PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OP LIGHT. 533 
tion, to be itself colourless and insensible. Suppose the fluid to be illuminated by 
light of given intensity and given refrangibility entering at the face next the eye, and let 
the eye E from a given position look in the direction of a given point P in the nearer 
surface of the vessel. In short, let everything be given except the strength of the 
solution. For the sake of simplicity regard the eye as a point, and make E the vertex 
of an indefinitely thin conical surface surrounding the line EP. Call this conical 
surface C, and let c be the surface within the fluid generated by right lines coinciding 
with the refracted rays which would be produced by incident rays coinciding with 
the generating lines of the surface C. This latter surface we may if we please regard 
as cylindrical, since we shall only be concerned with so much of the fluid contained 
within it as lies at a distance from P less than that at which the light entering the 
eye in consequence of internal dispersion ceases to be sensible; and in the cases to 
which the present investigation is meant to apply this distance is but small compared 
with PE. Let the fluid within c be divided into elementary portions by planes 
parallel to the surface of the fluid at P, and at distances from P proportional to the 
strength of the solution. It is evident that an element of a given rank, reckoned 
from P, will contain a constant number of sensitive molecules, and the incident light 
in reaching this element has to pass through a thickness of the medium such that a 
plate of the same thickness, and having a given area, contains a given number of 
sensitive or absorbing molecules. The same is true of the dispersed light which 
proceeds from the element and enters the eye. Nowit seems natural to suppose that 
if the strength of a solution be doubled, trebled, &c., or reduced to one-half, one-third, 
&c., the quantity of light absorbed will be the same provided the length of the path 
of the light be reduced to one-half, one-third, &c., or doubled, trebled, &c. This 
comes to the same thing as supposing that each absorbing molecule stops the same 
fractional part of the light passing it, whether the solution be more or less dilute. 
We should similarly be inclined to suppose that each sensitive molecule would give 
out the same quantity of light, when influenced by light of given intensity, whether 
it belonged to a stronger or a weaker solution. If we admit these suppositions, it is 
plain that the quantity of dispersed light which reaches the eye from the element 
under consideration will be independent of the strength of the solution. This being 
true for each element in particular will be true for the aggregate effect of them all, 
and therefore the quantity of light exhibited by dispersive reflexion will be indepen- 
dent of the strength of the solution. It may be readily seen that the result will be 
the same if we take into account the finite size of the pupil. 
187. Now this is by no means true in experiment. On examining in a pure spec- 
trum a highly concentrated solution of sulphate of quinine, a copious dispersion was 
observed to commence a little below the fixed line G. It remained very strong as 
far as H, and beyond. In the weak solution first mentioned in this paper, it will be 
remembered that the dispersion seemed to come on about G|^H. The reason of this, 
or at least one reason, is evident, and was very prettily shown by the form of the 
