As a result of a test of the completeness of the registration of deaths in Hawaii the territory was admitted to the registration area for deaths for 1917, thus extending beyond the Continental United States the area from which the Bureau of the Census annually collects and publishes mortality statistics. The population and land area of Hawaii are not included in the figures of the above table.

The states in which the registration of deaths is still too unsatisfactory to warrant inclusion in the registration area are: Alabama, Arizona, Arkansas, Delaware, Florida, Georgia, Idaho, Iowa, Mississippi, Nebraska, Nevada, New Mexico, North Dakota, Oklahoma, South Dakota, Texas, West Virginia, and Wyoming. (1918.)

United States registration area for births. A registration area for births was not established until 1915. For this year the Bureau of the Census published its first annual report of birth statistics based on registration records. The birth statistics published in connection with the regular decennial reports from 1850 to 1900 inclusive were based on enumerator's returns.

The registration area in 1915 included only ten statesMaine, New Hampshire, Vermont, Massachusetts, Rhode Island, Connecticut, New York, Pennsylvania, Michigan, Minnesota, and the District of Columbia. In these states the registration of births is believed to include upwards of ninety per cent of the actual numbers. This registration area includes only 10 per cent of the area and 31 per cent of the population of the country. In spite of this unfavorable showing a beginning has been made, and inasmuch as the standard birth certificate has been adopted for 85 per cent of the population and as public sentiment in regard to the importance of vital statistics is rapidly gaining ground, it is likely that the registration area for births will rapidly extend. No state is admitted until the accuracy of its records have been submitted to test.

In 1916 Maryland was added. In 1917, Virginia, North Carolina, Kentucky, Indiana, Ohio, Wisconsin, Washington, Utah and Kansas were added, bringing the population included up to 53.1 per cent.

Need of national statistics. More and more it becomes obvious that there is need of a national system of keeping records of vital statistics, with uniform state laws, and with proper provision for the local use of the data registered. The excellent work done by the Bureau of the Census has done much to emphasize this need. Likewise interstate barriers must be broken down in the interest of suppressing diseases dangerous to the public health. The

U. S. Public Health Service keeps a record of cases of diseases from data furnished by the states and publishes the same in its weekly Public Health Reports. This is only a part of what is needed. If the time ever comes when the United States establishes a real National Health Department the maintenance of an adequate system of vital records will be one of supreme importance.

EXERCISES AND QUESTIONS

1. Compare the methods of numeration used in taking the U. S. census of 1910, with those used in 1900, 1890, 1880, etc.

2. How do these methods compare with those used in England, France, Sweden?

3. Would there be any advantage in making the census date, as of January first "?

4. What advantages would come from the adoption of a uniform census date for the entire world?

5. How accurately is the population of China known?

6. To what extent is the keeping of accurate census records and records of vital statistics an index of national progress?

7. How can improvements be made in ascertaining the facts concerning morbidity?

CHAPTER V

POPULATION

Estimation of population. It is only for the census years that populations can be known with certainty. For the intercensal years, the years between two censuses, it is necessary to depend upon estimates. This is also the case for the postcensal years, namely, the years following the last census. These estimates are only approximately true, a fact which must not be forgotten, but they are sufficiently near the truth for many practical uses.

Estimations of population may be made in various ways. The natural growth of population is like that of money at compound interest except that the interest is being added constantly instead of semi-annually or quarterly. Mathematicians call this geometrical progression. With a given constant rate of interest money in the bank increases more and more each year. It is the same with population. In geometrical progression the basis of our population estimates is the annual rate of increase. When dealing with very large populations, and especially when dealing with populations not influenced by emigration or immigration, this method is the most accurate one to use. It has several practical disadvantages, however, and in the present shifting condition of the world's population there are not many places where the natural growth of population is the only factor to be considered.

A simpler method is that of arithmetical progression, which assumes a constant annual increment between two