the preceding year, 19.61, and compares favorably with the average for the past decade, 19.92. The improvement in the death rate since 1904 is due wholly to a diminution in the death rate of whites, which fell from 16.08 to 15.16, and which is only 94 per cent of the average annual death rate for the ten-year period. The death rate for the colored population rose from 27.92 to 28.81, the average for the past ten years being 28.52.

Estimated population, deaths, and death rates in the District of Columbia, by race and by calendar years, from 1896 to 1905, inclusive.

NOTE. The death rates given in this table differ in some particulars from the figures given in the corresponding table for several years past. This is due to the fact that the estimated population on which these death rates are calculated is based on the results of the recent police censuses and not upon the results of the Federal census of 1900.

MORTALITY BY SEX.

The following statement shows the death rates for males and for females during 1905. No corresponding figures for recent years are available for purposes of comparison, the number of each sex in the population for such years being unknown.

Deaths and relative death rates of males aud females during the calendar year 1905.

The death rate for males, as appears from the foregoing statement, is higher than that for females, whether the population be considered as a whole or the white and the colored elements be considered separately. The ratio between the white and the colored death rates among males is strikingly similar to the corresponding ratio among females, among the former 1 to 1.90, and among the latter 1 to 1.94. The ratio between the death rate for white females and that for white males is 1 to 1.31. The corresponding figures for the colored race is 1 to 1.28.

Of the white population 49 per cent is male, with a death rate of 31 per cent higher than that of the female population. Of the colored population only 46 per cent is male, and the male death rate is only 28 per cent more than the death rate for females of the same race. It is evident, therefore, that the relatively high colored death rate is not due to a preponderance of males among our colored population or to an excessively high death rate among colored females, but that on the other hand the sex composition of the colored population is such as tend to lower the colored death rate.

MORTALITY IN RELATION TO AGE.

Age at death. The average age at death of decedents, both of white and of colored, fell last year; the former about three months, the latter about nineteen months. In the absence, however, of information as to possible variations in the age composition of the population, these figures have but little significance.

Average age at death of decedents, distributed with reference to race and sex, during 1904

and 1905.

Mortality by age periods.-Reference to the following table shows that notwithstanding the increased population there was a decrease in the number of deaths among white people during 1905 as compared with 1904, and that this decrease was most marked during the first five years of life, while among colored people the number of deaths increased, the greatest increase being among the very children just referred to, viz, children less than 5 years old.

Relative mortality in 1904 and 1905, arranged with reference to age.

The following table shows how the death rates during 1905 varied according to age periods. The death rate per thousand per annum for children under 1 year of age can not well be omitted and yet it is published with some reluctance because of the probability of its being used as the basis for unwarrantable conclusions. This death rate is based upon the census returns, which have and can have reference only to the number of persons living on a specified day. When we calculate an annual death rate on the basis of the census returns we assume that such returns represent the average number of persons living in the community during each day of the census period; not the same persons, but the same number of persons. Probably no material error arises under this method in so far as relates to any age period after the first year of life, but during the first year, owing to the relatively small resisting powers in the earlier part of the period and to the relatively large number of individuals who must pass in review during the year, serious error may arise in the interpretation of the results.

Relation of death rates during 1905, for certain age periods, arranged with reference to race.

This can best be shown by undertaking to calculate the death rate per thousand per annum of children, say under 1 day old, on the basis of a census enumeration. The average number of children born in this district daily during 1905, according to recorded birth returns, was 17.57. This is certainly somewhat less than the actual number of births, but if an enumeration had been made on any one day of all children then under one day old the returns would have shown the number of such children to be probably not more than 25. But the mortality record for 1905 shows that during that year 110 children under 1 day of age died. Calculated on the basis of the enumerated population, therefore, the death rate per thousand per annum of children under 1 day old was 4,400, which clearly is not expressive of actual conditions. The fallacy arises from undertaking to calculate an annual death rate of the basis of an enumeration relating to but a single day, when the very existence of the object that forms the basis of the enumeration covers a period less than a year.

On the assumption that a census represents average conditions, for every child under 1 day old enumerated on a single day, at least 365 such children will pass in review during the year, and therefore the number of possible deaths of such children during the year represents approximately the number of children enumerated during any one day multiplied by 365, and the multiplicand represents a fair basis for the calculation of the death rate. But it represents also the number of children born during the year. And therefore the number of children born during the year is commonly recognized as the proper basis for the calculation of the death rate per annum of children under 1 year old. As we must depend for information as to the number of

children born, upon reports of births filed by physicians and midwives, and as probably many births are not reported, even the death rate thus obtained is over rather than under the true death rate.

Understanding the significance of figures relating to infantile mortality, the following data are of interest and importance:

Death rate of children under 1 year of age, during the year 1905, based on the reported births.

1. Mean number of births reported in 1904 and 1905

2. Number of children under 1 year of age dying during 1905

3. Death rate of children under 1 year of age during 1905, based on mean reported births

The death rates for children less than 1 year old, for children from 1 to 4 years, inclusive, and for persons from 5 to 20 years old, inclusive, as appears from the statements on pages 11 and 16, are in each class disproportionately high among the colored people as compared with corresponding death rates among white people. The ratio between the death rates for whites and for colored people more than 20 years old, while largely in favor of the whites, is much less so than the death rates for earlier age periods. The relatively high death rates during the earlier years of life is a factor of more or less importance in causing the high general death rate of the colored race, especially in view of the fact that the colored population is made up more largely of infants, children, and youths than is the white population, as is shown by the following statement. This, however, does not explain why the colored death rate is high, but only raises the further inquiry as to why the colored death rate during infancy, childhood, and adolescence should be so much higher relatively than the white death rates during corresponding age periods, while the ratio in adult life is so much less.

Percentage distribution of population by age periods and by race, as shown by the police census of April 12, 1905.

A fair conclusion, it is believed, from the age distribution of the mortality among colored people, is that the character of labor performed by colored people is not the chief factor, directly at least, in causing the high death rate of the race as a whole, for relatively the worst death rates do not occur during the working years of life. Neither is it to be believed from the figures now before us that vice and excesses of various kinds are responsible for the high colored death rate, since such factors do not operate to their fullest extent during infancy, childhood, and early adolescence, and yet during these very

age periods we find the relative death rates higher than at a later period. Possibly the relatively lowering of the death rate after the twentieth year may be due to the elimination of the weak of the race during the earlier period, but it is not apparent why the forces tending to the survival of the fittest should not continue to operate after the twentieth year, nor why the relative effect of such force should be distributed differently among whites and colored.

Infant mortality.-The number of deaths among children under 1 year of age and among those under 4 years of age are shown in the following statement, in their relation to race, reported births, and total population. There was an absolute decrease in the number of deaths among white children notwithstanding a considerable increase in the population. Among colored children there was an increase out of all proportion to the increase in the colored population. As a net result, there was a slight increase in deaths among children under 1 year of age and a slight increase among those under 5 years of age, taking the population as a whole.

Variations in the number of deaths among children under 1 year of age and among children under 5 years of age from 1896 to 1905, inclusive, with data as to population and reported births.

Deaths among children less than 2 years of age due to diarrheal diseases followed the same course as the general juvenile mortality. Among whites there was a considerable improvement and among colored children the death rate increased, and as a consequence the number of deaths of this class increased somewhat, but not in proportion to the increase in the population. Figures covering the past five years appear in the statement below:

Deaths and death rates from diarrheal diseases and inflammation of the bowels among children under 2 years of age from 1900 to 1905, inclusive.