530 The Rev. T. E. Robinson's Experimental Researches on 



out the magnetism, as may be inferred from its being less than half^o. 154. 

 The same may be observed in Table viii., No. 45, for which magnet I had a 

 remarkable proof of it, trying the effect of a single pair of spires with 1^=13-56 : 

 L was only 12-97, only 0-071 of No. 45, while the -f is 0-272. In the second 

 and third all are represented in which -^ is positive. In the fourth the highest 

 16 of the Table were taken: the agreement here is the worst, caused by the 

 aberrance of the highest one, which is very discordant, owing to the magnet 

 having been, for some time before it was taken, only slightly excited. In such 

 cases I have always found a great excess. The fifth is probably too far from 

 the maximum, and the sixth would be much improved by omitting No. 58, 

 which is not sufficiently excited. 



From the fragmentary character of this Paper, and those which preceded 

 it, it may be useful to conclude with a brief summary of the principal results 

 which they contain. 



An electro-magnet, whose circuit is closed, differs from one whose poles are 

 not connected, in having little free magnetism, but a much higher polarization of 

 its molecules. 



In this condition the magnetism of each molecule is in equilibrium under 

 the action of four antagonist forces : — 



1. The direct action of the helices on each molecule. 



2. The action of the other magnetized molecules transmitted by induction 

 through the iron : these tend to keep the two polarities apart. 



3. The coercive force, which in general seems to resist any magnetic change. 

 And 



4. The mutual attraction of the polarities, — oppose their separation. 



If, when the magnet is thus excited, the current which passes through the 

 helices be stopped, the force 1 vanishes, 4 prevails over 2, which therefore 

 decreases, but which is now assisted by 3 opposing the decrease, as it had done 

 the increase of polarity, and a new magnetic equilibrium occurs, which conti- 

 nues until the keeper is raised, when almost the entire of the force 2 is destroyed. 

 This "residual excitation," whose force I denote by A, is of a given amount as 

 long as the previous excitation can produce a force greater than half the maxi- 

 mum power of the magnet ; below that it decreases. It is not proportional to 

 the maximum ; and if a current be passed while the magnet is in this state, the 



