cast opposes radiation from the ground, which becomes heated on the surface: thus the air passes away in greater proportion to the higher regions than if the vapours were not condensed into clouds. Reciprocally, when the winds blow from the east, the sky is serene, and the cooling of the ground very considerable; whence arises an increase of pressure. In the interior of the continent, whither the sea-breezes arrive, loaded with a less quantity of vapour, the sky is generally more pure, the heating of the ground less marked, and the barometer more tranquil.

From this fact, that the decrease of temperature with latitude is more rapid as we recede farther from the equator, a second consequence may be drawn. If we admit that the two winds, which cause the barometer to rise and fall, always bring air coming from countries situated in the circumference of a circle, at the centre of which the observer is placed, we can comprehend that for hot winds blowing from the south the barometer falls less below the mean than it rises by north winds, which are relatively colder. As all these changes oscillate about the mean, the barometer ought to fall more slowly than it rises; observation confirms this anticipation. If we calculate the cases in which the barometer rises, setting out from any hour to the same hour of the next day, this number is to that of the case in which it falls as 10:11; that is to say, that the time required by the barometer to rise a certain quantity is to that during which it falls the same quantity as 10 11. Now, as it rises with north winds, the latter draw along a mass of air proportionate to that which the south winds can draw as 11:10. This fact explains to us a circumstance, in the table of the relations of the winds (p. 48), that is difficult of comprehension. We have found, indeed, that the north winds blow less frequently than the south, in the relation of 10 11,8; if, then, they draw along as much air as those of the north, the atmosphere would finally be carried towards the pole: but we have just seen that the north winds, compared with those of the south, draw along a mass of air greater in about the same proportion, namely, as 10: 11,3.

STATE OF THE BAROMETER DURING RAIN.—

Toricelli had long ago remarked that the barometer was low at the approach of rain; it is admitted as positive that the diminution of pressure ought to bring rain; whilst the weather ought to remain fine, as long as the barometer is high. If this coincidence did not occur, then would there be lamentations without end on the inaccuracy of barometers in general, or of accusations against him who should be par

ticular in observing it. It would be more wise to lament that a prejudice on this point could become rooted in the generality of minds.

The law that presides over all the oscillations of the barometer, which only indicate differences of temperature between countries that are not far apart, likewise finds an explanation here. If the fall of the column generally precedes rain, this is due to the particular position of Europe; in fact, the S.W. winds, which are the warmest, make the barometer fall; they likewise bring rain: hence the observed coincidence. The cold winds of the N.E., on the contrary, raise the barometric column, and almost always accompany a clear and serene sky.

For a long time, philosophers vainly endeavoured to explain the relation by which the two phenomena are connected; Deluc was the first to point it out in general terms, and, although his hypothesis does not induce a searching investigation, it is yet generally adopted. A cubic decimetre of the vapour of water not being so heavy as a cubic decimetre of air, Deluc explains all barometric oscillations by the greater or less proportion of the vapour of water contained in the atmosphere. Indeed, when a certain quantity of air absorbs a certain quantity of vapour, it expands; the atmosphere at this point is higher than in the neighbouring points, a part of the air passes away on all sides, and the pressure of the remaining part is less on account of the proportion of vapours that it contains. This principle being established, he deduces from it a host of consequences, of which the following are the most important :

1st. When the air, loaded with vapours derived from the sea, traverses the continent, the atmospheric pressure diminishes throughout the whole of the course it pursues, and the barometer falls. 2d. If these masses of moist air accumulate in any country, the vapours finally rise into the higher regions of the atmosphere, where they form clouds. The barometer then falls still lower; not because the clouds diminish the weight of the atmosphere, but because the proportion of vapour continues constantly increasing. 3d. The vesicles of the cloud finally collect together, and then the rain falls. 4th. When the sky is clear and the air moist, the barometer falls, if the dew is abundant. 5th. The barometer falls with south and with west winds, because they bring to us moist air; it rises, on the contrary, under the influence of dry east and north winds; it rains also with the former, whilst it is fine weather with the latter. 6th. If the sky is clear with south winds,

or clouded with north winds, the barometer does not indicate it. 7th. Should the arrival of air, loaded with vapours, cease during rain, the latter then draws the vapours toward the earth, the dry air flows in from all sides, the pressure increases, the barometer rises, and we may conclude that the rain will not last long. 8th. Should the barometer begin to rise merely because the wind that is loaded with vapours ceases to blow, the rain may then continue as long as the clouds are sufficiently dense to dissolve into water; but if the wind veers to the N.E., this dry wind dissolves the vapour, and the clouds are instantly dissipated. 9th. When the vapours that are accumulated in any region ascend in the atmosphere, they condense into clouds; a wind may then rise, blowing only in the higher regions of the atmosphere, and driving the clouds towards a country where the barometer is high it will rain there without the mercury falling, because that wind does not arrive loaded with vapours. It therefore rains in this country although the barometer is high, and it does not rain in that where the clouds were formed, although the barometer is low. 10th. As the barometer indicates the state of the entire column of the air, and the hygrometer merely that of the air in the place of observation, the range of the two instruments may

be

very different. 11th. Heat expands the air, and diminishes its weight; it acts much more energetically on vapours. The more the winter mean differs from that of summer, and the more the proportion of the vapour of water differs in the two seasons, the greater also are the barometric oscillations. For if, during summer, the air is hot and at the same time loaded with vapours, the barometer must fall: also in the north, where the difference between the temperature of winter and that of summer is very great, the barometer oscillates greatly, whilst it is almost motionless in the neighbourhood of the equator.

This theory was received with much favour, because it embraced the entire mass of phenomena better than all those that preceded it. However, its author himself so modified it subsequently that he did not fear to maintain that the air was metamorphosed into vapours of water, and even into water itself, under the influence of certain affinities, to pass again into the state of air under different circumstances. The consequences remained in detail the same. But the fundamental idea of Deluc's hypothesis is contrary to the most simple notions in physics and chemistry; for when the elements of the air combine, nitric acid, and not water, is produced. De Saussure, a fellow-countryman and con

temporary of Deluc's, had already shewn that barometric oscillations do not depend solely on vapours; his arguments have been corroborated by all the subsequent labours of philosophers on this subject, and yet Deluc's hypothesis is produced in almost all treatises on natural philosophy and Meteorology, from the end of the last century and the commencement of the present. De Saussure's objections rarely find a place in them. I think that, without being persuaded of the truth of Deluc's assertions, his contemporaries adopted them to avoid the labour of refuting them. Deluc built up a system of Meteorology easy of comprehension and of explanation; de Saussure, on the contrary, only gave meteorological fragments, disseminated throughout his Travels in the Alps, and his Essay on Hygrometry. It was difficult to collect and to arrange them, in order to oppose them to Deluc; they preferred neglecting them.

After having determined the quantity of vapour contained in the air, at different degrees of the thermometer and the hair-hygrometer, de Saussure made known a great number of facts that do not accord with Deluc's theory; for, if vapours acted as he contended, the barometric variations would be enormous. Suppose, for instance, that the dew-point was at 25°, the tension of the vapour would be in equilibrio to a column of mercury 23mm in length; if all this vapour were then precipitated in the state of water, which never happens, the barometer would rise the same amount. But, in our countries, we never observe such difference in the quantity of the vapour of water, whilst the extremes of barometric oscillations far exceed 23mm. Moreover, it is in countries and in the season where the heat is greatest, and the evaporation very active, that we should observe the greatest oscillations, namely, in summer, and in the neighbourhood of the equator; now experience shews precisely the contrary.

Deluc's hypothesis rests on a principle the fallacy of which has been proved by Dalton, Gay-Lussac, and others. At equal tensions a volume of moist air weighs less than an equal volume of dry air; but, when water quietly evaporates in the open air, the vapours ascend through the interstices of the aerial particles, without having any influence by their weight or their elasticity on the movements of the air. The atmospheric pressure is, therefore, increased by the weight of the vapour of water, all other things being equal. The barometer ought to be higher in moist than in dry air. Observation seems contrary to this assertion, since the barometer is lowest with winds loaded with vapours. But

S.W. winds, which bring rain, are also the hottest of all: they tend to raise the barometric column by the pressure of their vapour, and to lower it by their temperature. This last influence being the more energetic, the pressure diminishes; and it is by their temperature that sea-breezes in our climates make the barometer fall. In other countries they act differently: thus Flinders has shewn, in a work on the barometric oscillations on the coasts of New Holland, that beyond the tropics the dry winds, blowing from the shore, make the barometer fall; which is very well explained by Peron's remarks on the high temperature of these winds. At the mouth of La Plata, the barometer is higher during east sea-breezes than with west winds blowing from the land.

In these researches, we should in the outset distinguish the state of the barometer during continued rains from that which accompanies short and isolated showers. If the latter are frequent, and are due to clouds that approach the zenith, we may calculate on the barometer's rising several tenths of a millimetre; this often happens at the approach of storms. Sometimes the barometer falls again to its original height, when the cloud has departed. During storms we may affirm that the period of their greatest violence is passed when the barometer ceases to rise or begins to fall; this is because the rain that falls cools the lower strata of the atmosphere, and that masses of air flow in from all parts toward this spot. It also happens that the barometer rises regularly for several days: in this case the south winds have been driven away by the north winds; and, at the place where they meet, the mixture of strata of air, of different temperatures, produces a condensation of the vapours; the barometer then rises, under the influence of these cold winds: we see this during storms in the winter. If the storm comes from the south, and the barometer falls, it will rise again after the first flashes of lightning.

But generally, during rainy weather, the barometer is about 5 below its mean; a height corresponding to south and S.W. winds. M. de Buch compared the heights of the barometer during rainy weather at Berlin, and obtained the following results