As to the causes for this decreased rate, opinions seem to vary according to the view point of the person expressing the opinion. For example, the Board of Education officials were inclined to attribute it to the greater educational advantages enjoyed by the mothers through the establishment of their infant welfare centres. Improved sanitation is considered by others to have had great effect. Others again consider that the national campaign against venereal disease had probably had a considerable share in it; others the lessened employment of women in factories, and yet others the weakened strength of alcoholic beverages. Practically all these suggestions, as anything more than simply contributory to causation, are discounted by the fact of 1912 being so much lower than the years preceding and following it up to 1919.

The mortality rate in the case of illegitimate is very considerably higher than in the case of legitimate babies. In London during 1919 the total deaths of legitimate babies in 1,000 births was 77; in the case of illegitimate babies the total number was 233. This was the case in every area in the United Kingdom as the following table shows:

Infantile deaths from all causes in proportion to 1,000 births.
All Infants Legitimate Illegitimate
232.91

London

County Boroughs

Other Urban Districts.

Rural Districts

All Urban Districts

85.30

76.57

The rates for illegitimate children followed the same lines roughly as the legitimate ones, viz.: the death-rate increased from south to north, being generally lowest in the residential towns of the south, and highest in the industrial towns of the north. For infants of both sexes jointly the mortality varied during 1919 from 108 deaths per 1,000 births in the county boroughs of the north to 64 in the rural districts of the south. In rural districts, however, it is quite common for infant mortality during the first four weeks of life to be higher than the urban mortality. This is of course because environment conditions other than those of birth itself have not had time to tell. This increased rural mortality is very rare after the first month of life. The birth risk in the rural districts of the north was certainly at its maximum, for much the highest rate for the first day was returned by those areas. There was a similar rural excess during 1919 in Wales, though not in other parts of England, so that if defective midwifery in remote districts is the cause, it does not seem to apply to the Midlands and the south of England. In 1917, on the other hand, first-day mortality was in excess in all the rural districts with the excess still greatest in the north.

Two former causes of infantile deaths that have been tending to disappear altogether, are rickets and tuberculous meningitis.

Indeed the marked fall in the mortality of age period 0-5, which necessarily includes mortality of children under one year, from tuberculosis generally, is very noticeable, and was greatly accelerated during 1919. Formerly it was the highest at any age, but it is now exceeded by most of the adult life periods.

Infectious Diseases.-The infectious diseases which influence the infantile death-rate are measles, scarlet fever and whooping-cough. With regard to measles, the death-rate of 1919 was by far the lowest ever recorded. It was much the greatest in the north of England, and since 1912 it has been noted as being the greatest in the large towns. Thus Cumberland, Durham, Northumberland and the North Riding, all in the north, showed the highest mortality, and in spite of the low mortality-rate in the country generally, Middlesbrough and Rotherham remained very high. The total deaths from measles in children under one year in 1919 was 650, or 1.05% of the infantile deaths. In the case of scarlet fever, the same high mortality occurs in the north in all areas whether country or town, compared with the south and with Wales; Birkenhead, Bootle, St. Helens and Liverpool head the list as they did in the previous year. The total deaths from scarlet fever in infants under one year for 1919 was 32, or 051 % of all infantile deaths.

The death-rate from whooping-cough was the lowest ever recorded. There was a marked excess in the case of females, which is constant. The proportion of infantile deaths to the total number of deaths from whooping-cough naturally fell with the diminished birthrate, but it is a curious fact that it was higher in rural districts than in small towns, and in small towns than in county boroughs. The cause of this persistent characteristic it was impossible at present to discover. It was not shared by measles, which was the only other infectious disease causing infantile mortality to any marked extent. The total number of deaths from whooping-cough under one year was 1,054, or 1.7% of all infantile deaths.

Enteritis and Diarrhoea.-Diarrhoea as a cause of death is gradually disappearing from the British returns. This is also the case with tubercle and convulsions. The mortality ascribed to tubercle in 1919 was less than one-third, and convulsions less than one-half of that so returned 14 years earlier. In the case of these two latter diseases it is probably very largely one of nomenclature, whereas in that of diarrhoea it presumably represents in the main an actual decrease of mortality. During 1919 there was a lower infant mortality from diarrhoea and enteritis than ever before, except in 1912, when there were remarkably favourable weather conditions, even better than those in 1919. In 1911, a year with an almost tropical summer, the infantile deaths from these causes were 31,900 compared with 6,039 in 1919, the total infantile deaths being 114,600 and 61,715 respectively.

Syphilis shows a steady decrease compared with the startling rise in 1917, but it had not yet in 1919 attained its pre-war level as a cause of infantile mortality. As in the two previous years, mortality from syphilis was in considerable increase in the north.

Developmental and Wasting Diseases, notwithstanding an increase in 1919 in the first month, showed in that year the mortality of 10.3 per 1,000 births, the lowest ever recorded for England and Wales. By far the most important increase was that due to premature birth. This accounted for 59% of the total increase of mortality during the first four weeks of life. Injury at birth also showed a considerable rise, which increase was entirely confined to male infants, whose mortality had risen from III to 1.34, whereas that of females had remained stationary at 0.85. As the British Registrar-General says: "It looks as if the infants born during 1919 were for some reason exceptionally susceptible to death within the month." This cause of infant mortality was, like so many others, excessive in the north of England, and mainly in the rural districts where it caused a mortality of 1.91 as compared with 1.11 for the country at large. Rural mortality from birth injury was high, and as the health and general condition of the mothers is likely to be better than in towns, it looks as if defective midwifery might be at least a contributory cause. As higher mortality is attributed to premature birth in the north as a whole than elsewhere, the facts of the greater industrial employment of married women in the north might be pointed to as a cause, but mortality from this cause declined during the early part of the war, reaching a minimum in 1916, and increased after the war together with the increased employment of women. Speaking generally, it looks as if a baby born north of the Wash had less chance of surviving than if it were born in the south.

"Overlying."-This cause of infant mortality deserves separate mention. It was very much lower in 1919 than it had ever been before, being 0.76 per 1,000 births, or less than half of what it was a few years before the war. But Dr. W. A. Brend throws some doubt on this as a genuine cause of death. See Health and the State, 1917, as well as Inquiry into the Statistics of Deaths from Violence, 1915, in which he shows that there is no connexion between overcrowding and deaths from overlying, and that the seasonal variation follows that of bronchitis and pneumonia, being considerably higher in the first and fourth quarters of the year than in the second and third. This is borne out by the figures furnished by the Medical Officers of Health for Liverpool and Birmingham for 1919 and 1920. Thus in Liverpool in 1919 there were 18 deaths from this cause in the first and fourth quarters, and 7 in the second and third; and in 1920, 17 and 6 respectively. In Birmingham in 1919, the figures were 23 for

The pop. in the birth registration area for 1915 was 31% of the total estimated pop. of the United States. In 1919 this had increased to about 58%. During a slightly longer period of time, 149 cities were included in the registration area.

The standard accepted by the U.S. Census Bureau is based upon what is known as the Model Law, and its requirements are: 1. Registration of births within 10 days.

2. Use of standard birth certificates.

3. Checks on registration, chiefly by (a) tracing records of deaths of infants under one year of age to see whether birth was recorded, and (b) tracing records of births reported in newspapers.

4. The work of an efficient state registrar possessing full power and responsibility to enforce the law, in direct connexion with local registrars.

5. Prompt monthly returns of the original certificates from the local registrars to the state registrar, with report of "no births" or no deaths" where such is the case, and official statement of completeness of registration or report of delinquents.

6. Enforcement of penalties for non-compliance.

It was reported in 1921 that no state had obtained complete registration of births. The statistics in the area, however, were generally assumed to be approximately correct. It is evident, therefore, that infant mortality statistics in the United States, up to the end of 1921, were based entirely upon births reported in the registration area and that the total mortality must be estimated. It was believed, however, that the infant death rates in the states having unsatisfactory registration of births do not differ essentially from those recorded in the birth registration area.

Infant Mortality in the United States and Some Foreign Countries or Provinces (per 1,000 living births). Denmark (1919)

Connecticut Indiana Kansas Kentucky Maine Maryland

Total Urban

Rural

87

89

84

70

64

79

86

86

87

79

88

74

70

88

65

Massachusetts

The statistics of U.S. birth registration area show that the ratio of sex to infant mortality in that country is about the same as that in other countries. In all years reported, the death-rate of male infants is appreciably higher than the death-rate of female infants. This applies to both cities and rural communities, in 1919 the city rate for males being 98.7 as against a rural rate of 92.9, the city rate for females being 79.3 as against a rural rate of 74.7.

Rates for 10 Largest Cities of the United States 1914-20.

Note: The figures given are the latest available. It is possible that the relative position of the United States would vary if all the statistics were for 1919.

There was a steady and persistent decline in the rates in the birth registration area during 1916-21 with the exception of the year 1918, when there was an increase over the preceding year, due very largely to an epidemic of influenza. The further reduction of the infant death-rate to 87 during the year 1919 supports the belief that the factors which had to do with the general reduction of the rate were exercising a cumulative and progressive effect.

Infant Mortality Rates, U.S. Birth Registration Area, 1915-9. Deaths per 1,000 births 1918 1917 1916 1915

Boston
Pittsburgh

Cleveland.
Philadelphia

Buffalo

Chicago 3 Baltimore.

115.2 107.7 113.8 116.2
122.5 115.3 110.8
116.4 110.6 107.0 104.9 95.4 90.8 86.0
117.6 106.2
126.0 89.8 88.6
121.5 108.2
IOI.O
113.9 103.7 121.5 109.8
122.4 104.3 112.8 103.6 100.7 96.8 104.2

132.7 102.5 111.9 106.4 104.3 91.0 85.5
154.6 119.8 118.2 119.3 147.8 97.0 104.2.

As the above table shows, New York City has had the lowest rate of any of the 10 largest cities of the United States in the years 1914-20 inclusive, except for St. Louis in 1915, 1916, 1917 and 1919. St. Louis is not in the birth regitration area because its birth registration reports are not accepted by the U.S. Census Bureau. Rates in American Cities.-Philip Van Ingen states that a report from 432 or 87.8% of the cities of the birth registration area, with a population of 30,063,288 or 95.2% of the total urban population of this area, shows the infant mortality rates for the five years 1915-20, grouped according to population, to be as follows:

Causes of Infant Mortality in the United States.-It is probable that the variety of race groups in the United States has had a marked effect upon infant mortality. Statistical studies on this point are not readily available but the report of Dr. W. H. Guilfoy of New York City is worthy of mention. Dr. Guilfoy shows that out of every thousand infants born of mothers of RussianPolish or Austro-Hungarian nationality, over 920 survive the first year of life; of Italian mothers 897; of native mothers 894; of German mothers 885; of Irish mothers 881. Other significant results of this analysis show that the nationality of the mother seems to be a predominant factor in deaths from congenital diseases under one year of age per 10,000 births reported. This rate in children born of Italian mothers was 295, of Russian mothers 320, Austro-Hungarian mothers 284 and native-born mothers 544. This clearly indicates that measures for the reduction of infant mortality in the United States in the future must take into considerat on the high death-rate from congenital causes of infants of native-born parentage. The effect of nationality upon deaths by infectious diseases is shown in the fact that children of Italian mothers present the highest mortality in the group, with a rate of 58; children of Irish mothers rank next with a rate of 57; children of native-born mothers 38; children of Austro-Hungarian mothers 36. In respiratory diseases, race seems again to play an important part. The death-rate of infants of Italian mothers from acute respiratory diseases is more than three and one-half times that of children of German mothers, almost three times that of children of Russian, Austro-Hungarian and Irish mothers, and a little more than double that of nativeborn mothers. In diarrhoeal diseases, the racial aspect is shown as follows: The death-rate of infants of English mothers is 91 per 10,000 births, children of native-born mothers 80, children of Irish mothers 72, children of Italian mothers 70, children of AustroHungarian mothers 52, and children of Russian mothers 30.

The influence of race, as shown by these figures for New York City, would seem to indicate that the highest death-rate from congenital causes in infancy is among children of native-born mothers, the highest among infants from infectious diseases among children of Italian mothers, the highest rate from respiratory diseases among children of Italian mothers, and the highest rate from diarrhoeal diseases among children of English and native-born mothers.

Figures for the United States Registration area for 1919 are of interest in this connexion:

Rates Classified According to Country of Birth of Mother, 1919. Total rate

86.6

77.7 112.6

89.3 99.I

66.8

73.2

87.4

78.1

87.7 124.4

73.6 104.8 134.3

Age Groups as Factors in Infant Mortality.-The well-known fact that people are susceptible to their environment in inverse proportion to their age is graphically demonstrated in the case of the infant death-rate by examination of age subdivisions of the first year of life. It is evident that the new-born infant is extremely susceptible to its environmental conditions and that intrauterine factors have an effect in making the infant mortality rate unusually high during the first few days or first month of life. It must be remembered that the intrauterine infant's environment is its mother and that anything that affects her health inevitably reacts upon the infant.

Rates in the U.S. Registration Area 1919 by Subdivisions of First Year. United States (total)

Under one day

One day

Two days

Three to six days

One week.

Two weeks.

Three weeks but under one month

One month.

Two months

Three to five months

Six to eight months. Nine to eleven months

87.0 15.1

4.3 3.3

6.0

5.5

3.5

2.8

6.5

5.9

13.7

10.3

7.9

Sanitation and Environment.-Statistics regarding the effect of sanitation and environmental conditions on the infant death-rate are difficult, when not impossible, to obtain. Clinical and practical experience must be drawn upon to prove that lack of proper sanitation and poor hygiene cause infant deaths. It is generally accepted by child hygienists that the main factors in high rates are poverty and ignorance. The more definite causes of infant mor

tality due to lack of sanitation may be classified as those of social, economic and general environment. Although the direct relation of sanitation to the infant death-rate cannot be proved statistically, it has been proved many times by the marked fall in the infant death-rate when sanitary conditions in a community have shown improvement. The sanitary conditions affecting the infant deathrate may be classified from another point of view as decent housing, proper standard of living, opportunities for recreation and fresh air, clean water supply and clean milk supply.

Studies made by the Children's Bureau at Washington show that the infant death-rate is definitely affected by overcrowding, and that the number of people living in a room can be shown to have a direct statistical relation to the rate. Overcrowding has a direct relation to the economic condition of the family and is reflected in its general standard of living. Such factors, therefore, are not easily separated, but statistical studies have shown uncleanliness, overcrowding, lack of ventilation and lack of decent hygiene in the home are directly responsible for many infant deaths. In the same way, poverty can be shown to be allied to the rate by the fact that the rate bears a close and regular relation to the amount of wages received by the family. In the Johnstown report of the Children's Bureau, U.S. Department of Labor, definite figures are given in this regard. The results of that investigation showed that when the father earned less than $521 per year, the infant deathrate was 255.7; where the father earned more than $1,200, the infant death rate was 84. But wages must be considered again in relation to social and sanitary factors, as a decent standard of living may be maintained on a low wage-rate while a high wage-rate does not necessarily include conformity to hygienic requirements.

Poverty reacts upon infant mortality in still another direction. Insufficient earning capacity of the father usually forces the mother into industry. Statistics relating to the health of mothers who are industrially employed during their child-bearing life or during the period of pregnancy would seem to indicate that employment of these women, in itself, has no deleterious effect upon the infant. More detailed studies and more careful analysis of the studies already made would seem to indicate that the high rate in towns where women are industrially employed to any great extent is due not so much to the effect of the mother's industry upon the child, as to the conditions of poverty in the family that have forced the mother into industrial pursuits. In order to show conclusively that employment of women is a factor of importance in increasing the infant mortality rate, a further study should be made as to the effects of certain types of industry upon pregnant women. Probably one of the most harmful results of the employment of women, so far as infant mortality rates are concerned, is the fact that returning to the industry too soon after confinement is not only harmful to the health of the mother in relation to future pregnancies, but reacts disastrously upon the infant in that the latter usually is deprived of breast feeding.

Type of Feeding.-It has long been recognized that infant deathrates from diarrhoeal diseases are very markedly affected by the feeding employed. It has been proved beyond doubt that the infant death-rate from diarrhoeal diseases can be much reduced by the wider use of breast feeding and by the provision of safe, clean milk for use in artificial feeding. The relation of breast and artificial feeding to infant mortality is graphically shown in certain studies made in New York City. One such study, covering deaths of 1,065 infants from diarrhoeal diseases, showed that 17% of those who died had been breast-fed exclusively, while 83% had been artificially fed, either with cows' milk or some form of prepared infant food. In order to determine the extent of breast feeding among well children, a further study was made covering about 4,000 children between 3 and 12 months of age. In this study it was found that 79% of the babies were breast-fed exclusively while the remaining 21% were fed with bottled milk, or bottled milk and breast feeding combined. The experience of the Bureau of Child Hygiene of the Department of Health of the City of New York has shown that about 80% of the tenement population of that city nurse their babies exclusively, and that four-fifths of the high death-rates of infants from diarrhoeal diseases occur in that group of the infant population that is not breast-fed.

Milk.-Reduction of the infant death-rates in the various communities of the United States has followed very closely the improvement of the milk supply and the tendency towards general pasteurization of milk. The use of raw milk which has not been sufficiently protected in its production, transportation and in its care in the home is undoubtedly one of the most important factors in the causation of high death-rates from diarrhoeal diseases which occur so commonly among artificially fed infants. For these reasons any efforts which are directed towards obtaining a safer supply of milk for children may be classed as measures for the reduction of infant mortality.

Congenital Diseases.-Under the group classified as "congenital diseases" in the following table have been listed all deaths of infants from prematurity, feeble vitality and accidents of labor. This group, furnishing as it does over one-third of the total deaths during the first year of life, is of immense importance. Some cities, notably Boston and New York, have demonstrated that by the employment of public health nurses for the supervision of women

during their period of pregnancy; the observance by these women of all matters pertaining to timely hygiene; proper supervision and care in confinement, including adequate obstetrical and nursing care and provision for nursing supervision of the infant during the first month of life, it is possible to reduce the infant death-rate from congenital causes in the first month of life at least one-half, and, in many instances, two-thirds. Such results seem to show that the present high rates from congenital causes are unnecessary. Percentage for Various Disease Groups, U.S. Birth Registration Area, 1919.

Total

Diarrhoeal Diseases.-The causes of infant mortality from diarrhoeal diseases already have been incidentally discussed. They may be summed up, however, as wrong methods of feeding, lack of hygiene, depressed vitality due to heat and lack of observance of the ordinary methods of hygienic care during infancy.

Respiratory Diseases.-The infant death-rate from respiratory diseases is largely the result of broncho-pneumonia, secondary to measles or whooping-cough. The effect of influenza upon the infant death-rate in the United States has been marked. The results, however, have been due not so much to infant deaths from influenza as to the fact that the mother has died from the disease and the infant, owing to deprivation of breast feeding, has been unable to resist the disease. Experience in public health work has seemed to show that the occurrence of respiratory diseases, in common with the occurrence of contagious diseases in infancy, is due very largely to improper methods of living and is closely allied to lack of ventilation of living-rooms and overcrowding of families.

Contagious Diseases.-This group furnishes only a small proportion of the total infant death-rate. Deaths in this classification are mainly those due to measles and whooping-cough, both of which must be considered as extremely dangerous diseases in infancy.

National Maternity and Infant Welfare Law.-Reference has been made to the efforts of the Children's Bureau of the Department of Labor, Washington, and the various state and other bureaus of child hygiene to reduce the infant death-rates of certain localities. The 67th Congress of the United States passed a bill "to promote the welfare and hygiene of maternity and infancy." This bill, signed by the President, was to be operative during five years. It provides that each state shall receive $10,000 outright, an additional $5,000 provided it appropriates an equal sum, and thereafter a pro rata share of approximately $1,000,000, based upon the population of the various states, provided, however, that the state in question raises an amount equal to this additional appropriation. The money thus given is to be used for the purposes stated in the bill, that is, promotion of maternity and infant welfare work. The general central administration of the act is to be carried out by the committee, composed of the surgeon-general of the U.S. Public Health Service, the U.S. Commissioner of Education and the chief of the Children's Bureau of the U.S. Department of Labor, the chief of the Children's Bureau being the executive officer. The purpose of this bill is to reduce the maternity and infant death-rates by helping the states to establish work of their own for the purpose.

BIBLIOGRAPHY.-Birth Statistics for the Registration Area of the United States, 1919, Bureau of the Census, Department of Commerce, Washington, D.C.; W. H. Park, Public Health and Hygiene; Statistical Report of Infant Mortality for 1920, American Child Hygiene Assn.; Wm. H. Guilfoy, M.D., The Influence of Nationality upon the Mortality of a Community; Physicians' Pocket Reference to the International List of Causes of Death, Bureau of the Census, Department of Commerce, Washington, D.C. (S. J. B. *)

INFANTRY (see 14.517).—To appreciate the lessons learned from the experience of infantry in the World War in relation to the past as well as to the future, it is necessary to emphasize one particular aspect of infantry evolution-the gradual decrease in size of the unit which one man can command. It is desirable also to visualize what "command" really implies. A corporal is said to "command "the squad of recruits which he is training on a barrack square; he does it by shouting words of command to them. Marshal Foch also "commanded " the Allied armies in western Europe in 1918; he did it, however, without raising his voice above its usual pitch. Between the Marshal and the corporal were a host of intermediate commanders of every sort, kind and description, but we are concerned here mainly with infantry commanders and especially with those in the junior ranks. For theirs is the hardest task in a battle, and it is upon them that success depends. The wisest plans, the most thorough prepara

tions, the most brilliant guidance avail nothing unless the fight is won by the fighters-by the platoons." The minds of superior officers therefore are devoted-especially in peace-time-to hard thinking on the problem, of what they can possibly do or invent to make junior infantry commanders superior to the adversaries whom they are likely to meet in action. A general's command implies much forethought as well as some experience in its holder, and thus his "command" again has a different meaning. He is a trainer, and it is with respect to this part of "command" that we shall chiefly be concerned.

If we turn to the past for a moment, we find that the Greeks invented and trained their phalanx and the Romans their legion, and with these two systems the infantry arm dominated the known world for several centuries. Each of these tactical formations was based upon a most precise drill, executed almost daily by junior commanders. Moreover this drill was in each case suited to the age and the esprit de corps of the period. In battle the voice of the infantry superior could be heard and was instantly obeyed, both in the phalanx and the legion. Then ensued the Dark Ages and comparative chaos, which was dominated by feudal horse soldiers, until archery made infantry again supreme. Precise drill was at the root of the success of the archers, and fire orders were strictly enforced. If we take Crécy (1346) as an example, we find that Edward III. initiated fire orders himself, though he left the command of the front-line to his son, the youthful Prince of Wales in charge of the archers. A careful survey of the ground at Crécy from the commanding position of its windmill, in which King Edward was posted, enables one to see why it was possible for him to issue fire orders and instructions to the archers posted below him. They were only a few hundred yards from him, but he could see better than they could when the crowd of French cavaliers would offer the best target to the British longbowmen. These bowmen had under several reigns been disciplined and drilled with precision in the use of their weapons, and that is why they defeated the gallant but undisciplined mob of horsemen who attacked them without method at Crécy. It is also obvious that the diminutive size of that battlefield enabled the commander-in-chief, posted behind his reserve, to initiate fire orders and see their effect-in fact, he performed duties which now appertain to platoon commanders. Moreover, the size of the whole battle-field corresponded with a sector allotted to one battalion, or at most to two, in 1918. Thus the process of devolution of the physical command occupied some five centuries, chiefly because it took all that time to alter infantry armament from bows to Lewis guns; partly also because each generation of professional soldiers clung with punctilious tenacity to the admirable drill of a previous age. Similar tenacity is visible to-day, but changes are in the air. Frederick the Great (1740) attained parade-ground precision even during the shock of encounter, and won his battles by means of remorseless drill, stepping to music and machine-like fire tactics with inaccurate muskets. Such is the force of tradition in any army that in 1914, German companies in Flanders were trying to copy Frederick's tactics with the aid of song instead of a band to inspire their Parademarsch within close range of British infantry. But their opponent's rifles were accurate in 1914.

Frederick, however, did not teach one system on the barrack square and then practise a totally different one on active service; but that is what the British infantry did before the S. African War (1899) and what some soldiers would like it to do again. Their line of reasoning is that, as every war alters tactical formations, it is not of much avail to learn in peace-time tactics which will assuredly be discarded in the next war.

British Infantry in 1913.-In Oct. 1913 the British infantry underwent a drastic change, in spite of much opposition. The old "Eight Company " system was abolished, and the continental system was adopted of dividing the battalion into four companies, each 200 strong. This change gave a peace strength of about 100 men per company actually available for training. This admirable reform, which fortunately was accomplished on the verge of the World War, was effected by amalgamating two of the old companies to form one of the new and enlarged compa