Segregation and recombination are normal processes
constantly going on in all sexually reproducing
populations. They are the main genetic processes involved
in population variation, as has been seen. They can,
without mutation, carry the population beyond its former
range of variation by utilization of the pool of potential
genetic variation. These processes are, however, gradual in
a population as well as fluctuating and, in the main,
reversible. There is no evidence known to me of
segregation or recombination in nature occurring in a
single step and carrying an individual well outside the
former normal range of variation, even to the extent that a
large mutation can do this. If such an event did occur, its
results would almost always be quickly damped or lost by
backbreeding of the individual or its descendants. The
fluctuating variation of a population does not seem to
include any self-contained mechanism for permanent,
one-step changes in the mean and range of that variation.
The chances of multiple, simultaneous mutation seem to be
even smaller, indeed negligible. Postulation of a mutation
rate of .00001 and of each mutation's doubling the chances
of another in the same nucleus would correspond with the
most favorable circumstances warranted by laboratory
evidence. Under these postulates, the probability of five
mutations in the same nucleus would be approximately
10^-22. With an average effective breeding population of
100 million individuals and an average length of generation
of one day, again extremely favorable postulates, such an
event would be expected only once in about 274 billion
years, or about a hundred times the probable age of the
earth. Obviously, unless there is an unknown factor
tremendously increasing the chance of simultaneous
mutations, such a process has played no part whatever in
evolution. The chances of five different mutations
occurring separately in a population and eventually being
found together in one zygote may be appreciable under
favorable circumstances, but this is a gradual, not a
one-step process; it is part of the normal mechanism of
nonsaltatory evolution.
The reasons why some mutations have small and others
large effects are to be found in the process by which
mutations acquire phenotypic expression, i.e., in
development of the individual. It is known that some
mutations affect rates of development and times of
termination of development, changes in [sic] which may
have distinct or even radical effects on adult form. It is
also known that some genetic effects, presumably subject
to mutation as all genetic effects seem to be, may appear
only or, at least, most strongly in one stage of
development. As De Beer (1951) has emphasized, these
facts might make possible quite sudden and radical changes
in adult structure. Caenogenesis (appearance of youthful
characters not affecting the adult) followed by
paedogenesis or neoteny (retardation of development,
cessation of development at an earlier stage than in the
ancestry, or both) would result in paedomorphosis,
appearance of the formerly juvenile characters in the adult.
This might seem to be abrupt if the earlier juveniles were
unknown and if the neoteny resulted from a single large
mutation (a point on which De Beer does not commit
himself).^
