Such a peculiarity as this may almost deserve to be spoken of in the terms applied by Sir J. Herschel to the distribution of land and water upon our earth, it is 'massive enough to call for mention as an astronomical feature.' I propose to examine two theories which have been suggested in explanation of this feature of the earth's envelope. These theories are founded on local peculiarities, and the feature considered appears as a dynamical one-that is, as a peculiarity resulting from states of motion in the aerial envelope. I shall endeavour to establish a theory founded on a consideration of the earth's mass as a whole, and presenting the atmospheric feature in question as a statical one.

The first theory I have to notice is one founded on the configuration of land and water upon the northern and southern hemispheres of the earth's globe. In the northern hemisphere, and more especially in that part of the northern hemisphere in which barometrical observations have been most persistently and systematically conducted, there is much more land than in the southern hemisphere. Now barometrical observations are referred to the sea-level, and observations made in Europe and America may be considered as referred to the level of the northern parts of the Atlantic Ocean. It is argued that the North Atlantic, compared with southern oceans, is little more than 'a large lake, having elevated banks east and west.' ' Practically, the air there is a portion of the solid globe, so that the uncon-fined air will rest upon and rise above the former, as if it were solid and a portion of the earth; so that the altitude of the air over the North Atlantic will be increased some hundreds of feet, and the barometer at the sea-level will be pressed upon, not only by the free air clear of the earth's banks, but also by the air confined in the basin, much as if the air were at the bottom of a mine.' 1

Presented in the above form, the theory that the higher northern barometer is due to the contour of the northern hemisphere scarcely deserves serious comment.

1 From a letter addressed to the editor of the Athenaum by Dr. H. Muirhead.

To speak of the confined air of the North Atlantic Ocean is surely unreasonable. An ocean 2,000 miles across, swept by more frequent storms than are experienced in any other part of the globe, cannot be very aptly compared to 'the bottom of a mine.' An inelastic fluid flowing steadily over a rugged surface shows no trace, or but the slightest trace, of the nature of that surface, by any variations of its own level. But it is still less conceivable that an elastic fluid should be influenced in the manner suggested. In fact, if this happened, we should no longer be enabled to determine the heights of mountains by barometric observations; for according to the theory the air should extend to a greater height above mountains than above plains; and as regards comparison between a barometer at the foot of a mountain and one at the summit, we might argue that the barometer in the valley, compared with a barometer at the same level in a plane district, 'is pressed upon, not only by the air clear of the mountain tops, but also by the air confined within the valley,' so that the altitude over the valley is greater by some hundreds of feet than the altitude over a plain at the same level as the valley; and thus, before we could determine the height of the mountain above the level of the plain, we should have to determine the exact effects due to the confinement of the air in the valley. We know that, on the contrary, the average barometric pressure in the most confined valley does not differ appreciably from the average pressure over the most widely extended plain at the same level.

We may, however, reasonably inquire whether the presence of continents in the northern hemisphere might not operate in another manner. If we place any mass within a vessel containing fluid, it is clear that we increase the fluid pressure over every point of the vessel's bottom, because this pressure depends wholly on the depth of the bottom below the level of the fluid, and the level rises when any solid substance is placed within the vessel. Now if we suppose a globe covered all over by water to be surrounded by a perfectly uniform atmospheric envelope, the mean pressure of this envelope at the water-level would certainly be increased if continents were supposed to be raised in any manner above the surface of the water; and if the atmosphere over one half of such a globe were supposed to be prevented in any way from mixing freely with the atmosphere over the other half, then it is clear that the mean pressure at the water-level would be greatest on that half-globe over which the most extensive and highest continents had been raised. On the assumption, then, of some such arrangement over our own earth-an arrangement, that is, which should prevent the northern air from mixing with the southern - one might see in the northern continents a true cause of increased barometric pressure at the sea-level of the northern hemisphere.

We have, however, not only no evidence that such an arrangement exists, but very strong evidence of an atmospheric circulation which carries air from hemisphere to hemisphere, and mixes in the most intimate manner the whole mass of gases which form the earth's atmospheric envelope. The whole question of the circulation of the air is investigated in Maury's interesting work on the Physical Geography of the Sea, and he appears to establish in the most convincing manner the interchange of air between the northern and southern hemispheres.

And even if we could assume that the atmospheric covering of any portion of the earth's surface was in any way prevented from passing freely to other regions, yet the cause assigned would be inadequate to account for the difference of barometric pressure actually existing between the two hemispheres. All the land above the sea-level in the northern hemisphere, if uniformly distributed over the surface of that hemisphere, would be raised to a height of less than 200 feet above the present sea-level, and the actual difference of level corresponding to the observed difference of barometric pressure is more than four times as great.