Dr. Tyii<Jall on the Laws of M<tgnetism. 293 



the opposite polarity implied by the fundamental law, the magnet 

 and the mass of soft iron mast be a certain distance asunder. 

 The case of contact, in point of fact, reduces itself to the attrac- 

 tion which the magnet exerts upon itself, and this leads us to 

 the threshold of a subject which 'vviU be -better discussed in a 

 future memoir. ' . l; i ..U -ilcUvUn^ ^ 



"b^' '!■wlil''not^^'^;m•ii''for a'short' tite tV'tlie'bBnyid^ration of 

 the divergences which exhibited themselves on changing our 

 mode of experiment (18.). It has been satisfactorily shown (21.) 

 that they cannot be accounted for on the supposition of ' lagging 

 magnetism.' Let us imagine the ball placed at the sixth layer 

 of the gun-cotton solution. From this downwards, as it ap- 

 proaches the pole, we have a nearer and nearer approximation to 

 the law of simple proportionality. "WHien a single layer separates 

 the ball from the pole, the law is nearly fulfilled ; when this is 

 removed, the approximation is still nearer; but is the ball then 

 in contact ? Will not the unavoidable roughness which remains 

 after in^ibbing with fine sand-paper and polishing with emeiy still 

 make some difference ? If we lay a plate of glass upon a convex 

 lens of large radius, so as to produce Newton's rings, a consi- 

 derable pressui'e is needed to render the central spot permanently 

 black, or in other words, to establish perfect contact*. How 

 much more will such a pressure be required in our case, where the 

 surfaces operated with are comparatively so defective ! Besides 

 this, the manipulation of the surface must alter in some measui-e 

 the constitution of the thhi outside layer, and render it different 

 from that of the mass of the metal. This reniark applies to both 

 magnet and ball. An improvement of the contact would, accord- 

 ing to our reasoning, be accompanied by improved results. The 

 contact might be made more perfect by mechanical pressure — 

 with the hand, for instance; but on removing the hand a shaking 

 of the ball is xuiavoidable, which entirely nullifies the previous 

 pressiire. Besides, this pressure may not, and in all probability 

 will not, be exerted in the ])recise direction of the magnetic force ;, 

 and in this case, when the hand is raised, the ball will right 

 itself and thus defeat us. The best means of improving the con- 

 tact seems to be furnished by the magnet itself. If the ball be 

 squeezed tight, in the first instance, by strong magnetic power, 

 the pressure thus exerted is sure to be in the right direction. 

 This, in fact, is what has been done in our first experiments ; 

 and these, as we have seen, exhibit 'aimb^t; striking coincidence 

 with the laws affirmed in each case. 



139. The imperfections of th<; Hurface ajipcar, therefore, to 

 throw the tme jtole, or attracting plane, some distance beneath 

 * See Phil, Mag. for December' 185(y, p. ^^i '"' ""^""^^^ 



