INTENSITY RELATIONS IN THE SPECTRUM OF HELIUM. 149 



In connection with the interpretation of this table, it is necessary to remark, in the 

 first place, that the actual numbers themselves give no information in the absolute 

 sense, or in the relative sense down one column, as to the relative intensities of the 

 three lines in question, for the photographic plate is not equally sensitive in three 

 regions. But the actual changes in the numbers from one column to another give 

 decisive information, since the intensity of \5876 is reduced to a uniform scale. 

 These changes are very small, though quite definite, even taking into consideration 

 the fact that the numbers are derived from an exponential type of formula, and 

 they cannot be regarded as within the error of observation and consequent calculation. 

 To at least a close approximation, however, the relative intensities in the Diffuse 

 series do not vary with the distance from the cathode. 



The small variations which do occur present no striking regularity, and it is evident 

 that the behaviour of the last three photographs, already stated to be on a different 

 plate to the others, is not appreciably different in these regions, so that we have 

 further justification for the supposition that the two sets of photographs are directly 

 comparable. There is a small amount of evidence in the table, although it is not 

 decisive, that a slight energy transfer to the longer wave-lengths takes place with 

 increasing distance from the cathode, but if it be real, it is yet so small as to be a 

 comparatively unimportant phenomenon. There is no effective transfer of energy 

 along the Diffuse series of Helium with increasing cathode distance. 



Small variations in the numbers are to be expected, for it is difficult to maintain 

 complete uniformity in the experimental conditions over a long period, and the 

 various photographs were necessarily taken at different instants. But sucli variations 

 in the conditions from one photograph to another apply to all the series alike, and 

 from the uniformity of the numbers in Table V. we may assume with confidence that 

 they are small. 



According to this conclusion regarding the absence of an energy transfer along the 

 series, it is not difficult to show that the vanishing of X4026 on VIII. is to be 

 expected. For on the basis of 10 for the photographic intensity of X5876, the 

 average value for X4026 is 3'0. The actual photographic intensity of X5876 on VIII. 

 is, from a preceding table (Table III.) 2'67. That of X4026 should therefore be, on 



this scale, 



2'67x3'0/10 = 0"80. 



Accordingly, for this line, if h^ be its height, 



'" 1 = 0'80 



which is less than unity, and therefore h is negative. This signifies that the 

 exposure is insufficient to show the line even on theoretical grounds. In fact, on the 

 scale in Tables III. and IV. the minimum photographic intensity which can be visible 

 is not zero but unity. This particular scale, according to the definition of 



