302 THE INDIAN JOURNAL OF VETERINARY SCIENCE AND ANIMAL HUSBANDRY [VI, III

varying degrees of knowledge about each. These illustrations have been used to
show what kinds of knowledge are needed,.and how one should proceed to make a
business-like estimate of the probable gains to be derived from various methods of
selection, progeny-testing, etc. Not until such a valid estimate is made, is one in
a position to decide whether a given procedure is likely to provide gains that
will more than offset its costs. In preparing these estimates, there is
work in plenty to be done in measuring reproductive rates, selection differentials
reasonably attainable, the genetic portion of the observed variance, etc.

In addition to the genetic variance that can be expressed by an additive
scheme, there is a portion (perhaps a very large portion in stocks that are already
improved) of the variance which is genetic in a broad sense, but cannot be
expressed additively because the gene combination interact in other ways. Such
gene interactions include inhibitory, multiplicative, complementary and epistatic
effects, and also what may be a very common class of cases where that which is
optimum in the breeder's opinion is genetically intermediate.

[NLS note: a graphic appears here - see image of page]

         SCALE OF THICKNESS OF BACK FLESH.
Diagram illustrating the case where the maximum of desirability may
be produced by an intermediate combination of genes. In such a case
selection usually produces noticeable results in the first few generations
it is practised, but rather quickly approaches a limit beyond which
progress is not made, although hereditary variability still remains.
Further progress can often be made if some combination of inbreeding
and outbreeding is practised along with continued selection.