772 Mr. H. Pealing on Distribution and Quality of 



the frame. The ratio =- was found to be 1 : 1*00; and when 



J 90 



the experiment was repeated the ratio was found to be 

 1: 1*004, no correction being made. This agreed very well 

 with the value found by Barkla and Ayres * after making 

 the necessary corrections. 



There are two corrections which do not vary with the 

 hardness of the rays. 



(a) The observed value of I 90 will always be greater than 

 that given by theory because the radiator subtends a consi- 

 derable angle to the window of the electroscope, so that when 

 the axis of the electroscope is 90° to the direction of the beam 

 the rays entering the window from the edges of the carbon 

 make angles considerably less or more than 90° with the 

 beam. The average of (1 + cos 2 a) for all the rays was calcu- 

 lated and found to be 1*015. The errors due to this cause in 

 the direction a = 150° or a = 40° are negligible. 



(b) The window of the electroscope subtends a considerable 

 angle in the vertical direction with the radiator. This will 

 make the value of I 150 observed less than theory predicts. It 

 was calculated that the average value of (1 + cos 2 a) for ail 

 the rays when the axis of the electroscope made an angle of 

 150° with the rays was 1*735. 



We should expect that owing to all these corrections the 



value of t — observed for the scattered rays should be 1*75 

 I90 



when the beam is soft and 1*71 when the beam is hard. 



As it would have been very laborious to have obtained the 



variation of -A for a large number of directions, the investi- 



gation was confined almost wholly to two values of a, namely 



150° and 40°. The results for ~ and ~ should give results 



-1-90 190 



typical of all values of u greater than 90° and less than 90 '* 

 respectively. 



These angles give ratios whose values by theory should be 

 1*75 and 1*58 respectively. 



VI. Results. 



(a) For values of ^~. 



J 90 



Experiments were tried with the bulb very soft. It was 

 found invariably that the ratio observed was lower than 

 1*75; the results varying from about 1*55 to 165, according 



* Loc. cit. 



