trigonometrical operation, is found to be, at a mean rate, about sixty semidiameters of the earth, or in round numbers, about 240,000."

This is one of those demonstrative supports of the Solar System which we are required to believe, or to be denounced by the oracles of it, as, "the worst of heretics." It has the merit, certainly, of appearing extremely plausible, and so far it suits their system; but that is all that can be said in its favour; in other respects it may, with propriety, be classed with the rest of their inapplicable experiments and fanciful theories.

The refraction of the air, concerning which philosophers are entirely in the dark, as their own writings show,* renders this mathematical theory quite useless: besides which, may be mentioned the difficulty of noting the exact time of the moon's passage through the zenith; the rapid change in her declination; the unavoidable inaccuracy of instruments and time-pieces, used in making observations, and even the liability, in nice observations of this kind, to be deceived by the eye itself. These are obstacles which no human art can surmount. Besides, the moon moves through an angular space equal to what they estimate the whole parallax to be, in less than four minutes of time: it may be further remarked too, that an observer elevated to the short distance, of 969 yards above the level of the sea, would see the centre of the lunar disk until it reached the rational horizon, in which case she would seem to

"It would be endless to notice the different opinions respecting both the terrestrial and the astronomic refractions which are to be met with in the writings of various authors on the subject; and it would be equally useless to notice all the tables of its quantity given by them, some of which differ very much from others," Dr. Rees's New Cylopædia, Article, Refraction.

him to have no parallax. For, let a (Fig. 3,) represent the station of an observer on the earth's surface, at the level of the sea, viewing the moon while setting in the sensible horizon H; if the same observer were elevated to A; it is evident that he would then be enabled to see the centre of the moon until it reached S, the rational horizon. Suppose the moon's distance from the earth to be 239980 miles, the semidiameter of the earth 3985, and BSO, the angle of the moon's parallax, 57′ 5′′; we have then the angle SAO, 89° 2′ 55′′, and the angle SOA 90°. Let a line be drawn from the point B, where SA touches the surface of the earth, to O, the centre, And then the triangles SOA and OAB will be similar.

So that an observer at A, elevated 969 yards above a, the level of the sea, would see the centre of the moon at S, in the rational horizon; and consequently it would appear to his view full six hours and twelve minutes, even without the elevating aid of refraction. But I need not enlarge upon the useless theory in question; the proposer himself was sufficiently aware of its inefficiency for the purpose; for he observes, "The true quantity of the moon's horizontal parallax cannot be ascertained by this method, on account of the varying declination of the moon, and the inconstancy of the horizontal refractions, which are perpetually changing, according to the state of the atmosphere at the time: for the moon continues but for a short time in the equinoc

1

tial, and the refraction, at a mean rate, elevates her apparent place near the horizon half as much as her parallax depresses it."

Here I may be allowed to ask, how do these philosophers know what quantity the parallax depresses the appearance, before that parallax has been discovered? "But," says the proposer of the above-stated theory, "astronomers have thought of" (not practised,)" another method, which is free from these objections; and if practised by able observers, with good instruments, it is sufficient for determining the parallax, and distance of the moon, to a considerable degree of precision. I shall mention the most simple case first, and this will render the general method more clear and satisfactory. Suppose two observers placed under the same meridian at A and B, (Fig. 4,) at such a distance from each other, that the one at A sees the moon M, in his horizon, whilst the other at B, sees her in his zenith; then will the distance of the moon OM, and the horizontal parallax OMA, be easily determined. For the arc AB, which measures the angle O, is equal to the difference of latitude of the two observers; the side OA is equal to 3960 miles, the same as before; and the angle OAM is a right angle, &c. This," he adds, "is the simplest solution the problem admits of; but, as it may not be easy to perceive how the two observers can be placed in the manner required, I shall give you a more 'general method," &c.

In this last solution, the author begins by asserting, that it is free from those objections which the first is liable to; one of which, he very properly remarked, was the horizontal refractions; surely that attaches to this second method, since one of the observers is supposed to see the moon in the horizon; however, it is not necessary to dwell upon this, since it is admitted