580 FRANK A. HAYS 
In order to make the litters in the different service groups 
comparable with each other, whatever their number, it was 
thought first that litters of different numbers of individuals 
could be standardized to a mean litter number. Jackson (13, 
p. 17) in comparing the standard deviation of individual rate 
with the standard deviation of the entire race, reduced all indi- 
viduals to a common basis by multiplying the body weight of 
each rat by a factor obtained by dividing the mean of the total 
population at a given age by the mean of the given litter. Since 
the object we have in view is not the study of individuals, 
this formula cannot be used. Further attempts were made to 
obtain a factor for reducing large and small litters to a compar- 
able basis, but so far with no success. Again, it was thought 
possible that the coefficient of correjiation between number of 
individuals in the litter and average weight per individual might 
be made use of to reduce the litters to a comparable basis, 
but without any satisfactory results. Again, a comparison of 
different-sized litters in the several service groups by con- 
structing graphs upon a base line representing the different 
litter numbers and the vertical line representing the variable 
weights at birth, a second chart to show the time required to 
make eight times the birth weight, and a third chart to show the 
time required to make twenty-four times the birth weight were 
attempted. By this means the data could be much condensed, 
but such a system proved to be impracticable and was discarded. 
Finally, it was deemed best to compare litters of the same num- 
ber of individuals. Accordingly, the growth rate in the different 
service groups must be shown by a whole series of charts, the 
graphs on each chart representing a certain litter number. In 
each case the chart shows the number of litters which are lumped 
in each graph. Thus each chart gives a direct comparison of 
the growth rate of the five service groups, namely, Ist, 5th, 
10th, 15th, and 20th, the comparison being always between 
litters of the same number. 
As a further measure of divergence in rate of growth between 
the service groups, the coefficient of variability of weight for all 
litters in each of the five service groups is valuable, presented 
