728 Prof. Bragg and Mr. Kleeman on the 
there is a violent disturbance of radioactive equilibrium, they 
must emit streams containing nearly equal numbers of 
particles. They may, however, be much narrower than the 
quadrilateral GHPN which belongs to the first change. The 
width of a quadrilateral represents, of course, the number of 
particles in the stream, relative to the other streams: the 
longer upright side represents the range in air of particles: 
trom the top layer of radioactive material, the shorter upright 
side the range of the particles from the bottom layer. The 
sides AB, CD, EF should therefore have the same slope. 
In an actual experiment the corners must be rounded off, 
because the ionization chamber must have an appreciable . 
. 
| 
depth. Moreover, the cones must have an appreciable width. 
We were, in fact, obliged to use rather wide cones, in order 
to obtain convenient deflexions in the Kelvin and White 
electrometer. We bound together various little bundles of 
tubes of thin copper, and placed them vertically over the flat 
silver dishes on which we evaporated a few drops of solution ef 
radium bromide. Fig.1 shows the general arrangement. AB is 
Fie. J. 
— 70 QUADRANT ELECTROMETER. 
Arrangement of experiment :—Spherical ionization chamber. Conductors 
shaded, insulators plain. The gauze C made with B an external 
ionization chamber, which prey ented any external ionization from 
entering the measuring chamber AB. It was only used occasionally, 
and seemed to be unnecessary. 
the ionization chamber, A being thin aluminium, B a sheet of 
gauze. The lead plate to which the latter was soldered was 
raised to a positive potential of 250 to 400 volts. A was in 
