Monday, December 22, 2014

An eighteen pages long perfect solution fallacy (that special issue on paraphyly)

(The following is the fifth 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”.)

As I am looking at the next contribution to Annals of the Missouri Botanical Garden's special issue on how awesome it would be if only we would accept paraphyletic groups as they did in the 1950ies, it seems as if I should start with the following disclaimer.

I want it to be understood that I have the highest respect for the life works of everybody involved, and for their publications other than the one we are currently dealing with. I do not wish to offend, but merely to critically discuss the scientific merits of 'evolutionary' systematics versus phylogenetic systematics, and specifically whether the arguments presented by the former school of thought make any sense. It also needs to be understood that my opinions expressed here are my own and are not necessarily those of my employer, nor those of my line manager, of my colleagues, of my friends, of my relations or, for that matter, of my pot plants. The same applies to all my posts, of course.

With that out of the way: the contribution I will discuss today, The case against the transfer of Dryandra to Banksia (Proteaceae) is … not a terribly well written paper.

For background, several years ago it was found that both molecular and morphological data showed Western Australian Dryandra to be phylogenetically nested within more widespread Banksia. It is no exaggeration to say that a Dryandra is just a Banksia with a shorter inflorescence, and consequently most taxonomists and systematists decided to unite the two genera. However, the “loss” of Dryandra as a distinct genus has left many people profoundly unhappy. One of these people appears to be the author of the present contribution.


This paper is packed to the brim with numerous unrelated claims that cladists would consider to have been refuted by the 1980ies at the latest, but it would surely look impressive to anybody who is new to the controversy. A small selection:
  • “... have discussed the different concepts of monophyletic – the original meaning of Haeckel to define simply a group with a common ancestor ...” (Yes, but a polyphyletic group also has a common ancestor. Any two randomly chosen species on this planet have a common ancestor.)
  • “Phylogenies of groups above the species level are, with rare exceptions, unverifiable hypotheses.” (Ah, like all of science then, because it is in the very nature of scientific hypotheses that they are falsifiable but unverifiable.)
  • “Mapping morphological characters onto a phylogenetic tree derived from molecular data is an empty procedure, since those characters have not been tested for convergence with the data from which the tree has been built.” (Oh come on now – that mapping is the test for convergence that you are asking for! Also, if one were to use the morphological data to make a tree and then look at character distributions one would be open to the charge of circularity.)
And so on.

And barely any attempt is made to actually address the arguments made by phylogenetic systematists. For some strange reason, a single obscure special purpose paper on the genus Hibiscus (Pfeil & Crisp, 2005) is attacked over and over again despite its apparent irrelevance to Banksia, but any and all phylogeneticist literature discussing general principles is either ignored or brushed off as “biased”. How convenient!


At least from my perspective this paper is rambling, repetitive, unstructured and waaay to long for comfort. There are actually subheadings that read “Other Aspects” and “Furthermore...”, which is certainly a first among the scientific publications I have read so far. One wonders if this is all some big joke on the author's part, and I found myself scanning ahead for “And Another Thing...” and "By The Way...", but sadly those didn't come up.

The main argument, or how to shoot into one's own foot

Despite its lack of structure, this paper does have one argument that one could consider central because it pops up all over the place. It is that the sinking of Dryandra into Banksia was wrong or at a minimum premature because (1) there are some mistakes or unclarities in the crucial papers published by Austin Mast and Kevin Thiele, and (2) sampling was not at 100% of the known species, and thus more evidence is needed before any conclusions can be drawn.

There are several issues with this argumentation. First, this is straight out of the denialist's handbook: side issues such as vouchering or irrelevant minor incongruence between nuclear and chloroplast data are used to argue that the author's position wins by default, and to distract from the balance of evidence. But the balance of evidence was always on the side of Dryandra being nested in Banksia, and funnily enough the author admits that right at the start: “In fact, the latter 'tree' [showing Banksia as paraphyletic] is the traditional view of the relationships of Dryandra and Banksia.”

Second, it is all well to criticise Mast and Thiele for failing to sample more species, but science has made even more progress since their studies of 2007 and 2008. Just last year, Cardillo & Pratt (2013) published a new phylogeny of Banksia. I spoke with Marcel Cardillo a few days ago, and apparently they have sampled significantly more than 90% of the species of Banksia including Dryandra. And of course the results are again the same: the latter is nested within the former. How much more evidence will finally be enough to draw conclusions?

Third, whenever the author is arguing that things may be more complicated than presented in the papers of Mast and Thiele, he is not really helping his case. The complaint that “only a small proportion of Dryandra taxa was sampled” is utterly irrelevant because such additional sampling would not suddenly make Banksia monophyletic (only sampling less of Banksia could achieve that).

In fact it only has the potential to get worse: on pages 40 and 41 the author discusses what he perceives to be mistakes in Mast and Thiele's interpretation of character evolution, and his corrections are all either making the two taxa sound even more similar (“their 'Involucre of conspicuous bracts' as a synapomorphy for Dryandra is also incorrect. All species of Banksia and all species of Dryandra have an involucre of bracts subtending the inflorescence.”) or else raising the possibility of Dryandra being polyphyletic!

So here we have the odd situation of somebody arguing for Dryandra to be treated as a distinct genus on the basis of it potentially being such an unnatural grouping that not even fellow 'evolutionary' systematists would be able to accept it. (And once more one has to ask if these contributions to the special issue on paraphyly were critically peer reviewed at all.)


This paper is also, it has to be said, profoundly anti-scientific. One can find in it, for example, the following complaint: “We also have the situation of the same author being involved in the transfer of species from one genus to another and then back again a few years later after further research...” Ah, so being willing to change one's opinion as more data come in is a bad thing? Funny, I thought that is how science is supposed to work!

Repeatedly we are assured that nobody has to accept a taxonomic decision just because it has been made by the Council of the Heads of Australasian Herbaria. That is true; in related news, people are free to use leeches to treat cholera even as qualified medical professionals advise against it.

The last part of the paper then proudly cites various laypeople (Wildflower Societies etc.) who have rejected the phylogenetics-inspired merger of Dryandra into Banksia, as if that were any kind of argument. There are many people who reject the Theory of Evolution. So what?

Paraphylists failing by their own standards, again

There is one claim in this paper that I find particularly funny. The author writes, “separately, Banksia and Dryandra are readily distinguished morphologically – even people with just basic training have no difficulty recognising a Banksia or a Dryandra, even if it is a species that they have not yet seen.”

Two or three years ago I gave a seminar to the Science Group of the Friends of the Australian National Botanic Gardens, dealing with the question why we try to make taxa monophyletic. Because it is one of the most controversial cases in this country, I deliberately chose Dryandra and Banksia as one of several examples. It is nice to have more than just pictures, so the botanic garden supplied me with a few pots of living plants for illustration.

When a garden ranger carried in the Dryandra, he looked at me and said this: “What a nice plant. I have never seen this species before. Is it a Banksia?” I laughed and replied that his instincts served him well. Yes, these two genera are really extremely similar, and “even people with just basic training have no difficulty recognising” ... that they belong together. (Unless, I guess, they have grown up with two separate genera and are hostile to change.)

I rest my case.


Cardillo M, Pratt R, 2013. Evolution of a hotspot genus: geographic variation in speciation and extinction rates in Banksia (Proteaceae). BMC Evolutionary Biology 13: 155.
George AS, 2014. The case against the transfer of Dryandra to Banksia (Proteaceae). Annals of the Missouri Botanical Garden 100: 32–49.
Pfeil BE, Crisp MD, 2005. What to do with Hibiscus? A proposed nomenclatural resolution for a large and well known genus of Malvaceae and comments on paraphyly. Australian Systematic Botany 18: 49–60.

No comments:

Post a Comment