THE CAVENDISH LECTURE 1 3 



to man, and that environment, not constitution — nurture, not 

 nature — is the sole thing that social reformers need pay attention 

 to. It is astonishing that two statisticians of such repute should 

 not have recognised that practically no result of any importance 

 could be deduced from the fact that in the same neighbourhood 

 and for the same period the death-rates for two classes of the 

 community would both be high or both low ! How could the 

 deaths of infants in this year possibly influence the deaths of 

 children of three and more in the same year ? Darwin's theory 

 means this, that if individuals are reared under a constant 

 environment, and a larger percentage of them are killed off 

 in the first year of life, then a smaller percentage of those 

 remaining will die in the later years of life, because more of the 

 weaklings have been killed off. Dr. Newsholme's position is, if I 

 grasp it, this : that a high death-rate means a high rate of sickness, 

 and this sickness produces a general weakening in those that 

 do not die, so that they fall an easier prey in later years. In other 

 wcjrds he wholly rejects Darwin's position that a hard winter or 

 general scarcity would leave a stronger race of birds behind. 

 Whether Newsholme or Darwin held the truer view could not 

 possibly be tested upon Newsholme and Yule's data. You must 

 follow each array through its life and see whether increased infan- 

 tile death-rate means decreased child death-rate for the same group. 

 This method has actually been followed by Dr. Snow, working in 

 my laboratory. He has taken English (Fig. i) and Prussian data 

 (Fig. 2), and he has endeavoured to follow the same group from 

 birth through childhood. The difficulties of the problem arise 

 from ensuring that the environment shall be alike for all groups 

 dealt with. Dr. Snow has attacked the equality of environment 

 from a number of sides, but the general conclusion is substantially 

 the same : IVhen the itifa?itile death-rate is intensified, then the 

 death-rate in childhood of the survivors is /esse?ied, not increased. 

 In other words, a heavy infantile death-rate does select the weaker 

 individuals and leave a stronger population physically to the later 

 years of life. It is difficult, indeed, to believe that it could 

 possibly be otherwise, and I have every confidence in my 

 view that Darwinism applies absolutely to the case of man. 



Dr. Newsholme states definitely that it is a general rule that a 

 high infantile mortality means a high child mortality. Now the 

 Registrar-General's office has provided us with a series of life 

 tables which lead to the following results : 



