VOL. XLI.] PHILOSOPHICAL TRANSACTIONS. 373 



equal to 1 6' 4 1" when most open, and to l6'5"when narrowest. Also, the 

 least semi-angle of the cone being supposed l6' 5", its axis will be less than 

 the least distance of the moon from the earth, or 54 terrestrial diameters ; and 

 therefore the point of the lunar shade will not extend to the earth. So that, if 

 there were an atmosphere about the moon, in which the horizontal refraction is 

 8", there would be no total eclipse on the earth. There is, therefore, either 

 no atmosphere about the moon, or else it produces a horizontal refraction less 

 than 8*. 



But certain total eclipses of the sun are observed with some duration of the total 

 darkness. For instance, in the eclipse of 1724, the duration of total darkness 

 amounted to "i*" l6^ The moon, at that time, ran over l'l5''in her horary motion, 

 and her shadow, with a parallel motion, a space 54 times greater on the earth's disk, 

 that is, equal to 1° ^' 30" ; from which, if there be deducted the diurnal motion of a 

 habitation, by which the duration of the eclipse can be prolonged, it gives the dia- 

 meter of the shadow equal to 47'30", or 45 1 73 toises, or 11 Paris leagues. Hence, 

 by calculation, it is found, that the axis of the cone of the lunar shade is greater 

 by at least one diameter of the earth, than the moon's distance from it, which 

 was then the least, the moon being then about the perigeum. Further, from 

 the given diameters of the luminaries, observed in the same degree of anomaly, 

 the axis of the cone of the lunar shade is found at least equal to 55 semidi- 

 ameters : hence it follows, that the spot of the lunar shade on the earth's disk, 

 and the axis of the cone, are found to be exactly the same, as the distances of 

 the moon and the observed diameters of the luminaries seem to require. There 

 is therefore no atmosphere about the moon, or, if there is any, it produces no 

 sensible refraction. 



But to leave no room for doubt, we may consider the reason for those phae- 

 nomena, observed in solar eclipses, which have given cause to suspect a lunar 

 atmosphere. 



And first, that very faint light, observed in total eclipses, is no proof of any 

 refraction in a fluid about the moon ; for, by Maraldi's experiments, success- 

 fully repeated by M. Fouchy, it appears that the shadows of bodies that have 

 no atmosphere, when they are exposed to the sun, are bright about the axis of 

 the cone; and the more so as it is farther from the body itself. And the situa- 

 tion of an observer, in a total eclipse, is about the axis of the cone of the lunar 

 shade, and near its vertex. It is no wonder therefore, that the middle of the 

 shadow exhibits a kind of gloomy light, which may also be augmented by the 

 rays being reflected by an illuminated air surrounding the shadow about the 

 middle. 



2dly. The lucid annulus about the moon, in total eclipses, by no means 



