Re: macroevolution or macromutations?

From: Richard Wein (rwein@lineone.net)
Date: Fri Jun 16 2000 - 11:29:17 EDT

  • Next message: Richard Wein: "Re: macroevolution or macromutations?"

    From: Cliff Lundberg <cliff@cab.com>

    >Some of the advantages are:
    >--The Cambrian explosion is explained; a mechanism for rapid formation
    >of complex vertebrates and arthropods is outlined.

    But that begs the question (of whether your mechanism is an improvement on
    the conventional one). I see no reason why evolution by your mechanism
    should be any more rapid.

    >--The nature of post-Cambrian evolution (reduction and specialization
    >of segments, no new body plans, no rapid evolution) is explained.

    But this description of post-Cambrian evolution is debatable. And, as I've
    said before, even if it's true, it says nothing about the mechanism of
    Cambrian/pre-Cambrian evolution, which, according to you, is where all the
    novel parts arose.

    >--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.

    Again, this begs the question of whether your explanations are any better
    than the conventional ones. I think they're a lot worse!

    >There are lots of criteria by which an argument can be judged, but I don't
    see
    >why being ad hoc is a problem.

    Because *anything* can be explained by a sufficiently ad hoc explanation, no
    matter how wrong it is. But I suppose this is of a piece with your rejection
    of parsimony as a criterion in model selection. In rejecting parsimony as a
    relevant criterion, you're rejecting the prevailing scientific method.

    >You have not furnished counter-examples to the general claim that new
    >segments have not been added since the Cambrian.

    You yourself provided the counter-example of additional segments in snakes.
    I provided the examples of extra legs in fruit flies and extra toes in
    mammals. You rejected the latter as being the reactivation of atavistic
    genes, but offered no evidence to support that argument. I don't rule out
    the possibility that you're right about this. I just see no good reason to
    accept it in the absence of evidence.

    >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.

    I have no problem with reduction and specialization, I just don't think
    they're the *only* processes operating. Does Gould agree with your claim
    that new segments have not been added since the Cambrian? (Even if he does,
    that doesn't necessarily make it so, but I will be more
    impressed.)

    >>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.

    I don't say it can do *anything*. But duplication of an existing part is not
    far-fetched. And why should such a mutation be restricted by the number of
    segments that have been present in the past?

    >>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.

    First of all, is such a heritable Siamese-twinning ever seen in humans? If
    not, your argument that this mechanism is observable in our own species is
    invalid.

    Reading your article again, I see you say : "Some define "parabiosis" as the
    artificial joining of embryos; herein the term refers to embryos naturally
    conjoined due to a heritable mutation." But you fail to mention what sort of
    mutation you have in mind. (BTW this lack of discussion of genetic
    mechanisms is noticeable throughout your article.)

    It seems there are two main possibilities, which I'll call MC (multiple
    copies) and SCCC (single copy + control code). MC involves duplicating the
    genome for an entire segment, so an N-segment organism will have N copies of
    the genome. SCCC means having only one copy of the genome for a segment,
    plus some control code to cause the genome to be expressed N times.

    MC is simpler in the short run, but you can't get very far like that. On
    your page "Vertebrate and Arthropod Progenitors; Their Form and Origin", you
    write: "The segmented biserial radials come to serve as rowing organs; some
    of them, in some lineages, evolve into legs and pincers." SCCC is essential
    by this time, otherwise there could be no symmetry of limbs. Symmetry under
    MC would require identical independent mutations in each limb. Given the
    obvious difficulty of switching from MC to SCCC once you get started, I
    would suggest that vertebrate evolution had to await the rise of an SCCC
    lineage.

    Of course, addition of extra segments under SCCC is evolution by the
    "Dawkinsian" model! So I think the real difference between us is not over
    whether the Dawkinsian model is ever valid, but over how complex an organism
    can grow and still evolve by this model. Of course, it's easier to accept
    that this can work a for simple organism than for more complex ones, which
    would explain why it's so rare in the post-Cambrian. But you haven't given
    any grounds for establishing any particular complexity barrier beyond which
    it can't operate. On the other hand, restricting it to only the simplest
    organisms, as you do, leads you into other problems...

    - You show a hypothetical step-by-step formation of the protovertebrate.
    But it's not clear that the intermediate steps have any selective advantage.
    I can accept that size is an advantage. But why do your three-pronged forms
    have an advantage over simple linear or four-pronged forms? What advantage
    does a crescent with a little bar in the middle have over other
    three-pronged forms? (In fact, it's almost a reversion to a linear form.)
    What advantage does the final form in Figure 1 gain from having some
    of its radials curved forwards?

    - You write: "Circulatory and nervous systems fuse and communicate." This
    seems like a major, and difficult, step, which is unnecessary under the
    Dawkinsian model. You also have other organs fusing to form, for example,
    the multi-chambered heart. But fusing of organs in this way seems to me more
    problematic than the Dawkinsian duplication that you reject.

    - You then proceed to duplicate the entire organism to produce some
    proto-limbs (Fig. 2). Again, what selective advantage does this new
    variation have? There is no musculature in place yet to operate these limbs.
    Neither is there a central nervous system to control them. The circulatory
    and nervous systems of the limbs have not yet fused with those of the main
    body. (If you propose that the nervous systems are fused by the same
    mutation that created the limbs, then this is a complex Dawkinsian operation
    of the sort you reject. In fact it's worse, because Dawkinsian operations
    generally add parts in the same relation to existing parts, such as an extra
    vertebra, while you're adding parts in a completely new relation.)

    Sorry, but it seems to me that you're just sticking parts together to get
    the shape you're aiming for, without thinking about the genetic mechanisms
    and selective pressures that are required.

    >>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.

    What do you mean by "gradually elaborated"? I doubt that you could find any
    reputable evolutionary biologist who believes that individual vertebrae
    originated independently, rather than by duplication.

    >>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.

    A basic striped pattern isn't complex. If you imagine material
    being continuously extruded from a machine, all that's needed to give it a
    striped pattern is for the addition of pigment to the material to be
    switched on and off at regular intervals. Now, of course, an animal's skin
    doesn't grow in this way, and I don't suggest it would be quite a simple as
    this, but the analogy shows that a striped pattern is not complex in
    principle.

    It seems to me that much of your problem with the conventional view is that
    you're thinking almost entirely in terms of gross morphology, and not
    sufficiently in terms of the genotype. The genome almost certainly doesn't
    contain a blueprint for a striped pattern, or even instructions for coding
    each stripe individually. It's more likely that there's a small amount of
    code that sets up something like my pigment switch.

    >Really, the tie-in to the skeleton makes much more sense.

    Your supposed tie-in is extremely poor. If the stripes were originally tied
    to skeletal segements, how did they become untied in places and migrate to
    locations where they don't correspond to skeletal segments?

    And how could the skeletal structure be reflected in the skin? What is the
    physiological connection between skeletal segments and skin locations? (And

    bear in mind that we're talking about ancient skeletal segments, not
    present-day ones. I assume your explanation of multiple stripes per leg bone
    is based on the assumption that the present day leg bones have fused from
    mutiple smaller segments. By the way, since you like to use ontogeny to
    support your arguments, does embryology show any sign of
    fusing of limb bones from smaller segments?)

    >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.

    Unfortunately, I don't have access to the sources you cited. Could you
    briefly summarize the relevant points, please?

    Richard Wein (Tich)



    This archive was generated by hypermail 2b29 : Fri Jun 16 2000 - 11:27:01 EDT