Control and Reorganization in Evolution

[From Rick Marken (940913.2250)]

Martin Taylor (940912 12:00) --

In a stable population, that one person [with the "good" mutation]
will have an expected number of surviving offspring in the next
generation greater than one.

Maybe. Maybe not. Shit happens, even to good mutations and their
offspring. Besides, nobody said natural selection couldn't work. It just
works very much more slowly than "E. coli" evolution to produce the
kind of adaptation that requires mutation-based change.

I think the fast changes produced by Darwin's Hammer type selection
(like the beak of the finch or the color of the moth) are _not_ based on
mutation; they are based on "selection" from already existing
variations that make it possible for the population to deal with certain
kinds of environmental change. Finches, for example, already come equipped
with various sized beaks; catastrophic drought drastically changes the
distribution of beak lengths in the population -- but it doesn't produce
a new kind of finch. I bet that after a long spell of no drought the
proportion of long and short beaked finches evens out again. Same
with moths; get rid of the birds and the distribution of light and dark
moths will even out regardless of the background color. These kinds of
changes are like distrubance resistance in a control system -- but
operating over a population rather than an individual. The range of
phenotypes on a certain dimension (like beak size) probably represents
the evolutionarily developed output variations (over the population
of individuals) that allow control in terms of survival of some
proportion of the individuals and, hence, resistance to disturbances to
crucial variables (droughts that disturb the type of food available to the
finches or changes in background color that disturb the contrast
between moth and background). Disturbances to these variables were
probably confronted during the evolutionary history of the population
and output (phenotypic) variations on certain dimensions evolved to
handle them.

Reorganization-type evolution (mutation-based ) probably occurs when
there are environmental changes (disturbances) that the population
does not have variations to deal with. Such disturbances cannot be delt
with by a redistribution of phenotypes in the population. It can only be
delt with by developing a new dimension of variation in the
population. This can only be accomplished by mutation. I imagine that
the population would be decimated if it had to wait for the "right"
mutation to occur. So "controlled" evolution is probably what goes on
in mutation-based evolution -- leading to the pronounced "punctuations"
in the fossil record. But, I agree, some simulations are necessary in order
to see what might be going on.

The soldier who dies in a successful defence of his tribe probably has
done more to ensure that his genes propagate than he would have by
impregnating every woman in the village.

Well, I still say "Make love, not war".

Peace

Rick

[Martin Taylor 940914 16:50]

Rick Marken (940913.2250)

(Preamble: I'm away from this evening for about 10 days more or less,
so expect from me no responses to postings over that period. Back for
sure Sept 26 and maybe a few days earlier.)

In a stable population, that one person [with the "good" mutation]
will have an expected number of surviving offspring in the next
generation greater than one.

Maybe. Maybe not. Shit happens, even to good mutations and their
offspring.

Remember what "expected number" means. You are absolutely right, except
for the "maybe not." I dealt with the "shit happens" bit numerically
in the posting to which I think you are responding. "Expected number"
means not that any individual with the good mutation will have that
number, but that over a large enough population with the same good
mutation the total number of offspring will be close to the expected
number times the number of potential progenitors in the population.

Besides, nobody said natural selection couldn't work. It just
works very much more slowly than "E. coli" evolution to produce the
kind of adaptation that requires mutation-based change.

Nobody said it didn't. The question is whether e-coli type evolution
actually exists. Maybe, maybe not. In an earlier posting today I suggested
that we might test whether the e-coli version would really work faster.
I think it would, but I'd like to be reassured. If it really does, then
it would be plausible that natural selection had evolved methods for
directed evolution or e-coli type rate-selected evolution. If it
doesn't then we could stop worrying about it.

I think the fast changes produced by Darwin's Hammer type selection
(like the beak of the finch or the color of the moth) are _not_ based on
mutation; they are based on "selection" from already existing
variations that make it possible for the population to deal with certain
kinds of environmental change.

Nobody that I know of disagrees with this, either. All the GA experiments
that I know of demonstrate that the main role of mutation is to ensure
that there remains some variation in a population, not to enhance the
immediate fitness. For that, recombination is much faster, combined with,
I suppose, direct selection among the existing variety, as with the moth
colour and the finch beak.

I bet that after a long spell of no drought the
proportion of long and short beaked finches evens out again. Same
with moths; get rid of the birds and the distribution of light and dark
moths will even out regardless of the background color.

Quite so. There would presumably be random drift in the relative numbers
of light and dark, or of long and short beaks.

Reorganization-type evolution (mutation-based ) probably occurs when
there are environmental changes (disturbances) that the population
does not have variations to deal with.

Actually, it happens all the time. We are all subject to mutations, many
of them. Some theories of cancer have to do with cumulative mutations
in some somatic cells and their offspring within one organism. Sex cells
are not immune, but most mutations lead to non-viable eggs or sperm, or
to non-viable embryos. Only a very small proportion are not lethal, and
fewer are beneficial in an environment to which the parents are well
adapted. I'm not sure what your point is. Is it that the effects of
mutation are noted in a species only when the environment gets too difficult
for the various recombinations to produce viable offspring? That's
probably an exaggeration with a germ of truth.

So "controlled" evolution is probably what goes on
in mutation-based evolution -- leading to the pronounced "punctuations"
in the fossil record. But, I agree, some simulations are necessary in order
to see what might be going on.

As you know, I did a truly trivial simulation in HyperCard (which I can
send you). I assumed a population subjected to one pressure--resource
limitation that limited the total number of organisms in the population
to be the same in each generation. I used a constant mutation rate, which
varied in different runs from .1 (i.e. 1/100 of the organisms experienced
a mutation in any generation) to .001, and a probability that a mutation
was "good" that varied from .25 to .001 (any mutation that wasn't "good"
was "bad"). A good mutation was defined by increasing by 5% the probability
that the entity would have offspring (all offspring were identical to the
progenitor after any mutation in the progenitor). A bad mutation was
defined by decreasing the offspring probability by 5%. That's it.

The result of this "experiment" did not depend on any of the probabilities,
except as to the rate at which things happened. The population, after
a few generations, almost always consisted of 80-85% of one genotype,
with a few "better" and a few "worse". The same genotype would be dominant
for a few or for many generations unpredictably, and then a "better"
genotype would become dominant in very few generations, so that the
graph looked almost like a step function. There was never a regression
to a situation in which a worse genotype retook dominance. Punctate
evolution with a vengeance, and with an absolutely constant mutation rate.

"Controlled" evolution might well lead to "punctate" evolution, and even
without controlled evolution, shifts in environmental stress might well
lead to more rapid evolution, but neither are required for punctate
evolution to be evident. Steady mutation in a constant environment
will do it.

The soldier who dies in a successful defence of his tribe probably has
done more to ensure that his genes propagate than he would have by
impregnating every woman in the village.

Well, I still say "Make love, not war".

Me, too. But what's good for the individual isn't necessarily best for
his jeans.

See you in 10 days or so.

Martin

Tom Bourbon [940915.0645]

[From Rick Marken (940913.2250)]

. . .

I think the fast changes produced by Darwin's Hammer type selection
(like the beak of the finch or the color of the moth) are _not_ based on
mutation; they are based on "selection" from already existing
variations that make it possible for the population to deal with certain
kinds of environmental change. Finches, for example, already come equipped
with various sized beaks; catastrophic drought drastically changes the
distribution of beak lengths in the population -- but it doesn't produce
a new kind of finch. I bet that after a long spell of no drought the
proportion of long and short beaked finches evens out again. Same
with moths; get rid of the birds and the distribution of light and dark
moths will even out regardless of the background color.

Right on, Rick. In fact, I recall that as far back as the 60s people were
raising that same point anout the case of the moths. In the examples they
described, though, the full range of variation became evident when the
amount of heavy particulate pollution in the region declined. I know
of similar results from research on the variety of shades and patterns of
coloration within a species of rattlesnake here in Texas. (The coloration
in adult specimens "matches" the conditions and appearance of the soil in
their particular locations in the state; mismatches stand out and more often
become prey.)

Selection from a diverse population probably accounts for quite a few
instances of "rapid evolution."

Later,

Tom