It is clear from the foregoing review that any further contributions to the theory of population must come from the side of statistics; that only by careful statistical investigation can the laws which govern the increase or the decrease of population be determined. The true method has evidently been applied in the exhaustive studies of M. Levasseur in France. It should also be applied to the statistical data furnished by other countries, but especially should these investigations be made in the United States. There are presented here contrasts of geography, race, nationality, of industrial and social conditions not to be found in any other part of the world, and they are on such a scale of magnitude as to render them peculiarly well adapted to statistical research. Few people realize the wealth of material contained in our census and other statistical reports. It has frequently been used to show detached facts or to illustrate special topics, but not often to throw fresh light upon economic or sociological theory.

For a complete study of the birth-rate it is obviously necessary that there should be many more comparisons of one group of statistical facts with another than those which have already been made by Dr. Billings and Dr. Edson. Many other conditions indicative of the general advance of civilization and of individual evolution should be compared with the birth-rate. Further statistical research may prove that their theory of the cause of the diminishing birth-rate is insufficient.

The present investigation is a preliminary study of a few of the many facts found in the United States census reports. Its object is to show the relation of the birth-rate in different parts of the United States to certain phenomena which, it is thought, may have some influence upon the number of births. The statistics used are taken from the tenth census reports. The age of these figures is no bar to their use in such an investigation. The relation of connected phenomena to one another will appear in them as clearly as in figures of more

recent origin. The complete vital statistics of the eleventh census are not yet available.

The birth-rates of the white and the colored population are compared, the relation between the birth-rate and the deathrate from nervous diseases is shown for both sexes, and also the relation between the birth-rate and the density of population, the agricultural wealth, the manufactured wealth, and the mortgage indebtedness.

The vital statistics of the tenth census were tabulated according to a different plan from that of all the other statistics of that census. The unit of locality used was not the State or Territory, but the county. As it was, however, a work of too great magnitude to show the relations of each cause of death to the sex, age, etc., in each of the 2605 counties of the United States, and as the numbers for many of the counties would have been too small to permit of any useful deductions, Dr. Billings decided to make the more elaborate compilations for groups of counties within the limits of each State. The selection of the counties that formed these "State groups" was made by Mr. Henry Gannett, the geographer of the census. The groups were selected in most cases according to the topographical features of the State, and evidently could be consolidated by States for comparison with the tables of past or future censuses, with those of the State censuses, or with the other statistics of the same census. They were also consolidated into larger "grand groups," whose boundaries were determined by topographical peculiarities and not by State lines.* There are in the United States twenty-one of these "grand groups," made up from 111 "State groups.” †

*Tenth Census. Vital Statistics I, p. xiv.

+ Grand Group I, the North Atlantic Coast region, includes the following State groups; Maine 1, New Hampshire 1, Massachusetts 1, Rhode Island, Connecticut 1. Grand Group II, the Middle Atlantic Coast region, includes New York 1, New Jersey 1, Maryland 1, Delaware, District of Columbia, Virginia 1.

Grand Group III, the South Atlantic Coast region, includes North Carolina 1, South Carolina 1, Georgia 1.

Grand Group IV, the Gulf Coast region, includes Florida, Alabama 1, Louisiana 1, Mississippi 1, Texas 1.

In the diagrams or charts graphically illustrating the results obtained in the tables, such curves as have heretofore been used to represent sequent phenomena in the same group, and to compare different sets of sequent phenomena, are employed (1) to compare co-existent phenomena in the same group, (2) to compare the same phenomena in different groups, and (3) to compare the relations between the different co-existent phenomena in one group with the relations of those in other groups.*

Grand Group V, the northeastern hills and plateaus, includes Maine 2, New Hampshire 2, Vermont, Massachusetts 2, Connecticut 2, New York 2.

Grand Group VI, the Central Appalachian region, includes New York 3, New Jersey 2, Pennsylvania 1, Maryland 2.

Grand Group VII, the region of the Great Northern Lakes, includes New York 4. Ohio 1, Michigan 1, Indiana 1, Illinois 1, Wisconsin 1.

Grand Group VIII, the Interior Plateau, includes New York 5, Pennsylvania 2, Virginia 2, North Carolina 2.

Grand Group IX, the Southern Central Appalachian region, includes Virginia 3, West Virginia 1, North Carolina 3, South Carolina 2, Kentucky 1, Tennessee I, Georgia 2, Alabama 2.

Grand Group X, the Ohio River belt, includes Ohio 2, Indiana 2, West Virginia 2, Kentucky 2.

Grand Group XI, the Southern Interior Plateau, includes South Carolina 3, Georgia 3, Alabama 3, Mississippi 2, Tennessee 2.

Grand Group XII, the South Mississippi River belt, includes Kentucky 3, Tennessee 2, Mississippi 3, Louisiana 2, Arkansas 1.

Grand Group XIII, the North Mississippi River belt, includes Missouri 1, Iowa 1, Illinois 2, Wisconsin 2, Minnesota 1.

Grand Group XIV, the Southwest Central region, includes Missouri 2, Arkansas 2, Louisiana 3, Texas 2.

Grand Group XV, the Central region, plains and prairies, includes Ohio 3, Kentucky 4. Tennessee 4, Indiana 3.

Grand Group XVI, the Prairie region, includes Missouri 3, Iowa 2, Illinois 3, Kansas 1, Nebraska 1, Wisconsin 3, Minnesota 2, Dakota 1.

Grand Group XVII, the Missouri River belt, includes Missouri 4, Iowa 3, Nebraska 2, Dakota 2.

Grand Group XVIII, the region of the Western Plains, includes Dakota 3, Montana 1, Wyoming 1, Nebraska 3, Kansas 2, Colorado 1, New Mexico 1, Texas 3. Grand Group XIX, the heavily timbered region of the Northwest, includes Michigan 2, Wisconsin 4, Minnesota 3.

Grand Group XX, the Cordilleran region, includes Montana 2, Washington 1, Wyoming 2, Idaho, Oregon 1, Colorado 2, Utah, Nevada, California 1, Arizona, New Mexico 2.

Grand Group XXI, the Pacific Coast region, includes California 2, Oregon 2, Washington 2.

• After this part of the present investigation had been completed, a similar comparison appeared in an article on "The Life and Labour of the People of London," by Charles Booth, Esq., President of the Royal Statistical Society of London. It was delivered November 21, 1893, and was published in the Journal of the Royal

The interconnection of the conditions compared in the present investigation is shown by noting the conditions that cohere and the conditions that are opposed in the same group and in the different groups; that is, by noting in how many groups two given conditions are both above or both below the averages of the same conditions for the United States, and in how many groups these conditions oppose each other, one being above the average for the United States and the other below. If such coherence or opposition is found in a large majority of the groups, some causal relation may evidently be inferred. The curves, of course, in themselves mean nothing; they are simply a means of directing the eye to certain points.

The number of deaths from nervous diseases has been chosen for comparison with the birth-rate, because it is in. general a measure of the degree of civilization. Dr. Edson has clearly shown that the higher the civilization, the greater is the intensity of life, and the heavier is the strain upon the nervous system; consequently, the number of deaths from nervous diseases will be proportionally greater in the more highly civilized countries. Therefore, if civilization checks the birth-rate, as is affirmed by Mr. Spencer and others, we should expect the death-rate from nervous diseases to rise as the birth-rate falls, and vice versa. These statistics of the tenth census have been compiled on the basis of deaths from known causes, instead of on the usual basis of the living population, and therefore any comparison with similar statistics of other countries is impossible.

In this investigation the figures for the grand groups, as given in the tabulations of the tenth census, are used without further computation; but, since only the aggregates are

Statistical Society, December, 1893. In the twenty-seven registration districts of London, Mr. Booth makes a suggestive study of certain conditions that may influence the increase or decrease of population: poverty, crowding, early marriages, surplus of unmarried men, high birth-rate, and high death-rate. In his tables he shows the interconnection of these conditions "by arranging the London registration districts in order of each of these conditions in turn, from maximum to minimum, and by comparing these orders."

Number

given for the State groups, it was necessary to make many new computations. The figures for the States and Territories were obtained by combining those of the State groups.

CHART I.

BIRTH-RATE of white and COLORED IN CERTAIN GRAND GROUPS,

per 1000

The first study is a comparison of the white and the colored birth-rates in the ten grand groups in which the distinction of color is made; namely, in all grand groups in which the colored population forms twenty per cent or more of the total population. The birth-rates are estimated on the basis of the number of women of child-bearing age, that being, as all statisticians agree, a more scientific birth-rate than one estimated on the basis of the total population.