on the Rigi, and on the Faulhorn; the observations of the hours 10 and 17 were found by interpolation.*

The observations made in 1841 and 1842 by MM. BRAVAIS, WACHSMUTH, PELTIER, and myself, on the Faulhorn, assign to the barometric variation a diurnal range, differing in many respects from that obtained by the author on the same mountain. In order that the reader may compare, at a glance, these several series of observations, we will give the table of them, as constructed by M. BRAVAIS. The numbers followed by two points (:) were obtained by interpolation.

The first of these five series of observations does not accord with the following four; the latter agree with each other, and establish the existence of a maximum about 10 P.M., and a minimum about 6 A.M.; it appears, moreover, that the maximum of 10 A.M. has receded to three in the evening; and the minimum that follows it about 5 P.M. is very feeble; the slightest disturbance is sufficient to cause this retrograde range of the barometer to disappear. As the last four series are based on 106 days of observations made under varied circumstances, while the former series is the mean of twenty-five days, we must place less confidence in this series. Moreover, the general mean of the five series, such as we give it in the last column, cannot be very far from the truth. New observations are necessary in order to dissipate all the doubts that may remain as to the reality of the fall of the barometer between 3 and 5 P.M., at a vertical height of 2700 metres above the level of the sea.-M.

20

4,41 4,13 10,28

2,13

7,36

4,77

4,38 4,16 10,22 2,20

4,24

4,29 4,23 10,06 2,12 7,89 23 4,19 4,34 9,86 1,87 7,99 3,97

(Vide Appendix, fig. 26.)

These correspondent observations shew that the laws of diurnal variation change as we ascend in the atmosphere. At Zurich, the barometer falls from mid-day to 5 P.M.; we find an analogous change on the Rigi; but the oscillations are smaller, and the difference between the two barometers continues diminishing, and attains its maximum between three and four o'clock; the two barometers then rise, but that of Zurich much more than that of the Rigi, and the difference between the two barometers increases. During the night

the atmospheric pressure diminishes uniformly at the two stations, and the difference between the two barometers remains constant. However, the morning minimum falls at Zurich between three and four o'clock, and on the Rigi between five and six: the barometer then rises much more at Zurich than on the Rigi; at Zurich, the maximum is at about eight o'clock in the morning; on the Rigi the column rises uninterruptedly until about mid-day. The corresponding observations at the Faulhorn lead to the same results; the tropical hours differ; and while, in the evening, the difference between the two barometers of the mountain and the valley is 173mm,02, it is 174mm,82 in the morning.

In order to appreciate the influence of height, let us still consider the mean barometric oscillation; let us look for the two minima and the two maxima of Zurich, and deduct the mean minimum from the mean maximum. Let us then consider what was the barometric height of the upper station at the tropical hours of the lower, and subtract the mean from the hour of the maxima. The correspondents of Zurich and the Faulhorn will serve us to construct the following table :

HOURS OF THE SMALLEST AND GREATEST HEIGHT OF THE BAROMETER AT ZURICH AND ON THE RIGI.

At Zurich, the difference between the mean maxima and minima is 0mm,644; on the Rigi, it is only 0mm,237. Simultaneous observations at Geneva and Zurich give us the mean diurnal oscillation, Om,897; whilst on the Faulhorn the corresponding difference was only 0mm,268. Thus, at a certain elevation above the level of the sea, the diurnal oscillation would be null. In 1833, during my stay on the Faulhorn, the weather was constantly very bad; the mean diurnal oscillation at Berne, Basle, Geneva, and Zurich, was 0,656; on the Faulhorn, the corresponding difference was 0mm,178: whilst, in the plain, the barometer fell from

the maximum of nine o'clock in the morning, to the minimum of three o'clock in the evening, about 0mm,68; it rose on the Faulhorn from 550mm,83 to 551mm,12, that is, about Omm,29; thus, then, the phenomenon was the reverse on the mountain. Whilst on the plain the barometer generally falls during the day, it does the contrary on an elevated summit.

If we examine the mean diurnal oscillation at different points of the terrestrial surface, we shall have a correction to make in order to reduce these points to the level of the sea. The observations made on the Alps supply us with the elements of this correction. Let us admit that on the seashore the barometer is at 761mm,33; let D be the mean oscillation: let us ascend without changing latitude, and suppose that the barometer is at b millimetres below 761mm,35, the diurnal variation will be d, and we shall have the equation

d=D-a. b.

a being a coefficient to be determined by observation, the series made in 1832 on the Rigi, compared with that of Zurich, gives us the value of a

a=0,003694;

that made in 1833 on the same summit, compared with those of Basle, Berne, and Zurich, gives

that of the Faulhorn, in 1832, compared with Geneva and Zurich, gives

a = 0,003674;

finally, the series of 1833 on the same mountain, compared with those of Basle, Berne, Geneva, and Zurich, gives

a = 0,002758;

a series made by Buckwalder on the Sentis, and by Horner at Zurich,

a=0,003630;

the observations of de Saussure on the Col du Géant, compared with those of Geneva and Chaumouni,

a = 0,004053;

those made in the winter by Eschmann on the Rigi, corresponding to those at Zurich, give

a = 0,002856;

We admit that in the Alps the value of this coefficient is

a=0,003507.

From a comparison of series made at Halle, Dresden, Jena, Prague, Zittau, Gotha, Freyberg, and Altenberg, we deduce

a=0,003628;

those of many points in tropical America give

a = 0,002441.

we shall therefore adopt the mean of all these values, namely,

Thus, then, in order to reduce the diurnal observations to the level of the sea, we shall first seek the difference between the mean maxima and minima; we then multiply the number of millimetres which the barometer is below 761mm,35 by 0,003413, and we add this product to the difference found.*

MEAN DIURNAL VARIATION AT DIFFERENT

LATITUDES.—The following table gives the size of the

The process employed by the author, for correcting the value of the diurnal oscillation, does not appear to be entirely free from objection: dis the diurnal variation on the mountain, and D the diurnal variation on the sea-shore; let us, in addition, take H for the mean height of the barometer on the mountain, expressed in millimetres. M. KAEMTZ establishes the following relation :

d D 0,003413 (760-H).

The factor (760H) is represented by b in M. KAEMTZ's calculation. But what do d and D represent? We must not mistake this: these quantities represent the mean rise from 4 to 10 o'clock, in the higher and lower stations. Take, for example, the barometric observations at the Faulhorn (vide the table in the preceding note); we have

Therefore D = 0,285 +0,70 = 0,985.

This is the value of the oscillation at the sea-shore, deduced from the formula. But, instead of operating thus, M. KAEMTZ takes for the value of d the excess of the mean maximum over the mean minimum; so that, in the case before us, according to M. KAEMTZ's rule, we should have found

Thus, after all, if we wish to employ the mode of reduction proposed by