180 C. Bar us — Effect of Magnetization on the 



In these results the thermal and magnetic effects are super- 

 imposed on the continuous viscous motion due to inequalities 

 in the wires, the upper wire being more viscous than the lower 

 wire. In every case, however, this normal viscous detorsion 

 is perceptibly retarded whenever the upper wire is magnetized. 

 It follows conformably with the above results that the viscosity 

 of iron is apparently diminished by magnetization ; that the 

 amount of this diminution is no larger than is quite in keeping 

 with the heating effects due to the passage of current through 

 the helix. My results therefore fail to give any satisfactory 

 proof that magnetized and unmagnetized iron differ in viscosity 

 by more than the equivalent of one or two degrees centigrade, 

 at ordinary temperatures. 



I am far from wishing to assert, however, that Tomlinson's 

 results are temperature phenomena. It is possible that the 

 marked tendency of soft iron to assume permanent set, and con- 

 sequently the relatively large viscous motion immediately after 

 strain is imparted, makes the vibration method particularly sen- 

 sitive in registering the viscous effect of magnetizing iron.* 

 But by using static methods and observing in the interval of 

 gradual or 'purely viscous deformation, no satisfactory evidence 

 of such an effect is discernible. Moreover, magnetization 

 changes the rigidity of iron and therefore necessarily jars a 

 twisted wire. There result such changes of viscosity as are 

 produced by any sudden vibratory disturbance ; changes which 

 of course are purely mechanical, but which obscure the direct 

 result of magnetic induction beyond recognition. 



RIGIDITY OP MAGNETIZED IKON AND STEEL. 



Iron. — The effect of longitudinal magnetization on iron or 

 steel twisted within the elastic limits is marked detorsion, in- 

 creasing in amount with the intensitjr of magnetic field, increas- 

 ing also with the rate of twist, at a retarded rate in both 

 instances, toward a maximum. If . the sense of the magnetiza- 

 tion be reversed (i. e., if the helix current be changed in direc- 

 tion) the amount of detorsion is in general unchanged. f 



In tables 1, 2, 3, discontinuous but perfectly regular varia- 

 tion of <p, on passing from unmagnetized to magnetized iron 

 or steel are strikingly apparent. Table 4 contains special 

 results for large rates of twist (r), and has been drawn up to 

 exhibit the independence of <p of any possible irregularities in 

 the position of the helix as well as the change of sign of tp 

 with the sign of r. The table shows also that the sign of <p is 



* Of., moreover, Wiedemann's remarks on the vibration method, Wied. Ann., 

 vi, 485, 1879. 



f The above phenomena have been elaborately discussed by Wiedemann 

 (Galvanisinus, pp. 683-698). 



