Popular Science Monthly/Volume 50/January 1897/Our Present Knowledge of the Antarctic Regions
By Prof. ANGELO HEILPRIN.
IT can safely be said that we to-day know less about the antarctic regions than of any other portion of the earth's surface. We speak vaguely of an antarctic continent stretching across the southern pole, and some have even gone so far as to locate its boundaries, and to give an estimate of its superficial area. This has been placed almost anywhere between four and six millions of square miles—therefore larger than, or nearly twice the size of, the semi-continent of Europe. But no one is in possession of the facts which would prove the existence of such a continent, although it is by no means unlikely that it exists; and if it does, we know practically nothing of the possibilities of its flora or fauna. Up to the beginning of the past year perhaps the most striking definition that could be given of so-called Antarctica was that it was a region whose land area was entirely destitute of a flora and of a strictly terrestrial fauna. Not a vestige of moss, not a shred of lichen had up to that time been discovered; not an animal, excepting aquatic birds, had been found to give life to the few patches of open country that had been seen, or to the ice that almost everywhere covered it. The observations of the Norwegians Kristensen and Borchgrevink, made in the early part of 1895, to an extent modify this dreary conception, for at least one form of cryptogamous vegetation has been found within the Antarctic Circle—on Possession Island and on the opposite Victoria Land, near Cape Adare.
If we bar out the work of the past three years (1893–’95) it can be said that nearly all the knowledge that we possess of this Antarctica dates from a period a half century back and more—to the period of the researches of Bellany, Biscoe, Dumont d'Urville, Wilkes, and Sir James Clark Ross, and to no explorer are we indebted for more information than to the last-named. These investigators have determined the existence of certain patches of land, in most cases defined by prominent mountain swellings, which appear here and there behind a great barrier or wall of ice, to which the name of "Antarctic Barrier" has generally been given. Such land areas—perhaps not in all cases positively demonstrated to be distinct from sea ice—are Victoria Land (due south of New Zealand), Wilkes Land (not improbably a series of island elevations opposite Australia, and known under the various names of Adélie Land, Clarie Land, Sabrina Land, etc.), and Graham Land (somewhat east of south of the extremity of South America). The most extended piece of coast or land line is that which has been traced southward in Victoria Land by Ross from about the seventieth to the seventy-ninth parallel of latitude, or over about six hundred geographical miles. It is only here, and in Graham Land (with the adjoining parts of Palmer Land, Louis Philippe Land, Joinville Island, Alexander Land), that our knowledge becomes at all definite.
Ross found the whole eastern coast front of Victoria to be paralleled by one or more mountain ridges of very considerable elevation, and bearing upon themselves a large number of clearly defined volcanic cones. Mount Melbourne, seemingly the highest point (with an elevation of nearly fifteen thousand feet), is described as having a prodigious summit crater. Mount Erebus (12,400 feet), the most southerly of all active volcanoes, was in violent eruption at the time of Ross's visit (January, 1841); a little to the east of it (in approximately latitude 78° 30′ south) is the extinct cone of "Terror," 10,900 feet. Beyond Mount Terror the Parry Mountains, also of very considerable elevation, and which continue to be the most southerly piece of land area that has ever been sighted, follow the generally southern trend to at least the seventy-ninth parallel of latitude, and not impossibly for a long piece beyond.
Geographers who define the contours of the presumed antarctic continent usually deflect its course eastward beyond Mount Terror so as to make it conform to the east-and-west ice barrier which barred Ross's passage farther southward; but it is significant that Ross says, "If there be land to the southward [of the barrier], it must be very remote, or of much less elevation than any other part of the coast we have seen, or it would have appeared above the barrier." This statement becomes of special importance, because elsewhere the land was clearly defined by its mountains at distances of ninety, one hundred and thirty, and even one hundred and fifty miles.
The region explored by Ross has only once been visited since—by Kristensen and Borchgrevink and their associates of the Antarctic. The Antarctic succeeded in following the route of Ross to about the seventy-fourth parallel of latitude, when, with open water still to the south, a return was made, owing to an absence of whale supply. Few facts of any consequence were added by this journey, the most important being, perhaps, the discovery of shreds of lichen on Possession Island and on the mainland of Cape Adare opposite, this being the first landing on what may properly be designated mainland. Borchgrevink confirms in almost every particular the observations of Ross, and from the two accounts we learn that Victoria Land is a region of lofty mountains, largely and perhaps almost entirely of a volcanic nature, and almost entirely buried within a mantle of snow and ice. The covering of snow and ice is not sufficiently massive to obliterate the relief of the land—differing in this respect from the interior of Greenland—and the contours of valley and mountain are well and clearly retained. Giant glaciers descend toward and into the sea, terminating in vertical cliffs of ice of one hundred, one hundred and fifty, and two hundred feet in height. A vast ice barrier of vertical cliffs, whether of glacial formation or otherwise, and retaining a nearly uniform elevation of one hundred and fifty to two hundred and fifty feet—with a reduction at one point to nearly eighty feet (or less—defines a considerable part of the north and south coast line; beyond the seventy-eighth parallel of latitude this ice barrier trends eastward for at least three hundred miles, but it is not known that any approximate coast line lies back of it with a similar trend.
Westward of the one hundred and seventieth parallel of east longitude, and situated close upon the Antarctic Circle—now to the south of it, then to the north—and forming, as it were, a continuation of Victoria Land through some seventy degrees of longitude, are a number of designated land patches (such as Clarie Land, Adélie Land, Sabrina Land), which, with the uniting ice-cliff barriers, constitute the coast line of the antarctic continent of Wilkes—sometimes also known as Wilkes Land. How much of this continuous frontage of some two thousand miles is really land no one knows. The mountain undulations mark some parts as being indisputably such; yet a reasonable doubt may be entertained regarding some of the presumed land masses of Wilkes, and it is known that one of his mountain chains was sailed over by Ross in the region of the Bellany Islands. A reference likewise to the admirable illustration of Clarie Land in D'Urville's monumental work on The South Polar Regions makes one suspicious as to the true nature of this cóte, and forces one to inquire if it is not merely a portion of the great Antarctic Barrier.
Still farther west lie Kemp Land (probably island) and Enderby Land or Island, and finally, almost due south of the South American continent, the complex of Graham and Palmer Lands, with Terre Louis Philippe, Isle Joinville, and the more recently discovered or named King Oscar II Land, which was traced in 1893 by Larsen to nearly the sixty-ninth parallel of latitude, he himself attaining 68° 10′. This series of lands, which are closely contiguous with the South Shetland Islands, is also loftily mountainous, ranging to perhaps nine thousand feet, and with volcanic cones as a dominant feature. Lying east of King Oscar II Land, which is seemingly only a portion of Graham Land, are a number of small islands, some of which, as Christensen and Lindenberg, were volcanoes in active eruption at the time of Larsen's visit. This tract of archipelago lying south of the American continent is much less snow-bound than the region about Victoria Land, large areas of bare rock being exposed both on the islands and on the mainland, especially the volcanic slopes. Not improbably the heat of the volcanic cones has much to do with keeping an exposed surface, although it can hardly be supposed that this exposure is entirely due to this cause.
The geographical and geological study of the region under consideration resolves itself into four or more lines of inquiry: 1. Have we a continent in Antarctica? 2. What is the nature of the ice covering? 3. Of what construction are the rock masses? 4. Has Antarctica ever been united with any of the major divisions of the earth's surface which we recognize as continents?
The first inquiry hardly recognizes a positive answer as yet. Wilkes was certain that in the land masses seen by him, or thought to have been seen, we had the positive marks of a vast united continent; Ross, although he had seen more continuous coast line than any other investigator, was exceedingly doubtful on this point, and considered the evidence insufficient for positive determination; Murray, the distinguished geographer of the Challenger Expedition, has gone even beyond Wilkes, and constructed the contours of what appears to him to be true Antarctica, the outlines of which are in the map on the next page. These may be approximately correct or not a matter about as difficult to disprove as to prove but it is certain that the materials upon which this construction is based are hardly sufficient to warrant the mapping. Yet it is almost positive that a vast land area—perhaps two, three, or more of them—underlies the capping of snow or ice; but whether it is entitled to the designation of continent remains to be demonstrated by future exploration. Ross strongly emphasizes the doubt as to whether all the eminences or appearances reported to be land are really such, and he himself admits—cautious observer though he was—to having been deceived on more than one occasion. Murray, again, warns us that much of the ice barrier described by Wilkes is not the true barrier (which is presumed to be the boundary to a not distantly lying terra firma), but merely the cemented pack. It is a significant fact that none of the explorers refer to a distant elevated ice cap, such as everywhere bounds the horizon of the observer looking into Greenland, or to a mountain chain which is far removed from what is assumed to be the continental border.
The ablest discussion of the physical and geographical relations of Antarctica is furnished by the late Dr. Petermann in the Mittheilungen of 1863. He there argues strongly in favor of at least a partially open South Polar Sea, whose position is located central to the great ice masses which radiate out from it, and in
Sketch Map of the Antarctic Tract, giving the more Important Points that have been named by Navigators.
seasons of disruption—the months of November to April, the southern summer—press northward as the great Antarctic Barrier, the circle of pack ice which at varying latitudes has been the bar to passage of the different exploring expeditions. The special reasons advanced for this construction of the antarctic region are essentially two: (1) The comparatively low summer temperature of the south as compared with that of the arctic regions, an indication of oceanic rather than of continental conditions; and (2) the irregularity or instability of the pack ice, the front varying in position (in a north-and-south displacement) by hundreds of miles. Thus, in approximately the meridian in which, in 1823, Weddell reached the surprisingly high latitude of 74° 15′ south, the famous navigator Cook, nearly half a century earlier, was stopped in latitude 60°; and in 1855 Captain Grant found himself confronted by the impenetrable flat-topped barrier, three hundred to five hundred (?) feet in height, in latitude 56° 50′—40′ west longitude. Both of the arguments here stated have their force, but in how far they prove their case future exploration or penetration alone can show. Ice movements similar to those of the south take place in the arctic regions, and they are largely determined by the winds and currents which sweep over or govern a virtually open sea; but it should be noted that the ice pack of the north is very different from what is commonly designated the "barrier" of the south, with its stupendous wall precipices of one hundred and fifty to three hundred (or five hundred?) feet elevation; such fronts in the arctic regions belong exclusively to isolated icebergs or to the terminal faces of ice sheets (glaciers) which debouch into the sea and terminate at no great distance from the mainland. The bounding ice walls of the northern face of Melville Bay are examples of this kind. The main pack, or that which blocks navigation in the north, is a surface, regular or irregular, which rises but little above the level of the sea, except where it is tossed up into shingles and hummocks, or into those irregular eminences which have been by some identified with the so-called "palæocrystic" ice. On the other hand, a counterpart of the southern barriers is to be found in the land terminations of some of the giant glaciers of the interior, whose "Chinese walls" have been so graphically described in the explorations of Grinnell and Grant Lands (Greely).
It is unfortunate that the term "Antarctic Barrier" should ever have come into use, as it has been made to cover a variety of structures, and has led to confusion in the interpretation of the special features which it designates. There is no question, as Murray has pointed out, that much of the so-called "barrier" of Wilkes is merely ordinary pack ice—some of it, indeed, in the brash or broken condition; therefore, considerable allowance must be made in the acceptance of that assumed girdle which is supposed to define a continent. A point that probably favors the (Petermannian) view of partially oceanic conditions within the ice is the presence of strong northwardly trending currents, which have been observed by both earlier and later explorers. Thus, at his farthest southing, in 1894–’95, in latitude 74° 10′, Captain Kristensen, of the Antarctic, met with such currents trending almost due northward opposite Victoria Land, and the question naturally suggests itself, Is this a direct current, or one that is deflected northward after taking a westerly course along the edge of the southern barrier of Ross?
Of the arguments that have been advanced in favor of considering Antarctica as a vast continent buried deep beneath its covering of snow and ice, the most plausible are those which relate to the construction and form of the oceanic bottom within the region of the southern ice and the character of the ice itself. More explicitly stated, they are: (1) The shallowing of the sea toward the so-called antarctic tract—an approach to the borders of a continent—and the occurrence of what are stated to be subcontinental or terrigenous deposits, conditions that are well emphasized by Murray; and (2) the heavy massing of ice, which could seemingly not be other than of glacial origin. Ross found the depth of water opposite the barrier which stopped his farthest passage southward reduced to two hundred and fifty and one hundred and fifty fathoms, so that manifestly there was here a true shallow; somewhat similar results were obtained at a few other points along the barrier front. But it can be pertinently asked. In what special way would the approaches to an archipelago differ from those of a continent? With this special evidence of shallowing before him, Ross still believed in the probability of non-continental conditions, and he was in a measure justified in his belief by the fact that at many other points not far from the front of the barrier the lead indicated depths of from four thousand to six thousand feet, and even more.
The massiveness of the ice is in a condition which, so far as it is known to us, belongs exclusively to glacial formation; i. e., none but land ice is known to assume this form. The evidence which it offers, therefore, favors the notion of the existence of large terrestrial areas or gathering basins. Yet it is by no means impossible, or even improbable, that with the low summer temperatures which prevail in the antarctic tracts and the continuousness of fogs and clouds, the surface of the sea might of itself, through ages of precipitation and of comparatively little melting, build itself up in mountains of ice, hundreds or even thousands of feet in thickness. This view has, indeed, been held by some physicists, and no facts that are accessible to us are really incompatible with it. The uniformity of the table surface of the ice, which appears to be uninterrupted in places for hundreds of miles, combined with the fact that it only occasionally shows an undulating or rising surface back of it to mark out a land relief, is in itself a suspicious circumstance. This is very different from what we find in Greenland, the largest area of positive glaciation with which we are acquainted, and which certainly carries with it the constructional type of a continent. Whether seen from the east, south, west, or northwest, the relief line is plain and continuous, and over the greater part of it, in clear weather, the great dome of receding ice-cap is well visible. And yet from this ice accumulation hundreds of glaciers are given off whose terminal or sea walls are of much the same height as the greater part of the Antarctic Barrier. Naturally, it can he assumed that Antarctica is much less mountainous than Greenland—may, in fact, be a gently rising or almost flat plain—and that the great length of its glaciers, which marks off a termination possibly a hundred miles or more distant from the actual border of the land, is that which prevents the land contours themselves from being seen. But are there just grounds for a contention of this kind?
The distinctiveness of the antarctic climate as compared with the arctic is found in the relations of both the summer and the winter temperatures. The high summer heat of the north, which in the few months of its existence has the energy to develop that lovely carpeting of grass and flowers which gives to the low-lying lands even to the eighty-second parallel of latitude a charm equal to that of the upland meadows of Switzerland, is in a measure wanting in the south; in its place frequent cold and dreary fogs navigate the atmosphere, and render dreary and desolate a region that extends far into what may be properly designated the habitable zone. The fields of poppies, anemones, saxifrages, and mountain pinks, of dwarf birches and willows, are replaced by interminable snow and ice, with only Here and there bare patches of rock to give assurance that something underlies the snow covering. Man's habitations in the northern hemisphere extend to the seventy-eighth parallel of latitude, and formerly extended to the eighty-second; in the southern hemisphere they find their limit in Fuegia, in the fifty-fifth parallel, fully three hundred and fifty miles nearer to the equator than where, as in the Shetland Islands, ladies in lawn dresses disport in the game of tennis. And still seven hundred miles farther from the equator, in Siberia, Nordenskjöld found forests of pine rising with trunks seventy to one hundred feet in height. Yet it must not be supposed that there are not, as is perhaps commonly assumed, gleams of warm sunshine in this inhospitable south; indeed, we have yet to learn to what extent the far south is warm or cold. Thus, Captain Kristensen, the gallant commander of the Antarctic, who made the first landing on what is assumed to be the mainland of Antarctica, asserts that on January 5, 1895, when nearly on the sixty-eighth parallel of latitude, "the sun at noon gave so much heat that I took my coat off, and the crew were lying basking in the sunshine on the forecastle" (Transactions of the Royal Geographical Society of Australasia, Victorian Branch, March, 1896, page 87); and Biscoe, writing on the 16th of January, 1831 (on approximately the sixtieth to the sixty-third parallel of latitude), states that "the temperature of the water was 34°, of the air in the shade 45°, in the sun 77°, with a corresponding general warmth to the feelings of the crew." The highest reading of the thermometer for the month of January was noted by Kristensen to be 40° F., and the lowest, 27°; fifty-three years earlier (1842) Ross found for the same month 39° and 27·5°, with a mean of 32°, thus indicating an equality almost without fluctuation.
The fact that the high south has not yet been penetrated in the winter months leaves us in uncertainty as to the winter temperatures that may prevail there; but some indications of this temperature are to be found in the records which have been obtained in the circumantarctic tract. Ross registered the absolute minimum, for the year 1842, in the Falkland Islands to be only 19·2 (-5·7° R.); but still more significant is the reading of the minimum thermometer which was left by Foster in 1829 on Deception Island, and recovered by Captain W. H. Smiley (as reported by Wilkes) in 1842, or after an interval of thirteen years. The registry was found to be -5° F. (-16·45° R.). It is true that Deception Island lies well without the Antarctic Circle, and that its insular condition must measurablv reduce the rigors of a winter climate; but even these conditions permit us to form some just estimate of what "lies beyond," and of making some interesting comparisons with corresponding localities (so far as latitude is concerned) in the north. Thus, at Fort Reliance, in North America, the mercury descends to -70° F., and at Jakutsk, in Siberia, nearly one degree nearer to the equator, to -75°; and, if we are to fully believe the registry at Verkhojansk, for the winter of 1893, the unprecedentedly low temperature of -90° was reached. But one need not make comparisons with these especially cold localities, as it is well known that at the sites of the principal commercial cities of the world the mercury at times descends to from -5° to -15° (New York and Philadelphia, 1866, 1895). On January 23, 1823, the mercury in Berlin descended to -31° F., and in Paris, on January 25, 1795, to -21°. It is perhaps just to conclude from these and other facts that the extreme winter climate of the Austral Ocean, on or about latitude 63° south, is no more severe than that of southern France, and hardly more so than that of northern Italy. And while it is doubtless true that a considerably lower marking of the thermometer would be found in the much more extreme regions of the south, or nearer to the pole, it is practically certain that nothing comparable to the cold of the opposite face of the globe exists.
In summing up the various facts that have been noted, it may be admitted that they argue rather against than in favor of continental conditions, but they are by no means sufficient to make a demonstration; indeed, the fact that such a large mass of apparently continuous ice exists circumpolarly speaks against the conclusion, and it is probably the one most significant circumstance which has led to favor the view of continentality. It must be observed, however, that the continuousness of the so-called antarctic continent rests upon somewhat insecure and far from confirmatory evidence. The open sea (with only three ice islands visible from the mast) that confronted Weddell at the seventy-fourth degree of latitude speaks volumes in its, own behalf, and is evidence of a kind which is rather strengthened than otherwise—as proving the insularity of the ice—by the subsequent findings of Biscoe, D'Urville, Ross, and Grant, who found it impossible to penetrate to within eight hundred or a thousand miles of the position reached by Weddell.
In the same year with Weddell, Captain Benjamin Morrell, Jr., sailing from New York, reached in Weddell's track (March 14th), latitude 70° 14′ south, and he also describes the sea as being "entirely free of field ice," and stated that "there were not more than a dozen ice islands in sight. At the same time the temperature both of the air and the water was at least thirteen degrees higher (more mild) than we had ever found it between the parallels of sixty and sixty-two south. We were now in latitude 70° 14′ south, and the temperature of the air was 47°, and that of the water 44°." Morrell significantly adds, "I have several times passed within the Antarctic Circle, on different meridians, and have uniformly found the temperature both of the air and the water to become more and more mild the farther I advanced beyond the sixty-fifth degree of south latitude"; and further: "I regret extremely that circumstances would not permit me to proceed farther south, when I was in latitude 70° 14′ south, on Friday, the 14th day of March, 1823, as I should then have been able, without the least doubt, to penetrate as far as the eighty-fifth degree of south latitude" (op. cit., page 67).
Not much weight can be attached to the latter part of this statement, as it is well the experience of polar navigators how suddenly the presumably "open seas" close up; yet it would be nothing short of an assumption to place a barrier where it has in fact not been seen to exist. The circumstance that navigators who have followed Morrell found impenetrable barriers north of Weddell's and Morrell's positions is no evidence of what lies southward of them, and so far as anything that we now know of to the contrary, King George IV Sea may extend quite to or even across the pole. Both Weddell and Morrell experienced winds from the south in this water; the former, writing on the 18th of February, remarks: "Not a particle of ice of any description was to be seen, the evening was mild and serene, and our situation might have been envied had it not been for the reflection that probably we should have obstacles to contend with in our passage through the ice northward." As regards the various eminences which have at different times been designated "land" (such as Kemp Land, Enderby Land, etc.), and which have so often been united together to form a continuous land mass, there is every reason to believe that many of them are merely islands, and even islands of inconsiderable extent. Thus, it is certain that if the positions assumed by Morrell and Biscoe for their respective vessels are true, then Biscoe must have sailed over a portion of what was subsequently designated Enderby Land (Enderby Island), The deceptive forms of ice have by nearly all polar navigators been at one time or another taken for rising land surface, a condition which makes doubtful the references to many of the "land "masses that have been made known to us. It is by no means unreasonable to assume, with Petermann, that Antarctica may yet prove to be a disjointed association of land and ice masses, purely archipelagic in form; in this sense, Victoria Land, which is to-day the most extensive tract known, may be merely a correspondent of the insular form of New Zealand.
The only important addition to our knowledge of true Antarctica that has been made since Ross's voyage belongs to the close of the year 1893, when Larsen penetrated, in the region of the Graham Land complex, to latitude 68° 10′ south, and brought back with him a "departure" in the geological concept of the region under consideration. The finding of Tertiary fossils (Cytherea, Natica, etc.) on Seymour Island (Cape Seymour) is the opening vista in an investigation which has heretofore been considered closed, and at once affords, to use a business term, a basis for consideration. Not less significant is the finding at the same locality of an abundance of tree remains (conifers, Araucaria?). These fragments at least show that some part of Antarctica was of the same kind of construction as the continents generally, and their special fades immediately suggests a South American relationship. Previous to 1893 the only rocks known from the ice-bound region of the far south were granites, gneisses, (and related schists), the strictly eruptive and trappean rocks, and certain red sandstones (Piner's Island—Triassic?) from a very limited area. Most (and perhaps nearly all) of the higher mountains are distinctly of a volcanic nature, and many of them bear huge craters on their summits. Ross found Erebus in eruption at the time of his visit (1841), and Larsen found the mountains of Christensen and Lindenberg Islands similarly active in 1893–’94. Kristensen and Borchgrevink, who sailed over a portion of Ross's course in 1894–’95, attaining off Victoria Land, with clear water ahead of them, latitude 74° south, confirm in almost every detail the observations of their predecessor, adding some additional facts regarding the large glaciers which descend from the heights of the Sabine Mountains. They were the first to set foot on the mainland (or main island) of Antarctica, and to them science also owes the first discovery within this realm of a rock-covering vegetation (lichens?—on Possession Island and Cape Adare).
In its relation to other continents there is reason to believe that Antarctica, whether as a continent or in fragmented parts, had a definite connection with one or more of the land masses lying to the north, and the suspicion can hardly be avoided that such connection was, if with nothing else, with at least New Zealand (and through it with Australia) and Patagonia. In the fragmented parts of Graham Land archipelago and the outlying South Orkney and South Georgian Islands, we seem to have the bond of connection with the South American main; or, more specifically, a line of curvature of the great Andean chain, which, in its broken parts, can still be traced far beyond its present continental termination. If this concept is a true one, it places before us a parallel to the Andean curvature in the northern part of the South American continent, where the mountain system is deflected off into the broken mass of the Lesser Antilles; to the Aleutian flexure of the Cordilleran system of North America; and to the "Apennine-Atlas" and "Carpathian-Balkan" flexures of the Alpine mountains, the nature of which has been so clearly stated by Suess. In fact, it is hardly possible that any very extensive meridional or latitudinal mountain chain could have been forced up through contractional force without some such deflection being represented in one or more parts of its course; and where these deflections are found they are almost certain to be areas of breakage. The disruption of the Andean system is still (or has until recently been) taking place, as is evidenced in a portion of the Chilian archipelago.
When we look for the evidence of connection such as has been indicated, we find it in the fossils of Cape Seymour, already referred to, in a part of the living fauna of the continents of the southern hemisphere, and in the vestiges of a past life which these continents reveal. Thus, among land animals whose history favors this view, are the South American ostriches, whose close and only immediate allies are the ostriches and other large ratite birds of the African and Australian regions. The union of these birds in the southern continents, whatever may have been the exact place of their origination, gives evidence of migration, and this migration in the case of non flying birds could only have taken place along united land areas. It may be assumed, and has been assumed, that these large and seemingly closely related birds may have been independently developed in the regions which they now occupy; but in view of certain facts presented by other groups of animals, especially the quadrupeds, it is probably safer to assume that their origin was a common one, and that they were distributed over a land surface which had union with South America, Africa, and Australia, or New Zealand. Such a uniting land mass was not improbably ancient Antarctica.
Much more positive evidence bearing upon the question under consideration is afforded by the composition of the South American Tertiary mammalian fauna. Up to and inclusive of the Miocene period this fauna was a distinctively South American one, or at least not North American. The hoofed animals representing various orders (Toxodontia, etc.) were of a type at that time almost unknown in the north; there were no true artiodactyls or perissodactyls, and no insectivores, cheiropters, creodonts, or carnivores. The essentials of the fauna were recruited from the orders of edentates, marsupials, and rodents; the first named have specifically South American types, as the sloths and armadillos, while the last named, also specifically South American, lacked the more common northern forms, such as the hares, rats, squirrels, marmots, and beavers. The middle Tertiary fauna thus gives evidence of the isolation of this portion of the continental tract, and leans to the probability that the fauna then and there existing was derived from a southern source. It is only in the later or Pliocene period that we have an introduction of the northern types of quadrupeds, such as the true carnivores, deer, horses, mastodons, etc. Whether or not the region where this southern South American fauna originated was the region of South America itself, or another land mass that was at one time united with it, is not made clear by the evidence of the faunal elements that have been noted; but significant in this connection, as has been pointed out by Prof. W. B. Scott, is the relationship of the Miocene marsupials, which incorporate within themselves a number of distinctively Australian types, a condition that has been properly emphasized by the investigator last mentioned as giving strong evidence of a former union between the two main southern continents.
In explanation of the anomaly, assuming a connection between South America and Australia, that none of the South American mammalian types reached Australia, Prof. Scott suggests an early connection between the latter region and Antarctica (and through it with South America, thus permitting of a broad dispersion of the marsupialian types), and then a final severance before the appearance of the placental form of quadrupeds. Whether this explanation will ultimately prove to be the correct one or not, the geological construction of the region makes it very probable that a disruption of the Austral tract took place before that of the South American.
Prof. Scott, in a review of the relationship of the southern land masses (Science, February 28, 1896), thus states his position: "In conclusion, it may be observed that the facts of paleontology may best be explained on the assumption that the antarctic land mass has at one time or another been connected with Africa, Australia, and South America, which formerly radiated from the south pole as North America and Eurasia now do from the north pole. While this seems a highly probable assumption, much remains to be done before the history of the southern continents is as well known as that of the northern ones, and in particular many questions must remain open until the Tertiary mammals of Africa and Australia shall have been recovered."
This evidence from paleontology may thus be taken to strongly supplement that which comes from the side of pure geology—evidence indicating a much further extension southward of the South American continent, and of a former union between it and a past and still partially existing Antarctica.