To proceed now to the application-it is difficult not to be struck at first sight with the evident marks of system, which these curves exhibit, from the beginning to the end of the series. Were it possible to obtain, at successive equal intervals of time, the profile of the waves which roll after each other on the surface of the ocean, and were we to reduce these to a scale in like manner, it is not to be doubted that the group would present elevations and depressions indifferently, in all parts of the scale of time; and the intersections of the curves would soon produce confusion in the picture. But it is not so here-the wave occurs too often in the same place; and the intermediate depressions are too regular, for us to adınit, that what is called chance has any considerable share in producing them. The most prominent feature of the piece may be said to be, the nearly constant elevation of the curve at the approach of Full Moon-a very contrary result, certainly, to that found in the year 1798, and sufficient, at first view, to invalidate the partial conclusion I then came to, that the true atmospherical tide consisted, in part, of large depressions at this quarter. These elevations, however, will be found to have their apex, for the most part about two days before the Full, and to be going off at the time of the phasis. That they are properly connected with its approach, may be fairly inferred, from the manner in which the curve No. 2 rises at this time from a great depression, as if prevented from taking an upward tendency by some unusual cause, and become more elastic in consequence of being thus strongly bent downward. If we now turn to the New Moon, on the left of the plate, we perceive its, approach marked, by depressions chiefly in the fore part of the year, and by ele vations in the latter part. Yet the actual time of this position, or rather a day or two after it, exhibits a strong tendency in the diurnal variation to return to the mean of the period: and the same observation applies to both the other quarters; which have also some peculiar opposite variations connected with them. The latter are conspicuous in the elevations which belong to the third quarter of the first three periods, and in the depressions which attach to it in the last three each however with an exception attached : see Nos. 4 and 13. Enough has perhaps been pointed out, to satisfy the reader that in this year, there was a decided connexion between the Lunar positions, and the mean daily movements of the Barometer; which deviated in the same direction about the same point of a Lunar revolution, whether the mean of the season occupied the higher, lower, or middle part of the scale. A certain relation was long since found to obtain between the movements of the Barometer and the va riations of Temperature in the atmosphere: and very early in the course of these enquiries I perceived, on tracing the curve of the diurnal mean temperature on the same scale, and referable to the same mean line, with that of the Barometer, that the connexion was almost constant between them. It is manifested in two different ways, which may be termed conjunction and opposition; since in the one, the curve of the mean temperature accompanies (or precedes or follows by a short interval) that of the Barometer, and in the other the two vary in opposite directions, often with a very near coincidence in time. See Fig. 12, .. p. 235. Two degrees of Fahrenheit are equivalent in these variations to a tenth of an inch in the Barometer. Such are the proportions observed in this figure, the parts of which are copied from some of the many periods I have traced in this way. When the two curves run in opposition through a period, they cross at intervals, and form a succession of rhombs, differing in magnitude according to the extent of variation in either or both of the curves: when the two run in conjunction, the resemblance in the number and extent of the changes is often so close that the one might easily be mistaken for the other. There are also many periods in which both the kinds of relation appear; and some in which neither is very obvious. In Plate 5, the variations of the daily mean temperature through the Solar year 1806-7 are traced in curves, bearing the proportion already mentioned to those of the Barometer, and constructed in other respects on precisely the same plan as in Plate 4. The corresponding numbers on the curves in each Plate will serve to connect those of the Temperature with the Barometrical ones for the same periods. The place of the mean line of each period in the Thermometrical scale, is indicated in the curve at the bottom of the plate. These curves present features in some respects less striking than those of the Barometrical variation; but which, when attentively examined, indicate equally the existence of a system of variations, governed by the Moon's attraction, as a secondary cause, subject on the whole of the year, to the more powerful influence of the Sun as he varies in declination. The greater variations of Temperature, it may be first remarked, appear for the most part during this year in the intervals of the lunar phases: and there is |