248 
bustion due to the greater supply of air from the holes 
at the bottom. The flame immediately becomes of an 
intense yellow colour due to the vapour of sodium. In 
this we have further evidence of the connection between 
the colour of the light which we get from a vapour and 
the spectrum of that vapour. It is usual to place the salt 
to be examined in a platinum spoon, and insert it in the 
flame, but the utmost constancy is insured by adopting 
an arrangement of Mitscherlich’s shown in the accom- 
panying drawing, (Fig. 24) in which a platinum wick is kept 
continually moistened by a solution of the salt, generally 
the chloride, the spectrum of which is required to be 
examined. You will imagine, @ Aréori, from what I haye 
already said, that as in the case of sodium vapour, the 
colour of the light is orange, the line of the vapour will 
appear in the yellow or orange part of the spectrum, and 
you will not be mistaken, For you 
will see on examining this flame 
with a spectroscope, that we obtain 
a spectrum consisting of a brilliant 
yellow line upon an almost black 
background ; if, however, the flame 
is observed by means of a very 
narrow slit, this line will appear 
double, that is it really consists of 
two extremely fine lines which are 
very close to each other, and if the 
slit be wide the images overlap one 
another, 
If we then pass on to another 
substance, and take some lithium 
instead of sodium, we obtain a bril- 
liant carmine tinted flame, which 
on examination by the spectro- 
scope, is found to give a spectrum 
consisting of one splendid red, and 
a fainter orange line. Potassium 
gives a violet coloured flame, and 
yields in the spectroscope a red 
line and a violet line. If, again, 
we take a salt of strontium, which 
was one of the ingredients in red 
fire, it colours the flame crimson, 
and by the eye the flame can 
scarcely be distinguished from the 
colour of the lithium flame, but in 
the spectroscope there is no possi- 
bility of doubt, the spectrum of 
strontium, consists of a group of 
several lines in the red and orange, 
and a fine line in the blue end of 
the spectrum. 
If a higher temperature than that 
of the Bunsen flame is required the 
blow-pipe flame may be resorted 
to; in this, the quantity of air and 
coal-gas is varied at pleasure, and 
a very high temperature may be 
obtained. 
We might proceed thus to examine all the elementary 
substances one by one, but to observe the spectra of the 
metals, it will be found necessary to use a higher tempera- 
ture still, and for this purpose the electric arc or spark is 
employed. If a temperature only slightly greater than 
that of the blow-pipe flame is used, the spark from an 
induction. coil worked by five Grove cells may be taken 
as shown in Fig. 21, the Leyden jar not being em- 
ployed ; a few metallic lines will then be seen, and a 
background consisting generally of bands of light here 
and there. : 
If a higher temperature still is used, the jar may be 
thrown into the circuit, upon which the spark will become 
more intense, according to the power of the coil and size of 
the jar ; or the electric are may be employed. The spectra 
Fic. 24.—Geissler’s tube, 
showing electric discharge. 
y 
> 
NATURE 
[ Fan. 30, 1873 
thus obtained are much more complex; the spectrum 
of iron, for instance, when examined at this high tempe+ 
rature, is found to consist of no less than 46olines, many of 
which are situated in the green part of the spectrum. 
With regard to solid and vaporous bodies, the electric 
lamp affords a very handy method when properly em- 
ployed, of examining and exhibitimg the spectra of these 
bodies to large audiences. ‘ 
But there are a great many gases which the spectro- 
scopist also has to study, and to study with the greatest 
Fic. 25.—Mitscherlich’s arrangement for flame spectra. 
care ; and] here, I am sorry to say, the electric lamp 
utterly fails us. The light which we get from a gas by 
the electric discharge is so feeble that it is quite impos- 
sible to throw its spectrum on the screen, so as to be 
observed by large audiences, for we cannot render stron- 
tium incandescent in the way in which we render incan- 
ab 
Fic. 26.—Herapath’s Blow-pipe. 
descent sodium and the other substances I have brought 
before you. But we have other means of examining the 
spectra. I have here some tubes containing hydrogen 
and other gases at different pressures, and when we wish 
to study the spectrum of a gas, we do it in this way: we 
enclose it in a tube, and send a current through it by 
means of aninduction-coil. If we pass a stream of elec- 
