249 G. H. KNIBBS. 
dren) then equal increases in the rate of infantile mortality 
tend to be followed by equal though relatively small 
increases in the birthrate.' 
(ii) The influence of infantile mortality on the birthrate 
must always be very small, 
5. In order to test these propositions, the appropriate 
graphs for a number of countries have been prepared. In 
these the infantile mortality rate is taken as corresponding 
to the year following the birthrate. Widely divergent as 
are individual points, the results on the whole are definite 
enough. The graphs are shewn in Fig. 1, and the results 
are as shewn in “‘ Table I’? hereunder. 
In order to get a fairly wide range of values of infantile 
mortality and birthrate the period covered has in most 
cases been 25 years: but in addition to this the data for 
all countries have been plotted for a single year, this imply- 
ing the relation which might be assumed to exist for the 
group of countries, considered asa whole, at the beginning 
of this century. 
It may be remarked as curious, that the relation for the 
various countries tabulated for 1901 (1902) is sensibly 
identical with that for the period 1887 —1905 (1888 — 1906) 
for the Commonwealth of Australia. 
6. Looking through the table it isseen that the birthrate 
for Switzerland and Jamaica are sensibly unaffected by the 
rate of infantile mortality, while for Ireland, England and 
Wales, and Scotland, it actually decreases with increase of 
infantile mortality. The fundamental assumption conse- 
quently, however correct on the whole, may clearly be 
modified by other factors, the tendency of which in some 
other direction, may preponderate sufficiently to mask that 
+ See last paragraph Section 8 hereinafter in regard to the limitation 
of this proposition. 
ey 
