We have just now alluded to an experiment with a glass globe:--If, then, we take either a solid glass globe, or a hollow globe filled with water, and suspend it so high in the solar rays above the eye, that the spectator, with his back to the sun, can see the globe _red_;--if it be lowered slowly, he will see it orange, then yellow, then green, then blue, then indigo, and then violet; so that the drop at different heights, shall present to the eye the seven primitive colours in succession. In this case, the globe, from its form, will act in some measure like a prism, and the ray will be separated into its component parts. The following figure will more particularly illustrate this point. Suppose A (fig. 35.) to represent a drop of rain--which may be considered as a globe of glass in miniature, and will produce the same effect on the rays of light--and let SD represent a ray from the sun falling upon the upper part of the drop at D. At the point of entering the drop, it will suffer a refraction, and instead of going forward to C, it will be bent to N. From N a part of the light will be reflected to Q--some part of it will, of course, pass through the drop. By the obliquity with which it falls on the side of the drop at Q, that part becomes a kind of prism, and separates the ray into its primitive colours. It is found by computation that, after a ray has suffered two refractions and one reflection, as here represented, the least refrangible part of it, namely the _red_ ray, will make an angle with the incident solar ray of 42° 2´, as SFQ; and the violet, or greatest refrangible ray will make with the solar ray, an angle of 40° 17´, as SCQ; and thus all the particles of water within the difference of those two angles, namely 1° 45´--(supposing the ray to proceed merely from the centre of the sun)--will exhibit severally the colours of the prism, and constitute the _interior_ bow of the cloud. This holds good at whatever height the sun may chance to be in a shower of rain. If he be at a high altitude, the rainbow will be low; if he be at a low elevation, the rainbow must be high; and if a shower happen in a vale, when the spectator is on a mountain, he will sometimes see the bow in the form of a _complete circle_ below him. We have at present described the phenomena only of a single drop; but it is to be considered that in a shower of rain there are drops at all heights and at all distances; and therefore the eye situated at G will see all the different colours. All those drops that are in a certain position with respect to the spectator will reflect the red rays, all those in the next station the orange, those in the next the green, and so on with regard to all the other colours. [Illustration: _figure 35._]