PHYSICS. 245 



a concentric circle of the disk. To study the free magnetism 

 distributed over them, he used a small soft iron cylinder of a 

 few centigrams weight, fixed in the centre to the rod of an 

 areometer floating in water. The force required to detach this 

 was estimated by the weight of water which had to be let off 

 from the cylindrical vessel containing the areometer before the 

 contact was broken. The precise instant of contact and de- 

 tachment was indicated by an electric signal. In this way 

 it was proved that the quantities of free boreal and austral 

 magnetism were equal in the two portions of contrary name 

 in the same plate. For plates of different diameters the forces 

 of detachment depend simply on one specific coefficient, va- 

 riable with the nature of the steel and with the thickness. 



Gaugain has recorded the curious fact that a bar of steel 

 magnetized at 400 or 500 C. not only loses its magnetism 

 gradually as it is cooled until it becomes zero, but that 

 magnetism of contrary sign appears, and increases until the 

 bar reaches the temperature of the air, never becoming, 

 however, as intense as the original magnetism. On again 

 heating it, the same effects are produced in the inverse order, 

 and they may be reproduced many times without remagnet- 

 izing. To account for this result, the author proposed the 

 hypothesis that the bars which presented the phenomena 

 consisted of two layers of magnetism of contrary name, 

 which were differently modified by the variations in the 

 temperature of the bars. To test the question, experiments 

 were made with a steel tube enclosing a steel rod, forming a 

 magnetic system. If the tube be magnetized, the rod insert- 

 ed, and then withdrawn, the latter is found magnetized like 

 the tube. But if, before withdrawing the rod, the system 

 be heated to 300 and allowed to cool, the tube has lost near- 

 ly the whole of its magnetism, and the rod has become op- 

 positely magnetized. The same results are obtained if the 

 rod be magnetized in place of the tube. Moreover, if both 

 tube and rod be magnetized at the ordinary temperature, or 

 at 300 to 400, and be at once separated, they are magnet- 

 ized alike. But if the system be cooled before withdrawing 

 the rod, their magnetism is opposite. If the magnetization 

 has been effected at 300 to 400, reheating the system in- 

 creases the direct magnetism of the nucleus, and at the same 

 time diminishes the inverse magnetism of the tube. 



