If N. E. and E. are taken as the true representatives of the trade-winds, we see that they amount to 49 per cent. of all winds in the rainy season of Northern Venezuela, and to 68 per cent. in the dry season. In Guiana the trades are more regular, but it seems that it lies somewhat to the S. of the thermal equator; as December, January, and February are rainy months, the proportion of N. E. winds is then greatest. (See Plates 5, 6, and 7.)

In Tropical Brazil we have only the observations in Rio Janeiro, from which we deduct the PERCENTAGES given in the following table. To gain a better insight of the winds of this country, it is necessary to obtain observations made on the Atlantic Ocean near the coast.

On the Atlantic Ocean, near the coast of Brazil, the winds are more northerly in the rainy season, from December to February, or in the summer of the southern hemisphere. As to Rio Janeiro, the winds are influenced by the locality. The sea-wind (S. E.) generally begins at 9 A. M. and blows till sunset, while landwind and calms prevail night and morning.

The northerly direction of the winds off the coast of Brazil points to a barometrical minimum in the interior to the west of the Organ Mountains in the campos, as it should do, considering the great heat of the summer of the S. hemisphere, and the sparse covering of trees on the campos. Easterly winds are much more regular and strong in Northern Brazil1 than on the coast near Rio Janeiro. In the latter region they are to be considered as sea-breezes rather than trade-winds. In the La Plata States and on the Atlantic Ocean near them, the few observations we have give the following results in PERCENTAGES:

1 Burton gives a description of the strength of the E. winds on the lower S. Francisco River in "Highlands of Brazil,"

The only observations of a year's duration made in the interior are those at Mendoza and Parana by Burmeister.1 He remarks as to Mendoza, that calms largely prevail, strong winds are very rare. In Parana, on the contrary, as on the coast of the La Plata States (Buenos Ayres, Montevideo) the winds are violent and atmospheric changes frequent and sudden. In this respect it reminds us of the climate of the Atlantic coast of the United States, though extremes of heat and cold similar to those of North America are never experienced. Two winds are especially noted as strong, the Pampero (S. W.) and the Su-Estada (S. E.). Though Parana is near to Buenos Ayres, the yearly period of the winds is nearly opposite. (See Plates 5, 6, and 7.)

The winds on the Straits of Magellan and on the west coast of S. America are very different from those of the eastern part of this continent, as shown in the following table of PERCENTAGES.

The changes of the winds along the W. coast of South America are very regular and gradual; we can follow them for about 60° of latitude. In the extreme south, at Punta Arenas westerly winds are known to prevail especially from December to February, the warm season, while from June to August the number of N. and N. E. winds increases. These are land-winds. The mean direction is found to be northwesterly at all seasons, and the ratio of resultant great (see next page). We are here in the belt of westerly (or northwesterly) winds of the

89 July, 1875.

See his "Klima der Argentinischen Republik."

southern hemisphere, which are very strong and prevail all around the globe, especially from 40° to 60° S. In Puerto Montt these winds also prevail, especially in the cold season, June to August, while the quantity of southerly (cold polar) winds increases in December and February.

In Central Chili (Valparaiso and Santiago) we meet opposing winds in winter and summer. They are northerly in the cold season, southerly in the warm. This is a feature of the sub-tropical belt, which is especially well developed in the southern hemisphere, owing to the great extent of sea. But in Chili the winds are S. and S. W. from December to February, instead of S. E., the direction of the true trades. But this is easily explained by the high chain of the Andes, which does not permit an extensive circulation of air from the S. E. Besides this, air is drawn towards the land from the sea, which is to the westward. The seasons of Central Chili are in keeping with the sub-tropical winds; the summer months are rainless. The further we advance to the N. the greater time the polar winds (S. E., S., S. W.) blow and the longer is the rainless season. About 27° S. the rain ceases altogether, and this belt stretches along the coast of Bolivia and Peru to 5° S. Santiago has regular sea and land winds, especially from December to March, as is shown by the tri-horary observations of the U. S. expedition under Capt. Wilkes. (See tables, Zone 25.) At midnight, 3, and 6 A. M., the winds are nearly N., from 9 A. M. to 9 P. M. they are nearly S. W. There is no gradual passage of one into the other, but a calm separates them in the morning and evening.

From the latitude of Central Chili to the Isthmus of Panama we can supply the deficiency of land-observations by those made at sea, near the coast. The prevailing wind in zones 25 and 24 (25°-35° S.) is S., especially in the last, where from that quarter more than half of all the winds blow. As we advance towards the north the wind is deflected to the S. E. by the influence of the earth's rotation. Between 5° and 10° S. (Zone 20) 90 per cent. of all the winds blow from the S. E. in the cold months of the year, giving the ratio of resultant 96. This gradual change in the direction of the wind is clearly seen on Plates 5 and 6. There is scarcely any trade-wind region in the southern hemisphere where they are so largely prevailing, and none in the northern hemisphere. Yet it is necessary to remember that the observations between 5° and 10° S. were taken further from the coast (85°-98° W.) than on the other parallels, and thus the proportion of S. E. winds is greater, and of S. less. The nearer to the coast, the less frequent are the S. E. winds, because of the proximity of the Andes on the E., and also because the land is here much warmer than the sea, on account of the extremely cold Peruvian current.

As we approach the equator, the S. winds again increase. Between 0° and 5° S. this increase is probably caused by the position of the cold marine current, which is deflected to the westward. But southerly winds here cross the equator, and are by the earth's rotation deflected to the S. W. Already between 0° and 5° N. there is a great proportion of S. W. winds, though the S. winds still prevail. The mean direction is to the W. of S., as shown on Plates 5 and 6. Between 5° and 10° N., even southerly winds prevail during nine months, especially from June to August. Only in the winter of the northern hemisphere the wind is N. W., and then even with a small ratio of resultant. The equatorial boundary of the northern

trades is thus seen to lie much N. of the equator in the Eastern Pacific. A great body of air is thus drawn in to about 10° L. N., and forms what is called a S. W.

monsoon.

In other regions this is also the case; these S. W. monsoons reach a much higher latitude, about 12° N. on the coast of Africa, 17° N. in the interior of this continent, and even 30° N. in India. The following table gives the mean direction of the wind at stations in South America:

There are four wind-belts stretching across the Atlantic Ocean: the northern belt of westerly winds (principally S. W.); the northern trade-winds (N. E.); the southern trade-winds (S. E.); and the southern belt of westerly winds (principally N. W.). The first and the last of these are also called belts of variable winds in opposition to the constant trade-winds.

As the Atlantic Ocean is the great highway of civilized nations, its meteorology is better known than that of any other ocean. Though narrow when compared to the Pacific and Indian Oceans, the winds have sufficient space on the Atlantic, as it has very few islands, and no mountain-chain in its vicinity at all comparable to the Andes, which exercise so great an influence on the winds of the Pacific. This being the case, the winds of the Atlantic can be regarded as typical for the (See Plates 5, 6, and 7.)

oceans.

The most important boundaries of the different systems of winds which occur in the Atlantic are the so-called outer (or polar) and the inner (or equatorial) limits of the trades. We give below these limits, according to the best source of informa

direction.

resultant.

Mean

resultant.

tion, the "Pilot Chart of the Atlantic Ocean," edited by the Meteorological Office in London.

The N. E. trade is much more to the north in the eastern part of the ocean than it is near the coast of America, and on the meridians of 55° to 50° W. its polar limit is still further south. We do not know accurately the equatorial limits of the N. E. trade; on these meridians they must, however, fall on the continent of S. America. The trade-wind belt seems to be more narrow about 40° W. than further eastward, except in the months from July to September.

The equatorial belt of calms and variable winds between the N. E. and S. E. trades is much broader and better marked in the eastern part of the ocean than in the middle. About 20° W. its mean breadth attains 12° in September, and even in January 30, while at 35° W. its breadth is only from January to March, so that frequently ships sail from one trade into another without passing through intervening calms. It should be observed that the direction of both trades is much