558 



Mr. A. Schuster. 



decomposed at the kathode recombine in this dark interval. If, as 

 seems more probable to me, it should be found that the negative ions 

 diffuse more rapidly, the recombination will in part take place at the 

 anode. If the conditions in the tube are such that the gas may divide 

 into layers, such that in alternate strata the decompositions outnumber 

 the recombinations, and vice versa, stratifications will form. 



Such is the general outline of the theory, which may have to be 

 modified in detail, but which, I believe, has a strong element of truth 

 in it. 



The possibility of a volume electrification is denied by some of 

 Maxwell's disciples, who look on a current of electricity as on a flow of 

 an incompressible liquid in a closed circuit. But there is nothing, as 

 far as I can see, in the conclusions I have drawn from the gas dis- 

 charges which is inconsistent with the fundamental tenets of 

 Maxwell's theory, however much they may disagree with the acces- 

 sory embellishments with which that theory is occasionally adorned. 

 There may be a volume electrification without interfering with the 

 equation of continuity of an incompressible liquid as long as we 

 admit the possibility of displacement currents and displacements in 

 conductors, and I see nothing improbable in this. The ordinary 

 equations for the currents in a non-homogeneous solid (or any solid if 

 inequalities of temperature are taken into account) give a volume 

 electrification which can only be destroyed by the introduction of 

 a quantity which is analogous to hydrostatic pressure, and the sole 

 purpose of which is to destroy all electrifications except at the surface 

 of bodies. We know of no physical phenomena w r hich can justify the 

 introduction of such a quantity, which seems to me unnecessary. The 

 existence of a volume electrification can be shown to exist when a 

 current passes from one liquid to another floating on its surface. 

 Chemical effects are observed in the region in which the liquids 

 begin to mix, and these can be explained by the electrification which 

 accompanies each change in electric conductivity. Maxwell's equa- 

 tions assume conductors to be homogeneous throughout ; whenever 

 we are dealing with average effects only, this assumption is justified. 

 We deduce, in a similar way, the equations which represent the 

 transmission of light by assuming that each transparent body is 

 replaced by a homogeneous medium having certain properties. But 

 although this simplification is allowable in discussing some of the 

 phenomena, there are others in which it becomes necessary to go a 

 step further, and, considering the structural constitution of the body, 

 to take into account separately the effects of the medium separating 

 the atoms, and the effects of the atoms themselves. In all branches 

 of physics we are gradually forced by the advance of knowledge to 

 abandon the assumption of homogeneousness, and if that is done, 

 no further difficulty stands in the way of bodily electrifications ; for 



