Velocities of Ttvo Groups of Rays and their Absorption Coefficients. 83 
I should like further to point out that each member of the second group 
of logarithmic curves begins to rise again very near to where _p = O, indicating 
the presence of a third group of corpuscles whose speed is less than that of the 
second group. Further, if the shape of the logarithmic ionisation density 
curves can be associated with Kaye's* logarithmic absorption (or rather selec- 
tive transmission) curves, and if from further experiment accurate numerical 
relationships can be deduced as below as to the relative energies of sets 
of corpuscles liberated from a metallic surface, then great light will be 
thrown upon the mechanism of absorption by solids generally. 
I have shown elsewhere,t from a theory put forward by Barkla,^ viz. 
that for each quantum of energy absorbed there is emitted, if the wave- 
length is short enough, one high-speed electron together with quanta of 
K, L, radiations, etc., that in the region of an L absorption band the 
sudden fall in the constant of proportion between the absorption of energy 
per atom ionised and the fourth power of the atomic number of the absorber 
^ + + . . ■ ^ ^.g 
'^r + . . • 
if the absorber have an atomic weight approximately that of gold, v^^^ v^^^ etc., 
are the frequencies of X-ray spectra corresponding to the L,M, etc., emission 
lines of the absorber. It was for this reason that silver was chosen as 
radiator and gold as screen for emitting particles, for as the K wave-length 
for silver is intermediate between the K and L wave-lengths for gold, then 
only the L, M, etc., characteristic groups of particles would be produced. 
If the energies of the particles bore a simple relationship to the energy 
absorbed in their production, then it was presumed that the ratio of their 
absorption coefficients would be neither too great nor too small to be detected 
by this method. 
(The corresponding ratio for K and L corpuscles from gold would 
probably lie between 7 and 8.) 
It is perhaps too soon to draw any definite conclusions as to the exact 
processes involved, but I shall show that the ratio shown in Table II is not 
in conflict with the general photo-electric equation. 
The fundamental law of photo-electric activity is \mv'^ = hv - w, in which 
^mv^ represents the maximum kinetic energy of the liberated electrons, Ji is 
Planck's constant, v the frequency of the exciting radiations, and iv a con- 
stant which measures the work necessary to be done to get an electron out 
of the sphere of influence of the parent atom. Independent experiments of 
Hughes § and Millikan|| have shown that the equation is valid so far as 
* ' I'hil. Trans. Eoy. Soc./ a, ccix, p. 137 (1919) ; also ' X Rays/ p. 126 (1917). 
t ' Trans. Roy. Soc. S. Af./ vi, Pt. 4, p. 321 (1917). 
X ' Nature/ March 4, 1915, p. 7. 
§ ' Phil. Trans./ a, vol. ccxii, p. 205 (1912). 
II ^Phys. Rev./ vol. vii, pp. 18, 355 (1916). 
