Richard Wein wrote:
>Well, I'm not dissatisfied with the current model, but I also don't think
>that the current model is the final word. The question is whether your
>proposed model (or any other proposed model) is an improvement on the
>current model. I don't really see that your model offers any significant
>advantages over the current model, and I see some major disadvantages.
Some of the advantages are:
--The Cambrian explosion is explained; a mechanism for rapid formation
of complex vertebrates and arthropods is outlined.
--The nature of post-Cambrian evolution (reduction and specialization
of segments, no new body plans, no rapid evolution) is explained.
--The evolution of the vertebrate form and its components is explained.
--Solutions for specific problems follow from the model, ranging from
lines of flexure in embryonic skin to zebra stripes to the evolution of jaws.
>>I don't disagree that homologous segments or Siamese twins are formed
>>originally from the same genetic information, whether from the very same
>>sequence or from copies of that sequence. I do disagree about whether a
>>control gene can create brand-new segments; the evidence implies that
>>brand-new segments have not been formed during post-Cambrian evolution.
>
>That remains to be seen, since you dismiss evidence to the contrary by ad
>hoc arguments (as already discussed).
There are lots of criteria by which an argument can be judged, but I don't see
why being ad hoc is a problem.
You have not furnished counter-examples to the general claim that new
segments have not been added since the Cambrian. I remind you that this
is not my invention; the principle of reduction and specialization within
sets of homologous structures is an old one. Gould has commented on
it without contesting its truth.
>But let's suppose you're correct that no new segments have been formed
>during post-Cambrian evolution. The question still remains: why do you
>prefer Siamese twinning to duplication of segments by the Dawkinsian model
>for *pre-Cambrian* evolution?
It's just logically and mechanically much more feasible. Siamese-twinning
is understandable, demonstrable, observable, even in our own species.
We can see how it happens. But a mutation that inserts a new segment
within, e.g., the vertebral column, producing a higher number of segments
than had ever been known in that species' range of variation, that seems
a hard thing to do. Unless you invoke the deus ex machina of a "control
gene" that can do anything.
>But Siamese twinning isn't heritable (is it?). Genetic duplication (by the
>Dawkinsian model) is a better explanation precisely because it's heritable.
>Perhaps it's much less likely to occur in the first place, but, once it does
>occur, the change will be retained.
In my scheme Siamese-twinning is heritable.
>Evolutionary biologists today no longer propose the kind of gradualism that
>you keep attacking (if they ever did).
AFAIK it is generally believed that segmented organisms were and can be
gradually elaborated through evolution, despite the establishment of the
principle of reduction and distortion of homologs in paleontology.
>it's accepted that small changes in
>genotype can result in large changes in phenotype. I don't think that random
>blotches gradually became more and more linear and parallel until they
>formed a striped pattern. I suspect that the first, crude striped pattern
>resulted from a single mutation in a control gene.
I suppose anything is possible for a "control gene", but even a crude striped
pattern seems a pretty complex (and symmetrical) thing for random mutation
to produce. Really, the tie-in to the skeleton makes much more sense. As you
may recall from the article, this is not my idea, I cite the articles in which
this
idea is argued for. I only bring it up as it supports my 'skeletal archetype'
theory IMO.
--Cliff Lundberg ~ San Francisco ~ 415-648-0208 ~ cliff@cab.com
This archive was generated by hypermail 2b29 : Fri Jun 16 2000 - 04:05:54 EDT