This brings me to Dr. Carpenter's second objection, that if the removal of equatorial water draws in polar water from the bottom, the whole intermediate stratum should first rise towards the surface. I do not hold the view thus demolished, but simply that the inflow is from below. The question whether the inflow would be from above or below was dealt with by me in a paper on 'Oceanic Circulation' in the Student for July 1868. I do not urge this as a proof that Dr. Carpenter's objection is invalid. My reasoning may admit of being refuted. But I wish to show that the objection is not a new one to me. The inflow may be from below without being from the bottom. If it were from the bottom it would not have the effects I have ascribed to it, that is, it would not result in a west-wardly-flowing current. What I conceive is that since the whole tropical and equatorial area is a region of excessive evaporation (as surely no physicist will deny), there is over the whole region a depression of the ocean level. This depression may be, or rather must be, exceedingly minute; but the total quantity of water thus, as it were, wanting, must be enormous. The difference must by the laws of fluid equilibrium be supplied, and though the immediate supply in equatorial regions may come from tropical regions, the actual source of the total supply must be sought for in higher latitudes. That the water drawn in under these circumstances would traverse the surface of the Atlantic, is by no means proved by the fact that the eminent mathematicians cited by Dr. Carpenter consider that an in-draught to replace water 'swept off from the surface,' by trade-wind action would be a surface current. The two cases are wholly dissimilar. I must, however, admit that my case is one of extreme difficulty regarded as a problem in hydrodynamics. It is so difficult that I do not believe it can be solved even after the very imperfect fashion in which hydrodyna-mical problems have hitherto perforce been dealt with. When the physics of hydrodynamics have been treated by mathematicians like the physics of astronomy, or rather when they can be so treated, it may be possible to deal with this problem. Unless I greatly mistake, however, in such a then, we shall find a never.

I do not see how the action of the cause I have considered is affected by the circumstance that the equatorial heat does not show any effects below 200 fathoms; for the cause is in its very nature a surface one. But I would remark that so far as continuity of action is concerned, the equatorial heat seems at least on a par with the polar cold. For as the aqueous vapour rises it finds its way to regions where the atmospheric circulation is at work to carry it away (it is only the surplus quantity which is condensed into clouds, and even these are in great part carried away); and thus the process of evaporation can hardly be exhausted. Even at night, though in a modified manner, the evaporation must continue. But the action of the polar cold, though it is continuous in the sense that the increase of cold extends to great depths, yet has this great difficulty to contend with, that the descending water must perforce wait until room is made for it by the slow removal, the creeping away, as it were, of that which it replaces. That this cause, per se, can ever become one of sufficient activity1 to generate a complete system of vertical oceanic circulation seems at the least open to grave question. It appears to me also that when applied to the North Pacific this theory fails. Very little water can pass through Behring's Straits, and beyond Behring's Straits there is an island-locked and shallow sea of enormous area, altogether unlike the deep North Atlantic.

I would further point out that the interesting fact above mentioned, namely that the equatorial heat exerts no perceptible effect at a depth exceeding 200 fathoms, is in reality almost a necessity for my theory. For if the whole of the equatorial ocean were heated, and, therefore, of reduced specific gravity, the water arriving from higher latitudes would flow to the bottom, and so have to force up the intervening strata, in order to produce the observed effects; and this may be regarded as impossible. As it is, such colder and heavier water would be in dynamical equilibrium within a very short distance of the surface.

Next, as to the question of rainfall. Dr. Carpenter considers that I have overlooked the considerations (1) that the rainfall of Europe and North America may be accounted for by the evaporation in the Mid-Atlantic, beyond the region of the trade-winds, say between 20° and 40° north latitude; and (2) that there is an enormous rainfall in the region of equatorial calms, which Sir John Herschel attributes to the deposit of waters taken up by the N.E. and S.E. trades. To this I must reply that in my essay on Rain in the 'Intellectual Observer' for December 1867, I have weighed the whole question of rainfall at least with great care, and with constant reference to the best sources of information. One circumstance I there note which seems at a first view (or rather viewed as Dr. Carpenter appears to consider the matter) much more fatal as an objection to my theory than either of those noted by Dr. Carpenter; viz., that according to the observations of Humboldt and others, the annual rainfall is at a maximum at the equator, and diminishes with increase of latitude. But the whole question is, where does all this rain come from ? If it comes from tropical and equatorial evaporation it will surely not be argued that what falls in or near the place of evaporation itself, represents the total amount of such evaporation. It is unquestionable, I conceive, that the rainfall is only the excess of the aqueous vapour poured so copiously into the air from the whole of this region. It is the quantity which the air, as it were, rejects. It is a matter of little importance where the rainfall of higher latitudes comes from, though it should be noticed that the views of Dove, Kaemtz, and other leading meteorologists respecting the winds and rains of high and low latitudes, support my remark about the great rivers.