390 PROF. KNOTT ON MAGNETISM AND TWIST IN IRON AND NICKEL, 



calculation is perhaps interesting as being, I believe, the first of its kind, namely, a 

 numerical comparison of what are at first sight different phenomena depending on the 

 relations of magnetic and mechanical stress and strain. The result of the comparison, in 

 my opinion, demonstrates the sufficiency of Maxwell's explanation of the Wiedemann 

 effect in terms of the simpler Joule effect. Thus Maxwell's explanation has, to a first 

 and simple approximation, stood the test of numerical calculation; whereas it is impossible 

 even to imagine how to begin in applying such a test to Wiedemann's theory. 



Another serious objection to Wiedemann's theory is, that it gives a so-called explana- 

 tion of a particular kind of magnetic strain, but furnishes no insight into the mechanism 

 of simpler magnetic strains. And then, again, it seems to me — in making this statement 

 I may simply be showing how little I understand the mechanism of the frictionally 

 restrained rotating molecules — but it seems to me that, according to Wiedemann's theory, 

 the direction of twist in an iron or nickel wire should depend on the order in which the 

 circular and longitudinal magnetising forces are applied to the wire. Thus, let the wire, 

 hanging vertically, be magnetised with north pole downwards, and then let a current 

 be passed down it. Then the originally vertically polarised surface molecule facing the 

 spectator will tend to rotate like the hands of a watch. But, if the wire is first magnetised 

 circularly by a current flowing down it, and then subjected to the influence of the longi- 

 tudinal field, the originally horizontally polarised surface molecule will tend to rotate 

 contrary to the hands of a watch. Now, it is difficult to see how such contrary tendencies 

 can possibly cause a similarly directed twist. 



I have recently found by experiment that the amount of twist, due to a given com- 

 bination of magnetising forces, does depend upon the order in which the forces are 

 applied ; but. except in a very particular case, the direction of twist never does. As the 

 experiments, however, are not quite completed, I reserve their discussion for a second 

 paper. 



