500 The Rev. T. R. Robinson's Experimental Researches on the 



greatly weakened by the induction which transmits it to the poles. As far as I 

 know, these disturbing causes have not been studied ; and I liope that the fol- 

 lowing results will not be without their value to the construction as well as 

 theory of the electro-magnet. 



The electro-dynamic laws, discovered by Ampere, Biot, and others who have 

 followed in their steps, may seem sufficient to establish its theory ; and it is 

 desirable to ascertain whether they succeed in doing so, for the simple case of a 

 circular or helical current. 



Let BEG be a differential slice of the magnet, E one of its elements, H a 

 current element below its 



plane, whose direction is 

 perpendicular to the plane 

 HA <?, the power of H to 

 attract and, as we may infer, 

 to masrnetise E is, as its 

 magnitude dc, as its energy, 

 fxF (F being the force of 

 the current), as the inverse 

 square of the distance HE^ 

 and as the cosine of the an- 

 gle CHE ; or putting HA 

 =z, AC — X, CE — y, DE — r, AD = b, and BDE — 6, the magnetism of E, 



_ , .F.dc.Ev{x^ + r) 

 '^^ - {x' + f + z')i ■ 



A difficulty occurs here however. The polarity given to the element E is 

 such, that its axis is perpendicular to HE ; but, as in the case before us, the 

 current surrounds the magnet, it is evident that unless z = Q, each portion of it 

 must produce a different axis. Is this physically possible in the molecules of 

 iron ? If so, their polarities must be very in-egularly distributed. Or do the 

 axes coalesce into one resultant, whose intensity may be estimated by the com- 

 position offerees? This seems most probable; and, therefore, assuming it, 

 and resolving dM' in the direction of z, 



fiFEdc X x 



dM = 



{x' + 2/- -h z')r 



