Thus, there can be no possible question that the depths of the Atlantic are occupied by a vast current much colder than the deep sea temperature due to the latitude, and, therefore necessarily flowing from the arctic towards the tropical seas.

Such are the broad facts of the Atlantic circulation. We have a surface circulation whose general features are such as have been described, and are generally admitted, though a dispute has arisen as to a question of nomenclature; and then we have a general submarine 'set' of water from the arctic regions towards the tropics, the existence of which is also generally admitted.

But now we again approach a subject of controversy, and one which is certainly better worthy of discussion than that which we considered above. It relates, in fact, to the question how this wonderful system of oceanic circulation is brought about.

Passing over several crude theories which have long since been disposed of, we come first to the theory that the system of oceanic circulation is due to the action of the trade-winds. This theory has been maintained by Franklin and others in former times, by Sir John Herschel in our own, and is warmly advocated in the present day, by many whose opinions are not to be rashly contradicted.

Against this theory it has been urged by Captain Maury' - 'with singular wrongheadedness' according to the Edinburgh Reviewer, but as it seems to me with no small degree of reason-that the trade-winds are neither powerful enough nor persistent enough to account for the great equatorial currents, or therefore for the Gulf Stream. Maury says, 'with the view of ascertaining the average number of days during the year that the north-east trade-winds of the Atlantic operate upon the water between the equator and 25 degrees north latitude, log-books containing no less than 380,284 observations on the force and direction of the wind in that ocean were examined. The data thus afforded were carefully compared and discussed. The results show that within these latitudes-and on the average-the wind from the north-east is in excess of the wind from the south-west only 111 days out of the 365. Now, can the north-east trades, by blowing for less than one-third of the time cause the Gulf Stream to run all the time, and without varying its velocity either to their force or prevalence.' Our reviewer not only dwells on the wrongheadedness of this argument-wholly irresistible as it appears - but asserts that' the trade-wind origin of the Gulf Stream is about as certain as the rotundity of the earth.' It could have been wished that in place of abusing Captain Maury for wrongheadedness, the reviewer would have devoted a few lines to the demolition of Maury's argument.

Maury himself advanced the relative lightness of the equatorial water as the true reason of the oceanic circulation. But granting that the expansion of the equatorial water under the sun's heat, as well as the resulting buoyancy, would cause an overflow of equatorial water polewards, this overflow would be an exceedingly slow movement, and it would result in an eastwardly instead of a westwardly flow, for the very same reason that the counter trade-winds travelling polewards assume an eastwardly direction.

In the Student for July 1868, I advanced another explanation. I urged that the sun's action on the equatorial and tropical regions of the Atlantic, raising immense quantities of water by evaporation, causes an influx of water from below. ' There can be no question,' I then wrote, 'that under-currents arriving in this manner, whether from the north or from the south'

(that is from arctic or from antarctic regions) ' would acquire a strong westerly motion (just as the trade-winds do). Thus they would generate from below the great equatorial westerly current. In this upflow of cool currents having a strong westerly motion, we find the mainspring of the series of motions. The water thus pouring in towards the equator is withdrawn from beneath the temperate and arctic zones, so that room is continually being made for that north-easterly surface-stream which is the necessary consequence of the continual flow of the great westerly equatorial current against the barrier formed by the American continent. . . . . Captain Maury's views seem only to require the mainspring or starting-force towards the west which has been here suggested, to supply a complete, efficient, and natural explanation of the whole series of phenomena presented by the great ocean currents.'

Four or five months later, and while the results on which Dr. Carpenter subsequently based his theory of the oceanic circulation were as yet unpublished, I drew attention in the columns of the Daily News to the comparatively limited extent of the influences due to polar cold. This cause, I pointed out, ' scarcely has any influence in latitudes lower than the parallel of 50 degrees.'

In the year 1869 Dr. Carpenter was first led to advocate the theory that the continual descent of cold water in the Arctic Seas is the mainspring of the system of oceanic circulation. He reasoned that the Arctic Seas being exposed to great cold, the surface water 'descends in virtue of its reduction in temperature and increase of density, its place being taken, not by the rising up of water from beneath, but by an inflow of water from the neighbouring area; and since sea-water becomes continually heavier in proportion to its reduction of temperature, this cooling action will go on without the check which is interposed in the case of fresh water.'1 Thus the water becoming denser and heavier will descend, and 'there will be a continual tendency to the flowing off of its deepest portion into the warmer area by which the polar basin is surrounded; producing a reduction in the level of the polar area, which must create a fresh indraught of surface-water from the warmer area around to supply its place. This, in its turn, being subjected to the same cooling action, will descend and flow off at the bottom, producing a fresh reduction of level and a renewed indraught at the surface.'