1 In passing I may notice that I did not suppose Sir J. Herschel to be humorous in reference to the intensity of the polar action, but in his use of the word 'emphasis.' I should not have touched on the point, did I not thoroughly sympathise with the emphatic utterance of speculative or theoretical opinions.

Now we have in the phenomena of the zone of calms a crucial test of Sir J. Herschel's theory as to the origin of the equatorial rains. It appears to me that this test altogether negatives Herschel's theory. If the moisture to which these equatorial rains are due came from the trade-wind regions, we should certainly not expect the fall of these rains to be associated in any marked degree with the progress of the equatorial day; or, if at all, then the cooler parts of the day, when the point of saturation is lower, would be the time of precipitation. With the mid-day heat would come a cessation of precipitation. As a matter of fact the contrary is the case. The sun (we are told by Dove, Kaemtz, Humboldt, Maury, Buchan, and many more) rises commonly in a clear sky in equatorial regions. As the day proceeds clouds form, and towards mid-day they grow dense. It is at noon that heavy showers fall, and towards evening the skies again become clear. Now, any one who has noticed what happens on calm summer days in any well-watered region can see that the equatorial phenomena represent the same processes on a greatly enlarged scale. On a summer's day in such regions we see how scattered cumulus clouds begin to form in early morning, become larger and more numerous as the day proceeds,, and in the afternoon begin to be transformed into cumulo-stratus. The explanation is simple. The sun's heat has caused aqueous vapour to rise into the air, until there is so much that not very far above the earth's level the saturation point is reached. The further rise of the vapour is followed by the process of condensation into clouds, much heat being given out in the process, causing the air to expand in the neighbourhood of the clouds so formed, and thus giving to these clouds their peculiar rounded tops. (At least this feature seems better explained thus than by De Saussure's theory.) Now suppose the conditions changed to those existing at the equator. The supply of vapour is very much greater, the saturation point is very much higher near the sea-surface, and the contrast between the conditions prevailing there and in the region where condensation begins is very much more marked. The air above the equatorial and tropical seas contains, in the form of invisible aqueous vapour, an enormous quantity of water; this vapour rises and extends itself, its place being continually supplied by fresh evaporation. What must happen when the process has continued for several hours, but precisely what is observed to happen ? There is an overflow, so to speak, resembling, only much more marked, that which causes the formation of our summer clouds. Enormous cloud-masses are formed, which cannot be carried away by the atmospheric circulation (very high above the calm zone), so fast as they are formed. Hence follows excessive accumulation, presently resulting in precipitation, accompanied by remarkable electrical phenomena.

But to suppose that the whole quantity of water evaporated at the equator and in tropical regions, is precipitated there in the form of rain, corresponds to such a supposition as that the water overflowing a dam includes all that has risen to the level of the dam.

I should not be greatly concerned if the result of the experiments I spoke of should not accord with my prediction. But merely to put ice in water capable of melting it, is not in any sense to represent the conditions of the actual case. The addition of water from the ice as it melts is not in accordance with these conditions. It cannot surely be maintained that the oceanic circulation depends on the addition of water from the melting of ice; and yet I apprehend that the melting of ice is no unimportant feature of Dr. Carpenter's experiment. At any rate, the ice does melt, and the movement comes to an end when all the ice has melted away. Let the ice be packed outside the arctic end of the canal, so as merely to produce a refrigeration corresponding to what actually takes place with water carried into arctic latitudes, and I conceive that a very feeble circulation would result. Under the actual circumstances, the melting of the ice produces effects much more nearly corresponding to those due to rainfall than to the mere effects of arctic cold. The very activity of the circulation shows that the water which moves towards the ice does not undergo refrigeration. Water does not cool quite so quickly. It is the melted ice-water which descends; and nothing takes place in the arctic regions which corresponds to this continued addition of water to that already circulating. Otherwise, the arctic ice would be continually diminishing, which, of course, is not the case.

It will be gathered that I agree entirely with the opinion which Sir W. Thomson expressed, as to the reason why heat is necessary for Dr. Carpenter's experiment. Heat is necessary, because the ice must be melted to make the experiment succeed. But comparing the effects of heat and refrigeration (not of heat and the continual inflow of ice-cold water), I conceive that heat would be found altogether the more effective.

Lastly, as to the wind theory of the Gulf Stream, Dr. Carpenter remarks that, so far as he knows, I am 'the only man of science in this country agreeing with Capt. Maury in attributing the Gulf Stream to some other cause than the impelling force of the trade winds.' He must be aware that there are not half a dozen students of science in this country who have expressed definite opinions on the subject after a thorough and independent inquiry into the evidence. Amongst those who maintain the wind theory there is not one, so far as I know, with whom Dr. Carpenter is in agreement. Mr. Laughton disputes the very principle of Dr. Carpenter's reasoning, holding that the change of temperature from equator to poles proceeds too slowly mile for mile to produce the effects which Dr. Carpenter indicates. Mr. Croll, in like manner, has expressed his complete dissent from Dr. Carpenter's reasoning. So also has Mr. Findlay. I believe these gentlemen to be mistaken, and I conceive that I have been able to put my finger on the precise point where their respective lines of reasoning fail. But, if Dr. Carpenter is to take general consent as an argument, and to maintain that I am wrong because he knows of no one who agrees with me, I may as well point out that he is entering into a very questionable alliance, so far as his special views are concerned. So far as I know, all the continental students of science who share our common views as to vertical circulation, reject the wind theory as solely sufficing to account for the Gulf Stream. Again, he sets Sir J. Herschel's opinion (thirty years ago) that 'the Gulf Stream is entirely due to the trade winds' as almost conclusive against me. It is, at least, not new to me, since it is cited in every paper I have written on the subject. But is there no evidence to show that Sir J. Herschel abandoned the view he formerly entertained? I would ask what Sir John Herschel implies when, in his letter to Dr. Carpenter, he writes, 'The action of the trade and counter-trade winds, in like manner, cannot be ignored; and henceforward the question of ocean currents will have to be considered under a twofold point of view.' The word 'henceforward' implies very distinctly that Sir J. Herschel was entertaining a new opinion-that is, an opinion new to him; and I think Dr. Carpenter would find it difficult to demonstrate that this new opinion would not have enforced the omission of the word entirely from the sentence quoted by Dr. Carpenter.