winds taken in conjunction with the configuration of the coast and its submarine approaches. The trade-wind regions correspond pretty closely with westward-flowing currents, while in the equatorial calm belts there are eastward-running counter-currents, these lying north of the equator in the Atlantic and Pacific, but south of the equator in the Indian Ocean. In the region of the westerly winds on the poleward side of 40° N. and S. the currents again flow generally eastward. A cyclonic circulation of the atmosphere is associated with a cyclonic circulation of the water of the ocean, as is well shown in the Norwegian Sea and North Atlantic between the Azores and Greenland. Where the trade-winds heap up the surface water against the east coasts of the continents the currents turn poleward. The north equatorial current divides into the current entering the Caribbean Sea and issuing thence by the Strait of Florida as the Gulf Stream, and the Antilles current passing to the north of the Antilles. Both currents unite off the coast of the United States and run northward, turning towards the east when they come within the influence of the prevailing westerly winds. In a similar manner the Brazil current, the Agulhas current and the East Australian current originate from the drift of the south-east trades, and in the North Pacific the Japan current arises from the north-east trade drift. The west-wind drifts on the poleward side carry back part of the water southward to reunite with the equatorial current, and thus there is set up an anticyclonic circulation of water between 10° and 40° in each hemisphere, the movement of the water corresponding very closely with that of the wind. The coincidence of wind and current direction is most marked in the region of alternating monsoons in the north of the Indian Ocean and in the Malay Sea.
The accordance of wind and currents is so obvious that it was fully recognized by seafaring men in the time of the first circumnavigators. Modern investigations have shown, however, that the relationship is by no means so simple as appears at first. Wemust remember that the ocean is a continuous sheet of water of a certain depth, and the conditions of continuity which hold good for all fluids require that there should be no vacant space within it; hence if a single water particle is set in motion, the whole ocean must respond, as Varenius pointed out in 1650. Thus all the water carried forward by any current must have the place it left immediately occupied by water from another place, so that only a complete system of circulation can exist in the ocean. Further, all water particles when moving undergo a deviation from a straight path due to the forces set up by the rotation of the earth deflecting them towards the right as they move in the northern hemisphere and towards the left in the southern. This deflecting force is directly proportional to the velocity and the mass of the particle and also to the sine of the latitude; hence it is zero at the equator and comes to a maximum at the poles. When the wind acts on the surface of the sea it drives before it the particles of the surface layer of water, and, as these cannot be parted from those immediately beneath, the internal friction of the fluid causes the propelling impulse to act through a considerable depth, and if the wind continued long enough it would ultimately set the whole mass of the ocean in motion right down to the bottom. The current set up by the grip of the wind sweeping over the surface is deflected by the earth's rotation about 45° to the right of the direction of the wind in the northern hemisphere and to the left in the southern. The deeper layers lag behind the upper in deflection and the velocity of the current rapidly diminishes in consequence. The older theory of the origin of drift currents enunciated by Zöppritz in 1878 was modified as indicated above by Nansen in 1901, and Walfrid Ekman subsequently went further. He showed that at a certain depth the direction of the current becomes exactly the opposite of that which has been imposed by deflection on the surface current, and the strength is reduced thereby to only one-twentieth of that at the surface. He called the depth at which the opposed direction is attained the drift-current depth, and he found it to be dependent on the velocity of the surface current and on the latitude. According to Ekman's calculation with a trade-wind blowing at 16 m. per hour, the drift-current depth in latitude 5° would be approximately 104 fathoms, in latitude 15°, 55 fathoms, and in latitude 45° only from 33 to 38 fathoms. A strong wind of 38 m. an hour would produce a drift-current depth of 82 fathoms in latitude 45°, and a light breeze of 3 m. an hour only 22 fathoms. It follows that a pure trade-wind drift cannot reach to any great depth, and this seems to be confirmed by observation, as when tow-nets are sunk to depths of 50 fathoms and more in the region of the equatorial current they always show a strong drift away from the side of the ship, the ship itself following the surface current. Ekman shows further that in a pure drift current the mean direction of the whole mass of the current is perpendicular to the direction of the wind which sets it in motion. This produces a heaping-up of warm water towards the middle of the anticyclonic current circulation between 10° and 40°, and on the other hand an updraught of deep water along the outer side of the cyclonic currents. The latter phenomenon is most clearly shown by the stripes of cold water along the west coasts of Africa and America, the current running along the coast tending to draw its water away seawards on the surface and the principle of continuity requiring the updraught of the cool deep layers to take its place. For this reason the up-welling coastal water is coldest close to the shore, and hence it only appears on the Somali coast during the south-west monsoon. On the flat coasts of Europe the influence of on-shore wind in driving in warm water, and of off-shore wind in producing an updraught of cold water, has long been familiar to bathers. In a similar way updraughts of cold water to the surface occur in the neighbourhood of the equator, especially in the Central Atlantic and Pacific.
When a drift-current impinges directly upon a coast there is a heaping up of surface water, giving rise to a counter-current in the depths, which maintains the level, and this counter-current, although subject to deflection on account of the rotation of the earth, is deflected much less than a pure drift-current would be. Such currents, due to the banking up of water, have a large share in setting the depths of the sea in motion, and so securing the vertical circulation and ventilation of the ocean.
The difference in density which occurs between one part of the ocean and another, shares with the wind in the production of currents. Vertical movements are also produced by difference of temperature in the water, but these can only be feeble, as below 1000 fathoms the temperature differences between tropical and polar waters are very small. If we assign to a column of water at the equator the density S4t = 1.022 at the surface and 1.028 at 1000 fathoms, or an average of 1.025, and to a column of water at the polar circle a mean density of 1.028, there would result a difference of level equal to (1.028 − 1.025) × 1000 = 3 fathoms in a distance from the equator to the polar circle of some 4600 m. A gradient like this, only 1 in 1,350,000, could give rise only to an extremely feeble surface current polewards and an extremely feeble deep current towards the equator. If there were strong currents at the bottom of the ocean the uniform accumulation of the deposit of minute shells of globigerina and radiolarian ooze would be impossible, the rises and ridges would necessarily be swept clear of them, and the fact that this is not the case shows that from whatever cause the waters of the depths are set in motion, that motion must be of the most deliberate and gentlest kind. In exceptional cases, when a strong deep current does flow over a rise, as in the case of the Wyville Thomson Ridge, the bottom is swept clear of fine sediment.
Strongly marked differences in density are produced by the melting of sea-ice, and this is of particular importance in the case of the great ice barrier round the Antarctic continent. O. Pettersson has made a careful study of ice melting as a motive power in oceanic circulation, and points out that it acts in two ways: on the surface it produces dilution of the water, forming a fresh layer and causing an outflow seaward of surface water with very low salinity; towards the deep water it produces a strong cooling effect, leading to increase of density and sinking of the chilled layers. Both actions result in the drawing in of
