The Development of Population Genetics 329such regularity was guaranteed by the same circumstances that made a statistical
assemblage of particles like a gas bubble obey the gas law (without appreciable
deviation) [Fisher, 1958, 34–40]. He went on to compare his fundamental theorem
to the second law of thermodynamics: Both are properties of populations that
are true regardless of the nature of their individual units, and both are statistical
laws, each of which requires continual increases in a measurable quantity (in one
case entropy, and in the other fitness). And, he adds, “it is more than a little
instructive that so similar a law should hold the supreme position among the
biological sciences” [Fisher, 1958, 39]. The key for understanding Fisher’s approach
is, once again, the use of idealizing assumptions similar to those used in gas theory.
The power of his model derived from elimination of parameters such as migration,
isolation, genetic recombination and gene interaction. He was able to separate, in
a way that could not be done in empirical studies, the key features in populational
variation: genetic factors, environmental factors and sampling errors. Fisher’s
quantitative expression for genetic variance enabled him to show the amount of
variation any particular genetic trait could exhibit in a population distinct from
environmentally caused fluctuations.
Building on Hardy’s work, Fisher pointed out that in Mendelian inheritance
there was no inherent tendency for variability to diminish over time. Alternative
genes were conserved unless their proportions were changed by selection, chance
or mutation; and as Fisher’s calculations showed, selection was the factor most
effective in changing gene frequencies. Mutation could no longer be thought of as
the force of evolutionary change, for even very small selective effects could over-
power it. We know that the Malthusian parameter m, which represents fitness, has
a natural-selection component that must always equal W, the genetic variance in
fitness. So for Fisher, m did not simply measure changes in population numbers,
but rather the change in the total reproductive value of population members; in
other words, it measured the rate of population growth in terms of total repro-
ductive value. Because Fisher was able to give a quantitative expression for W,
it was possible to determine the effects of selection as a parameter in the overall
determination of m.
Wright wrote a review of the Genetical Theory of Natural Selection in 1930
which set out in very brief form some of his disagreements with Fisher, particularly
the characterization of selection. He went on to develop his views and opposition
to Fisher in a series of papers written over the next ten years. In conclusion I
would like to highlight the nature of these differences, many of which remain the
subject of debate today.
5 DEBATING THE DETAILSIt was in the 1930 review and his famous paper of 1931 “Evolution in a Mendelian
Population” that Wright presented his ideas about the shifting balance theory of
evolution and the notion of an adaptive landscape. If one can imagine a field
of visible joint frequencies of all genes spread out in a multidimensional space,