169 
whence we obtain 
log ^ = 2^1og^i 
Oj\ 
log = 
Ct2 
log 
a. 
qt^ogk^ 
and by addition we obtain 
Cl C 
log— + log— + &C. = 2^(log^i + log ^2 + ‘^0.) 
Oi 
«2 
or 
= constant. 
log ( K) 
In my last paper I stated that the law of absorption of 
light given by Herschel appears to have been obtained 
a priori ; 1 have not found in his memoirs any experimental 
confirmation of it. The form of the expression has a some- 
what formidable appearance, inasmuch as it involves the 
measurements of infinite varieties of light. But suppose 
that in the formula '2ak*, h is the same for every species of 
light; then we may write T = ^*Da or T = ^*I, if I denote the 
incident light. In such a case the emergent light will be of 
the same nature as the incident light, and will diflfer only in 
intensity. Suppose the incident light to be white, the 
emergent light will be a white of less intensity, — that is, 
will be a grey, approaching to blackness, as the length of 
the column increases. A fluid medium, affecting white light 
in this way, we might call a soluble black, and my aim, in 
seeking to obtain such a fluid, was to apply it to the con- 
firmation of the law. In a previous note I state that I had 
tried to obtain such a body; what I got was not wholly 
satisfactory, but I thought that with it I might obtain some 
approximate results. The solution I used consisted of 500 
cub. c. of the previously mentioned fluid with 500 cub. c. of 
distilled water. 
The mode in which I proposed to operate was as follows : 
take two white lights of different intensities, say Wi and Wa, 
and look at them through the liquid. Suppose the lengths 
of the columns when equality of intensity is obtained to be 
ti and U, then 
