The differences between the mean results given above are shown by the following table, in which the signs show the excess or deficiency of the southern as compared with the northern hemisphere :

Dr. Hann has carefully discussed the question as to the comparative mean temperatures of the two hemispheres in a paper published in the proceedings of the Vienna Academy, the substance of which is given in his Klimatologie, pp. 89, et seq.; and it is difficult to refuse assent to his conclusion that so far as the available evidence goes, it shows that the mean temperature of both hemispheres is equal.

I find, then, that the same train of reasoning by which Mr. Croll has sought to explain the occurrence of glacial periods by changes in the eccentricity of the earth's orbit, and the precession of the equinoxes, leads us to conclusions respecting the climatal condition of the different parts of the earth, at the present amount of eccentricity, which are altogether opposed to the results of observation; and I am driven to the conclusion that the causes which he has adduced have not the predominant influence which he has attributed to them, and that there must be other agencies to which he has not assigned their due importance, but which are adequate to counteract the efficiency of those which, as observation proves, fail to achieve the effects anticipated from them.

I am far from pretending to be able to analyze completely the complex agencies which, by their mutual action, determine the climate of different parts of the earth, but I may briefly refer to two of them. Foremost of these is the relative distribution of land and sea, for a due appreciation of which we are indebted to the great work of Sir Charles Lyell. It is unnecessary here to discuss how far his view of the probable amount of change in past geological epochs may, in the present state of our knowledge, be subject to limitation. Mr. Wallace, who is the most strenuous supporter of the modern doctrine of the permanence of the present continents and ocean basins, recognizes the theoretical correctness of Lyell's views, and admits that changes of level great enough to cause profound modifications

of climate have actually occurred. Notwithstanding recent objections, it appears to me that Darwin's hypothesis as to the subsidence of a great tract in the Southern Pacific is that which best accounts for the existence of the countless coral islands in that region; nor is the probability of a nearly continuous barrier of volcanic islands across the Atlantic to be completely dismissed. That such changes would have largely affected the climate of the earth cannot, I think, be doubted.

If I may venture to express my own view on this difficult subject, I must say that, although it has not been overlooked by the able men who have discussed it, the paramount importance of aqueous vapour as an agent for modifying climate has not yet been fully recognized. Mr. Croll has constantly discussed the phenomena of ocean-currents, as if their chief function were to affect climate by heating or cooling the surrounding air, which is thence diffused over the land surfaces, and he has devoted little attention to the effects of evaporation from the sea, and the subsequent condensation in some other region of the vapour produced. When we remember that as much heat is consumed in the conversion of one cubic mile of water into vapour as would raise the temperature of nearly ninety-seven cubic miles of water by 10° Fahr., we get some measure of the vast power of vapour as a vehicle of heat. Admitting, as I am disposed to do, that 166,000 cubic miles of water are annually conveyed northward by the Gulf-stream, and suffer an average loss of 20° Fahr. before returning to the torrid zone, I must point out that the entire heat requisite to maintain this great volume of water at the higher temperature would be consumed in the conversion of 3433 cubic miles of water into vapour. In point of fact, I believe that more than one-half of the quantity specified is expended in evaporation, and that the cooling of the waters of the Gulf-stream is mainly due to this agency. To follow the vapour thus produced, to ascertain where it is condensed, and where the heat disengaged in the act of condensation becomes available to raise the temperature of the air, is a task which is beyond our present resources; but it is one which must be performed before we can reason with any confidence as to the ultimate distribution of the heat carried by the Gulf-stream or any other ocean-current. Whatever part of the vapour produced by evaporation from the Gulf

stream goes to supply the rainfall of Western Europe, or to form snow in the arctic regions, acts as a vehicle to transfer heat from the tropics to the temperate and frigid zones. But it is more than probable that a large part of the vapour in question is carried back to the torrid zone, and that some of it is even restored to the southern hemisphere. The south-eastern branch of the Gulf-stream flows, at least partially, into the area of the north-east trade-winds. These winds reach the lower region as cold and very dry winds. As they advance towards the equator, and are gradually warmed, their capacity for aqueous vapour constantly increases, and there can be no doubt that in both hemispheres the trade-winds bear with them a large share of the vapour which goes to supply the heavy rainfall of the tropics.

In the Pacific region we have direct evidence to this effect, in the fact that in Hawaii, and elsewhere, the side of the islands exposed to the trade-winds is that of heavy rainfall, and is generally covered with forest. No sufficient data exist for estimating the amount of vapour thus carried back to the tropics from high latitudes on both sides of the equator, nor the amount of heat set free by its condensation; but we may form some conception of its probable amount by considering that at the moderate estimate of a mean annual rainfall of seventy-two inches for the portion of the globe between the tropics, this amounts to a yearly fall of 88,737 cubic miles, and that we can scarcely reckon the share of this great volume of water supplied by evaporation from the same part of the globe at more than one-half. Still less is it possible to calculate the amount of vapour annually transferred from the northern to the southern hemisphere, which goes to neutralize the apparent effect of the diversion of portions of the equatorial waters to the north side of the line. In the Atlantic basin it is probable that the larger part of the rainfall in the region including and surrounding the Gulf of Mexico and the Caribbean Sea is supplied by vapour carried from the temperate zone by the north-east trade-winds. There is some reason to believe that a portion of the rainfall of the great basin of the Amazons, south of the line, is also supplied from the same source. Several travellers report that during the rainy season the prevailing winds are from the west and north-west, the latter being especially predominant at

Iquitos, about 4° S. latitude, and 1600 miles from the mouth of the river.

In tropical Australia the rainy season falls during the prevalence of the north-west monsoon, and we cannot doubt that this is mainly supplied by vapour carried from the northern hemisphere. Another region wherein the same phenomenon is exhibited on a large scale is the central portion of Polynesia, extending from the Feejee to the Society Islands over a space of at least twenty degrees of longitude. Over that wide area, as far as about twenty degrees south of the line, the regular south-east trade-wind prevails only in the winter of the southern hemisphere, while during the rest of the year, especially in summer, north and north-east winds have the predominance. Taking the mean of three stations in the Feejee Islands, of which the returns are given by Dr. Hann, I find in round numbers the very large amount of 150 inches for the mean annual rainfall, of which 105 fall during the seven months from October to April, while the five colder months from May to September supply only forty-five inches of rain. There can be little doubt that the larger part of the 105 inches falling during the warm season is derived from the northern hemisphere.

I by no means seek to account fully for the apparent contradiction between the results of theory, as developed by Dr. Croll, and the actual distribution of heat over the earth as proved by observation; but I venture to think that I have shown reason to doubt the possibility of drawing absolute conclusions as to the results of astronomical changes until we shall have fuller knowledge than we now possess of all the agencies that regulate climates.

Before concluding these remarks, I will notice one other branch of the argument in regard to which I am unable to concur with Mr. Croll. As we have seen, the essential point in his theory as to the modus operandi of changes of eccentricity, and the relative position of the poles, on the distribution of temperature, is that the currents of the equatorial zone are driven towards the pole which has the summer in aphelion, and that the cause of this shifting of the currents depends on the greater strength of the trade-winds in the hemisphere which has the winter in aphelion; the strength of the trade-winds in turn depending on the amount of difference of temperature

between the equatorial and the colder zones. Taking the surface of the earth generally, the trade-winds of the southern are probably stronger than those of the northern hemisphere, and, if it were true that the south temperate and frigid zones were colder than those of the other hemisphere, it would be allowable to argue that the greater difference of temperature as compared with the equatorial zone was the cause of the greater strength of the trade-winds. But we now certainly know that the southern hemisphere between latitudes 45° and 55° is considerably warmer than the corresponding zone of the northern hemisphere, and we have good grounds for believing that the mean temperature of the whole hemisphere south of latitude 45° is higher, and certainly not lower, than that of the same portion of the northern hemisphere. We are therefore not justified in explaining the greater strength of the southern trade-winds by a greater inequality of temperature between the equator and the pole.

In my opinion the cause of this predominance of the southern trade-winds is to be sought in the fact that the southern is mainly a water hemisphere, while the northern is in great part a land hemisphere. In the south, the great currents of the atmosphere flow with scarcely any interruption, except that caused by Australia, where, in fact, the trade-winds are irregular, and lose their force. In the northern hemisphere the various winds originating in the unequal heating of the land surface interfere with the normal force of the trade-winds, and weaken their effect.

In connection with this branch of the subject, I may remark that the belief in the greater cold of the southern hemisphere mainly rests on the fact that all the land hitherto seen in high latitudes has been mountainous, and is covered by great accumulations of snow and ice. But this does not in itself justify the conclusion that the mean temperature is extremely Ow. It is true that the fogs which ordinarily rest on a snowcovered surface much diminish the effect of solar radiation during the summer in high latitudes, but this is compensated by the great amount of heat liberated in the condensation of vapour. The only part of the earth which is now believed to be covered with an ice-sheet is Greenland, but the mean of the observations in that country shows a temperature higher by at