434 Mr. A. A. Campbell Swinton. 



observations it was connected to a mercury pump, so that the degree 

 of exhaustion could be varied as desired. 



FIG-. 2. 



3 6 7 63 K) II 12 



The following are the main effects observed. 



(1) When the anti-cathode intersects the cathode stream at the 

 focus, the dimensions of the active area are independent of the 

 degree of exhaustion. For all other positions beyond the focus it is 

 larger the lower the exhaustion and vice versa. These observations 

 are of course only possible between the limits of exhaustion with 

 which K-ontgen rays are produced. 



(2) When the anti-cathode intersects the cathode stream beyond 

 the focus, the active area is larger the greater the distance between 

 cathode and anti-cathode. For instance, with the tube illustrated in 

 fig. 2, exhausted to a good Rontgen ray vacuum, it was found 

 that the active -area gradually increased from about 0'15 cm. 

 diameter with 4 cm. distance between cathode and anti-cathode up 

 to about 2'3 cm. diameter as the distance was gradually increased 

 to 14 cm. The increase is less the higher the vacua, but is always 

 very considerable. 



(3) When the anti-cathode intersects the cathode stream con- 

 siderably beyond the focus, the active area is found to consist of a 

 well defined and very intense central nucleus, surrounded by a much 

 fainter but quite appreciable halo. Both of these increase in size as 

 the distance between cathode and anti-cathode is increased. 



In some cases the halo consists of a well marked hollow ring with 

 a dark space between it and the central nucleus. In other cases two 

 distinct concentric rings are visible surrounding the nucleus. 

 Moreover, the nucleus itself, when very large, shows distinct signs 

 of being made up of one or more concentric rings, sometimes with a 

 still smaller nucleus within them. These observations correspond 

 with and amplify what the writer has already noticed by direct 

 observation of the visible luminescence of a carbon screen arranged 

 to intersect the cathode stream.* 



(4) With an anti-cathode inclined at an angle of 45 to the 

 axis of the conical cathode stream, it is found that those portions 

 of the stream which impinge most normally upon the anti-cathode 



* ' Eoy. Soc. Proc.,' vol. 61, pp. 8184. 



