20 THE ROYAL SOCIETY OF CANADA 



observers have recorded, besides, that these wave-lengths are fre- 

 quently present in the spectra of many of the metals when sparked 

 in hydrogen or helium. It is, however, practically certain that these 

 wave-lengths are due to carbon and to carbon alone. It would follow 

 from this that in spite of all the precautions taken the helium used 

 in the present investigation was not entirely free from carbon in 

 some form. Since the wave-lengths X = 1931 Â.U. and X = 1561 A.U. 

 were not recorded in the spark spectrum of h^^drogen, and since the 

 wave-length X = 1657 A.U. came out with only weak intensity in that 

 gas it would indicate that the hydrogen was contaminated with carbon 

 to a less degree than the helium. 



Interest also attached to the wave-lengths X = 1640.2, 1215.1, 

 1086.1, 1026.0, and 972.7 Â.U. These, as Hicks^ has pointed out, 

 are the first six members of the series spectrum of the helium ion 



whose frequencies are given by 1^ = 4 N( -- -^ V Of these wave-lengths 



X = 1640.2 A.U. did not appear on any of the plates of the spectrum of 

 helium obtained by us, neither was it obtained by Millikan in the 

 spark spectrum of carbon. It has, however, been observed by Lyman 

 with powerful disruptive discharges in helium. It would appear, 

 therefore, that this wave-length can be emitted by helium provided 

 the gas is subjected to a sufficiently powerful stimulus. Owing to 

 the absence of this wave-length from the spectra of helium obtained 

 by us we are forced to conclude that the excitation used was not 

 sufficiently intense. Possibly its existence was masked by the action 

 of the ionized helium. 



Regarding a wave-length at or near X = 1215 A.U. it will be 

 seen that one was obtained by us at X = 1215.8 A.U. with the dis- 

 charges in helium and hydrogen and with the lead spark in helium. 

 A wave-length at X = 1215.7 A.U. was also found by Millikan on his 

 plates of the spark spectrum of carbon. 



From the evidence given above it will be recalled that the hydro- 

 gen as used by us was probably only contaminated with carbon to 

 an extremely small degree, if at all. The wave-length X = 1215.8 À.U. 

 obtained by us with hydrogen and with helium came out very strongly 

 with both gases, the intensity being 12. Our conclusion regarding 

 this wave-length is, then, that it is emitted by both helium and 

 hydrogen and that in Millikan's experiments with carbon it may have 

 had its origin in hydrogen occluded in his electrodes or in the walls 

 of his spectrograph, or it may have originated in tri-protonic helium 



6 Hicks, Nature. 104, p. 393, Dec. 18, 1919. 



