of the American shores on the latitude of London, the temperature is 10° below the normal. The stoppage of all ocean and aërial currents would therefore lower the temperature there only 20°.

It is at the equator and the poles that the great system of ocean and aërial currents produces its maximum effects. The influence becomes less and less as we recede from those places, and between them there is a point where the influence of warm currents from the equator and of cold currents from the poles exactly neutralize each other. At this point the stoppage of ocean-currents would not sensibly affect temperature. This point, of course, is not situated on the same latitude in all meridians, but varies according to the position of the meridian in relation to land, and ocean-currents, whether cold or hot, and other circumstances. A line drawn round the globe through these various points would be very irregular. At one place, such as on the western side of the Atlantic, where the arctic current predominates, the neutral line would be deflected towards the equator, while on the eastern side, where warm currents predominate, the line would be deflected towards the north. It is a difficult problem to determine the mean position of this line; it probably lies somewhere not far north of the tropics.

CHAPTER III.

OCEAN-CURRENTS IN RELATION TO THE DISTRIBUTION OF HEAT

Influence of the Gulf-stream on the Climate of the Arctic Regions.-Absolute Amount of Heat received by the Arctic Regions from the Sun.-Influence of Ocean-currents shown by another_Method.-Temperature of a Globe all Water or all Land according to Professor J. D. Forbes.-An important Consideration overlooked.-Without Ocean-currents the Globe would not be habitable. Conclusions not affected by Imperfection of Data.

Influence of the Gulf-stream on the Climate of the Arctic Regions. Does the Gulf-stream pass into the arctic regions? Are the seas around Spitzbergen and North Greenland heated by the warm water of the stream?

Those who deny this nevertheless admit the existence of an arctic current. They admit that an immense mass of cold water is continually flowing south from the polar regions around Greenland into the Atlantic. If it be admitted, then, that a mass of water flows across the arctic circle from north to south, it must also be admitted that an equal mass flows across from south to north. It is also evident that the water crossing from south to north must be warmer than the water crossing from north to south; for the temperate regions are warmer than the arctic, and the ocean in temperate regions warmer than the ocean in the arctic; consequently the current which flows into the arctic seas, to compensate for the cold arctic current, must be a warmer current.

Is the Gulf-stream this warm current? Does this compensating warm current proceed from the Atlantic or from the Pacific? If it proceeds from the Atlantic, it is simply the

warm water of the Gulf-stream. We may call it the warm water of the Atlantic if we choose; but this cannot materially affect the question at issue, for the heat which the waters of the Atlantic possess is derived, as we have seen, to an enormous extent from the water brought from the tropics by the Gulf-stream. If we deny that the warm compensating current comes from the Atlantic, then we must assume that it comes from the Pacific. But if the cold current flows from the arctic regions into the Atlantic, and the warm compensating current from the Pacific into the arctic regions, the highest temperature should be found on the Pacific side of the arctic regions and not on the Atlantic side; the reverse, however, is the case. In the Atlantic, for example, the 41° isothermal line reaches to latitude 65° 30′, while in the Pacific it nowhere goes beyond latitude 57°. The 27° isotherm reaches to latitude 75° in the Atlantic, but in the Pacific it does not pass beyond 64°. And the 14° isotherm reaches the north of Spitzbergen in latitude 80°, whereas on the Pacific side of the arctic regions it does not reach to latitude 72°.

On no point of the earth's surface does the mean annual temperature rise so high above the normal as in the northern Atlantic, just at the arctic circle, at a spot believed to be in the middle of the Gulf-stream. This place is no less than 22.5 above the normal, while in the northern Pacific the temperature does not anywhere rise more than 9° above the normal. These facts prove that the warm current passes up the Atlantic into the arctic regions and not up the Pacific, or at least that the larger amount of warm water must pass into the arctic regions through the Atlantic. In other words, the Gulf-stream is the warm compensating current. Not only must there be a warm stream, but one of very considerable magnitude, in order to compensate for the great amount of cold water that is constantly flowing from the arctic regions, and also to maintain the temperature of those regions so much above the temperature of space as they actually are.

No doubt, when the results of the late dredging expedition

are published, they will cast much additional light on the direction and character of the currents forming the northeastern branch of the Gulf-stream.

The average quantity of heat received by the arctic regions as a whole per unit surface to that received at the equator, as we have already seen, is as 5.45 to 12, assuming that the percentage of rays cut off by the atmosphere is the same at both places. In this case the mean annual temperature of the arctic regions, taken as a whole, would be about-69°, did those regions depend entirely for their temperature upon the heat received directly from the sun. But the temperature would not even reach to this; for the percentage of rays cut off by the atmosphere in arctic regions is generally believed to be greater than at the equator, and consequently the actual mean quantity of heat received by the arctic regions will be less than 5-45-12ths of what is received at the equator.

In the article on Climate in the "Encyclopædia Britannica" there is a Table calculated upon the principle that the quantity of heat cut off is proportionate to the number of aërial particles which the rays have to encounter before reaching the surface of the earth—that, as a general rule, if the tracts of the rays follow an arithmetical progression, the diminished force with which the rays reach the ground will form a decreasing geometrical progression. According to this Table about 75 per cent. of the sun's rays are cut off by the atmosphere in arctic regions. If 75 per cent. of the rays were cut off by the atmosphere in arctic regions, then the direct rays of the sun could not maintain a mean temperature 100° above that of space. But this is no doubt much too high a percentage for the quantity of heat cut off; for recent discoveries in regard to the absorption of radiant heat by gases and vapours prove that Tables computed on this principle must be incorrect. The researches of Tyndall and Melloni show that when rays pass through any substance, the absorption is rapid at first: but the rays are soon "sifted," as it is called, and they then pass onwards with but little further obstruction. Still, however, owing to the dense fogs

which prevail in arctic regions, the quantity of heat cut off must be considerable. If as much as 50 per cent. of the sun's rays are cut off by the atmosphere in arctic regions, the amount of heat received directly from the sun would not be sufficient to maintain a mean annual temperature of -100°. Consequently the arctic regions must depend to an enormous extent upon ocean-currents for their temperature.

Influence of Ocean-currents shown by another Method. That the temperature of the arctic regions would sink enormously, and the temperature of the equator rise enormously, were all ocean-currents stopped, can be shown by another methodviz., by taking the mean annual temperature from the equator to the pole along a meridian passing through the ocean, say, the Atlantic, and comparing it with the mean annual temperature taken along a meridian passing through a great continent, say, the Asiatic.

Professor J. D. Forbes, in an interesting memoir,* has endeavoured by this method to determine what would be the temperature of the equator and the poles were the globe all water or all land. He has taken the temperature of the two meridians from the tables and charts of Professor Dove, and ascertained the exact proportion of land and water on every 10° of latitude from the equator to the poles, with the view of determining what proportion of the average temperature of the globe in each parallel is due to the land, and what to the water which respectively belongs to it. He next endeavours to obtain a formula for expressing the mean temperature of a given parallel, and thence arrives at "an approximate answer to the inquiry as to what would have been the equatorial or polar temperature of the globe, or that of any latitude, had its surface been entirely composed of land or of water."

The result at which he arrived is this: that, were the surface of the globe all water, 7107 would be the temperature of the equator, and 12°-5 the temperature of the poles; and were the

* Trans. of Roy. Soc. Edin., vol. xxii., p. 75.