68 



THE ROYAL SOCIETY OF CANADA 



W The spectra were recorded on Schumann plates made by the 

 Adam Hilger Co. With the quartz spectrograph no difficulty prevent- 

 ed the securing of clear, sharply defined spectrograms, but with the 

 fluorite spectrograph, the fluorescence of the prism and lenses produced 

 a heavy general fogging of the plate. Since this did not occur with 

 other arc sources of equal intensity in the visible region, it is possible 

 that this may be due to the strong ultra-violet emission in the mag- 

 nesium arc between 2700 A.U. and 3000 A.U. 



Typical spectra are reproduced in Fig. 1 . The upper spectrum is 

 that of the magnesium spark in air, and below it is the spectrum of 

 the magnesium arc in air, both taken with the quartz spectrograph. 

 The lower illustration is the spectrum of the magnesium arc in vacuo 

 taken with the fluorite spectrograph. Wave length scales are at- 

 tached for reference. 



It may be noted that as was observed by Saunders^ and as men- 

 tioned in the previous work, reversal was readily obtained with the 

 arc in air at 2852 A.U. and 2026 A.U., the frequencies of which are 

 given by j; = (1-5,S) — (2,P) and v= (1-5,S) — (3,P), while no reversals 

 were obtained with the arc in vacuo. 



The wave lengths with the quartz spectrograph were obtained 

 from a calibration curve constructed by using the following prominent 

 lines in the mercury arc and the zinc, cadium and aluminium sparks. 



The calibration curve for the fluorite spectrograph was obtained 

 by measurement of the following prominent carbon, tin and lead 

 vacuum arc lines. 



1 Saunders. Astro.-Piiys. Jl. Vol. 43, No. 3. April, 1916. 



