in zoology. See MUSTELA.

WEATHER denotes the state of the atmosphere with regard to heat and cold, wind, rain, and other morsels.

The phenomena of the weather must have at all times attracted much of the attention of mankind, because their subsistence and their comfort in a great measure depended upon them. It was not till the seventeenth century, however, that any considerable progress was made in investigating the laws of meteorology. How desirous forever the ancients might have been to acquire an accurate knowledge of this science, their want of proper instruments entirely precluded them from cultivating it. By the discovery of the barometer and thermometer in the last century, and the invention of accurate electrometers and hygrometers in the present, present, this defect is now pretty well supplied; and philosophers are enabled to make meteorological observations with ease and accuracy. Accordingly a very great number of such observations have been collected, which have been arranged and examined from time to time by ingenious men, and consequences deduced from them, on which several different theories of the weather have been built. But meteorology is a science so exceedingly difficult, that notwithstanding the united exertions of some of the first philosophers of the age, the phenomena of the weather are still very far from being completely understood; nor can we expect to see the veil removed, till accurate tables of observations have been obtained from every part of the world, till the atmosphere has been more completely analysed, and the chemical changes which take place in it ascertained. From the meteorological facts, however, which are already known, we shall draw up the best account of the weather we can. We shall treat of the different phenomena in the following order—heat and cold, wind, rain, thunder, alterations in the gravity of the atmosphere.

I. Though there is a considerable difference in every part of the world between the temperature of the atmosphere in summer and in winter; though in the same season the temperature of almost every day, and even every hour, differs from that which precedes and follows it; though the heat varies continually in the most irregular and seemingly capricious manner—still there is a certain mean temperature in every climate, which the atmosphere has always a tendency to observe, and which it neither exceeds nor comes short of beyond a certain number of degrees. What this temperature is, may be known by taking the mean of tables of observations kept for a number of years; and our knowledge of it must be the more accurate the greater the number of observations is.

The mean annual temperature is greatest at the equator (or at least a degree or two on the north side of it), and it diminishes gradually towards the poles, where it is least. This diminution takes place in arithmetical progression, or, to speak more properly, the annual temperature of all the latitudes are arithmetical means between the mean annual temperature of the equator and the pole. This was first discovered by Mr Mayer; and by means of an equation which he founded on it, but rendered considerably plainer and simpler, Mr Kirwan has calculated the mean annual temperature of every degree of latitude between the equator and the pole. He proceeded on the following principle. Let the mean annual heat at the equator be \( m \) and at the pole \( n \); put \( \phi \) for any other latitude; then the mean annual temperature of that latitude will be \( m - n \times \sin \phi^2 \). If therefore the temperature of any two latitudes be known, the value of \( m \) and \( n \) may be found. Now the temperature of north lat. 40° has been found by the best observations to be 62.1°, and that of lat. 50°, 52.9°. The square of the sine of 40° is nearly 0.419, and the square of the sine of 50° is nearly 0.586. Therefore

\[ m - 0.41n = 62.1 \quad \text{and} \\ m - 0.58n = 52.9 \]

therefore

\[ 62.1 + 0.41n = 52.9 + 0.58n \]

as each of these, from the two first equations, is equal to \( m \). From this last equation the value of \( n \) is found to be 53 nearly; and \( m \) is nearly equal to 84. The mean temperature of the equator therefore is 84°, and that of the pole 31°. To find the mean temperature for every other latitude, we have only to find 88 arithmetical means between 84 and 31. In this manner Mr Kirwan calculated the following table.

This table, however, only answers for the temperature of the atmosphere of the ocean. It was calculated for that part of the Atlantic ocean which lies between the 80th degree of northern and the 45th of southern latitude, and extends westwards as far as the Gulf-stream, and to within a few leagues of the coast of America; and for all that part of the Pacific ocean reaching from lat. 45° north to lat. 40° south, from the 20th to the 275th degree of longitude east of London. This part of the ocean Mr Kirwan calls the landward; the rest of the ocean is subject to anomalies which will be afterwards mentioned.

Mr Kirwan has also calculated the mean monthly temperature of the standard ocean. The principles on which he went were these: The mean temperature of April seems to approach very nearly to the mean annual temperature; and as far as heat depends on the action of the solar rays, the mean heat of every month is as the mean altitude of the sun, or rather as the sine of the sun's altitude. The mean heat of April, therefore, and the fine of the sun's altitude being given, the mean heat of May is found in this manner:

As the fine of the sun's mean altitude in April is to the mean heat of April, so is the fine of the sun's mean altitude in May to the mean heat of May. In the same manner the mean heats of June, July, and August, are found; but the rule would give the temperature of the succeeding months too low, because it does not take into the heat derived from the earth, which possesses a degree of heat nearly equal to the mean annual temperature. The real temperature of the Weather. these months therefore must be looked upon as an arithmetical mean between the astronomical and terrestrial heats. Thus in latitude $51^\circ$, the astronomical heat of the month of September is $44.6^\circ$, and the mean annual heat is $52.4^\circ$; therefore the real heat of this month should be $\frac{44.6 + 52.4}{2} = 48.5$. Mr Kirwan, however, after going through a tedious calculation, found the results to agree so ill with observations, that he drew up the following table partly from principles and partly by studying a variety of sea journals.

### Table of the Monthly Mean Temperature of the Standard from lat. $80^\circ$ to lat. $10^\circ$.

| Lat. | Jan. | Feb. | Mar. | Apr. | May | June | July | Aug. | Sept. | Oct. | Nov. | Dec. | |------|------|------|------|------|-----|------|------|------|-------|------|------|------| | $60^\circ$ | $22.5$ | $23.$ | $23.5$ | $24.$ | $24.5$ | $25.$ | $25.5$ | $26.$ | $26.5$ | $27.$ | $27.5$ | $28.$ | | $61^\circ$ | $29.$ | $30.$ | $30.5$ | $31.$ | $31.5$ | $32.$ | $32.5$ | $33.$ | $33.5$ | $34.$ | $34.5$ | $35.$ | | $62^\circ$ | $36.$ | $37.$ | $37.5$ | $38.$ | $38.5$ | $39.$ | $39.5$ | $40.$ | $40.5$ | $41.$ | $41.5$ | $42.$ | | $63^\circ$ | $48.$ | $49.$ | $50.$ | $50.5$ | $51.$ | $51.5$ | $52.$ | $52.5$ | $53.$ | $53.5$ | $54.$ | $54.5$ | | $64^\circ$ | $55.$ | $56.$ | $56.5$ | $57.$ | $57.5$ | $58.$ | $58.5$ | $59.$ | $59.5$ | $60.$ | $60.5$ | $61.$ | | $65^\circ$ | $62.$ | $63.$ | $63.5$ | $64.$ | $64.5$ | $65.$ | $65.5$ | $66.$ | $66.5$ | $67.$ | $67.5$ | $68.$ | | $66^\circ$ | $69.$ | $70.$ | $70.5$ | $71.$ | $71.5$ | $72.$ | $72.5$ | $73.$ | $73.5$ | $74.$ | $74.5$ | $75.$ | | $67^\circ$ | $76.$ | $76.5$ | $77.$ | $77.5$ | $78.$ | $78.5$ | $79.$ | $79.5$ | $80.$ | $80.5$ | $81.$ | $81.5$ | | $68^\circ$ | $82.$ | $82.5$ | $83.$ | $83.5$ | $84.$ | $84.5$ | $85.$ | $85.5$ | $86.$ | $86.5$ | $87.$ | $87.5$ | | $69^\circ$ | $88.$ | $88.5$ | $89.$ | $89.5$ | $90.$ | $90.5$ | $91.$ | $91.5$ | $92.$ | $92.5$ | $93.$ | $93.5$ | | $70^\circ$ | $94.$ | $94.5$ | $95.$ | $95.5$ | $96.$ | $96.5$ | $97.$ | $97.5$ | $98.$ | $98.5$ | $99.$ | $99.5$ | | $71^\circ$ | $100.$ | $100.5$ | $101.$ | $101.5$ | $102.$ | $102.5$ | $103.$ | $103.5$ | $104.$ | $104.5$ | $105.$ | $105.5$ | | $72^\circ$ | $106.$ | $106.5$ | $107.$ | $107.5$ | $108.$ | $108.5$ | $109.$ | $109.5$ | $110.$ | $110.5$ | $111.$ | $111.5$ | | $73^\circ$ | $112.$ | $112.5$ | $113.$ | $113.5$ | $114.$ | $114.5$ | $115.$ | $115.5$ | $116.$ | $116.5$ | $117.$ | $117.5$ | | $74^\circ$ | $118.$ | $118.5$ | $119.$ | $119.5$ | $120.$ | $120.5$ | $121.$ | $121.5$ | $122.$ | $122.5$ | $123.$ | $123.5$ | | $75^\circ$ | $124.$ | $124.5$ | $125.$ | $125.5$ | $126.$ | $126.5$ | $127.$ | $127.5$ | $128.$ | $128.5$ | $129.$ | $129.5$ | | $76^\circ$ | $130.$ | $130.5$ | $131.$ | $131.5$ | $132.$ | $132.5$ | $133.$ | $133.5$ | $134.$ | $134.5$ | $135.$ | $135.5$ | | $77^\circ$ | $136.$ | $136.5$ | $137.$ | $137.5$ | $138.$ | $138.5$ | $139.$ | $139.5$ | $140.$ | $140.5$ | $141.$ | $141.5$ | | $78^\circ$ | $142.$ | $142.5$ | $143.$ | $143.5$ | $144.$ | $144.5$ | $145.$ | $145.5$ | $146.$ | $146.5$ | $147.$ | $147.5$ | | $79^\circ$ | $148.$ | $148.5$ | $149.$ | $149.5$ | $150.$ | $150.5$ | $151.$ | $151.5$ | $152.$ | $152.5$ | $153.$ | $153.5$ | | $80^\circ$ | $154.$ | $154.5$ | $155.$ | $155.5$ | $156.$ | $156.5$ | $157.$ | $157.5$ | $158.$ | $158.5$ | $159.$ | $159.5$ |

Lat. From this table it appears, that January is the coldest month in every latitude, and that July is the warmest month in all latitudes above 48°. In lower latitudes, August is generally warmest. The difference between the hottest and coldest months increases in proportion to the distance from the equator. Every habitable latitude enjoys a mean heat of 60° for at least two months; this heat seems necessary for the production of corn. Within ten degrees of the poles the temperatures differ very little, neither do they differ much within ten degrees of the equator; the temperature of different years differs very little near the equator, but they differ more and more as the latitudes approach the poles.

The temperature of the earth at the level of the sea is the same with that of the standard ocean; but this temperature gradually diminishes as we ascend above that level till, at a certain height, we arrive at the region of perpetual congelation. This region varies in height according to the latitude of the place; it is highest at the equator, and decreases gradually nearer the earth as we approach the poles. It varies also according to the season, being highest in summer and lowest in winter. M. Bouguer found the cold on the top of Pinchinca, one of the Andes, to extend from seven to nine degrees below the freezing point every morning immediately before sunrise. He concluded, therefore, that the mean height of the term of congelation (the place where it first freezes during some part of the day all the year round) between the tropics was 15,577 feet above the level of the sea; but in lat. 28° he placed it in summer at the height of 13,440 feet. Now, if we take the difference between the temperature of the equator and the freezing point, it is evident that it will bear the same proportion to the term of congelation at the equator that the difference between the mean temperature of any other degree of latitude and the freezing point bears to the term of congelation in that latitude. Thus the mean heat of the equator being 84°, the difference between it and 32 is 52; the mean heat of lat. 28° is 72.3, the difference between which and 32 is 40.3. Then \( \frac{52}{72.3} : \frac{40.3}{32} = 1 : 20.72 \).

In this manner Mr Kirwan calculated the following table:

| Lat. | Mean height of the Term of Congelation | |------|----------------------------------------| | 0 | 15577 | | 5 | 15457 | | 10 | 15007 | | 15 | 14398 | | 20 | 13719 | | 25 | 13030 | | 30 | 11592 | | 35 | 10664 | | 40 | 9016 |

If the elevation of a country above the level of the sea proceeds at a greater rate than six feet per mile, we must, according to Mr Kirwan *, for every 200 feet of elevation, diminish the annual temperature of the standard in that latitude as follows. If the elevation be at the rate of 6 feet per mile, \( \frac{1}{4} \) of a degree; 7 feet \( \frac{1}{4} \); 13 feet \( \frac{1}{2} \); 15 or upwards \( \frac{1}{2} \).

According to him † also, for every 50 miles distance from the standard ocean, the mean annual temperature in different latitudes is to be depressed or raised nearly at the following rate:

- From lat. 7° to lat. 35° cooled \( \frac{3}{4} \) of a degree. - 35° - \( \frac{3}{4} \) - 30° - 0 - 25° warmed \( \frac{1}{2} \) - 20° - \( \frac{1}{2} \) - 10° - 1°

The cause of the heat of the atmosphere is evidently the cause of sun's rays; this has been observed and acknowledged in all ages. The heat which they produce is less according as they fall more obliquely; hence the temperature constantly diminishes from the equator to the pole, because their obliquity constantly increases with the latitude. But if the heat depended on the solar rays alone, it would disappear in the Weather, the polar regions during winter when the sun ceases to rise. This, however, is by no means the case; the mean temperature, even at the pole, is $31^\circ$; and we find within the arctic circle as hot weather as under the equator. The reason of this is, that the sun's rays heat the earth considerably during summer; this heat it retains and gives out slowly during winter, and thus moderates the violence of the cold; and summer returns before the earth has time to be cooled down beyond a certain degree. This is the reason that the cold weather does not take place at the winter solstice, but some time after when the temperature of the earth is lowest; and that the greatest heat takes place also some considerable time after the summer solstice, because then the temperature of the earth is highest. For pure air is not heated by the solar rays which pass through it, but acquires slowly the temperature of the earth with which it is in contact. This is the reason why the temperature decreases according to the elevation above the level of the sea (a).

Since the atmosphere is heated by contact with the surfaces of the earth, its temperature must depend upon the capacity of that surface for receiving and transmitting heat. Now this capacity differs very much in land and water. Land, especially when dry, receives heat with great readiness, but transmits it through its own substance very slowly. Dr Halles found, that in 1724, when the air and surface of the earth were both at $38^\circ$, a thermometer placed only two inches below the surface stood at $85^\circ$; another 16 inches below the surface, at $70^\circ$; and another 24 inches deep, at $68^\circ$. The two last-mentioned thermometers retained the same temperature till the end of the month, though the temperature of the air frequently varied, and then fell only to $63^\circ$ or $61^\circ$. The earth, at about 80 or 90 feet below its surface, constantly retains the same temperature; and this is nearly equal to the mean annual heat of the country. Hence the mean annual temperature of any country may be found out pretty accurately, by examining the heat of deep wells or springs. Water, on the contrary, receives heat slowly, on account of its transparency; but what it does receive, is very quickly transfused through the whole mass.

Land is often heated and cooled to a much greater degree than sea is. Dr Raymond often found the earth in the neighbourhood of Marseilles heated to $170^\circ$, but he never found the sea above $77^\circ$; in winter the earth was often cooled down to $14^\circ$, but the sea never lower than $45^\circ$. The sea atmosphere, therefore, ought to preserve a much more uniform temperature than the land atmosphere; and we find this in fact to be the case. The cause of the greater equality of water than land is evident. In summer the surface of the sea is constantly cooled down by evaporation; and in winter, whenever the surface is cooled, it descends to the bottom from its increased gravity, and its place is supplied by warmer water. This process goes on continually, and the winter is over before the atmosphere has been able to cool down the water beyond a certain degree. It must be remembered also, that water has a greater capacity for heat than land has, and therefore is longer either in heating or cooling.

These observations will enable us to explain the difference which takes place between the annual temperature of the atmosphere above the ocean and that of places at some considerable distance from it. As the sea is never heated so highly as the land, the mean summer temperature at sea may be considered, all over the world, as lower than on land. During winter, when the power of the sun's rays in a great measure ceases, the sea gives out heat to the air much more readily than the earth; the mean winter temperature, therefore, at sea is higher than on land; and in cold countries the difference is so great that it more than counterbalances the difference which takes place in summer; so that in high latitudes the mean annual temperature ought to be greater at sea than on land. Accordingly from lat. $50^\circ$ to $35^\circ$, to find the temperature of a place, the standard temperature for the same latitude ought, according to Dr Kirwan, to be depressed half a degree for every 50 miles distance; for the cold which takes place in winter always increases in proportion to the distance from the standard. At a less distance than 50 miles the temperatures of land and sea are to be blended together by sea and land winds, that there is little difference in the annual mean. In lower latitudes than $35^\circ$, the rays of the sun, even in winter, retain considerable power; the surface of the earth is never cooled very low, consequently the difference between the annual temperatures of the sea and land becomes less. As we approach nearer to the equator, the power of the solar rays during winter increases so that the mean winter temperature of the land atmosphere approaches nearer and nearer to that of the sea, till at last at the equator it equals it. After we pass lat. $30^\circ$, therefore, the mean annual land temperature gradually exceeds that of the sea more and more till at the equator it exceeds it a degree for every 50 miles distance.

Such then, in general, is the method of finding the mean annual temperature over the globe. There are, however, several exceptions to these general rules, which come now to be mentioned.

That part of the Pacific ocean which lies between north lat. $52^\circ$ and $66^\circ$ is no broader at its northern extremity than 42 miles, and at its southern extremity than 1300 miles; it is reasonable to suppose, therefore, that its temperature will be considerably influenced by the surrounding land, which consists of ranges of mountains covered, a great part of the year, with snow; and there are besides a great many high, and consequently cold, islands scattered through it. For these reasons Mr Kirwan concludes, that its temperature is at least 4 or 5 degrees below the standard. But we are not yet furnished with a sufficient number of observations to determine this with accuracy.

It is the general opinion, that the southern hemisphere, beyond the 42nd degree of latitude, is considerably colder than the corresponding parts of the northern hemisphere. The cause of this we shall endeavour to assign in the article Wind.

Small seas surrounded with land, at least in temperate climates, are generally warmer in summer and colder in winter than the standard ocean, because they are a good deal influenced by the temperature of the land. The Gulph of Bothnia, for instance, is for the most part frozen in winter; but in summer it is sometimes heated to $70^\circ$, a degree of heat never to be found in the opposite part of the Atlantic. The German sea is above three degrees colder in winter, and five degrees warmer in summer, than the Atlantic. The Mediterranean Sea is, for the greater part of its extent, warmer both in summer and winter than the Atlantic.

(a) It was some time ago the favourite opinion of philosophers, that the heat of the earth was derived from a mass of fire in its centre. But there does not seem any probability in the opinion, as the heat of the earth does not increase the deeper we go, but remains constant nearly at the mean heat of the place. In the mine of Joachimsthal in Bohemia, one of the deepest existing, Mr Monnet found the temperature at the depth of 1700 feet to be $50^\circ$. The temperature of the earth has even been found to diminish the deeper we go, though never lower than $36^\circ$. The eastern parts of North America are much colder than the opposite coast of Europe, and fall short of the standard by about 10° or 12°, as appears from American Meteorological Tables. The causes of this remarkable difference are many. The highest part of North America lies between the 45th and 50th degree of north latitude, and the 100th and 110th degree of longitude west from London; for there the greatest rivers originate. The very height, therefore, makes this spot colder than it otherwise would be. It is covered with immense forests, and abounds with large swamps and morasses, which render it incapable of receiving any great degree of heat; so that the rigour of winter is much less tempered by the heat of the earth than in the old continent. To the east lie a number of very large lakes; and farther north, Hudson's Bay; about 50 miles on the south of which there is a range of mountains which prevent its receiving any heat from that quarter. This bay is bounded on the east by the mountainous country of Labrador and by a number of islands. Hence the coldness of the north-west winds and the lowness of the temperature. But as the cultivated parts of North America are now much warmer than formerly, there is reason to expect that the climate will become still milder when the country is better cleared of woods, though perhaps it will never equal the temperature of the old continent.

Islands are warmer than continents in the same degree of latitude; and countries lying to the windward of extensive mountains or forests are warmer than those lying to the leeward. Stones or sand have a less capacity for heat than earth has, which is always somewhat moist; they heat or cool, therefore, more rapidly and to a greater degree. Hence the violent heat of Arabia and Africa, and the intense cold of Terra del Fuego. Living vegetables alter their temperature very slowly, but their evaporation is great; and if they be tall and close, as in forests, they exclude the sun's rays from the earth, and shelter the winter snow from the wind and the sun. Woody countries, therefore, are much colder than those which are cultivated.

Thus we have endeavoured to ascertain the mean temperature of every climate, and to assign the causes by which that temperature is governed. Mr Kirwan, in his admirable Treatise on the Temperature of Different Latitudes, has done much to reduce this part of meteorology to regularity, and to subject it to calculation; and he has in some measure succeeded. To enable our readers to judge how far his rules agree with facts, we shall subjoin a table of the mean temperature of a variety of places drawn up from actual observations.

### Table of the Mean Temperature of different Places

| Latitude | Year of Observation | Place | Mean Heat of the Ther. | |----------|--------------------|-------|------------------------| | 43° 50' | 36 | Lucca† | 60.9 | | 43° 51' | 5 | Nîmes* | 62.3 | | 44° 50' | 16 | Bordeaux* | 56.3 | | 45° 22' | 7 | Padua* | 53.8 | | | | St Gotthard* | 30 | | 45° 28' | 16 | Milan§ | 54.9 | | 46° 31' | 10 | Lausanne|| | 43.5 | | 46° 35' | 15 | Poitiers* | 52.7 | | 47° 12' | 13 | Chinon* | 53.6 | | 47° 14' | 11 | Béziers* | 51.3 | | 48° 27' | 12 | Chartres* | 50.7 | | 48° 31' | 12 | St Brieux* | 52.1 | | 48° 50' | 28 | Paris* | 52.4 | | 48° 56' | 6 | Ratisbon* | 49.1 | | 49° 59' | 22 | Montorenci* | 50.9 | | 49° 26' | 6 | Mannheim* | 51.5 | | 49° 46' | 24 | Neufchatel* | 50.9 | | 50° 17' | 14 | Arras* | 48.2 | | 50° 51' | 5 | Breda* | 51.2 | | 51° 31' | 19 | London¶ | 50.6 | | 51° 41' | 7 | Copenhagen* | 51.1 | | 52° 4 | 8 | Hague* | 51.8 | | 52° 30' | 15 | Lynden§§ | 48.3 | | 52° 32' | 11 | Berlin* | 49.1 | | 53° 11' | 13 | Frankfort* | 52.5 | | 55° 45' | 4 | Moscow* | 42.1 | | 57° 3 | 3 | Nain* | 27.5 | | 59° 20' | 15 | Stockholm* | 44.3 | | 59° 56' | 18 | Peterburg‡‡ | 39.5 | | 60° 27' | 10 | Abo* | 41.9 |

As to the daily variations of the temperature of the atmosphere, they are owing to a variety of causes; many of which are probably unknown. Some of them, however, variations are the following:

1. Wind. It is evident that winds flowing from cold countries heat; and that whatever has a tendency to produce such winds must be the cause of unusual cold or heat.

2. Evaporation. Water always absorbs a quantity of heat when it affinnes the state of vapour. Hence the coldness of marshy countries, and the cold which we often experience during and after violent rains. Hence also we may expect a cold winter after a rainy summer, because the usual evaporation carries off the heat of the earth.

3. Vapour, when condensed, gives out a quantity of heat; a country, therefore, may be heated by the condensation over it of vapour brought from a distance. Hence the unfreinless often felt before rain.

4. Vapours, when they remain long over any country, may produce cold by obstructing the passage of the sun's rays to the earth. To this cause Dr Franklin ascribed the very severe winter which followed 1783; a year remarkable for the thick fog which overspread Europe and America. America during several months.—5. When, from any of these causes, the winter has been severer than usual, prodigious quantities of ice may accumulate about the pole, which may contribute something perhaps towards lowering the temperature of several succeeding years.

II. The winds evidently have a very great influence on the weather; the causes which produce them, therefore, ought to be examined with the greatest attention. Were we able to regulate their motions, we might, in a great measure, mould the climate of any country according to our pleasure; were we able to foresee them, it would be of the greatest importance to navigation and agriculture. In the torrid zone, where they are regular, the mean annual temperature remains almost always the same; their irregularity increases as we approach the pole, and in the same manner the difference between the mean annual temperature increases with the latitude.

Wind is produced chiefly by the action of the sun on the atmosphere; there are many other causes, however, and some perhaps of which we are yet ignorant. But we shall reserve this part of our subject, on account of its importance and extent, for a separate article.

III. We come now to the most difficult part of our subject, the phenomena and causes of rain. It has been long known, that water is constantly rising from the whole surface of the globe, in the form of vapour, and mixing with the atmosphere. Evaporation has been ascribed to various causes; but the greater number of philosophers have for some time past acquiesced in the theory first advanced by Dr Halley, that it was produced by a real solution of water in air, just as sugar or salt is dissolved in water. This theory is supported by a great many very plausible arguments, which at the first view seem to establish its truth.—These arguments, however, are not all of them so conclusive as they appear. Thus it was thought, that because evaporation was promoted by heat, and retarded by cold, it bore an exact resemblance to the solution of salts in liquids; but it is now known that evaporation is not so much retarded by cold as was at first supposed; that in some circumstances it is even promoted by it; and that it does not depend so much upon the absolute degree of heat or cold, as upon the difference of temperature between the atmosphere and the evaporating surface. Besides, water evaporates much more rapidly in a vacuum than in the open air, which could not possibly be the case if evaporation were owing to the solution of water in air.

Evaporation, then, cannot be owing to solution of water in air; it is produced by the combination of a certain quantity of caloric with the particles of water, by which it is converted into an elastic fluid lighter than air, which therefore immediately ascends and mixes with the atmosphere. This was long ago shown by Dr Black to be the way in which steam or the vapour arising from boiling water is produced. The same principles were afterwards applied by Mr De Luc to spontaneous evaporation; and the proofs upon which this theory rests are quite convincing. But though evaporation is not produced by air, vapour would very soon condense and return to its former state by contact with colder bodies, unless it were attracted and supported by air.

We are indebted to the experiments of Saussure and De Luc for much of our knowledge of the qualities of vapour. It is an elastic invisible fluid like common air, but lighter; being to common air, according to Saussure, as 10 to 14, or, according to Kirwan, as 10 to 12: it cannot pass beyond a certain maximum of density, otherwise the particles of water which compose it unite together, and form small, hollow, visible vesicles, called vesicular vapour; which is of the same specific gravity with atmospherical air. It is of this vapour that clouds and fogs are composed. This maximum increases with the temperature; and at the heat of boiling water is so great, that steam can resist the whole pressure of the air, and exist in the atmosphere in any quantity. See Meteorology, No. 7—23.

Evaporation, at least in our climate, is about four times greater during the summer than the winter half-year; other vapours being equal, it is so much the more abundant the greater the difference is between the temperature of the air and that of the evaporating surface; to such a degree, the nearer they approach to the same temperature; and least of all when they actually arrive at it. Whenever the atmosphere is more than 15 degrees colder than the evaporating surface, little evaporation takes place at all. Evaporation is powerfully promoted by winds, especially cold winds blowing into warm countries, or warm winds blowing into cold countries*. Tracts of land covered with trees or vegetables emit more vapour than the same space covered with water. From the experiments of Mr Williams, the quantity appears to be one third more†. But the method in which these experiments were made (the same objection lies against several of Dr Hailes's experiments, the original discoverer of the fact) prevented him from ascertaining exactly the quantity of vapour emitted by plants. He made the plants grow in a box well closed up from the air, measured the quantity of water with which he supplied them, and at the end of the experiment weighed the box and the plants themselves. By this means he knew pretty accurately the quantity of water which the plants had absorbed, and which had afterwards disappeared; and all this he concluded had been emitted by the plants in the state of vapour. But it is well known that plants have the power of decomposing water, of retaining the hydrogen, and throwing off the oxygen. A part of the water then was decomposed and changed into air; and the quantity of this ought to have been ascertained and subtracted. Still, however, the quantity of vapour emitted by vegetables is very great. Evaporation is promoted by heat, and is therefore much greater in the torrid zone than in our latitudes. There, too, the difference between the quantities in summer and winter is much less than in our climate, because the difference between the temperature of the two seasons is less. Animals also are continually throwing off vapour by insensible perspiration; the quantity of which is exceedingly different, according to the climate, season, and temperament, and cannot therefore be calculated exactly. According to Keil, a simple man perspires 31 ounces of vapour in 24 hours, and consequently 767 pounds of water in a year. The quantity of vapour then which is emitted by animals alone must be very great.

From an experiment made by Dr Watson in England, during summer, when the earth had been burnt up by a month's drought without rain, it appears that 1600 gallons of water were evaporated from a single acre in 12 hours.—If we were to suppose that this represented the mean daily evaporation all over the globe, it would be easy to calculate the quantity of water annually evaporated from the whole of its surface. And if we consider the state of the earth when the experiment was made, the situation of England nearer the pole than the equator, and the evaporation constantly going on from animals and vegetables, which is not taken in, we will surely not think the mean amount too great. 1600 gallons in 12 hours is 3200 in 24 hours. Let us call it only 3000, which is equal to 693,000 cubic inches. An acre contains 272,640 square inches; so that the daily evaporation from every square inch will be about 1.11 of a cubic inch. This in a year will amount to somewhat more than 40 cubic inches for every square inch. From the expe- experiments of Mr. Williams*, it appears, that in Bradford in New England the evaporation during 1772 amounted to 42 inches; but from the way that his experiments were conducted, the amount was probably too great. These experiments, however, serve to show, that our calculation is not perhaps very remote from the truth. 40 inches from every square inch on the surfaces of the globe makes 107,942 cubic miles, equal to the water annually evaporated over the whole globe.

Were this prodigious mass of water all to subside in the atmosphere at once, it would increase its mass by about a twelfth, and raise the barometer nearly three inches. But this never happens, no day passes without rain in some part of the earth; so that part of the evaporated water is constantly precipitated again. Indeed it would be impossible for the whole of the evaporated water to subside in the atmosphere at once, at least in the state of vapour.

M. De Saufure has shown, that when the thermometer is at 66°, a cubic foot of air cannot contain more vapour than what is equivalent to 8 grains of water. If more than this be added, it will pass its maximum, be converted into vesicular vapour, and at last fall down in drops of rain. At the temperature of 32° a cubic foot of air can contain only 4 grains, and the quantity it can contain is increased, 1109 of a grain by every additional degree of heat. Supposing then that the whole atmosphere was saturated with water, it would not amount to the hundredth part of the quantity of water evaporated annually.

The quantity of vapour existing in the atmosphere is indicated by the hygrometer. Water has the property of arriving at a state of equilibrium in hygroscopic substances; that is, supposing a certain quantity of water attached to a hygroscopic substance, if another hygroscopic substance be brought into contact with it containing less water, some of the water attached to the first substance will leave it, and attach itself to the other, till both contain the same proportion of water. Air is a hygroscopic substance, and so is every thing of which hygrometers are made. Now the hygrometer never points at extreme moisture while the air continues transparent, and consequently contains nothing but invisible vapour; the atmosphere therefore, while transparent, never contains the greatest possible quantity of vapour.

The higher regions of the atmosphere contain less vapour than the strata near the surface of the earth. This was observed both by M. De Saufure and M. De Luc, who mentions several striking proofs of it. See Meteorology, p. 10, &c.

At some height above the tops of mountains the atmosphere is probably still drier; for it was observed both by Saufure and De Luc, that on the tops of mountains the moisture of the air was rather less during the night than the day. And there can be little doubt that every stratum of air descends a little lower during the night than it was during the day, owing to the cooling and condensing of the stratum nearest the earth. Vapours, however, must ascend very high, for we see clouds forming far above the tops of the highest mountains.

Rain never begins to fall while the air is transparent; the invisible vapours first pass their maximum, and are changed into vesicular vapours; clouds are formed, and these clouds gradually dissolve in rain. Clouds, however, are not formed in all parts of the horizon at once; the formation begins in one particular spot, while the rest of the air remains clear as before; this cloud rapidly increases till it overshadows the whole horizon, and then the rain begins.

It is remarkable, that though the greatest quantity of vapours exist in the lower strata of the atmosphere, clouds never begin to form there, but always at some considerable height. It is remarkable, too, that the part of the atmosphere at which they form has not arrived at the point of extreme moisture, nor near that point even a moment before their formation. They are not formed then, because a greater quantity of vapour had got into the atmosphere than could remain there without raising its maximum. It is still more remarkable, that when clouds are formed, the temperature of the spot in which they are formed is not always lowered, though this may sometimes be the case. On the contrary, the heat of the clouds themselves is sometimes greater than that of the surrounding air. Neither then is the formation of clouds owing to the capacity of air for combining with moisture being lessened by cold; so far from that, we often see clouds, which had remained in the atmosphere during the heat of the day, disappear in the night, after the heat of the air was diminished.

The formation of clouds and rain, then, cannot be accounted for by a single principle with which we are acquainted. It is neither owing to the saturation of the atmosphere, nor the diminution of heat, nor the mixture of airs of different temperatures, as Dr Hutton supposes; for clouds are often formed without any wind at all either above or below them; and even if this mixture constantly took place, the precipitation, instead of accounting for rain, would be almost imperceptible.

It is a very remarkable fact, that evaporation often goes on for a month together in hot weather without any rain. This sometimes happens in this country; it happens every year in the torrid zone. Thus at Calcutta, during January 1785, it never rained at all; the mean of the thermometer for the whole month was 66½ degrees; there was no high rainfall, wind, and indeed during great part of the month little wind at all.

The quantity of water evaporated during such a drought must be very great; yet the moisture of the air, in appearance, instead of being increased, is constantly diminishing, and at last disappears almost entirely. For the dew, which is at first copious, diminishes every night; and if Dr Watson's experiment formerly mentioned be attended to, it will not be objected that the quantity of evaporation is also very much diminished. Of the very dry state to which the atmosphere is reduced during long droughts, the violent thunder-storms with which they often conclude is a proof, and a very decisive one. Now what becomes of all this moisture? It is not accumulated in the atmosphere above the country from which it was evaporated, otherwise the whole atmosphere would in a much less period than a month be perfectly saturated with moisture. If it be carried up daily through the different strata of the atmosphere, and waited to other regions by superior currents of air, how is it possible to account for the different electrical state of the clouds situated between different strata, which often produces the most violent thunder-storms? Are not vapours conductors of the electric fluid; and would they not have daily restored the equilibrium of the whole atmosphere through which they passed? Had they traversed the atmosphere in this manner, there would have been no negative and positive clouds, and consequently no thunder-storms. They could not have remained in the lower strata of the atmosphere, and been daily carried off by winds to other countries; for there are often no winds at all during several days to perform this office; nor in that case would the dews diminish, nor could their presence fail to be indicated by the hygrometer.

It is impossible for us to account for this remarkable fact upon any principle with which we are acquainted. The water can neither remain in the atmosphere, nor pass through it. it in the state of vapour. It must therefore assume some other form; but what that form is, or how it assumes it, we know not.

It will immediately occur to every body, that vapour is decomposed in the atmosphere, and changed into oxygen and hydrogen gas. But is it true that a greater quantity of oxygen exists in the atmosphere after a long drought than immediately after rain? Have such prodigious quantities of hydrogen been found in the atmosphere as must always exist in it if these hypotheses were true? Has any hydrogen ever been found in analyzing atmospheric air? Or if hydrogen, from its lightness, ascends to the higher regions of the atmosphere, what causes it to descend at particular times, contrary to that lightness, in order to come into contact with oxygen? Do not clouds often form on mountains round the habitations of men? Yet has the presence of hydrogen been ever ascertained by any phenomena? Would it not produce dangerous conflagrations when it came into contact with fire? But has this been the case in a single instance? If this hypothesis were true, could rain take place at all without a conflagration in the atmosphere? Yet has any such conflagration been ever observed? The hypothesis, then, that vapour is changed into oxygen and hydrogen in the atmosphere, and that rain is produced by the reunion of these elements, cannot be admitted, though it is not improbable that some small part of it actually undergoes this change. See Wind.

We do not take notice of M. De Luc's conjecture about the composition of the atmosphere, because it is not supported by a single proof, and because he refuses to believe the analysis of the atmosphere resulting from the very decisive experiments of Scheele, Lavoisier, and Priestley, though he has seen them often performed, and has nothing to urge against their force. There is no philosopher to whom meteorology lies under greater obligations than to M. De Luc. His discoveries have been many and important, his experiments ingenious, and his application unwearied; but his conjectures are like those of every other man who attempts to fathom the wisdom of the Almighty. Were we possessed of an understanding equal to that of the Author of Nature, we might expect, with reason, to dive by our conjectures into the mysteries of his operations; but in our present state they are vain.

Evaporation goes on longest without producing rain in the torrid zone, where the heat is greatest; it goes on longest also in every place in summer, when the heat is also greatest: heat therefore seems to be an agent.

There are then two steps of the process between evaporation and rain, of which at present we are completely ignorant: 1. What becomes of the vapour after it enters into the atmosphere? 2. What makes it lay aside the new form which it must have assumed, and return again to its state of vapour, and fall down in rain? And till these two steps be discovered by experiments and observations, it will be impossible for us to give a rational or a useful theory of rain.

It has for some time past been the opinion of philosophers, that electricity is the principal agent in producing rain; and M. Bertholon affirms us, that by raising proper conductors to draw off the electrical matter from the atmosphere, the quantity of rain may be diminished at pleasure. That the electric fluid acts a very important part in nature, cannot be doubted, and it is not improbable that it may be the agent in producing rain. This supposition indeed is supported by many facts. Dew at least exhibits a great many electrical phenomena; it is attracted by points, and attaches itself to some substances, while it avoids others. Whenever there are no clouds, the electricity of the atmosphere is always positive; but the formation of clouds produces considerable changes in the state of its electricity. The atmosphere also gives signs of electricity constantly during rain; and clouds are evidently attracted by mountains.—In what manner, however, the electrical fluid produces rain (if it is the agent at all) is still unknown. Some philosophers affirm us, that clouds are induced to dissolve in rain by becoming negative, others by becoming strongly positive, and both support their opinion by experiments. We do not see the analogy, however, between clouds and plates of metal covered with drops of water. And even if their opinion were well founded, the production of the clouds themselves would remain to be accounted for.

The mean annual quantity of rain is greatest at the equator, and decreases gradually as we approach the poles.

Thus at * Granada, Antilles, 12° N. lat., it is 126 inches. According to the latitude,

* Cape François, St Domingo 19° 46' 120

† Calcutta 22° 23' 81

* Rome 41° 54' 39

England 33° 32

‡ Pittsburgh 59° 16' 16

On the contrary, the number of rainy days is smallest at the equator, and increases in proportion to the distance from it. Appendix.

From north latitude 12° to 43° the mean number of rainy days is 78; from 43° to 46° the mean number is 103; from 46° to 50° it is 134; from 51° to 60°, 161.

The number of rainy days is often greater in winter than in summer; but the quantity of rain is greater in summer than in winter. At Pittsburgh, the number of rainy or P. Colte snowy days during winter is 84, and the quantity which falls is only about five inches; during summer the number of rainy days is nearly the same, but the quantity which falls is about 11 inches.

More rain falls in mountainous countries than in plains. Among the Andes it is said to rain almost perpetually, while in Egypt it hardly ever rains at all. If a rain-gauge be placed on the ground, and another at some height perpendicularly above it, more rain will be collected into the lower than into the higher; a proof that the quantity of rain increases as it descends, owing perhaps to the drops attracting vapour during their passage through the lower strata of the atmosphere where the greatest quantity resides. This, however, is not always the case, as Mr Copland of Dumfries discovered in the course of his experiments. He observed also, that when the quantity of rain collected in the lower gauge was greatest, the rain commonly continued for some time; and that the greatest quantity was collected in the higher gauge only either at the end of great rains, or during rains which did not last long. These observations are important, and may, if followed out, give us new knowledge of the causes of rain. They seem to show, that during rain the atmosphere is somehow or other brought into a state which induces it to part with its moisture; and that the rain continues as long as this state continues. Were a sufficient number of observations made on this subject in different places, and were the atmosphere carefully analysed during dry weather, during rain, and immediately after rain, we might soon perhaps discover the true theory of rain.

Rain falls in all seasons of the year, at all times of the day, and during the night as well as the day; though, according to M. Toaldo, a greater quantity falls during the day than the night. The cause of rain, then, whatever it may be, must be something which operates at all times and seasons. Rain falls also during the continuance of every wind, but oftentimes when the wind blows from the south. Falls of rain often happen likewise during perfect calms. It appears from a paper published by M. Cotte in the Journal de Physique for October 1791, containing the mean quantity of rain falling at 147 places, situated between north latitude 11° and 65°, deduced from tables kept at these places, that the mean annual quantity of rain falling in all these places is 34.7 inches. Let us suppose then (which cannot be very far from the truth) that the mean annual quantity of rain for the whole globe is 34 inches. The superficies of the globe consists of 17,981,412 square miles, or 686,407,498,471,475,200 square inches. The quantity of rain therefore falling annually will amount to 2,333,371,650,812,930,156,870 cubic inches, or somewhat more than 9,175,1 cubic miles of water. This is 16,191 cubic miles of water less than the quantity of water evaporated. It seems probable therefore, if the imperfection of our data warrant any conclusion, that some of the vapour is actually decomposed in the atmosphere, and converted into oxygen and hydrogen gas.

The dry land amounts to 52,745,253 square miles (see the article Sea, n° 1); the quantity of rain falling on it annually therefore will amount to 50,960 cubic miles. The quantity of water running annually into the sea (see Sea, n° 3) is 13,140 cubic miles; a quantity of water equal to which must be supplied by evaporation from the sea, otherwise the land would soon be completely drained of its moisture.

The quantity of rain falling annually in Great Britain may be seen from the following table:

| Years of observation | Places | Rain in inches | |----------------------|--------|---------------| | 3 | Dover § | 37,52 | | | Ware, Hertfordshire § | 23,9 | | | London † | 17,5 | | | Kimbolton ‡ | 23,9 | | | Lyndon § | 22,210 | | | Chatworth, Derbyshire § | 27,365 | | | Manchester § | 43,1 | | | Liverpool § | 34,41 | | | Lancaster § | 45,3 | | | Kendal § | 61,223 | | | Dumfries § | 30,127 | | | Branxholm, 44 miles south-west of Berwick ¶ | 31,26 | | | Langholm ¶ | 36,73 | | | Dalkeith ¶ | 25,124 | | | Glasgow * | 31 | | | Hawkhill ** | 28,966 |

In this country it generally rains less in March than in November, in the proportion at a medium of 7 to 12. It generally rains less in April than October in the proportion of 1 to 2 nearly at a medium. It generally rains less in May than September, the chances that it does so are at least as 4 to 5; but when it rains plentifully in May (as 1.8 inches or more), it generally rains but little in September; and when it rains one inch or less in May, it rains plentifully in September *.

IV. Thunder has been explained at such great length in the article Electricity, that we shall content ourselves at present with a few remarks.

Thunder is exceedingly frequent in the torrid zone, and it seems to decrease gradually till we approach latitude 60°, or perhaps farther north. During the year 1785, for instance, there were 90 thunderstorms at Calcutta. According to Professor Muschenbroek, it thunders at Utrecht at a medium 15 times annually: in this country the medium is considerably below that number. Thunder, too, seems to be very common in some polar regions. The Abbé Chappe informs us, that he observed thunder much more frequently at Tobolski and in other parts of Siberia than in any other country. Muschenbroek, however, affirms, we know not upon what authority, that it never thunders at all in Greenland and at Hudson's Bay. Thunderstorms happen almost always during the summer, and very seldom in winter. During the year 1785 above mentioned, it never thundered at Calcutta in January, November, nor December. In this country a thunderstorm during winter is exceedingly rare.

The phenomena of thunder are now no longer a secret, since the great Franklin discovered the identity of lightning and electricity; a discovery inferior to none in the annals of philosophy. But though we can explain the nature of thunder in general, and the manner in which it is produced, there are several difficulties still remaining, which future experiments and observations only can remove. Air is an electric per se, and cannot therefore when dry conduct electrical matter from one part to another. We know from the experiments of Dr Franklin and others, that the atmosphere constantly contains in it a quantity of electric matter. If a stratum of dry air were electrified positively, it would occasion a negative electricity in the neighbouring stratum. Suppose now that an imperfect conductor were to come into contact with each of these strata, we know from the principles of electricity that the equilibrium would be restored, and that this would be attended with a loud noise, and with a flash of light. Clouds which consist of vesicular vapours mixed with particles of air, are imperfect conductors; if a cloud therefore come into contact with two such strata, a thunder clap would follow. If a positive stratum be situated near the earth, the intervention of a cloud will, by serving as a stepping-stone, bring the stratum within the striking distance, and a thunder clap will be heard while the electrical fluid is discharging itself into the earth. If the stratum be negative, the contrary effects will take place. It does not appear, however, that thunder is often occasioned by a discharge of electric matter from the earth into the atmosphere. The accidents, most of them at least, which were formerly ascribed to this cause, are now much more satisfactorily accounted for by Lord Stanhope's Theory of the Returning Stroke. Neither does it appear that electricity is often discharged into the earth, as the effects of few thunderstorms are visible upon the earth; that it is so sometimes, however, is certain. The experiments of Mr Sansfire have demonstrated, that electrical matter is carried into the atmosphere by simple evaporation; so that there is no difficulty in understanding how particular strata of air may be supplied with a sufficient quantity of electrical fluid to be charged positively; and we know that in that case a negative state must be produced in the neighbouring stratum. In what particular manner, however, this electrical matter is accumulated in particular strata of air, and how it comes to be separated from the vapour to which it was united, remain still secrets. They are intimately connected with the causes of evaporation and rain, whatever they may be, and probably the discovery of the causes of either would lead to that of the other.

V. The gravity of the atmosphere was first demonstrated by Torricelli, the disciple of Galileo (see Pneumatics, the Gravity 25). A column of air, the basis of which is a square of the inch, weighs at a medium 15 pounds. The weight of the atmosphere is measured by the barometer. It is greatest at at the level of the sea, because there the column of air is longest; there the mean height of the barometer is 30 inches. This Sir George Shuckburgh found to be the case in the Mediterranean and the Channel, in the temperature of 55° and 60°; Mr Bouguer, on the coast of Peru, in the temperature of 84°; and Lord Mulgrave, in latitude 80°. The mean height of the barometer is less the higher any place is situated above the level of the sea, because the column of air which supports the mercury is the shorter. The barometer has accordingly been used for measuring heights. It indicates, too, with a great deal of accuracy, all the variations in the gravity of the atmosphere; falling when the atmosphere is lighter, and rising when it is heavier, than usual. These changes have attracted the attention of philosophers ever since the discovery of the barometer; and many attempts have been made to explain them, some of which have been mentioned under the word Barometer. These variations come naturally to be examined here, because the causes which produce them, whatever they are, must have a great deal of influence on the weather.

Between the tropics the variations of the barometer are exceedingly small; and it is remarkable, that in that part of the world it does not descend above half as much for every 200 feet of elevation as it does beyond the tropics*. In the torrid zone, too, the barometer is elevated about two-thirds of a line twice every day; and this elevation happens at the same time with the tides of the sea §.

As the latitude advances towards the poles, the range of the barometer gradually increases, till at last it amounts to two or three inches. This gradual increase will appear from the following table:

| Latitude | Places | Range of the Barometer | |----------|--------|------------------------| | * Kirwan, Irish Trans., vol. iii. p. 22 | Peru | 0°20° * | | 47° | Calcutta | 0°77 † | | 40°55′ | Naples | 1°00° * | | 51°8′ | Dover | 2°47° ‡ | | 53°13′ | Middlewick | 3°00° § | | 53°23′ | Liverpool | 2°89° ‡ | | 59°56′ | Pittsburgh | 3°45° ‡ |

In North America, however, the range of the barometer is a great deal less than in the corresponding European latitudes. In Virginia, for instance, it never exceeds 1°1° †.

The range of the barometer is greater at the level of the sea than on mountains, and in the same degree of latitude the extent of the range is in the inverse ratio of the height of the place above the level of the sea.

From a table published by Mr Cotte in the Journal de Physique ‡, it seems exceedingly probable that the barometer has always a tendency to rise from the morning to the evening; and that this tendency is greatest between two o'clock in the afternoon and nine at night, at which hour the greatest elevation takes place; that the elevation of nine o'clock differs from that of two by 4°ths, while that at two differs from the morning elevation only by 1°th; and that in certain climates the greatest elevation takes place at two o'clock. We shall insert a part of the table on which these observations are founded, which we have reduced to the English standard.

| Places | Years of observation | Mean height of Barometer | |--------|---------------------|-------------------------| | Arles | 6 | 29°9347 | | Arras | 6 | 29°6688 | | Bordeaux | 11 | 29°7121 | | Cambrai | 13 | 29°8756 | | Chinon | 12 | 29°7719 | | Dunkirk | 8 | 29°9199 | | Hagenau | 10 | 29°5048 | | Laon | 7 | 29°3354 | | Lille | 6 | 29°9165 | | Mayenne | 7 | 29°7172 | | Mainz | 5 | 29°6167 | | Montmorenci | 22 | 29°6536 | | Mulhouse | 7 | 29°1873 | | Obernheim | 12 | 29°4842 | | Paris | 67 | 29°8002 | | Poitiers | 12 | 29°7276 | | Rouen | 11 | 29°8007 | | Rome | 3 | 29°8607 | | St Maurice | le Gerard | 29°8160 | | Troyes | 10 | 29°6885 |

The range of the barometer is greater in winter than in summer. Thus at Kendal the mean range of the barometer for five years, during October, November, December, January, February, March, was 7°982; and for the six summer months 5°447*.

In serene and settled weather it is generally high; and low in calm weather, when the air is inclined to rain; it sinks on high winds, rises highest on easterly and northerly winds, and sinks when the wind blows from the south‡. At Calcutta §, however, it is always highest when the wind blows from the north-west and north, and lowest when it blows from the south-east.

The barometer falls suddenly before tempests, and undergoes great oscillations during their continuance.—Mr Copland || of Dumfries has remarked, that a high barometer is attended with a temperature above, and a low barometer with one below, the monthly mean.—Such are the variations of the barometer as far as they have yet been observed. Let us now endeavour to account for them as well as we can.

It is evident that the density of the atmosphere is least at the equator, and greatest at the poles; for at the equator the centrifugal force, the distance from the centre of the earth, and the heat, all of which tend to diminish the density of the air, are at their maximum, while at the pole they are at their minimum. The mean height of the barometer at the level of the sea, all over the globe, is 30 inches; the weight of the atmosphere, therefore, is the same all over the globe. The weight of the atmosphere depends on its density and height: where the density of the atmosphere is greatest, its height must be least; and, on the contrary, where its density is least, its height must be greatest. The height of the atmosphere, therefore, must be greatest at the equator, and least at the poles; and it must decrease gradually between the equator and the poles, so that its upper surface will resemble two inclined planes, meeting above the equator their highest part*.

During summer, when the sun is in our hemisphere, the mean 43°, &c. mean heat between the equator and the pole does not differ so much as in winter. Indeed the heat of northern countries at that time equals the heat of the torrid zone; thus in Russia, during July and August, the thermometer rises to $89^\circ$. Hence the rarity of the atmosphere at the pole, and consequently its height, will be increased. The upper surface of the atmosphere, therefore, in the northern hemisphere will be less inclined; while that of the southern hemisphere, from contrary causes, will be much more inclined.

The very reverse will take place during our winter.

The density of the atmosphere depends in a great measure on the pressure of the superincumbent column, and therefore decreases, according to the height, as the pressure of the superincumbent column constantly decreases. But the density of the atmosphere in the torrid zone will not decrease so fast as in the temperate and frigid zones; because its column is longer, and because there is a greater proportion of air in the higher part of this column. This accounts for the observation of Mr Caffan, that the barometer only sinks half as much for every 200 feet of elevation in the torrid as in the temperate zones (b). The density of the atmosphere at the equator, therefore, though at the surface of the earth it is less, must at a certain height equal, and at a still greater surplus, the density of the atmosphere in the temperate zones and at the poles.

In the article Wind we shall endeavour to prove, that a quantity of air is constantly ascending at the equator, and that part of it at least reaches and continues in the higher parts of the atmosphere. From the fluidity of air, it is evident that it cannot accumulate above the equator, but must roll down the inclined plane (c) which the upper surface of the atmosphere assumes towards the poles. As the surface of the atmosphere of the northern hemisphere is more inclined during our winter than that of the southern hemisphere, a greater quantity of the equatorial current of air must flow over upon the northern than upon the southern atmosphere; so that the quantity of our atmosphere will be greater during winter than that of the southern hemisphere; but during summer the very reverse will take place. Hence the greatest mercurial heights take place during winter, and the range of the barometer is less in summer than in winter.

The density of the atmosphere is in a great measure regulated by the heat of the place: wherever the cold is greatest, there the density of the atmosphere will be greatest, and its column shortest. High countries, and ranges of lofty mountains, the tops of which are covered with snow the greatest part of the year, must be much colder than other places situated in the same degree of latitude, and consequently the column of air over them much shorter. The current of superior air will linger and accumulate over these places in its passage towards the poles, and thus occasion an irregularity in its motion, which will produce a similar irregularity in the barometer. Such accumulations will be formed over the north-western parts of Asia, and over North America; hence the barometer usually stands higher, and varies less there, than in Europe. Accumulations are also formed upon the Pyrenees, the Alps, the mountains of Africa, Turkey in Europe, Tartary, and Tibet. When these accumulations have gone on for some

---

(b) Should it not be examined whether the number of parts which the mercury sinks for every 200 feet of elevation be not proportioned to the latitude of the place?

(c) It is of no consequence whether the surface of the atmosphere actually forms an inclined plane, or, becoming rarer in a very slow ratio (as is probably the case), ascends much higher than the place at which the equatorial currents begin to flow towards the poles; for still the different heights of air of the same density in different parts of the atmosphere will in fact form an inclined plane, over which these currents will roll, notwithstanding the very rare air which they may displace. The falling of the barometer which generally precedes rain remains still to be accounted for; but we know too little about the causes by which rain is produced to be able to account for it in a satisfactory manner. Probably a rarefied state of the atmosphere is favourable to the production of rain; we know, at least, that it is favourable to evaporation. Supposing the observations which we made upon the changes which vapour undergoes in the atmosphere well founded, may not the vapour in its new form accumulate at a considerable height in the atmosphere? and is not the height at which clouds are always formed a proof of this? May not this substance, whatever it is, when by some means or other it returns to the state of vapour, pass its maximum, and begins to fall in drops of rain, and consequently is no longer supported by the atmosphere, cause the barometer to fall suddenly, at least till new air rushes in to supply its place?

Thus we have endeavoured to describe the various phenomena of the weather, and to account for them as far as the present state of our meteorological knowledge enables us to go.

It will be expected that we should not pass by unnoticed that branch of meteorology which has in all ages attracted the attention of mankind, and in which, indeed, every other part of the science, as far as utility is concerned, evidently centres; we mean the method of prognosticating the weather. All philosophers who have dedicated their attention to meteorology, have built upon the hope of being able to discover, by repeated observations, some rules concerning the periods of the seasons and the changes of the weather, convinced that such discoveries would be of the highest utility, especially in agriculture; for by foreseeing, even in part, the circumstances of the seasons, we would have it in our power to prevent at least a part of the losses arising from them, as by sowing, for instance, the kind of corn best adapted for the rain or the drought which is to ensue.

The influence of the moon on the weather has in all ages been believed by the common people; the ancient philosophers embraced the same opinion, and engrained upon it their pretended science of astrology. Several modern philosophers have thought the opinion worthy of notice; among whom Messrs Lambert, Cotte, and Toaldo, deservedly take the lead. These philosophers, after examining the subject with the greatest attention, have embraced the opinion of the common people, though not in its full extent. To this they have been induced both by the certainty that the moon actually has an influence on the atmosphere as it has on the sea, and by observing that certain situations of the moon in her orbit have almost constantly been attended with changes of the weather either to wind, to calm, to rain, or to drought.

There are ten situations in every revolution of the moon in her orbit, when she most particularly exert her influence on the atmosphere, and when consequently changes of the weather most readily take place. There are (1) the new and (2) full moon, when she exerts her influence in conjunction with or opposition to the sun; (3 and 4) the quadratures; (5) the perigee and (6) apogee (for the difference in the moon's distance from the earth is about 27,000 miles), the two positions of the moon over the equator, one of which, Mr Toaldo calls (7) the moon's attending, and (8) the other the moon's defending equinox, the two lunilices as M. de la Lande has called them, (9) the boreal lunilice, when the moon approaches as near as she can in each lunation to our zenith, (10) the austral, when she is at the greatest distance from it, for the action of the moon varies greatly according to her obliquity. With these ten points Mr Toaldo compared a table of 48 years' observations for Lombardy, and found the result as follows:

| Lunar Points | Attended with a change of weather | Attended with no change | Proportion reduced to the lowest terms | |-----------------------|-----------------------------------|------------------------|--------------------------------------| | New moons | 522 | 82 | 6 : 1 | | Full moons | 506 | 92 | 5 : 1 | | First quarters | 424 | 189 | 2½ : 1 | | Last quarters | 429 | 182 | 2½ : 1 | | Perigees | 546 | 99 | 7 : 1 | | Apogees | 517 | 130 | 4 : 1 | | Ascending equinoxes | 465 | 142 | 3½ : 1 | | Descending equinoxes | 446 | 152 | 2½ : 1 | | Southern lunilices | 446 | 154 | 3 : 1 | | Northern lunilices | 448 | 162 | 2½ : 1 |

And after examining a number of other tables of observations, and combining them with his own, he found the proportions between those lunar points on which changes of the weather took place, and those which passed without any change when reduced to the lowest terms, to be as in the last column of the above table: so that we may wager six to one, that this or that new moon will bring a change of weather, and five to one that a full moon will be attended by a change, and so on. Several of these lunar points often coincide with one another, occasioned by the inequality of the moon's periodical, anomalistical, and synodical revolutions, and by the progressive motion of the apses. Thus the new and full moon sometimes coincide with the apogees, the perigees, &c. These coincidences are the most efficacious. Their changing power, according to Mr Toaldo, is as follows:

- New moon coinciding with the perigee: 33 : 1 - Full moon coinciding with the perigee: 10 : 1

It ought to be remarked, that these changes of the weather seldom or never take place exactly when the moon is in these lunar points, but some time before or after; just as the tide, say the philosophers who contend for the influence of the moon, is not at its height till after the moon has passed the meridian.

The power of the moon over the ocean and the atmosphere is displayed in a particular manner during the apses, in consequence of her different distances from the earth during these two situations. Now the apses advance about 1° in the zodiac every year, and complete a revolution in about eight years and ten months. It is probable that the seasons and the constitutions of years have a period nearly equal to this revolution, and that therefore nearly the same seasons return every ten years. This periodical return of the seasons, as Pliny (d) seems to inform us, was observed by the ancients. And to confirm this Mr Toaldo found, that in Lombardy the quantities of rain which fell during periods of nine successive years were nearly equal; but that this was not true of other periods, for instance, of six, eight, or ten years. By comparing in like manner the quantities of rain published by the Royal Academy of Sciences at Paris, from 1699 to 1752, he found, that

(d) "Tempestates ardoros fuos habere quadrinis annis.—Ootonis vero augeri easdem centesima revolventi se luna."

Lib. 18. c. 25. During the revolution of the apses, there are four remarkable points, the two equinoctial and two solstitial points; in which, when the moon is in perigee, her effect will be most powerful on the weather. The moon passes from one equinoctial point to another in about four years; in them its power is greatest; it is probable, therefore, that when an extraordinary year happens, a return of another may be expected in about four years. As the apses after their revolution return again in the same order as before, it is probable that the return of the seasons will be nearly the same in every series of nine years.

Such, according to Mr Toaldo, is the period at the end of which we are to expect a return of the seasons. Mr Cotte, however, though he does not deny the influence of the revolution of the apses, places greater confidence in the lunar period of 19 years; at the end of which, the new and full moons return to the same day in the Julian year. He supposes, that in like manner the seasons correspond with one another every 19 years. The similarity, he informs us, is striking between the temperatures of the years 1701, 1720, 1739, 1758, and 1777. That of 1758, upon which we have observations much detailed by M. du Hamel, has a remarkable coincidence with 1777; there was scarcely any difference in the temperatures of the corresponding months. The years 1778, 1779, and 1780, have been hot and dry, and they correspond with years which have had the same character. The years corresponding with 1782, especially 1725 and 1764, have been singularly cold, humid, and late, as was the case with 1782.

Such is an imperfect view of the opinions of those philosophers who have endeavoured to establish the influence of the moon over the weather. The most important of their maxims for prognosticating the weather are the following:

1. When the moon is in any of the ten lunar points above mentioned, a change of the weather may be expected. The most efficacious of these points are the conjunctions and apses.

2. The coincidence of the conjunctions with the apses is extremely efficacious; that of the new moon with the perigee gives a moral certainty of a great perturbation.

3. The new and full moons, which sometimes produce no change on the weather, are such as are at a distance from the apses.

4. A lunar point commonly changes the state into which the weather was brought by the preceding point. For the most part the weather never changes but with some lunar point.

5. The apogees, quadratures, and southern lunifices, commonly bring fair weather, for the barometer then rises; the other points tend to make the air lighter, and thereby to produce bad weather.

6. The most efficacious lunar points become stormy about the equinoxes and solstices.

7. A change of weather seldom happens on the same day with a lunar point, but sometimes before and sometimes after it.

8. At the new and full moons about the equinoxes, and even the solstices, especially the winter solstice, the weather is commonly determined to good or bad for three, or even six months.

9. The seasons and years have a period of eight or nine years corresponding with the revolution of the lunar apses, and another of 19 corresponding to the lunar period.

Would it not be worth while to publish a meteorological calendar yearly, marking the time, to which the lunar points correspond, at which changes of the weather may be expected, especially when any of these points coincide; and marking the probability of a change at any particular time? and might not this be attended by a diary of the weather for the 9 or 19 corresponding years? By this means, if there is any probability in the opinion that the moon has influence over the weather, men would be enabled to foresee changes with a considerable degree of probability; and at any rate, we would be able, by the united observations of a whole nation, to determine whether there be any truth in the opinion; and if there be, as its universality would lead one to suppose, succeeding observations would gradually correct the imperfection of our present rules, and enable us to bring our prognostics of the weather to the greatest exactness.

We are not so sanguine, however, as Mr Toaldo and P. Cotte on this subject. Even allowing the influence of the moon on the weather to be as great as they could desire, supposing, which is very far from being the case, that it is not influenced by any other cause, we do not see how the seasons could return in the same order every 9th or 19th year. The motions of the heavenly bodies (especially the moon) are, strictly speaking, incommensurable. The lunar apogee returns to the same situation in eight years ten months (without reckoning hours and minutes); at its first return it will be two months or signs removed from the same situation with the sun; at the end of the second period, four months; and at the end of the third, six months; so that if the season was winter at the beginning, after three revolutions it will be the middle of summer. Now, how in this case can the same seasons return? Supposing the equinoctial points to produce constantly great changes on the weather, if one of them during the first revolution happened in winter, in the second it would happen in spring, and the third in summer; so that what would during the first revolution produce a particular winter, would in the second act upon the spring, and in the third on the summer. Would it in these cases produce similar changes on the weather? Surely not. And whether it did or not, would the same seasons return in every revolution? In six complete revolutions, indeed, or 53 years, the lunar perigee returns to the same situation as at first, very nearly, in the same season: it might be expected then that the seasons would perform a complete revolution every 53 years, and that the 54th would exactly resemble the first, and so on. This may possibly be the case, but it is by no means probable; for when Mr Toaldo compared the quantity of rain which fell at Paris during 1699, 1700, 1701, 1702, &c., with what fell in 1752, 1753, 1754, &c., though the first years in each series corresponded pretty exactly, the difference being only eight lines, there was no such resemblance between any of the following years.

Neither are we convinced that the influence of the moon can have such an effect on the weather as the above mentioned philosophers suppose. The moon only acts, as far as we know at least, by producing tides in the atmosphere; for the refined speculations of Mr Toaldo about its electrical influence we cannot admit, as the electricity of the atmosphere is less during the night, when the moon's influence should be greatest, than during the day. Now we do not see how these tides, supposing them greater than they are, can be adequate to the effects ascribed to them.

Mr Kirwan has lately endeavoured to discover probable rules for prognosticating the different seasons, as far as regards Britain and Ireland, from tables of observations alone. On perusing a number of observations, taken in England from 1677 to 1789, he found,

1. That when there has been no storm before or after the vernal equinox, the ensuing summer is generally dry at least five times in six. 2. That when a storm happens from an easterly point, either on the 19th, 20th, or 21st of May, the succeeding summer is generally dry four times in five.

3. That when a storm arises on the 25th, 26th, or 27th of March (and not before), in any point, the succeeding summer is generally dry four times in five.

4. If there be a storm at south-west or west-south-west on the 19th, 20th, 21st, or 22nd of March the succeeding summer is generally wet five times in six.

In this country winters and springs, if dry, are most commonly cold; if moist, warm; on the contrary, dry summers and autumns are usually hot, and moist summers cold. So that if we know the moistness or dryness of a season, we can judge pretty accurately of its temperature.

From a table of the weather kept by Dr Rutty, in Dublin, for 41 years, Mr Kirwan endeavoured to calculate the probabilities of particular seasons being followed by others. Though his rules relate chiefly to the climate of Ireland, yet as probably there is not much difference between that island and Britain in the general appearance of the seasons, we shall mention his conclusions here.

In 41 years there were 6 wet springs, 22 dry, and 13 variable; 20 wet summers, 16 dry, and 5 variable; 11 wet autumns, 11 dry, and 19 variable. A season, according to Mr Kirwan, is counted wet when it contains two wet months. In general the quantity of rain which falls in dry seasons is less than five inches, in wet seasons more: variable seasons are those in which there falls between 30lb. and 36lb. a lb. being equal to .175639 of an inch.

The order in which the different seasons followed each other was as in the following table:

| Season | Times | Probability | |-------------------------|-------|-------------| | A dry spring | | | | A wet spring | | | | A variable spring | | | | A dry summer | | | | A wet summer | | | | A variable summer | | | | A dry spring and dry summer | | | | A dry spring and wet summer | | | | A wet spring and dry summer | | | | A wet spring and wet summer | | | | A wet spring and variable summer | | | | A dry spring and variable summer | | |

Hence Mr Kirwan deduced the probability of the kind of seasons which would follow others. This probability is expressed in the last column of the table, and is to be understood in this manner: The probability that a dry summer will follow a dry spring is $\frac{1}{2}$; that a wet summer will follow a dry spring $\frac{1}{3}$; that a variable summer will follow a dry spring $\frac{1}{4}$; and so on.

This method of Mr Kirwan, if there is such a connection between the different seasons that a particular kind of weather in one has a tendency to produce a particular kind of weather in the next, as it is reasonable to expect from theory, may in time, by multiplying observations, come to a great degree of accuracy, and may at last, perhaps, lead to that great desideratum, a rational theory of the weather. As we wish to throw as much light as possible on this important subject, we shall add to these a few maxims, the truth of which have either been confirmed by long observation, or which the knowledge we have already acquired of the causes of the weather has established on tolerably good grounds.

1. A moist autumn with a mild winter is generally followed by a cold and dry spring, which greatly retards vegetation.—Such was the year 1741.

2. If the summer be remarkably rainy, it is probable that the ensuing winter will be severe; for the unusual evaporation will have carried off the heat of the earth. Wet summers are generally attended with an unusual quantity of seed on the white thorn and dog-rose bushes. Hence the unusual fruitfulness of these shrubs is a sign of a severe winter.

3. The appearance of cranes and birds of passage early in autumn announces a very severe winter; for it is a sign that it has already begun in the northern countries.

4. When it rains plentifully in May, it will rain but little in September, and vice versa.

5. When the wind is south-west during summer or autumn, and the temperature of the air unusually cold for the season, both to the feeling and the thermometer, with a low barometer, much rain is to be expected.

6. Violent temperatures, as storms or great rains, produce a sort of crisis in the atmosphere, which produces a violent constant temperature, good or bad, for some months.

7. A rainy winter predicts a sterl year.—A severe autumn announces a windy winter.

Thus we have endeavoured to describe the various phenomena of the weather, and to explain them as far as the infant state of our knowledge of the atmosphere furnished us with principles.

Notwithstanding the imperfection of our present knowledge of this subject, the numbers and the abilities of the philosophers who are at present engaged in the study cannot fail at least of being crowned with success; and perhaps a rational and satisfactory theory of the weather is not so far distant as we at present suppose. It is a pity, however, that in a science attended with so much difficulty as meteorology is, various artificial difficulties should have been thrown in the way, which contribute very much to obstruct its progress. There are no fewer than four thermometers. Weather meters used at present in different parts of Europe; and the observations made by each of them must be reduced to one common standard before it is possible to compare them with one another. This is a tedious enough business, but it is nothing at all to the reduction of observations of rain and of the barometer to one common standard. Every nation has its own peculiar measure; and the French, to add to the difficulty, have reckoned by lines, and twelfths of lines, instead of by decimal parts of an inch. Whether, however, this be the case at present or not, we know not, as we have seen no meteorological tables drawn up in France later than 1792.

Philosophers ought certainly to fix upon some common standard of weights and measures, otherwise the labour in meteorology, and even in chemistry, must soon become intolerable. The only other possible way to remedy this evil would be, to construct accurate tables, in which the various weights and measures used by philosophers are reduced to one common standard. This has already been done in part; but no table of this kind which we have seen is sufficient to remedy the evil: few of them descend to decimal parts of small weights or measures; yet without this they seldom can save the trouble of calculation.