250 Mr. J. F. W. Herschel on the aberrations of 
condition, we may content ourselves as already remarked, 
with annihilating A / for the most luminous rays, but the 
former must be satisfied as rigorously as possible. To fix 
the colour of a ray, we may, either fix its position in a spec- 
trum cast by a prism of a standard substance, or the length 
of its fits of easy transmission and reflexion in vacuo. The 
latter method is on all accounts preferable. The whole dif- 
ference then between the lengths of the fits of an extreme 
red and violet ray being taken for unity, let c be difference 
between those of a ray of any assumed colour and those of 
the most luminous ray in the spectrum, c being positive for 
rays nearer the red end of the spectrum, and negative for 
those nearer the violet. Then in different media, the refrac- 
tive indices [x, [x', &c. for that colour will be functions of c of 
a form depending on the nature of the media, and which per- 
haps is not the same for any two media in nature. What this 
form is in any one medium is at present altogether unknown, 
but in all, we may represent it by 
^ l){/»C-j-^ 2 -J- 7 Y 3 + &c.} 
p being the quantity usually termed the dispersive power of 
the medium, and which even in the most dispersive bodies 
hitherto observed does not exceed 0 4, and is generally a very 
small fraction, while q , r, &c. are numerical co-efficients, 
whose influence was perceived shortly after the discovery of 
the different dispersive powers of bodies, by Clairaut, and 
whose real existence the experiments of Blair, Brewster, 
&c. seem to have placed beyond a doubt. The presumption 
is that they decrease rapidly in magnitude, and are altogether 
insensible in the higher terms of the series. 
