288

LUNAR CELESTIAL PHENOMENA.

circumstances of pressure, yet even then we shouldhave only the same density at the moon’s level as atthe earth’s. That density could only be due to thepressure of the superincumbent parts of the atmo-sphere. Diminishing with height above the moon’smean surface, according to the laws of gaseous pres-sure, it would extend as high above the moon’s surfaceas our air above the earth’s, even on the suppositionof its having so remarkable a specific gravity comparedwith that of common air.

We see, then, that if we were to suppose the atmo-spheric pressure at the moon’s surface equal to that atthe earth’s, we should have to suppose either that thisatmosphere is composed of gases of very great specificgravity, or else that it extends to a much greaterheight than our own atmosphere. In either case, it isobvious that we should expect to find very markedeffects produced by such an atmosphere.

In the first place, when the moon was carried byher motion over a star, the place of the star would beaffected by refraction, not only when the moon’s edgewas very close to the star, but for some considerabletime before. If the lunar atmosphere were actuallyas dense near the moon’s mean surface as our air is atthe sea-level, then a star would not be occulted at all,even though the moon passed so directly over thestar’s true place on the heavens that the geometricalline joining the star and the observer’s eye passedthrough the moon’s centre. This is easily seen. For the