1896.] on Electric Besearch at Low Temperatures, 261 



certain metals at low temperatures there are sudden changes of 

 direction which indicate a change in the sign of the Thomson effect 

 in that metal at that temperature, and probably, therefore, some 

 important molecular change at the corresponding temperature. 



In the case of the 19 and 29 per cent, nickel-steel alloys there is 

 an interesting thermo-electric phenomenon. If a loop of wire of this 

 material is partly dipped in liquid air, the portion cooled becomes 

 thermo-electrically dififerent from the remainder, and gives a strong 

 thermo current if connected to a galvanometer and warmed at one 

 point, where the changed and unchanged portions meet. 



Leaving the further elaboration of these points, we must next 

 notice some of the facts with respect to the magnetisation of iron at 

 low temperatures. Professor Dewar mentioned, in a discourse on the 

 scientific uses of liquid air, some results obtained on cooling small 

 steel magnets. These effects we have since again explored at greater 

 length. 



Let me show you, in the first place, the effect of cooling a small 

 steel permanent magnet to the temperature of liquid air. We will 

 first take a magnet made of a fragment of knitting needle or ordinary 

 carbon steel and examine the effect of low temperature upon it. 

 Placing the magnet behind the small suspended magnetic needle of a 

 magnetometer we obtain a deflection of the magnetometer needle, 

 which is a measure of the magnetisation of the magnet causing the 

 deflection. On bringing up a small vessel of liquid air and immers- 

 ing in it the magnet under test we notice at once a sudden decrease 

 in the deflection of the magnetometer needle. This indicates that a 

 notable percentage of the magnetisation of the magnet has been 

 removed. On taking away the liquid air bath and allowing the magnet 

 to heat up again we find that there is a still further decrease in mag- 

 netisation. On cooling it again with liquid air the magnetisation 

 then increases, and from and after that time the effect of the cooling is 

 always to increase the moment of the magnet, and the effect of heat- 

 ing it up again always to decrease the moment of the magnet. Hence 

 we see that the effect of the first immersion in liquid air is to give 

 a shock to the magnet which deprives it permanently of a consider- 

 able percentage of its magnetism ; but when once it has survived this 

 treatment, then cooling it strengthens the magnet, and warming it 

 weakens it. 



This is not by any means always the case. If we take a magnet 

 made of the 19 per cent, nickel-steel, the peculiar characters of which 

 were explained a few moments ago, we shall find a very different 

 state of affairs. Here we see the first effect is, as before, to remove 

 a very considerable percentage of the initial magnetisation ; but after 

 that stage is passed, then cooling this nickel-steel magnet always 

 weakens it still more, and warming it up again strengthens it. The 

 subsequent effect of cooling is therefore in the opposite direction in 

 the carbon-steel and in this nickel-steel. These changes of moment 

 can best be represented by a diagram of lines as in Fig. 15. 



