reasons why isobarometric charts have not hitherto been prepared. My leisure hours for a long time have been employed in preparing the materials for the thirteen Isobarometric Charts, for the twelve months and the year. Three of these are given in this Chapter-viz., the Charts for July, January, and the year.

85. The charts are drawn from observations made at 360 places, thus distributed over the globe-167 in Europe, 51 in Asia, 23 in Africa and adjoining islands, 63 in North America, 35 in South America, West India Islands, and Atlantic, and 21 in Australasia and the Antarctic Ocean. Of the European stations, 12 are in Scotland, 14 in England, 27 in Austria, 12 in Italy, 10 in France, 10 in the Netherlands, 9 in Norway, 57 in the Russian Empire, &c. The list might have been largely increased; thus a larger number might have been given from the Scottish stations, but the 12 given were judged sufficient to represent the mean atmospheric pressure of that country.

86. For the British Islands the means were uniformly taken for the ten years from 1857 to 1866, so that they might be strictly comparable with each other; and the means of several European places are for the same years. In the United States of America the means are uniform for the six

years from 1854 to 1859. In selecting the stations, respect was paid to the obtaining of a good mean, that is, observations for a sufficient number of years, to show as nearly as possible the true mean. At Bombay, for example, from 1847 to 1860, the lowest mean for July was 29.598 inches in 1851, and the highest 29.673 inches in 1853. This regularity in the pressure of the same month in different years is a feature common to all tropical countries, and hence, in such places, a few years were accepted as a good mean. On the other

hand, since the mean pressure at Reykjavik, Iceland, during January 1867, was 29.913 inches; during February, 29.359 inches; and during March, 30.037 inches,-it is evident that a good many years are required to represent the mean of the months. Hence, a very subordinate place, if any at all, has

been given to observations from the places at which the observations did not extend over a considerable number of years.*

87. For many means I am indebted to the labours and writings of Buys Ballot, Carl Jelinek, Dové, Quetelet, and Kuppfer, and particularly to Secchi's admirable abstracts which have appeared from time to time in the Bulletino Meteorologico.' To the Royal Society of Edinburgh and to Professor C. Piazzi Smyth, Royal Observatory, Edinburgh, I return my most cordial thanks for giving me, at all times, access to their invaluable libraries.

88. In reducing to sea-level, Table IV. was used for all places that did not exceed 800 feet in height. For higher situations the reduction was made by means of Dippe's method, as detailed in Guyot's Mathematical and Physical Tables,' D, p. 60.

89. In the charts the isobarometric lines are drawn for every tenth of an English inch in the difference of the pressure. The lines representing 30 inches of pressure and upwards are coloured red; and the lines representing pressures lower than 30 inches are coloured blue. Thus the red lines show at a glance those portions of the globe where the pressure is above the average (30 inches). In the centre of Africa, from which there are no observations, and where, consequently, the lines are hypothetical, they are made dotted lines.

90. MEAN ATMOSPHERIC PRESSURE FOR JULY.-It will be seen from Plate I., representing the mean pressure for July,

* The following are a few of the places, with the number of years, for which the means are given :-Sitka, 15; Algiers, 10; Hobart Town, 25; St Louis, Mauritius, 13; Bogoslovsk, 26; Nijni-Tagilsk, 21; Barnaul, 19; Nertchinsk, 18; Pekin; 14; Calcutta, 11; Tiflis, 14; Baku, 17; Alagir, 15; Jakobshavn, 10; Reykjavik, 13; Hammerfest, 13; St Petersburg, 19; Archangel, 18; Zlatoust, 28; Lugan, 22; Christiania, Cracow, and Kursk, 27; Brussels, 33; Gand, 26; Geneva, 25; Ahun, 34; Verona, 73; Bologna, 45; Milan, 25; Turin, 74; and most of the Austrian stations from 14 to 18, &c.

that the lowest pressures are distributed over the continents, and the larger the continental mass the greater the depression; and that the highest pressures are distributed over the ocean between 50° N. and 50° S. lat. Of these high pressures the highest occur in those parts of the ocean which are most completely enclosed by the continents.

91. Areas of Low Pressure.-The most extensive of the areas of low pressure is that broad tract which extends almost from the Gulf of Guinea in a north-easterly direction through Africa and Asia to the north of Siberia, reaching the greatest depression, 29.468 inches, near Irkutsk. A region of low pressure also occurs in North America. The observations showing this depression near its centre are scanty; but as the North American observations show clearly a decreasing pressure from the coast towards the interior of the continent till it falls to 29.686 inches at Great Salt Lake City, there can be no doubt as to its existence. The pressure is under the average in the north of South America, and thence westward across the Pacific till it join on to the low pressure of Southern Asia. Round the north pole, where, at this season, the sun never sets, the pressure is no doubt under the average.

92. In the Antarctic Ocean the pressure is abnormally low throughout the year, with perhaps a slight tendency to an increase in winter (July). This singular permanent depression has been calculated from a mean of upwards of 100,000 sea observations.

93. Areas of High Pressure.-There are two areas of high pressure of very unequal magnitude. The first includes the eastern part of North America, the Atlantic between latitudes 50° N. and 50° S., the south-west of Europe, the west and south of Africa, the greater portion of the Indian Ocean, and Australia, from which a belt of about 20° in breadth passes eastward across the Pacific and central parts of South America, where it joins the high pressure of the South Atlantic. The regions of highest pressure occur in the centre of the North Atlantic between lat. 25° and 40°, and in the South Atlantic between 10° and 35°. In these spaces the

pressure rises to a maximum of 30.348 inches at 35° N. lat., and to the same height at 20° S. lat.

94. The other region of high pressure is comparatively insignificant, stretching from California westwards across the Pacific to probably 150° E. long.

95. The effect which the peculiar distribution of the atmosphere during the summer months has on the surfacewinds and on the temperature, will be better discussed in Chapters VI. and X., which treat of these subjects. It only remains here to say a few words on the chief causes which bring about this distribution-viz., the temperature primarily, and, secondarily, the moisture of the atmosphere.

96. As the sun is nearly at the greatest distance north of the equator in July, the greater part of his heat falls on the northern hemisphere. That portion of the heat which falls on the land is wholly absorbed by the thin superficial layer exposed to the heating rays, which therefore becomes quickly heated, together with the air resting upon it. On the other hand, the portion of the solar heat that falls on the ocean is not arrested at the surface, but a considerable part of it penetrates to some depth below the surface. Owing to the low specific heat of land as compared with water, the land is more quickly and more highly heated than the ocean by the sun. Further, as the atmosphere over land is much drier than that over the sea, fewer of the solar rays being absorbed where the air is drier, more heat falls upon the land, and consequently its temperature rises higher than that of the sea. From the above causes the summer temperature of continents greatly exceeds that of the ocean in the same latitudes. An examination of Plate IV. will show the abnormally high temperature which prevails in Asia, north of Africa, and North America during summer. Thus, the temperature of each of these regions is very high, and the air being in consequence specifically lighter, it ascends, as from a furnace, in a vast column thousands of miles in diameter. In this way the summer atmospheric pressure of continents is diminished. The amount of the diminution over nearly all the continent

of Asia is about half an inch. This is perhaps the finest example that can be adduced of the enormous mechanical power of the sun's heat, which first raises, and then removes about one-sixtieth part of the whole atmosphere from so large a portion of the earth's surface.

97. What comes of the air thus removed? It is evident that the columns of heated air will continue to ascend till they reach a height at which the pressure or tension of the air surrounding the ascending currents is less than that of the aerial columns themselves. In popular language, the warm air will ascend just as long as it continues to be buoyed up by the denser air which surrounds it. Hence when it attains a height at which the surrounding air is less dense than itself, it will no longer continue to ascend, but will flow laterally over those regions which offer the least resistance to its course-in other words, it will flow into those places where at that height the density is least.

98. Now those regions where at great heights the density of the air is least, are precisely those places where the mean temperature of the whole stratum of air reckoned from the level of the sea upwards is least, for the obvious reason that cold air being dense and heavy, the greater bulk of its mass is concentrated in the lower beds of the atmosphere, thus leaving the upper beds less dense and heavy. It follows from this that the greater proportion of the air which is removed from the continents of the northern hemisphere in summer, will flow over as an upper current into the southern hemisphere, and increase the atmospheric pressure there. The massing of the isobarometric lines which indicate a pressure above the average to the south of the equator at this season, is conclusive proof of the justness of this reasoning.

99. But at this time the mean temperature of the air over the ocean is much lower than that over the land: we should, therefore, expect to find high barometers in the northern hemisphere over the ocean, and particularly in that ocean which is most completely surrounded by heated land, which may thus pour over into it, by the upper currents of the