1912-13.] Sensitive Magnetic State induced in Various Ways. 251 
in the solenoid, hung from the upper brass wire with its ends about 
equally distant from the ends of the solenoid. Also the upper brass wire 
could be moved through a small distance up and down at the clamp till 
that position was obtained which brought the pole* of the specimen most 
nearly opposite to the magnetometer needle. 
To apply longitudinal stress to the specimen, a large can was obtained, 
capable of holding about a cubic foot of water, and fitted with a gauge, and 
also a tap near the bottom. This can was hung on the hook attached to 
the lower brass wire, and a gentle stream of water was allowed to flow into 
it, until the desired weight, which was read off on the gauge, was obtained. 
In this way any desired stress could be applied with an almost entire 
absence of vibration — a very important point, as it has been shown that 
vibration greatly diminishes the sensitive state. To decrease the weight 
by any fixed amount, it was only necessary to turn on the tap near the 
bottom of the can, and allow the required amount of water to flow out. 
The five specimens tested were all steel wires, 50 cm. in length and about 
yq inch in diameter. This thickness of wire, which is moderately large, 
was chosen so that the effect of small stresses might be the better examined. 
A weight of 1 kilogramme applied to such a wire gives a stress of 19'74 kilos 
per square cm. Of the five specimens used, four, supplied by Messrs W. N. 
Brunton & Son, form a graded carbon series having respectively 0T5 per 
cent., 055 per cent., 0*8 per cent., and 1*2 per cent, carbon in their composi- 
tion. The fifth is interesting as being a specimen of the same wire as was 
used in Part I., and the results obtained by the two methods are therefore 
for it directly comparable. 
Each specimen was, before being tested, carefully annealed as in the 
previous case, and then stretched several times by the application and 
removal of the maximum weight to be employed. Unless this was done, 
it was found that the specimen was not in a stable condition, and that 
repetition of the same processes did not produce the same results. 
The specimen was then placed in position in the solenoid, carefully 
demagnetised, subjected to a certain strain, tested, demagnetised, and tested 
again, and from the two curves obtained the percentage sensitive state due 
to increase of stress from 0 to w was determined. It was then de- 
magnetised again, the strain was removed and the sensitive state due to 
decrease of stress from w to 0 obtained, and this was repeated for various 
values of w from 0 up to 20 kilogrammes. 
The effect of the cycle 0-u>-0 was also determined, and also the effect 
of increasing or decreasing the stress by successive differences of 2*5 and 
of 5 kilos. 
