becomes inappreciable again. The effect is the same as a perfectly elastic ball bouncing on a perfectly smooth surface, and consequently the angle to the surface at which the corpuscle comes up is equal to the angle at which it departs.

Refraction.—To explain refraction, it is supposed that when the corpuscle comes very close to the surface of the transparent substance it is attracted by the denser substance, e.g. glass, more than by the lighter substance, e.g. air. Thus a particle moving along the dotted line in air (Fig. 4) would reach the point A before the attraction becomes appreciable, and therefore would be moving in a straight line. Between A and B the attraction of the glass will be felt and will therefore pull the particle round in the path indicated. Beyond B, the attraction again becomes inappreciable, because the glass will attract the particle equally in all directions, and therefore the path will again become a straight line. We notice that by this process the direction of the path has become more nearly normal to the surface, and this is as it should be. Further, by treating the angles between the two paths and the normal mathematically we may deduce the laws of refraction which have been obtained experimentally. One other important point should be noticed. Since the surface has been attracting the particle between A and B the speed of the particle will be greater in the glass than in the air.

FIG. 4

Ejection and Refraction at the same Surface.—A difficulty very soon arises from the fact that at nearly all transparent surfaces some light is reflected and some refracted. How can the same surface sometimes repel and sometimes attract a corpuscle? Newton surmounted this difficulty by attributing a polarity to each particle, so that one end was repelled and the other attracted by the reflecting and refracting surface. Thus, whether a particle was reflected or refracted depended simply upon which end happened to be foremost at the time. By attributing suitable characteristics to the corpuscles, Newton with his superhuman ingenuity was able to account for all the known facts, and as the corpuscles were so small that direct observation was impossible, and as Newton's authority was so great, there was no one to say him nay.

Wave Theory. Rectilinear Propagation.—True, Huyghens in 1678 had propounded the theory that light consists of waves of some sort starting out from the luminous body, and he had shown how readily it expressed a number of the observed facts; but light travels in straight lines, or appears to do so, and waves bend round corners and no one at that time was able to explain the discrepancy. Thus for nearly a century the theory which was to be universally accepted remained lifeless and discredited. The answer of the wave theory to the objection now is, that light does bend round corners though only slightly and that the smallness of the bend is quite simply due to the extreme shortness of the light waves. The longer waves are, the more they bend round corners. This can be noticed in any harbour with a tortuous entrance, for the small choppy waves are practically all cut off whereas a considerable amount of the long swell manages to get into the harbour.

Interference of Light. Illustration by Ripples.—The revival of the wave theory dates from the discovery by Dr. Young of the phenomenon of interference of light. In order to understand this we will consider the same effect in the ripples on the surface of mercury. A tuning-fork, T (Fig. 5), has two small styles, S S, placed a little distance apart and dipping into the mercury contained in a large shallow trough. When the tuning-fork is set into vibration, the two styles will move up and down in the mercury at exactly the same time and each will start a system of ripples exactly similar to the other. At any instant each system will be a series of concentric circles with its centre at the style, and the crests of the ripples will be at equal distance from each other with the troughs half-way between the crests.

FIG. 5.

The ripples from one style will cross those from the other, and a curious pattern, something like that in Fig. 6, will be formed on the mercury. S S represents the position of the two styles, while the plain circles denote the positions of the crests and the dotted circles the positions of the troughs at any instant. Where two plain circles cross it is evident that both systems of ripples are producing a crest, and so the two produce an exaggerated crest. Similarly where two dotted circles cross an exaggerated trough is produced. Thus in the shaded portions of the diagram we get more violent ripples than those due to a single style. Where a plain circle cuts a dotted one, however, one system of ripples produces a crest and the other a trough, and between them the mercury is neither depressed below nor raised above its normal level. At these points, therefore, the effect of one series of ripples is just neutralised by the effect of the other and no ripples are produced at all. This occurs in the unshaded regions of the diagram.

The mutual destruction of the effects of the two sets of waves is "Interference."

FIG. 6.

Now imagine a row of little floats placed along the line EDCBABCDE. At the lettered points the floats will be violently agitated, but at the points midway between the letters they will be unmoved. This exactly represents the effect of two interfering sources of light S, S, sending light which is received by a screen at the dotted line EDCBABCDE. The lettered points will be brightly illuminated while the intermediate points will be dark.

In practice it is found impossible to make two sources of light whose vibrations start at exactly the same time and are exactly similar, but this difficulty is surmounted by using one source of light and splitting the waves from it into two portions which interfere.

Young's Experiment.—Dr. Young's arrangement is diagrammatically represented in Fig. 7.

Light of a certain wave length is admitted at a narrow slit S, and is intercepted by a screen in which there are two narrow slits A and B parallel to the first one.

FIG. 7.

A screen receives the light emerging from the two slits. If the old corpuscular theory were true there would be two bright bands of light, the one at P and the other at Q, but instead Dr. Young observed a whole series of parallel bright bands with dark spaces in between them. Evidently the two small fractions of the original waves which pass through A and B spread out from A and B and interfere just as if they were independent sources like the two styles in the mercury ripples experiment.

Speed of Light in Rare and Dense Media.—The discovery of interference again brought the wave theory into prominence, and in 1850 the death-blow was given to the corpuscular theory by Foucault, who showed that light travels more slowly