number in the country. Changes which are quite proportionate in the total population do not produce proportionate results in the percentages. Take a simple example. Suppose in 1890 you have a city population of 50 foreign-born out of a total population of 100, or 50 per cent., and a country population of 50 out of a total population of 500, or 10 per cent. During the decade the foreign-born population remains stationary and the nativeborn population doubles in both city and country, which indicates no change in the city tendency of either the foreign-born or native-born. Then in 1900 you have a foreign-born population of 50 out of 150, or 33.33 per cent. in the city, and 50 out of 950, or 5.26 per cent. in the country. This shows a loss of 16.67 per cent. in the city population and 4.74 per cent. for the country; or if one attempts to use the figures proportionately the foreignborn city population has apparently decreased a third and the country population about a half. Or, following the author's method, the city population in 1890 was 5 times as large, in proportion, as the country population, and in 1900, 6 times as large. But, however the figures are used, the results are absolutely without value, for they all show a change of some kind, whereas by the supposition there has been no change whatever in the city tendencies of either the native-born or the foreign-born. And when the attempt is made to multiply or divide sets of figures obtained in this way it simply complicates the difficulty. It is evident then that while percentages of total population may be used broadly, as in the discussion of point I, to show certain general relations, they are wholly untrustworthy for close comparisons and computations such as these.

The simple and rational method to determine the point in question is to find what per cent. of the total foreign-born population are living in cities at one date and compare it with the per cent. of total foreign-born population living in cities at another date. In 1890, 61.4 per cent. of the foreign-born in the United States were in cities of at least 2,500; in 1900, 66.3 per cent. This shows a change in distribution between country and city, in favor of the city, of 4.9 per cent. of the total. If it is desired to know whether this change is greater or less than the corre

sponding change of the native-born, secure exactly the same data for the native-born and compare the two. In 1890, 31.4 per cent. of the native-born were in cities of at least 2,500, and in 1900, 36.1 per cent., a change in distribution of 4.7 per cent. of the total. Of the total foreign-born population in the United States, 4.9 per cent. more were in the cities in 1900 than in 1890; of the native-born 4.7 per cent. This should be sufficient proof that the foreign-born are going to the cities more than the nativeborn, especially when it is borne in mind that a loss of 4.9 per cent. out of a rural foreign-born population of 38.6 per cent. is much more significant than a loss of 4.7 per cent. out of a rural native-born population of 68.6 per cent.

The figures given in the article, then, can hardly be depended on to show what they seem to show. But even if the figures were accurate, they would not prove the point conclusively, as Prof. Willcox himself recognizes owing to the fact that, as he proceeds to show, the foreign-born population in 1900 had been in the United States a longer time on the average than in 1890, and so a larger per cent. of those who were eventually to reach the country had had time to get there. The year 1890 came at the close of a period of large immigration; the year 1900, at the close of a period of small immigration. This will be made evident by a glance at the following table:

1

1 Report Commissioner-General of Immigration, 1906, pp. 43 and 56. Figures missing for 1896 and 1897.

The eleventh census found a larger number of foreign-born who had been in the United States a short time than did the twelfth census. Hence a comparison of the two years is not conclusive. Another important element of which Prof. Willcox takes no account, barely mentioning it once in the course of his article, is the large number of foreigners who return to their native land in times of industrial depression here. I believe it is Miss Claghorn who has styled the Italians the safety-valve of our labor market. These departures always bear a higher ratio to the arrivals in times of depression than in times of prosperity. A list of the departures is included in the above table. The figures do not go back of 1890, so we can not tell how numerous were the departures in the previous years. For our purposes it seems best to take only the non-cabin passengers as most nearly representing the class we are dealing with. Subtracting the departures from the arrivals, we find a gain of foreign-born of 323,048 in 1890; 41,871 in 1895; 98,442 in 1898, and 183,954 in 1899. If these departures were evenly distributed between city and country dwellers, they would make little difference in the reckoning. But, while there are no figures to prove it, it seems wholly probable that the great majority are from the cities. The foreigner who leaves in a period of industrial depression is the laborer who finds himself out of a job, not the agriculturalist who has established himself on a piece of land, or the farm laborer, the demand for whose labor is not largely dependent on industrial conditions. So this large proportion of departures undoubtedly cut down the foreign city population just previous to the census of 1900. Just what the showing would be if we had the figures for the years immediately preceding 1890 can not be definitely stated. But we are safe in concluding that we should find a much slighter diminution in the foreign city population from this cause than in the last years of the following decade. At any rate it is clear that conclusions based on the census figures for 1890 and 1900 in regard to the foreign-born city population are untrustworthy.

Accordingly the author takes up another line of argument. The last census reported 201,128 foreign-born in the United

States whose residence here was given as less than one year. The reports of the Bureau of Immigration show that 431,501 immigrants were admitted to this country in the twelve months preceding the date of the census. Increasing this number by 14.2 per cent., which the author has decided upon earlier in the paper as the proper proportion to be added to represent the unreported immigration from Canada and Mexico, we have a total of 492,000 immigrants admitted in the year ending June 1, 1900. This number is more than twice as large as that reported by the census. The only reason for the discrepancy which Prof. Willcox offers is that the missing number is included in the million odd foreign-born whose length of residence was not reported. Taking the number which was reported, the author estimates what proportion of them entered in the cities of various sizes and compares it with the number of foreign-born of less than one year's residence who were found in the seaboard cities at the time the census was taken. By this means he shows that at least four-fifths of those arriving in New York City and threefifths of those arriving at other seaboard ports had left the port of arrival; more than a quarter of the total arrivals were found within six months in some city of over 25,000, not a seaboard city; and that more than half of the total arrivals were found in the country districts. This, he claims, proves "with a conclusiveness hitherto unattainable that the congestion of the foreign-born in our large cities, particularly the seaboard cities, is in no sense an evidence that the arrivals linger or stagnate there."

This is perhaps the most direct piece of evidence in the article, but there are two grave faults to be found with it. First, a set of figures which admittedly covers less than half of the units in question and is arranged in groups simply by an estimate, even though a careful one, and which moreover applies to only one fiscal year-such a set of figures is too incomplete and uncertain to serve as the basis for a conclusive proof of anything, particularly of a statement which is contradicted by common observation and by other lines of investigation. Secondly, it seems very probable that the unreported and unreckoned half of the arrivals

for that year were in fact in cities to a much larger extent than in the country. The individuals whom the enumerators had difficulty in finding would not be those in the thinly settled country districts, where their arrival would be noted by their neighbors, but those in the densely crowded sections of the seaboard and other large cities, where they were surrounded by a host of others of the same race.

These figures are based on table 29 of the derivative tables in the 1900 census reports. A little further study of this table reveals a state of affairs quite detrimental to Prof. Willcox's whole line of argument. Figures are given of the number of foreign-born in the cities of at least 25,000, and of the cities under 25,000 and the country districts, arranged according to their length of residence in the United States. Reducing these figures to percentages of the total population of each group, we have the following results:—

Of the foreign-born who have lived in the United States the specified length of time the following per cent. were found in cities of at least 25,000:

Residence in years. Per cent.

...

Less than I

I

2

3

4

5 6-9 10-14 15-19 Over 20 45.7 52.7 56.2 57.4 56.9 59.2 56.6 57.2 50.7 45.4

Thus we see that with increased length of residence in the United States there is a general increase in the proportion living in cities until the group 15-19 is reached, when there is a decided falling off. The evidence of the figures is not absolutely conclusive on account of the large number of foreign-born whose length of residence was unknown, but as far as it goes it is diametrically opposed to the author's whole contention that the foreign-born in cities are merely transients there because they landed there, and that if they are given time enough they will get out. When the length of residence exceeds fifteen years it refers not to the present generation of immigrants, but to the previous one, and the showing in this respect is also quite contrary to Prof. Willcox's argument.

As an example of the location of the recently immigrating races in the United States, the author chooses the Italians.