198 THE TREND OF THE RACE 



show that in a region that has a high infant death rate, the death 

 rate for older children is higher than it is in some other region 

 with a low infant death rate. What we want to know is whether 

 the child death rate is less than it would have been tmder the same 

 conditions if the infant death rate had not been so high. If it 

 should be found that a high infant mortality is generally followed 

 some years later by a reduced child mortality of the same group 

 and under the same environment the evidence would point to the 

 selective value of early mortality. 



An investigation of this problem was made by Mr. E. C. Snow 

 whose memoir on The Intensity oj Natural Selection in Man con- 

 tains evidence of much painstaking and critical labor even though 

 it may leave something to be desired in the way of lucidity of expo- 

 sition. The data for one study were taken from the Reports of the 

 Registrar General for England and Wales, and those for another 

 were obtained from the vital statistics of Prussia. Correlations 

 were worked out for various districts of England and Prussia 

 between the mortality of early life (1-3 years in different cases) 

 and the mortality of subsequent age intervals. After many 

 corrections for environmental differences and the variable sizes of 

 the cohorts, the data were found to show a negative correlation 

 between the death rates of early periods and those of later periods 

 of life. In other words, a relatively high death rate in the first 

 period renders the death rate of the survivors in the subsequent 

 period less than it otherwise would have been. Such a result is 

 not inconsistent with the conclusion stated previously, that cer- 

 tain regions have a relatively high death rate for several succes- 

 sive years. There may be a more severe selection all through life 

 in one group than there is in another. 



It would be a matter of interest to ascertain, though the 

 problem would present many difficulties, whether the death rate 

 tends to be less selective, or in other words more indiscriminate 

 as we approach the period of birth. A priori, this would seem 

 to be very probable. There may be some truth in Dr. D. S. 

 Jordan's statement that *'a strong child can be killed almost as 

 readily as a weak one when it is very young," and it is when 



