232
HISTORY OF PHYSICAL ASTRONOMY.
the sun, the rays of light proceeding from those parts of the sun whichare apparently contiguous to the lunar disk, will necessarily graze themoon’s surface, and if a lunar atmosphere exists, they ought, by the in-flexion which they suffer in passing through it, to distort the visible por-tion of the solar disk, and to cause the phase of the eclipse correspondingto any assigned instant to be different from what it would be according tothe theories of the solar and lunar motions. From observations of a solareclipse which took place in 1748, Euler concluded that the horizontal re-fraction of the lunar atmosphere amounted to 20" *. His researches,however, were incomplete, inasmuch as he took no account of the effect ofirradiation, which, by causing an apparent enlargement of the solar disk,might be expected to exercise a perceptible infl uence on the phase of theeclipse. The latter is, in fact, modified by two distinct causes, which areso interwoven in their operation, that it is impossible to investigate the ef-fects of the one without taking simultaneous cognizance of those of theother. In order to ascertain whether the phenomena of solar eclipseswere capable of affording any indications of a lunar atmosphere, Du Se-jour, a French astronomer of the last century, undertook a rigorous inves-tigation of the subject, founded on observations of the solar eclipse of1764. He introduced the effects of irradiation and inflexion into his re-searches in the shape of two unknown quantities, and then, by means ofequations of condition, which it was necessary that they should satisfy, hedetermined the values of them which accorded best with the totality of theobservations. The conclusion at which he finally arrived was, that theeffect of irradiation amounted to 3", and that the inflexion which the solarrays suffered in passing through the lunar atmosphere amounted to an equalquantity. This gives l-"5 for the mean horizontal refraction of the lunaratmosphere. The mean refraction of the terrestrial atmosphere at thehorizon amounts to 34', a quantity which is 1400 times greater than l"-5.It hence follows that the lunar atmosphere is 1400 times rarer than com-mon atmospheric air, and consequently it exceeds in this respect the mostperfect vacuum which has been hitherto formed by means of the air-pump.It appears, therefore, from the phenomena of solar eclipses, that, if a lunaratmosphere really exists, it is quite inconsiderable when compared with theterrestrial atmosphere.
Auzout first remarked that the moon could not be surrounded by anatmosphere without possessing also a twilight f. This effect ought tomanifest itself in a faint light, extending to a short distance within theunenlightened part of the lunar disk, and thereby rendering a narrow stripof it visible, although not directly illuminated by the solar rays. Astro-nomers having failed to discover any indications of such a phenomenon, itwas concluded, that the moon could not be encompassed by a fluid capableof dispersing and reflecting the sun’s rays. The appearance, however, wasat length recognised by Schroeter, who was enabled«to deduce from it theexistence of an atmosphere of a small refractive power *. When the moonexhibited a very slender crescent, he discovered a faint crepuscular light, ex-tending from each of the cusps along the circumference of the unenlightenedpart of the disk. Its length was 1' 20", and its greatest breadth, 2". Bymeans of these data he found that the height of the lunar atmosphere, con-sidered in so far as it was capable of affecting the brightness of a star, or caus-
* Mem. Acad. Berlin, 1748.
f Mem. Acad, des Sciences, tome vii., p. 106.
f Phil. Trans., 1792.