South of the tropic in India the pressure is so much lower on the land than on the sea, that the yearly direction is S. or S. W., with a ratio of resultant, increasing towards the south.

Calcutta S. 2° E. .162. Madras S. 30° W. .18. Colombo S. 61° W. .29. Farther to the west, at Bombay, the mean yearly direction is N. 45° W. .42, thus showing a flow of air from the west, or a much higher pressure on the part of the Indian Ocean between India and North Africa, as also seen on Plates 3 and 14.

As will be shown hereafter, the prevailing winds are also W. and N. W., in Syria and Mesopotamia, especially in summer, but to a less degree in the mean of the

year.

WESTERN ASIA.

In Western Asia, that is, in the part of the continent west of India and south of the Caucasus and Black Sea, numerous observations of the winds have not been made. Yet they are needed much more than, for example, in India and eastern Asia, because the latter countries have such a marked climatic type that a very few stations are enough to give us an idea of the whole. Not so western Asia, where there is no regularity and uniformity of climate, and where many local causes have influence on the wind at the few stations established there. The following table gives the PERCENTAGES of winds in this region:

The first five places have extremely prevailing west and northwest winds in summer-a flow of air towards the depression in Central Asia. At Beirut, Jerusalem and Aleppo, they may be said to be sea-winds, but this is certainly not the case at Mosul and Bagdad, as the Persian Gulf lies to the S. E. of them. Here the winds in the summer are directed from the land to the sea, as also during the same time at Madras. This movement is thus shown not to be local, caused by the difference of temperature between land and sea, but it is part of the general movement towards the depression in Central Asia and India. The meteorological effects of these winds in Syria and Mesopotamia are very different from those of the S. W.

monsoon in India-they bring dry and clear weather. This is easily explained by their origin and direction: in Mesopotamia they come over the land, in Syria from a colder part of the sea.

In all these stations there is a greater number of N., N. E., and E. winds in winter than in summer, Jerusalem alone excepted. In this the influence of the high pressure of the more northerly parts of Asia is clearly seen, and is also shown by Plate 14. Yet it seems that the higher pressure prevailing over northern Africa in that season, and generally about 30° N., has also an influence on the winds: the S. W. in Jerusalem and Beirut, and the W. at Bagdad have probably this origin. Generally the winds are not as constant in winter as they are in summer.

On the Caspian local monsoons prevail. Ashur-Ade, an island in the S. E. corner of the sea, has E. winds in winter and W. in summer. Lenkoran, on the

western shore, has prevailing S. E. in summer and N. W. in winter.

At Aralikh at the foot of the Ararat, the prevailing winds are S. E. in winter, that is, from the interior of the continent, and W. in summer.

On the eastern shore of the Black Sea, we see again very strongly marked monsoon winds, from the land (E) in winter, from the sea (S. W., W.) in summer. The winds here are nearly opposite to those of Lenkoran. It will be noticed that the monsoon character is more marked on the eastern shore of the Black Sea; the reason is, that here the local monsoons correspond to the general movement of the air over this part of Asia, while at Lenkoran they are nearly opposite to it. (See Plates 5 and 6.) The winds of Tiflis are too much influenced by the locality to show the general flow of air over the region.

Trebizonde has prevailing E. and N. W. winds in winter and summer. The country around is very mountainous, and nearly all winds come from one of these two directions. It is very difficult to reach a conclusion on the character of the winds when they are so much influenced by locality.1

It seems that the winds at Erzeroom are also much influenced by locality, as it is situated in a rugged mountainous country. It may be that at this elevation, above 6000 feet, the winds are not the same as in the lowlands. E. and N. E. are prevailing here the whole year. I must remark that at all continental stations of western Asia, north, east, and south of Erzeroom the winds are either W. or N. W. in summer (Tiflis, Aralikh, Ooroomiah, Mt. Seir, Bagdad, Mosul, Aleppo). (See Plate 5.) This is with the exception of Alexandropol, which is also a high station (4800 feet) on the plateau of Armenia. At Constantinople the local position is such as to allow scarcely any other wind than N. E. and S. W.; the N. E. are dominant. But it would be rash to conclude that this is the trade-wind. We have seen that to the N. W. of the Black Sea there is a region of prevailing N. W. winds. The country is so walled in by mountains, especially south of the Black Sea, in Asia Minor, that the air must escape through the narrow aperture of the Bosphorus; thence the N. E. winds at Constantinople.

Some meteorologists think that the "trades" are dominant in Western Asia, reaching as far as Constantinople. This idea is founded on the observations at Erze

1 See the remarks of the observer at Erzeroom, Rev. N. Benjamin, in the tables, p. 371.

room, Trebizonde and Constantinople, given by Prof. Coffin in his "Winds of the Northern Hemisphere." I have shown that so far as the summer is concerned, we cannot accept this conclusion. At Constantinople and Trebizonde the winds are too strongly influenced by locality, and Erzeroom is too high to warrant an application to the lower regions. Besides this, at other stations, better situated, the winds are westerly. As to the winter, and especially the autumn, I have no difficulty in admitting prevailing N. and E. winds in Trans-Caucasia and Asia Minor, but these are winds which have not the constancy of the trades. (See Plates 6, 7 and 14.) I should say that the erroneous opinion in regard to the extension of the trades cannot be imputed to Prof. Coffin. The number of stations was so small when he wrote his book (1853), that he wisely refrained from a conclusion.1

NORTH AFRICA.

In Africa north of the equator the winds are as given in percentages in the following table:

To express the general features of the climate of Africa north of the equator, it may be said that N. of 17° N. northerly winds prevail the whole year, especially in the south of this zone, and south of 17° N. the winds are north in the winter and south in the summer. (See Plates 5 and 14.)

From

The division-line of about 17° is the zone of lowest pressure in summer. the north air is drawn towards it from the Mediterranean, producing a wind similar to the trade in its constancy and other features. From the S. air is drawn in from the equatorial parts of the Atlantic and Indian oceans, and, coming over a great extent of warm sea, it brings clouds and rain as in the case of the Asiatic monsoon.

1 For example, p. 137 of the "Winds of the Northern Hemisphere."
2 Mean of Algiers, Oran, Mostaganem, Setif, Oum-Theboul.

N. E.

December to February.

S. E.

S. W.

N. W.

We have seen already that on the Atlantic the division-line between the N. E. trade and the S. W. monsoon of the African coast runs about 12° N., the trade losing its regularity even at 14° N. in July. On the continent this line runs more north. The country to the north, having the whole year N. winds, is rainless, or nearly so; it is the Sahara or Great Desert; south is the Soudan, the country of Agriculture, where vegetation is more and more luxuriant the more we advance southward. This is caused by the longer continuance of the rain They reach in the middle of the summer to about 17°, but in spring and autumn the division-line is more to the south, and south of this line there are southerly winds and rains.

The African traveller Rohlfs remarks that "in the beginning of July we traversed the Titümna or Great Steppe between 16° and 17° where a luxuriant vegetation is found. I noticed a remarkable change in the direction of the wind, instead of the N. E., E. and S. E. we had before, the S. W. was prevailing now. Later, when we came to the country with tropical rains (Kouka) the S. W. was still prevailing, though the rain-clouds came from the S. E." The woodland (Mimosa trees) began at about 150 N. on the route he traversed. Very similar are the conditions in Nubia. Irregular tropical rains fall as far north as 19° N., further there is a country of prairies or savannah (openings) and still further south the woodland begins. The observations at Sennaar show very well the character of the climate in S. Nubia: N. winds in winter, S. winds in summer, both largely prevailing. (See Plate 7.) The remarkable frequence of calms when the sun passes the zenith is also to be noticed. So, for example, at Schimmedru, 18° 57' N., there were 37 per cent. of calms in April, 62 per cent. in May, and 47 per cent. in June. At Kouka, 12° 52', N. Rohlfs observed 46 per cent. of calms in July, 66 per cent. in August, and 51 per cent. in September. The sun is at its zenith at Schimmedru in May and at Kouka in August.

Gorée exhibits the change of monsoon in Western Africa.

Further S., on the coast of Guinea, the winds are from the same direction the whole year; this is the region of the S. W. winds. The daily period is very well marked the whole year, the winds being N. W. in the night and morning, and S. W. in the middle of the day.

On the shores of the Mediterranean the direction of the wind is not the same as in the desert. Especially in Algeria, where the N. W. is most frequent winter and summer. (See Plates 5, 6, and 7.) At the coast stations of Egypt (Alexandria and Port Said) W. and S. W. prevail in winter, and N. and N. W. in summer. There is a belt of highest pressure in winter, and, besides this, there are winds from the land to the sea in the cold season. Farther south, N. W. and N. winds prevail the whole year, as at Cairo, Ismailia, Suez.

1 See Petermann's Mittheilungen, Ergänzungsheft, N. 25.

SOUTH AFRICA.

The winds of South Africa are very little known, except in the British Colonies in the extreme south.

It has been said already that along the west coast of S. Africa there were S. W. winds, that is, from the cold marine current towards the land. On the E. coast of Africa easterly winds prevail, although from December to February they are rather N. E., as the southern hemisphere is much heated then, and the pressure is higher on the north.

The mean direction and amount in percentage are as follows:

In Natal the general character of the winds is tropical, they are still E., but more regular in the summer season of the southern hemisphere (December to February).

In the Cape Colony the winds are regularly sub-tropical: polar (S.) from December to February, and equatorial (N., N. W.) from June to August. The regular yearly movement of the belt of highest pressure which forms the polar limit of the trades is seen here, in the extreme S. of Africa. In the warm season (December to February) it moves southward further towards the pole, so that the Cape Colony has then S. winds. In the winter (June to August) it recedes northward towards the equator. A reference to the map of isobars (Plate 14) will show that in July (midwinter) the pressure is very high in S. Africa, the isobar of 30.2 inches going from the Atlantic to the Indian Ocean, in latitude about 30°. In January, on the contrary, a pressure of 30 inches is found nowhere on the continent of S. Africa, nor on the Indian Ocean, but is restricted to the region of the cold marine current on the Atlantic.