238 
DR. T. R. MERTON AND PROF. J. W. NICHOLSON ON 
which should render their spectra especially suited to theoretical investigation. 
Researches have been mainly directed to the study of the series relations in the 
spectra, and production of “ enhanced ” or “ spark ” lines under conditions of powerful 
excitation. The present investigation is devoted to a study of the spectra of 
Hydrogen and Helium, with special reference to the relative intensities of the lines 
under different conditions, a quantitative knowledge of which must necessarily be of 
importance in any discussion of the relation of spectra to the constitution of the 
atom. 
It has been the custom to record the intensities of spectrum lines on an arbitrary 
scale, ranging generally from 10 for the strongest lines to 1 or 0 for the weakest. 
The intensities have usually been assigned purely at the discretion of the observer, 
and without much regard to the conditions of observation, whether visual or 
photographic. In any case, the apparent intensities are affected by the optical 
system by means of which the spectrum is produced, and whereas in visual 
observations the sensitiveness of the eye to different wave-lengths should be taken 
into account, matters are even more complicated in the case of photographs of spectra, 
where the sensibility of the photographic plate varies very considerably with the 
wave-length in a manner which is not precisely defined, and which appears in fact to 
be to a great extent irregular. These remarks apply to the visible and ultra-violet 
regions of the spectrum. Measurements in the infra-red are usually carried out with 
the thermopile or bolometer, and in this case quantitative measurements of the 
intensities of lines are obtained. It is, however, the visible and the less refrangible 
part of the ultra-violet spectrum in which accurate measurements of intensity, on a 
precise and quantitative basis, are perhaps most urgently needed at the present time. 
For example, in the case of celestial spectra, the problem of the relative intensities 
of lines has become of considerable importance, and is probably, at present, the 
greatest obstacle in the direction of the further elucidation of the phenomena which 
occur in new stars. In these spectra, observations are of necessity limited to the 
visible and the less refrangible ultra-violet regions. 
The great advances which have been made already in the study of the variations 
in spectra under different conditions of excitation must owe their success largely to 
the magnitude of the effects in question ; for they have depended to a great extent 
on methods of observation which cannot take account of the more subtle and less 
conspicuous variations which would probably, if detected, provide the key to the 
elucidation of many fundamental problems of spectra, such as the distribution of 
intensity in spectral series and the relations of different series to one another. At 
the present time, when a radiating gas emits a series spectrum, we have no exact 
knowledge of the relative amounts of energy thrown into the individual members of 
the series, or even whether these amounts are definite functions of the “ term 
number,” or capable of variation according to circumstances, although a certain 
amount of evidence has hitherto favoured the latter alternative. 
