Def of evol: the phenotype is not an allele

arlin at arlin at
Fri Nov 1 22:11:31 EST 1991

Note added in proof:  I see that the debate is heating up and I am
glad to see that more people are joining into the discussion.  Alas,
I am already behind-- this posting refers only to the recent comments
of Mr. Schultz.  I have not taken a close look at Larry's arguments
as yet, but I think we are about to make a breakthrough regarding
non-allelic genetic changes.  This posting is on the 
subject of isolation.

The argument that isolation is conceptually distinct from allele
replacement will resume below.  But first, to prevent further postings
of verbose examples of sympatric speciation, I want to make perfectly
clear that I understand how sympatric speciation can occur, by
explaining what is left out by Mr. Schultz, who writes:

>In article <1991Oct30.015630.1801 at> arlin at writes:

>>However, the process of replacing 100 "A" beads at a locus (in
>>a population of 100 necklaces) with a 100 "B" beads at the same locus
>>will never by itself split 100 necklaces into two populations of 50
>>necklaces each.  One must invoke a conceptually distinct process of
>Let's say each individual produces one of two possible mating pheromones
>(coded at another locus), and the "A" allele says "mate without regard 
>to pheromone" while the "B" allele says "mate only with same-pheromone
>individuals."  You start out with 100% A and random mating, and end up 
>with 100% B and two isolated populations.
>Of course, the big question is how and why does B spread?  This is the
>precisely the sort of problem confronted by advocates of (non-
>chromosomal) sympatric speciation.  My point is not that this mode of
>speciation is common, but that isolation can, physically, result from
>single-locus allele replacements.  Isolation is not conceptually 
>distinct, but can be considered as a phenotype of the allele.

If B individuals only mate with B individuals, then who does the first
B individual mate with? Unless two B individuals arise independently
in the same generation, the B allele will not be propagated, but will
die out each time it arises. This is not a criticism of Schultz's
general position, since the problem can be corrected in several ways.
For instance, one solution is to suggest that the B allele arises in
the germ line of an individual that has already developed phenotype A
(lets say its a matter of mate choice). The B-bearing A-phenotype
individual could mate with an A individual, and if this union passed
the B allele on to male and female children, they would have the B
phenotype but could mate with each other, to create a separate
population.  A second way to get around the problem is to assert that
B individuals *prefer* to mate with B individuals, but will mate with
A individuals (rather than insisting that B's *only* mate with B's):
the first B individual will be forced to mate with an A, but the
B-type offspring of this union or some successive union will be able
to mate with the preferred mate, another B type.

My argument is not that sympatric speciation cannot occur, but that it
cannot occur by allele replacements alone. Given this, it is odd that
Mr. Schultz omits from the section that follows the final sentence of
my paragraph-- whichout which the argument is incomplete-- and
thus attacks a strawman:

>>To reiterate this point one more time (since it is indeed conceptually
>>difficult) : if you wrote a computer program that carried out allele
>>replacements in a population and set it running, you would never come
>>back to your computer and notice that two populations or more were in
>In fact this can and has been done.  You start out with a panmictic
>population, introduce an allele whose phenotype is some sort of 
>assortative mating (e.g. "mate only with other carriers of this allele," 
>or "mate only with individuals who share an allele at some other 
>locus"), introduce a mammoth selective force, and eventually you get
>a "population" of two non-interbreeding subpopulations.  The splitting is
>gradual as it accompanies the spread of your assortative mating allele.  

The complete paragraph reads as follows:

>To reiterate this point one more time (since it is indeed conceptually
>difficult) : if you wrote a computer program that carried out allele
>replacements in a population and set it running, you would never come
>back to your computer and notice that two populations or more were in
>existence.  In order for this to happen, you would have to add a new
>mechanism to the program's repertoire, with an instruction such as 
>"when the allele at locus 25 changes from *A* to *F* in a single 
>individual, separate the population into two populations, one with 
>all *A* at locus 25, and one with the *F* organism."

To the claim that reproductive isolation is distinct from a genetic
allele replacement, Mr. Schultz argues that:

>Isolation can be "included" as the result of an external force 
>(random or not), or as a phenotype of an assortative mating 
>allele.  The former is the rule in nature, and I know of no definitions
>of evolution that exclude the operation of external forces . . . 

But the phenotype of an allele is not an allele, nor can behavioral
ramifications of expressing the phenotype be construed as an allele
replacement in any precise way.  In general, if "allele replacement"
is such a fuzzy concept that it can be expanded to include such
diverse processes as 1) symbiotic association (see L. Moran's posting
of 24 October); 2) isolation resulting from phenotypic expression of a
variant allele (above); and 3) changes due solely to mutation and not
cumulative differential reproduction (when a macromutation creates a
new species represented by a single individual, as described in the
previous round of postings by Schultz and myself), then I would
suggest that "allele-replacements" can be expanded to encompass
everything including the kitchen sink, and that it is a foregone
conclusion that "allele-replacements" will subsume all of organic

If, on the other hand, the allele-replacement paradigm is a precisely
defined model of genetic change [in which deterministic or
non-deterministic differences in the reproduction of individuals in a
population alter the frequency of a nucleic acid sequence that has
arisen by mutation at a defined chromosomal location, allowing it to
eventually replace the previous nucleotide sequence] then a definition
of evolution based solely on allele replacements can be judged
incomplete due to its failure to refer to the isolation (biological or
geographic) necessary for speciation.  In addition to failing to
address cladogenesis adequately, the allele-replacements definition of
evolution does not even encompass all of the types of genetic change
that can take place in a stable population (some changes such as
concerted evolution in multigene families, creation of new loci,
chromosome splitting, etc, are non-allelic changes).

The resistance to these simple concepts befuddles me, since it is easy
enough to correct the allele-replacements definition with a more
complete definition that refers to isolation (which allows speciation)
and to "changes in the genetic composition of a population" (which
allows non-allelic changes).  It would indeed be nice if evolution
boiled down to some rigorous singular kernel like "allele replacement"
 . . but it doesn't.

Arlin Stoltzfus

Arlin at

More information about the Mol-evol mailing list