the unionized skilled trades, such as glass blowing and building construction, are distinctly seasonal trades, and therefore show a high proportion of unemployment. Then the strikes and stoppages, due to union influence, are responsible for much unemployment. The figures presented in this paper do not furnish any adequate basis for a conclusion as to whether union or non-union workers show the higher percentage of unemployment. All that can be said is that the New York figures cannot be used as strictly corroborative of the figures from other sources.

To make the comparison still more difficult, there is a slight difference in the methods of statement between the New York figures and the United States Census and Commissioner of Labor figures. The New York figures state average unemployment for the year. Thus, if January shows 20 per cent. unemployment and July 10 per cent., the New York method of figuring the unemployment for the year would be to add the 20 per cent. and the 10 per cent. and divide by 2, giving an average unemployment for the year of 15 per cent. In the cases of the Census and the Commissioner's report, however, the unemployment is the total for the year, not the average. Thus in the above illustration, if 20 men in a hundred were unemployed in January and 10 men in a hundred in July, and if the 20 men and the 10 men formed different groups, no member of the group of 20 belonging also to the group of 10, the unemployment would be 30 per cent. for the year. The difference in these two statements is caused by the fact that the New York figures are compiled from the labor union returns,-the question asked is, "What per cent. of your men are unemployed?"while the census and Commissioner of Labor figures are compiled from individual investigation, the question asked is, "Were you unemployed?" The resulting difference is apparent. Aside from these slight differences in method the two groups of figures are fairly comparable. It will be noticed, in the succeeding tables, that the percentage of unemployment given by the various authorities is very similar for similar industries and geographical locations.

IV. The United States Geological Survey issues an annual report in which there is a statement of the number of work days and idle days in the coal-mining industry of the United States.

V. The Illinois (annual) Coal Report contains similar material for Illinois. In addition, detailed figures are given for the individual mines. These coal-mine figures give the unemployment not of men but of an industry, and they, therefore, indicate the probable number of days that a miner can go to work if he wishes to, but they do not, like the other figures dealing with unemployment, take into consideration sickness, accident, or any other cause that might keep a man from working. They represent opportunity to work and not work actually performed.

These five authorities, taken as a whole, are by no means satisfactory, but they represent the total available resources of one who seeks to learn the extent of unemployment in the United States, and an attempt will now be made to analyze and compare them. Unemployment in 1900 is shown in Table II. Four of the five authorities cited contribute to this group of figures, but the census figures do not give the length of time unemployed, as was explained in a previous paragraph. It will be noted from the table that unemployment is most severe among the miners. The Census shows that nearly half of them were unemployed at some time during the year, while the figures from Illinois show 126 idle days, and those from the United States Geological Survey show 96 idle days, out of a possible 306 working days,-in each case about a third of the maximum days during which employment is possible. In the lower part of the table a contrast is presented between the census figures and the New York Bureau figures in three specific industries,printing, tobacco work, and textile work. The census figures cover a much larger number of employees than the New York figures, and in the first two industries show a much higher percentage of unemployment. As previously explained, however, this should be the case because of a difference in method of compilation. When the difference in method is taken into

account, it will appear that the census figures and the New York figures correspond very closely.

Table II shows that, excepting the miners, about one-quarter of those gainfully occupied during 1900 were unemployed. This conclusion is borne out by the census and the New York figures, which show unemployment in "Industry" to the extent of 27 per cent. and 20 per cent. respectively.

The 1902 unemployment figures are presented in order to bring out the relation of the United States Commissioner's report of 1903 to the general problem. This report, as already indicated, is entitled to great consideration, owing to the intensive method of the investigation underlying it. It will be noted that the New York Bureau figures and the coal-mine figures in Table III are almost identical with the same figures in Table II. From this it might fairly be inferred that unemployment was about as extensive in 1902 as in 1900. In view of this stability of unemployment in two cases where comparison is possible and in view of the careful manner in which the United States Commissioner's report was compiled, the unemployment of half of the heads of families coming under the United States Commissioner's investigation is most startling.