RURAL.

URBAN.

TABLE 36. Relation of the Death-rate to Density of Population, by Groups of

In the accompanying table the density is expressed by the average number of persons living upon each square mile of area. The same might also be expressed in terms of acres to each person, as in the twenty-third annual report (paper on "Geographical Distribution of Disease").

The density has increased as follows:

NOTE. In the supplement to the twenty-fifth annual report of the Registrar-General of England, Dr. Farr gives a definite statement of the relation which subsists between the density of population and the deathrate, which is briefly as follows, as determined by observation, dividing the population into five groups: :

"1. Where the death-rate was 14, 15 or 16 per 1,000 the density of the population averaged 86 persons per square mile.

2.

Where the death-rate was 17, 18 or 19 per 1,000 the population was 172 per square mile.

3.

Where the death-rate was 20, 21 or 22 per 1,000 the population was 255 per square mile.

4.

Where the death-rate was 23, 24 or 25 per 1,000 the population was 1,128 per square mile.

5.

Where the death-rate was 26 and upwards the density of the population averaged 3,399 persons per square mile."

In 1855 Massachusetts held a position as regards density about midway between the first and second of the foregoing groups (the density then being 136 per square mile), while in 1895 it had advanced to a position between the third and fourth groups (with a density of 300 per square mile). Yet during these four decades the death-rate remained practically the same, 19.4, 19.6, 19.4, 19.6, a fact that may be taken to show that improved sanitary conditions have very nearly counterbalanced the unfavorable influence of increasing density.

Deaths by Ages.

The mortality at different ages constitutes an important index of the sanitary condition of a community and hence requires careful consideration.

"As tests of the sanitary condition, the death-rates of infants under one year and of children under five years of age are more important than the rates at any other groups of ages. Investigation has shown beyond doubt that the ages of young children are very incorrectly returned at the census enumeration, owing to the general and indiscriminate use of two methods for describing age. At the census enumerations the numbers of infants under one year of age are understated." (Memorial volume on "Vital Statistics," London, 1885, page 114.)

The foregoing comment by Dr. Farr applies with equal justice to each of the enumerations of the United States census and to those of Massachusetts. (See report of the Board for 1894, page liv.)

Hence, in the following table the infantile death-rate is not calculated upon the census figures for living children under one year but upon the registered births, in accordance with the method advised by Dr. Chalmers in his "New Life-table of Glasgow," page

14.

No attempt has been made to carry this estimate beyond the fifth year of life, since the disturbing effect of migration seriously affects the accuracy of any calculation upon this basis when applied to a rapidly increasing New England community. The newly arrived foreign-born infants in the first year of life constitute less than 1 per cent. of the children of that age, and the foreign-born children among those under five years of age constitute less than 4 per cent. of all children of that age; but the foreign born among the population from thirty to fifty years of age constitute fully 40 per cent. of the population of those ages. It is chiefly for this reason that the construction of an accurate life-table from the population and mortality tables of an American State is rendered practically impossible. TABLE 38.-Infant Mortality, Massachusetts, 1856-95, Forty Years.

66

Strictly speaking, the death-rate of infants under one year is obtained by comparing the deaths of such infants occurring in a year with the mean number of infants under one living throughout a year, and this number must " lie between the annual number of births and that number diminished by the deaths under one. It would be nearer the latter than the former number on account of the excess of deaths in the first months of life." (Dr. Farr.)

For the sake of uniformity in comparison with the statistics of other states and countries, the deaths under one are here compared with the births. In the foregoing table the births in the first line (1856) are those which occurred between July 1, 1855, and June 30, 1856, inclusive, and so on through the table, the births in the last line being for the year ending June 30, 1895.

The deaths under one in the same table are those of the calendar years ending Dec. 31, 1856, 1857, etc. The births in this whole period were 1,762,874, and the deaths under one year were 276,260, which is equivalent to a death-rate of 156.7 per 1,000 living births, as a mean of the whole period.

Dividing the whole time into ten and twenty year periods, the infantile death-rate of the first twenty years was 150.3 and that of the second twenty years was 161.2 per 1,000 births. The difference in the rates of the ten-year periods is much greater, chiefly on account of the disturbing effect of the epidemic years 1872 and 1873, when cholera infantum prevailed to an unusual degree. The infantile deathrates of the four ten-year periods were, respectively, 134.2, 164.2, 157.8, 163.6. By this it appears that the infantile death-rate has increased considerably when we compare the first twenty years with the last twenty, and especially when the first and last ten years are compared. If the figures for the two years 1872 and 1873 were to be omitted from the table, the infantile death-rate of the remaining eight years in the ten-year period 1866-75 would be reduced to

156.5.

The death-rate of infants under one has diminished with considerable uniformity during the last five years (1891-95) from 169.0 per 1,000 births in 1891 to 158.2 in 1895.

Infant Mortality of Cities.

For the purpose of presenting the infant mortality of the cities of Massachusetts, a ten-year period has been selected, and since only eleven out of the thirty-two cities embraced in this summary existed as incorporated cities at the beginning of the period under con sideration, the ten years 1881-90 are selected for examination. The later ten-year period 1886-95 would have been selected, but unfortunately the presentation of the deaths by ages in the different cities in the State appears to have been abandoned in the registration reports since 1890.

In the following table the cities are arranged with reference to their comparative rank so far as infant mortality is concerned.

Of the first seven cities in this table all except Boston are cities in which the cotton and woollen industries largely predominate. In the first three the tenement-house population is also unusually large. In the twenty-third report of the Board, page 836, attention was also called to the fact that the mortality from cholera infantum, which is one of the chief factors of the death-rate under one year, bears a direct relation to the ratio of married women employed away from their homes, and this ratio reaches its maximum in Fall River, where this class constitutes 3.1 per cent. of the total population.

There are also several smaller towns, not included in this list (Webster, Ware, Adams and Millbury), in which the same class of manufactures is conducted, and which have a considerable tenementhouse population, with a high rate of infant mortality.

The cities of Haverhill, Marlborough, Brockton and Lynn, each having an infantile death-rate below that of the State at large, are all shoe manufacturing towns, in which the tenement-house population is comparatively small.

Near the bottom of the list are the cities of Malden, Everett, Medford and Newton, in which the residential population largely predominates. The manufacturing population constitutes but a small fraction, and the infantile death-rate is much less than that of the State.

The death-rate of infants for the State at large was 160 per 1,000 births, while that of the combined urban population was 175 per 1,000 and that of the rural population was 129.