510 PROFESSOR STOKES ON THE CHANGE OF REFRANGIBILITY OF LIGHT. 
except when the active and the dispersed light have very nearly the same refrangi- 
bility. But such changes could not be observed even with a pure primitive spectrum, 
because in the place where they occur the primitive and derived spectra overlap ; 
and independently of this, the brilliancy of the primitive spectrum would prevent all 
exact observation of the derived. It is true, that in the case of chlorophyll, or some 
of its modifications, changes of intensity having apparently somewhat the same 
nature were observed when the active and the dispersed light were widely separated 
in refrangibility. But the sensibility of this substance is difficult, if not impossible, to 
observe in the case of a washed paper or a green leaf, except by one of the methods 
not yet described, so that it is not to be expected that such fluctuations could be 
made out. Besides, it is to be remembered that the fluctuations observed in the case 
of solutions of chlorophyll, were fluctuations in the rate at which dispersed light was 
produced, not fluctuations in the sum total of the dispersed light produced by the 
time the active light was exhausted. Fluctuations of the former kind by no means 
imply fluctuations of the latter; and indeed, the circumstance, that maxima of 
activity in the solution correspond to minima of transparency, would seem to show 
that the total quantity of light dispersed, considered as a function of the refrangibility 
of the active light, is not subject to these fluctuations, or at least not to anything 
like the same extent. Now the total quantity of red light dispersed by a green leaf, 
or by a paper washed with a solution of chlorophyll, must depend upon the sensibility 
of this substance and upon its transparency conjointly, and therefore it is likely 
enough that such maxima and minima would not be observed, even were the 
dispersed light much stronger than it is. 
107. Suppose now the slit by which the light enters to be placed in a horizontal 
instead of a vertical position, so as to lie in the plane of refraction. Corresponding 
to light of any given refrangibility, the image of the slit formed after refraction 
through the prisms and lens will now be a narrow parallelogram, which may be 
regarded as a horizontal line. The series of these lines, succeeding one another in a 
horizontal direction, and consequently overlapping, forms the spectrum incident on 
the body examined. This spectrum is now no longer pure, but only approximately 
so, a point, however, which, as we have seen, is not of much consequence. But by 
this trifling sacrifice two very great advantages are gained. The first is increase of 
illumination. When the slit is vertical, the spectrum received on the body occupies 
a rectangle having for breadth the length of the image of the slit ; but when it is 
horizontal, the same, or very nearly the same quantity of light is concentrated into a 
rectangle having the same length as before (the length of the image of the slit being 
disregarded compared with that of the spectrum), but having for its breadth only 
the length of the image of a line drawn across the slit. Hence the intensity of the 
incident light is increased in the ratio of the breadth to the length of the slit. The 
second advantage is purity in the derived spectrum, a point of much consequence, 
because sometimes the composition of this spectrum presents very remarkable 
