Rontgen Radiation in Air. 609 



falls below unity. This can be explained by assuming (1) that 

 in the secondary beam the softer rays are less strongly repre- 

 sented than they would be if perfect scattering took place, so 

 that more rays are cut off at P than at S : (2) that the hard 

 rays are somewhat transformed into a softer radiation, so that 

 when the hard rays alone are operating, the secondary rays 

 appear more absorbable than the primary. 



To determine how these results were affected by the nature 

 of the bulb and coil, a small six-inch spark coil was used with 

 a current of about one ampere and a fairly soft bulb. The 

 following results were obtained : — 





T 



A.BLE II. 





Thickness of 

 .bsorbing plate 

 at A in cms. 





*v 



Percent, of secondary 

 rays cut off by Al plate 

 •04 cm. thick. 









1-045 



60-2 



•01 





1-025 



51-9 



•02 





•993 



53-2 



•0-1 





•982 



39-6 



On comparing this table with the preceding, we see that 

 the penetrative power is not so great to start with, while it 

 decays more rapidly. Thus when 39" 6 per cent, is cut off b} r 

 the Al plate, P = -982 with the soft bulb, while it is 1-09 (for 

 a nearly similar absorption) with the harder bulb. If our 

 assumptions (1) and (2) above are to hold, the radiations 

 from the soft bulb should be more homogeneous than those 

 from the other, which would account for the small initial P, 

 and should also be of a softer character, taken as a whole, 

 which would explain the rapid decrease in the percentage cut 

 off (3rd column) with increasing thickness of absorbing plates 

 at A. 



To test this hypothesis, curves were drawn in which the 

 ordinates represented deflexions in the primary electroscope 

 per unit time of working the bulb, and the abscissae repre- 

 sented thickness of Al in centims. placed at P, to cut down the 

 beam entering the primary. The secondary electroscope was 

 used to standardize the intensity of the main beam ; and it was 

 found that the deflexion in it per unit time kept fairly con- 

 stant as the experiment proceeded, the maximum variation 

 being 3 per cent. 



If the beam were perfectly homogeneous, each thin plate of 

 thickness dx would cut off the same percentage of the radia- 

 tion falling on it, so that if I is the intensity at any time Xldx 

 would be stopped by the plate, so that after passing through 



