Sea and Indian Ocean, where a temperature as low as 32° Fahr. was found at a depth of 2656 fathoms. In short the North Atlantic, and probably the intertropical seas also, may be regarded, Dr. Carpenter considers, as divided horizontally into two great layers or strata-an upper warm, and a lower cold stratum. All these facts I, of course, freely admit; nor am I aware that their truth has been called in question by any one, no matter what his views may have been as to the mode in which they are to be explained.

The Explanation of the Facts. We have next the explanation of the facts, which is simply this:-The cold water occupying the bottom of the Atlantic and of intertropical seas is to be accounted for by the supposition that it came from the polar regions. This is obvious, because the cold possessed by the water could not have been derived from the crust of the earth beneath: neither could it have come from the surface; for the temperature of the bottom water is far below the normal temperature of the latitude in which it is found. Consequently the inference seems irresistible that this depression must be produced and maintained by the convection of cold from the polar towards the equatorial area." Of course, if we suppose a flow of water from the poles towards the equator, we must necessarily infer a counter flow from the equator towards the poles; and while the water flowing from equatorial to polar regions will be warm, that flowing from polar to equatorial regions will be cold. The doctrine of a mutual interchange of equatorial and polar water is therefore a necessary consequence from the admission of the foregoing facts. With this explanation of the facts I need hardly say that I fully agree; nor am I aware that its correctness has ever been disputed. Dr. Carpenter surely cannot charge me with overlooking the fact of a mutual interchange of equatorial and polar water, seeing that my estimate of the thermal power of the Gulf-stream, from which it is proved that the amount of heat conveyed from equatorial to temperate and polar regions is enormously greater than had ever been anticipated, was made a considerable time before he began to write on the subject of oceanic circulation*. And in my paper "On Ocean-currents in relation to the Distribution of Heat over the Globe"+, I have endeavoured to show that, were it not for the raising of the temperature of polar and high temperate regions and the lowering of the temperature of intertropical regions by means of this interchange of water, these portions of the globe would not be habitable by the present existing orders of beings.

*Trans. of Glasgow Geol. Soc. for April 1867. Phil. Mag. for Feb. 1867 and June 1867 (Supplement).

† Phil. Mag. for February 1870.

Phil. Mag. S. 4. Vol. 47. No. 310. Feb. 1874.

H

The explanation goes further:-" It is along the surface and upper portion of the ocean that the equatorial waters flow towards the poles, and it is along the bottom and under portion of the ocean that polar waters flow towards the equator; or, in other words, the warm water keeps the upper portion of the ocean and the cold water the under portion." With this explanation I to a great extent agree. great extent agree. It is evident that, in reference to the northern hemisphere at least, the most of the water which flows from intertropical to polar regions (as, for example, the Gulf-stream) keeps to the surface and upper portion of the ocean; but, for reasons which I have stated in my last paper*, a very large proportion of this water must return in the form of under currents; or, which is the same thing, the return compensating current, whether it consist of the actual water which originally came from the equator or not, must flow towards the equator as an under current. That the cold water which is found at the bottom of the Atlantic and of intertropical seas must have come as under currents is perfectly obvious, because water which should come along the surface of the ocean from the polar regions would not be cold when it reached intertropical regions.

The explanation hypothetical.-Here the general agreement between us in a great measure terminates; for Dr. Carpenter is not satisfied with the explanation generally adopted by the advocates of the wind theory, viz. that the cold water found in temperate and intertropical areas comes from polar regions as compensating under currents, but advances a hypothetical form of circulation to account for the phenomenon. He assumes that there is a general set or flow of the surface and upper portion of the ocean from the equator to polar regions, and a general set or flow of the bottom and under portion of the ocean from polar regions to the equator. Mr. Ferrel (Nature,' June 13, 1872) speaks of that "interchanging motion of the water between the equator and the pole discovered by Dr. Carpenter." In this, however, Mr. Ferrel is mistaken; for Dr. Carpenter not only makes no claim to any discovery of the kind, but distinctly admits that none such has yet been made. Although in some of his papers he speaks of a "set of warm surface-water in the southern oceans toward the Antarctic pole" as being well known to navigators, yet he nowhere affirms, as far as I know, that the existence of such a general oceanic circulation as he advocates has ever been directly determined from observations. This mode of circulation is simply inferred or assumed in order to account for the facts referred to above. "At present," Dr. Carpenter says, "I claim for it no higher character than that * Phil. Mag. for October 1871, p. 267.

of a good working hypothesis to be used as a guide in further inquiry" (§ 16); and lest there should be any misapprehension on this point, he closes his memoir thus:-"At present, as I have already said, I claim for the doctrine of a general oceanic circulation no higher a character than that of a good working hypothesis consistent with our present knowledge of facts, and therefore entitled to be provisionally adopted for the purpose of stimulating and directing further inquiry."

I am unable to agree with Dr. Carpenter on this latter point. It seems to me that there is no necessity for adopting any hypothetical mode of circulation to account for the facts, as they can be quite well accounted for by means of that mode of circulation which does actually exist. It has been determined from direct observation that surface-currents flow from equatorial to polar regions; and their paths have been actually mapped out. But if it is established that currents flow from equatorial to polar regions, it is equally established that return currents flow from polar to equatorial regions; for if the one actually exists, the other of necessity must exist. We know also on physical grounds, to which I have already referred, and which fall to be considered more fully in a subsequent part of this paper, that a very large portion of the water flowing from polar to equatorial regions must be in the form of under currents. If there are cold under currents, therefore, flowing from polar to temperate and equatorial regions, this is all that we really require to account for the cold water which is found to occupy the bed of the ocean in those regions. It does not necessarily follow, because cold water may be found at the bottom of the ocean all along the equator, that there must be a direct flow from the polar regions to every point of the equator. Water brought constantly from the polar regions to various points along the equator by means of under currents will necessarily accumulate, and in course of time spread over the bottom of the intertropical seas. It must either do this, or the currents on reaching the equator must bend upwards and flow to the surface in an unbroken mass. Considerable portions of some of those currents may no doubt do so and join surface-currents; but probably the greater portion of the water coming from polar regions extends itself over the floor of the equatorial seas. In a letter in Nature,' Jan. 11, 1872, I endeavoured to show that the surface-currents of the ocean are not separate and independent of one another, but form one grand system of circulation, and that the impelling cause keeping up this system of circulation is not the trade-winds alone, as is generally supposed, but the prevailing winds of the entire globe considered also as one grand system. The evidence for this opinion, however, will be considered more fully in the next part of this paper.

Although the under currents are parts of one general system of oceanic circulation produced by the impulse of the system of prevailing winds, yet their direction and position are nevertheless to a large extent determined by different laws. The water at the surface, being moved by the force of the wind, will follow the path of greatest pressure and traction,-the effects resulting from the general contour of the land, which to a great extent are common to both sets of currents, not being taken into account; while, on the other hand, the under currents from polar regions (which to a great extent are simply "indraughts" compensating for the water drained from equatorial regions by the Gulf-stream and other surface-currents) will follow, as a general rule, the path of least resistance.

The Cause assigned for the hypothetical mode of circulation.Dr. Carpenter assigns a cause for his mode of circulation; and that cause he finds in the difference of specific gravity between equatorial and polar waters, resulting from the difference of temperature between these two regions. "Two separate questions," he says, "have to be considered, which have not, perhaps, been kept sufficiently distinct either by Mr. Croll or by myself :first, whether there is adequate evidence of the existence of a general vertical oceanic circulation; and second, whether, supposing its existence to be provisionally admitted, a vera causa can be found for it in the difference of temperature between the oceanic waters of the polar and equatorial areas" (§ 17). It seems to me that the facts adduced by Dr. Carpenter do not necessarily require the assumption of any such mode of circulation as that advanced by him. The phenomena can be satisfactorily accounted for otherwise; and therefore there does not appear to be any necessity for considering whether his hypothesis be sufficient to produce the required effect or not.

An important consideration overlooked.-But there is one important consideration which Dr. Carpenter seems to have overlooked-namely, the fact that the sea is salter in intertropical than in polar regions, and that this circumstance, so far as it goes, must tend to neutralize the effect of difference of temperature. It is probable indeed that the effect produced by difference of temperature is thus entirely neutralized, and that no difference of density whatever exists between the sea in intertropical and polar regions, and consequently that there is no difference of level nor any thing to produce such a general motion as he supposes. This I am glad to find is the opinion of Professor Wyville Thomson.

"I am greatly mistaken," says that author, "if the low specific gravity of the polar sea, the result of the condensation and precipitation of vapour evaporated from the intertropical area,

do not fully counterbalance the contraction of the superficial film by arctic cold. . . . . Speaking in the total absence of all reliable data, it is my general impression that if we were to set aside all other agencies, and to trust for an oceanic circulation to these conditions only which are relied upon by Dr. Carpenter, if there were any general circulation at all, which seems very problematical, the odds are rather in favour of a warm undercurrent travelling northwards by virtue of its excess of salt, balanced by a surface return-current of fresher though colder arctic water." (The Depths of the Sea,' pp. 376 & 377.)

This is what actually takes place on the west and north-west of Spitzbergen. There the warm water of the Gulf-stream flows underneath the cold polar current. And it is the opinion of Dr. Scoresby, Clements Markham, and Lieut. Maury that this warm water, in virtue of its greater saltness, is denser than the polar water. Mr. Leigh Smith found on the north-west of Spitzbergen the temperature at 500 fathoms to be 52°, and once even 64°, while the water on the surface was only a degree or two above freezing*. Mr. Aitken, of Darroch, in a paper lately read before the Royal Scottish Society of Arts, showed experimentally that the polar water in regions where the ice is melting is actually less dense than the warm and more salt tropical waters. Nor will it help the matter in the least to maintain that difference of specific gravity is not the reason why the warm water of the Gulf-stream passes under the polar stream--because if differences of specific gravity be not the cause of the warm water underlying the cold water in polar regions, then difference of specific gravity may likewise not be the cause of the cold water underlying the warm at the equator; and if so, then there is no necessity for the gravitation hypothesis of oceanic circulation.

There is little doubt that the superheated stratum at the surface of the intertropical seas, which stratum, according to Dr. Carpenter, is of no great thickness, is less dense than the polar water; but if we take a column extending from the surface down to the bottom of the ocean, this column at the equator will be found to be as heavy as one of equal length in the polar area. And if this be the case, then there can be no difference of level between the equator and the poles, and no disturbance of static equilibrium nor any thing else to produce circulation.

Under currents account for all the Facts better than Dr. Carpenter's Hypothesis.-Assuming, for the present, the system of prevailing winds to be the true cause of oceanic currents, it necessarily follows (as will be shown hereafter) that a large quantity of Atlantic water must be propelled into the Arctic Ocean; and such, as we know, is actually the case. But the Arctic The Threshold of the Unknown Region, p. 95.