If we omit all above 24s., we have: Average, 13s. 4d.; median, 13s. 7d.; quartiles, 10s. 5d., 16s. 6d.

The difference in these averages is not significant.

The table is best written in percentage.

2s. to
4s.		4s. to
6s.		6s. to
8s.		8s. to
10s.		10s. to
12s.		12s. to
14s.
24791317
14s. to
16s.		16s. to
18s.		18s. to
20s.		20s. to
22s.		22s. to
24s.		Above
24s.
181375½2 2½ per cent. earning.

Note, that if these wages were repeated week by week through the year the average worker would make about £35.

III.—CHANGE OF WAGES BETWEEN 1885 AND 1900.

Where wages are continually fluctuating week by week and month by month, while, in addition, there are depressions and inflations affecting various groups of workers for one or two years, it is a matter of very great statistical difficulty to determine whether wages have on the whole been stationary, rising, or falling. Even if we had a complete account year by year these difficulties would remain; but as it is we are dependent on the records of only seven firms—good, bad, or indifferent—since 1885, 1887, 1894, 1895, 1896, 1897, and 1898, respectively. No amount of further research would make such records more than very insufficient, for it is very rarely that the figures are preserved for any length of time. What changes there are may very likely be due to peculiarities of a particular firm, to its success, or to changes in character of work, and only in case of agreement in all the figures could we generalise. Our conclusions, then, will be chiefly negative.

There is no sufficient evidence that wages in 1899 are above or below wages about 1895, 1890 or 1885; the only difference appears to be due to individual busy or slack years.

In the two cases (C. and G.) where machine rulers are separated their wages have risen from 6s. to 8s. in 1897-99.