been anticipated by Dr. Smith, in a note upon article 164 of his optics, where he acknowledges that the horizontal sun and moon appear at different times, of different magnitudes. In order to remove this difficulty, he supposes, that in such cases, the images formed upon the retina are larger or smaller. He inclines to the opinion, that the largest horizontal moons, happen generally at her perigee in the warmest summer evenings, the barometer being low, and the thermome ter high; and on the contrary, that the smallest horizontal moons, happen generally at her apogee in the coldest winter mornings, the barometer being high, and the thermometer low. According to this account the largest suns and moons, when setting or rising, should be seen through a clear and thin atmosphere in summer, and the smallest through a thick and dense one in winter, as these are known to be the results, if they depend solely upon the refraction of the rays of light. I am entirely convinced, however, from my own uniform experience, that not only is this account erroneous, but that the direct contrary is the fact. At all seasons of the year, I have remarked, that the sun varied in his dimensions at setting, and that too in quick succession, sometimes scarcely appearing to be diminished, and at other times exhibiting his greatest magnitude. As far as my observation has extended, too, I have become convinced that the sun's disk is always larger at setting, when we have reason to believe, that the air is filled with exhalations. Just before a spell of wet weather, it is largest; and before a long succession of dry and clear weather, it is uniformly smallest, the very reverse of what Dr. Smith supposes to be true.

As to the second objection to the theory, that the increased size of the sun and moon at setting, depends upon a view of the interjacent distance and objects, I would proceed to

remark.

If this be the sole cause of their augmented size, how happens it that the sun or moon, when rising or setting, if

viewed through a window, where both the flatted concave of the sky, and the view of interjacent fields are excluded, still appear larger? After having made several observations, all of which went to confirm me in the impression, that this presented a very great difficulty in the system of Dr. Smith, upon recurring to Bishop Berkeley's essay at a new theory of vision, I find that he has in part anticipated the objection I have stated. "With reference to this opinion, I shall only observe," says he, " that if the prospect of interjacent objects, be that which suggests the idea of farther distance, and that idea of farther distance, be the cause that brings into the mind the idea of greater magnitude, it should hence follow, that if one looked at the horizontal moon, from behind a wall, it would appear no bigger than ordinary." The objection, as stated by Bishop Berkeley, is not entirely conclusive; and may be evaded by alleging, that although the wall will cut off the whole prospect of interjacent space, yet one of the signs of distance still remains, the apparent figure of the sky. But how shall we, upon Dr. Smith's principles, account for the fact mentioned above, that even when the rising or setting sun is viewed through a window, or from any situation in which all prospect of the sky above, or earth beneath is excluded, it still exhibits the same augmented size? It will not answer to refer this to the force of habit, for habits cannot lead us to be deluded by fallacies in vision, when the very signs themselves, by which those fallacies are occasioned, are removed. It appears to me, therefore, that so many contradictory phenomena are presented to the theory of Dr. Smith, that we are compelled to resort to other principles in addition to his, in order to solve them all. These principles, I am inclined to think, have been stated and ably defended by Berkeley, although greatly mixed with errors, and tinged as all his writings are, with his fanciful doctrine of immaterialism. In his essay at a new theory of vision, the

most valuable of his productions, he says-" Now in order to explain the reason of the moon's appearing greater than ordinary in the horizon, it must be observed, that the particles which compose our atmosphere, intercept the rays of light proceeding from any object to the eye, and by how much the greater is the portion of atmosphere, interjacent between the object and the eye, by so much the more are the rays intercepted; and by consequence, the appearance of the object rendered more faint, in proportion as it sendeth more or fewer rays into the eye. Now, between the eye and the moon, when situated in the horizon, there lies a far greater quantity of atmosphere, than there does when the moon is in the meridian. Whence it comes to pass, that the appearance of the horizontal moon is fainter, and therefore, it should be bigger in that situation, than in the meridian, or in any other elevation above the horizon." This is the solution of Bishop Berkeley, which receives strong confirmation from many phenomena, and which seems to be indispensable to account for the changes or variations, which are continually taking place in the dimensions of the horizontal sun and moon; and more especially from the appearances exhibited by them, when neither the concave surface of the sky, nor the interjacent planes are perceived. Dr. Smith has undertaken to refute these doctrines of Bishop Berkeley, but I think without success. The arguments with which he assaults them, we shall now state, and endeavour to show that they are insufficient. First, he maintains, "experience shows that various degrees of the moon's faintness, make no sensible variations of her apparent magnitude, because the moon appears much fainter in the day-time than in the night, and therefore, according to our author's principle, should appear larger by day than by night, at the same height, which I could never perceive, though I have often viewed the moon for this purpose." This is by no means a just inference from Bishop Berkeley's doctrine. It is not every

kind of faintness in appearance, which suggests to the mind great distance and magnitude. It is only that kind of faint ness, which has been usually found indicative, or significative of great distance and magnitude. When, for instance, we view from a distance, mountains, forests, or any large objects, the rays of light being intercepted in their passage to us, give to them that dim and faint appearance, which suggests the idea of great remoteness and dimensions. But this is not the case in all kinds of faintness. The moon, when seen in the day-time, when her light fades before the sun, sheds but a dim light indeed, but it is not that kind of dimness, which we have been accustomed to contemplate as the sign of distance. The same answer may be given to many other appearances, that seem to contradict this system. As for example, if it be asked why does not the sun, when beheld through a darkened glass, or on one of our Indian summer-days, when he is enveloped in smoke and darkened, appear much larger at such times, as he is more faint, and may be contemplated on such occasions with the naked eye? The answer is, that these are not those faint appearances, which are significative of great distances and magnitudes, and therefore, no more convey such conceptions to the mind, than mere arbitrary sounds, which have not been agreed upon by mankind, as signs of ideas, could awake those perceptions in them.

The two next objections of Dr. Smith, admit of a similar explanation. 2dly. "I observe," says he, "that the horizontal moon, being much fainter than the horizontal sun, viewed by the naked eye, should, in consequence of that principle, appear much larger than the horizontal sun." This consequence does not rest upon sufficient ground, because greater degrees of faintness, are not always so good signs of distance as less degrees. A man seen through the dim lights of a city, does not appear larger than he is, but seen through a thick fog he does.

3dly. "I observe," continues Dr. Smith, "that the moon when totally eclipsed, appears much fainter than she does at the same elevation, when not eclipsed; but does not appear larger than usual." We have already furnished the answer to this objection.

Upon the whole, we are inclined to think, that the increased dimensions in the disk of the horizontal moon and sun, is to be accounted for, only by having recourse to the causes assigned both by Dr. Smith and Bishop Berkeley. Neither of the causes separately serves to explain all the phenomena, and it appears indispensably necessary from the nature of the case, that they both should conspire in producing the results. The sun and moon, we are sure, move apparently in that concavity of the sky above us which is flatted, and the parts of which, that rest upon the limb of the horizon, evidently seem more remote than those which are above our heads. While their real dimensions, therefore, remain the same, they are perceived at one time to be more remote than at another. In such case, from the known laws that influence the fallacies of vision, they must appear larger when resting upon the horizon. But this cause does not serve to explain all the phenomena, and the principle of Bishop Berkeley, must undoubtedly have its influence also in producing those effects. We know that when objects are seen at a distance, the rays of light proceeding from them are dispersed on their passage, and many of them never reach the eye. This gives to such objects a peculiar appearance of dimness and faintness. This appearance, therefore, becomes to us the sign of distance and magnitude. Now the sun and moon are large objects, which as they approach the horizon have to shoot their rays, before they reach our eyes, through a larger tract of atmosphere, and atmosphere more dense from its immediate vicinity to the earth. Under these circumstances, we know that these objects must always ap