EDWARD BUNNELL PHELPS, M. A., F. S. S. 171 



of deaths at all ages in that area. Such a comparison for the 

 10 years, 1900-1909, I have worked out and present in Table II 

 attached to this paper (and also in Chart II), and, without tak- 

 ing up in detail the figures which may there be found, the 

 general effect of that tabulation may be here shown by the cita- 

 tion of the average ratios on these lines for the two five-year 

 periods, 1900-1904 and 1905-1909. In the former quinquennial 

 period, the deaths under 1 in the Registration Area amounted 

 to 19.2 per cent, of the total deaths at all ages in that area; in 

 the latter quinquennial period, the ratio had risen to 19.5 per 

 cent, thus showing an apparent increase rather than decrease. 

 As was stated in the note from Dr. Wilbur's current report 

 which I have just quoted, however, this method of comparison 

 is "a very crude means of judging of the condition," and its 

 credibility depends upon the assumption that age distribution 

 and population conditions were substantially similar during the 

 periods of comparison. But practically no other means of even 

 attempting to measure the rise or fall of the infant mortality 

 rate for the Registration Area as a whole is possible, and there 

 are some reasons for believing that, in the main, the ratio of 

 infant deaths to deaths at all ages affords a fairly reliable index 

 of the infant mortality situation under normal general condi- 

 tions. In the case of the comparison of the ratios of 1900-1904 

 and 1905-1909, the different conditions must be noted. From 

 1900 to 1905, inclusive, the Registration Area remained prac- 

 tically unchanged, no additions of area being made, whereas in 

 1906 the States of California, Colorado, Maryland, Pennsyl- 

 vania and South Dakota were added to the Registration Area, 

 thus increasing the population of that area by more than 7,000,- 

 000, or nearly 20 per cent. The addition of the five States in 

 question materially increased the urban population of the Reg- 

 istration Area, and as the infant death rate in the cities is in 

 general considerably larger than that of the rural districts this 

 radical change in the make-up of the Registration Area might 

 confidently be expected to send up the infant death rate of the 

 area in question. But from 1900 to 1905, inclusive, the mor- 

 tality statistics of the Registration Area dealt with precisely the 

 same territory, hence are fairly comparable, and the registration 

 returns for infant and child mortality for that period, as pre- 

 sented in Table II attached to this paper (and in Chart II), 

 are worthy of careful study by all interested in the problem of 

 infant mortality. In the last of the six years in question, 1905, 

 the deaths under 1 year were fewer by more than 6,000 than 

 those in the first year of the period, 1900, the deaths between 

 1 and 5 years showed an even larger decrease, one of more 

 than 10,500, and of course the total of deaths under age 5 



