504 



PROFESSOR KNOTT ON SOME RELATIONS BETWEEN 



the second and third cases, for which certainly the field-reversal twist with steady current 

 is much greater than the current-reversal twist for steady field. 



Current. 



Field. 



Twist. 



3-12 



118 



23 



2-2 



82-9 



2-45 



19 



71-8 



2-5 



1-81 



68-3 



2-3 



1-27 



47-5 



1-25 



•99 



37-5 



•90 



•79 



29-8 



•73 



•66 



25-4 



•63 



•53 



199 



•42 



•44 



16-6 



•33 



•35 



132 



•29 



•09 



3-3 



•04 



This result, namely, the obtaining of a twist when both magnetising forces are 

 simultaneously reserved, is of considerable interest. It demonstrates in a very simple 

 manner what all the other experiments here discussed also show, that the law of the 

 after-effect for circular magnetisation differs largely from that for longitudinal magnetisa- 

 tion. Suppose, for example, that a nickel wire is subjected for the first time to the 

 combined influence of a line current and a longitudinal field. Whatever be the directions 

 of the magnetising forces the initial twist produced will have the same numerical value, 

 provided the wire is quite symmetrical with regard to its axis. Nevertheless, if the 

 currents, after producing this initial twist, are simply reversed, so that their directional 

 relation is unchanged, the twist changes. Thus, if we are to explain the Wiedemann 

 effect in terms of the elongations and contractions associated with magnetisation, we 

 must suppose that the lines of magnetic induction in the wire are not simply reversed 

 with the direction of the magnetising forces. In short, we must assume a magnetic 

 aeolotropy. The phenomenon just described seems to me to be a demonstration of the 

 existence of such an aeolotropy, which on general grounds is a plausible enough 

 hypothesis. 



12. Magnetic Twists in Cobalt. — So far as I am aware, no attempt has been made 

 to obtain the Wiedemann effect with cobalt. The practical impossibility of obtaining a 

 cobalt wire is perhaps a sufficient explanation of this. Nevertheless it seemed very 

 desirable to find out if there was any hint at all of the existence of the phenomenon in 

 this third magnetic metal. It struck me that if Maxwell's explanation were the correct 

 one, the effect ought to be visible with a thin rod of rectangular section. No doubt a 

 cylindrical rod or wire is much to be preferred ; but the same strain effects ought to be 

 produced to some extent in rods of any form of section. Accordingly, from a piece of 

 rolled sheet cobalt supplied to me two years ago by Professor Tait, I cut a strip 36'1 

 cm. long, and with a rectangular section of 0'45 mm. by *068 mm. This I set hanging 



