Evolution without Selection

Chris Cogan (ccogan@sfo.com)
Thu, 28 Oct 1999 12:03:17 -0700

The following is a draft of an essay I'm working on. Copyright 1999, by
Chris Cogan
(Of course, members of this list may do the usual list-things with it, but I
include the copyright notice in case I decide to try to publish a later
version of it in "the real world.)

When I pointed out recently that evolution would be faster if there were no
selection, someone responded that, logically, there would be no evolution at
all without selection.

I'm not sure of the person's reasoning (he didn't give any), but other
people hold similar views, so I thought I'd develop the point further.

Let's say that we have Godlike powers so that for every replication of the
genome we can save it and reproduce it in the future, and suppose we start
with the simplest possible self replicating genome. Suppose we reproduce it,
with exactly the kinds of "random" (i.e., purposeless) variations that occur
in natural genome replication, and that we do this for many, many, many,
generations, saving every genome thus created for, let us say, three billion

Now, admittedly, many of these variations would still be simple. But, since
genome replication does often produce more-complex genomes, it would also
include a progressively larger number of more-complex genomes, which, in
time, would lead to even *more* more-complex genomes.

In effect, there would be it a "fanning out" of genomes so that, graphed
two-dimensionally, the range of genome complexity would be increasing
through time, so that, at later times there would normally be some species
(of genome, at least) that would be more complex than the most complex
species at prior.

Obviously, if all variations are saved in each generation, there will be
more material for variation in the next generation, and so on. That is, if
we do any kind of culling, even if only for one generation, there will,
other things being equal, be fewer opportunities for variation in the next
reproductive cycle. In fact, if selection is sufficiently stringent, we can
totally prevent evolution simply by removing all replications that are even
slightly different from the very first genome, allowing only perfect copies
to reproduce.

(Natural selection can, very indirectly, influence variations by changing
the availability of resources, so that, in the real world, the decrease in
the population of one species allows for an increase in population and
ecological niches for another species, which may therefore be allowed to
produce more variations than it otherwise would. But this does not affect
the main point: New variations depend on the variation mechanism itself, not
on the selection or culling mechanism.)

Put another way: If all variations are saved and also reproduced with
variations, there will have to be more variations than there will be if many
variations are removed before they can replicate.

Consider an example: A bear-like animal that evolves into whales. I will
call the bear-like animal a "blike" for the purposes of this example. The
blike, let us suppose, lives in and at river deltas, mainly by catching and
eating fish in the river. It spends much time in water, so swimming well is
a good thing for it, but swimming poorly is not, so blike variations that
can't swim are selected out.

Eventually, the need to swim and the consequent swimming-improvement
modifications that make it adapted to swimming make the modified blike so
much less suited to life on land that it simply stays in the water for all
of its life. Now, swimming well becomes even more important, so variations
in which the blike's limbs are more fin-like and less leg-like give a
decided advantage over those blikes with less fin-like limbs.

Given enough time, and continued advantageousness of whale-like
characteristics (such as moving the nostrils further up and back, etc.), a
whale will evolve from the blike.

But, suppose there is a complete lack of selection in the case of the
initial blike, such that all replications (and thus all variations) of its
genome are also allowed to replicate. In this case, evolution to whalehood
would proceed faster because:

1. There will be many more opportunities for more radical variations to

2. All variations that happen to be in the direction of whales would be
saved, including many that would normally get culled out early on only to be
needed later.

Of course, this kind of evolution would produce not only whales but also a
vast number of other organisms that would be reached by variations starting
from the blike. If one mutation was created that produced a two-headed
blike, it would be reproduced with variations (usually). Eventually, more
heads would appear. Why? Because there would be no selection to get rid of
the two-headed blike, thus leaving things set up for yet more heads to
appear. Without selection, nature would be absolutely teeming with genetic
"monsters" as well as "normal" organisms, and some of these "monsters" would
be intermediate between organisms that normally could not evolve across the
gap between them. Selection eliminates unfit organisms whether they are
intermediates or not, thus reducing the overall "progress" of evolution
toward more-complex organisms.

To come at my point another way, imagine putting a huge tank around a bushy
tree and filling it full of plaster of Paris, and then somehow dissolving
all of the tree out so that what's left is a mold of the tree. Now think of
each small space in the mold as an ecological niche into which life may
evolve, and the plaster of Paris as selection, preventing organisms from
evolving into any volume of space filled with plaster.

Eventually, life will fill all the niches in the mold. But, if there were no
mold at all, life would simply spread in all directions, including into the
volume of space that would have been occupied by the plaster of Paris, until
some boundary is reached (and then that boundary will act as a selection
mechanism). Without the restrictions of selection, life would be free to
spread in all directions and thus to reach most points more quickly than it
would if it is forced to follow only those paths that are biologically
viable in the real world.

Logically, evolution without any form of selection or culling will have to
produce more variations than evolution with selection. More-complex
organisms are the result not of selection as such, but of variation during

My point is not to be confused with genetic drift, in which a genetic
population gradually changes because of the internal "economics" of genetics
and genetic replication with variations. What I'm talking about is total
lack of any kind of selection mechanism, so that every variation that occurs
during replication is preserved and used to replicate with variations yet

I have belabored my point so much, and in so many different ways, because so
many people seem not to understand the significance of variation in
evolution. The remark that evolution would logically produce nothing without
selection demonstrates how deeply this fact is misunderstood (and how poorly
some people understand evolutionary theory generally).

So, what is the role of selection? It is this: It introduces restrictions on
evolution so only some genome variations will be perpetuated. Selection is
the force that introduces sharper differentiation among surviving
variations. Where random variation would produce an almost perfectly dense
range of variations between any two organisms, selection steps and
eliminates those that are not suited to surviving in their respective
environments. Because the environment (during DNA replication and onward)
disfavors some variations more than others, the incredibly rich
diversification that would occur under variation alone is sharply reduced.

Physics, chemistry, environmental forces, and availability of resources all
impose selection, or culling, of life so that only some variations are
allowed to reproduce. This reduces the total number of distinguishable
variations in organisms (genomes, actually), but makes for sharper and
deeper local (and temporary) distinctions among the genetic types that do

This reduction in diversity also slows overall "progress" toward the most
complex genomes, because it eliminates most potential intermediates between
simple genomes and those that are vastly more complex. This is why flies
don't evolve into bears: bear-ward steps in fly evolution are mostly bad for
flies, therefore bears have to be arrived at via other paths. Without
selection, flies could evolve into bears (and millions of other things).

Summary: Sloppy copying produces variations. Overall, evolution would
proceed to more-complex (and other) variations more quickly without

--Chris Cogan

Now is the time for all good people to come to.