Sewall Wright 93and not by the effect of the individual mutations[Haldane, 1937]. A second idea is
his showing that the amount of reproductive excess required to carry through an
allele substitution is a function mainly of its initial frequency[Haldane, 1957]. Both
of these have the Haldane touch of finding an unexpectedly simple relationship,
but both have turned out to be too restricted for most modern applications.
Fisher’s best-known contribution to animal breeding and evolution is his show-
ing by the use of least-squares theory how selection, either natural or man-made,
can improve a population despite complicated interactions among genes[Fisher,
1918 ]. The central idea in his view was his “Fundamental Theorem of Natural
Selection”, namely that the rate of change of mean fitness of a population at any
instant is determined by the genic variance of fitness at that time. Genic vari-
ance is the variance of the additive component of the phenotype, as determined
by least-squares. Fisher was also the first to introduce stochastic processes into
genetics. His block-buster was his “The Genetical Theory of Natural Selection”,
regarded by some as the greatest book on evolution since Darwin[Fisher, 1926].
Whether this is true, there is no doubt as to the richness of the ideas; each re-
reading brings new insights. Fisher was often obscure, but nobody could match
his elegance both in mathematics and in English prose. If you read the book, use
the variorum edition, which has a very useful appendix that clears up a number
of Fisher’s obscurities[Fisher, 1926].
Wright was always interested in population structure. His inbreeding coeffi-
cient, discussed earlier, and its extension to hierarchical population structure are
among his greatest accomplishments, now widely used in the study of population
structure, including humans[Wright, 1951]. One of his aims was to write in gen-
eral form an equation describing the change of gene frequency under the influence
of mutation, migration, selection, and random processes. It later turned out that
his equation, in differential form, was already in use by physicists — the Fokker-
Planck equation applied to diffusion and other processes that involve deterministic
and random elements[Wright, 1968-78].
None of the three pioneers was concerned with mathematical rigor. After World
War II the field of population genetics became more mathematical. The French
mathematician, Gustave Mal ́ecot, developed Fisher and Wright’s ideas in a math-
ematically rigorous form[Mal ́ecot, 1948]. Motoo Kimura introduced the Kol-
mogorov partial differential equations to solve problems that had eluded Wright.
Kimura’s most important papers have been collected in a single volume[Kimura,
1994 ]. This work laid the foundation for the realization that a large part of DNA
changes in evolution is driven by mutation and random processes and is relatively
uninfluenced by selection[Kimura, 1983]. This then led to a “molecular clock” by
which rates of evolutionary change can be calibrated.
Although most of Wright’s work in population genetics was theoretical, he en-
tered several collaborations with experimentalists, especially Th. Dobhansky. This
was the beginning of a trend toward theory-driven experimentation in population
genetics that still exists. Wright soon tired of this and, sometimes with difficulty,
disentangled himself from these alliances, and went back to his theoretical studies