﻿Mr. Norman Campbell on Delta Rays. 53 



rays leave the electrode and that the action of the field is to 

 drag some of the rays out o£ the electrode at which they are 

 generated, the main feature is explained. For, if the effect 

 of positive potentials on B is not to stop the rays from B 

 generated in B from reaching A, but to aid the rays gene- 

 rated in A in reaching B, then it is clear that the branch 

 of the curve corresponding to potentials positive on B must 

 be determined by the properties of A and not by the properties 

 ofB. 



In the case of soot this new hypothesis has considerable 

 a priori probability. For Baeyer has shown that the reflecting 

 properties of a material depend on the condition of its 

 surface rather than upon its chemical nature. For instance, 

 metallic platinum is a good reflector, platinum-black a bad 

 reflector. It is natural, therefore, to suppose that the absence 

 of reflexion is due to the entanglement of the electrons in 

 the rough surface of the material, and that the slow-moving 

 electrons generated at the surface of the soot-covered electrode 

 will not be able to escape from it completely unless they are 

 aided by an electric field. We should expect, as we find, 

 that it would require a greater electric field to drag a given 

 proportion of the rays out of a surface of soot than out of a 

 metallic surface, and, consequently, that the curve in fig. 1 

 for soot should be flatter than those for metals. 



7. When one of the electrodes is covered with soot some 

 explanation of this nature appears absolutely necessary. The 

 question next arises whether it can be applied to explain the 

 differences between the curves obtained with different metals. 

 Since no metal surface is perfectly smooth and no metal is a 

 perfect reflector for slow-moving electrons, it is to be 

 expected that the results with metals will also be affected to 

 some extent by the cause which has been suggested in 

 connexion with soot. It is true that the curve for aluminium 

 is always steeper than those for the other metals, whereas 

 Baeyer found the reflecting power of that metal to be some- 

 what lower than that of the others, but there may be causes, 

 such as electric fields in the surface, which would affect 

 electrons generated in the surface but not electrons moving 

 up to the surface and reflected by it. On the other hand, 

 the supposition that the form of the curves is determined 

 largely by the condition of the surface of the electrodes would 

 readily explain the difficulty which has been experienced in 

 reproducing accurately a series of readings after small 

 alterations have been made in the apparatus. 



On these grounds it seems to me much more probable 

 that the differences obtained with electrodes of various' 



