threshed out as to call for no comment here. But, to a certain extent, the defects of one country's registration system would be offset by the comparative perfection of that in force in some other country, in a tabulation of world-wide scope, and in the case of the twoscore countries and States dealt with in the last table it is not improbable that the percentage of error is substantially uniform. Even were the inquiry restricted to the two English-speaking sections whose registration systems are generally regarded as freest from defects-to wit, England and Wales on one side of the Atlantic and the State of Massachusetts on this side of the water-the variations in their infant mortality rates in the last twenty-five years and the averages for the entire period differ but slightly, the infant death-rates in England and Wales for the five latest five-year periods having been in the order of 139, 145, 151, 156, and 138, and those of Massachusetts for the same periods having been 160, 161, 161, 153, and 138. The widest range of five-year variation in the case of England and Wales was 18 per 1,000 births, and that in the case of Massachusetts 22 per 1,000 births. Their respective averages for the twenty-five year period were 146 and 154. In the tabulation of infant mortality in the principal countries of Europe compiled and published in 1890 by Dr. Bertillon and reproduced in transposed form on a previous page of this paper, the latest date of observation was the year 1883,-that is to say, twenty-five years ago,—and some of the figures dated back to 1862. In those days the registration of vital statistics in many-if not most-of the countries of Europe was far less advanced than it has become of late years, and, taking into account the well-known fact that approximate completeness in the registration of deaths almost inevitably precedes that of registration of births, it might naturally be assumed that the apparent rates of infant deaths to births would have been much larger in the case of the records of 1862-1883 than in those of 1881-1905, the divisor in the previous calculation presumably having been much farther removed from the correct figure. As to how well founded that assumption proves, the following comparison of the deathrates in the Bertillon table and that compiled by the writer of this paper will indicate. Some considerable apparent decreases in the infant mortality rates of certain countries are to be noted in the following table, TABLE V. A COMPARISON OF THE INFANT MORTALITY IN THE PRINCIPAL COUNTRIES OF EUROPE IN THE LAST TWENTY-FIVE YEARS WITH THE EARLIER PERIODS NAMED IN BERTILLON'S TABLE AND the DecreaSE OR INCREASE AND RELATIVE RANK IN THE CASE OF EACH OF THE COUNTRIES. and on the face of the returns it would seem that those countries were to be congratulated on having somehow succeeded in devising ways and means of reducing this phase of the mortality problem in which the vast majority of countries have notably failed, whether through lack of serious attention to the subject or for other reasons. But is such the case? Let us go behind the returns, locate the countries which have scored the largest apparent decreases, and consider for a moment whether those countries might naturally be expected to be found in the forefront of the movement for the reduction of infant mortality. Of the sixteen countries named in the table, eight show an apparent annual decrease of more than 10 deaths per 1,000 births since the early 80's, the decrease ranging from 10.9 in the case of Norway up to the remarkable figure of 47 in the case of Roumania, and the other countries in the order of the largest seeming decrease being Switzerland, Italy, Austria, the Netherlands, Sweden, and Finland. And yet none of these countries has ever attained any particular prominence in the crusade for the protection of children's lives. Austria is the only one of the eight which could be even seriously considered as among the great European powers, and, as will be noted, none of the indisputably first-class powers appears in the list. In England, France, and Prussia the decrease in the death-rate was merely nominal, and, as the registration systems of those countries would probably be regarded as superior to that in any of the eight countries which have scored the apparent large decreases in the infant death-rate, it would seem not only possible, but extremely probable, that the decrease in the last-named countries was more apparent than real-in other words, was a decrease in figures only, very likely due to the material increase in the registration of births and the consequent decrease in the ratio of deaths under age 1 to births. In England, France, and Prussia the registration of births was probably much more complete thirty or forty years ago than in most of the smaller countries of Europe, and, if such was the case, there naturally would be much less fluctuation in the mortality rates in the case of those leading countries. Therein probably lies the explanation of most of the apparent large decreases in the last twenty-five years. In default of positive evidence it would be absurd to believe that the little country of Roumania, with its limited resources, had succeeded in effecting a reduction of its infant death-rate by nearly 20 per cent. in the last twenty-five years, and thereby materially distanced every other country in the world. Furthermore, as shown in Table III, the record of its infant deathrates by five-year periods proves that the rate has been almost continuously in the ascendant for that same period. The British Registrar-General's office was unable to obtain any infantile death-rates from Austria up to 1896, or from Italy up to 1891, and in those obtained since those dates there is no sign of any sharp decline. Norway, Sweden, Finland, the Netherlands, and Switzerland have apparently shown material decreases in the infant death-rate in most of the five-year periods of the last quarter of a century, but improved registration of births probably accounts for that fact in most, if not all, of those cases, and it would therefore seem that the pronounced differences in some cases between the death-rates of the old-time Bertillon table and the up-to-date table presented in connection with this study are unworthy of any serious attention. In the face of the surprising uniformity of the mortality rate in question in the world at large, and especially in countries having thoroughly established registration systems, only the most irrefutable evidence will convince any student of infant mortality of a permanent reduction in the infant death-rate in any country up to this time. In so far as the infantile death-rate in the United States as a whole-either now or at any previous time-is concerned, there are absolutely no authentic data. In his contribution to the Eleventh Census Report on Vital and Social Statistics, published in 1896, Dr. Billings accounted for that fact by authoritatively stating (Part I, p. 21) that "we have no fully complete and accurate registration of births in any part of the United States. The most accurate registration is probably in Massachusetts, in which it is estimated that the deficiency is not greater than 2 per cent." Again, in Volume III of the Twelfth Census Reports (Vital Statistics, Part I, p. xlix) the late William A. King, Chief Statistician for Vital Statistics, commenced his discussion of births with the admission that "the data relating to births are the most incomplete and unsatis factory of any treated in this report. Were it not considered desirable to give such results as bear upon the question for the information of students of the statistics, the subject might be dismissed with the statement that they are entirely inadequate to determine, directly, the general birth-rate of the country, or, what is of equal practical importance, the relative birth-rate of different classes of population. A number of the States and cities have laws requiring the registration of births, but it is doubtful if there is a single place in which births are registered as fully as deaths." DIAGRAM II.-THE INFANT MORTALITY RATES OF ENGLAND AND WALES, |