This is the blog for IAB

The mission of the International Association of Bryologists (IAB), as a society, is to strengthen bryology by encouraging interactions among all persons interested in byophytes.

Tuesday, October 13, 2009

BRYONET: levels of selection


David Wagner wrote:

"...natural selection. And that natural selection is just as likely to
select for stability based on adaptive value of plesiomorphies as it
is to select for change due to advantages of apomorphies. Natural
selection may produce species that have unusual breeding systems
inimical to monophyly. Natural selection coupled with a dynamic
environment serves to emphasize the distinctness of the entities we
can call species. Recognizing species as the basic unit upon which
natural selection acts distinguishes the species rank as different
from all others."

I hope I might politely change the subject away from classification
and take off from what Dave says, perhaps for the benefit of some
scholars out there in bryonet-land who would be interested in the
hierarchical evolutionary process.

I think there are levels of selection.

The most familiar selection to those of us who teach evolution
classes is what we evoke in explaining such organismal adaptations as
hairpoints on leaves in Grimmia laevigata and its relatives or a
short generation time in Pluridium subulata and its relatives. This
is selection that happens among individuals mostly within
populations. In ideal sexual populations, you could alternatively
view it as among alleles (forms of a gene) that are in individuals in
a population. Those individuals that are more well-endowed survive
and reproduce better than individuals less well-endowed. Theory and
many experiments have show that this sort of familiar selection very
quickly optimizes most quantitative traits. Most of the time most
plants are expected to have traits not too far from the optimum for
that kind of plant as it lives in its usual habitat. When the trait
rose to prominence, the selection would have been directional
selection. But later selection would have ended up being stabilizing
selection. I generally imagine there is plenty of stabilizing
selection out there. (Brent might like to think this is not the case,
but I guess when pressed he would admit that there was probably
stabilizing selection in the past or at least eliminating extreme
detrimental mutants in a form of normalizing selection, and so on.)
Anyway, when I say "kind of plant" I do not only mean "kind" at the
species level specifically. A vast group could be adapted to a sort
of niche (all the species of Sphagnum), and the characters of that
vast group might be conservative because they are (or have been in
the past) subject to stabilizing selection. Within that general sort
of niche, there might be finer portions of the niche space claimed by
smaller groups (different species of Sphagnum) There might also be
fine-scale stabilizing selection that acts on different ecotypes,
keeping them different, for example a population on the coast that is
bathed in maritime fog versus an ecotype that is predominant inland.
Blah, blah, blah... (Yes, I know that all selection is not external
niche-based selection, but the same argument applies to selection
that is judged internally as the coadapted genome...) blah, blah, blah.

There are also higher levels of "selection" once you have a whole
bunch of what is usually referred to as "species" in the regional
species pool. This kind of selection goes under many names, including
such things as species sorting, species filtering, or even ecological
succession. The use of the phrase "species selection" is unfortunate
because the things that are being selected in a local community are
avatars (i.e. the local representatives of a clade generally not
consisting of closest relatives: see my previous email) and going
beyond the local population, then they are various groups some
monophyletic and others paraphyletic. "Clade selection" is also used
often, and it has some advantages terminologically, but not the
advantage of actually being strictly speaking monophyletically
correct. The selection we are talking about is specific to the level
at which it acts because of the criterion of selection that sorts the
groups that differ by that criterion. Consider the simple case of
just one character with two states, like dioicy versus monoicy. If we
start out with a diverse bunch of things that are dioicous and then
monoicy evolves repeatedly, monoicy is derived and dioicy is
plesiomorphic. (Ignore for the moment that in the real world there
are reversals to dioicy). We would end up with two states and
variously related plants possessing those two states. Now, let's
imagine that in a desert region, there is high-level selection for
monoicy, and in a forest region there is high-level selection for
dioicy. The two floras would come to differ. The desert region would
have relatively more subspecies, species, subsections, sections, etc.
of monoicous plants, and the forested region would have relatively
more subspecies, species, subsections, sections, etc., of dioicous
plants. But, as with the ancestral and derived forms of alleles, the
dioicous ramification had a head-start over the various monoicous
derived clades. Thus, it is at least slightly incorrect to emphasize
monophyly in the name of the macroevolutionary process. Blah, blah.

The microevolutionary process of organismal adaptation by selection
among individuals differs in many important ways from the
macroevolutionary process of a biota evolving by selection among
clades. Furthermore, the lower level process feeds the upper level
process. Adaptation at the lower level corresponds to "mutation" at
the higher level. It is only because of adaptations at the lower
level that clades have come to differ in biologically meaningful
ways. An important difference is the opportunity for sex. At the
lower level, there is usually some way for recombination to happen
among loci, and this means that loci can coadapt with one another
very quickly. At the macroevolutionary scale(s), there is not a
regular opportunity for recombination. At the lower level, for
quantitative traits, we rarely worry much about the effects of
mutation mostly going in one direction (mutational drive). At the
upper level(s), the corresponding process, annoyingly called
"speciational" drive may be very important in affecting the balance
of different kinds of organisms in the flora. For example, there may
be a great deal of speciational drive that changes forms that are
unspecialized as to rock type into forms that are specialized
calciphiles. This could account for a large number of endemic
calciphiles whose calfiphily is convergent. Blah, blah, blah...

Okay, anyway, I'm getting to the point. When Dave says stuff like,
"Recognizing species as the basic unit upon which natural selection
acts distinguishes the species rank as different from all others," it
sort of drives me nuts. I think I might know what he feels, but would
my students who also read bryonet? Let me try to unravel into my
language what I think he is saying. First, I suppose he means that
different avatars fill different niches in the local community, and
selection among avatars keeps most of them in each their own familiar
niche (except when a small subset ends up on an empty island for a
long time), and that then there is stabilizing selection among
individuals of each avatar that is very similar in all the
communities where that clade lives throughout a fairly large
geographic range. Second, I think that he means that there are gaps
between different stable adaptive states and so much of the evolution
that ends up being taxonomically important happens in a punctuational
manner. But surely, he doesn't think that the gaps are all the same
size. Sometimes when there is a punctuational shift it moves the new
clade a long long ways from the old adaptive norm (and the resulting
thing is considered a monotypic genus), and sometimes it moves it
just to another rise on the same adaptive cordillera (and the
resulting thing is considered a subspecies). Third, I think that he
might mean that at some point, things are indivisible because they
are all the same. Again I agree that this is be basically true if one
sticks to one geographic region and studies organisms that move
around that region pretty well (it also greatly "helps" if the region
was glaciated), but anyone who has tried to come up with a treatment
of all the specimens from around the world in a genus of any size,
will realize that many widespread species vary geographically. One
could break them up into segregate species, but being a splitter is
all a matter of degree, unless you get down to where there are no
widespread species. Even if you went to an absurd amount of splitting
(what Richard Zander wishes to avoid), you would end up with entities
that were made not equal in other respects, such as the degree of
distinctiveness when compared to their closest other "species".

Finally, I think all this stuff of which I write about avatars and
niches and hierarchical selection is way-totally interesting, and
need not be motivated by any desire to figure out how we are going to
formally name species or clades. Classification and nominclatural
rules are a little bit interesting. The evolutionary play on the
ecological stage is riveting, at least that's what I think.

Respectfully, Paul Wilson
California State University Northridge

Bryonet is a service of the International Association of Bryologists (IAB)
and is administered by Janice Glime <>
through Michigan Technological University
IAB Officers:
President Jeff Duckett <>
Secretary-Treasurer Jim Shevock <>
First VP: Masanobu Higuchi, Second VP: Jiri Vana
To join IAB ($11 per year), contact Jim Shevock <>
IAB website:

No comments:

Post a Comment