Their Application to Live Stock. 
61 
rule it will be advisable not to attempt to work with too many 
characters at a time, as this tends to complicate the experi- 
ments considerably. Where practicable, two pairs only should 
be chosen, as this will limit the numbers to be raised from the 
cross-breds to reasonable dimensions. 
The data obtained from the first crosses will as a general 
rule be sufficient to determine whether one of a given pair is 
dominant or recessive, or whether the result of their union 
is to give an intermediate. A third possibility that the cross- 
bred may show a reversionary form is not considered here, 
as such cases appear to be rare amongst animals in which 
the breeder will have any interest. The next step is to mate 
together the cross-breds. Many breeders will feel that this is 
a mistaken policy, for the idea is prevalent that the offspring 
of such a mating is degenerate. 
This is doubtless due to the fact that among the offspring, 
particularly where the parents differ in numerous characters, 
many types appear, and this has been taken as an expression 
of the fact that the original types have been hopelessly broken 
up — a matter, of regret in the days when pedigree was far more 
important than Mendelists now consider it to be. Nevertheless, 
as it is in this generation that the sorting out and the recombi- 
nation of the characters present in the original parents occurs, 
this mating is essential. As large a number of cross-breds as 
practicable should be used in order to obtain a sufficient number 
of animals to choose from. Thus, if two pairs of characters 
are under consideration, on an average the four possible com- 
binations in a stable form will only occur amongst sixteen 
animals, and this number must be more than doubled in order 
to secure males and females of each type to test the purity 
and establish the breed. The animals not considered desirable 
should be disposed of, and those chosen for further testing 
should be examined character by character. A comparison 
with the experiment described in Fig. 3, or the case following 
it, will at this stage help to show the chances of obtaining the 
type in a fixed condition. If it shows a combination of two 
recessives, then, mated with its like, it will breed true. If a 
recessive and an intermediate, it will be fixed as regards one 
feature, the recessive, but when bred with its like it will throw 
intermediates, together with the characters as seen in the 
parents. For instance, in the case of the sheep-breeding 
experiment, if we assume that two of the animals are pure 
as far as the horned character goes, but have speckled faces, 
these, if bred together, would give in their offspring horned 
black, white, and speckled faced individuals, and pure races 
with black or white faces could be bred from these, but not 
a race with speckled faces. 
