ACCORDING TO THE UNDULATORY THEORY. 487 
So long as we take 26 less than from three to four times the length 
of the wave of red light, we obtain only one place of the spectrum, or 
both its ends absorbed; but if we increase 6 still more, that is if we 
advance fig. 2 still more forwards, we perceive that more maxima and 
minima appear in the spectrum, the more indeed the greater 26 is 
taken. If we suppose 2 6 = 0:004 of an English inch, we obtain about 
the same number of absorptions as by iodic gas. 
I have endeavoured to produce artificially those kinds of retardations 
which the phenomena of absorption presuppose, and have been so for- 
tunate as to produce in a very simple manner any kind of the phzno- 
mena of absorption I chose. The simplest, and, as I have found, the 
easiest, manner of performing this experiment successfully is the 
following: Bend a piece of a thin plate of mica so that it forms the 
surface of a perpendicular cylinder; then place at some distance a 
lighted candle at the same elevation. The flame which is reflected 
towards the eye from the cylindrical surface must now appear as a slen- 
der vertical line. This light is reflected partly from the front surface 
of the mica, partly, once or more than once, from its hind surface ; 
the retardation of the last part is to that of the first in proportion to a 
distance whose magnitude depends on the thickness of the mica. If 
the thickness of the mica is at all considerable in proportion to the 
length of the wave of light, that is about 0°001 inch and more, the re- 
flected light appears quite uncoloured. But if we divide this light into 
colours by means of a prism, and observe the spectrum through a tele- 
scope, it appears, from the most external red to the most external violet, 
filled with stripes, which are quite black, and more numerous the 
thicker the plate of mica is. After having shown how we can explain 
a great number of phenomena of absorption by the supposition of a 
single retarding cause, I will endeavour to show how we may explain 
all the rest by a further supposition of many other similar retarding 
causes. If we suppose light of the intensity a, which has been sub- 
jected to the action of a retarding medium, and through that brought 
down to the intensity 
a.(1—r)? 
a 
1 — 27! 0082 0 20 + rt 
and subjected anew to a fresh retardation, which of itself would have 
caused the intensity 
2d ert 
= ’ 
af) — 21? cos 2 x 2 + r'4 
it is evident that the result A’ of both retardations must be 
