REFRACTION, in astronomy, or Rs FRACTION of the stars, is an inflexion of the rays of those luminaries, in passing through our atmosphere, by which the apparent altitudes of the heavenly bodies are increased. This refraction arises from hence, that the atmosphere is unequally dense in different stages or regions ; rarest of all at the top, and densest of all at the bottom ; which inequality, in the same medium, makes it equivalent to several unequal mediums, by which the course of the ray of light is continually bent into a continued curve line. And Sir Isaac Newton has shown, that a ray of light, in passing ffom the highest and rarest part of the atmosphere, clown to the lowest and densest, undergoes the same quantity of refraction that it would do in passing immediately, at the same obliquity, curt of a vacuum into air of equal density with that in the lowest part of the atmosphere.
Hence arise the phenomena of the cre pusculum or twilight ; and hence also it is, that the moon is sometimes seen eclipsed, when she is really below the horizon, and the sun above it.
That there is a real refraction of the stars, &c. is deduced not only from phy. sical considerations, and from arguments a priori, but also from precise astronomi cal observations ; for there are number less observations, by which it appears that the sun, moon, and stars rise much sooner, and appear higher, than they should do according to astronomical cal culations. Hence it is argued, that as light is propagated in right lines, no rays could reach the eye from a luminary be. low the horizon, unless they were deflect ed out of their course, at their entrance into the atmosphere : and therefore it ap pears that the rays are refracted in pass ing through the atmosphere. Since the stars appear higher by refraction than they really are, to bring the observed or apparent altitudes to the true ones, the quantity of refraction must be subtracted. Accordingly the ancients, as they were not acquainted with this refraction, rec koned their altitudes too great. Refrac tion lengthens the day, and shortens the night, by making the sun appear above the horizon a little before his rising, and a little after his setting. Refraction also makes the moon and stars appear to rise sooner, and set later than they really do. The apparent diameter of the sun or moon is about 32'; the horizontal refraction is about 33' ; whence the sun and moon ap pear wholly above the horizon when they are entirely below it. Also from observa tions it appears, that the refractions are greater nearer the pole than at lesser la titudes, causing the sun to appear some days above the horizon, when he is really below it ; doubtless from the greater den sity of the atmosphere, and the greater obliquity of the incidence.
Stars in the zenith are not subject to any refraction : those in the horizon have the greatest of all : from the horizon the refraction continually decreases to the zenith. All which follows from hence, that in the first case, the rays are perpen dicular to the medium ; in the second, their obliquity is the greatest, and they pass through the largest space of the low er and denser part of the air, and through the thickest vapours; and in the third, the obliquity is continually decreasing. The air is condensed, and consequently refraction is increased, by cold ; for which reason it is greater in cold countries than in hot ones. It is also greater in cold weather than in hot, in the same country ; and the morning refraction is greater than that of the evening, because the air is rarified by the heat of the sun in the day, and condensed by the coldness of the night. Refraction is also subject to some small variation at the same time of the day in the finest weather.
The horizontal refraction, being the greatest, is the cause that the sun and moon appear of an oval form at their rising and setting ; for the lower edge of each being more refracted than the up per edge, the perpendicular diameter is shortened, and the under edge appears more flattened also. Again, if we take with an instrument the distance of two stars when they are in the same vertical and near the horizon, we shall find it considerably less than if we measure it when they are both at. such a height as to suffer little or no refraction ; because the lower star is more elevated than the higher. There is also another alteration made by refraction in the apparent dis tance of stars when two stars are in the same parallel of declination, their appa rent distance is less than the true ; for since refraction makes each of them light er in the azimuth or vertical in which they appear, it must bring them into parts of the vertical where they come nearer to each other ; because all vertical circles converge and meet in the zenith. This contraction of distance, according to Dr. Halley, (Philos. Trans. numb. 368) is at the rate of at least one second in a degree ; so that, if the distance between two stars in a position parallel to the ho rizon measure 3P, it is at most to be reckoned only 29°, 59', 32".