4 BULLETIN 1121, U. S. DEPARTMENT OF AGRICULTURE. 
was entered as mature under the month in which the male reached 4 
months of age or under the month following mating if he was already 
3 or more months old. The mating was dropped in the month follow- 
ing the death or disposal of the female. The number of months of 
mating in the experiment, divided by 12, gives the number of mating 
years with sufficient accuracy for use in calculating the average num- 
ber of litters per year, young per year, and young raised per year, 
which are given in the last three columns of Table 14. 
Table 15 presents data on the percentage of the young born alive 
(or at least found alive) in each experiment during the years 1916- 
1919. The total percentage is given in the next to the last column. 
This is not necessarily the best measure of the relative standing of 
the experiments in success in bearing living young under fixed con- 
ditions, as has been discussed in Part I. The percentages are accord- 
ingly given separately for each size of litter and an index is given 
(last column) for the purpose of showing the standing of the experi- 
ments free from the influence of size of litter. In the present paper 
this index is derived by assigning weights of 1, 3, 3, and 1 to the rec- 
ords of litters of 1, 2, 3, and 4, respectively. 
The desirability of using such an index may be seen by comparing © 
the records of the first two inbred families (Nos. 2 and 13). Family | 
13 is superior to Family 2 in the percentage born alive in each size | 
of litter except in litters of 4, yet in total percentage born alive it | 
comes out markedly inferior. The explanation is merely that Family | 
13 produced a greater number of large litters. The index brings out | 
better the true relation under constant conditions. 
So far as the crossbreeding experiments are concerned, it makes | 
very little difference whether the total percentage or the index is 
used. This is because size of litter has less effect on mortality of 
the young than among the inbreds. | 
There are, of course, other conditions which affect the percentage 
born alive, the uniformity of which must be considered before making | 
a final interpretation of the results. The more important of these > 
conditions will be taken up later after consideration of the other | 
tables. | 
The data on the percentage raised to 33 days of the young born 
alive are presented in a similar way in Table 16. The index given in 
the last column is derived in the same way as described above. 
The product of these percentages, the percentage raised to 33 days 
of all young born, is treated in the same way in Table 17. 
In the case of birth weight (Table 18) the effect of size of litter 
is so great that the actual average in an experiment means very | 
little. The average birth weight in one of the inbred families (No. 
35), for example, is greater than that in Experiment AC (72.6 com- 
pared with 72.1). This, however, is not as significant as the fact 
