August 31, 1900.] 



SCIENCE. 



327 



of these same corpuscular charges. It 

 was first pointed out by Newton that if 

 the density of the ether continually in- 

 creased as we move away from a particle of 

 matter, that we should obtain a gravita- 

 tional effect. Later it was shown by other 

 writers, notably by Kelvin, that the same 

 result would follow if the density decreased. 

 No way of accounting for this continuous 

 variation of density has as yet been sug- 

 gested. Again, it was shown by Maxwell 

 that on any stressed medium theory of grav- 

 itation, the stresses must be enormous, 

 whilst the estimates given by Kelvin of the 

 elastic constants for the ether were not such 

 apparently as to permit of this. But the 

 writer showed, in the paper above referred 

 to, that the elasticity of the ether is im- 

 mensely great, i. e., 6 x 10™. Now if we 

 calculate, as I have done in one of the papers 

 referred to, what the diameter of the cor- 

 puscle must be, in order that it shall give 

 the inertia effect, and fx-om that calculate 

 the electrostatic stress at the surface of a 

 corpuscle, we find that it is of the order 

 10™, and this stress acting on a medium 

 whose elastic coefficients are as given, I 

 have found, can produce a change of den- 

 sity sufficient to give the observed gravita- 

 tional attraction. 



We thus find that both inertia and gravi- 

 tation are electrical effects and due to the 

 fact that the atom consists of corpuscular 

 charges. The constant ratio between quan- 

 tity of inertia and quantity of gravitation, 

 for a given body, is thus explained. We 

 may state the theory thus : 



The inertia of matter is due to the electro- 

 magnetic inductance of the corpuscidar charges, 

 and gravitation is due to the change of density of 

 the ether surrounding the corpuscles, this change 

 of density being a secondary effect arising from the 

 electrostatic stresses of the corpuscular charges. 



A fuller paper on this subject is in course 

 of preparation, but will be delayed for some 

 time by pressure of other work. 



I may here mention that I have found 

 that the equation 



3I/L'= M/LTx T'/L\ 



given in the paper in the Physical Review, 

 above referred to, and stated to represent a 

 phenomenon not yet discovered, really rep- 

 resents Kerr's electrostatic optical effect, 

 and the above gravitational efifect, and 

 that this effect therefore varies directly 

 with the elastic coefficient of the dielectric. 

 As this is one of the remaining links neces- 

 sary to complete the full chain of proof of 

 the theory there given, this latter is thus 

 put upon a still firmer footing.* 



The weight of matter in a gaseous state 

 should be very slightly greater than in the 

 solid state, and iron should weigh slightly 

 less when dissolved. It is doubtful, how- 

 ever, whether the expei-imental conditions 

 are not too difficult. If the measurement 

 could be made it would give an independ- 

 ent method of arriving at the size of the 

 corpuscle. 



The writer has pointed out that the Kel- 

 vin-Maxwell theorem, deduced from the 

 phenomenon of the electromagnetic rotation 

 of light, that whenever we have a magnetic 

 field we have also a rotation of the medium, 

 is incorrect, in that it assumes that light 

 consists of a certain kind of periodic mo- 

 tion for which there is no evidence. The 

 question arises : In spite of the fact that 

 the supposedly general theorem is incorrect, 

 is there any actual material rotation con- 

 cerned in the electromagnetic rotation of 

 light ? The answer I would give is ' yes, 

 but not as a cause, merely as an effect.' 

 According to the theory advanced by the 

 writer,t the rotation is a consequence of 

 light absorption, and can only take place 

 in an absorbing medium. When the light 

 waves strike the atoms, if the period of vi- 



* A Determination of the Nature of the Electric and 

 Magnetic Quantities. Phys. Bev., January, 1900. 



