24 THE ROYAL SOCIETY OF CANADA 



call to our aid the theory put forward by Rutherford as to the possi- 

 bility of disrupting carbon atoms into tri-protonic helium to account 

 for the wave-lengths found by us in the spectrum of helium which 

 closely corresponded to wave-lengths found by Millikan in the spark 

 spectrum of carbon. 



From the investigation generally, our conclusion is that Millikan 

 has good warrant for ascribing to carbon the origin of a number of 

 the wave-lengths obtained by Lyman in photographing the spectrum 

 of helium. Our experiments lead us to conclude that it is exceedingly 

 difficult to keep helium when it is used in a vacuum grating spectro- 

 graph entirely free from contamination with carbon in some form. 

 It may be that carbon was present in the helium used by Lyman. 



The wave-lengths found by us for hydrogen when compared with 

 the values obtained by Lyman for the spectrum of this gas, agree very 

 closely with the latter in the majority of cases. In 27 cases, however, 

 our wave-lengths are also within 2 A.U. of the values given by him 

 for the spark spectrum of mercury. We consider that this limit, 

 however, is greater than the probable error of our results, and we are 

 inclined to the view that the w^ave-lengths tabulated by us for hydro- 

 gen are correctly attributed to this gas. The wave-lengths X = 1662.6 

 Â.U., X = 1672.4 Â.U., X = 1760.0 Â.U., X = 1795.7 Â.U., X = 1806.2 

 A.U. and X = 1832.7 A.U. are within 0.5 Â.U. of mercury spark wave- 

 lengths given by Lyman, and it may be that in these cases the radia- 

 tions should be attributed to mercury. Against this view, however, 

 there is the fact that no trace was obtained on our hydrogen plates of 

 the strong mercury wave-lengths X = 1942.2 A.U. and X = 1849.6 Â.U. 

 which would indicate that the hydrogen used by us was practically, 

 uncontaminated with mercury vapour. 



V. Conclusions 



The conclusions which may be drawn from the present investi- 

 gation are: 



1. That the wave-lengths X = 1931 Â.U., X = 1657 Â.U., and 

 X = 1561 A.U. frequently obtained in the spectrum of helium, as well 

 as others of less intensity, have their origin in the atoms of carbon. 



2. That the series of wave-lengths whose frequencies are given 

 by i' = 4N I -^—-) exists for the spectrum of the spark discharge 

 in helium. 



