402 



Mr. A. LI. Hudies on the 



adopted a magnetic field was used to measure the velocities. 

 The apparatus is shown in fig. 4. 



& U-Lctx4K.CC 



tb iAxxJL urrruXssi; 



A flat cylindrical box is divided into quadrants by thin 

 brass partitions. In three of these, slits 10 mm. by i mm. 

 were cut. The radius of the circle passing through the 

 centres of the slits was 1*1 cm. The illuminated plate P 

 was connected to an electrometer and capacity by means of 

 which the total leak from the surface could be measured 

 when required. The surfaces inside the box were all covered 

 with soot. 



The source of light of short wave-length was a discharge 

 in hydrogen in the tube G. This was separated from the 

 brass box by a fluorite window F which transmitted light 

 down to \ 1330. The superposed long wave-leno-th ultra- 

 violet light was obtained from a mercury arc in a fused 

 quartz tube placed close to the quartz window Q. The 

 shortest wave-length transmitted by the fused quartz was 

 X1930. 



The distribution of velocities for each of the two sources 

 of light was first obtained. This was done by measuring the 

 charge acquired by M for different magnetic fields per 

 pendicular to the plane of the box. Although the light was 

 as intense as could be obtained under the conditions, yet the 

 charge communicated to M was always small. In some 

 cases the leak was measured with the tilted electroscope at a 

 sensitiveness of 600 divisions per volt. 



The velocities were obtained by the formula 



llev = 



R 



= l-76xl0 7 , E = l-1 



cm, 



The distributions of velocities are given [ u g. 



the total 



