atmospheric pressure, so that we need not wonder that delicate persons feel the change, when the pressure on their body is increased or diminished in this manner by a few hundredweights.

These fluctuations in the barometer are by no means of equal extent in different places. Their amount increases greatly with the latitude, being much greater at St. Petersburgh or Stockholm than in London, while in warm latitudes it is greatly diminished, and within the tropics becomes not only so small as to escape the notice of unscientific persons, but assumes a totally different character from that which it has in cold and temperate climates. This difference we will proceed to explain.

49. The fluctuations occurring in extra-tropical latitudes are so fitful as utterly to defy all attempts to reduce them to rule, that is, to any periodical order. Indeed, this must be obvious from their generally acknowledged connection with the wind and weather, the uncertainty of which are proverbial. But in the torrid zone, where this uncertainty, as regards two elements at least, namely, wind and temperature, is replaced by the most clock-like regularity, the barometric changes, although so small as to be scarcely perceptible to an indifferent observer, partake of the same perfect regularity. Unaffected by the utmost extremes of wet or dry, it gives no indications of the approach of the equinoctial deluges, or the solstitial drought; it is thrown aside by the planter as a useless instrument, nevertheless its small daily oscillations go on with the regularity of a clock. Scarcely once in a man's life, at any one spot, does the mercury undergo a decided disturbance, and that not greater than occurs in England before a slight thunder-shower, but such a disturbance is the sure and rapid precursor of one of these stu

ATMOSPHERIC TIDES.

77

pendous atmospheric convulsions, which, in the extent of their disastrous effects, are scarcely exceeded by the greatest earthquakes. The importance of observing the barometer at sea, within the tropics, cannot be over-rated. There are many instances recorded of ships saved from otherwise certain wreck, by the rapid fall of the mercury giving notice of an approaching hurricane. But to return to the ordinary oscillations; it is found that, totally unconnected with changes of weather, these oscillations occur like the tides of the ocean, twice in every twenty-four hours, only with this difference, that they are purely solar tides, or those arising from the action of the sun only, not soli-lunar, or those arising from the combined action of the sun and moon. They have no variations of spring and neap, and instead of their rise and fall occupying half a lunar day*, it occupies only twelve hours, or half an ordinary solar day, the two maxima of pressure always occurring at the same hours, namely, at 9 o'clock, A.M., and the two minima each at about 3 o'clock, P.M. This is, therefore, a tide not produced by gravitation, for if it were, the sun's effect would, as in the oceanic tides, be masked by the greater effect of the moon; but being dependent solely on the sun's position, this atmospheric tide must be due not to his attraction, but to his heat. The manner in which this power acts so as to produce two opposite and equal protuberances of the atmosphere, is beset with difficulties, and cannot yet be said to be even partially understood, but the facts are nevertheless of the greatest interest to science.

50. In speaking of the semi-diurnal oscillations, we must always be understood to refer to what takes place at the level of the sea, because on an eminence their effect is necessarily partly disguised by the super-addi

* The mean length of the lunar day is 24h 50′ 21′′.

tion of that diurnal oscillation deduced above, the period of which being 24 hours, it must have the effect of making these 12-hour tides appear alternately unequal, and this inequality is greater the higher we ascend.

The extent of the semi-diurnal oscillation is found to be greatest at the equator, where it averages more thanth of an inch, and it diminishes to 16ths of an inch in lat. 30°; being, however, in both situations greater in the hottest months than in the coolest. Beyond the limits of the trade winds, these oscillations being still further diminished, are exceeded in extent (and therefore entirely masked) by the irregular fluctuations so familiar in our climate. These fitful variations, unknown within the tropics, increase prodigiously in extent as we recede therefrom; and as the regular oscillations, on the contrary, diminish, they are of course soon lost in, or confounded with, these irregular

ones.

51. But although thus disguised so as to be no longer recognisable by simple observation, this regular tide still flows and ebbs amidst all the irregularities of the pressure in temperate climates. The proof of this is simple and elegant. We have only to examine a barometric register kept at stated hours (at least twice a day) for a long period, to take the average of each set of observations made at the same hour, and by comparing these averages, it is found that the mean height at about 9 o'clock, A.M., is greater, and that at about 3 o'clock, P.M., less than at any other hour. In this climate, and in summer, the mean of a few weeks is sufficient to elicit this fact. In winter, or in higher latitudes, the greater extent of the irregular fluctuations requires a longer series to enable them to neutralize each other. In this way it is found that, however dis

DAILY ATMOSPHERIC TIDE.

79

guised by irregular disturbances, the regular atmospheric tide still obtains throughout the temperate zones, although greatly diminishing in extent from the tropics to the Polar circles, near which it becomes altogether imperceptible. Among the many theories proposed to account for these oscillations, that of Professor Daniell is the most satisfactory. It is too complex to be introduced here, but we may observe that he deduced from it not only the entire disappearance of these tides at a certain latitude far short of the Pole, as above stated, but also that in higher latitudes a much smaller tide ought to appear at exactly the contrary hours, that is, flowing from 9 to 3, and ebbing from 3 to 9, and an examination of the register kept for this purpose in Captain Parry's Second Arctic Expedition, has actually given such a result, although its amount is so very small, only a thousandth of an inch, that considering the great extent of the irregular fluctuations, it is very doubtful whether the observations were contiued long enough to give a fair average.

52. This principle of obtaining the mean of many observations made under circumstances bearing a certain resemblance, and comparing this mean with the mean of many others made without that particular resemblance, is the key-stone of the meteorological arch, the sole method of accumulating that mass of materials by which the infant science of meteorology is to be nourished. It is only by such averages that the diurnal tide, observable only on eminences (50), can be shown to exist. Professor Daniell elicited it from the mean of a very moderate number of observations, even on so small an elevation as Box Hill, in Surrey, and the registers kept at greater heights show it distinctly, provided they be on narrow summits or ridges; but on extensive table lands it is not observable, because the

air resting on them has not time to flow off during the night, nor the surrounding air to overflow them during the day. This alternate flux and reflux, however, produces at the border of all elevated tracts the hill and valley breezes already noticed (43), namely, a breeze from the hills during the night, and towards them during the day, similar in character to the land and sea-breezes, and like them most observable in low latitudes, because the daily and nightly alternations of temperature are there greatest.

53. The observations on Box Hill just noticed, also displayed another oscillation, which, from the time of Newton, had often been sought for in vain, namely, the atmospheric tides, strictly so called, produced by lunar attraction. It is obvious that such tides, however great, could never affect the height of the barometer at the sea-level, because the greater height of air accumulated under the moon would, on the whole, weigh no more than the shorter column elsewhere, because every particle of the former column (and also of the mercury weighing it) has its weight diminished by the moon's attraction opposing that of the earth. But the effect ought to be seen by the upward diminution of pressure being slower under the moon; and, accordingly, the mean of several observations made when the moon was near the meridian, gave a less difference between the pressures at the foot and top of the hill than was obtained from the mean of several others made when the moon was near the horizon. It is to be regretted that similar observations have not been made at greater elevations, as they have an important bearing on the use of the barometer for measuring heights. This subject will be noticed presently (56), but we must first make a few remarks on the use of the barometer as a weather-glass.