Secondary Réntgen Radiation. 547 
order to get the rays normal to the absorbing plate, and so 
greatly diminishing the energy of secondary radiation enter- 
ing the electroscope, it was put immediately behind the 
aperture inscreen C. The effective thickness of the absorbing 
plate was then much greater than before, as most of the 
radiation passed through it in a direction making a consider- 
able angle with the normal. Before any accurate results 
could be obtained of the absorption of this secondary radia- 
tion, it was necessary to determine the effective thickness of 
the plate. 
First using a primary beam consisting of rays of moderate 
penetrative power, such as had been experimented upon 
previously, it was possible, by comparing the rates of ioniza- 
tion in electroscope M when no absorbing plates were used, 
when an aluminium plate was placed in the primary beam by 
aperture P, and when the same plate was placed in front of 
the electroscope, to calculate, on the assumption that the 
beam was homogeneous in chara icter, the effective thickness 
of a plate placed before the secondary electroscope. By 
using thin plates, the error due to the primary beam being 
a mixture of radiations differing in character was reduced to 
a minimum. 
To obtain the very absorbable primary beam a copper plate 
was placed in the induction-coil box opposite the aperture, 
with the X-ray bulb as close as possible without emitting 
any direct radiation through the aperture Q. The secondary 
radiation from copper then constituted the new primary beam, 
and the tertiary radiation from air was the new secondary 
radiation. The boundaries of this primary beam were ap- 
proximately the same as before, so that, neglecting the small 
error due to want of homogeneity, the effective thickness was 
as before. From the diminution in the rate of ionization in 
M when an absorbing plate was put at P, the corresponding 
diminution when the plate was placed at R was calculated on 
the assumption that the secondary and primary radiations 
again had the same penetrative power. This was then com- 
pared with the actual absorption produced by the plate at R. 
As this test would only hold with accuracy provided the 
primary and secondary beams consisted of rays of one 
character, the thickness of the absorbing plate at P which 
produced. the same absorption as a plate of given thickness 
at R was determined experimentally for the moder rately pene- 
trating rays, and the same comparisons were made as before. 
This method was still gpen to the objection that the effective 
thickness of the plate at RK determined experimentally for 
rays coming direct from an X-ray bulb was not the same as 
2P 2 
