EFFECT OF STRESS ON MAGNETIZATION. 51 



since we may put f=ff' 

 If H is zero, we have 



de m rif m — n 



lience neglecting the change of elastic constants by magnetization, 

 we have, from (o) , 



/ 3/ \ VI / r)/ \ m-n / dl \ . 



\ dT Jtc~ nßin — 7i) \ 'Oc ) jfj^ n{om—n) \ Of I H,e,g 



Hence by (2) and (4), we get 

 lien ce finally we get 



(w). = (Tri <«' 



As to the twist, .1. J. Thomson obtained a relation, whicli 

 strictly speaking, holds in the case of a thin tube, i. e. 



'^-^m„\^-"{':,)\ ^- 



oc 1 r bi 



n — 



Here again in his original work, the factor j is dropped ; <• is 

 not the twist r per unit length, but it is connected with r by the 



