250 G. H. KNIBBS. 
the actual. This indeed is suggested by the fact, that 
while the weighted mean gives 
8 = 2276-03334. eee (10) 
the results for various countries for 1901-2 gives 
0 = 19°4 --"0°083 jazceeee ore (10a) 
which, if it means anything at all, would imply that the 
correction is becoming more significant. For individual 
countries the factor b could be taken from Table I, and the 
birthrate corresponding to an absence of infantile mortality 
as given by 
b. =10"— Ui ae (11) 
in which b is the tabular factor, andthe rate of infantile 
mortality for the preceding year. 
12. The actual deductions made as to the influence of 
infantile mortality on the birthrate of any country should 
perhaps be regarded merely as a general indication. The 
theoretical deduction gives the elementary law, but to 
obtain the actual coefficient b it would be necessary to be 
assured that no other source of variation in the relation 
investigated, was operating. A thoroughly satisfactory 
investigation demands the proper subdivision of population 
into homogeneous groups, and the elimination of other 
influences. The present investigation should be regarded 
as an indication of the way, rather than as a final solution. 
a 
