372 Prof. G. D. Liveing on the Line-spectrum of 



every case lie found it to be a faint continuous spectrum 

 extending from red to blue, fading out at both extremities, 

 and brightest in the green. No line but the yellow line of 

 sodium was to be seen, except when he used jets of brass. 

 Even in this case it was not until the metal had become a 

 little oxidized that any lines were visible. When this occurred 

 several bright lines were seen in the red, green, and blue, and 

 the brightest of them he found to be identical with those 

 produced by copper chloride in the flame of a Bunsen burner. 



In 1868 (Proc. K. S. xvi. p. 421) Frankland stated that he 

 had observed the combustion of hydrogen in oxygen under a 

 pressure gradually increasing up to ten atmospheres ; and 

 that at the last-named pressure the spectrum of the flame was 

 perfectly continuous. He does not say what it was at the 

 lower pressures. Salet also (Compt. Rend, lxviii. p. 404) says 

 that hydrogen burnt in oxygen gives a continuous spectrum. 



Some years ago, when examining the spectrum of the arc 

 in hydrogen, given by a powerful dynamo, I was struck by 

 the fact that no line of hydrogen, bright or reversed, could 

 be detected, except at the moment of breaking the arc. This 

 led me to look at the oxyhydrogen flame ; when I found, to 

 my surprise, that the hydrogen lines were absent from the 

 flame. Subsequently (Proc. R. S. xxx. p. 157) Dewar and I 

 found that the alternating current from a De Meritens mag- 

 neto-electric machine, arranged for high tension, produced an 

 arc in hydrogen which showed H a fairly bright and H^ as a 

 faint diffuse band. Secchi (Compt. Rend, lxxvii. p. 177) 

 stated that the negative pole of an arc showed lines of hydro- 

 gen and of the metals in the carbons, and that a drop of 

 water thrown on the carbons caused the hydrogen-lines to 

 appear, and many others. Subsequently we found (Proc. R. 

 S. xxxv. p. 75) that when a small drop of water was let fall 

 into the arc given by a dynamo machine, both H a and Hp 

 were visible in the spectrum for an instant. 



Since then I h#ve made a very great many observations of 

 the spectrum of the oxyhydrogen flame, both directly and by 

 photography, but have never met with a trace of the hydrogen- 

 lines in the flame. The jets used, the proportions of the two 

 gases, and their pressures have been varied in all sorts of 

 ways, but with the same negative result as regards the line- 

 spectrum of hydrogen. In particular, when mapping the 

 spectrum we have ascribed to water (Trans. R. S. 1888, A. 27) 

 I took a great many photographs of the spectrum of the oxy- 

 hydrogen flame, at atmospheric pressure, in the region from 

 about \ 350 to \ 450, with exposures of 3 hours, and in a few 



