304 



screen. Again, part of a was due to radiation returned from 

 the open air above the ionization chamber. Some of these 

 radiations might be appreciably interfered with by placing 

 the radiating sheet at B or at the top of the chamber. We 

 were, however, able to satisfy ourselves by special experiments 

 that the want of symmetry was quite real, and that as a 

 matter of fact no valid objection could be made. But we 

 abandoned the first arrangement for a second which, as we 

 expected, would show the want of symmetry more clearly, 

 and which proved better than the first in every M-ay. The 

 first method was exactly the same as that used by Madsen 

 in examining the secondary v rays ; but it was clear that the 

 enormous difference which these rays showed was not going 

 to be repeated in the case of the X-rays. 



Our new arrangement was, as shown in fig. i., or, in- 

 verted, in fig. ii. Two cylinders of brass, each 2 in. long, 

 but of different diameters — 4 in. and 2 in. — were fixed to a 

 connecting piece 7^,7), shown in plan in fig. iv. The latter 

 resembled a light brass wheel with four spokes, and various 

 thin screens cut in the form of flat rings could be attached 

 to it, filling up all the spaces between the spokes. In fig. ii. 

 the double cylinder is shown as arranged for the measure- 

 ment of incidence secondary radiations : the radiating sheet 

 was placed at C, supported by a sheet of celluloid lying flat 

 on the top of the cylinder. A hole was cut in the centre 

 of the celluloid sheet big enough to allow the primary beam 

 to pass through without touching the edges : and a fluores- 

 cent screen was used to make sure that this was the case. 

 The radiating sheets were of thin metal, about H in. square. 

 In fig. ii. the cylinder is shown as arranged for the measure- 

 ment of emergence secondary radiations : it hardly requires 

 further explanation. 



We expected that this arrangement would show up the 

 want of symmetry better than the former, because the por- 

 tions of the emergence and incidence beams under compari- 

 son would be more nearly normal to the plate. Looking 

 upon the radiations as material we should naturally expect 

 the intensity of the secondary radiation to decrease gradu- 

 ally as its direction increased in inclination to the forward 

 direction of the primary ray. The emergence rays lie, in in- 

 clination, between deg. and 90 deg. ; the incidence between 

 90 deg. and 180 deg. In our first arrangement we compared 

 the emergence rays between about 40 deg. and 90 deg. with 

 the incidence rays between about 90 deg. and 140 deg. There 

 should be a larger ratio of emergence to incidence with the 

 newer arrangement, since the emergence rays between about 

 30 deg. and 50 deg. would be compared with the incidence 



