The forest seems to act, in fact, as a breakwater, against which the violence of the storms expends itself for a time, till it can gather fresh force. Almost all the French observers are agreed as to the origin of this development and protection. We must recall the fact which we mentioned in our chapter on thunderstorms, that the windcirculation of thunderstorms is always confined to the low strata of the atmosphere only.

Hail is produced when two clouds are superimposed at a certain distance. A storm is never isolated. The ordinary French ones are often formed by partial derivations taken from the south-west winds, and occasioned by the passage of cyclones. When the thick mass of cloud which marks the existence of a storm meets a valley, the lower clouds are diverted from the general route; a portion follows more or less exactly the contours of the valley. Thus it happens that the clouds passing at a great height from the ground cross those which, entangled in the valley, have been deflected by the successive bends of the river into a direction different from the south-west. It is then that hail falls.

In the same way, when the lower clouds meet such an obstacle as a forest, or even a mountain, eddies are formed. The masses of cloud come back on themselves, and seem to be repelled and dispersed by the forest. When the clouds have succeeded in passing the obstacle, their force is exhausted, and they only precipitate rare or inoffensive hail, and do not regain their intensity for some time afterwards.

From this we can readily understand the manner in which hills and valleys develop their respective rains.

In a large, deep cyclone, hills deflect the moist currents upwards on a large scale, and the greatest rain falls in the mountains. In shallow secondaries, with slight general wind-force, the rivers and forests give rise to local eddies, which for some reason develop the precipitation of rain, and especially of hail.

TIDAL SHOWERS.

Tidal showers are of very little practical importance, but they may advantageously be mentioned as belonging to the class of rain which hugs the valleys, and not the hills. These showers are so called because they are brought up by the tide, either along the coast or up tidal rivers. How the rising water should develop rain, we cannot explain, but the character of the influence is very obvious. On a cloudy day, when showers or heavy masses of vapour are flying about, it is frequently observed that after the tide turns to rise, and the stream is running upwards, the weather begins to get worse, so that what was merely a mass of cloud before will now precipitate rain. This rain is quite local, and does not extend far from the river-banks.

In calm weather, a wind also often comes up with the tide, or if the flow of the tide assists the general direction of the wind, the latter will be much increased in force and gustiness. If the day is really fine, of course the tide will not bring up any rain, though it may modify the wind.

From this description, the general nature of tidal action on weather will be sufficiently obvious. For some

reason or other the rise of the tide increases the intensity of the existing system of weather. If this is tending towards precipitation, the tide will give just the last impulse which is required, and rain will fall. If the ascensional impulse is strong enough of itself, it will rain independently of the tide; and if the natural impulse is downwards, as in an anticyclone, no tide is sufficient to invert the general character of the weather and cause rain. Tidal influence on weather is found all over the world. Professor Hazen has found a marked increase of thunderstorms with a rising as opposed to a falling tide in the United States; and the author has observed a well-defined tidal variation of the trade wind in tropical Fiji.

We can, therefore, sum up the contents of this chapter very easily. All over the world local influences modify, but do not make, the general character of the weather. When the latter is weak, local weather may be the prominent feature of the climate of any place; when it is strong, then local influences may be entirely obliterated.

CHAPTER XI.

DIURNAL VARIATION OF WEATHER.

In this chapter we propose to explain how to collate the variations of weather that are found in many places to depend on the hour of the day, with the great principles of the relation of weather to the distribution of atmospheric pressure which lie at the bottom of all modern meteorology. In many places the direction of the wind changes regularly at certain hours, or cloud and rain gather at the same time day after day; how can all this be reconciled with the laws of the dependence of wind on gradient, and of weather on the shape of isobars?

The cases which arise in practice are endless. Every country, every season, has its characteristic diurnal weather, and a complete account of these variations in some climates would more than fill the whole of this volume. We must, therefore, content ourselves with a statement of the general principles which are found by observation to regulate all diurnal variations, and with illustrations of a few typical examples from various parts of the globe.

INDEPENDENCE OF DIURNAL VARIATIONS AND GENERAL

CHANGES.

The great principle which underlies all diurnal weather is that diurnal variation modifies but never alters the general character of the weather, which is determined by the distribution of surrounding pressure. In England the amount of cloud or rain in a cyclone will vary at different hours, but the kind of cloud and quality of the rain will never be altered. In like manner, the land and sea breezes at Bombay will veer or back with the sun, but the general character of the wind due to the monsoon of the season will never be lost. This law of weather not only enables us to explain many phenomena of weather which would otherwise present a chaos of discordant observations; but it also serves to guide our research into the great problems of weather-forecasting. When once we know that we may safely neglect all considerations of diurnal variations when we wish to study the motion of depressions and the consequent changes of weather, our task is thereby much simplified. If we were to discuss the statistical values of meteorological elements, we should find that all the voluminous results which have been obtained by the method of averages are of no use in forecasting, and that the diurnal values of wind or rain which have thus been obtained have nothing to do with weather-changes.

DIURNAL TEMPERATURE.

In this independence of the general changes of weather, diurnal are very like local variations; but there