arbitrary. Suppose, the working force of a mill in 1882 consisted of 1000 men, of whom 750 were skilled mechanics whose wages averaged $3.50 a day, and 250 were unskilled laborers hired at $1.25 a day. The average wage for all mill workers was, accordingly, $3.00 per day. Suppose, further, that in the thirty years that have elapsed since, the business of the mill has grown and two new departments have been added, with 1000 men in each. But owing to the installation of new machinery the same 750 skilled mechanics have been distributed over the three plants, and the additional force of 2000 men consists solely of unskilled laborers. Suppose, the wages of the skilled mechanics have been raised from $3.50 to an average of $5.00 per day, and the wages of unskilled laborers from $1.25 to an average of $2.00 per day. The average for the three plants, however, would be $2.75 per day, i. e., twenty-five cents less than thirty years ago, notwithstanding the substantial gain in the wages of all employees. The same defect is inherent in the latest refinements of the average, the "median, the "quartile," the "decile," etc.

Moreover, our wage statistics present a huge mass of fragmentary and heterogeneous data, which in their present undigested form "are well-nigh inaccessible." The use

Nearing: Wages in the United States, p. 7. The defects of our wage statistics are well stated by Professor Nearing in the following paragraphs:

"At every turn the need arose for an accurate, concise statement of the wages being paid in the various parts of the United States, yet to date no study has been made which supplies the need. Ryan's Estimate is old, and at best incomplete; Mrs. Moore's statement, like the statement in the 1903 Report of the Commissioner of Labor, is of standards of living primarily, and only incidentally of wages. In neither case is the ground covered sufficiently to warrant valuable wage deductions. The Wage Study accompanying the Census of 1900 is old, and rather inadequate, as the compilers themselves point out. The available data on the subject of wages exist chiefly in the reports of State bureaus of labor, and are unfortunately of such a nature as to render comparison with data of a decade since (in the few cases where such data exist) most unsatisfactory. . . . New York wage statistics

of wage statistics, such as they are, for an analysis of the effects of immigration on wages is restricted by lack of comparable statistics of occupations by nativity.'

On the other hand, a rise or a fall in money wages is no indication of an increase or decrease of the resources of the wage-earners, unless coupled with comparative statistics of the cost of living. The various index numbers of prices, however, admit of a wide margin of error. An illustration is furnished by the curve plotted by Mr. Streightoff from the figures of the United States Bureau of Labor on "real wages. " It appears that during the period from 1890 to 1907, the purchasing power of full-time weekly wages was at its maximum in 1896, when, according to the statistics of the Massachusetts Bureau, the ratio of unemployment was as high as during the crisis year 1908;3 the country had relate to members of labor unions only; the average wage statistics of Pennsylvania are incomplete—even those cited are wretchedly compiled and presented; Illinois has published no recent statement of wages except in department stores; Missouri, Michigan, and Indiana publish little or no wage data. The statistics for Ohio are excellent, but very diffuse and unconcentrated. . . . Therefore, of the ten leading industrial States, three present worthy wage data; the statistics of two are far from satisfactory; while five of the ten States furnish no current wage material of value to this study. Deplorable as is the lack of statistics in these great industrial States, the conditions in the country at large are infinitely worse. Of the forty-seven States of the Union, not more than five publish up-to-date wage statistics. Of the remaining States, a score publish statistics of average wages only, which, in some cases, are so unrepresentative as to be valueless. (Pp. 9-15.)

"The last Bulletin of the Federal Bureau or Labor relating mo wages was published in 1908. . . . The material as a whole permits of practically no deduction, save that wages are considerably higher in the West than in any other section of the country, and that the wages in some trades are very much higher than in others." (Pp. 138-139.)

For example, though the average wages of coal miners can be computed from census statistics for every State, it is impossible to ascertain for many States the purcentage of Slavs among coal miners, because coal miners are combined in census statistics with metalliferous miners and quarrymen.

Streightoff, loc. cit., Chart XI.

See Ch. VI., Table 23 and Diagram IX.

not recovered from the effects of the crisis of 1893-1894, and the industrial situation was again disturbed by the uncertainty of a Presidential campaign fought on one of the most vital economic issues, the money question. No trust can be placed in statistics which lead to conclusions so glaringly at variance with facts still fresh in people's memory.

Overlooking, however, the inadequacy of our wage statistics, let us examine the material, such as it is, bearing upon the relation between immigration and wages.

I

We have seen that in the 40's the wages of Irish street laborers in Brooklyn were insufficient to provide for rent, and they were compelled to live in shanties. Bad as the housing accommodations of the Italian street laborers may be to-day, they nevertheless earn enough to pay rent, which is indisputable proof of an increase in "real wages.

A generation later, a statistical inquiry into the earnings of 75,000 wage-earners in the State of Massachusetts led the Bureau of Labor Statistics to the conclusion that "the average earnings of a majority of the skilled laborers in this State do not reach the average cost of the necessities of life," with the result that "the children of the poor are taken away early from school, and brought into the labor market; the son to the factory, store, or shop, and the daughter to the life and wages of a factory or cash girl, or of a serving woman." Evidently the skilled mechanics forty years ago did not fare better than the wage-earners of our own day.

In the same report there is a comparison of earnings and expenses in Massachusetts for 1800, 1830, and 1860. It is estimated that in 1800 the master mason alone of all craftsmen earned more than his expenses, whereas master carpenters and master painters could not pay their expenses; journeymen carpenters, masons, and painters were in the same category. In 1830 a journeyman mason earned barely

1 New York Weekly Tribune, May 2, 1846. Quoted in Documentary History of American Industrial Society, vol. viii., pp. 225–226.

Third Annual Report of the Massachusetts Bureau of Statistics of Labor (1871-1872), pp. 531-532.

enough to support a family of four, but the earnings of a journeyman carpenter were still insufficient to provide for a family of the same size. The wages of a laborer were estimated at $226 a year, which was equivalent to a little over one half of the estimated expenses of an average family of four persons. In 1860 neither a master carpenter, nor a master painter earned enough to support a family; no journeyman in the building trades was able to support a family solely on his own earnings. The earnings of a common laborer remained, as thirty years before, at a little over one half of the estimated cost of supporting a family.'

The Bureau of Statistics of New Jersey in its early days published a number of workmen's budgets. A compilation of the data for 1885 will be found in the Appendix, Tables XIX. and XX. The workmen were either native, or immigrants from Northern and Western Europe. Immigration from Southern and Eastern Europe was still too insignificant to affect the labor situation. It appears that of all wage-earners in specified occupations only glass-workers and blacksmiths earned enough to support on their wages an average family of about five persons. Other skilled mechanics, such as machinists and carpenters, needed the assistance of members of their households to support a family of the same size, while workers in textile mills could not meet expenses even with the assistance of members of their families. Among unskilled laborers there were some whose earnings were sufficient to support their families, but their expenses averaged only $1 a day. Those families whose expenses averaged about $1.50 a day or more were barely able to keep above water with the aid of the children's earnings. None of the Irish laborers could make both ends meet, although their expenses were somewhat below those of the other English-speaking laborers.

There are similar budget data in the report of the Ohio Bureau of Labor Statistics for 1885. In a few trades, the

'Third Annual Report of the Massachusetts Bureau of Statistics of Labor, (1871-1872), pp. 514–517. See Appendix, Table XVIII.

average earnings were insufficient to provide for the support of the wage-earner and his family. The average deficit per family for each occupation is shown in Table 87:

TABLE 87.

AVERAGE ANNUAL DEFICIT PER WORKING FAMILY IN OHIO, BY OCCUPATIONS, 1885.'

Among the skilled mechanics the stone-cutters and the machinists were on the border line between surplus and deficit; the cabinet-makers, wood-carvers and cigar-makers depended upon outside sources, in addition to their wages; likewise the iron-workers and the miners. The proportion of immigrants from Southern and Eastern Europe in those occupations five years later varied in Ohio between 2.1 and 5.3 per cent.

One fact may be taken as firmly established by the preceding statistics, fragmentary and insufficient as they are for other purposes, viz., that in the days of "the old immigration" the wages of unskilled laborers, and even of some of the skilled mechanics, did not fully provide for the support of the wage-earner and his family in accordance with their usual standards of living. The shortage had to be made up by the wife and children.

If the tendency of the new immigration be to lower the rate of wages or to retard the advance of wages, it should be expected that wages would be lower in great cities where the recent immigrants are concentrated, than in rural districts where the population is mostly of native birth. All * See Appendix, Table XXI.