266 



SCIENCE 



[N. S. Vol. XXX. No. 765 



other from hydrogen; a similar result was 

 obtained when carbon tetrachloride, or 

 mercury methyl, was used instead of 

 methyl-iodide. These and similar results 

 lead to the conclusion that the atoms of the 

 different chemical elements contain definite 

 units of positive as well as of negative 

 electricity, and that the positive electricity, 

 like the negative, is molecular in structure. 

 The investigations made on the unit of 

 positive electricity show that it is of quite 

 a different kind from the unit of negative, 

 the mass of the negative unit is exceedingly 

 small compared with any atom, the only 

 positive units that up to the present have 

 been detected are quite comparable in mass 

 with the mass of an atom of hydrogen ; in 

 fact they seem equal to it. This makes it 

 more difficult to be certain that the unit of 

 positive electricity has been isolated, for 

 we have to be on our guard against its 

 being a much smaller body attached to the 

 hydrogen atoms which happen to be present 

 in the vessel. If the positive units have a 

 much greater mass than the negative ones, 

 they ought not to be so easily deflected by 

 magnetic forces when moving at equal 

 speeds; and in general the insensibility of 

 the positive particles to the influence of a 

 magnet is very marked; though there are 

 eases when the positive particles are much 

 more readily deflected, and these have been 

 interpreted as proving the existence of 

 positive units comparable in mass with the 

 negative ones. I have found, however, that 

 in these cases the positive particles are 

 moving very slowly, and that the ease with 

 which they are deflected is due to the small- 

 ness of the velocity and not to that of the 

 mass. It should, however, be noted that M. 

 Jean Becquerel has observed in the absorp- 

 tion spectra of some minerals, and Professor 

 Wood in the rotation of the plane of polari- 

 zation by sodium vapor, effects which could 

 be explained by the presence in the sub- 

 stances of positive units comparable in 



mass with corpuscles. This, however, is not 

 the only explanation which can be given of 

 these effects, and at present the smallest 

 positive electrified particles of which Ave 

 have direct experimental evidence have 

 masses comparable with that of an atom of 

 hydrogen. 



A knowledge of the mass and size of the 

 two units of electricity, the positive and 

 the negative, would give us the material 

 for constructing what may be called a 

 molecular theory of electricity, and would 

 be a starting-point for a theory of the 

 structure of matter; for the most natural 

 view to take, as a provisional hypothesis, is 

 that matter is just a collection of positive 

 and negative units of electricity, and that 

 the forces which hold atoms and molecules 

 together, the properties which differentiate 

 one kind of matter from another, all have 

 their origin in the electrical forces exerted 

 by positive and negative units of electricity, 

 grouped together in different ways in the 

 atoms of the different elements. 



As it would seem that the units of posi- 

 tive and negative electricity are of very 

 different sizes, we must regard matter as a 

 mixture containing systems of very differ- 

 ent types, one type corresponding to the 

 small corpuscle, the other to the large posi- 

 tive unit. 



Since the energy associated with a given 

 charge is greater the smaller the body on 

 which the change is concentrated, the 

 energy stored up in the negative cor- 

 puscles will be far greater than that 

 stored up by the positive. The amount 

 of energy which is stored up in ordi- 

 nary matter in the form of the electro- 

 static potential energy of its corpuscles is, 

 I think, not generally realized. All sub- 

 stances give out corpuscles, so that we may 

 assume that each atom of a substance con- 

 tains at least one corpuscle. From the size 

 and the charge on the corpuscle, both of 

 which are known, we find that each cor- 



