CONTEMPORARY ADVANCES IN PHYSICS 171 



The question is much more difficult to answer than it seems: for 

 even if the cosmic rays are of the nature of gamma-rays, the effect ivhich 

 they produce in ionization-chamhers {and in the kinds of apparatus yet to 

 be described) is almost entirely due to fast electrons. 



Consider a beam of X-rays (of the same nature as gamma-rays but 

 softer) projected into a gas. It is known by experiment that they 

 free electrons from atoms in two distinct ways : some of the corpuscles or 

 photons are entirely absorbed in atoms, and tightly-bound electrons 

 are ejected (photoelectric effect) ; some on the other hand make elastic 

 impacts against loosely-bound electrons, communicating a part of 

 their energy to these and retaining a part (Compton effect). As the 

 frequency of the X-rays is increased, the former effect becomes less 

 usual; entering into the range of gamma-ray frequencies, we find it 

 fading out; it is reasonable to suppose that if there are corpuscles of 

 light of frequency even higher than that of the gamma-rays, they ionize 

 only by Compton-effect or "Compton scattering." It remains of 

 course conceivable that such corpuscles may ionize in ways unknown 

 at lower frequencies, as for instance by shattering the nuclei of atoms 

 into fragments some of which are fast-flying electrons or protons. 



Now though these impacts are properly called "elastic," their 

 laws are peculiar, owing to the differences between corpuscles of light 

 and corpuscles of matter.^" The electron acquires some, usually a 

 considerable part, of the energy of the photon; this latter (or as some 

 would prefer to say, a new photon) goes off' more or less obliquely, with 

 lessened energy and lessened frequency. Given the wave-length of the 

 light, the laws of the impact determine the energy acquired by the 

 electron, the energy retained by the photon, and the direction of 

 departure of the photon, as functions of the direction in which the 

 electron is projected. If we know the distribution-in-direction of the 

 projected electrons, we can determine their average energy. For 

 electrons ejected by X-rays this distribution is known. But if the 

 cosmic rays are (or include) gamma-rays whereof the absorption- 

 coefficient has such values as those given (e.g.) by Millikan and by 

 Regener for /x, they are much harder than any otherwise-known gamma- 

 rays or X-rays, hence presumably of much lesser wave-length. Then, 

 for the distribution-in-direction of the electrons which they eject, we 

 must resort to an untested theory. This theory indicates that a 

 beam of short-wave gamma-rays proceeding through matter should 

 provide for itself an escort of electrons, most of them moving in 

 directions inclined at very small angles to the beam itself, and many 



1" See for instance my Introduction to Contemporary Physics, pp. 145-163, or the 

 sixth article of this series. 



