26. THERE are methods of separating the heterogeneous rays of the sun’s light by reflection, which perfectly conspire with and confirm this reasoning. One of which ways may be this. Let A B (in fig. 129) represent the window shutter of a darkened room; C a hole to let in the sun’s rays; D E F, G H I two prisms so applied together, that the sides E F and G I be contiguous, and the sides D F, G H parallel; by this means the light will pass through them without any separation into colours: but if it be afterwards received by a third prism I K L, it shall be divided so as to form upon any white body P Q the usual colours, violet at _m_, blue at _n_, green at _o_, yellow at _r_, and red at _s_. But because it never happens that the two adjacent surfaces E F and G I perfectly touch, part only of the light incident upon the surface E F shall be transmitted, and part shall be reflected. Let now the reflected part be received by a fourth prism Δ Θ Λ, and passing through it paint upon a white body Ζ Γ the colours of the prism, red at _t_, yellow at _u_, green at _w_, blue at _x_, violet at _y_. If the prisms D E F, G H I be slowly turned about while they remain contiguous, the colours upon the body P Q shall not sensibly change their situation, till such time as the rays become pretty oblique to the surface E F; but then the light incident upon the surface E F shall begin to be wholly reflected. And first of all the violet light shall be wholly reflected, and thereupon will disappear at _m_, appearing instead thereof at _y_, and increasing the violet light falling there, the other colours remaining as before. If the prisms D E F, G H I be turned a little farther about, that the incident rays become yet more inclined to the surface E F, the blue shall be totally reflected, and shall disappear in _n_, but appear at _x_ by making the colour there more intense. And the same may be continued, till all the colours are successively removed from the surface P Q to Ζ Γ. But in any case, suppose when the violet and the blue have forsaken the surface P Q, and appear upon the surface Ζ Γ, Ζ Γ, the green, yellow, and red only remaining upon the surface P Q; if the light be received upon a paper held any where in its whole passage between the light’s coming out of the prisms D E F, G I H and its incidence upon the prism I K L, it shall appear of the colour compounded of all the colours seen upon P Q; and the reflected ray, received upon a piece of white paper held any where between the prisms D E F and Δ Θ Σ shall exhibit the colour compounded of those the surface P Q is deprived of mixed with the sun’s light: whereas before any of the light was reflected from the surface E F, the rays between the prisms G H I and I K L would appear white; as will likewise the reflected ray both before and after the total reflection, provided the difference of refraction by the surfaces D F and D E be inconsiderable. I call here the sun’s light white, as I have