these extra ten due to suicide or to Caucasian markmanship?
Very serious, too. is the mistake of studying arrests for homicide, without taking into account the relative efficiency of the local authorities. Mr. (Shipley's failure to realize this is shown, not only by his indiscriminate grouping of cities in all parts of the country, but specifically by his discussion of homicide in Colorado. He compares the small number of arrests for this offense in Denver with the large number of murders in the state, properly enough emphasizing the lawless condition of the mining-camps; but he fails to ask how many Denver murderers elude a comparatively small police force, and escape into the wilderness on the many railroads radiating from the city. The question of relative police efficiency might have been studied from his Table I., giving the number of arrests for homicide in various cities. The population of New York can not be much worse than that of neighboring cities, with their factories, docks and tramp-infested freight-yards. Yet New York reports 13.23 homicide arrests per 100,000, against 9.16 in Newark and only 4.51 in 'anarchist' Paterson, Jersey City and Hoboken also reporting low figures.
Again, the attitude of the community ought to be considered, if arrests for homicide are made the criterion for this form of crime. Is it conceivable that the word 'lynching' should be absent from an inquiry into American murders? Not once does Mr. Shipley allude to the hundreds, nay thousands, who have participated in the lynching of negroes, of horse-thieves and 'rustlers'; were the members of these mobs enumerated, or even the crimes counted singly, how would our statistics look? Let us now examine his study of the relation of the immigrant to homicide. First comes a diagram, Fig. 1, 'showing changing character of immigration into the United States' by three curves, showing the percentage of each year's total immigration, derived from northwestern Europe, southeastern Europe and 'all others.' Below is a homicide curve, which is strikingly parallel to the southeastern Europe curve. But had these curves been drawn to represent the actual numbers of each group of nationalities, instead of their percentage out of a variable whole, the parallelism would have vanished. Furthermore, the southeastern curve represents natives of Italy, Russia, Poland and Austro-Hungary, whereof the Italians form about one third. Another table, Fig. 3, gives the ratio of murderers among 1,000,000 Italians as 50.2; hut that of the Poles and Magyars is shown at very much less than the French, and the large masses of Russian and Galician Jews are bunched with unclassified nationalities in a group rated at a trifle over 1 per 100,000. The peoples representing two thirds of that incriminating curve actually produced less than the average number of murderers. Bearing in mind that the criminal tendency of an immigrant could not exhaust itself in the first year after his arrival, it is evident that the homicide curve, if strongly influenced by the immigration of the last fifteen years, the period in which the Russian and Italian immigration has outnumbered the Germanic, ought to have run steeply upward, in a hyperboloid form, instead of remaining almost horizontal, and showing less of an upward tendency than during the preceding decade.
Further to incriminate the Italians, Fig. 5 arranges the states and territories in the order of homicidal statistics, and gives a graphic representation of the number of natives of northern Europe, as contrasted with a combined group of Italians, Mexicans and Chinese, living in each; the southwestern and Rocky Mountain states show the greatest mortality by violence, and by combining these three, otherwise unrelated, nationalities, their
