was the case in January, 1855, as I shall proceed to show.

If a barometrical minimum in the temperate zones be compared to a long valley, a maximum to a mountain ridge, it will sometimes happen, if the area of observation be not too widely extended, that we shall find ourselves on the slope of the hill. A steeper slope than that which the observations on New Year's Day of the above named year would represent for the centre of Europe, is not often met with.

The following table gives the depression of the barometer, in inches, below its mean level for the month, on January 1:—

At Geneva there was an excess of 0.086 inches; at St. Bernard, of 0.097; at Lyons, of 0.264; and at Lisbon 0.388; the maximum during the month. In the vicinity of New York the barometer stood 0.53 inches above its mean level; at Norfolk and Charleston 0.27, while at San Francisco its level was unusually low (29-13 inches), and there was a hurricane from the SW. At Benicia there was a similar SW. storm, and the barometer stood nearly 0.62 inches below its mean level.

Throughout the whole of the Prussian and Austrian districts of observation (in the latter of which the minimum occurred, not on the 1st but on the 2nd, at all stations excepting Bodenbach, Prague, Pilsen, and Czernowitz), the year 1855 set in with a relatively high temperature, high winds from the W. and NW., and heavy falls of rain and snow. In Berlin the character of the showers was precisely that of violent thunder storms, and I expected to observe thunder and lightning every moment. They were reported to have occurred at many places between Silesia and Hamburg. Part of the island Wangeroge was washed away by the violence of the sea, and the strongest embankments along the north coast of Germany were hardly able to withstand the fury of the waves. The storm was at its greatest height at Vienna at 9 A.M.; at Berlin at about noon. In the forest of Lambach, near Kremsmünster, where the storm was most violent at 2.15 P.M., on the 1st, 30,000 trees were blown down in a space of about 1,000 acres, 1,200 close to the observatory in a space of 300 acres. In Jaslo the roof of the court house was stripped off on the morning of the 2nd, and at Trautenau, on the southern slope of the

Riesengebirge, men and laden wagons were blown down. In Zara there was a calm.* The southern current had previously predominated over the whole area to such a degree that the barometrical mean of December had been beneath its value for seven years, from 1848 to 1854, 0-308 inches at Arys; 0.387 at Königsberg; 0.296 at Stettin; 0.309 at Berlin; 0-209 at Gütersloh; and 0.178 at Cologne. Its depression below the mean level for several years had been 0.214 inches at Cracow; 0·194 at Vienna; 0.220 at Prague; 0.236 at Kremsmünster; 0.168 at Milan. This is remarkable, inasmuch as the barometrical mean for the month of December had already been 0.178 inches too low.

A barometrical difference of 2·04 inches between Upsala and Lisbon would produce a difference of level on a water surface of 27.7 inches. It was only natural to expect, as really was the case, that the air from the area in which the pressure was not only not diminished, but even increased, would press in with great violence in order to fill the vacuum which existed over the Baltic. This cold and densé current became so universally prevalent

*The hourly observations give the following results at some of the stations:

The figure 10 indicates a hurricane, the highest point of the scale of force of the wind.

that the temperature of the latter half of January and of February was unusually low.

I must refer to my 'Representation of the Phenomena of Temperature by the aid of Five-day Means,' to show the distribution of this extreme cold, and the gradual restoration of the equilibrium of the barometrical pressure.

On the very day on which this cold current forced its way into the warm atmosphere of Europe, a southerly current forced its way to the northwards in California. Between the two barometrical minima, there is interposed a district where there was a barometrical maximum, extending over the whole Atlantic ocean and embracing the eastern coasts of the United States.

A letter from Greenland says:-'We had such fine 'warm days in February (which was desperately cold in Europe) and March, that we were almost tempted to ask if we were really in Greenland.'

6

6

An unprejudiced examination of the facts which have been stated in the preceding pages leads us to the conclusion that there are two different types of motion of the air in a storm: firstly, when its direction does not change, excepting in so far as it is modified in the course of its advance by the rotation of the earth; and secondly, when it possesses a rotatory motion about a calm centre which is advancing. It surely cannot be denied that violent storms may be produced by the sudden irruption of the cold air of the polar current into the warm and rarefied air of an equatorial current*, or that the return trade wind, confined between meridians which are gradually

First rise after very low
Indicates a stronger blow.

closing in, bursts open for itself, in its fury, a channel to the north. Many of the American snowstorms appear to belong to the first of these two classes of storms, but only a very few of these have been investigated. Henry (Agricultural Report,' 1858, p. 484) has recently given in his adhesion to this view.

In order to determine with accuracy the true character of a storm in any individual instance, it is in general requisite that we should obtain simultaneous observations from an area of considerable extent. The reason of this is that the gyration of the vane on one side of a cyclone is the same as that produced by the ordinary currents of the atmosphere, in accordance with the law of gyration. Redfield attempted to illustrate the origin of the rotatory motion of considerable masses of air, by attaching a circular disk of paper loosely, at one point, to a globe which he set in rotation. If the disk be placed near the pole it will be carried with a less velocity than if it be placed near the equator, and thus a tendency to rotatory motion will be imparted to it, which will be in the direction opposite to that of the hands of a watch, when the disk is placed anywhere on the northern hemisphere.

It is easy to see that the cause of this phenomenon is the cohesion of a solid body; a condition which cannot be applied to the particles of the atmosphere, which possess perfect freedom of motion. Taylor (Herschel's 'Meteo'rology,' p. 67) attempts to prove that if the air rises from any point, and the air about that point flows in from all sides towards it, the result will be to produce, not as Espy and Hare assert, a direct centripetal confluence, but a whirlwind in the sense of the cyclone theory. Ferrel also agrees with this view (p. 43).

We may admit that the conditions which predispose the air to rotatory motion may be very various, but the