hourly variation of the barometer each day is well marked, the following are the extreme variations from the daily mean pressure, with the periods of their occurrence in the four months of the year which represent the seasons. They are the mean of the three years, 1862-64 :

Thus the greatest maximum occurs at Calcutta at 9.30 A.M., and the greatest minimum at 4.30 P.M.; the least maximun at 10.30 P.M., and the smallest minimum at 3.30 A.M. Hence the two maxima occur when the temperature is about the mean of the day, and the minima when it is at the highest and the lowest respectively thus suggesting a connection. between the daily barometric oscillation and the daily march of temperature.

65. The surface of the globe is always divided into a day and night hemisphere, separated by a great circle which revolves with the sun from east to west in twenty-four hours. Hence these two hemispheres are in direct contrast to each other in respect of heat, and consequently evaporation, which depends on the temperature. The hemisphere exposed to the sun is warm, and the hemisphere turned in the other direction is cold. Owing to the short time of each revolution, the period of greatest heat is not at noon, when the sun is in the meridian, but about two or three hours thereafter; similarly, the period of greatest cold occurs about four in the morning. As the hemisphere under the sun's rays becomes heated, the air, expanding upwards and outwards, flows over upon the other hemisphere where the air is colder and denser.

There thus revolves round the globe from day to day a wave of heat, from the crest of which air is constantly flowing towards the meridian of greatest cold on the opposite side of the globe.

66. The barometer is influenced to a large extent by the elastic force of the vapour of water invisibly suspended in the atmosphere, in the same way as it is influenced by the dry air (oxygen and nitrogen). But it is probable that the vapour of water exerts a pressure on the barometer in another way. Vapour tends to diffuse itself equally through the air; but, as the particles of air offer an obstruction to the watery particles, the vapour is accumulated or pent up in the lower stratum of the atmosphere about 9 or 10 A.M., when evaporation is most rapid, and, being impeded in its ascent, its elastic force is increased by the reaction, and the barometer consequently rises. When the air falls below the temperature of the dew-point, part of its moisture is deposited in dew; and since some time must elapse before the vapour of the upper strata can diffuse itself downwards to supply the deficiency, the barometer falls-most markedly at 10 P.M., when the deposition of dew is greatest.

67. Hence, as regards temperature, the barometer is subject to a maximum and minimum pressure each day,—the maximum occurring at the period of greatest cold, and the minimum at the period of greatest heat. And as regards vapour in the atmosphere, the barometer is subject to two maxima and minima of pressure-the maxima occurring, the first at 10 A.M., when, owing to the rapid evaporation, the accumulation of vapour near the surface is greatest, and the second about sunset, or just before dew begins to be deposited, when the absolute amount of vapour in the atmosphere is greatest; and the minima in the evening, when the deposition of dew is greatest, and before sunrise, when evaporation and the absolute quantity of vapour in the atmosphere is least.

68. Thus, taking both causes into consideration, the maximum in the forenoon is brought about by the rapid evapora

tion arising from the dryness of the air and the increasing temperature, together with the overflow of air in the upper regions of the atmosphere from the wave of heat which has been going on for some hours. But as the vapour becomes more equally diffused and the air more saturated, evaporation proceeds more languidly; the air becomes also more expanded by the heat, and flows away to meet the diurnal wave of cold advancing from the eastwards. Hence the pressure falls to the afternoon minimum. From this time the temperature declines, the air approaches more nearly the point of saturation, and the pressure being further increased by accessions of air from the warm wave, now considerably to the westward, the evening,maximum is attained. As the deposition. of dew proceeds, the air becomes drier, the elastic pressure of the vapour is greatly diminished, and the pressure falls to a second minimum about 4 A.M.

69. If the pressure on the barometer due to that of the aqueous vapour of the atmosphere be subtracted from the whole pressure, the pressure of the dry air which remains shows the smaller maximum and minimum less decided. This peculiarity is the more strongly marked the farther any place is from the sea and the higher the mean temperature— that is, at those places where the influence of temperature is most felt and its effects on the daily barometric pressure least disturbed by large evaporating surfaces, or by moist winds. Thus the smaller maximum and minimum are apparent in the curve of dry air at Calcutta ; at St Petersburg they are very faintly marked; whilst at Nertchinsk, in Siberia, they have totally disappeared. At Nertchinsk the curve of dry air for the mean of July 1861 and 1862 attained the maximum 27.050 inches at 5 A.M., whence it fell steadily to the minimum 27.020 inches, and then rose uninterruptedly to the maximum.

70. The amount of the daily barometric variations, as the accompanying table will show, diminishes from the equator towards either pole, for the obvious reason that they depend, directly or indirectly, on the heating power of the sun's rays:

TABLE SHOWING THE DAILY VARIATIONS AND RANGE OF THE BAROMETER IN DIFFERENT LATITUDES.

It is

Thus, while at the equator the daily fluctuation is 0.125 inch, in Great Britain it is only a sixth part of that amount. very small in the high latitudes of St Petersburg and Bossekop; and in still higher latitudes, at that period of the year when there is no alternation of day and night, the diurnal variation probably does not occur.

71. Since the whole column of the atmosphere, from the sea-level upwards, expands during the heat of the day, thus lifting a portion of it above all places at higher levels, it is evident that the afternoon minimum at high stations will be less than at lower stations, especially when the ascent from the one to the other is abrupt. Thus, at Padua, in Italy, the afternoon minimum is 0.014 inch, but at Great St Bernard it is only 0.003 inch. On the other hand, the cooling and consequently condensing of the air during the cold of night lowers a portion of the atmosphere below the level of St Bernard, and hence its minimum when the temperature is at the minimum of the day falls 0.010 inch; whereas at Padua the same minimum amounts only to 0.004 inch.

72. The daily variation is less in winter when the temperature is low, than it is in summer, unless when rains in summer occur to diminish it. Thus, at St Petersburg, the variation in

January is only 0.009 inch, whereas in July it is 0.023 inch. At Nertchinsk it is 0.028 inch in January, but in July it is 0.044 inch. Rains, as just stated, diminish it; thus, at Calcutta it is 0.127 inch, but in the rainy season in July it is only 0.093 inch, and these proportions were observed in each of the years I have examined. It will be also observed that it is much larger in the dry climate of Nertchinsk than it is at St Petersburg. In the generally wet climate of Great Britain there appears, from Glaisher's Table for Greenwich, to be little difference between the winter and the summer months.

73. The mean pressure at Calcutta of the three Januarys (1861-63), as determined by observations made every hour, was 30.009 inches. At the same place and for the same time, the mean at 9 A.M. and 9 P.M. was 30.046 inches. Hence, if observations had been made at these hours only, the mean, deduced from them, would have been 0.037 inch too high. It is one of the chief uses of hourly observations to find from them the corrections to be applied to observations made at the different hours of the day, in order to reduce them to the true Such corrections, when found, serve an important end, especially in tropical regions where the daily variation is great; but they may also be, nay, frequently are, applied in cases where they are not applicable, thus leading to a good deal of confusion.

mean.

74. The legitimate use of these corrections is in their application to long averages when it is necessary to compare together the atmospheric pressure in different regions which differ in the hours of observation, or in the amounts of the variations at the same hours. Their illegitimate use, which should be discouraged, is in applying them to the means of single months or of single years, and publishing the results as the true means, without note or comment of what has been done. Thus, at St Petersburg for July, the correction deduced from the hourly observations of three years is +0.002. Applying this correction to each of the means of the 9 A.M. and 9 P.M. observations of the three Julys, we obtain as the "corrected" means, 29.906, 29.885, and 29.798 inches. The