420 Sir J. J. Thomson on the Origin of 



be emitted by a single electron placed near a unit positive 

 charge would not be of: this character. Assuming the usual 

 law of force tor two charges, if the atom is to be in a steady 

 condition the electron must describe an orbit round the 

 positive charge. The time of rotation of the electron will 

 depend on its distance from this charge, and if this varies 

 continuously the times of rotation will do so also, and 

 the spectrum whose frequencies are determined by these 

 times would be a continuous one. If the charges are regarded 

 merely as centres of inverse square forces there are no 

 reasons for retaining some of these orbits as possible and 

 neglecting the others. Mr. Bohr, in his theory of spectra, 

 supposes that the only orbits which are possible are those 

 where the ratio of the energy of the electron to its angular 

 velocity is an integral multiple of a definite unit. This, 

 however, is not the consequence of dynamical considerations; 

 it is arithmetical rather than dynamical, and if it is true it 

 must be the result of the action of forces whose existence 

 lias not been demonstrated. The investigation of such forces 

 would be a problem of the highest interest and importance. 



By the use of this principle and a further one, that when 

 an electron passes from one orbit to another it gives out 

 radiation whose frequency is proportional to the difference 

 of the energy of the electron in the two orbits, Mr. Bohr 

 obtains an expression which gives with quite remarkable 

 accuracy the frequencies of the lines in the four-line spectrum 

 of hydrogen. It is, I think, however, not unfair to say 

 that to many minds the arithmetical basis of the theory 

 seems much more satisfactory than the physical. 



The vibrations which give rise to the spectrum do not on 

 this theory correspond in frequency with any rotation or 

 vibration in the atom when in the steady and normal state. 

 That in the normal atom there is something which can vibrate 

 with the frequency of the lines, or at any rate with that of some 

 of the lines in the spectrum, seems to be proved almost irre- 

 sistibly by the experiments of Professor Wood and Mr. Be van 

 on the absorption spectra of the vapours of the alkali metals. 

 These vapours give as absorption spectra fine well-defined 

 black lines coinciding in position with the lines in the 

 principal series of the spectrum of the metal. Thus, for 

 example, Professor Wood obtained 48 of these lines in 

 the absorption spectrum of sodium vapour, and Mr. Bevan 

 24 for potassium, 30 for rubidium, 24 for csesium. The 

 sharpness and intensity of these absorption lines produced 

 by comparatively cold vapour are so great that it is very 

 difficult to believe that they are not due to a resonance 



