39 
1919-20.] The Absorption of X-Rays. 
charges, then the value in the second last column should be ’000546 or 
000273 respectively. The J bands are so faint and difficult to measure 
that the experimental values can be said to agree with the larger of the 
above figures within the error of observation. 
It is straining the laws of probability altogether too far to assume 
that the results in the foregoing table are chance coincidences. Even 
if the mean values '972 and 1*01 were accidental, we should still have 
to explain why the results for the J bands come so close to the value 
for a particles. Also, if the agreement for the K bands was accidental, 
we would expect a progessive falling away from unity as the atomic 
weight increases; but the result for silver, the element on the list with 
the highest atomic weight, agrees well with the mean, the large variations 
in A m , \ — \ m , and /u m all compensating one another. It would be interesting 
to get measurements on the K bands of elements heavier than silver and 
lighter than iron ; we could then test (8) over a wider range. 
The agreement, such as it is, of the first seven results with unity 
verifies the wave-lengths of X-rays, which have hitherto been determined 
from (i) the theory of the crystal lattice, and (ii) the quantum theory, 
in a new and independent manner. The agreement also tells against 
the Bohr atom model, for there is no place in that model for electrons and 
atoms vibrating in the manner contemplated. But it suggests a con- 
nection with Soddy’s theory of the relation between atomic disintegration 
and the periodic classification of the elements ; if electrons and alpha 
particles are ejected from the atom, why should we not detect them 
vibrating in the atom ? 
§ 3. The difficulty in accepting the foregoing results seems to be 
that the foregoing theory apparently makes all the atoms absorb the 
X-rays, whereas in reality it is only a few atoms that do the absorbing. 
This difficulty can be diminished, but not removed, by two different 
arguments. 
In the first place, it is nowhere assumed that all the atoms do the 
absorbing. N is not the number of electrons actually present in unit 
volume, but the limiting value of the ratio of number of electrons to 
element of volume, as the latter is made indefinitely small, at the place 
where the absorption is taking place. It is not necessary as regards 
the absorption (or the scattering) that the X radiation should go through 
the substance as a uniform wave-front; it may go through in streaks 
separated by dark spaces. 
Again, although in light as a matter of convention we deal with uniform 
waves, they constitute an ideal case. The uniform spherical wave spreading 
