once these various preconditions had been established

So, once these various preconditions had been established, it was

possible to make rational technological decisions. Finally, it is necessary to mention two particularly Japanese characteristics of technology choice. The first is, as has already been stressed several times,

the particularly high capacity of the Japanese cotton-spinning industry

for adapting to market conditions once the choice of technology had

been made: for example, the deliberate use of cheap female labour;

the introduction of combination spinning and weaving mills; the import

of Indian raw cotton under a special discount contract; or the export

of coarse yarn to China and so on. The Japanese experience suggests that to make successful technology choices, it was important

not only to make the right choice from a purely technological viewpoint, but also to choose on the basis of market conditions and to

adapt to those market conditions.

The second is that in Japan, the switch from the mule to the ring

frame was very rapid and moreover all-embracing; decisions were

first made by leading companies and information spread rapidly,

leading to successive decisions made by small mills. For this process

to happen, a flourishing exchange of information between engineers

and, at the same time, a strong competitive spirit between them were

needed. Examples of the competitive spirit abound: there were the

tactics of Yamanobe and Taniguchi when they were purchasing ring

frames, as was mentioned earlier; or there was the case of developing

cotton-blending techniques, when each company competed with the

others to develop the technique under a cloak of secrecy. An example

of the exchange of information is the organization of spinning engineer gatherings by the chief engineers of the companies affiliated to

Boren. The importance of these meetings for the exchange and diffusion of knowledge should not be overlooked.37 Another important

characteristic of the Japanese cotton-spinning industry is the very

early creation of a framework of cooperation between companies.

There are many examples: companies received help from other

companies in training core workers and lent engineers to other 

For more detail see Kiyokawa (1973), especially the footnotes and references.

More exactly, the number of mules increased again in about 1897 by

more than 30000 spindles. These machines, however, were not installed

for producing low count yarns as before, nor were they for experimental

purposes: they were for producing high count yarns (74-78s), as in the

case of Nippon Boseki. Therefore they are excluded from this study.

Certainly by around 1935, the world's cotton-spinning industries (excluding England) had completed the switch from mules to ring frames,

but at the beginning of the twentieth century the mule was still dominant

in the majority of countries: see Dainippon B6seki Rengokai (1937) for

example. In America, the country responsible for the development of

the ring frame, the number of ring frames had already reached 13 400 000

per cent of total spindles) by 1900, but it was not until after the turn

of the century that mule spindles started to decline in absolute numbers.

Even in the 1910s there was still a strong demand for mules in India,

which concentrated likewise on the production of coarse yarns. This

phenomenon was, however, due to the fact that the Indian cotton industry relied heavily on English engineers; for more detail see Kiyokawa

(1983).

In most cases in Japan, it took roughly 12 months for the machinery to

arrive after placing the order (when importing from England). This period

was several months or even a year shorter than in India (Bombay).

Perhaps this implies that Japan's trading functions and customs clearing

were relatively efficient.

The statistical data used for calculating the figures in Table 4.1 and

Figure 4.2 were from individual company data for each of the three

years. The 1884 figures for Mie B6seki are based on data for May 1883;

and for companies with missing data in their 1890 and 1893 reports (such

as Kanegafuchi B6seki), the available data from the closest proleeding

preceding month was used instead. For the sources of these data see the

sources listed under Table 4.1.

The 'separator' is the same as the 'anti-ballooner' in England. Compared

with material on the mule; there is relatively little material on the history

of the ring frame. Nasmith (1890), the Indian Textile Journal, September

1894 (pp.281-2), April 1902 (pp.194-5) and Indian Textile Journal (1954),