Tertiary Mammals and Continental Drift.
A Rejoinder to George G. Simpson.

by Alexander Du Toit (1944)

Editor Charles H. Smith's Note: Original pagination indicated within double brackets. Notes are numbered sequentially and grouped at the end, with the page(s) they originally appeared at the bottom of given within double brackets. Reprinted by permission of the American Journal of Science. Citation: American Journal of Science 242 (1944): 145-163.

[[p. 145]] ABSTRACT. Doctor Simpson's criticisms are opportune, pertinent and helpful, although partial, through being confined to the rôle of the mammals, while ignoring the overshadowing geologic and tectonic evidence--Tertiary and older. The Cretaceo-Tertiary Land-bridges of du Toit are the logical outcome of the Earth's structural pattern whether Drift be admitted or not, and must therefore be viewed as essentially orogenic in character and generally linear, narrow, impersistent and imperfect, particularly during their later history. Simpson has merely shown that such inter-continental linkages could not have persisted as long, nor have permitted terrestrial migration as freely, as believed, but he has by no means disproved the probability of Continental Drift, which has been based on a wealth of other evidence, much of it long antedating the evolution of the Mammals.


    Critics of Continental Drift are apt to forget that this conception--like the accepted science of Geology--is still being shaped and must experience revisions and modifications during its evolution, more especially in regard to inter-continental linkages during the Tertiary. They are equally liable to ignore the fact that current ideas on such linkages are often unhelpingly vague and too frequently lack a strictly geologic basis.

    When hypotheses of Drift are being criticised, it is essential that the particular scheme of former continental spacing and movement be specified, since there is as yet no true consensus of opinion on the subject and several different systems have thus far been advanced. The citations made from Wegener are usually from the English translation of the third German edition of 1922, regardless of the fact that several revisions have appeared thereafter, the latest being the French version, dated 1937, of the fifth and last German edition. (9) In this latter the chapter on Paleontologic and Biologic Arguments embodies much new matter as well as extensive quotations from recent literature, all of which the reader is earnestly invited to study. For that reason this particular aspect will only be dealt with briefly.

     [[p. 146]] Realising the value to the problem of specialised knowledge, the writer issued in 1937 an invitation to authorities to coöperate, a challenge which has been accepted by Dr. G. G. Simpson, speaking on behalf of Tertiary Vertebrate Paleontology. In a trenchant article in The American Journal of Science entitled "Mammals and the Nature of Continents," (4) he has set out to prove that the evidence of mammalian distribution, while limited in scope and time, is adverse to hypotheses of transoceanic or drifting masses, but favorable to current views of stable continents. In contravention it may be questioned whether the mammalia have not created almost more biogeographic difficulties than they have solved. Even Matthew (3; p. 299) is forced to admit a discordance between the vertebrate and invertebrate evidence. Regrettably they have been robbed of much of their value through their late appearance on the scene, well after continental rupturing is deduced to have occurred, yet while such supposed movements were in full swing.

    Since Simpson refers at length to the writer's particular views, but has at times failed to present them correctly, or has omitted their other vital aspects, it becomes necessary to go into greater detail than would otherwise have been necessary, so as to clear up important differences of statement or interpretation.

    It is true that paleontologists have thus far not shown any leanings towards the new ideas and seemingly remain "orthodox" in outlook, but such is understandable for, as the writer remarked in 1927 (6; p. 118), "geological evidence almost entirely must decide the probability of this hypothesis for those arguments based upon zoö-distribution are incompetent to do so." After studying Simpson's paper and consulting those authorities cited therein that are available in South Africa, the author sees no reason for modifying that apparently sweeping pronouncement. For argument, the finding of identical mammalian species in, say, Madagascar and Rhodesia would merely indicate a former land-connection, but in the absence of other information would not prove whether Madagascar had at the time lain 10 or 1000 miles distant from Africa or had been attached thereto via Natal, Mozambique, East Africa or Somaliland. Those common fossils would only say "yes" and "when," but not "where," or not with the precision desired. Furthermore, as contrasted with other methods, paleontologic comparison suffers from [[p. 147]] certain disabilities, as set forth below. Within those limitations excellent use can be made of this valuable arm, and in the case of the mammalia, from the Eocene onwards. Perhaps the above may help to explain the remarkable paradox that, while paleontologists as a body are antagonistic towards Drift, the protagonists can claim that their views are actually supported by paleontologic data. For that matter Simpson has done signal service by pointing out serious errors not only in initial identifications or comparisons, but in the perpetuation of those mistakes through lack of critical study of the original sources, in which the writer must incidentally admit to having shared.

    Study of the many, varied and conflicting opinions expressed by more than one generation of biogeographers reveals most strikingly the serious difficulties, surprises and dilemmas encountered by them, which collectively leave a strong impression that some elusive, though vital, principle must have been missing from their particular ideas of past continental linking. It is more than significant that, of those welcoming the hypothesis of Drift, so goodly a number have been biologists. One of the exceptions is Matthew (3; pp. 202-9), who practically eliminates land-bridges and is seemingly content to explain faunal relations on the Noah's ark principle, though even so he runs into difficulties. The notion of random, and sometimes two-way, "rafting" across the wide oceans, which he favors, evinces, however, a weakening of the scientific outlook, if not a confession of doubt from the viewpoint of organic evolution.


    a) Paleontologic Record. The "imperfection" of the record is universally deplored. Everywhere, though more particularly in the southern lands, paleontologic knowledge has lagged, and must continue to lag, far behind geologic. Not only are the terrestrial less known than the marine Tertiaries, but they have been, as Matthew points out (3; p. 178), eroded from over wide areas, and the gaps in our knowledge concerning their life must be considerable, whereas the corresponding marine faunas are relatively well known. Conclusions from the meager finds recorded must largely be tentative. Furthermore, unlike present-day life, the scanty vertebrate fossils have in the main been collected from few and widely separated deposits in which the recorded proportions of the determinable genera and species could be very different from their actual frequency in [[p. 148]] nature. Such assumes importance when faunal resemblances are tabulated on a percentage basis, as done by Simpson (4; p. 20). The scantiness of such material cannot be overlooked nor the likelihood of unexpected discoveries that would appreciably modify current ideas. Negative evidence must be quite unreliable therefore.

    b) Instability of Determinations. Although writing without "special competence" in this field, the author has been struck by the fragmentary nature of so much of the material labeled generically or specifically, the periodic changes in nomenclature brought about by subsequent revision, the departures in opinion regarding affinities expressed by men of eminence, etc., so well exemplified by Simpson's account of the history of Neotragoceros (4; pp. 24-5). The observer is led to speculate inhowfar the genera and species listed to-day, and of course their affinities, would remain appreciably unaffected by future work and better material. All this is apart from that fundamental question, which cannot be discussed, "What actually constitutes a Species?"

    c) Wrongful ascription. To the examples quoted by Simpson can be added the incorrect identification of forms because of their association with admittedly related fossils. One example of outstanding importance will be cited, namely the recording by Amalitzky in 1899 of typical members of the Glossopteris Flora together with "Karroo Reptiles" in the Permian of Russia, an association that has never been questioned, at least in print. While the Karroo affinities of the Russian vertebrates are beyond doubt, the writer's examination in 1937 of several of the types identified as Gangamopteris and Glossopteris by Amalitzky, or of duplicates so labeled by him, showed that those particular specimens, while outwardly very similar, yet failed to agree in their venation with those type genera of the Glossopteris Flora. That Amalitzky should have been so misled is under the circumstances quite excusable, but this instance gives rise to speculation whether other analogous cases of paleontologic inexactitude might not exist unsuspected.

    d) Specific Determination. In the early days of research specific determinations in far-off lands were, and understandably, made with the well known forms of Europe. With progress and the realisation of faunal provinces, the institution of local species became inevitable, a policy which may perhaps have been overdone. A disturbing psychologic factor, which [[p. 149]] is quite distinct from the "personal equation" must here be pointed out: namely the natural tendency to compare new finds with material described from corresponding formations nearest to hand. It is probably difficult to keep this "distance factor" under subjection, since any tendency to make specific comparison with rather similar forms from half way round the globe could so readily be regarded as incautious, with the contingency that the closer relationships of far-separated faunas might for long escape notice. In not a few cases again the more distant faunas, particularly if the relevant literature be in some foreign language, are liable to be overlooked, with the same ultimate result.

    With the continents viewed as no longer fixed, this aspect becomes one of actual concern, and indeed, if the feasibility of drift be conceded, reconsideration of the long-distance correlation of faunas generally would appear to be inevitable. Until a considerable amount has been done in that direction the existing paleontologic evidence can be accepted only with reserve by those favoring drift.


    The degree of similarity between any two faunas unquestionably presents great difficulties. Simpson has exposed several weaknesses in the writer's statements and arguments on this subject, some of which are frankly admitted, due primarily to the inherent tendency to stress the points of similarity for the very reason that the areas concerned happen to be so far apart to-day. Such relative accentuation finds a parallel when a paleontologist speaks of "highly fossiliferous Cambrian" when he means "highly fossiliferous for the Cambrian." Simpson criticises the writer's statement "2) faunal differences at the ends of a link, as von Ubisch (quoted by Wegener) has pointed out, ought to be appreciable owing to the differentiation that must have occurred along the bridge itself, whereas a close specific relationship or even identity of forms may indeed characterise the life of the now-opposed lands," in which certain, and not all, forms were so obviously indicated. He then says "If, as the logical alternative, he means to say that identical faunas have been found on separate continents, then he is even farther from the truth," thereby enlarging the idea of certain closely specific or identical forms (not faunas) into one of identical faunal assemblages.

     [[p. 150]] With no desire to detract from the true value of paleontologic comparisons, but the weaknesses exposed in the preceding section in mind, a protagonist of drift may be pardoned for feeling some reasonable doubt whether the mathematical comparisons of similar or related species of distantly-spaced faunas, as compiled from current literature by Simpson (4; p. 20), Arldt or others, are as real as he is asked to believe. The instance cited of the apparent lack of identical species among the Triassic reptiles of Brazil and South Africa should not strictly be pressed in view of the limited amount of collecting, the high ratio of new genera and species so far found, and conceivably the "distance factor." Even under the closest fitting of the continents permissible, the distance between the two areas could not have been less than 1300 miles.

    Accepting, however, as a basis of discussion Simpson's percentages (4; p. 20) as applied to faunal relationships, they but serve to emphasise the rapid falling off of specific similarities with distance, even within the confines of a single landmass, a peculiarity which is quite well marked among the living faunas. With a long and narrow linkage the "con-specific ratio" falls to a surprisingly low figure to become negligible or zero in extreme cases, even when a past union of the particular lands happens to be indicated by other credible evidence. Striking is the lack of common species from among the closely allied Permian reptiles of Russia and South Africa, areas 6500 miles apart.

    Now the Cretaceo-Tertiary land-bridges envisaged under Drift (see next section) are essentially linear or arcuate, narrow and unstable, ranging in character, it is thought, from continuous corridors to isthmian links in the sense of Bailey Willis. In not a few instances they would have paralleled Simpson's Case D with a community of families of 67 per cent, of genera 24 per cent and of species 5 per cent, while the respective percentages might well have been lower. In certain instances such bridges would have run with the Earth's climatic zones (Behring Strait; Central Atlantic); in others across them (Malaya-Australia-New Zealand-Antarctica). For post-Eocene times they could in general have corresponded closely with the "sweepstake routes" so aptly defined by Simpson. Even in the case of the Afro-Eurasiatic connections (Spain-Morocco; Egypt-Syria) the limited stratigraphic data are not necessarily in favor of wide bridges across the Mediterranean "Tethys."

     [[p. 151]] Under those circumstances his contention, that a far higher proportion of common species should be detectable under Drift than is brought out by the recorded fossil mammalia, is greatly weakened. It may be questioned too whether the mammals, by their very mobility and peculiar organisation, could be expected to have migrated with the same presumed uniformity as the ancestors of the living lower orders--snakes, scorpions, earth-worms, land and fresh-water mollusks, plants, etc.--which, because of their number and variety are surely of not less importance in this problem, especially as their evidence proves to be not always in accord with that of the vertebrates (3; p. 299). In his latest edition (9; Chap. VI) Wegener effectively quotes authorities such as Michaelsen, von Ubisch, Okland, Huus, Jaschnov, Meyrick, Irmscher, etc., in those other fields.

    Despite criticism, the statement "Migration along a link need not be equally effective in both directions" (7; p. 294) is reaffirmed, a cogent illustration being that of Russia and Southern Africa with similar reptiles, amphibians and fresh water mollusks, but different plants during the Permian. Simpson admits (4; p. 17) possible selection during intermigration between continents now separated, though only along limited corridors. Conceivably, with only a very brief spell of linking, something rather like "one-way traffic" could have operated; of such Madagascar would have formed an example. The factors controlling migration must be extremely complex, and, in the case of crustal wrinkles thrust up from the ocean-bed or lavas and scoriae heaped up to above sea-level the laws of biologic spreading within continental areas might not hold good. With fixed continents round-about routes have often had to be postulated by biogeographers--to well within the polar circles sometimes; these, if biologically possible, must have tended towards the elimination of the less adaptable forms, and thus towards selection.

    Simpson has had to pass over the abundant non-mammalian life and thus ignore so much of the very evidence that has compelled biogeographers to formulate their individual schemes of past land-connections. Of such living forms admittedly many, probably most of them, have as yet no well known fossil representatives, although of high significance none the less. To argue that such southern disjunctive distribution is due to colonisation from the north through forms not yet discovered in the Holarctic region, is neither scientific nor fair. Attention [[p. 152]] can merely be drawn to the remarkable distribution of certain organisms, as done by Wegener and others, not forgetting their parasites (7; p. 294). The plants too have received insufficient attention, seeing that their thermal sensitivity makes them useful through indicating, for example, the difficulties inherent in the assumption that the cold temperate floras of the southern continents, which show unexpected resemblances, must necessarily have been derived from the Holarctic vegetation that flourished north of the climatic barrier set by the wide Tropical Zone. The same argument applies to the fresh-water decapods.


    The way of the bridge-builder is strewn with pitfalls. Simpson deplores that, opponent as the writer is of the numerous land-connections of current conception, he (du Toit) has nevertheless accepted several bridges differing from the latter, so it is stated, in no respects, and faulty too in that they could only be broken once, whereas paleontologic evidence would suggest a repeated "make and break."

    That is very far from being the case, since under the hypothesis proposed by the writer such connections--save possibly in the extreme northern Atlantic--are viewed as having been essentially of an orogenic kind, developed within restricted zones of lateral compression, produced by one crustal block impinging on a second, or else thrown up from an epicontinental sea or even from an ocean by marginal or advance folding in front of a drifting block or blocks. Longitudinal tension and/or erosion, or sinking in the post-orogenic period, would have served to reduce and even to sever such connections. The geologic record is punctuated by such far-extended squeezings (7; p. 48), the Tertiary being in fact marked out by at least three world-wide compressive phases and interludes.

    This "segmental oscillation," which Simpson cannot accept, is written large, though in other characters, in the Tertiary fold-girdles, which should dispose of his complaint that under Drift the "make and break," so rightly stressed by him, could not have been cyclic. Is such not perhaps Joleaud's idea of the "accordion movement" under a different guise?

    More important still is the fact that such "fold-" or "drift-linkages" have not been flung out at haphazard, but at spots and times closely fixed by the visible and dated crumplings of the crust, and have not been conjured up, as Simpson would [[p. 153]] hint, to explain actual or supposed resemblances of the terrestrial life of the continents in question. Indeed the situations of such "bridge-heads" are regarded as pre-determined, their prolongations across the lands being in most cases actual major fold-ranges. Even under current views such fold-zones are regarded by a large body of geologists as persisting beneath the present oceans and as having at times in the past projected as island-chains along certain stretches thereof. Accepting Drift, they are merely shorter to start with. It can be asked, and rightly, why, if such be the case, there should be any recourse to Drift. The bitter fact is that current Geology, sublimely unconscious of its impotence, is wholly unable with the continents remaining fixed to account for the tectonics of those fold-zones in a physical, logical and convincing fashion! For an adequate explanation some drift would have to be invoked.

    Of no less importance through ruling out any bridge-head is the presence along a coast-line of fringes of purely marine strata--or even their probable former presence, if adequately supported by other lines of evidence. As an example, the little-disturbed Lower Cretaceous to late Tertiary marines bordering the Southern Atlantic exclude from consideration lengthy stretches of African and American shores. On the contrary widespread unconformities or thick non-marine intercalations point to uplift and to possible temporary connections. Paleogeographers have all too frequently overlooked such basic restrictions when locating their bridges.

    By the writer the Hypothesis of Drift is regarded as essentially established by the Paleozoic and early Mesozoic evidence; the rest is largely a logical corollary, which in the main is consistent in outline, but has not yet been worked out in detail, for that would be a mighty task. The continents are pictured as having proceeded to "part" during the later Mesozoic as shallow gulfs were extended into and between them under the initial crustal stretching, but actual rupture of the sial shell did not generally take place until the close of the Cretaceous, to pass through a series of climaxes during the Tertiary. Such should dispose of the erroneous idea that at the beginning of the Tertiary the continents were still usually in contact along their edges--making allowance for the widths of the continental shelves--with opportunity for free biological migration.

    The true rôle of the mammalia will be the fixing of land- [[p. 154]] bridges not in the horizontal, but in the vertical plane, through revealing the uprisings and downwarpings of the crests of such linkages, themselves positioned by other, and mainly older, criteria.


    That those lands, including Africa and India, had some inter-migration during the Tertiary is proved by the mammalian evidence marshalled by Simpson and needs no stressing. His arguments (4; pp. 8-10) in favor of a connection across Behring Strait are indisputable, while his recognition of its apparently cyclic nature is in agreement with the known triple folding that crosses from Siberia to Alaska, and is not in conflict therefore with the scheme of linking advocated by the writer (7; pp. 186, 298).

    On the most vital part of the problem, namely the connection between North America and Europe, Simpson firmly opposes all bridges across the Central Atlantic but concedes some linking farther north, possibly of late date (4; p. 6). Now the bridging of this ocean, with a span of between 900 and 1500 miles across its narrower part, much of it over 1000 fathoms deep, while perhaps a simple matter for the biologist, is a far tougher proposition for the geophysicist, who cannot so readily admit the sinking of non-tectonic crustal blocks of that width in defiance of isostatic principles. Only a narrow linkage of either tectonic or volcanic origin can accordingly be visualised. Outside the presumably local wrinklings of southern England and western Spitzbergen no relics of Tertiary fold-ranges are represented along the North Atlantic, while the Tertiary faulting, e.g. Greenland and the British Isles, although common, is nowhere to our knowledge of truly large vertical displacement. Under current views the only assumption remaining would seem to be a volcanic chain extending say from the British Isles via the Faroes and Iceland to Greenland and Canada.

    Under Drift the North Atlantic is regarded as having been widened by intermittent tension, a movement least and slowest in the north, with magma erupting from fissures produced in the stretching ocean floor, and building up piles of volcanic material that were in turn ruptured towards the close of drifting, a process that may still be in progress. The problem is admittedly more or less equally explicable under theories of stable or moving continents, but the latter has the merit of forming the logical consequence of a particular scheme of [[p. 155]] Earth structure which seems to hold good for the pre-Tertiary eras. The practical outcome, however, is that, when current views are applied to the Northern Continents, the assumptions required are not any simpler, or more probable from the geophysical standpoint.


    Simpson's exposure of the flimsy nature of Joleaud's paleontologic evidence for Tertiary Central Atlantic connections is welcome, and the writer candidly admits his lapse through citing Gregory in this matter. All the same, the feasibility of double linkages, of an orogenic type, between Central America, Cuba and Spain and Venezuela and Morocco respectively during the Cretaceo-Tertiary, happens to be a consequence of the scheme of drift proposed and is furthermore independent of any paleontologic similarities between those four opposed continents.

    Those bridges are regarded as having been pressed up under the rhythmic approach of the North American-Eurasiatic and the South American-African masses, but lowered by lengthwise tension under the westerly drift of the Americas as the meridional "rift" of the Atlantic widened. Their persistence above sea-level must have depended essentially on the relative intensities both in space and time of the periodic but contrasted north-south compressions and east-west tensions. The shallow-water marine faunas of those latitudes show that for the earlier Tertiary the state amounting to a line of submarine ridges was undoubtedly attained, while certain resemblances in the terrestrial life, referred to below, suggest that the state of island-chains, or even of isthmuses, was reached. The feasibility of effective transoceanic connections after the Miocene was, however, doubted by the author (7; p. 204) and he should therefore have ruled out as a possible migrant so much younger a genus as Hipparion. South Africa has nevertheless yielded Hipparion-like teeth, though their specific characters would appear not incompatible with a northern derivation. While the affinities of the mid-Tertiary fauna of South-west Africa are according to Stromer with North and Central Africa, the specialized hyracoid Protypotheroides is allied to the Miocene Protypotherium of Patagonia.

    Conceding these two structural "fronts"--Central America-Spain and Venezuela-Morocco--approaching each other while being stretched unequally at their western ends, accompanied [[p. 156]] by their attendant advance and diagonal folds, the migrational paradoxes of the Central Atlantic that have been forced upon biologists, and that Simpson has found (4; p. 25) so hard to credit, become easier to understand. Indeed, with such long, sinuous, warping and disrupting chains quite unexpected interchanges of terrestrial life could hardly have failed to take place, as remarked by H. B. Baker. As illustration, ignoring Drift, taking a chart of the Caribbean region and delineating simple rises or falls of sea-level, peculiar sunderings or linkings of the various islands would follow, to become more surprising still were even a little crustal warping to be introduced into the process. One should not overlook too the tremendous part played by volcanicity in the building up of the Tertiary ranges, so finely displayed for example in the island-chains of the West and East Indies. Cross-connections or short-circuits might thus be established, as in the case of the Lesser Antilles.

    Even with the continents fixed geologic evidence is against any direct connection between North and South America from the Mesozoic down to the late Miocene. How then were the marsupials of South America derived from a pre-Tertiary stock in Europe? The South American monkeys too in their brain development are surprisingly similar to those of the Old World. Each truly demands an early Tertiary trans-Atlantic bridge! While Simpson has been so far successful in disproving any free intermigration of the mammalia across the mid-Atlantic during the latter half of the Tertiary, there remains in favor of an earlier bridge or bridges, as well claimed by numerous biogeographers, quite a number of lower animals and plants, the distribution of which forms a stubborn problem under views of fixed continents.


    While Simpson severely questions the mechanics advocated by the writer in regard to these continents, a study of du Toit's restoration (7; Fig. 7) and of an artificial globe will show that the maximum displacement of Australia would not have been much more than 2000 mules. Even assuming his extreme proposition, that all such movement had been concentrated into the post-Oligocene period, Australia in order to reach its present position would only have had to "speed" at under six inches per annum. There is also the alternative that, although the parent mass had been broken into its three parts at an earlier [[p. 157]] date, intermittent contacts could still have been maintained between those portions by the circum-Pacific Tertiary fold-zone running through Patagonia, Grahamland, West Antarctica, New Zealand, New Guinea and the East Indies, which last is a global fact and quite independent of any special theory of Earth structure. The existence of such a lengthy ridge finds support from the great resemblances of the Miocene marine littoral faunas of Patagonia, Australia and New Zealand, as stressed by van Ihering, Ortmann and Hedley.

    The curving of that fold-zone around Australia shows how that block, twisting anti-clockwise, attempted to by-pass Indo-Malaya, thereby permitting the Asiatic wave-front to press out into the Indian Ocean with the development of crustal swirls, while the Australian wave-front escaped into the Pacific. Conditions tending towards the production and destruction of linkages between Asia and Australia were accordingly to hand. Umbgrove (8; p. 34) has recorded the surprising shallowness of the seas over the vast East Indies region throughout the Tertiary--in strong contrast to the Mesozoic--with intense folding during the mid-Miocene, large parts having certainly been land in the Eocene and Pliocene. The intensity and rapidity of the Pleistocene movements, that led to the upheaval of coral reef-terraces for many hundreds of feet, to the tearing out of astonishingly deep oceanic basins, etc., have been fully set forth by the Dutch geologists, although little known generally. Under the universal lowering of sea-level in the Pleistocene New Guinea, Australia and much of the East Indies --"Sundaland"--must for a time have become joined by low-lying ground.

    As against all this evidence paleontologists and biologists alike have tended to stress the biologic isolation of Australia from the early until the very late Tertiary, a relationship which must surely have been less likely with the continents fixed than with Australia far away and only drifting towards Asia. In regard to the marsupials--and in such the writer stands corrected--Simpson denies any special resemblance between those of South America and Australia, though he does not appear to be so happy about the family Dasyuridae or Thylacinidae of the order Polyprotodontia and the one advanced order Diprotodontia1, possessed by both continents. [[p. 158]] Suggestively their respective parasites are identical according to Breslau (9; p. 105). He reluctantly admits that, giving that evidence its maximum weight, any southern connection could only be "something of the order of discontinuous evanescent island chains." One is tempted to speculate whether any part is played by the "distance factor."

    The contrasted, yet homologous relations of South America and North America on the one hand and of Australia and Asia on the other can advantageously be considered together. Marsupials appeared in the Mesozoic in North America and Europe, but neither they nor their Tertiary descendants seem as yet to be known from Asia or Africa despite the fact that all those lands were interconnected during the early Tertiary. Now according to Simpson they could not have reached South America or Australia from the north, since those continents were under his, as well as under current, views then unconnected with the northern lands. Had such been the case those southern continents should logically have received other northern vertebrates. Strikingly, early Tertiary marsupials have not yet been recorded from Asia or Australia though their future discovery there is not denied. For the present, until the paleontologist can explain away such paradoxes, he can scarcely maintain that the balance of existing knowledge negatives any Tertiary connection between South America and Australia--from both of which Southern Africa had already become detached.

    Admittedly in those two continents the resemblance among the marsupials only extends to families and those lands are not therefore con-specific, while the monotremes are absent from South America and the edentates from Australia. Accepting Drift, however, Australia could never have been closer to South America than 3000 miles--as measured across Antarctica--from either side of which central mass those continents ultimately broke away. Any very close faunal resemblances between those far-sundered and isolated lands could well be lacking, an explanation anticipated by Wegener (9; p. 106).

    The significance of other terrestrial life should on the other hand not be overlooked, such as the presence of giant birds, fossil or living in the "Gondwanaland" region--in Patagonia, Seymour Island, Antarctica, South Africa, Madagascar, Australia and New Zealand--while the distribution of the fresh-water fishes is highly suggestive (Eigenmann). Striking are [[p. 159]] the fresh-water crayfishes of South America, Australia, New Zealand and Madagascar--but not found in Africa--separated as they are by the equatorial zone from those of the northern hemisphere (Ortmann). Matthew's explanations are far from satisfying. The scorpion family Bothriuridae is restricted to South America and Australia (Hewitt). Those curious worms, the Phreodrilidae, are confined to the colder parts of the South Temperate Zone, including Kerguelen and the Falkland Islands. The botanical affinities are much more marked between South America and Australia than between either of them and Africa (Hooker, Hemsley, Bentham and Andrews).

     In the face of the above faunal relationships, which could be greatly extended, there would seem to be a prima facie case for a former linking or linkings between South America and Australia via Antarctica during the Cretaceo-Tertiary, an opinion stoutly maintained by not a few eminent biologists and geologists.


    In regard to the separation of this island from Africa the mechanics are not so puzzling as Simpson would suggest (4; pp. 27-9). Whether Drift be accepted or not, the controlling factor was clearly the lengthy geosyncline, developed in the Paleozoic, that ran down the eastern margin of Africa and covered the western half of Madagascar. Various geologists have contributed towards the elucidation of its history. The formations along both sides of the Mozambique Channel dip towards the latter and indicate through their dominantly marine and occasionally terrestrial phases the deepenings and shallowings of that trough from the Lower Permian onwards, with intermittent but progressive lengthening, until in the late Cretaceous it penetrated to the South Atlantic (see 7; Figs. 11 & 12).

    The affinities of the later Cretaceous marine faunas in this region show that, while this trough was in existence, peninsular India must have been prolonged southwards to Madagascar to form a joint barrier such that mollusks from the Mediterranean and North-west India were able to enter western Madagascar only in limited numbers but failed to reach round to the Madras area and beyond, where a different fauna held sway. Marine older Tertiaries fringe western Madagascar dipping towards the Channel, and the island must then have [[p. 160]] been cut off from Africa, but stratigraphic gaps at the Oligocene and about the Miocene show that, through slight shallowing of the geosyncline at those times, Madagascar could again have become linked to the mainland. Its geologic history during the Pliocene is unknown, though during the Pleistocene the sea-level fluctuated more than once. Simpson's statement (4; p. 27) that "it contains early Tertiary groups without their later Tertiary relatives abundant on adjacent continents, mid-Tertiary groups with neither their early nor their late Tertiary allies, and late Tertiary to Recent groups such as are always, elsewhere, accompanied by members of other groups [are] here quite lacking," is quite consistent with the stratigraphic evidence.

    From the data available the writer has deduced the intermittent stretching of the crust during the Jurassic and Cretaceous with slow widening of the geosyncline (7; Fig. 12) and the outpouring of lavas, its actual rupture in the Eocene, the gradual drifting of the Madagascar block to the south--in the wake of Antarctica--with tilting to the west-north-west along with vast extrusions of lavas in East and Northeast Africa and the Deccan as Australia and India drew away from Africa. In the case of Pemba Island its separation from the mainland is set by Stockley (5; p. 239) as late Miocene or early Pliocene. Any closer dating of those separations would have to be furnished by the terrestrial fossils, and among them the mammalia, though the evidence provided is not only limited but confusing. From its home opposite Kenya and Tanganyika--duly fixed by abundant and consistent data--Madagascar would have had to travel about 600 miles, though in a direction largely parallel to the mainland, a point of high importance when possible re-connections with the latter are in question.

    An Upper Cretaceous connection between South America, Madagascar and India is suggested by the dinosauria (von Huene; Matley), though Simpson does not explain how that evidence can have been misinterpreted. Even allowing for isolation and lack of knowledge concerning its fossil representatives, the living Malagasy fauna remains as much a puzzle as ever. Matthew (3; p. 203) views the insectivore Centetidae as the earliest colonisation, perhaps pre-Tertiary, and the lemurs, rodents and viverrines as about mid-Tertiary. On the basis of the more primitive types of lemurs Gregory (1; p. 372) [[p. 161]] regarded Madagascar as isolated from India by the Oligocene, before India had been occupied by the more advanced lemurs. After the beginning of the Tertiary the contacts with Africa were manifestly slight and any linking must have been, as suggested earlier, brief.2

    H. F. Standing has shown that among the lemurs of Madagascar there are closer analogies with the monkeys of South America than with those of the Old World, which brings out the fact that there are positive resemblances with South America, as long pointed out by Wallace, Lydekker, Gregory and many others, which are explicable under the particular scheme of linking discussed earlier. To mention only a few cases there are the boas, iguanas and certain of the earth-worms (Acanthodrilidae) which suggestively are unknown north of the equator.

    Matthew (3; pp. 203, 206-7), whom Simpson follows with approval (4; pp. 28-9), explains such peculiar relationships as due to occasional introductions from Africa by "rafting," an hypothesis which in view of the limited distances involved--400 miles across the narrowest part of the present channel--is quite conceivable. If true, there was apparently no return rafting, while the African monkeys obviously failed to make the journey. The fact, however, that analysis of the Malagasy fauna shows it to embrace diverse elements, which, as Simpson himself points out, would seem to mark out distinct phases of the Tertiary, while those in turn can be correlated with deduced crustal movements in that region, all forces one to associate such elements with terrestrial linking rather than chance rafting. One can even doubt the alleged natatory abilities of the Malagasy hippopotami (Lemoine) or the dinosaurs (Matthew). The flora of Madagascar does not provide much help, being strictly endemic and much diversified, but demands a closer study from botanists.

    To sum up, the faunal relationships between Madagascar and Africa prove not less puzzling with fixed than with moving [[p. 162]] continents, though explicable by the latter hypothesis under the reasonable assumption of very brief post-Cretaceous connections. Those with South America and Australia are difficult to interpret with stable continents since in several critical cases either the ocean or the equatorial zone would have prevented migration from the Holarctic region. On the contrary the relationships would follow logically from Drift.


    Doctor Simpson's positive assertion, that "the known past and present distribution of land mammals cannot be explained by the hypothesis of drifting continents," can be met with a not less emphatic contradiction.

    Anyone who sponsors a new hypothesis must unhesitatingly investigate every conflict with facts and modify his views accordingly, and Simpson's contribution--long overdue--is therefore valuable through clearing away sundry errors and misapprehensions, reviewing faunal relations from a new angle and throwing fresh light on a complex and fascinating subject. At the same time he does not seem to have appreciated the multitudinous and weighty geologic evidence upon which had been based the writer's synthesis and in which several of the inherent difficulties had been anticipated. Throughout his paper too he has entirely ignored the vital stratigraphic and tectonic aspects.

    What he has really succeeded in proving is that du Toit's Tertiary Orogenic Linkages could not have persisted above the level of the widening oceans down to so late a period as assumed, or that they could not have been continuous enough to permit migration with the freedom hitherto supposed. This consequence, which is fully accepted, clarifies the position and marks a real advance in Tertiary paleogeography.

    On the contrary he has not thereby disproved the Hypothesis of Drift, which is regarded as resting upon a far wider basis--one rooted in Paleozoic times at least--while, paradoxically, most of his statements of facts prove, when analysed, far from incompatible with the idea of moving continents, and in particular with the writer's scheme, modified slightly to meet such objections. It is re-iterated that, only under Drift can unstable, though ordered, land-bridges be made available at about the particular places and times required for the inter-migration of terrestrial life--linkages which in addition give [[p. 163]] the minimum number of possible routes and which automatically exclude from consideration a host of other connections that have been postulated, largely on paleontologic or biologic grounds, though without full recognition of possible conflicting geologic or other evidence.

    The acid test of that Hypothesis will, it is felt, depend among other things on its ability simply and logically to account for the harvest of fossil forms yet to be unearthed. In the meantime fresh evidence of every kind and new ideas will be as welcome as ever in our attempt to establish on a firmer basis this great revolutionary and far-reaching theory of Earth Evolution, which we owe to the genius of Taylor and Wegener.


1. Gregory, J. W.: 1921, The Rift Valleys and Geology of East Africa, London.
2. Lydekker, R.: 1911, Zoölogical Distribution. Encyclopedia Brittanica, 11th ed., Vol. 28, pp. 1002-1018.
3. Matthew, W. D.: 1915, Climate and Evolution. Ann. New York Acad. Sci., Vol. 24, pp. 171-318.
4. Simpson, G. G.: 1943, Mammals and the Nature of Continents. Amer. Jour. Sci., Vol. 241, pp. 1-31.
5. Stockley, G. M.: 1942, The Geology of the Zanzibar Protectorate and its relation to the East African Mainland. Geol. Mag., Vol. 79, pp. 233-240.
6. Toit, A. L. du: 1927, A Geological Comparison of South America with South Africa. Carnegie Inst. Wash., Publ. 381.
7. _____: 1937, Our Wandering Continents. Oliver and Boyd, Edinburg and London.
8. Umbgrove, J. H. F.: 1938, Geological History of the East Indies. Bull. Amer. Assn. Petr. Geol., Vol. 22, pp. 1-70.
9. Wegener, A.: 1937, La Genèse des Continents et des Océans. Libraire Nizet et Bastard., Paris.

Pinelands, Cape Town, Union of South Africa.


Notes Appearing in the Original Work

1. The sole South American genus Caenolestes could admittedly be an aberrant polyprotodont, as believed by Broom and Diderer. [[on p. 157]]
2. The writer's reference to Gregory queried by Simpson (4; p. 28 footnote) comes from (1; pp. 312 and 372) and applies to the Afro-Malagasy mass; that ascribed to Lydekker (2; p. 1009) arose from his cryptic remark that the Malagasy fauna "is much more distinct from the Ethiopian fauna than is the latter from the fauna of either the Oriental or the Holarctic region." [[on p. 161]]

*                 *                 *                 *                 *

Return to Home