Bell Curve's Fatal Flaw
Joseph J. Schall
CurveS are interesting, but seem to have missed a critical error in the
arguments used in this recent book, as well as others on the same topic.
This error deals with heritability and its meaning. My thought suggested
thought experiments are based on Richard LewontinUs brilliant essays on
this topic.
Heritability is a measure of the fraction of phenotypic variation that
can be accounted for by genetic variation. In the case of IQ, we can ask
what fraction of the VARIATION in IQ can be accounted for by GENETIC
VARIATION in a population. Thus if heritability (h2) for IQ in a human
population is 0.80, this means that 80% of the variation in IQ scores is
accounted for by genetic variation....thus, people with a high IQ, on the
average are different genetically than individuals with low IQUs. h2
says NOTHING about individuals and CANNOT be used to compare groups.
Even if h2 for IQ score is 0.80 for both African-Americans and White
Americans, the two conclusions drawn in the Bell Curve do not follow.
First error: The 15 point difference between Blacks and Whites on IQ
tests is of genetic origin. Heritability deals only with VARIATION, not
with the placement of the mean. The following thought experiment shows
this. Suppose 100 people were chose at random from the New York
population. Their skin color would vary in darkness, in part because of
genetic variation in that population (h2 would be fairly high for skin
color). Then, spit the group into two groups of 50 randomly, send one
group to spend the winter on sunny Aruba island and the other group stays
in New York for the winter. Next spring bring the two groups back
together. The group that spend the winter in Aruba will be darker on the
average than the group in New York. They cannot be different genetically
because the two groups were chose randomly from the original 100 people.
The difference is of environmental origin, even though h2 for skin color
is high! This thought experiments shows the general case....heritability
tells us nothing about differences between groups.
Second error, related to the first: Will special educational programs
help move the IQ score of Black Americans if h2 for IQ is high? And,
suppose the IQ difference really IS of genetic origin?
Again, even if h2 is high for IQ score, and even if the difference IS of
genetic origin, educational and other programs ALMOST CERTAINLY will
cause changes in IQ scores. Again, h2 is a measure of variation, not the
mean. Consider this thought experiment: Bring 10 African pigmies to
your thought lab, and 10 people from Samoa. The first group will be
quite small, and lower in weight than the second group. Heritability for
weight is greater than zero for each group, and the two groups differ in
weight in part because of a genetic difference....pigmies are small and
Samoan are large in part because of their genetic nature.
Now, starve your Samoans (it is only a thought experiment!), and force
the pigmies to eat a very high caloric diet...lots of Ben and JerryUs.
Result: the Samoans will start to shrink, and the pigmies will start to
swell. After a time, the pigmies might on the average even be heavier
than the Samoans.
Again, this shows that the mean....the placement of the distribution of
any trait.....can change even if the trait has a high heritability. And,
the mean for a trait can change in two groups, even if the two groups
differ genetically for that trait, if the environment differs.
My argument here is NOT that IQ measures real intelligence, or that IQ
really has a h2 that is high, or that the races really differ genetically
for this trait. My point here is that the arguments used in the Bell
Curve suffer a fatal flaw...not a matter of opinion, but a matter of how
the arithmetic is done.
These problems were pointed out many years ago by Michael Lerner who
worked out many of the details of measuring heritability and by Lewontin
who is one of the great geneticists of our time. EVERY quantitative
geneticist knows of this error. The mystery to me is why NO journalist
or reviewer seems to have found this flaw in the argument.
There is no way to measure the relative importance of genes and
environment on individuals, it must be done with populations. And, the
only way known now to do this is by calculating heritability. The only
way known now to determine if two groups differ, is to raise them in a
Rcommon gardenS with identical environments. Thus, if some scholar
really cares if Blacks and Whites differ genetically for IQ (why would
anyone care anway!?), then we would have to be sure that Blacks and
Whites have similar environments for at least one generation. I donUt
think that is what the authors of the Bell Curve have in mind.
Bob Wheeler
> The discussions posted here concerning the recently published RBell
> CurveS are interesting, but seem to have missed a critical error in the
> arguments used in this recent book, as well as others on the same topic.
> This error deals with heritability and its meaning. My thought suggested
> thought experiments are based on Richard LewontinUs brilliant essays on
> this topic.
<snip>
>
> First error: The 15 point difference between Blacks and Whites on IQ
> tests is of genetic origin.
<snip>
> Second error, related to the first: Will special educational programs
> help move the IQ score of Black Americans if h2 for IQ is high? And,
> suppose the IQ difference really IS of genetic origin?
<snip>
These arguments are very interesting -- sorry to delete them to save
bandwidth. However, I can't seem to find were in The Bell Curve the
errors occur. Could you cite a page or section? Many thanks.
Bob Wheeler
Harry Erwin
known to be massive. This makes it likely that groups coming from
environments where IQ is depressed (by cultural factors, exposure to
lead, discrimination, etc.) are under more selective pressure for higher
inherent general intelligence that groups whose environments are more
benign. It would be interesting to investigate this.
--
Harry Erwin
Internet: her...@gmu.edu
Stephen Bogner
Not really my field. However...
The assumption that higher intelligence confers a "fitness" advantage wrt
selective pressure appears to me to be unwarranted. Intelligence will only
create an advantage if it contributes to reproductive success, and there would
appear to be very few environments where the differential in mortality between
the "smart" and the "stupid" is measurable, let alone decisive. On the other
hand, as has been pointed out by others, there does appear to be a negative
correlation between education level and number of offspring. Perhaps
intellegence is not such an advantage as has been supposed...
Regards, Steve.
.............................................................................
Stephen Bogner (DRES/DTD/MES/Vehicle Concepts Group) sbo...@dres.dnd.ca
(403) 544-4786 DRE Suffield; Box 4000; Medicine Hat, Alberta; Canada T1A 8K6
"Always leave your clothing and weapons where you can find them in the dark."
- from the notebooks of Lazarus Long
Paul S. Winalski
In article <395m2a$m...@portal.gmu.edu>,
her...@science.gmu.edu (Harry Erwin) writes:
On the other hand, the wards of the current US welfare programs are under no
selective pressure whatsoever for higher inherent general intelligence. You
don't have to think very much to sit around and collect welfare checks.
I believe this is one of the main points that The Bell Curve's authors are
trying to make: the economic structure of modern welfare programs is such that
they put barriers in the path of upward mobility out of poverty. The trend
is towards the splitting of society into two behaviorally isolated populations:
the workers, for whom high intelligence is becoming more and more of a
requirement, and a permanent welfare underclass, for whome intelligence is
irrelevant.
--PSW
Tim Benham
>The discussions posted here concerning the recently published RBell
>CurveS are interesting, but seem to have missed a critical error in the
>arguments used in this recent book, as well as others on the same topic.
>This error deals with heritability and its meaning. My thought suggested
>thought experiments are based on Richard LewontinUs brilliant essays on
>this topic.
>Heritability is a measure of the fraction of phenotypic variation that
>can be accounted for by genetic variation. In the case of IQ, we can ask
>what fraction of the VARIATION in IQ can be accounted for by GENETIC
>VARIATION in a population. Thus if heritability (h2) for IQ in a human
>population is 0.80, this means that 80% of the variation in IQ scores is
>accounted for by genetic variation....thus, people with a high IQ, on the
>average are different genetically than individuals with low IQUs. h2
>says NOTHING about individuals and CANNOT be used to compare groups.
You seem to contradict yourself here. If h2 tells us that individuals
with high IQ differ genetically from individuals with low IQ then it
is telling us something about individuals.
>Even if h2 for IQ score is 0.80 for both African-Americans and White
>Americans, the two conclusions drawn in the Bell Curve do not follow.
>First error: The 15 point difference between Blacks and Whites on IQ
>tests is of genetic origin. Heritability deals only with VARIATION, not
>with the placement of the mean. The following thought experiment shows
>this. Suppose 100 people were chose at random from the New York
>population. Their skin color would vary in darkness, in part because of
>genetic variation in that population (h2 would be fairly high for skin
>color). Then, spit the group into two groups of 50 randomly, send one
>group to spend the winter on sunny Aruba island and the other group stays
>in New York for the winter. Next spring bring the two groups back
>together. The group that spend the winter in Aruba will be darker on the
>average than the group in New York. They cannot be different genetically
>because the two groups were chose randomly from the original 100 people.
>The difference is of environmental origin, even though h2 for skin color
>is high! This thought experiments shows the general case....heritability
>tells us nothing about differences between groups.
I don't see that your thought experiment has proved anything. Obviously
one can alter skin color or IQ by manipulating the environment, so what?
If skin color in NY has a high h2 then the dark skinned subgroup probably
has more dark skin genes than the light skinned subgroup. By taking two
groups and deliberately manipulating them you've invalidated the conditions
under which h2 was computed.
Claiming that, because one can induce a skin color difference between
two groups where there was none, it follows that h2 is irrelevant to
differences in skin color between groups is absurd.
>Second error, related to the first: Will special educational programs
>help move the IQ score of Black Americans if h2 for IQ is high? And,
>suppose the IQ difference really IS of genetic origin?
>Again, even if h2 is high for IQ score, and even if the difference IS of
>genetic origin, educational and other programs ALMOST CERTAINLY will
>cause changes in IQ scores. Again, h2 is a measure of variation, not the
>mean. Consider this thought experiment: Bring 10 African pigmies to
>your thought lab, and 10 people from Samoa. The first group will be
>quite small, and lower in weight than the second group. Heritability for
>weight is greater than zero for each group,
It might be zero for both groups and the difference could be 100% genetic.
>and the two groups differ in
>weight in part because of a genetic difference....pigmies are small and
>Samoan are large in part because of their genetic nature.
>Now, starve your Samoans (it is only a thought experiment!), and force
>the pigmies to eat a very high caloric diet...lots of Ben and JerryUs.
>Result: the Samoans will start to shrink, and the pigmies will start to
>swell. After a time, the pigmies might on the average even be heavier
>than the Samoans.
>Again, this shows that the mean....the placement of the distribution of
>any trait.....can change even if the trait has a high heritability. And,
>the mean for a trait can change in two groups, even if the two groups
>differ genetically for that trait, if the environment differs.
No-one would deny this, but your example is inapt. Do think that
whites and asians should be subjected to the mental equivalent of
starvation to reduce the differences in average IQs? Given the
sort of access to food we associate with modern free societies
Samoans will always be enormously bigger than pygmies on average.
...
>There is no way to measure the relative importance of genes and
>environment on individuals, it must be done with populations. And, the
>only way known now to do this is by calculating heritability. The only
>way known now to determine if two groups differ, is to raise them in a
>Rcommon gardenS with identical environments.
Identical environments don't exist.
>Thus, if some scholar
>really cares if Blacks and Whites differ genetically for IQ (why would
>anyone care anway!?),
Because if it were true then differences in those outcomes which
correlate highly with IQ would be expected even in a context of
environmental equality. Differences in such outcomes would not be
prima facie evidence of discrimination or environmental inequality.
> then we would have to be sure that Blacks and
>Whites have similar environments for at least one generation. I donUt
>think that is what the authors of the Bell Curve have in mind.
--
People who like this sort of thing
will find this the sort of thing they like.
Tim J.Benham ben...@cs.adfa.oz.au
Paul S. Winalski
In article <Cyqxt...@edsdrd.eds.com>,
edsdrd.eds.com!m...@edsr.eds.com (Michael H Coffin) writes:
|>
|>Perhaps because they don't use these arguments. Having now trudged
|>carefully through all but the last few chapters of The Bell Curve, I
|>have not seen these arguments made anywhere. In fact, Murray and
|>Hernstein carefully point out these possible misinterpretations of
|>their arguments.
For those of us without the time to trudge just now, could you post a brief
summary of the argumentst that Murray and Hernstein *do* use?
--PSW
Harry Erwin
: Paul S. Winalski (wina...@gemcil.enet.dec.com) wrote:
: : On the other hand, the wards of the current US welfare programs are under no
: : selective pressure whatsoever for higher inherent general intelligence. You
: : don't have to think very much to sit around and collect welfare checks.
: On the other hand, the selective pressures toward higher intellegence
: (or any other trait) take an awfully long time to produce noticeable
: results. It takes *much* more than a generation or so (in a large population)
: to produce hereditary differences worth bothering about.
Response to Winalski: the stats seem to show that lower socioeconomic
status confers lower fitness. Hence, there would be a selective gradiant
to eliminate genetic features confering lower socioeconomic status.
Response to Taibi: actually about 50 generations for fixation if the
feature has significantly higher fitness. In homo, we're talking
1000-1500 years.
Robert Flory/Cec Haas-Flory
>Having now trudged carefully through all but the last few chapters of
The Bell Curve, I have not seen these arguments made anywhere. In fact,
Murray and Hernstein carefully point out these possible
misinterpretations of their arguments.
>
I have followed this thread with waning intrest. Finding myself totally
tangled up in the knot I have several thoughts:
1. Has anyone else really read the book?
2: Does this mean that more oriental angels can dance on a pinhead than
white or black ones?
3: What about us POOR whites, we rated below orientals didn't we? It
seems that everyone is discounting that difference. Why is one worth
concern and another not of concern?
Michael H Coffin
Paul S. Winalski <wina...@gemcil.enet.dec.com> wrote:
>
>For those of us without the time to trudge just now, could you post a brief
>summary of the argumentst that Murray and Hernstein *do* use?
>
I think you're better off reading the book than relying on a second-hand
version. But...
They make several arguments. One is
1) IQ is between 40 and 80 percent heritable, according to
many, many recent studies. This is now so well established
that experts hardly argue about it anymore. [I'm paraphrasing
Murray and Hernstein. I'm not an expert.]
2) Over the last century, the US has become highly efficient at
finding smart people, giving them a very good education, and
putting them in elite jobs. Whereas in 1894 you would find a
lot of very smart people making moderate wages and doing
mundane jobs, you now find that 90%+ of very smart people are
making high wages and working at intellectually challenging
jobs.
3. Most because of (2), exceptionally smart people now usually
live in neighborhoods with other exceptionally smart people, go
to work with other exceptionally smart people, and marry other
exceptionally smart people. This is leading to a two-tiered
society, and the process is likely to accelerate, given the
heritability of IQ. This is something we should be concerned
about.
Another argument goes something like
1. (above)
2. Many social problems and pathologies are strongly related to
IQ. In fact, studies seem to indicate that IQ is a better
predictor of `success' than anything else. Better, for
example, than parent's socio-economic status or race.
3. Current social programs are predicated on the notion that the
poor and disadvantaged differ from the rest of us only by
having less money money. The theory is that the unemployed can
simply be retrained as computer programmers and
bio-technicians. Murray and Hernstein suggest that most of the
unemployed are in fact incapable of doing jobs that require
high cognitive abilities, and we ought not to ignore that
fact in setting social policy.
Another argument goes something like this
1. The attempt to `equalize' things is not only futile, but
actually harmful. For example, University affirmative action
programs give African Americans a strong advangtage in
admissions. At Berkeley, for example, the average African
American has an SAT score 288 points lower than the average
White, which is in turn lower than the average Asian.
There is reason to believe that at typical state universities,
the gap is even larger.
2. This is such a large gap that there is almost no overlap
between the qualifications of African Americans and Whites at
Berkeley. The *average* African American SAT score at Berkeley
is in the bottom decile of SAT scores at Berkeley.
3. In concrete terms, this means that most of the 18-year-old,
impressionable students entering Berkeley will be put in a
situation where African Americans will be near the bottom
of most of their classes. They are likely to notice this
and form lasting opinions.
4. The African Americans are likely to notice it, too, and the
result will not be particularly good for their self esteem.
There are many other threads running through the book. I really suggest
that reading it before spending too much time discussing it.
-mike
Paul S. Winalski
In article <39ip4i$j...@portal.gmu.edu>,
her...@science.gmu.edu (Harry Erwin) writes:
|>
|>Response to Winalski: the stats seem to show that lower socioeconomic
|>status confers lower fitness. Hence, there would be a selective gradiant
|>to eliminate genetic features confering lower socioeconomic status.
The stats that I've seen show higher birth rates and rising populatios
among the poor than among the rich. That says lower socioeconomic status
confers higher fitness. Remember that "fitness" in the evolutionary sense means
"reproductive fitness" (relative probability of passing genes onto the next
population).
--PSW
Alex Merz
Harry Erwin <her...@science.gmu.edu> wrote:
>
>Response to Winalski: the stats seem to show that lower socioeconomic
>status confers lower fitness. Hence, there would be a selective gradiant
>to eliminate genetic features confering lower socioeconomic status.
That depends on how you circumscribe the populations compared.
If you compare rich Sweden to poor, say, Kenya, you'll see a negative
correlation between birth rate and wealth.
As standards of living rise from abject poverty to a standard of living
that allows basic health and a steady diet you find that the birth
rate _drops_.
Moreover, you'll find that low death rates correlate well with
low birth rates & shrinking rates of population growth.
(None of which demonstrates _any_ relationship between wealth
and native intellegence - regardless of how one defines intellege
-A.
Alex Merz
>
>3: What about us POOR whites, we rated below orientals didn't we? It
>seems that everyone is discounting that difference. Why is one worth
>concern and another not of concern?
Well, there is a long history of scientists (yes, that's right - the
Exalted Priesthood) using _flawed_ and _biased_ arguments to support
the notion that the dominant culture is dominant _because of inherent
genetic differences between the dominant and non-dominant groups_.
For example, we might examine the use of "science" to justify the
creation of an Aryan 'master race'; to argue that blacks are inferior to
whites and that this justify slavery; or to wage a war of 'ethnic
cleansing'. Or that the Victorian English social order was a
natural result of a Darwinian selecive process.
In this country Asians are, as a group, doing pretty well. No one is
arguing that we should write 'em off as a group. People _are_, however,
saying these things about blacks and latinos. And some are even saying
that doing so would have a scientific basis.
It is my opinion that scientists have an ethical obligation to
vigourously and carefully criticize such proposals, and to point
out where they are flawed.
It's strange how much this (US) culture is starting to look like
the strange and cruel world described so vividly by Dickens.
I suppose it's only natural that those who are benefitting from
this transformation would try to find a justification for their
policies; and it's only natural that they'd resurrect 'social
Darwinism.
But it certainly is sad.
-Alexey Merz
ps- you might prefer to use the term 'asian' - unless, of course,
you think of yourself as 'occidental'.
Solomon Taibi
: On the other hand, the wards of the current US welfare programs are under no
: selective pressure whatsoever for higher inherent general intelligence. You
: don't have to think very much to sit around and collect welfare checks.
On the other hand, the selective pressures toward higher intellegence
(or any other trait) take an awfully long time to produce noticeable
results. It takes *much* more than a generation or so (in a large population)
to produce hereditary differences worth bothering about.
And what evidence is there that inherent general intellegence is
increasing among people who *do* work for a living?
Just how long do you expect the present political setup to last?
A million years?
--
Solomon Taibi
tai...@zodiac.cs.newpaltz.edu
State University of New York College at New Paltz
New Paltz, New York 12561, United States of America
Chris Colby
: The assumption that higher intelligence confers a "fitness" advantage wrt
: selective pressure appears to me to be unwarranted.
It also doesn't seem to fit the data. If the heritability of IQ is as
high as it is reported, this is good evidence that selection is not
acting on IQ. Natural selection reduces genetic variation in a
population. After selection acts on a trait the alleles within the pop-
ulation would all(*) be the same (or have similar phenotypic consequences),
so the environment would be the only source of variance.
(*) OK, not _all_, but you know what I mean.
Chris Colby --- email: co...@biology.bu.edu or co...@acs.bu.edu ---
"'My boy,' he said, 'you are descended from a long line of determined,
resourceful, microscopic tadpoles--champions every one.'"
--Kurt Vonnegut from "Galapagos"
Guy F. Barbato
I'm thrilled at J. Schall's comments, it's about time somebody thought about
what the term 'heritability' really means...I'm also happy that somebody
remembers Professor Lerner....his small book on Genetic Homeostasis still has
more Ph.D. projects in it than any other book i've read (including ed wilson's
Sociobiology)...More to the point, in *this* old chicken breeders mind,
outside of the context of a selection experiment, the concept of heritability
has no meaning at all (i think i just wrote that in response to another
post)...and the point of J. Schall is well taken, even in
selection/breeding experiments, heritability estimates from one population
have no relevance to measured differences in average trait epression from
a different population ......
much exciting work on the inheritance of IQ is available, an excellent
example is the work of Robert Plomin, who is using the twin model, not just to
tease out genes influencing intelligence, but is using differential estimates,
from different twin populations to discern the functional distinction of
shared and non-shared environment...that, to me, is the key question in the
intelligence debate....not just who is on what side of the Bell Curve, but
what is it in the environment that allows individuals to excel!!
Michael H Coffin
Joseph J. Schall <jsc...@moose.uvm.edu> wrote:
>First error: The 15 point difference between Blacks and Whites on IQ
>tests is of genetic origin. Heritability deals only with VARIATION, not
>with the placement of the mean. The following thought experiment shows
>this. Suppose 100 people were chose at random from the New York
>population. Their skin color would vary in darkness, in part because of
>genetic variation in that population (h2 would be fairly high for skin
>color). Then, spit the group into two groups of 50 randomly, send one
>group to spend the winter on sunny Aruba island and the other group stays
>in New York for the winter. Next spring bring the two groups back
>together. The group that spend the winter in Aruba will be darker on the
>average than the group in New York. They cannot be different genetically
>because the two groups were chose randomly from the original 100 people.
>The difference is of environmental origin, even though h2 for skin color
>is high! This thought experiments shows the general case....heritability
>tells us nothing about differences between groups.
Strange to call this an `error' since the authors point this out
themselves, and at great length. The example they use to illustrate
the point is a plant (a pea plant, if I remember correctly, I don't
have the book handy). Saying that `the size of peas is 80%'
heritable' means that 80% of the variation seen in the sample
population can be attributed to heredity. However, Murray and
Hernstein point out, if you plant one pea plant in a garden in the
midwest and another in Death Valley, California, one will shrivel and
die and the other will produce peas. This difference has nothing to
do with genetic differences between the two plants. The two can even
be clones.
The point is that *extreme* variations in environment can have a large
effect even if, in the population as a whole, most of the variation is
due to heredity. Murray and Hernstein point this out on several
occasions, and, as far as I can see, are scrupulous in not making
arguments that rely on such fallacies.
> [...]
>These problems were pointed out many years ago by Michael Lerner who
>worked out many of the details of measuring heritability and by Lewontin
>who is one of the great geneticists of our time. EVERY quantitative
>geneticist knows of this error. The mystery to me is why NO journalist
>or reviewer seems to have found this flaw in the argument.
Perhaps because they don't use these arguments. Having now trudged
carefully through all but the last few chapters of The Bell Curve, I
have not seen these arguments made anywhere. In fact, Murray and
Hernstein carefully point out these possible misinterpretations of
their arguments.
-mike
Ashwani Vasishth
>Stephen Bogner (sbo...@dres.dnd.ca) wrote:
>
>: The assumption that higher intelligence confers a "fitness" advantage wrt
>: selective pressure appears to me to be unwarranted.
>
>It also doesn't seem to fit the data. If the heritability of IQ is as
>high as it is reported, this is good evidence that selection is not
>acting on IQ. Natural selection reduces genetic variation in a
>population. After selection acts on a trait the alleles within the pop-
>ulation would all(*) be the same (or have similar phenotypic consequences),
>so the environment would be the only source of variance.
>
>(*) OK, not _all_, but you know what I mean.
>
Is this, in fact, the case? Does this not assume that *selection*
is some smooth, if variably sloped, directional progression? Some
sort of funneling toward an *ideal* state? And does this not, in turn,
assume that the criteria of fitness (and I do mean biological) are
fixed in time and space?
I am led to think otherwise (confident that I will be corrected
if I go astray) on two counts. First, if I have understood Gould
& Eldgredge/s arguments for allopatric speciation (on the basis
of geographical isolation?) and punctuated equilibria, then
(does this work at the level of the allele?) two groups could
start from the same *place* and evolve quite separately. Then,
given mobility and mixing, should we expect reduced variation?
Second, besides the above, there is the work of Margulis & Lovelock on
their modified formulations of the Gaia hypothesis (and I do not
mean the touchy feely versions, but the work around their Daisy
World model [eg: Lynn Margulis & James Lovelock, 1989, **Gaia
and Geognocy** in _Global Ecology: Towards a Science of the
Biosphere_, Academic Press]). If *the environment* is fully coadaptive
to the organisms in it (and process-function ecology would deny
the distinction in the first place), then must not any meaningful
criteria of fitness also be fully evolutionary? So is there some
one place to narrow in on?
>Chris Colby --- email: co...@biology.bu.edu or co...@acs.bu.edu ---
>"'My boy,' he said, 'you are descended from a long line of determined,
>resourceful, microscopic tadpoles--champions every one.'"
> --Kurt Vonnegut from "Galapagos"
>
regards.
ashwani. vasi...@usc.edu (213) 737-7875
Chris Colby
: >Natural selection reduces genetic variation in a
: >population.
: Is this, in fact, the case?
Yes, this is, in fact, the case. Selection is differential reproductive
success of (classes of) genetic variants in the gene pool. On average,
selection reduces the amount of genetic variation at the locus under
selection. There are some exceptions. Balancing selection, and
frequency-dependent can maintain higher levels of variation at a
locus than predicted by the neutral theory.
: Does this not assume that *selection*
: is some smooth, if variably sloped, directional progression? Some
: sort of funneling toward an *ideal* state?
No, all it assumes is that some alleles have a higher relative fitness
than others. These alleles increase in frequency with respect to less
fit alleles. This reduces genetic variation.
: And does this not, in turn,
: assume that the criteria of fitness (and I do mean biological) are
: fixed in time and space?
No, alleles may have a high relative fitness at some times and a
low relative fitness at others. It depends on what other alleles
are in the gene pool at the time and the environment. Evolution
can be like a game of paper/scissors/rock. The "scissor" allele
will have a high relative fitness in a population of "paper"
alleles, but a low relative fitness in a population of "rock"
alleles.
: ashwani. vasi...@usc.edu (213) 737-7875