It is impossible but that on a subject so difficult and complicated as that of oceanic circulation, different views should be entertained by students of science. And it is clear that in the present stage of the inquiry no useful purpose could be fulfilled by making the problem a matter for controversy. Dr. Carpenter himself has shown that much more is to be gained by observation than by reasoning on imperfect knowledge. If I venture to remark that his deep-sea researches have led to the most important contribution which has been added for many years to our information respecting oceanic circulation, he will not, I trust, consider that I am passing beyond the bounds of controversial courtesy. But I am, indeed, not anxious to treat the matter as one for controversy in any sense. It will be perceived by those who have read my remarks on the subject, that I have rather put them forward as suggestions than as indicating theories which can be maintained with any degree of assurance, far less with conviction. Nor does it seem to me likely that one explanation can suffice to account for all the phenomena recognised in oceanic circulation. This is a case, if ever such case were, in which more causes are in operation than one; so that it may very well happen that excellent arguments can be adduced in maintenance of different views. ' If, therefore, I enter on the defence of what I have already written on this subject, it is not with the wish to show that one particular explanation of oceanic circulation is correct, and all others erroneous. If I am desirous of dealing with the considerations urged by Dr. Carpenter, it is not because they seem to him to militate against the views I have to some extent advocated. What I wish to show is that I have not addressed your readers on the subject of oceanic circulation without making myself familiar with the facts which bear upon that subject, and at the very least, with those comparatively fundamental facts to which attention has been invited.

1 This paper was written in reply to comments by Dr. Carpenter on the former paper. The nature of these comments will be inferred from my reply; in fact I quote the most important passages.

And here I would remark that one who writes so much and so often as I have had occasion to do on this and kindred subjects, is placed to some degree at a disadvantage. He cannot, on the one hand, assume that the readers of any particular essay have also read all that he has written on the subject; yet, on the other, he cannot assume that none have done so, and that he is therefore free to repeat (in a more or less modified form) much that he has formerly urged. I was, perhaps, somewhat too careful in writing for your pages to avoid touching at any length on any parts of the subject which I had more particularly dealt with elsewhere; and accordingly I have laid myself open to a method of attack, which in reality involves the suggestion that I have written without due consideration even of the elements of my subject. I have no doubt that Dr. Carpenter has no wish to imply this directly, yet indirectly it is implied in every paragraph of his reply. I shall be able to show, however, that every one of the points touched on by Dr. Carpenter had been fully considered by me-and, for the most part, several months before he had turned his attention to this subject.

First, there is the remark that I have left out of view the circumstance that if there is excess of evaporation in the intertropical area, the excess ought to show itself, as in the Mediterranean, in an increase of specific gravity, whereas the specific gravity of the equatorial water is lower than that of tropical water. Now, it is unquestionably true that the effect of evaporation is to increase the specific gravity of sea water; but it is equally true that the effect of the heat which causes the evaporation is to diminish the specific gravity. The point is considered in my essay entitled 'Is the Gulf Stream a Myth?' in the first series of 'Light Science for Leisure Hours.' 'We recognise,' I there say,' two contrary effects as the immediate results of the sun's action. In the first place, by warming the equatorial waters it tends to make them lighter; in the second place, by causing evaporation it renders them Salter, and so tends to make them heavier.' And I proceed to inquire which cause is likely to be the more effective, arriving at the conclusion that the water is made lighter. The case, indeed, appears to me to be altogether different from that of the Mediterranean Sea cited by Dr. Carpenter. In the Mediterranean we have the same heating action as on the Atlantic in the same latitudes, but not the same relatively enormous quantity of water freely communicating with the region so heated. We have, then, in the Mediterranean evaporation as everywhere else, and evaporation to the same degree, appreciably, as elsewhere in similar latitudes; but evaporation not compensated as in the open Atlantic by the effects of free communication with surrounding water. Hence we have in the Mediterranean an increase of saltness; in other words, an increase of specific gravity. And precisely because this increase takes place in the Mediterranean, whereas the water of the Atlantic in the same latitudes, exposed to the same average degree of heat, is not rendered heavier, it may be maintained not unreasonably that the water of the equatorial Atlantic being unconfined, will in like manner not be rendered heavier by evaporation. It seems to me that we have here a positive argument of great weight in favour of my views. But independently of this I would ask whether it can be questioned that enormous evaporation does take place over the equatorial area. This is what I contend for, and I should have imagined that few would undertake to deny the proposition.

In passing, I must remark that I do not adopt the distinction between equatorial and tropical water which Dr. Carpenter appears to recognise. I have in view the evaporation over an enormously larger area than he considers-no less an area, in fact, than the whole ocean between latitudes 40° north and south of the equator (at the equinoxes, and varying according to the season). It by no means follows that because the equatorial current does not cover this enormous area, therefore the relation which I have suggested as the mainspring of oceanic circulation has not that extent. On the contrary, while it is on the one hand certain that there is an excess of heat over this enormous area, it is on the other almost a necessity of my theory that the resulting current should be found running along the middle only of the great region of evaporation.