Re: Levels of selection (Was: Can cognition override natural

Hi Bill,

Sorry for the delay in my response.

in article [EMAIL PROTECTED], William Morse at
[EMAIL PROTECTED] wrote on 11/22/03 10:42 AM:

>> The system cannot self-organize if there is no communication among the
>> components (subpopulations in this case).  If you explore a
>> computational model of evolution in a spatially explicit context, you
>> can ASSUME lack of communication between subpopulations and divergence
>> happens.  Of course you have only learned what you assumed in that
>> case.  If you allow for limited gene flow among subpopulations in a
>> stepping-stone model, or among locations in a more general model of
>> isolation by distance, then you can observe emergence of spatial
>> boundaries between gene pools that you did not assume. This way you
>> can learn about how such local subpopulations can evolve without just
>> a circular confirmation of an assumption.  Divergence among
>> subpopulations in this way requires gene flow.
> 
> Fascinating! (I snipped the remainder which was more about the level of
> selection argument) One of my (and many others - I am not original in
> this)  "benchmarks" for evolutionary explanations (as some of you may be
> aware) is the cichlid fishes of Lake Victoria. Since there has not really
> been time for allopatry to work its speciation magic, the question is how
> do you explain the observed diversity ? What you seem to be claiming
> above is that you have a model that does this.

The "model" I described above is very general.  The one that I have been
working with depends critically on the constraints of space, which may or
may not be a critical factor in the African cichlid story.  I suspect it is
a factor, but only because of the social viscosity of so many cichlid
species.  You might want to see the papers by Sergei Gavrilets on parapatric
speciation for a formal analysis of these kinds of ideas.  The model can, of
course, be generalized further to allow for the same sort of process to
unfold in the absence of spatial segregation.  In essence, the constraints
of space can be replaced by other constraints (e.g., social constraints)
that maintain the structural integrity of the mating network.

> One of the reasons I am particularly intrigued is that I have read
> elsewhere (and I wish I could remember where)  that much speciation
> occurs not at the fringes of a population but in the middle.

This has been dubbed the centripetal model of speciation, as opposed to the
centrifugal (or peripatric) model.

> This makes 
> intuitive sense from one standpoint - one might assume that marginal
> populations are in fact only marginally adapted to their local
> environment, and might not be likely to survive if isolated from
> replacement from the main population. I have previously argued that (at
> least for all but large  macrofauna) relatively severe weather (the 1000
> year storm) can provide "mini-allopatric" conditions. But there is likely
> to  be at least some gene flow within recovering populations, so a model
> that can show divergence even with gene flow is enticing.

Fight.  If I can get enough funding to support this aspect of my research, I
hope to explore the interactions between environmental heterogeneities
(spatial and temporal) and the generic tendency toward self-organization in
a homogeneous setting.  After all, both appear to be going on in nature.

Cheers,

Guy