These are crude rates, relating all deaths to total population. Inasmuch as the populations in these two areas differ in their age, sex, and racial composition, crude death rates, taking no account of these important factors in mortality, do not necessarily represent the comparative healthfulness of the two areas.

The following table shows the movement of the crude death rate in these areas since 1898.

Death rate by districts. Table 12, pages 413, 115, shows the death rate in the chief subdivisions of the State and in each city for 1917 and the five years' period 1912–1916.

Deaths by months. A study has been made of deaths in New York State by month of occurrence. The thirty years 1888 to 1917 have been divided into six periods of five years each. The average daily number of deaths occurring in each of the 12 months during each of the six five-year periods has been determined, and stated in ratio to the average daily number for the year during each period, taken as 100.

This method eliminates the inequality in length of the months and the comparison is probably not invalidated by the improvement in promptness and completeness of death registration during the last thirty years. It is likely that deficiencies in death registration were more or less equally distributed throughout the year; hence, the average daily number of deaths in each month can be related to the average daily number for the year, taken as unity, and the ratios for the same month in different groups of years directly compared.

One possible disadvantage of this method may be noted. The average daily number of deaths for each month being stated in proportion to the average daily number for the whole year, it follows that a reduction in the average daily number of deaths for the year, due to a marked decline at some one season, may result in an apparent increase at some season where the average daily number of deaths may have remained unchanged.

RATIO OF AVERAGE DAILY NUMBER OF DEATHS IN EACH MONTH
DAILY NUMBER FOR YEAR (=100), NEW YORK STATE: 1888-1917, BY 5 YEAR PERIODS

TO AVERAGE
TO

120

110

110

1903-1907 100

90

100

90

80

80

120

120

110

110

RATIO OF AVERAGE DAILY NUMBER OF DEATHS IN EACH MONTH TO AVERAGE DAILY NUMBER FOR YEAR (-100), NEW YORK STATE: 1888-1917, BY 5-YEAR PERIODS.

This table and the graphs facing page 380 are worthy of careful study. It is instructive to follow the movement of each month in successive periods of time and to compare the relative positions of the months within each period.

Some of the more apparent results of such study may be mentioned. (1) The most conspicuous change has been in the position of July, which, during the first two periods examined, was frequently characterized by writers as the month of greatest mortality, but which has since undergone continuous decrease in rank sufficient to make it virtually the month of lowest relative mortality. This decrease-added to, in much less degree, by August has brought the summer mortality below the level for the year in both the last two periods. (2) The position of the months of January, February, March, and April has been above that for the year in every period. (3) The positions relative to the months of April, June, September, November and December during the last three 5-year periods, and of October during all the six periods, have remained practically unchanged.

Death rate by months. For the twenty years 1898 to 1917, it has been possible to recalculate the annual death rate. by months, and to bring these into harmony with the published annual death rates. By this method, the position of any month is not necessarily affected by changes in others.

Chart (A), opposite, shows, in terms of annual rate, the crude death rate per 1,000 population at all ages for each month in each of the years 1898 to 1917; while (B) shows rates by months for each of the four 5-year periods therein.

Virtually the same features are apparent as those already observed by the proportionate method. The elimination of the July "peak," together with a less reduction in the rate for August, has brought the summer rates below the level for the year in both the last two 5-year periods.

Causes of death. The table on page 386 shows for New York State the rates per 100,000 total population at all ages from the important causes of death during 1917 compared with 1916 and the quinquennium 1911-1915.

It should be observed that these rates for specific causes are ratios of deaths to total population at all ages. Unfortunately, it is impossible to present for a series of years rates by specific age periods. Such rates, relating deaths at any age to the lives actually at risk at that age and thus taking account of lives saved at early ages which increase the number of those living at later years of life, would furnish a much more certain measure of comparative mortality. In the absence of such death rates at specific ages, it is uncertain what part of the increased rates from those causes which affect adult and later age is due to changed age composition of the population.

The graphic chart facing page 380 shows the movement of death rates from certain important causes over a period of years. It is so constructed that equal vertical distances represent equal ratios rather than equal absolute differences. The vertical distance from 1 to 2 is the same as that from 10 to 20, or 100 to 200, - each representing a 1:2 ratio. Lines plotted on such a chart show rates of change rather than numerical differences. If two lines change at the same rate, they will be parallel; if their rates of change are unequal, the one with the greater rate of change will be more nearly perpendicular.

From the table and chart it is apparent that the death rates for those diseases against which public health efforts have in the main been directed-the common infectious diseases, typhoid