452 BELL SYSTEM TECHNICAL JOURNAL 



the oiitical spectra were produced chiefly by maintaining an electrical 

 discharge in a gas, the X-ray spectra invariably by bombarding a 

 solid body with exceedingly fasl-mo\ing electrons or with other 

 X-ra\s: the optical frequencies could be diffracted and refracted, 

 the X-rays not at all or almost imperceptibly little; the optical fre- 

 quencies were all inferior to 3.10'^, the X-ray frequencies all clearK- 

 more than a thousand limes as great. Since then, rays of almost 

 all the intermediate frequencies and with intermediate properties 

 have been generated in a variety of ways, and the distinction is no 

 longer trenchant, except between the extremes. To make it so, one 

 must seek a theoretical reason — and perhaps there is none to be found. 



There is, however, apparently good ground for introducing a 

 theoretical distinction. I have pointed out heretofore that the energy 

 which an atom loses, when it radiates one of the lines of its "optical" 

 spectrum, is less than the ionizing-energy. Or, turning this state- 

 ment around and amplifying it a little: the energy which an atom 

 absorbs, when it absorbs one of the rays of its optical spectrum, is 

 less than what is required to detach the loosest electron from it. 

 Therefore it is possible to assume, at least as a trial hypothesis, that 

 the energy is spent in lifting the loosest electron partway out — a 

 hypothesis fortified 1)\ ilu' t.ui ih.it, w hen the atom has just absorbed 

 some energy in this niaiiner, the eleitron can be detached by sup])lying 

 the atom with enough extra encrg>- to bring the total amount up to 

 the ionizing-energ\-. But if we take one of the typical X-ray fre- 

 quencies, and multiply it by It to ascertain how much energy the atom 

 gains in the process of absorbing that frequency, we find that the 

 cjiiantity lii> exceeds the ionizing-energy tremendously. This circum- 

 stance makes it quite out of the question to imagine that the X-ra\s 

 are due to changes in the position or the motion of the loosest electron 

 alone. We may therefore define the X-ray frequencies as those which 

 cannot be explained as due to transitions of the loosest electron, 

 from one motion or position to another, unaccompanied by other 

 changes. By this definition, every frequency v for which the quantum- 

 energy liv is greater than the ionizing-energy, goes into the X-ra\ 

 spectrum. For the remaining frequencies the question is more 

 dubious, perhaps never quite to be settled unless and until com- 

 plete theoretical classification of all the lines is attained. In this 

 section, howe\'er, I shall speak only of frerjuencies hundreds or thou- 

 sands of limes greater than the ionizing-frequency. 



Clazing upon typical X-ray emission spectra one sees that the\' 

 consist of groups of lines with wide inter\-als between. Going from 

 higher frequencies towards lower, the groups are known successively 



