Vle CONG. INTERN. REPROD. ANIM. INSEM. ARTIF., PARIS, 1968, VOL. lI

Inbreeding in artificial insemination.

A . ROBERTSON
Institute of Animal Genetics, West Mains Road, Edinburgh, Great Britain.

Before we discuss the effect artificial insemination has had upon the
structure of our animal populations it might be informative to consider
first the situation before the days of A.l. Most discussions of inbreeding
are founded upon the pioneer work of Sewall Wright, who very clearly
analysed its consequences and proposed methods by which its probable
effect could be estimated from pedigrees. His ideas may be summarized
briefly as follows. Due to the appearance of the same ancestor in the
pedigree of both the sire and the dam of an animal, the animal has a
m easurable chance of receiving genes from both its parents which are
copies of a gene carried by this common ancestor. This probability
(usually specified in terms of the chance of having received replicate
genes from ancestors within a specified number of generations from the
animal concerned) is usually known as Wright's coefficient of inbreeding.
If we take as a simplerexample the mating between two animals who have
the same sire, then half the genes in both will come from this common
sire. Considering now the genes at a particular locus that they pass on
to their progeny, on one quarter of occasions both genes will have come
from the common sire and on one half of these a replicate gene will be
transmitted by both parents. Thus the probability that the offspring is
homozygous because of his common grandsire is 1/8. If complete pedi-
grees are available it is in principle simple but in practice difficult, if
the pedigree goes back for many generations, to calculate the inbreeding
coefficient. It perhaps should be stressed that the coefficient is in no
w ay an absolute figure but really applies to a pedigree rather than to an
animal. It does not measure the homozygosity of the animal with this
pedigree but the increased probability of homozygosity (compared to
that of the remote ancestors in the pedigree) due to the inbreeding. In
practice, calculations of inbreeding over many generations are far too
time consuming, and Wright himself invented a method of working out
the average degree of inbreeding of a population by looking not at whole
pedigrees, but at the chance of occurrence of common ancestors in two
lines drawn back at random, one from the father and the other from the
m other from the animal in question (1). This method has been used on
a great many breeds - in particular Wright himself did a famous study on
the Shorthorn breed, in which there is a well-known concentration of
blood some 150 years ago so that the inbreeding at the present time
compared to the remote base population is about 25%. Wright's method
can be made more accurate for a given amount of labour by writing down
several lines within the same pedigree and by including both parents at
each place in a randomly chosen line (2). More recently the use of
computers has enabled us to work out the inbreeding coefficients in com-

1317

32