The genetical structure of populations
Annals of Eugenics, 15 (1):
323--354 (1949)DOI: 10.1111/j.1469-1809.1949.tb02451.x
Abstract
I propose to discuss certain genetic aspects of population structure. The term is used to include
such matters M numbers, composition by age and sex, and state of subdivision. The best starting
point is the consideration of the situation in a large random-breeding population, in which structure
in the last sense is absent.
... It has probably occurred to the reader that the coefficientof inbreeding may mean very different
things in different cases. (1) There may be divisionof the population into completely isolated small
strains, within each of which there is random mating. The inbreeding coefficient of individuals
relative to the total is here due merely to the relationship of all members of the same strain and
disappears at once with random mating among strains. (2) There may be fiwquent mating of close
relatives but no permanent separation of strains. Here again random mating at once reduces the
inbreeding coefficient to zero. (3) The sires used may be rather limited in number and derived from
even more limited numbers of grandsires and great-grmdsires. In this cam there may be little
apparent close inbreeding at any time, contrary to (2), and no division into strains, contrary to (l),
but the value of F, relative to the foundation stock, keeps rising and cannot be much reduced by
random mating. There are other possibilities. Clearly we need something more than a single
value of F to give an adequate description of structure.