Very large hailstones, weighing several ounces, have been sometimes observed. Some have been picked up in the Orkneys as large as goose eggs, but each of these very large stones is a mass of smaller ones which have come together during their fall.

Stories of hailstones weighing pounds, or even as large as an elephant, like that reported to have come down near Seringapatam, in Tippoo Sahib's reign, are, of course, ludicrous. The germ of truth there is in them arises from the possibility of a number of stones falling successively into a hole like a well, and freezing into a single mass.

The formation of hailstones, however, whatever be their size, is very difficult to understand. No sufficient explanation has yet been offered of their structure, or of the agency which enables them to remain suspended in the atmosphere before they fall.

In the explanation already given of glazed frost, the occasional presence in the atmosphere of water cooled below its freezing-point has been mentioned. Liquids in such a condition solidify instantly on agitation or contact with a foreign body. We may, therefore, suppose the origin of the initial hail particle to be a drop of water, in a state of superfusion, caused to congeal by some sudden impulse of wind, which produces agitation. A number of such drops might easily congeal together into a single mass, but it is hardly conceivable that structures of such complexity as some of those shown above could be the product of any instantaneous action, and therefore these must have been suspended in the atmosphere for some time.

The oldest theory of the formation of hail is that of Volta, who supposed that, as hail is generally asso

L

ciated with thunder and lightning, the hail-pellets were in a state of constant oscillation between two oppositely electrified clouds. Condensation, according to him, was always going on, and the stones grew in size until they became so heavy that they broke through the lower stratum and fell to the earth.

Dove broached a different theory, which is that hailstorms are always whirlwinds, but with their axis almost horizontal instead of vertical. The air then sweeps the growing hailstone round and round from hot to cold stratum alternately; water settles on it in the hot, and is frozen in the cold, layer. Ultimately, either by its own weight or by centrifugal force, it comes outside the influence of the rotatory motion of the whirl and falls to the ground.

In these islands we, fortunately, know but little of destructive hailstorms. On the Continent we are almost everywhere met by the frequent advertisements of Hail Insurance Offices, which are comparatively rare here. This immunity of ours is due to our insular climate, and its consequent freedom from extremes of temperature.

The geographical repartition of rainfall, and its seasonal distribution in different latitudes, will be considered later.

The law of the fall in the daily period has not been very much studied, as automatic gauges are comparatively modern inventions, and the main fact which is known on the subject is that as, between the tropics, the wettest portion of the year is the hottest, so there also the early afternoon, the hottest part of the day, is the wettest.

Caldcleugh' says of Rio Janeiro that it used to be the fashion to state in invitations for the afternoon whether the guests were to assemble before or after the thunder-storm, which came on regularly at a certain

hour. In fact, in the tropical regions the clouds frequently form in the forenoon, the rain falls in the afternoon, and all the night through the sky is cloudless.

In a recent paper on the rainfall of Cherrapunji, by Professor John Eliot, read before the Meteorological Society, and printed in their 'Quarterly Journal' (vol. viii., p. 41), it is stated that, during the wet season there, about twice as much rain falls by night as by day.

In Europe the distribution of the fall has not been studied for long or for many stations. Prague is that which affords the longest series of hourly records, but material also exists, and has been discussed, for Greenwich, Vienna, Modena, St. Petersburg, Zechen, Berne, and Coimbra, as well as for some other observatories. The result shows that the law is a very complicated one. Almost all these stations exhibit three maxima and three minima in the twenty-four hours, and at most of them the absolute maximum, the wettest portion of the day, occurs at two or three o'clock in the afternoon. The figures for Berne, however, form an exception, for there the afternoon maximum entirely disappears, and is replaced by one, very strongly marked, at 10 P.M. The subject demands further study.

1 Daniell, Meteorological Essays and Observations, 1st edition, 1823, p. 335.

CHAPTER IX.

THE WIND.

WE have already discussed the conditions of the atmosphere as regards its temperature and its pressure, and we have now to consider what the immediate effects of these conditions are in disturbing its equilibrium and determining its motion.

Air in sensible motion is wind, and wind is produced by differences of pressure, which are themselves ultimately, and in the main, attributable to differences of temperature.

We must first say something about the way in which the direction and force of wind are estimated.

The direction of the wind is recorded according to the points of the compass, and for ordinary observations the estimation according to the eight principal points (N., N.E., E., S.E., &c.) is close enough. Care must always be taken to enter the wind to true bearings, for as in these islands the north end of the compass needle lies to the west of the geographical north, it is evident that all compass' bearings will be affected by this variation, as it is called, of the magnetic needle, which practically amounts to two points. The variation changes slowly from year to year; it is at present diminishing in the British Isles, and its value varies from about 19° in the south-east of England to about 25° in the north-west of Ireland.

lines of equal variation run N.N.E. and S.S.W.

The

The

mode of recording the direction numerically is given in the note.1

The instrument employed for giving the direction of the wind is the ordinary wind-vane. In using this apparatus care should be taken that the vane is long enough to be easily affected by the wind, that the bearings are kept well oiled, and that the cross, indicating the cardinal points, is set by the true meridian.

1 The compass card is divided into 32 parts called points. The number of degrees in the entire circumference being 360, the points in their numerical order, with their angular difference from North, are given in the following diagram:

From this it will be seen that a difference of a point, in the bearing of an

object, amounts to 11° 15'.