REFRACTION, REFRANGIBILITY. Refraction is the turning of a ray of light, heat, or other imponderable substance from its direction, when it falls obliquely on the surface of a medium differing in density from that through which it had previously moved. The differently coloured rays of light have different degrees of refrangibility, as evi denced by the common prismatic spectrum; in other words, the re fractive indices of different lights vary for a given medium. The fundamental law of refraction and the optical effects of this law are dismissed under the heads LIGHT and therms, and a table of refractive indices is given in the article OPTICS, PRACTICAL. For the different refraneibilities of the rays, thu articles DISPERSION and ACHROMATIC may be consulted, and for one of the most striking phenomena thence arising see RAINBOW. On the subjects of Fraunhofer's lines and of double refraction, see DISPERSION and UNDULATORY THEORY.
The doctrine of refraction, as distinguished from reflection, is called dioptrice, and the caustics formed by the continued intersections of refracted rays emanating from a luminous point, are termed dia.
caustics ; properly speaking, these are surfaces, but by confining the investigation to the plane of refraction, they are generally treated as curves. A diacaustio curve, like a catacaustic, has the property of being rectifiable. They are noticed in the articles above quoted, but they are rather objects of analytical dexterity than of practical use.
So long as the medium into which the refracted ray enters remains of uniform density, the ray will pursue a straight course, but every alteration of density in the medium gives rise to a corresponding deviation in the path of the ray. Now the air is a medium of which the density continually increases as its altitude above the surface of the earth diminishes ; its density is also altered in the same stratum by inequality of temperature, and frequently from the aqueous and other vapours which it holds. Hence arise the ordinary terrestrial refraction and the phenomena of Mirage, Fate .Morgana, arc., which are treated under their respective heads.
A ray of light proceeding from a star which is not vertical, on enter• ing the atmosphere is bent towards the radius drawn from the eartha centre to its point of incidence, and upon its successive incidences on the lower strata it continues to bend towards the successive radii, thus describing a curvilinear trajectory through the air. The star is visible in the direction of the tangent to this curve, at the point at which it meets the eye of the spectator; hence the apparent altitude of the stars is increased by refraction, and thus the sun, moon, &c., are visible before the real time of rising and after that of setting.
From the cause. above assigned for atmospheric refraction, It follows let the nearer the direction of the ray is to the plane of the horizon, ;he greater is its refraction, and the retraction Is nothing when tho ray Is vertical. This is the cause of the apparently oval forms of the sun met moon in the horizon ; for the eon's angular dlsmeter being taken it 32 minutes, its lower limb is elevated through horizontal refraction more than its upper by 4 minutes 54 seconds.
Atmospheric refraction of the solar rays after sunset, combined with subsequent reflection, is the cause of twilight, and also of the light thrown on the moon's surface when eclipsed by the earth.
The amount of refraction of rays proceeding from a celestial body would be proportional to the tangent of its zenith distance, if the atmosphere were homogeneous, and the zeuith distance sufficiently small to allow us to regard the refracting strata of air as hounded by parallel surfaces. Not only, however, is the atmosphere defective in homogeneity, but its state is continually altering, M shown by the barometer, thermometer, and hygrometer. A partly empiric formula, the result of numerous observations made by Bradley, gives a good correction relative to the first two of these instruments, namely :— Similar formulas' have been the objects of analytical research to Laplace and other modern mathematicians; their results are however not well adapted for insertion in this work.