303 



in this paper were made with that object. We find that in 

 general the effect does exist, that it is sometimes very pro- 

 nounced, and that is in keeping with expectation based on 

 Madsen's study of the secondary y rays. Hard y rays show 

 a very large difference between the quantities of emergence 

 and incidence radiation : for soft y rays the difference is 

 smaller. Since X-rays are to be looked on as a very soft 

 form of y rays, the difference should be smaller still ; and 

 this is what we have found to be the case. 



The general form of the apparatus which we have used 

 is shown in fig. i. Variations of the upper portion of it are 

 shown in figs, ii and iii. A small pencil of X-rays passed 

 upwards through apertures in lead plates at .4 and B, and 

 then along the axis of the ionization chamber and out into 

 the 023en. In our first experiments the upper part of the 

 apparatus was arranged as in fig. iii. The primary rays did 

 not pass through the effective part of the ionization cham- 

 ber, being separated therefrom by the cylindrical screen S,S, 

 which could be made of various thicknesses and various mate- 

 rials. But if a thin sheet of any substance was laid over 

 the hole at B, secondary X-rays spread out therefrom, and 

 some passed through the screen *S',,S' and caused a deflection 

 in the electrometer. The difference between the deflections 

 fa) without and Hj ) with the sheet at B was taken as a 

 measure of the emergence secondary X-ray radiation. When 

 the sheet was removed from B, and the same or a similar 

 sheet placed in the plane of the top of the screen so as tf- 

 be struck from below by the primary rays, then the measure 

 of the incidence secondary radiation was obtained as the 

 difference between the deflections (a) without and fc) with 

 the sheet so placed. 



In this way it was easy to show that the expected want 

 of symmetry actually existed, particularly with aluminium, 

 celluloid, or paper as the radiators, substances of small 

 atomic weight. But the experiments were open to some extent 

 to the objection that a was too large compared with h—r, 

 and that possibly the excess of emergence over incidence was 

 an apparent effect due to actual variations of a under dif- 

 ferent circumstances. The current a was in fact due to seve- 

 ral causes. There was a small natural ionization leak even 

 when the X-rays were not acting: there was an effect due 

 to primary X-rays which had penetrated the walls of the 

 chamber though they were made of zinc, one-eighth of an 

 inch thick. But the greatest part of a was due to a diffu- 

 sion of soft rays about the primary beam, much of which 

 came through the hole at B at such an angle as to penetrate 

 the screen S,S; it could be largely cut out by thickening the 



