scribed is considerable. The sea round the Irish coast is about 10° warmer than in February, while along the English Channel, and as far up as the Wash, a uniform temperature of nearly 60° exists. In the last-named district the annual variation is nearly 20°. Passing farther north, we find stations giving a mean reading of nearly 55° on the North Sea coasts, and of fully that temperature on those exposed to the Atlantic. Here then the annual range is about 15°.

Of course these figures do not make any claim to precise accuracy, but they are sufficient to show that the mean temperature of the sea-surface is most constant at the entrance of the Channel, and least so off the mouth of the Thames. These are also the districts which are marked by great contrasts between the range of air temperature, which is small in Cornwall but comparatively large in the south-east of England, where the conditions most nearly approach those of the continent of Europe.

321

CHAPTER XVI.

THE DISTRIBUTION OF RAIN.

HAVING dealt with the subjects of the Temperature and Pressure of the Air, of the Winds, and of the Temperature and Currents of the Sea, the remaining factor which affects climate is the Distribution of Rain; and of this a brief account will be given.

The early rain charts of the globe exhibited certain belts, coloured in different tints, to show differences in the amount of fall. The average amount all over the globe was put down at about sixty inches, which was apportioned between the different geographical zones as follows:-The torrid zone was credited with 100 inches, the temperate zones with 30 inches, and the frigid zone with 15 inches. All such statements, however, rest on very questionable authority, for between the tropics the measured rainfall varies from 260 inches at Mahabuleshwur, on the Ghauts near Bombay, to two or three inches at Ascension, and to even less at some guano islands in the mid Pacific, to which I shall allude subsequently; while over the entire surface of the ocean there are no measurements at all that are of any value. As ships are in motion, we can only obtain from their logs entries of the number of times that rain was recorded, but not of its amount at any fixed station.

Y

The first advance on the simple belt rain map was to shade the coast-lines in certain places of a deeper tint; but if we compare a map on this principle, showing the total annual amount of rain, for example, that of Professor Loomis in the American Journal of Science and Art,' for January 1882, with the most recent map of seasonal rain distribution, that of Wojeikoff in vol. i. of the Zeitschrift für wissenschaftliche Geographie,' it is hard to realise that they represent only different aspects of the same phenomenon, the fall of rain.

It is impossible, in the present state of our knowledge, to give a true picture of these several elements, the amount, and the distribution in space and in time, of rainfall. If we were to attempt to draw monthly rain-charts we should at once be met by the difficulty that information for many portions of the globe is not attainable. The regions, in the northern hemisphere alone, in which practically no stations exist, are the interior and east of Asia; the whole of British North America, except the province of Canada; the entire Arctic Regions; and the Atlantic and Pacific Oceans. Curves could therefore be drawn with confidence over a very limited area. It is not with rainfall as with temperature or pressure, where corrections of more or less exactitude can be applied to reduce readings to their equivalents at sea-level. The distribution in vertical height of the amount of rain varies in different latitudes and with differences in the position of the mountain chains with regard to the prevailing wind, so that two stations situated close together may differ widely in the amount and in the seasonal distribution of their rain. Wojeikoff gives a very telling instance of this from two stations in Java, Batavia and Buitenzorg.

The rainfall at the former place amounts to 81 inches, and the proportion between the quantity collected in the wettest month, January, and the driest, July, is 8 to 1. At Buitenzorg, situated only 25 miles from Batavia, and at an elevation of about 900 feet above it, the rainfall is 148 inches, and the proportion between the wettest month, March, and the driest, June, is only 2 to 1. The former station has true Monsoon rains, the latter has a fall which is nearly uniform in all the months. How then can the district of the island of Java, which contains these two stations, be fairly represented in any rain-chart of the globe? And yet in such charts, for many parts of the world, owing to the paucity of observing stations, the rainfall at one spota purely local phenomenon-has been assumed as a correct indication of the precipitation for an immense

area.

I shall not, therefore, attempt any graphical representations of rain distribution, but shall give a brief description of its principal features.

In Chapter VIII. (p. 137), the three great agencies which are efficacious in bringing about precipitation of moisture have already been stated: ascending currents; the contact of warm air with the cool surface of the ground; and the mixture of masses of air of different temperatures. I shall now touch on the several factors which influence the distribution of the fall.

Of these the wind has to be considered first. Winds moving from high to low latitudes are generally dry, those moving in the opposite direction are generally damp. Winds blowing from off the shore on to the sea-surface, especially if the coast be bold and the land elevated, so that the wind is forced to descend, are

essentially dry-the action being similar to that which gives its character to the Föhn of Switzerland-while winds blowing from the sea deposit their moisture on the coast. The South-east Trade wind, itself essentially dry at sea, becomes a rain-bringer to the mountains of eastern Brazil, and even to the eastern slopes of the Andes. For this reason, that the winds in low latitudes are Easterly, the eastern sides of tropical mountain ranges are, as Wojeikoff points out, often better watered and wooded than the western-the reverse of what is the experience in the temperate zone.

As already explained, this action is only, in a very slight degree, attributable to the contact of the warm air with the colder surface of the ground, for such an action can only affect directly the lowest strata of the atmosphere, unless very high mountains line the coast. The real cause is to be found in the enforced transference of the air into higher, and therefore colder, levels.

The equatorial zone of calms over the Atlantic and Pacific is a region where it is alleged that there are heavy and constant rains. These are due to the ascending movement of the air in the calm zone, as explained in Chapter XIII., which produces a constant canopy of dense cloud from the condensation of its moisture. This moisture is again restored to the earth in constant torrent-like showers, often accompanied by thunder and lightning. Caldcleugh's account of the regular occurrence of thunderstorms in Brazil has been already noticed (p. 146).

This statement of the perpetual rainfall near the equator may probably be accepted as a general assertion, but that the rain is not incessant in very low latitudes all round the globe is proved by the fall at Ascension