REMEDIAL THEORIES AND EXPERIMENTS 223 



tion carried on by interested officials. Infants' milk 

 depots have passed the experimental stage ; and there 

 is no good reason why the cities should shirk their 

 responsibility in this matter and leave the work to 

 private enterprise to be done well or ill, completely 

 or incompletely, according to the will or the mental 

 or physical equipment of those undertaking it. 



IX 



There has been so much statistical juggling with 

 regard to the influence of infants' milk depots upon 

 the infantile death-rates, alike by friends pleading 

 their cause and opponents attacking them, that I am 

 tempted to leave the matter alone, except in so far 

 as it has been touched upon in preceding pages. I 

 know of no subject in connection with which figures 

 have been used with more recldessness than that of 

 the influence of infants' milk depots upon the mor- 

 tality of infants. So many factors, which no statis- 

 tician can properly evaluate, enter into the infant 

 mortality rate that unless the greatest care is exer- 

 cised wrong conclusions are almost inevitable. For 

 example, in a certain city there has been, during a 

 period of five years say, a marked decline in the 

 number of infant deaths, and it happens that the 

 period of decline is coincident with the period during 

 which an infants' milk depot has been in operation. 

 It is not unnatural that in such circumstances the 



