304 BELL SYSTEM TECHNICAL JOURNAL 



can be tested by any possible way. Just so were the three phenomena 

 cited in these Rules discovered. All of them involve either the 

 emission or the absorption of radiation, and so do all the others which 

 I could have quoted in addition, if this account had been written 

 three years ago. Reserving to the end the one new phenomenon 

 that transcends this limitation, I must explain the relation between 

 this problem and the contemporary Theory of Atomic Structure. 



The classical notion of a source of radiation is a vibrating electron. 

 The classical conception of an atom competent to emit radiations of 

 many frequencies is this: a family or a system of electrons, each 

 electron remaining in an equilibrium-position so long as the system 

 is not disturbed, one or more of the electrons \-ibraling when the 

 system is jarred or distorted. A system with these properties would 

 have to contain other things than electrons, otherwise it would fly 

 apart; it would have to contain other things than particles of posi- 

 tive and particles of negative electricity intermixed, otherwise it 

 would collapse together. One would have to postulate some sort 

 of a framework, some imaginary analogue to a skeleton of springs 

 and rods and pivots, to hold the electrons together in an ensemble 

 able to vibrate and not liable to coalesce or to explode. This would 

 not be satisfying, for in making atom-models one wants to avoid 

 the elaborate machinery and in particular the non-electrical com- 

 ponents; it would be much more agreeable to build an atom out of 

 positive and negative electricity associated with mass, omitting all 

 masses or structures not electrified. Nevertheless, if anyone had 

 succeeded in devising a framework having the same set of natural 

 frequencies as (say) the hydrogen atom exhibits in its spectrum — if 

 anyone expert in dynamics or acoustics had been able to demonstrate 

 that some peculiar shape of drumhead or bell, if anyone \ersed in 

 electricity had been able to show that some particular arrangement 

 of condensers and induction-coils has such a series of natural vibrations 

 as some one kind of atom displays — then, it is quite safe to say, that 

 framework or that membrane or that circuit would today be either 

 the accepted atom-model, or at least one of the chief candidates for 

 acceptance. Nt)body e\er succeeded in doing this; it is ilu' consensus 

 of opinion today that the task is an impracticable one.'* 



' It is difficult to put this statement into a more precise form. Rayleigh was of 

 the opinion that the hydrogen sjX!i-trum could not be regarded as the ensemble of 

 natural freciucncies of a mechanical system, because it is the general rule for such 

 systems that the second power of the frequency conforms to simple algebraic formulae, 

 while in the hydrogen s|)cctruni it is xhe first power for which the algebraic expression 

 is simple. He admitted, however, that it was possible to find "e.\ccplional " mechan- 

 ical systems for which the first power of the frequency is given by a simple formula; 

 which goes far to vitiate the conclusion. Another aspect of the formula (6) for 



