Transverse Fields and Allied Phenomena. 411 



tigators, and recently in the exhaustive researches of C. Gr. 

 Knott,* and others. In Knott's experiments longitudinal and 

 circular fields are superposed and the data are made tributary 

 to Maxwell's^ and Chrystal's^: theories, in which the stresses 

 of the Joule effect are sufficient to account for the torsional 

 phenomena. Drude§ uses similar reasoning to compute Kirch- 

 hoff's third constant. 



The immediate effect of circular magnetism is a decrease of 

 the rigidity of the wire (iron, nickel) through which the cur- 

 rent flows. The case therefore corresponds at once to type II, 

 figure 2, or to lines of stress relatively far apart as compared 

 with molecular distances. 



The presence of an electric current, however, is accompanied 

 with an accession of heat in the wire through which the cur- 

 rent flows, and the diminution of viscosity resulting! manifests 

 itself in my apparatus and perhaps inevitably, as a diminution 

 of rigidity. Hence in circular magnetization the heat effect is 

 superimposed on the magnetic effect of the same sign and in 

 thin wires it is difficult to separate them satisfactorily, if at all ; 

 for even if the wire carrying current is submerged in water 

 there is still grave room for doubt, as will appear below. 



For these and similar reasons I endeavored to investigate 

 results with thin wires subjected to the strong transverse fields 

 in the air gap of an electromagnet, for which conditions no 

 data have as yet been forthcoming. In spite of the toil spent 

 upon the work, however, my endeavors have not brought out 

 sharp results, except in so far as they furnish a superior limit 

 for the change of rigidity sought ; and even this is a small and 

 uncertain residuum raising a doubt as to whether rigidity is at 

 all influenced by strong transverse fields. 



Having completed this work I thus found myself under the 

 necessity of taking up the question of the effect of circular 

 magnetization on rigidity for comparison. The results were 

 again such as to place doubt on the occurrence of such an 

 effect. 



3. Method of Experiment. — The apparatus used was vir- 

 tually the same as that of my earlier experiments on the sub- 

 ject, except that the helix formerly surrounding the upper 

 wire is replaced by the fissure-like air gap of the strong tubular 

 electromagnet, A, figure 3. The upper of two identical soft 

 iron wires, ab and ca, is placed in this gap to be strongly mag- 

 netized when the current flows through A. Any twist may be 

 applied to the system ab, cd, by the torsion heads D and E, 



* Knott: Trans. Roy. Soc. Edinb., (2), xxxvi, pp. 485-535, 1891. 



f Maxwell: Electr. and Mag., ii, p. J 09. 



% Chrystal: Encyclop. Britanica, Art. Magnetism, p. 2*70. 



§ Drude: Wied. Ann., lxiii, p. 9, 1897. 



