THE SPEOTRUM-MICEOSCOPE. 
73 
both distinctly is the most refined ttest I have yet found for a 
spectrum-microscope. Then there is another much broader 
absorption band in the green, which is not well seen until the 
thickness or strength of the solution is such, that the bands in 
the red and orange have so far coalesced, that the spectrum con- 
sists of a well-defined bright greenish-yellow band at one end, 
and a broad blue at the other. Now, the relation between this 
spectrum and that of blue cobalt glass, fig*. 4 d, is very curious 
and instructive. Instead of having, as it were, two double 
absorption bands, there are two single. The rest of the spec- 
trum is similar, but is as it were dislocated, or, to use a 
geological expression, faulted and thrown down, as will be 
seen on comparing it with c. The result is, that in the case 
of the glass there is a far broader band of red, which is 
visible when the thickness of the object is sufficient to cause 
the two absorption bands to coalesce ; and then, besides the 
broad blue, there is a bright band in the green ; whereas, in 
the deep blue solution, it is in the yellow, and there is no ex- 
treme red band. Day-light is too poor in extreme red rays to 
show these facts to advantage, and they are also obscured by 
the black Fraunhofer lines. These different spectra, obtained 
with various modifications of cobalt, serve very well to explain 
what I mean when I speak of characteristic and uncharacter- 
istic spectra. Not, indeed, that we can draw a definite line 
between them — it is more a question of degree — but we may 
say that c is so remarkably characteristic that one cannot 
expect that any other substance would give the same result, 
whilst a is so little characteristic that the colouring matter of 
various red fruits shows precisely the same spectrum. As 
the spectrum of a glass, d is also sufficiently characteristic; 
and as that of a crystal, b is so much so that it would 
probabty be very difficult to find any other that would give a 
similar result. These four spectra will also serve to illustrate 
the effect produced by different molecular states of the same 
substance. The change from b to a may probably indicate 
that the hydrated crystallized chloride of cobalt combines with 
additional water when it dissolves in water ; whilst, when dis- 
solved in a concentrated solution of a salt like chloride of 
calcium, it loses its former water of crystallization. On the 
contrary, when the anhydrous oxide is dissolved in glass, the 
spectrum is still further modified. 
After this description of the spectrum of glass coloured 
with oxide of cobalt, the application of the spectrum-micro- 
scope to the examination of blow-pipe beads suspected to 
contain cobalt will be readily understood. Supposing we 
have got a borax bead coloured more or less blue, we may 
place it on the stage of the microscope and compare its 
