more easterly in the western than in the eastern part of the ocean. This will be readily seen by a reference to the map. (Plates 3, 5, 6, and 7.)

There are in other places much greater differences in the limits of the S. E. trade. Near the coast of America the winds are so irregular that the seamen do not consider them true trades, thus on the meridian of 30° W. the polar limit is set down at 160 S. to 21° S. according to the seasons. Near the coast of Africa (10° E.) the polar limit is south of 30° S. at all seasons. The S. E. trade advances much beyond the equator, except in the months of February, March, and April. In September it goes to 6° N. under 40° W.

The narrowing of the ocean in its equatorial part between Cape S. Roque in S. America and Cape Verde in Africa does not allow of a determination of the equatorial limits of the trade east of 17° W. and west of 40° W., as it is known that the trades blow regularly only on the ocean.

The greater breadth, however, of the S. E. trade and its regularity near the equator are well known.

Along the coast of S. Africa there are prevailing S. W. winds the whole year. They exist also on the ocean. This African monsoon is caused by the rarefaction of the air in the interior of the continent, and, in the months from July to September, extends far beyond the equator, and occupies much of the zone between the S. E. and N. E. trade. Violent rains and thunder-storms are experienced at this season in this region of S. W. winds. There is no doubt that the S. E. trade is drawn far beyond the equator and gradually changed into a S. and then a S. W. wind. Having passed over a broad expanse of warm sea it is copiously loaded with vapor. From January to March the Pilot Charts give the southern boundary of the S. W. winds at 21 N. and 15° W. It crosses the equator under 10° W.; 7° S. under the meridian of Greenwich; 10° S. under 4° E.; 20° S. under 10° E. Thus the belt of S. W. winds has the greatest breadth opposite the Bay of Biafra, and is much narrower North and south. From April to June the S. W. winds advance to 19° W. opposite Sierra Leone, while the boundary is nearer to the coast of Africa further southward.

From July to September the belt of S. W. winds occupies a great space off the west coast of N. Africa, between 17° and 32° W. and 60-110 N. If the boundary were traced for every month, it would be found to coincide much more closely with the inner limits of the N. E. and S. E. trade; as it is, it is near enough, as the southern limit of the N. E. trade is 12° N. in September, near the coast of Africa, while the S. W. monsoons begin about 1110 N.

It is important also to obtain a knowledge of the minor characteristics of the winds of the Atlantic, and this can best be done by studying the percentage of winds in the different regions of the ocean, as presented in the following table:

North of the regularly-established trades, there is a zone with prevailing northerly winds, especially in summer, in the eastern part of the ocean, as seen on maps, Plates 5, 6, and 7. To this zone the Azores belong.

At Funchal, Madeira, the trade-wind is well established in summer, but northerly winds prevail in winter, though not regular enough to be called trades.

The northerly winds of summer between 30° and 40° N. are N. W. rather than N., showing the influence of the heated surface of Southern Europe and Northern Africa. The African monsoon is to be observed in Z. 17, N. 31, and on Plate 5; the prevailing wind is N. from December to February and S. from June to August.

Under the same latitude in the middle of the ocean the N. E. trade is well established at both seasons. (See Plates 5, 6, and 7.) In zone 18 (0°-5° N.) the S. E. trade begins to prevail.

Along the coast of Africa the S. E. trade is very southerly, especially from June to August in latitude from 5° to 15° S. It must be remembered that a cold marine current flows along this course, and, therefore, the conditions must be like what prevail near the western coast of S. America. (See maps, Pl. 5, 6, and 7.)

The wind blows along this cold current, while on the coast it blows from the cold current to the land; this gives the S.W. winds of South Africa from 0° to 20° S. The only difference from S. America is, that no such high chain of mountains rises here near the coast. The belt of land under the influence of the sea-winds is more

extensive in Africa, and more heated, the ascending current is, therefore, more powerful, and thus the air from over the cold current is attracted with more force. The mean direction of the wind in the tropical part of the Atlantic is as follows:

This table is so arranged as to show the corresponding latitudes north and south opposite to one another. It will be seen how much more regular are the southern trades, especially between 0°-15°.

In the northern hemisphere the trades are well established between 10°-15° N. in the middle and western part of the ocean; while near the African coast the winds are very variable, or better to say this latitude is divided in summer between the N. E. trade and the S. W. monsoon. In the corresponding latitude south, the S. E. trade is blowing regularly the whole year.

In latitude 5°-10° N. the S. E. trade is already established in the middle of the ocean from June to August and the African monsoon in full force further east. In the corresponding latitude in the southern hemisphere the trade is very regular. It is also blowing between 0°-5° N. with the exception of the months from December to February, when the mean direction is E. N. E. in the western part of the ocean, probably owing to the heating of a part of S. America, towards which the air is drawn from the ocean. (See also Map, Plate 6.)

The more easterly direction of the trades in the western part of the ocean is well marked, especially as concerns the S. E. trade. It is probably due to the rotation of the earth, which gives the winds more easting the further they advance. There is no doubt that the winds of the Atlantic which blow near the coasts of America have traversed a great part of the ocean, and thus acquired more easting. As to the winds which blow in the eastern part of the ocean, they do not come from so far. The African continent rather attracts the winds than otherwise. It has before been shown that from 5° N. to 20° S. southwesterly winds blow the whole year on the ocean near the coast of Africa, as exhibited on Plate 7. Thus the trade which blows further to the west cannot come from Africa. It originates on the Atlantic Ocean itself, over the cold antarctic current flowing at some distance from the African coast.

Barometric observations are numerous on the Atlantic Ocean, and are important as giving us the key to the winds. (See Plate 14.) Unfortunately their tabulation and reduction is not all that can be desired. They are calculated without regard to longitude, and from 5° to 5° of latitude only. Thus we do not know the difference of pressure in the eastern and western parts of the ocean, although it must be great, especially in latitude from 20° to 35° N. and S. as shown by the great difference in the polar limits of the trades.

The Meteorological Institute of the Netherlands has undertaken the calculation of the barometric means of the Atlantic Ocean for every degree of latitude, distinguishing also, in the southern hemisphere, the outward and homeward voyages. This would give two sets of figures, one for the eastern and one for the middle part of the ocean, as the ships going to the East Indies take a course more to the westward, while on returning they go nearer to the coast of Africa. This expected publication will shed light on many obscure problems.

The most complete barometrical table for the Atlantic we now possess is published in the Pilot Charts. It is calculated from 5° to 5°, for every month. I have calculated from it the pressure of the two contrasting seasons, and have given in the following table the pressure observed on some islands and coast stations reduced to sea-level. (See also Plate 14.) The mean pressure is at 32° Fahr.

The polar boundaries of the N. E. and S. E. trades are marked by a high pressure (at 30° to 35° N. and at 20° to 30° S.), while the space intervening between the two trades-the belt of equatorial calms and variable winds, has a comparatively low pressure. It should be remarked that this low pressure remains on the northern hemisphere, changing from 10° to 15° N. in our summer and from 0° to 5° in our winter. The air from north and south is attracted towards this belt of low pressure, and, as the conditions of the tropics are very uniform, the winds also are very regular,

A comparison of stations in the west and east of the ocean will show that pressure is generally higher in the east (as in Madeira compared with Bermuda, in Christiansburg compared with Cayenne, and Cape Town compared with Buenos Ayres). This is an additional cause for the easting of the trade-winds near the American continent.

Pressure is extremely low in the higher latitudes of the southern hemisphere.

Between 55° and 60° it is lower than around Iceland, the lowest known in the northern hemisphere. The great permanence and strength of the westerly winds in the southern temperate zone is explained by this. (See Plates 5, 6, and 14.)

NORTHWESTERN EUROPE.

The islands to the N. W. of Europe have still the climate of the Atlantic Ocean. Only one of them, the largest and most northerly, Iceland, has some of the characteristics of the polar zone.

Near Iceland, on account of the heated current of the gulf-stream, is the lowest pressure of the northern hemisphere, and though it is especially marked in autumn and winter it is also conspicuous at the other seasons. As is to be expected from a country in such a position, the winds are very changeable, according to the shifting of the centre of lowest pressure to the north and south. The equatorial winds, S. W., and the polar, N. E., prevail in turn.

The Faröe islands have prevailing S. W. winds at all seasons.

This is even

more the case at the Shetland islands, and in Great Britain generally, as is shown by percentages in the next table.

There is little difference between the winds in winter and summer, from Faröe islands to southern England. A very slight one only can be detected in the greater number of W. and N. W. winds in summer. This applies not only to Great Britain, but also to the greatest part of northern and central Europe. It is due to two causes: First, the belt of highest barometer is more northerly in summer than in winter; and second, part of the air is attracted towards the depression of Central Asia.

In Great Britain the influence of the last cause is very small, as Central Asia is too distant, and the depression about Iceland so near, that it must act very powerfully even in summer. But the further we advance eastward the greater is the influence of the depression in Central Asia, and consequently the greater the difference between the direction of the wind in winter and summer. The next table gives the mean direction of the wind in Great Britain and Iceland.

L