let us examine with statistical data the view that the ring

Finally, let us examine with statistical data the view that the ring

frame was adopted to resolve the problem of a scarcity of skilled

labour, or to take advantage of cheap female labour. As was discussed

above, there is no doubt that operation of the mule required skilled

labour. As a result mules were not operated at high productivity

levels in 1884, since there were serious shortages of experienced

labour: one spindle produced only 24 momme of yarn in one day (100

momme = 0.827 lb or 375g). The ring frame would allow the immediate substitution of unskilled women for skilled men, who were

difficult to find, but it does seem highly doubtful that this was the

main reason for choosing the ring frame because the female worker

ratio in mule mills followed the same trend as the female worker

ratio in the industry as a whole. For example, although the female

worker ratio in mule mills in 1884 was evidently far lower than in

ring frame mills in 1890, by around 1890 the ratio in mule mills had

risen. Moreover, given that production per spindle increased at the

same time by 2.5 times, it does not seem that the lack of skill of male

workers was the only reason for low productivity. In addition, in

neither of the years 1890 and 1893 was the difference in the female

worker ratio between mule mills and ring frame mills statistically

significant (at a 95 per cent level, as are other significance tests below),

and therefore it is difficult to believe that the female worker ratio

was determined merely by the machines' technological characteristics. In other words, it does not necessarily follow that the replacement of skilled male workers by unskilled female workers arose

because of the switch from mules to ring frames, and therefore it is

doubtful that the reason for switching to ring frames was to employ

female labour intensively. In short, the development of the widespread use of cheap female labour occurred once the ring frame had

been chosen: indeed, the development was merely one of the adjustments made to use the ring frame more effectively. This aspect,

however, is related to other variables of the production structure and

so needs to be examined in a wider context (i.e., by using multivariate

analysis). 

Now let us examine a little more closely the changes in production

structure that accompanied the introduction of the ring frame. There

are three further points arising from a study of Table 4.1 that should

be mentioned in relation to the foregoing discussion. The first is that,

as is expected, the wages of men working in mule mills were higher

than the wages of women working in ring frame mills (the difference

is statistically significant). Moreover, because there is no significant

difference between the wages of women working in a mule mill and

women working in a ring frame mill, it seems that the mule was by

and large operated by skilled men (in 1890 the difference between

wages of men in mule mills and wages of men in ring frame mills was

significant at a 90 per cent level, but in 1893 it was statistically insignificant). Over time there was a tendency for the wages in mule mills

and wages in ring frame mills to equalize. While wages were rising

for women working with either type of spinning machine in the relatively large-scale mills such as Kanegafuchi, Osaka Boseki, Mie Boseki

and Naniwa Boseki, a tendency for intercompany wage differentials

to widen had also started.

The second point is that although it is thought that the price of

ring-spun yarn was higher than the price of mule-spun yarn, as was

the case for imported mUle-spun and ring-spun yarn, the difference

in price is not statistically significant. The third point is that, as has

already been mentioned, there was clearly a large difference (statistically significant) in average yarn count and output per spindle

between the mule and the ring frame. Let us focus on average yarn

count. It should be stressed that, although there was a clear rise

(statistically significant) in the ring frame's average count over the

years 1890-3, even by 1893 it was languishing at between 17 and 19

in the 'proper' mills such as Osaka Boseki, Kanegafuchi and Settsu

Boseki.

Let us turn to operating hours. Since 1890 happened to be a recession year, some of the mule mills (four out of 11) were using a twoshifts system. In 1893, however, all mills except for two (out of 38)

had introduced a night work shift. The night work shift was first used

by Kuwahara Boseki in 1883. Then its use spread first through the

relatively large-scale mills, such as Osaka Boseki, Mie Boseki and

Temman Boseki, followed by the small scale 'Model Ten' spinning

companies. By 1889 at the latest, the night work shift had become

a firmly established practice and it could be viewed as one of the

distinctly Japanese types of adaptation. 

The production structure is determined not only by the technological factors peculiar to a particular type of spinning machine, but also

by factors common to the whole industry. As those common factors

are correlated with other factors, the nature of the production structure can only be understood adequately using multivariate techniques.

The nature of the production structure of both the mule mills and the

ring frame mills can be measured and identified by the generalized