16 BULLETIN 1121, U. S. DEPARTMENT OF AGRICULTURE. 
first cross, though marked improvement in the next generation. The 
percentage born alive, birth weight, size of litter, and frequency 
of litters come here. In the cases of percentage raised of those born 
alive, and gain, there is on the other hand marked improvement in 
the first cross. 
THE PERCENTAGE BORN ALIVE. 
A number of points of interest are brought out in Figure 8 in refer- 
ence to the mortality at birth. It is shown that there is considerable 
variation among the inbred families, but that no one of them has as 
good a record as the random-bred stock (B). There is only a slight 
improvement, too small to be relied upon, when two families are 
90 
4 
466 CLN7. 
Le) 
75 
Fia. 8.—The percentage born alive, 1916-1919. Indices correcting for effects of size of litter and seasonal 
conditions: 2, 13, 32, 35, 39, inbred families; OI, other inbred families; A, average ofall inbreds; CO, first 
cross between inbred families; CA, crossbred male from C0, unrelated inbred female; AC, inbred male, 
unrelated crossbred female from C0; CC, crossbred male, unrelated crossbred female; Cl, crossbred 
brother and sister from C0; C2, brother and sister from C1; CL, selection for large litters, small weight; 
CG, selection for small litters, large weight; B, random-bred stock. 
crossed (CO). When, however, the crossbred females produce young, 
whether mated with an unrelated crossbred (CC), with a brother 
(C1), or with an unrelated inbred (AC), there is a substantial im- 
provement which goes beyond the best of the inbred families and 
reaches practically the level of the random-bred stock. In contrast 
with the good record of Experiment AC is the relatively small im- 
provement over the inbreds in the reciprocal cross CA, in which 
inbred females are mated with unrelated crossbred males. It is clear 
from these results that crossbreeding increases the number of young 
born alive, but that the breeding of the dam, not that of the young 
themselves, is the critical factor. The most significant line of cleavage 
between high and low records falls between Experiments A, CO, and 
