PoLLOK — The Vacuum Tube Spectra of some Metallic Vapours. 205 



dryers aud pump is made by having the tube beyond the stop-oock drawn 

 out a little, covered with a piece of thin india-rubber tubing, and capable of 

 slipping into the cup of an ordinary mercury junction, so that the tubes can 

 be rapidly taken to pieces, cleaned, filled, and exhausted. The illustration 

 shows the tube on its stand, with the wires connecting the electrodes to 

 the coil, and a Meeker burner in position for vaporizing the chloride 

 contained in the tube. 



The experiments were all performed in substantially the same way: about 

 five centigrams of the substance to be examined were placed in the lower limb 

 of a clean dry quartz vacuum tube ; the stopper with its electrode was then 

 attached to this end, and connected to the electro-negative pole of the coil. 

 The tube and stopcock with its electrode were attached to the other end, 

 and to this, the upper electrode, was connected the electro-positive pole of the 

 coil. The connecting tube was slipped into the cup of the mercury junction, 

 shown at the side, and the vacuum tube was then exhausted by two or three 

 strokes of tlie pump, warmed a little, and exhausted again for the removal 

 of nitrogen by the traces of water vapour that are invariably present, and 

 are found very convenient for this purpose. A Meeker or Bunsen burner was 

 then placed under the capillary portion of the tube to keep it hot, and 

 the side bulb containing the substance warmed by a Bunsen or Meeker burner 

 according to the volatility of the material under examination, the burner 

 being best held in the hand, and applied as required, so as to maintain 

 a steady luminescence in the tube. 



If large quantities of vapouT are evolved, bands are generated in the 

 spectra, and this seems to be universally the case ; but as the author's present 

 desire is to get a preliminary knowledge of the tines common to the spark 

 and vacuum tube spectra of the metals and their compounds, this part of the 

 inquiry was not followed up, and care was taken to avoid excessive heating, 

 so as not to obscure the lines by the development and extension of bands. 

 The author hopes to make a more critical examination of the bands later ; 

 and the generalisations regarding lines do not apply to bands. 



When possible, a photograph was first taken of the spark spectrum of 

 the metal, using metallic electrodes, in air, with a Hemsalech self-induction 

 coil for the removal of air lines, and a spherical condensing lens to throw 

 the image of the spark on the slit, thus preserving the character of the 

 continuous and discontinuous lines, which of course is lost if a cylindrical con- 

 densing lens is used. Wratten and Wainwright plates were used throughout. 

 The exposure for the metallic spark spectra was one minute. With the 

 vacuum tubes the same spherical condensing lens was used, to throw the 

 image of the capillary tube on the slit, and an exposure of five minutes was 



