179. Circulation caused by Differences of Temperature. The powerful sun of the Tropics heats and expands the waters of the ocean to a considerable depth; while the cold of the poles renders the water there dense and heavy. The balance of pressure is thus destroyed; the heavier polar water sinks down and flows in below the waters lying nearer the equator, while these lighter waters flow over towards the poles to restore the level. Such is the theory of what we should expect to take place; and, so far at least as the flow of the cold water below is concerned, observation seems to confirm the theory. Wherever the ocean is open through its whole depth to the polar seas, there a stratum of ice-cold water is found at the bottom, even under the equator; and this must necessarily come from the polar regions. As the oceans are most open towards the South Polar Sea, it is chiefly from it that the cold bottom currents come; so that they look like tongues shot out by the lower strata of the Antarctic Ocean. The return flow of warm water towards the poles is more difficult to trace, as it is interfered with and disguised by the wind currents.

180. This Theory disputed. In fact, some deny that differences of temperature have any sensible effect on the circulation of the waters. They hold that all the currents in the upper strata of the ocean can be traced directly or indirectly to the impulse of winds; and they account for the flow of glacial waters from the antarctic seas along the deeper channels of the oceans, by assuming that there is a greater amount of evaporation in the northern hemisphere than in the southern; that this excess of vapour is carried in the upper strata of the atmosphere, and precipitated in the southern seas; and that the deep current from south to north sets in to restore the equilibrium. The objection to this theory

is, that there is no evidence in the way of observed facts that there is an excess of evaporation in the northern hemisphere and an excess of rainfall in the southern; there are even reasons for presuming that the excess is the other way. The advocates of the temperature theory point to the difference of temperature between the polar and tropical seas as an observed fact, and argue that where there is such a difference, we know from the laws of fluid pressure that a circulation of the kind described must take place, unless prevented by a more powerful cause. There is here no such cause known to exist; and if they are unable rigidly to demonstrate the upper part of the circuit due to the temperature differences, it is because it is mixed up with the more decided movements produced by winds.

181. Horizontal Currents.-The great features of the horizontal circulation are thus sketched by Sir Wyville Thomson (Address to the Geographical section of the British Association, August 22, 1878): The causes of natural phenomena, such as the movements of great masses of water, or the existence over large areas of abnormal temperature conditions, are always more or less complex, but in almost all cases one cause appears to be so very much the most efficient, that in taking a general view, all others may be practically disregarded; and speaking in this sense, it may be said that the tradewinds and their modifications and under-currents are the cause of all movements in the stratum of the ocean above the neutral layer. This system of horizontal circulation, although so enormously important in its influences upon the distribution of climate, is sufficiently simple. Disregarding minor details, the great equatorial current, driven from east to west across the northerly extensions of the ocean by the trade-winds, impinges

upon the eastern coasts of the continents. A branch turns northwards and circles round the closed end of the Pacific, tending to curl back to the North American coast from its excess of initial velocity; and in the Atlantic, following a corresponding course, the Gulf Stream bathes the shores of Northern Europe, and a branch of it forces its way into the Arctic basin, and battling against the palæocrystic ice, keeps imperfectly open the water-way by which Nordenskjöld hopes to work his course to Behring's Straits. The southern deflections are practically lost, being to a great extent, though not entirely, dissipated in the great westerly current of the southern anti-trades.'

182. Current of the Atlantic.-The equatorial current in the Atlantic, with its deflections and secondary currents, being the best known, may be taken more in detail as typical of the whole set of phenomena. It sets from the west coast of Africa towards the east coast of Brazil, and is owing to the trade-winds (see par. 250), which push the waters of the Atlantic westwards. The depth of such a drift-current is not great-not exceeding 50 fathoms-and the velocity is less than a mile an hour.

183. Indraught-currents.-But here we must note an indirect effect of this action of the wind. The waters being blown away from the African coast, a depression is caused, and to fill it up, two indraughtcurrents set in, the one from the north along the coast of Portugal and Morocco, the other from the Cape of Good Hope along the west coast of Africa; and both these currents, it may be added, are aided by winds blowing in the same direction.

184. Equatorial Current divides.-When the equatorial current itself reaches the coast of Brazil, it divides

into two branches. One proceeds southward, turning gradually eastward, and a part of it curling fairly round, flows northward along the African coast until it falls in with the northern indraught formerly spoken of, and thus makes the complete circuit. The rest of this branch finds its way into the Southern Ocean, and merges in the easterly drift caused by the anti-trades. The other and much larger branch of the equatorial current is deflected northward into the Caribbean Sea and the Gulf of Mexico. The water thus driven into this pent-up sea now rushes with accumulated force through the strait or gulf between Florida and the Bahamas, and forms the famous Gulf Stream.

185. The Gulf Stream.-The Gulf Stream, after issuing from the Florida Strait, proceeds at first northward, parallel to the American coast; but after coming to Cape Hatteras, it turns gradually eastward, passing over the southern extremity of the Bank of Newfoundland, and all the while expanding in breadth and becoming shallower. The temperature of the

stream when it starts is from 83° in summer to 77° in winter, and even after travelling 3000 miles to the north, as high as the Banks, there is a difference in a winter day between its water and that of the surrounding ocean of 20° to 30°. Along its whole course a cold polar current underlies it; and this polar current comes in between the western border of the stream and the coasts of Florida, Georgia, and the Carolinas, the line of separation between the two being so abrupt that it is known as the 'cold wall.' The velocity of the stream at its outset is from 50 to 60 miles a day; but this velocity becomes greatly reduced as it proceeds.

186. It has usually been held that the Gulf Stream extends across the Atlantic to the shores of Northern

Europe, and is the cause of the mild and moist climate enjoyed by the western parts of that continent. The opinion, however, is beginning to prevail that, as a distinct current, the Gulf Stream ceases in the middle of the North Atlantic, its waters being by this time thinned out to a mere film, and its original velocity and distinctive heat having been dissipated. That warm waters from tropical seas are brought to the coasts of Britain, and even into the Polar Seas beyond, is proved by drift-wood, seeds, and fruits from the West Indies being frequently cast ashore on the Hebrides, the north of Norway, and Spitzbergen. But this is accounted for by the general flow of the surface-water towards the poles, forming part of the vertical oceanic circulation; a flow which receives an eastward deflection as it proceeds northward. This general set of the surface-water is further promoted by the prevalence of south-westerly winds, which maintain a pretty constant north-east drift over the whole surface of the north-eastern portion of the Atlantic. In this way, although the Gulf Stream may have lost its original impetus, a large portion of the superheated water which it brings into the centre of the Atlantic must be carried to the shores of Europe and into the Arctic Sea.

187. The Labrador Current.-While the climate of the west of Europe is thus rendered mild by having its shores washed by waters from warm seas, and by the moist and south-westerly winds that predominate, the opposite coast of America is greatly depressed by a current from the Greenland seas which flows southwards along the shores of Labrador, carrying with it fields of polar ice and icebergs, and accompanied by dry and piercing winds from the north.

188. The Sargasso Sea.-Part of the waters of the