Re: Origin of feathers

["John V Jackson" <jjackson@interalpha.co.uk>]

-----Original Message----Date: 09 April 1998 08:27


>In a message dated 98-04-08 23:45:33 EDT, swo@execpc.com writes:
>
><< The "behavior evolves first" hypothesis is one of the main tenets of
> modern zoology. And it is not always untestable! In this case, yes, I
> agree, it is hard to test. >>
>
>To confirm this hypothesis in any particular case, you would have to be
able
>to observe the behavior prior to the appearance of the feature in question,
>follow the evolution of the population through the appearance of the
feature,
>and then observe the behavior after the appearance of the feature.

Any feature, behavioural or physical, will only be selected for if it
confers an advantage.  Webless birds that find themselves on water often
will be more likely to multiply any "web" gene in the pool :-) than
waterless birds.  The important thing is that it can go both ways - each can
precede and facilitate the other.  Basic genetic evolution is so blind it
can't even tell the difference between genes for behaviour and body
structure (at least in terms of selective advantage).

>I can certainly imagine circumstances in which the evolution of a feature
would
>evoke a behavior that was absent prior to the feature's appearance.

So blind that "evoke" would have to be an indirect description of the
process.

Of course, a *need* or use for a feature has no direct connection with its
appearance.  So if animals needed to get rid of sulphur they would just have
to wait for the mutation, genetic rearrangement whatever, to appear, just as
if they had a use for skin friction, an important component of drag, they
would have to wait for increased roughness of the body surface.  (Every now
and then of course, evolution throws up a jackpot, of the order of hairs
becoming branched.)

They've done some staggeringly subtle experiments both observed  and
controlled, in the evolution of lizards and butterflies over the years.  I'm
starting to believe generalisation of evolutionary principles from insects
to dinosaurs can be valid, but you've got to be very careful.  If we accept
that, then perhaps Hinton's significant result in the field of neural nets
may be admissible.   He demonstrated or at least exemplified, that evolution
of flexible learning behaviour can speed overall evolution (even in other
aspects of the "creature"), since it allows more distant minima of the
optimisation surface to be explored (think of a creature's genetic state as
being a point on a hilly landscape, and its evolution as the rolling of a
ball down to the lowest valley).  For this to have any meaning, behaviour
would have to precede body form.

>In constructing an evolutionary scenario (since we cannot go back in time
and
>observe populations of organisms evolve), it is necessary to provide a
>plausible, compelling cause for each step in the scenario. "For display" or
>(even more remote) "for flight" are simply not compelling reasons for the
>initial appearance of pre-feathers, because they stem from characteristics
of
>the behavior of the organism that can only be known after the evolution of
>pre-feathers, not before. You are using the future to compel the past. But
if
>you say that pre-feathers evolved as a mechanism for sulfur excretion, you
>have provided a compelling reason for their appearance. The cause--the
>reason
>for the initial appearance< of the pre-feathers--here lies >before< their
>appearance, not after.

A need is never an initial cause in evolution.  And every body surface has
drag of some kind, and there will always be some variation of it within the
population.  There is not however always some background element of
solid-form sulphur excretion through the skin, and no variability to drive
any selection for it.

>Another example: Say that a lineage of small animals, previously accustomed
to
>a ground-dwelling lifestyle, becomes arboreal. Two causes are postulated:
(1)
>the animals went into the trees to escape ground-dwelling predators; (2)
the
>animals went into the trees to find food. Which of these is the more
>compelling? (Here I'm not concerned with the >real< reason, just in
choosing
>between the two proferred reasons.) I assert that (1) is more compelling,
>because the danger from ground-dwelling predators exists prior to the
arboreal
>lifestyle, whereas the availability of food in the trees cannot be known to
>the animals until they go up there and find it. Hypothesis (2) is an
example
>of the future compelling the past, and should be rejected.

If only humans were as sensible as your small animals, then perhaps I
wouldn't spend so much time fruitlessly barking up the dinosaur tree!
Animals look anywhere they can get to - they don't all say "it's a tree so
there's nothing up there".  In any case, since trees, being organic, hold
more insects per unit area than bare ground, they would already be very
likely to be going up them anyway.  It's the jumping out of trees that's
more likely to be caused by escaping predators than seeking food.

>Once the animals get into the trees, perhaps the presence of a good food
>supply helps compel them to stay there, but it could not have gotten them
up
>there in the first place.


They weren't subtle enough to make that initial mistake.


John V Jackson    jjackson@interalpha.co.uk

(Wannabeasaurus   beecee-effia - & proud of it!)