On Mon, 8 Mar 2004 04:12:39 +0000 (UTC),
William Morse <
[email protected]> wrote:
>
[email protected] (Larry Moran) wrote in
>
news:[email protected]:
[snip]
>> How does one distinguish between a just-so story and a
>> valid explanation? If it turned out that Asian elephants
>> had big ears and African elephants had small ones do you
>> think that someone would come up with an adaptionist
>> explanation? I do. That's the problem. With sufficient
>> imagination one can conjure up an adaptionist explanation
>> for almost anything that happens in biology. This isn't a
>> very good reason for preferring adaptionist explanations
>> over chance and accident - expecially since we know for a
>> fact that chance plays the major role at the molecular
>> level. Why shouldn't it play a major role at the
>> morphological level as well?
>
> How major a role?
If we compare homologous genes from different species the
vast majority of differences appear to be neutral with
respect to fitness. Thus, it appears that random genetic
drift is responsible for most of evolution when measured by
actual changes in the fixation of new alleles. This doesn't
take into account the role of negative selection in
eliminating detrimental alleles but that's not what we're
discussing. One question is whether positive natural
selection accounts for a significant proportion of change at
the molecular level. This doesn't seem likely.
What about the mechanism of evolution for alleles that have
an obvious phenotype? Is it only natural selection or drift
involved? I don't know what you might consider a "major"
role. Right now I'd be happy just to see adaptionist admit
to the possibility that some morphological traits can be
fixed by accident.
> Darwin's logic is fairly incontrovertible here - if
> organisms produce more offspring than can survive (which
> they do), and if there is heritable variation between
> those offspring (which there is) then _if_ that variation
> results in a differential fitness (which appears to be the
> crux of our disagreement) there _will be selection_ for
> the morphological trait.
Yes, this is the crux of the disagreement. The question is
whether a given morphological trait is associated with a
difference in fitness. I argue that this should not be
automatically assumed. It is equally likely that (many)
morphological traits are effectively neutral.
> Drift plays the major role at the molecular level because
> most changes at that level _are_ selectively neutral. I
> can change an alanine for a valine far from a binding site
> and selection won't (at first) see it. But AFAIK drift
> hasn't made much of a dent in the binding site of
> hemoglobin, to give only one obvious example.
Nobody is arguing that positive natural selection doesn't
exist at the molecular level. However, I hope we can agree
that most change in the sequences of genes is not due to
adaptation. The question is whether this observation should
apply to nucleotide substitutions that give rise to visible
phenotypes or whether it is confined to substitutions that
don't have any obviously visible phenotype.
> Now while it is true that adaptationists are prone to
> conjuring up explanations for observed morphological
> traits, that doesn't make those explanations wrong, it
> just makes them untested.
Correct, they aren't "wrong." However, a problem arises when
adaptionists completely ignore the possibility that the
morphological trait may not be due to adaptation at all.
There seems to be an implicit assumption that all
morphological traits must be explained by adaptation. In
fact, most adaptionists seem to think of "evolution" as a
synonym for "natural selection." I argue that this bias may
prevent adaptionists from considering alternative
explanations and this isn't a good thing.
> Ideally one would like to alter the ear genes for African
> elephants, stick a bunch of small-eared ones in an
> otherwise identical savanna environment, and see if they
> survive as well as big eared ones, but even if we could
> perform the experiment we would have to wait 20000 years
> for results, and even postdocs might get a little antsy
>
Luckily for us mother nature has performed some of
> these experiments for us. Pangolins, aardvarks, anteaters,
> and echidnas are all mammals but otherwise quite unrelated
> - except that they share the same ecological niche and
> share numerous morphological traits. Take a good look
> sometime at pictures of the heads of deer and kangaroos,
> which also only distantly share a common ancestor, and use
> quite different mechanisms of locomotion (score one for
> contingency!), but share a similar niche. A zoologist can
> look at the teeth of an animal and tell you what it eats,
> can look at its skin and tell you what climate it
> inhabits, can look at its eyes and tell you whether it is
> predator or prey. So how are these morphological traits
> _not_ adaptations?
Everyone agrees that there are morphological and molecular
traits that are adaptive. Just because these exist is no
reason to assume that ALL morphological and molecular
changes are adaptive.
> The argument for drift is a much tougher row to hoe.
> You've got all the tools of molecular biology that prove
> that drift occurs, but on a morphological basis how do you
> prove a negative? And even if you do, where is the glory?
> The adaptationist gets to come up with a great (if
> unprovable) story, while the stochasticist is left with
> endless strings of purines and pyrimidenes which mostly
> cannot (yet) be linked to morphology and behavior.
I agree that just-so stories are seductive. Does that mean
they are good science?
> (snip intro to this)
>
>> So, is it possible that variation in skin color is due to
>> selection in some parts of the world (Northern Europe)
>> but drift elsewhere?
>
> Certainly (although my argument is that "some parts of the
> world" includes all of Europe, Africa, Asia and Australia)
> . Drift happens. Selection happens. Calamity (aka ****)
> happens
If Arlin Stoltzfus were still contributing to
> sbe he would be telling us that mutation pressure happens.
> Evolution doesn't care which of the above is leading to
> changes in gene frequencies. What I was trying to define
> with my "default assumptions" was traits that were
> _primarily_ shaped by one or the other
> - but there are no magic force fields that protect
> reproducing organisms from any of the above effects.
Why does there have to be a "default assumption" when we
know for a fact that both random genetic drift and natural
selection play important roles in evolution?
> (snip)
>
>> (A minor quibble ... natural selection also has a major
>> stochastic component. Most beneficial mutations are
>> eliminated from the population. The best you can do is
>> calulate the "probability" that a given allele will
>> become fixed if you know the selection coefficient.)
>
> Agreed. One of the reasons that zebra mussels have been
> able to spread so rapidly in North American fresh waters
> is their ability to fix to a substrate. This ability is
> common in mussels throughout the world, but there are no
> native North American freshwater mussels that can do this.
> Even I have difficulty in coming up with a Just So story
> to explain this
> - it may well be due to "bad luck".
I'm glad we can agree on something ...
Larry Moran