The Development of Population Genetics 317progress in the theory of evolution since no tabulation of individual instances could
possibly lead to definite conclusions.
However, what the new emphasis on Mendelian germ structure implied was the
rejection of the “descent with modification’ account of heredity that characterised
Darwinism. In this latter context ancestry was a crucial feature of heredity and
the biometricians went some way toward providing a quantitative analysis of the
resemblance of individuals based on common ancestry. The law of ancestral hered-
ity (in the form of a multiple regression equation) stated the contribution from
each ancestor to the constitution of the individual. No mechanisms for explaining
heredity were postulated, nor were they desirable. It could be completely under-
stood in terms of descent or lineage and correlation coefficients. In contrast to
this Mendelism focussed on the genetic structure of the individual, particularly
the composition of the zygote with the structure of the parents uniquely deter-
mining the structure of the offspring. Once this was established ancestry was of
no importance. And, as R.A. Fisher later pointed out, the methods for predicting
the appearance of certain traits were fundamentally different.
With its emphasis on statistical analysis of large populations the biometri-
cians/Darwinians could fairy accurately predict the ways in which populations
would evolve. Because Mendelism used a different kind of analysis it could ac-
curately predict the offspring of particular parents but not of populations as a
whole. However, if Mendelism was to provide the mechanism of heredity that was
absent in Darwin’s theory then any fusion of these two would require an account
of how the structure of Mendelian heredity can be analysed for populations. In
other words, because selection acts on populations and not individuals we need an
account of how to reconcile the individualistic nature of Mendelism with the more
broadly based theory of selection. In addition we need a way of understanding,
from the point of view of selection, how genetic structure changes; that is, how to
account for mutation. Answers to these questions emerged in the period between
1918 and 1930 in the work of R.A. Fisher, J.B.S. Haldane (1892–1964) and Sewall
Wright (1889–1988). But, before this there were developments that would prove
crucial for Mendelism, especially the formulation of the Hardy-Weinberg law.
In 1902 George Udny Yule (1871–1951) showed that two populations A and a,
that are the same size will reproduce the same population structure in successive
generations (1A:2Aa:1a) provided there is random mating. Prior to this Yule had
studied engineering at University College and Physics under Hertz in Bonn. He
accepted a position as Pearson’s assistant in his lab at UCL 1893 and there began
his work in statistics. In the 1902 paper Yule assumed the total dominance of
one character which resulted in the population remaining in equilibrium with a
ratio of 3:1 dominants: recessives. A similar result, but without the assumption
of dominance, was produced by Pearson in 1904. But, both Pearson and Yule
considered only two Mendelian factors present in equal frequencies. It was the
geneticist R.C. Punnett, sceptical of the dominance assumption, who in 1908 asked
Godfrey Hardy (1877-1947), an English mathematician, to investigate the problem.
He showed that regardless of the initial frequencies of the forms AA, Aa, aa, a