276 PROCEEDINGS OP THE AMERICAN ACADEMY 



a magnet, — in either case, each individual atom has its own magnetic 

 field, which is, necessarily, always with it. It is really its reaction 

 upon the ether. If such atoms be elastic, as there is the best of rea- 

 sons for believing, then it follows that impact must set them into 

 periodic vibratory motion, that is, periodic change of form at a rate 

 depending upon its degree of density and elasticity. Such changes 

 of form set up corresponding periodic waves in the ether, as changes 

 in the magnetic field, and these are transmitted outwards with a 

 rate depending upon the properties of the ether to transmit such 

 motions, not upon the source of the disturbance. 



Such vibratory motions among atoms and molecules we call heat, 

 and such periodic waves in the ether we call light, and thus Maxwell's 

 idea of light being an electro-magnetic phenomenon is altogether 

 in accordance with the experiments. For waves of the lengths of 

 light waves, it is essential that the vibrating body be small and highly 

 elastic. Maxwell's idea was, that the opposite phases of ether waves 

 could produce opposite electrical effects, so that each half-vibration 

 represented either positive or negative conditions; and these implied, 

 though I have not noticed the statement, that they must have origi- 

 nated with vibrating magnetic atoms or molecules. It has been diffi- 

 cult or impossible heretofore to imagine how ether waves could be 

 set up by vibrations of the elements, though the idea that the atoms 

 of matter are magnets is not new at all, and has a good degree of 

 probability. If one is to picture to himself at all how this kind of a 

 phenomenon can occur, he is bound to have in mind some form 

 for an atom that shall at the same time be a consistent magnetic 

 form. If atoms are magnets, it is welluigh inconceivable that they 

 should be spheres or cubes, or tetrahedra or disks, or any of the or- 

 dinary geometric forms, for such would be very poor forms to exhibit 

 magnetic properties. But a ring presents a very different case, as a 

 ring magnet is the most perfect form possible. There is this to be 

 said of such a form, however. It does not present what we com- 

 monly call a magnetic field ; it is a closed circuit. Nevertheless, I 

 would ask if it is probable that the ether external to a magnet of that 

 form should be quite unaffected, quite neutral. I should supjiose not, 

 but, on the contrary, should look for some sort of stress there, though 

 it might be of somewhat different nature, and have somewhat different 

 properties, from an ordinary magnetic field. But if such were the 

 case, it follows that any magnetic change in the ring magnet itself 

 would be followed by a corresponding change in the external field, and 

 vibratory motions would necessarily set up waves in that field. Such 



