NATURAL SELECTION IN MAN 191 
this figure can be considered only as a rough approximation to 
the truth. 
It is a remarkable fact that while the death rate of most civi- 
lized countries has been falling for the past hundred years the 
infant death rate in general should have suffered little improve- 
ment and in some countries actually increased up to the beginning 
of the 2oth century. During the past few years much greater 
attention has been devoted to the subject, and a variety of organ- 
izations have been active in checking the inexcusable loss of infant 
life which has been so long suffered to go on, and as a consequence 
infant mortality in many localities has very rapidly fallen. In 
the same country enormous differences in the infant death rate 
still exist in different towns and sections not far removed from 
each other, as may be illustrated by the infant mortality rates of 
the following towns of Massachusetts in 1912: 
Chicopee stains eueseseeeuues 177 
North Adams.................. 113.1 
Wealthamiine:. isi ues Seen ce 86.8 
Brookliné 2 2csc.seseses onde 55 
These conditions are usually associated with the economic 
status of the inhabitants. The death rate is higher in urban 
than in rural districts, and it increases in cities with the greater 
density of the population. 
In all places infant mortality is very much higher among the 
poor. In fact Mr. Ashby states that “poverty is perhaps the first 
and greatest predisposing factor in infant mortality.” Duncan 
and Duke in their valuable survey of the infant mortality of 
Manchester, N. H., find that the rate of infant deaths rapidly 
falls as the income of the father rises. Where the annual earnings 
of the fathers are less than $450, the infant mortality rate was 
found to be 242.9. Fathers earning from $650 to $850 lose 
162.6 per thousand of their children, while those earning $1,250 
and over, lose only 58.3. Among the foreign born mothers of 
Manchester the death rate was 183.5, while among the native 
