8 THE ROYAL SOCIETY OF CANADA 



The arc was struck between blunt carbons, held at right angles, the 

 vertical one being filled either with metallic salt or with the metal 

 itself. Owing to the high melting point of the metals investigated 

 it was found necessary to use currents of from 10-20 amperes at 200 

 volts. The vertical carbon was made positive and it was found that 

 a few seconds after the arc struck metallic vapour was passing up in 

 dense clouds in front of the arc which was maintained at the back 

 edges of the carbon. The length of exposure varied fi om 30-60 seconds. 

 Small supplies of vapour in the arc gave the emission spectrum only. 



A calibration curve based on standard wavelengths in the spectra 

 of mercury, zinc, cadmium and magnesium was employed to obtain 

 the absorption wavelengths, measurements being taken for some 

 known line as zero.' A Hilger comparator was used in measuring the 

 plates. 



III. Absorption Spectrum of Thallium 



Previous work by Guthrie^ on the absorption spectrum of thallium 

 has shown that pure thallium gives four absorption bands at 3230 

 A.U., 3092 A.U., 2330 A.U. and 2380 A.U. and on adding mercury 

 additional bands appeared at 3776 A.U., 2768 A.U. and 2580 A.U. 

 It was considered advisable to carry this work further into the ultra 

 violet in the hope of obtaining absorption over a series of wavelengths. 



With thallium the vertical carbon was filled with the chloride. 

 In the heat of the arc the salt became dissociated and free metal was 

 obtained in the form of vapour with a current of from 8-12 amperes 

 at 200 volts. 



The spectrograms showed absorptions at several places which 

 agreed with those given by Dunz^ for the series y= (2,p2) — (m.di) and 

 y=(2,p2)-(m,s). 



These are : For series v = (2,p2) — (md ) 



^ Guthrie Dissertation — Baltimore — 1908. 



* Dunz — Inaugural Dissertation — Tubingen — 1911. 



