stuessy (2010), and many others, we prefer the latter, less rigid, evolutionary, and also morphologically plausible approach: we accept paraphyly in DNA-supported cases, where extant, relatively plesiomorphic and rather unchanged ancestor-like taxa in their geological past appear to have given rise to strongly deviating and clearly more apomorphic descendant taxa (see Schmidt-Lebhuhn [sic], 2011: fig. 2).
It might be a bit hard to understand what is going on here, so I will try to illustrate the differences between an 'evolutionary' classification which to promote is the point of the special issue this paper appeared, the approach taken in the paper itself, and the cladist approach that this paper is taking such care to argue against. First, a possible 'evolutionary' classification:

As we can see, the red taxon is paraphyletic because it contains some but not all descendants of its most recent common ancestor. Now a phylogenetic or cladist solution:

Here, all taxa are monophyletic, AKA natural groups, because they are entire branches of the tree of life. (The black parts would be left unassigned at this taxonomic level.) Finally, one of the possible solutions under the approach apparently taken in the present paper. Be careful or you may miss the difference to cladism:

Yes, that appears to be it. The argument against phylogenetic systematics in this paper is that it is too “rigid” because the authors want to mentally assign the ancestor X of green, red and blue, which they do not actually have anywhere but infer to be most similar to green, to the green taxon; and they want to assign the ancestor Y of red and blue, which again they do not actually have anywhere but which they infer to probably have looked quit a bit like red, to the red taxon.

To repeat: none of these ancestors will actually appear in the classification. And even if one of them were found as a fossil, we could not be really sure that it is the ancestor, and we should most prudently restrict ourselves to saying that this looks like a fossil very close to what we suspect the common ancestor would have looked like.

In other words, unless I severely misunderstand the paper, the future classification of the Rubieae as desired by the authors will be completely indistinguishable from a cladist one. (Note again that these are just possible options. Another solution would be to assign every single species to its own genus, or to sink everything into one. The point is simply that whatever solution they chose, it will apparently be one that a cladist could just as well arrive at.)

Yes, assigning an ancestor to the same group as only some of its descendants as opposed to all of them is rather non-cladistic (and, I feel, illogical), so the authors of this paper are not cladists at a philosophical level. It also has to be admitted that they show some confusion in other areas. In the abstract there is talk of taxa “that do not contain all descendant species of an ancestral clade”, a phrasing that given the meaning of the word clade makes as much sense as a rectangular circle, and the genus Kelloggia is described as placed at the base of the Rubieae “in spite of the presence of a plesiomorphic 50-bp sequence” instead of precisely because of its presence. Both these sentences might be taken to betray some misconceptions about basic concepts of phylogenetics.

But still their classification will, for all practical purposes, quite sensibly be phylogenetic, and I remain puzzled why a purely phylogenetic classification of the extant Rubieae was considered sufficiently helpful to the cause of 'evolutionary' classification to be included in this special issue.


Ehrendorfer F, Barfuss MHJ, 2014. Paraphyly and polyphyly in the worldwide tribe Rubieae (Rubiaceae): Challenges for generic delimitation. Annals of the Missouri Botanical Garden 100: 79-88.

No comments:

Post a Comment