[eve] IONIZATION WITHIN CLOSED VESSELS 79 
The agreement between his results for # rays and my results with 
y rays is sufficiently good to indicate that multiple reflection of the 
secondary 8 rays, originating at the walls where the 7 rays strike them, 
is the principal cause of the ionization within the testing vessels. The 
effect is controlled by the atomic weight of the material of the walls of 
the vessel. 
Moreover, the y rays which strike the concave or farther side of the 
vessel produce “incident” g rays to an extent also dependent on the 
atomic weight. 
But as Bragg has shown, this is not the case with the “emergent ”’ 
secondary radiation arising on the convex side where the y rays first 
strike. In this case the ionization plotted against the atomic weight gives 
a U-shaped curve, so that aluminium is about equal to lead as an ionizer. 
So far as the “emergent‘’ electrons are concerned, we might ex- 
pect an aluminium vessel to give as much ionization as a lead vessel, 
with penetrating y rays and suitable thickness of walls; but such a 
result would be difficult or impossible to realise, because the “incident ” 
radiation and the multiple reflection together quite overwhelm and mask 
this emergent radiation effect. 
The influence of the thickness of the walls is not very pronounced in 
the case of the 7 rays; but of course very thin or very thick walls alike 
give small results. The curve, ionization plotted against thickness, 
rises rapidly from the origin to a maximum and proceeds with a slow 
decrease, and is represented by the difference of two exponentials. 
The results obtained then agree well with the view that the Rontgen 
rays lonize mainly or wholly by electrons projected with velocities of 
the same order as those of the cathode rays; and that the yrays from 
radium C also ionize mainly or wholly by electrons projected with 
velocities of the same order as those of the / rays from radium C. 
With the X rays these electrons originate chiefly in the air in the 
testing vessel, but with the y rays for the most part from the walls. 
Some experiments made with very thin walled electroscopes in- 
dicated that with Rontgen rays the ionization within iron or zine ves- 
sels about half a millimetre thick is only 15 or 16 per cent of that in 
the free air. 
But with y rays the ionization in such vessels is about 140 per cent 
of that in the free air. The experiments in each case were made in the 
centre of a fairly large room with brick walls. 
The accompanying curve gives the relation between the ionization 
current in the electroscope, to an arbitrary scale, and the width in em. 
of the spark gap of the induction coil. 
McGill University, Montreal. 
June 15, 1912. 
