﻿Mr. Norman Campbell on Delta Rays. 51 



while the upper part varied. Again, since the curve 

 with both electrodes covered with soot is flatter than 

 that with the electrodes covered with metal, it would be 

 expected that the lower half of the curve would be flatter 

 than the upper half. Both expectations are completely 

 falsified ; the most casual examination shows that if the 

 carves are to be divided into two halves, one of which re- 

 sembles more the curve for soot and the other resembles 

 more the various metals, it is the upper half which must be 

 associated with the soot and the lower half which must be 

 associated with the metals. No more striking refutation of 

 the theory could possibly be obtained. 



5. Now the theory which has been employed hitherto was 

 based upon the fundamental assumption that, when the 

 electrodes were at the same potential, all the rays emitted by 

 one were received by the oth^r ; and that the effect of 

 raising A (or B) to a positive potential relative to B (or A) 

 was merely to prevent some of the rays emitted by A (or H) 

 from reaching B (or A). If this assumption is correct, it is 

 clear that the part of the curve corresponding to positive 

 potentials of A (or B) must be determined by the speed of! 

 the rays leaving A (or B). But it was pointed out in the 

 last paper that it is possible that the electrodes possess so 

 high a reflecting power for the rays incident upon them 

 that the variation of the number ot rays received by them is 

 determined less by a variation in the number of rays which 

 can cross the space between the electrodes under different 

 electric fields than by the number which reach the electrode, 

 but are reflected by it, under different electric fields. In 

 this case, since the rays emitted by A are reflected by B, the 

 part of the curve corresponding to positive potentials on A 

 would be determined less by the properties of A than by the 

 properties of B. Accordingly it appears at first sight as if 

 the results which have just been recorded might be explained 

 on the assumption that the influence of reflexion upon the 

 form of the curve is so important as to mask completely any 

 effects due to the direct stopping of the rays by an opposing 

 field. 



There are many objections to this view which will probably 

 occur at once to the reader. Most of them rest upon 

 improbabilities concerning actions of which we know very 

 little, and are not therefore conclusive, though their accumu- 

 lated weight is considerable. But one objection appears to 

 me quite insuperable, and it alone will be put forward. 



It is known from Baeyer's work that soot shows very 

 little reflexion. Accordingly the curve obtained when both 



' E2 



