The Coming Ice Age

A large number of geologists are of the opinion that during the whole of the Tertiary period the climate of the northern temperate and arctic latitudes was uniformly warm, without a trace of intervening frigid periods. I have before explained why the climate was made warm in the southern hemisphere during the Tertiary epoch, and how on the closing of that age, and subsequently, a considerable portion of the ocean waters had moved from the northern hemisphere into the southern.

Therefore, the northern seas during Tertiary times covered a much larger area than have obtained during periods following that mild epoch. So, when the low lands of Europe were submerged, the Baltic, Caspian, and other neighboring seas, now land-locked, were a portion of an enlarged Atlantic. Consequently, the westerly winds blew over a much wider North Atlantic than during the later periods.

Thus the high sea-level caused by such winds on its European side was greater than has since been obtained with the Atlantic of less breadth. This high sea-level, composed largely of drift water from the ancient Gulf Stream, had convenient access to the enlarged Arctic Ocean, which then covered the low plains of Northern Europe and Siberia. And owing to the trend of elevated lands north-eastward, which then formed the southern shores of the Arctic Ocean in those regions, the warm waters of the high sea-level of the Eastern North Atlantic found an easy passage into the arctic seas; for, while they moved over the European and Siberian seas to the north-east, they had the assistance of the westerly winds well into the arctic seas, from which position they were attracted across the Arctic Ocean to the low sea-level abreast Labrador and Davis Strait.

The Gulf Stream of Tertiary times comprised a much larger area than it now obtains; for with Florida and a large portion of the Gulf States submerged, and a wide, shallow sea covering the Mississippi valley and the Great Lake region, the tropical waters of the enlarged Gulf of Mexico moved from their vast high sea-level to the low sea-level abreast British America and Labrador, without being confined to the narrow Florida channel. Thus with an enlarged Gulf Stream in possession of a wide and clear passage leading northward, in connection with a mild period in the southern hemisphere, giving warmth to the southern oceans, the resources of the ancient Gulf currents for warming the northern regions were so ample and inexhaustive they were fully able to maintain a mild climate on the shores of the European seas, and also on the shores bordering the Arctic Ocean, during the Tertiary epoch.

Furthermore, the Humboldt current, which had its rise in the mild southern seas of that age, mingled its warmth with the equatorial current of the Pacific, which in turn gave its warmth to the Japanese current. Therefore, the latter stream under such conditions was competent to maintain a mild climate on the North Pacific coasts.

The origin of a cold period in the northern hemisphere was largely owing to the changed condition of the northern oceans following the close of the Tertiary epoch. The movement of the ocean waters into the southern hemisphere lessened the area of the Arctic and North Atlantic Oceans, and brought them to their present reduced limits, and also diminished the volume of the Gulf currents.

This great geographical change, in connection with a cold period progressing in the southern hemisphere, and so increasing the coldness of the Japanese current, and the cold antarctic currents, previously explained, which set northward on the bottom of the sea through the torrid latitudes even into the North Pacific and North Atlantic Oceans, were altogether sufficient to cause conditions favorable for the advancement of a cold period in northern latitudes. Besides, with reduced northern oceans and a diminished Gulf current, conditions were favorable for an independent circulation of the arctic waters, such as is being carried out at the present time. Hence an explanation of the movements of the ocean waters of to-day will explain the conditions which caused the northern ice periods in times past, as well as those to come in a future age. Although the conditions are such that the independent circulation of the arctic waters cannot be so well performed as the independent circulation of the southern ocean, still the open arctic channels are able to prevent the tropical Gulf Stream water from largely entering the higher northern latitudes. For it is certain that the prevailing westerly winds blow the surface waters of the North Atlantic away from the eastern shores of North America from Georgia to Labrador.

Consequently, the low sea-level thus caused attracts the waters of the Arctic Ocean southward through Baffin’s Bay and Davis Strait, and likewise down the east coast of Greenland, thus surrounding that large island with an arctic temperature, and so causing it to become a land of glaciers, which are constantly launching icebergs into the sea to cool the waters of the northern oceans. The tropical waters of the high sea-level of the Gulf of Mexico also seek the low sea-level abreast the American coast, thus causing the Gulf Stream. This great ocean current, being the main conveyer of tropical heat into the high latitudes of the North Atlantic, calls for particular notice. The great gravity currents, of which the Gulf Stream is one of the most conspicuous, are moved by small gradients.

Hence the gradient which causes the Gulf Stream waters to move out of the Florida passage is small. The levellings which have been made place the surface waters of the Gulf of Mexico as being about one metre higher than the Atlantic abreast New York, the pressure of the higher Gulf waters toward the low level of the Atlantic being nearly equal in the narrow Florida channel from the surface to the bottom of the stream. Therefore, according to descriptions given by Commander Bartlett, the warm stream moves like a river over the hard level floor of the channel; but to the northward of the Bahamas, abreast Cape Hatteras, the stream spreads out in fanlike form, and flows over a bed of cold water of great depth.

A bed of cold water is found to cover the bottom of all the deep oceans that are accessible to the antarctic seas, through which the cold water is mostly supplied, as I have before pointed out.

But the cold water which underruns the Gulf Stream is probably furnished by the arctic waters which move down Davis Strait and the east coast of Greenland. The Gulf Stream, as it widens and becomes more shallow, is, through its exposure to the westerly winds, gradually converted into a drift current; and in this way its surface waters are forced over abreast the shores of Western Europe, where it imparts its warmth to a wide region, and also causes a high sea-level. A portion of the waters of this high sea-level turn southward to replenish the waters which have been moved by the trade winds from the eastern tropical North Atlantic over into the Caribbean Sea and Gulf of Mexico, while its northern and smaller portion mingles with the Arctic Ocean waters north of Europe. These latter waters, having escaped from the westerly wind-belt, and acquired a high sea-level, and also made cool on mingling with the icy arctic seas, lose a part of their bulk on becoming chilled by sinking and returning in under-currents to the seas from which they were forced by the south-westerly winds; while the larger remaining surface waters set across the Arctic Ocean over to the northern coast of Greenland, and so down the east and west coasts of that large island to the low sea-level abreast the American coast, where the cold waters not only crowd the Gulf Stream from the shore, but they also sink under it, and form the vast bed of cold water over which the Gulf currents flow. This cold underflow of water southward probably joins the deep antarctic currents south and south-east of the Bermuda Islands, and returns to the tropical latitudes a portion of the water that is carried into the Arctic Ocean by the Gulf Stream.

There are times during the late summer and early fall months when the arctic channels are considerably obstructed by icebergs, and the low sea-level of Davis Strait and Baffin’s Bay, with the assistance of occasional south-east winds, is able to attract the temperate waters of the Atlantic as far north as the Arctic Circle. Also from the same cause the icy waters which flow down the east coast of Greenland are attracted along its southern and south-western shores into Davis Strait.

Yet at the same time the icy waters which flow from Smith’s Sound and other arctic channels move in a counter-current down the westerly side of Baffin’s Bay and Davis Strait, and so carry the icebergs and field-ice past Labrador and Newfoundland well on to the borders of the Gulf Stream. And, according to Lieutenant Maury, the westerly gales of the winter months force the temperate waters of the Atlantic, which pertain to the Gulf Stream, several degrees away from the south-east coast of Greenland. Therefore, during such seasons the surface waters of the returned arctic currents, which flow down the east coast of Greenland and Davis Strait, are drifted past Southern Greenland and Iceland, and so onward into the arctic seas, north of Europe. Thus the arctic waters maintain an independent circulation sufficient to largely exclude the Gulf Stream from the arctic seas, and surround Greenland with an arctic temperature; and it is on this account glaciers have formed on Greenland and other arctic shores, and such glaciers are probably increasing, as every iceberg launched from the frigid lands and floated to the lower latitudes lowers somewhat the temperature of the North Atlantic, and so causes conditions favorable for larger accumulations of ice on the arctic shores.

Yet it is probable that an ice period extending over the northern temperate zone could not be perfected by this process alone, should the tropical and southern oceans maintain their present temperature. But, with the assistance of a frigid period in the southern hemisphere to cool the ocean waters, and thus lower the temperature of all tropical currents, including the Gulf Stream and Japan currents, an ice age could be brought about in the northern hemisphere equal in intensity to the glacial periods of the past.

And, when we know that a considerable portion of the heat carried into the northern latitudes by tropical streams is largely derived through the mingling of the waters of such currents with the warm waters of the southern tropical oceans, it is evident that the ice periods of the northern and southern hemispheres were concurrent; although the culmination of the northern frigid period would be somewhat later than the perfected southern ice age, on account of the northern seas requiring the assistance of the cold oceans of the southern hemisphere to perfect a northern ice age.

The small area of the northern seas, compared with the southern oceans, and the wide mingling of the ocean waters of the hemispheres, make it evident that the comparatively scanty northern seas could not bring about or maintain either a frigid or mild period in opposition to the superior oceans of the southern hemisphere.

On the consummation of an ice period in the northern hemisphere heavy glaciers covered the larger portion of its continents and islands, which added so much weight to the northern lands as to attract the waters of the southern oceans into the northern latitudes, as I have before explained.

Thus, when the ice was mostly melted from the lands of the southern hemisphere, the heavy ice-sheets that remained on the extensive northern lands would still continue to attract the warm waters of the southern seas into the northern oceans; and in this way the Japanese and Gulf currents would gain a higher temperature and greater volume, and thus add to their ability for melting the northern glaciers wherever they were able to flow, and so hasten the growth of a mild era in the northern hemisphere.

And it seems reasonable to suppose that there was more water in the northern hemisphere on the ending of its ice period than at this age; yet it appears that it was returned to the southern hemisphere during a short period by the prevailing winds in the manner which I have previously explained.

Therefore, there are but few traces of such flowage to be found in the glacial drift, especially with the scarcity of marine life after the rigor of a frigid age.

An article in Science, July 5, 1895, written by Agnes Crane, states that Professor Joseph Prestwich has recently contributed a suggestive memoir on this subject to the Philosophical Transactions of the Royal Society. It treats of the evidence of a submergence of Western Europe and the Mediterranean coasts at the close of the glacial period; and in a previous paper communicated to the Geological Society of London, in 1892, the author gave evidence, deduced from personal observation, of the submergence of the south of England not less than a thousand feet, at the close of the glacial epoch.

Since that time the flood of water which flowed all of the low lands of the high northern latitudes has been returned to the southern seas, because of the force of the prevailing winds in connection with the great oceans which open so widely toward the south, the force of the winds being assisted through the attraction caused by the difference of temperature in the surface waters of the vast southern temperate oceans and the antarctic seas, and in this manner bringing about the geographical conditions of to-day which favor the return of another ice age.

It is said by those who attribute the great currents of the ocean to the rotation of the earth that the winds have little to do in causing such currents as the Gulf Stream. But my impression is that the southern portion of the Gulf Stream waters, after being drifted by westerly winds over abreast Europe, are attracted to the low sea-level in the vicinity of the Canary Islands, to be moved by the trade winds toward the equatorial calm belt and the West India Islands. And during my many months’ cruising over these seas I have had my attention directed to the singular action of the surface waters, while being impelled by the trade winds toward the West India sea; for during the first fifteen hundred miles of their passage they are moved by the prevailing easterly winds without much apparent resistance or unusual disturbance. But on nearing the longitude of Cape St. Roque, and having acquired a high sea-level from which there is no easy or wide outlet, the impelled surface waters begin to rebel against the forceful winds, and cause a remarkable commotion in the shape of tide-rips and white-capped ripples, which extend from the equator in a northerly direction to the latitude of about 19° north, thus crossing the central portion of the north-east trade-wind belt, with a breadth of over three hundred miles, as shown on map No. 2.

This disturbed region where the winds and waters conflict is the probable fountain-head of the Gulf Stream. The reason why the surface waters of this disturbed portion of the Atlantic do not flow peacefully along through the West India passages into the Caribbean Sea and Gulf of Mexico is because of their narrow outlet at the Florida channel. For it is mainly through this narrow channel that the vast waters of the tropical high sea-level are attracted to the low ocean-level of the Western North Atlantic.

Thus it seems that the great fountain-head of the Gulf Stream is situated between the wide tide-rips and the Caribbean Islands. The waters from this high ocean-level enter the Caribbean Sea mainly through the several passages south of Guadeloupe; while the northern portion of the raised waters set mostly toward the north-west, and so unite with the eastern portion of the Gulf currents after they enter the Atlantic. Still, the great high sea-level which presses against the Windward Islands, being somewhat higher than the Caribbean Sea, forces its waters through the island passages in quantities sufficient to supply the Gulf Stream; and there are times when the winds are so strong and favorable that all of the passages east of Cuba conduct water into the Caribbean Sea, the cold under-waters entering the deeper channels as well as the warm surface waters. Yet the currents setting through these numerous channels are subject to fluctuations, and so also is the Gulf Stream which they supply.

That portion of the high sea-level south of Guadeloupe receives considerable assistance as a feeder for the Gulf Stream through being connected on the south by the great high sea-level abreast Brazil and the great high sea-level of the equatorial calm belt. The latter high level is caused by the trade winds, which generally blow briskly down the coast of Sahara, and also further off shore, and ending south of the Cape Verde Islands somewhat abruptly in the equatorial calm belt.

The south-east trades which blow over the Eastern and Middle South Atlantic terminate on the southern side of the calm region. Therefore, the two trade winds impel the surface waters of the tropical Atlantic from opposite directions directly toward the calm belt, and so raise its waters above the common level of the sea.

This is the opinion of the writers of the South Atlantic Directory. Still, it is probable that the high ocean-level of the calm belt is but slightly raised above the common level of the sea, on account of the trade winds having to contend against the tendency of the warm tropical surface waters to move toward the polar latitudes. The calm belt expanse which extends from Africa, where it attains its greatest width, gradually narrows as it extends westward to the longitude of Cape St. Roque, where it attains its highest sea-level, on account of the borders of its narrowing space being impelled westward by the trade winds.

The movement of the waters of this high ocean-level is mostly toward the west, forming a portion of the equatorial current of the Atlantic. The reason of its western movement is on account of its raised waters being able to supply a portion of the Gulf Stream with water which is sent off in a westerly current along the South American coast, west of Cape St. Roque into the Caribbean Sea; while, on the other hand, it joins with the great high sea-level abreast Brazil, and so unites with its great southern current. The gradient of the high sea-level of the calm belt on its southern side probably extends south of the equator, on account of the south-east trades being weak in latitudes near the equator; while on the north side the north-east trades generally blow brisk and end more abruptly, so producing a gradient of less width than that of the South Atlantic side.

It does not appear that the seas of the high northern latitudes gain an undue proportion of the tropical Atlantic waters, because of the south-east trades extending north of the equator, on account of such winds being weak, and the waters of the high sea-level of the Western North Atlantic having narrow and otherwise obstructed passages leading to its northern seas. Yet the high sea-level of the equatorial calm belt is always ready, whenever a favorable grade is formed by a monsoon or otherwise, to run off its surplus water obtained by winds and rain; and I have noticed, while cruising in these seas, that it happens at times during the northern winter months when the north-westerly gales drive the surface waters of the North-western Atlantic toward the tropical zone, and at the same time a strong north-east monsoon is prevailing along the southern coast of Brazil, the westerly currents setting past the Amazon River are reversed, and set to the south-east, while such conditions last.

For, when the summer solstice is in the south, and the north-east monsoon moves southward along the coast of Brazil, much equatorial water moves off in that direction; and during the same season the cooled Sahara has an outward flow of air toward the south, which moves more or less water from the coast of Guinea, which is easily accomplished, because the warm surface waters of that coast are inclined to join with the south equatorial stream. Consequently, the waters move from their high sea-level north of Cape Palmas, and so form the Guinea current.

The high sea-level of the equatorial calm belt of the Atlantic contains a large portion of the conserved heat of the tropical Atlantic, which at this age sends off a somewhat limited supply of warm water to the Gulf Stream, and also to the Brazil current. But, whenever the Cape Horn channel is closed or much obstructed, so causing a great low sea-level in the Southern Atlantic, the tropical waters heaped against Brazil, and the raised waters of the great calm region being one continuous high sea-level, would mostly be attracted to the vast low sea-level of the southern ocean. Hence it will be seen how large a portion of the conserved heat of the tropical Atlantic would be used to warm the high southern latitudes during a warm period in the southern hemisphere, and at the same time the head-waters of the Gulf Stream would obtain the same height as now. For we now see much of the force of the north-east trade winds lost, while maintaining so large a high sea-level to the windward of the West India Islands, which is probably capable of supplying a stream of double the capacity of the gulf current which passes through the Florida channel.

And it appears, while viewing the vast reservoirs of warm water apparently gathered by trade winds to subdue the cold of the high latitudes, that much of the energy of such winds is now lost to the world, while maintaining a vast and pent-up high sea-level which has a difficult outlet to the northern seas, and no strongly attractive low sea-level to move its waters into the oceans of the high southern latitudes. The wide waters which are banked up to the windward of the West India Islands, and cause the wide tide-rips, set mostly to the westward into the Caribbean Sea through the passages south of Guadeloupe, while the northern portion of the raised waters set mostly toward the north, and thus form the eastern boundary of the Gulf Stream, and comprise the inner circle of the great current that encircles the Sargasso Sea.

I have been informed by an old Barbuda fisherman that “the weeds which float on the surface of the Sargasso Sea grow in large quantities on the bottom of the shoal waters to the north and eastward of that island and Antigua.” Consequently, the currents of that region carry such weeds as become detached from their places of growth into the higher latitudes, where the westerly winds in the winter season drift them eastward south of Bermuda, until finally the central area of their gathering, where the most dense collection of weeds is found, is situated near the tropic of Cancer, and about 55° west longitude, as shown on map No. 2.

This position is also the centre of the great circular currents which encompass the Sargasso Sea. The comparatively few weeds which enter the Gulf Stream abreast Florida are currented to the northward of the Bermuda Islands, and from thence drifted by the westerly winds to the south-west of the Azores before entering the trade-wind belt. The weeds, on their long drift from their native shoals, hold their freshness, and continue to grow while floating on the sea for a considerable time, but at length lose their renovating properties, and in certain areas of the sea acquire an appearance of age and decay.

The Gulf Stream, and such other tropical waters as are attracted northward to the low sea-level abreast the North American coast, pass into the westerly wind-belt, and so gradually become drift currents, while being forced by the winds over to the European side of the ocean, as we have previously shown.

The vast movement of the North Atlantic waters encircling the great Sargasso Sea has often been pointed out by writers on the subject. But the central and most dense portion of the vast sea of weeds has always been placed on the charts several degrees of longitude east of its true position.

It is fifteen years since I wrote of the Gulf Stream and arctic currents as being attracted to a low sea-level caused by the westerly winds. But, as far as I know, writers on the Atlantic currents have had nothing to say of the great low sea-level caused by the westerly winds blowing the surface waters of the North Atlantic away from the eastern coast of North America, from Georgia to Newfoundland, and thus attracting the arctic and Gulf Stream waters in opposite directions, fifteen hundred miles along the North American coast. For, were it not for this low sea-level, the Gulf Stream would not be able to move so far northward as it now flows, but would spread out, were there no unevenness in the sea-level of the Atlantic, and become a drift current far south of its present northern limits. The United States government has caused surveys to be made of the Gulf Stream, and the interesting discoveries thus obtained have all been laid before the public. Still, such surveys cover but a portion of the whole round of the vast movement of the Gulf Stream water, and do not refer to the vast high sea-level of the calm belt as being one of its feeders, or to the wide disturbance of the surface waters of the tropical North Atlantic in their conflict with the trade winds, while being forced to the vast high sea-level of the Caribbean Sea and Gulf of Mexico, and so giving head to the Gulf Stream.

Thus from the foregoing explanations it will be seen that the ability of the prevailing winds to move the surface waters of the ocean away from the weather shores of continents over against the opposite leeward shores in the different wind-belts of the globe, and so cause both high and low sea-levels, is the main reason why there is an interchange of surface water between the tropical and colder zones sufficient to carry heat from the tropics to the cooler regions, and thus largely affect the temperature of the higher latitudes.

The unmistakable traces of cold periods having occurred in both hemispheres have given rise to an ingenious astronomical theory to account for their origin. According to this theory the ice periods in the two hemispheres were consecutive; and it is admitted by its supporters that, should it be shown that the frigid periods in the northern and southern hemispheres were concurrent, the astronomical doctrine would have to be abandoned.

It is impossible for a person who is acquainted with the great surface currents of the several oceans to conceive how a mild period could be maintained in the northern hemisphere with a frigid period existing in the southern hemisphere. A frigid period in the latter hemisphere necessitates a cold temperature for the superior oceans of the globe south of the equator. With this vast area of water reduced to a chilling temperature, it seems impossible for the inferior waters of the northern latitudes to maintain sufficient warmth to favor a mild period in the northern hemisphere, especially with both hemispheres receiving an equal annual amount of the sun’s rays. The great Humboldt current, having its rise in the southern ocean west of Cape Horn, would during a southern frigid period greatly lower the temperature of the vast equatorial stream in the Pacific Ocean. Consequently, the Japanese stream, which branches off from the equatorial current into the North Pacific, would be cooled to such a degree that it would be unable to maintain the mild climate on the shores of the North Pacific which extensive lands now enjoy. Furthermore, during a cold period in the southern hemisphere the temperature of the Gulf Stream would also be greatly lowered by the great South-eastern Atlantic return current, which is caused by the south-east trade winds impelling the surface waters of that region into the equatorial latitudes, such waters being replenished from the common level of the southern ocean, and so mingling the cool waters of that sea with the equatorial waters of the Atlantic during a frigid period in the southern latitudes. And it may be said that during such times the frigid Antarctic Ocean would send its cold under-currents to cool the inferior northern oceans. Even to-day the northern and southern hemispheres, through the intermingling of the waters of the northern and southern oceans, largely maintain a like temperature in their temperate zones. Therefore, when we consider the certain traces of ice-sheets having formed on South Africa and Southern Australia, and to have overrun South America above the latitude of 40° south, thus strewing the oceans of the southern temperate zone with ice that are now largely free from it, it seems that the maintenance of warm oceans in the northern hemisphere during the time of a frigid period in the southern hemisphere would be impossible.

In order to make this statement more plain, I will again refer to the importance of the great Humboldt current for cooling the waters of the North Pacific during the perfection of a southern ice age. For during such times the ocean strewed with ice west of Cape Horn, where the Humboldt current takes its rise, would impart its coldness to the Humboldt stream, while it was floating icebergs toward the equator. The equatorial current of the Pacific being a continuation of the Humboldt stream, its waters would partake of its coldness. The Japanese current, being a large offshoot from the equatorial stream, would also possess a lower temperature than it obtains at this age. Yet at this date, with the southern ice-sheets confined to the antarctic lands, it does not possess heat sufficient to prevent glaciers from flowing down to the tide-water from mountains in Alaska.

Consequently, the Japanese stream could not maintain a mild climate on the North Pacific coasts while a cold period was being completed in the southern hemispheres. Therefore, under the conditions above set forth the support of a mild period in the northern hemisphere during the existence of a frigid period in the southern hemisphere could not be carried out.

From what has been explained, it will be seen that the growth of an ice period is necessarily slow, especially in its early stage, and also that the storage of ice is carried on in both hemispheres at the same time; but I will call further attention to the southern hemisphere, because it possesses greater resources than the northern for the production of an ice age.

The independent circulation of the southern ocean waters, as before shown, turns away the tropical currents, and thus largely prevents their warm waters from entering the high southern latitudes. Consequently, the heat from the sun’s rays, and all other sources of heat included, are not sufficient to prevent ice from gathering on lands within the antarctic circle. This increasing storage of ice is only another name for the accumulation and spreading of cold, and so the increasing chillness goes on. The snow falls, and thus adds to the extension and thickness of the ice-sheets; and at the same time the spreading snow-fields reflect the heat received from the sun’s rays into space, while the cold is retained and increased in the growing glaciers.

The spreading ice-sheets having covered the land are able to flow into the surrounding seas, where their outer edges become detached and form icebergs, which float out to sea, and so scatter over the adjoining oceans. Thus their coldness is mingled with and largely preserved by the sea, while the surface water, which is carried into the southern latitudes from the northern oceans by the prevailing winds, and also such surface waters as are attracted into the antarctic seas because of the difference of temperature of the antarctic waters and the more northern seas, are on gaining the frigid latitudes made cool, and returned to the more northern seas in cold under-currents, and so chilling the vast under-waters of the great oceans of the globe, and eventually their wide surface waters also; and so the coldness increases until the ice-sheets which at first formed on polar lands are enabled to spread slowly toward the equatorial regions so long as the independent circulation of the southern ocean is maintained.

But at length the depth of the great southern ocean is diminished because of the water evaporated from its surface, and precipitated in the shape of hail and snow over the vast continents and islands of the high northern latitudes, thus adding sufficient weight to the northern lands to attract the waters of the southern seas and still further lessen their depth. Thus during such times the Cape Horn channel is so reduced as to be obstructed by the heavy glaciers and icebergs of an ice age.

Consequently, a great change is wrought in the circulation of the southern seas. For, when the Cape Horn channel is closed, the westerly winds employ their strength to force the ocean’s surface waters away from the glaciers which have filled the diminished channel. This potent action of the winds necessarily creates a great low sea-level on the Atlantic side of the obstructed strait, sufficient to attract the tropical waters heaped against Brazil by the trade winds, and the waters of the high sea-level of the equatorial calm belt, and also the equatorial waters which set along the east coast of Africa, well into the southern seas.

It will thus be seen that the conditions for the circulation of the tropical ocean waters have met with a great change.

But the temperature of the waters has been lowered by the coldness of a frigid period; and, consequently, their capability for conveying heat to the high latitudes has largely diminished. Therefore, their first inroads in the higher latitudes make small impression on the icy seas, so the early process for melting ice is exceedingly slow. But the icy southern ocean, deprived of its independent circulation, in the course of time yields to the warming invasion of the tropical waters, whose wide and increasing spread is eventually able to bring about a mild period, according to the natural methods which I have explained in the preceding pages.

And it may be said that a mild period succeeding a glacial age gained sufficient warmth to melt the ice-sheets from all lands excepting the highest mountains. For it is probable that there are lands situated in the antarctic circle sufficiently elevated even during late Tertiary times to have been above the snow-line. Therefore, the glaciers on such lands could not have melted away during mild periods succeeding an ice age. For, as has been explained, a portion of the waters of the southern seas had moved into the northern hemisphere. Consequently, the antarctic lands were raised higher above the sea-level than at this age. Hence the area of lofty land was increased above the snow-line. And, according to Dr. James Croll’s estimate, the ice-sheet at the south pole is at this age several miles in thickness. Therefore, its upper surface is above the line of perpetual snow, and could not be melted away during the warm eras succeeding glacial periods.