Li lies in the Spectra of Stars of the /B Criicis Tupe. 47 



With an ordinary discharge it gave the pure helium spectrum, but 

 with the highly disruptive discharge the helium spectrum vanished 

 mtirehj, and was replaced by the unknown lines and the spectrum of 

 oxygen. The helium spectrum could be obtained at will by reverting 

 to the ordinary discharge. 



This helium tube had platinum electrodes, and these last observations 

 finally banished any idea that the gaseous contents of the tubes or the 

 metallic electrodes could be the origin of the substance searched for, 

 and the conclusion that the glass of the tubes contained the substance 

 sought was now irresistible. Yet in some of the spectra from the 

 helium tube, the H and K lines of calcium were absent when those of 

 oxygen were present, showing that the lime of the glass did not neces- 

 sarily account for the presence of oxygen. 



After various fruitless experiments, sparks were taken between the 

 platinum terminals of a broken up vacuum tube on which still adhered 

 some of the blue fusible glass, commonly used in sealing in platinum 

 wire in glass. The spectrum of this spark in air showed the unknown 

 lines. 



Beads of glass made from ordinary soda glass tubing, were then 

 fused on platinum wires, and the spark from these was examined. 

 The unknown lines again appeared. The substance sought was now 

 strongly suspected to be the element silicon. The siliceous diatomace- 

 ous earth " kieselguhr " was next used as the most convenient source of 

 silica, and beads of sodium silicate were made by fusing this material 

 with sodium carbonate on platinum wire. The result of the examina- 

 tion of the spark was that the unknown lines were again found. 

 The next step was to replace the kieselguhr by pure rock crystal 

 obtained from the South African Museum by Dr. Gill. Sodium 

 silicate made from the pure rock crystal, also furnished the unknown 

 lines, whilst the sodium carbonate alone failed to give them. 



These experiments left little room for doubt that the element sought 

 was silicon. Nevertheless, it was very desirable to confirm the result 

 in another way, by examining the spectrum of a gaseous siliceous com- 

 pound. 



Platinum wires were sealed into the ends of a piece of wide glass 

 tubing, f inch internal diameter, the ends of the wires leaving a gap 

 of only f inch for the passage of the spark. The tube was also fur- 

 nished with an inlet and outlet tube for the gas. No capillary tube 

 was used in order to avoid the hot spark coming into direct contact 

 with glass. The tube was then filled with silicon tetrafluoride, and 

 after the gas had been passing for some time, it was sealed off at 

 atmospheric pressure. 



An ordinary discharge passed through the gas without jars or air 

 gap gave a banded spectrum of the compound itself. 



The disruptive discharge obtained by using four jars and an air gap, 



