INFANT MORTALITY: 
 
 ITS RELATION TO SOCIAL AND 
 
 INDUSTRIAL CONDITIONS 
 
 BY HENRY H. HIBBS, JR. 
 
 FORMERLY OF THE DEPARTMENT OF RESEARCH 
 
 BOSTON SCHOOL FOR SOCIAL WORKERS 
 
 AND THE DEPARTMENT OF SOCIOLOGY 
 
 UNIVERSITY OF ILLINOIS 
 
 NEW YORK 
 
 DEPARTMENT OF CHILD-HELPING 
 RUSSELL SAGE FOUNDATION 
 
 MCMXVI 
 

CH 24 
 
 INFANT MORTALITY: 
 
 ITS RELATION TO SOCIAL AND 
 
 INDUSTRIAL CONDITIONS 
 
 By henry hThibbs, jr. 
 
 FORMERLY OF THE DEPARTMENT OF RESEARCH 
 
 BOSTON SCHOOL FOR SOCIAL WORKERS 
 
 AND THE DEPARTMENT OF SOCIOLOGY 
 
 UNIVERSITY OF ILLINOIS 
 
 SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS 
 
 FOR THE DEGREE OF DOCTOR OF PHILOSOPHY 
 
 IN THE 
 
 FACULTY OF POLITICAL SCIENCE 
 
 COLUMBIA UNIVERSITY 
 
 NEW YORK 
 
 DEPARTMENT OF CHILD-HELPING 
 RUSSELL SAGE FOUNDATION 
 
 MCMXVI 
 
.14 rt 
 
 323 
 
 ' r 
 
 THIS SERIES OF PAPERS HAS BEEN 
 COPYRIGHTED BY THE PERIODICALS 
 IN WHICH THEY FIRST APPEARED; 
 ALL RIGHTS RESERVED 
 
 3ltt 
 
 MAY a Mm 
 
 THE RUMFORD PRESS 
 
 CONCORD, N. H. 
 
 1916 
 
INTRODUCTION 
 
 This series of papers, all of which have been published pre- 
 viously in periodical form, is the outcome of a house-to-house 
 investigation of infant mortality in four wards of Boston 
 made in 1910-11 and 1911-12 by the Research Department 
 of the Boston School for Social Workers under a grant from 
 the Russell Sage Foundation. During the second year the 
 writer was in charge of the field work. He was later com- 
 missioned to write the report of the investigation for publi- 
 cation. In undertaking this task it appeared that no par- 
 ticular service would be rendered by publishing another study 
 of the subject which dealt only with local conditions and which 
 was prepared with only incidental reference to the data 
 gathered elsewhere in other investigations. It did appear, 
 on the other hand, that there was a place for a book dealing 
 with the problem of infant mortality in the United States and 
 based on all the studies of the subject that are available. 
 A number of such investigations have recently been made in 
 different parts of the United States but, so far, the results 
 of these have never been brought together and compared 
 with a view to finding the common relationships and laws 
 involved. This series of papers is an attempt to meet this 
 need. The sources of the study, therefore, are very numerous. 
 Since, however, they are cited in footnotes in the various 
 chapters they need not be mentioned here. 
 
 In view of the prominence given to previously unpublished 
 data, gathered in the house-to-house investigation in four 
 wards of Boston, it will be necessary to describe briefly the 
 scope and method of this investigation. The four wards 
 were 6, 8, 13, and 17, the boundaries of which correspond 
 roughly with the sections commonly known as the North End, 
 the West End (both of these lying in the oldest parts of the 
 city). South Boston (lower half), and Roxbury (in part). 
 The investigation was based on the calendar year 1910. 
 Transcripts of the original records of all the births and infant 
 
iv Infant Mortality 
 
 deaths which occurred in these wards were made from the 
 copies on file in the office of the registry department of the 
 city and visits were made to the homes of the mothers by 
 Fellows in the Research Department of the School for Social 
 Workers.* Interviews were obtained with the mothers of 
 2,063 infants.! 
 
 The infant mortality rates in this investigation, like those 
 in the investigation by the U. S. Children's Bureau into 
 infant mortality in Johnstown, Pa., were obtained by com- 
 paring the number of infants (whose parents could be located) 
 who were born in the four wards in 1910 with the number 
 of these same infants who died before they were one year of 
 age, whether the death occurred in 1910 or 1911. The com- 
 parison was upon the basis of deaths to 1,000 births, still- 
 births not being included. 
 
 The writer is indebted to Dr. Jeffrey R, Brackett, director 
 of the School for Social Workers, and Dr. T. W. Glocker, 
 formerly director of the Research Department, for the use of 
 the data collected in the Boston investigation. 
 
 As the title indicates, the series of papers deals with the 
 relation of social and industrial conditions to infant mortal- 
 ity. Certain conditions, however, that may be properly 
 classed as industrial or social have been necessarily omitted. 
 Among these omitted factors may be mentioned the charac- 
 ter of the milk supply, J the influence of artificial feeding, § and 
 the form of the medical attendance. It has also been nec- 
 essary to omit all consideration of infant mortality in insti- 
 tutions, of the mortality of illegitimate infants, and of still- 
 births. 
 
 Finally the reader should be warned that the various fac- 
 tors that produce infant mortality are so closely interrelated 
 and so difficult of measurement!] that conclusions stating the 
 exact extent of the influence of each are not to be expected. 
 It has thus been impossible to determine except very roughly 
 
 *Mi8s Helen C. Schindler, Miss Margaret 0. Cook, Miss Helen L. Spencer, Miss Charlotte I. Claflin, 
 Miss Bertha C. Lovell, Miss Alice M. Mclntire, Mr. A. H. Lord, Mr. T. Eaton, and Mr. P. E. Shannon. 
 
 fThis represents visits to about 4,000 homes, as about half the families sought for could not be found. 
 
 J See, however, p. 63. 
 
 §There have been many excellent studies made of this subject. In consequence no extended treatment 
 of it was considered necessary in this essay. See references on p. 49 and also pp. 25-28, 79-80, and 112. 
 
 llSee for example pp. 59, 88-92, 95-9G, and 113. 
 
Introduction v 
 
 the relative influence of the various factors involved.* It 
 has been possible, however, to show clearly that the social 
 and industrial conditions studied do exercise an influence on 
 infant mortaHty and to indicate roughly the extent and im- 
 portance of their influence. If this warning is carefully 
 heeded the reader will not be hkely to feel that the writer 
 has been led to over-emphasize the influence of certain con- 
 ditions at the expense of others. 
 
 The greater part of the actual writing of the book was done 
 in the statistical laboratory of Columbia University. The 
 writer wishes to express his sincere appreciation to Prof. R. E. 
 Chaddock, director of the laboratory, for this privilege but 
 especially for his unstinting advice and help throughout the 
 writing and preparing of the book for publication. 
 
 The writer wishes also to acknowledge the courtesy of the 
 editors of the Quarterly Journal of Economics and the Quart- 
 erly Publications of the American Statistical Association and 
 the secretaries of the Southern Sociological Congress and the 
 American Academy of Medicine in allowing the papers to be 
 reprinted in book form.f 
 
 •See for example pp. 21, 42-45, 81, and 115. 
 
 tThe papers were written, however, with a view to their publication in book form. The work as a 
 whole, therefore possesses a unity that it would not have had if each chapter had been written indepen- 
 dently. 
 
 The papers are reprinted without changes except for slight modifications of titles. The reader will 
 understand, therefore, that references in footnotes to " other articles" on infant mortaUty by the writer 
 may be read as references to papers now included in this book. 
 
TABLE OF CONTENTS 
 INTRODUCTION 
 
 I. 
 THE PRESENT POSITION OF INFANT MORTALITY: ITS 
 RECENT DECLINE IN THE UNITED STATES 
 
 The problem of infant mortality; why made an object of special 
 study; the infant mortality rate. Infant mortality in foreign 
 countries; in the United States; significance of the wide variation 
 in rates. The recent decline in infant mortality; in foreign 
 countries; in Massachusetts and Boston; in the registration area. 
 The decline in the infant mortality rate compared with that in the 
 general death rate for all ages. Comparison of the decrease in 
 the death rate for certain age groups in the registration area since 
 1900. Conclusion 3 
 
 II. 
 THE INFLUENCE OF PRENATAL CONDITIONS ON INFANT 
 MORTALITY 
 
 Study of the causes of infant mortality not to be confined solely to 
 period of infancy. Subdivisions of the periods of infant devel- 
 opment; importance of the germinal period before conception 
 and the intra-uterine period. The influence of prenatal condi- 
 tions partly hereditary and partly environmental or social. 
 Preventive methods. Extent of the influence of prenatal condi- 
 tions : as seen in the mortality of infants in the early weeks and 
 months of life; in the death rate from the diseases of early infancy; 
 conclusion. Comparison of the recent decline in the relative in- 
 fluence of prenatal and postnatal conditions. How the lesser de- 
 cline in the influence of prenatal conditions is to be accounted 
 for. The future 19 
 
 III. 
 INFANT MORTALITY AND THE SIZE OF THE FAMILY 
 
 The influence of the birth rate on the rate of infant mortality. The 
 relation between the rate of child mortality and the order of 
 birth; between the rate of infant mortality and the number of 
 children previously born to the mother, the order of birth, and 
 the number of the mother's previous pregnancies. Other factors 
 which must be considered in interpreting the relationship between 
 
viii Contents 
 
 the rate of infant mortality and the size of the family; conclusion. 
 The relation of the length of the interval between the mother's 
 pregnancies to infant mortality; importance of the adequate 
 spacing of births 37 
 
 IV. 
 THE MOTHER AND INFANT MORTALITY 
 
 Proportion of infant deaths to births varies among families as widely 
 as among states, cities, or other subdivisions of population; the 
 "domestic " factors of infant mortality. The influence of the age 
 of the mother. The influence of the character and intelligence 
 of the mother. To what extent infant mortality is a question of 
 motherhood. Preventive methods; the extent to which these 
 depend upon the mother for their effectiveness; education for 
 parenthood 53 
 
 V. 
 INFANT MORTALITY AND URBAN, HOUSING, AND LIVING 
 CONDITIONS 
 
 Wide variation in infant mortality rates for different countries, 
 states, cities, etc., and its significance. The influence of urban 
 conditions of life on infant mortality. The relative decline in 
 infant mortality in urban and rural districts; significance of the 
 greater decline in the cities. The influence of congestion and 
 overcrowding. The influence of housing and living conditions; 
 housing conditions proper; home sanitation, cleanliness, ventila- 
 tion, care of home, etc 77 
 
 VI. 
 THE INFLUENCE OF ECONOMIC AND INDUSTRIAL CON- 
 DITIONS ON INFANT MORTALITY 
 
 The older view that women's work in itself caused high infant mor- 
 tahty: recent investigations lead to doubt as to its direct influ- 
 ence. Infant mortality in industrial cities. Extent of the 
 employment of mothers in gainful occupations. Statistical 
 evidence of direct influence on infant mortahty. Indirect influ- 
 ences; on general home standards of the community, on the 
 vitality and education of potential mothers. Influence of house- 
 work. Influence of poverty. Conclusion; low wages the fun- 
 damental economic and industrial factor in infant mortality; 
 other adverse economic and industrial influences 103 
 
I. 
 
 THE PRESENT POSITION OF INFANT MORTALITY: 
 ITS RECENT DECLINE IN THE UNITED STATES 
 
 Reprinted from the Quarterly Publications of the American 
 
 Statistical Association, New Series, No. 112, 
 
 December, 1915, pp. 813-826. 
 
THE PRESENT POSITION OF INFANT MORTALITY: 
 ITS RECENT DECLINE IN THE UNITED STATES. 
 
 '^The term infant mortality, according to the generally 
 accepted usage, is employed to designate the deaths of infants 
 under one year of age. The problem is measured by an infant 
 mortahty rate which is an expression of the proportion of 
 infants dying under one year of age to 1,000 births or, when 
 birth statistics are not available, to 1,000 population under 
 one year of age. The deaths of infants at this early age is 
 made the object of special study because they constitute 
 such an enormous proportion of the deaths at all ages. In 
 no other period of life do deaths occur with such frequency. 
 This is evident upon examination of any of the bulletins on 
 mortality statistics issued annually by the Bureau of the 
 Census for the registration area of the United States. Thus, 
 Bulletin 109 shows that 27 per cent, of the total number of 
 deaths which occurred in the registration area in 1910 were of 
 children under five years of age and 19 per cent., or almost 
 one fifth, were of infants under one year of age. In 1911, 
 1912, and 1913 the deaths of infants constituted 18 per cent, 
 of the deaths at all ages. 
 
 That the death rate for infants and young children should 
 be greater than for other persons is not, however, surprising. 
 As a recent English writer says, "The young of all animals 
 are more susceptible than the adult to the influence of the 
 environment and the approach of death. Hence, it is inevit- 
 able that, even under the most favourable circumstances the 
 deaths of infants will furnish a large contribution to the bills 
 of mortahty."* It is not the mere fact of excess but the 
 tremendously greater excess of deaths which occur during the 
 first year of life that constitutes the problem of infant mor- 
 tality. 
 
 Infant Mortality in Foreign Countries. The world-wide 
 significance of the problem will be evident upon examination 
 
 * George Newman, M.D.: Infant Mortality. London, 1906, p. v. 
 
 3 
 
4 American Statistical Association. 
 
 of the following figures showing the infant mortaUty rate per 
 1,000 births for the foreign countries for which statistics 
 are available during the five year period from 1906 to 1910:* 
 
 Chile 315 Finland 117 
 
 , j Hungary 204 Switzerland 115 
 
 I / Jamaica 191 The Netherlands ... 114 
 
 Ceylon 189 Scotland 112 
 
 I Prussia 168 Denmark 108 
 
 /l Servia 154 Ireland 94 
 
 I Italy 153 Sweden 78 
 
 I Belgium 141 Australian Common- 
 Ontario 127 wealth 78 
 
 France 126 Norway 70 
 
 England and Wales 117 New Zealand 70 
 
 Thus, in one third of the countries of the world for which 
 statistics are available the infant mortality rate was over 150, 
 while in about one half it was over 125. In only 5 of the 
 countries, 3 in Europe and 2 in Australasia, was the rate less 
 than 100 deaths per 1,000 births. Expressed in another way 
 this means that, out of every 1,000 children born in countries 
 like Hungary, Prussia, and Italy, from 150 to 200 die before 
 reaching the end of the first year of fife; out of every 1,000 
 born in countries like England, Scotland, Switzerland, and 
 The Netherlands, from 110 to 120 die before reaching this age; 
 while in countries like Ireland, Sweden, Norway, and Aus- 
 tralia, from 70 to 100 die before they are a year old. 
 
 Infant Mortality in the United States. Figures comparable 
 with these exist for only a few of the states and cities of the 
 United States. In 1911 the Bureau of the Census reported 
 that only in the cities of Washington and New York and in 
 the states of Pennsylvania and Michigan and the six New 
 England States could the registration of both births and deaths 
 be regarded as sufficiently complete (amounting to at least 
 90 per cent, of the total) to make possible the calculation of an 
 accurate rate of infant mortality based on the ratio of births 
 to deaths. This area comprises the provisional "registration 
 
 ♦Seventy Third and Seventy Fourth Annual Reports of the Reg'iatrar General for Births, Deaths, and 
 Marriages in England and Wales (pp. xciv and 105-15 respectively). 
 
The Present Position of Infant Mortality. 5 
 
 area" of the United States for births and deaths. The infant 
 mortality rates per 1,000 births for these states and cities in 
 1910 were as follows : 
 
 Rhode Island 158 Connecticut 127 
 
 New Hampshire . . . 146 Michigan 124 
 
 Pennsylvania 140 
 
 Maine 135 Washington, D. C. 152 
 
 Massachusetts 131 New York, N. Y. . 125 
 
 Besides these states and cities where the registration of 
 both births and deaths are regarded by the census office as 
 being sufficiently or "fairly complete," there are others where 
 the registration of deaths only is regarded as sufficiently 
 complete to be included in what is known as the registration 
 area for deaths. In 1910, 22 entire states and a large number 
 of cities in non-registration states were included in this area. 
 Since complete birth statistics are not available, it is not 
 possible to calculate an infant mortality rate for this area 
 in the ordinary manner — by computing the ratio of deaths 
 to 1,000 births. To overcome this difficulty, the Bureau of 
 the Census in its report on mortality statistics for 1910 em- 
 ployed an infant death rate based on the proportion of deaths 
 to 1,000 population under one year of age in 1910. This 
 method is confessedly inaccurate because the enumeration of 
 the population under one year of age is never complete and 
 entirely accurate. Yet it is the best method available for 
 studying the distribution of infant mortaHty in the United 
 States and, although its crudities should serve as a caution 
 against drawing too fine conclusions from its use, its defects 
 should not be over-emphasized. The following table shows 
 the death rate per 1,000 population under one year of age 
 for the registration states :* 
 
 Ohio 115.9 
 
 Michigan 127.5 
 
 Maine 140.4 
 
 New York 143.6 
 
 Connecticut . . . 143.7 
 
 New Jersey. . . . 148.8 
 
 Utah 
 
 . 82.3 
 
 Washington . . . 
 
 . 84.3 
 
 Kentucky 
 
 . 87.9 
 
 Montana 
 
 . 90.4 
 
 California 
 
 92.2 
 
 Minnesota 
 
 . 92.4 
 
 * Bureau of the Census: Bulletin 112, p. 24. 
 
6 American Statistical Association. 
 
 Missouri *96.7 Pennsylvania.. 149.7 
 
 Colorado 104.5 Maryland 152. 1 
 
 Indiana 106.9 Massachusetts. 160.8 
 
 Wisconsin 108.0 New Hampshire 164.9 
 
 Vermont 109.4 Rhode Island. . 181.5 
 
 Thus the infant death rate per 1,000 population varied 
 ■from less than 85 in the Western states, Utah and Washington, 
 to 165 and 182 in the two New England states, Rhode Island 
 and New Hampshire. It was 127.6 for the entire group of 
 registration states considered as a whole. 
 
 It would be of interest to compare the rates for this group 
 of states with those for foreign countries given in a previous 
 table. But this is, of course, impossible since infant death 
 rates based on population can not be compared with the true 
 infant mortality rate based on births. In the 1911 bulletin 
 on ''Mortality Statistics," the Bureau of the Census estimated 
 on the basis of the figures quoted here and others, that the 
 1 infant mortality rate per 1,000 births for the United States 
 / as a whole was about 124. Comparing this estimate with 
 .' the computed rates for the foreign countries given in the 
 preceding table, it will be seen that the rate of infant mortality 
 in the United States is lower than in such countries as Chile, 
 Hungary, Jamaica, Prussia, Servia, and Italy; about equal 
 to the rate for the province of Ontario and for France; higher 
 than the rate for England and Wales, Scotland, Finland, 
 Switzerland, and Denmark; and considerably higher than the 
 rate for Ireland, Sweden, Norway, New Zealand, and the 
 Australian Commonwealth. 
 
 The figures quoted in the previous table also are of value 
 in that they show the relative position of infant mortality in 
 the different states and sections of the United States. The 
 states included in the registration area are arranged in an 
 ascending order according to their infant death rates. An 
 examination of this table at once reveals the fact that in gen- 
 eral the lowest rates are to be found in the Western and the 
 highest in the Eastern states, with the rates for the North- 
 Central (or Middle- Western) states in between. Thus, the 
 
 * Figures for deaths for 1911, first year of operation of state law. 
 
The Present Position of Infant Mortality. 
 
 average infant death rate for the 5 Western and Mountain 
 states included was 91, for the 6 North-Central states, 
 108, and for the 9 New England and Middle-Atlantic states, 
 149.* 
 
 The same conclusion that in 1910 infant death rates were 
 lowest in the Western part of the registration area and highest 
 in the Eastern, with the Middle- West in between, is also to be 
 drawn from an examination of the following table showing 
 the infant death rate per 1,000 population under one year of 
 age for the larger cities of the registration area : f 
 
 Oakland, Cal 
 
 94.8 
 
 Dayton, Ohio 
 
 146.8 
 
 Seattle, Wash 
 
 100.4 
 
 Cleveland, Ohio. . . 
 
 147.2 
 
 Portland, Ore 
 
 105.3 
 
 Cincinnati, Ohio . . 
 
 149.8 
 
 Los Angeles, Cal.. . 
 
 110.7 
 
 Jersey City, N. J. . 
 
 153.2 
 
 San Francisco, Cal. 
 
 113.6 
 
 New Orleans, La. . 
 
 154.9 
 
 Toledo, Ohio. . 
 
 125.0 
 
 Atlanta, Ga 
 
 155.3 
 
 Cambridge, Mass. . 
 
 126.1 
 
 Bridgeport, Conn. . 
 
 155.5 
 
 St. Paul, Minn. . . . 
 
 130.8 
 
 Philadelphia, Pa. . . 
 
 162.2 
 
 Birmingham, Ala. . 
 
 133.0 
 
 Albany, N. Y 
 
 162.9 
 
 Louisville, Ky 
 
 134.0 
 
 Boston, Mass 
 
 165.5 
 
 Denver, Col 
 
 134.7 
 
 Worcester, Mass.. . 
 
 168.0 
 
 Grand Rapids, Mich. 
 
 134.8 
 
 Kansas City, Mo. . 
 
 170.4 
 
 New Haven, Conn. . 
 
 134.9 
 
 Milwaukee, Wis. . . 
 
 172.0 
 
 Nashville, Tenn. . . . 
 
 135.1 
 
 Providence, R. I. . . 
 
 173.7 
 
 St. Louis, Mo 
 
 135.8 
 
 Syracuse, N. Y. ... 
 
 176.4 
 
 Chicago, 111 
 
 139.5 
 
 Pittsburg, Pa 
 
 179.6 
 
 Omaha, Neb 
 
 140.0 
 
 Buffalo, N. Y 
 
 180.9 
 
 Columbus, Ohio. . . 
 
 140.4 
 
 Washington, D. C. 
 
 194.6 
 
 Spokane, Wash 
 
 142.4 
 
 Detroit, Mich 
 
 204.8 
 
 Indianapolis, Ind. . . 
 
 144.8 
 
 Baltimore, Md 
 
 209.6 
 
 Newark, N. J 
 
 145.8 
 
 Richmond, Va. . . 
 
 229.3 
 
 New York, N. Y. . . 
 
 146.2 
 
 Fall River, Mass. . 
 
 259.5 
 
 Paterson, N. J 
 
 146.7 
 
 Lowell, Mass. . . . 
 
 261.0 
 
 * As only 2 of the 16 Southern states were included in the registration area in 1910, no comparison of 
 the incidence of infant mortality in this with other sections of the country is possible, 
 t Bureau of the Census: Bulletin 112— Mortality Statistics, p. 24. 
 
8 American Statistical Association. 
 
 The average infant death rate in 1910 per 1,000 population 
 under one year of age was 115 for the 7 Western and Moun- 
 tain cities included in the table, 149 for the 14 North-Central 
 (or Middle- Western) cities, 165 for the 7 Southern cities, 162 
 for the 9 Middle-Atlantic cities, and 181 for the 8 New Eng- 
 land cities. 
 
 ^ This study of the position of infant mortality in the United 
 States and foreign countries shows the seriousness and world- 
 wide significance of the problem. It also shows how the 
 infant mortality rate varies in different parts of the civilized 
 world. Thus, the rate has been found to be much lower in 
 Australia than in Europe. Among the European countries it 
 was lowest in Norway, Sweden, Ireland, and Denmark and 
 highest in Russia, Prussia, Hungary, and Italy, Turning to 
 a single country, the United States, and substituting the use 
 of the infant death rate per 1,000 population under one yea,T 
 of age for that of the infant mortality rate per 1,000 births, 
 the same wide variation was revealed, the ratio of infant 
 deaths to population being considerably less in the Western 
 than in the Eastern parts of the registration area. Further 
 examination of the tables also showed that the ratio varies 
 just as widely when the cities of any state or country are 
 compared. The examination of the report of the health 
 department of almost any city that requires the registration 
 of births and deaths will reveal the same variation by wards — 
 and even by blocks, if figures are given for such small areas. 
 
 This wide variation in rates of infant mortality for different 
 countries, states, and cities constitutes a fact of fundamental 
 importance in the study of the subject. Out of it arise ques- 
 tions that at once bring us face to face with the relationship 
 between social and industrial conditions and infant mortality. 
 Why this wide variation in the geographic distribution of 
 infant deaths? Why is the infant death rate lower in one 
 country than another, in certain cities of the same country 
 than others, in certain wards of one city than in others? 
 Why, indeed, should the death rate for little children in the 
 first year of life so far exceed the rate for older children and 
 adults? All of these questions require for their answer some 
 knowledge of the causes of infant mortality and their rela- 
 
The Present Position of Infant Mortality. 
 
 tion to industrial, domestic, and social conditions. But with 
 this aspect of the problem this paper cannot deal.* 
 
 The Recent Decline of Infant Mortality in Foreign Countries, 
 Since 1881, the first year for which statistics are available for 
 most countries, there has been a noticeable decline in infant 
 mortality in most foreign countries and cities and, since 1900, 
 in most of the states and large cities included in the regis- 
 tration area of the United States. The following table shows 
 this decline for the principal foreign countries for which sta- 
 tistics are available: 
 
 PER CENT, OF DECREASE IN THE INFANT MORTALITY RATE PER 1,000 BIRTHS FOR 
 THE PRINCIPAL FOREIGN COUNTRIES FOR WHICH STATISTICS ARE AVAILABLE 
 BETWEEN 1881-85 AND 1906-10. (a) 
 
 Country. 
 
 1881-1885. 
 
 1906-1910. 
 
 Per Cent, 
 of Decrease. 
 
 EUHOPE 
 
 250 (b) 
 207 
 
 185 (b) 
 157 
 156 
 
 167 
 139 
 181 
 171 
 162 
 
 116 
 117 
 135 
 
 94 
 
 99 
 
 125 
 90 
 
 204 
 168 
 153 
 154 
 141 
 
 126 
 117 
 114 
 115 
 117 
 
 78 
 112 
 108 
 94 
 70 
 
 78 
 70 
 
 18.4 
 
 
 18.8 
 
 Italy 
 
 17.3 
 
 
 1.9 
 
 
 9.6 
 
 
 24.6 
 
 
 15.8 
 
 The Netherlands 
 
 31.5 
 
 
 32.7 
 
 
 27.7 
 
 
 32.8 . 
 
 
 4.3 
 
 
 20.0 
 
 
 0.0 
 
 
 29.3 
 
 Australasia: 
 
 37.6 
 
 
 22.2 
 
 
 
 (a) Seventy Third and Seventy Fourth Annual Reports of the Registrar Genera 1 for Birthe, Deaths, 
 and Marriages in England and Wales (d. xciv and pp. 105-15 respectively) . 
 
 (b) Figures tor 1881-85 not available: those given are for 1891-95. 
 
 The rate of infant mortality for every country included in 
 the tabic declined during this period of 30 years with the single 
 exception of Ireland where, although the rate for both periods 
 remained the same, it was at a very low point— 94 deaths 
 
 * In other recent articles the writer has discussed thb question of the relation of social conditions to 
 infant mortality. See "Infant MortaUty and the Size of the Family," Quarterly Publications of 
 THE American Statistical Assocution, September, 1915; "Infant Mortality and Urban, Housing, 
 and Living Conditions," Journal of Sociologic Medicine, October. 1915; "The Relation of Economic and 
 Industrial Conditions to Infant Mortality," Quarterly Journal of Economics, November, 1915; and "The 
 Influence of Prenatal Conditions on Infant Mortality," Proceedings of the Southern Sociological Congress, 
 1915. 
 
10 American Statistical Association. 
 
 per 1,000 births. The most notable decrease was in New 
 Zealand and Australia. In the former the rate fell from the 
 already low point of 90 deaths per 1,000 births to 70 — a 
 decrease of 22.2 per cent. — and in the latter from 125 to 78 — 
 a decrease of 37.6 per cent. The decline was also notable in 
 Norway, Sweden, and Denmark and to a lesser extent in 
 England and Wales. The absolute decrease was also great 
 in Switzerland, the Netherlands, France, and Finland, but 
 the rate was very high for each of these countries at the 
 beginning of the period.* 
 
 The Registrar-general for England and Wales, from whose 
 annual reports the preceding table was compiled, also gives 
 figures showing the decline in the rate of infant mortahty in 
 the principal foreign cities since 1881-85. Space will not 
 permit quotation of these in detail but the fact should be 
 noted that in each of the cities included, with one exception 
 (Trieste, Hungary), the rate of infant mortality decHned 
 during the period under consideration. The most notable 
 decrease was in the three Dutch cities, Amsterdam, The 
 Hague, and Rotterdam; the two Australian cities, Sydney 
 and Melbourne; and the cities of Norway and Sweden, Stock- 
 holm and Christiania, in each of which the rate fell to a point 
 below 100 deaths per 1,000 births — a record, as shown in the 
 preceding table, also attained by each of the countries in 
 which these cities are situated. The absolute decrease was 
 also great in the two Prussian cities, Munich and Berlin, and 
 the Hungarian city of Budapest, but the rate for each of 
 these cities was very high at both the beginning and the end 
 of the period. 
 
 The Decline in Infant Mortality in the United States. Unfor- 
 tunately no series of infant mortality rates at all comparable 
 with those just shown for foreign countries can be presented 
 
 * In view of the fact that, as has been frequently pointed out, the apparent decline in the rate of infant 
 mortality in any country in a period of years may be affected by the increase in the per cent, of birth* 
 which are registered, the figures given in the table may not in all cases be strictly comparable. For instance, 
 if in three countries, in each of which the proportion of births registered in 1881 was 90 per cent., it should 
 happen that the proportion registered should gradually increase in each but unequally so that in the first 
 92 per cent, of the births which occurred in 1910 were registered, and in the second 95 per cent., and in the 
 third 99 per cent., the decline in the rate of infant mortality between these two years would not be strictly 
 comparable unless the factor of varying perfection in birth registration were allowed for. This difficulty 
 probably is not of sufficient importance to require its consideration here even if sufficient material bearing 
 on the comparative efficiency of birth registration in foreign countries in the last thirty years were available. 
 
The Present Position of Infant Mortality. 
 
 11 
 
 for the United States. Figures are available, however, for 
 Massachusetts and Boston for the same period, 1881-85, to 
 1906-10 and for three later years, 191 1-13. Also, the per cent, 
 of decrease in the infant death rate per 1,000 population under 
 one year of age between 1900 and 1911 has been calculated by 
 the Bureau of the Census for the registration area and the 
 larger registration cities. The following table shows the de- 
 cline in the infant mortality rate for Massachusetts and Boston 
 since 1881 and the per cent, of decrease in the rate between 
 1881-85 and 1909-13: 
 
 PER CENT. OF DECREASE IN THE INFANT MORTALITY RATE PER 1,000 BIRTHS FOR 
 THE COMMONWEALTH OF MASSACHUSETTS AND THE CITY OF BOSTON BETWEEN 
 1881-85 AND 1909-13. (a) 
 
 Years. 
 
 Massachusetts. 
 
 Boston. 
 
 1881-85.... 
 1880-90. . . . 
 1891-95. . . . 
 1896-1900. . 
 1901-05. . . . 
 1906-10.... 
 1909-13 (b) . 
 
 Per cent, of decrease. 
 
 160 
 161 
 161 
 153 
 138 
 133 
 121 
 
 24.4 
 
 186 
 178 
 167 
 151 
 138 
 133 
 120 
 
 35.5 
 
 (a) Compiled from the Massachusetts ammal reports on births, deaths, and marriages and the annual 
 reports of the Health Department of Boston. 
 
 (b) Figures for five year period are not available. 
 
 It will be noted that the infant mortality rate in this period 
 of thirty-three years decreased over 24 per cent, in Massa- 
 chusetts and about 36 per cent, in Boston. It will also be 
 noted that the decrease was especially marked during the past 
 
 few years. 
 
 The nearest approach to an accurate determination of the 
 position of infant mortality in the other states and cities of 
 the United States is to be found in a table recently presented 
 by the Bureau of the Census, and herewith reproduced in 
 part, which shows the per cent, of decrease in the infant death 
 rate per 1,000 population under one year of age between the 
 census year 1900 and the calendar year 1911 for the states 
 and large cities of the registration area. It should be noted, 
 however, that the rates given in this table are infant death 
 rates calculated upon the basis of infant deaths to 1,000 pop- 
 ulation under one year of age and not according to the usual 
 method of the ratio of deaths to 1,000 births. 
 
12 
 
 American Statistical Association. 
 
 PER CENT. OF DECREASE IN THE INFANT DEATH RATE PER 1,000 POPULATION UNDER 
 1 YEAR OF AGE BETWEEN THE CENSUS YEAR 1900 AND THE CALENDAR YEAR 1911 
 FOR THE STATES INCLUDED IN THE REGISTRATION AREA IN 1900 AND FOR CITIES 
 OF 400,000 POPULATION OR OVER, (a) 
 
 Area. 
 
 Census 
 Year: 
 1900. 
 
 Calendar 
 Year: 
 1911. 
 
 Per Cent. 
 
 of 
 Decrease. 
 
 States included in the registration area in 1900 (b) 
 
 Rhode Island 
 
 Massachusetts 
 
 New Hampshire 
 
 New Jersey 
 
 New York 
 
 Connecticut 
 
 Maine 
 
 Vermont 
 
 Michigan 
 
 Cities of 400,000 population or over in 1910. (c) 
 
 Baltimore 
 
 Philadelphia 
 
 Detroit 
 
 Boston 
 
 New York 
 
 Cleveland 
 
 Pittsburg 
 
 St. Louis 
 
 San Francisco 
 
 Buffalo 
 
 Chicago 
 
 159.3 
 
 197.9 
 177.8 
 172.0 
 167.4 
 159.8 
 
 156.8 
 144.1 
 122.1 
 121.3 
 
 235.1 
 201.9 
 201.2 
 194.1 
 189.4 
 185.5 
 
 179.8 
 162.4 
 152.2 
 150.9 
 146.6 
 
 129.5 
 
 138.6 
 143.3 
 150.3 
 131.5 
 128.8 
 
 130.9 
 110.9 
 102.0 
 111.4 
 
 189.2 
 141.9 
 168.8 
 160.9 
 130.6 
 123.7 
 
 141.4 
 123.8 
 104.8 
 140.6 
 123.3 
 
 (a) Twelfth Annual Report of the Bureau of the Census on Mortality Statistics for the year 1911, p. 24, 
 
 (b) Includes District of Columbia. 
 
 (c) Space does not permit the quoting of rates for smaller cities. 
 
 From this table it will be noted that in this period of 11 
 years the ratio of infant deaths to 1,000 population under one 
 year of age decreased nearly one fifth (19 per cent.) in this 
 group of registration states. The largest decrease shown in 
 the rate for any of the states was in that for Rhode Island 
 (30 per cent.) and the least in that for Michigan (8 per cent.). 
 In all the cities included in the table the infant death rate 
 also showed a decline — ranging from 33 per cent, in Cleveland 
 to 7 per cent, in Buffalo. The fact that this comparison 
 relates to only two individual years and that complete 
 returns of deaths of infants under one year of age may 
 not always have been made, coupled with the fact that the 
 number of infant deaths per 1,000 population under one year 
 of age does not furnish as satisfactory a basis for the study of 
 infant mortality as the number of such deaths per 1,000 births, 
 tends to diminish somewhat the value of the figures given in 
 
The Present Position of Infant Mortality. 
 
 13 
 
 the table. Yet, in spite of these Hmitations, these figures, 
 taken in conjunction with those previously given for Massa- 
 chusetts and Boston, show that in all probability there has 
 been a marked reduction in infant mortahty in this country 
 in recent years.* 
 
 The Decline in the Infant Mortality Rate Compared with 
 that in the Gereral Death Rate for All Ages. The extent of the 
 decline in the mortahty rate for infants under one year of 
 age can not be fully appreciated until it is compared with 
 the decline in the death rate for other age periods. The fol- 
 lowing table compares the decline in the infant mortality 
 rate shown in preceding tables with the decline in the general 
 death rate for all ages during the same periods: 
 
 PER CENT.OF DECREASE IN THE INFANT MORTALITY RATE PER 1,000 BIRTHS AND IN 
 THE GENERAL DEATH RATE FOR OF ALL AGES PER 1,000 POPULATION BETWEEN 
 1881-85 AND 1891-95 AND BETWEEN 1896-1900 AND 1906-10, FOR THE PRINCIPAL 
 FOREIGN COUNTRIES, (a) 
 
 Country. 
 
 Per Cent, of Decrease 
 
 Between 1881-85 and 
 
 1891-95 
 
 General 
 Death Rate. 
 
 Infant Mor- 
 tality Rate. 
 
 Per Cent, of Decrease 
 
 Between 1896-1900 and 
 
 1906-10 
 
 General 
 Death Rate. 
 
 Infant Mor- 
 tality Rate. 
 
 Hiin^kry 
 
 Prussia 
 
 Italy 
 
 Servia 
 
 Belgium 
 
 France 
 
 England and Wales 
 
 The Netherlands 
 
 Switzerland 
 
 Finland 
 
 Sweden 
 
 Scotland 
 
 Denmark 
 
 Ireland 
 
 Norway 
 
 New Zealand 
 
 The Australian Commonwealth 
 
 3.9 
 
 
 10.4 
 
 10.2 
 
 1.0 
 
 17 6 
 
 6.6 
 
 
 8.3 
 
 -M5.2 
 
 +8.7 
 
 +35.6 
 
 2.4 
 
 +4.9 
 
 +26.7 
 
 +0.4 
 
 +2.3 
 
 7.2 
 
 3.6 
 
 +7.9 
 
 16.9 
 
 8.4 
 
 8.8 
 
 16.8 
 
 7.0 
 
 9 3 
 
 +30.4 
 
 7.7 
 
 10.5 
 
 8.4 
 
 5.1 
 
 11.2 
 
 11.2 
 
 3.1 
 
 +7.1 
 
 10.6 
 
 +1.1 
 
 +2.2 
 
 16.5 
 
 +2.7 
 
 +7.8 
 
 4.4 
 
 2.3 
 
 1.0 
 
 11.5 
 
 7.3 
 
 3.3 
 
 +1.0 
 
 15.3 
 
 12.8 
 
 15.7 
 
 6.8 
 16.4 
 8.9 
 3.1 
 10.8 
 20.8 
 
 25.0 
 24.5 
 19.6 
 15.8 
 22. S 
 13.2 
 
 18.2 
 11.3 
 27.1 
 12.5 
 30.4 
 
 A plus sign (+) denotes an increase. 
 
 (a) Compiled from the Seventy Third and Seventy Fourth Annual Reports of the Registrar General 
 for Births, Deaths, and Marriages in England and Wales for the years 1910 and 1911. The Twelfth 
 Annual Report of the Bureau of the Census on Mortality Statistics for 1911 quotes in detail the general 
 death rates from which the per cents, of decrease in this table were compiled. 
 
 ♦ This is the conclusion arrived at in the Twelfth Annual Report of the Bureai of the Census on Mortality 
 Statistics for the year 1911, p. 24, and expressed in the following words: "There has been a marked reduc- 
 tion in the infant death rate in recent years." 
 
14 American -Statistical Association. 
 
 The first thing to be noted upon examination of the above 
 table is the much greater per cent, of decrease in both the 
 general death rate and the infant mortaUty rate in the last 
 than in the first half of this period of thirty years. Thus, 
 from 1881-85 to 1891-95 the infant mortahty rate decreased 
 in only 8 of the 15 countries for which rates could be obtained, 
 while from 1896-1900 to 1906-10 it declined in every country 
 included in the table. Moreover, the average per cent, of 
 decrease for all countries in the latter period was twice as 
 great as in the former. The same variation is also shown in 
 the decline of the general death rate in the two periods but 
 to a somewhat lesser extent. 
 
 By comparing the extent of the decline in the infant mor- 
 tality rate with that in the general death rate it will be seen 
 that in the first half of the period the greater decline occurred 
 in the general death rate, while during the second half the 
 greater decline occurred in the infant mortaUty rate. Thus, 
 from 1881-85 to 1891-95, a greater per cent, of decrease in 
 the infant mortahty rate occurred in only 4 of the countries 
 included in the table, while from 1896-1900 to 1906-10 a 
 greater per cent, of decrease failed to occur in only 2 countries. 
 
 A similar comparison extending over the same periods can 
 be made for Massachusetts, and it shows the same results. 
 Thus, from 1881-85 to 1891-95, the general death rate for 
 Massachusetts decreased 0.5 per cent, and the infant mor- 
 tality rate increased 0.6 per cent., while during the period 
 from 1896-1900 to 1906-10 the former rate decreased 10.5 
 per cent, and the latter 13.1 per cent. 
 
 That this greater decline in the infant mortality rate than in 
 the general death rate during recent years is probably typical 
 for this country is shown in the following table, which com- 
 pares the per cent, of decrease between 1900 and 1911 in the 
 general death rate and the infant death rate for the states 
 included in the registration area in 1900: 
 
The Present Position of Infant Mortality. 
 
 15 
 
 PER CENT. OF DECREASE IN THE INFANT DEATH RATE PER 1,000 POPULATION UNDER 
 1 YEAR OF AGE AND THE GENERAL DEATH RATE FOR ALL AGES PER 1,000 POPULA- 
 TION BETWEEN 1900 AND 1911, FOR THE STATES INCLUDED IN THE REGISTRATION 
 AREA IN 1900. (a) 
 
 State. 
 
 General 
 Death Rate. 
 
 Infant 
 Mortality Rate. 
 
 All States (b) 
 
 14 
 
 25 
 13 
 17 
 14 
 17 
 15 
 9 
 13 
 11 
 
 19 
 
 Rhode Island 
 
 30 
 
 Maine 
 
 23 
 
 
 21 
 
 New York 
 
 19 
 
 
 19 
 
 Connecticut 
 
 17 
 
 
 16 
 
 
 13 
 
 
 8 
 
 
 
 
 (a) Twelfth Annual Report of the Bureau of the Census on Mortality Statistics for the year 1911, 
 pp. 22 and 25. The general death rates are " corrected on the basis of the standard million of England and 
 Wales." 
 
 (b) District of Columbia incladed in both rates and Indiana in the general death rate in addition to the 
 states mentioned. 
 
 During this period of 11 years a greater decline in the infant 
 death rate than the general death rate for all ages occurred 
 in all of the 9 states included in the above table except 2, New 
 Hampshire and Michigan. In the former the decline in the 
 two rates was exactly equal. All evidence seems to point, 
 therefore, to the conclusion that the decline in the infant mor- 
 tality rate during the last 10 or 15 years has been greater than 
 that in the general death rate for all ages. 
 
 Before leaving this subject it will be advisable to compare 
 the decline in mortality by age. This is possible from the 
 figures given in the following table comparing the per cent, 
 of decrease in the death rate for persons of different ages be- 
 tween 1900 and 1911 for the group of registration states as 
 constituted in 1900: 
 
 PER CENT. OF DECREASE IN THE DEATH RATE PER 1,000 POPULATION FOR CERTAIN 
 AGE GROUPS BETWEEN 1900 AND 1911, FOR THE STATES INCLUDED IN THE REG- 
 ISTRATION AREA IN 1900. (a) 
 
 All ages 13 
 
 Under 1 year 22 
 
 1 to 4 years 35 
 
 5 to 9 years 32 
 
 10 to 14 years 27 
 
 15 to 19 years 27 
 
 20 to 24 years 26 
 
 25 to 34 years 23 
 
 35 to 44 years 9 
 
 45 to 54 years. 3 
 
 55 to 64 years +4 
 
 65 to 74 years +3 
 
 75 years and over 
 
 A plus sign (+) denotes an Increase. 
 
 (a) Twelfth Annual Report of the Bureau of the Census on Mortality Statistics, p. 22. 
 
16 American Statistical Association. 
 
 An examination of these figures shows that the death rate 
 for all age groups under 55 decreased between 1900 and 1911. 
 The greatest decrease was for the age group 1 to 4 years, the 
 per cent, of decrease falling off with each succeeding age group 
 until the period from 55 to 64 years was reached, this and the 
 next group showing a small increase. The death rate above 
 75 years was practically the same in each period. The per 
 cent, of decrease in the mortality of the first year of life was 
 noteworthy (22) but it was exceeded by that of the years of 
 both childhood and adolescence, being about equal to that for 
 the age group 25 to 34 years and greater than that for all suc- 
 ceeding groups. 
 
 It has thus been shown that since 1881 the rate of infant 
 mortality has been declining in practically all European coun- 
 tries for which statistics are available, in the Australian Com- 
 monwealth and New Zealand, and in Massachusetts and Bos- 
 ton in this country. This decline in infant mortality has 
 been especially marked in the last twelve or fifteen years dur- 
 ing which period it has practically everywhere exceeded the 
 decline in the general death rate for all ages. It has also been 
 shown that between 1900 and 1911 a marked decline in the 
 infant death rate per 1,000 population under one year of age 
 occurred in the states and large cities of the registration area. 
 Here, too, the decline in the infant death rate was with one 
 or two exceptions found to be greater than that in the general 
 death rate. On comparing the per cent, of decrease in the 
 death rate for infants under one year of age with that for 
 other ages it was found that the decline in infant mortality 
 was less than that for children or young persons under 25 
 years of age but about equal to that for persons in the age 
 group 25 to 35 years and greater than that for persons over 
 35. It appears, therefore, that the decline in the infant death 
 rate has in general been greater than the decline in the adult 
 death rate; but the decline in the infant death rate has not 
 been as great as the decHne in the death rate for persons in 
 the years of childhood and adolescence. 
 
II. 
 
 THE INFLUENCE OF PRENATAL CONDITIONS ON 
 INFANT MORTALITY 
 
 Reprinted from the Proceedings of the Southern Sociological 
 
 Congress, "The New Chivalry- Health," 
 
 1915, pp. 176-190. 
 
II. 
 
 THE INFLUENCE OF PRENATAL CONDITIONS ON 
 INFANT MORTALITY 
 
 The term "infancy" is usually defined as that period 
 extending from birth to the end of the first year of life, and 
 the problem of infant mortality as the excessively high 
 ratio of deaths to births during this period. It is obvious, 
 however, that any study of the causes of infant mortality 
 cannot be confined solely to the period of infancy. Children 
 come into the w^orld as the result of a long period of fetal, 
 embryonic, and germinal development, and their ability to 
 withstand the vicissitudes of life and to adjust themselves 
 to their new environments is largely determined by the 
 conditions surrounding them in their prenatal existence. 
 Just as the adult carries with him into manhood the effects 
 of the conditions with which he came into contact in infancy 
 and childhood, so the new-born infant brings with him into 
 postnatal life the results of a long period of prenatal and 
 even preconceptual development; and just as in accounting 
 for adult mortality careful consideration must be given to 
 the period of infancy and childhood, so in any attempt to 
 account for the problem of infant mortality a careful study 
 must be made of the conditions surrounding the infant in 
 its intra-uterine and preconceptual development. 
 
 There are, then, two distinct periods concerned in the 
 problem of infant mortality — the postnatal, or period of 
 infancy proper, and the prenatal, or period before birth. 
 Each of these is in turn further subdivided, as is shown in 
 the following chart: 
 
 SUBDIVISIONS OF THE PRENATAL AND POSTNATAL PERIODS OF 
 I INFANT DEVELOPMENT 
 
 r. The Prenatal Period. 
 
 1. The Germinal Period before Conception. 
 
 2. The Intra-uterine Period. 
 
 19 
 
20 THE NEW CHIVALRY — HEALTH 
 
 a. The Embryonic Period. (The first three months 
 
 of intra-uterine life.) 
 
 b. The Fetal Period. (From the end of the embry- 
 
 onic period to birth.) 
 
 II. The Postnatal Period. 
 
 1. The Period of Early Infancy. (The first three 
 
 months of life.) 
 
 2. The Period of Later Infancy. (The last nine months 
 
 of the first year of life.) 
 
 To understand the causes of infant mortality it is neces- 
 sary to keep these periods of infant development clearly in 
 mind and to take into account the character of the parental 
 stock that is united at conception, the circumstances 
 through which the infant's organism passes before birth 
 while in the mother's womb, the conditions of life to which 
 it must adjust itself immediately after birth, and the influ- 
 ences which affect its growth during the later months of 
 infancy after the adjustment to the demands of life has once 
 been made. 
 
 -^ During the period of early infancy, after the child has 
 been expelled from the mother's womb by birth, and adapta- 
 tion and adjustment to the new environment begun, its sur- 
 vival is determined not only by the conditions with which 
 it then comes into contact but also by the strength, adapta- 
 bility, and fitness to meet the normal demands of life which 
 it may have developed in the long period of germinal, embry- 
 onic, and fetal growth. The infant may be born already 
 diseased or malformed, it may have acquired before birth 
 a predisposition to disease, it may be congenitally so weak 
 that it is unable to cope with its new environment, its new 
 temperature, and its new method of receiving food and air. 
 The failure of such children to make the necessary adjust- 
 ment to vital conditions and to survive creates to a very 
 large extent, as will be shown later, the mortality of the 
 early weeks and months of infancy. 
 
THE INFLUENCE OF PRENATAL CONDITIONS 21 
 
 THE INFLUENCE OF PRENATAL CONDITIONS ON INFANT MOR- 
 TALITY PARTLY SOCIAL AND PARTLY HEREDITARY 
 
 Closer examination of the chart will bring out the part 
 played by heredity and by environmental conditions in 
 determining the extent of the influence of prenatal condi- 
 tions on infant mortality. In the broadest sense the term 
 "heredity," or the transmission of physical or mental char- 
 acters from parent to offspring, may be said to include all 
 that is implied in the term, "the prenatal factors of infant 
 mortality." On the other hand, this use of terms does not 
 emphasize sufficiently the importance of the environmental 
 influences which affect the growth and development of the 
 infant's organism in the intra-uterine period between con- 
 ception and birth. Better is it to say that the prenatal 
 factors of infant mortality are the product of both heredity 
 and the environment working together — the product of 
 heredity in that it is through this process that the organism 
 is brought into existence, and the product of the environ- 
 ment in that environmental influences acting through the 
 mother affect the development of the infant's organism from 
 conception to birth, 
 
 HThe prenatal influence of the environment may be seen, 
 for instance, in the influence of the nutrition of the mother 
 during pregnancy on the rate of infant mortality. As the 
 developing organism in the uterus is entirely dependent for 
 its food upon its mother, it is to be expected that the rate 
 of infant mortality will be affected by the character of the 
 food she obtains during pregnancy, as well as by the kind 
 which the infant obtains directly after its birth. Dr. 
 Herman Schwarz, in the first annual report of the New York 
 Free Outdoor Maternity Clinic, shows how this occurred in 
 the case of 2,540 infants whose mothers came under the 
 observation of the clinic during the first nine years of its 
 work.* Thus, the rate of mortality was 144 deaths per 
 1,000 births among the infants whose mothers' food was 
 poor, 124 where it was fair, and only 62 where it was good. 
 In the same way the character of the work which the 
 
 *1910, p. 45. 
 
V 
 
 22 THE NEW CHIVALRY — HEALTH 
 
 mother does during the later stages of pregnancy and the 
 amount of rest she takes before confinement have an impor- 
 tant prenatal influence on infant mortality. To quote from 
 a recent writer on this subject : "A number of investigations 
 have shown that the state of the infant at birth is greatly 
 affected by the conditions under which the mother has lived 
 during the previous months. The children of working 
 women who are able to rest during the later months of preg- 
 nancy are to a marked degree larger and finer than the 
 children of working women who have pursued their occupa- 
 tion to within a short time of their confinement, even 
 though the women who thus pursue their work may be 
 entirely healthy and robust. Moreover, such rest is a pow- 
 erful agent in preventing premature birth. This is an 
 important matter, for . . . the child who is born before its 
 time comes into the world in a relatively unprotected state, 
 and is unduly liabLe to perish or else to lead a permanently 
 enfeebled life, . . , Opportunity for completing its develop- 
 ment is of immense and lifelong importance to the newborn 
 infant, while the rest is also of benefit to the mother, who 
 cannot with impunity stand the double strain of work and 
 of nourishing the future child within her. Yet the impor- 
 tance of such rest for women in its bearing on the elevation 
 of the race and the lightening of social burdens is still under- 
 stood by few."* 
 
 Moreover, although this aspect of the question is not 
 often emphasized, this adverse effect of the mother's work 
 during the later stages of pregnancy may result from the 
 work of mothers who are employed only in the performance 
 of their own household duties, as well as from the work of 
 those who are employed in gainful occupations. While it is 
 true that the prenatal effect of gainful employment on infant 
 mortality is much more serious in most cases than the effect 
 of work which the mother does in her own home as a part 
 of her household duties, it does not follow that all the empha- 
 sis should be laid upon gainful employment. All mothers 
 need rest and care before confinement and any program for 
 
 *Havelock Ellis, "The Problem of Race Regeneration." London, 
 1911, pp. 18-19. 
 
THE INFLUENCE OF PRENATAL CONDITIONS 23 
 
 the control or reduction of the influence of prenatal condi- 
 tions on infant mortality should include within its scope 
 better care for all classes of expectant mothers, whether rich 
 or poor, and whether employed in gainful or non-gainful 
 occupations. 
 
 PREVENTIVE METHODS : THE CONTROL AND REDUCTION OF THE 
 
 INFLUENCE OF PRENATAL CONDITIONS ON INFANT 
 
 MORTALITY 
 
 The distinction between the influence of heredity in the 
 strict sense and of the environment during the prenatal 
 stage of infant development has been clearly drawn in the 
 campaign for the prevention of infant mortality. On the 
 one hand are the efforts that are being made to prevent the 
 marriage of the unfit, the diseased, and the incapable, and 
 on the other the efforts that are being made to improve the 
 conditions under which expectant mothers must live and 
 to raise the standards of "mother-care" during pregnancy. 
 The former efforts to control the influence of hereditary 
 tendencies usually have primarily in view the improvement 
 of the race stock, the conservation of a desirable strain, or 
 the elimination of an undesirable strain, while in the latter 
 efforts to control the conditions under which the infant's 
 organism develops between conception and birth the chief 
 end in view is to enable each child that is to come into the 
 world to bring with it unimpaired all the capacities that 
 heredity bestowed upon it at conception and to give to its 
 organism a fair opportunity to develop between conception 
 and birth. This form of prenatal work is simply an effort 
 to conserve and protect the developing organism before birth 
 in the same way that we have long felt that the growing child 
 should be protected after birth. As one writer has expressed 
 it, "it is not enough to begin the social care of the child at 
 birth. It has been living for nine months before birth, and 
 it is now recognized that the conditions of its life must be 
 guarded by society during that supremely important form- 
 ative period." We must recognize the fact that from the 
 standpoint of infant mortality the trouble is by no means 
 only with the fitness of the parents for bringing children 
 
24 THE NEW CHIVALRY — HEALTH 
 
 into the world; too frequently it is with the fitness of the 
 environment in which the mother must live during preg- 
 nancy. 
 
 This is not the place for an account of the specific meth- 
 ods that have been proposed or which are now being applied 
 in the reduction and control of the influence of prenatal 
 conditions on infant mortality. The fact must be kept 
 clearly in mind, however, that the importance of this aspect 
 of the problem is not generally realized. One distinguished 
 authority has gone so far as to say that "to-day the dregs 
 of the human species — the blind, the deaf-mute, the degen- 
 erate, the nervous, the vicious, the idiotic, the imbecile, the 
 cretin, and the epileptic — are better protected than preg- 
 nant women."* This is, of course, an extreme statement, but 
 it is probably true that these dependent and degenerate 
 classes do receive more care and especial consideration 
 because of their condition than expectant mothers do he- 
 cause of theirs. Expectant mothers, because they are expect- 
 ant mothers, should receive more care and consideration 
 than they are given to-day. 
 
 THE INFLUENCE OF PRENATAL CONDITIONS ON THE MORTALITY 
 OF INFANTS IN THE EARLY WEEKS AND MONTHS OF LIFE 
 
 An idea of the extent of the influence of prenatal condi- 
 tions on infant mortality can be obtained from the following 
 figures comparing the per cent of infant deaths which oc- 
 curred in the registration area of the United States in 1910 
 at certain age periods :t 
 
 Under 1 week 23.5 
 
 1 week to 1 month 14.1 
 
 1 month to 3 months 18.3 
 
 3 to 6 months 19.3 
 
 6 to 9 months 14.1 
 
 9 to 12 months 10.7 
 
 *Quoted in Havelock Ellis's "The Problem of Race Regeneration." 
 London, 1911, p. 20. The writer does not cite the author of the 
 statement. 
 
 tU. S. Bureau of the Census: "Mortality Statistics," 1910, p. 533. 
 
THE INFLUENCE OF PRENATAL CONDITIONS 25 
 
 Thus, the deaths of infants are not distributed evenly 
 throughout the first year of life, 38 per cent occurring during 
 the first month and over half during the first three months. 
 Moreover, the number of children dying during the first 
 week of infancy was almost exactly equal to the number 
 dying during the entire last six months of the period. In 
 this truly astounding fact we see the influence of prenatal 
 conditions upon infant mortality. The failure of such a 
 large proportion of the babies who are born to survive more 
 than a few weeks or months can be ascribed only to the 
 influence of conditions affecting them before birth, during 
 what we have called the prenatal period of infant develop- 
 ment. 
 
 THE INFLUENCE OF PRENATAL CONDITIONS ON THE INFANT 
 , DEATH RATE FROM THE DISEASES OF EARLY INFANCY 
 
 But the influence of prenatal conditions on infant mor- 
 tality is also to be seen in the large proportion of deaths of 
 infants which result from premature birth, congenital 
 debility, congenital malformations, and other causes of 
 death having their origin in conditions affecting the infant's 
 organism before birth. It will therefore be necessary to 
 examine the causes of the deaths of the 154,373 babies who 
 died in the registration area of the United States in 1910. 
 The following figures compiled from Bulletin 109 of the 
 Bureau of the Census show the per cent of the total number 
 of infants who died from each cause : 
 
 Diseases of the digestive system 32.1 
 
 Diseases of early infancy (including premature birth, 13.1 per 
 
 cent, congenital debility, 7.8, and injuries at birth, 2.4) .... 25.5 
 
 Diseases of the respiratory system (including broncho-pneu- 
 monia, 6.9, pneumonia, 5.5, and acute bronchitis, 2.7) 15.8 
 
 General diseases (including all forms of tuberculosis, 1.6, syph- 
 ilis, 1.1, and whooping cough, measles, scarlet fever, diph- 
 theria, and croup, 4.0) 9.2 
 
 Diseases of the nervous system (including convulsions, 2.6, and 
 
 meningities, 1.5) 5.5 
 
 Congenital malformations 4.9 
 
 All other causes '^•4 
 
26 THE NEW CHIVALRY — HEALTH 
 
 V' 
 
 It will thus be seen that 32" per cent of the infant deaths 
 occurring in this representative year in the registration 
 area of the United States were reported as being caused by 
 the diseases of the digestive system, 26 per cent by the dis- 
 eases of early infancy, and 16 per cent by the diseases 
 of the respiratory system — nearly 74 per cent in all being 
 reported as due to these three groups of causes. Before 
 drawing any conclusions as to the significance of these facts 
 it will be necessary to inquire in more detail as to what age 
 periods of infancy are most affected by each of these great 
 causes of infant deaths. This is possible from the following 
 data taken from the annual report of the health department 
 of the city of Boston for 1910:* "Fifty-five per cent of the 
 deaths of infants under three months of age which occurred 
 in Boston during this representative year were reported as 
 having been caused by the diseases of early infancy (especi- 
 ally premature birth, congenital debility, and injuries at 
 birth), and 9 per cent by a similar cause, congenital mal- 
 formations, while the deaths reported under all these head- 
 ings combined caused only 11 per cent of the total number 
 of deaths which occurred during the last nine months of 
 infancy. In comparison the diseases of the digestive system 
 were reported as causing 16 per cent of the deaths of the 
 first three months of infancy, almost half of those of the 
 second three months, 35 per cent of those of the third three 
 months, and 40 per cent of those which occurred during the 
 last three months of the period. In a similar manner the 
 deaths from the diseases of the respiratory system increased 
 from 11 per cent during the first three months of infancy to 
 26 per cent during the last, and those from general diseases 
 from 5 per cent during the first three months to 19 per cent 
 during the last three months of the period." 
 
 In other words, congenital malformations and the dis- 
 eases of early infancy are the prime factors in the mor- 
 tality of the first three months of infancy and the diseases 
 of the digestive and respiratory systems the prime factors 
 
 *Such figures are not available for the registration area of the 
 United States. 
 
THE INFLUENCE OF PRENATAL CONDITIONS 27 
 
 in the mortality of the last nine months. What, then, is the 
 nature of these diseases which so completely dominate the 
 mortality of the early weeks and months of life and what 
 are the conditions determining their influence? 
 /^ The diseases of early infancy comprise one of the thirteen 
 classes of causes of death in the international classification. 
 Within it are included those infants who die of such causes 
 as congenital debility, premature birth, injuries at birth, 
 and the like. The group also includes such causes of death 
 as "marasmus," "inanition," "wasting disease," "constitu- 
 tional weakness," "congenital malnutrition," and the like. 
 It is only necessary for the lay reader, even if he has but a 
 small knowledge of the meaning of medical terms, to glance 
 at these titles to see that practically all of them imply that 
 the child was not born with sufficient strength to withstand 
 the normal demands of life for more than a few weeks or 
 months at most.* Clearly the deaths of infants so soon 
 after birth as a result of such causes — whether they grow 
 out of prematurity of birth, an accident or injury at birth, 
 or to simple congenital inability to survive — must be due to 
 something besides the influence of the environment into 
 which they are born. It is not the postnatal environment 
 or its diseases which really bring about their death, but 
 their absolute unfitness to withstand the normal demands of 
 life itself. Such deaths are clearly attributable to prenatal 
 conditions, to heredity, and to influences affecting the child's 
 organism before birth. 
 
 THE EXTENT OF THE INFLUENCE OF PRENATAL CONDITIONS ON 
 INFANT MORTALITY 
 
 There are two other causes of death which are so closely 
 associated in origin with the diseases of early infancy, being 
 
 *The only exception is in the case of lack of care; but since the 
 number of infant deaths reported as due to this cause is so small, the 
 exception is of little consequence. In Boston only two and in the 
 registration area of the United States only 132, or in each case only 
 about one-tenth of one per cent of the total number of infant deaths 
 in 1910, were reported as due to this cause. 
 
28 THE NEW CHIVALRY — HEALTH 
 
 also the result of conditions affecting the child's organism 
 before birth, that their influence must also be counted with 
 them in order to fully estimate the influence of prenatal 
 conditions on the rate of infant mortality — congenital mal- 
 formations and syphilis, the former including infant deaths 
 resulting from some deformity acquired by the child's 
 organism before birth (such as congenital intestinal obstruc- 
 tion, harelip, malformation of the brain or heart, and the 
 like), and the latter in the great majority of cases resulting 
 from syphilitic infection before birth. In 1910 congenital 
 malformations were reported as causing 5.3 per cent of the 
 deaths of infants under one year of age in Boston and 8.6 
 per cent of those under three months, while syphilis was 
 reported as causing 1.0 per cent of the deaths under one year 
 and 1.2 per cent of those under three months of age. 
 
 Combining the deaths from these two causes with those 
 from the diseases of early infancy, it will be seen that at 
 the very least 35 per cent of the deaths occurring during 
 the first year and 60 per cent of those during the first three 
 months of life and a much larger proportion still of those 
 which occur during the first month are largely the result of 
 prenatal conditions. This proportion, astounding as it may 
 seem, cannot be regarded as much too large, because any 
 infants included whose deaths may not have been "largely 
 due" to conditions affecting their organisms before birth 
 are probably compensated for by the returns of other infant 
 deaths actually due to congenital debility, syphilis, congenital 
 malformation, or premature birth, under some other heading. 
 
 But even according to the most conservative estimate 
 the influence of prenatal conditions on infant mortality must 
 be regarded as of great importance not only through its 
 contribution to the death rate from chief causes but from 
 contributory causes as well; and not only upon the rate of 
 infant mortality but on child mortality also ; for many chil- 
 dren, although their chances of life in the period of infancy 
 are greatly reduced by congenital influences, still survive 
 until after the end of the first year of life, thus tending to 
 increase the death rate for later years. 
 
THE INFLUENCE OF PRENATAL CONDITIONS 29 
 
 COMPARISON OF THE RECENT DECLINE IN THE INFLUENCE OF 
 
 PRENATAL AND POSTNATAL CONDITIONS ON 
 
 INFANT MORTALITY 
 
 During recent years the infant death rate for the first 
 weeks and months of infancy lias not declined as rapidly as 
 the rate for the latter and middle months of the period. It 
 is necessary to examine the actual figures in some detail 
 since, as will be clear from the discussion in the previous 
 section, they indicate that the influence of the prenatal fac- 
 tors of infant mortality is not decreasing as rapidly as that 
 of the postnatal factors, or as rapidly as the infant mortality 
 rate. 
 
 The following table shows the per cent of decrease in the 
 infant mortality rate per 1,000 births for England and 
 Wales between 1888-92 and 1908-11 by age: 
 
 Under Under 3 to 6 6 to 12 
 
 Years 1 Year 3 Months Months Months 
 
 1888-92 145.4 70.5 29.5 45.4 
 
 1893-97 153.4 73.7 32.0 47.7 
 
 1898-1902 152.2 74.3 32.3 46.1 
 
 1903-07 131.1 67.3 25.9 37.9 
 
 1908-11 116.1 62.1 21.9 32.1 
 
 Per Cent of Decrease 20.2 11.9 25.8 29.3 
 
 An examination of this table shows that in England and 
 Wales during the twenty-four years under consideration the 
 rate of infant mortality for the last six months of the first 
 year of life declined about 29 per cent and that for the sec- 
 ond three months about 26 per cent, while the rate for the 
 first three months declined only about 12 per cent. The 
 registrar-general for England and Wales in his annual 
 reports on births, deaths, and marriages from which the 
 above figures were compiled also gives similar data for the 
 city of London. The decline in the rate of mortality for the 
 first three months of infancy was only 16.4 per cent, while 
 the decline for the second three months was 29.1 and 31.6 
 per cent respectively. Thus the decline in the infant mor- 
 tality rate for the first three months of age— the period so 
 
30 THE NEW CHIVALRY — HEALTH 
 
 largely influenced by the prenatal factors of infant mor- 
 tality — was in England and Wales less than half and in 
 London somewhat more than half as great as that for either 
 the second three months or the last six months of infancy. 
 Similar figures are available for this country only for the 
 last four years. The following table shows for 1910-11 and 
 1912-13 the average annual number of infant deaths in the 
 registration area at certain age periods and in addition the 
 crude death rate for each age period per 100,000 population 
 of all ages.* So crude a rate is, of course, open to serious 
 objections; but it is better than nothing, and is therefore 
 included. 
 
 Average Annual Crude 
 
 Number of Deaths Death Rate 
 
 Age at Death 1910-11 1912-13 1910-11 1912-13 
 
 Under 1 year 151,848 153,446 268.4 248.0 
 
 Under 1 month 60,491 66,472 107.0 107.5 
 
 1 to 3 months 27,198 26,496 48.1 42.8 
 
 3 to 6 months 28,062 26,520 49.6 42.9 
 
 6 to 9 months 20,452 19,016 36.2 30.7 
 
 9 to 12 months 15,645 14,940 27.7 24.2 
 
 An examination of the table shows that during the last 
 four years the death rate per 100,000 total population for 
 the first month of infancy remained practically stationary. 
 The rate for each of the later age periods decreased ; but the 
 decrease in the rate for the period 1 to 3 months (11 per 
 cent) was less than that for either of the later periods (13 
 to 15 per cent) . 
 
 This same tendency may also be illustrated with figures 
 for Boston. Thus during the same period the infant mor- 
 tality rate per 1,000 births for the first month of infancy 
 
 ^Compiled from the Eleventh to the Fourteenth Annual Reports of 
 the Bureau of the Census on Mortality Statistics. The four years here 
 considered do not furnish a sufficiently long period to establish inde- 
 pendently a definite conclusion as to the shifting of the age incidence 
 of infant mortality in this country, but when they are supplemented 
 by the figures already given for England and Wales and London they 
 may be regarded as furnishing a fairly adequate indication of existing 
 tendencies. 
 
THE INFLUENCE OF PRENATAL CONDITIONS 31 
 
 remained practically stationary (46.9 in 1910-11 and 46.0 
 in 1911-12) ; while for each subsequent age period it 
 decreased, the decline in the rate for the age period 1 to 3 
 months (20 per cent) being greater than that for any other 
 age period (11 per cent for the period 3 to 6 months, 15 for 
 the period 6 to 9, and 19 for the period from 9 to 12 
 months).* The marked decline in the rate for the period 
 from 1 to 3 months is especially interesting, since it may be 
 partly due to the efficient "prenatal work" with expectant 
 mothers that has been carried on in Boston throughout the 
 four years. Perhaps if this work had not been done the 
 decline in the rate for this age group would have been no 
 larger than that which occurred in the registration area. 
 
 Thus such figures as are available in this country reveal 
 tendencies which are in exact accord with those shown by 
 figures extending over a much longer period for England 
 and Wales and for London. Now if this conclusion be 
 accepted its significance is not difficult to discover, for, as 
 has already been shown, the mortality of early infancy, and 
 particularly the mortality of the first month, is largely 
 determined by deaths from causes growing out of influences 
 affecting the child's organism before birth — or, in other 
 words, from the influence of prenatal conditions — while the 
 mortality of the later months of infancy is largely deter- 
 mined by the influence of conditions affecting the child's 
 organism after birth — or, in other words, to the influence 
 of postnatal conditions. Any changes in the incidence of 
 mortality in these periods, therefore, is an indication that 
 similar changes have occurred in the relative influence of 
 the prenatal and postnatal factors of infant mortality. The 
 conclusion seems well founded that during recent years the 
 influence of prenatal conditions on infant mortality has not 
 declined as rapidly as the influence of postnatal conditions. 
 
 In accounting for this stability in the rate of mortality 
 for the early weeks and months of infancy, especial con- 
 sideration must be given to the fact that deaths in the first 
 
 *These figures were compiled from the Annual Reports of the 
 Health Department of Boston. 
 
32 THE NEW CHIVALRY — HEALTH 
 
 weeks and months of life are not so readily affected by the 
 action of preventive methods as those in the later months 
 of infancy, and that the influence of the prenatal factors of 
 infant mortality are not as easily controlled as the influence 
 of the postnatal factors. Yet, while this is true, it should 
 not be forgotten that not until the last few years were the 
 methods employed in bringing about a reduction in the 
 general death rate for all ages, as well as in the rate for the 
 middle and later months of infancy, such as could be 
 expected to have much influence on prenatal conditions. The 
 movement for the improvement in the water and milk supply 
 and for the elimination of contagious and infectious dis- 
 eases, for instance, both of which have been important 
 factors in the decline of the death rate for adults and young 
 children as well as of infants over two or three months of 
 age, have had in comparison only an incidental effect on the 
 mortality rate for the first two or three months of life. In 
 the campaign for the direct purpose of preventing infant 
 mortality for a long time attention was confined primarily 
 to the prevention of infectious and contagious diseases and 
 to the reduction of the death rate from digestive diseases. 
 It was not until a comparatively few years ago that any 
 organized, systematic efforts were made to lessen the num- 
 ber of infant deaths brought about by the action of prenatal 
 conditions. Milk stations, valuable as they have been, were 
 not important factors in "prenatal infant mortality" until 
 they became "infant welfare stations" and added prenatal 
 instruction to their work and prenatal nurses to their staflLS. 
 The method that has been adopted and found most effect- 
 ive in coping with this phase of infant mortality is, as has 
 already been shown, what is called prenatal work with 
 expectant mothers. It cannot be described here more fully 
 than has been done already except to say that it is simply 
 an attempt to give direct instruction to expectant mothers 
 at different times during pregnancy, usually as early as 
 possible before delivery, in the proper care of themselves 
 and their baby to come. This kind of work is usually carried 
 on in the poorer quarters of the city, but this is not the only 
 
THE INFLUENCE OF PRENATAL CONDITIONS 33 
 
 or even the most important phase of "prenatal care." The. 
 increasing interest that is being taken in the whole question 
 of the effect of prenatal conditions on mortality and the 
 increasing emphasis that is being laid upon the prenatal 
 care of all mothers regardless of their economic or social 
 condition will have and is having an important influence on 
 the problem. It is to be hoped and expected, therefore, that 
 the next five or ten years will show a greater decline in 
 the influence of prenatal conditions on infant mortality than 
 has occurred in previous years. 
 
III. 
 
 INFANT MORTALITY AND THE SIZE OF THE 
 FAMILY 
 
 Reprinted from the Quarterly Publications of the American 
 
 Statistical Association, New Series, No. Ill, 
 
 September, 1916, pp. 629-641. 
 
III. 
 
 INFANT MORTALITY AND THE SIZE OF THE 
 FAMILY. 
 
 The influence of the birth rate on the rate of infant mortality 
 has been frequently pointed out. Newsholme in his recent 
 report to the Enghsh Local Government Board notes "the 
 connection often observed between a high birth rate and a 
 high rate of infant mortality."* Phelps, also, in his study of 
 Infant Mortality and Its Relation to Women's Employment, 
 states that in Massachusetts cities a "more direct relationship 
 exists between infant mortality and the birth rate . . . 
 than infant mortality and the employment of women," 
 although in neither case, as his figures bring out, is the relation- 
 ship invariable or even close, f The Director of the Statis- 
 tical Service of France, on the other hand, has shown from an 
 investigation made in 1907 by the Superior Council of Statis- 
 tics that whatever relationship there may be between the 
 birth rate and the infant mortality rate in cities or countries 
 there is a very close relationship between the birth rate in 
 families — or, what is practically the same, a close relationship 
 between the size of the family and the rate of child mortality. 
 It is not possible from the figures given in his study to separate 
 the statistics for infants under one year of age, but by consid- 
 ering only the period from 1901-1907 it will be possible to se- 
 cure a group with a maximum age at death of six years. 
 This is done in the following table, which shows the mortality 
 rate for infants whose fathers were employed in the public 
 service of France according to their order of birth, or, in other 
 words, according to the number of previous births in the family : 
 
 *Great Britain, Local Government Board, Supplement to the Thirty-ninth Annual Report, p. 49. 
 
 tPhelps, Edward B., "Infant Mortality and Its Relation to Women's Employment — A Study of Mass- 
 achusetts Statistics." In Volume XIII, Part 1, of the Bureau (Now Department) of Labor's Report on 
 Condition of Women and Child Wage Earners in the United States, p. 38. 
 
 37 
 
38 
 
 American Statistical Association. 
 
 TABLE I. 
 
 MORTALITY RATE PER 1,000 BIRTHS AMONG CHILDREN WHOSE FATHERS WERE EM- 
 PLOYED IN THE PUBLIC SERVICE OF FRANCE CLASSIFIED ACCORDING TO ORDER 
 OF BIRTH, 1901-1907. (a) 
 
 Order of Birth. 
 
 Number of Births. 
 
 Infant Mortality Rate. 
 
 
 25,102 
 
 21,384 
 
 14,675 
 
 9,750 
 
 6,430 
 
 4,251 
 
 2,783 
 
 1,842 
 
 1,215 
 
 826 
 
 455 
 
 687 
 
 113.7 
 
 
 121.5 
 
 Third bom 
 
 139.4 
 
 
 148.5 
 
 Fifth bom 
 
 165.2 
 
 
 173.2 
 
 
 186.5 
 
 Eighth bom 
 
 204.5 
 
 
 187.7 
 
 
 236.2 
 
 
 248.4 
 
 
 276.6 
 
 
 
 Total 
 
 89,400 
 
 138.3 
 
 
 
 (a) Compiled from Lucien March's "Some Researches Concerning the Factors of Mortality." Journal 
 of the Royal Statistical Society, Vol. LXXV, Part 5, p. 519. (Cf. Appendix.) 
 
 This table shows a strikingly close relationship between 
 child mortality and the order of birth, the rate having in- 
 creased from 113.7 per 1,000 births among the first born 
 children to 276.6 among those who were twelfth born or over.* 
 As has just been stated the relationship shown in this table is 
 between the order of birth and the rate of mortality for 
 children under six years of age. The New York Free Out- 
 door Maternity Clinic has recently published data collected 
 during the first nine years of its work which show the relation- 
 ship between infant mortality and the number of children to 
 which the mother has previously given birth — a classification 
 for all practical purposes identical with that in Table I, 
 based on the order of birth. 
 
 *The only exception to the regular variation of the mortality rate with the order of birth was in the 
 ninth group but this is not a serious exception, as it may be due to chance. 
 
 Dr. Alice Hamilton of Hull House, Chicago, in a recent article, "Excessive Child-bearing as a Factor 
 in Infant Mortality" (Proceedings of the Conference on the Prevention of Infant Mortality, New Haven, 
 1909, pp. 74-80) , shows a similar relationship between the order of birth, or size of the family, and the 
 mortality rate for children under three years of age. Thus, among 1,600 infants bom to mothers classi- 
 fied according to the number of children in their family the mortality rate per 1,000 births was as follows: 
 
 4 children or less 118 
 
 6 children or more 267 
 
 7 children or more 280 
 
 8 children or more 291 
 
 9 children or more 303 
 
 Among the Berlin working class Hamburger's investigation has also shown how serious a cause of in- 
 fant mortality large families may be. (Kinderzahl und Kindersterblichkeit. Die Neue Generation, 
 August, 1909). Quoted in Havelock Ellis' "The Task of Social Hygiene," pp. 150-1. Space does not 
 permit the quoting of his figures in full. 
 
. Infant Mortality and the Size of the Family. 
 
 39 
 
 TABLE II. 
 
 MORTALITY RATE PER 1,000 BIRTHS FOR INFANTS BORN TO MOTHERS COMING 
 UNDER THE OBSERVATION OF THE NEW YORK FREE OUT-DOOR MATERNITY 
 CLINIC, CLASSIFIED ACCORDING TO THE NUMBER OF PREVIOUS CHILDREN 
 BORN TO MOTHER, (a) 
 
 Number of Previous Children Bom to Mother. 
 
 Number of Births. 
 
 Infant Mortality Rate. 
 
 Less than 4 
 
 1,182 
 
 1,064 
 
 294 
 
 97 
 288 
 369 
 428 
 355 
 354 
 203 
 152 
 135 
 90 
 69 
 
 77 
 
 4 to 8 
 
 127 
 
 8 and over 
 
 170 
 
 
 
 10 
 
 1 
 
 66 
 
 2 
 
 81 
 
 3 
 
 96 
 
 4 
 
 127 
 
 5 
 
 189 
 
 6 
 
 163 
 
 7 
 
 105 
 
 8 
 
 207 
 
 9 
 
 166 
 
 10 and over 
 
 275 
 
 
 
 Total 
 
 2,540 
 
 124 
 
 
 
 (a) Compiled from the First Annual Report of the Free Out-door Maternity Clmic, Covering the First 
 Nine Years of the Clinic's Existence, New York, 1910: Chapter 2, Fart 2, Report of the Pediatric De- 
 partment with a Study of Early Infant Mortality, by Herman Schwarz, M.D., pp. 42-3. 
 
 Table II shows a very close relationship betvs^een the num- 
 ber of children to which the mother has previously given birth 
 and the rate of infant mortality during the year under consid- 
 eration. Thus, the mortality rate rose from 77 deaths per 
 1,000 births for infants born to mothers who had previously 
 given birth to less than four children to 170 for the infants 
 whose mothers had given birth to eight or more.* 
 
 In the investigation by the Children's Bureau of infant mor- 
 tality in Johnstown, Pa., the data gathered bearing on the 
 relationship between fertility and the size of the family and 
 infant mortality were classified according to the order of birth ; 
 but, as has already been shown, the terms "order of birth" 
 and "number of the mother's previous pregnancies," though 
 not absolutely identical, may be considered as practically so 
 for our purpose. 
 
 *The exceptions to the otherwise continuous relationship of these mortality rates with the number of 
 children to which the mother had previously given birth which appear when, as in the lower part of the 
 table, each group is considered separately, will be discussed in connection with a later table showing similar 
 data gathered in the Boston investigation. 
 
40 
 
 American Statistical Association. 
 
 TABLE III. 
 
 MORTALITY RATE PER 1,000 BIRTHS FOR INFANTS INCLUDED IN THE JOHNSTOWN 
 
 INVESTIGATION CLASSIFIED ACCORDING TO THE ORDER OF THEIR BIRTH, (a) 
 
 Order of Birth. 
 
 Number of Births. 
 
 Infant Mortality Rate. 
 
 
 622 
 400 
 241 
 137 
 91 
 
 138.3 
 
 
 143.2 
 
 
 177.0 
 
 
 181 5 
 
 
 201.1 
 
 
 
 Total 
 
 1,491 
 
 149.9 
 
 
 
 (a) TJ. S. Children's Bureau: Infant MortaUty — Results of a Field Study in Johnstown, Pa., Based on a 
 Calendar Year, by Emma Duke. Washington, 1915, p. 51. Hereafter this report will be referred to by 
 the briefer title— Infant Mortality: Johnstown, Pa. 
 
 Thus, in Johnstown a relationship appears to exist between 
 fertility and the size of the family and the rate of infant 
 mortality similar to that shown in previous tables for other 
 investigations : * 
 
 It is obvious that in this relationship between infant mor- 
 tality and the order of birth, or the number of previous children 
 to which the mother has given birth, there are two distinct 
 factors present — that of the size of the family in which the 
 child lives after birth and that of fertility in its narrowest 
 sense, including primarily the physical influence of childbear- 
 ing upon the mother and upon the chances of survival of sub- 
 sequent infants to which she gives birth. Better expressed, 
 this relationship is brought about by the influence of both 
 prenatal and postnatal conditions on infant mortality. With 
 this in view an attempt was made in an investigation made in 
 1910 and 1911, by the Research Department of the Boston 
 School for Social Workers, to take account of both of these 
 factors and as far as possible to measure the influence of each.f 
 
 *For other figures showing the relationship between infant mortality and the order of birth in the family 
 see— R. J. Ewart, "The Aristocracy of Infancy and the Conditions of Birth," Eugenics Re\aew, Vol. Ill, 
 p. 166. 
 
 fThe data collected in this investigation have not been before published. A brief description of the 
 investigation would, therefore, be desirable if space permitted. Visits were made to the homes of 2,063 
 infants (stillborn infants not included) who were born in 1910 in Wards 6, 8, 13, and 17 of the city of Bos- 
 ton. The birth and death records were copied from the files of the Registry Department of the city and the 
 visits to the homes made by fellows in the Research Department of the School for Social Workers during 
 the academic years 1910-1911 and 1911-1912. During the second year of the investigation this field work 
 was supervised by the writer under the general direction of Dr. T. W. Glocker, director of the Depart- 
 ment of Research. The writer in using these data for publication wishes gratefully to acknowledge the 
 interest and cooperation of the fellows who made the visits to the homes, as well as that of the di- 
 rector of the Department of Research. He is also indebted to Dr. J. R. Brackett, Director of the 
 School for Social Workers, for permission to use the data in this way. 
 
Infant Mortality and the Size of the Famihj. 
 
 41 
 
 The method used was that of dividing the infant deaths into 
 two groups, those reported as due to the diseases of early 
 infancy and congenital malformation and those reported as 
 due to all other causes — the former group being largely domi- 
 nated by prenatal and the latter by postnatal influences — 
 and showing the relation of the number of the mother's pre- 
 vious pregnancies to the infant mortality rate for each group. 
 These figures are given in the following table, which shows 
 the infant mortality rate per 1,000 births from all causes, 
 from the diseases of early infancy and congenital malforma- 
 tion, and from all other causes for infants classified according 
 to the number of their mother's previous pregnancies:* 
 
 TABLE IV. 
 
 MORTALITY RATE PER 1,000 BIRTHS FROM ALL CAUSES, FROM THE DISEASES OF 
 EARLY INFANCY AND CONGENITAL MALFORMATION COMBINED, AND FROM ALL 
 OTHER CAUSES FOR INFANTS BORN IN 1910 AND VISITED IN THE HOUSE-TO- 
 HOUSE INVESTIGATION IN BOSTON, CLASSIFIED ACCORDING TO THE NUMBER 
 OF THEIR MOTHER'S PREVIOUS PREGNANCIES. 
 
 
 Number 
 of Births. 
 
 Infant Mortality Rate per 1,000 Births from— 
 
 Number of the 
 
 Mother's Previous 
 
 Pregnancies. 
 
 All Causes. 
 
 Diseases of Early 
 
 Infancy and Congenital 
 
 Malformations. 
 
 All Other 
 Causes. 
 
 
 1,533 
 
 475 
 
 53 
 
 371 
 
 390 
 
 309 
 
 265 
 
 198 
 
 177 
 
 118 
 
 84 
 
 96 
 
 53 
 
 120.0 
 134.7 
 226.4 
 
 91.6 
 100.0 
 139.2 
 150.9 
 141.4 
 175.1 
 110.1 
 107.1 
 114.6 
 226.4 
 
 29.4 
 33.7 
 75.5 
 
 24.3 
 25.6 
 35.6 
 41.5 
 20.2 
 45.2 
 16.9 
 11.9 
 52.1 
 75.5 
 
 90.7 
 
 
 101.1 
 
 
 228.6 
 
 
 
 67.4 
 
 1 
 
 74.4 
 
 2 
 
 103.6 
 
 3 
 
 109.4 
 
 4 
 
 121.2 
 
 5 
 
 129.9 
 
 6 
 
 93.2 
 
 7 
 
 95.2 
 
 8 and 9 
 
 62.5 
 
 10 and over 
 
 228.6 
 
 Total (a) 
 
 2,061 
 
 126.2 
 
 31.5 
 
 94.6 
 
 
 
 (a) Information was not obtained in two instances. 
 
 Table IV shows that the infant mortality rate varies in 
 direct ratio with the number of the mother's previous preg- 
 nancies. Thus, among the infants born to the mothers with 
 less than five previous pregnancies, 120 died in every 1,000 
 
 *This classification of the mothers differs primarily from that used in the preceding table quoted from the 
 report of the New York Free Out-door Maternity Clinic in that it is based on the number of the mother 8 
 previous pregnancies including stillbirths instead of upon the number of livmg births escludmg sti Ibirtha 
 and, secondarily, in that, as previous pregnancies are dealt with, the one resultmg m the birth of the m- 
 fant under consideration is not counted. It was adopted largely because of the difficulty of distrnguishing 
 deaths soon after birth from stillbirths. 
 
42 American Statistical Association. 
 
 births, in comparison with 135 among those whose mothers 
 had had from five to ten previous pregnancies, and 226 among 
 those who had had ten or more.* The rate of mortality from 
 the diseases of early infancy and congenital malformation 
 combined and from all other causes varied in the same manner, 
 thus showing that the influence of the number of the mother's 
 previous pregnancies on infant mortality is both prenatal and 
 postnatal, t There are, then, two distinct factors to consider 
 in accounting for this relationship, first, the size of the family, 
 and second, fertility and the effect of childbearing on the 
 health and strength of the mother and her ability to bear 
 strong and healthy children. 
 
 *Closer examination of the lower part of the table where each pregnancy group is considered separately 
 shows that the relationship is not entirely continuous. Thus, although the rate increases in almost con- 
 tinuous succession with the number of the mother's previous pregnancies up to the sixth group, it then be- 
 gins to decrease and continues to do so until the group of mothers having eight and nine previous preg- 
 nancies is reached when it begins to rise again , until in the last group the highest rate of all appears. (The 
 rate for the last group when subdivided was 171 for the infants born to those mqthers ha^'ing 10 or 11 
 previous pregnancies and 333.3 for those having 12 or more.) Practically the same condition is also seen 
 to exist when the previous table compiled from the figures of the New York Free Out-door Maternity Clinic 
 is examined in a similar manner. 
 
 Yet, these exceptions do not disprove, aa might appear at first sight, the tendency shown when the 
 pregnancy groups were combined. The drop in the mortahty rate for the infants bom to the mothers 
 who had already had six or seven previous pregnancies is probably due to the fact that a large number of 
 pre\aous pregnancies not only decreases the chances of life of the infants born but, in the case of the weaker 
 mothers, also tends to make childbearing impossible or else so perilous that it is voluntarily refrained from. 
 It is, therefore, to be expected that the proportion of infant deaths to births will begin to decrease after 
 the fourth and fifth pregnancy group when the weaker mothers begin to drop out of the ranks of the child- 
 bearing, thus leaving in these groups a larger proportion of strong and healthy mothers whose children 
 will be relatively better fitted to survive. 
 
 That the low mortality rate for infants in the sixth and seventh group is due to the inclusion of a larger 
 proportion of strong and healthy mothers in them is shown by an examination of the last two columns 
 of the table where it will be seen that the drop in the rate from the diseases of early infancy and congenital 
 malformations — diseases largely due to the condition of the mother during and before pregnancy — is greater 
 than that from all other causes of death, these being largely the result of conditions arising after birth 
 and not so directly connected with the intra-uterine period of the child's development. But even though 
 from the standpoint of childbearing the physical condition of the mothers in the sixth and seventh groups 
 may be so good that it renders their children relatively immune from the effects of continuous childbearing, 
 in time, if they continue, their children will also be affected, as is shown by the renewed rise in the rate 
 for the groups following the seventh. 
 
 These exceptions, therefore, cannot be regarded as ^^tiating the tendencies shown when the pregnancy 
 groups are combined or as materially weakening the conclusion that the rate of infant mortality varies 
 strikingly with the number of the mother's previous pregnancies and the number of children to which she 
 has previously given birth. 
 
 fThis follows from the fact — a fact that space does not permit us to submit detailed proof of — that the 
 deaths of infants during the first week and month of life and to a lesser extent during the first three months 
 are largely the result of conditions which affected the child's organism before birth and while it was de- 
 veloping in the mother's womb, or in other words largely the result of prenatal conditions, while the deaths 
 of the later months of infancy are largely the result of conditions which affected the child after its birth, 
 or in other words to the influence of postnatal conditions. 
 
Infant Mortality and the Size of the Family. 
 
 43 
 
 In accounting for this relationship between the size of the 
 family and infant mortality several considerations should be 
 borne in mind. It is to be expected, for instance, that con- 
 gestion both as shown by the average number of persons per 
 room and the number of persons sleeping in the bedroom 
 with the infant will be greater among large than small families, 
 and this was found to be true in the Boston inquiry.* Povertj' , 
 too, is generally agreed to be worse, other things being equal, 
 where the number of small children in the family is large, t 
 Moreover, those parents who bring into the world larger 
 families than their neighbors deem themselves able to rear 
 properly are frequently improvident, with a low standard of 
 life, and in addition, are often characterized by a lack of in- 
 telligence or of sufficient knowledge of the simpler laws of 
 hygiene. This is, of course, not true of the parents of all 
 large families but of a sufficient number of them to raise the 
 mortality rate for the class. In the Boston investigation it 
 was found that of 341 mothers of whose character and intel- 
 ligence the investigators felt competent to express an opinion, 
 
 ♦PER CENT. OF INFANTS VISITED IN THE BOSTON INVESTIGATION, CLASSIFIED AC- 
 CORDING TO THE NUMBER OF THEIR MOTHERS' PREGNANCIES WHO LIVED IN 
 HOUSEHOLDS WHERE THE AVERAGE NUMBER OF PERSONS PER ROOM WAS LESS 
 THAN TWO, TWO, OR THREE OR MORE. 
 
 
 Less 
 than Two. 
 
 Two. 
 
 Three 
 or More. 
 
 TotaL 
 
 Pregnancies. 
 
 Per Cent. 
 
 Number. 
 
 1 
 
 84.0 
 59.5 
 34.2 
 18.7 
 17.7 
 
 14.0 
 35.1 
 44.9 
 61.0 
 61.7 
 
 1.4 
 5.4 
 20.9 
 20.3 
 20.6 
 
 100.0 
 100.0 
 100.0 
 100.0 
 100.0 
 
 349 
 
 2, 3 or 4 
 
 908 
 
 5 or 6 
 
 350 
 
 7 or 8 
 
 192 
 
 
 141 
 
 
 
 Total 
 
 52.3 
 
 37.6 
 
 10.1 
 
 100.0 
 
 1,940 
 
 tBertillon in the following figures has shown this relationship of poverty to the size of the family in 
 Paris (Nombre D'Enfants par Families, Journ, de la Soc. de Statisque de Paris, April, 1901, p. 134. Quoted 
 in Bailey's "Modern Social Conditions," p. Ill), in the following table showing the number of children 
 per 100 families in Paris, classified according to the economic resources of their parents (1896): 
 
 Very poor 156 
 
 Poor 14* 
 
 Comfortable • 131 
 
 Very comfortable ■ ■ 129 
 
 Rich 129 
 
 Very rich 127 
 
 Entire city 1*0 
 
44 American Statistical Association. 
 
 a somewhat larger proportion of the mothers of large families 
 were rated as unsatisfactory in these respects than of those 
 with small families.* The mothers of the larger families also 
 ranked lowest when graded according to their knowledge, or 
 rather their observance, of the laws of hygiene, this being 
 especially evident when they were graded according to their 
 standards of cleanliness and of ventilation.f 
 
 But all this leaves unanswered the question whether, other 
 things being equal, the number of children in the family into 
 which the infant is born has any direct postnatal influence upon 
 the mortality rate. Doctor Newsholme, an eminent English 
 authority, feels that it does not. He says, though he gives 
 no data to sustain his opinion, that "large families evidently 
 do not necessarily imply a tendency to high infant mortality. 
 The connection often observed between a high birth rate 
 and a high rate of infant mortality probably is due in great 
 part to the fact that large families are common among the poor- 
 est classes, and these classes are specially exposed to the 
 degrading influences producing excessive infant mortality." t 
 This view, however, seems very onesided. While it is un- 
 doubtedly true, as has already been shown, that "the de- 
 
 *These figures were as follows: 
 
 Of 341 mothers of whom an opinion was given 81 were unfavorable; classifying these according to the 
 number of the mother's pregnancies (including the one resulting in the birth of the infant under consid- 
 eration) they were found to include: 
 
 8 per cent, of those who had had 1 pregnancy, 
 
 29 per cent, of those who had had 2, 3, or 4 pregnancies, 
 
 30 per cent, of those who had had 5 or more pregnancies. 
 
 tThus, 781 out of 1,817 mothers from whom information was obtained said that they did not ventilate 
 their bedrooms at all at night. 
 
 These included: 
 
 40 per cent, of those who had had 1 pregnancy, 
 
 44 per cent, of those who had had 2, 3, or 4 pregnancies, 
 
 46 per cent, of those who had had 6, 6, 7, or 8 pregnancies, and 
 
 32 per cent, of those who had had 9 or more pregnancies. 
 
 Thus, with one exception the per cent, of mothers who said that they did not ventilate their bedrooma 
 at night increased with the number of the mother's pregnancies. This exception, which occurs with the 
 mothers who had 9 or more pregnancies, may be due to chance, as the group included only forty cases. 
 
 Dr. Herman Schwarz in the First Annual Report of the Free Out-door Maternity Clinic (New York, 
 1910), p. 41, gives similar figures verifying the statement that the mothers of the larger families rank lower 
 in general intelligence and in knowledge of hygiene than those of smaller families. 
 
 Thus, out of 679 mothers, 67 were graded as unsatisfactory in intelligence and 612 as satisfactory , the 
 average number of children bom per family being 3.8 among the former and 3.1 among the latter. 
 
 Of 670 mothers 491 were graded as having an unsatisfactory and 179 as having a satisfactory knowledge 
 of infant hygiene, the average number of living births per family being 3.9 among the former and 3.5 among 
 the latter. 
 
 tGreat Britain, Local Government Board, Supplement to the Thirty-ninth Annual Report, p. 49. 
 
Infant Mortality and the Size of the Family. 45 
 
 grading influences producing excessive infant mortality" 
 also tend to produce excessively large families — or else both 
 the degrading conditions and the excessively large family are 
 produced by the same deeper-lying causes— it is also just as 
 true that an excessive number of children in the family in it- 
 self brings about conditions in the home that lower the infant's 
 chances of survival. The term "large family" is, of course, 
 relative. In some cases, where the parents have sufficient 
 resources, interest, and leisure from other duties, what might 
 otherwise be regarded as an excessively large family would 
 under such circumstances be regarded only as normally large. 
 On the other hand, it is difficult to understand how, in cases 
 where the parents have not the resources, the interest, or the 
 ability to provide for more than three children properly, the 
 chances of survival of subsequent infants born into the family 
 will not be lessened, other things being equal, by the birth of 
 more than this number of children. 
 
 It is not difficult to understand how the popular miscon- 
 ception that the death rate in large families is low has arisen. 
 The large families we meet so impress us with the number of 
 children who are living that we forget the number who have 
 died while, again, we never notice the small families that would 
 have been large if so many of the children had not died during 
 infancy and childhood. The writer has been unable to find 
 any evidence whatever to support this popular belief that large 
 families have low infant mortality rates, while there is abun- 
 dance of evidence to show that the rate of infant mortality 
 increases with the size of the family, and no small amount 
 of evidence to show that the mortality rate for children and 
 even adults is greater in large families than in small.* 
 
 The Influence of the Length of the Interval Between the Mother's 
 Pregnancies. — As has already been intimated, the influence — 
 both prenatal and postnatal — of fertility and the size of the 
 family on infant mortality is partly determined bj^ the length 
 of the interval between the mother's pregnancies or deliveries. 
 
 *See especially the article by March in the Journal of the Royal Statistical Society, Vol. LXXV, pp. 
 519 ff., previously quoted, and Dr. R. J. Ewart's two articles in the Eugenics Review on " The Aristocracy 
 of Infancy and the Conditions of Birth," Vol. Ill, pp. 142-70, and "The Influence of Parental Age on Off- 
 spring," Vol. Ill, pp. 201-232. In the latter of these articles Doctor Ewart shows how the mean height of 
 children is also affected adversely by the order of birth (p. 213). 
 
46 
 
 American Statistical Association. 
 
 Where this interval is large the influence of large families is 
 less but where it is small it is much greater. The relation 
 between the average interval between the pregnancies of the 
 mothers visited in the Boston investigation and the rate of 
 infant mortality is shown in the following table : 
 
 TABLE V, 
 MORTALITY RATE PER 1,000 BIRTHS FOR INFANTS BORN TO MOTHERS VISITED IN 
 THE HOUSE-TO-HOUSE INVESTIGATION IN BOSTON, CLASSIFIED ACCORDING TO 
 THE AVERAGE INTERVAL BETWEEN PREGNANCIES. 
 
 Average Interval between Mother's Pregnancies. 
 
 Number of Births. 
 
 Infant Mortality Rate. 
 
 1 year or less. . . 
 1| years 
 
 2 years 
 
 3 years 
 
 4 years 
 
 5 years and over 
 
 Total 
 
 138.4 
 147.2 
 127.9 
 128.8 
 106.2 
 185.0 
 
 132.1 
 
 (a) Information was not obtained in 67 instances. The 371 infants who were bom during their mothers' 
 first pregnancies are not included in this table. 
 
 Table V may at first sight seem to show no relationship 
 between the average interval between the mother's pregnancies 
 and the rate of infant mortality. Thus, although the rate 
 decreases as the length of the interval increases until the period 
 of five years and over is reached (from 143 where the average 
 interval was less than two years to 128, 129 and 106 where it 
 was two, three, and four years respectively) it then increases 
 to 185 deaths per 1,000 births with the group of mothers the 
 average interval between whose pregnancies was five years 
 and over. Moreover, although the mortality rate for the first 
 two groups combined (one year and less and one and one half 
 years) is greater than that for the next three the rate for the 
 first group proves to be higher than that for the second when 
 they are separated (138 and 147 respectively). Yet, closer 
 examination of the results shows that these apparent excep- 
 tions are not of sufficient importance to influence the con- 
 clusion. 
 
 In the former case, the rise in the mortality rate with the 
 group where the interval between pregnancies was five years 
 or over can be accounted for in two ways; first, on the suppo- 
 sition that the number of cases included in the group is too 
 
Infant Mortality and the Size of the Family. 47 
 
 small (67) to allow any weight to be given it in drawing con- 
 clusions, or, second, on the supposition that such an excep- 
 tionally large interval between pregnancies as five years or 
 more is, among families of the class for the most part visited 
 in this investigation, generally the result, not of choice but of 
 weakness or physical incapacity — a condition which would be 
 likel}' to effect the strength and resistance of such children as 
 might be born and thereby to raise the mortality rate for the 
 class. 
 
 In the latter case, the fact that the rate for the group of 
 infants born to mothers with an interval between pregnancies 
 of one year or under is less than that for those where the in- 
 terval was one and one half years does not disprove the exist- 
 ence of a relationship between the infant mortality rate and 
 the length of the average interval between the mother's preg- 
 nancies because the rate for the first group where the average 
 interval was less than one year would have been much higher 
 if it were not composed so exclusively of mothers who had had 
 few previous pregnancies and whose families were, therefore, 
 small — a class which, as has already been shown, tends to have 
 a very low infant death rate. In fact 70 per cent, of the 
 mothers in this group had had only one previous pregnancy. 
 
 To fully appreciate the influence of this small interval 
 between pregnancies the mothers in the group must be further 
 classified according to the number of their previous pregnan- 
 cies. When this is done it will be seen that the mortality'- rate 
 for the infants born to mothers in the group who had had 1, 
 2, or 3 previous pregnancies was 129; for those who had had 
 4, 5, or 6 it was 333 ; while for those who had had over 6 pre- 
 vious pregnancies it reached the enormous proportion of 500 
 deaths per 1,000 births. A careful examination of the table, 
 therefore, indicates, when account is taken of the number of 
 the previous pregnancies and the size of the families of the 
 mothers included in each group, that there is a striking rela- 
 tionship between the rate of infant mortality and the aver- 
 age interval between the mother's pregnancies. 
 
 Doctor Ewart in a recent article in the Eugenics Review 
 emphasized this need for the "adequate spacing of births" by 
 showing that the physical development of the children who 
 
48 American Statistical Association. 
 
 survive is retarded by a short interval between births. Thus, 
 as he shows in the following table, the average height and 
 weight of over 800 children at the end of the sixth year of age 
 was greater where the interval between births was large than 
 where it was small ; 
 
 TABLE VI. 
 RELATIONSHIP BETWEEN THE LENGTH OF THE INTERVAL BETWEEN BIRTHS AND 
 THE MEAN HEIGHT AND WEIGHT AT THE END OF THE SIXTH YEAR OF AGE AMONG 
 866 CHILDREN OF MIDDLESBOROUGH, ENGLAND; 1911. (a) 
 
 Interval Between Births. 
 
 Mean Height in Inches. 
 
 Mean Weight in Pounds. 
 
 
 38.6 
 39.9 
 40.3 
 41.7 
 
 37.2 
 
 2 and under 2| j'ears 
 
 38.8 
 39.1 
 
 3 years and over 
 
 39.4 
 
 (a) R. J. Ewart, M.D., " The Influence of Parental Age on Offspring," Eugenics Review, Vol. Ill, p. 211 . 
 
 In commenting on this table Doctor Ewart says: "The 
 female is used to the greatest extent that her fertility will 
 allow; births at intervals of eleven months being quite com- 
 mon. . . . The birth interval is so short that the mother 
 is unable to bring her whole vitality to bear. Thus one child, 
 as it were, spoils the next." He then asks this very pertinent 
 question: "Which is the most desirable, three children of a 
 mean height of 39.5 inches or two of 41.0 inches, with all the 
 other attributes of mankind altered in the same proportion? 
 
 . . . As regards the individual there is no hesitancy as 
 to the answer; but from the point of view of economic produc- 
 tion it is quite possible that three inefhcients may be better 
 and do more work than the two efficients. Racial supremacy, 
 however, is not a question of numbers, and concerns individ- 
 ual fitness only." 
 
 It must not be forgotten, moreover, that the mothers 
 themselves do not escape without injury from the strain put 
 upon them by too frequent childbearing; but this lowering of 
 the vitality and strength of the mother as a result of an in- 
 sufficiency of time between the two pregnancies for complete 
 recovery from the strain of the first creates a condition which 
 will be likely, other things being equal, to harmfully affect 
 both the prenatal and postnatal development of subsequent 
 
Infant Mortality and the Size of the Family. 49 
 
 children. In fact, both the mother and the children suffer 
 when the length of the interval between births is too small. 
 It is manifestly impossible for most mothers to properly nour- 
 ish themselves, a new born baby, and a child within the 
 uterus at the same time. 
 
 This point is especially important, for artificial feeding is 
 more often resorted to by these mothers who are attempting 
 to rear two babies at once — one within the womb and the 
 other just born. Among the mothers visited in the Boston 
 enquiry, for instance, 25 per cent, said that they had resorted 
 to the use of bottle feeding because the quantity or the quality 
 of their breast milk was reduced by a subsequent pregnancy. 
 Among the Italian mothers this reason was given in over half 
 the cases — a truly astounding proportion. Thus, when it is 
 remembered that bottle feeding decreases the chances of sur- 
 vival of the baby from three to five times (as all authorities 
 agree),* it can readily be understood how the coming of one 
 child spoils the chances of survival of the previous one. 
 
 It is thus apparent that the influence of fertility and the 
 size of the family, especially when combined with the influence 
 of the length of the interval between pregnancies, constitutes 
 an important factor in infant mortality. As Doctor Ewart 
 says: "The wastage of life, and production of immature prog- 
 eny with its consequent misery and suffering to the mother, 
 can, to a much larger extent than is generally believed, be 
 traced directly to the unfortunate fact that the fertility of 
 women between their twenty-fifth and thirty-fifth years 
 exceeds their power to reproduce healthy offspring." t How- 
 ever dangerous "race suicide" and the declining birth rate 
 may be there can be little doubt that excessively large famihes 
 is no remedy, and however desirable a high rate of births may 
 be it is mere waste to bring children into the world faster than 
 the laws of nature decree to be desirable. 
 
 *See, for instance, the report of the investigation by the Children's Bureau of Infant Mortality; Johns- 
 town, Pa., pp. 38^, Davis' "Statistical Comparison of the Mortality of Breast-fed and Bottle-fed Babies" 
 in the Am. Journ. of Diseases of Children, March, 1913, pp. 234-47, and the U. S. Bureau of Labor's In- 
 vestigation of Infant Mortality in Fall River, Mass. 
 
 tPrev.cit.,p.215. 
 
IV. 
 THE MOTHER AND INFANT MORTALITY. 
 
 Reprinted from the Quarterly Publications of the American 
 
 Statistical Association, New Series, No. 113, 
 
 March, 1916, pp. 66-79. 
 
IV. 
 THE MOTHER AND INFANT MORTALITY. 
 
 Even the most hasty examination of almost any report on 
 vital statistics will show how widely rates of infant mortality 
 vary in different countries, states, cities, and other commu- 
 nities. This is a phenomenon of great significance to students 
 of infant mortality because it brings us face to face with the 
 question of causes. Why is the infant mortality rate lower in 
 one country than in another, in one city of the same country 
 than in another, in one ward or locality of the same city than 
 in another? Thus, the wide variation in infant mortality in 
 different areas and localities immediately challenges investi- 
 gation into the relation of housing and living conditions, of 
 sanitation and congestion, of rates of wages, and of social 
 conditions in general, to the proportion of the infants born to 
 those who die in such areas. In the same way any similar 
 variation found to exist, when the infant death rates of a large 
 number of families are compared, will suggest an inquiry into 
 the relationship existing between the proportion of infant 
 deaths in families and the age of the mother, the size of the 
 family, the character and intelligence of the parents, and 
 domestic conditions in general. It is with one phase of this 
 aspect of the problem that this article will deal — the relation 
 of the mother, her age, her character, her intelligence, and 
 her knowledge of infant hygiene, to infant mortality and to 
 her baby's health. 
 
 It is obvious that this variation in the proportion of infant 
 deaths in families cannot be shown by the use of the usual 
 infant mortality rate based on the proportion of deaths to 
 1,000 births because the number of both births and deaths in 
 any family is too small for such a comparison. The only 
 feasible method is the very indirect one of comparing the 
 number of infant deaths that have previously occurred in 
 families with the infant mortality rate for such families in 
 any given year. If the families in which a large proportion 
 of infant deaths have previously occurred contribute a larger 
 number of deaths to the total infant mortality of the year in 
 
 53 
 
54 
 
 American Statistical Association. 
 
 question than those f amihes in which no previous infant deaths 
 have occurred, it will follow that in the former group of families 
 infants tend to die in larger proportions than in the latter; 
 or, in other words, that the proportion of infant deaths to 
 births is year after year larger in some families than in others. 
 This was the method followed in an investigation of infant 
 mortaUty in Wards 6, 8, 13, and 17 of Boston by the Research 
 Department of the Boston School for Social Workers.* The 
 results of this phase of the inquiry are set forth in the following 
 table where the mortality rate for infants born in 1910 and 
 classified according to the number of infant deaths that oc- 
 curred among the children previously born to their mothers 
 is given — the number of previous children born into the family 
 being taken into account by an additional classification of the 
 mothers according to the number of their previous pregnancies: 
 
 TABLE SHOWING (A) NUMBER OF BIRTHS AND (B) DEATHS PER 1,000 BIRTHS IN 
 1910 IN FAMILIES VISITED IN BOSTON CLASSIFIED ACCORDING TO THE NUM- 
 BER OF DEATHS DURING INFANCY WHICH OCCURRED AMONG CHILDREN BORN 
 IN PREVIOUS YEARS AND THE NUMBER OF THE PREVIOUS PREGNANCIES OF 
 THE MOTHERS, (a) 
 
 Families Classified According to the Number of Deaths During 
 Infancy Which Occurred Among Children [Bom Previous 
 to 1910. 
 
 Families Classified According to the 
 Number of the Mother's Pregnancies 
 Occurring Previous to 1910. 
 
 Total. 
 
 1.2, 
 or 3 
 
 4,5, 
 or 6 
 
 Over 
 
 (A) NUMBKR OF BiBTHS IN 1910 IN: 
 
 All families (b) ._ 
 
 Families with no previous infant deaths 
 
 Families with one or more previous infant deaths 
 
 Families with — 
 
 One previous infant death 
 
 Two previous infant deaths 
 
 Three or more previous infant deaths 
 
 (B) Deaths pee 1,000 Births in 1910 in: 
 
 All families 
 
 Families with no previous infant deaths 
 
 Families with one or more previous infant deaths 
 
 Families with — 
 
 One previous iufant death 
 
 Two previous infant deaths 
 
 Three or more previous infant deaths 
 
 1671 
 
 961 
 
 487 
 
 1192 
 
 807 
 
 297 
 
 479 
 
 154 
 
 106 
 
 331 
 
 134 
 
 118 
 
 115 
 
 19 
 
 62 
 
 33 
 
 1 
 
 10 
 
 134 
 
 126 
 
 126 
 
 117 
 
 114 
 
 125 
 
 173 
 
 188 
 
 179 
 
 154 
 
 194 
 
 153 
 
 191 
 
 158 
 
 210 
 
 303 
 
 
 300 
 
 223 
 88 
 135 
 
 79 
 34 
 22 
 
 146 
 125 
 148 
 
 177 
 318 
 
 (a) This table is intended to show the relationship between the infant death rate in 1910 in the 
 families visited and the infant death rate during previous years in the same families. Since it is not 
 possible to compile a true infant mortality rate based on the ratio of deaths to 1,000 bbths for so small 
 a unit as the family this double classification, according to the number of previous infant deaths and 
 the niunber of the mother's previous pregnancies — thb last being a rough approximation of the number 
 of births — ^has been substituted. 
 
 (b) Information was not obtained in 21 instances. The 371 infants who were bom during their mother's 
 first pregnancy are omitted from this table. 
 
 * The writer is indebted to Dr. J. R. Brackett, director of the School for Social Workers, for the use of 
 these figures which have not been previously published. The investigation was made in the academic 
 years 1910-11 and 1911-12. During the second of these years the field work was carried on under the 
 diTPPt.ion of the writer sunervised by the director of the Research Department, Dr. T. W. Glocker. 
 
The Mother and Infant Mortalitij. 55 
 
 An examination of this table shows that the rate of mortaUty 
 for infants born to mothers, none of whose previous children 
 have died during infancy, is strikingly less than the rate for 
 those whose mothers have already had one or more infant 
 deaths. Moreover, the rate rises as the number of previous 
 infant deaths in the family increases and varies in the same 
 manner when the infants are also classified according to the 
 number of their mothers previous pregnancies.* The con- 
 clusion that infant mortality rates for families vary as widely 
 as those for states and cities appears, therefore, to be justified. 
 The table also shows that the rate of infant mortality in 
 any community is determined, not by the relatively uniform 
 occurrence of deaths in all families, but by their relatively fre- 
 quent occurrence in certain especially unfortunate famiUes, and 
 that the high mortality rate for the period of infancy is to be 
 laid at the door of these relatively few families where the pro- 
 portion of infant deaths to children born is especially large. 
 
 This, however, is merely another way of saying that the 
 proportion of infant deaths to births in families varies in 
 the same way that it does in other divisions of the popula- 
 tion. Having shown this — and it would probably have been 
 admitted a priori by many — it will be necessary to study at 
 length the nature of some of the domestic and social condi- 
 tions which cause the regular occurrence of a high ratio of 
 infant deaths to births in certain families, while others escape 
 with a smaller proportion of deaths or none at all; and to 
 ascertain, if possible, why in any district parents may be found 
 who have successfully reared every member of a large family, 
 while side by side with them are other families in which numer- 
 ous infant deaths have occurred. Among these "domestic 
 factors" of infant mortaUty one of the most important is the 
 mother, her age, her health, her character and intelligence, 
 and her devotion to the child. 
 
 The Influence of the Age of the Mother at the Birth of the 
 Child. The relation of the age of the mother at the birth of 
 her child to the rate of infant mortality has been frequently 
 commented upon, though not always made the subject of 
 careful study. It is often asserted on the one hand that 
 young mothers, because their strength has not been depleted 
 
 ♦ The one exception in the case of the mothers with one, two, or three previous pregnancies and two 
 previous infant deaths is probably due to chance as there were only nineteen births and three deaths in- 
 cluded in the group. 
 
56 American Statistical Association. 
 
 by previous childbearing or by the cares of a family, tend, 
 other things being equal, to give birth to healthy children and 
 that the rate of mortality among their infants is, therefore, 
 very low. On the other hand, it is probably as often claimed 
 that young mothers are too immature to give birth to healthy 
 and well formed children and, in addition, often incapable 
 through the lack of previous experience of giving them the 
 needed care after birth, as a consequence of which the mor- 
 tality among their infants is relatively high. The following 
 table summarizes the data bearing on this aspect of the subject 
 which have been collected in five recent investigations: 
 
 MORTALITY RATE PER 1,000 BIRTHS AMONG INFANTS INCLUDED IN FIVE INVESTIGA- 
 TIONS, CLASSIFIED ACCORDING TO THE AGE OF THEIR MOTHERS. 
 
 Age of Number Infant 
 
 the Mother. of Births. Mortality Rate. 
 Boston (Mass.) Intestiqation. (a) 
 
 All ages 2,025 125 
 
 Under 21 years .' 145 90 
 
 21 to 25 years 559 109 
 
 26 to 30 years 573 131 
 
 31 to 35 years 440 132 
 
 36 to 40 years 241 149 
 
 Over 40 years 67 164 
 
 Fau. Rivee (Mass.) Investigation, (b) 
 
 All ages 746 202 
 
 Under 20 years 29 103 
 
 20 to 29 years 386 189 
 
 30 to 39 years 257 206 
 
 40 and over 36 222 
 
 Unknown 38 — 
 
 Johnstown (Pa.) Investigation, (c) 
 
 Allages 1,463 134 
 
 Under 20 years 95 137 
 
 20 to 24 years 454 121 
 
 25 to 29 years 391 143 
 
 30 to 39 years 449 136 
 
 40 years and over 74 149 
 
 Ewart's Investigation (England), (d) 
 
 Under 19 years 152 171 
 
 20 to 24 years 536 132 
 
 25 to 29 years 396 166 
 
 30 to 34 years 316 170 
 
 35 to 39 years 150 220 
 
 Over 40 years ! 36 330 
 
 Birmingham (Eng.) Investigation, (e) 
 
 All ages 3,773 176 
 
 Under 25 years 936 207 
 
 25 to 35 years 1,982 167 
 
 35 years and over 855 163 
 
 (a) No information was obtained in 38 instances. For source of data see first note. 
 
 (b) Dublin: Infant Mortality in Fall River, Mass. — ^A Survey of the Mortality among 833 Infants Born 
 in June, July, and August, 1913. Quarterly Publications of the American Statistical Association, New 
 series, No. 110, June, 1915, p. 515. 
 
 (c) U. S. Children's Bureau: Infant Mortality: Johnstown, Pa., Washington, 1915, p. 35. 
 
 (d) R. J. Ewart, M. D.: The Aristocracy of Infancy and the Conditions of Its Birth. The Evgenict 
 Review, Vol. Ill, p. 166. The writer does not state where his data were collected but from a later article 
 in the same journal it is to be inferred that they were collected in Middlesborough, England. 
 
 (e) Health Department of the City of Birmingham, England: Report on Industrial Employment of 
 Married Women and Infant MortaUty, 1911 (p. 11) and 1912 (p. 11). 
 
The Mother and Infant Mortality 57 
 
 Interpretation of the data presented in this table is some- 
 what difficult because of the complex factors involved and 
 because the results of the five investigations differ considerably. . 
 In the first two, Boston and Fall River, the rate of mortality 
 was lowest for the infants born to the very young mothers 
 under 20 and 21 years of age. In both also the rate increased 
 regularly with the age of the mother after the twentieth year, 
 the highest rate occurring among the infants whose mothers 
 were over 40. In two others, Johnstown and Ewart's, the 
 rate was higher for the children of the very young mothers 
 under 19 or 20 than for those whose mothers were between 
 the ages of 20-30 and 30-35 years. These two investigations 
 agree with the first two, however, in showing that the rate is 
 highest of all among the infants whose mothers were 40 years 
 of age or over. Finally, when the figures for Bu-mingham are 
 examined, exactly the opposite tendencies are revealed, the 
 rate for the older mothers beuag lower than that for the 
 younger. 
 
 There seems to be a general agreement in the results of the 
 first four of these investigations in showing that the chances 
 of survival are greater for infants born to mothers between 
 the ages of 20-25, and 35-40 than for those born to mothers 
 over 40. The results for Bu-mingham do not share in this 
 agreement but in interpreting these figures the fact must be 
 considered that no separate data were given for the very young 
 mothers under 20 or the very old mothers over 40 years of 
 age. The results of the five investigations are not in sufficient 
 accord to justify any conclusion as to the rate of mortality 
 among the infants of very young mothers. This question 
 must, therefore, be left in doubt— at least as far as the avail- 
 able statistics are concerned. 
 
 On the other hand, in drawing conclusions from these 
 figures the mere fact that the results of the five investigations 
 differ is in itseK significant. It indicates, in the writer's 
 opmion, that the age of the mother is not a factor in infant 
 mortality of primary importance. The figures do not show 
 that mothers can obtain the knowledge necessary for bearing 
 and rearing babies only from experience acquired with age. 
 In fact, except for the very young mothers, these figures seem 
 to show that, as far as infants are concerned, one may fairly 
 
58 American Statistical Association. 
 
 expect as high a degree of efficiency in motherhood from the 
 young and inexperienced mothers as from the older — a con- 
 clusion which from the standpoint of prevention it is not 
 necessary to point out. 
 
 The Influence of the Character and Intelligence of the Mother. 
 But if the age of the mother is not a factor in infant mortality 
 of primary importance, her health, her character, and her 
 intelligence certainly are. It is difficult for the careful student 
 of this problem to avoid the conclusion that the real under- 
 lying factor of infant mortahty and the chief consideration in 
 the health and welfare of babies is the strength, character, 
 health, and intelligence of the mother. For instance, it is 
 the mother's health and strength that determines whether it 
 is physically possible for the baby to start life with the tre- 
 mendous advantage that comes from breast feeding and it is 
 her character and intelligence that largely determine how the 
 child shall be fed when either method is possible. Again, in 
 the recent campaign that has arisen for the control of the 
 influence of prenatal conditions on infant mortality it is evi- 
 dent that the mother is the all-important consideration. Cer- 
 tainly, while the child is developing in the mother's womb her 
 health and strength, the care she takes of herself, and her 
 general standards of hygiene are among the most important 
 considerations in determining whether or not the baby after 
 its birth will get a fair start in life or die in a few weeks or 
 months from prenatal causes.* 
 
 But even after the physical connection between mother and 
 child is severed at birth the infant is little less dependent upon 
 its mother, for she determines to a very large extent the char- 
 acter of everything with which it comes into contact — food, 
 sunlight, cleanliness, and even the kind of air it breathes. In 
 both the Boston and Johnstown investigations a close relation- 
 ship between bad housing conditions and infant mortality 
 was found to exist, but it should not be forgotten in inter- 
 preting these figures that the closeness of this relationship is 
 also to a large extent determined by the mother, f From even 
 the briefest investigation it will be evident that even in the 
 
 * See the writer's article on " The Influence of Prenatal Conditions on Infant Mortality," Proceed- 
 ings of the Southern Sociological Congress, 1915. 
 
 t See the writer's article on " Infant Mortality and Urban Housing, and Living Conditions," Journal 
 ofSocdogic Medicine (American Academy of Medicine), October, 1915. 
 
The Mother and Infant Mortality. 59 
 
 poorest districts, where the housing and Hving conditions are 
 exceedingly bad, many instances in which parents have reared 
 large families without the occurrence of a single death are to 
 be found, while side by side with these in the same neighbor- 
 hood or even in the same house are to be found other families 
 in which numerous deaths of infants have occurred. Such 
 instances as these, where other differences between the families 
 concerned appear to be so small, seem logically to be the 
 result not of poverty, or the cleanly or sanitary condition of 
 the home, or even of the physical strength of the parents, but 
 of the amount of intelligent attention and care which the 
 mother bestows upon her baby. In such cases it is not only 
 the " resistance of the child " but the "resistance of the mother" 
 that determines the extent of the influence of the other factors 
 on the problem. 
 
 The influence of the character and intelligence of the mother 
 on the rate of infant mortality is exceedingly difficult to meas- 
 ure statistically. It is especially difficult to distinguish this 
 influence from that of the other closely related factors and 
 one can never feel certain that the desideratum, "other things 
 being equal," has been reasonably well attained. It is 
 not to be expected, therefore, that the relationship between 
 the character and intelligence of the mother and the rate of 
 infant mortality can in the nature of the case be so accurately 
 measured as to show the extent of the influence of the former 
 on the latter, although the application of the statistical method 
 does show the existence of such a relationship. 
 
 Thus, in the Boston investigation it was found that of the 
 341 mothers, upon whose general intelligence and standards of 
 child care the investigators felt competent to express an 
 opinion, the rate of mortality (99.6) was lower for the 254 
 infants of whose mothers a favorable opinion was given than 
 the rate (103.4) for the 87 infants whose mothers were 
 criticised unfavorably. These figures are of somewhat doubt- 
 ful value, however, as they are based on an opinion formed 
 upon only one visit and include only the somewhat extreme 
 cases upon each end of the scale, all the other mothers in 
 between being omitted, as the investigators did not feel com- 
 petent to express an opinion upon them. A similar attempt 
 
60 American Statistical Association. 
 
 was made by Dr. Schwarz to grade the mothers coming under 
 the observation of the New York Free Out-Door Maternity 
 CUnic according to their general intelhgence, which was more 
 successful, as it was based on a much larger number of cases. 
 Thus, among 2,326 infants whose mothers were rated as un- 
 satisfactory in general intelligence the rate of mortality was 
 126 deaths per 1,000 births in comparison with a rate of 100 
 for the 210 infants whose mothers were rated as satisfactory 
 in this respect.* These figures are also of doubtful value as 
 the standard of intelligence used must have been high to have 
 excluded such a large proportion of the mothers from the 
 satisfactory grade. 
 
 A number of attempts have also been made to show the 
 relationship of the intelhgence of the parents to infant mor- 
 tality in which literacy is used as a test of intelligence. Thus, 
 Dr. Schwarz, in the study referred to above, shows that the 
 rate of mortaUty among 1,29.7 infants, both of whose parents 
 were literate, was 111 deaths per 1,000 births, while among 
 702 others, only one of whose parents was literate, the rate 
 was 113, and among 458, where both were illiterate, it was 
 172. t In a similar manner it was shown in the Johnstown 
 investigation that the rate of mortahty was higher among 
 infants born to foreign-born mothers who could not read and 
 write any language or who could not speak English than among 
 those infants whose mothers could meet these tests. Thus, of 
 the 229 mothers who could not read and write, the infant mor- 
 tahty rate was 214, while among the 419 mothers who could 
 read and write it was only 148. Among the 401 infants, whose 
 mothers could not speak English, the mortality rate was 187 
 in comparison with a rate of 146 among those whose mothers 
 could speak English. { In so far, therefore, as literacy and 
 ability to speak English are indices of intelligence these figures 
 indicate, though probably not adequately, the heavy handicap 
 under which babies are placed because of a lack of intelli- 
 gence on the part of their mothers. 
 
 * First Annual Report of the Free Out-Door Maternity CUnic, Covering the First Nine Years of the 
 Clinic's Experience. New York, 1910; Chapter 2, Part 2, Report of the Pediatric Department, with a 
 Study of Early Infant Mortality, by Herman Schwarz, M. D., p. 41. 
 
 t Loc. cit., p. 39. 
 
 t U. S. Children's Bureau: Infant Mortality: Johnstown, Pa. Washington, 1915, p. 34. 
 
The Mother and Infant Mortality. 61 
 
 Other writers have shown the existence of this relationship 
 between hteracy and infant mortahty by correlating the per- 
 centage of illiteracy and the rate of infant mortality in a 
 number of cities. For instance, Phelps, in his recent study 
 of the relation of infant mortality to the employment of 
 mothers, concludes, after a careful examination of the avail- 
 able statistics for the cities and states of New England and 
 especially of the cities of Massachusetts, that a high female 
 illiteracy is "with fair uniformity coexistent with a high infant 
 mortality rate." * Newman also notes the relationship from 
 a study of English figures, although he does not find it to be 
 as close as the writer just cited. f 
 
 On the whole, however, the extent of the influence of the 
 character and intelligence of the mother upon infant mortality 
 is not capable of statistical demonstration or of exact measure- 
 ment by any method so far devised. Nevertheless observa- 
 tion and experience show that it is an all-important factor in 
 the problem — the underlying factor which determines to a great 
 extent the influence of other measurable factors. In conclusion, 
 therefore, it may be said of this article, as Dr. Newman says 
 of his own excellent work on the subject: 
 
 This book will have been written in vain if it does not lay 
 the emphasis of this problem upon the vital importance to the 
 nation of its motherhood. Wherever we turn, and to what- 
 ever issue, in this question of infant mortality, we are faced 
 with one all pervading need— the need of a higher standard 
 of . . . motherhood. Infant mortality in the early weeks 
 of life is evidently due in large measure to the physical condi- 
 tion of the mother, leading to prematurity and debility of 
 the infant; and in the later months of the first year infant 
 mortality appears to be due to unsatisfactory feeding of the 
 infant. But from either point of view it becomes clear that 
 the problem of infant mortahty is not one of sanitation alone, 
 or housing, or indeed of poverty as such, but is mainly a 
 question of motherhood. No doubt external conditions as those 
 named are influencing maternity, but they are, in the main, 
 affecting the mother, and not the child. Improved sanitation, 
 better housing, cheap and good food, domestic education, a 
 
 * Edward B. Phelps: Infant MortaUty and Its Relation to Woman's Employment— A Study of Massa- 
 chusetts Statistics. In Volume XIII, Part 1, of the Bureau of Labor's Report on Condition of Woman 
 and Child Wage Earners in the United States, p. 48. 
 
 t George Newman, M. D.: Infant Mortality. London, 1906, p. 222. 
 
62 American Statistical Association. 
 
 healthy life of body and mind . . . exert but an indirect 
 effect on the child itself, who depends for its life in the first 
 twelve months, not upon the state or the municipality, nor 
 yet upon this or that system of creche or milk-feeding, but upon 
 the health, the intelligence, the devotion and maternal instinct 
 of the mother. And if we would solve this great problem of 
 infant mortality, it would appear that we must first obtain a 
 higher standard of . . . motherhood. Without a mo- 
 ment's hesitation I place this need as the first requirement. 
 Other things . . . are important, but this is the chief 
 thing.* 
 
 To this conclusion the present writer heartily subscribes. 
 The most important factors in infant mortality are the 
 strength, the health, the character, and intelligence of the 
 mother. 
 
 Preventive Methods: Education for Parenthood. Space would 
 not permit, even if such a thing were possible, an enum- 
 eration of all the agencies and influences in modern life 
 which are helping to raise the standard of motherhood or of 
 all those other agencies and influences which are helping to 
 lower it. It will be desirable, however, to conclude this dis- 
 cussion of the relation of the mother to infant mortahty with 
 at least a brief survey of some of the chief methods whereby 
 the standard of motherhood, and also of fatherhood, may be 
 raised. Probably the most important of these methods is 
 education — the instruction and training of mothers and poten- 
 tial mothers in the care and rearing of children. 
 
 In a recent pamphlet, issued by the Children's Bureau on 
 Baby-saving Campaigns, the methods employed by certain 
 American cities in the prevention of infant mortality were 
 pointed out — the work of health departments, registration of 
 births and deaths, provision of a pure milk supply, milk 
 stations and baby clinics, encouragement of maternal feeding, 
 ''little mother leagues" and classes, visiting nurses, prenatal 
 work with expectant mothers, improvement of housing and 
 living conditions, fight against flies, garbage accumulation, 
 and dust, fresh air camps and hospitals, educational work 
 through the distribution of circulars, pamphlets, etc.f A 
 
 * Ibid, pp. 257-258. 
 
 t U. S. Children's Bureau: Baby Saving Campaigns. Washington, 1911. See table of contents. 
 
The Mother and Infant Mortality. 63 
 
 careful examination of each of these methods of preventing 
 infant mortality will at once reveal to what a large extent 
 each depends for its effectiveness upon the mother. At first 
 sight it would appear that the chief function of a milk station 
 is to provide cheap and pure milk and yet, as the report just 
 referred to points out, "it has been the experience of prac- 
 tically all milk dispensaries that it is useless to send pure, 
 clean milk into a dirty home to be handled by an ignorant, 
 dirty mother or older child. It is necessary to reach the 
 mothers, and not only teach them how to care for the baby's 
 milk, but to convince them of the necessity of cleanliness. 
 . . . In manj'- cities it is believed that the principal good 
 to be derived from milk stations consists in the opportunity 
 given for those in charge to come in contact with the mothers 
 and with the home surroundings. . . . Many cities place 
 practically all the emphasis upon visiting nurses and the 
 instruction of the mothers in the homes." * Dr. Josephine 
 Baker, director of child hygiene of the Department of Health 
 of the City of New York, also expressed the same idea in an 
 address before the International Congress of Hygiene and 
 Demography at Washington in 1912: "Without overlooking 
 the value of pure milk, I believe this problem must primarily 
 be solved by educational measures. In other words, the 
 solution of the problem of infant mortality is 20 per cent, 
 pure milk and 80 per cent, training of the mothers. The 
 infants' milk stations will serve their wider usefulness when 
 they become educational centers for prenatal instruction and 
 the encouragement of breast feeding and teaching of better 
 hygiene." f 
 
 In the case of visiting nursing, it is, of course, not necessary 
 to point out that it is the function of such nurses not to care 
 for the child themselves but to instruct the mother in the 
 proper care of children. 
 
 Recently the work of visiting nurses has been extended to 
 include the instruction of expectant mothers during the period 
 of pregnancy. This "prenatal care," as it is often called, is 
 
 * Prev. cit., pp. 22-23, and 32. 
 t Transactions, Vol. 3, p. 149. 
 
64 American Statistical Association. 
 
 one of the newest and one of the most important direct means 
 of reducing infant mortaUty by the education of the mothers, 
 for the instruction is begun not after the child is born but soon 
 after it has been conceived. 
 
 Finally, among the methods of instructing mothers in the 
 proper care of children after the birth of the child or after 
 conception, should be mentioned the educational work that 
 is being carried on through the distribution of circulars, 
 pamphlets, etc., on the care of infants and children as well as 
 upon general hygiene and health. The volume of such publi- 
 cations, as well as the character of the material and the method 
 of presentation and distribution, is improving rapidly. State 
 and city boards of health and many private societies and 
 associations have issued many exceedingly valuable pamphlets 
 written in such a way as to appeal effectively to the. mothers 
 interested. Some of these have been printed in foreign lan- 
 guages as well as in English. The Federal Public Health 
 Service and the Children's Bureau have also issued similar 
 publications. The latter bureau is issuing a Care of Children 
 Series, two numbers of which have already appeared. The 
 first, " Prenatal Care," deals with the care of mothers during 
 pregnancy, while the second, " Infant Care," gives the mother 
 specific instruction as to the care of the baby after birth. 
 Both of these pamphlets are written in popular, non-medical 
 terms and are admirably suited to their purposes. 
 
 The trouble with the publications of the Children's Bureau, 
 as well as those of the Public Health Service, is that the amount 
 of funds available has heretofore been so small that they have 
 not been as widely distributed as they should be. If these 
 bulletins on the care of children, for instance, were as widely 
 distributed as the Department of Agriculture's very successful 
 series of Farmers' Bulletins, we should feel more confident 
 that the government is doing its part in the prevention of 
 infant mortality than is possible at present. On the other 
 hand, the work of many city health departments has not been 
 so limited by a lack of funds. In a number of cities educa- 
 tional pamphlets and folders are sent to each mother upon 
 the registration of the birth of the child, either through the 
 
The Mother and Injajit Mortality. 65 
 
 mails or by the hands of a visiting or health nurse. Effective 
 use has also been made of newspapers for the instruction of 
 mothers in a number of cases.* 
 
 But this instruction of mothers in the care of themselves 
 during pregnancy and of their babies after their birth is not 
 the only means of reducing infant mortality by means of 
 education. This kind of instruction may be called instruc- 
 tion or education of parents; but there is another kind of 
 education that begins at childhood and which may be called 
 education J or parenthood. To say that "our educational 
 system needs revision and that it should be made to conform 
 more nearly to the actual requirements of our complex daily 
 life" is merely to repeat a statement that has been made many 
 times already; and yet there can be little doubt that one of 
 the most effective means of lowering the rate of infant and 
 child mortality, as well as the general death rate for all ages, 
 would be the education of children and young men and women 
 for parenthood. Just how much of this education should be 
 given in the home and how much in the schools, and just how 
 much should be obtained by the study of the laws of health 
 and hygiene or by the study of domestic science and home 
 management, it is impossible to say at present. What can 
 be said at present is that the need of education for parenthood 
 cannot be denied and that, in the overhauling and revision of 
 our system of education which is going on at present, efficient 
 methods for meeting this need must be worked out.t 
 
 But education for parenthood should include more than 
 instruction in the science and art of infant and child care, or 
 in home management or even in the ideals of parenthood; it 
 should also include instruction, direct or indirect, in "the 
 selection of parents." To take the case of the girl, for instance, 
 education for parenthood, 
 
 * U. S. Children's Bureau: Baby-saving Campaigns, pp. 44-45 and Appendix. 
 
 t One of the most interesting of the recent attempts to instruct and educate children directly in the care 
 of babies and in the laws and rules of infant hygiene has been carried on by the Department of Health of 
 the City of New York in its lectures and classes for the " httle mothers " who attend the city schools. little 
 Mothers' Leagues have also been organized which have reached a membership of over 17,000. These 
 leagues were described by Dr. Baker in her address before the International Congress of Hygiene and 
 Demography previously referred to. The idea has also been put in operation in other cities. Another 
 movement of the same kind that should at least be mentioned is the establishment of continuation schools 
 for homemaking. 
 
66-74 American Statistical Association. 
 
 instead of concerning itself with the care of her baby, will 
 be at work when she is choosing the baby's father. In all 
 times and places, woman's primal and supreme function is or 
 should be that of choosing the father of the future. This great 
 idea should be recognized, implicitly and explicitly, in the ed- 
 ucation of every girl; she is or may be partly responsible for 
 the future of mankind. She herself, mind and body, is holy, 
 for she is the temple of the life of this world to come. She 
 must honor and care for herself accordingly; and this twofold 
 aspect of her present and future duty, in caring for herself and 
 in choosing her co-creator of the future, must be instilled into 
 her mind with the solemnity, the sanctity, and the authorita- 
 tive sanction of a religious dogma. * 
 
 This quotation, in suggesting how the standard of parent- 
 hood may be raised by eugenics, both on its positive side of 
 the encouragement of the marriage of the j&t and on its nega- 
 tive side of the discouragement or prevention of the marriage 
 of the unfit, the unhealthy, and the weak, opens up an inter- 
 esting field for discussion but one which, even if space would 
 permit, it would hardly be proper for us to enter upon here. 
 
 * Saleeby, C. W.: The Methods of Race Regeneration. New York, 1911, p. 36. 
 
V. 
 
 INFANT MORTALITY AND URBAN, HOUSING, AND 
 LIVING CONDITIONS 
 
 Reprinted from the Journal of Sociologic Medicine (American 
 
 Academy of Medicine), October, 1915, 
 
 pp. 306-329. 
 
V. 
 
 INFANT MORTALITY AND URBAN. HOUSING, AND 
 LIVING CONDITIONS. 
 Even the most hasty examination of almost any report on vital 
 statistics will show how widely infant mortality rates vary 
 among different countries, states, cities, and other communities. 
 A recent report of the Registrar-general for England shows that 
 the mortality rates for infants varied from 204 deaths per 1,000 
 births in Hungary for the five-year period from 1906 to 19 10 to 
 about 70 in New Zealand and from 313 in Moscow to 85 in Sydney 
 and 92 in Stockholm. In a similar way the annual reports of the 
 Btureau of the Census on mortality statistics of the registration area 
 of the United States show how year after year certain states and 
 cities have higher infant death rates than others. In these cities 
 also the rates are not equal everywhere, some of the wards and 
 blocks having rates much higher than the average for the city as 
 a whole while others have rates much lower. Such wide and 
 constant variations are not due to chance and the many attempts 
 that are now being made to discover their causes and to explain 
 why certain types of communities, cities, states, and countries 
 have a high infant mortality rate while others, often not far 
 distant, have a moderate or even a very low rate furnish one of 
 the most interesting and helpful fields of discussion in vital sta- 
 tistics. All the causes of infant mortality are concerned in this 
 question but in accounting for strictly community variations, 
 community and home conditions are especially important. It is 
 with some of these conditions that this article will deal. 
 
 THE INFI.UENCE OF URBAN CONDITIONS OF LIFE ON INFANT MOR- 
 TALITY. 
 
 One of the most striking and constant differences in mor- 
 tality rates existing in the whole field of vital statistics is to be 
 found when urban and rural districts are compared. The differ- 
 ence appears whenever the deaths of persons of all ages in the two 
 areas are compared; but it is especially great in the case of in- 
 
 77 
 
78 
 
 fants and young children, those classes of the population most 
 affected by the influence of adverse environmental conditions. 
 Thus,' Bulletin 109 of the Bureau of the Census shows that in 
 1910 the death rate per 100,000 population for infants was 36 
 per cent, higher in the cities (326) than in the rural parts of the 
 registration area (240). The per cent, of excess of the death rate 
 for children under five years of age in the cities (460) was 38 per 
 cent, higher than in the country (333) while the per cent, of excess 
 of the rate for persons of all ages was only 19 per cent, higher in 
 the cities (1,590) than in the country (1,340). Confining our- 
 selves strictly to the problem of infant mortality it will be neces- 
 sary to examine certain aspects of this excessiv urban mortality 
 in detail, with a view particularly to ascertaining to what extent 
 the excess of deaths in the cities is due to preventable conditions. 
 The following table shows the death rate in 191 1 per 1,000 popula- 
 tion under one year of age for infants classified by age at death 
 in the cities and rural parts of the registration states with the per 
 cent, of excess of the rate for the cities over that for the rural dis- 
 tricts.^ 
 
 Age at death. Cities. 
 
 Under i year 135 ■ 9 
 
 Under i week 34-7 
 
 I week to i month 18.8 
 
 I to 3 months 24.6 
 
 3 to 5 months 24.6 
 
 6 to 8 months 18.6 
 
 9 to 12 months 14.4 
 
 An examination of this table shows that the infant death rate 
 per 1,000 population under one year of age is higher in the cities 
 than in the rural parts of the registration states at each age period 
 but that the per cent, of excess of the rate for the cities increases 
 with the age at death, being least during the first week and month 
 and greatest during the last six months of infancy. Thus, the 
 infant death rate for the cities in the age period from six to twelve 
 months exceeded that for the country districts by about 70 
 
 1 Twelfth Annual Report of the Bureau of the Census on Mortality Statistics for the 
 year 1911, pp. 537-8. Population in 1911 estimated by the Bureau of the Census. 
 
 Rural 
 
 Per cent. 
 
 districts. 
 
 of 
 
 excess. 
 
 94.1 
 
 
 44 
 
 28 
 
 7 
 
 
 21 
 
 14 
 
 9 
 
 
 26 
 
 15 
 
 9 
 
 
 55 
 
 15 
 
 5 
 
 
 59 
 
 10 
 
 8 
 
 
 72 
 
 8 
 
 4 
 
 
 71 
 
79 
 
 per cent., while for the period from three to five months the per 
 cent, of excess was 59, for the period from one to three months 
 55, and for the periods from one week to one month and under 
 one week only 26 and 21 per cent., respectively. The greatest 
 excess of the urban over the rural death rate thus occurs during 
 the later and not the earlier months of infancy. Before pointing 
 out, however, that the excessiv death rate is due to preventable 
 conditions, the causes of the deaths of infants in the two areas 
 should be compared. This is done in the following table which 
 shows the infant death rate in 191 1 per 1,000 population under 
 I year of age from certain important causes for the cities and 
 the rural districts of the registration states, with the per cent. 
 of excess in the rate for the cities over that for the rural districts:^ 
 
 Rural Per cent. 
 Cause of death. Cities. districts, of excess. 
 
 All causes 135-9 94-1 44 
 
 Diarrhea and enteritis 371-2 191 .6 71 
 
 Premature birth 210.8 157-7 34 
 
 Congenital debility 168.5 119 -2 41 
 
 Injuries at birth 45 - 1 26.4 71 
 
 Congenital malformations 66 . 6 57-6 16 
 
 Bronchopneumonia iii-S 52.4 113 
 
 Pneumonia 59-3 51 -o 16 
 
 Acute bronchitis 37-0 21.5 72 
 
 Convulsions 30.8 26.4 17 
 
 Meningitis 18.3 16.7 10 
 
 Tuberculosis — all forms 25.1 16.6 5 
 
 Measles, diphtheria and croup, whooping 
 
 cough, and scarlet fever 5° - o 45-8 9 
 
 Syphilis 22.1 4.6 38 
 
 All other causes 142.8 i54-0 8^ 
 
 From every cause of death of sufficient importance to warrant 
 its separate inclusion in the table the death rate per i ,000 popula- 
 tion under one year of age was greater in the cities than in the rural 
 
 1 Prev. cit., pp. 537-38. 
 
 ' Per cent, of excess of rate for rural districts over that for cities. 
 
8o 
 
 districts. The difference in the rates from each cause varied con- 
 siderably. The per cent, of excess of the death rate for the cities 
 over that for the rural districts from broncho pneumonia, 
 diarrhea and enteritis, acute bronchitis, and injuries at birth 
 was greater than that for all causes combined. The per cent. 
 of excess from congenital debility, premature birth, congenital 
 malformations, pneumonia, convulsions, meningitis, the five 
 infectious diseases, measles, diphtheria, croup, whooping cough, and 
 scarlet fever, syphilis, and all forms of tuberculosis was less than 
 that for all causes combined. Grouping these causes of death 
 in accordance with the international classification it will be seen 
 that in general the higher infant death rate in the cities than in 
 the rural parts of the registration states is the residt of a higher 
 death rate from all important causes. The diseases of the di- 
 gestiv and respiratory systems (pneumonia excepted) play a 
 larger part, however, in making up the excessiv urban death 
 rate than do the diseases of early infancy (injuries at birth ex- 
 cepted), congenital malformations and the diseases of the ner- 
 vous system.^ This fact that the diseases of the digestiv and 
 respiratory systems have the predominant influence in causing 
 the higher infant death rate in urban than rural districts has 
 frequently led to the statement that the excessiv rate for cities 
 is due to "bad air" and "bad feeding" or "impure milk" diseases. 
 Such a statement does not represent the entire truth, however, 
 for, as this table shows, the infant death rate from all important 
 causes is higher in the cities than in the country. Rather should 
 it be said that "bad air" and "impure milk" diseases play the 
 most important part in bringing about the higher rate of mor- 
 tality in urban than rural districts. 
 
 It has now been shown that, altho the death rate is greater 
 in urban than in rural districts in each age period of infancy, the 
 
 I Of the two exceptions to this generalization, injuries at birth and pneumonia, the 
 former may be due to differences in certifying and diagnosing causes of deaths in urban 
 and rural districts — altho this is by no means certain. In any case, the per cent, of infant 
 deaths in the registration states from injuries at birth was so small (3 per cent.) in com- 
 parison with the much greater proportion caused by the other diseases of early infancy 
 (16 per cent, from premature birth and 12 from congenital debility) that the tendency 
 ihown by the above classification is not vitiated thereby. The exception in the case of 
 pneumonia is more important, however, as the deaths from this cause constituted a rather 
 large proportion of the deaths from respiratory diseases and it should prevent the pushing 
 of the conclusion drawn from the classification too far. 
 
8i 
 
 per cent, of excess of the rate for the cities over that for the country- 
 districts increases in direct ratio with the age of the child, being 
 least during the first weeks and months of infancy and greatest 
 during the last. In a similar manner it has also been shown that, 
 altho the infant death rate from every important cause of death 
 is greater in turban districts, the per cent, of excess of the rate for 
 the cities over that for the country districts is much greater in the 
 case of the diseases of the digestiv and, to a lesser extent, of the 
 respiratory system than in the case of the diseases of early infancy 
 and congenital malformations. Both of these facts — that the 
 infant death rate in cities is especially high during ^the last six 
 months of infancy and that it is especially high from digestiv 
 and respiratory diseases— indicate that the higher mortality 
 rate among infants in cities than in rural districts is largely due 
 to the influence of "postnatal" conditions affecting the infant 
 after birth. Now, as the writer has shown in another article,^ 
 the deaths of infants resulting from the influence of postnatal 
 conditions are more easily prevented than those due to the in- 
 fluence of prenatal conditions. The conclusion, therefore, seems 
 well founded that the higher rate of infant mortality in urban 
 than rural districts is to a large extent the result of preventable 
 causes. 
 
 Further than this it is very difficult to go in the search for the 
 causes of the higher mortality rate among infants in urban than 
 in rural districts for it is well nigh impossible to put one's finger 
 on the specific factors which bring about the difference. Conges- 
 tion and overcrowding, impure air and bad sanitation certainly 
 play a large part; but they are not the only, or perhaps, even the 
 most important factors concerned. When urban and rural mor- 
 tality is compared there are many things to consider. In cities 
 there is probably a wider and more extensiv interest in the pre- 
 vention of infant deaths than in the country. Medical attendance, 
 hospitals, free dispensaries, and other artificial means of curing 
 disease and prolonging life are better developt as a rule in cities 
 than in the country, while in some cities the water supply* and 
 
 * Southern Sociological Congress: Proceedings of the annual meeting for 1915 — 
 Health Crusaders. Nashville, 1915. Space does not permit the writer to quote from this 
 article more in detail. 
 
sanitation are also better than in small villages and country places. 
 On the other hand, other conditions are present in large towns 
 and cities which counteract these advantages. The milk supply- 
 is usually not so good as in the country while the natural aids 
 to survival and the prevention of sickness are less eflfectiv. The 
 effect of bad sanitation, poverty, impiu^e air, a poor water supply 
 and the neglect of the elementary requirements of health is also 
 probably greater in the cities. Excessively large families also 
 probably exert a greater influence on infant mortality in cities 
 than in the less-crowded country communities while a smaller 
 number of children would constitute a "large family" in the cities 
 than in the country.^ The employment of women in extra- 
 domestic occupations is also more extensiv in the cities and large 
 towns and artificial feeding is probably more often employed there 
 also. Finally, domestic and home conditions in general are prob- 
 ably on the whole more favorable to infant life in the country 
 than in cities. As Dr. Newman says, "the homelessness of the 
 people is one of the worst features of town life, and is operating 
 injuriously on infancy. Of that I do not think there can be any 
 doubt in the mind of a careful observer of the life of the poor in 
 a great city. He may not be able to put his finger upon any one 
 item which is affecting the mother and killing the infant, for an 
 infant is a complex organism, and bears within itself a tempera- 
 ment, a physique, and a heredity composed of a vast array and 
 medley of influences inextricably interwoven. But he will be 
 able to say that the general conditions of domestic life in a city 
 tenement are such as to make the rearing of infants a difficult and 
 doubtful undertaking. "2 
 
 THE R^IvATlV decline; IN INFANT MORTAI.ITY IN URBAN AND 
 RURAL DISTRICTS. 
 
 On the other hand, conditions in cities have greatly improved 
 in the last generation as a result of the great advances in sanita- 
 tion and preventiv medicine and of the widespread efforts that 
 
 1 See the writer's article in a recent number of the Quarterly Publication of the Ameri- 
 can Statistical Association on Infant Mortality and the Size of the Family. (September, 
 1915.) 
 
 * George . Newman, M.D., Infant Mortality — A Social Problem. London, 1906, p. 
 180. 
 
83 
 
 have been made to prevent disease and lengthen Hfe; and on the 
 whole this improvement has probably been greater in the cities 
 than in the country districts. This would lead one to expect 
 that the markt decline^ which has occurred in the infant mor- 
 tality rate in recent years will be found upon comparison to have 
 been greater in the cities than in the country. Unfortunately 
 such a comparison cannot be made for most countries or for 
 any part of this country because figures extending over a suffi- 
 ciently long period to avoid the influence of superficial and tem- 
 porary conditions are lacking. The only countries for which such 
 figures are available are Germany and Prussia and these are given 
 in the following table which shows the decline in the infant mor- 
 tality rate per i ,000 births for the cities and the country districts 
 of Prussia between 1876-80 and 1901-03 and of Germany between 
 1886-90 and 1907-08: 
 
 Country 
 Years. Cities. districts. 
 
 Prussia. 2 
 
 1876-80 211 183 
 
 1881-85 211 186 
 
 1886-90 210 187 
 
 1891-95 203 187 
 
 1896-1900 195 185 
 
 1901-03 . 180 176 
 
 Germany.' 
 
 1886-90 210 187 
 
 1896-1900 195 185 
 
 1901-06 181 178 
 
 1907-08 156 164 
 
 An examination of this table shows that in both countries the 
 decline in infant mortality in the period considered was greater 
 in the cities than in the country districts. In Prussia the per cent, 
 of decrease in the rate for the cities was 14.7 per cent., over three 
 
 I See the Twelfth Annual Report on Mortality Statistics for the year 1911 (pp. 24-26) 
 by the U. S. Bureau of the Census. 
 
 » Zeitschrift des Koeniglich Preussischen Statistischen Landesamts, Bd. XV, s. xvii. 
 Infants of illegitimate birth not included in either case. 
 
 • Statistik des Deutschen Reichs. Quoted in Dr. Helen MacMurchy's Spec. Rep. 
 on Infant Mortality to the Leg. Assembly of Ontario, Toronto, 1911, p. 26. She does not 
 cite the source of this data. Years 1891-96 were not included. 
 
84 
 
 times as great as in the country (3.8), while in the German Em- 
 pire the decline in the cities (25.7 per cent.) was almost twice as 
 great as that in the country (12.3 per cent.). Moreover, in Prussia 
 the falling off in infant mortality was so much greater in the 
 cities than in the country that by the end of the period 1901-03 
 the rates had become practically equal, while in Germany the de- 
 cline in the cities was so much greater that at the end of the 
 period 1907-08 the rate for the cities had reacht a point somewhat 
 lower than that for the country — a very unusual condition to say 
 the least. 
 
 The lack of statistics prevents the continuance of this com- 
 parison for other European countries so that no general conclu- 
 sion as to the relativ decline of infant mortality in urban and rural 
 districts can be drawn for the continent as a whole. ^ An indirect 
 method of comparison is possible, however, from which a very 
 strong probability can be estalDlisht that tendencies in other 
 European countries in this respect are not different from those 
 in Prussia and the German Empire, i. e., that of comparing the 
 decline in the infant mortality rate for the large cities of each coun- 
 try with that for the country as a whole. Applying this method 
 of comparison to the countries and cities included in the section 
 on vital statistics of the annual reports of the Registrar-general 
 for England and Wales,^ the following table showing the percent- 
 ages of decrease in the infant mortality rate during the period 
 from 1881-85 to 1906-10 for the countries and their large cities 
 has been drawn up. 
 
 1 Somewhat similar figures can be compiled from the Sixty-fifth to the Seventy-third 
 Annual Reports of the Registrar-general for Births, Deaths, and Marriages in England 
 and Wales showing the decrease in the infant mortality rate for rural and urban counties 
 between 1897 and 1910. These figures are of no value in this connection, however, since 
 it is not urban and rural districts that are being considered but a group of "counties which 
 are mainly urban in character" with another group "in which the rural character greatly 
 predominates, altho the group contains some considerably urban communities." Since 
 1910 this method has been given up in favor of another which furnishes "a more satis- 
 factory basis of comparison," i. e., "one in which the representative urban area is entirely 
 urban and the representativ rural area entirely rural." See also annual report for 1906, 
 Vol. 69, pp. xxvii and Ixxi. 
 
 ' Seventy-third and Seventy-fourth Annual Reports of the Registrar-general for 
 Births, Deaths, and Marriages in England and Wales, pp. xciv and 105-15, respectively. 
 Rates for 1881-85 and 1906-10 are five-year averagfes. 
 
85 
 
 Australian Commonwealth 37.6 
 
 Sydney 50.9 
 
 Melbourne 44-4 
 
 Denmark 20.0 
 
 Copenhagen 36 . i 
 
 England and Wales 15.8 
 
 London 24.0 
 
 France 24.6 
 
 Paris 34 . 6 
 
 Hungary^ 18.4 
 
 Budapest^ 24.1 
 
 Ireland 0.0 
 
 Dublin 17.0 
 
 Belfast 4.7 
 
 Italy 17.3 
 
 Milan 17.3 
 
 The Netherlands 31 
 
 Amsterdam 55 
 
 Rotterdam 49 
 
 The Hague 53 
 
 Norway 29 
 
 Christiania 39 
 
 Prussia 18 
 
 Breslau 32 
 
 Berlin 41 
 
 Hamburgh 32 
 
 Dresden- 35 
 
 Munich- 42 
 
 Scotland 4 
 
 Glasgow 13 
 
 Edinburgh 6 
 
 Sweden 32 
 
 Stockholm 55 
 
 Thus, in every country of Europe and Australia for which sta- 
 tistics are available with one unimportant exception a greater 
 decrease occurred in the rate of infant mortality for the large 
 cities than for the country considered as a whole. The single 
 exception was in the case of Italy where the decline in the rate 
 for the country as a whole and for the one large city included was 
 the same. 
 
 Similar data are available for only one state in this countrj^ 
 Massachusetts, where the. decline in the rate for the state as a 
 whole can be compared with that for Boston, the chief city. 
 Thus, between 1881-85 and 1906-10 the infant mortality rate 
 for Boston dechned from 186 to 133, a decrease of 28.5 per cent., 
 while the rate for Massachusetts declined from 160 to 133, a 
 decrease of only 16.9 per cent. Starting at the beginning of the 
 period with a rate of 26 deaths for i,oco births higher than that 
 for Massachusetts the decline in Boston was so much greater 
 that by the end of the period both rates were equal.^ 
 
 1 Per cent, of decrease between 1891-95 and 1906-10. 
 
 * These three German cities, tho not in Prussia, are added for comparison. 
 
 » Figures for Massachusetts are from the 17th Annual Report of Births, Deaths, and 
 Marriages for the year 1911, p. 180, and for Boston from the 40th Annual Report of the 
 Department of Health for the year 191 1, p. 171. This same method of comparing the de- 
 cline in the infant death rate for the states included in the registration area in 1900 which 
 occurred between the census year 1900 and the calendar year 1911 with the decline in the 
 
86 
 
 Obviously these figures cannot be accepted as an entirely ade- 
 quate substitute for a direct comparison of the decline of infant 
 mortality in urban and rural districts such as was given for the 
 European statistics and Massachusetts. They do show clearly enuf 
 that in all countries for which statistics are available the infant 
 mortality rate has declined more during this period of thirty years 
 in the large cities than in the country districts and small towns. 
 The exclusion of the cities from the returns for the entire country, 
 which would be necessary in compiling the rate for the rural dis- 
 tricts, would make the contrast between the large cities and the 
 rest of the country all the more striking. It is not certain, on 
 the other hand, altho it is highly probable, that the decline in 
 the rate for the large cities is a fair index in this respect of ten- 
 dencies existing in all cities or urban districts as a whole. Yet, 
 taken in conjunction with the figures for Prussia and Germany 
 comparing the decline in the two areas directly, they do create 
 a very strong probability, if not the certainty, that the decline 
 
 cities of 100,000 inhabitants or over in each state can be made from the table given in the 
 Eleventh Annual Report of the Bureau of the Census on Mortality Statistics for 1911 
 (p. 25). The per cent, of decrease in the infant death rate per 1000 population under one 
 year of age for these states and cities during this period was as follows: 
 
 Michigan 8 Connecticut 17 
 
 Detroit 16 New Haven 13 
 
 Grand Rapids 27 Bridgeport 20 
 
 New Jersey 21 Massachusetts 19 
 
 Jersey City 27 Boston 17 
 
 Newark 33 Worcester 17 
 
 Paterson 38 Fall River. . '. 21 
 
 Lowell 27 
 
 New York 19 Cambridge 33 
 
 Buffalo 7 
 
 Albany 18 Rhode Island 30 
 
 New York 31 Providence 36 
 
 Syracuse +11 
 
 An examination of these figures shows that in three of the six states a smaller per cent, 
 of decrease occurred in the infant death rate for the state at large than for any of the cities, 
 while in the three others the per cent, of decrease in the rate for the state was greater than 
 that of about half the cities in each case. These figures are of practically no value for the 
 purpose because they extend over too short a period of time to avoid the risk of the over- 
 influence of temporary and superficial factors and because they are based on one year 
 periods instead of five year periods, as was the case in the tables giving the figures for the 
 European countries and for Massachusetts. These two objections make the use of these 
 figures as an indirect means for comparing the decline in infant mortality in urban and 
 rural districts of little value. 
 
87 
 
 in infant mortality which has occurred during the last thirty years 
 has been greater in urban than in rural districts. 
 
 It has now been shown that the rate of infant mortality is 
 generally higher in urban than rural districts and that the ex- 
 cessiv death rate for the cities results, to a much larger extent, 
 from the influence of postnatal than prenatal conditions and to 
 a larger extent from preventable than nonpreventable causes. 
 In view of this it has also been stated that the recent decline which 
 has occurred in infant mortality should be expected to result in 
 a greater proportional reduction in the rate for urban than rural 
 districts. The lack of statistics extending over a sufficiently 
 long period has made it impossible to show this directly except for 
 two countries but the use of an indirect method of comparing the 
 per cent, of decrease in the rate for an entire country with that for 
 its chief cities showed that the decline in infant mortality during 
 the last thirty years in the principal foreign countries was greater 
 in the large cities than the rural districts. This also proved to 
 be the case in the one state of this country where such a com- 
 parison could be made, Massachusetts. In the two countries 
 where a direct comparison was possible, Prussia and Germany, 
 it was shown that the per cent, of decrease in the infant mortality 
 rate for the urban was greater than that for the rural districts. 
 From this direct and indirect method of comparison the conclu- 
 sion is to be drawn that in countries where the rate of infant mor- 
 tality has been declining for a considerable period this decline 
 has probably been greater in urban than rural districts. 
 
 The significance of the greater decline in the number of infant 
 deaths in the cities than in the country is obvious. It is de- 
 cidedly encouraging when considered from the standpoint of 
 the prolongation of life and the conservation of the public health ; 
 for this markt decline in the mortality of infants in cities has been 
 brought about by methods which if applied in a similar manner 
 in the country and in small towns would probably yield as grati- 
 fying results. Such figures as these just quoted demonstrate 
 beyond a doubt that the improvement in sanitation, in housing, 
 and in living conditions in general, combined with the great ad- 
 vances in preventiv medicine, has been so successful in the cities 
 
88 
 
 that a wider extension of the campaign to the country is im- 
 perativ.^ 
 
 THE INFlvUBNCS OF CONGESTION AND OVERCROWDING ON INFANT 
 
 MORTAIvlTY. 
 
 -i One of the conditions producing a higher rate of mortaUty 
 in tu-ban than rural districts is, as has already been pointed out, 
 congestion and overcrowding. The influence of this factor is 
 very difficult to measure and this must be fully appreciated at 
 the outset. It is a factor, moreover, the influence of which 
 can never be entirely separated from that of the other closely 
 related conditions, such as poverty, the size of the family, and 
 the character, strength, and intelligence of the mother. These 
 two difficulties combined with others that might be mentioned 
 make the study of the influence of congestion and overcrowding 
 on infant mortality a very difficult one and necessitate the use 
 of the most refined methods if conclusions are to be drawn with 
 any assurance of accuracy. 
 
 Broadly speaking, there are two methods of approach to the 
 problem; first, by the study of congestion and overcrowding as 
 measured by the density of population per acre or square mile 
 in a given area; and, second, as measured by the number of 
 persons per room in individual households. It is obvious that 
 the first of these methods of studying congestion in relation to 
 area is very crude and defectiv. Certainly when the areas com- 
 pared are whole countries or even cities, no conclusiv results can 
 be expected, as such areas are not even approximately equally 
 congested in all their parts or sufficiently alike in other respects 
 to justify the making, of such a comparison. Neither can the 
 wards of a city when compared be expected to show any significant 
 relationship between density of population per acre and the rate 
 of infant mortality. Such comparisons have often been made 
 but always with unsatisfactory' results. In Boston, for instance, 
 practically no relationship whatever can be found to exist be- 
 
 1 For a good example of preventiv work which reaches rural as well as urban districts 
 see the recent report of the U. S. Children's Bureau on the New Zealand Society for the 
 Health of Women and Children — an Example of Methods of Baby-Saving Work in Small 
 Towns and Rural Districts. This publication can be obtained from the bureau free of 
 charge. 
 
89 
 
 tween density of population per acre on occupied land in the va- 
 rious wards and the rate of infant mortality.^ Wards 6 and 8 
 where density of population is greatest usually have very low, 
 and sometimes the lowest, rates of infant mortality in the city, 
 while in the other wards the relationship between the two condi- 
 tions is no closer. 
 ^ The reason for this is not that no relation exists between con- 
 gestion and infant mortality but that density of population per 
 acre is too crude a method of measuring degrees of congestion and 
 overcrowding to bring out the true relation to mortality. Ward 
 lines were not intended to mark off sanitary or overcrowded areas 
 and they are very rarely even approximately uniformly congested 
 in all their parts, different sections, blocks, and tenements pre- 
 senting an infinit variety of degrees of congestion. The popula- 
 tion, moreover, is usually far from homogeneous in nationality, 
 economic condition, and standards of life — to mention only a few 
 of the diversities. It is very doubtful whether it is possible to 
 divide a city into such homogeneous and evenly congested dis- 
 tricts that, when classified according to density of population 
 per acre, any conclusiv relationship to the rate of infant mor- 
 tality will be revealed. As has already been pointed out conges- 
 tion is so inextricably interwoven with the other factors of the 
 problem that a much better method of measuring it than density 
 of population per acre is necessary. To study the problem with 
 any hope of accurate results it will be necessary to carry the in- 
 vestigation into individual households and to study the relation 
 of congestion to infant mortality family by family. 
 
 It should also be mentioned, before leaving this fase of the sub- 
 ject, that it is very doubtful whether density of population per 
 acre has any markt influence on infant mortality in itself. As 
 Dr. Newsholme, the eminent English public health officer, has 
 pointed out, "given houses properly constructed and drained, 
 and given cleanly habits on the part of the tenants, increast 
 aggregation of population on a given area has no influence in 
 raising the death rate, except in so far as it is accompanied by 
 
 ' See the Report of the Massachusetts Homestead Commission, Boston, 1913 (House 
 Document No. 2000), p. 20. 
 
90 
 
 overcrowding in individual rooms, an evil which is by no means 
 necessary under the circumstances named. In other words, 
 there is no causal relationship between density of population 
 per se and a high mortality. The true index of density is the num- 
 ber of persons to each occupied room." 
 
 Realizing the inadequacy of density of population per acre as 
 an index of congestion, Dr. George Newman, Medical Officer 
 of Health of the Metropolitan Borough of Finsbury, Central 
 London, in studying the problem of infant mortality in his own 
 district, used a method which closely approximated, altho it 
 does not fully meet the requirements of "the true index" as pre- 
 scribed above by Newsholme. In this investigation, the results 
 of which are summarized in the following table, he correlated the 
 rate of infant mortality with the number of rooms which the 
 family occupied:^ 
 
 Size Number Infant 
 
 of tenement. of births. mortality rate. 
 
 The borough 2,886 148 
 
 One room 532 219 
 
 Two rooms 1,216 157 
 
 Three rooms 468 141 
 
 Fom- rooms and over 464 99 
 
 Unknown 206 39 
 
 An examination of this table shows that in this district of 
 Central London the number of rooms which the family occupies 
 is an important factor in infant mortality. This method of 
 studying the problem, however, does not take account of the 
 number orf persons in the family. Yet, in spite of this limitation, 
 the table can be taken as furnishing a fairly accurate general 
 indication of the influence of congestion and overcrowding on 
 the death rate among infants, since it would appear that even the 
 smallest family would find a one-room tenement very congested 
 quarters and that the family of average size would find them- 
 selves very much overcrowded in a two-, and moderately over- 
 crowded, in a three-room tenement. 
 
 In tabulating the results obtained in an investigation of infant 
 mortality in Wards 6, 8, 13, and 17 of Boston, recently made by 
 
 * Geo. Newman, M.D.: Infant Mortality. London, 1905, p. 184. 
 
91 
 
 the Research Department of the Boston School for Social Workers, 
 the use of the average number of persons per room in the home 
 as an index of congestion and overcrowding was first considered; 
 but several objections to this method soon appeared.^ In the 
 first place, lodgers and boarders often cause a higher degree of 
 congestion in the family than their actual numbers indicate, 
 as the rooms which they occupy are not always as crowded as 
 those which are left for the children and the father and mother. 
 Some families also set aside one room for use as a parlor or for 
 some other similar purpose, even tho it be at the expense of 
 greatly overcrowding the other rooms, while others do not — ^and 
 no account of this difference is taken when the number of persons 
 and the number of rooms is combined in a ratio. But the chief 
 objection to the use of the average number of persons per room 
 as an index of congestion in the homes of the infants visited is 
 a practical one arising from the fact that the visit was not made 
 until the child was a year old which often made it impossible for 
 the mother to recall just how many persons were living with the 
 family, including lodgers, boarders, relatives, grown children, 
 and others, during the first year of the child's life. In other cases 
 the number of persons in the family is not the same thruout the 
 entire year, while in still others the family may have moved 
 to a new home where the number of rooms was not the same as 
 in the old. All these objections and others that might be men- 
 tioned tend to lessen the value of the average number of persons 
 per room as a measure of the extent of congestion and over- 
 crowding in the infant's home during its first year of life. 
 
 There is another method of measuring the extent of congestion 
 and overcrowding in the home, however, which is not subject, 
 to such a large extent, to the objections just named and which, 
 altho it has not been often used, was found in both the Boston 
 and Johnstown investigations to provide the best working method 
 for attaining this end — the number of persons sleeping in the room 
 
 1 The writer is indebted to Dr. J. R. Brackett, Director of the School for Social 
 Workers, for the use of data collected in this investigation which will be quoted later on. 
 These figures have not been previously publisht. The investigation was made in the 
 academic years 1910-11 and 1911-12. During the second of these years the field work 
 was carried on under the direction of the writer supervised by the director of the research 
 department, Dr. T. W. Glocker. 
 
92 
 
 with the infant. The advantages of this method over the others 
 are plain. It considers the use to which the rooms are put — if 
 the actual congestion in the bedrooms and the rest of the house 
 is increast by the withholding of one room for use as a parlor or 
 if lodgers occupy a larger proportion of the rooms in proportion 
 to their number than the other members of the family this condi- 
 tion will, in the great majority of cases, be reflected in the number 
 of persons sleeping in the room with the infant. The number of 
 persons sleeping in the room with the infant is also more easily 
 and more accurately obtainable in a great many cases than the 
 number of persons in the home, particularly when the family 
 "takes in lodgers" and when relatives and grown children live 
 with the family at certain times but not regularly. When the 
 infant's sleeping room only is considered account can also be 
 taken of the size of the room, a thing which is practically im- 
 possible when all the rooms in the home are taken into account. 
 As has been frequently pointed out, the influence of congestion 
 and overcrowding is so closely interwoven with that of the other 
 domestic and social factors of infant mortality that in the study 
 of the problem the method of measurement chosen is of funda- 
 mental importance. 
 
 The following table shows the relationship between congestion 
 and overcrowding, as indicated by the number of other persons 
 sleeping in the room with the infant, and the rate of mortality 
 for infants visited in both the Boston^ and Johnstown investi- 
 gations : 
 
 Boston Investigation.'' 
 
 Number other persons Number Infant 
 
 sleeping in room of mortality 
 
 with infant. births. rate. 
 
 Total 1,899 92.2 
 
 Two or less 1,315 86.7 
 
 Three 403 81.9 
 
 Four or more 181 i54-7 
 
 1 See previous note. Visits were made in this investigation to the homes of all the 
 infants born in the four wards in 1910 with a few exceptions. Answers were obtained in 
 2,063 cases. 
 
 * Information was not obtained in 164 instances. 
 
93 
 
 Johnstown Investigation. ^ 
 
 N umber other persons Number Infant 
 
 sleeping in room of mortality 
 
 with infant. births. rate. 
 
 Total 1,382 87.8 
 
 Two or less 600 66 . 7 
 
 Three, four, or five 725 97 . 9 
 
 Six or more 57 122.8 
 
 Thus, in both the Boston and Johnstov^m investigations the 
 infant mortahty rate was found to vary in direct ratio v^dth the 
 number of other persons sleeping in the room with the infant — 
 or, in other words, with the degree of congestion and overcrowd- 
 ing in the household. Thus, in Boston the rate was over three times 
 as high in the homes where four or more other persons slept in 
 the room with the infant as in those where only two other per- 
 sons or less slept in the room, while in Johnstown the infant 
 mortality rate was almost twice as high where six or more persons 
 slept in the room with the infant as where two other persons 
 or less slept in the room. 
 
 It is worth while to study separately the influence of con- 
 gestion and overcrowding among the Italian families visited in 
 Boston, since extreme overcrowding was so much more common 
 among them than in any other nationality. Over twice as large 
 a proportion of the infants of Italian parentage who were included 
 in the investigation slept in a room with four or more other persons 
 than was the case among infants of any other nationality (18.2 per 
 cent, in comparison with 7.2 per cent, among the Irish, the next in 
 order) . The effect of this high degree of congestion and overcrowd- 
 ing on the rate of infant mortality is reflected in the following figures : 
 
 Of the 709 infants of Italian parentage^ included in this investigation, 414 
 slept in a room with two other persons or less, 166 with three other persons, 
 and 129 with four or more. Among the infants in the first group who slept 
 in a room with two other persons or less the rate of mortality was 86.9, among 
 those who slept in a room with three other persons it was higher, 102.4, 
 while among those who slept in a room with foiu^ other persons or more the 
 rate was highest of all, 155 deaths per 1,000 births. 
 
 ' U. S. Children's Bureau: Infant Mortality: Johnstown, Pa. Washington, 1915, 
 p. 25. No information obtained in 7 instances. No infants who did not live at least one 
 month are included. 
 
 ' Reference is to infants both of whose parents were born in Italy. The term 
 "nationality" was also used above in the same sense. 
 
94 
 
 Turning now to a more detailed study of the effect of con- 
 gestion and overcrowding in Boston, it is significant that where 
 the overcrowding was so great that it was necessary for four or 
 more other persons to sleep in the same room with the infant 
 the rate of mortality showed a decided variation with the size 
 of the room, being lowest when the room was large and enormously 
 higher where it was small. This will be seen upon examination 
 of the following table which, altho it is based on a rather small 
 number of cases, is probably in approximate accord with actual 
 conditions. It shows the mortality rate per i,ooo births for in- 
 fants visited in the Boston investigation who slept in bedrooms 
 with four or more other persons, the rooms being classified ac- 
 cording to their relative sizes. ^ 
 
 Number Infant 
 
 Size of bedroom. of births. mortality rate. 
 
 Total 162 142 
 
 Large 48 42 
 
 Medium 84 155 
 
 Small 30 267 
 
 ~^ The evil effects of congestion and overcrowding are greatly 
 increast by the presence of boarders, lodgers, and other persons 
 in the home who are not members of the family circle. This is 
 due both to the fact that such persons, by each occupying a larger 
 proportion of the household than any member of the family 
 frequently increase congestion to an extent out of proportion to 
 their actual numbers, and that their presence in the home, es- 
 pecially in the poorer sections where the number of rooms in the 
 average tenement is small, prevents privacy and necessitates the 
 substitution of an abnormal form of community life for that of 
 the normal family group. This latter condition is moral- and 
 social, using the term in a broad sense, as well as sanitary, 
 and it probably exerts a greater direct influence upon the 
 parents and the older children in the family than upon the 
 baby. Yet this effect, thru its influence upon the mother, her 
 standards of child care, and the amount as well as the character 
 of the attention she can give to her children, reacts no less power- 
 fully, tho indirectly, on the infant. In view of the fact that 
 
 1 No information as to size of bedroom was obtained in 19 cases. 
 
95 
 
 "probably one-fourth of the foreign -bom famiUes in the United 
 States" "take in" boarders and lodgers — not to speak of the 
 nativ-born families — the influence of this factor both directly 
 and indirectly must be considerable.^ How important its in- 
 fluence really is, however, the difficulty of applying the statistical 
 method to this fase of the problem has so far made it impossible 
 to determine. 
 
 THE INFLUENCE OF HOUSING AND LIVING CONDITIONS ON INFANT 
 ^ MORTALITY. 
 
 The study of housing and living conditions in relation to in- 
 fant mortality may for convenience be divided into three parts: 
 first, the character and location of the house or apartment and 
 the equipment of the home; second, the care of the home by the 
 occupants; and, third, the number of persons living in the home 
 in proportion to its size and the degree of relationship of such 
 persons to the head of the family. The last of these three as- 
 pects of the problem has just been discust in the preceding sec- 
 tion on the influence of congestion and overcrowding and the 
 practice of taking in boarders and lodgers on infant mortality. 
 Consideration must now be given to the other two. 
 
 In the last section attention was drawn at the outset to the neces- 
 sity of exercising great care in the selection of a method of measur- 
 ing degrees of congestion and overcrowding and of applying it 
 in actual practice. The difficulties, however, of measuring the 
 extent of overcrowding and congestion are not nearly so great 
 as those incident to any attempt to grade homes according to the 
 kind of care which the family takes of them. For this reason it 
 is almost impossible to devise a working method for accurately 
 measuring the extent of the influence of the hygienic and sanitary 
 condition of the home on infant mortality. In both the Boston 
 
 • In the recent investigation by the immigration commission it was found that 27.2 
 per cent, of the foreign-born families investigated kept boarders or lodgers. This practice 
 was especially common among the Lithuanians (70.3 per cent.), the Poles (35.5), the 
 Slovaks (41.0), the Russian Hebrews (32.1), and the South Italians (22.4). Of the nativ- 
 born white families 13.0 per cent, kept boarders or lodgers and of the nativ-born Negro, 
 33.7 per cent. See Reports of the Immigration Commission, Immigrants in Cities — A 
 Study of the Population of Selected Districts in New York, Chicago, Philadelphia, Boston, 
 Cleveland, Buffalo, and Milwaukee. Washington, 1911, Senate Document No. 338, 
 Vol. 1, p. 81. 
 
96 
 
 and Johnstown investigations attempts were made to measure 
 the influence of this factor by using cleanHness and ventilation 
 as indices of home sanitation and domestic hygiene. The method 
 used in grading the homes visited in the Boston inquiry, however, 
 was so rough that the data collected could not be used in de- 
 termining the effect of bad sanitation on infant mortality, altho 
 this information does furnish a valuable rough indication of the 
 extent to which bad sanitary conditions were present in the homes 
 visited. 
 
 In Johnstown, on the other hand, sufficiently refined methods 
 of measuring degrees of cleanliness and of adequacy of ventilation 
 in the homes visited were employed to enable the results to be 
 used in determining the relationship between infant mortality 
 and domestic hygiene and home sanitation. The following table 
 summarizes the results of this investigation:^ 
 
 Number Infant 
 
 of mortality 
 
 Grading of home. births. rate. 
 
 CL^A>fLINESS. 
 
 Total 1,463 134 • O 
 
 Clean 943 "3 -5 
 
 Moderately clean 354 163 . 8 
 
 Dirty 166 186.7 
 
 Ventilation.^ 
 
 Total 1,386 87.8 
 
 Good 604 28 . 1 
 
 Fair 392 91.8 
 
 Poor 390 169 . 2 
 
 An examination of this table shows a close relationship be- 
 tween the rate of infant mortality and the cleanliness and venti- 
 lation of the home. Thus, the rate for the infants who lived in 
 dirty homes was over a third higher than that for the infants 
 whose homes were clean. The difference between the well and 
 poorly ventilated homes was even greater; only 28 deaths per 
 1,000 births occurred in the well ventilated homes while 169 
 
 1 U. S. Children's Bureau: Infant Mortality. Johnstown, Pa. Washington, 1915, 
 pp. 23 and 26. 
 
 ' Only infants who lived at least one month included here. No information was ob- 
 tained in 3 instances. 
 
97 
 
 occurred in those which were poorly ventilated. The influence 
 of bad home sanitation on the rate of infant mortality appears, 
 therefore, to be markt. 
 
 As has already been shown, these figures are also of value as 
 furnishing an indication of the kind of care which the mothers 
 visited take of their homes. They may also be compared with 
 the similar data gathered in the Boston and other investigations. 
 In reference to ventilation an examination of the table shows that 
 only 604, or somewhat less than half of these Johnstown homes 
 visited were well ventilated, altho an additional 392, or 28 per cent., 
 were fairly well ventilated. In Boston, on the other hand, the 
 proportion of inadequately ventilated homes was probably even 
 greater, since 785, or 43 per cent, of the 1,825 mothers from whom 
 information was obtained admitted that they never kept a window 
 open at all at night except in warm weather.^ 
 A When graded according to cleanliness homes everywhere, as 
 far as can be determined from the available statistics, show a 
 much more favorable condition than in respect to ventilation. 
 Thus, in Johnstown only 1 1 per cent, of the homes were graded 
 as dirty (24 per cent, being graded as moderately clean, and 65 
 per cent, as clean) while 44 per cent, were found to be poorly venti- 
 lated. In Boston also only 15 per cent, of the homes included in 
 the inquiry were found to be dirty at the time of the agent's 
 visit^ while, as has just been shown, 43 per cent, of the homes, 
 on the statement of tjhe mothers themselves, were never venti- 
 lated at night in winter. 
 
 Similar data collected by the United States Immigration Com- 
 mission in its investigation of the population of selected districts 
 in seven large cities, altho the question of ventilation was not 
 considered, show about the same proportion of badly cared for 
 and dirty homes (16 per cent.). The grading of the 10,123 house- 
 holds or apartments visited in this investigation was not based 
 exclusively on cleanliness but on "the degree of care which char- 
 
 1 From the other 1,040 mothers who said that they did keep a window open at night 
 accurate enuf information could not be obtained to determine whether or not the actual 
 amount of ventilation secured was adequate to meet the needs of the family and the baby. 
 
 * The method of grading the homes visited in the Boston investigation was much 
 rougher than the one used in Johnstown and, therefore, less exact. 
 
98 
 
 acterized them at the time of the agent's visit," altho "in de- 
 termining the degree of care both cleanHness and tidiness were 
 taken into consideration." The detailed results were as follows: 
 of the 10,123 homes visited 45 per cent, appeared to be well cared 
 for and 39 per cent, fairly well cared for, while 13 per cent, appeared 
 to be badly and 3 per cent, very badly cared for. In other words, 
 84 per cent, were well or fairly well cared for and 16 per cent, badly 
 or very badly cared for.^ 
 
 "^Besides the care of the home by the occupants and its sanitary 
 and hygienic condition, the equipment of the house or apartment 
 and housing conditions proper may also directly affect the health 
 of the infant and its chances of survival. Indirectly they exert an 
 influence upon the problem of infant mortality by increasing the 
 difficulty of the mother's caring properly for the home or for her 
 children. This fase of the problem was inquired into in some de- 
 tail in the Johnstown investigation, the results of which are sum- 
 marized in the following table showing the relation of "housing 
 accommodations" to infant mortality:^ 
 
 Number Infant 
 
 of mortality 
 
 Housing accommodations. births. rate. 
 
 Total 1,461' 1340 
 
 Water supply in house 1.173 117.6 
 
 Water supply outside house 288 197 • 9 
 
 City water available i,333 132 .0 
 
 City water not available 128 148 .4 
 
 Water closet 739 108 .3 
 
 Yard privy 722 159-3 
 
 Dry homes 808^ 122.5 
 
 Moderately dry homes 336 139-9 
 
 Damp homes 319 156.7 
 
 Bath in home 496 72.6 
 
 No bath in home 965 164.8 
 
 Yard clean 801 99 . 9 
 
 Yard not clean 632 169 .3 
 
 No yard 28 ... 
 
 1 Reports of the U. S. Immigration Commission: Immigrants in Cities, Vol. I, pt. 5, 
 p. 102. 
 
 2 U. S. Children's Bureau: Infant Mortality: Johnstown, Pa. Washington, 1915, 
 p. 23. 
 
 3 Information as to dryness of home secured in 2 additional instances — total, there- 
 fore, equals 1,463. 
 
99 
 
 An examination of this table shows a close relationship be- 
 tween housing accommodations and infant mortality. Thus, 
 the mortality rate was considerably higher for infants living in 
 damp homes, in homes where there was no supply of city water, 
 and in homes where there was no water closet than for those who 
 lived in dry homes equipt with city water connection, water 
 closet, a bath, etc.^ 
 
 -^The character of housing' and living conditions is determined 
 partly by municipal regulations, partly by the standards of land- 
 lords, partly by the standards of the community in which the 
 family lives, and partly by the wishes of the tenants themselves. 
 Which of these exercises the greater influence it is difficult to say. 
 Municipal regulations and the standards of landlords probably 
 have their chief influence upon housing conditions in the strict 
 sense but even here the choice and wishes of the tenant are very 
 important factors, as the well-known fact that the tenants who 
 pay the highest rents do not always receive the best accommoda- 
 tions would indicate.^ The primary responsibility of the tenants 
 for the interior of the apartment or dwelling, the extent to which 
 it is overcrowded, its sanitary condition, and the adequacy of its 
 ventilation will not be seriously disputed even tho it be realized 
 that they are themselves in turn influenced by other conditions 
 outside the home. It is here that the problem becomes inex- 
 tricably bound up with the resources, the character and the in- 
 telligence of the parents. Living conditions in general, but es- 
 pecially domestic hygiene and sanitation, and the fitness of the 
 
 1 It may well be questioned whether the character of the housing accommodations 
 is not also an index of the economic condition of the family; in which case the effect of bad 
 housing accommodations on the mortality rate shown above may be due partly to the in- 
 fluence of poverty. On the other hand too much must not be made of this as the dis- 
 turbing influence of poverty may be offset by some other disturbing factor acting in the 
 opposite direction. See the writer's article in the November issue of the Quarterly Journal 
 of Economics on "The Influence of Economic and Industrial Conditions on Infant Mor- 
 tality." 
 
 2 Commenting on this aspect of the problem in Boston the report of the immigration 
 commission shows how "inertia and racial cohesion may keep the tenants in their present 
 location in spite of the fact that they can secure better accommodations for less money 
 in other parts of the city" and remarks that it "is reasonable to suppose that the owners 
 of the houses, rather than lose profitable tenants, would make necessary improvements 
 if the demands on the part of the tenants were persistent enuf." See the Reports of the 
 U. S. Immigration Commission: Immigrants in Cities. Washington, 1911, Vol. I, p. 
 466. 
 
lOO 
 
 home as a place for a baby to grow up in are to a very large ex- 
 tent the result of that factor, the importance of which can hardly 
 be too strongly emphasized — the strength, the character, and the 
 intelligence of the mother. The chief effect of urban conditions 
 of life, t>ad housing and sanitation, and poor living conditions upon 
 infant mortality is in all probability not direct but indirect, in 
 that they lower the efficiency of the mother as a mother. These 
 external conditions no doubt exercise an important direct effect 
 on the child but their chief influence upon the problem of infant 
 mortality is indirect; that is, housing and living conditions exert 
 their primary influence on the child thru the influence which they 
 have on the health, strength, and character of the mother. 
 
VI. 
 
 THE INFLUENCE OF ECONOMIC AND INDUSTRIAL 
 CONDITIONS ON INFANT MORTALITY 
 
 Reprinted from the Quarterly Journal of Economics, 
 November, 1915, pp. 127-161. 
 
VI 
 
 THE INFLUENCE OF ECONOMIC AND 
 
 INDUSTRIAL CONDITIONS ON INFANT 
 
 MORTALITY 
 
 I. Introductory. Older and Modern Views 
 
 The relation between the rate of infant mortality and 
 the proportion of women employed in gainful occupa- 
 tions was for the first time given serious consideration 
 in the middle and early part of the last century, when 
 the changes and readjustments following the industrial 
 revolution had largely worked themselves out and the 
 methods of the modern science of public health were 
 brought to bear on conditions in the factory and indus- 
 trial towns of England. The subject was discussed 
 officially about 1860 in the reports of the investigations 
 of Sir John Simon and his associates into the Sanitary 
 State of the People of England, when it was shown, to 
 quote the words of the report itself, " that in proportion 
 as adult women were taking part in factory labour or in 
 agriculture the mortality of their infants rapidly in- 
 creased; that in various registration districts which had 
 such employment in them the district death rate of in- 
 fants under one year of age had been from two and a 
 quarter to nearly three times as high as in our own 
 standard districts; and that in some of the districts 
 
 Note. Reprinted from the November, 1915, issue of the Quarterly 
 Journal of Economics, published by the Harvard University Press. 
 
104 INFANT MORTALITY 
 
 more than a few of the infants were dying of ill-treat- 
 ment." ^ Since the time of Simon and his associates 
 much has been written upon the subject of women's 
 work and its relation to infant and child mortality. 
 Until recently the fact that in cities and communities 
 where a large proportion of women are employed in 
 gainful occupations the rate of infant mortality is 
 generally excessively high has usually been pointed to 
 and accepted as conclusive evidence of the influence of 
 the employment of mothers on infant mortality. 
 
 Recent writers, on the other hand, have shown that 
 this relationship is not necessarily one of cause and 
 effect and that the method of studying the influence of 
 the employment of mothers on infant mortality by cor- 
 relating the proportion of women engaged in gainful 
 occupations with the infant mortality rate is faulty and 
 inconclusive. Phelps, particularly, in his recent study 
 of Infant Mortality and Its Relation to Women's Em- 
 ployment in Massachusetts, has clearly shown that 
 other adverse conditions present in the industrial cities 
 of that state can just as well be held accountable for the 
 high rate of infant mortality as the employment of 
 women in industry. In commenting on the data sup- 
 porting his conclusions, which space does not permit us 
 to quote here, he says: 
 
 It has often been customary, in approaching the subject of the 
 employment of married women in its relation to infant mortality, to 
 ignore the many other complex social and economic factors having a 
 bearing upon the problem. The preceding tables show clearly that 
 . . . certain of these factors which have in the past been ignored in 
 the consideration of the problem are with fair uniformity coexistent 
 with a high infant mortahty rate; these being (1) a high proportion 
 of foreign-born, (2) a high female illiteracy, and (3) a liigh birth 
 rate. These factors operate with equal force over large or small 
 areas . . . accompanying the infant death rate with almost perfect 
 
 1 Quoted in George Newman, Infant Mortality, A Social Problem. London, 1906, 
 p. 92. 
 
INFANT MORTALITY 105 
 
 regularity. The [other] factor, . . . the proportion of women en- 
 gaged in extra-domestic occupations ... is found, statistically 
 speaking, associated very uncertainl3% to say the least, with the in- 
 fant mortality rate. ... It will be seen that this result clearly 
 disproves the contention that the extra-domestic emplojnuent of 
 women is the dominant factor in determining the infant death rate, 
 so far as these Massachusetts cities are concerned.^ 
 
 "^ Bearing in mind all the other factors of infant mor- 
 tahty which a high birth rate, a high proportion of 
 foreign-born, and a high rate of illiteracy imply — large 
 families, poverty, low standards of life, ignorance, bad 
 housing and sanitation, and so on, — but to which the 
 statistical method cannot so easily be applied for large 
 areas, little room for doubt remains that there are many 
 other adverse conditions in industrial cities which can 
 with much less probability of error be held accountable 
 for their excessive infant mortality rate than the em- 
 ployment of women in gainful occupations. 
 
 The direct influence of the employment of mothers in 
 gainful occupations in any community on infant mor- 
 tality is largely determined, not by the proportion of 
 females ten years of age and over employed in such 
 work, or even by the proportion of married women, but 
 by the proportion of mothers who are at work during the 
 infancy of their child or were at work during pregnancy. 
 Little accurate information is available on this point, 
 yet enough to show that the proportion of such mothers 
 employed in gainful occupations does not account for the 
 excessive mortality among infants in industrial cities. 
 
 II. The Statistical Evidence 
 
 In a house-to-house investigation recently made by 
 the Research Department of the Boston School for 
 Social Workers, of infant mortality in Wards 6, 8, 13, 
 
 1 Infant Mortality and Its Relation to Women's Employment, in vol. xiii, of 
 the Bureau of Labor's Report on Condition of Women and Child Wage-Earners in the 
 United States. Washington, 1912, pp. 48-49. 
 
106 INFANT MORTALITY 
 
 and 17 of Boston ^ it was found that of 1,810 mothers 
 from whom information on this point was secured, 101, 
 or 5.6 per cent, were employed in some gainful occupa- 
 tion other than keeping boarders or lodgers either during 
 pregnancy or for some time during the infancy of the 
 child or during both periods. In a similar investigation 
 by the federal Children's Bureau in Johnstown, Pa., out 
 of 1,463 mothers visited only 3.1 per cent " went outside 
 their homes to earn money." ^ Other figures are given 
 in the Bureau of Labor's Report on Condition of Women 
 and Child Wage-Earners in the United States. In the 
 volume on the Cotton Textile Industry it is shown that 
 of the married women employed in the cotton mills of 
 New England and the Southern states only 19 per cent 
 had children under three years of age.^ In a similar 
 manner other volumes show that the per cent of such 
 mothers employed in the men's ready-made clothing 
 industry (home finishers not included) was only 9.9,* in 
 the glass industry, 14.1,^ and the silk industry, 17.3 per 
 cent.^ The number of mothers with infants under one 
 year of age is not given, but it must, of course, have 
 been very much smaller — not more than a third of the 
 number who had children under three at most. 
 
 These figures are also indirectly corroborated by the 
 returns of the Massachusetts census of 1905. In this 
 census it was found that of the 573,673 mothers in the 
 state who had children of any age, only 63,400, or 11.1 
 per cent, were engaged in gainful occupations. The 
 
 1 The writer is indebted to Dr. Jeffrey R. Brackett, Director, for the use of these 
 and other figures quoted later from the results of this investigation which have not 
 been previously published. The visits were made to the homes of infants born in 1910 
 by fellows in the research department during the academic years 1910-11 and 1911-12. 
 During the second of these years this field work was done under the direction of the 
 writer, supervised by the director of the research department. Dr. T. W. Glocker. 
 
 * U. S. Children's Bureau, Infant Mortality: Johnstown, Pa. Washington, 1915, 
 p. 47. 
 
 > Vol. i. Washington, 1910, pp. 1016-1022. « Vol. iii. 
 
 « Vol. ii. 8 Vol. iv. 
 
INFANT MORTALITY 107 
 
 children of these mothers were not classified according 
 to age or the number in each family, so that it is impos- 
 sible to estimate from these figures the approximate 
 proportion of mothers at work who had children under 
 one year of age. Obviously, however, the proportion 
 must have been small indeed. 
 
 Some additional data bearing on the proportion of 
 women employed in gainful occupations during preg- 
 nancy is available from the Bureau of Labor's investiga- 
 tion of infant mortality in Fall River, Mass. Of the 580 
 children dying under one year whose families were inter- 
 viewed, the mothers of 45.9 per cent were at work out- 
 side the home during pregnancy. Not quite half of these 
 continued to work until less than three months of con- 
 finement.^ The proportion of mothers employed during 
 pregnancy thus was about equal to the birth rate among 
 working mothers. 
 
 It appears from the available data that the proportion 
 of mothers employed in gainful occupations in the cities 
 and industrial communities of the United States who 
 have children under one year of age ranges from 3 or 4 
 to 8 or 9 per cent. Probably it rises rarely much above 
 14 or 15 per cent, even in cities like Fall River where a 
 large proportion of the female population is engaged in 
 gainful occupations. The proportion of married women 
 employed during pregnancy is greater than that of 
 mothers employed during the first year after confine- 
 ment, being probably about equal to the birth rate 
 among married women who are habitually employed. 
 Thus the available data, altho meager, seem to be suffi- 
 cient to show that in this country the proportion of 
 mothers employed in gainful occupations, while large 
 enough to constitute in itself an important social prob- 
 lem is by no means large enough to account for the 
 excessive infant mortality of industrial communities. 
 
 • U. S. Bureau of Labor Report, vol. xiii, pp. 72 and 111. 
 
108 INFANT MORTALITY 
 
 The conclusion can be drawn that in industrial cities 
 and communities the employment of mothers is not the 
 chief or dominant direct factor in the mortality of in- 
 fants, that its direct influence has in the past often been 
 exaggerated, and, finally, that it is simply one of the 
 adverse conditions in such communities that produce 
 the high rate of infant mortality. How important an 
 adverse influence it is remains yet to be determined. 
 
 As has just been shown, only 3.1 per cent of the 
 mothers visited in the investigation by the Children's 
 Bureau into Infant Mortality in Johnstown, Pa., were 
 found to have been engaged in gainful occupations out- 
 side the home. In the report of this investigation, how- 
 ever, all mothers who gained money by keeping boarders 
 or lodgers were classed with the mothers who went out 
 to work as " employed mothers." Thus, the data col- 
 lected in this investigation apply, not to gainfully em- 
 ployed mothers in the strict sense, but to mothers 
 earning money by keeping boarders or lodgers. The 
 results of the inquiry are summarized in the following 
 table showing the mortality rate for infants visited in 
 the investigation, classified both according to the em- 
 ployment of the mother and the annual earnings of the 
 father. 1 
 
 Mother Mother 
 
 Gainfully Not Gainfully 
 T J. , i i-i Employed Employed 
 
 Infant mortality: 
 
 Total 188.0 117.6 
 
 Annual earnings of father, under $521 247.6 263.2 
 
 " " " $521 to $624 150.9 160.7 
 
 " " " 625 to 779 127.1 102.3 
 
 " " " 780 or over, or ample 2 166.7 93.1 
 
 * U. S. Children's Bureau, Infant Mortality: Johnstown, Pa. Washington, 1015, 
 p. 49. 
 
 2 The word " ample " was used to designate cases " where information concerning 
 the father's earnings was not available and the family showed no evidences of actual 
 poverty." 
 
INFANT MORTALITY 109 
 
 Examination of this table shows that the mortahty 
 rate among the infants born to the gainfully employed 
 mothers was much higher (188) than the rate for those 
 infants whose mothers had no money earning occupa- 
 tions (118). At first sight this would seem to indicate 
 that the work which the gainfully emploj^ed mothers 
 were engaged in had a decidedly adverse influence on the 
 health and mortality of their babies. This conclusion 
 cannot be drawn, however, because closer examination 
 of the table at once shows that the fundamental differ- 
 ence between the two classes of famihes represented is 
 not in respect to the employment of the mother but in 
 respect to the amount of the annual earnings of the 
 husband. By comparing the two groups as a class with- 
 out subdividing them according to the amount of the 
 annual earnings of the father, such effect as the employ- 
 ment of the mother may have on the problem is masked 
 by the influence of poverty. To overcome this difficulty 
 it will be necessary to consider only groups at least fairly 
 homogeneous in respect to the father's income. Examin- 
 ing, therefore, the lower columns of the table it will be 
 seen then in the first two groups, which contain almost 
 60 per cent of the gainfully employed mothers, the infant 
 mortality rate was higher in the families where the 
 mother was not gainfully employed than in those where 
 she was. Only in the upper groups, where the father's 
 income was $625 or over, and where the proportion of 
 mothers who were gainfully employed was very small, 
 was the rate higher among the employed mothers than 
 among those not employed. This table, therefore, fails 
 to prove that the employment of mothers in money 
 earning occupations has a direct adverse influence on 
 infant mortality. On the contrary, it would rather seem 
 to indicate that in the poorer families where the earn- 
 ings of the father are small the employment of mothers 
 
110 INFANT MORTALITY 
 
 in such gainful occupations as these may have a bene- 
 ficial influence on infant mortality — by mitigating the 
 evil effects of poverty. 
 
 The fact should be kept in mind, however, that the 
 money-earning occupation of the Johnstown mothers 
 was mainly that of keeping boarders and lodgers. It 
 will be necessary to turn to the results of two other 
 investigations to study the influence of the employment 
 of mothers in factories and other places outside the 
 home — one recently made in Birmingham, England, 
 and the other in Fall River, Mass. 
 
 The Birmingham investigation was confined to St. 
 Stephen's and St. George's wards of Birmingham, the 
 two wards which contain the largest proportion of 
 mothers engaged in gainful occupations.^ Of the 3,777 
 mothers visited in the three years of the investigation, 
 1908-10, 1,657 were employed in gainful occupations, 
 1,441 being employed in factories and 675 elsewhere. 
 The infant mortality rate was 173 among the children 
 whose mothers were gainfully employed and 179 among 
 those whose mothers were not so employed. In only one 
 of the three years, 1909, was the rate higher for the 
 children of the employed mothers (179) than for those 
 whose mothers were not so employed (169). In no case 
 was the difference in the rates great enough to indicate 
 any direct relationship between infant mortality and the 
 employment of mothers. 
 
 The most comprehensive American inquiry is that 
 recently made by the federal Bureau of Labor into 
 Infant Mortality and Its Relation to the Employment 
 of Mothers in Fall River, Mass.^ It differed in method 
 
 1 Health Department of Birmingham (England), Report on Infant Mortality in 
 St. Stephen's and St. George's Wards, 1911 (p. 7) and 1910 (p. 10). These families 
 need not be classified according to the amount of the father's earnings, as they were 
 sufficiently ahke in this respect to be compared as a class. 
 
 2 U. S. Bureau of Labor, Report on Condition of Women and Child Wage-Earners 
 in the United States. Vol. xiii, part 2. 
 
INFANT MORTALITY 111 
 
 from the other three considered in that only infants who 
 died during the year under consideration, 1908, and 
 whose parents could be found by the agents of the 
 bureau, were included. No births being included, there- 
 fore, the use of infant mortality rates based on the pro- 
 portion of deaths to births was impossible. 
 
 The report of the investigation is divided into parts; 
 the first, on Mother's Work Before Childbirth in Rela- 
 tion to Stillbirths and Infant Mortality, and the second, 
 on Mother's Work After Childbirth in Relation to 
 Infant Mortality. The following quotation from the 
 report sunmiarizes the conclusions of the first part. 
 
 Summarizing the results of the study of the effect upon the chil- 
 dren of the mother's emplojTiient before childbirth, the conclusion 
 must be reached that in Fall River ... no marked differences are 
 discoverable between the cliildren of mothers at home and mothers 
 at work outside the home.^ A slightly larger per cent of stillbirths 
 was reported for the mothers at home, but the per cent of the still- 
 births winch could be traced to the mother's work was the same for 
 mothers at home and mothers at work. The percentage of total 
 deatlis due to diseases of early infancy (indicating prematurity, im- 
 maturity, or defects) was higher for the children of mothers at home 
 than for the cliildren of mothers at work. . . . The mothers at work 
 showed a slightly liigher percentage of cliildren not well and strong 
 at birth. It would appear then that the conditions which were found 
 existing do not indicate that the work of the mother in the cotton mill 
 before childbirth was producing results noticeably different from the 
 work [housework] of mothers at home. It must be borne in mind, 
 however, that the two classes, mothers at work and mothers at home, 
 are not sharply defined and that the group, mothers at home, in- 
 cludes a considerable number of women who were formerly engaged 
 in millwork and whose physical condition may still be affected in 
 some degree by such earlier emplojTnent. ^ 
 
 The following statement summarizes that part of the 
 investigation dealing with the effect of the employment 
 of the mother after confinement on infant mortahty. 
 
 1 Of the 314 mothers at home only 6 were engaged in a gainful occupation See 
 p. 101. The group of " mothers at home " can, therefore, be considered as one of 
 mothers not employed in gainful occupations. 
 
 » Ix)c. cit., pp. 119-120. Italics added. 
 
^ 
 
 INFANT MORTALITY 
 
 Only 83, or 14.4 per cent of all the children dying under one year 
 concerning whom information was secured, were found to have been 
 deprived of the mother's care because of her going to work. This 
 per cent represents the extent of the possible effect of the mother's 
 absence from home. But the extent to which the nursing of the 
 child was affected is smaller than even this figure indicates, for in 
 only 41 cases, or 7.9 per cent of all those whose feeding was reported, 
 was the mother's nursing in any way affected by her absence from 
 home, and in the 42 other cases she either failed to nurse because of 
 disinclination or inability, or had discontinued nursing for reasons 
 not in any way connected with her return to work. 
 
 But while the number and per cent of cliildren affected by the 
 mother's absence from home was small, yet the causes of death 
 among this number as compared with the causes among cliildren 
 whose mothers remained at home, show strikingly the fatal effect 
 of the mother's absence and the lack of her care and nursing. Thus, 
 the proportion of deaths from diarrhoea, enteritis, and gastritis 
 among the children whose mothers went to work (62.7 per cent) 
 was over 80 per cent in excess of that of the cliildren whose mothers 
 remained at home (34.6 per cent). The real significance of tliis 
 excess will not be fully realized until we recall . . . that for Fall 
 River as a whole the death rate under one year from diarrhoea, 
 enteritis, and gastritis, was two or three times what it was in manj^ 
 other localities. 
 
 The much liigher mortality among the children of the mothers 
 who went to work after childbirth is plainly due chiefly to the great 
 extent of the absence of breast feeding and of the improper feeding 
 and the additional evil influence of the withdrawal of the mother's 
 care. Among the mothers at home only 34 per cent of the children 
 were nursed exclusively; while 24 per cent were given solid food, 
 and for 16 per cent condensed milk was the principal food. Among 
 the children of the mothers who went to work only 1.2 per cent were 
 nursed exclusively, while 40 per cent were given solid food, and for 
 30.5 per cent condensed milk was the principal food. 
 
 The causes of the excessive infant mortality in Fall River may 
 be summed up in a sentence as the mother's ignorance of proper 
 feeding, of proper care, and of the simplest requirements of hygiene. 
 To this all other causes must be regarded as secondary.^ 
 
 The results of these investigations, then, clearly 
 demonstrate that the employment of mothers in gainful 
 occupations is not the chief or even one of the more im- 
 portant direct factors in infant mortality. They also 
 
 1 Loc. cit., pp. 168-169. 
 
INFANT MORTALITY 113 
 
 indicate that even in industrial cities like Fall River the 
 proportion of mothers employed in gainful occupations 
 at any particular time is by no means large enough to 
 exercise directly the influence on infant mortality that 
 has frequently been ascribed to it in the past. 
 
 On the other hand, these studies do not demonstrate 
 that because the gainful employment of mothers is not 
 the chief direct factor in the problem that it is a factor 
 of little importance. How important a factor it is the 
 studies, so far as we have followed them, do not indicate. 
 Moreover, they are so limited in nature and scope that 
 they can throw light only on the direct influence of the 
 gainful employment of the mother during pregnancy, or 
 after confinement during the infancy of the child, on 
 infant mortality. They cannot adequately take into 
 account the influence of the mother's employment dur- 
 ing childhood and young motherhood ; neither can they 
 allow for the compensating influence of the mother's 
 employment in housework at home, which when the 
 family is large is often, as will be shown later, as hard 
 and exacting as many forms of gainful employment; 
 and, finally, to omit mention of other difficulties and 
 limitations inherent in house-to-house investigations of 
 the kind, they do not adequately consider the indirect 
 influence of the employment of married women and 
 mothers and the continued absence of housewives from 
 their homes during a large part of the day on the home 
 standards and the standards of infant and child care of 
 the neighborhood and community. The subject is not 
 so simple as our treatment of it so far would seem to 
 indicate. 
 
114 INFANT MORTALITY 
 
 III. Indirect Influences 
 
 So far we have been dealing primarily with the direct 
 influence of the employment of mothers in gainful occu- 
 pations on infant mortality — that is, the effect which 
 the work in which the mother is engaged has on the 
 chances of survival of her own infant during the first 
 year after confinement. We have only incidentally 
 made reference to the indirect effect which the employ- 
 ment of a large number of mothers in gainful occupa- 
 tions in a community may have on the mortality of their 
 neighbor's children. This aspect of the problem must 
 now be considered with some care. 
 •J As has already been shown, the continual absence of 
 the mother from the home either because she is engaged 
 in some gainful occupation or for any other reason tends 
 to lower the efficiency of the home as " a place for babies 
 to grow up in "; and the same effect also follows, tho 
 perhaps, to a less degree, when the mother is engaged in 
 some gainful occupation within the home. Such em- 
 ployment of the mother may lower or impair the eflS.- 
 ciency of the home in a number of ways, many of which 
 have been pointed out already. It may deprive the 
 baby of the mother's care without furnishing a satisfac- 
 tory substitute therefor; it may necessitate the use of 
 bottle feeding; it may prevent the mother from keeping 
 the baby or the home in as cleanly condition as she 
 might otherwise be able to do; and it may strengthen 
 the tendency for the mother to lapse in the observance 
 of the ordinary rules of hygiene and child care. More- 
 over, — what is especially important in this connection, 
 — the influence of the employment of mothers in gainful 
 occupations on the efficiency of the home may reach out 
 to the homes of other mothers who are not engaged in 
 any gainful occupation and may never have been, and 
 
INFANT MORTALITY 115 
 
 thus help to lower the general domestic and hygienic 
 standards of the community. 
 
 This tendency was clearly evident in the four wards 
 included in the Boston inquiry, where it was found that 
 in spite of numerous individual exceptions the homes of 
 the mothers who were not engaged in any gainful occu- 
 pation did not as a class present any striking differences 
 from the homes of those who were. The same condition 
 has also been noted in other similar investigations.^ 
 This similarity in the homes of these two classes of 
 mothers is probably the result of a number of influences, 
 among the most important of which is the indirect in- 
 fluence of the employment of mothers and married 
 women in gainful occupations on the general home 
 standards of the community. All the families who 
 live in a community help to create the general home 
 standards for that community. The mothers who 
 work do not create one standard for their families and 
 those who do not work another somewhat higher for 
 theirs. 
 
 '-^ In the light of this it is not surprising that the mor- 
 taUty rate for infants born to mothers who are engaged 
 in gainful occupations does not vary markedly from the 
 rate for those infants born to mothers who are not so 
 employed. The truth of the matter probably is, not 
 that the gainful employment of mothers does not affect 
 the chances of life of infants born to such mothers, but 
 that it does not affect their chances exclusively. The effect 
 doubtless falls upon them and their homes first and most 
 severely. But a condition which exists in a large pro- 
 portion of the homes of a neighborhood, or of an entire 
 city, will probably in time affect other homes also ; and 
 the case of the gainful employment of married women 
 
 1 Thus, in the Birmingham report (1910, p. 5) it is stated that " the home condi- 
 tions of those industrially employed do not differ to any large extent from those not 
 80 employed." 
 
116 INFANT MORTALITY 
 
 outside the home is no exception. While not subject to 
 statistical demonstration, it is highly probable that in 
 the cities of England and the United States where a 
 large proportion of married women are engaged in gain- 
 ful occupations, a condition has resulted that has 
 lowered the standards of home life and child care of the 
 entire city, at least of the factory neighborhoods. The 
 influence of this factor of infant mortality is therefore 
 not individual, in the sense that it affects only or princi- 
 pally those mothers who, during pregnancy or while 
 having children under one year of age, are engaged in 
 gainful occupations; it is social, in the sense that it 
 affects all women regardless of the fact of occupation. 
 " No man liveth unto himself and no man dieth unto 
 himself." 
 
 ^\( Besides its influence upon home and community 
 standards the employment of girls and young women in 
 gainful occupations may exercise an important indirect 
 influence on the rate of mortality of infants by sapping 
 the strength and vitality of potential mothers and by 
 affecting the training and education of the mothers of 
 the next generation. The rate of infant mortality may 
 be affected by the employment of women before mar- 
 riage as well as during pregnancy or during the first year 
 after confinement. Dr. Robertson, Medical Officer of 
 Health of Birmingham, England, in the report on infant 
 mortality in that city from which we have already had 
 occasion to quote, lays great emphasis on this point. He 
 says, " I regard as probably one of the most important 
 influences of the industrial employment of women the 
 obvious fact that girls and young women who are in 
 industrial work for many hours daily can have but little 
 time to make themselves practically familiar with the 
 very numerous and often apparently unimportant mat- 
 ters which make all the difference between a well-ordered 
 
INFANT MORTALITY 117 
 
 home and one which lacks the mfluence of a capable 
 mother." ^ 
 
 " Since the employment of women and mothers in gain- 
 ful occupations thus affects the rate of infant mortahty . 
 indirectly in several ways, it is manifestly impossible to 
 measure the influence of the employment of mothers on 
 infant mortality by comparing the rate for the children 
 whose mothers were employed during pregnancy, or 
 during the first year after confinement, with the rate for 
 those children whose mothers were not employed during 
 these periods. To measure with any degree of accuracy 
 the influence of any factor on the problem, the effect of 
 that factor must be " isolated." It is impossible to 
 " isolate " the factor of employment because it affects 
 to a greater or less degree the entire community. There 
 is still another reason, however, that seems also to show 
 that the statistics quoted in the early part of this article, 
 which on their face seem to minimize the importance of 
 the employment of women in gainful occupations as a 
 factor in infant mortahty, are not conclusive. This 
 other neglected factor of infant mortality must now be 
 considered in detail. 
 
 IV. Influence of Housework 
 
 In past discussion the assumption has too often been 
 made that the only kind of employment that can have 
 an appreciable influence on the mortality rate of infants 
 is gainful employment. Such an assumption, however, 
 as will be evident from even a cursory examination of 
 its basis, is false. Work is work and employment is 
 employment, whether it be housework or factory labor; 
 or whether it be in the mother's own home or in some 
 other woman's home, whether it be for hire or simply to 
 keep the mother's own house in order and her own 
 
 1 Loc. cit., Report for 1910, p. 16. 
 
118 INFANT MORTALITY 
 
 family clothed and fed. Tho gainful employment is 
 likely on the whole to have a more harmful influence on 
 infant mortality than the employment of mothers in the 
 performance of their own household duties, it does not 
 follow that the influence of the latter can safely be dis- 
 regarded. 
 
 A good example of the similarity of the influence pn 
 infant mortality of the employment of mothers in gain- 
 ful occupations and in the performance of their own 
 household duties is seen in the length of time the two 
 classes of mothers left off work before confinement and 
 began again afterwards. It was shown by the Fall 
 River/ Boston, and Johnstown investigations that a 
 considerable proportion of the mothers employed in 
 gainful occupations did not stop work until less than two 
 weeks before confinement (9 per cent in Fall River) and 
 that a large proportion did not stop until a month before 
 confinement (21 per cent in Fall River). This failure to 
 stop work a sufficiently long time before confinement is 
 generally recognized as a factor in infant mortality. 
 But it is not a factor the influence of which is confined 
 solely to the children of the gainfully employed mothers. 
 Both the Boston and Johnstown investigations showed 
 that a much larger proportion of the mothers employed 
 only in the performance of their own household duties 
 continued to work very close up to confinement. More- 
 over, the results of these investigations also show that 
 both classes of mothers began work again too soon after 
 confinement. 2 Thus the effect of this factor of infant 
 mortality, altho probably exercising a more serious in- 
 fluence in the case of gainfully employed mothers, is 
 present in the case of the children of mothers not gain- 
 fully employed. 
 
 1 Loc. cit., p. Ill, 
 
 2 For Johnstown, see loc. cit., pp. 44-45. Of the mothers visited by the Research 
 Department of the Boston School for Social Workers 21 per cent began work less than 
 one week after confinement and more than 60 per cent less than two weeks. 
 
INFANT MORTALITY 119 
 
 It is not possible to compare the character of the work 
 done by mothers who are employed simply in the per- 
 formance of their own household duties with that of the 
 mothers who are gainfully employed, in such a way as to 
 measure the relative effect of each kind of work on infant 
 mortality. Yet practically all of the recent investiga- 
 tions that have dealt with this phase of the subject 
 have brought out the fact that the conditions under 
 which mothers perform theu- household duties at home, 
 and the amount and character of this work, are often 
 such as to have an injurious effect on their own health 
 and that of their babies. Thus in the Boston investiga- 
 tion it was found that the conditions under which the 
 mother worked at home were often no better than those 
 of the factory and in many cases the work itself was no 
 hghter. Dr. Robertson also found a similar state of 
 affairs to exist in his study of infant mortality in two 
 wards of Bu-mingham, England, to which reference has 
 already been made. The same condition is also alluded 
 to in the report of the Fall River investigation. Thus 
 " the character of the work . . . seemed to be as im- 
 portant an apparent cause of stillbirths among the 
 mothers who were engaged at their own housework as 
 among those who were employed in mills." Moreover, 
 a sHghtly larger proportion of the children of " mothers 
 at housework only " (54 per cent) were reported as " not 
 well and strong at birth " than of " mothers at mill- 
 work " (53 per cent). In commenting on these figures 
 the writer says: 
 
 " The significance of these figures appears to be not in the slight 
 excess of children not well and strong at birth, but in the fact that 
 for the mothers at home the percentage is practically as high, 
 plainly indicating that if there is an injurious effect of mill work, 
 there must also be in many of these cases an effect almost in the 
 same degree injurious resulting from the work at home." ^ 
 
 » Loc. cit., pp. 104, 110-111. 
 
120 INFANT MORTALITY 
 
 It is, thus, evident that the influence of the employ- 
 ment of mothers on infant mortaUty is not confined 
 exclusively to children whose mothers are gainfully 
 employed. The essential thing is the work which the 
 mother has to do, not whether she is paid for her work or 
 not. Where the hours are long and the conditions under 
 which the work is done are inconvenient or unsanitary, 
 where the work itself is heavy and exacting, and where 
 it is continued close on to the day of confinement and 
 begun again soon afterwards we may expect that in the 
 long run the influence on infant mortahty will be bad, 
 whether the particular work done be the mother's own 
 housework, some form of factory labor, or some other 
 work for which she receives a definite wage. 
 
 All this is equivalent to sa;ying that the figures pre- 
 sented in the previous section on the statistical evidence 
 are inconclusive. If, first, the influence of gainful em- 
 ployment, so far from being confined exclusively to 
 infants born to mothers who are gainfully employed 
 during pregnancy or during the first year after confine- 
 ment, affects indirectly, through its influence on home 
 and community standards of child care, the chances of 
 survival of infants whose mothers are not gainfully 
 employed; and if, second, the influence of employment 
 — that is, of the work which the mother does, is active 
 in both groups, it is hardly to be expected that a com- 
 parison of the mortality rate of infants born to mothers 
 who are employed in gainful occupations with the rate 
 for those whose mothers are employed simply in the per- 
 formance of their own household duties will yield any 
 conclusive results. But still another difficulty with this 
 method of measuring the influence of the employment 
 of the mother on infant mortality remains yet to be 
 considered. This arises from the close interrelationship 
 of the influence of the employment of the mother and 
 poverty. 
 
INFANT MORTALITY 121 
 
 V. Influence of Poverty 
 
 The relationship between poverty and infant mor- 
 tahty was first made the subject of statistical study by 
 Charles Booth as a part of his investigation of East 
 London. He classified the thirty-three residence dis- 
 tricts of the city which were selected for study into three 
 groups according to the proportion of the population 
 living in poverty and in crowded quarters and compared 
 the infant mortality rates for the groups. In the first 
 group, which contained the largest proportion of the 
 " poor and overcrowded," the rate was 169 deaths per 
 1,000 births; in the second group, containing the largest 
 proportion of people of " the comfortable central class," 
 it was considerably less (148) ; while in the third group 
 which contained the largest proportion of the " upper 
 classes," it was lowest of all (132).^ 
 
 Later, in 1898, Rowntree made a similar investigation 
 of " vital statistics of typical sections of the population 
 of York, England," including infant mortahty. He 
 divided the population of the city into four classes and 
 compiled the rate of infant mortality for each class. 
 The rate for the " poorest working class " was highest of 
 all, 247 deaths per 1,000 births; for the " middle work- 
 ing class " considerably lower, 184; for the highest 
 working class lower still, 173; and for the " servant 
 keeping class " lowest of all, 94.^ 
 
 The objection to the method employed in both these 
 investigations, altho not of sufficient importance to 
 seriously weaken the conclusions drawn, is that the unit 
 considered was not the individual family but selected 
 areas. Later investigations have avoided this objection 
 by comparing the infant mortality rates for different 
 
 ' Charles Booth, Life and Labour of the People of East London, 1903. Final vol- 
 ume, pp. 26-27. 
 
 » B. S. Rowntree, Poverty; A Study of Town Life, 1901. 
 
122 INFANT MORTALITY 
 
 families classified by income. This method was used in 
 the investigation by the Health Department of the City 
 of Birmingham, England, in the investigation to which 
 reference has been made several times already. Among 
 the infants whose fathers were " out of work or earning 
 less than one pound per week " the mortality rate during 
 the two years of the investigation was 204 deaths per 
 1,000 births; while for the infants whose fathers earned 
 more than one pound per week the rate was considerably 
 less, 137.^ Thus, in these two wards of the city, both of 
 which were " occupied almost entirely by poor people," 
 the infant mortality rate varied markedly with the 
 wages of the father. 
 
 Before leaving the results of the British investigations 
 of this phase of the subject attention should at least be 
 called to the tabulation which the registrar-general of 
 England and Wales made in his last annual report show- 
 ing the relationship between infant mortality and the 
 father's occupation. ^ The results, altho they cannot be 
 quoted in detail, showed quite clearly the effects of 
 poverty and its accompaniments on infant mortality, 
 since everywhere the rate of mortality was higher among 
 the babies whose fathers were employed in the poorly 
 paid occupations than among those whose fathers were 
 better paid. 
 
 The best as well as the latest American investigation 
 of the relation between poverty and infant mortality 
 was made by the federal Children's Bureau as a part of 
 its comprehensive study of infant mortality in Johns- 
 town, Pennsylvania. The data gathered in this inquiry 
 are of especial value in this connection, since all the 
 babies born in the city during the year of the investiga- 
 tion were included, not simply those born in a particular 
 
 1 Report on Infant Mortality in St. Stephen's and St. George's Wards, 1912, p. 11. 
 
 2 Registrar-general for Births, Deaths, and Marriages in England and Wales, 
 Annual Report for 1912. 
 
INFANT MORTALITY 
 
 123 
 
 section or in relatively poorer families. The results are 
 summarized in the following table, which shows the 
 mortality rate per 1,000 births for the infants included 
 in the investigation, classified according to the annual 
 earnings of the father: ^ 
 
 Number Infant 
 
 of Births Mortality Rate 
 
 Total 1,431 130.7 
 
 Annual earnings of father, under $521 219 255.7 
 
 " " " $521 to $624 165 157.6 
 
 " " " $625 to $899 385 122.1 
 
 " " " $900 to $1,199 138 101.4 
 
 " " $1,200 or more ... 48 83.3 
 
 " Ample " 476 84.0 
 
 Thus the infant mortality rate varies closely with the 
 amount of the annual earnings of the father. For the 
 infants whose fathers earned less than $521 annually the 
 rate was twice as great as that for infants whose fathers 
 earned between $625 and $799 annually, and three times 
 as great as the rate for those infants whose fathers 
 earned $1,200 or more. The relationship is a very close 
 one — as close a one as one can well expect to find in 
 vital statistics. The question remains for determina- 
 tion: is the relationship between poverty and infant 
 mortality as close a one as the relationship between the 
 employment of the mother and infant mortality ? 
 
 As has been shown already, families in which the 
 mother is gainfully employed differ radically in respect 
 to the father's income from those in which the mother is 
 not so employed. Thus, 48 per cent of the husbands 
 included in the Johnstown investigation who received 
 less than $521 annually had wage-earning wives, in com- 
 parison with 33 per cent of those who received from $521 
 to $624, and 22 per cent of those who received from $625 
 to $779 a year. Only 9 per cent of the husbands who had 
 
 1 U. S. Children's Bureau, Infant Mortality: Johnstown, Pa. Washington, 1915, 
 p. 46. 
 
124 INFANT MORTALITY 
 
 an annual income of $900 to $1,199 a year had wage- 
 earning wives and only 2 per cent of those whose income 
 was $1,200 a year or over.^ It is, of course, to be ex- 
 pected that the great majority of wage-earning mothers 
 should have husbands whose annual earnings are small. 
 But the significance of this fact in relation to the in- 
 fluence of the employment of mothers in gainful occupa- 
 tions on infant mortality has by no means always been 
 fully appreciated. It may seem that in one sense the 
 employment of mothers in gainful occupations is not so 
 much a cause of the high rate of infant mortality as it is 
 a sort of remedy for the adverse influence of poverty. 
 Before going further into this subject, however, it will 
 be necessary to compare, as far as the difficulty of dis- 
 tinguishing the effect of the one from the other makes 
 possible, the relative influence of these two factors on 
 the problem. Such a comparison is possible from the 
 figures quoted in the following table from the reports of 
 the Birmingham, England, and the Johnstown, Pa., 
 investigations showing the mortality rate for infants 
 classified both according to the annual earnings of the 
 father and the employment of the mother: 
 
 Johnstown ^ 
 
 Mother Mother 
 
 Gainfully Not Gainfully 
 
 Employed Employed 
 
 Total 188 118 
 
 Annual earnings of father, under $521 248 263 
 
 " " $521 to $624 151 161 
 
 " " " 625 to $779 127 102 
 
 " " " 780 or over, or ample 167 93 
 
 Birmingham ^ 
 
 Total 176 170 
 
 Out of work or less than one pound weekly . . . 208 195 
 
 More than one pound weekly 118 152 
 
 1 Loc. cit., p. 48. 2 Ibid., p. 49. ' Ibid., 1912, p. 11. 
 
INFANT MORTALITY 125 
 
 A careful examination of the figures seems to indicate 
 that the influence of poverty in these families is much 
 greater than that of the employment of the mother. 
 In Birmingham the infant mortality rate varies very 
 slightly with the employment of the mother, while the 
 variation with the earning capacity of the father is 
 marked. This latter fact is especially significant when 
 it is remembered that the families visited did not vary 
 greatly as to income, the two wards included being 
 " occupied almost entirely by poor people." In Johns- 
 town, also, the amount of the father's annual earnings 
 seems to vary more closely with infant mortality than 
 with the employment of the mother. Moreover, as has 
 been shown already, it is very probable that the higher 
 rate shown in the table for the children of the gainfully 
 employed mothers is produced, not by the influence of 
 gainful employment at all, but of poverty. It must not 
 be forgotten that the mothers who were gainfully em- 
 ployed were for the most part living in poverty while 
 those who were not so employed were in comparison 
 relatively well-to-do. 
 
 As Dr. Robertson says in commenting on the results 
 of the Birmingham inquiry, " the hfe of the mother 
 among the poorer classes is always a strenuous one if the 
 family is large. ... It does not matter much whether 
 the mother is industrially employed or not ... if 
 poverty is great the infant suffers. . . . From the 
 tables [given in his report] it is seen that the influence of 
 poverty ... on the infant mortality rate is far greater 
 than that of industrial employment." To this conclu- 
 sion the present writer must subscribe, for all the data 
 presented tend to emphasize the fundamental impor- 
 tance of the relationship of poverty to infant mortality. 
 
 The relation of poverty to hygienic and home condi- 
 tions remains to be pointed out, altho this is not the 
 
126 INFANT MORTALITY 
 
 place for an adequate discussion of this phase of the 
 problem. As an experienced English medical officer of 
 health writes to Dr. Newman, — " Infant mortality in 
 Lancashire is, I am sorry to say, as much a financial as a 
 hygienic question. ... A weaver's wages will not 
 allow of the wife's remaining at home, considering rents 
 and rates, and so both go — which is the rule — and a 
 hand to mouth existence results even for themselves, 
 let alone the little ones. . . . Much good may be done 
 by hygienic tuition, but I am certain that the root of the 
 whole matter with us is, as I have said, comparatively 
 low wages and high rents and rates," ^ — or, as one 
 would say in America, low wages, and a high cost of 
 living. 
 
 VI. Conclusion 
 
 It appears, then, that the fundamental cause of the 
 / excessive rate of infant mortality in industrial communi- 
 I ties is poverty, inadequate incomes, and low standards 
 /' of living with their attendant evils, including the gainful 
 I employment of mothers. The employment of the mother 
 I, in gainful occupations is simply the remedy for these 
 ' evils or " adverse conditions " which the working people 
 in industrial communities have adopted. Undoubtedly, 
 this recourse has had an important effect on the prob- 
 lem, in many cases actually tending to reduce the rate of 
 infant mortality, while in others having just the oppo- 
 site effect. The primary question in considering the 
 social causes of infant mortality is whether the employ- 
 ment of mothers and married women in extradomestic 
 occupations is, from the viewpoint of society as a whole, 
 a good remedy for poverty and an acceptable means 
 of mitigating its influence on the health and mortality 
 of babies and young children. From the point of view of 
 
 1 Quoted in Newman's Infant Mortality. London, 1906, pp. 137-138. 
 
INFANT MORTALITY 127 
 
 the individual poor or poverty stricken family, the fact 
 cannot be escaped that this effect may be both good and 
 bad: bad, in that it causes the baby to be artificially 
 fed, forces the mother to be absent from home, and in 
 other ways lowers her efficiency as a mother; good, in 
 that it increases the family income and decreases the 
 influence of poverty. We are, thus, forced to conclude 
 that the fundamental economic and industrial factor of 
 infant mortality is low wages. The fundamental remedy 
 is obviously higher wages. Other remedies, such as 
 legislation restricting or regulating the employment of 
 mothers before and after confinement,^ day nurseries, 
 the instruction of mothers and school girls in domestic 
 economy, and the like, all have their place; but the 
 chief thing remains the provision of an adequate family 
 income. 
 
 • In commenting on this phase of the problem, Dr. Robertson, in his Report on 
 the Industrial Employment of Married Women and Infant Mortality in St. Stephens 
 and St. George's Wards, Birmingham. England (for the year 1910, p 21) says: " It 
 appears to be a question in this Birmingham area whether the additional poverty which 
 would be occasioned by preventing mothers from working for, say, six months after a 
 birth, would not be the greater of the two evils." 
 
VITA 
 
 The writer was born in Smithland, Ky., November 25, 
 1887. In 1908 he graduated from Wilhamsburg (Ky.) In- 
 stitute. He received the A. B. degree from Brown University 
 in 1910 and the A. M. degree in 1911. In 1910-12 he held a 
 Fellowship in the research department of the Boston School 
 for Social Workers. While in Boston he was a resident of the 
 St. Mary's House for Sailors and also of South End House 
 and in addition was a member of Conference 7 of the Asso- 
 ciated Charities. In 1911-12 he attended a course in sociol- 
 ogy under Prof. T. N. Carver at Harvard University. In 
 1912-13 he taught history and social science in Tarleton 
 College (Texas) and in 1914-15 sociology and statistics at the 
 University of Illinois. In 1915 he was a lecturer in the Sum- 
 mer School for Social and Religious Workers conducted by the 
 Biblical Department of Vanderbilt University and the Ameri- 
 can Interchurch College. He was registered at Columbia 
 University in 1913-14 and 1915-16, attending courses under 
 Professors Giddings and Tenney in sociology, under Profes- 
 sors Seligman and Seager in Economics, under Professor 
 Devine in Social Economy, and Professor Chaddock in Sta- 
 tistics.