SOME COMF.MPOR.IRV .-IDr.-tNCF.S IN PHYSICS— III 303 



fxiitfil from a soliti targi-t by very fast fUrtrons, it is legitimate to 

 Siiy for these radiations which form the single line s()ectra of metallic 

 atoms, that 



Exiilution oj the ray forming a single-line spectrum, by the collision 

 of an electron against an atom, occurs as if the energy in the radiation 

 were concentrated in units of amount hv, and one such unit were createa 

 out of the total energy which the electron surrenders. 



There are yet se\eral phenomena which I might treat by the same 

 inductive methcx.!, arriving after each exposition at a Rule which 

 would resemble one or the other of those which I have thus far written 

 in italics; but it is no longer expedient, I think, to pass in each instance 

 through the simie elaborate inductive detour. These three phe- 

 nomena which I have discussed already combine into an impressive 

 and rather fornudable obstacle to the classical manner of thinking. 

 Here is a mercury atom, which receives a definite quantity of energy 

 U from an electron, and distributes it in radiation of a definite fre- 

 quency U/h. Here again is a multitude of atoms locked together 

 into a solid, and when an electron conveys its energy U to the solid, 

 it redistributes that energy in radiation of a definite frequency U/h. 

 (It is true that many other radiations issue from the solid, but they 

 are all explicable if one assumes that the electron may deliver over 

 its energy in stages, and there is no radiation of the sort which would 

 controvert the theory by virtue of its frequency exceeding U/h.) 

 And when that radiation of frequency U/h in its turn strikes a metal, 

 it is liable and able to release an electron from within the metal, 

 conferring upon it an energy which is apparently equal to U. Ap- 

 parently there is some correlation between an energy U and a fre- 

 quency U, h, between a frequency ;' and an energy hv. Apparently 

 a bl(xk of energy of the amount U tends to pass into a radiation of 

 the frequency U/h; apparently a radiation of the frequency v tends 

 to deliver up energy in blocks of the amount hv. The three italicized 

 Rules coalesce into this one: 



Photoelectric emission, and the e.Kcitation of A'-rav5 from solids by 

 electrons, and the excitation of single-line spectra from free atoms, occur 

 as if radiant energy of the frequency v were concentrated into packets, 

 or units, or corpuscles, of energy amounting to hv, and each packet were 

 created in a single process and were absorbed in a single process. 



If the neutralizing as if were omitted, this would be the corpuscular 

 theory rediviva. It is good policy to leave the as if in place for awhile 

 yet. But conservatism such as this need not and should not deter 

 anyone from using the idea as basis for every prediction that can 

 be founded upon it, and testing every one of the predictions that 



