INTENSITY RELATIONS IN THE SPECTRUM OF HELIUM. 159 



the band spectrum appear together. It seems possible that this mode of observation 

 will be of use in the resolution of complex spectra into series. 



The somewhat narrow region to which the band spectrum is confined would seem 

 to imply that the conditions of excitation which give rise to this spectrum fall 

 between very restricted limits. 



(IX.) The Spectra of Mixed (rases. 



A considerable amount of qualitative information has already been obtained by 

 various investigators, who have examined the effect, on the spectrum of a gas, 

 produced l>y impurities, or by a definite mixture with another gas. We have already 

 referred to this work in our introductory section and in a preceding communication, 

 in which we described some strictly quantitative results shown by the spectrum of 

 Hydrogen when this gas is mixed with a certain amount of Neon. The effect on the 

 spectrum of Neon, which may be expected to be in some sense complementary, of the 

 large admixture of Hydrogen was not investigated on account of our lack of know- 

 ledge of the series relations in the Neon spectrum. In order to determine in what 

 sense complementary effects occur in the spectra of the two mixed gases, it is necessary 

 to select two gases whose series relations are known, and the present section of this 

 communication details the experimental results obtained by our method of measurement, 

 with a view to the discussion given in later sections. 



In our experiments with mixed gases, we have not been able to eliminate as a source 

 of error the possibilities of effects arising from small differences in pressure in the 

 different tubes, but as the effects observed do not correspond to those which would 

 accompany an alteration in pressure, we feel that the observed phenomena may be 

 described as particular to the mixed gases. 



We have made experiments on this subject in two cases, which present strikingly 

 dissimilar phenomena. In the first case, the gas consisted mainly of Helium, with 

 only a very small admixture of Hydrogen sufficiently small, in fact, to justify the 

 statement that practically only a trace of Hydrogen was present. In the second case, 

 a considerable addition of Hydrogen to the Helium was made, so that the tube 

 actually contained a " mixed gas," in the sense that the orders of magnitude of the 

 quantities present were the same. For purposes of comparison, we also examined the 

 spectrum, under like conditions, of the purest Helium which could be obtained, and 

 which we have already mentioned. We shall refer to this as " pure Helium," and to 

 a similar spectrum obtained with the purest available Hydrogen as that of "pure 

 Hydrogen." 



Allied to this investigation is another on the spectrum of Helium under very low 

 pressure, and it is convenient to record and reduce the observations relating to this 

 question with those relating to mixed gases, the mode of reduction being identical. 

 In the subsequent discussion, they may also be taken together for the sake of brevity. 



