Mr. M. Smith on the Phcenomena of Lunar Eclipses. 171 



the moon's parallax, inasmuch as that the object producing 

 it (viz. the earth) would, if viewed from the moon, appear 

 bounded by a curve, though of less convexity, as the parallax 

 increases ; hence the augmentation could not amount to a mi- 

 nute, although the parallax were increased to any extent: it 

 may therefore, as before observed, be considered a constant 

 quantity, or only affected by the variation of the sun's semi- 

 diameter. 



Let us now consider the reason why the moon when totally 

 eclipsed is on some occasions entirely invisible, while on other 

 occasions she appears considerably illumined with a sort of 

 ruddy light. The cause usually assigned for this variation is 

 the different distances of the moon from the earth : but I can 

 affirm from my own experience that this solution is unsatis- 

 factory, having seen the eclipsed portion of the moon very di- 

 stinctly both when she was near her apogee and her perigee. 

 I apj)rehend that this variation will be found to depend entirely 

 on them oon's -deolination ; and that she will be always very 

 visible when near the equator, and quite invisible when near 

 either of the tropics. To explain the cause of this, let us in- 

 quire what appearance a lunar eclipse would present to a 

 spectator in the moon. This lunarian observer would evi- 

 dently have a total solar eclipse ; and, on directing his view to- 

 wards the place of the sun, he would see nothing but a lu- 

 minous ring nearly two degrees in diameter, but so very slen- 

 der as probably not to exceed a minute in thickness : this ring 

 would be the effect of the earth's atmosphere refracting the 

 rays of the sun ; and its degree of brightness would, I imagine, 

 be less than that of the sun when in the horizon. Now if the 

 moon were vertically over the earth's equator, the two poles 

 would be in the circumference of the earth's disc, and the lu- 

 minous ring would be vastly brighter in those parts conti- 

 guous to the poles, by reason of the great refraction of the 

 earth's atmosphere in the polar regions : if, on the contrary, 

 the moon were in the tropic, one of the poles would be turned 

 away from the moon, and the other would fall considerably 

 within the disc ; so that the luminous ring would not appear 

 particularly bright in any part, and therefore would not il- 

 lumine the moon sufficiently to render her visible at the earth. 

 If this solution be admitted, it follows that the visibility of the 

 moon in eclipses depends on the position of the lunar nodes, 

 the moon being always most visible when the nodes coincide 

 with the ecjuinoctial points, and least visible when they are in 

 the solstitial points. The eclipse of June 10, 181G, was an ex- 

 ample of this, — although the air was very clear, the moon so 

 entirely disappcari'd that when totally eclipsed her place could 



Y 2 not 



