The average number of days the mines were worked might not be identical with the average number of days of employment for each mine worker. But the average tonnage mined per day worked serves as a measure of the amount of work furnished every miner per day. The average for the nine-year period is a little over three and a half tons per employee per day, the fluctuations from year to year are insignificant. The number of days in operation accordingly represents the number of days of employment. It can be seen from Diagram X. that the curves representing immigration and days of employment run almost parallel. The deviations from that course are slight, and one of them can be accounted for by temporary conditions unrelated to immigration. The increase of the average number of days worked in 1902 was due to the anthracite coal strike, which increased the demand for bituminous coal. In 1903, after the settlement of the strike, the number of days worked again dropped to the level of 1901. Of course, no mathematical accuracy must be expected from these curves. On the one hand, the number of immigrant laborers comprises a great many who found employment in other industries than bituminous coal mines; on the other, the number of days is not a weighted average and has only the value of an approximation. On the whole, however, the tendency of the two curves is unmistakable; the number of days of employment rises and falls as immigration rises and falls.

The statistics of the New York Labor Bureau are collected annually through correspondence with officers of labor unions and show the number of days of employment in organized trades. While these statistics relate primarily to the skilled crafts only, yet indirectly they reflect the conditions in the industrial field as a whole. Nowadays there are few skilled crafts that do not enter as a part into a larger industrial system. Unemployment of the engineer

1908, Table IX; 1909, p. 57. Reports of the Department of Mines of Pennsylvania, Part II., for the years 1901 to 1909.

or fireman means unemployment for a number of factory hands.

In Diagram XI. days of employment are plotted along with the number of immigrants, exclusive of dependents,' who gave New York as their destination on landing.2 The two upper curves represent the average number of days of employment during the first and the third quarter of every year from 1897 to 1909,3 the lowest heavy line represents immigration of breadwinners. Contrary to the general assumption, the rise of the immigration curve is not followed by a decline of the curves representing duration of employment. On the whole, the three curves move in a uniform direction; the number of days of work increases as immigration increases, and declines as immigration declines. In the fall of 1900 (a presidential year), there was less work on an average than in the fall of 1899, but the middle curve shows that conditions improved in the spring of 1901. These fluctuations were reflected in the total immigration for the fiscal year ended June 30, 1901, which remained almost stationary. A divergence between the employment and immigration curves strikes the eye in 1907. The spring of that year was marked by a decline of employment compared with the preceding year, and the opportunities in the fall showed no progress compared with the previous fall, whereas the immigration curve was still rising. The effects of the curtailment of the days of employment were reflected in the immigration curve next year. On the other hand immigration continued to decline when the condition of the labor market began to improve. The conclusion that can be drawn from this divergence is that it takes some time before the conditions of the labor market are reflected in the immigration movement. As stated in a preceding

"No occupation (mostly women and children)," in Immigration Bureau terminology.

2

* For detailed figures see Appendix, Table XXIII.

3 Annual Report of the New York Bureau of Labor Statistics, 1909, vol. ii, p. xvii., Table 5.

XI. Days of employment in organized trades in the State of New York and number of immigrant breadwinners destined for

New York, 1897-1909

chapter, the arrivals of immigrants in the United States at any given time are the result of preparations made some months before their embarkation on the other side. Viewed as a whole, however, the diagram strongly contradicts the assumption that immigration results in the curtailment of the days of employment. During the ten normal years 1897-1906, which preceded the crisis of 1907, the number of working days increased with the increase of immigration. It could not have been a fortuitous coincidence. No one claims that the arrival of the immigrants was the cause of the increase of the per capita share of work. By the method of exclusion there is room for no other inference than that immigration has merely responded to the increased demand for labor.

The preceding analysis may be summed up in the following proposition:

Unemployment and immigration are the effects of economic forces working in opposite directions; that which produces business expansion reduces unemployment and attracts immigration; that which produces business depression increases unemployment and reduces immigration.

Yet it may be said that while immigration is not a contributory cause of unemployment, restriction of immigration would nevertheless reduce unemployment. An answer to this argument is furnished by the example of Australia, where immigration does not keep up with emigration, and yet unemployment is an ever-present problem, precisely as in the United States. Australia is a new country with abundant natural resources. Its area is as great as that of the continental United States (exclusive of Alaska), while its population at the census of 1906 was a million short of the United States figure for 1800. The Australian statistics of unemployment essentially differ from ours. The XII. Census counted all breadwinners who were idle at any time during the twelve months preceding the date of enumeration. The statistics of the New York Bureau of Labor comprise all wage-earners who were unemployed during

ΤΟ

the first or the third quarter of the year. The Australian statistics, on the other hand, give the number unemployed on the date of enumeration. A comparison of the Australian ratio of unemployment with the New York ratio must therefore be favorable to Australia and unfavorable to New York. Still the comparison is highly instructive. The Australian ratio in 1901 varied from 3.96 per cent for South Australia to 6.73 per cent for New South Wales.' In the State of New York the total amount of unemployment for the three summer months, July, August, and September, fluctuated during the years 1897-1907 between 1.9 per cent and 6.5 per cent. It thus appears that Australia with an excess of emigration over immigration is suffering from unemployment at least as much as the State of New York, which is teeming with immigrants. It is evident that unemployment is created by the modern organization of industry even in the absence of all immigration.

Unemployment not being the result of overpopulation, it necessarily follows that limitation of the number of wageearners can promise no relief against unemployment. To be effective, any proposed remedy must attack the problem of unemployment, not collaterally, through restriction of immigration, but directly.

It is interesting to note the remedy which economic necessity has suggested in one great seasonal industry, viz., in coal mining. More coal is mined in the fall and winter than in the spring and summer. The operator who mines for the general market must have more men at work in the winter than in the summer. The mines being usually located at a distance from the large cities, the operator cannot promptly secure the necessary supply of labor, as he could do in New York, Chicago, or any other great city.

Victor S. Clark: "Labor Conditions in Australia," Bulletin of the Bureau of Labor, No. 56, p. 180.

2 Annual Report of the New York Bureau of Labor Statistics, 1909, vol. ii., p. xvii., Table 5.