Friday, April 10, 2015

Chance dispersal, 'normal' dispersal, and long distance dispersal ... confused yet?

Having now picked up Michael Heads' second contribution to the recent issue of Australian Systematic Botany, Biogeography by revelation: investigating a world shaped by miracles, I am glad that I read the other one first. On the one hand, the present paper is really just a 23 pages long criticism of a single book, Alan de Queiroz' The Monkey's Voyage: How Improbable Journeys Shaped the History of Life; on the other, its argumentation is remarkably redundant with the first paper. It even contains another discussion of the ratite birds! Was it really necessary to write both of these papers, and for the same journal issue at that, considering that this criticism of de Queiroz is merely a special case of the criticism of mainstream biogeography that has been expressed in the first one?

And because it is all about one book, the paper is also, at least in my eyes, a remarkably uninteresting contribution to the discussion. I have no intention of reading The Monkey's Voyage, so I cannot judge if any quote mining is going on or not. But let us for assume the sake of argument that Heads' criticisms are, in this case, right on the mark; that de Queiroz really is that most elusive of straw men, a biogeographer who not merely sees an important role for founder effect but who actually rules out the possibility of vicariance as a speciation mechanism. Would that make the panbiogeographic approach, here strangely called 'vicariance theory', any more defensible?

Of course not. If you could find a person who irrationally rejects the possibility of vicariance a priori, that would still not in any way whatsoever make the a priori rejection of speciation after long distance dispersal less irrational. So in the grand scheme of things there is really not much point in addressing anything the present paper argues about de Queiroz.

Similarly, a very large part of the paper is taken up by case studies where the author argues for vicariance scenarios. Again, let us assume for the sake of argument that vicariance is the correct answer in all these cases, every single one of them. Now would that make the panbiogeographic approach any more defensible?

Again, no. The real question of interest is whether a panbiogeographic analysis is science, and because it always builds the conclusion into its starting assumptions (the ancestor was already distributed everywhere) there are good reasons to argue it isn't.

As much as Heads tries to frame the controversy as 'dispersal theorists' versus 'vicariance theorists', to the best of my knowledge at least there are really hardly any, if any, dispersal theorists anywhere, at least not any fitting his characterisation. Who I have met and read are numerous mainstream biogeographers who accept vicariance, range extension, long distance dispersal, local extinction, range shifts and (depending on who you ask and what geographic scale we are talking) sympatric speciation as possible biogeographic processes, deciding between them case by case depending on the available evidence.

And on the other side "the panbiogeographic approach, as illustrated here, explains allopatry by vicariance and overlap by dispersal" (Heads, 2015). Notabene: One single process to explain speciation, all others are forbidden. There is no symmetry to these two positions.

Still, there are a few things that occurred to me when reading this paper that didn't when I read the first one.

As in the first paper, Heads argues that vicariance theory is on the rise thanks to the tireless efforts and good arguments of panbiogeographers. His evidence is that the number of publications mentioning 'vicariance' in title or keywords is increasing, and in the present paper he actually has a diagram showing a steep increase in such publications from 1970 to today.

Again, to me this is besides the point anyway because it would be hard to find a mainstream biogeographer who rejects vicariance. But I had another suspicion and so I went and obtained the same data for other search terms, only using Scopus instead of Google Scholar. Here is the graph for 'dispersal':

Note also that the 'vicariance' graph maxes out at less than 2,000 while this one ends above 4,000. Zounds! Sure looks as if dreaded despicable dispersalism is gaining ascendancy. But how can it at the same time as vicariance is doing so? Here is another one, the graph for 'phlogiston':

Even weirder. Does that prove that phlogiston theory is gaining traction in recent times? No. It merely shows that there are now more scientific publications coming out per year than there were in the 1970ies, and perhaps that more of them are in searchable databases. Arguing about the acceptance of theories based on a glorified Google Trend is ... not productive. Remember also that a paper titled "vicariance theory is wrong" would have been scored as a data point in favour of vicariance theory on Heads' graph.

In a previous discussion of panbiogeography I mused that its logic is self-defeating: it rules out dispersal to explain all speciation through vicariance, but for that to be possible the ancestor must first have dispersed all over the place, and generally across such huge areas as to be unrealistic for a single species. (By the time you get to eastern Sibera, you will have evolved so much that you are not the same species as your relatives in Patagonia.) If the ancestor could disperse so readily, why not the descendants?

I speculated then that panbiogeographers might locate all dispersal in the time of the super-continents like Gondwana and Laurasia and merely consider it impossible since their break-up. To my great surprise, Heads goes one step further in the present paper and writes explicitly that panbiogeography accepts even long distance dispersal:
Despite this, [de Queiroz] finished every chapter by citing observations of animals or plants moving long distances, as if this indicated chance dispersal. He argued that 'we know that long-distance dispersal occurs because people have actually seen it happen' (p. 288). But the observed dispersal is 'normal dispersal' (without speciation), which often takes place over very long distances, as in sea birds, migrating birds, sharks and others.
One might add fern spores, pappose daisy seeds, ballooning spiders and various other groups. But the point is this: Heads accepts that organisms can disperse across very long distances. But he does not accept that they can disperse across very long distances and then, when they are in the new place and isolated from the parental population, evolve into a new species, because that would be speciation through founder effect.

Let's try that again: Yes, they can get across the ocean alright, but speciation can ... somehow ... only happen if they got here across land.

Does anybody understand how that is supposed to work? Because I surely don't.

No comments:

Post a Comment