RECENT ADVANCES IN SCIENCE 19 



first or innermost path, and it therefore must absorb sufficient 

 energy to eject it from this path to one of the outer ones before 

 the atom can become luminescent. In the hydrogen tube 

 an atom would acquire this energy by collision with a free 

 electron possessing sufficient energy. An interesting question, 

 therefore, arises as to the relative intensities of the various 

 lines at different pressures. Thus as the pressure in the tube 

 is reduced, and the mean free path of the free electrons in- 

 creased, their velocity at the moment of an encounter with 

 an atom would be increased also ; and so it might be expected 

 that a greater number of the luminescent atoms at a given 

 instant would be in the state in which the electron was in 

 paths 4, 5, or 6, as compared with those in the state corre- 

 sponding to path 3. If that were so, it might also be expected 

 that the relative intensity of the light corresponding to the 

 H p , H y , H s lines, as compared to the H a light, would increase 

 with diminution of pressure. The author, however, concludes 

 from an examination of his own and other experimental work 

 that this view is not justified, and that the energy distribution 

 in the lines of the Balmer series is independent of the energy 

 of the colliding electron (or molecule, if excitation is due to 

 impact of canal rays), which is responsible for the luminescence 

 of an atom. By admission of water vapour it is possible to 

 vary the free path of the atom itself, and here the results 

 seem to justify the statement that the intensity distribution 

 is affected by the mean free path of the luminescent atom 

 and the nature of the molecules in its vicinity whose fields 

 of force act on it. If the mean free paths of two atoms, one 

 emitting H a light and the other H^ light, are less than the 

 distances which the atoms travel during emission, the intensity 

 ratio will remain constant, and be independent of the par- 

 ticular pressure used, so long as it is consistent with the 

 assumption made. Again, if the pressures were so low that 

 the mean free paths were all longer than the distances travelled 

 during emission, a constant ratio of intensities would also 

 result, although a different constant from the former. But 

 within an intermediate range of pressures which permitted 

 one of the atoms to complete its emission in a shorter length 

 than its free path, but destroyed the radiating power of the 

 other by a premature collision, due to the relative shortness 

 of its free path, one would expect the ratio Up/U a to alter. 



