524 Mr Campbell, Discontinuities in Light Emission. 



than 1. On the first assumption, which leads to 'q^=o), the value 

 of ft) is 3'1, a quantity which agrees well in order of magnitude 

 with that found by totally different experiments. (See § 7.) It 

 also agrees, at least so far as order of magnitude is concerned, 

 with that to be expected from Planck's theory of radiation, if, and 

 only if, the bundle of energy theory is true. If the spherical 

 wave theory is true, and the energy in a single light disturbance 

 is spread over a sphere surrounding the source, the above value 

 for ft) must be multiplied by more than 100 in order to take 

 account of the angle subtended at the source by the photoelectric 

 cell. 



It appears to be the chief importance of this work that it 

 has shown that it is possible, by the fluctuation method of 

 V. Schweidler, to obtain directly measurements of quantities 

 agreeing well with those deduced on other and much more 

 indirect methods. 



Summary. 



1. The paper is a continuation of that presented lately to the 

 Society under the same title. The difficulties which had hitherto 

 prevented the attainment of results are described, 



2, 3. A description of new sources of light designed to over- 

 come these difficulties, and an explanation of their nature. 



4. Reasons are given for the failure of the new devices to 

 attain the desired object. 



5. With this failure all hope of carrying out the main purpose 

 of the research was abandoned. 



6. It was thought, however, that some results of value might 

 be obtained by the observation of the fluctuations of two indepen- 

 dent sources of light. The method of measurement is described. 



7. The variation of the fluctuations due to such sources with 

 the total intensity of the light is first examined. It is pointed 

 out how far such measurements could be used to distinguish 

 between the rival theories of light. The conclusion is reached 

 that the experiments recorded here afford no grounds for dis- 

 tinguishing between these theories. They can only fix a limit 

 to certain values which is in accordance with deductions from 

 other work. 



8. 9. The variation of the fluctuations with the instrumental 

 constants is examined. It is concluded that the agreement 

 between experiment and the theory worked out in a previous 

 paper is satisfactory, except in cases where the value of the 

 resistance R is small. Reasons are given for thinking that these 



