270 Scientific Proceedings, Royal Dublin Society. 



the lines in a spectrum. To detect the presence of a trace of a substance 

 the most persistent or ultimate lines of that substance obtained under the 

 same conditions require to be known ; that is, those lines which are the last 

 to become extinct as the quantity sparked is diminished. The absence from 

 a spectrum of these lines for a particular element proves conclusively the 

 absence of tliat element, no matter how many other lines in the spectrum 

 appear to coincide with otliers of that element. As an example, suppose 

 the spectrum of a certain sample of zinc has been obtained, and the wave- 

 lengths of the lines measured, if none of these correspond with the ultimate 

 lines of cadmium, that is conclusive proof that there is not the smallest 

 trace of cadmium present, no matter what lines apparently correspond with 

 the other lines of cadmium. In order to make spectrographic analysis com- 

 plete it is obvious that the ultimate lines of all elements under all conditions 

 require to be known. 



The vacuum-tube was first invented by Geissler, and greatly improved 

 by Plucker and Hittorf,' who found that when a capillary tube was placed 

 between two wider portions, the wliolo exhausted, and a current of electri- 

 city passed through, the brilliancy of the illumination was enormously 

 increased in tlie narrow portion. These tubes have long been extensively 

 used for obtaining the spectra of gases, but they are now equally useful for 

 tlie spectrographic analysis of solid substances, owing to the recent improve- 

 ments by Dr. Pollok. His tubes, which somewhat resemble those of Plucker, 

 were used throughout the present investigation. They are composed entirely 

 of quartz, and are open at each end, having the electrodes sealed into glass, 

 and attached by rubber tubing, as described in his paper, " On the Vacuum- 

 Tube Spectra of the Vapours of some Metals and Metallic Chlorides."'' 

 Tlie general arrangement of the apparatus in the present experiments, and 

 the method of working, were substantially the same as those described in the 

 paper referred to. Tlie fact that the tubes are made of quartz enables the 

 whole of the ultra-violet region of the spectrum to be photographed with 

 facility. 



The presence of a much smaller quantity of an element may be detected 

 by the vacuum-tube spectra than could possibly be determined by the ordinary 

 balance. 



The apparatus used consisted of a vacuum-tube, spectrograph, coil, 

 condenser, pump, and driers. The spectrograph was one of Sir Walter Hartley's 

 design, as used in his researches described in the Sclent. Trans. Boy. Dub. Sec, 

 1882, Vol. I., pp. 231-238, and having quartz lenses of 15 inches focal length. 



1 Phil. Trans. Roy. Soc, 1865, vol. civ., pp. 1-29. 

 - Scient. Proc. Roy. Dub. Soc, 1912, vol. xiii., p. 20^. 



