486 COSMOS. 



During the long dispute as to the probability or impro- 

 bability of an atmospheric envelope round the Moon, accurate 

 occultatioii observations have proved that no refraction takes 

 place on the surface of the Moon, and that consequently the 

 assumption made by Schroter 30 of the existence of a lunar 

 atmosphere and a lunar twilight are disproved. " The com- 

 parison of the two values of the Moon's diameter which may 

 be respectively deduced from, direct measurement, or from the 

 length of time that it remains before a fixed star during the 

 occultatioii, teaches us that the light of a fixed star is not 

 perceptibly deflected from its rectilinear course at that moment 

 in which it touches the Moon's edge. If a refraction took 

 place at the edge of the Moon, the second determination of 

 her diameter must give a value smaller by twice the amount 

 of the refraction than the former ; but on the contrary, both 

 determinations correspond so closely in repeated determina- 

 tions, that no appreciable difference has ever been de- 

 tected.'' 31 The ingress of stars, which may be particularly 

 well observed at the dark edge, takes place suddenly, and 



* Schroter, Selenotopographi&chte Fragments, th. i. 1791, 

 p. 668; th. ii. 1802, p. 52. 



31 Bessel, Uebcr eine angenommenc Atmosphare dcs Mondes 

 in Schumacher's Astron. Nachr. No. 263. pp. 416-420. 

 Compare also Beer and Madler, der Monde, 83 and 107, 

 pp. 133 and 153; also Arago, in the Annuaire for 1846, 

 pp. 346-353. The frequently mentioned proof of the existence 

 of an atmosphere round the Moon, derived from the greater or 

 less perceptibility of small superficial configurations and ' 4 the 

 Moon-clouds moving round in the valleys," is the most un- 

 tenable of all, on account of the continually- varying condition 

 (darkening and brightening) of the upper strata of our own 

 atmosphere. Considerations as to the form of one of the 

 Moorfs horns on the occasion of the solar-eclipse on the 5th of 

 September, 1793, induced William Herschelto decide against 

 the assumption of a lunar atmosphere. (Philos. Transact. 

 vol. Ixxsiy. p. 167.) 



