284 
‘* In vapour of turpentine the stream was narrow and dull; 
fluorescence very faint. At each of the pointed terminals 
was an intense green star, even when they were platinum ; 
and from the negative one red hot globules were projected, too 
obscure to be melted platinum, and therefore probably carbon 
from decomposition. 
“In vapour of alcohol, gauge =0'-7, the stream, which (as in 
the other vapours) was much contracted, was blue, with a tinge 
of green. The terminals (balls in this instance) were covered 
with green sparks, which on the positive one at least do not 
depend on combustion of the brass. Here also was little fluores- 
cence, and as there is no note of dark bands, I suppose they 
were not conspicuous. 
‘¢ As air was present in these vapours, it may be concluded 
that the fluorescence actually observed is due to it, and not to 
the vapours. 
‘* 3. The receiver was filled with coal gas from a burner. 
The sulphuric acid probably absorbed some hydrocarbon from 
it as it became brown. It was exhausted to 0°08. The light 
was livid white, giving the idea of an excess of the more 
refrangible rays, though the prism showed much red. There 
was not more fluorescence than could be explained by the com- 
mon air with which commercial coal gas is often adulterated. 
It, however, exhibited the true nature of the dark bands; 
they are intervals between the luminous menisci of which the 
entire column of light is made up, the centre of whose cur- 
vatures is the bright point on the positive terminal from which 
the discharge breaks out. This structure is far more beauti- 
fully exhibited in— 
“4, Hydrogen. It was obtained from Liege zine and 
diluted sulphuric acid, and passed through solution of potassa. 
The terminals were, in the first case, the point with its card 
disk already described, and an inch ball. When the gauge 
was 1*15 (the Ruhmkorff being excited by three Groves), 
the discharge passed as a crimson spark 3'-5 long and +4; thick. 
