December 18, 1896.] 



SCIENCE. 



899 



prove Branly's conclusions. The experi- 

 ments of the latter were repeated as nearly 

 as possible under the original conditions, 

 and indications of a loss of positive elec- 

 tricity were in fact found. But it was in 

 all cases possible to ascribe these results to 

 disturbances, as in Elster and Geitel's pre- 

 vious experiments. And when the experi- 

 ments were modified so as to more com- 

 pletely remove the sources of disturbance, 

 the apparent loss of positive electricity be- 

 came insignificant. In view of the fact 

 that the X-rays, whose discharging action 

 is in many respects similar to that of ultra- 

 violet light, are capable of dissipating both 

 charges, it seems probable that some effect 

 is also produced by light upon a positive 

 charge. The experiments of Elster and 

 Geitel indicate, however, that this action 

 must be extremely small. 



THEORIES PKOPOSBD. 



111. Although it has been shown, I think 

 conclusively, that the dissipation of a nega- 

 tive charge by light is accomplished by con- 

 vection, the theories heretofore cited do not 

 explain how this convection is brought 

 about ; and they also leave unexplained the 

 still more important fact that negative 

 charges only are affected. The theories 

 proposed by J. J. Thomson* and by Elster 

 and Geitelf respectively seem satisfactory, 

 however, in regard to both these difficulties. 



112. Thomson bases his explanation of 

 the phenomenon upon the hypothesis of 

 Helmholtz that "bodies attract electricity 

 with different degrees of intensity." This 

 conception was shown by Helmholtz to be 

 able to explain electrification by friction and 

 the differences of potential produced by 

 contact. If the attraction of a metal for 

 positive electricity is greater than that of the 

 dielectric surrounding it, the tendency is for 

 the metal to become positively charged. 



* Kecent Researches in Elect, and Mag., p. 64. 

 t W. A. 55, p. 697. 



But "when a conductor, which does not 

 disintegrate, is surrounded by air in its 

 normal state, * * * the conductor cannot 

 get charged * * * for the electricity of op- 

 posite sign to that which would be left on 

 the conductor has no place to which it can 

 go." "The case is, however, different when 

 the conductor is exposed to the action of 

 ultra-violet light, for then, as Lenard and 

 Wolf's experiments prove, one or both of 

 the following effects must take place : (1) 

 disintegration of the conductor; (2) chemi- 

 cal changes in the gas in the neighborhood 

 of the conductor which put the gas in a 

 state in which it can receive a charge." 



113. The first effect mentioned, namely, 

 disintegration of the metal, might be pro- 

 duced by the absorption of the light rays, as 

 explained in § 71. On the other hand, it 

 is quite conceivable that light, and es- 

 pecially the ultra-violet rays, may produce 

 electrolytic dissociation in the gas, and so 

 render the latter capable of taking a charge. 

 This hypothesis is strengthened by the 

 experiments of Arrhenius,-'' who found that 

 certain gases became conductors when 

 illuminated. That some such change in 

 gases is produced by X-rays is now proven 

 beyond question. 



114. The theory suggested by Elster and 

 Geitel depends upon the assumption that 

 the electro-positive character of the metal 

 relatively to the dielectric causes the for- 

 mation of two charges, viz : A positive 

 charge on the metal, and an equal negative 

 charge in the adjacent portions of the gas. 

 These two charges being extremely close to- 

 gether, their outside effect disappears. When 

 light waves fall upon the surface in such a 

 manner as to give a component electric dis- 

 placement (electro-magnetic theory) in the 

 direction of the normal, electric oscillations 

 will be set up in the metal. These oscilla. 

 tions may be sufficiently powerful to destroy 



*Wied. Ann. 32, p. 545 ; 33, p. 638 ; Phil. Mag. 



28, p. 75. 



