$2 



Mr. 0. D. Child on th 



le 



questions can be answered definitely until further experi- 

 mental work kasjbeen done; and, since I do not have apparatus 

 making it possible to repeat Strutt' s experiments, I am taking 

 the liberty of pointing out the desirability of further investi- 

 gation of this subject. 



In regard to the first question, the following statement may 

 be made. Jn the experiments of Strutt, a current of a few 

 milliamperes at a high voltage was passed through mercury 

 vapour which had a slight admixture of some other gas. The 

 presence of the additional gas was found to be necessary in 

 order to produce the effects observed. In the experiments by 

 myself there was a current of several amperes passing through 

 a gas that was in all probability pare mercury vapour. The 

 necessity of some impurity in the former experiments would 

 indicate that two different phenomena were being studied. 

 On the other hand, the fact that the luminosity lasts for 

 practically the same length of time indicates that the cases 

 are identical. TVe should be in a better position to decide 

 this question it' the spectrum of the luminosity observed by 

 Strutt could be obtained. Unfortunately, the light was too 

 faint to allow r this to be done. 



Again, we could better decide regarding the identity of 

 the two phenomena, if it were certain that the light examined 

 by Strutt was not caused by recombination — that is, if his 

 experiments prove his conclusions. Certainly an explanation 

 different from that given by him can be suggested, as is shown 

 by the following discussion. 



Fi>r. 1. 



In Strutt' s experiments a discharge is passed through a 

 tube at the right of the part shown in fig. 1. A. is the anode 

 'for this discharge. In this anode there is a small opening, and 

 gas from the discharge is blown through this opening into the 

 tube shown in the drawing. P is the connexion to the pump; 

 g and h are electrodes to which an E.M.F. may be applied, so 



