942 
EEV. T. E. EOBINSON ON SPECTEA OE ELECTEIC LIGHT. 
a platinum wire is sealed long enough to dip in the mercury and serve as the negative 
electrode, ( 3) a tube also dipping in the mercury, and with a platinum electrode at top. 
If the bottle be filled one-third with pure mercury, and the stopcock be connected by a 
caoutchouc tube* with the transfer and the apparatus exhausted, the tube can be filled 
with mercury, which on erecting it falls, leaving a vacuum, through which discharges 
can be passed by the electrodes. If the tube has heexi ])erfectly cleaned by fining it with 
nitric acid and washing it with distilled water, dried by sulphuric acid in a vacuum, and 
finally wiped with a morsel of linen which has been boiled in distilled water, after being 
fastened to a flexible wire, on inclining the tube the mercury will fill it without leaving 
the least speck of air, and will often adhere with considerable force. It, however, always 
falls at the first discharge. In this case the space is filled with mercurial vapour alone. 
If a few drops of any volatile fluid be introduced into the tube, by filling it with mercury 
the excess is expelled, and the vacuum contains only its vapour highly rarefied. With 
phosphorus, the tube filled with water was warmed till it fused and adhered to the 
upper part ; the water was removed, and the tube put in its place. The apparatus was 
then repeatedly filled with dry nitrogen till all traces of moisture disappeared. If a 
platinum wire of sufiicient length be introduced below, both electrodes are platinum ; 
and if the upper part be bent so that the descending branch remains full of mercury, 
both are of that metal. 
Prisms . — For three-fourths of the observations I used a prism by Meez, with an angle 
45° 35' •4, By sets of from eight to twelve it gives, for the deviations of Frauxiiofer’s 
lines, 
' o / 
A . . 
. 32 20-00 . . 
. y. 1-6230 
B . . 
. 32 32-04 . . 
. . 1-6259 
C . . 
. 32 38-88 . . 
. . 1-6285 
D . . 
. 32 56-50 . . 
. . 1-6336 
E . . 
. 33 20-26 . . 
. . 1-6405 
F . . 
. 33 41-67 . . 
. . 1-6467 
G . . 
. 34 23-06 . . 
. . 1-6587 
H . . 
. 35 0-06 . . 
. . 1-6692 
It is therefore nearly identical with Fraunhofer’s flint No. 2, but in dispersive power 
it is far inferior to those used by Masson and Plucker. After working with it a long 
time, I found that several bright but cloudy bands which it showed were resolved into 
two, three, or more by a prism of bisulphuret of carbon having an angle of 60°. The 
indices of this fluid change so much with temperature, that I did not venture to employ 
it; and I obtained from Mr. Duboscq a prism of 60° 3'-54 angle, which, though nearly 
of the same density as the Merz, is more effective, in the proportion of 3 to 2. Deter- 
mining with it Fraunhofer’s lines, and comparing with their deviations in Merz, I 
* This, with which I was supplied by Messrs, Siltee, is far superior to vulcanized tube, which always 
leaks. The glass apparatus was made with great precision and intelligence by Mr. Casella. 
