452 



produced, the distribution and properties of the internal medium 

 may still be such as to cause the cube to be attracted least, or repelled 

 most strongly when the force acts on the line of compression, and 

 thus, if the substance be diamagnetic, to cause it to agree, in its 

 deportment, with experimental results. On the other hand, if these 

 hypothetical properties of the internal medium be discarded as arti- 

 ficial or inadmissible, then at present I see no way of escaping the 

 conclusion of Professor Tyndall's argument. 



With regard to the explanation given by Professor Williamson, it 

 will be observed that he pursues a path quite different from that of 

 Professor Tyndall, when he considers the effects produced by com- 

 pressing a number of particles surrounded by a magnetic medium. 

 This compression, he states, may alter the attraction of the mass 

 by a magnet in two ways ; " first, by altering the density of the 

 matter ; secondly, by altering the density of the medium." By the 

 term 'density of matter' is usually understood the relation which 

 exists between the quantity of matter which a body contains, and 

 the volume of the space enclosed by its external surface. But in 

 the present case, where a comparison is instituted between the 

 matter of the body and the medium which is supposed to fill all its 

 pores, we must, I imagine, understand by the term ' density of 

 matter/ the relation which exists between the sum of the masses of 

 the particles, and the sum of their volumes ; but if so, then, the par- 

 ticles themselves being incompressible, it is clear that compression 

 could not alter the ' density of matter.' 



As to the second effect of compression, viz. an alteration of the 

 density of the medium, it may be quite conceivable, although I do 

 not find that Professor Williamson has any where taken it into con- 

 sideration. The effects of compression may, therefore, be more cor- 

 rectly described as either first, a diminution of the interstices of 

 a body, without altering the density of the medium which fills 

 them ; or secondly, a diminution of the interstices, accompanied by 

 an alteration of the density of the medium within them. 



Let us assume, as Professor Williamson has virtually done, that 

 the first of these effects takes place ; then, if we admit that " in a 

 cubical mass of carbonate of iron the material particles are more 

 magnetic than the medium which they displace, and the force with 

 which it is attracted is proportional to this excess," we can by no 



