Sewall Wright 89consequences of various pedigrees involving mating of relatives and the consequent
decrease of heterozygosity. One problem was repeated mating of brother and
sister. The usual method was to grind out the detailed consequences for a few
generations and hope to find a pattern. In later generations, the problem soon
beomes unwieldy. One well-known geneticist got it wrong by not carrying the
process through enough generations. When Wright was starting his graduate work
at Harvard, his roommate, H. D. Fish, was working on this problem. Fish had
covered not only his desk but much of the floor with calculations. Wright thought
there must be a better way, and he found it.
He devised the now-familiar algorithm which uses a simple process of tracing
paths of ancestry from one parent to a common ancestor and back to the other
parent. The whole process is easily learned and routinized, so it is readily adapted
to computer use for complicated pedigrees, including tens of thousands of records
for an entire breed of livestock. The inbreeding coefficient is a measure of the
proportion by which the heterozygosity of an individual is reduced by inbreed-
ing. Although Wright derived his algorithm by using correlation coefficients, this
derivation has been largely replaced by using the concept of “identity by descent
(IBD)”. Two homolgous alleles are IBD if they are both descended from an al-
lele in a common ancestor, or one is descended from the other. This has been
extensively used in animal and plant breeding and recently, with the coming of
molecular chromosome markers and computers, the concept is now widely used in
the study of human genetics.
Wright made two important applications of inbreeding theory while he was at
the USDA. The first was experimental. He inherited a colony of guinea pigs that
had been maintained, usually with close inbreeding, for many generations. He
analyzed the effects of inbreeding on a variety of important traits, such as size,
vigor, and fertility. The analysis was masterful and these papers, published in
1922, have stood the test of time; they can be read with profit today[Wright,
1922b].
The second application was to the study of breeds of livestock. Wright sub-
divided the amount of inbreeding into components such as local consanguineous
mating and herd differences. He found that almost all the increased homozygos-
ity was due to random differences between herds rather than non-random mating
within a herd. He also found considerable genetic differentiation among herds. His
analysis of the extensive records of Shorthorn cattle is classic.
Wright extended this work into what are called F-statistics. The inbreeding
coefficient is generalized to include population structure, particularly of a hierar-
chical sort. These measures are widely used in the study of natural populations,
including humans. Both the cattle studies and later theory are summarized by
Wright in a 1951 paper[Wright, 1951].
These studies led Wright to what he regarded as his major accomplishment,
the “shifting balance theory” of evolution. Wright noticed that most of the breed
improvement arose by random differences among the herds, followed by exporting
bulls from the best herds and thereby upgrading the whole breed. His evolutionary