produced l>y Beta Rays. 785 



independent of the intensity of the Rontgen rays. The 

 largest possible error in the determination of the current for 

 a given P.D. was 3 per cent, of the maximum current when 

 Rontgen rays were used and 1 per cent, when a rays were used. 



4. The Rontgen rays used were heterogeneous primary 

 rays. No change in the measurements due to variations in 

 the hardness of the bulb could be certainly detected. Ex- 

 periments were made when the electrodes were covered with 

 gold, copper, and paper coated with graphite. The maximum 

 saturation current (measuring the total number of § rays) 

 for copper was about one half of that for gold ; for the paper 

 the current was so small that it could not be detected with 

 certainty and was certainly not greater than 1 per cent, of 

 that for gold. 8uch a result is, of course, to be expected; 

 for the /3 radiation excited by Rontgen rays in gold and 

 copper is known to be very much more intense than that 

 excited in paper. However, whether the electrodes were 

 covered with gold or copper, the relation between the current 

 and the potential difference was very nearly the same. It 

 may be concluded, therefore, that the quality of the £ rays 

 excited by ft rays, like that of the £ rays excited bv a rays, 

 is independent of the penetration of the primary rays and of 

 the nature of the material they strike. 



5. The results of the comparison of Rontgen rays with 

 a rays are shown in the figure below. The abscissae repre- 

 sent P.D.'s between the electrodes ; the unit is 1*02 volts. 

 The current was found to be completely saturated for a P.D. 40. 

 If iy is the current between the electrodes when one elec- 

 trode is at a potential V higher than the other, the ordinates 



represent the values of -^ ?— (see the previous papers). 



The continuous curve represents the results obtained 

 with Rontgen (i. e. /3) rays ; the points marked x represent 

 the results for a rays. It will be seen that the two series 

 of results are exactly the same for all P.D.'s greater than 2. 

 No significance important for our purpose can, I think, 

 be attributed to the slight divergence of the curves for small 

 P.D.'s*. In previous papers the uncertainties affecting this 



* From the figure it might not be seen that there are divergencies for 

 small potentials; tho steepness of the curve conceals errors in the 

 ordinates. .For values of V less than 1, the ordinate of the a ray curve 

 is about 10 per cent, greater than that of the Rontgen ray curve, while 

 for values of V greater than 2 the difference between the ordinates is 

 never more than 1 per cent. The curve makes errors in V of equal 

 value with errors in the ordinate, whereas experimentally the values of 

 V are certain to 1 in 3000. 



