70 
POPULAR SCIENCE REVIEW. 
more and more of tlie spectrum is obscured, until we may 
have a spectrum consisting merely of blue, shading off into 
obscure green, or of red shading off into obscure orange. 
Such spectra afford little positive evidence of what the object 
may be, since a great variety of simple or mixed substances 
might give just the same results. On the contrary, some 
objects, even when the colour is very faint, almost totally cut 
off* the light in well-defined bands, of which the general 
character and position may be so characteristic that there is 
little fear of confounding the substance with any other, even 
if its ordinary colour be disguised by coloured impurities. 
Sometimes these so-called absorption bands are broad and 
indistinct, but in other cases remarkably narrow and well 
defined, and are then so much the more characteristic. Some- 
times only one exists, but in other cases they are numerous, 
and they appear to me to be of special interest, as showing 
some intimate relation between the molecules of the body and 
vibrations of light of particular velocities — the molecules, per- 
haps, having a natural tendency to vibrate in unison with par- 
ticular waves of light, and to convert them into heat. If this 
view be correct, we might expect that they would furnish us 
with a test for otherwise very inappreciable molecular changes, 
and such appears to me to be the fact. It is, indeed, somewhat 
analogous to what would occur if the vibration of the air cor- 
responding, for example, to the notes A, B, C, D, were to 
pass through a number of strings all tuned to B. They would 
be set in motion by the vibrations corresponding to B, and the 
intensity of that note would be diminished by the force 
required to move the strings ; whereas A, C, D, would not 
be affected in that manner, and we should have, as it were, an 
absorption band at B. If then we found that the absorption 
band was at C, we should know that some change had 
occurred to alter the pitch of the strings from B to C. Excel- 
lent examples of similar changes are met with in coloured 
spectra. 
The objects most easily obtained, and which furnish us with 
the greatest variety of spectra, are coloured crystals, coloured 
solutions, and coloured glasses. The spectrum-microscope 
enables us to examine the spectra of very minute crystals, of 
very small quantities of material in solution, and of small blow- 
pipe beads. As previously named, the thickness of the object 
makes a very great difference in the spectrum. For example, 
an extremely thin crystal of ferridcyanide of potassium cuts 
off all the blue rays, and leaves merely red, orange, yellow, 
and more or less green; but on increasing the thickness the 
green and yellow disappear, and when very much thicker little 
^lse but bright red light is transmitted. In all such cases, 
