a marked change is observed when a second piece of glass is made
to approach the reflecting face, so as to be separated from it only
by a very thin layer of air. The reflection is then found no longer
to be total, part of the light finding its way into the second piece of
glass. Newton concluded from this that the corpuscles are attracted
by the glass even at a certain small measurable distance.
3. New Hypotheses in the Corpuscular Theory.—The preceding explanation of reflection and refraction is open to a very serious objection. If the particles in a beam of light all moved with the same velocity and were acted on by the same forces, they all ought to follow exactly the same path. In order to understand that part of the incident light is reflected and part of it transmitted, Newton imagined that each corpuscle undergoes certain alternating changes; he assumed that in some of its different “phases” it is more apt to be reflected, and in others more apt to be transmitted. The same idea was applied by him to the phenomena presented by very thin layers. He had observed that a gradual increase of the thickness of a layer produces periodic changes in the intensity of the reflected light, and he very ingeniously explained these by his theory. It is clear that the intensity of the transmitted light will be a minimum if the corpuscles that have traversed the front surface of the layer, having reached that surface while in their phase of easy transmission, have passed to the opposite phase the moment they arrive at the back surface. As to the nature of the alternating phases, Newton (Opticks, 3rd ed., 1721, p. 347) expresses himself as follows:—“Nothing more is requisite for putting the Rays of Light into Fits of easy Reflexion and easy Transmission than that they be small Bodies which by their attractive Powers, or some other Force, stir up Vibrations in what they act upon, which Vibrations being swifter than the Rays, overtake them successively, and agitate them so as by turns to increase and decrease their Velocities, and thereby put them into those Fits.”
4. The Corpuscular Theory and the Wave-Theory compared.—Though Newton introduced the notion of periodic changes, which was to play so prominent a part in the later development of the wave-theory, he rejected this theory in the form in which it had been set forth shortly before by Christiaan Huygens in his Traité de la lumière (1690), his chief objections being: (1) that the rectilinear propagation had not been satisfactorily accounted for; (2) that the motions of heavenly bodies show no sign of a resistance due to a medium filling all space; and (3) that Huygens had not sufficiently explained the peculiar properties of the rays produced by the double refraction in Iceland spar. In Newton’s days these objections were of much weight.
Yet his own theory had many weaknesses. It explained the propagation in straight lines, but it could assign no cause for the equality of the speed of propagation of all rays. It adapted itself to a large variety of phenomena, even to that of double refraction (Newton says [ibid.]:—“. . . the unusual Refraction of Iceland Crystal looks very much as if it were perform’d by some kind of attractive virtue lodged in certain Sides both of the Rays, and of the Particles of the Crystal.”), but it could do so only at the price of losing much of its original simplicity.
In the earlier part of the 19th century, the corpuscular theory broke down under the weight of experimental evidence, and it received the final blow when J. B. L. Foucault proved by direct experiment that the velocity of light in water is not greater than that in air, as it should be according to the formula (1), but less than it, as is required by the wave-theory.
5. General Theorems on Rays of Light.—With the aid of suitable assumptions the Newtonian theory can accurately trace the course of a ray of light in any system of isotropic bodies, whether homogeneous or otherwise; the problem being equivalent to that of determining the motion of a material point in a space in which its potential energy is given as a function of the coordinates. The application of the dynamical principles of “least and of varying action” to this latter problem leads to the following important theorems which William Rowan Hamilton made the basis of his exhaustive treatment of systems of rays.[1] The total energy of a corpuscle is supposed to have a given value, so that, since the potential energy is considered as known at every point, the velocity v is so likewise.
(a) The path along which light travels from a point A to a point B is determined by the condition that for this line the integral ∫v ds, in which ds is an element of the line, be a minimum (provided A and B be not too near each other). Therefore, since v = μv0, if v0 is the velocity of light in vacuo and μ the index of refraction, we have for every variation of the path the points A and B remaining fixed,
| δ∫μ ds = 0. | (2) |
(b) Let the point A be kept fixed, but let B undergo an infinitely small displacement BB′ (= q) in a direction making an angle θ with the last element of the ray AB. Then, comparing the new ray AB′ with the original one, it follows that
| δ∫μ ds = μΒq cos θ, | (3) |
where μΒ is the value of μ at the point B.
6. General Considerations on the Propagation of Waves.—“Waves,” i.e. local disturbances of equilibrium travelling onward with a certain speed, can exist in a large variety of systems. In a theory of these phenomena, the state of things at a definite point may in general be defined by a certain directed or vector quantity P,[2] which is zero in the state of equilibrium, and may be called the disturbance (for example, the velocity of the air in the case of sound vibrations, or the displacement of the particles of an elastic body from their positions of equilibrium). The components Px, Py, Pz of the disturbance in the directions of the axes of coordinates are to be considered as functions of the coordinates x, y, z and the time t, determined by a set of partial differential equations, whose form depends on the nature of the problem considered. If the equations are homogeneous and linear, as they always are for sufficiently small disturbances, the following theorems hold.
(a) Values of Px, Py, Pz (expressed in terms of x, y, z, t) which satisfy the equations will do so still after multiplication by a common arbitrary constant.
(b) Two or more solutions of the equations may be combined into a new solution by addition of the values of Px, those of Py, &c., i.e. by compounding the vectors P, such as they are in each of the particular solutions.
In the application to light, the first proposition means that the phenomena of propagation, reflection, refraction, &c., can be produced in the same way with strong as with weak light. The second proposition contains the principle of the “superposition” of different states, on which the explanation of all phenomena of interference is made to depend.
In the simplest cases (monochromatic or homogeneous light) the disturbance is a simple harmonic function of the time (“simple harmonic vibrations”), so that its components can be represented by
The “phases” of these vibrations are determined by the angles nt + ƒ1, &c., or by the times t + ƒ1/n, &c. The “frequency” n is constant throughout the system, while the quantities ƒ1, ƒ2, ƒ3, and perhaps the “amplitudes” a1, a2, a3 change from point to point. It may be shown that the end of a straight line representing the vector P, and drawn from the point considered, in general describes a certain ellipse, which becomes a straight line, if ƒ1 = ƒ2 = ƒ3. In this latter case, to which the larger part of this article will be confined, we can write in vector notation
| P = A cos (nt + ƒ), | (4) |
where A itself is to be regarded as a vector.
We have next to consider the way in which the disturbance changes from point to point. The most important case is that of plane waves with constant amplitude A. Here ƒ is the same at all points of a plane (“wave-front”) of a definite direction, but changes as a linear function as we pass from one such wave-front to the next. The axis of x being drawn at right angles to the wave-fronts, we may write ƒ = ƒ0 − kx, where ƒ0 and k are constants, so that (4) becomes
| P = A cos (nt − kx + ƒ0). | (5) |
This expression has the period 2π/n with respect to the time and the perion 2π/k with respect to x, so that the “time of vibration” and the “wave-length” are given by T = 2π/n, λ = 2π/k. Further, it is easily seen that the phase belonging to certain values of x and t is equal to that which corresponds to x + Δx and t + Δt provided Δx = (n/k) Δt. Therefore the phase, or the disturbance itself, may be said to be propagated in the direction normal to the wave-fronts with a velocity (velocity of the waves) v = n/k, which is connected with the time of vibration and the wave-length by the relation
| λ = vT. | (6) |
