620 
Proceedings of the Royal Society of Edinburgh. [Sess. 
In the second column are entered the percentage “ sensitive states ” 
induced by the successive annealings, while in the third column these are 
expressed as a percentage of the effect produced by the first annealing. The 
figures in all cases correspond to a magnetising field of 10 C.G.S. units. 
Similar reductions took place in the enhancement of the residual magnetism 
and the coercive force. These results are not appreciably affected by any 
“ ageing ” of the material through the repeated exposures to the high tem- 
peratures employed in the annealing process. A few similar measurements 
have been made with the hard steel mentioned in the analyses at the com- 
mencement of this paper. These indicate that the fatigue effect is decidedly 
less marked in the case of this high carbon steel than in that of the steel wire. 
Several specimens which had been reannealed and tested from twelve 
to sixteen times were laid aside for fifty-four days, and then once more 
annealed and tested. They exhibited no indication of recovery from the 
fatigued condition. 
Repeated Annealing . — Fresh specimens of the steel were annealed at 
900° C., some twice, some four times, and some six times, without any 
intermediate testing or application of any magnetic field whatever. They 
showed no difference in percentage “ sensitive state ” beyond what could be 
accounted for by experimental error and accidental jarring. In particular, 
they gave no sign of an increased “ sensitive state ” induced by repeated 
annealings, and showed that the fatigue effect referred to above only conies 
into play when a specimen is changed from the “ sensitive ” into the 
normal magnetic condition. 
“ Sensitive States ” induced by low temperatures . — So far, the “ sensitive 
state ” had been induced in the specimens by heating them up to temperatures 
varying from 100° C. to 900° C., and leaving them to slowly cool. It 
appeared probable that an effect might be obtained by starting with a 
specimen at an extremely low temperature, allowing it to rise to room 
temperature, then cooling again to the initial condition, and there testing it. 
This was tried in the following manner. 
A specimen of the stout steel wire A (fig. 2) was placed in a glass tube 
BCD, closed at one end B, and open at the other D. The portion C D was 
bent up at right angles to the main length B C. The magnetising solenoid 
E E was fitted with cork bungs F F F. The glass tube passed through two 
of these at such a height as to make the axes of the specimen and coil 
coincident. Liquid air was poured in at D, and the specimen thus cooled 
down to about —190° C. As the corks were bad conductors of heat and the 
air inside the solenoid was at rest, it was possible to keep the temperature 
of the specimen perfectly steady. An alternating current was sent through 
