From the beginning of the United States Army's death records to the end of the nineteenth century its death rates were greater than those of the males 15 to 34 years of age in the extremely healthy Swedish population at the same date. The death rate in each population group sank with little interruption during these 80 years, but that in the Army much faster than that in Sweden. As a result in the early decades the Army death rate was more than three times that of the Swedish group, at the end of the century the difference. was less than half as great, and for the seven years 1911-1917 the American Army death rate (see Table 2) was less than that of males 15 to 34 years of age in Sweden for any period of which we have record. Apparently the death rate of the American Army now in time of peace is about the same as that of male Swedes 15 to 34 years of age, and less than that of males 20 to 34 years of age living in the registration States.

Table 2 (American troops only) also gives a clue to the number of deaths in 1917 which can fairly be ascribed to the abnormal conditions of that year. The lowest recorded death rate in the Army, including officers and enlisted men and Porto Ricans and Philippine Scouts as well as American troops, was 4.39 per 1,000 in 1910. If this death rate be applied to 678,579, the mean strength of the Army in 1917, it indicates that 2,979 deaths would have occurred had the minimum death rate prevailed. The difference between this computed number and the number of deaths which actually occurred is 1,180, and this is apparently an approximate measure of the number of deaths in 1917 above the normal and so of the loss of life attributable to the unfavorable conditions resulting from the sudden and enormous expansion of the Army between March, 1917, and January, 1918.

2. RATES OF ADMISSION TO SICK REPORT SINCE 1820.

When anyone in the Army is physically unable to report for duty he is enrolled as sick. The decision whether he is or is not sick is made in the first instance and ordinarily by himself in applying for exemption from duty. It is then reviewed by the medical officer, who may decide that he is able to do duty but who probably in most cases accepts the man's judgment. If the man is adjudged sick, he may be allowed to remain in quarters or he may be sent to the hospital. Which course is taken depends upon many considerations. Probably the line between those who are and those who are not admitted to sick report is drawn at various posts and in successive years with greater uniformity than the line between those who are admitted to hospital and those who are treated in quarters. For this reason the number and proportion of those admitted to sick report is a safer index of health than the number and proportion of those admitted to hospital. The unit in this tabulation is the admission not the person admitted, a fact which explains why the rate so often runs above 100 per cent.

Persons admitted to sick report fall into two main classes, those whose condition is due to external agents, producing accidents or wounds, for which the medical and sanitary authorities of the Army ordinarily have no responsibility, and those suffering from various diseases. The proportion of admissions to sick report for disease is probably a better index of health conditions than the proportion of

admissions to sick report for all causes, but in the early annual reports of the Surgeon General the distinction between these two classes was disregarded. The figures of Table 4, therefore, refer to admissions for all causes.

TABLE NO. 4.-Rate of admission to sick report in United States Army (American troops) per 1,000 mean strength by decennial periods, 1820-1909, and single years, 1910-1917.

1 Information for 1833, 1834, and 1836 not obtainable.

* Statistics during the period of the Mexican War, 1847 and 1848, not reliable.
* Statistics during the period of the Civil War, 1862-1866, not reliable.

Table 4 shows that the health of the Army measured by the proportion of sick has been steadily and rapidly improving since the middle of the nineteenth century and thus supports and confirms thẹ conclusion already drawn from Table 1. The improvement measured by the proportion sick is less than the improvement measured by the death rate; the average rate of admission to sick report between 1910 and 1917, inclusive, being a little more than one-fourth of the maximum while the death rate in the same period was only one-seventh of the maximum. In another respect also the results of the two measurements differ; the highest death rate was in the twenties and the highest sickness rate in the forties of the nineteenth century. The most striking coincidence in the two tables is found in the steady and rapid decrease both of sickness and of death in the Army since the sixties, probably due to various sanitary improvements growing out of discoveries in the germ theory of disease.

3. RATES OF DISCHARGE BECAUSE OF DISABILITY SINCE 1820.

A soldier who has been certified by a medical officer to be permanently incapacitated for military service is examined by a board of medical officers and on the basis of its report is or is not discharged because of disability. In the older records of the War Department the distinction between disability due to disease and disability due to external causes was disregarded and for that reason in the following table all discharges for disability are included. But during the five years, 1912-1916, of all discharges for disability nearly nine-tenths were due to disease and only about one-tenth to external causes. So the changes in the total discharges shown by Table 4 would probably be much the same if the discharges due to external causes were excluded.

TABLE NO. 5.-Rate of discharge for permanent disability in United States Army (American troops) per 1,000 mean strength by decennial periods, 1820–1909, and single years, 1910-1917.

1 Information for the four years 1833, 1834, 1836, and 1837 not obtainable.
2 No reliable statistics for the Mexican War period, 1847 and 1848.
3 No reliable statistics for the Civil War period, 1861-1866.

The discharges due to permanent disability remained until after 1890 at about a constant rate or certainly with no tendency to a steady decrease. Examination of the figures for single years' shows that the Spanish-American War was accompanied and followed through several years by a high discharge rate. The rates for 1895-1897 averaged below 10 and were the lowest on record; they were followed by a rate of 11.6 in 1898 and 27.4 in 1899; for eight years thereafter the rate did not fall below 20. It seems likely that many persons were then taken with the Army who under ordinary conditions in time of peace would have been rejected as below the physical standards and that a considerable proportion of them, after being accepted, could not meet the exacting demands of army life. The sharp increase in the proportion of men discharged for disability in 1917 when it was almost double what it had been on the average during the preceding seven years confirms this theory. Probably many men volunteered or were summoned for examination and were passed as physically sound by the various boards who would not have been accepted if the pressure to get men and the little time available for examining each man had not combined to lower the standard actually used by the examining authorities.

4. NONEFFECTIVE RATES SINCE 1870.

The noneffective rate means the average number of men off duty per 1,000 mean strength. It is found by dividing the total time in days within the year of the enlisted men who were off duty because of ill health by the number of days in the year to get the average number of persons off duty on any day and dividing this quotient by the mean strength to get the rate.

Noneffectiveness is of two kinds, one in which the cause is disease, another in which the cause is a wound or other injury. As the duty of the medical force centers about protection from disease, probably the noneffective rate due to disease would be the better form to use as a measure of the health of the Army, but because in earlier reports the two kinds of noneffectiveness were not invariably distinguished they have been treated together in this paragraph. The figures for

1 Report of the Surgeon General, 1913, p. 25.

noneffectives, unlike those previously reviewed, run back only to the years just after the close of the Civil War.

TABLE NO. 6.Noneffective rate in United States Army (American troops) per 1,000 mean strength by decennial periods, 1870-1909, and single years, 1910–1917.

1

Table 6 shows that there was no decrease but perhaps a slight increase in the noneffective rate between 1870 and 1910. An examination of the rates for single years shows the maximum rate (828 per 1,000) was in 1898 and that it fell from that figure almost without interruption until 1914 when it was less than one-third of what it had been 16 years earlier and below the rate of any previous year. We may state the results from this analysis of changes in the health of the Army in the following conclusions:

1. The death rate has been falling from the first and practically without interruption.

2. Sickness measured by the admissions to sick report has been falling since the forties.

3. Permanent disability leading to a discharge from the Army increased sharply at the time of the Spanish-American War.

4. Noneffectiveness leading to temporary exemption from active duty also increased sharply in the same period.

5. The former but not the latter increased notably in 1917 indicating that both of these forms of imperfect health become much more common in the Army in time of war.

All four methods of analysis concur in indicating that during the period between 1910 and the entrance of the United States into the present war the health of the Army was better than at any previous period.

5. MONTHLY VARIATIONS IN THE HEALTH OF THE ARMY IN 1917.

That the results reached by different methods of testing the health of the Army do not agree in detail but do have a general similarity has already been made clear. In dealing with more special problems it seems hardly worth while to use all of these methods and if that is not to be done a selection must be made. On the whole and for the purposes of this report the death rate seems to be the best single test of the Army's health. It is better understood by the public and the sources of error in this field seem less important than in the others. Holding this view the further study of health in the army will stress this method of measurement.

Report of the Surgeon General, 1913, p. 20.

The calendar year 1917 was occupied mainly with preparations for the participation of the United States in the great war which had been forced upon her. The mean strength of the Army in January, 1917, was only 217,272 officers and men all branches included and in March just before the declaration of war it had fallen to 194,165 officers and men, decreasing thus by more than one-tenth in two months as a result mainly of the return of certain mobilized elements of the National Guard to civil life. But by December, 1917, it had risen to 1,538,203 officers and men. That is, in the nine months between March and December the size of the Army was multiplied about eight times.

The year was mainly a period of organization and training. In comparison with the size of the Army the amount of active service was slight and the casualties resulting directly from the war were few. Only one-twentieth of the deaths occurred in Europe and of this small fraction two-thirds were due to disease. The new men brought into the Army were young and had been recently subjected to medical examination and selection. These influences were favorable to a low sickness and death rate. But on the other hand, when large numbers. some of them carrying in their systems the seeds of disease, are taken out of their home life and crowded together in camps and cantonments with the rapidity and under the conditions which prevailed almost of necessity in the last part of 1917, it is hardly possible to prevent epidemics of contagious diseases or to keep the mortality from such diseases during the early months down to the level maintained either before or after the period of rapid expansion.

The health of the Army month by month in 1917 is measured in the following table by showing the death rate for each month of that year: TABLE NO. 7.-Death rate in Army (officers and enlisted men, all stations 1), by months.

1 Total Army, including American and native troops. * Of these 60 were among cas ́s admitted to sick report before 1917 and 9 were deaths among Porto Ricans or Philippine Scouts with date of admission unreported.

For purposes of comparison in default of similar figures for 1916 the corresponding figures for 1915 have been introduced. But the deaths on which these monthly rates for 1917 are based are those which