"John Wilkins"  wrote in message
news:c8098r$237g$1{at}darwin.ediacara.org...
> Perplexed in Peoria  wrote:
>
> > "John Wilkins"  wrote in message
> > news:c7ph34$1gh$1{at}darwin.ediacara.org...
> > Let me know if I paraphrase it correctly.  "Even though a
trait may be a
> > species-level trait, in that there is no variation for that trait within
> > the species, Jim's proposed mechanism of species-level selection for
that
> > trait would also work as a mechanism of population-level (or
group-level)
> > selection if the species did vary for that trait.  Group-level selection
> > is still group-level selection, even if the group in question is the
> > entire species. Therefore, we don't need the concept of species-level
> > selection."
>
> Sort of. It is not that a trait *does* vary within species, but that it
> *can*, and if it does then it can go to fixation (i.e., spread to the
> entire species) through population-level dynamics. If that is true then
> there is not a presumption that because something is species wide - an
> invariant character - it had to have occurred through dynamics at the
> species level (barring mutations, all my cells that have a genome have
> the same genome, and this is not due to selection against genetic
> chimeras in the past; for some individuals are chimeras).
> >
> > If that is what you are saying (approximately), then all I can do is to
> > point out (again) that the species level is special in that there are no
> > disruptive gene flows to disrupt the frequency- dependent equilibrium
> > which I take to be crucial to a polymorphism being heritable.  But,
since
> > you have heard that before, I suppose we will just have to agree to
> > disagree.  Unless, that is, I have completely misinterpreted your
> > position.
>
> Some species are cohesive gene pools. Many are not. It follows that the
> category "species" itself is not definable in terms of cohesive gene
> flows. Hence the idea that a species exists because there has been some
> selection for species-wide traits in the past is less secure; you still
> need to show that enough species are cohesive. I don't think that the
> evidence supports this.
>
> Perhaps I had better say what I think species *are* then. In my view, a
> species is an effect, not a cause. It is real enough, and in the case of
> sexual species reproductive isolation keeps it real (and novel isolation
> makes more species), but isolation is a property of a relation between
> two breeding individuals, and additively of populations. Hence, unless
> the individual effect here (groups isolated by various mechanisms)
> results in a suitably cohesive entity (which it mostly doesn't) species
> cannot act as "players" in any process that cannot be eliminated in
> favour of a more causal story at a lower level.
>
> I do not deny that [some] species *may* be subjected to a selection
> process at that level. I merely deny that anyone has given good reason
> yet to show they have. The problem is threefold:
>
> 1. To show that species are cohesive enough to "act" as an individual
> (the precondition for entering into a selection process);

Cohesion may be partly "in the eye of the beholder".  I think that species
are cohesive.  You do not.  I suspect that the difference between us
is in the time scale contemplated.  You are thinking of short-term
cohesion.  I am thinking of the long term.

I now describe why the time scale makes a difference.  In discussions
of speciation, one sometimes draws a line on a map and asks
how much gene flow crosses that line.  The rule of thumb is that if
there much more than one genome of cross-flow per generation,
then there is no tendency to speciate.  But, if there is much less than
that, speciation is likely.  Therefore, we can safely assume that any
species will have at least a cohesion of one genome per generation.

You would argue, correctly I think, that one genome per generation
is just not enough cohesion to make the species a pop-gen individual.
The individuality lies at lower demic levels.  After all, the pop gen
effect of a single genome is just noise in a large demic population.
How much gene flow is needed between two demes before we can
say that they cohere into a single deme?  Intuition says that the flow
has to be at least SQRT(N) per generation.

But now ask: "Why are we talking about genomes per _generation_?
Why not genomes per millenium, or some other time scale?"  Well,
for selection at the level of individual organisms, the generation, which
is roughly the life span of an organism, is clearly the right time unit.
But for selection at the level of a species, shouldn't we look for a
time scale comparable to the life span of a species, or perhaps to
the inverse of the species birth rate?  I believe that we should, and
on this time scale, the flow quantities are greater than SQRT(N) for
most species.  Species ARE cohesive on a species-selection time
scale.

On the other hand, lower level demes, while clearly coherent, are
just not long-lived enough to act as individuals in the kind of
selection I am talking about.  They are momentary clusterings
(comparable to the transient micro-crystaline arrays of our
"water of the gods" dialog) that flicker in and out of existence.

> 2. To show that those that are have, in fact, properties of a causal
> nature at that level (the precondition for *actually* entering into a
> selection process); and

Polymorphisms, such as {male,female} or {A, B, AB, O} clearly
exist at demic levels (not in any sense at the individual organism
level) and are frequently uniform across demes.  Hence, to my
way of thinking, they qualify as species-level properties.  Whether
they are causal or not is a question that must be addressed on a
case-by-case basis.  I hope that I have at least made it plausible
that they are frequently causal.

Once the existence of species-level selection has been established,
I think it might be beneficial to extend the concept to cover some
traits that could alternatively be thought of as individual level traits.
That is, I am not swearing undying loyalty to your epistemological
preferences, but am willing to consider the arguments of Tyler and
McGinn.  In this, I follow the lead of Hamilton, who did not originally
attribute parental care to kin selection (since, as Edser points out,
it can also be explained by non-inclusive fitness selection if you
define reproductive success as offspring raised to maturity).  But
once the concept of kin selection is established, parental care
is more naturally handled as kin selection rather than as individual
selection.

What individual-level traits do I think might be more profitably be
treated as species traits?  Well, I think that any trait that is fixed -
that no longer has any variation within the species - is a likely
candidate.  One example might be the dietary requirement for
vitamin C of most primate species.  In fact, I would suggest that
whenever the loss of a structural gene becomes fixed within a
species, then that trait should be "promoted" from the organism
level to the species level.  Notice that this "promotion" wouldn't
make sense at lower demic levels, because the fixation of the
trait may be a transient phenomenon due to potential gene flow
from elsewhere in the species.  Also notice that "promotion"
would be inappropriate for fixed dominant genes, because these
fixations are also transient due to mutation.

> 3. To show that enough of these beefed up species interact in ways that
> causes a selection process to result (showing that they *do* enter into
> a selection process).
>
> I will grant 1 and 2 in some isolated cases - although so far they are
> all at the demic level. But I haven't seen reason to think that 3
> obtains or that there is even an epistemic advantage to think it might
> obtain.

I am not certain that I understand you here, but I suspect that you
are trying to carry too much connotation with you in transferring
the word "selection" from the individual level to the species level.
Are you suggesting that in order to have "selection", you have to
identify a "population"?  I disagree, or at least deny that the process
that I have been calling "species-level selection" involves a population
of species any smaller that the set of all species on earth.

The slogan of neo-Darwinism is sometimes stated as "Genes mutate,
organisms are selected, species evolve".  But in the process of
species-level selection that I have been defending, there is no
coherent individual that evolves.  Unless you want to say that the
biosphere (Gaia?) evolves...
---
þ RIMEGate(tm)/RGXPost V1.14 at BBSWORLD * Info{at}bbsworld.com

---
 * RIMEGate(tm)V10.2áÿ* RelayNet(tm) NNTP Gateway * MoonDog BBS
 * RgateImp.MoonDog.BBS at 5/14/04 5:37:50 PM