206 Scientific Proceedings, Royal Dublin Society. 



given in each case, a pliotograph being first taken with the coil alone, and 

 then with a Leyden jar or condenser in circuit. In Plate XV., metallic spark 

 spectra are marked 8, vacuum tube spectra without tlie Leyden jar V, and 

 vacuum spectra with the Leyden jar V. 



The cliaracter of the discharge was interesting, the colour being regulated, 

 of course, by the dominant lines in the visible part of the spectrum, but the 

 appearance was greatly affected by the introduction of the Leyden jar, the 

 change being generally from a voluminous discharge of uniform colour, to 

 a tliin crackling discharge of varying colour, bordered with green, the green 

 lines that came out being those of chlorine, which for the most part did not 

 show without the condenser, and did not show with equal facility with all 

 compounds. 



The vapours of metals and their chlorides, such as cadmium, zinc, and 

 mercury, conducted the current with great ease, while the vapours of metal- 

 loids and their chlorides, such as arsensic and antimony, allowed the spark to 

 pass with difficulty, and only when a precisely suitable quantity of vapour 

 was present. 



The tables give a comparison of the relative intensities of tlie more pro- 

 minent lines of the vacuum tube spectra of the metals with and without a 

 condenser, compared with the spark spectra of the same metals. Plate XV. 

 shows reproductions of tlie spectra themselves, together with a rough scale 

 of wave-length to facilitate identification of the lines. The wave-lengths and 

 intensities taken for the spark spectra of the metals are those published by 

 Dr. Marshall Watts, and are mostly from the measurements of Eder and 

 Valenta, Exner aud Haschek, and Kayser and Runge. 



The general conclusions that may be drawn from a study of the spectra 

 of the vapours of the metals and their chlorides, so far examined, are : — - 



1. That the lines shown by the vacuum tube spectra of metallic vapours 

 consist of all, or some, of the lines shown by the metallic spark spectrum of 

 the same element. 



2. Tfiat those lines whicli show as continuous lines in the spark spectrum 

 of an element are invariably the strong lines of the vacuum tube spectrum, 

 and discontinuous spark lines often do not show, particularly when no 

 Leyden jar is used. 



3. That certain lines are enhanced by the introduction of a Leyden jar, 

 and that such lines are very frequently prominent discontinuous lines in the 

 spark spectrum of the metal. 



4. That some lines show more prominently without a Leyden jar than 

 with it, but this is a rather exceptional phenomenon. 



