Chambers's Journal, July 1868.

Major rennell was the first, I believe, to whom we owe the comparison of ocean-currents to rivers. He spoke of them as ocean-rivers, and pointed out how enormously their dimensions exceed those of such streams even as the Amazon and the Mississippi. Some of the ocean-currents are from 50 to 250 miles in breadth, and flow more swiftly than the largest navigable rivers. The banks and bottom of these currents are not land, but water; and so deep are the currents that they are turned aside by shoals and banks whose tops are '40, 50, or even 100 fathoms beneath the surface of the ocean.' The outlines of ocean-currents are sharply defined, insomuch that 'often,' says Captain Maury, 'one half of a vessel may be seen floating in the current, while the other half is in common water of the sea.' The border-line of the Grulf Stream can be traced by the eye. Yet more remarkable is the distinction between the moving water and that which is at rest, when large masses of sea-weed carried along by the former enable one to recognise the rapidity with which it moves.

Of all the ocean-currents the most important, perhaps, in its bearing on the destinies of men and nations, is the great Grulf Stream. I propose to examine the course and habitudes of this current, and then to inquire a little into the vexed question of its cause.

Major Rennell traced the Gulf Stream from a supposed source in the Indian and Southern Oceans. Modern geographers and physicists prefer to look for the rise of the current somewhere near the Cape of Good Hope. 'The commencement and first impulse of the mighty Gulf Stream is to be sought,' writes Humboldt, 'southward of the Cape of Good Hope.' It appears to me, however, that the true source of the great stream is to be looked for in the equatorial zone of the Atlantic. When we come to inquire into the cause or causes which give birth to the Gulf Stream, we are led, as I imagine, to this region rather than to any other (though, perhaps, in a stream which forms part of a continuous system of circulation, we can hardly speak of any one portion as the source); I shall therefore trace the stream, and the system to which it belongs, from the great equatorial waters which move, as Columbus was the first to discover, ' with the heavens (las aguas van con los cielos), that is, from east to west, following in this the apparent motions of the sun, moon, and stars.'

The map of the Atlantic Ocean on p. 217, is constructed upon one of those forms of isographic pro-jection described in my Essays on Astronomy. It is important, in dealing with the subject of currents, that the question of area should be considered, and, therefore, that our illustrative charts should represent such areas correctly. This Mercator's charts are far from doing. The portion of the Atlantic Ocean between England and the United States of America is unduly magnified, and still more is this the case with the portion between Sweden and Greenland. On the other hand the portion between Africa and the Gulf of Mexico is unduly diminished. Thus it is scarcely possible to form from such charts just notions of the actual character of the oceanic circulation whereof the Gulf Stream forms a part. (Compare the charts illustrating the Essay on the Climate of Great Britain, pp. 264, 265.)

We see, in our map (p. 217),1 that there is a great equatorial stream extending in its eastern portion far to the south of the equator, but passing to the north also even here, and still further to the north between the coasts of Africa and South America. Near here the great equatorial current divides into two portions. One passes southward and then returns towards the east, according to some authorities, but, according to others, continues its course southward until it is lost in the Antarctic Ocean. We shall follow the northern bifurcation, however. The course of this portion of the Atlantic current system has been far more exactly traced out. Taking a north-westerly course, the great current pours itself against the barrier formed by the Leeward and Windward Islands. Passing between these islands, it sweeps around the shores of the Gulf of Mexico, a portion, however, of its volume passing probably outside the West Indian Islands, to rejoin the other outside the promontory of Florida. At this point the stream has become, probably, somewhat diminished in volume, but being still more diminished in breadth, it flows as a deep, strong, and swift stream, known among sailors as 'The Narrows of Bernini.' From hence the stream, now become the true Gulf Stream, grows gradually wider, less deep, and less swift. Off Hatteras it is already twice as broad as in the Florida Straits, and as it stretches with a wide easterly sweep across the Atlantic towards the shores of Ireland and the Hebrides, the current not only reassumes something of its original extent of surface, but again bifurcates; a wide but somewhat sluggish stream is sent southward towards the shores of north-western Africa, to rejoin the equatorial stream. The main portion of the current, however, passes with a north-easterly course up the Atlantic valley, between Iceland and Sweden to the Polar seas. It seems uncertain whether Rennell's current, which passes around the Bay of Biscay, and the current which streams southward past the shores of Spain, are forks of the Gulf Stream. They are usually represented in maps as independent currents, and in Captain Maury's large map of the Gulf Stream the great southern bifurcation already mentioned is represented as a current impinging upon the flank of the stream which flows past Spain and north-western Africa. Yet, if these streams have not their source in the Gulf Stream, it will be found no easy problem to assign their origin; and I cannot but think that the Biscay and Guinea currents, as well as the current which flows into the Mediterranean through the Straits of Gibraltar, are as truly bifurcations of the Gulf Stream as the current which laves the shores of Ireland and Sweden.