66 THE ROYAL SOCIETY OF CANADA 



It is important to notice that in obtaining the spectra of metals, 

 the secondary gas spectrum is always present. Thus, by having the 

 gas spectrum immediately above, and on the same plate, the lines due 

 to the gas that might otherwise be attributed to the metal, could be 

 easily picked out. When an intermittent arc, obtained by using a 

 small current, was used, the time of exposure was longer and the 

 secondary spectrum always came up strong, while when a high current 

 was used, the arc remained much more steady and the secondary 

 spectrum was much weaker relatively to that of the metal. 



The length of the exposure ranged from 5 to 10 minutes for the 

 metals. The steadier the arc remained, the shorter the time required. 

 When being used, the arc could always be observed through a glass 

 window sealed in the end of the casing of the arc. 



The apparatus was connected up to a hydrogen tank, so that when 

 not in use it could be filled with hydrogen gas. The hydrogen gas, 

 together with phosphorous pentoxide kept the interior free from 

 moisture. 



Results. 



With the apparatus described above, the vacuum arc spectra 

 of lead, tin, iron, nickel, cobalt, and thallium were investigated; also 

 the spectra of copper, aluminium, zinc, and carbon, which were 

 studied by Ainslie last year, were repeated, and the wave lengths of 

 lines carefully measured again. 



In calculating the wave lengths of the lines of these different 

 spectra, certain lines previously determined were used as standards. 

 From these lines, by means of graphical interpolation, the various 

 lines were carefully measured. In working with a prism spectroscope 

 it is necessary to use quite a large number of lines as standards, in 

 order to get accurate results by graphical means. It is impossible 

 to get results with this instrument by referring to a few standard 

 lines, such as Saunders did in using a grating spectroscope. For 

 the region from 1850 down, hydrogen and aluminium lines measured 

 by Lyman, and for the region above 1850, carbon monoxide by Lyman, 

 and copper and thallium lines by Eder and Valenta, were used. 



Copper. 

 The copper vacuum arc spectrum was obtained by using a current 

 of about 9 amperes. With this current the arc was almost con- 

 tinuous. The results obtained agree fairly well with those of Eder 

 and Valenta for the region covered by their work and from A 1750 

 down they agree with the values of Handke^ obtained by using a 

 copper spark. 



1 Lyman, Spectroscopy of the Extreme Ultra-Viotet, p. 122. 



