348 Mr. H. A. Rowland on Magnetic Distribution. 



cular arrangement at each point is determined solely or princi- 

 pally by electric force at the point, we cannot easily suppose the 

 new structure in a dielectric, originally isotropic, to be any thing 

 else than uniaxal, symmetrical at each point with reference to 

 the line of force through the point. And, even in the case of an 

 seolotropic body, we may assume, as a simple and sure approxi- 

 mation to the truth, that the effect of electric force is to super- 

 induce a uniaxal structure upon the primitive structure. 



The uniaxal structure thus induced by dielectrization has been 

 experimentally detected and characterized by birefringent action 

 in three cases. As a matter of fact, it appears to be negative in 

 glass and quartz, but positive in resin. 



The electric force has probably a certain resistance to over- 

 come, something analogous to coercive force in the case of mag- 

 netism. A sensible time is therefore required for the develop- 

 ment of the uniaxal structure by electric action, and for its dis- 

 appearance after the electric action has ceased. Under an in- 

 tense and long-sustained electric force, the new structure of the 

 dielectric may assume the character of a very stiff and perhaps 

 permanent set, analogous to permanent or subpermanent mag- 

 netism (18). We shall see afterwards, as might indeed be 

 expected, that there is nothing similar to this in the phenomena 

 presented by dielectrified liquids. 



Contrary electrizations rapidly succeeding one another exert 

 contrary actions of electric polarization, but conspiring actions 

 of molecular arrangement : they are therefore as effective as a 

 continued electrization in one direction; and aRuhmkorfFs coil 

 is as effective as an electrical machine of equal strength. 



I have made some experiments, and have had a good many 

 reflections, bearing on other explanations of the phenomena ; 

 and I think it not unlikely that strains due to the mutual actions 

 of intensely charged shells of the dielectric, or strains due to 

 changes of temperature, may have something to do with the 

 facts. But in the meantime I offer the preceding remarks as a 

 sketch of what appears to me to be the only probable theory. 

 Glasgow, September 22, 1875. 



XLI. Studies on Magnetic Distribution. By Henry A. Row- 

 land, of the Johns Hopkins University, Baltimore, Md. 3 U.S.A. 



[Continued from p. 2//.] 



V. 



LET us now consider the case of that portion of the bar which 

 is covered by the helix. First of all, when the helix is sym- 

 metrically placed on the rod, equations (5) and (6) will apply. As 



