742 



SCIENCE. 



[N. S. Vol. XII. No. 307. 



by the results obtained on investigating the 

 relations between F and h. 



A number of additional pieces of corrob- 

 orative evidence were also given, i. e.: 



3. The relation between the magnetic 

 constant a and the elasticity. 



4. The relation between this constant and 

 elastic strain. 



5. The relation between this constant and 

 permanent strain. 



6. The relation between this constant and 

 hysteresis. 



7. The relation between k and the den- 

 sity of substances. 



Several phenomena were also predicted, 

 i. e.: 



A. A change in the velocity of light, along 

 a slope of electric potential. 



B. A relation between refractive index 

 and piezo-electric effect in doubly refracting 

 substances. 



These have not yet been confirmed, but 

 arrangements ai-e being made to investi- 

 gate the former. 



This same result, originally obtained by 

 qualitative mathematics, can also be ob- 

 tained by Lagrangian methods. By con- 

 sidering the way in which permeability and 

 specific inductive capacity are affected in 

 the case of stressed iron, and in the case of 

 Kerr's phenomenon, as influenced by the 

 elasticity of the material, it can be shown 

 that a change in /^ involves the first power 

 of the magnetic intensitj' and that a change 

 of k involves the second power of the elec- 

 tric intensity. This proof will be given 

 later. It, however, in reality, adds nothing 

 to the proof already given, which in the 

 opinion of the writer is of such a character 

 that we may say that the nature of elec- 

 tricity and magnetism is now definitely 

 and finally determined, though no doubt it 

 may be years before the absolutely decisive 

 nature of the proof is generallj' appreci- 

 ated. 



Next, it follows from the writer's experi- 



ments on the relation between Hand fi that 

 the presence of matter does not alter the 

 elasticity of the ether by as much as one- 

 fourth of one per cent. Also, knowing now 

 that i: is a density, we are enabled to say 

 that aberrational and other optical phe- 

 nomena show that the density of the ether 

 is not appreciably altered by the presence 

 of matter, otherwise the (n'' — 1) and sim- 

 ilar formulae would not hold. From these 

 facts we see that the actual volume of the 

 atom, compared with the space occupied by 

 it, must be quite small. 



The diameter of the mercury atom I have 

 shown to be 2.75 (±0.2)x IQ-', and in 

 1899 I showed that the actual cross section 

 of the space actually occupied by the atom 

 must be less than one four-hundredth of 

 the space occupied by the atom to the ex- 

 clusion of other atoms, and that the atoms 

 ' must have a configuration analogous, in its 

 efiects, to that of structures of thin plat- 

 inum wire, suspended in oil.' 



Later, J. J. Thomson, from his beautiful 

 and wonderfully ingenious work on electric 

 discharges in gases, was able to show that 

 the atom is made up of a number of smaller 

 bodies, which he calls corpuscles. 



On comparing the results of Thomson, 

 Ewers, Kaufmann, Lenard, Lorenz, Wie- 

 chert and Simon, we arrive at the conclu- 

 sion that there are about 1,000 corpuscles 

 in a hydrogen atom, and that the weight of 

 a corpuscle is therefore about 1.5 x 10 ""gm. 



Since, then, there are about 200,000 cor- 

 puscles in a mercury atom, and their cross 

 section is less than one four-hundredth part 

 of the cross section of the mercury atom, 

 we find that the diameter of the corpuscle 

 is certainly less than 2.10 ~" cm. 



From J. J. Thomson's formula for the 

 electrically produced inertia of a charged 

 sphere, we find, as was shown by Thomson 

 (and independently by the wi'iter), that if 

 the diameter of the corpuscle is approxi- 

 mately 10~" cms., the ionic charge which 



