peculiarities of the curve with such completeness as to exhibit the variations impressed on the curve by season and by geographical position, particularly as regards masses of land and extended sheets of water. In lower latitudes a comparatively short time is required; but even here, where a general regularity in the phenomena is perhaps the most striking fact in the meteorology of equatorial regions, the variations which do occur from year to year ought to be carefully observed from their important bearing on the whole theory of the movements of the atmosphere. The summer months of the northern hemisphere, as regards the diurnal oscillations of the barometer, that is the period when the influence of the sun is at the maximum as regards its effects on these phenomena, are May, June, and July, and the winter months, November, December, and January, both corresponding with the sun's declination; that is, the effects are not cumulative, as in the case of the temperature of the air or that of the sea, by which the critical periods are retarded from one to two months.

Among the many interesting features of the curves which were pointed out may be noted the enormous influence of latitude and of land and sea respectively in determining the amount and time. of occurrence of the different phases of the oscillations, and a diagram was exhibited showing the curves of a large number of places, from which it appeared that as regards the summer the A.M. maximum occurs at any time from 6-7 A.M. to 2 P.M., and the P.M. minimum from 3 to 8 P.M.-the stations selected showing a regular gradation between these extremes, a gradation dependent on geographical position. The tendency of assimilation of the curves for certain elevated stations and those for strictly sea-side stations was pointed out, and attention was drawn to the striking fact that the summer curves of inland stations within lat. 30° N. and S. essentially differed from those of higher latitudes-a difference which the varying declination of the sun with season failed to obliterate.

An examination of the different theories yet propounded shows that none of them are in accordance with the facts which have been collected. It would not be difficult by a proper selection of stations to bring proof in support of any of these theories. The truth is, however, that as more facts are obtained the difficulty of framing a satisfactory theory can scarcely be said to be materially

lessened, even though it becomes easy to arrive at a close approximation to the curves of a place from its geographical position alone, before determining the curves by working them out numerically.

An examination of these curves by the harmonie analysis and a similar examination of the temperature, hygrometric, wind, magnetic, and electric curves, will in all likelihood be required before the true theory of the diurnal barometric oscillations can be stated.

3. On the Air dissolved in Sea-Water. By J. Y. Buchanan.

4. Why the Barometer does not always indicate real Vertical Pressure: A continuation of the Paper laid before the Royal Society of Edinburgh in July 1875, in which, in addition to several other points, this was attempted to be shown. It is now more fully written out. By Robert Tennent.

The barometer only indicates real pressure when the atmosphere is in a state of perfect rest. It may then be represented a existing in vertical columns, but when it moves over a resisting surface its lower surface currents will be greatly retarded, while those aloft will move comparatively free and unimpeded. In this state it may be represented as moving in columns inclined in the direction towards which it moves. The atmosphere may thus be conceived as being divided into a number of spheroidal concentric layers, each of which is possessed of a different rate of speed, and moves more rapidly than the one beneath it: an increasing amount of friction will take place betwixt the layers as they approach the surface where its influence is greatest. What takes place may be represented in this way. Let RS be the resisting surface, and let

A A', B B', C C', and D D' represent different layers of air moving

at different rates of speed. Those being most rapid where the arrow heads are most numerous. Let these layers also exhibit equal masses of air, but of different volumes, which increase in size from the ground upwards, accompanied also by much greater mobility, as shown by Tyndall in his experiments at Chamouni and the summit of Mount Blanc. An important point is this. The surface current D D has only a horizontal source of supply from the direction D' from which it is fed, while the upper current A A' has not only a horizontal source of supply from A', but it also derives supply from the slower moving current B B' beneath it, which will be drawn upwards, or "lifted," in the direction of the small inclined arrows a b c d. To enable this upper current to supply itself in this way, while it possesses the same amount of mass as that beneath it, it must have a greater velocity, and therefore greater momentum than that of the current beneath it, from which it draws its supply. Each of the lower currents, as they approach the surface, are also supplied in the same manner, but in a decreasing ratio, from those beneath them, until the lowest layer is reached, which is that which is most retarded, not only owing to its proximity to a resisting surface, but also to the scarcity of supply, which can now only be derived from a horizontal source, and not from beneath, as was the case with those above it. A tendency in the air to accumulate aloft will now take place, by "lifting," in the direction of the small inclined arrows a b c d. Pressure is thus diminished at the surface, while it is abnormally increased higher up.

The effect of this will be that the surface barometer will exhibit diminished or fictitious pressure, while the real weight of the atmosphere remains unaltered. A partial vacuum is found on the lee-side of a wall, over the top of which a strong wind blows: it withdraws the air there to such an extent, that it causes removal to exceed restoration, but it does not affect the real vertical mass of the column overhead except to an infinitesimal degree. In this, as in the former case, real pressure cannot be ascertained by the surface barometer. It is only to be obtained from the result of an observation of a series of barometers placed vertically above each other, and not very far apart. From the results thus obtained, it would be found that the normal upward diminution of pressure

which takes place when the atmosphere is at rest would be greatly altered when its upper portion is in rapid motion.

Retarded surface currents and rapid upper currents which move in inclined columns and produce these effects, can only be found with an imperfect fluid, and on a resisting surface, into which the element of friction enters. On a frictionless surface this could not take place. The atmosphere would there move in vertical, not in inclined columns, no "lifting" would take place, pressure would be real or statical; its upward diminution would be normal, as when it is at rest, and horizontal movement would not take off vertical pressure.

Barometric pressure must hence be regarded in two points of view-1st as being a cause; and 2d, as being an effect. When, as in the first case, it is real or statical, it operates as a cause due to gravitation, which is unresisted. When, as in the second instance, there is an introduction of the dynamical element, surface gravitation is diminished by "lifting," and it must then to a certain extent be regarded as an effect. The practical conclusion from this is obvious. On weather charts the constant rise and fall of the barometer, which is there reported, is to a large extent simply due to the passage of air over a resisting surface. Over a surface devoid of friction these mechanical effects would be entirely removed its rise and fall would be greatly reduced, and might be considered as being solely dependent on the effects of heat and vapour. The gradients and isobars which are represented by these movements of the barometer would consequently require also to be similarly corrected.

The barometer does not indicate the real weight of the atmosphere, it only exhibits the amount of its elasticity, from which its real weight can only be deduced when the dynamical element of motion does not enter into its currents. The two cases above mentioned may illustrate this point. The surface barometer there indicates fictitious pressure, or in other words, the amount of pressure due to the elasticity of the air, but not to that of its real weight, which is there diminished by "lifting," and as lifting can increase or diminish in amount, so also can the elasticity of the air, while its real weight remains unaltered.

As a general rule, in the British Isles, equatorial winds are

accompanied by these rapid upper movements, while polar winds move with a greater uniformity in the velocity of their various layers, and sometimes even those on the surface move more rapidly when copiously supplied from a vertical source. There is thus a remarkable difference in their mode of inflow. Equatorial winds, as they increase in force, are hence accompanied by "lifting” and a fall of the barometer. Polar winds are not attended by "lifting," and if their supply is copious and partly from a vertical source, their increase in force is accompanied by a rise of the barometer.

The range of the thermometer is equally great, both above and below its mean; but with the barometer the extent of its range above the mean is not more than one-half of that which takes place when it is below it. When it is below the mean, equatorial winds generally prevail, which are accompanied by lifting and extensive range. When above the mean, polar winds prevail, which are not attended by lifting or such extensive fluctuations. Hence, as a general rule, equatorial winds exhibit fictitious or dynamical pressure, while polar winds possess more nearly, real or statical pressure, being less accompanied by rapid upper currents, and by the mechanical oscillations due to the passage of air over a resisting surface.

Observations of a general description and illustrations will probably be afterwards introduced to exemplify the above conclusions. It is to this difference in the mode of inflow that an attempt has been made to explain the causes why depressions move in an easterly direction.

5. Laboratory Notes. By Professor Tait.

(a) On an Effect of Heat on Electro-Static Action. By means of a very delicate galvanometer, transient currents were detected when one of two plates of the same, or of different, metals, separated by a sheet of mica or glass, was suddenly heated.

(b) On Dr Blair's Scientific Aphorisms in connection with the Ultra-Mundane Particles of Le Sage.

Accident has recently called my attention to a work entitled Essays on Scientific Subjects, by Robert Blair, Regius Professor of

VOL. IX.

31