3. THIS seems already to explain the reason of the coloured spectrum made by refracting the sun’s light with a prism, though our author proceeds to examine that in particular, and proves that the different coloured rays in that spectrum are in different degrees refracted; by shewing how to place the prism in such a posture, that if all the rays were refracted in the same manner, the spectrum should of necessity be round: whereas in that case if the angle made by the two surfaces of the prism, through which the light passes, that is the angle D F E in fig. 126, be about 63 or 64 degrees, the image instead of being round shall be near five times as long as broad; a difference enough to shew a great inequality in the refractions of the rays, which go to the opposite extremities of the image. To leave no scruple unremoved, our author is very particular in shewing by a great number of experiments, that this inequality of refraction is not casual, and that it does not depend upon any irregularities of the glass; no nor that the rays are in their passage through the prism each split and divided; but on the contrary that every ray of the sun has its own peculiar degree of refraction proper to it, according to which it is more or less refracted in passing through pellucid substances always in the same manner[312]. That the rays are not split and multiplied by the refraction of the prism, the third of the experiments related in our first chapter shews very clearly; for if they were, and the length of the spectrum in the first refraction were thereby occasioned, the breadth should be no less dilated by the cross refraction of the second prism; whereas the breadth is not at all increased, but the image is only thrown into an oblique posture by the upper part of the rays which were at first more refracted than the under part, being again turned farthest out of their course. But the experiment most expressly adapted to prove this regular diversity of refraction is this, which follows[313]. Two boards A B, C D (in fig. 130.) being erected in a darkened room at a proper distance, one of them A B being near the window-shutter E F, a space only being left for the prism G H I to be placed between them; so that the rays entring at the hole M of the window-shutter may after passing through the prism be trajected through a smaller hole K made in the board A B, and passing on from thence go out at another hole L made in the board C D of the same size as the hole K, and small enough to transmit the rays of one colour only at a time; let another prism N O P be placed after the board C D to receive the rays passing through the holes K and L, and after refraction by that prism let those rays fall upon the white surface Q R. Suppose first the violet light to pass through the holes, and to be refracted by the prism N O P to _s_, which if the prism N O P were removed should have passed right onto W. If the prism G H I be