But let the square No. 2, seen through a crown-glass prism g, be supposed to be displaced by refraction three compartments, exhibiting the coloured borders to a certain extent; again, let the square No. 3, seen through a flint glass prism h, in like manner be moved downwards three compartments, when it will exhibit the coloured borders by about a third wider than No. 2.
295.
Again, let us suppose that the square No. 4, has, like No. 2, been moved downwards three compartments by a prism of crown-glass, and that then by an oppositely placed prism h, of flint-glass, it has been again raised to its former situation, where it now stands.
296.
Here, it is true, the refraction is done away with by the opposition of the two; but as the prism h, in displacing the square by refraction through three compartments, produces coloured borders wider by a third than those produced by the prism g, so, notwithstanding the refraction is neutralised, there must be an excess of coloured border remaining. (The position of this colour, as usual, depends on the direction of the apparent