local, detailed studies. While most of Antarctica hasbeen geologically mapped on a reconnaissance scale,the remaining unknown areas must be investigated.At the same time, more detailed mapping of certainareas is necessary, both to clarify the structuralframework of the continent and to test fully thecorrectness of du Toit's predictions.

The basement rocks of East Antarctica requirenuch more investigation. Those in the Transantarc-ic Mountains appear to be younger than those in'oastal East Antarctica; if correct, this fact is im-ortant in understanding the tectonic evolution ofhe East Antarctic shield. Detailed comparativetudies of the ancient rocks of coastal East Antarctica

Ztheir assumed counterparts on the matching

oasts of Africa, India, and Australia are needed.Radiometric age determinations on these igneous andmetamorphic rocks will be a great help in testing theoriginal continuity of basement rock terranes in Ant-arctica with those in the other southern lands.

Detailed studies of the Beacon Group are just be-ginning, and it deserves examination along the lengthof the Transantarctic Mountains to establish lateralchanges in composition, age, and origin of thesestrata. Thorough knowledge of the coastal localitiesis especially important to allow careful comparisonwith the Gondwana successions in Africa, India, andAustralia. Additional fossil discoveries can be ex-pected in the Beacon Group and in the stratifiedrocks of West Antarctica.

Two kinds of studies of the geologic structure inAntarctica will enable further testing of the Gond-wanaland reassembly. On the one hand, additionalknowledge of the areal geology will lead to betterdefinition of the tectonic provinces of the continent,such as the Ellsworth Mountains fold belt, and tomore accurate comparison of these provincial bound -aries in Antarctica and the other Gondwana frag-ments. On the other hand, thorough study of thestructural style within each province is necessary totest the supposed continuity between that provincen Antarctica and a tectonic province in another'outhern land.

Finally, continued investigation of paleomnagiietisrn11 Antarctica should provide valuable insights into

its geologic history. Although this technique is corn-nonly a complicated one in l)1'a(t1('e, it (an, underavorahie circumstances, he used to establish paleo-

latitudes. Thus, the paleomagnetism locked in ant-arctic rocks, together with the mnagnetic-anorrialvbelts frorn the ocean floor, offer hope of direct con-firmation, not only of the reality, but also of thechronology of continental drift.

Reference

du Toit, Alexander L. .1937. Our Wandering ContinentsEdinburgh, Oliver and Boyd. 366 p.

The Fossil Tetrapods ofCoalsack Bluff

EDWIN H. COLBERT*

Museum of Northern Arizona

Tetrapods and Continental Drift

The evidence for an ancient continent of Gond-wanaland, the fragments of which now constitutethe several Southern Hemisphere land masses andthe peninsula of India, is various. Much of it is veryconvincing. Moreover, the numerous lines of independent data would seem to fit together reasonablywell, giving us an integrated concept of the supposedGondwana continent and its fragmentation. Yet,almost all of the evidence is equivocal in one way oranother, resting to a greater or lesser degree on cer-tain basic assumptions, the validity of which some-times may be called into question.

The theory of Gondwanaland and continentaldrift must be in accord with all of the evidence if itis to be valid. In this regard, the records of the rocksand fossils are particularly important. As the re-nowned geologist Walter Bucher remarked in 1964:"Ultimately the proof [for drift] must come from thegeologic and paleontologic record."

One of the problems that has faced the paleontol-ogists, particularly the vertebrate paleontologists, isthat drift has not been essential to explain the dis-tribution of fossils on the southern continents. Thisis certainly true for the land-living vertebrates ofCenozoic age: their distributions, from the beginningof the Cenozoic era to the present, are on the wholevery readily explained by the present intercontinen-tal land connections, notably the isthmian link be-tween the two Americas and the undoubted Bering.Strait "bridge" that periodically connected the NewWorld with the Old throughout the past 60 millionyears or more of geologic history. For older fossilvertebrate faunas, however, the concept of a former(;ondvarsa1and that subsequently fragmented, itsseveral fragments—now the Southern Hemisphereconti nents---dri fting to their present positions, hasbecome increasingly attractive within the past twodecades. The very close resemblances between theTriassic reptiles of South Africa and South America,for example, would seem to indicate a close conti-guity of these continents, with a single, unified rangefor the land-living animals then inhabiting whatare now widely separated land masses. Yet, although

*Cura tor Emeritus, The American Museum of NaturalHistory, and Professor Emeritus, Columbia University.

May-June 1970 57

contiguity of the continents and an unbroken rangeof distribution for the reptiles typical of them inTriassic times would most readily explain the veryclose taxonomic relationships—even identities—thatare becoming increasingly apparent among thesefossils, one cannot completely rule out the possibilityof Triassic intercontinental migrations by the longway around—by a trek up through Africa and acrossAsia, across the Bering bridge, down through NorthAmerica, across the isthmus of Panama, and intoSouth America. Improbable, one might say, but, none-theless, a distant possibility.

In the light of these considerations, Antarctica'srole has acquired unusual importance with respectto the theory of Gondwanaland and continentaldrift. Fossils of land-living vertebrates in Antarcticawould be especially crucial since, for such animals,there would be no "long way around" to get fromAfrica or other Southern Hemisphere land masses towhat is now the south polar continent.

But, until a little more than two years ago, nofossil vertebrates of significant geologic age had beenfound on the Antarctic Continent. The role of Ant-arctica within the problem of Gondwanaland anddrift was quite unknown, so far as the evidence ofland-living vertebrates was concerned.

In December of 1967 Peter Barrett, then of OhioState University, discove a fragment of a fossiljaw in the Triassic Fremouw Formation at GraphitePeak in the Transantarctic Mountains, near theBeardmore Glacier. The specimen (submitted to mefor study) proved to be the back portion of a mandib-ular ramus of a labyrinthodont amphibian, a fossilrepresentative of a group of animals that inhabitedthe continents of the earth in great profusion duringlate Paleozoic and Triassic times. The specimen wastoo fragmentary to enable a more precise identifica-tion.

For the time being, this identification sufficed,however. The discovery of the jaw fragment andthe clue to its zoological affinities created worldwideinterest among scientists of many callings, as well asamong the general public. Here was a fossil of greatsignificance, not only because it afforded the firstglimpse of truly ancient land-living vertebrates inAntarctica, but also because it might be a harbingerof things to come.

A Plan of Operation

Having worked on this isolated fossil, I stronglyurged that an expedition, composed of vertebratepaleontologists, should return to Graphite Peak tosearch for more materials. It seemed logical to sup-pose that where there was one specimen, there wouldbe more. To improve the odds of finding them, the

search should be conducted by men trained in thetechniques of vertebrate paleontology, men whowould be devoting their full time and energies in thefield to fossil bones, and nothing else.

Accordingly, after much planning by the Instituteof Polar Studies of Ohio State University, workingin close cooperation with the National ScienceFoundation, a search for Triassic vertebrates in Ant-arctica was organized as part of the U.S. AntarctiResearch Program for 1969-1970. Four paleontol-ogists were to prosecute this work as members oflarge group of geologists and paleontologists whowould be carrying out comprehensive studies in thBeardmore Glacier region of the TransantarctiMountains. The party as a whole, consisting of about20 scientists, was to be under the leadership of Dr.David H. Elliot of Ohio State, a man of much ext-perience in antarctic geology and exploration. Thfour vertebrate paleontologists were: Edwin H. CoiLbert, of the American Museum of Natural History,New York, and the Museum of Northern Arizona,Flagstaff, leader; William J . Breed, of the Museumof Northern Arizona; James Jensen, of BrighamYoung University, Provo, Utah; and Jon S. Powell,of the University of Arizona, Tucson. Maps werecarefully studied, and a series of likely localities forTriassic vertebrates were selected (largely on thebasis of Elliot's knowledge and experience) for ex-ploration with close helicopter support.

Coalsack Bluff

There is no need to dwell on the frustrations anddelays that occurred between the departure of theparty from the U.S.A. and its arrival at the Beard-more camp. It was a bad spring weatherwise in Ant-arctica, and schedules inevitably were held up. Butthe group finally did-arrive—on November 22, 1969—and established itself in camp in the midst of ahowling storm.

The next morning, most of the personnel of thparty remained in camp to get settled, for arrangeL

-ments upon arrival the day before were necessarilhurried and sketchy. But David Elliot and one otwo others decided to go and take a lok at CoalsacBluff, the nearest outcrop of rocks, located aboutthree or four miles from camp.

And now we see how the discovery of fossils s4frequently depends on luck and serendipity. Elliohad sited the Beardmore Camp near Coalsack Bluflargely because it was a good place for supply planeto land. He went out that first morning to look at theBluff largely because it was near at hand, and alsobecause he wanted to prove to James Schopf, thepaleobotanist in our group, that certain strata withinthe Bluff belonged to the Triassic Fremouw Forma-

58 ANTARCTIC JOURNAL

tion. He proved that to his satisfaction, and he foundtwo small pieces of what looked like bone scrap.

The impatient explorers returned to camp at noon,and Elliot showed the two scraps to me. I was ableto confirm that they were pieces of fossil bone.

After lunch, a group of us—the vertebrate pale-ontologists and other interested members of theparty—returned to Coalsack Bluff. There, in a seriesof low cliffs that stretched for perhaps the betterpart of a mile, were numerous fossil bones. Thesediments, dipping at low angles, were cross-beddedchannel sands, deposited in streams, and the fossilsoccurred within them as isolated bones. It was evi-dent that here were the remnants of animals thathad perished and been washed down the streams,their carcasses being dismembered and the bonesscattered as a result of the vigorous action of thetumbling water. Thus, there were no articulatedskeletons, which was a pity, but there were numerousbones, and on the whole they were nicely preserved.

Virtually all of the fossils at Coalsack Bluff werecollected from a single stratigraphic horizon, butfrom two topographic levels. The bulk of the collec-tion was made along the low cliffs, already described,where isolated bones were found at one spot afteranother. We simply started at one end of the ex-posures and worked our way along, day after day,chiseling out the fossils, hardening them and, whenneeded, treating them by specially devised methodsto preserve them. A part of the collection, however,was made several hundred feet downslope from thecliffs, within a down-faulted monoclinal block inwhich the sediments were almost perpendicular.Here, the fossils occurred in sufficiently large con-centrations to be quarried. Unfortunately, most ofthese fossils were fragmentary and not as nicelypreserved as the specimens high on the cliff.

Incidentally, we subsequently learned that we werefantastically lucky in having found the CoalsackBluff deposit. Explorations at other locations in theBeardmore region failed to reveal any concentrationsof bones, although scattered specimens were discov-ered. Such occurrences are characteristic of fossilvertebrates deposited in freshwater sediments: greatstretches of rock exposures will be barren of fossils,or almost so, but here and there, where the conditionsfor burial and preservation were just right, there willbe remarkable accumulations of fossils. Undoubtedly,other fossil pockets comparable to Coalsack Bluffexist, but it may take a great deal of hard prospecting'to find them.

The discovery of bones at Coalsack Bluff deter-mined the activities of the vertebrate paleontologistsfor the remainder of the 1969-1970 antarctic fieldseason. In short, we spent most of our time collectingfossils instead of searching for them, as had beenour original intention. This was lucky for us because,

as a result of unexpected helicopter troubles, muchof the reconnaissance work that had been scheduledfor the Beardmore party was restricted or cancelled,while we could carry on our work without the heli-copters. About 450 fossil specimens were collected,catalogued, and packed for shipment back to theUnited States.

The Coalsack Bluff Fossils and Gondwanaland

At an early stage in our collecting work, it becameapparent that we were dealing with a characteristicSouthern Hemisphere tetrapod fauna. There werenumerous bones of labyrinthodont amphibians. Therewere also numerous bones of various reptiles: mam-mal-like or therapsid reptiles, so characteristic of theTriassic of southern Africa and South America (andof peninsular India as well), and also thecodonts-those Triassic reptiles that gave rise to the dinosaurs,crocodiles, flying reptiles, and birds.

It seemed evident that we were working in rocksof probable early Triassic age. Consequently, I kepthoping to find Lystrosaurus, a fossil reptile that isextraordinary abundant in—and characteristic of—the earliest Triassic rocks of South Africa. We hadsome bones that looked suspiciously like Lystrosaurus,but field identifications can be tricky, especially whenone is dealing with bones that are not highly diagnos-tic. Then, on December 4, we at last recognized Ly-strosaurus, collected by James Jensen. It was only aportion of a right maxilla from the skull, containingone of the big teeth typical of this reptile, but itseemed to me to be enough. The shape of the bone,and its topography, were characteristic, and I couldnot think of anything else this fossil might be. Sub-sequently, on our return to the U.S.A., it was possibleto compare this fossil and various other antarcticbones with specimens of Lystrosaurus from SouthAfrica in collections in Berkeley and New York.These comparisons confirmed the identification be-yond all doubt. In Antarctica we had found Lystro-saurus, essentially identical to Lystrosaurus murrayifrom the Lower Triassic sediments of South Africa.

Photo by Wm. J. Breed

Lysfrosaurus jawfragment:rightmaxilla from theSkull showing thetip of the righttooth. This wasthe specimen thatmade identifica-tion of Lystrosau-rus possible in the

field.

May-June 1970 59

Lystrosaurus skeleton, on display in South Africa.Photo by Wm. J. Breed from an article by A. W.Crompton

Restoration of Lystrosaurus. Reprintedfrom The Age of Reptiles with the per-mission of W. W. Norton and Co. Copy-

right 1965 by Edwin H. Colbert.Drawing by Margaret M. Colbert

At present, this is the only generic identificationthat it has been possible to make with certaintyamong the materials from Antarctica. The collectionwill require much close study. But the presence of Ly-strosaurus in the Fremouw Formation of the Trans-antarctic Mountains is an occurrence of unusualsignificance.

This rather strange reptile—about as large as amedium-size dog, with a heavy body, short limbs andbroad feet, and a remarkably specialized, beakedskull, in which all teeth except a single large "ca-nine" tooth on each side have been suppressed—isso characteristic of the lower phase of the UpperBeaufort beds in South Africa that these sedimentshave been designated the "Lystrosaurus Zone." Ly-strosaurus skeletons, virtually identical with the SouthAfrican fossils, are found in the Lower Triassic Pan-chet beds of the Indian peninsula and in the Tung-hungshan beds of Sinkiang, China. Now, Antarcticacan be added to these localities.

Lystrosaurus was a denizen of the land. Althoughit shows certain specializations, such as the highposition of the nostrils, indicating that it may havebeen aquatic to the extent of living in and aroundrivers and lakes—as do modern tapirs—there is noevidence, either in its anatomy or in its sedimentaryoccurrences, that it was capable of venturing acrossoceanic barriers, even relatively narrow ones. More-over, Lystrosaurus is associated with dry-land reptilesin Africa. India, and Antarctica. Therefore, the dis-

tributions of Lystrosaurus and associated reptiles andlabyrinthodont amphibians—which were freshwater-pond and stream dwellers—must necessarily beviewed as indicating former dry-land connectionsbetween the regions in which these fossils are nowfound.

It is quite possible to envision the migration ofLystrosaurus from southern Africa to India and towestern China within the confines of existing con-tinental relationships. But the presence of Lystro-saurus and associated reptiles and amphibians in Ant-arctica confronts us with a fact that cannot be relatedto modern continental distributions. There are twopossibilities: either these ancient tetrapods reachedAntarctica by way of a long, isthmian land-bridge,or they got there when Antarctica and Africa werepart of a large, single continent.

The land-bridge hypothesis does not seem probableon the face of present geological knowledge. An ob-vious route would be by way of the Scotia Arc, be-tween the tip of South America and the AntarcticPeninsula. But there is no evidence that any part ofthe Scotia Arc dates back to the beginning of Tri-assic times; it is essentially a continuation of theAndean system. Therefore, all probabilities point toclose ligation between Africa and Antarctica, andhere the distribution of what may be called the Ly-strosaurus fauna is crucial.

If the southern continents are brought togetheraccording to the most modern interpretations of theconfiguration of Gondwanaland, the areas contain-

60 ANTARCTIC JOURNAL

).

((-'S

Above: Laurasia and Gondwanaland: a map prepared byE. H. Colbert, showing Triassic tetrapoci localities. The orien-tation of Antarctica-Australia follows Hurley (ScientificAmerican, 1968). Below: Gondwanaland: a map preparedby E. H. Colbert, showing Lower Triassic tetrapod localities.Orentation as above. Legend: B—Blina Formation; C-Cynognathus zone of Upper Beaufort beds; F—FremouwFormation; G—Gosford Formation; K—Knocklofty Forma-tion; L—Lystrosaurus zone of Upper Beaufort beds; P-Panchet Formation; PV—Puesto Viejo Formation; Y—Yerra-path Formation; (A)—labyrinthodont amphibian; (K)—kan-nemeye rid; (KC)—Kannemeyeria, Cynognathus; (L)—Lys-trosaurus; (P)—Parotosaurus; (R)—rhytidosteid labyrintho-

dont.

ing the Lystrosaurus fauna—now so widely separated—are brought into reasonably close proximity. Sucha distribution would explain, among other things,the presence of Lystrosaurus in these several regions.It would explain the occurrences of Lystrosaurus andassociated tetrapods as an uninterrupted range ofdistribution—a range that was subsequently inter-rupted and dispersed by the rifting of Gondwana-land and the drifting of its component parts to theirpresent positions. Moreover, it would explain howthe Lystrosaurus fauna, composed of amphibians andreptiles that almost certainly lived in tropical or sub-tropical habitats, can now be found in Antarctica.Antarctica in the Lower Triassic would seem to havebeen in much lower latitudes than at present.

Conclusions

The discovery of a Lystrosaurus fauna in Antarc-tica would seem to establish beyond any reasonabledoubt the fact that the Antarctic Continent was oncepart of a larger Gondwana continent and was con-tiguous to Africa. Although many of the evidencesfor continental drift are based on certain assump-tions, there are no questionable assumptions, it seemsto me, concerning the distribution of the Lystrosaurusfauna. One can say with confidence that the amphib-ians and reptiles that constituted this fauna wereunable to swim long distances across oceanic depths,and that they perforce must have spread by way ofdry land. Such being the case, the only explanationfor the Lystrosaurus fauna in Antarctica is to pred-icate the ligation of the south polar continent withAfrica. The evidence for Gondwanaland (and Lau-rasia, or Pangaea) has been accumulating in im-pressive dimensions during the past two decades. Thepresence of Lystrosaurus in Antarctica constitutes afinal—and remarkably strong—link in this chain ofevidence.

References

Barrett, Peter J. 1969. Stratigraphy and Petrology of theMainly Flu viatile Permian and Triassic Beacon Rocks,Beardmore Glacier Area, Antarctica. Ohio State Univer-sity. Institute of Polar Studies. Report No. 34 (mimeo-graphed). 132 p.

Barrett, Peter 3 . , R. J . Baillie, and E. H. Colbert. 1968.Triassic amphibians from Antarctica. Science, 161: 460-.462.

Bonaparte, J . F. 1967. New vertebrate evidence for African-South American connexions during the lower or middleTriassic. Paleontology, 10: 554-563.

Colbert, Edwin H. 1966. The Age of Reptiles. New York,The Norton Library, W. W. Norton and Co., Inc. 228 p.

Colbert, Edwin H. 1970 (in press). Antarctic fossil verte-brates and Gondwanaland.

Romer, Alfred S. 1968. Fossils and Gondwanaland. Ameri-can Philosophical Society. Proceedings, 112: 335-343.

Smith, A. Gilbert and A. Hallam. 1970. The fit of the south-ern continents. Nature, 225: 139-144.

May-June 1970 61