mean temperature of the year, because two or three observations are sufficient for knowing this mean temperature. However, we should not forget a remark first made by Wahlenberg. Having observed in the neighbourhood of Upsal a great number of springs, some of which preserved a constant temperature, whilst that of others was variable, he found that, at a mean, the constant springs had a higher temperature than the others. This is, because they come from a great depth. Experiments made on Artesian wells shew, in the most evident manner, that temperature increases with depth. The water of some of these wells, and of almost all mineral springs that are not thermals, gives a higher degree of temperature than that which corresponds to the depth of their reservoir: hence, it is often very difficult to decide whether a spring may be made use of in meteorological researches.*

Roebuck was probably the first to advise the observation of springs in order to obtain the mean temperature of a place. He observed that those of London and Edinburgh have a temperature that very closely approaches that of the annual mean. After him, John Hunter again directed attention to this subject. But it is principally the observations of MM. de Humboldt, Wahlenberg, de Buch, Erman, and Kupffer, made in almost all parts of the world, which have demonstrated the interest of researches of this kind.

The experiments of M. WALFERDIN, in the Artesian wells of the Paris basin, are contrary to the opinion of M. KAEMTZ, who insists that the temperature of the waters playing from an Artesian well are higher than that which they ought to have in respect to the depth of the reservoir. M. WALFERDIN made use of his inverting thermometer. These instruments, of which a description and figure are given in M. POUILLET'S Elements de Physique, t. ii. p. 507, and fig. 366, were protected from pressure; and the author always employed several simultaneously. Their agreement, which is often marvellous, is a guarantee to the correctness of the results. The following are instances:

To deduce the law of the increase of temperature in proportion to the depth, the constant temperature of 11°,7, which is that of a thermometer placed 28 metres deep, in the cellars of the Observatory, has been taken as a

The differences that are found between the temperature of springs and the mean of the years are due to the climacteric conditions peculiar to each locality. In Western Europe there is an equality; in Western Norway, on the contrary, the springs appear to be a little colder than the air. In proportion as we recede from the sea-shore, in that part of Europe which is on the north of the chain of the Alps, the springs are hotter than the air; and the difference is greater as we penetrate more deeply into the interior of the Continent. In almost all Italy, and between the tropics, the springs are colder than the mean of the air.

In coun

M. de Buch was the first to explain these apparent anomalies, by attending to the mode of the formation of springs. If it never rained, the soil would have, at a certain depth, the mean temperature of the air; if the same quantity of rain fell every month, and if we admit that this rain is at the temperature of the air, the mean of the springs would be equal to that of the air. This is the case in England, where as much rain falls in winter as in summer. tries, on the contrary, where the summer rains exceed those of the winter, the mean temperature of the water that falls is higher than that of the air, and the springs are in the same condition. So, also, in Sweden and Germany, the springs are many degrees warmer than the annual mean. The contrary occurs in countries where it rains much in winter, as Norway and Italy. In tropical countries, the temperature falls rapidly at the commencement of the rainy season; but, in the localities where it rains in intervals throughout the year, there is an identity between the heat of the springs and that of the air.

WITH THE

DECREASE OF TEMPERATURE HEIGHT. In proportion as we ascend a mountain, it is

starting point. Thus it is proved that in the chalk strata, which forms the lower part of the Paris basin, the temperature increases 1° for every 31 or 32 metres. The last of these numbers, obtained by MM. ARAGO and WALFERDIN, in experiments made by them at the well of Grenelle with most minute carefulness, has yet been disputed by philosophers; but it was easy to shew its accuracy, since the water flows to the surface of the soil. We know, indeed, that it comes from a depth of 548 metres. If the notation of 26°,43 established for 505 metres, and the law of 1° of increase for every 32,3 metres, which has been deduced from it, be correct, we must find, from this latter datum for 43 metres, the difference between the depth of 505 and 548 metres, 1o,33, which, when added to 26°,43, obtained for 505 metres, makes 27°,76. Now, the water that comes to the surface has a temperature of from 27°,65 to 27°,70, and this minimum difference corresponds, as M. WALFERDIN was assured of by experiments of another kind, to the diminution of temperature that flowing water experiences in ascending from a depth of 548 metres, to the upper orifice of the well.

We see that it was difficult to find a more striking agreement between the latter temperatures obtained before the starting forth and those of the water that now runs on the surface of the soil.-M.

found that the temperature falls. Cases may undoubtedly occur in which this fall is nothing, or in which it is even warmer above than below; but these exceptions are rare, and are to be traced to the direction of the winds and to the season. Sometimes, indeed, warm south winds prevail above whilst the north wind blows on the plain. To know the laws of the decrease of temperature with the height, we must take the mean of a great many observations. The law according to which temperature decreases, as to the limits of the atmosphere, is yet unknown; however, within the limits that have hitherto been examined, we commit no great error in admitting that the same differences of level correspond to the same differences of temperature. If, then, we know the first of these quantities we shall divide it by the second, and the quotient will indicate the number of metres that we must ascend in order that the temperature may fall one degree.

Long series of correspondent observations, made at great differences of level, shew that this decrease varies with the season and with the hour of the day. The observations that de Saussure continued for seventeen days, on the Col du Géant, 3428 metres above the sea, whilst simultaneous observations were being made at Geneva (407 metres), and at Chamouni (1044 metres), have made the horal influence evident. According to the observations of de Saussure, and those which I made on the Rigi (1810 metres), while observations were being made at Basle, at Berne (548 metres), at Geneva, and at Zurich (459 metres), the following are the heights in metres which we must ascend in order to obtain a decrease of one degree :

DIFFERENCE OF LEVEL, CORRESPONDING TO A FALL OF 1° OF THE THERMOMETER, AT ALL HOURS OF THE DAY.

De Saussure made observations during the night; being alone, I could not read the barometer longer than from five in the morning till ten in the evening, and the laws of nocturnal decrease are deduced from those of the day. Although these tables present a few anomalies, yet they render the diurnal period clearly evident. About five in the evening is the time when the decrease of temperature is most rapid, and towards sunrise it is slowest. The difference that corresponds to these two times, deduced from observations, is equal to about the third of the mean height that we must ascend in order to obtain a fall of one degree. The difference of the two means, 164,7 and 149,1, is derived from the differences that the meteorological phenomena presented in the course of the two series.*

* If the diurnal variation of the thermometer followed the same laws on mountains and on plains, the two thermometers would rise and fall simultaneously; and their ranges, remaining parallel, the difference of their indications would be constant. The decrease of temperature would not vary according to the hours of the day. But, if the two diurnal curves of the temperature of the two stations are constructed graphically (taking the time as the abscissa), we immediately recognise the want of parallelism. In the following table M. BRAVAIS has chosen for his lower station the mean of corresponding observations made at Milan, Geneva, and Zurich. The upper

The annual period is not less marked in our climates; the simultaneous meteorological series made at Geneva and on

station was on the summit of the Faulhorn, 2673 metres, in the canton of Berne. The observations lasted forty-four days; their mean epoch corresponds to August 12, 1841.

DECREASE OF TEMPERATURE BETWEEN MILAN, GENEVA, ZURICH, AND

During winter the results appear to be somewhat different. In January 1827, M. ESCHMANN remained for eleven days on the Rigi, while HORNER was making observations 1370 metres lower. The following are the hourly

means:

DECREASE OF TEMPERATURE BETWEEN ZURICH AND THE RIGI, IN WINTER.

The differences of temperature are more feeble during the night than by day, which is precisely the reverse of what happens in summer. The habitually hazy state of the lower strata during winter may account for this difference. Moreover, the different terms, the means of which are represented by the figures in the column marked difference, are very irregular; the thermic state