EFFECTS OF INBREEDING AND CROSSBREEDING. 29 
fact that these experiments gave very much less increase than AC 
confronts us with something of a puzzle. The difference is too great 
to be dismissed as due to chance (5.8X PE). The probable explana- 
tion is a reciprocal physiological relation between frequency and size 
of litter. It has, in faet, been found in the inbred stock that when 
one litter follows immediately after another (1. e., in 9 or 10 weeks, 
the gestation period), it is smaller than the average, and that a 
large litter predisposes toward a delay before the appearance of the 
next litter. (See Part I, Bulletin 1090.) These negative relations 
were not very strong, and in the more vigorous random-bred stock 
the relations were positive. We doubtless have, however, as sug- 
gested in the case of the relations between mortality at birth and 
mortality between birth and weaning, an unstable balance between 
opposed influences. It was shown in Part I that the records of the 
inbred and control stocks rose and fell in parallelism from year to 
year in these two elements of fertility (as well as in all other elements 
of vigor). In the present paper it is shown that there is considerable 
agreement between frequency and size of litter in the rise and fall 
from season to season during 1916 to 1919 (Figs. 4 and 5). Thus 
external conditions tend to produce a positive correlation. ‘There 
may also be common genetic factors which tend the same way. 
The apparent conflict between the evidence from random-bred and 
inbred stocks, referred to above, merely means that in the random- 
bred stock the causes of positive correlation were not completely 
- balanced by the reciprocal physiological relation suggested above, 
while it was overbalanced in the inbred stock. One would expect 
to find the physiological relations more important in the weaker 
inbred stock. 
In comparing AC with CC or Cl we have experiments in which the 
inherent characters of the dam and young are essentially the same. 
Owing, as it appears, to the influence of the crossbred sire, CC and Cl 
produced litters distinctly more regularly than did AC. We have 
here a situation in which a negative physiological correlation could 
reveal itself uncomplicated by any positive correlation. The high 
record of AC relative to CC and C1 in size of litter and the opposite 
relations in frequency of litter are the expected results on this 
hypothesis. 
Similarly the relatively high record of CA in frequency of litter 
may exert a slight depressing influence on its record in size of litter. 
The argument that the sire exerts some influence on size of litter is 
strengthened, but hardly enough to be relied upon. 
We may conclude that crossbreeding causes a marked increase in 
the size of litters produced by females. Crossbreeding of the sire 
may have some influence, but too slight to be demonstrated by the 
present data. The heredity of the young appears to be wholly with- 
out influence. It may be ‘added that selection of the dams (and 
