276 Prof. Barkla and Mr. Ayres on the Distribution of 



position a = 90°, gave an average value of cos 2 a for the rays 

 which was estimated as '035, consequently the observed 



intensity was not truly J v / 2 but 1*035 I r/2 . The values T — 



J-tt/2 



., , . , „ ,, . 1+ cos 2 a + *lcos 2 « 

 were thus obtained from the expression -i.riOK ~> 



the third term in the numerator being due to polarization. 



Column 4 gives the observed ratio -.— ; 



-U/2 



Column 5 gives the percentage difference between calcu- 

 lated and observed values. 



The observed values in column 4 are the mean of several 

 readings, but in only very few observations did the reading- 

 differ by more than 4 per cent, from the value given. 

 Except in the case of angles 20° and 30° the agreement 

 between theory and experiment is remarkably good. In 

 these cases the correction for scattering from air was large, 

 and the results were much less reliable. The fact that the 

 discrepancy appears only for these small angles is signi- 

 ficant. It seems just possible that these results are vitiated 

 by an irregular refraction effect. Experiments are being 

 made to test these further. We are not yet in a position to 

 say definitely if for the scattered radiation there is here an 

 appreciable discrepancy between theory and experiment. 



It is interesting to compare the intensity distribution of 

 the scattered radiation with that found for a true secondary 

 radiation (the homogeneous fluorescent X-radiation) from a 

 substance such as copper. In this case the intensity is 

 uniform in all directions as shown by Table II. 



Table II. 





la 





I* 



Angle a. 



lfr/2 



Angle a. 



Ijr/2 



20 9 



•99 



120° 



1-03 



30 



1-02 



130 



•97 



40 



101 



140 



102 



50 



•99 



150 



•99 



60 



1-02 



ltS5 



101 



70 



•98 







80 



•99 



-50° 



•99 



90 



1-00 



-120 



•99 



100 



TOO 



-150 



103 



110 



•99 . 







