Absorption of « Rays. 723 
certain cone of @ rays is tested at various points, and the 
inverse square law does not enter into the question. 
With the assistance of Mr. R. Kleeman | have carried out 
a set of experiments of this kind. Since the results obtained 
by Rutherford and Madame Curie would lead us to expect 
a simpler result in the case of polonium than radium, it 
would have been preferable to have employed the former, 
but the latter was alone available. 
In the case when all the rays are initially of uniform 
velocity, the curve obtained ought to show, when the radium 
is out of range of the ionization chamber, an effect due 
entirely to @ and y rays, which should slowly increase as 
the distance diminishes. When the erays can just penetrate, 
there should be a somewhat sudden appearance of the 
ionization, and, for a short distance of the approach, equal 
to the depth of the chamber, the curve should be a parabola. 
Afterwards it should become a straight line. 
This is exactiy realized; and so far the hypothesis is 
verified. But a further effect appears. As the radium is 
gradually brought nearer to the chamber, the straight line 
suddenly changes its direction ; and indeed there appear to 
be two or three such changes. Thus the curve is really a 
rectilinear polygon, with the corners rounded off. 
Moreover, the slope of the last side, representing effects 
close to the radium, is nearly four times the slope of the first 
side; whilst it seems probable that the slopes of two inter- 
mediate sides are two and three times that of the first, 
respectively. 
For all this there is a ready explanation. The atom 
passes through several changes, and it is supposed that at 
four of these an a atom is expelled. Probably the & particles 
due to one change are all projected with the same speed. We 
ought therefore to expect four different streams of @ particles, 
differing from each other only in initial energy. If the 
radium and its products are in equilibrium, the number of 
a particles due to each change is the same. Thus, if q, is 
the range of one stream, a, of another, and so on, the 
ionization should, when two streams reach the chamber, be 
nb b nb b 
 (a-A— je = (a-h— y 
or (ay +4 —2h—2). 
Thus the slope of the curve should in this case be a 
: mle 
whereas if only one stream enters it should be nb/p. When 
