as far as the entire surface of the earth is considered; for it is sooner or later followed by condensation, whereby the greater part of the absorbed heat is again returned. When a piece of land or water parts with its heat by radiation into space, that warmth can never be restored to any part of the earth's surface; but whatever heat the water loses by evaporation, becomes latent in the vapor so produced, and is ultimately transferred by condensation to some other part of the globe; and hence evap oration does not constitute an agent in causing a diminution of general terrestrial temperature. Let us now suppose a sheet of water at the equator nearly surrounded by fixed boundaries, so as to form a species of immense lagoon. Its temperature, from the causes here referred to, will rapidly augment. The heat which it has acquired during the day shall have penetrated so deeply as to be incapable of being radiated backwards into space during the night, with the same facility as on the surface of a sandy plain or from the summits of a mass of vegetation. Its temperature should thus continue to accumulate up to a certain limit imposed by the conditions of evaporation, and it might ultimately attain a mean temperature superior to any which is now met at the surface of intertropical seas.

3. These views are strikingly illustrated by the phenomena accompanying the origin of the Gulf Stream. The mass of water which rushes into the Gulf of Mexico, along the southern shores of the Carribbean Sea, has already acquired a certain elevated temperature from the action of sunshine in the southern torrid zone in its passage from Cape St. Roque. In moving around the Caribbean Sea and the Mexican Gulf, these waters still continue under the influence of a tropical sun, and are constantly increasing in temperature. The islands and coasts which they happen to bathe, have no part in directly promoting this augmentation. On looking over the isothermal chart of the Caribbean Sea and Gulf of Mexico, prepared by Mr. Charles Deville,* it becomes manifest that in general the temperature decreases in going towards the land. In some places the mean annual temperature of the water close to the land is 24°5 Cen. tigrade; further out at sea it is 25°, and still further from the land it is 25°5. In other places it gradually augments from 26°, in going from the land, up to 27°-4. These results are unconnected with the influence of latitude, and they are still less explicable by the influence of centrifugal force, in driving the cooler and heavier waters towards the edges of the great current, in its semi-rotatory movement around the gulf. For in this case the law of decrease of temperature in going from the

* Annuaire de la Société Météorologique de la France, tom i, p. 160. Reduced to degrees of Fahrenheit's scale, these numbers, arranged in the same order as in the text, are 76°1, 77o·0, 77o·9, 78°.8, 81°.3.

land, should not hold on approaching the coasts of large islands situated towards the centre of the moving mass of waters. But, in such instances, it is also manifested; for on the north and south coasts of the Island of Cuba we find the isothermal lines of 26°-2 and 26°5, while the isothermals of 26°-7 and 26°-8 are situated outside them respectively.* In Mr. Deville's chart these are closed isothermals, similar to those which I have indicated on the surface of the British Islands; but as the lowest isothermals in my map are the most remote from the sea, those in his chart which exhibit the highest temperature are farthest from the land. It is thus apparent that the intertropical sea may become a storehouse of heat, by retaining much of what it receives from the sun, which, but for the physical properties of water, it would, like the intertropical land, lose by radiation into space. It is important to bear this conclusion in mind in any inquiries respecting the influence of the distribution of land and water on general climate, especially as the influence of the land seems to have been hitherto principally considered as a calorific agent.

The heating action of intertropical land has been so often discussed by writers on climate, that it is unnecessary to do more than to point out its principal agency in the production of aerial currents, by which exchanges of temperature may be promoted between different parts of the earth's surface.

In contrasting the mean temperature of the sea with that of the land in tropical climates, the want of nocturnal observations, as referred to by Melloni, is peculiarly felt. While the temperature of the one is nearly constant, that of the other is liable to considerable fluctuations; and, as our records are principally derived from diurnal observations, the results are probably too favorable to an excess of land temperature. This conclusion is confirmed by the results exhibited in Mr. Deville's map, and, in some measure, by the fact of the higher mean temperature of the entire oceanic covering of our planet compared to its atmospheric coating.

In comparing the calorific influence of the land on distant regions with the agency of the sea, it should therefore be remembered, that while the latter stores up heat and acts by night as well as by day, the action of the land is effective only as long as the sun's rays are impinging upon it.

4. Let us endeavor to apply these conclusions to the question of the influence of the distribution of land and water upon general terrestrial temperature. As the amount of solar heat received by any point on the earth's surface is a function of the latitude, it follows that the distribution of land and water at different latitudes must be studied in order to obtain its influence

* Equivalent respectively to 79°-16, 7907, 80°06, and 80°-24 of Fahrenheit's scale.

on temperature. This distribution may be supposed to take place in an endless variety of ways, of which the following three cases are the most important:

(1.) Preponderance of land towards the poles, and of water towards the equator. (2.) Preponderance of land towards the equator, and of water towards the poles. (3.) Equable distribution of land and water in polar and equatorial regions.

At the present day three-fourths of the earth's surface are covered with water, so that all the dry land has been truly characterized as an assemblage of large and small islands placed in a great ocean. If we suppose, with Sir Charles Lyell, that in the question now under consideration, the proportion of sea to land is the same as at present, each of the above three cases is susceptible of two principal divisions, according as the islands composing the land happen to be few and large, or numerous and small. If all the dry land on the globe were collected into a single vast continent, the climatological conditions of the earth, all other things remaining the same, would be very different from what would take place if the land were broken up and spread out in numberless islands. Whatever may be the supposed distribution of land and water, it is manifest that its chief influence on the general temperature at the surface of our planet, should result from the action of aerial and oceanic currents.

In the first case above referred to, the belt of equatorial ocean would probably acquire a high temperature, and although the circumpolar islands would possess very rigorous climates in their interior, portions of their coasts might be washed by heat-bearing currents, just as the northwestern coast of Europe is washed by the Gulf Stream at the present day. The superiority of the mean temperature of the ocean might, in this case, be so great that the distribution of heat over the islands should present remarkable instances of the laws found to hold good in the British Isles, and almost all of the isothermals on the land would be closed curves.t

In the second case, the ocean would acquire much less heat from the sun, and it would exercise a cooling influence on the belt of intertropical land. But as whatever evidence we possess seems to indicate that intertropical seas owe their elevated temperature not so much to the influence of thermal exchanges with the air which has passed over the adjacent land, as to the direct influence of sunshine, we may conclude that upon the whole the heat-bearing currents would, in this case, be less influential than in that which has just been considered. The heated air flowing from the equatorial land should, by the agency of winds, in some

Principles of Geology, chap. vii, 9th ed., p. 101. + See Atlantis, No. ii, p. 399.

SECOND SERIES, VOL. XXVII, No. 81.-MAY, 1859.

measure mitigate the temperature of the polar regions, but we have no reason for believing that this influence would be superior to that of the heat-bearing water currents in our former

instance.

If now we suppose the land to be equally distributed in islands between the equatorial and polar regions, we shall have conditions more or less favorable to the existence of oceanic as well as of aerial heat-bearing currents, and it seems not impossible that, under such circumstances, the entire surface of the globe may enjoy the highest possible amount of general warmth by being best circumstanced for the accumulation, retention, and distribution of the heat it receives from the sun. In this case, as well as in the first which has been considered, warm currents from the equatorial seas might freely bathe the coasts of islands in higher latitudes, thus producing similar characteristic cases of insular climate. The mean temperature of such seas being higher than that of the air over the land, the isothermal lines of the islands should be partly or entirely closed curves, having shapes dependent upon the outlines of the islands. The greater the difference of atmospheric and water temperature, the more strictly should the isothermals conform to this law. Thus it is manifest that a nearly circular island, with a surface equal to that of Labrador, and lying in the same latitude, would present a much greater diversity of climate between its interior and its coasts, if the latter were bathed by sea water having a temperature of 80° Fahrenheit, than if that temperature amounted only to 40°. As the manner in which the warm air over the water would exchange its heat with the air over the land should take place undoubtedly by circulation, it would not be easy to assign a distinct law for the difference of temperature between the interior and the coast of the island; but it seems evident that this difference should, up to a certain limit, increase with the temperature of the heat-bearing oceanic currents. A group of islands situated in high latitudes, and surrounded by currents possessing a high temperature, while receiving but a small amount of heat from sunshine, should present a series of closed isothermals, and while their interiors would be cold, their coasts might enjoy an extremely genial climate.

5. If such conditions existed at former geological epochs, we may fairly expect to find some evidence of their existence by comparing the characters of the organized beings by which the interior and the coasts of such islands were inhabited. Such geologists as have hitherto studied the diversities in structure of the fossil remains which have come under their notice, appear to have attended principally to the climatic influence of the eleva tion of the interior parts of such islands. Professor Ramsay,*

* Memoirs of the Geological Survey of Great Britain, vol. i, p. 824.

in his memoir on the denudation of Wales, after pointing out the great elevation above the sea, which portions of that region had formerly possessed, calls attention to the resulting varieties of climate that must have prevailed. "If," he says, "the climate of our latitudes, when the coasts were washed by the new, red, and liassic seas, were tropical, as is generally supposed, still on the heights indicated on the vertical sections, we have ample space for tropical and temperate zones, each probably abounding in its own appropriate forms of life. And here, in connection with this subject, it may be remarked, that in Mr. Brodie's recent work, 'A History of the Fossil Insects of the Secondary Rocks of England,' it has been stated that, with certain excep. tions, the minute size of the great mass of the insect remains seems to indicate a very cold, or at all events, a temperate climate."

This appeared to Professor Ramsay not to be in harmony with. the other fossil evidence, which proves that most of the creatures whose remains are preserved in the strata of the secondary series inhabited a tropical climate. If the interior temperature of the land, whose inhabitants apparently existed under such different conditions of climate, depended not only on the coördinate of height above the sea, but also on that of distance from the coast, in the manner here described, a more complete explanation would be afforded of these remarkable phenomena. The discovery by Mr. Strickland, in the alluvial sand of Worcestershire, of the bones of a hippopotamus, accompanied, not only by the bones of other mammalia, but by twenty-three species of fresh water and land shells, of which nineteen are existing British species, seems to show that, even at a period so recent as that of the deposit from which these remains were taken, remarkable differences of climate may have existed over a comparatively small area of land.* The strong presumptions furnished by the fossil flora, and other evidences connected with the history of earlier geological formations in favor of the existence of numerous islands scattered over an ocean enjoying a tropical temperature, should lead us to expect more of such results as are here noticed, instead of feeling surprise at the discrepancies which they seem to exhibit.

6. I shall now attempt to illustrate some of the preceding general views from the actual condition of the earth's surface. The higher mean temperature of the northern, compared to the southern hemisphere, is clearly proved and universally acknowl edged. This superior warmth is usually ascribed to the greater amount of land in the former compared with the latter. It has been apparently assumed that the surface of the dry land exercises upon the whole a far more energetic influence, in tending * Geological Society's Proceedings, June, 1834, p. 94; and Lyell, p. 76, 9th edition,