[Taxacom] Hedges /Kumar (eds) The Timetree of Life

[Sergio Vargas]

  Hi Jason,

I just wanted to add some comments to your discussion on 
panbiogeography, at least as currently used:
> Cladistic biogeography has weaknesses, but it still recognises dispersal and vicariance (as do most biogeographers) as valid mechanisms to explain current distributions. To which we could add extinction of course, but to prove this you need factual, objective data (fossils again). On the other hand not only does panbiogeography function in the absence of phylogenetic information (more in the next paragraph), but in addition it can function in the absence of dispersal (or explain it away). Surely organisms can move around and sometimes hit the insular or contienetal jackpot. Calling names like dispersalists seems to me like a tactic to draw attention away from the limitations of the panbiogeographical methodology.
I agree with you that cladistic biogeography recognizes both dispersal 
and vicariance, it uses vicariance as a null model and when the null 
model doesn't fit postulate dispersal (pre- or post- speciation) to 
explain discordance. I find this really powerful. Now, panbiogeography 
can use phylogeny as well. Page published a method to incorporate 
phylogenetic information into track construction. To the best of my 
knowledge, this method has been never used. This I think is because 
people tend to use panbiogeography as an escape when they don't have 
phylogenies at hand. But yes, it can work without a reference phylogeny 
for the group of organisms under study. Regarding dispersal, I disagree 
with what you say about panbiogeography working without dispersal or 
explaining dispersal away. Criozat term mobilism clearly refers to 
dispersal, but definitely not to random dispersal. I think, 
panbiogeography tries to look for the causes of "inmobilism", hence 
panbiogeography's fixation with vicariance. I think the core of the 
controversy between panbiogeographers and dispersalists is that the 
panbiogeography rejects random dispersal as an explanation for 
distribution patterns. I see no problem with this. Random dispersal is, 
at best, naive. If dispersal accounts for a historical pattern it means 
dispersion must have been a non-random process, not a unique event. I 
like to think cladistic biogeography accepts this idea about dispersal. 
I think the criticism goes to papers pointing to random dispersal as the 
cause of a pattern; not sure these qualify as cladistic biogeography...

There is another problem with panbiogeography in its current 
incarnation. This has nothing to do with Croizat or the way he did his 
panbiogeography but with the use of clique analysis for the 
determination of generalized tracks. If you check Croizat's work (not an 
easy task: hard to read and to follow), it is clear that he operated 
with graphs summarizing the distribution of a taxon. Yet, track 
compatibility analysis operates on a compatibility matrix, and no longer 
on the graphs. Moreover the graphs are never used if you use this 
approach because when you code the individual tracks for compatibility 
analysis you loose all the information on the edges of the graph. 
Anyways... the problem of using compatibility is that you turn 
panbiogeography, without necessity, into a maximum-vicariance technique. 
This was, I think, not intended by Croizat who criticized chance 
dispersal, but acknowledge its own flavor of dispersal under the term 
mobilism. I could be completely wrong though... would not be the last 
one not understanding Croizat's writings ;-)

cheers

sergio