Vie CONG. INTERN. REPROD. ANIM. INSEM. ARTIF., PARIS, 1968, VOL. II

be expected in an ideal population - a "random breeding population” in
w hich all matings are equally likely and all animals have an equal chance
of contributing progeny to the next generation. This is not a particularly
good model for a domestic animal population because most such popula-
tions before the days of A.l. had a kind of social structure. The simplest
w ay to describe this is in terms of an internal nucleus within which all

m ales for the whole population are bred but which may take in females
from the rest of the population. The rate of inbreeding can then be pre-
dicted from the number of males used to breed sires each generation. In
the days of natural service a more satisfactory kind of description would
be in terms of three or even four layers forming a pyramid within the
breed. There is then a flow of genes, by the usage of males, from the
higher to the lower levels every generation.

Atwo layer system or, in the other terminology, one with an
internal bull breeding nucleus - is a very good model for a dairy cattle
population depending on progeny testing in artificial insemination for
improvement. The standard rate of inbreeding found under natural
service in the past - 0.4% per generation - would then arise in such a
system in which there were eight males used to breed males each gener-
ation. This would correspond roughly to an input of two such bulls into
the bull breeding nucleus every year.

Letme digress a little to discuss another aspect of inbreeding in
animal populations. This concerns the setting up of some sub-populations
in which the degree of inbreeding is deliberately kept as low as possible.
This is done in order to minimise the genetic change from generation to
generation in order to use such populations in the measurement of genetic
change in other selected strains. The principles of this are now well
w orked out and control populations in poultry have existed for some
twenty years. Methods for the measurement of genetic change in large
animal populations have so far been only statistical and, because of
difficulties in the interpretation of some of these results, | think it would
be worthwhile taking deliberate steps to measure genetic change in pop-
ulations of dairy cattle using artificial insemination and in populations
of pigs in national breeding programmes.

What has happened to the structure of breeding since the introduction
of artificial insemination? It has in a way cut right across the social
structure and enabled bulls bred in fashionable herds at the top of the
breeding pyramid to be used in herds at all levels. One consequence of
this (not particularly unexpected because of the previously existing breed
structure) was the demonstration in Great Britain that bulls bred in these
top herds were not in fact genetically superior for milk yield to the bulls
w hich were already being used in the commercial herds (6). The same
data were used to show that within the commercial herds there was little
connection between level of production and genetic merit. This lack of
genetic difference between herds, confirmed by direct experimentation

1319