174 ROYAL SOCIETY OF CANADA 



AVe liîive found il a mutter of great im])ortance for olttainini;- clear 

 and brilliant photographs, to remove the least trace of water vapour from 

 the tube. The ]»resence of water vapour has the etfect of making the 

 focus point much less sharp. an<l the sparking very irregular. The phos- 

 iihorescence of the glass is much less brilliant, with occasional sectorial 

 flashes from the back of the kathode. Shadows of bones on the fluoros- 

 cope are much less clearlj^ defined, and appear to be nearly' as transparent 

 as the flesh. If much water va])Our is present, the platinum plate 

 appears to be more highly heated for the same power, than if the air in 

 the tube is dry. The vacuum at the sparking limit, appears to be much 

 hioher, as measiired by the McLeod gauge, 002 to 005 mm. as compared 

 with -010 to "OoO mm. for dr;y aii\ but these figures, owing to the extreme 

 slowness of diffusion, and the absor))tion of water vapour by the phos- 

 phoric anh3-dritle in the pump, do not necessarily represent the actual 

 vacuum existing in the tube. That these effects are to be attributed to 

 the presence of water vapour, is rendered |)robable by the fact that they 

 are alwavs oVjserved on exhausting a new tube, if precautions have not 

 been taken to dry it. and that they disa])])ear if a small quantity of di-y 

 air is admitted and the tube re-exhau!;;ted. The presence of water may 

 also be verified by spectroscopic observations. 



We have observed these appearances in X-ray tubes of various 

 makers. Tht;y have generally been explained by other observers as being 

 due to X-rays of different kinds, or to dirterent degrees of vacuum, or to 

 different kinds of discharge. No doubt these explanations arc often 

 true. An increase in the vacuum certainly increases the relative trans- 

 parency of the bones, and the ettects are often considerably modified by 

 any change in the character of the discharge. But so far as our exper- 

 iments go, the ])resence of water vapour is a much more serious source of 

 disturbance, and attords in many cases a mucdi more likely explanation of 

 the irregularities. 



Method of Operating.— For the sake of more completely investi- 

 gating the effects with different gases, and under different conditions, we 

 have i)referred to keep the tube permanently connected with the pump. 

 Inci<lentally this method possesses the advantage that it is i)os>ible to work 

 the tube continuously at its point of highest efficiency for any length of 

 time. With a good vacuum and a powerful discharge, the air in the tube 

 ajipears to get used up so rapidly that the spark soon refuses to pass 

 otherwise than outside the tube. With a sealed tube, it is generally 

 necessary to stop at intervals and warm the tube. We found it prefer- 

 able, however, to sup})ly small quantities of dr3^air occasionally through 

 a tap connected with t he pump. On one occasion, the tube was operated 

 almost continuously in this manner for nearly two hours, including one 

 exposure of over an hour without any intermission. The admission of 

 air was so adjusted that the discharge took an alternative path by a six 



