534 PROFESSOR STOKES ON THE CHANOE OF REFRANGIBILITY OF LIGHT. 
space to which the dispersed light was confined. On looking down from above^ so 
that this space was seen in projection, it appeared in the case of the weak solution 
to have approximately the form of the space contained between one branch of a rectan- 
gular hyperbola, one asymptote, and a line parallel to the other, the first asymptote 
being the projection of the anterior surface, and the line parallel to the other being the 
course of the least refrangible of the active rays which were capable of producing a 
sensible quantity of dispersed light. The breadth of the illuminated space, which among 
the most highly refrangible rays was almost insensible, continually increased, until 
the space ended in a blue beam which went quite across the vessel. But in the case 
of the strong solution the illuminated space had throughout an almost insensible 
breadth, except just close to its lower limit, that is, the limit corresponding to the 
least refrangible of the active rays, where it ended in a sort of tail or plano-concave 
wedge, which penetrated to a moderate distance into the fluid. Hence one reason, 
though perhaps not the only reason, why the strong solution showed a copious 
dispersion from G to G|^H, where the weak solution showed hardly any, is plain 
enough. But in the region of the invisible rays beyond the violet, the dispersion 
was plainly more copious with the weak than with the strong solution. It appears 
then that in such a case the sensitive molecules do not act independently of each 
other, but the quantity of light emitted by a given number of molecules is less, in 
proportion to the light (visible or invisible) consumed, than when a solution is more 
dilute. We should expect a priori that when a solution is tolerably dilute further 
dilution would make no more difference in this respect. This seems to agree very 
well with experiment. For when a pretty dilute solution and one much more dilute 
are compared with respect to the quantity of dispersed light given out in a given 
portion of the incident spectrum, they appear to be alike. I suppose the comparison 
to be made with respect to such a portion of the incident spectrum, or in the case of 
solutions of such strength, that the dispersed light is confined to a space extending to 
no great distance into the fluid in either solution. Under these circumstances the 
comparison may be made easily enough. 
188. In the actual experiment, the elementary portions of light coming from the 
elementary strata of fluid situated at different distances from the anterior surface 
enter the eye together. Let us however trace the consequences of the very natural 
supposition, that in passing across a given stratum of fluid the quantity of light 
absorbed, as well as the quantity given out by dispersion, is proportional, cceteris 
paribus, to the intensity of the incident light. The incident light is here supposed 
to be homogeneous, and to belong indifferently to the visible or invisible part of 
the spectrum. In crossing the elementary stratum having a thickness dt, let the 
fraction qdt of the incident light be absorbed, and the fraction rdt dispersed in such 
a direction as to reach the eye ; and of the latter portion let the fraction sdt be 
absorbed in crossing a stratum having a thickness dt, s being different from q on 
account of the change of refrangibility. Then by a very simple calculation similar to 
