236 SCIENCE AND METHOD. 



mutually attracting each other are generally not the 

 same, how can this coefficient depend upon these two 

 velocities ? 



Upon this subject we can but form hypotheses, but 

 we are naturally led to enquire which of these hypo- 

 theses will be compatible with the Principle of 

 Relativity. There are a great number, but the only 

 one I will mention here is Lorentz's hypothesis, which 

 I will state briefly. 



Imagine first of all electrons in repose. Two 

 electrons of similar sign repel one another, and two 

 electrons of opposite sign attract one another. Accord- 

 ing to the ordinary theory, their mutual actions are 

 proportional to their electric charges. If, therefore, 

 we have four electrons, two positive, A and A', and 

 two negative, B and B', and the charges of these four 

 electrons are the same in absolute value, the repulsion 

 of A upon A' will be, at the same distance, equal to 

 the repulsion of B upon B', and also equal to the 

 attraction of A upon B' or of A' upon B. Then if A and 

 B are very close to each other, as also A' and B', and 

 we examine the action of the system A -i- B upon 

 the system A' + B', we shall have two repulsions and 

 two attractions that are exactly compensated, and the 

 resultant action will be nil. 



Now material molecules must precisely be regarded 

 fis kinds of solar systems in which the electrons circulate, 

 some positive and others negative, in such a way tJiat 

 the algebraic sum of all the charges is nil. A material 

 molecule is thus in all points comparable to the system 

 A + B I have just spoken of, so that the total 

 electric action of two molecules upon each other 

 should be nil. 



