138 THE ROYAL SOCIETY OF CANADA 



As every precaution was taken to maintain uniform conditions 

 throughout the experiments, it seems reasonable to suppose that in 

 so far as I and II are concerned the ratio of the intensity of the radia- 

 tion of wave lengths X = 5460.97 Â.U. and X = 4358.66 Â.U. that 

 issued from the quartz tube to the intensity of the radiation of these 

 wave lengths issuing from the glass tube was the same when both 

 tubes contained mercury vapour as when they were empty. It seems 

 fair then to draw the inference that the increase invariably observed 

 in the intensity of the radiation of wave lengths X = 5460.97 A.U. and 

 X = 4358.66 A.U. issuing from the quartz tube when the tubes con- 

 tained mercury vapour was due to some cause other than those 

 referred to in I and II. As glass is opaque to all radiations shorter 

 than X = 3000 A.U. while quartz is transparent to radiations with 

 wave lengths as short as X= 1800 A.U. one is forced to the conclusion 

 that the observed increase in the intensity of the wave lengths X = 

 5460.97 Â.U. and X ^ 4358.66 Â.U. was due in some way to wave 

 lengths shorter than X = 3000 Â.U. 



It is known that mercury vapour strongly absorbs radiations of 

 wave lengths at or near X = 2536.72 A.U. and at or near X = 1849.6 A.U. 

 On this account one would naturally seek to associate the observed 

 increased emission of the radiations in question with the absorption 

 by the mercury vapour in the quartz tube of radiations of the wave 

 lengths X = 2536.72 Â.U. and X = 1849.6 Â.U. From the work of 

 R. W. Woods Steubing- and others, it is known that mercury vapour 

 exhibits strong fluorescence when light of the wave lengths X = 2536.72 

 Â.U. and X= 1849.6 Â.U. is allowed to fall on it and that the fluores- 

 cence spectrum produced under these circumstances includes a 

 symmetrical structureless band extending from the red down to the 

 wave length X = 3700 Â.U. with its maximum at 4850 Â.U. 



It was therefore thought at first that the observed increased 

 intensity of the radiation of wave lengths X = 5460.97 A.U. and 

 X = 4358.66 Â.U. might have been due to the fluorescence of the 

 mercury vapour produced by either or both of the wave lengths 

 X = 2536.72 Â.U. and X = 1849.6 Â.U. Numerous trials showed 

 however, that in order to obtain distinct photographs of the X = 4850 

 A.U. fluorescent band from mercury vapour, which was made to 

 fluoresce as strongly as possible, exposures of long duration extending 

 from 3 to 5 hours were required when the fluorescent light was focussed 

 directly on the slit of the spectrograph and when no absorption 

 screen was inserted in the path of the light. Even with such exposures 



I R. W. Wood, Physik. Zeit. X. 425 u. 466, 1909. 

 ^Steubing, Physik. Zeit. X. p. 787, 1909. 



