oners and juvenile delinquents in institutions." To indicate the sense in which the Census Bureau uses certain terms which are possible of misinterpretation, attention is called to the following statement of the kinds of offenses census agents were instructed to include under these terms:

Assault includes those offenses designated "assault," "assault and battery," criminal assault," "felonious assault," "indecent assault," "assault on police," "mayhem," "assault with intent to murder," and all other assaults except assaults to rob or rape.

All other offenses against the person include all offenses other than those specifically described, such as "abduction," "abortion,' "abortion," "cruelty to children," "false imprisonment," "peonage," etc.

Larceny includes all offenses designated as "larceny," "larceny from the person" (as distinguished from robbery), larceny from buildings," "larceny as bailee," "attempt to commit larceny," "pocket picking," "embezzlement," "buying and receiving stolen goods," "obtaining money or property under false pretenses,' and kindred offenses which involve the unlawful conversion of, or attempt to convert, the property of another to the use of the accused.

All other offenses against property include such offenses as "swindling," "malicious injury to, and destruction of, property," and "unlawful interference with another's business."

Disorderly conduct includes all offenses consisting of disturbance or disorderly conduct not attributable to the use of intoxicating liquors and spoken of in local records as "disorderly conduct," "idle and disorderly," "disturbing the peace," "disturbing religious meetings," "disturbing schools," "disturbing public meetings," "common brawlers," "carrying concealed weapons," "using profane, obscene, or abusive language," and all other offenses of a like nature. In a few cities arrests of drunken persons are made on the ground of "disorderly conduct" and so appear in the statistics; but a distinction between the two offenses seems desirable.

All other offenses against society include such offenses as "perjury," "bribery," "malfeasance in office," "extortion," "violation of election laws," "concealment of crime," "conspiracy," "rescuing prisoner," "suppressing evidence," "resisting officer," "violation of Sunday laws," "violation of fish and game laws,' "violation of license laws," "nonsupport of family," and offenses of like character.

The relative rank of the individual cities in regard to the total number of arrests reported is of chief importance when considered in connection with the population, which is discussed in the text for Table 48. Even when so considered, the total number of arrests made is rather a reflection of the varying local laws and conditions and of the activity of the police department than a criterion of morals.

An increase of 50 per cent or over in the total number of arrests reported in 1907 as compared with 1905 is shown for each of the following cities: Pittsburg, Pa.; Indianapolis, Ind.; Seattle, Wash.; Albany, N. Y.; Oakland, Cal.; Manchester, N. H.; San Antonio, Tex.; and Tacoma, Wash. Of the 31 cities which exhibit decreases in the total number of arrests in 1907 as compared with 1905, those reporting the most noteworthy decreases are shown in the following comparative. statement, the cities being arranged in the order of per cent of decrease:

TABLE XXX.-Total arrests in specified cities showing decreases in number of arrests: 1907 and 1905.

The decrease shown above for Chicago is due to the establishment of the municipal court, before which cases are now directly brought. That noted in Covington, Ky., is caused chiefly by a change in the method of keeping the police records-cases known in police parlance as "safe-keeping" being now omitted altogether from the report. It is worthy of mention that in each of these cities except Covington the decrease in total arrests is coincident with a decrease in arrests for drunkenness. Table XXXI shows, for the 12 cities having the largest decreases in the percentage of arrests for drunkenness, the total number of such arrests for 1907 and for 1905 and the per cent of decrease, the cities being arranged in the order of the percentages:

TABLE XXXI.—Arrests for drunkenness in specified cities showing decreases in number of arrests on this ground: 1907 and 1905.

Of the total number of arrests, 35.2 per cent are charged to drunkenness, 19.2 per cent to miscellaneous offenses against society, and 18.3 per cent to disorderly conduct, none of the other offenses specified in the table contributing a proportion as high as 7 per cent. There is but little variation in the relative importance of arrests for offenses against the person and offenses against property for the several groups of cities and for all cities combined. Of the arrests for offenses against society, however, those for drunkenness and disorderly conduct show a number of differences in relative importance, due largely to varying methods of recording arrests for drunkenness, disorderly conduct, and offenses against chastity, as explained in the text for Table 49, which deals exclusively with the arrests of females.

TABLE 48.

Number of arrests per 10,000 inhabitants and per policeman. This table, which is based upon Table 47, shows the average number of arrests per 10,000 inhabitants, and the average number per policeman, thus furnishing the data for additional comparisons. The number of arrests per policeman indicates to a certain extent the activity of the police department, but correct conclusions can be drawn from these averages only after a study of local laws and conditions. The calculations are based, first, upon the entire active numerical strength of the police department, including officers, detectives, and patrolmen, and, sec

ond, upon the number of patrolmen on beats and posts, i. e., those men who make the greater part of all

arrests.

The exhibit of the number of arrests in proportion to population discloses the fact that the average for "all offenses" for Group IV is strikingly high as compared with the corresponding averages for Groups I, II, and III, which are comparatively uniform. The cities of Group IV showing high individual averages are the following in the order named: Birmingham, Ala., Fort Worth, Tex., East St. Louis, Ill., Little Rock, Ark., Jacksonville, Fla., Chattanooga, Tenn., Mobile, Ala., Knoxville, Tenn., and Macon and Augusta, Ga. Of all the cities reported, Birmingham, Ala., in Group IV, shows the highest average, while the averages for Atlanta, Ga., in Group II, Norfolk, Va., and Savannah, Ga., in Group III, and Washington, D. C., and New Orleans, La., in Group I, are also very high. It is a noteworthy fact that the cities named are located in sections of the country where the percentage of colored population is large.

The following comparative statement for the cities of Group I, showing the relative standing of each. city in population and in the proportion of arrests to population, emphasizes the fact that the number of arrests in proportion to population does not increase with the size of a city, for in no instance, save that of Cincinnati, is the standing identical in both particulars:

TABLE XXXIII.-Comparative statement of rank according to population, and according to number of arrests in proportion to population, for cities of Group I: 1907.

the averages for the cities in these divisions extend over a greater range than is found in other divisions. Of the several principal divisions, the figure for the North Atlantic most nearly approaches that for the 158 cities reported for continental United States, these averages being 526.4 and 582.4, respectively.

The most marked variation between subdivisions of the same principal division is that between the Northern South Atlantic and Southern South Atlantic states, the figure for the former being 798, and that for the latter 1,510.9, while the most noteworthy difference between states of the same subdivision is found in the case of Kentucky and Alabama. Each of these states had 3 cities with a population of over 30,000, but the average for Kentucky is only 317.3, while that for Alabama is 1,781.1-the largest for individual state. The minimum average for any any one state, 300.5, is found in Wisconsin. Of the conditions causing such differences as the foregoing, the diversity in state laws and especially in the methods of their enforcement may be mentioned as important. Examination of Table 48 shows that there is frequently a wide divergence between figures for cities in the same state; in Illinois, for instance, Chicago reported 301, while East St. Louis reported 2,178.7; in Iowa, Davenport reported 540.4, and Des Moines, 1,010.6; in Massachusetts, Everett reported 184.1, and Boston, 937; and in Minnesota, St. Paul reported 282.1, and Duluth, 660.8. Such variations as these are due partly to transient population, and to strict excise laws in neighboring cities.

TABLE 49.

Arrests of females, classified by offenses.-This table, in connection with Table 47, furnishes data for a comparison of the number of arrests reported for the two sexes. Of the 158 cities for which statistics were secured, 125 reported the sex of the person arrested, but the arrests of females, classified by offense, were reported by only 88 cities.

The greater portion of the arrests of females are found in the subdivision designated "arrests for offenses against society," this general class comprising 87.6 per cent of the entire number of arrests, while "offenses against property" and "offenses against the person" account for 7.3 and 5.1 per cent, respectively. The classification of "offenses against society" is, however, somewhat indefinite, owing to the different methods of reporting adopted by the several cities. In many instances arrests for drunkenness, disorderly conduct, vagrancy, and offenses against chastity are to a certain extent intermingled in the statistics, owing to the practice of entering offenses of the latter class under one of the other three heads referred to, as is done in Albany, N. Y.; and to the even more common practice of entering drunkenness as disorderly conduct, as in Harrisburg, Pa., or vice versa, as in Chicago, Ill.

An important feature of Table 49 is the column. showing the percentage of total arrests, male and female, represented by the arrests of females. These data are, however, subject to limitations due to incomplete reports and varying methods. The percentages for Groups I and IV are larger than those for Groups II and III, but the reasons therefor are not apparent. Of the several cities, Salt Lake City, Utah, shows the highest percentage, 33.7, and Johnstown, Pa., the lowest, seven-tenths of 1. Considered by groups, the extremes are as follows: In Group I, Cincinnati, Ohio, 19.5, and Milwaukee, Wis., 8.3; in Group II, Nashville, Tenn., 20.5, and Grand Rapids, Mich., and Portland, Oreg., each 4; in Group III, Salt Lake City, Utah, 33.7, and Erie, Pa., 3; and in Group IV, Spokane, Wash., 32.8, and Johnstown, Pa., seven-tenths of 1. The large percentages shown for some cities are due in part to the practice of arresting women in houses of ill fame at stated intervals and imposing upon them a fine, which is in effect a license fee.

TABLE 50.

Arrests of children, classified by offenses.-The arrests of children reported in this table, like those of females, presented in Table 49, are included in the total arrests shown in Table 47. The completeness of this exhibit is, however, limited by local conditions, less than onethird of the cities reporting the arrests of children by offenses, while not quite one-half classify arrests by the age of the person arrested. Further, as shown by the footnotes to Table 50, the age limit selected by the several cities as the line of separation between children and adults varies in a number of instances from that fixed by the Census Bureau, namely, 16 years. account of these variations in the age limit, percentages based upon the figures in Table 50 are of little value for comparison, as in most cases the higher percentages are due to the inclusion as "children" of persons over 16-this being strikingly illustrated in the case of some of the cities of Group II, where the designation includes all persons under 20.

In a few of the cities it has been possible to obtain statistics in reference to the sex of the children arrested, the figures for the different cities, arranged in order of size, being as follows: Chicago, Ill., 1,966 males and 19 females, age limit under 17; St. Louis, Mo., 1,749 males and 77 females, age limit not given; Buffalo, N. Y., 1,333 males and 18 females, age under 15; New Orleans, La., 824 males and 119 females, age under 16; Washington, D. C., 1,971 males and 186 females, age under 16; Nashville, Tenn., 2,034 males and 794 females, ages 10 to 20; Duluth, Minn., 70 males and 13 females, age under 17; and South Bend, Ind., 240 males and 35 females, age under 21. In 3 cities-Washington, D. C., Nashville, Tenn., and Mobile, Ala.—a segregation by color was also obtained, the figures being as follows: Washington, D. C., 1,004 white and 1,153 colored, age under 16; Nashville, Tenn., 875 white and 1,953 colored, ages 10 to 20; Mobile, Ala., 645 white and 690 colored, age under 20.

The distribution of the arrests of children by offenses, where obtainable, shows the greater number to have been made for disorderly conduct, followed by arrests for miscellaneous offenses against society, and for larceny. These designations include petty thieving and violations of city ordinances.

TABLE 51.

Juvenile courts and results of trials of juveniles.— The report on official statistics of cities of over 30,000 population in 1905 contained the first census report on the trial of juveniles. The inquiries as originally prepared for both that year and 1907 related to municipal courts only; the published tables, however, include not only all available data on the trial of juveniles in such courts in cities of over 30,000 population, but also, for certain cities, data on the trial of juveniles in state and county courts. The tables for 1905 and 1907 are incomplete, but it was considered best to present all data returned to the Bureau of the Census rather than to restrict the report to the few cities with municipal courts for the trial of juveniles. In some instances the jurisdiction of the state and county courts for the trial of juveniles extends beyond the limits of the city, but this office has no data on the territory within the jurisdiction of the several courts. In using the data in Table 51 in respect to the number of juveniles appearing before the court, it must, therefore, be borne in mind that in some cities the number reported as appearing before the court includes children from territory outside the city.

Table 51 contains the data secured from 69 cities, while for 1905 only 37 cities were reported. Five cities reported in 1905-Cleveland and Cincinnati, Ohio; Jersey City, N. J.; Evansville, Ind.; and Woonsocket, R. I.-are omitted from the 1907 report through lack of information. The following juvenile courts have been established since the report for 1905 was issued: Boston, Mass., September 1, 1906; Wash