Published online by Cambridge University Press: 06 May 2016
In the analysis of the causes of variation between cows of a given breed in milk production, differences between herds pose some difficult problems. The nature of this variation is of importance from several points of view in the discussion of breed improvement. The situation is complicated by the fact that the herds of any breed type do not form a uniform group but can be classified roughly into a sort of ‘social structure’ (Robertson and Asker, 1951). The discussion of variation between all herds of a given breed type can then be broken down into the variation between strata and that between herds in the same stratum. From the analysis of production data from herds using artificial insemination (Robertson and Rendel, 1954; Korkman, 1953) there is evidence that the genetic component is a small proportion of the total variance in these herds; and also that in Great Britain the genetic difference in yield between the dominant group of herds (which provide many of the bulls used in A.I.) and the herds using A.I. is small. This is in agreement with studies on breed structure. In this paper we shall discuss the genetic differences between members of the small group of important herds which dominate each breed. These usually supply a high proportion of the bulls used. If we have good estimates of the breeding values of such bulls from the production records of their progeny, we can then examine the different breeding herds according to the breeding value of the bulls bred in them. Evidence of this type has been presented for Jersey bulls in New Zealand (Castle, 1952) but without statistical analysis of the results, although inspection suggests that there are definite differences between studs.