253 
1906-7.] The Strength of Twisted Threads. 
relation to the strength, although in a general way a strong thread stretches 
more than a weak one of the same quality. 
There is only one machine at present in the market adapted for testing 
the relation between twist and yield-point strength by the second method 
which has been described. This apparatus was designed by Mr G. R. Smith 
of Bradford, and is shown in fig. 3. A thread A is stretched between two 
clamps B and C. C forms the end of one arm of a bell-crank lever pivoted 
at G, whose other arm F carries a can E. Water may be run into can E 
from a reservoir D, and thus the thread may be twisted under any tension. 
The can is provided with a gauge. H is an adjustable counterpoise to 
balance F when E is empty. The amount which the thread stretches under 
tension is taken up by turning wheel I, which communicates motion by 
a train of wheels and special pinion to a cylindrical rack K. K moves 
the clamp B so that the lever F always remains horizontal. The twist in 
the thread may be varied by turning wheel M, the turns being indicated on 
dial T, to which motion is communicated by a worm W carried on bar K. 
Arm F carries at the end a rubber pad P, which drops on the water nozzle 
and stops the flow when the thread yields or breaks. 
The method of testing the relation in question is as follows : — Take out 
all the twist from the thread. If it is a single, it will probably not carry 
any appreciable load at all after untwisting. Then put on one turn 
per inch and run in water very slowly to E from reservoir D, and take 
up the stretch by moving B. As soon as the yield point is noticed, read 
off the amount of water in E by the gauge. Proceed in the same way for 
two turns per inch and more until the thread wears out. Suppose this 
occurs at four turns per inch. Start a new series with one turn per inch, 
and the thread will probably break at five or six turns. Start a third series 
with two turns, and so on. The observations were restricted to 10-inch 
lengths of thread, because in the machine used rack K did not admit of 
greater movement than 4 inches. With excessive twist the stretch 
sometimes amounts to 35 or 40 per cent, of the original length. But as 
woollen, worsted, and cotton fibres are much less than 10 inches, it is 
probable that the results do not differ greatly from those obtained on 
lengths of 20 to 30 inches. 
In the later experiments, the above apparatus was modified. The 
reservoir D was removed, arm F was graduated, and the tension produced 
by a sliding weight on arm F. When one weight was not sufficient to 
cause the thread to yield, it was dropped gently into E, and a similar weight 
was started on its outward journey from pivot G along the arm F. This 
modification was found to give much better results, because the hydraulic 
