a window, and a piece of wire be held over it, there will be seen two shadows of the wire, one from each of the surfaces; if a second plate be laid upon the former, there will be three shadows, one from the uppermost surface, one from the lowest, and a third between them from the contiguous surfaces which together form but one image. If either of the surfaces in contact be slightly curved, so as to be a portion of a large sphere, then rings of colours observed by Newton are seen, and are found to arise from reflection at the surfaces of contact, as is proved by their being intercepted by the middle shadow of the wire. At the same time may also be seen other rings, occasioned by light transmitted through the surfaces in contact and reflected back to the eye from the undermost surface; but as these rings are not the subject of Mr. Knox's experiments, he painted the under surface of the lower plate black, in order to prevent their interference with his observations. Under these circumstances, a second image of the rings, formed by reflected light, is seen by means of two reflections from the parallel surfaces of the upper plate of glass; and in a favourable light further repetitions of the same rings may be seen by several successive reflections. Under these circumstances, in addition to these appearances, Mr. Knox observed certain fringes to proceed on each side from the points of intersection of a set of primary circles, with their reflected images appearing as parallel lines at right angles to the line joining their centres, and divided into two sets, coloured in opposite directions from a central line. When a piece of glass is laid upon a convex surface, the secondary set is equal to the primary, and in that case the fringes are straight lines; but when by due combination of surfaces the primary and secondary sets differ in size from each other, then these fringes assume a circular form, coloured according to the same law as the straight fringes, being divided into two classes by a middle curve, towards which the violet edges of the curves on each side are turned. To these curves, which have not been observed before, Mr. Knox gives the name of intersectionary rings.
In the same manner as one set of rings is produced by the intersection of primary and secondary reflected circles with each other, they may also be produced by the intersection of other sets, either of transmitted or reflected rings, and may be rendered numerous by a number of surfaces of various forms; or, on the contrary, may be exhibited in their most simple state by the intersection of primary fringes with each other. For if two slips of plane glass be applied to each other at a small angle, the fringes of colour then appear as straight bars of prismatic colours parallel to each other. And if a third plate of glass be placed upon the uppermost, with a slight inclination situated transversely to the former, the bands thus produced are at right angles to the preceding; and by their intersections present a set of intersectionary fringes parallel to each other, and bisecting the angle between the primaries from which they originate. These fringes, as in the former cases, are divided by a central band into two sets, of which the colours are oppositely placed, and on this
