74 
POPULAR SCIENCE REVIEW. 
spectrum with that of a wedge-shaped piece of glass, known 
to be coloured with oxide of cobalt. If the bead be coloured 
with oxide of cobalt, it is easy to so adjust the wedge-shaped 
glass that the two spectra are precisely the same. But since 
the characteristic light and dark bands are in the green, 
yellow, and orange part of the spectrum, it is quite clear that 
any other substance which did not absorb those rays might be 
present in such quantity as to entirely alter the colour of the 
bead, and yet the bands characteristic of the cobalt might be 
so well seen as to leave no kind of doubt of its presence. 
Oxide of iron and oxide of manganese, which more than any 
other substances are likely to interfere with the ordinary 
method which relies on colour without spectrum analysis, are 
of this character. The bead may be so far coloured green or 
pink that the blue colour of the cobalt is quite disguised, and 
yet the dark and light bands characteristic of cobalt may be 
easily seen. 
Few substances give a more striking spectrum than per- 
manganate of potash. A solution so dilute as to be a pale 
pink gives five well-marked absorption bands, at about equal 
intervals, in the green, and one or more very obscure in the 
blues. Since this solution is decomposed by organic matter, and 
even by long contact with flint glass, it should be sealed up in 
tubes of glass free from oxide of lead. The crystals are too 
highly coloured to show these bands, but if pressed with an 
agate pestle on glass, so as to be very thin, the spectrum 
of the crystalline salt may be seen to be exactly the same as 
that of the solution. I know no better example of the value 
of being able to compare two spectra together, side by side, 
than that furnished by crystals of perchlorate of potash coloured 
with permanganate, as compared with their solution in water. 
These two salts are isomorphous, and readily crystallize 
together; but some care is required to obtain the crystals 
moderately free from fluid-cavities, which render them 
opaque, and to arrange the relative amount of tlie two sub- 
stances, so that they may be dark enough to show the 
absorption bands distinctly, and not so dark as to cause them 
to be united into one broad band. Sometimes on cooling 
from a hot solution, small thin crystals are formed, which are 
admirably suited for examination ; and by slow spontaneous 
evaporation in a vessel covered up, so as to prevent the access 
of particles of dust, larger and sufficiently transparent crystals 
can be obtained. They should be of a clear pink colour. Now, 
though the two salts are isomorphous, and though the per- 
manganate, in a crystalline condition, or dissolved alone or 
along with the perchlorate, when thin enough or dilute enough 
to transmit a bright pink light, gives a spectrum like fig. 5 a. 
