EDWARD BUNNELL PHELPS, M. A., F. S. S. 169 



namely, Maryland, figures in those statistics. Not only are mor- 

 tality statistics for nearly one-half of the population of Conti- 

 nental United States therefore unavailable, but as our national 

 statistics for the Registration Area as yet present no birth re- 

 turns except for Census-taking years it is a physical impos- 

 sibility to compute the ratio of deaths under 1 year to living 

 births on the one universally recognized basis, even for our 

 Registration Area. To be sure, the majority of the Registration 

 States do issue annual compilations of their respective vital stat- 

 istics, and most if not all of these compilations now include 

 tabulations of living births, and deaths at the various ages. Even 

 the oldest and most reliable of these State systems of registra- 

 tion of births and deaths, that of Massachusetts, has been unable 

 to round out complete annual returns of births, and in the 

 Sixty-Seventh Report of Births, Marriages and Deaths in Mass- 

 achusetts, for the year 1908, a frank admission of and partial 

 explanation for this fact are made in these words (p. 142) : 



"Although the law applies to the registration of births, as well 

 as to that of marriages and deaths, it is probable that the stat- 

 istics of the births are less accurate than those of either of the 

 other two classes. From the nature of things, marriages and 

 deaths must be registered, in order that the former may be 

 solemnized, or that interment be possible in case of deaths ; but 

 in the case of the births, the inadequacy of penalty for neglect, 

 ignorance of the law, as well as topographical conditions, tend 

 to an incomplete registration. It is therefore likely that the 

 number of births returned in Massachusetts in 1908 was less 

 than the actual number which occurred; hence a lower birth 

 rate, and comparisons between births and deaths inaccurate." 



In many, if not most, of the other States which purport to 

 present annual birth statistics, the registration of births is far 

 more defective than in Massachusetts, and as an inevitable 

 result of the incomplete returns of births the infant mortality 

 rates or ratios of deaths under age 1 to living births pre- 

 sented in the annual reports of these States can only be taken 

 in a Pickwickian sense, so to speak. The divisor being too small 

 in every case in some cases materially under the proper figure 

 of course the resultant, and apparent infant mortality rate, is 

 above the actual rate. As the years roll by, the birth registration 

 is doubtless improving in most cases, the margin of error is 

 therefore continuously changing, and hence attempted compari- 

 sons of the apparent infant mortality rates of recent years with 

 those of earlier years are more or less misleading. But, as this 

 most glaring defect in our American system of registration 

 of vital statistics, and most serious obstacle in the way of secur- 

 ing correct figures of infant mortality in this country, has 



