94 James F. Crow
[Provine, 1986].
6 THE SHIFTING BALANCE THEORYWright regarded his evolution work as his most important contribution. The gen-
eral idea goes back to 1920s and, although it underwent a number of refinements,
the basic principle remained unchanged. In later writings, he called this the “Shift-
ing Balance Theory”. For a very full analysis, see[Wright, 1968-78].
The basic difficulty, as Wright saw it, is that natural selection cannot ordinarily
change a sexual population from a state of high average fitness to one of higher
fitness if this means passing through a stage with lesser fitness. The shifting
balance theory is an attempt to get around this difficulty.
According to Wright, the theory came mainly from three observations.
(1) In his analysis of guinea pig hair pattern and coat colors Wright was im-
pressed by the fact that genes in combination are often not what would be pre-
dicted from their individual effects. An example that he often mentioned was a
rosette hair pattern, in which the combined effect of two genes was opposite to
what would have been predicted from their individual effects.
(2) His observations of inbred lines of guinea pigs showed that inbreeding, in
addition to bringing about decline in vigor and fitness, also led to differentiation
among the inbred lines. The variability among the lines was much greater than
that in the foundation stock from which the inbreds were derived.
(3) The third observation came from the history of domestic livestock, partic-
ularly shorthorn cattle. He saw that advances in the quality of the breed did not
seem to arise from recurrent selection by breeders. Rather it sometimes happened
that, for no apparent reason, a particular herd demonstrated superior qualities.
Bulls exported from this herd transmitted these qualities to the whole breed so
that the breed gradually attained the quality of the favored herd. Then the cycle
could be repeated, as another herd drifted into having desirable qualities.
These observations are almost a summary of the shifting balance theory. Wright
liked to use a peaks and valleys metaphor to describe the process as he saw it. In
such a three dimensional graph the abscissas are usually allele frequencies and the
ordinate is the mean fitness of a population with corresponding allele frequencies.
The loci are treated as independent, making the abscissas orthogonal. The idea is
easily generalized, if not easily visualized, to more than three dimensions.
Wright thought of the shifting balance process in three phases.
Phase 1. He envisaged a large population broken up into partially
isolated subpopulations (like herds of cattle). These subpopulations are
small enough that allele frequencies within a population drift randomly.
Somewhere among the subpopulations, one may happen to drift into a
favorable combination of allele frequencies.
Phase 2.Once a subpopulation comes into the domain of attraction of
a higher peak, the gene frequencies will change such that the subpopu-