54 Sahotra Sarkar
alleles were supposed to be only subject to weak selection for two reasons: “(1)
the effect of the factor on development may be very slight, or (2) the factor may
effect changes of little adaptive importance”. Therefore, he concluded, “variation
of adaptive importance should be due to numerous factors, which individually are
difficult to detect” (p. 24). The assumptions that weak selection was paramount
and that evolution was most likely to occur in large populations became the two
major tenets of Fisher’s theory of evolution.
Meanwhile, Wright had just published his first analysis of selection In 1921, in
a set of five papers, “Systems of Mating”.^10 Unlike Fisher, Wright worked with
explicit genetic models (with two alleles at each of one or two loci). However,
he, too, assumed a population at equilibrium where the Hardy-Weinberg rule
held and, for the two locus case, ∆ 0 = 0. He first worked out several of the
correlation coefficients for populations at equilibrium — these were only special
cases of Fisher’s [1918] more general treatment [Wright, 1921a]. However, he gave
a general and systematic treatment of inbreeding [Wright, 1921b]. He analyzed
assortative mating for a simple two-locus model [Wright, 1921c] but the analysis of
selection remained superficial [Wright, 1921d]. When only one locus was involved,
he rederived the results of Jennings [1916] and Wentworth and Remick [1916]. For
two loci, his results demonstrated that selection decreased the variability within
a population. The most lasting contribution of this set of papers was the first
systematic presentation of his method of path coefficients, his novel — and peculiar
— method for calculating the correlations between variables.^11
(^10) By this point, significant work on population genetics had also begun in the Soviet Union. S.
Chetverikov, in Moscow, published an important theoretical paper in 1926, and the group around
him even began investigating the genetics of several natural populations of various Drosophila
species that were native to the Moscow region [Adams, 1968; 1980]. In the 1920’s this work
remained unknown in Britain or the United States, where most of the significant development of
population genetics theory took place. Chetverikov’s 1926 paper was only available in Russian,
and a projected English translation was never published probably because of his mysterious arrest
by the Soviet secret police (OGPU) in 1929, and subsequent three-year exile to Sverdlovsk and
Voronezh [Adams, 1980]. Among Western geneticists, only Haldane seems to have been aware of
the Russian work. He met Chetverikov at the Fifth International Congress of Genetics in Berlin in
- Subsequently Haldane visited the Soviet Union, as Vavilov’s guest, and there is little doubt
that he became familiar with Chetverikov’s work. Eventually, Haldane had Chetverikov’s papers
translated into English by a Mrs. A. Sproule, and these were then circulated in his group [Adams,
1980]. However, that was in the 1930’s, after he had moved to University College (London) from
Cambridge. In 1932, inThe Causes of Evolution, the only reference to Chetverikov was to his
experimental work. The significance of Chetverikov’s theoretical paper remained unrecognized
during the critical period in the late 1920’s when Fisher, Haldane and Wright were systematizing
their alternative views about the process of evolutionary change.
(^11) In the early 1920s, Wright developed quantitative models of selection. A long manuscript re-
porting this work remained unpublished, as Wright dealt with his heavy teaching responsibilities
at the University of Chicago [Provine, 1986]. A large portion of it became obsolete as, starting
in 1924, Haldane began publishing his series of ten papers, all but one of which were entitled,
“A Mathematical Theory of Natural and Artificial Selection”. The rest of Wright’s manuscript,
after significant development, became Wright’s classic “Evolution in Mendelian Populations”, his
alternative to Fisher and, incidentally, to Haldane.