522 Professor John W. Nicholson [May 2^ 



WEEKLY EVENING MEETING, 



Friday, May 2, 1919. 



General E. H. Hills, C.M.G. D.Sc. F.R.8., 

 Secretary and Vice-President, in the Chair. 



Professor John W. Nicholson, M.A. D.Sc. F.E.S. 



Energy Distribution in Spectra. 



Our knowledge of the structure of very fine spectrum lines is now 

 on a secure and nearly comprehensive basis, from the point of view 

 both of theory and of experiment, and it is very remarkable that sa 

 little is known of the distribution of energy in these lines, either in 

 individual lines, or as between the different lines contained in the 

 spectrum of any atom. The analysis of the spectra of atoms, theo- 

 retically and in the laboratory, is now recognised as the most critical 

 test to which any theory of atomic structure can be subjected, and 

 we have recently had theories of the atom entirely based upon the 

 wave-lengths of the radiations which they emit. A question of equal 

 importance, to w^hich I now wish to direct your attention, is the 

 relative amounts of energy which the atom throws out in the different 

 wave-lengths, for this is well known to vary in some cases very 

 greatly, according to the circumstances under which the atom is 

 excited. I shall describe, in the first instance, a method designed 

 by Dr. Merton for investigating the distribution of energy among 

 spectrum lines, or in the breadth of an individual line, with great 

 accuracy. It is possible by this means to obtain the long-desired 

 object of an absolute scale of spectral intensity, independent of all 

 the ordinary difficulties determined by such matters as the unequal 

 behaviour of the photographic plate for light in different regions of 

 the spectrum. Dr. Merton and I have been working together on 

 this subject for the past three years, and I shall conclude the present 

 discourse with an account of some of the more interesting results 

 which have been reached, after an explanation of the method. 



The intensities of spectrum lines have usually been recorded on 

 an arbitrary scale, ranging between 10 and zero, the numbers 

 assigned being at the discretion of the observer, and varying so 

 greatly among different observers as frequently to be of little value 

 for exact knowledge. They depend also very much on the nature of 

 the observation, wiiether visual or photographic, and in the latter 

 case, on the region of the spectrum to which the line belongs. The 

 sensitivity of a photographic plate varies with the wave-length of 



