66 Profs. J. Dewar and J. A. Fleming, Changes produced in 



marked. Steels having these high percentages of nickel are, as Dr. J. 

 Hopkinson has pointed out,* remarkable for the wide range of tem- 

 perature within which they can exist in two states, one considerably 

 magnetic, and one practically non-magne'tic or but feebly magnetic. 

 In these two states their mechanical and other physical properties are 

 entirely different. In the experiments here mentioned, the nickel 

 steel samples were in the magnetic condition. They are put into this 

 condition by dipping for one moment in liquid air, and are only trans- 

 formed back into the feebly magnetic condition by heating to a 

 cherry-red heat. The 29 per cent, sample of nickel steel being in the 

 magnetic condition was magnetised by contact with the poles of the 

 efectromagnet. On cooling it in liquid air it immediately lost about 

 20 per cent, of its moment, on warming up again to 5 C. it lost about 

 5 per cent, more, and from and after that point remained in a condi- 

 tion in whiph cooling the magnet to 185 C. caused its moment to 

 become about 10 per cent, less than it was at 5 C. Hence the 29 per 

 cent, nickel steel exhibits the same quality but in a less marked 

 degree than the 19 per cent., in that its magnetic moment is decreased 

 by cooling to -185 C., and recovers again on heating up to 5 C. 

 In this respect the two samples of nickel steel differ from all other 

 samples of steel which we have examined, in that they have a negative 

 temperature coefficient for magnetic moment change with tempera- 

 ture, after the first change on cooling has taken place. 



Pure Nickel. In order to see if this peculiarity extended to pure 

 nickel, we examined the behaviour of a small magnet made with 

 Mr. Mond's pure nickel, but we found that such a nickel magnet, 

 magnetised to saturation, behaved exactly as did a carbon steel 

 magnet (see fig. 15). The effect of the first cooling to the tempera- 

 ture of liquid air was to diminish the magnetic moment. On allow- 

 ing the magnet to heat up again to the ordinary temperature the 

 moment diminished still more, and from and after that time the 

 behaviour of the magnet was perfectly normal, that is to say, its 

 magnetic moment when at 5 C. was less than its magnetic moment 

 at 186 C., but only by about 3 or 4 per cent, of the latter value. 



Silicon Steel. A. sample of silicon steel, containing 2'67 per cent, 

 of silicon, behaved in a normal manner (see fig. 16). The magnet 

 experienced a permanent diminution of moment on cooling for the 

 first time, and after that, its magnetic moment when cold was greater 

 than its magnetic moment when warm. 



Soft Iron. In order to determine if similar changes of magnetic 

 moment could be produced in the case of soft annealed iron, small 

 magnets of Swedish iron were prepared, formed of a short length, 

 about 15 mm., of soft iron, or a small slip of annealed transformer 

 iron. On magnetising these in a strong field, and testing them with 

 * ' Koy. Soc. Pi-oc.,' 1890, vol. 47, p. 138. 



