1878.] on the Reversal of the Lines of Metallic Vapours. 133 



The iron tube had its ends covered with glass plates, and was heated 

 in a furnace supplied with charcoal. By this means he did not succeed 

 in observing any new reversal of bright lines, but he noticed an un- 

 known absorption line which sometimes appeared when zinc was expe- 

 rimented on. He confirmed, however, the channelled- space absorp- 

 tion spectra observed by Roscoe and Schuster, in the cases of potas- 

 sium and sodium, and recooded channelled-space spectra in the case of 

 antimony, phosphorus (?), sulphur (previously observed by Salet) and 

 arsenic (probably). "As the temperature employed for the vola- 

 tilization of the metals did not exceed bright redness, or that at which 

 cast-iron readily melts, the range of metals examined was necessarily 

 limited ; and in order to extend these observations to the less fusible 

 metals, as well as to ascertain whether the spectra of those volatilized 

 at the lower temperature would be modified by the application of a 

 greater degree of heat," a new series of experiments were undertaken 

 by Lockyer and Roberts, in which the combined action of a charcoal 

 furnace and the oxy-hydrogen blowpipe was employed. By this 

 means they obtained no new reversal of a metallic line, but they 

 observed channelled-space spectra in the cases of silver, manganese, 

 chromium, and bismuth. They observed, however, that the metal 

 thallium gave the characteristic bright green line, the light of the arc 

 not being reversed. 



In the above-mentioned experiments the coolness of the ends of the 

 tube, which acted as condensers of the metallic vapours, and the 

 inequality of density and temperatures necessarily produced by the 

 maintenance of a current of hydrogen in the tube, appear to us to 

 account for the complication and uncertainty of the results of the 

 observations. 



In order to examine the reversal of the spectra of metallic vapours, 

 we find it more satisfactory to observe the absorptive effect produced 

 on the continuous spectrum emitted by the sides and end of the tube 

 in which the volatilization takes place. For this purpose we find it 

 convenient to use iron tubes about half-an-inch in internal diameter, 

 and about 27 inches long, closed at one end, thoroughly cleaned inside, 

 and coated on the outside with borax, or with a mixture of plumbago 

 and fireclay. These tubes are inserted in a nearly vertical position in 

 a furnace fed with Welsh coal, which will heat about 10 inches of the 

 tube to about a welding heat, and we observe through the upper open 

 end of the tube, either with or without, a cover of glass or mica. To 

 exclude oxygen, and avoid as much as possible variations of tempe- 

 rature, we introduce hydrogen in a gentle stream by a narrow tube 

 into the upper part only of the iron tube, so that the hydrogen floats 

 on the surface of the metallic vapour without producing convection 

 currents in it. By varying the length of the small tube conveying the 

 hydrogen, we are able to determine the height in the tube to which the 



