The most obvious inference from the preceding table is that the amount and proportion of enforced idleness are very much smaller in summer (second and third quarters, April 1 to September 30) than in the winter quarters (first and fourth.) As might be expected, the inclemencies of the weather cause much more idleness in December and March than in June and September, thus:

The weather conditions accounted for 46 and 15.5 per cent of all the idleness at the end of March, 1900 and 1901, respectively, and only 0.5 and 2.2 per cent at the end of September in the same years. This cause really affects but two groups of trades, the building trades and the lake trades (grain handlers, marine engineers and firemen).

Trade disputes, and personal causes, like sickness, accident and old age, account for a good deal of the reported idleness; but the principal cause in every instance is inability to find employment. It is, then, lack of work on the part of working. men able and willing to work that creates the problem of the unemployed.

In order to compare 1901 with previous years, it will be convenient to obtain a yearly average. In this case the simple arithmetical average of the percentages will be found nearly as accurate as the geometrical average of the basic figures and is therefore to be preferred. The results are as follows:

The averages for 1901 are for three quarters, as the fourth quarter is not included in the present official year; if the fourth quarter were included the estimated average for 1901 would be those stated by the bracketed figures. In either case 1901 appears to be a more favorable year for employment than any one of the four preceding years, with the possible exception of 1899.

These results are confirmed by more complicated calculations based on the number of days worked by all members of labor organizations. Thus in the third quarter of 1901, 8,341 members were continuously idle and 258,807 members were employed one or more days. The number of days worked by each member being reported, it is a matter of simple multiplication to find the aggregate number of days worked by all members. For the third quarter of 1901 that aggregate was 18,125,904 days. But if the 267,148 members had been fully employed during the quarter they would have accomplished 20,570,396 days' work. They therefore lost 2,444,492 days or 11.9 per cent of full time. This is as accurate a statement of the actual amount of idleness in, this period as it is possible to obtain by any practicable method. In the first quarter of the year the amount of idleness was of course larger; the average for the two quarters was 17.2 per cent. Following this method of calculation for the preceding years, we reach the following results:

TABLE 13.

PERCENTAGE OF TIME LOST BY UNION MEMBERS IN THE FIRST AND THIRD QUARTERS, 1897-1901.

It appears therefore that the proportion of members continuously idle varies in different years in about the same degree as the aggregate proportion of time lost, which depends on such complex calculations that it cannot be so freely used. We may therefore return to the consideration of Table 12.

The low percentage of unemployment in 1899 was due to exceptionally favorable conditions that existed in the summer of that year, especially in the building and clothing trades, which comprise nearly one-half of all the union members in the state and therefore exercise a decisive influence upon the totals. Thus during the second and third quarters of 1899 only .5 per cent and .3 per cent, respectively, of the members of clothing trades organizations were unemployed as compared with 4.5 per cent and 5.7 per cent in the corresponding quarters of 1901 (and 15.6 and 10.6 per cent in 1900). But taking the two quarters together, the building trades exhibited more activity in 1901 than in 1899, and hence the very favorable figures for the aggregate of all trades in the present year. The percentages of unemployment in each are given in Table 14.

The causes of unemployment at the end of March and September are stated in the tables and have already been summarized in Table 11. The following table (15) gives the causes of the idleness in each industry, or group of trades, at the end of September; in most industries "no work" or "slack trade" explains most of the idleness, but in the clothing trades strikes and lockcuts figure prominently, in the tobacco trades sickness, and in public employment all the idleness is due to sickness, accident or old age.

Tables IV and VI show the number and percentage of unionists idle in the various towns and cities of the State. In some of the smaller places, with only one or two organizations, the percentages were often very large-sometimes rising, for example, to 100 per cent. It will be found upon investigation, however, that such figures are chiefly due to the seasonal cessation from

work of artisans in the building trades. The following table shows the degree of unemployment in the seven principal cities:

In the winter quarters the smaller cities have more idleness than the metropolis, but less in the summer. The explanation is the distribution of industries-the dock workers in Buffalo, the clothing trades in New York City, etc.

III. Duration of Employment.

[Summary tables N-O in Appendix III, detailed table VIII in Appendix IV.] As observed in one of the preceding paragraphs under "Unemployment," all trade unionists who are employed even for a single day are counted among those at work. Hence, the tables which show the duration of employment indicate a large variation in the number of days worked each quarter. The following table shows for the aggregate membership this variation in the amount of employment, which is indicated for groups of trades in the summary tables N-O and for the individual trades in the detailed Table VIII.

TABLE 17.

MEMBERS OF LABOR UNIONS CLASSIFIED ACCORDING TO NUMBER OF DAYS OF EMPLOYMENT.