152 



Prof. A. LI. Hughes on Characteristic 



(alternately increasing and decreasing) was found and 

 plotted. In this way slight accidental variations dis- 

 appeared^ while the break showed up distinctly. It is 

 fortunate that the lines above and below the break-point are 



Fiar. 3. 



140 



160 



180 



200 



220 



240 



so nearly straight, otherwise it would have been almost impos- 

 sible to detect a slight discontinuity in the slope of the curve. 

 Both types of apparatus (I. and II.) were used for boron. 



Five curves (each the mean of four, six, or eight consecu- 

 tive runs) were obtained for boron giving breaks at 150, 

 150, 145, 153, and 145 volts. We may take the mean value 

 to be 148 volts. This is regarded as the energy of the 

 electron necessary to call out the K-radiation of boron, and 

 by the quantum relation it corresponds to a wave-length 

 X 83-5. 



A number of observations were carried out in the region 

 10 to 40 volts in the hope of locating the L-radiation. 

 Fig. 4 shows the average of six consecutive runs, each 

 showing a break between 24 and 26 volts, the break in the 

 mean curve being at about 24*5 volts. This corresponds to 

 a wave-length X 505. 



Carbon. — Carbon was used in the form of natural graphite, 

 the specimen from which the targets were cut was said to be 

 99'5 per cent, carbon. ■ Experiments were carried out in the 

 same way as on boron (apparatus II. being used). Typical 

 curves are shown in fig. 5. The breaks which are taken 

 to correspond to the K-radiation of carbon occur in four 

 different curves at 210, 220, 213, 216 volts, giving a mean 

 at 215 volts. This corresponds to a wave-length \57*5. 



