310 BULL SYSTEM TECHNICAL JOURNAL 



kind of atom. Of all the [jlieiioinena which might serve to illuminate 

 this difficult question of the relations between radiation and atoms, 

 this is the one which has been least studied. The experimental 

 material is scanty and dubious. There is no reason to suppose that 

 light of a lower frequency than the one I have called Vo is able to 

 ionize; but it is not clear whether perceptible ionization commences 

 just at the frequency Vo, although it has been observed at frequencies 

 not far beyond. The energy of the released electrons has not been 

 measured. 



The removal of deep-lying electrons, the electrons lying close to 

 llic nuclei of massive atoms, is much better known; and the data 

 confirm in the fullest manner the idea that radiant energy of the 

 frequency' v is absorbed in units amounting to hv. When a beam of 

 X-rays of a sufficiently high frequency is directed against a group of 

 massive atoms, various streams of electrons emanate from the atoms, 

 and the electrons of each stream have a certain characteristic speed. 

 The kinetic energy of each electron of an>' particular stream is equal 

 to //;', minus the amount of energy which must be spent in extracting 

 the electron from its position in the atom; for this amount of energy 

 is independently known, being the energ>' which a free-flying electron 

 nnist possess in order to drive the bound electron out of the atom, 

 which is measurable and has been separately measured. Here again 

 I touch upon a subject which has been treated in an earlier article 

 of this series — the second — and to prevent this article from stretching 

 out to an intolerable length, I refrain from further repetition of what 

 was written there. The analogy of this with the photoelectric effect 

 will escape no reader. Here as there, we observe electrons relea.sed 

 with an energy which is admittedly not hv, but hv minus a constant; 

 the idea that this constant represents energy which the electrons 

 have already spent in escaping, in one case through the surface of the 

 metal and in the other case from their positions within atoms, is 

 fcirtified by independent measurements of these energies which give 

 \alucs agreeing with these constants. 



We have considered various items of evidence tending to sliow 

 that radiant energy is born, so to speak, in units of the amount hv, 

 and dies in units of the amount hv. Whether energy remains sub- 

 divided into these units during its incarnation as radiation remains 

 unsettled; to .settle this question absolutely, one would have to de\ise 

 some way of testing the energy in a beam of radiation, otherwise than 

 by aljsorbing it in matter; and such a way has not yet been di.sco\ered. 

 There is, however, another quality which radiant energy possesses. 



Conceive a stream of radiation in the form of an extremely long 



