Evidence" by Wegener and His Followers

DRIFTING CONTINENTS AND LATE PALEOZOIC GEOGRAPHY* BY P. J. DARLINGTON, JR.

MUSEUM OF COMPARATIVE ZOOLOGY, HARVARD UNIVERSITY Communicated August 20, 1964 As a result of 19 months spent in Tasmania and eastern Australia and 7 weeks at the southern tip of South America, followed by a year of reading and writing on the biogeography and history of the southern end of the world, I have gradually become convinced that continental drift is not just an intellectual exercise but a probable reality. I was not easily convinced of this. I am a conservative, sixty-year-old biologist and also a professional biogeographer (if there is such a thing), and I have been and still am repelled by the exaggerations and unintentional misrepre- sentations of biogeographic "evidence" by Wegener and his followers. Neverthe- less I am now a Wegenerian, of sorts. I think that some continents probably have drifted, although not in quite the usual Wegenerian pattern. I give this fragment of my own history to justify reviewing a subject which, if there is any reality in it, is of great interest and importance to many persons. Convincing evidence-evidence that has convinced me-of movement of continents comes from the matching shapes of Africa and South America and their relation to the Mid-Atlantic Ridge, from new evidence of convection currents in the earth's mantle that might move continents, from the distribution of glaciation and of floras in the late Paleozoic, and from the new (and still very incomplete) record of paleomagnetism. Two questions in the general hypothesis of continental movement are critical and can perhaps be answered. First, when did Africa and South America separate, if they were united? And second, what were the positions of the southern continents when they were glaciated late in the Paleozoic, 200 million or more years ago? Time of Separation of Africa and South America.-The fitting of edges of Africa and South America' is virtual proof that these continents were united, if the hypothesis of continental displacement is accepted at all. Wilson2 suggests as a "highly speculative" hypothesis that convection in the earth's mantle, rising along the line of the Mid-Atlantic Ridge, has separated the continents and is pushing them apart, and that the rate of movement calculated from the ages of volcanic islands, which increase in age with distance from the Ridge, indicates that separation occurred about 120 million years ago, about the end of the Jurassic. How- ever, a suboceanic ridge exists in the Pacific too, apparently similar to the Mid- Atlantic Ridge and similarly bordered by islands of graded ages.2 3 Convection currents may be pushing South America from both sides, and both currents (if they exist) may be diving downward under South America rather than moving the continent. Therefore, rate of movement of South America and time of separation from Africa cannot safely be calculated from Wilson's data, although the data may indicate the latest time that separation can have occurred. Paleomagnetism does not show longitudes (see page 1089) and therefore does not show how far apart Africa and South America were in the past. However, if basic assumptions are correct, paleomagnetism does show rotations of continents. If Africa and South America were united, they have not only separated but have also 1084 Downloaded by guest on September 27, 2021 VOL. 52, 1964 GEOLOGY: P. J. DARLINGTON, JR. 1085

rotated so that once-parallel coast lines now diverge southward at an angle of about 450, and paleomagnetic data seem to show that the divergence was already about 220 early in the Jurassic.4 This suggests that the continents separated much earlier than supposed by Wilson. Biogeographic evidence of time of separation of Africa and South America might have been decisive, but is not. For example, angiosperms (the existing dominant land plants) might have shown a uniquely close relationship between the floras of Africa and South America if these continents were joined when the angiosperms dispersed, which was primarily during the Cretaceous. However, angiosperms do not show this. As many plant taxa, most of them angiosperms, are shared by the Indo-Malaysian region and tropical America as by Africa and tropical America.' The angiosperms have an additional significance. 1\Jost families of trees that now dominate tropical rain forests are angiosperms, and if the rain-forest trees dispersed without benefit of contact betweeen Africa and South America, the insects and other animals in the rain forests may have done so too. This is just one example, but an important one, from a very large, largely controversial body of biogeographic evidence. In sum, this evidence seems to me to show no definite time of separation of Africa and South America and no clear indication that the continents were ever united at all. If these continents were united, as I think they probably were, they apparently separated too long ago to leave any recognizable traces of the union in the distributions of existing plants and animals. In summary of time of separation of Africa and South America (if they were united): the latest time allowed by geologic dating (ages of Atlantic islands) is probably late Jurassic; paleomagnetism suggests that the continents had rotated independently before the Jurassic; and other evidence (not all reviewed here) seems consistent with separation in the Triassic, or earlier. Southern Supercontinents?-Wegenerians have been possessed by a fixed idea: that, before drift, the continents must have formed a great supercontinent, an all- embracing Pangaea or at least a southern Gondwanaland. This idea seems to find support in the apparent former union of Africa and South America. How- ever, union of these continents (if they were united) need not require that others were united too. Actual evidence of union should be looked for in each case. Cri- tical evidence comes especially from late Paleozoic time, from the late Carboni- ferous and early Permian periods (the Permo-Carboniferous) when the southern continents were heavily glaciated. The rest of the present paper will be concerned chiefly with the situation in the southern hemisphere during this small segment of geologic time. Supposed evidence of union of southern continents has been derived from fitting of the continents' shapes, from continuity of geosynclines and other geologic structures, from the distribution and directions of movement of Permo-Carboniferous ice, from the distribution of ancient floras, and from paleomagnetism. Much of this evidence does strongly suggest that most continents lay farther south on the world than they do now, but no evidence clearly indicates a general union of southern continents. A hypothetical grouping of the continents without union, which seems to satisfy all requirements now known, is suggested in Figure 1. Evidence of Shapes. The southern continents, excepting Africa and South Amer- ica, do not fit together well. This fact is emphasized by the many different ways Downloaded by guest on September 27, 2021 1086 GEOLOGY: P. J. IARLINGTON, JR. PROC. N. A. S.

FIG. 1.-Suggested arrangement of southern continents inl late Paleozoic time. Only Australia is placed, inl relation to the pole, by paleomagnetic data for the time in question (see text, p. 1089). West Antarctica, which is a relatively recent, geologically active area, is shown as an archipelago, as it may still be, under the ice cap. Wegenerians have tried to make the continents fit, unsuccessfully. The contrast between the precise fit of Africa and South America and the poor fit of other southern continents strongly suggests that the others were not united or that, if they were, they separated long before Africa and South America did. Geologic Evidencee.- Wegenerians have stressed the supposed continuity of a synclinal system, the "Samfrau Geosyncline," connecting southeastern South America, the southern tip of Africa, western Antarctica, and eastern Australia, and supposedly indicating that these pieces of land were once aligned and in contact.6 Two serious criticisms of this argument can be made. First, part of the supposed geosyncline may not exist. Harrington7 shows no geosyncline where the hypothesis requires it in South America. And second, continuity of geosyndlines does not require continuity of land. A geosynclinal, mountain-forming system crosses the ocean gap between South America and western Antarctica, for example. However, this system does suggest a long-standing relationship between the tip of South America and the Antarctic Peninsula which should be a guide to orientation of the continents in the past. Wilson2 suggests that secondary (not mid-oceanic) ridges on the ocean bottom may connect former contact-points of continents, and he offers a mid-Mesozoic "supercontinent" assembled according to these guides. The results disqualify Downloaded by guest on September 27, 2021 VOL. 52, 1964 GEOLOGY: P. J. DARLINGTON, JR. 1087

the method, I think, so far as the southern continents are concerned. The continents are brought together in an unnatural way, with the Antarctic Peninsula rotated out of its probable relation to South America, and with the southern quarter of South America overlapping the Precambrian shield of Antarctica! In the space available I can consider only small samples of the nonglacial geologic evidence (above), and by way of summary can say only that, so far as I can judge, no such evidence indicates an actual union of southern continents, except of Africa and South America. Glacial Evidence.-Late in the Paleozoic, in late Carboniferous or early Permian time, ice sheets were widespread on southern Africa and southern Madagascar, southern South America, and southern Australia including Tasmania. Ice occurred north into what is now the southern edge of the tropics on all these continents. Antarctica was glaciated too, at least in part, and so was India, which now lies mainly in the tropics north of the Equator. But more-northern continental areas were not glaciated at this time. The only satisfactory explanation of this distribution of ice seems to be that all the continents (except Antarctica) lay farther south than they do now, the southern ones in colder and the northern ones in warmer latitudes than now. Other explanations are unconvincing.8 However, the distribution of the ice is not evidence that the southern continents were united. Rather, the formation of continental ice sheets suggests the presence between continents of ocean gaps wide enough to allow massive evaporation and, as a result, heavy pre- cipitation on the land. The arrangement of continents suggested in Figure 1 seems to satisfy the glacial requirements. All the principal glaciated areas are brought near or within about 450 of the South Pole, into latitudes where glaciation is possible, judging by what happened in the northern hemisphere in the Pleistocene. Water gaps between continents allow evaporation to supply water for ice on land. And the eccentric position of Antarctica seems greatly to increase chances of glaciation on surrounding continents. A massive ocean current probably always circles Antarctica from west to east, if the land allows it. With the land arranged as in Figure 1, the current would be cooled by passage through the polar region and then would be deflected away from the pole toward the continents that were in fact glaciated. Ice sheets on land gouge rock surfaces, deposit glacial "drift," and move boulders sometimes hundreds of miles from their places of origin. These evidences of glaciation may show directions of movement as well as extent of ice in the past, although indications of direction are sometimes ambiguous. Ice might conceivably flow from one continent to another, if the continents were united, and might leave unmistakable evidence that it had done so. This is the theoretical basis of a diagram by King9 showing radiation of late Carboniferous ice from Antarctica onto several conjoined southern continents. If this diagram of ice radiation were based on fact, it would virtually prove that the continents were united. However, it is not based on fact. The central part of the diagram shows directions of movement of ice on Antarctica now. Directions of movement of Antarctic ice in the Permo- Carboniferous are unknown, and persons who believe that continents have changed position are not entitled to assume that ice flowed in the same directions in the past as now. Ice did not move across Tasmania into Australia. Tasmania was glaciated from a local center in the northwest, and the ice moved eastward and partly Downloaded by guest on September 27, 2021 1088 GEOLOGY: P. J. DARLINGTON, JR. PROC. N. A. S.

southward.10 This fact, not fully appreciated until recently, is alone almost enough to destroy the validity of King's diagram. Movement of ice from the south northward onto the southern tip of Africa, as indicated on the diagram, would seem to bring the hypothetical southern ice sheet into head-on collision with the main mass of South African ice, which moved southward. I do not think South African geologists have discussed this difficulty. And directions of movement of ice on South America and India are doubtful and disputed.8 This detailed criticism of King's diagram is justified by its reverse significance. The diagram shows the pattern of ice movement that Wegenerians think ought to be found if the southern continents were united. And I agree. Very extensive glaciation did occur, and if the southern continents were united then, a pattern of ice radiation like that diagrammed would be expected and should be unmistakable. But the pattern probably does not exist and certainly is not unmistakable. This seems to be real evidence that the southern continents were not united. Biogeographic Evidence.-Distribution of floras: Two different, major floras existed in different parts of the world during late Carboniferous and Permian time.8 One was a chiefly southern flora which accompanied or followed glaciation wherever glaciation occurred; on South Africa and Madagascar, southern South America, southern Australia and Tasmania, and Antarctica, and also on India. It was a species-poor flora dominated by Glossopteris and allied genera. The other flora (actually several others) occupied parts of what is now the north-temperate zone. This flora was more diverse (consisting of more species of plants) and was apparently adapted to a relatively warm, uniform, humid climate. The geographic separation of these floras has led some botanists to postulate an impassable ocean barrier between the northern and southern halves of the world (with India in the southern half), but this explanation is unlikely and unnecessary. Some other plants and animals, including terrestrial reptiles, crossed the supposed barrier while the floras remained distinct. And these floras were adapted to different climates. The real explanation of their distribution probably is that they occupied climatically differentiated areas. The ice and the floras of the Permo-Carboniferous are best thought of not as separate items of evidence but as reinforcing each other. Both are primarily climatic indicators, not direct indicators of latitude or of continuity of land. To- gether, I think they prove that South Africa, southern South America, southern Australia, and India were much colder in the Permo-Carboniferous than they are now or than they were in the Pleistocene, and that the northern continents were warmer. That this distribution of climate was due to a more-southern position of all continents (except Antarctica) is a secondary inference, but a strong one. Wegenerians assert, quoting each other to prove it, that Glossopteris and its allies could not cross ocean gaps and that all the lands where the plants occurred must have been united. However, the fruiting structures of at least some glossopterids were leaflike and have been compared to sail planes, and the seeds were apparently very small, about 1/2 mm in diameter." These plants may have been wind-dispersed and, since they were often associated with glaciation, they may also have been carried on floating ice. I do not pretend to know how they really did disperse, but their distribution is not good evidence of continuity of land. On the other hand, terrestrial plants and animals cross narrow water gaps much more freely Downloaded by guest on September 27, 2021 VOL. 52, 1964 GEOLOGY: P. J. DARLINGTON, JR. 1089

than wide ones, and the distribution of the Glossopteris flora does suggest that gaps between southern continents were narrower than now. I think dispersal of this flora might have occurred across the water gaps indicated in Figure 1, especially if strong winds and ocean currents circled Antarctica then as they do now. Biogeographic Evidence.-Coal: Coal and apparent (Glossopteris leaves and coni- fer needles have been found about 30 from the present South Pole.'2 And a Glos- sopteris flora existed on Tasmania and coal was formed there'0 at a time when paleomagnetism places the Tasmanian corner of Australia on or within the Antarctic Circle." Whether the continents were in their present positions or in positions indicated by paleomagnetism, the Glossopteris flora existed and coal was formed in polar or subpolar regions. Coal on Antarctica indicates only that that continent was warmer than now, not that it was in a different position. Permo-Carboniferous Revolutions of Climate.-Paleomagnetism shows Australia almost motionless, with its Tasmanian corner near the South Pole, from the Car- boniferous to the Jurassic.'3 During most of this time even Tasmania was ice- free and forested. But during part of this time ice sheets formed not only on Tas- mnania but across the whole southern half of Australia, and then the ice sheets disappeared again. If paleomagnetism is evidence of movement of continents, it is also evidence that Australia moved little during this time and that the glaciation of this and presumably of other southern continents was brought on and ended by revolutions of climate, not (as often claimed) by movements of land. Actually, glaciation such as occurred in the southern hemisphere probably required both a more-southern position of the land and a climatic revolution comparable to that of the Pleistocene. So far as occurrence of coal and of a glacial cycle are concerned, the southern continents apparently may have lain far south (as in Fig. 1) and motionless throughout the time under consideration. Paleomagnetic Evidence.-If certain assumptions are correct, as I think they probably are, a paleomagnetic determination from a given rock formation establishes an arrow pointing at the position of the North or South Pole when the rock was formed, like a man pointing an arrow at a mark on the ground. (This is a gross oversimplification, but true for present purposes.) The dip of the paleomagnetic arrow shows the distance to the pole, and the direction of the arrow shows the orientation (not the east-west position) of the continent concerned. But the continent carrying the arrow may be anywhere around the earth at the distance from the pole and with the orientation indicated, just as a man may place himself anywhere on a circle around a mark on the ground and keep his arrow always pointing at the mark. Therefore, longitudes cannot be determined from paleomagnetic data. This fact is not always emphasized by Neowegenerians. Practically, paleomagnetic evidence' is further limited by enormous gaps that still exist in the known record before the Tertiary. Determinations for the (mid-)- Mesozoic are scattered and not very satisfactory. They show Africa and South America near their present latitudes, Australia and India far southward, and Antarctica slightly off the South Pole. Determinations for the Carboniferous and Permian are satisfactory apparently only for Europe, North America, and Australia. Australia, including Tasmania, is therefore the only place where the paleomagnetic record can be compared directly with the record of glaciation and of floras in the Downloaded by guest on September 27, 2021 1090 GEOLOGY: P. J. DARLINGTON, JR. PROC. N. A. S.

southern hemisphere in the Permo-Carboniferous. The comparison gives exciting new information about the limits and tolerances of southern coal floras and about the revolutions of climate that apparently accompanied glaciation (preceding pages). It is unfortunate that, although paleomagnetism places India far south of its present position, about halfway between the Equator and the South Pole, probably in the Jurassic, no rocks have yet been found in India suitable for paleomagnetic determinations between the Cambrian and the Jurassic. 15 Nevertheless, in spite of the limits and gaps, the paleomagnetic record does agree in general with the joint record of glaciation and of distribution of floras. Both records strongly suggest that most continents lay farther south in the past than now, and that northward movements have occurred. Why should the two independent records suggest this, unless in fact northward movements have occurred? I think they probably have. Conclusions.-Evidence given in as much detail as space allows (I have made some extreme oversimplifications!) suggests that Africa and South America probably were united but that they separated not later than the Triassic and perhaps earlier, so long ago that no clear traces of the union are visible in distribution of existing life. Other southern continents were probably not united. All continents (except Antarctica) probably lay farther south on the world late in the Paleozoic than they do now, perhaps in positions suggested in Figure 1, and no changes of position are required to account for the distribution of southern floras and of coal or for the onset and ending of glaciation during the Permo-Carboniferous. Since then, a general northward movement of continents (except Antarctica) has apparently occurred. The paleomagnetIc record suggests that different pieces of land moved northward at different times and for different distances. However, the continents may have moved in nearly straight lines from the positions indicated in Figure 1, to the latitudes indicated by paleomagnetism in the Mesozoic, to present positions, with no deviation and not much rotation. A general northward movement of continents, excepting Antarctica, might be correlated with the existing ridge-rift system of the earth. Apparent ridge-rift lines on the ocean bottom almost surround Antarctica3 and may mark a line of rising and spreading convection currents in the earth's mantle that may have moved the other continents northward. However, I am not a geophysicist, in fact not even a geologist, and I do not want to promote any particular geophysical hypothesis.

* My work in Tasmania and Australia was supported in part by a John Simon Guggenheim Memorial Foundation fellowship, and that in South America, by National Science Foundation grant GB-93. 1 Carey, S. W., in Continental Drift (Hobart: Geology Department, Univ. of Tasmania, 1958), pp. 218-225, Fig. 21; Dicke, R. H., Science, 138, 659 (1962), Fig. 3. 2 Wilson, J. T., Nature, 198, 928 (1963), Fig. 6; Sci. American, 208, 86-100 (1963). 3Heezen, B. C., and M. Ewing, in The Sea (New York: John Wiley & Sons, 1963), vol. 3, The Earth Beneath the Sea, pp. 388-410; Girdler, R. W., Nature, 198, 1037 (1963), Fig. ic. 4Creer, K. M., Ann. Geophysique, 14, 373-390 (1958). van Steenis, C. G. G. J., Blumea (Rijksherbarium, Leiden, Netherlands), 11, 235-542 (1962). Tropical angiosperm floras are compared on pp. 247-253, 314. 6du Toit,_ A. L., Our Wandering Continents (Edinburgh: Oliver and Boyd, 1937), with map of "Gondwana" facing p. 58; Doumani, G. A., and W. E. Long, Sci. American, 207, 180 (1962). Downloaded by guest on September 27, 2021 VOL. 52, 1964 ANATOMY: BYERS AND PORTER 1091

7Harrington, H. J., Bull. Am. Assoc. Petrol. Geologists, 46, 1773-1814 (1962). See especially map on p. 1775. 8 Schwarzbach, M., Climates of the Past (London, Princeton, etc.: D. Van Nostrand Co., 1963). See p. 150 for criticism of Brooks' attempted explanation of Permo-Carboniferous glaciation; pp. 140 (India) and 147-148 (South America) for references to opposing ideas of direction of movement of Permo-Carboniferous ice; p. 134 for diagrammatic map of distribution of Permo-Carboni- ferous floras. 9 King, L. C., in Descriptive Palaeoclimatology, ed. A. E. M. Nairn (New York and London: Interscience Publishers, 1961), p. 310, Fig. 1. 10 Banks, M. R., "The geology of Tasmania," J. Geol. Soc. Australia, 9, part 2 (1962). See pp. 195 (Fig. 30b), 214 for center and movements of Permo-Carboniferous ice on Tasmania; pp. 204, 211, 217, 223-224 for Permian and Triassic coal deposits. 11 Plumstead, E. P., Palaeontographica, 100, 1-25, P1. 1-14 (1956). See p. 7 for size of seeds and p. 21 for "sail plane" structures and wind dispersal of Glossopteris. 12 Barghoorn, E. S., in Science in Antarctica, NAS-NRC Pub. no. 839 (1961), part 1, pp. 5-9. 13Runcorn, S. K., Continental Drift (New York and London: Academic Press, 1962). See p. 26, Fig. 22, for positions of Australia in relation to the South Pole indicated by paleomagnetism; p. 33, Fig. 24, for attempted reconstruction of southern hemisphere for mid-Mesozoic time; but note that paleomagnetism shows only (apparent) latitudes, not longitudes, and does not justify swinging Australia around the pole or grouping all the continents on one side of the pole. 14 Cox, A., and R. R. Doell, Bull. Geol. Soc. Am., 71, 645-768 (1960). This paper includes a good discussion of paleomagnetic theory and a 55-page table of actual determinations. See p. 762 for relatively unsatisfactory nature of Mesozoic determinations; pp. 761-762, Fig. 35, for Carbon- iferous and Permian determinations for Europe, North America, and Australia. 15 Athavale, R. N., et al., Geophys. J., 7, 304-313 (1963).

ORIENTED MICROTUBULES IN ELONGATING CELLS OF THE DEVELOPING LENS RUDIMENT AFTER INDUCTION* BY BRECK BYERS AND KEITH R. PORTER

BIOLOGICAL LABORATORIES, HARVARD UNIVERSITY Communicated August 7, 1964 Experimental and histological studies of lens induction have consistently failed to reveal the mechanism of interaction between the optic vesicle and the presumptive lens ectoderm. Even recent electron microscope examination of this phenomenon in mammals by Cohen' and in birds by Hunt2 and by Weiss and Jackson,3 though effective in describing in greater detail changes in cell organelles and base- ment membranes, has failed to show an interpretable induction mechanism or, for that matter, any structural changes seemingly involved in such manifestation of induction as shape changes in the affected cells. To some extent these initial in- vestigations were hampered by the inadequacies of the then available fixing reagents for preserving fine structure. Thus, with the more recent introduction of glu- taraldehyde and related compounds and the demonstration of their efficacy in preserving elements of fine structure previously lost, it became important to ex- plore again this particular phenomenon-and to look especially at the early stages of cell elongation. The results describe a prominent development of "microtubules" in the cortical regions of presumptive lens ectoderm cells subsequent to induction and especially as the cells palisade to form the lens placode. Additional observa- Downloaded by guest on September 27, 2021