THE UNIVERSITY 
 MI ILLINOIS 
 
 v-w- 
 
 LIBRARY 
 
 Y’*‘ vf 
 
 55Q 
 
 D95<? 
 
 GE0L0GY1 
 
 V ^ ■.>**>* 
 
-1 
 
 
 The person charging this material is re¬ 
 sponsible for its return to the library from 
 which it was withdrawn on or before the 
 Latest Date stamped below. 
 
 Theft, mutilation, and underlining of books 
 are reasons for disciplinary action and may 
 result in dismissal from the University. 
 
 UNIVERSITY OF ILLINOIS LIBRARY AT URBANA-CHAMPAIGN 
 
 L161 —0-1096 
 
CARNEGIE INSTITUTION OF WASHINGTON 
 
 Publication No. 381 
 
 UNIVtni. 
 
 ! \ )' 
 
 Of i i _ ! I V 1 
 
 IQ 
 1 .3 
 
 1927 
 
PRESS OF J. B. LIPPINCOTT COMPANY 
 PHILADELPHIA, PA. 
 
A GEOLOGICAL COMPARISON OF SOUTH 
 AMERICA WITH SOUTH AFRICA 
 
 BY 
 
 ALEX. L. DU TOIT, D.Sc., F.G.S. 
 
 WITH A PALAEONTOLOGICAL CONTRIBUTION 
 
 BY 
 
 F. R. COWPER REED, M.A., Sc.D., F.G.S. 
 
 Published by the Carnegie Institution of Washington 
 
 Washington, 1927 
 

 
 
 
 
 
 
 - . 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
CONTENTS 
 
 •TT? 
 
 D <? ^ 
 
 PAGE 
 
 Chapter I. Introduction . i 
 
 Chapter II. Geological Framework of the South Atlantic. 6 
 
 Section A. The western side of Africa. 6 
 
 Section B. The Falkland Islands. n 
 
 Section C. South America. 13 
 
 Section D. Generalized comparisons between the two continents. 15 
 
 Chapter III. Argentina and Bolivia. 18 
 
 Section E. The sierras of the District of Buenos Aires. 18 
 
 Section F. The southern pre-cordilleran region. 27 
 
 Section G. The pre-andine belt of northern Argentina and of 
 
 Bolivia. 55 
 
 Chapter IV. Pre-Gondwana Basement of Eastern South America. 59 
 
 Section H. The Devonian system. 59 
 
 Section I. The Early Palaeozoic beds. 62 
 
 Section J. The pre-Cambrian formations. 64 
 
 Chapter V. Gondwana System in Brazil, Uruguay, and Paraguay. 66 
 
 Section K. The Parana Basin. 66 
 
 Section L. The northeastern region of Brazil. 94 
 
 Section M. Eastern Brazil. 95 
 
 Section N. The Cretaceous of the interior. 96 
 
 Chapter VI. Geological History of the Afro-American Land-mass. 97 
 
 Chapter VII. Bearing on the Displacement Hypothesis. 109 
 
 Chapter VIII. Conclusion. 118 
 
 Bibliography. 121 
 
 Appendix. Upper Carboniferous Fossils from Argentina. 129 
 
 Triassic Fossils from Rio Claro, Parana, Brazil. 150 
 
 Index. 151 
 
 ILLUSTRATIONS 
 
 Plates I to XVI follow page 158 
 
 Geological map of part of South America. In pocket at end of volume 
 
 TEXT-FIGURES PAGE 
 
 1. Succession on the northeastern side of the Sierra de la Ventana. 20 
 
 2. Geological sketch-map of a small area along eastern base of the Sierra Chica de 
 
 Zonda to the south of San Juan. 29 
 
 3. Section, about 75 meters in height, in ravine north of the Rio Grande, Sierra 
 
 Chica de Zonda. 30 
 
 4. An infolded patch of tillite in Palasozoic beds close to Leoncito Encima, Barreal 33 
 
 5. General succession at Potrerillos. 49 
 
 6. Section, 10 km. in length through Taquara, Rio Grande do Sul, showing the 
 
 Triassic volcanics resting on an undulating surface of the false-bedded Rio do 
 Rasto sandstones. 9 1 
 
 7. Suggested continental restoration under the Displacement Hypothesis. 116 
 
 v 
 
 1 
 
A GEOLOGICAL COMPARISON OF SOUTH AMERICA 
 
 WITH SOUTH AFRICA 
 
 By 
 
 Alex. L. Du Toit, D. Sc., F. G. S. 
 
 With twelve plates, one map, and seven text-figures 
 

 
 
 
 
 ■ 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 ' 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
CHAPTER I 
 
 INTRODUCTION 
 
 With the passing of time and the growth of knowledge there has 
 come about a fuller realization of the extremely prominent part 
 played by the ancient continent of ‘ ‘ Gondwanaland ” in the past 
 history of this earth. It does not detract from its importance that 
 owing to lack of data the limits set to the land-masses during the 
 various epochs are not always properly known, or, more correctly, 
 the limits as seen in the relics preserved in the existing continental 
 masses of the globe. 
 
 We are, on the other hand, still forced to speculate as to how these 
 various sections, as we now find them, were bound together at the 
 beginning of the Mesozoic and how they ultimately came to be 
 parted by thousands of kilometers of ocean. 
 
 During the early part of the century and with the limited 
 knowledge available, it sufficed to assume that the fragments, as 
 we see them in South America, Africa, Madagascar, India, Australia, 
 and Antarctica, were linked together at some periods in the late 
 Palaeozoic and in the Mesozoic by equally extensive land connections, 
 or else by relatively narrow “land bridges,” which at later stages 
 became submerged by the ocean. Geological and geophysical 
 researches within the past decade have, however, rather severely 
 shaken our confidence in such orthodox views, which until then had 
 commended themselves by their very simplicity. Astonishing simi¬ 
 larities—stratigraphical, lithological, and palaeontological—between 
 regions now parted by ocean came to obtrude themselves, becoming 
 all the more striking in that the fossil remains contained in these 
 formations were not marine, but entirely terrestrial. Almost identical 
 climatic vicissitudes and sequences could be deduced for the several 
 sections—glacial to start with and arid toward the close—though 
 impossible to explain under current theories, followed by volcanicity 
 on a truly enormous scale. Major tectonic structures found their 
 homologues across the wide intervening seas. Furthermore, the 
 marine Tertiary deposits fringing the present continental masses not 
 only indicated how different must have been the distribution of land 
 and water during that late era, but hinted at an unexpectedly recent 
 date for the origin of the existing oceanic basins. 
 
 The hypothesis of “continental disruption” or the “displace¬ 
 ment hypothesis, ” brought forward independently by F. B. Taylor 1 
 and A. Wegener , 2 that the present continents have owed their posi- 
 
 1 Taylor (1910). 
 
 2 Wegener (1912). 
 
 1 
 
2 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 tions in great part to the breaking-up of much larger land-masses in 
 the Cretaceo-Tertiary through the “drifting apart” of the crustal 
 fragments, has on the contrary helped to bring numerous other 
 irreconcilable facts into harmony and would almost seem to have 
 provided us with the key to the riddle of the structure of the earth 
 and more particularly of the history of “Gondwanaland. ” 
 
 In 1916, 3 though in ignorance of the details of Wegener’s doctrine, 
 the writer put forward an hypothesis somewhat different from those 
 of Taylor and Wegener, the outcome of the study mainly of the 
 glacial deposits of the Karroo system, supplemented by a personal 
 acquaintance with their Australian equivalents. Later developments 
 having seemingly rendered these ideas more and more worthy of 
 serious consideration, it became one of his aims to visit South Amer¬ 
 ica in the hopes of being able, while gathering information about the 
 Gondwana deposits of that continent, to discover any evidence 
 favorable or adverse to the “displacement hypothesis. ” 
 
 Following the sympathetic representation by Dr. Fred. E. 
 Wright, of the Geophysical Laboratory of the Carnegie Institution 
 of Washington, and Dr. Reginald A. Daly, of Harvard University, 
 the president (Dr. John C. Merriam) and trustees of the Carnegie 
 Institution of Washington, appreciative of the scientific importance 
 of such a mission, most generously offered a grant in aid, and, making 
 use of my six months of leave from official duties, fortunately then 
 available, I was enabled to cross the Atlantic and spend five months 
 in Brazil, Uruguay, and Argentina. To the Carnegie Institution of 
 Washington, in thus promoting research in the Southern Hemisphere, 
 I am accordingly under the deepest of obligations, while I have fur¬ 
 ther to thank them also for their generosity in publishing the results 
 of these investigations, rather unfortunately delayed by reason of 
 pressing official and other duties. 
 
 Taking into account the enormous size of the region traversed 
 and the time necessarily consumed in traveling, the opportunities 
 for original investigation were naturally limited, and it was often 
 with keen regret that a particular line of research had to be dropped 
 just as it was disclosing results of consequence. Nevertheless, the 
 knowledge so acquired has permitted the copious scattered literature 
 to be read with a certain amount of discrimination and has enabled a 
 closer comparison of the Gondwana system of South America to be 
 made with that of South Africa than has hitherto been possible. 
 Papers in the English language on the geology of any part of South 
 America are relatively few, while those in Spanish or Portuguese 
 dealing comprehensively with any one of its republics are equally 
 rare. Consequently, instead of limiting this account to the points 
 investigated personally, the opportunity has been taken, although at 
 
 * Du Toit (1916). 
 
INTRODUCTION 
 
 3 
 
 the risk of undue length, of reviewing in addition the geology of 
 South America in so far as it may have a bearing on the problem of 
 Gondwanaland and of a former connection between Africa and the 
 New World. The extraordinarily divergent conclusions reached on 
 many vital questions by various eminent South American geologists 
 is eloquent proof indeed as to the need of such a summary for 
 the benefit of English-speaking persons. Furthermore, the notable 
 advances made in recent years have rather impaired the value of that 
 monumental work by Suess, The Face of the Earth, which even the 
 French edition with its new maps and copious references to recent 
 literature just fails to restore. 
 
 In this, the first comprehensive summary of the geology of the 
 eastern part of South America, considerable liberty has been taken 
 in the interpretation of the accounts of areas which naturally could 
 not be visited, and doubtless there will be found not a few statements 
 or conclusions to which the local geologists can and will take excep¬ 
 tion. It is nevertheless hoped that such differences in the presenta¬ 
 tion and interpretation may not be of too violent a character; for 
 such the insufficiency of the data, even for the better known parts 
 of this enormous territory, must in part be held responsible. Con¬ 
 siderable difficulty was experienced in the preparation of the accom¬ 
 panying geological map; drastic alterations are required to Branner’s 
 map of Brazil, 4 while that by Leme, 5 though more up to date, is on a 
 smaller scale and not sufficiently detailed; Walther’s map of Uru¬ 
 guay 6 is valuable, but there is no general map of Argentina, while 
 much of Paraguay, Bolivia, and the adjacent part of Brazil are 
 geologically unexplored. 
 
 I found my South American confreres keenly interested in the 
 problems herein discussed, though the size of their countries, the 
 smallness of the staffs, the pressure of economic work, and the lack of 
 coordinated effort between the geological surveys of the various 
 republics and states have hindered the carrying out of any compre¬ 
 hensive scheme of work thereon. Furthermore, no one had any 
 first-hand acquaintance with any of the other countries possessing 
 Gondwana beds. I have reason to believe that interest in these 
 problems has been appreciably stimulated as the consequence of 
 this visit. While immediate results therefrom could hardly be 
 expected, attention might be drawn to a paper by Professor 
 Walther 7 on the Borings for Coal in Cerro Largo, Uruguay, and 
 one by Senor E. T. Arocena 8 on the Glacials in the Department of 
 Durazno as the outcome of certain observations made by the writer 
 in that republic. 
 
 In order to aid comparisons with the African continent, a brief 
 
 4 Branner (1919). 6 Walther (1919). 8 Arocena (1926). 
 
 6 Leme (1924). 7 Walther (1924). 
 
4 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 account is given of the geological features of its western side, the 
 literature thereon being rather scattered, with much uncertainty 
 pertaining to the territories from Angola northward, while a sum¬ 
 mary of the geology of the important Falkland Islands has also been 
 introduced for reasons that will appear in the sequel. 
 
 The schematic representation embodied in Figure 7 will assuredly 
 be subjected to active and probably to hostile criticism on the score 
 of the fantastic and apparently improbable character of the dis¬ 
 placement theory, but a close and impartial study thereof is invited, 
 so that the numerous congruences and '‘coincidences” in the stra¬ 
 tigraphy, structure, etc., of the two great land-masses now parted 
 by the South Atlantic may be duly scrutinized and pondered over, 
 the weaknesses of the arguments advanced duly pointed out, and 
 more reasonable alternatives suggested in explanation of the facts or 
 their interpretation. 
 
 Leaving Cape Town on June 12,1923, Rio de Janeiro was reached 
 a fortnight later, where the director of the Geological Service, 
 Dr. Euzebio Paulo de Oliveira, with the permission of the Brazilian 
 government, kindly detailed Dr. L. F. Moraes Rego to act as my 
 guide and instructor, under whose able direction a tour was made of 
 nearly six weeks’ duration, the department obligingly providing 
 rail transport. The program embraced a short visit to Ouro Preto, 
 Bello Horizonte, and Diamantina to inspect the ancient formations 
 of Minas and the Boa Vista Diamond “pipe.” The journey from 
 Sao Paulo to Ityrapina and Bauru provided sections of the Triassic 
 “traps” and the Cretaceous of the interior, while a trip through 
 Ponta Grossa, Iraty, Roxo Roiz, Marechal Mallet, Rio Claro, Porto 
 Uniao, and Rio Negro to Curityba enabled the Devonian as well as 
 the full succession of the.Gondwana system to be studied. Taking 
 the steamer from Paranagua to Porto Alegre, Taquara, Cachoeira, 
 Ferreira, Santa Maria, and Sant’ Anna were visited, enabling good 
 sections of the Triassic sediments and volcanics to be inspected. 
 
 In Uruguay the director of the Section of Geology and Water 
 Boring, Senor E. Terra Arocena, generously placed his records at my 
 disposal, while Dr. Karl Walther, professor of Geology at the Insti¬ 
 tute of Agronomy, Montevideo, kindly supplied me with valuable 
 information. Fraile Muerto provided confirmation of the existence 
 of Carboniferous glacial beds, originally reported therefrom by 
 Guillemain, while Melo, Paso del Cerro, Santa Rosa, and Sauce were 
 also visited. 
 
 In Buenos Aires, through the courtesy of the Minister for Mines, 
 the director of the Geological Survey, Dr. J. M. Sobral, whole¬ 
 heartedly detailed several of his staff to accompany me in turn on a 
 
INTRODUCTION 
 
 5 
 
 tour of nearly two months, namely, Dr. Franco Pastore, Dr. Augusto 
 Tapia, and Dr. Juan J. Nagera, the government defraying the cost 
 of rail transport. The route lay through Villa Dolores, Bajo de Velis, 
 Cordoba, La Cumbre, Capilla del Monte, Marayes, San Juan, Men¬ 
 doza, Cacheuta, Uspallata, Barreal, and across the Andes to Val¬ 
 paraiso in Chile and back to Buenos Aires; this enabled the various 
 horizons in the Gondwana system to be inspected. Trips were made 
 to Olavarria, Sierra de la Ventana, and La Plata to study the equiva¬ 
 lents of the “Cape Fold ranges. ” 
 
 The return journey was made by Montevideo, Santos, Sao Paulo, 
 Rio de Janeiro, and Pernambuco. Through the kindness of Dr. 
 Miguel A. de Lisboa, director of Irrigation, a rapid journey was 
 made by rail to Campina Grande and thence to the irrigation dams 
 under construction at Gargaleira and Parelhas in Rio Grande do 
 Norte, every facility being placed at my disposal by the Brazilian 
 government through the agency of Doctor Netto. The voyage there¬ 
 after was made via |Madeira| arriving at Cape Town on December 
 io, 1923. 
 
 Among the numerous other persons to whom I have been particu¬ 
 larly indebted are Drs. T. H. Lee, Paulino F. de Carvalho, and 
 Axel Lofgren, of the Brazilian Geological Service, Dr. Joviano A. 
 Pacheco, head of the Geographical and Geological Commission 
 of the State of Sao Paulo, Drs. Pablo Groeber, Roberto Beder, Juan 
 Rassmuss, Anselmo Windhausen, Ricardo Wichmann, and Hausen, 
 of the Geological Survey of Argentina, Dr. Juan Keidel, formerly 
 director of that survey, Dr. H. Schiller, of the Museum of La Plata, 
 Dr. C. Hosseus, of the University of Cordoba, and Mr. H. J. Hawley, 
 geologist to the Standard Oil Company. To my former South African 
 friends, Mr. Bernard W. Ritso, M.I.C.E., late director of the Section 
 of Water Boring in Uruguay, Mr. David Draper, F.G.S., manager of 
 the Boa Vista Diamond Mine, to Mr. F. W. Scott, M.I.C.E., engi- 
 neer-in-charge at Gargaleira, and Mr. B. H. Heatlie, of Buenos Aires, 
 I am under great obligations. Furthermore, to Mr. L. S. Goldsmith, 
 of the Geological Survey, Pretoria, is due the lettering on the Geo¬ 
 logical map. 
 
 My thanks are due to Dr. R. S. Bassler, of the United States 
 National Museum, Washington, for his opinion on the invertebrate 
 fossils collected in Brazil and Argentina, whilst, finally, I am exces¬ 
 sively indebted to my friend Dr. F. R. Cowper Reed, of the Sedgwick 
 Museum, Cambridge, England, for his labors on these collections, 
 the results of which are embodied in the valuable Appendix to this 
 report. His noteworthy discovery of the Triassic affinities of the 
 Estrada Nova mollusca was, however, only made while this work was 
 in the press, and it has been found impossible at this late stage to do 
 full justice thereto and avoid some minor inconsistencies in the text. 
 
CHAPTER II 
 
 GEOLOGICAL FRAMEWORK OF THE SOUTH ATLANTIC 
 SECTION A. THE WESTERN SIDE OF AFRICA 
 
 For convenience we may divide this region into two portions: 
 
 (i) that from the extreme south of the Cape up to the Cunene 
 River, of which the geology is now firmly established, while forma¬ 
 tions younger than the Triassic are scarcely represented, and (2) 
 that from the Cunene River northward, for which reliable data are 
 often ’wanting, while the coastal strip, just as in the opposed section 
 of Brazil, is made up largely of marine Cretaceous and Tertiary 
 deposits. 
 
 (1) The southern section , 9 embracing part of the Cape Province 
 and Territory of Southwest Africa (formerly German Southwest 
 Africa), displays a great stretch of Archaean rocks from the Cunene 
 River southward to Van Rhynsdorp, having a northeasterly strike 
 in the Swakopmund-Windhoek region. Younger, though probably 
 of Cambrian or pre-Cambrian age, is the Nama system, with a basal 
 series of conglomerates and quartzites reposing unconformably on the 
 crystallines, followed by limestones, dolomites, and shales, and these 
 by red sandstones, red and green shales, and gray and red quartz¬ 
 ites. Though nearly flat-lying inland, e.g., at Kuibis, this formation, 
 on approaching the coast, suddenly displays folding along axes 
 following the ocean margin closely, from Liideritz to the mouth of the 
 Olifants River continued with a south-southeasterly to southeasterly 
 strike through the Western Province of the Cape. These crumplings , 
 which were accompanied by intrusions of granite, are of pre- 
 Devonian age, but in the south a renewal of movement took place 
 along them in late Carboniferous times. 
 
 Upon a worn-down surface of these rocks was deposited the 
 triply-divided Cape system some 3,300 meters in thickness, widely 
 distributed over the southern third of the Cape Province. The 
 Lower Devonian Table Mountain sandstone at the base, some 1,600 
 meters thick, is composed of unfossiliferous, whitish, hard sandstones 
 and grits characterized by tiny pebbles of white or gray vein-quartz 
 and sometimes of red jasper, while current-bedding is conspicuous. 
 An important feature is the occurrence of a fine glacial horizon some 
 700 meters above the base, the material having been derived from 
 some locality lying out in the present Atlantic. 
 
 Then follows the Bokkeveld series of shales and sandstones, with 
 its Lower-Middle Devonian “Austral” invertebrate fauna and above 
 
 • For fuller details see Du Toit (1926). 
 
 6 
 
GEOLOGICAL FRAMEWORK OF THE SOUTH ATLANTIC 
 
 7 
 
 it the Witteberg series (upper Devonian-lower Carboniferous) of 
 white quartzites, with some poorly preserved plant remains. The 
 source of all these Palaeozoic sediments lay in the north and north¬ 
 west. In the neighborhood of the Olifants River mouth these beds 
 are lying nearly flat, but on proceeding southward folding is dis¬ 
 covered that follows the much older south-southeasterly directed 
 crumplings, while near Ceres a second series of foldings appears with 
 a general west-east trend continuing to the Indian Ocean; the moun¬ 
 tainous belt thus formed is known as the “Cape Fold ranges. ” 
 
 South of latitude 33° the Witteberg passes up without a break 
 into the upper Carboniferous glacials of the Dwyka series, which 
 constitutes the lowest member of the Karroo system, and which, par¬ 
 taking in the above-mentioned foldings, possesses a thickness of 
 over 400 meters and seems to have been largely deposited in standing 
 water by ice-sheets originating in the north. North of the thirty- 
 third parallel the tillite is found undisturbed, but transgresses across 
 the Cape system and older groups to rest upon the Archaean, while 
 the floor beneath it becomes uneven and in good exposures reveals 
 typically striated roches moutonnZes. The horizontal tillite is a true 
 ground-moraine and can be traced through Calvinia and Bushman- 
 land to appear again beyond the Orange River, whence it con¬ 
 tinues northward in Southwest Africa up the Fish River Valley to 
 Mariental, but has not yet been recognized north of that point. 
 
 The succeeding beds, sometimes sharply defined, but often 
 carrying small erratics, compose the upper Dwyka Shales and pass 
 up into a zone of white-weathering carbonaceous and pyritic shales 
 with occasional casts of the small reptile Mesosaurus , and, together 
 •with some dolomite and chert, constitute the horizon of the “White 
 band, ” the thickness of these post-glacial beds being from 150 to 160 
 meters, approximately. These strata are recognizable from Robert¬ 
 son in the south to Mariental in the north, but fragments of Meso¬ 
 saurus have been obtained from the Kaokoveld in latitude 20° 45'. 
 All this is instructive when comparison is being made with the almost 
 identical Iraty shales of Brazil with this similar zone fossil. No coals 
 occur in the Dwyka shales, though the Brazilian seams appear to be 
 on the corresponding horizon. 
 
 The White band passes up in the southern Karroo into the Ecca 
 series, a thick group of greenish, bluish, and sometimes purplish 
 shales and flagstones with calcareous nodules and dark-green or gray 
 sandstones that have yielded silicified wood and the Gangamopteris 
 flora, but to the northeast and to the north entirely different facies 
 in the series have been developed. For instance, in Calvinia the 
 group is composed almost entirely of dark-bluish, soft shales, while 
 in the southern Kalahari and to the northeast of Keetmanshoop the 
 strata are now largely red, the formation consisting of deep-red and 
 
8 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 green-blue shales with brown-weathering calcareous bands and soft 
 yellowish and reddish sandstones, a facies closely comparable with 
 the second and third stages of the Paganzo system of western Argen¬ 
 tina and also somewhat like the Estrada Nova group in Uruguay. In 
 the Transvaal, on the other hand, the overlap of the several sub¬ 
 divisions of the Dwyka and Ecca has taken place on to the thin 
 glacial conglomerate or else on to the pre-Karroo rocks, while a 
 fluviatile and thinner phase of the Ecca has led to the production of 
 the Transvaal and Natal “Coal Measures,” a group of pale grits, 
 conglomerates, arkoses, and sandstones, with dark micaceous shales 
 and seams of coal often from 2 to 8 meters in thickness. The flora is 
 the ‘ ‘ southern ’ ’ one, allied to that of the Karharbari of India or the 
 Greta stage of New South Wales, but with some “northern” ele¬ 
 ments, a Lower Permian age being indicated. 
 
 The succeeding Beaufort series, well over 1,000 meters thick, in 
 the southwest, is made of greenish mudstones and shales, often varie¬ 
 gated and carrying calcareous concretions, together with yellowish 
 felspathic sandstones, the beds being conspicuously red and maroon 
 near the top. Small local depositional unconformities are common 
 and the beds are decidedly of the “interior basin” type. A fairly 
 abundant reptilian and amphibian fauna and a scanty flora allied to 
 that of the Damuda of India and Newcastle of New South Wales, 
 collectively indicate an age covering the upper Permian and lower 
 Triassic. This formation is unrepresented to the northeast in the 
 central Transvaal and to the northwest in Southwest Africa. 
 
 The Stormberg series embraces the plant- and coal-bearing 
 Molteno group of grits, sandstones, and gray and blue shales, the 
 striking deeply-colored group of the Red beds and massive contrasting 
 cream or pink Cave sandstone and finally the Drakensberg volcanics, 
 the whole assemblage representing the upper Triassic, Rhaetic, and 
 just possibly the Liassic. While conformable to the Beaufort in the 
 south, this series transgresses across the Ecca or even the pre-Karroo 
 rocks in the Transvaal, Southwest Africa, and Southern Rhodesia, so 
 that the Red beds, Cave sandstone (and its equivalents, the Bushveld 
 and Forest sandstone) and basalts come to cover wide tracts without 
 the intervention of the Molteno or older groups, a feature paralleling 
 the conditions to be found in South America in the case of the Trias- 
 Rhaetic beds. The series is well developed in the Kaokoveld, where 
 it makes the coast-line for quite a considerable distance. 
 
 The Molteno flora with its characteristic genus Thimifeldia, is 
 extremely like that of the “Rhaetic” of western Argentina, though, 
 the vertebrates (saurischia and pseudosuchia ) of the Red beds and 
 Cave sandstone confer an age not younger than Rhaetic for these 
 succeeding sediments. The Basaltic lavas cover broad areas and 
 must have had a much wider extension formerly; they exceed 400 
 
GEOLOGICAL FRAMEWORK OF THE SOUTH ATLANTIC 
 
 9 
 
 meters in depth in each of the main areas, the maximum thickness 
 known being about 1,300 meters. Dolerite sills and dikes, closely 
 related to the lavas, are most abundantly represented in the Karroo 
 beds from Sutherland and Beaufort West in the south right up into 
 Southwest Africa, a feature closely paralleled in the Parana Basin 
 in Brazil. 
 
 Cretaceous (Neocomian-Wealden) beds and Tertiaries (Eocene- 
 Pliocene) are represented within the Coastal Belt of the south of the 
 Cape, but no undoubted marine Cretaceous is known along the 
 Atlantic shore south of the Cunene River nor directly opposite in 
 South America. 
 
 (2) In the less known northern area Cretaceous and Tertiary 
 beds, largely marine in origin, fringe the coast from Lobito Bay in 
 Angola up to French Congo, widening out where traversed by the 
 Cuanza and Congo Rivers, while in the former region the strata are 
 actually flexed along axes concave towards the west and are also 
 faulted. Dipping beneath them in places are the Dombe sandstones 
 and in the Congo section the “Gres sublittoraux” of Fourmarier, 10 
 which have been regarded by Choffat as Lower Cretaceous, but by 
 others as of Triassic age. The Albian of Benguella is of importance 
 because of the presence of certain forms common or allied to the 
 Sergipe Cretaceous of Brazil, as pointed out by Kossmatt, Newton, 
 Gregory, and others. 
 
 The plateau to the east of Benguella, formed of crystallines and 
 belts of ancient sediments, is crowned with the Bihe sandstone, 11 
 which is probably merely the fringe of the Lubilache series that 
 covers such an enormous area within the Congo Basin, with its 
 western boundary passing through Leopoldville northward to the 
 borders of the Cameroons, a distance of over 1,500 km. This forma¬ 
 tion, several hundred meters thick, is composed'of horizontal friable 
 white, yellowish, or reddish sandstones, often displaying curious sur¬ 
 face silicification. On the eastern side of this immense basin they 
 repose upon, but quite possibly to some extent pass laterally into, 
 the shales and sandstones of the Lualaba series. The age of these 
 groups, as determined from some fish and crustacean remains, is 
 Triassic or Rhaetic and they correspond therefore with the Stormberg 
 series farther to the south. In the west the Lubilache generally rests 
 with unconformity upon the slightly folded strata ascribed to the 
 Kundulungu series, a group of deep-red shales and micaceous flag¬ 
 stones alternating with gray or light-red sandstones and becoming 
 more felspathic and coarser toward the summit. There are serious 
 differences of opinion as to the age of this formation as seen in the 
 Lower Congo or in French Congo, which can not be settled, owing 
 
 10 Fourmarier (1924). 11 Gregory (1916), p. 523. 
 
10 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 to the absence of fossils, but a Permian or Carboniferous age is just 
 possible, and if so, the group would correspond to the lithologically 
 similar Estancia beds of Bahia and Sergipe. 
 
 The Kundulungu in turn rests disconformably upon the system 
 “ Schisto-calcaire, ” a group unconformable to the pre-Cambrian 
 complex, having a basal conglomerate stated by Delhaye and Sluys 12 
 to be largely of glacial origin, followed by over 1,000 meters of lime¬ 
 stones and dolomites, occasionally oolitic, but as yet not proved 
 fossiliferous. They either correspond with the Otavi dolomite 
 (Nama) of Southwest Africa—probably Proterozoic—or are of early 
 Palaeozoic age and hence possibly comparable with the Bambuhy 
 series of eastern Brazil. These beds display strong foldings directed 
 nearly north-northwest, crossed by a set at right angles thereto, 
 along the latter of which a renewal of movement occurred probably 
 in the Permian as well as in still later times. 
 
 From Cameroons the coast of Africa turns abruptly westward 
 and a region 13 is reached displaying many points of similarity with 
 the Pernambuco-Maranhao territory of Brazil. Along the low-lying 
 coast, extending from Cameroons to Accra, with a tongue reaching 
 into northern Nigeria, is a curving belt of marine Tertiaries, in which 
 zones so far back as the Eocene are to be found, resting upon either 
 Cretaceous or ancient formations and passing beneath younger 
 deposits. Bituminous shales and lignites are known from the earlier 
 tertiaries in several areas and can be compared with the Eocene 
 bituminous shales with fish remains in Alagoas and perhaps Maran- 
 hao. Much of the western part of Africa was inundated by the 
 Cretaceous ocean, and wide tracts of the hinterland are covered by 
 marine beds belonging to various stages of that epoch, just as is so 
 marked a feature of Ceara, Piauhy, and Maranhao. On both sides of 
 the Atlantic these strata have locally been affected by post-Creta¬ 
 ceous movements. 
 
 Among the older formations might be mentioned the Buem series 
 on the eastern side of the Volta River in Togoland, composed of con¬ 
 glomerates, arkoses, quartzites, and shales, of which certain con¬ 
 glomerates have been regarded as of morainic origin by Koert and 
 hence of Permo-Carboniferous age; Lemoine classes them, however, 
 in the Devonian. Conspicuous is the tectonic line of the Atacora 
 Range south of the Niger in Dahomey—the “Saharides” of Suess— 
 in which a belt of folded quartzites and schists, flanked by crystal¬ 
 lines, ascribed to the Silurian, but probably older, strikes north- 
 northeastward from the Gold Coast and thence at intervals into the 
 Sahara, where the folds now trend more nearly northward; associated 
 is the Tarkwa series in Gold Coast Colony with northeastward strike 
 and certainly pre-Palasozoic. 
 
 12 Delhaye and Sluys (1921). 
 
 13 Lemoine (1913). 
 
GEOLOGICAL FRAMEWORK OF THE SOUTH ATLANTIC 
 
 11 
 
 This brings us to the consideration of the undoubted Palasozoic 
 rocks of western Africa and central Sahara. On the borders of Sierra 
 Leone and French Guinea, resting unconformably on schistose rocks, 
 is an undulating sandstone formation carrying Silurian (Wenlock) 
 graptolites , 14 apparently the beginning of the huge area of horizontal 
 sandstones on the upper Niger building up the plateau of Homboi 
 (“Bandiagara sandstone”) and south of the great bend of the Niger, 
 but which have not as yet been proved fossiliferous. To the north¬ 
 east the Silurian of Tuareg and Air is folded and overlain by gently 
 dipping Lower Devonian possessing a fauna including in addition to 
 European forms, not only some from North America, but certain 
 from the Bokkeveld as determined by Haug . 15 Future work will 
 doubtless add largely to the known area of Palaeozoics in this part of 
 Africa. Of great importance too is the occurrence of the Middle 
 Devonian in a small down-faulted block at Accra on the Gold Coast, 
 with a “Hamilton” fauna , 16 while slates, provisionally placed in the 
 Lower Carboniferous because of certain poorly preserved mollusca— 
 but possibly Devonian—are known at Sekondi to the west of Accra. 
 
 Undoubted pre-Mesozoic marine beds have not yet been found to 
 the southeast in Cameroons, and the presence of this wide, though 
 interrupted belt extending from Sierra Leone into the Sahara, in 
 which marine Palaeozoic beds have been developed, is of immense 
 importance, inasmuch as it shows correspondence with the broad, 
 shallow syncline of marine Silurian, Devonian, and Carboniferous 
 strata within the lower Amazon Valley in which similar faunal ele¬ 
 ments are actually present. 
 
 SECTION B. THE FALKLAND ISLANDS 
 
 These remarkable islands, of which a brief geological sketch can 
 not be omitted, constitute a most striking link between South Africa 
 and South America, despite the fact that they are situated only a few 
 degrees from Patagonia and well down below the fiftieth parallel. 
 Our knowledge concerning them has been enormously extended 
 through the work of Andersson 17 and after him Halle , 18 with H. A. 
 Baker 19 just recently, while the Devonian molluscan fauna has been 
 discussed by J. M. Clarke 20 and the Gondwana flora described by 
 Halle, Seward and Walton . 21 
 
 The dominant Devono-Triassic succession has been affected by 
 two main tectonic structures. The more important of these, marked 
 by strong folding and some faulting, traverses the northern part of 
 East Island, trending west to west-northwest and curving more to the 
 
 14 Dixey (1925), p. 213; Lecointre and Lemoine (1925). 
 
 16 Haug (1905). 
 
 16 Kitson and Morley Davies (1925). 19 Baker (1923). 
 
 17 Andersson (1907). 20 Clarke (1913)* P- 55 » P* 3 2 ^. 
 
 18 Halle (1912). 21 Seward and Walton (1923). 
 
12 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 northwest through the northern side of West Island, with some 
 parallel subsidiary flexuring in the center of that land; the other 
 structure crosses nearly at right angles, striking south west ward along 
 both shores of Falkland Sound. In the southern side of East Island 
 and the southwestern side of West Island the beds, on the contrary, 
 undulate slightly or are flat. 
 
 The lower Devonian sandstones are restricted to the southwestern 
 side of West Island, resting unconformably upon crystallines at 
 Cape Meredith and having a thickness of many hundred meters. 
 The series is extraordinarily like the Table Mountain sandstones of 
 the Cape Province, consisting of a group of unfossiliferous, pale, 
 false-bedded sandstones and quartzitic sandstones with some red 
 micaceous shales at the very base, while small pebbles of white vein- 
 quartz are common, making in places conglomerates. It passes 
 beneath a group of slaty rocks, shales, flagstones, and sandstones that 
 has yielded a typical “austral” Devonian fauna, overlain by similar 
 beds with lepidodendroid plant remains and thus like the Bokkeveld 
 series of the Cape. Of the mollusca, some 20 species are common 
 to the latter and some 2 5 to the equivalent Devonian of South Amer¬ 
 ica, including in each case several trilobites, besides other closely 
 allied forms. This group is succeeded by quartzites and quartzitic 
 sandstones and interlaminated shales that correspond with the 
 Witteberg series. Baker has pointed out these surprising resem¬ 
 blances with the Cape system, not only lithologically, but in regard 
 to thickness as well, so far as such can be made out in the folded 
 ranges of East Island. 
 
 The Gondwana beds are referred to as the Lafonian system, 
 having a thickness of possibly over 3,000 meters and occupy the 
 southern half of East Island, but are brought down by folding in a 
 few localities in West Island. At the base are glacial boulder-beds 
 about 650 meters thick, identical in their characters with those of the 
 , Dwyka tillite, usually folded and cleaved, with striated erratics often 
 traversed by regular jointing; underneath them Baker has discovered 
 some quartzite surfaces with striations running from north to south, 
 presumably glaciated floors. Some passage beds with small inclu¬ 
 sions lead up into the Lafonian sandstone, and this in turn into 
 “banded siltstones, ” apparently seasonally banded “varve” rocks 
 representing glacial material deposited in fresh water. 
 
 Next comes a huge group of alternating brown, yellow, gray, and 
 green mudstones, shales and similarly colored sandstones and hard 
 greenish sandstones in which occur members of the Glossopteris 
 flora. The form identified by Halle as Gangamopteris is regarded by 
 Seward and Walton 22 as more probably a Glossopteris , with which 
 occur G. browniana, G. indica, Voltzia , and several species of 
 
 22 Seward and Walton (1923), p. 324. 
 
GEOLOGICAL FRAMEWORK OF THE SOUTH ATLANTIC 
 
 13 
 
 Phyllotheca and Dadoxylon, evidently an Ecca-Beaufort flora. The 
 highest beds, exposed on the eastern side of Falkland Sound, have 
 yielded Neocalamites carrerei (Zeill.), suggestive of a Triassic age. 
 
 Dolerites, just like those injected into the Karroo system at the 
 close of the Rhaetic, pierce the strata at a number of places, signifi¬ 
 cantly avoiding the northern folded belt, but frequently possessing a 
 northeast-southwest trend. Not improbably the westerly to north¬ 
 westerly directed foldings correspond in age with those striking 
 nearly east-west in the Cape, which we know were initiated at the 
 very close of the Palaeozoic, but were renewed in mid-Cretaceous 
 times, and from which dolerite is absent. The Western Province 
 of the Cape displays in addition another set of foldings trending 
 southwest or west-southwest. 
 
 The precise place of the Falkland Islands, stratigraphically, 
 lithologically, and structurally, with reference to the Cape and to 
 Argentina, will be discussed at a later stage (Chapter VI), when it 
 will be shown that they display characters consistently intermediate 
 between those of the two countries, though somewhat more closely 
 allied with the former. 
 
 SECTION C. SOUTH AMERICA 23 
 
 Commencing in the north, the outstanding feature is the great 
 syncline of Lower Pakeozoics, much covered by Tertiary and Quater¬ 
 nary deposits, in the Amazon Valley, trending eastward from Manaos 
 to Para, in which direction its axis deviates towards east-northeast. 
 In this trough marine strata of Silurian, Devonian, and Carbonif¬ 
 erous ages are all involved, resting upon the Archaean, which bounds 
 the structure for immense widths, both to north and south. 
 
 Eastward in Maranhao, Piauhy, Ceara, Parnahyba, and Pernam¬ 
 buco there is a wide fringe of Tertiaries, and no Palaeozoics are known 
 to occur, if we ignore the rather doubtful series of Serra Grande, while 
 inland in the first two States Permian and Triassic Gondwana sedi¬ 
 ments spread over a wide area, reposing mainly on crystallines. In 
 Alagoas, Sergipe, and the adjoining part of Bahia undoubted Permian 
 and possibly upper Carboniferous form belts rest on the (probably) 
 pre-Devonian basement and display a certain amount of folding. 
 Save for them and the similar patch in the southern end of Bahia, 
 Gondwana beds do not reach the coast of Brazil, except in the neigh¬ 
 borhood of Torres. 
 
 Inland from Bahia is an extensive trough of folded lower Palaeo¬ 
 zoic beds—mainly slates and limestones, probably Ordovician—as 
 well as much older formations, striking southwestward from near the 
 great bend in the Rio Sao Francisco, following up the valley of that 
 river almost to the boundary of Minas, which feature is continued 
 
 23 SeeBranner (1915) (1919). 
 
14 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 beyond with similar trend through Sao Paulo and Parana in the more 
 closely folded and altered condition to reappear in Uruguay. This 
 lengthy tectonic structure, running approximately parallel to the 
 Atlantic coast, has been termed the “Brasilides” by Keidel. 
 
 Occupying the whole of the Upper Parana drainage area is the 
 great but shallow “Parana Basin,” made of Gondwana beds, called 
 the Santa Catherina system, deposited upon a platform of crystal¬ 
 lines and folded Ordovician, but in parts of Parana and Matto 
 Grosso upon nearly horizontal lower Devonian strata like those of 
 the Cape; its boundaries have been determined by erosion on all 
 sides except to the west of the Paraguay and Lower Parana Rivers, 
 where the beds disappear beneath the geologically recent Neogene 
 formations. A large part of east-central Brazil is, furthermore, 
 covered by horizontal Cretaceous and perhaps Tertiary sandstones 
 of continental type, in places concealing the Santa Catherina beds. 
 
 Passing over the wide stretch in Eastern Argentina, beneath 
 which these Gondwana strata have been proved by deep borings, 
 those beds reappear in the west at intervals within a huge region from 
 Bolivia southward to Cordoba and Mendoza, where they, together 
 with Palaeozoics beneath and Cretaceo-Tertiaries above, have been 
 involved in the Andine movements, with general strike either north 
 or a little east of north; here they are known as the Paganzo system. 
 Farther to the south similar strata, from Devonian to Permian in 
 age, build the ranges between Buenos Aires and Bahia Blanca with 
 southeasterly trend, structures termed the “ Gondwanides ” by 
 Keidel 24 and regarded by him as the continuation of the Southern 
 Fold ranges of the Cape, involving as they do precisely similar 
 and equivalent formations. 
 
 The equivalent systems of Santa Catherina and Paganzo gener¬ 
 ally begin with a glacial formation, which corresponds with the 
 Dwyka tillite of South Africa and is followed by plant- and often 
 coal-bearing lower Permian beds with the “ Glossopteris flora” and by 
 higher Triassic and Rhaetic rocks, but an important intraformational 
 break exists, extending perhaps over most of the region. At the top 
 come the enormously widespread basaltic effusions of Parana, 
 though not restricted to that region, being also represented in 
 Maranhao and in certain of the remnants in the pre-Cordillera as 
 well as in Patagonia, where, however, the bulk of the volcanics 
 appear to be of acid composition, largely overlain by nearly flat 
 lying Cretaceous and Tertiary beds. 
 
 The western limits of Gondwanaland are not exactly known, as 
 the Andine belt is largely composed of Mesozoic marine beds and 
 volcanics, all intensely folded and injected with igneous matter. 
 
 24 Keidel (1916 and 1922). 
 
GEOLOGICAL FRAMEWORK OF THE SOUTH ATLANTIC 15 
 
 SECTION D. GENERALIZED COMPARISONS BETWEEN 
 THE TWO CONTINENTS 
 
 Confining attention in each case to a strip some 45 degrees in 
 length by 10 in breadth, we shall now proceed to compare the two 
 stretches, namely, the tract extending from Sierra Leone to Cape 
 Town on the one side with that from Para to Bahia Blanca on the 
 other; furthermore, whenever the direction of any structural feature 
 has to be mentioned, such will be given with reference not to the 
 meridian, but to the coast-line near at hand. 
 
 In each continent: 
 
 (1) The foundation rocks consist of crystallines of pre-Cambrian 
 age and certain belts of infolded pre-Devonian sediments of various 
 though mostly undetermined ages, but generally corresponding 
 lithological characters, 
 
 (2) In the extreme north, marine Silurian and Devonian beds, 
 only slightly disturbed, rest unconformably upon this complex, 
 occupying a broad syncline trending obliquely to the coast-line, 
 namely, between Sierra Leone and Gold Coast and underlying the 
 estuary of the Amazon. 
 
 (3) Farther to the south, belts of Proterozoic and early Palaeo¬ 
 zoic strata, mainly quartzites, slates, and limestones, strike nearly 
 parallel to the coast, being gently flexed in the north, but becoming 
 more disturbed toward the south, where they are invaded by granitic 
 masses, for example, in the region between Liideritz and Cape Town 
 and that between the Rio Sao Francisco and Rio La Plata. 
 
 (4) Corresponding to the nearly flat-lying Devonian of Clan- 
 william is its all but identical counterpart in Parana and Matto 
 Grosso. 
 
 (5) More to the south we find the Devono-Carboniferous of the 
 southern Cape paralleled by the terrain appearing a little to the north 
 of Bahia Blanca, passing up conformably into the Carboniferous 
 glacials and Permian sediments, both successions having been 
 intensely crumpled under Permo-Triassic and Cretaceous move¬ 
 ments that display similar orientations. 
 
 (6) Traced northward, the tillites in each case become horizontal 
 and transgress across the Devonian to rest upon a glaciated pene¬ 
 plain formed by these and by older rocks; farther to the north they 
 fail. 
 
 (7) The glacials are in each case overlain by continental Permian 
 and Triassic strata with the “ Glossopteris flora, ” covering enormous 
 areas, followed by vast outpourings of basalts and penetrated on an 
 extensive scale by dolerites of presumed Liassic age. 
 
 (8) These Gondwana beds extend northward from the Southern 
 Karroo to the Kaokoveld and from Uruguay to Minas Geraes. 
 
 (9) Further great detached areas occur in the north, in each case 
 
16 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 some distance inland, in the Angola-Congo and the Piauhy-Maran- 
 hao regions. 
 
 (io) An intraformational break is widespread, though commonly 
 there is no angular unconformity between late Triassic and early 
 Permian beds. In certain areas, however, the former may rest with 
 visible discordance on tilted Permian or pre-Permian formations. 
 
 (n) Tilted Cretaceous beds occur on the coast only in the Ben- 
 guella-Lower Congo and Bahla-Sergipe areas. 
 
 (12) Horizontal Cretaceo-Tertiary, both marine and continental, 
 cover great extents between Cameroons and Togoland and in Ceara, 
 Maranhao, and also more to the south, while the extensive deposits 
 of the Kalahari can roughly be paralleled with the Neogene and the 
 Quaternary Pampean of Argentina. 
 
 (13) In setting down this generalized summary, the important 
 link constituted by the Falkland Islands must not be overlooked. 
 Their folded Devono-Carboniferous succession is all but indistin¬ 
 guishable from that of the Cape, while the Lafonian closely parallels 
 the Karroo system. Strati graphically and structurally the Falklands 
 have their place with the southwest of the Cape and not with Patagonia 
 (Chapter VI). 
 
 (14) From the palaeontological viewpoint, attention should be 
 focussed on: ( a ) the “austral facies” of the Devonian of the Cape, 
 Falklands, Argentina, Bolivia, and Southern Brazil, in contrast to 
 the “boreal facies” of northern Brazil and central Sahara; ( h ) that 
 unique reptilian genus Mesosaurus from the Dwyka shales of the 
 Cape and Iraty shales of Brazil, Uruguay, and Paraguay; ( c ) the 
 Gangamopteris-Glossopteris flora, with a small admixture of northern 
 forms within the Lower Gondwana beds in the south of each country; 
 (d) the Thinnfeldia flora of the Upper Gondwana of the Cape and 
 Argentina; ( e) the Neocomian (Uitenhage) fauna in the south of the 
 Cape and northwest of Neuquen in Argentina; (/) the Northern or 
 Mediterranean facies of the Cretaceous and Tertiary faunas north 
 of the Tropic of Capricorn; and (g) the South Atlantic-Antarctic 
 facies of the Eocene (San Jorge formation) of Patagonia. 
 
 (15) The geographical outlines of Africa and South America are 
 amazingly similar, not only in the main, but even as to detail; more¬ 
 over, excepting in the north, the fringe of Tertiaries is of small width 
 and the presence of those beds of little moment therefore. 
 
 On following the various groups embraced between the Devonian 
 and the Rhaetic of the two continents in a direction away from the 
 Atlantic shores an amazing generalization shows itself, namely, that 
 the variation in facies of these several formations commonly dis¬ 
 closes greater divergences within distances of say from 5 to 15 degrees 
 of arc than those actually noticed on comparing the developments 
 
Ti 
 
 European equivalents 
 
 Cape Province 
 
 Falkland Islands 
 
 Sierra de la Ventana 
 
 Liassic. 
 
 
 r 
 
 be 
 
 (h 
 
 <u 
 
 ’ Basalts. 
 
 
 
 
 Rhaetic. 
 
 
 Cave sandstone. 
 
 
 
 
 
 
 o 
 
 Red beds. 
 
 
 
 
 Triassic. 
 
 
 -4— 1 
 
 CO 
 
 k Molteno beds. 
 
 
 
 
 s 
 
 
 Beaufort. 
 
 • 
 
 
 
 
 0) 
 
 4-J 
 
 
 
 
 
 
 GO 
 
 >> 
 
 o 
 
 o 
 
 
 
 
 ’ Sandstones and 
 shales. 
 
 
 Permian. 
 
 fc 
 
 d 
 
 M 
 
 
 Ecca. 
 
 a 
 
 <D 
 
 C/3 
 
 >> 
 
 C/3 
 
 a * 
 a 
 
 • H 
 
 Shales, clay- 
 stones, etc. 
 
 Pillahuinc6 beds. 
 
 
 
 
 
 
 
 Upper Carboniferous. 
 
 
 Dwyka 
 
 ' White band. 
 
 Upper shales. 
 
 Glacials. 
 
 O 
 
 >1 
 
 Shales, sand¬ 
 stones, etc. 
 
 . Glacials. 
 
 Shales. 
 
 Glacials. 
 
 
 
 
 Lower shales. 
 
 
 
 
 Middle and lower 
 
 
 l 
 
 
 
 
 Carboniferous and 
 middle Devonian. 
 
 Lower Devonian and 
 uppermost Silurian. 
 
 Cape system 
 
 _/V________ 
 
 Witteberg quartzites 
 and shales. 
 
 Bokkeveld beds (fossil - 
 iferous) . 
 
 Table Mountain sand¬ 
 stones. 
 
 Quartzites, sand¬ 
 stones, and shales. 
 
 Fossiliferous shales 
 and sandstones. 
 
 Barren sandstones. 
 
 Greywackds. 
 
 Fossiliferous slates. 
 
 Sierra de la Ventana 
 quartzites. 
 
 
 
 
 
 
 
 
 * 
 
 Paganzo system 
 
European equivalents 
 
 Cape Province 
 
 Falkland Islands 
 
 Sierra de la Ventana 
 
 
 Liassic. 
 
 
 be 
 
 1-4 
 
 03 
 
 vO 
 
 8 
 
 Basalts. 
 
 
 
 
 
 Rhaetic. 
 
 
 Cave sandstone. 
 
 
 
 
 
 Triassic. 
 
 
 o 
 
 +-> 
 
 CO 
 
 Red beds. 
 
 . Molteno beds. 
 
 
 
 
 * 
 
 
 e 
 
 
 Beaufort. 
 
 
 
 
 
 
 03 
 
 4-» 
 
 03 
 
 >> 
 
 03 
 
 o 
 
 o 
 
 
 
 Sandstones and 
 shales. 
 
 
 a 
 
 03 
 
 03 
 
 03 
 
 Permian. 
 
 b 
 
 cj 
 
 M 
 
 
 Ecca. 
 
 a 
 
 03 
 
 -*-> 
 
 03 
 
 03 
 
 a ■ 
 
 Shales, clay- 
 stones, etc. 
 
 Pillahuinc6 beds. 
 
 0 
 
 N 
 
 § 
 
 be 
 
 Cj 
 
 P-t 
 
 Upper Carboniferous. 
 
 
 ca 
 
 44 
 
 White band. 
 
 Upper shales. 
 
 a 
 
 '£ 
 
 i 
 
 41 
 
 Shales, sand¬ 
 stones, etc. 
 
 Shales. 
 
 
 
 
 & 
 
 Q 
 
 Glacials. 
 
 
 . Glacials. 
 
 Glacials. 
 
 
 
 
 
 Lower shales. 
 
 
 
 
 
 Middle and lower 
 
 
 
 
 
 
 
 
 Carboniferous and 
 middle Devonian. 
 
 Lower Devonian and 
 uppermost Silurian. 
 
 6 
 
 03 
 
 4-> 
 
 03 
 
 03 
 
 <V 
 
 Oh 
 
 a 
 
 O 
 
 Witteberg quartzites 
 and shales. 
 
 Bokkeveld beds (fossil- 
 iferous). 
 
 Table Mountain sand¬ 
 stones. 
 
 Quartzites, sand¬ 
 stones, and shales. 
 
 Fossiliferous shales 
 and sandstones. 
 
 Barren sandstones. 
 
 Greywackds. 
 
 Fossiliferous slates. 
 
 Sierra de la Ventana 
 quartzites. 
 
 
 
 
 
 
 
 
 
 Table I .—Strati graphical Table Showing the Devono-Lias Succession in South America and South Africa 
 
 (The wavy line indicates an unconformity.) 
 
 San Juan—La Rioja 
 
 V. Red conglomerate, 
 etc. 
 
 IV. “Rhaetic.” 
 
 III. Stage. 
 
 11. Stage. 
 
 Shales, etc. 
 Glacials. 
 
 Cardiopteris beds. 
 Glacials. 
 
 Barreal 
 
 V. Red conglomerate. 
 IV. “Rhaetic.” 
 
 03 
 
 a; 
 
 ’G 
 
 <D 
 
 03 
 
 3 
 
 g Glacials. 
 
 o 
 
 Marine beds with Produc- 
 tus and Spirifer. 
 
 Slates, greywack6s, and sandstones. 
 
 o 
 
 • ’—i 
 
 <D 
 
 Q) 
 
 m 
 
 Northern Argentina and 
 Southern Bolivia 
 
 Marls. 
 
 Limestone and Dolomite. 
 
 Upper sandstones (Macha- 
 reti). 
 
 Mandiyuti. 
 
 Oquita. 
 
 Materal. 
 
 Productus beds. 
 
 Sica-sica beds. 
 Huamampampa sandstones. 
 Conularia and Crinoid beds. 
 Ida sandstones. 
 
 03 
 
 Cj 
 
 .£ 
 
 <v 
 
 rX 
 
 4- 5 
 
 a 
 
 O 
 
 Parand Basin 
 
 Serra Geral volcanics. 
 Botucatti sandstone. 
 
 Rio do Rasto. 
 
 Estrada Nova. 
 
 (upper part) 
 
 Estrada Nova. 
 
 (lower part) 
 
 Irat£. 
 
 f Palermo. 
 <u i 
 
 11: 
 
 'g r 
 
 jg j Glacials. 
 
 Bonito. 
 
 Ponta Grossa shales. 
 Furnas sandstone. 
 
 a 
 
 P-, 
 
 Maranhao 
 
 .|3 f Basalts. 
 
 *2 I Red sandstones. 
 
 (?) 
 
 Southwest Africa 
 
 North 
 
 bo 
 
 g (Basalts. 
 
 43 
 
 o 
 
 •M 
 
 CO 
 
 Kaoko beds. 
 
 { Shales and sand¬ 
 stones with 
 Psaronius. 
 
 South 
 
 bo 
 
 <D 
 
 43 
 
 O 
 
 W 
 
 Basalts. 
 
 Ecca. 
 
 0) 
 
 44 
 
 £ 
 
 Q 
 
 White band. 
 Upper shales. 
 Glacials. 
 
 Transvaal 
 
 bo 
 
 1-4 
 
 <v 
 
 rO 
 
 CO 
 
 S 
 
 03 
 
 03 
 
 03 
 
 o 
 
 o 
 
 i 
 
 ni 
 
 o 
 
 y 
 
 W 
 
 Basalts. 
 
 Bushveld sandstone 
 Bushveld marls. 
 
 Coal measures. 
 
 stf r 
 44 
 
 b 1 Glacials. 
 
 q! 
 
GEOLOGICAL FRAMEWORK OF THE SOUTH ATLANTIC 
 
 17 
 
 nearest the two opposed shores, although the ocean between is no 
 less than from 40 to 70 degrees in breadth. This observed arrange¬ 
 ment is considered to constitute the most telling argument in favor 
 of the conception that the two continental masses were in the past 
 geographically closer to each other , which is actually the keynote of 
 the Displacement or Disruption Hypothesis. Evidence for this 
 momentous assertion will be found in the following pages and the 
 problem reviewed in Chapter VII. 
 
 For convenience, the accompanying correlation table has been 
 introduced at this stage to indicate the stratigraphical succession in 
 the type areas to be dealt with in detail below. 
 
CHAPTER III 
 
 ARGENTINA AND BOLIVIA 
 
 SECTION E. THE SIERRAS OF THE DISTRICT OF BUENOS AIRES 
 
 Rising out of the plain of Pampean deposits (late Tertiary to 
 Quaternary), here close upon 200 meters above sea-level, are the two 
 well-known chains of ancient rocks lying to the south and southwest 
 respectively of Buenos Aires. 
 
 (1) The Sierra de Tandil 25 
 
 The first of these stretches from Mar del Plata on the Atlantic 
 coast in a direction a little to the west of northwest past Tandil to 
 Olavarrfa—a distance of over 300 km. with a mean breadth of about 
 50 km.—being composed of a series of small chains and isolated 
 hills with frequent interruptions and nowhere exceeding an altitude 
 of 450 meters. The features displayed are those of a buried and 
 partially exhumed mountain chain, and along its margins the belt is 
 bordered by inkers of the hard rocks projecting through the soft 
 mantle of the plain. 
 
 ! The formation consists largely of pre-Palasozoic granite, on which 
 rests unconformably a basal quartzitic member followed by dolomite, 
 quartzite, slate, white limestone, etc., to a thickness of fully 150 
 meters. In contrast to the distinctly younger beds of the Sierra de la 
 Ventana, the strata in the Sierra de Tandil are but gently flexed, and 
 are often lying nearly flat, the harder members producing table- 
 topped hills; on the coast the nearly horizontal quartzites have 
 determined Cape Corrientes. This flexuring follows two directions 
 mainly, namely, west-northwest to east-southeast and one almost at 
 right angles thereto, i.e. } north-northeast to south-southwest. The 
 latter represents the mid-Palaeozoic “Brasilides, ” as Keidel has 
 called them, which have been crossed and modified here by the first- 
 mentioned set of the Permo-Triassic ‘ ‘ Gondwanides, ’ ’ so finely 
 displayed in the belt to the southwest, which foldings, although of 
 considerable intensity in that quarter, seem to have faded out to the 
 northeast before reaching Uruguay. 
 
 Where seen by me in one of the outliers of the Dos Hermanos 
 (“The Two Brothers”), near Olavarrfa, the quartzites, resting on a 
 pedestal of red granite, are white, hard, and fine-grained, though 
 somewhat coarser at their base and not unlike the Kuibis quartzites 
 of the Nama system of Southwest Africa, that are also followed by 
 slates and limestones. They are, however, presumed to be of Silurian 
 
 25 Aguirre (1879); Hauthal (1896, 1901); Valentin (1894); and Nagera (1919), who 
 gives a bibliography. 
 
 18 
 
ARGENTINA AND BOLIVIA 
 
 19 
 
 or more probably of Ordovician age, because of impressions referred 
 to Arthropliycus harlani Hall, in certain of the quartzites at Balcarce 26 
 while Siemiradski reported finding Stromatopora and A try pa in the 
 dolomitic group, 27 though such does not appear to have been con¬ 
 firmed. 
 
 It should be observed that this belt stands wholly isolated in the 
 Pampean formation, that it is over 400 km. distant from the possibly 
 equivalent strata in Uruguay and 150 km. from the Palaeozoics of 
 the Sierra de la Ventana and Pillahuinco, and that the few bore-holes 
 that have penetrated the Pampean covering have only proved 
 granite; for example, that at Buenos Aires and at Telen in the district 
 of La Pampa to the west. 
 
 (2) The Sierra de la Ventana 
 
 The Fold-ranges, among which is embraced this most important 
 of the sierras, lie not far to the north of Bahia Blanca and occupy a 
 tract some 180 km. long by 50 km. at its widest. They include the 
 Sierra de Puan, Sierra del Chaco, and Sierra de la Ventana proper in 
 the southwest and west, and the Sierra de Bravard, Sierra de las 
 Tunas, and Sierra de Pillahuinco on the northeast, and make a 
 sweeping double curve with general strike from southeast to north¬ 
 west, bending to the west in the latter direction. In the heart of the 
 ranges the Rio Sauce Grande courses along a strike valley that 
 generally follows the Tillite zone. As with the Sierra de Tandil, the 
 Pampean formation laps around the base of the rising ground and 
 enters many of the valleys. 
 
 The geology has been described by Aguirre, Hauthal, and 
 Schiller, 28 and more particularly by Keidel, 29 who has not only given 
 the most detailed and accurate account, but by the discovery of the 
 Carboniferous tillite was first able to point to the wonderful simi¬ 
 larity with South Africa and show that the ranges, structurally as 
 well as geologically, appeared to form the continuation across the 
 Atlantic of the Fold-ranges of the Cape Western Province. The 
 writer’s researches bear out the value of Keidel’s work and percep¬ 
 tion, for in only one essential does his interpretation differ from that 
 of this Argentine geologist, namely, in the stratigraphical position of 
 the Pillahuinco beds. To the visitor from South Africa the resem¬ 
 blances with that country are simply astounding, such being the case 
 not only with the major elements, but in so many minor respects, as 
 well as in lithological details regarding the tillite and in the presence 
 and attitude of the Tertiary “high-level gravels.” 
 
 The cross-section from southwest to northeast passing through 
 
 26 Ndgera (1919a), p. 7. 
 
 27 Siemiradski (1893). 
 
 28 Aguirre (1891); Hauthal (1892, 1896, 1901, 1904); Schiller (1907). 
 
 29 Keidel (1916), also (1922). 
 
20 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 a> 
 
 two 
 
 d 
 
 > 
 
 the railway station of Sierra de la Ventana (formerly called Rio 
 Sauce Grande) would show (Fig. i) on the southwest a closely folded 
 
 and inverted succession of Palaeozoic strata 
 in which the softer beds display a strong 
 cleavage that is dipping at a high angle 
 to the southwest (shown by broken lines 
 in the diagram). 
 
 The lowest member —a thick, whitish, 
 quartzitic sandstone formation, the equiv¬ 
 alent of the Table Mountain sandstone of 
 the Cape—builds the main range, which 
 in its various aspects strongly recalls the 
 chains of the Langebergen, Outeniqua, and 
 other coastal barriers of the Cape, and 
 which culminates in the Cerro de los Tres 
 Picos (1,280 meters); farther along the 
 crest is the peak pierced by the famous 
 “ventana” or window, from which the 
 range takes its name. On its southwestern 
 foot the basement granite is reported to 
 crop out near Tomquist and Aguas Blan¬ 
 cas, the contact being an inverted one 
 apparently. The sloping ground to the 
 northeast, for a distance of some 5 or 6 
 km. down to the valley of the Rio Sauce 
 Grande, is made by the second member , a 
 series of cleaved slates and greywackes, 
 corresponding to the Bokkeveld and 
 Witteberg series of the Cape, in which 
 the folding can at times be made out, while 
 the valley floor, together with the rising 
 ground just beyond, is composed of a 
 cleaved greenish tillite, the representative 
 of the Dwyka glacials. 
 
 The stratigraphical position of the 
 strata to the northeast has hitherto been 
 in doubt, the folded quartzites and slaty 
 beds of the Sierra de Pillahuinco having 
 been regarded by Schiller as the equiva¬ 
 lent of those in the ranges to the south¬ 
 west, a view satisfactorily disposed of by 
 Keidel, who has nevertheless considered 
 them as probably of pre-Devonian age. 
 My observations, however, have shown an 
 upward passage into them from the tillite, 
 
 *d 
 
 o> 
 
 rt 
 
 & 
 
 to 
 
 d 
 
 X 
 
 <u 
 
 r<D 
 
 o 
 
 w 
 
 73 
 
 o 
 
 • H 
 
 u 
 
 <D 
 
 > 
 
 a 
 
 o 
 
 -M 
 
 d 
 o 
 
 ^ 8 
 8*0 
 
 a 
 
 tfl +» 
 'A a> 
 
 d 
 
 a s 
 
 d 8 
 
 g | 
 
 > ;d 
 
 •a g 
 
 <D ^4 
 "d o 
 
 £ V 
 
 w.g 
 o> 73 
 
 rG C 
 
 -f-> .th 
 
 «4-l K*~ > 
 
 °2 
 oj txo 
 
 *53 ^ 
 a 2 
 8 H 
 
 W 
 
 rt 
 
 a> 
 
 rd 
 
 u 
 
 o 
 
 d 
 
 <v 
 
 rd 
 
 d 
 
 o 
 
 d 
 
 o 
 
 • rH 
 
 in 
 
 8 
 
 O 
 
 o 
 
 d 
 
 tn 
 
 o 
 
 ►1 
 
 Pd 
 
ARGENTINA AND BOLIVIA 
 
 21 
 
 which goes to indicate that these beds are the equivalent of the 
 folded Ecca series of the southwestern part of the Cape Karroo 
 and of the Lafonian of the Falkland Islands, as will be discussed in 
 the sequel. While the entire succession is thus extraordinarily simi¬ 
 lar to those in the Cape and Falklands, its resemblances to the 
 equivalent beds represented in the pre-Cordillera of San Juan to the 
 northwest are, it should be noted, by no means so clearly marked. 
 
 (a) The Infra-tillite Succession —To the excellent account given 
 by Keidel 30 there is little to add. The basal group may possibly be 
 close on 1,000 meters in thickness, but the intense crumpling in this 
 section would prevent any exact measurements being made. In the 
 splendid exposures at the head of the Arroyo San Bernardino the 
 quartzites (Fig. i, I ) are isoclinally folded in a rather regular fashion, 
 the succession being of course an inverted one, the pressure, just 
 as in South Africa, having come from a southerly direction. Owing 
 to a mishap to a negative, I am unable to give a photograph of 
 these fine contortions, but one of Beder’s views, published by Keidel, 
 taken further to the west along the same range, is reproduced, which 
 will serve to bring out this feature very clearly (Plate I). 
 
 Most of the beds are hard, white, fine-grained quartzite, but 
 there are darker and faintly greenish varieties and a number of thin, 
 greenish-gray schistose quartzites and slaty kinds in thinner layers. 
 One conspicuous shale bed (Plate II, A, and Fig. i, ia) perhaps a few 
 hundred meters from the top of the series, is 8 meters thick, a dark- 
 green, silvery slate, with base well defined, which invites comparison 
 with the well-known ‘ * shale band ” of the Table Mountain sandstone, 
 but no glaciated pebbles were seen in it or just below it. The upper¬ 
 most hundred meters of strata are distinctly darker than those com¬ 
 posing the range proper, approaching a greywacke in some respects, 
 the rock being schistose, showing some false-bedding and including 
 a little grit with quartz fragments and pebbles, which was the only 
 coarse material seen. The schistose and darker character tends to 
 obscure the upward passage into the slate group, which, as Keidel 
 has pointed out, is rapid, though not obvious. That the quartzites 
 underlie the slates can be shown by a study of their folded junction. 
 
 To sum Up, the quartzites have marked similarities to the Table 
 Mountain series of the Cape; not to that in the Ceres or Clanwilliam 
 districts, but to the facies in the south, i.e., Mossel Bay, George, or 
 Knysna, where it is finer-grained, where quartz pebbles are rarer and 
 smaller, and where thin layers of dark slates are not unusual. 
 
 In the overlying, seemingly thick, group—called the ‘ ‘ esquistos 
 by Aguirre, the folding can now and again be made out, despite the 
 strong cleavage. The lowest beds are fine slates, with some bands 
 of greywacke appearing higher up, while the topmost beds consist 
 
 10 Keidel (1916), pp. 11-12; 21-22; 28-29. 
 
22 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 not of quartzites, as in the Cape and Falklands, but of highly cleaved 
 greywackes, mostly brownish in color and full of feldspar, finer 
 alternating with coarser varieties, the latter becoming in places quite 
 gritty, while a few pebbles of dark bluish quartzite were observed. 
 The marked cleavage obscures both the dip and the folding, but 
 the topmost bed, in contact with the tillite, is a hard, brown, un¬ 
 cleaved grit. 
 
 These strata are manifestly the equivalent of the middle and 
 upper divisions of the Cape system, much distorted Devonian fossils 
 having indeed been obtained in several places in the slaty basal por¬ 
 tion of the assemblage, while a specimen inclosed in a sandy rock 
 approaching a greywacke, which is exhibited in the La Plata Museum, 
 was derived from this locality. The greywacke, marked (. 2a ) in the 
 section (Fig. i), is evidently the same bed as is shown on Keidel’s 
 map of the territory lying a short distance farther to the northwest 
 along the same strike, and recalls the first or fossiliferous sandstone 
 of the Bokkeveld series and the Tibagy sandstone of Parana. The 
 uppermost group of greywackes and occasional slates, which has 
 so far yielded neither molluscan nor plant remains, is wholly differ¬ 
 ent from the white, fine-grained quartzites with green slaty interca¬ 
 lations of the Witteberg series or those occupying a similar horizon 
 in the Falklands, and in this very important respect the succession 
 here differs from those in South Africa and the Falkland Islands. 
 
 It is uncertain whether the phyllites and greywackes known from 
 the Sierra Pintada in the southeastern part of the Mendoza district 
 are of Devonian age, but, as will be set forth later on (Chapter VI), 
 the strata in Jachal to the north referred to this epoch and containing 
 the typical “austral fauna” in their upper part, are in their basal 
 section no longer white quartzites, but shaly sandstones and sandy 
 shales with bands of soft, dark sandstone and of a friable nature. 
 This marked difference in lithology of the basal portion of the 
 Devonian within the distance of less than 1,000 km. toward the 
 northwest, as compared with the slight change noted in the distance 
 of almost 6,000 km. from the Cape or 2,000 km. from the Falklands 
 can not be overemphasized and will be discussed in Chapter VII. 
 
 ( b ) The Tillite —In the neighborhood of Sierra de la Ventana 
 station the lower part of the glacial beds (4a) is but poorly exposed 
 over the gently undulating ground, but practically continuous out¬ 
 crops can be seen in the higher country to the east of the railway 
 bridge spanning the Rio Sauce Grande, while excellent artificial 
 sections are disclosed in the railway cuttings between kilometer 
 posts 529 and 536. The total breadth of the belt is about 9 km. 
 The prevailing strong cleavage (Plate II, B), directed southwestward 
 at an average angle of about 65°, coupled with the unsatisfied charac¬ 
 ter of the deposit, obscures the true behavior of the formation, which 
 
ARGENTINA AND BOLIVIA 
 
 23 
 
 accounts for the erroneous interpretations of the section by Hauthal 
 and others, but a careful study of the attitude of certain shaly or else 
 quartzitic layers goes to indicate that the formation, though seem¬ 
 ingly dipping southwestward beneath the greywackes in that direc¬ 
 tion, is actually resting upon the latter and as a whole is dipping 
 northeastward below the strata of the Sierra de Pillahuinco. On the 
 southwest the folds are isoclinal with axial planes dipping southwest- 
 ward at nearly 45°, but along the Rio Sauce Grande flattening has 
 occurred and thereafter the glacials are but slightly flexed, to pass 
 ultimately at quite a low angle northeastward beneath the sediments 
 of the Cerro Bonete. I agree with Coleman 31 that Keidel 32 has over¬ 
 estimated the degree of folding that is present in the railway section, 
 since, except for one important reversal of dip, the mass is inclined 
 regularly northeastward, though signs of strong compression are 
 nevertheless clear. The thickness is difficult to estimate here, because 
 of the uncertainty as regards that of the lower part of the forma¬ 
 tion, but it is undoubtedly well in excess of the figure given by 
 Keidel, namely, 60 meters; it would seem to be at the least 200 
 meters, perhaps more. 
 
 The base was studied on the left side of the Arroyo San, Ber¬ 
 nardino, a few kilometers west of the station, the nearest point where 
 the junction with the hard top of the greywacke group (3) is exposed. 
 A careful examination showed the tillite to be dipping at 45 beneath 
 the greywacke, but within the latter are what were taken to be two 
 narrow infolds of the base of the glacials, certain marginal and other 
 characters indicating that they were merely the repetition of one and 
 the same zone. As a rule, the change from greywacke to dark-blue, 
 fine-grained tillite with small inclusions takes place within a distance 
 of a few centimeters, while some small pebbles occasionally occur 
 just within the greywacke itself. At one place a wedge of soft 
 schistose strata appears just within the latter, including some iso¬ 
 lated pebbles like those in the tillite, all of which makes it practi¬ 
 cally assured that there is a perfectly conformable transition from the 
 arenaceous into the glacial formation. This conclusion, therefore, 
 confirms the opinions arrived at by both Hauthal and Keidel. 
 
 The condition thus parallels that in the south of the Cape (south 
 of latitude 33 0 ) where no discordance has been detected, though the 
 change from the one formation to the other is often quite abrupt, but 
 differs from those in the Tanqua Karroo or in the Falklands, where 
 the tillite rests upon a slightly striated surface of the underlying 
 Carboniferous quartzites. 
 
 Of the glacial origin of the formation, a discovery first made by 
 Keidel and since confirmed by Coleman, there can not be the slightest 
 doubt. All the phenomena displayed duplicate down to minutest 
 
 81 Coleman (1918), p. 319* 82 Keidel (1916), P- 2 5 - 
 
24 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 detail those characterizing the Dwyka tillite in the south of the 
 Karroo. The rock is a hard greenish unstratified material, having 
 scattered irregularly and sometimes rather sparsely through it inclu¬ 
 sions that rarely exceed 0.5 meter in length. Among the host of rock 
 types represented, Keidel discovered a lump of fossiliferous limestone, 
 while noteworthy are certain quartzites and greywackes indistin¬ 
 guishable from those of the Sierra de la Vent ana, but it is important 
 to observe in the uncleaved condition. Keidel also mentions erratics 
 of material like the beds in the Sierra de Pillahuinco, but I could find 
 nothing that could with any confidence be ascribed to those forma¬ 
 tions exposed to the northeast. Well-striated and often faceted 
 boulders, such as are typical of glacial tills, can be obtained without 
 difficulty, such as have been figured by Keidel and Coleman. 
 
 Like the Dwyka along the south of the Karroo, the Rio Sauce 
 Grande conglomerate is not one uninterrupted, single body of 
 morainic matter, for there are several surfaces of discontinuity, which 
 evidently mark the cessation and subsequent renewal of glacial 
 activity. At kilometer 535 in a railway cutting is exposed the 
 sharply defined top to a tillite body with a thin, splintery shale rest¬ 
 ing upon the plane or break and passing up in turn into a normal 
 tillite. A little higher up there appears a “gravelly” band of the 
 kind of material termed in South Africa “gravel Dwyka,” charged 
 with small angular and sub-angular stones, while next in turn comes 
 an instance of a “boulder pavement” in which the inclusions within 
 the top of the lower body of morainic material have all been pressed 
 down to one level and striated in a direction, so far as can be made 
 out, north 65° west (true); this is overlain by normal tillite. The way 
 in which the top of each of these pebbles just rises above the surface 
 of discontinuity, and other features, show that the phenomena are 
 due to the overriding of the moraine by the ice itself and are not due 
 to subsequent tectonic movements. Boulder pavements of identical 
 character have been recorded in the Dwyka and in the Carboniferous 
 glacials of Cape Wynyard in Tasmania. Furthermore, there is a very 
 important intercalated zone of quartzitic rocks (Fig. 1, 4b ), probably 
 lenticular in habit, with a maximum thickness of 40 meters, which 
 splits the tillite mass, occurring on a horizon about 70 meters from 
 the top of the glacials building up the outlying peak of the Cerro de 
 Pillahuinco immediately to the west of the 534 km. post and seen 
 again on the railway at 533.50 and 533 and running in a belt 1 km. 
 to the east on the left side of the Arroyo Negro, where it can be seen 
 to pass below the upper tillite (4c). At 531.2 km. the base is found to 
 be an intensely hard, gray, small-pebble conglomerate, crammed with 
 inclusions—some certainly striated—resting through the medium 
 of a greenish, thin, and variable quartzite on the lower tillite, a break 
 
 occurring at this level. This is the band whence most of the specimens 
 
 \. 
 
ARGENTINA AND BOLIVIA 
 
 25 
 
 in museum collections labeled “tillite” come from, though it is 
 decidedly not representative. 
 
 The zone itself consists of this hard conglomerate, followed by 
 brownish to greenish quartzites and greywackes possessing feeble 
 or irregular bedding, with some softer bands allied to tillite. The zone 
 recalls some of the quartzite bands and lenses in the southern Dwyka, 
 for example, some near Matjesfontein station in the Southern 
 Karroo, being in the nature of interglacial sands, while at kilometer 
 531.75 are some of those curious quartzitic “pseudo-boulders, ” such 
 as have been recorded from the Cape and Natal, representing masses 
 of sandy material picked up by the ice and “balled up” during the 
 progress of the glacier. 
 
 The upper part of the glacial series is best seen along the road 
 from Sierra de la Ventana to Peralta, at a point nearly midway 
 between the Cerro de Pillahuinco and the Cerro Bonete. Here the 
 tillite (4 C ) becomes finer grained and more cleaved, inclusions get 
 isolated, smaller, and at length fail entirely, and the rock passes up 
 without a break into a zone (5) about 100 meters thick, of hard, 
 cleaved slates of blue-gray color characterized by little ocherous 
 concretions due to the oxidation of pyrite or marcasite, and these in 
 turn into the group of sandstones, greywackes, flagstones, and quartz¬ 
 ites (1 6 ) building the Cerro Bonete and the hills to the northeast and 
 north, which can for convenience be termed the “Pillahuinco beds. ” 
 
 (c) The Strata of the Sierra de Pillahuinco —This is the only point 
 in this neighborhood where I found the exposures sufficiently good to 
 show the passage between the glacials and the group building the 
 Sierra de Pillahuinco, the age of which has hitherto been in doubt. It 
 is true that Keidel recorded the superposition of the latter upon the 
 tillite at one point, which, however, he took to be due to thrusting, 
 as is not at all unlikely, for in the range lying a few kilometers to 
 the northwest from kilometer post 529 the arenaceous zone near the 
 base of the Pillahuinco beds seems to be thrown against the tillite 
 by faulting. 
 
 The strata that were studied between the Cerro Bonete and 
 Peralta station consist of hard gray-green sandstones and quartzites, 
 pale in the coarser and generally darker in the finer-grained vari¬ 
 eties ; there are also schistose greywackes and some thinner shaly and 
 flaggy bands still displaying cleavage, but the strata as a whole are 
 very different from the Devono-Carboniferous succession to the 
 southwest. To the east of the railway, Keidel has recorded red and 
 green slaty beds, often blotchy, and beds of clayey limestone and 
 marl similar to those observed by Schiller on the northern edge of 
 the Sierra de Pillahuinco, north of Peralta, intercalated with the 
 quartzites—probably higher zones within the series. 
 
 The folding in the Cerro Bonete is gentle, but, followed north- 
 
26 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 eastward, it becomes more intense. While asymmetrical there, as is 
 also the case just northwest of the railway, on proceeding northward 
 the folds, as Keidel has remarked, become overturned to the south¬ 
 west, though it is important to observe that the crumpling after¬ 
 wards decreases and that opposite Stegmann Station (507 km.) the 
 beds are now practically flat. The fact that the pre-Devonian strata 
 of the Sierra de Tandil are so little disturbed indicates that the 
 “ Gondwanides ” must be fading out towards the northeast. 
 
 The parallel is extraordinarily good with the Ecca series of the 
 southwestern Cape, say between Laingsburg and Prince Albert, where 
 the foldings die out in a northerly direction within a relatively short 
 distance. The southern Ecca, moreover, contains thick groups of 
 hard grayish and dark greenish sandstones, almost a quartzite in 
 places, and much flagstone in addition to hard green and purple mud¬ 
 stones and shales and sometimes impure limestones. The unfossil- 
 iferous nature of the Cape beds is another parallel, for in the brief 
 search possible near Peralta only worm castings were seen and some 
 poor impressions that might be parts of plants. A further examina¬ 
 tion of these beds is hence extremely desirable. In the corresponding 
 portion of the Lafonian system of the Falklands, Baker has also 
 recorded fine-grained and intensely hard greenish sandstones, and 
 sometimes typical slates. It should be noted that the carbonaceous 
 shales of the Dwyka or of the Iraty series do not appear to be repre¬ 
 sented, but they are certainly missing in the Falklands and probably 
 too in the district of San Juan to the northwest. 
 
 id) The High-Level Gravels —To my mind the most striking par¬ 
 allel with the Cape is formed by the consolidated gravels and breccias 
 resting on benches or terraces cut along the inner side of the quartz¬ 
 ite chain of the Sierra de la Ventana (Plate III, A). Indeed, viewed 
 even at short range, I had great difficulty in realizing that this was 
 another continent and not some portion of one of the southern 
 districts in the Cape, say Caledon, Oudtshoorn, or Uniondale. Taken 
 in conjunction with the other resemblances already mentioned, the 
 parallelism is so wonderfully close that one can not escape from the 
 conclusion that the geological histories of these two countries 
 must have been all but identical from mid-Palaeozoic down to 
 early Tertiary. 
 
 As Keidel 33 has so well described, and exactly as in the Cape, the 
 highest terrace is merely a shelf incised in the quartzites beneath the 
 crest of the range, running at a level of close on 800 meters above the 
 ocean. The second has an altitude varying between 450 and 550 
 meters, cut at about the junction of the quartzites and the younger 
 slates, and evidently had a wide distribution formerly, since the 
 spurs of slates and greywackes between the valleys leading down to 
 
 33 Keidel (1916), pp. 37-42. 
 
ARGENTINA AND BOLIVIA 
 
 27 
 
 the Rio Sauce Grande show all the characters of a dissected peneplain 
 transecting the highly tilted Palaeozoics and sloping from the range 
 gently toward the northeast. In places this terrace still bears cap¬ 
 pings of hard, bright-red conglomerates and breccias, as, for example, 
 at the head of the Arroyo San Bernardino below the Tres Picos (Fig. 
 i, 7) where the deposit is composed chiefly of subrounded to angular 
 lumps of quartzite coarsely stratified and set in a hard arenaceous 
 cement colored with oxides of iron. The material resembles exactly 
 in its character and habit the so-called “high-level gravels” or 
 “ferricretes” of the southern Cape; comparison might be made 
 between Keidel’s Plates IX and X and Plates XXIII and XXIV 
 in Rogers and Du Toit’s Geology of Cape Colony. 
 
 Obviously these are terrestrial scree and gravel deposits, which, 
 from analogy with South Africa, were formed on a peneplain carved 
 out during the early part of the Tertiary, and which in similar fashion 
 have since been dissected to a depth of from 100 to 200 meters, as 
 marked by accumulations of boulders or by patches of younger 
 gravels at various levels in the valleys. This in the Cape was due to 
 the elevation of the interior of the continent in about the Miocene or 
 Pliocene. In Argentina, however, this broken country subsequently 
 became in great part buried beneath the mantle of Pampean loess. 
 
 Attention might be drawn to the fact that in the neighborhood of 
 Bahia Blanca borings have disclosed at a depth of about 600 meters 
 beneath marine (probably lower) Tertiary a group of red clays and 
 grey and white marls sloping gently to the southeast, which, although 
 unfossiliferous, are regarded as upper Cretaceous from their resem¬ 
 blance to strata of that age developed in Patagonia. 34 
 
 SECTION F. THE SOUTHERN PRE-CORD 1 LLERAN REGION 
 
 As Windhausen 35 has pointed out, the fold-belt of the area just 
 described is doubtless prolonged to the northwest beneath the Pam¬ 
 pean formation, curving somewhat toward the north in that direc¬ 
 tion, for Gondwana beds appear from below Pleistocene basalts and 
 Triassic volcanics in the Sierra Pintada near San Rafael on the Rio 
 Atuel, where they exhibit a north-northwesterly strike. 
 
 Unfortunately, very little is known about this rather important 
 inlier, though Windhausen has mentioned the presence of a tillite and 
 also an unconformity at the base of the Gondwana beds, but the 
 intensity of the erosion that followed the late Permian foldings is 
 indicated by the unconformity between these and the succeeding 
 Triassic quartz-porphyries and tuffs. 36 
 
 It would appear that the “ Gondwanides ”—the “Hercynian 
 folds” of Groeber—extend northward into Mendoza and San Juan; 
 
 34 Wichmann (1918). 36 Windhausen, p. 35; Groeber (1918), p. 53. 
 
 35 Windhausen (1918), p. 35. 
 
28 GEOLOGICAL COMPARISON OP SOUTH AMERICA WITH SOUTH AFRICA 
 
 but in that direction the strata have subsequently become involved 
 in the younger Andine movements with a north-south strike and are 
 much affected by overturning and thrusting, whereby slices of differ¬ 
 ent formations down to the Tertiary have repeatedly been brought 
 into abnormal relationships with one another and the continuity of 
 the relatively narrow belts of Gondwana frequently interrupted. 
 The earlier north-northwesterly directed flexures can nevertheless 
 still be made out in parts of San Juan and Jachal. Throughout this 
 region the Gondwana, known as the “Paganzo, ” rests unconforma- 
 bly upon the Devonian or upon older formations. 
 
 (1) The Glacial Beds of San Juan and Jachal 
 
 At the foot of the Sierra Chica de Zonda, to the south of the town 
 of San Juan, various strata have, through the labors of Stappenbeck, 
 Bodenbender, and Keidel, 37 long been known to crop out, containing 
 examples not only of Carboniferous plants, but of the Glossopteris 
 flora, and also glacial beds. It becomes important, therefore, to con¬ 
 sider the precise relationships of the formations involved, in view 
 particularly of the opinion expressed by Keidel that there are two 
 distinct plant-bearing groups, the one pre-Gondwana with European 
 Carboniferous forms, the other overlying the glacials and considered 
 as of “Permian” age. It will accordingly be instructive to describe 
 in some detail the certain sections that I visited at the foot of this 
 great range some 25 km. to the south of San Juan, where the relation¬ 
 ship of the plant-bearing zone to the glacials is beyond all question. 
 The area examined is a few kilometers long, extending from a point 
 a little to the Rio Grande down to the Rio de la Mina, as shown in the 
 sketch map (Fig. 2). 
 
 The Sierra consists of a lofty barrier of gray-weathering Ordo¬ 
 vician limestones and along its eastern base extends a group of highly 
 tilted green slates and greywackes, with some thin fossiliferous lime¬ 
 stones, presumably of Silurian age. The Gondwana beds are exposed 
 in the foothills and in a number of small ravines, but on leaving the 
 slopes their outcrops soon become hidden beneath deposits of boul¬ 
 ders and gravels. Good exposures are found of the basal beds, but the 
 outcrops of the higher groups are limited and sometimes isolated. 
 Piecing together the various sections, which can be done without 
 much uncertainty, there is discovered a thickness of about 360 meters 
 of the Gondwana, or, as they have customarily been termed, ‘ ‘ Pagan- 
 zo” beds, resting unconformably upon the Lower Palaeozoic and 
 dipping eastward below a covering of tilted Calchaquenos (Tertiary) 
 and nearly horizontal Quaternary gravels. 
 
 The Gondwana embraces three distinct glacial horizons , one of 
 which forms the very base of the system, while the intercalated sedi- 
 
 17 Stappenbeck (1910); Keidel (1922). 
 
ARGENTINA AND BOLIVIA 
 
 29 
 
 ments between that and the second horizon yielded plant remains, 
 among which were the Carboniferous genera Cardiopteris and Rha- 
 
 Fig. 2—Geological sketch-map of a small area along the eastern base of the 
 Sierra Chica de Zonda to the south of San Juan. OS, Ordovician and 
 Silurian limestones and shales. 1-7 Paganzo system: 1, first glacial zone; 
 2, plant-bearing shales and sandstones; 3, second glacial zone; 4, sand¬ 
 stones; 5, third glacial zone; 6, shales with some sandstones; 7, red and 
 dark shales. C, Calchaquenos and Quaternary gravels. 
 
 copteris , but no marine fossils. The discovery of these forms, taken in 
 conjunction with the other plants found in this locality by previous 
 workers, tends to invalidate Keidel’s argument for the existence of 
 
30 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 two distinct formations, one with Carboniferous plants older than 
 the Paganzo, the other having glacials and containing the Glossop- 
 teris flora. For this reason it does not become necessary to make use 
 of the two-fold division made by him into the “Jejenes” and the 
 “Zonda” series, nor need the supposed relationship of these series be 
 argued. In only this respect do the very valuable observations and 
 conclusions of Keidel need to be modified, though unfortunately it 
 will affect his criticisms of the writings of Bodenbender, Hausen, 
 Penck, and others. 
 
 In the small ravine o. 5 km. north of the Rio Grande the base of the 
 Gondwana is excellently exposed, the first glacial zone (Plate III, B) 
 from 8 to 20 meters thick, resting upon an uneven surface of highly 
 tilted slates (A) and limestones (B) (Fig. 3). At the point X the 
 floor is actually vertical, and in the pocket below it lies true ground 
 
 Fig. 3—-Section, about 75 meters in height, in ravine north of the Rio Grande, 
 Sierra Chica de Zonda. For explanation, see text. 
 
 moraine (a), above which comes a thin, curving layer of green fissile 
 shale ( b ) with annelid or crustacean markings, which laps over the 
 limestone band (B), the top of which is nevertheless polished and 
 striated, the groovings (Plate IV, A), after making due allowance for 
 the subsequent tilting, being directed north-northwest-south- 
 southeast. The overlying glacials (c) range from a true, unbedded 
 tillite of normal character to a stratified fluvio-glacial conglomerate 
 having the inclusions lying with their longer axes parallel to the 
 stratification, though striated boulders are common. They pass up 
 into, or to some extent are even interlaminated with, dark carbon¬ 
 aceous shales and micaceous flagstones and thin sandstones with 
 occasional plant remains (d), grading up into a group of flagstones 
 (e) and yellowish-gray sandstones (/), sometimes micaceous, getting 
 coarser toward the summit, the succession being clearly shown in 
 Plate IV, B. 
 
 At a height of about 165 meters above the base the second glacial 
 zone makes its appearance (Plate V, A). In the neighboring Rio de 
 
ARGENTINA AND BOLIVIA 
 
 31 
 
 Jejenes, on the southern side of the ravine (Plate V, B ), its base is sharp, 
 the band incloses lenses of sandy stuff as shown in the nearby view 
 (Plate VI), and the inclusions are as a rule small and of hard rock, 
 though one of them, composed of gneiss and resting directly upon 
 the sandstone, constituting the floor, is nearly 3 meters in diameter. 
 The succession, it might be remarked, is a perfectly conformable one. 
 In the lower ground to the east, where this zone is proceeding to dis¬ 
 appear beneath higher beds, it is nearly 50 meters thick, composed 
 of a lower group of fliivio-glacial sandstones and tills and a thin bed 
 of shale above, passing up through boulder-shale into tillite (Plate 
 VII, A ); finely striated erratics can readily be picked up. An example 
 has been figured by Coleman and it is hence unnecessary to give any 
 additional illustrations. 
 
 The top passes through pebbly grit and sandstone into greenish 
 shaly beds, in which at 50 meters there appears boulder-shale, and 
 next the third glacial zone, about 12 meters thick, overlain by more 
 hard, greenish shales, and these, as seen to the south in the channel of 
 the Rio de la Mina, by alternations of sandstone and shale, with a thin 
 zone of soft black carbonaceous shales constituting the highest strata 
 visible. Although so thin, this upper tillite carries some exceptionally 
 large inclusions, mainly of Ordovician limestone, one slab of which, 
 a couple of meters in length, contained several large coiled cephalo- 
 pods. The conspicuous green coloring in some of the beds, absent 
 in the region to the northeast, serves to link this formation with 
 the equivalent Pillahuinco beds of the Sierra de la Ventana. 
 
 In this terrain, crossing the Rio Grande and the Agua de Jejenes, 
 the beds are dipping regularly eastward and are not dislocated, 
 there can not be the slightest doubt that the sequence is a normal 
 upward one which is resting unconformably upon the Silurian and is 
 embracing Paganzo beds only. Great weight, therefore, must be 
 attached to this section, since it provides us with the key to the 
 Permo-Carboniferous stratigraphy of a very wide region in the pre- 
 Cordillera, as will be discussed later on. 
 
 Proceeding southward in the direction of the Rio de la Mina, 
 irregularities make their appearance, due partly to the fact that the 
 floor beneath the system seems to have been highly diversified, and 
 partly to subsequent tectonic movements, by which anticlinal cores 
 of Ordovician limestone have been driven through the younger for¬ 
 mation. By such movements, what from their lithological peculiari¬ 
 ties must be regarded as slightly higher zones of the Paganzo, appear 
 in two fine continuous domes (Plate VII, B), from beneath the folded 
 unconformable Calchaquenos formation a little to the southwest of 
 the lime-kilns of Carpinteria not far from the Cerro Bola, a group of 
 carbonaceous and micaceous shales with nodules carrying plant 
 remains, soft micaceous sandstones, and reddish mudstones resting 
 
32 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 seemingly through the medium of a thrust-plane upon a core of Ordo¬ 
 vician limestone; this has been figured very roughly by Stappenbeck 
 (1910, p. 36, Fig. 1). 
 
 The overshadowing feature of the sections just described is the 
 direct evidence that they furnish upon the position of the plant¬ 
 bearing horizons in relation to the glacials, at the same time proving 
 the close association of European Carboniferous plant forms with 
 members of the Glossopteris flora, thereby bringing the succession 
 into harmony with the sequences observed more to the north and 
 northeast in the Sierra Umango, Sierra Villa Union, Sierra de los 
 Llanos, etc., namely, near the base of the Paganzo system, as will be 
 discussed later. 
 
 Seemingly the glacials are represented at intervals along the base 
 or flanks of this range,. as, for example, in the Cuesta de Pedemal 
 farther to the south, where Stappenbeck 38 reported thick conglomer¬ 
 ates overlain by beds with a “Culm flora” resting upon the Silurian; 
 but this and other cases require to be reexamined in the light of the 
 sections described from near Rio Grande. 
 
 Working southward from this “type” region, mention has to be 
 made of the fine section on the northwestern side of the Cerro Pelado 
 between Mendoza and Uspallata, where, according to Stappenbeck 39 
 the Paganzo, thrown down against Ordovician limestones and 
 dipping southeastward at a pretty high angle, has a development of 
 basal conglomerate, stated by Keidel to be glacial in its nature, rest¬ 
 ing on presumed Devonian and followed by conglomerates, arkoses, 
 sandstones, shales, bituminous shales, and impure coals from which 
 Sigillaria , Annularia, and Gangamopteris have been collected. This 
 section is of further interest in that here and also at Villa Vicenze 
 toward the north these beds, according to Keidel, 40 are overlain with 
 strong unconformity by the sediments and volcanics of the upper 
 Triassic. 
 
 The mention by Windhausen 41 of the “Permian” glacial con¬ 
 glomerate resting unconformably on older beds near San Rafael to 
 the south in the curious Sierra Pintada with its northwest-southeast 
 trend is important, as such would indicate a link between the glacials 
 of the Sierra de la Ventana and those of the Mendoza district. 
 
 In the museum of the Geological Survey at Buenos Aires stands 
 a photograph of an exposure of a tillite labeled “Arroyo del Chiquero, 
 High Cordillera, Mendoza,” which apparently refers to outcrops at 
 the base of the Sierra de Tigre, some 40 km. to the northwest of 
 Uspallata. This helps to connect with the important occurrence 
 recorded by Keidel at Leoncito Encima in the same valley further to 
 the north, 25 Ion. distant from Barreal and with that at the Quebrada 
 
 38 Stappenbeck (1910), p. 36. 
 
 39 Stappenbeck (1910), p. 51 and Fig. 2. 
 
 40 Verbal communication. 
 
 41 Windhausen (1918), p. 35. 
 
ARGENTINA AND BOLIVIA 
 
 33 
 
 del Salto immediately to the east of the latter village, referred to by 
 Stappenbeck, 42 though he did not then realize the glacial character of 
 the basement conglomerate. Compared with the succession in the 
 Sierra Chica de Zonda these two occurrences are outstanding, because 
 the presence of marine fossils in them indicates that within a length 
 of 45 km., though appreciably greater distance when the effect of 
 folding has been allowed for, the formation has changed in character 
 from continental to marine ; it has now been termed the Tontal series. 
 
 In the Arroyo de Cabeceras, about 3 km. above the homestead of 
 Leoncito Encima, I discovered the tillite folded in with the pre- 
 
 Fig. 4—An infolded patch (stippled) of tillite in Palaeozoic beds close to 
 
 Leoncito Encima, Barreal. 
 
 sumed Devonian slates and greywackes in an extraordinary fashion, 
 as showm in Figure 4, a feature due to the superposition of the pre¬ 
 vailing north-south crumplings of the pre-Cordillera upon a pre¬ 
 sumably older northwest-southeast folding. The base of the tillite 
 transgresses across the strike of the supporting Palaeozoics and the 
 unconformable contact can clearly be made out at several points, 
 where the surface below, when visible, is polished and shows striae 
 directed north-northwest-south-southeast (making allowance for 
 the considerable subsequent tilting of the floor) and which are not 
 strictly parallel, as in slickensided faces, but cross one another at a 
 very acute angle in the manner characteristic of glacial pavements. 
 The tillite is dark green, fine-grained, with small inclusions, rarely 
 more than 15 cm. in diameter, many of them being finely striated; 
 
 42 Stappenbeck (1910), p. 46. 
 
34 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 i 
 
 a rude bedding is present in places, brought out by some banded and 
 more quartzitic layers. The thickness preserved in this narrow ravine 
 would not be more than about 30 meters. 
 
 About 150 meters distant from the homestead of Leoncito Encima 
 is the outlier mentioned by Keidel. 43 Lack of time unfortunately 
 prevented any detailed examination of either the base or the topmost 
 beds, but some sandstones and thin shales were observed overlain by 
 tillite, and the latter in turn by gritty sandstone with pebbles, fol¬ 
 lowed by fluvio-glacial cong^merates. Important was the presence 
 of some particularly fine “boulder pavements” of the kind recorded 
 in the Sierra de la Vent ana, with erratics exposed on the curving, 
 ploughed-up surfaces, showing good striae directed north-northwest 
 and therefore in full agreement with the evidence obtained higher up 
 the valley, as well as in the Sierra Chica de Zonda. Fossils were 
 collected here by Keidel either as casts or in nodules of clayey lime¬ 
 stone, principally of brachiopods such as Spirifer, Dielasma, etc., but 
 not specifically determined by him, though he has noted forms resem¬ 
 bling Spirifer alatus and S. rugulatus, suggestive of the Permian; 44 
 Productus was absent. Whether these beds at Leoncito Encima are 
 really Permian, as supposed by Keidel, or Carboniferous like those at 
 Barreal, to be described below, can only be decided by the working 
 out of these fossils. 
 
 I was unable to trace the glacials continuously northward, but 
 feel fairly confident that it must be the same formation that is laid 
 bare immediately to the east of Barreal, in the Quebrada del Salto, 
 the beds being finely sectioned by a small ravine a kilometer to the 
 south of the latter, this dry watercourse rising in an anticlinal valley 
 scooped out of shales overlying the tillite and hemmed in by a ring 
 of hard gray and speckled quartzitic sandstones, thin-bedded red 
 flagstones, gray and bluish siliceous shales, and some soft dark shales 
 (Plate VIII, A). The tillite is greenish and homogeneous, passing by 
 degrees through sandy mudstones into olive shales with occasional 
 small pebbles, and from these into olive sandy shales without inclu¬ 
 sions. Not more than 25 meters of glacial material was exposed at 
 this spot, with the base not visible, though not improbably a search 
 to the south would disclose its floor. 
 
 Fossils occur not only in the overlying shales in profusion, but 
 less abundantly in the mudstones with erratics, though absent from 
 the tillite proper. Spirifer , Productus , Pseudamusium , and Spiriferina 
 are common, while Polypora and Fenestella are conspicuous low down 
 in the mudstones, thus duplicating the features seen in the lower 
 marine series of the Hunter River region of New South Wales, with 
 its glacial erratics and abundant polyzoa. 
 
 Some 150 to 200 meters above this fossiliferous zone, exposed 
 
 43 Keidel (1922), p. 256. 44 p. 334. 
 
ARGENTINA AND BOLIVIA 
 
 35 
 
 close by and just to the north in the Quebrada del Salto, is another 
 horizon of dark, reddish weathering, sandy shales from which numer¬ 
 ous mollusca were collected; this must be the locality discovered by 
 Stappenbeck, 45 from which he obtained a fauna with Spirifer supra- 
 mosquensis Nik., etc, correctly regarded by him as of upper Carbon¬ 
 iferous age. Some higher beds of hard sandstones and shales follow, 
 covered up by Calchaquenos to the north, while on the east the suc¬ 
 cession ends by faulting against “Rhaetic” beds. 
 
 The forms collected from these two horizons are fully discussed 
 by Dr. Cowper Reed in the Appendix, and it will suffice therefore to 
 quote his important decision that these Barreal faunas can without 
 hesitation be ascribed to the Carboniferous and probably to the base 
 of the upper Carboniferous. The tillite should belong, then, to about 
 the border-line between the Moscovian and theUralian, that is to say, 
 between the Westphalian and the Stephanian, wherefore it has to be 
 concluded that this succession near Barreal must be the marine 
 equivalent of the glacials with their plant-bearing zone of the Paganzo 
 system of the Sierra Chica de Zonda. Phases transitional between 
 them will now have to be sought for in the intermediate belts of 
 Paganzo mapped by Stappenbeck 46 in the Sierra Alta de Zonda and 
 on the western side of the Paramillo de Tontal, where, as recorded by 
 him, conglomerates, arkose, green and reddish slates, and sandstones 
 rest on “Devonian” greywackes. 
 
 Further to the north towards Jachal, illuminating sections have 
 been recorded by Stappenbeck and in more detail by Keidel. 47 For 
 example, fluvio-glacial beds are mentioned by the latter as resting 
 with slight discordance upon a thick group of unfossiliferous though 
 post-Devonian greywackes in the Quebrada de Talacasto (lat. 31 0 ). 
 In the Lomas de los Piojos, a little to the south of Jachal, reposing 
 unconformably on the lower Devonian with Leptocoelia flabellites or 
 on upper Silurian beds, is a tillite passing up into fluvio-glacial con¬ 
 glomerates and these into sandstones with occasional boulders, the 
 whole possessing a thickness of from 50 to 60 meters. At one spot 
 Keidel 48 found the morainic matter resting upon a well-striated 
 pavement of fossiliferous Devonian sandstones with the groov¬ 
 ings directed northwest-southeast (see his Plate VII, Fig. 2). 
 
 Evidence of glacial deposits is forthcoming from the Cerro Fuerte, 
 to the east of Jachal, and also from farther north, near Guandacol, 
 similar strata cropping out around the southern end of the Cerro 
 Villa Union and the Sierra de Umango, being well seen, for example, 
 on the western side of the Cerro Guandacol, as described by Boden- 
 bender and Hausen, 49 the latter recording a coarse, tillite-like con- 
 
 45 Stappenbeck (1910), p. 37. 
 * 6 Ibid., p. 43. 
 
 47 Keidel (1921). 
 
 is Ibid., p. 59. 
 
 49 Hausen (1921), p. 70. 
 
36 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 * 
 
 glomerate at the base, followed, but overlapped, by sandstones and 
 shales and by much younger stages of the Paganzo, including beds of 
 Rhaetic age. More detailed reference to these last and to their equiv¬ 
 alents in the ranges more to the east, e.g., the Sierra de Sahogasta, 
 Sierra del Valle Fertil, and Nevada de Famatina, will be made later. 
 Undoubted glacials have not yet been recorded to the east and north¬ 
 east of the last named, barring Penck’s assertion of an unbedded 
 tillite-like formation close to Angulos on the eastern side of the chain. 
 Instructive is his discovery in the Puna de Atacama of Cardiopteris 
 polymorpha in coaly shales resting on the conglomeratic sandstones of 
 the lowest division of the Paganzo, as in San Juan, while intercalated 
 basaltic flows are also present. 
 
 (2) The Plant-bearing Beds and Their Age 
 
 Though this discussion might more appropriately follow the 
 succeeding account of the Paganzo system, it is most conveniently 
 introduced here in order to demonstrate the close relation of the 
 fossiliferous bands to the glacial horizons of the Sierra Chica de 
 Zonda. 
 
 In the course of a brief search in the Rio Grande I obtained from 
 the carbonaceous shales (d t Fig. 3) between the first and second 
 glacial horizon and about 50 meters above the base of the system, 
 specimens of Cardiopteris polymorpha and Rhacopteris szajnochai , 
 while Bodenbender 50 collected (almost certainly from this band, 
 which is to be found cropping out again in the next ravine nearby, 
 that of the Arroyo de Jejenes) Rhacopteris szajnochai , and in addition 
 Gangamopteris cyclopteroides , Glossopteris browniana, Cordaites , species 
 of Sphenopteris , and fragments of a fish, Rhadinichthys argentinus 
 Tornq. So far as I can make out, just a little to the north of the Rio 
 Grande and west of Paradero kilometer 489 is Rinconada, whence 
 Bodenbender gathered Cardiopteris polymorpha , Adiantites antiquus , 
 Neuropteridium validum , and species of Sphenopteris . Some 5 km. 
 south of Jejenes and to the west of Carpinterla he also discovered 
 Glossopteris ampla , presumably in the same beds as Bergiopteris 
 insignis , Lepidodendron cf. australe, and Asterocalamites scrobiculatus, 
 while beyond at Retamito, in the Quebrada del Agua, in beds over- 
 lying thick conglomerates—not improbably glacials—resting upon 
 the Silurian, Bodenbender found Rhacopteris sp., Botrychiopsis 
 weissiana , Asterocalamites scrobiculatus , Lepidodendron pedroanum, 
 L. australe , L. cf. notlium , and Cordaites sp. Kurtz 51 has described 
 several other forms from these several localities, such as Rhacopteris 
 inequilatera , var. ovata, Archceopteris argentine , and Cardiopteris 
 elegans. 
 
 60 Bodenbender (1902), p. 203, (1911), p. 86; Stappenbeck (1910), p. 40. 
 
 61 Kurtz (1921-22). 
 
ARGENTINA AND BOLIVIA 
 
 37 
 
 More to the south, in bituminous shales above the glacials of the 
 Cerro Pelado, Gangamopteris is associated with Sigillaria and Astero- 
 calamites. Traveling northward, one finds at Trapiche, 62 between 
 Jachal and Guandacol, Cardiopteris polymorpha, with Lepidodendron 
 (. Lepidophloios) laricinus , Cordaites hislopi, and Neuropteridium 
 validum, the latter also at Cerro Bola, a little to the east, and at Sano- 
 gasta, while in the Cuesta Colorada (Escala de Famatina) Rhacop- 
 teris inequilatera has been gathered. Penck 53 has reported the finding 
 in the Puerta de Guanchin in the Puna de Atacama, of Cardiopteris 
 polymorpha, Botrychiopsis weissiana, Phyllotheca australis, and 
 Gingkophyllum grasseti in clay shales and thin impure coals overlying 
 the basal conglomerate, here 22 meters thick. In the Sierra de Vilgo 
 at the Cuesta de la Torre, Lepidodendron aculeatum occurs low down 
 in stage I of the Paganzo. At Saladillo, 54 near the southern end of 
 the Sierra del Velasco, black shales have yielded Lepidodendron selagi- 
 noides and L. veltheimianum . On the other hand, proceeding south¬ 
 eastward into the Sierra de los Llanos, we have recorded from the 
 Arroyo Totoral and elsewhere Neuropteridium validum, Cordaites 
 hislopi , Glossopteris retifera, G. indica, Phyllotheca deliquescens, and 
 other forms, while at Bajo de Velis, 55 in the Sierra de San Luis farther 
 to the south, has been obtained the well-known assemblage including 
 Neuropteridium validum, Gangamopteris cyclopteroides, Cordaites 
 hislopi, Noeggerathiopsis whittiana, Rhipidopsis densinervis, R. 
 ginkgoides, etc. 
 
 Keidel 56 has been to great pains in order to prove that there are 
 actually two distinct series resting unconformably upon the older 
 Palaeozoics, one with a flora including lower and upper Carboniferous 
 forms and hence of Carboniferous age, overlain by the second formed 
 of the glacials and the sediments with the Glossopteris flora, belong¬ 
 ing to the Gondwana system and of “Permian” age. 
 
 It must candidly be admitted that the evidence as to the precise 
 horizons from which these several identified collections were made is 
 not beyond criticism, and unfortunately my limited opportunity for 
 collecting did not enable me to dispel such doubts, though the ques¬ 
 tion could very readily be settled. It can nevertheless be affirmed 
 that along both the Rio Grande and the Arroyo de Jejenes the car¬ 
 bonaceous shales (d, Fig. 3) constitute a zone such as would natur¬ 
 ally be searched by the collector, although not the only fossiliferous 
 horizon present, since the upper part of the sandstone group, /, also 
 yielded some poorly striated stems. No plants were observed in the 
 sandstones between the second and third glacial zones and, sugges¬ 
 tively, no strata higher than the latter are represented in those two 
 
 52 Bodenbender (1911), pp. 82-83. 
 63 Penck (1920), p. 146. 
 
 M Bodenbender (1911), p. 81. 
 
 65 Kurtz (1895). 
 60 Keidel (1922). 
 
38 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 localities. To the south, however, along the Rio de la Mina, plant 
 fragments were noticed in shaly sandstones between 30 and 40 meters 
 above the third glacial zone. 
 
 It would consequently seem most likely that Bodenbender’s 
 specimens mentioned above from the Arroyo de Jejenes came from 
 the strata lying between the first and second glacial zones. Boden- 
 bender, Kurtz, and others nowhere remark that the plants collected 
 by them in any one locality near the base of the Sierra Chica de 
 Zonda were actually derived from distinct horizons, wherefore, until 
 more careful collecting has been carried out, we can only assume that 
 each lot was obtained from some stratigraphical unit rather than 
 from several horizons a good distance apart. 
 
 In analyzing these various collections, we observe further that in 
 one locality or another some typical member of the “northern flora’’ 
 is in apparent association with some member of the “southern, ” and 
 this in so many ways that we have difficulty in escaping from the 
 conclusion that all these various plants must probably have occurred 
 intermingled on horizons intimately associated with the glacial 
 series. It will, however, not be denied that the southern forms might 
 perhaps have come from slightly higher levels within that particular 
 series than those that yielded the northern elements, and it will 
 therefore be conceded that Gangamopteris , Glossopteris, etc., if not 
 actually accompanying Rhacopteris, Cardiopteris , etc., must occur 
 shortly above this lowest plant-bearing carbonaceous zone d , with 
 its northern Carboniferous flora. A search should enable this highly 
 important palaeobotanical relationship to be definitely established. 
 
 Granted that the data have been correctly interpreted, the pres¬ 
 ence of Gangamopteris, Glossopteris , Neuropteridium, etc., would 
 suggest that the whole of this definitely conformable succession, 
 down to its very base, resting discordantly upon pre-Carboniferous 
 strata, should be regarded as forming part and parcel of the Gond- 
 wana system. In this conclusion I find myself in accord with Boden- 
 bender, Stappenbeck, and Penck. The sections on the Rfo Grande 
 and Arroyo de Jejenes are particularly clear, and, while some uncer¬ 
 tainty attaches to the precise levels from which many of the plants 
 have been collected, the interpretation here advanced does not seem 
 to me to do violence to the facts. 
 
 It is most suggestive that in the chain of outcrops from the Cerro 
 Pelado due northward to the Puna de Atacama and eastward into 
 Famatina, the bulk of the plants appear to have been found in a zone 
 of carbonaceous shales with occasional thin layers of impure coal 
 that lies not far above the basal conglomerates and arkoses, beds 
 largely of glacial origin. From the section in the Sierra Chica de 
 Zonda it is clear that this zone (d, Fig. 3) is overlain by glacials as 
 well, but how far the latter extend northward is a matter for future 
 
ARGENTINA AND BOLIVIA 
 
 39 
 
 investigation. It is precisely, too, in this western section of Argen¬ 
 tina, that the northern plant forms with a Carboniferous aspect 
 are to be found, and they most decidedly indicate an age for the 
 inclosing strata not younger than upper Carboniferous (Stephanian or 
 Pennsylvanian) , possibly even the top of the middle Carboniferous. 
 
 On the other hand, Bodenbender 57 is quite definite that in the 
 region eastward, Sierra de Velasco, Sierra de los Llanos, etc., the 
 plant-bearing beds are of similar lithological character and are all 
 more or less on the same particular horizon, within Stage I, seemingly 
 in its upper half, and hence apparently on a slightly higher level than 
 in the west. It is significant too, that in the east the records of 
 northern plant forms all but fail, but that Lepidodendron aculeatum 
 occurs in the Cerro de Vilgo lower down than is usual, while near 
 Cerro Bola in the west Cordaites hislopi and Neuropteridium validum 
 are represented near the top of the group. Taken in conjunction 
 with the fact that some carbonaceous shales also occur in the Sierra 
 Chica de Zonda some distance above the third glacial horizon, there 
 is a definitely proved stratigraphical overlapping toward the east 
 (referred to below), causing higher zones to come to rest on the ancient 
 rocks, it would seem that in the west lower horizons may generally be 
 represented, which tend to be missing in the east. This view is sup¬ 
 ported by the triple development of the glacials in the west and their 
 absence, or perhaps their non-record, in the east. The first mentioned 
 condition is certainly the case at Bajo de Velis, where I was able 
 to examine the base of the Paganzo, for the beds with Gangam - 
 opteris rest directly on schists without the intervention of a glacial 
 conglomerate. 
 
 Some remarks are needed regarding the plants themselves as 
 indicators of age. The determinations were all made by Kurtz , 58 and 
 in a number of cases no descriptions were published, though recently 
 a few have posthumously appeared , 59 in most cases, however, unac¬ 
 companied by figures. I consequently have found it difficult to 
 criticize the determinations, while, when in Cordoba, I had only time 
 to examine the extensive collection of Rhastic plants in the museum 
 of the university. 
 
 It might be remarked at the outset that the majority of the mem¬ 
 bers of the Glossopteris flora are of little or no value in establishing the 
 absolute age of the beds. Recent work has been showing more and 
 more that certain genera and species thereof had a long range in 
 time and these sections in Argentina are of immense importance, 
 therefore, in demonstrating that Gangamopteris cyclopteroides , Glos¬ 
 sopteris browniana, and Cordaites hislopi extend back well into the 
 Carboniferous, as also appears to be the case in New South Wales. 
 
 67 Bodenbender (i9ii),p. 83. 
 
 68 Kurtz (18950); Bodenbender (1904), (1911). 
 
 69 Kurtz (1921-22). 
 
40 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 Of the undoubted “northern” forms, Asterocalamites scrobicu- 
 latus (Schloth), described originally in 1894-95 by Kurtz as Archx- 
 ocalamites radiatus , is found in Europe and North America, being 
 sometimes known as Bornia scrobiculata Stern, extending up into 
 the lower part of the upper Carboniferous (Pennsylvanian), but has 
 been recorded by Walkom under the name of Archxocalamites scro- 
 biculatus from Queensland, together with Glossopteris and Cordaites, 
 the only other country in which this remarkable association has yet 
 been recorded. Adiantites is a genus with a similar range, but from 
 Kurtz’s meager description thereof and the absence of a figure, direct 
 comparison with the Culm species A . antiquus (Ett) is not possible. 
 Archxopteris argentinx Kurtz is compared by him with A . wilkinsoni 
 Feist, from New South Wales, while with Archxopteris obtusa Lesq., 
 from the Pennsylvanian, Kurtz compares his Bergiopteris insignis. 
 Botrychiopsis weissiana Kurtz, with Cardiopteris- like pinnules, 
 appears to be intermediate between Cardiopteris and Neuropteris. 
 
 Cardiopteris polymorpha (Goepp) is a lower Carboniferous form 
 and does not seem to be known from the upper division; an undeter¬ 
 mined species comes from the Kuttung series (middle Carboniferous) 
 of New South Wales. C. elegans Kurtz stands close to C. polymorpha . 
 Rhacopteris inequilatera (Goepp) has been referred by Dun to 
 Aneimites ( Otopteris ) ovata (McCoy), with which Kurtz has placed 
 R. inequilatera var. ovata from Jejenes, including also A. austrina 
 Eth., the last from the Drummond series of Queensland, the first 
 named from the Kuttung series of New South Wales, which contains 
 other species of Rhacopteris. Lepidodendron veltheimianum is a lower 
 Carboniferous species, while L. pedroanum occurs in the Bonito group 
 of Brazil. The Sphenopterids may be upper Carboniferous forms, 
 but only one has been figured . 60 
 
 It might be remarked that the only records of such Palaeozoic 
 plants are from Lake Titicaca in Bolivia and from Paracas on the 
 coast of Peru by Seward 61 and Berry . 62 No specific names are given 
 by the former, but his Sphenopteris sp. recalls, however, S. boden- 
 benderi Kurtz, but appears to agree better with the description given 
 of S. fonsecx Kurtz, from Carpinteria, though no figure of the latter 
 has been published. Berry, however, has identified among others 
 Lepidodendron rimosum Stem, L. obovatum Stern, Catamites suckowii 
 Brong., Eremopteris whitei Berry, and E. peruiana Berry. He regards 
 the flora as probably Westphalian (middle Carboniferous) and points 
 out that on Lake Titicaca at least one of these forms occurs in strata 
 underlying the marine upper Carboniferous (Uralian). 
 
 In spite of some uncertainties in the identifications and the 
 necessity of revision, there can be no doubt that the northern ele- 
 
 60 Kurtz (1921-22), PI. XIV, Nos. 138, 138a, and 139. 62 Berry (1922). 
 
 61 Seward (1922). 
 
ARGENTINA AND BOLIVIA 
 
 41 
 
 ments in this western Argentine flora definitely place it on about the 
 border line of the middle and upper Carboniferous. So far back as 1895 
 Kurtz correctly pointed out this fact and also drew attention to the 
 parallelism with the plant-bearing Carboniferous beds of New 
 South Wales and Queensland. The discovery of associated glacials 
 both in San Juan and in the Seaham area of New South Wales 
 strengthens this comparison enormously. The floras of the lowermost 
 Paganzo and of the Kuttung 63 series of New South Wales have strong 
 resemblances, and in each case, and also in Peru, there is the com¬ 
 plete absence of species of the common Neuropteris , Pecopteris, and 
 Alethopteris groups, all of which become very abundant in the typical 
 middle and upper Carboniferous floras in various parts of the world. 
 The presence of the Glossopteris flora and its absence from the Kut¬ 
 tung, so far as we know, suggest an age slightly younger than that 
 of the latter for the lowermost beds in Argentina, which would 
 place them at about the beginning of the upper Carboniferous 
 (Pennsylvanian). 
 
 The flora of the beds situated to the east of San Juan, with its 
 paucity of Northern forms, can be compared with that of the lower 
 marine series of New South Wales that immediately follow the 
 Kuttung, or the Queensland ‘ ‘ Permo-Carboniferous ’ ’ with species of 
 Glossopteris, Sphenophyllum, Cordaites, and Sphenopteris. Just as in 
 the Hunter River area glacial conditions recurred, so was the case in 
 San Juan, where the basal glacials and lowermost plant-bearing 
 group were succeeded by more morainic matter, all, however, still of 
 Carboniferous age. 
 
 This strong parallel between Argentina and the eastern side of 
 Australia must be emphasized, since under the displacement hy¬ 
 pothesis the space separating these two regions would at this epoch 
 have been enormously less. This question wx>uld, consequently, 
 appear to be worthy of examination from other viewpoints in the 
 light of this hypothesis. 
 
 (3) The Name “Paganzo System” 
 
 Thus far attention has been focussed almost wholly upon the 
 basal portion of the system as developed in that section of the pre- 
 Cordillera from Mendoza northward into La Rioja with the object, 
 firstly, of bringing out the evidence for the existence of glacial 
 deposits, as in other parts of Gondwanaland, and, secondly, of show¬ 
 ing that certain of the scattered florules are actually sandwiched 
 between glacial horizons. Before reviewing the information available 
 regarding these and the still higher strata that are so extensively 
 developed in the region to the east and northeast—not only in the 
 pre-Cordillera, but in the central or Pampean ranges as well—some 
 
 63 Siissmilch and David (1920). 
 
42 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 discussion becomes necessary upon the limits to be set to the 
 “Paganzo beds. ” 
 
 This name was given by Bodenbender 64 from the village of Pagan- 
 zo to the south-southwest of La Rioja, where the several stages or 
 divisions of this important group of strata are so well developed, but 
 from which the so-called “Rhastic ” beds were definitely excluded. It 
 is true that Keidel 65 subsequently adopted a new classification, the 
 ‘ ‘ La Rioja system, ” a course in which he was doubtless influenced by 
 his separation of the basement strata in San Juan into the two series 
 of “ Jejenes ” and “Zonda. ” Now that it has been shown that such a 
 division is scarcely necessary, the term “La Rioja system” becomes 
 superfluous and there would be no genuine grounds for superseding 
 the original and well-established term “Paganzo, ” which has so firm 
 a place in geological literature. On the other hand, the exclusion of 
 the “Rhaetic” from the Paganzo is wholly indefensible, since, (a) 
 these younger beds generally include some strata of upper Triassic 
 age similar in many ways to the older groups of the Gondwana; ( b) 
 they are non-marine in character; (c) although they are generally 
 unconformable to the lower stages of the Paganzo, places are known 
 where no marked break has yet been recorded, or where as yet no 
 inferior limits can be set to the group; and (d) they admittedly cor¬ 
 respond with the upper part of the Santa Catherina system of Brazil, 
 with the Stormberg series of South Africa and with a part of the 
 middle Gondwana of the type region of India. 
 
 As will be realized from the study of the following pages, it would 
 be difficult to make out any adequate case for the exclusion of these 
 younger beds, excepting for their general unconformable position, 
 which relationship, it must be pointed out, is certainly present some¬ 
 where in each of the other sections of Gondwanaland. Since the 
 strata thus embraced range from Carboniferous to Rhaetic in age, it 
 is furthermore clear that a mere “series” would be insufficient to 
 include them—nothing less, indeed, than a “system,” so the Paganzo 
 must be given a rank equal to the systems of the Santa Catherina, 
 Karroo, or Gondwana, of which it is the Argentine representative. 
 
 In the following pages, therefore, though contrary to practice in 
 Argentina, the so-called “ Rhaetic ” will he regarded as forming the 
 uppermost member , i.e. y Stage IV> of the Paganzo system. 
 
 (4) The Distribution of the Paganzo 
 
 Deposited unconformably upon Palaeozoic and crystalline rocks, 
 and largely covered by Cretaceous and Tertiaries, these beds 
 have been involved in the Andine movements, wherefore dips are 
 often high, inversion may occur, while faulting and thrusting are 
 common. In consequence, the beds now make their appearance along 
 
 65 Keidel (1922), p. 267. 
 
 64 Bodenbender (1911), p. 47. 
 
ARGENTINA AND BOLIVIA 
 
 43 
 
 the flanks of the various ranges, occupying a position intermediate 
 between their Palaeozoic or crystalline cores (which generally compose 
 the mountain mass) and the Cretaceo-Tertiary (which commonly 
 form the lower ground), and therefore crop out most usually in the 
 form of narrow belts, not infrequently obscured over considerable 
 stretches by late Tertiary beds or by Quaternary gravels and sands. 
 The clarity of the exposures, due to the broken nature of the ground 
 and the aridity of these regions, is consequently offset by the fre¬ 
 quency with which the strata are hidden by these younger formations; 
 rarely is any wide area to be found occupied by the Paganzo, though 
 the system doubtless underlies many of the Tertiary-filled valleys 
 or intermontane plains. 
 
 The principal occurrences, as described by Brackebusch, 66 
 Stappenbeck, 67 and particularly by Bodenbender, 68 are to be found 
 between the extremity of the Gerro Villa Union and the Sierra del 
 Valle Fertil, on the western and southern sides of the Sierra de Vilgo 
 and Sierra de Sanogasta, at the southern end of the Sierra de la 
 Huerta, around the greater part of the edge of the Sierra de Chepes and 
 Sierra de los Llanos and in small areas in the Sierra Brava, Sierra de 
 Cordoba, and at Bajo de Velis in the Sierra de San Luis. To the north¬ 
 west they are represented in the Sierra de Umango and beyond in 
 the Puna de Atacama up to the border of Bolivia, while to the north¬ 
 east they reappear in Tucuman, Jujuy, and Oran, and thence have a 
 lengthy distribution far to the north, as will be set forth in section G. 
 Bore-holes in the northern and northeastern parts of Argentina sug¬ 
 gest that they may occupy a considerable tract beneath the thick 
 covering of Neogene, possibly even joining up underground with the 
 known extension of the Santa Catherina system into the territory 
 west of the lower Parana River. 
 
 (5) The Stratigraphy of the Paganzo 
 
 The system can be divided into four “pisos ” or stages , the first of 
 upper Carboniferous, the second and third of Permian, and the 
 fourth of Trias-Rhsetic age, though extending probably into the 
 Liassic, with the very uppermost portion thereof regarded tentatively 
 as constituting stage V. 
 
 Stage I —This generally ranges between 150 and 300 meters in 
 thickness, though probably well exceeding that value in San Juan, 
 where, as remarked above, lower horizons would seem to be present. 
 On the other hand, in certain places it is thin or even absent, owing 
 to the transgression of higher beds, occasionally even those of stage 
 II. Thus in the valley of the Rio Nogues to the east of Cerro Villa 
 
 66 Brackebusch (1891) and geological map (1892). 
 
 67 Stappenbeck (1913, I9 1 ?, 1920)- 
 
 68 Bodenbender (1897, 1902, 1911, 1912, 1916). 
 
44 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 Union the upper conglomerate, lying just below Stage II, rests on the 
 crystalline schists, while farther to the east, in the neighborhood of 
 Sanogasta (Nevado de Famatina), Stage II rests on granite or 
 quartz-porphyry. Furthermore, in the Puna de Atacama, Penck 69 
 has observed the red sandstone of Stage III reposing directly upon 
 granite in the Cerro Negro de los Andes (west of Tinogasta) and 
 Cerro Palca, more to the north. 
 
 In the extreme west there are developed at the base of the system 
 the glacials detailed earlier, having an intercalated zone with plants 
 of middle and upper Carboniferous genera and species, but, while 
 stratified conglomerates (particularly well developed in the Sierra 
 de los Llanos for example), arkoses, and gray, yellowish, and greenish 
 sandstones occur in this inferior position toward the east, Boden- 
 bender has stated that he had observed no evidences of glaciation in 
 them. Penck’s note of an unstratified boulder deposit, like a tillite, 
 close to Angulos, a little north of Famatina, indicates the possible 
 existence of glacial material in that quarter. At Bajo de Velis 
 glacials are absent at the bottom of the system. 
 
 Not improbably the prevailing conditions are rather like those 
 that have been made out in South Africa in passing from Natal to 
 the Transvaal, where, owing to the uneven nature of the (glaciated) 
 floor and to a temporary “post-glacial” exposure of the latter, coarse 
 fluviatile sediments have come to overlap on to, and rest upon, the 
 glaciated land surface, and the morainic deposit is only to be found 
 preserved in the deeper hollows in the latter. 
 
 Above this coarse basal phase in Argentina is a conspicuous zone 
 of carbonaceous shales with thin coaly streaks and layers, sometimes 
 pyritic in its nature. It is followed by sandstones (as in the Cerro 
 Villa Union), by conglomerates with porphyry pebbles, and by means 
 of a hard white sandstone passes up by degrees and quite conformably 
 into Stage II. Some calcareous and siliceous concretions are to be 
 found and also thin beds of limestone in the Sierra de los Llanos. 
 The carbonaceous zone alluded to is, as Bodenbender 70 has remarked, 
 a constant feature, and has yielded, as already stated,the Glossopteris 
 flora at a number of points, e.g., Sierra de los Llanos and Bajo de 
 Velis (Plate VIII, B), an association strikingly like that of the Bonito 
 series of Brazil succeeding the Itarare glacial series of that country, 
 and which, from external evidence very largely, we have to place at 
 the top of the Carboniferous. 
 
 Stage II —Not less than 300 and probably up to 500 meters in 
 thickness, it is distinguished from the older group by the brilliant 
 coloring of the beds—rose, red, gray, white, etc.—by clayey, cal¬ 
 careous, and siliceous strata, fine-grained red and partly white sand¬ 
 stones, thin quartzites, and bands of limestone up to some meters 
 
 69 Penck (1920), p. 134. 70 Bodenbender (1911), p. 64. 
 
ARGENTINA AND BOLIVIA 
 
 45 
 
 thick, and silicified marls. The calcareous matter may occur in the 
 form of concretions and lenses and is often dolomitic; partial silicifi- 
 cation has caused the strata (particularly in the middle of the group) 
 to resemble on weathering mottled felsites, while oolitic structure is 
 common; especially is this the case in La Rioja, though this phase is 
 not known in Catamarca. 
 
 This group is not fossiliferous, but in its lithological characters 
 recalls the divisions of the Iraty and Estrada Nova (Passa Dois) 
 series of Brazil and Uruguay, more particularly as developed in the 
 latter country. Its age is probably lower Permian. 
 
 Stage III —Transitional from the foregoing and from 300 to 500 
 meters in thickness, come quartzitic sandstones, dark arkoses, and 
 conglomerates (particularly towards the top), while the softer beds 
 are darker and there is less calcareous matter present, but lime¬ 
 stone—occasionally silicified—still occurs and gypsum is sometimes 
 found, as in the Puna de Atacama. On several horizons, both in this 
 and in Stage II, have been recorded contemporaneous basaltic flows, 
 also augite-porphyrite and occasionally tuff, for example, in the 
 Troya Valley (Puna de Atacama), Cerro Villa Union, Cerro Bola, 
 and at and to the north of Paganzo—forerunners obviously of the 
 marked volcanicity of the late Triassic. 
 
 Unusual interest attaches to its development in the Puna de 
 Atacama, where, according to Penck, 71 a thickness of over 900 meters 
 is represented with some melaphyres at the top, the red-brown sand¬ 
 stones being fine-grained and diagonally-bedded to platy, while the 
 calcareous thin-bedded, fine-grained beds carry siliceous layers and 
 concretions and gray, red, and violet “marls,” observations which 
 suggest that this succession actually includes part of the Trias- 
 Rhaetic Stage IV, and may even be in part equivalent to the Estrada 
 Nova and Rio do Rasto of Brazil. In a remarkable facies arising 
 through alteration by granite, these reddish beds have been con¬ 
 verted into gray, yellow, red, and violet quartzites containing a fos¬ 
 siliferous zone of white porcellanous shales in the lofty Tolarcito 
 Pass (lat. 27 0 15'), which has yielded a scanty marine fauna (Ana- 
 plophora , Avicula, etc.), indicating a Triassic age. 72 These are the 
 only fossils known from this stage, and should be compared with 
 those from the Estrada Nova (see page 150). 
 
 Stage IV, the Trias-Rhxtic —The so-called “Rhaetic beds”—as 
 usually indicated by the presence of the ‘ ‘ Thinnfeldia flora ’ ’—out¬ 
 crop in quite a number of places between Mendoza in the south and 
 Famatina in the north and between the Andes in the west and perhaps 
 Cordoba in the east, though occupying a more restricted area than 
 the older stages. In the northern half of Patagonia, however, their 
 equivalents have been found to make a wide distribution. 
 
 71 Penck (1920), pp. 132-133. 72 Penck (1920), p. 151. 
 
46 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 As mentioned already, the late Permian or early Triassic earth 
 movements (Gondwanides) led to the folding and to the subsequent 
 erosion of the lower stages of the Paganzo, and so enabled the suc¬ 
 ceeding Mesozoic formations to transgress across the edges of the 
 latter. At quite a number of places the unconformable relationship of 
 the younger beds can be studied, but in the northwest, in the but 
 little disturbed tract between the Sierra de Paganzo and the Cerro 
 Villa Union, there is reported to be an apparently unbroken succes¬ 
 sion from Stage III into Stage IV, and from the latter in turn into 
 beds that may just possibly be Jurassic, 73 tentatively referred to as 
 Stage V. 
 
 This region embraces by far the largest continuous area in the 
 pre-Cordillera, the other widely scattered localities being those at 
 Barreal and Jarillal in the valley of the Rio de los Patos west of San 
 Juan; the important elevated mass of the Paramillo overlooking the 
 Uspallata Valley; Salagasta, San Isidro, Challao, Potrerillos, and 
 Cacheuta among the foot-hills of the Mendoza district; and Marayes 
 on the southwest side of the Sierra de la Huerta; but not in the Sierra 
 de los Llanos or Sierra de San Luis. This stage is only doubtfully 
 present in the Sierra de Umango and is not known in the Nevado de 
 Famatina, though it has its equivalent farther to the northeast and 
 north in the districts of Tucuman, Jujuy, and onward into Bolivia, 
 to which reference will be made below. In the east, beds regarded as 
 Triassic are known in various parts of the Sierra de Cordoba, while 
 the deep bore-hole at Alhuampa on the Central Norte Railway in 
 Santiago del Estero proved nearly flat-lying Trias-Rhaetic (?) beds 
 from the depth of about 1,340 meters downward, having Permian 
 most probably (with Schizoneura and Phyllotheca ) below 1,700 meters 
 down to 2,111 meters—a link with the strata forming the vast 
 Triassic plateau of Uruguay and Paraguay. 
 
 In the main region between Cerro Villa Union and the Sierra de 
 Paganzo, Bodenbender has recorded this stage as extending along 
 the eastern side of the Cerro Bola, dipping eastward and covering 
 the wide stretch made by the Cerro Totorilillo, wrapping around 
 the northern end of Sierra del Valle Fertil, reaching almost to the 
 Sierra del Cerro Blanco, and occurring again on the eastern side of that 
 range of crystallines, where, at El Molle, it reposes upon conglomer¬ 
 ates taken, though not without some doubt, as the top of Stage III; 
 the same relationship is seen in the Cerro Morado a little to the west 
 and again on the eastern side of the Cerro Bola. A thickness of 
 between 400 and 500 meters is represented, consisting of gray arkoses 
 and yellow sandstones, from coarse to fine, with intercalations of 
 carbonaceous shales, carrying plant remains, chiefly Thinnfeldia, 
 and a seam of coal about 1 meter thick on the western side of the 
 
 73 Bodenbender (1911), p. 124. 
 
ARGENTINA AND BOLIVIA 
 
 47 
 
 Cerro Morado and impure seams at Marayes farther to the south. 
 In addition, intercalated flows of olivine-basalt are known on at 
 least two horizons, such volcanic rocks crowning some of the ridges, 
 such as the Cerro Morado. These beds are most probably of Upper 
 Triassic age. 
 
 The group is followed with complete conformity by Stage V, some 
 400 to 500 meters of strata, including marls with many calcareous 
 nodules and sometimes of concretions of barite, by sandstones, and 
 by whitish-yellow or grayish, white, red, and green clays, passing up 
 into light sandstone, and by brilliant red sandstones and clays with 
 gypsum. Trunks of Araucarian type have been recorded from a 
 point east of the Sierra de Cerro Blanco. Bodenbender has proposed a 
 Jurassic age for the lower part of this division and a Cretaceous one 
 for the upper part, but there seem to be no valid reasons why these 
 strata should not be Rhaeto-Lias, as is suggested by the sections 
 near Barreal. 
 
 At Hilario, a little to the north of that village, I found Stage IV 
 resting unconformably on tilted lower Palaeozoics, its base being 
 crowded with chips of green slates and greywackes derived there¬ 
 from, followed by sandstones with cherty layers and dark fissile 
 cherts, white shales, thin layers of volcanic ash, grits, and some con¬ 
 glomerates. In the Quebrada del Jarillal, a few kilometers to the 
 south, higher beds are finely exposed (Plate IX, A), with soft vari¬ 
 egated shales and dark shales carrying numerous well-preserved 
 plants ( Thinnfeldia , Txniopteris, Baiera , Cladophlebis, Phcenicopsis , 
 etc., and silicified wood, probably Dadoxylon ) and some torbanite- 
 like bands. Similar soft gray beds appear in the Quebrada del Salto 
 due east of Barreal, wherein, below a band of gray clays with Thinn¬ 
 feldia , were found some lenses of a green rock with what seemed to be 
 traces of marine fossils, though not identifiable. Both here and at the 
 Quebrada del Jarillal these plant-bearing beds are followed by vivid 
 red conglomerates and sandstones (Stage V, Plate IX, B), which, 
 though regarded by Stappenbeck 74 as Cretaceous, appear to be quite 
 conformable, there being an actual interlamination of pink and gray 
 clays with small-pebble conglomerate about the junction, a character 
 to be found duplicated in the Mendoza district. 
 
 Time only permitted of a single rapid traverse across the high- 
 lying Paramillo region, described originally by Stelzner 75 and Ave- 
 Lallemant 76 and briefly by Stappenbeck. 77 On the east, resting upon 
 tilted Devonian (?) slates, one finds gently inclined hard sandstones, 
 arkoses, and greenish shales in which Keidel 78 has discovered Cyzicus 
 (. Estheria ), like the beds characterizing the middle of the group at 
 
 74 Stappenbeck (1910), pp. 56, 66. 77 Stappenbeck (1910), pp. 60-61. 
 
 75 Stelzner (1878). 78 Verbal communication. 
 
 76 Av6-Lallemant (1890). 
 
48 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 Challao, overlain by andesitic lavas, and followed by similar sedi¬ 
 ments to a total thickness of several hundred meters. Conspicuous 
 is a splendid section along the main road, displaying a large number 
 of coniferous tree-trunks (probably Dadoxylon ), standing upright 
 (Plate X, A), rooted in a sandstone bed and surrounded by drab- 
 colored tuff, above which comes a sandstone which the stems do not 
 penetrate. This horizon was noticed by Ave-Lallamant and termed 
 by him “the sandstone with Araucarias. ” On the water-shed these 
 beds give place to a group of thin sandstones, dark shales, and 
 bituminous shales with tuffs and basaltic intercalations to a thickness 
 of perhaps 120 or 150 meters. The bituminous shales occur as thin 
 bands, weathering with a bluish-gray or white exterior and brownish 
 to blackish interior, giving out a wooden sound when struck with the 
 hammer; they are clearly allied to the torbanites, and are crowded 
 with the carapaces of Estheria and contain occasional scales of 
 Semionotus. Some thin limestones and cherts have been recorded 
 from these beds by Stappenbeck, while plants have been obtained at 
 a number of points, including Thinnfeldia, Tasniopteris, Baiera , etc. 
 
 That the basaltic layers are extrusive, though generally quite 
 compact, is shown by the presence of small vesicles either at top or 
 base, or in some instances by a zone of “pipe-amygdaloid” at the 
 bottom, and by the fact that the shales may rest in the unaltered 
 state on a slightly uneven surface of the igneous rock. In addition 
 to these lavas, intrusive sills of andesite cut these beds. 
 
 In the Cerro Colorado, lying to the west, the plant-bearing shales 
 and white and gray tufaceous sandstones pass up into red and green 
 clays and false-bedded red sandstones with clays, pierced and over- 
 lain by intrusive andesite, these strata being viewed by Stappen¬ 
 beck as Cretaceous, but being more probably of Rhsetic or Liassic 
 age. The total thickness of the sediments involved in the Para- 
 millo ascribable to Stages IV and V may perhaps possess a value of 
 1,000 meters. 
 
 At Challao, 7 km. to the northwest of Mendoza, the succession, 
 which I estimate to be not less than 450 meters in thickness, begins 
 with a massive basal conglomerate, deep red in color, resting on 
 tilted Palaeozoics, passing up through red sandstone and pale sand¬ 
 stone and greenish shales, arkoses, etc., into a group of well laminated 
 shales, banded cherty shales, dark “papery shales,” and torbanite- 
 like layers standing almost vertically, of which latter thin-bedded 
 strata some details are given by Stappenbeck, and ending in a thick 
 mass of sandstones of red or gray-violet tint, so-called “Cretaceous. ” 
 Noteworthy is the abundance of Estheria ( E . draperi and not E. 
 mangliensis ), not only in the shales and gray cherty layers, but in 
 the associated sandstone beds and carbonaceous bands. The latter 
 are much weathered and plant remains are hence difficult to obtain 
 
ARGENTINA AND BOLIVIA 
 
 49 
 
 03 
 
 > * 
 
 k 
 
 \ 
 
 
 V 
 
 x 
 
 vsv:-; 
 
 now; Cladophlebis mesozoica Kurtz has been recorded here, while 
 Keidel has discovered remains of Gyporella in a gray dolomite. 79 In 
 
 a block of yellowish sandstone from a horizon 
 about midway in the group and at least 200 
 meters below the Estheria beds—unfortunately 
 too large to be removed—I obtained on develop¬ 
 ment an impression of Glossopteris, probably G. 
 angustifolia, a long pinna of Marattiopsis , prob¬ 
 ably allied to M. muensteri, and a large equiseta- 
 ceous stem, which would indicate a Triassic age 
 for this part of the formation. 
 
 At Potrerillos, on the Transandine Railway 
 o kv "K * • between Mendoza and Uspallata, is to be found 
 
 a very similar and complete section in the Cerro 
 de las Cabras displaying fully 700 meters of beds 
 dipping southward at moderately high angles 
 (Fig. 5), wdiich, in the light of the foregoing de¬ 
 scriptions, I am inclined to interpret differently 
 to Stappenbeck. 80 The formation is underlain by 
 porphyritic and other volcanics (a) and com¬ 
 mences with deep-red conglomerates and small 
 pebble breccias some 150 meters thick ( b ), fol¬ 
 lowed by transitional beds of pink clays and 
 grits (c) into yellow sandstones, grits, gray shales 
 and mudstones (d), from the midst of which 
 Estheria and Thinnfeldia were collected at (e). 
 Conspicuous is next a red small-pebble conglom¬ 
 erate (/) and some distance above that a body 
 of 15 meters of black and gray shales with car¬ 
 bonaceous mudstones (g) yielding scales of 
 Semionotus , passing up without a break into 
 dark-maroon shales (h) with yellow sandstones 
 *'oo ^ (i), after which appear vermilion grits and small- 
 
 pebble conglomerates (j), seemingly the beds 
 regarded by Stappenbeck as Cretaceous. 
 
 There is, however, no break discernible in 
 the entire succession and the alternation of deep 
 red with the normal and often black strata, ex¬ 
 ceptionally well seen in the Cerro Bayo, would 
 show that this upper and predominantly red for¬ 
 mation must be perfectly conformable with the 
 beds below and consequently ought to be grouped 
 with the Paganzo, though it could very con- 
 
 79 Stappenbeck (1910), p. 62. 
 
 80 Stappenbeck (1917), p. 21. 
 
 CD 
 
 "T3 
 
 x 
 
 o 
 
 a> 
 
 a> 
 
 in 
 
 P 
 
 O 
 
 • rH 
 
 ctf 
 
 Q 
 cJ 
 >■ ■< 
 
 Ph 
 
 X 
 
 0 
 
 u 
 
 O 
 
 m 
 
 O 
 
 <5 
 
 u 
 
 O 
 
 Cu 
 
 a 
 
 o 
 
 *H 
 
 in 
 
 m 
 
 <V 
 
 o 
 
 o 
 
 3 
 
 in 
 
 y 1 ■ < 
 
 n 
 
 & 
 
 a> 
 
 a 
 
 a> 
 
 t 
 
 10 
 
 d 
 
 l-H 
 
50 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 veniently be considered as forming * ‘ Stage V. ” At Potrerillos it is 
 overlain with a break by the tilted Pliocene “ Calchaquenos. ” 
 
 At the southern end of the Cerro de Cacheuta, 20 km. to the 
 southeast, is the classic section at the back of the Mina de Petroleo. 81 
 Here the beds dip at 20° to the south, displaying at their base 
 porphyritic lavas and keratophyres, then 35 meters of tufaceous 
 sandstones, 15 meters of bituminous shales (lower), 20 meters of 
 sandstones, and 8 meters of thin-bedded gray shales with plants, 73 
 meters of coarse sandstones, gray conglomeratic sandstones, etc., 
 70 meters of bituminous shales (upper), while quite conformably 
 above come 130 meters of red, gray, and violet and gray sandstones 
 and shales with some gypsiferous bands (Stage V), overlain by a great 
 mass of Tertiaries. 82 Plants are best preserved in the gray shales and 
 scarcely in the highly carbonaceous bands—most abundantly 
 Thinnfeldia , Baiera , Cladophlebis, Stenopteris, etc. 83 Striking is the 
 abundance of Cyzicus (.Estheria ) draperi Jones, with which the bitum¬ 
 inous shales are crowded, and which organism has not improbably 
 been responsible for the petroleum seeping out from the outcrop of 
 the lower shales, also struck in bore-holes to the south beneath the 
 Tertiary covering, in the sandstones overlying those shales, where 
 the formation has been bent into a dome. The petroleum is un¬ 
 usually high in its content of paraffin wax. 
 
 Occurrences Outside the Pre-Cordillera —Most illuminating are 
 certain localities where the Trias-Rhaetic is present in the Cordillera 
 of Chile and Argentina and also in Patagonia. 
 
 Of particular interest is the isolated patch at La Ternera, 84 
 northeast of Copiapo (lat. 27 0 12', long. 69° 50'), where, in perfect 
 conformity beneath nearly vertical tuffs including a limestone bed 
 carrying a decidedly lower Liassic marine fauna, comes a 50-meter- 
 thick group containing conglomerates with porphyry pebbles, sand¬ 
 stones, shales with cherty bands, clay ironstone, and impure coals 
 in layers and streaks, the bottom in contact with a mass of porphy- 
 rite, apparently intrusive. From this group, which lithologically 
 shows close resemblances with the occurrences in the San Juan and 
 Mendoza districts, Solms-Laubach has described a flora with 
 Thinnfeldia , Txniopteris, etc., while Estheria also occurs, and fossil 
 wood. The horizon represented here must undoubtedly be the 
 Rhaetic, and, significantly, includes certain genera not known in 
 Argentina, such as Dictyophyllum, Lesley a } Clathropteris, Copiapxa , etc. 
 
 The presence of the upper Triassic ocean is proved to the south¬ 
 west by the finding on the Rio de Huasco of beds resting directly on 
 granite and crystalline schists (of which they carry pebbles) con- 
 
 81 Stappenbeck (1910), p. 65; (1917), pp. 22, 45. 
 
 82 From a measured section furnished me by Mr. H. J. Hawley. 
 
 83 Szajnocha (1888). 
 
 84 Steinmann and Solms-Laubach (1899). 
 
ARGENTINA AND BOLIVIA 
 
 51 
 
 taining Halobia , while farther down the coast, near the Rio Choapa, 
 Groeber 85 has discovered strata containing that genus and also Pseu- 
 domonotis , as well as older beds of Permian or more probably Car¬ 
 boniferous age carrying Productus. 
 
 Groeber 86 has summarized information on the supra-Triassic and 
 pointed out that at Puente del Inca, Rio de las Vacas, and the region 
 of Espinacito to the west of Uspallata the Callovian, Dogger, or Lias 
 overlap quartz-porphyries, porphyrites, and melaphyres belonging to 
 Stage IV (or V) which rest discordantly on possibly Devonian rocks, 
 while, as remarked earlier, they appear again in the Sierra Pintada 
 to the south and are of importance in the area between the Rio Atuel 
 and the Rio Neuquen, where they are covered by Liassic or Jurassic 
 marine beds, the “Paganzo” (with its original significance) being 
 absent; quartz-porphyries and tuffs here play a very important role. 
 Groeber regards this mainly volcanic assemblage of the Cordillera, 
 extending roughly north and south and having a thickness of up to 
 i,ooo meters, as wholly of upper Triassic age, but points out that 
 there are actually similar volcanics in the overlying Liassic as well. 
 Of immense importance is his discovery that near Las Lajas (lat. 
 39°, long. 71 0 ) on the Chilean border the first-mentioned group of 
 volcanics succeed marine middle or upper Triassic beds, which are 
 resting unconformably on the ancient platform, and that the Dogger 
 overlies these volcanics and has a basal conglomerate with pebbles 
 derived from the latter. 87 
 
 From the sub-Andine zone in Chubut (Patagonia) to the south¬ 
 east, Keidel 88 has described crystal tuffs, brilliantly colored tuffs, 
 calcareous shales, etc., crowded with Estheria draperi. Porphyrites, 
 agglomerates, and tuffs are mentioned by Rassmuss 89 at the head of 
 the Rfo Chubut, while similar beds are recorded by Windhausen 90 
 from the Rfo Chubut, the neighborhood of Sarmiento, and from the 
 Rfo Deseado, San Julian, and other places farther south in Santa 
 Cruz. Quartz-porphyries, rhyolitic rocks, and keratophyres are 
 known in various places, but it is somewhat difficult from a study of 
 the literature alone to obtain a clear idea as to the precise relation of 
 these effusions to the pyroclastic rocks, more especially as Rassmuss 
 remarks that bodies of intrusive granite and grano-diorite, probably 
 of Jurassic age, occur in the upper part of the Rfo Chubut. It is 
 certain, however, that this assemblage, with a maximum recorded 
 thickness on the coast of over 900 meters, rests upon the crystalline 
 basement and has been bent into moderate folds along arcs trending 
 roughly south-southeast to southeast as the result of movements 
 during the mid-Cretaceous, and are overlain unconformably by 
 
 “Groeber (1922). “Keidel (1917-18). 
 
 “Groeber (1918), pp. 49-59. 89 Rassmuss (1922). 
 
 87 Verbal communication. 90 Windhausen (1921, 1924). 
 
52 GEOLOGICAL COMPARISON OP SOUTH AMERICA WITH SOUTH AFRICA 
 
 almost undisturbed upper Cretaceous, Tertiary, or Quaternary; 
 they moreover show out at intervals over a great area stretching 
 southward into Terra del Fuego. 
 
 The presence of Cyzicus (.Estheria ) draperi and of silicified w r ood 
 enables comparison to be made with the so-called “Rhaetic” of the 
 Mendoza district. In the latter are to be found the beginnings of the 
 acid effusions and fragmental rocks that have such a development to 
 the south and southeast, though not represented in other parts of 
 South America. In South Africa, however, are the comparable and 
 immense acid outpourings of the Lebombo on the southeastern side 
 of that continent, with certain intrusive granophyric, felsitic, and 
 dioritic masses, all of Stormberg age (Rhaeto-Lias), while at Brisbane 
 in Queensland the base of the Triassic Ipswich series is made by a 
 bed of quartz-porphyry tuff resting on the vertical edges of schists, 
 while rhyolites, dacites and andesites are also known from this 
 neighborhood. 
 
 A matter to which attention should be drawn is the presence in 
 the southwestern corner of the Neuquen district of Neocomian beds 
 carrying molluscan fossils typical of the Uitenhage series of the 
 Cape, such as Holcostephanus atherstonei, H. rogersi, H. uitenhagensis , 
 Trigonia ventricosa, T. aff. rogersi , T. transitoria allied to T. hertzogi , 
 T. eximia allied to T. conocardiiformis, T. pusilla and other forms 
 related to T. van, Grammatodon jonesi, Exogyra imbricata , etc. 91 
 
 This locality, situated to the southwest of the arc of the 
 “ Gondwanides, ” shows therefore a very close analogy with the geol¬ 
 ogy of the Uitenhage district of the Cape, as will be discussed at 
 a later place. 
 
 The Occurrences in the Pampean Ranges —Constituting a link 
 between the west and the center are the exposures at Marayes along 
 the southern base of the Sierra de la Huerta, where, sandwiched 
 between two groups of chocolate- and maroon-colored conglomeratic 
 strata, come some 150 meters of greenish and buff sandstones with 
 thin shales and coals (the so-called “Rhaetic,” Plate X, B)\ the 
 basal group rests upon an uneven surface of the crystallines. In the 
 sections exposed at the Rickard Mine and at Carrizal there seemed 
 to me to be a conformable succession, but according to Rassmuss 
 what would appear to be an intraformational unconformity is present, 
 the ‘ ‘ Rhaetic ’ ’ transgressing across the lower group of the ‘ ‘ Paganzo ’’ 
 and coming to rest directly upon the crystallines when followed 
 toward the north. 92 The upper group of deeply colored rocks, fully 
 400 meters thick, is referred to by him as of Cretaceous age, but since 
 it recalls the succession already described in the region to the west, 
 and appears to follow conformably upon the Rhaetic within this 
 
 9X Letter written me by Dr. Pablo Groeber dated June 7, 1926. 
 
 92 Rassmuss (1922a). 
 
ARGENTINA AND BOLIVIA 
 
 53 
 
 trough, it can provisionally be considered as forming the uppermost 
 portion of the Paganzo system, Stage V. The assemblages can hence 
 be placed in the Trias-Lias succession. The coals at the Rickard 
 Mine and at Carrizal are very thin, high in ash, and extremely pyritic. 
 In addition to calcified coniferous stems, the middle group has 
 yielded the Thinnfeldia flora, this being the most southeasterly point 
 where such has been recorded in Argentina, 93 with the probable 
 exception of the strata with Cladophlebis mesozoica resting on the 
 granite in the Sierra de los Llanos. 94 
 
 Turning attention to the Central Pampean ranges, the eastern¬ 
 most outcrop of beds, that on lithological grounds have been ascribed 
 to the Triassic, are to be found in the Sierra de Cordoba, 95 forming 
 a series of patches resting upon an uneven surface of gneiss, Palaeo¬ 
 zoic (?) limestones, etc., in some places horizontally, in others tilted 
 and faulted. At Saldan, just north of Cordoba, the series consists of 
 ioo meters of red conglomerates, passing up into deep-red sandstones 
 with some red shaly layers (Plate XI, A). Ripple-marked and sun- 
 cracked surfaces are present and the sandstones are markedly cross- 
 bedded, recalling the upper Triassic seen elsewhere. In Los Terrones, 
 northeast of Capilla del Monte, are splendid sections through at least 
 200 meters of red conglomeratic beds, weathered out into striking 
 erosional forms (Plate XI, B). The gneiss surface beneath is highly 
 uneven and the junction here is not a faulted one, as maintained by 
 Rimann. 96 Some volcanic rocks overlie the strata at La Cumbre, but 
 the circumstances rather suggest an unconformable relationship and 
 a much younger age. 
 
 Provisionally these sediments can be allotted to the top of the 
 Paganzo, but, inasmuch as coals are reported to occur to the south 
 of Cordoba along the line of strike of the Saldan conglomerates, 
 plants should shortly be forthcoming to decide this question, which 
 is an important one, since these are the easternmost outcrops of 
 such beds in the republic. 
 
 The Thinnfeldia Flora —Some remarks might be made on the 
 composition of this flora and its bearing on the age of the inclosing 
 strata. 
 
 The several florules in western Argentina, occurring as they do 
 each in a group of gray or carbonaceous shales, sometimes accom¬ 
 panied by impure coals, sandwiched between groups of coarser vari¬ 
 egated sediments, represent with but little doubt practically one and 
 the same horizon situated not very far from the middle of Stage IV, 
 being followed by the red strata placed by me in Stage V, correlated 
 provisionally with the Botucatu of Brazil. 
 
 From those sections of Gondwanaland, that possess beds carry- 
 
 93 Geinitz (1876); Gothan (1925). 
 
 84 Vide Keidel (1922), pp. 288-290. 
 
 95 Bodenbender (1905). 
 
 96 Rimann (1918). 
 
54 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 ing the Thinnjeldia flora, the formation that is most suitable for age 
 determination is undoubtedly the Molteno beds of the Stormberg 
 series of the Cape, because this group fortunately happens to be 
 succeeded by strata containing vertebrate remains (saurischian and 
 pseudosuehian) of typical European Keuper and Rhaetic genera, and 
 its flora can consequently be regarded as not younger than upper¬ 
 most Keuper. 
 
 Comparisons of the South African and East Australian (especially 
 the Ipswich) assemblages with that of Argentina reveal marked 
 resemblances, which the few published descriptions of the latter fail 
 to do justice to, since many of those determinations were based on 
 mere fragments. We find as common species Thinnfeldia odontopter- 
 oides (Morr), T. lancifolia (Morr), and T. feistmanteli John. ( Cardi - 
 opteris zuberi Szaj.), while certain large fronds from Cacheuta placed 
 by Kurtz 97 under T. lancifolia are not far removed from T. narra- 
 beenensis Walkom from the Triassic Hawkesbury series of New South 
 Wales, and are also very similar to Danxopsis hughesi Feist, a form 
 known from the Stormberg, Australia, and perhaps Argentina 
 (Potrerillos). Txniopteris mareyesiaca Gein. is indistinguishable from 
 T. dunstani Walk, from Queensland; T. brackebushiana Kurtz from 
 Cacheuta has lately been identified in the Cape, while T. carruthersi 
 and T. m'clellandi (Old. and Morr), known from Queensland and 
 Southern Rhodesia, have been discovered at Barreal. Another recent 
 identification is Cladophlebis goppertiana (Schen.) both at Barreal 
 and in Natal (South Africa) the forms referred to Sphenopteris 
 elongata, and under Phcenicopsis elongata (Morr) the leaves described 
 as Podozamites elongatus , the last two being known from South Africa 
 and Australia. The form Chiropteris copiapensis Stein, et Solms, from 
 La Ternera in Chile and from Cacheuta, is also known from the 
 Molteno beds. 
 
 In addition, the widely distributed crustacean Cyzicus (. Estheria ) 
 is the same species as originally described from the Cave sandstone of 
 South Africa, E. draperi Jones, the associated smaller individuals then 
 called E. stowiana being now regarded as representing the youthful 
 stage. The fish-scales referred to Semionotus mendozaensis Gein. are, 
 moreover, comparable with those of S. capensis S.-Woodw., from the 
 Cave sandstone. 
 
 The above comparisons bring out the great similarities in the 
 floras of western Argentina, the Cape, eastern Australia, and Tas¬ 
 mania, apart from strong lithological resemblances in these beds, 
 and stage IV can accordingly be regarded as belonging with more 
 probability to the upper Triassic (Keuper) than to the Rhastic, in 
 which these beds have hitherto universally been placed. The flora of 
 the Upper Paganzo is quite different in its composition from that of 
 
 97 Kurtz (1921-22), PI. XIX. 
 
ARGENTINA AND BOLIVIA 
 
 55 
 
 the undoubtedly Liassic flora with Otozamites and Dictyophyllum 
 known close at hand in the upper part of the Rio Atuel, and differs 
 in several essentials from that at La Ternera, which, as shown by the 
 marine fossils above, is almost certainly not earlier than Rhaetic. The 
 small assemblage described by Berry 98 from far to the south in 
 Patagonia, in the Gran Bajo de San Julian associated with Estheria 
 draperi in the upper part of a thick mass of quartz-porphyry tuffs 
 containing a few coaly bands, 99 displays, on the other hand, Liassic 
 affinities. Gothan 1 has also expressed doubt as to whether the beds 
 at the Cerro Alto in Santa Cruz that have yielded the interesting 
 cone of Araucaria windhauseni are as low as the Triassic, as supposed 
 by Windhausen. 
 
 SECTION G. THE PRE-ANDINE BELT OF NORTHERN ARGENTINA 
 
 AND OF BOLIVIA 
 
 Of high importance are the recent discoveries pointing to a wide 
 distribution of the Permo-Triassic continental beds in Tucuman, 
 Salta, and Jujuy, stretching into Bolivia as far to the north as Santa 
 Cruz at least in the pre-Andine chains bordering the plain of the 
 Gran Chaco. The strata are exposed, for example, in the various 
 ranges, proceeding from south to north, of the Sierra de Lumbrera, 
 Sierra de Santa Barbara, Sierra del Alto, Sierra de Aguaragwe, Sierra 
 de Mandiyuti, Sierra de Charagua, and Sierra de Santa Cruz. 
 
 The southernmost outcrops, seen a little to the northeast of 
 Tucuman, are separated by quite a considerable stretch of more 
 ancient rocks from the patches of the equivalent Paganzo to the 
 northwest of Tinogasta in the Puna de Atacama or those of La Rioja 
 and Cordoba, and there is no actual connection between the latter 
 and the “Bermejo series, ” as this northern facies has been named by 
 Heald and Mather. 2 Nevertheless, a few representatives of the 
 Gondwana flora have been found, while the generally marked litho¬ 
 logical resemblance points to such a correlation. Further importance 
 attaches in that there may be an unbroken succession right through 
 into the Jurassic, whereas in the rest of South America almost 
 everywhere either a definite unconformity or else a pseudo-con¬ 
 formity is present at about the top of the Permian or near the base 
 of the Triassic. 
 
 The beds make their appearance in a series of narrow and often 
 faulted anticlines from beneath tilted Cretaceous and Tertiary, but 
 without the intense crumpling and the igneous injection that charac¬ 
 terizes the equivalent formations to the south and southwest, in which 
 structures at a number of points from Salta northward fossiliferous 
 
 98 Berry (1924). 
 "Wichmann (1922). 
 
 1 Gothan (1925). 
 
 2 Heald and Mather (1922); Mather (1922). 
 
56 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 shaly beds, of Devonian age with a typical austral marine fauna, 
 form the cores. These ranges constitute the eastern Cordillera. 
 
 The upper limit of the Bermejo series, through a distance of many 
 degrees, right up to the northwest and west of Potosi in fact, is 
 formed by the relatively thin calcareous member with marine 
 fossils—the “Limestones and dolomites with Melania ” of D’Orbigny 3 
 or the “Horizonte Calcareo-Dolomitico” of Bonarelli, 4 also called by 
 Heald and Mather the “Viticua limestone” from the range of that 
 name (lat. 20°. 15' S.) and the “Cajones limestone” in the Santa 
 Cruz area. 
 
 This valuable datum, from 5 to 40 meters thick, consists of white, 
 pink, gray, or bluish hard limestones and dolomites in very regular 
 thin layers, together with chert in layers or lenses—sometimes 
 oolitic—the rocks being much silicified in places. They pass upward 
 into a group of variegated Keuper-like marls with thin calcareous 
 intercalations, also fossiliferous, up to 150 meters in thickness in the 
 south. These various limestones are characterized by the presence 
 of Chemnitzia {Melania) potosiensis (D’Orb.), and have yielded a con¬ 
 siderable fauna, mainly gasteropods, which Bonarelli has shown to 
 point to an uppermost Triassic or Liassic age. 5 The sandstones over- 
 lying the marls are apparently unconformable and most probably of 
 Cretaceous age. 
 
 Perfectly conformable beneath the limestone come the uppermost 
 sandstones of the Bermejo series—the “lower sandstones” (arenis- 
 cas inferiores) of Bonarelli—but in the Puna de Jujuy the latter thin 
 out and the limestone, according to Bonarelli, come to rest either 
 directly or with a thin basal conglomerate on the Ordovician, the 
 Devonian being absent. 6 In the Sierra de Zapla, close to Jujuy, the 
 Bermejo is only some 400 to 500 meters thick, but more to the north 
 it ranges generally from 1,300 to 1,500 meters, attaining its maximum 
 of 2,500 meters in the Sierra del Alto (about lat. 22 0 S.). While the 
 series usually rests with unconformity upon the Devonian, in certain 
 places no marked discordance has been detected between it and that 
 formation below. 7 
 
 Long ago D’Orbigny had observed that over a wide tract to the 
 northeast of the area under consideration, in the neighborhood of 
 Lake Titicaca, the Devonian was followed by marine Carboniferous, 
 which has since been proved over an extensive region around Cocha¬ 
 bamba and also to the southwest of Santa Cruz on the “alta planicie ” 
 or high plateau of Bolivia. The fauna, including Productus cora 
 D’Orb., P. nebraskensis Owen, P. clarkianus Derby, Spirifer condor 
 D’Orb., Spiriferina cristata Schloth, Fenestella retiformis Schloth, is 
 that of the Uralian of the Urals, Salt Range of India, and Guada- 
 
 3 D’Orbigny (1842). 6 Bonarelli (1921), p. 74; Jaworski (1923). 7 Meyer (1914). 
 
 4 Bonarelli (1921). *Ibid., p. 21. 
 
ARGENTINA AND BOLIVIA 
 
 57 
 
 lupian of New Mexico, and has species in common with the Carboni¬ 
 ferous of the Amazon Valley and to a less degree with that of 
 the Sahara. 
 
 On the shores of Lake Titicaca, Douglas 8 observed red and choco¬ 
 late colored sandstones and breccias following the Carboniferous 
 limestones and shales conformably, but 50 km. to the southeast, at 
 Coniri, the highest beds of the former consist of coarse reddish con¬ 
 glomerates with pebbles, some of which are of fossiliferous Carboni¬ 
 ferous limestone. These deeply colored rocks possibly represent the 
 Bermejo series. A long distance to the southeast, not far to the south¬ 
 west of Santa Cruz, the undoubted Bermejo rests discordantly on the 
 arenaceous “Materal formation,” which is believed by Heald and 
 Mather 9 to be itself unconformable to the Devonian slates, and which 
 may perhaps be the continental facies of the Titicaca Carboniferous. 
 
 Thus delimited above and below, the Bermejo is considered to 
 range from the Permian to the base of the Jurassic, though, except 
 for the mollusca at the top of the series, the only fossils yet recorded 
 consist of Equisetites and Phyllotheca from Lipeon, northwest of 
 Oran in Argentina, and some Lingulas from above the glacials in 
 the Sierra de Santa Cruz. 
 
 Its lowest division —the Oquita formation 10 —is dominantly a gray 
 sandstone followed by sandy shales and these by cream or red sand¬ 
 stones, or in the Sierra de Aguaragwe by massive white sandstones, 
 w T hile the uppermost part is composed of persistent sandy shales of 
 various colors and thin-bedded sandstones. Certain shaly bands may 
 possibly be of glacio-lacustrine origin, seen, for example, in the 
 Parapiti Gorge (lat. 20° S.). The middle division is the striking 
 Mandiyuti conglomerate, which builds many of the rugged cliffs and 
 gives rise to picturesque scenery. It consists of massively bedded 
 conglomerate and grit, together with sandstones, and varies from 
 500 meters in the northern part of the Sierra de Aguaragwe on the 
 border of Argentina to 1,000 meters in the type locality in the 
 Sierra de Mandiyuti farther to the north (lat. 20° 30' S.); individual 
 beds as much as 80 meters in thickness occur in the midst of the 
 formation at several localities. The majority of the beds display 
 brilliant shades of red and purple. 
 
 The pebbles in the conglomerates and grits are of all sizes and 
 shapes and up to nearly a meter in length, mainly of hard rocks and 
 mostly well rounded and polished by stream or current action, but in 
 every locality are to be found many angular pebbles with faceted 
 faces, and the deposit is therefore regarded by Heald and Mather as 
 a fluvio-glacial accumulation. In the presumably equivalent beds in 
 the Sierra de Santa Cruz tough clays were found carrying beautifully 
 
 8 Douglas (1914), p. 30. 10 Mather (1922), p. 739. 
 
 9 Heald and Mather (1922), p. 560. 
 
58 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 faceted grains and pebbles, while true tillite was also observed by 
 them. These faceted inclusions scattered through both the sand¬ 
 stones and shales clearly indicate the presence of streams issuing 
 from a glacier-front or from glacier-capped highlands. 
 
 Though paralleled by Heald and Mather with the (Carboniferous) 
 glacials of the San Juan-Jachal area, it must not be overlooked that 
 they may not improbably be much higher in the succession, possibly 
 well up in the Permian, and that they might even be comparable with 
 the glacials of the upper Marine series of the Hunter River area in 
 New South Wales or perhaps those of the Congo, ascribable to a 
 Triassic (or probably late Permian) age according to Ball and Shaler. 
 
 The uppermost division of the Bermejo series ranges from 250 
 to 500 meters in thickness, mainly massive soft green to maroon 
 sandstones, with non-fissile shales and clays of red color, overlain 
 conformably by the “calcareo-dolomitic horizon.” Madgwick has 
 remarked on the highly false-bedded nature of these sandstones, 
 wherefore comparison can be instituted with the Rio do Rasto series 
 of the Santa Catherina system of Brazil, the nearest occurrences of 
 which lie less than 800 km. due east on the railway from Porto 
 Esperanga to Sao Paulo. 
 
 This upper Bermejo (areniscas inferiores) appears in the Cerro 
 Colorado, west of Rosario de la Frontera in Salta, beneath the lime¬ 
 stones to a thickness of 600 meters at least without the base being 
 seen, yet in the Alto de Munoz and to the south in the Cerro de 
 Campo this group rests directly on the pre-Cambrian; clearly the 
 lower Bermejo does not reach Salta. 11 The upper Bermejo, the lime¬ 
 stones, and the marls above together make up the “formation 
 petrollfera” of Bonarelli, in which much boring has been done at the 
 base of the various chains. 
 
 That the Permo-Triassic extends far to the east beneath the 
 covering of Cretaceo-Tertiary is indicated by the deep boring at 
 Alhuampa previously mentioned, in which at a depth of some 1,340 
 meters there was cut a group of dark-reddish micaceous sandstones 
 and clays with occasional conglomerates, down to about 1,700 meters 
 possibly Triassic, and below that down to 2,111 meters light feld- 
 spathic and micaceous sandstones and grits with dark micaceous 
 shales, which at 1,860 meters yielded impressions of Schizoneura and 
 Phyllotheca, indicating some stage of the Paganzo. It should be 
 noticed that the volcanic rocks, so well represented in the Parana 
 Basin and found in borings just east of the Parana River, and also the 
 flows associated with Stage IV in La Rioja, are absent from the 
 northern belt just described. 
 
 11 Stappenbeck (1921). 
 
CHAPTER IV 
 
 PRE-GONDWANA BASEMENT OF EASTERN SOUTH AMERICA 
 
 SECTION H. THE DEVONIAN SYSTEM 
 
 In all Brazil perhaps the most striking parallel with the Cape is 
 afforded by the central portion of the State of Parana with its 
 Devono-Gondwana succession. 12 Conspicuous here is the nearly 
 horizontal Furnas sandstone, making a belt curving from Faxina in 
 the north past Pirahy, Castro, and Ponta Grossa to Serrinha with an 
 outcrop of fully 300 km. in length, fading out at either end beneath 
 the Itarare glacial beds. The escarpment that it makes picturesque 
 because of its fine araucarias, is known in the north as the Serra das 
 Furnas and in the south as the Serra de Sant’ Anna. 
 
 The formation, which is perhaps just over 150 meters thick and 
 marvelously like the Table Mountain sandstone of the Cape western 
 Province, rests on an even surface cut equally across granite and 
 the tilted Assunguy series, the beds ranging from a white fine-grained 
 sandstone to a coarse white, false-bedded grit (Plate XII, A), con¬ 
 taining a little kaolinized feldspar and tiny quartz pebbles, the rocks 
 giving rise to a coarse sandy soil and to rather treeless plains. Either 
 isolated in the sandstone or occurring in thin washes are pebbles 
 up to about 7 cm. in length, smooth and oval to irregular in outline 
 chiefly made of white or flesh-colored vein-quartz, together with some 
 of quartzites, and occasionally of cherty rocks. At Ponta Grossa I 
 observed none of jaspers, such as are to be found in the Clan william 
 district of the Cape, but Doctor Ferraz 13 has ascribed to the Devonian 
 certain coarse sandstones with red jasper pebbles that cap the Morro 
 Bahu to the southeast (860 meters) not far from the coast a little to 
 the west of Itajahy. According to Leme, 14 diamonds have been found 
 in the sandstones near Tibagy, a feature that has not yet been 
 reported from the western Cape. 
 
 The concordantly following fossiliferous Ponta Grossa shales 
 appear for a distance of over 200 km. between the Furnas sandstone 
 and the overlapping Itarare glacials; from the descriptions available 
 the highest zones are represented in the central part of the great curve 
 in the neighborhood of Tibagy. At Ponta Grossa, where nearly 100 
 meters of these shales are exposed dipping gently beneath the Itarare, 
 the strata recall down to the minutest detail the lowermost shale 
 group of the Bokkeveld series of the Cape western province ranging 
 
 12 Clarke (1913), pp. 36-41; Woodworth (1912), p. 42; Oliveira (1925); Branner 
 (1915). PP- 311-312; (1919), P- 288; Leme (1924), pp. 159-162. 
 
 13 Ferraz (1921). 
 
 59 
 
 14 Leme (1924), p. 159. 
 
60 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 from fine carbonaceous shales with pyritic nodules through gray and 
 yellowish-weathering shales to soft, yellowish, sandy, and somewhat 
 micaceous beds; limestones are, furthermore, absent, though hard, 
 slightly calcareous nodules containing Lingulas are common. This 
 horizon has yielded abundant fossils at Ponta Grossa, Tibagy, 
 Jaguariahyva, etc., the palaeontological facies being closely com¬ 
 parable with the faunal assemblages of the Bokkeveld, Falkland 
 Islands, Argentina, and Bolivia, as detailed by Clarke 15 and Reed 16 
 and summarized in Chapter VI. 
 
 Above these beds, at Tibagy, and overlain by some more shales, 
 is a soft, gray feldspathie sandstone—the Tibagy Sandstone—up to 
 20 meters thick, which is characterized by casts of large Spirifers, 
 and corresponding stratigraphically with the well-known soft, red- 
 weathering “fossiliferous sandstone” of the Bokkeveld with its 
 molds of Spirifers and other mollusca, situated about 130 meters 
 above the base of that series. This is all the more remarkable, 
 because Reed 17 has just recently grouped the characteristic Tibagy 
 species, S. iheringi, under the name of S. ceres Reed var. iheringi 
 Kays, that is so abundant in the f ossilif erous sandstone of the Hex 
 River Valley and Warm Bokkeveld. 
 
 Except for the fact that the Table Mountain sandstone is thicker, 
 the lithology and succession in Parana compares almost exactly with 
 that of the Cape system in the northeast of Clanwilliam, where the 
 Dwyka tillite passes transgressively across the almost flat-lying 
 Bokkeveld and Table Mountain sandstone on to the folded pre- 
 Devonian slates and intrusive granite. 
 
 Of note is the recent discovery of such fossiliferous shales and 
 sandstones with Lingula , Orbiculoidea , etc., in Uruguay on the Rio 
 Negro at the Rincon de Alonso underlying the Itarare glacials; they 
 are nearly horizontal. 18 
 
 In Paraguay, between 60 and 80 km. to the east-northeast of 
 Asuncion, at the Cerro Aparipi, and at Arroyos y Esteros, Beder and 
 Windhausen 19 have recorded exposures of soft, medium-grained 
 micaceous sandstones and white, yellow, or red weathered shales 
 with secondary limonite; these strata lie horizontally and must well 
 exceed 70 meters in thickness. To the south they must thin out 
 beneath the Gondwana system, for in that direction the latter come 
 to rest directly upon the crystallines, but the distance to which they 
 extend toward the northeast is as yet unknown, that region being 
 geologically unexplored. Windhausen has identified among forms 
 typical of the Ponta Grossa shales, that characteristic Hamilton 
 brachiopod Tropidoleptus carinatus Conrad, known from the Erere 
 
 16 Clarke (1913). 
 
 16 Reed (1925). 
 
 17 Reed (1925), p. 52. 
 
 18 Arocena (1926), p. 16. 
 
 19 Beder and Windhausen (1918). 
 
PRE-GONDWANA BASEMENT OF EASTERN SOUTH AMERICA 
 
 61 
 
 sandstone of the Amazon Valley and from the Sicasica beds of Bolivia, 
 which accordingly suggests a mid-Devonian age and therefore a hori¬ 
 zon somewhat higher than that present to the east in Parana. 
 
 Just as on the western so on the northwestern side of the basin 
 the Devonian reappears at a point 20 km. north of the town of Rio 
 Bonito not far from Goyaz, according to C. L. Baker, 20 where shales 
 have yielded characteristic fossils. Better known, however, is the 
 plateau of the Chapada in Matto Grosso, overlooking the plain of 
 Cuyaba, wherein the Devonian rests unconformably upon tilted 
 slates and limestones of supposed Ordovician age and dips gently 
 northward beneath Triassic beds, the series extending in that direc¬ 
 tion toward Diamante and eastward to Lagoinha at least, 40 km. 
 from Sant’ Anna de Chapada. From the remarks of Evans, 21 Derby, 
 Leme, 22 and others it would appear that the great escarpment is 
 made by a group of soft red sandstones with basal conglomerates 
 followed by shaly beds with thin sandstones. Smith puts the thick¬ 
 ness of the first-named at 105 meters, with the fossiliferous shaly zone 
 in the next 50 meters, but Leme ascribes the value of 500 meters to 
 the series, remarking that the molluscan fauna occurs in a ferruginous 
 sandstone in the upper part of the group, a statement seemingly more 
 in accord with the observations of other visitors. The unfossiliferous 
 sandstones below are clearly the equivalent of the Furnas sandstone 
 and the fossiliferous beds above of the Pont a Grossa shales. 
 
 The fossils described by Derby and Von Ammon have been 
 exhaustively reviewed by Clarke, 23 who points out the great simi¬ 
 larity of the fauna to that of Parana, and to a lesser extent with that 
 of Bolivia, but contrast to the Maecuru fauna of Amazonas, and 
 remarks that the presence of the uncommon trilobite Phacops 
 braziliensis of boreal aspect would suggest that the Chapada Devon¬ 
 ian, if more closely studied, would disclose additional affinities with 
 the fauna of the Amazon Valley, which in view of its geographically 
 intermediate position is only to be expected. 
 
 A point worthy of note is the statement that the diamonds of this 
 region have been concentrated from these Palaeozoic sandstones and 
 conglomerates. Scattered observations suggest that the Chapada 
 Devonian passes to the northeast and east beneath the Triassic sand¬ 
 stones and basalts and the Cretaceous Bauru sandstones (‘'sand¬ 
 stones of the Taboleiros” of Evans), 24 and to the northwest of 
 Cuyaba beneath the Matto shales, Permian (?), while inliers to the 
 southwest in the Chiquitos of the Gran Chaco form links with the 
 Devonian of Bolivia. 
 
 An occurrence in Uruguay is worthy of mention, though ascriba- 
 
 20 Baker (1923), p. 67, 
 
 21 Evans (1894). 
 
 22 Leme (1924), p. 158. 
 
 23 Clarke (1913), pp. 41-46, p. 148. 
 
 24 Evans (1894), p. 98. 
 
62 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 ble only with doubt to the Devonian, but, if so, forming a link with 
 the strata in the Sierra de la Ventana. This is the limited outcrop 
 recorded by Walther 25 in the neighborhood of the station of Piedras 
 de Afilar on the Montevideo-Maldonado Railway, and quite near the 
 coast. Resting with presumed unconformity on the granite, though 
 this is not assured, and striking northwestward, the beds make a 
 narrow strip with a moderately high dip to the southwest, and give 
 rise to a sharp ridge. Though regarded by Walther as a faulted-down 
 outlier of the Sao Bento series, the descriptions of the rocks and the 
 appearance of the latter in hand specimens incline me after a dis¬ 
 cussion with Professor Walther to suggest a greater age for the 
 strata, which consist mainly of a hard, fine-grained, grayish to 
 faintly greenish white quartzite, together with some hard reddish 
 sandstones. While there is the possibility that the beds may repre¬ 
 sent a pre-Devonian formation, they resemble the quartzitic sand¬ 
 stones of the Sierra de la Ventana. On the assumption that such is 
 the case, it should be noted that their strike is nearly parallel to 
 the course of the “ Gondwanides ’ ’ within the region to the south¬ 
 west, and hence at right angles to the older “Brasilides” to be men¬ 
 tioned below. 
 
 SECTION I. THE EARLY PALAEOZOIC BEDS 
 
 The several terrains of early or of pre-Palaeozoic age, though 
 occupying great extents of country, are surprisingly little known, 
 and even their relative ages are under dispute. Much discrimination 
 will be required in reading Branner’s 26 account, for his extensive 
 areas classed under the Permian actually include beds ranging from 
 lower Palaeozoic to late Cretaceous. 
 
 Prominent are the various belts of pre-Devonian strata in the 
 lengthy stretch between the Rio de la Plata and Pernambuco, of 
 which some are probably of Ordovician age, while others may be 
 older. They, however, have a general lithological resemblance to the 
 folded Nama succession on the eastern side of the Atlantic, between 
 Cape Town and Luderitz, and also possess a strike that is more or 
 less parallel to the coast. This likeness is heightened by the fact 
 that in certain localities the granite by which these belts are flanked 
 is intrusive, just as in the Nama beds between Cape Town and 
 Namaqualand. 
 
 Termed many years ago the “Brazilian system” by d’Orbigny, 
 this tectonic structure is most probably'of Silurian age, extending 
 into eastern Argentina, where, by no means conspicuous, it is crossed 
 nearly at right angles by the Permo-Triassic “Gondwanides” (see 
 Fig. 7 )- 
 
 In Uruguay these beds—phyllites, crystalline limestones and 
 
 25 Walther (1919), p. 10, p. 77. 
 
 26 Branner (1919). 
 
PRE-GONDWANA BASEMENT OF EASTERN SOUTH AMERICA 
 
 63 
 
 dolomites, and quartzites—which with great probability correspond 
 with those of the Sierra de Tandil, occupy a well-marked belt com¬ 
 mencing on the coast a little to the west of Maldonado and trending 
 between north and north-northeast through the departments of 
 Minas, Treinta y Tres, and Cerro Largo, in the last-named of which 
 they pass beneath the Gondwana beds; within the belt local devi¬ 
 ations from the mean strike are not unusual. 27 Where I saw them in 
 the Arroyo Fraile Muerto, to the southwest of Melo, the calcareous 
 rocks consisted of blue-gray, fine-grained limestones and brown and 
 sometimes red-weathering dolomites with subordinate dark slates, 
 dipping at high angles, thus resembling those of the Nama system of 
 Namaqualand. No fossils have yet been found in them in Uruguay. 28 
 
 In southern Brazil the distribution of the limestones in the crys¬ 
 talline complex is all but unknown in Rio Grande do Sul, though a 
 narrow belt possesses, so far as I can gather, an easterly strike at 
 Ca£apava, but in Parana the folded slates, limestones, and quartzites 
 of the “Assunguy series,” veined and injected by granite and quartz- 
 porphyry, have again a northeasterly strike, appearing from beneath 
 the Furnas sandstone near Curitiba, near Castro at Iporanga, and 
 below Xiririca and Iguape in the Ribeira Valley, and trending 
 through Sao Roque and Jundiahy to the northwest of Sao Paulo. 
 By the Brazilian Geological Survey these beds have been placed 
 in the Ordovician. 
 
 To the northeast of the Sao Paulo border appears the similar 
 and apparently equivalent “Bambuhy series” of limestones and 
 slates, which stretches in a wide belt down the valley of the Rio Sao 
 Francisco and Rio das Velhas to about the thirtieth parallel. 29 
 Where I saw this group on the railway between Bello Horizonte and 
 Pirapora the thick gray and black limestones were followed by soft 
 micaceous shales and flagstones, the strata dipping usually at low or 
 moderate angles, though near Curvello they were folded along tw r o 
 axes directed northeast and northwest and the slates displayed a 
 cleavage and were traversed by quartz veins. 
 
 These beds were regarded by Branner as Permian, but Leme has 
 pointed out that they are probably continuous with the strata at Bom 
 Jesus de Lapa on the bank of the Rio Sao Francisco in the north, from 
 the limestones of which Derby described Favosites and Chastetes. 
 The long-reported occurrence of fossils in the limestones of the 
 Rio das Velhas and the recent statement of the finding of Spirifers 
 in limestone at Dores de Indaya, respectively north and northwest of 
 Bello Horizonte, and at Paracatu west of Pirapora, as cited by Leme, 
 all point to a Palaeozoic—probably an Ordovician or Silurian—age for 
 
 27 Walther (1919), pp. 28-38 and Plate XV. 
 
 28 01 iveira (1918), p. 8, (1925); Leme (1924), p. 138. 
 
 29 Leme (1924), pp. 139-143. 
 
64 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 this extensive formation with general north-northeasterly trend. It 
 should be noted that these foldings must be of pre-upper Carbonifer¬ 
 ous age, since the glacial beds at Areado in the Abaete Valley rest 
 undisturbed on the tilted Bambuhy series. 
 
 North of Bello Horizonte the series must rest on the granite, but 
 between Corintho and Diamantina it is in contact with and appears 
 to repose unconformably upon the ancient Espina^ or Cara^a 
 quartzites. It is true that this relationship has been called into 
 question, but to my mind the section seen in the ascent of the Serra 
 da Toccaia at Rodeador Station clearly shows the quartzites, here 
 bent into an anticline, passing beneath the gently inclined limestones. 
 
 It is possible that the strata in the Serra de Caetite and those 
 covering so wide an area farther to the northeast in Bahia, such as 
 the Tombador-Caboclos-Paraguassu series described by Branner 30 
 and Crandall, 31 resting on granite or the ancient tilted Jacobina 
 series, are the representatives of this formation. They are bent into 
 open folds, with axes usually trending north-northeast and north¬ 
 west, are overlain unconformably by the Lavras series, and this in 
 turn with probable disconformity by the Estancia beds and Salitre 
 limestones, which, as will be shown in Section M, belong to the 
 Permo-Carboniferous. 
 
 Immense interest attaches to the Lavras series of gray and pink 
 quartzites with conglomerates, inasmuch as they are most definitely 
 the source of the diamonds and carbonados of Bahia, with which the 
 similar gem-bearing quartzites of Grao Mogul in northern Minas 
 have been paralleled. If the stratigraphy of this region has cor¬ 
 rectly been interpreted in the foregoing account, then there is a 
 strong probability that the Lavras series may be of Devonian age. 
 Branner has grouped it in the Carboniferous, though without palae¬ 
 ontological evidence; Leme, on the contrary, emphasizes the sig¬ 
 nificance of the fact that elsewhere the diamond is known to occur in 
 Devonian sandstones, namely, in Parana and the Chapada. 
 
 SECTION J. THE PRE-CAMBRIAN FORMATIONS 
 
 A striking feature in the geology of Minas Geraes is the long belt 
 of somewhat metamorphosed sediments called by Derby the ‘ ‘ Minas 
 series, ” deposited, it is stated, unconformably on the granite, gneiss, 
 and crystalline schists, but complexedly faulted and thrust by move¬ 
 ments directed from the east and southeast. 
 
 Most fully developed between Bello Horizonte and Ouro Preto, 
 its basal member is the thick Cara^a or Espina^o quartzite, which 
 forms the “backbone” or watershed coursing northward through 
 Diamantina and thence northeastward toward Bahia. Above comes 
 the important “itabirite” or iron-ore formation, associated with 
 
 30 Branner (1910; 1919). 
 
 31 Crandall (1919). 
 
PRE-GONDWANA BASEMENT OF EASTERN SOUTH AMERICA 
 
 65 
 
 dolomite, slates, etc., and followed by a group of quartzites. These 
 itabirites are identical with the rocks called “banded ironstones,” 
 so marked a type in the primitive systems of South Africa. In Bahia, 
 according to Crandall, the Jacobina series is probably the repre¬ 
 sentative of these rocks of Minas. 
 
 Harder and Chamberlin 32 have regarded the Minas series as 
 Algonkian, a view which the strong local metamorphism fully justi¬ 
 fies; for instance, I observed kyanite in certain of the highest beds 
 present at the summit of the Itacolumi Peak near Ouro Preto. 
 
 Because of its alleged diamondiferous character, the Espina^o 
 quartzite has by certain geologists been correlated with the Lavras 
 quartzites of Grao Mogul and Bahia, but there are excellent reasons 
 for concluding that the former is a much older formation. The alle¬ 
 gation that the gem can actually be traced up to these quartzites in 
 the neighborhood of Diamantina is by no means secure, several high 
 authorities having wholly denied this association. It is true that 
 diamondiferous gravels rest upon a peneplain cut across the quartz¬ 
 ites round about Diamantina, but the gems may well have been 
 derived from other sources. My examination of the famous Boa 
 Vista Mine east of Diamantina showed most definitely that it was an 
 eruptive “pipe” piercing the Espina^ quartzite, the gem-bearing 
 but wholly “non-igneous” breccia being crowded with fragments 
 derived from the shattering of that formation, its somewhat schistose 
 character proving the antiquity of this peculiar phase of volcanicity; 
 not improbably its age is pre-Devonian. Draper’s 33 researches suggest 
 that several of the other diamondiferous occurrences in this part of 
 Brazil are of a similar kind. 
 
 It is hence likely that the gems in the Lavras series were in part 
 derived from the waste of the Espina^o quartzites because of these 
 diamond-bearing pipes. Again, it is a matter of more than ordinary 
 interest that in South Africa the diamond should be known as a 
 detrital mineral in at least two formations of “pre-Kimberlite” age, 
 namely, in the “bankets” of the (probably) pre-Cambrian Witwaters- 
 rand system in the Transvaal and in the upper Triassic sandstones 
 of Somabula in Southern Rhodesia. 
 
 32 Harder and Chamberlin (1915). 
 
 33 Draper (1911; 1920; 1921). 
 
CHAPTER V 
 
 GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 
 
 SECTION K. THE PARANA BASIN 
 
 This covers a region of close on a million square kilometers, with¬ 
 out taking into account certain outliers to the south and northwest, 
 and is therefore comparable in size with the Karroo “basin” in 
 South Africa, but, in contrast, the volcanic group at the top occupies 
 a relatively much greater area. 
 
 Within this immense territory the dips inward toward the north¬ 
 east-southwest axis are generally extremely slight, it being rare to 
 find inclinations of more than about 3 0 . Two feeble cross-ridgings 
 athwart the basin run northwestward through Parana and Rio 
 Grande do Sul respectively, while the base has been brought down 
 to below sea-level at Torres on the coast. In Uruguay, warping has 
 brought up the basement granite inside the basin in the Departments 
 of Riviera and Cerro Largo. On the southwest the very flat syncline 
 passes beneath the Neogene of the lower Parana along a line that 
 runs rather regularly north and south. 
 
 Stratigraphy 
 
 Termed by I. C. White 34 the “Santa Catherina system,” the 
 subdivisions and their nomenclature have been altered from time to 
 time, while a different classification is employed by the Geological 
 Commission of Sao Paulo to that adopted by the Federal Geological 
 Survey of Brazil. I furthermore discovered that in Uruguay the 
 classification used by Professor K. Walther and by the State Boring 
 Department was still practically the same as that put forward by 
 White, but that the boundaries chosen for the several subdivisions 
 seemingly did not each fall at precisely the horizons adopted by the 
 Brazilian Survey. Such differences, however, are only to be expected 
 with a number of geologists laboring in an enormous territory inde¬ 
 pendent of one another and unable to meet and discuss essentials 
 from time to time. Attention is drawn to this feature, since uncer¬ 
 tainties in the way of exact correlation had arisen that I found to be 
 due very largely to such causes and not to errors of observation 
 and record. 
 
 It is regrettable that the Brazilian Coal Commission should have 
 labored in an area in which the highly important basal glacial divi¬ 
 sion was so feebly developed, thus necessitating a revision in termin¬ 
 ology for the lower part of the system at the hands of Doctor Oliveira. 35 
 
 34 White (1906), p. 36. 35 Oliveira (1918), pp. 11-12. 
 
 66 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 
 
 67 
 
 I also find it impossible to subscribe to the general opinion that the 
 basal beds of the system are so young as Permian and am compelled 
 to place them in the upper part of the Carboniferous. The strati- 
 graphical break between the Triassic and Permian was for a long 
 time suspected and its recognition has come only after comparison 
 with other parts of Gondwanaland has indicated the presence of 
 such a gap in the succession. The determination by Doctor Reed 
 of a Triassic age for the upper part at least of the Estrada Nova 
 necessitates a radical alteration in the grouping of the beds and 
 further stratigraphical investigation is now urgently needed. 
 
 The following classification is to some extent tentative (the 
 wavy line indicates an unconformity or stratigraphical hiatus): 
 
 Sao Bento series < 
 
 Serra Geral eruptives 
 Botucatu sandstone (where separable) 
 Rio do Rasto group 
 Estrada Nova (upper part) 
 
 Rhsetic-Liassic 
 
 ► Triassic 
 
 Passa Dois series 
 
 Tuberao series 
 
 Itarare series 
 
 ) Estrada Nova (lower part) 
 k Iraty - group 
 [ Palermo group 
 l Bonito group 
 (Glacial beds) 
 
 Permian 
 
 > Upper Carboniferous 
 
 (1) The Succession at Fraile Muerto in Uruguay 
 
 Before entering into a discussion upon the stratigraphy of the 
 basin, it will be useful to give some details concerning this small 
 tract of country situated 380 km. northeast of Montevideo, on the 
 railway to Melo, because of the interesting stratigraphical points 
 that can be established there and because of the marked similarity 
 in the succession to that in San Juan in Argentina. 
 
 It was from the upper part of the Arroyo Fraile Muerto, at 
 Laguna la Tuna south of the railway that Guillemain 36 in 1911 re¬ 
 ported the presence of glacial beds. This area was examined in more 
 detail by Walther 37 and described by him in several instructive 
 writings, in which use was made of the important data arising out of 
 borings made by the Section de Perforaciones of Uruguay, mainly 
 within the stretch between the pre-Gondwana rocks forming the 
 Cerro Largo on the southeast and the railway from Fraile Muerto to 
 Melo on the northwest. 
 
 36 Guillemain (1911); also (1912). 
 
 37 Walther (1919). 
 
68 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 On the Arroyo Fraile Muerto, at the Casa de Comereio belonging 
 to Senor Gonsalez (the Paso Tla Lucia), the basement rocks are 
 schistose greywaekes evidently constituting an uneven surface, over- 
 lain by pale sandy tillite (lower) with finely striated boulders up to 
 80 cm. in diameter, which, apparently following the contours of the 
 floor, is dipping gently northeastward, succeeded by about 1.5 meters 
 of shales—hard, banded, and drab in color at the bottom, and soft 
 and blue higher up, followed with sharp junction by the upper 
 tillite, a clayey material with well-striated boulders of fine-grained 
 quartzite, granite, gneiss, quartz-porphyry, phyllite, limestone, and 
 dolomite, and this in turn by hard shales, exposed along the road 
 leading up to the rising ground toward the east. This compound 
 tillite constitutes th e first and lowest glacial horizon. 
 
 The base of the glacials is not to be seen here, but is exposed 
 about 2 km. farther downstream, just below the homestead of Paso 
 de la Cruz, the foundation rocks consisting of highly tilted and folded 
 blue-gray limestones and dolomites with dark slaty bands—probably 
 lower Palaeozoic. The floor is uneven, having been cut into hollows 
 that do not always follow the strike of the beds, but commonly the 
 direction of glaciation, and not improbably owe their origin to ice 
 action. The tillite is in places tucked in beneath such low ridges and 
 at one spot is banked up against a vertical wall of limestone, where 
 it is crammed with lumps of that rock and is furthermore calcareous 
 in its matrix and rather hard. Well-grooved surfaces can be found 
 with the striae directed north 40 0 west (true) and though “chatter- 
 marks” were not seen, it was noticed that on the northwestern side 
 of cross-joints in the pavement flakes had sometimes been dislodged, 
 leaving small gaps bounded by the vertical joint on the southeastern 
 side, which is sharp-edged, but shallowing out on the northwestern, 
 where the lip is sometimes weakly striated. This, to my mind, sug¬ 
 gests a movement of the glacier toward the northwest rather than in 
 the reverse direction, which will be observed to be in agreement with 
 that deduced in Argentina. 
 
 Whereas the rock filling the hollows is a tillite of normal charac¬ 
 ter, its matrix gets coarser on approaching the low ridges, and, when 
 overlying the latter, may become a hard, coarse grit with small 
 boulders, many of them striated, the material arching over and 
 dipping at high angles toward the tillite in the hollows. The phe¬ 
 nomena recall those to be seen in the ‘ ‘ kames * ’ of the Pleistocene glaci¬ 
 ation, wherein a glacial sand has become molded over protuberances 
 of the floor beneath the ice-mass. The glacial beds are not many 
 meters thick hereabouts, and on passing upward the irregularities 
 in stratification disappear and the material grades upward into grits 
 and sandstones that are lying flat, though sometimes false-bedded, 
 and carry only small pebbles of hard rocks. Features just like these 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 69 
 
 are well known in and have been recorded from the “northern” 
 Dwyka of South Africa. 
 
 A few kilometers to the west, along the railway at Cerro de las 
 Cuentes, the basement complex displays typical roches moutonnees 
 on which the glacials are resting, and the sections, which I was 
 unable to examine in detail, should well repay study. 
 
 Returning to the account of the sections at the Paso Tia Lucia, 
 the shales referred to as overlying the first glacial horizon are suc¬ 
 ceeded to the north by at least 15 meters of grits with very small 
 pebbles and by a bed of gritty stuff in which are set boulders up to 
 75 cm. across, well glaciated, which makes the second glacial horizon. 
 The homestead of Senor Gonsalez stands on this band, which, fol¬ 
 lowed by a conspicuous sandstone, is traceable toward the east up to 
 the main road, a distance of fully a kilometer, while, according to the 
 account given by Walther 38 a sandstone of this description carrying 
 boulders crops out between this locality and Cerro de las Cuentes 
 and plays an important part in the western end of the Cerro Largo 
 district. 
 
 At about 20 meters above this glacial band, seen at several points 
 to the east along the ridge which the main road to Fraile Muerto 
 follows and finally passing below higher strata at the Establecimiento 
 Quebrachal to the north, is the third glacial horizon , much thinner 
 and indicated only by scattered inclusions in the gritty sandstones, 
 mostly small and composed of hard rock, some of them certainly well 
 striated; lumps of silicified wood are also present in this sandstone. 
 
 Sandstones and shales follow thereon, all dipping northeastward 
 at 3 0 or 4 0 , giving a total thickness of fully 130 meters of beds from 
 the floor to a horizon where first are seen the dark sandy but 
 “papery” shales with occasional thin limestones and silicified wood 
 that continue to Fraile Muerto, being well exposed in the railway 
 cutting to the east of the station and also pierced by a bore-hole in 
 the village (No. 48) to the depth of 54 meters. 
 
 These are broadly equivalent with the Iraty group and thus 
 furnish the datum plane from which the stratigraphical succession 
 can be determined. The 130 meters of beds beneath this zone cor¬ 
 respond, therefore, with the Bonito and Itarare of Brazil, though it 
 is impracticable in Uruguay to draw a sharp line between the two 
 divisions, while even the precise base of the Iraty is a little in doubt. 
 I have taken the latter at the horizon where the highest of the sand¬ 
 stones and grits is to be found; below that level the strata, as dis¬ 
 closed by borings, consist of white, fine to coarse sandstones and 
 limonitic sandstones, dark shaly sandstones, sandy shales, and some 
 bituminous shales. All these beds that lie above the glacials will be 
 referred to in detail later on. 
 
 38 Walther (1919), p. 97. 
 
70 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 (2) The Boulder Beds in Brazil 
 
 Proceeding to the consideration of the “Parana Basin” we find 
 the Itarare series, so called by Oliveira, 39 consisting of a series of beds 
 of glacial origin, alternating with conglomerates, grits, sandstones, 
 and shales that are just like those of the succeeding Bonito or Coal 
 Measures, the first-named varying from unbedded, true ground- 
 moraine or tillite through boulder-shale to bedded shale with small 
 inclusions and even “pelodites” on the one hand, and through 
 “sandy” rocks with glacial erratics to sandstones (sometimes cross- 
 bedded), with waterworn pebbles (fluvio-glacial beds) on the other. 
 The succession, as well as the relative proportion of “normal” sedi¬ 
 ment to glacial material, appears to vary in different parts of the 
 country, just as is true of the Pleistocene glacial deposits of the 
 Northern Hemisphere, and, while in places the top of the group is 
 sharply defined, as, for example, to the northwest of Jaguariahyva 
 in Parana, the upper limit is commonly vague, the chief criterion 
 being the absence higher up of any pebbly bands displaying charac¬ 
 ters not purely fluviatile. It is obvious that in borings the upper 
 limit of the Itarare would often be rather indefinite, just as is the 
 case in Uruguay, while certain conglomeratic strata at present 
 mapped with the Bonito may in the future be found to be of fluvio- 
 glacial origin and therefore have to be placed with the Itarare. 
 
 The northeasternmost point from which this series has been 
 reported is in the Abaete Valley in Minas Geraes, where, according 
 to Rimann, 40 the glacials and the succeeding “Areado sandstone” 
 rest upon the tilted lower Palaeozoic Bambuhy series, the existence 
 of the tillite in that neighborhood having long previously been noted 
 by D. Draper. 41 This area is not improbably an outlier from the 
 main basin, for to the west, so far as can be gathered, the Triassic 
 sandstone rests directly upon mica-schists at the Agua Suja diamond 
 mine, and on granite between Sacramento and Franca. The tillite 
 appears more to the south near Mococa, is relatively thin at Monte 
 Santo, and runs in a widening belt southward through Casa Branca 
 past Campinas and Sorocaba, at the same time becoming thicker. 
 Excellent descriptions have been given by Woodworth 42 and Cole¬ 
 man 43 of the glacials in the region both to the northwest and west of 
 Campinas, while a fine photograph of the boulder-beds has been 
 reproduced by Branner. 44 
 
 From Sorocaba the belt curves south west ward to Tibagy and 
 then turns abruptly southeastward past Ponta Grossa, Serrinha, Rio 
 Negro, and onward, the group ranging in thickness from 250 to 350 
 meters in Parana, but becoming progressively thinner in the direc- 
 
 39 Oliveira (1918), p. 12. 42 Woodworth (1912). 
 
 40 Leme (1924), p. 178. 43 Coleman (1918),"pp. 312-317; (1926), pp. 156-162. 
 
 41 Branner (1919), p. 214. , 44 Branner (1919), PI. X. 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 
 
 71 
 
 tion of Santa Catherina. Woodworth’s 45 detailed section of the 
 group between Itaicy and Piracicaba shows a huge development— 
 between 700 and 1,000 meters, mainly of sandy tillite and sandstone, 
 with at least two important partings of shales, although much of the 
 material can not strictly be referred to as tillite, consisting as it does 
 of sandstones with occasional small boulders, while similar erratics 
 occur in certain of the shaly bands. The succession is consequently 
 rather like that described by various geologists at Bacchus Marsh in 
 Victoria, though fossils have not yet been recorded therefrom. 
 
 The accounts by Woodworth and by Coleman of the glacials at 
 Jaguaricatu, Sengens, Itarare, and Ponta Grossa give a very good 
 idea of the variable nature of the deposits, but much of the rock, 
 when fresh, is actually a dark-blue massive tillite, often weathering 
 spheroidally, with rather small inclusions, though at times boulders 
 of over a meter across can be seen; the rock thus recalls the Dwyka 
 conglomerate. From near the railway station of Elias Fausto, Doctor 
 Pacheco has reported some unusually large granite erratics, one of 
 them measuring fully 3 by 3 by 2 meters. The material at Ponta 
 Grossa, on the contrary, I found to be a yellowish soft sandstone 
 with only a few conglomeratic bands in which striated stones are 
 decidedly rare. Boulder beds in verity with abundant and large 
 inclusions were, however, observed between Rio Negro and Campo 
 do Tenente on the railway to Serrinha. 
 
 Whereas the floor beneath the Itarare is, so far as can be gathered, 
 rather even in the States of Sao Paulo and Parana, the features 
 visible at Serrinha and the description given of the glacials there by 
 Woodworth 46 indicate an irregular floor and indeed suggest to me the 
 presence of a “pre-glacial” gorge in the Furnas sandstone, cut right 
 down into the pre-Devonian slates beneath, the northern wall of 
 which is made by the sandstone cliffs a few hundreds of meters from 
 Serrinha, along the railway to Curitiba, the hollow being filled in 
 with a sandy tillite overlain by a fluvio-glacial sand. The general 
 aspect strongly reminds me of several well-established cases of such 
 a kind in Natal (South Africa), where the Dwyka tillite is resting 
 in ravines cut in the Table Mountain sandstone and in places lies 
 banked against cliffs of the latter, for example, on the Umtamvuna 
 River near the road bridge between Izingolweni and Bizana. The 
 high irregularity of the pre-glacial landscape is, moreover, obvious 
 between Serrinha and Lapa, a little distance to the south. 
 
 The sections between Rio Negro and Canivete on the Porto 
 Uniao Railway reminded me very much of the Dwyka in parts of 
 Natal and Zululand—the blue, solid tillite, blue uniform mudstone 
 
 L 
 
 with hardly any stratification, and bedded blue shales. The latter, 
 in thickness from a few decimeters upward, run regularly for good 
 
 45 Woodworth (1912), pp. 54-61. 46 Ibid., pp. 65-66. 
 
72 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 distances and probably mark, just as they do in Natal, the temporary 
 cessation of morainic deposition within an estuary, but opportunity 
 for detailed examination was wanting. The discovery by Doctor 
 Oliveira of marine fossils close to Teixeira Soares on the railway 
 southwest of Ponta Grossa and again at Rio Negro in black shales 
 sandwiched between two banks of boulder shales—so far as I can 
 gather in the upper part of the Itarare—indicates that the ice was at 
 a certain stage discharging into the ocean, just as was the case with 
 the Dwyka ice-cap between Keetmanshoop and Mariental in the 
 Warmbad district in Southwest Africa. The fossils, which are as yet 
 undescribed, consist of small forms of Lingula , Discina , Leda (?), a 
 pentamerid brachiopod, a gasteropod resembling Pleurotomaria, 
 other indeterminable forms, scales of ganoid fishes, and wings of 
 insects. 47 
 
 To the south of Rio Negro the Itarare is still thick—over 270 
 meters on the Lages-Florianopolis road—though there it includes 
 152 meters of beds at the base doubtfully grouped with the Gon- 
 dwana, but regarded by Oliveira as forming an integral part thereof. 
 Approaching the Tuberao Valley the Itarare must thin rapidly, being 
 there represented by 32 meters of strata, of which the well-known 
 massive “Orleans conglomerate” forms the bulk, a stratum in which 
 the pebbles are sometimes from 20 to 25 cm. in diameter, and perhaps 
 only a local deposit. Some 2 km. below the station of Lauro Muller, 
 resting directly on granite, are shales enveloping a large granite 
 erratic which is in contact with the floor, as figured by both White 48 
 and Woodworth. 49 From this point the boundary with the granite 
 runs southward past Cocal and the group ends near the mouth of 
 the Urussanga River. 
 
 In Rio Grande do Sul, the Itarare is thin or missing between 
 Porto Alegre and the Uruguayan border. In the Gravatahy coal¬ 
 field a little to northeast of Porto Alegre, according to data kindly 
 supplied me by Doctor Simch, the bore-holes pass directly through 
 shales and clays belonging to the Bonito into granite, which indi¬ 
 cates the overlapping of the higher group on to the Carboniferous floor. 
 Due west of Porto Alegre, between Xarqueadas and Butia, 50 a basal 
 conglomerate is erratically developed, resting on granite and con¬ 
 taining large blocks of that rock; important is the section revealed 
 by the old Isabel shaft near Sao Jeronymo, where the 42 meters of 
 beds referred to as the “Orleans conglomerate” 51 include shales, 
 some of them dark and bituminous. Between Suspiro and Jaguary, 52 
 on the Cacequy-Rlo Grande Railway, an unbedded gray shaly rock 
 1 meter thick with inclusions resting on the granite probably alone 
 
 60 White (1908), pp. 43, 47. 
 
 51 Ibid., p. 119. 
 
 62 Oliveira (1918), p. 27. 
 
 47 Oliveira (1918), p. 14. 
 
 48 White (1908), plate facing p. 28. 
 
 49 Woodworth (1912), Fig. 23. 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 
 
 73 
 
 represents the glacials, being overlain by conglomeratic sandstones, 
 sandstones, shales, and coals—the Bonito beds. 
 
 ( 3 ) The Boulder Beds in Uruguay 
 
 The distribution of the boulder beds in Uruguay is hardly known, 
 except in the vicinity of Fraile Muerto, as described above in (i), 
 but there can be no doubt as to the highly diversified character of 
 the “pre-glacial” surface in northeastern Uruguay. While the dip 
 of the formations is everywhere low, knobs of the crystalline floor 
 project through the covering of Gondwana beds at a number of places, 
 the surrounding strata being formed in some instances by the 
 Estrada Nova even. Such a granite inlier occurs along a tributary of 
 the Arroyo del Chuy, north-northeast of Melo, one just to the east 
 of that town, another to the south, forming the Cerro Verde, and two 
 in the Arroyo Quebrebracho to the northwest of the station of Cerro 
 de las Cuentas. The conspicuous Cuchilla de Cerro Largo has all the 
 appearance of being a glaciated range that must have remained 
 exposed long after the ice had disappeared and its undulating base 
 had become buried beneath glacial, fluvio-glacial, and normal 
 sediments. 
 
 The bore-hole of Zanja Honda, situated in the valley on the north¬ 
 western side of this great ridge, passed down into phyllites at about 
 70 meters, the section showing several meters of blue-gray conglom¬ 
 erate and breccia, which from the cores preserved I believe to be of 
 glacial origin, overlain by 14.8 meters of sandstones and these by 
 12 meters of the curious banded shales described by Walther 53 under 
 Woodworth’s name of “ desmopelodite, ” and which, from their 
 alternating light and dark laminae of differing texture (in thickness 
 from 12 down to below 1 mm.), have been compared with the 
 seasonally banded glacial clays of the Pleistocene. Similar beds were 
 described by Guillemain at Laguna la Tuna in the head of the 
 Arroyo Fraile Muerto, which not improbably correspond with the 
 “inter-glacial” shale parting in the lowest glacial horizon at Paso 
 Tia Lucia a short distance farther down the valley. These shales were 
 obviously laid down in tranquil water (limno-glacial deposits) and 
 display on their faces what are supposed to be crustacean tracks. If 
 the 33 meters of strata at Zanja Honda immediately above them 
 represent the Iraty, as stated by Walther, though this with good 
 reason is doubted by Arocena, 54 then a great overlap has taken place 
 within a distance of 25 km. with the suppression of the Bonito, which, 
 taking into account the physiographical conditions that must have 
 been present, would not be altogether unexpected. 
 
 While Walther’s account 55 shows that coarse sandstones with 
 
 53 Walther (1919), p. 91 and PI. X. 65 Walther (1919), pp. 97-98. 
 
 54 Arocena (1926), p. 5. 
 
74 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 boulders, possibly the second glacial horizon, play an important role 
 to the west of Cerro de las Cuentas, so far west indeed as the Arroyo 
 Cordobes, and while glacials have recently been discovered in the 
 region immediately beyond, it would appear that a little to the north 
 of Durazno the Triassic has overstepped the lower series. As regards 
 the deep boring at Tacuarembo, 59 it is just possible that glacial beds 
 here form an important group, my examination of the cores showing 
 that from 310 meters downward they are perhaps as much as 40 
 meters thick, including a dark gray clayey rock and a pale gray gritty 
 material with grains of quartz and rock fragments, passing down at 
 about 349 meters into a typical tillite containing pebbles of granite 
 with the granitic basement struck at about 350 meters. No similar 
 rocks have yet been noticed at the base of the system shortly toward 
 the east along the Tacuarembo River, where again the Triassic 
 appears to transgress on to the basement. 60 
 
 It is unfortunate that so little should be known about the north¬ 
 western side of the Parana Basin. No glacials have yet been reported 
 from Paraguay or southern Matto Grosso, though they are known to 
 occur close to Rio Bonito in the southern part of Goyaz, resting 
 apparently on the Devonian, and not unlikely are present in the 
 upper part of the Parnahyba Valley also. 
 
 ( 4 ) The Position of the Glacial Center 
 
 Although the rounded granite surfaces recorded close to Campi¬ 
 nas in Sao Paulo do not bear any definite striae, and roches moutonnies 
 have not been recorded elsewhere in Brazil, a certain amount can be 
 deduced regarding the probable position of the Carboniferous glacial 
 center from the inclusions in the tillite of Sao Paulo and Parana. 
 Coleman 61 has mentioned erratics of a peculiar coarse conglomerate 
 near Capivary containing pebbles of brilliant red jasper derived from 
 some ancient formation unknown in the country, and which I have 
 elsewhere remarked are very like those found in the Dwyka of 
 Southwest Africa that must have been derived from the Matsap 
 series of Griqualand West lying hundreds of kilometers to the east. 
 Recently, however, Dr. L. C. Ferraz 62 has recorded a practically 
 identical conglomerate in situ far to the south, in the Morro Bahu, 
 on the north side of the Itajahy River close to Blumenau, in Santa 
 Catherina, which, if not the source of the blocks in question, would 
 indicate at least the likelihood of a southerly origin. 
 
 Furthermore, Woodworth 63 has drawn attention to certain inclu¬ 
 sions of fine-grained white sandstone (whetstone), which are indica¬ 
 tive of an easterly derivation. In Uruguay, as has been pointed out, 
 the center seemingly lay toward the southeast, and a similar direc- 
 
 69 Walther (1919), p. 100. 61 Coleman (1918), p. 314. 63 Woodworth (1912), p. 63. 
 
 60 Arocena (1926). 62 Ferraz (1921). • . 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 
 
 75 
 
 tion is prompted by the Argentine occurrences, all of which suggest 
 an area lying out in the present Atlantic, and with which the sporadic 
 development of the glacials and the overlapping of the higher beds on 
 to the crystallines in the extreme south of Brazil would appear to 
 support. A certain amount of speculation on this problem will be 
 indulged in later on. 
 
 ( 5 ) The Tuberao Series 
 
 Doctor Oliveira has restricted the name Tuberao series to the 
 strata above the Orleans conglomerate or the Itarare glacials, 
 namely, to the Bonito group below and the Palermo group above, the 
 former about 180 meters thick, the latter about 80 meters, or 270 
 meters in all. 
 
 The Bonito consists of soft, yellowish and grayish-white sand¬ 
 stones, sometimes in massive beds, interbedded with bluish or gray 
 and occasionally micaceous shales, often plant-bearing; conglomer¬ 
 ates are rare and thin. Important is the occurrence therein in Brazil 
 of seams of coal upon at least five horizons that are identifiable over 
 pretty wide areas. The workable seams are those of the Bonito coal 
 about midway in the succession and the Barro Branco or Sao Jero- 
 nymo coal near the top; partings of shale are general, and the coals 
 are always high in ash and often extremely pyritic. They are well 
 developed in the Tuberao-Treviso area in Santa Catherina and again 
 along the Jacuhy Valley west of Porto Alegre in Rio Grande do Sul, 
 while coals have also been proved at intervals on the Rio Grande 
 Railway, such as at Suspiro, Rio Negro, Candiota, and between 
 Cerro Chato and Sao Joao do Herval down to the Uruguayan border, 
 and also in the opposite direction in the belt west of Ponta Grossa and 
 Tibagy in Parana and on the Rio das Cinzas northwest of Itarare. 
 A wealth of detail is to be found in the monumental report of the 
 Brazilian Coal Commission 64 and in the supplementary memoir 
 by Doctor Oliveira. 65 
 
 The Palermo group is formed mainly of soft shaly beds of a gray 
 or blue color (though red tints have occasionally been observed in 
 Parana) and passes up without a break into the Iraty shales. It is 
 marked off from the Bonito in Brazil by a calcareous horizon with 
 large nodules of chert, while silicified wood of the genus Dadoxylon 
 is common, but the group is difficult to delimit in Uruguay. 
 
 It might be noted that in two areas where coals are locally 
 developed in the basal part of the Bonito, 66 a thin seam rests directly 
 upon the top of the tillite forming part of the Itarare, for example, in 
 the valley of the Rio das Cinzas and at Butia to the southwest of Sao 
 Jeronymo, and seemingly also near Suspiro on the Rio Grande Rail¬ 
 way, thus recalling to some extent the features to be seen at Vereenig- 
 
 66 Oliveira (1918). 66 Ibid., pp. 34, 103. 
 
 64 White (1908). 
 
76 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 ing in the Transvaal, where the Ecca coals may rest directly upon 
 either tillite or pre-Devonian dolomite. 
 
 It is further interesting to observe that the Tuberao series, when 
 followed northward into the State of Sao Paulo, where it is termed 
 the “Tatuhy ” by Doctor Pacheco, is largely made by greenish sand¬ 
 stones, and that coals are absent, save for a seam at Jacuha, 20 km. 
 southwest of Campinas, overlain by a boulder bed supposed by him 
 to be of glacial origin, but which is supposed by others to be remanie, 
 as in the succession at Vereeniging. Between Limeira and Rio 
 Claro, in Sao Paulo, the Tatuhy, followed by the Iraty, has become 
 quite thin. 
 
 In the opposite direction the Gondwana beds make a basin near 
 the Uruguayan border and the Tuber ao is found to the south of the 
 Jaguarao River, 67 where, in three bore-holes at least, 92 meters of 
 beds were proved beneath the Iraty group with Mesosaurus , dis¬ 
 closing a seam of impure coal 1 meter thick that probably corresponds 
 to one of the upper coals of this series in Brazil. This being the first 
 locality in Uruguay where coal was struck, borings were thereafter 
 undertaken at a number of points in the Department of Cerro Largo 
 and also at Tacuarembo—the sections of which have been detailed 
 by Walther 68 —but without success, from which it would appear 
 that the limit of the area characterized by the coal-bearing facies of 
 the lower Gondwana has been reached somewhere about Melo, 
 though I am inclined to agree with Walther that the borings at 
 Buena Vista and Isla de Zapata were stopped too soon. The boring 
 in the town of Melo showed no less than 301 meters of strata ascri- 
 bable to the Tuberao without the crystalline floor being reached, 
 though it is not improbable that toward its base fluvio-glacial beds 
 like those described from Fraile Muerto were actually cut. 
 
 It is undoubted that to the southwest of Melo the thickness of 
 strata decreases considerably, which is in great part due to the 
 uneven character of the floor upon which the beds repose, as proved 
 by the isolated inliers of granite in the midst of the nearly horizontal 
 sediments, as already referred to. The equivalent of the Palermo is 
 more arenaceous than in the coal-fields of Brazil, making it practic¬ 
 ally impossible to separate this group from the Bonito. The same is 
 the case to the west, the Tacuarembo boring exhibiting only 84 
 meters of strata between the Iraty and the granite, of which perhaps 
 only just over one-half can be relegated to the Tuberao (Bonito 
 and Palermo). 
 
 Westward in Argentina the Bonito is apparently represented in 
 the San Cristobal bore-hole in Santa Fe at between 1,140 and 1,384 
 meters, some of the sandstones being reddish, like the Paganzo, a 
 feature still more marked in the Tostado bore-hole (1,600 meters) to 
 
 67 Walther (1919), pp. 92-93. 68 Ilid., pp. 99, 100; see also (1924). 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 77 
 
 the northwest, but the strata in the Alhuampa bore-hole, from 1,700 
 meters down to 2,100 meters, are whiter again. 
 
 Our knowledge concerning this series on the northwest and north 
 of the basin is of the scantiest. Information given me by Doctor 
 Beder and an examination of the specimens collected by him in 
 Paraguay show that carbonaceous shales occur about Santa Maria to 
 the west-northwest of Posadas and grits, sandstones, conglomerates, 
 and carbonaceous shales like those of the Bonito at Paraguary close 
 to Asuncion and similar strata at Villa Hayes, a little above that 
 town, on the right bank of the Paraguay River, the latter consti¬ 
 tuting the only solid rock exposed in the broad alluvial plain west of 
 that river. 
 
 (6) The Flora of the Bonito Group 
 
 D. White 69 and G. Lundqvist 70 have materially added to Zeiller’s 
 original description, while the former has also made it clear that the 
 plants occur on various levels from almost immediately above the 
 Orleans conglomerate upward. Those from the lower part of the 
 group in striking fashion consist solely of typical members of 
 the southern flora, such as Gangamopteris cyclopteroides {G. obovata 
 White), Glossopteris browniana, Cordaites hislopi, Phyllotheca gries- 
 bachi and P. muelleriana, but on a higher horizon in association with 
 Glossopteris indica, G. ampla, G. occidentals, and on some unknown 
 level Neuropteridium validum and Annularia australis, appear 
 Lepidophloios laricinus, Sigillaria brardii, Lepidodendron pedro- 
 anum, Sphenophyllum oblongifolium, Pecopteris spp., and perhaps 
 Lycopodiopsis derbyi, though the last is regarded by White as a 
 southern form. 
 
 It is instructive to observe that in western Argentina these latter, 
 as well as other northern Carboniferous forms, appear to have been 
 found, together with the southern types, in the lowest stage of the 
 Paganzo, so that in Brazil these “exotic” forms make their entry at 
 a somewhat later date apparently. In South Africa they entered at 
 a later period still during the Ecca, subsequent indeed to the deposi¬ 
 tion of the Iraty group. It is to be hoped that further systematic 
 collecting may be undertaken to afford additional information on this 
 important question. These observations rather lessen the probability 
 of White’s 71 supposition that those northern forms were able to pene¬ 
 trate the region only because of a progressive amelioration of 
 the climate. 
 
 In making age determinations we have to eliminate the prepond¬ 
 erant southern element as being of no direct value, though of course 
 that assemblage is closely related to the floras of the equivalent 
 Talchir-Karharbari series of India, the Greta Coal Measures of New 
 
 69 White (1908), pp. 361-373; also (1907), p. 617. 71 White (1908)^.399. 
 
 70 Lundqvist (1919). 
 
78 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 South Wales, and the Ecca Coal Measures of South Africa. Con¬ 
 sidering merely the northern forms, the five just mentioned, together 
 with Samaropsis {Cardiocarport) bar cello sa are all identical with or 
 very closely related to typically upper Carboniferous species , and in 
 my opinion form useful confirmation for a pre-Permian age, such as 
 has been arrived at from other considerations, though this view 
 differs radically from the conclusions reached by D. White and 
 certain others. 
 
 ( 7 ) The Passa Dois Series 
 
 This embraces the Iraty, some 70 meters thick, below', and the 
 Estrada Nova (or Corumbatahy), fully 150 meters thick, above, two 
 groups that are extraordinarily constant throughout the basin, 
 though not yet recognized in central or western Argentina. 
 
 The Iraty group 72 is composed chiefly of black, carbonaceous, and 
 often “papery” shales, in places pyritic, and which on being freshly 
 broken may give out an odor of petroleum; small veins of bituminous 
 matter—albertite or grahamite—are sometimes to be found. On 
 weathering they bleach to pale bluish, and finally to white or faintly 
 pink, flaky clay shales with secondary gypsum, exactly as in the case 
 of the “White band ” of the Dwyka shales. Characteristic are nodules 
 of chert and thin layers of limestone (particularly toward the base), 
 sometimes fibrous in structure. 
 
 In Uruguay the shales are not so argillaceous, while thin bands of 
 white sandstone and darker micaceous sandstone are also present, 
 making it difficult to delimit the group from the “Palermo ” beneath, 
 which fact explains the excessive value given by Walther for its 
 thickness, though such figure must certainly exceed 100 meters. In 
 Paraguay the beds are also more arenaceous. On the other hand, 
 the group becomes more calcareous to the northeast, and in the 
 neighborhood of Rio Claro in Sao Paulo, white and gray fine-grained 
 crystalline limestones play an important part, being actually quarried 
 as a source of lime. 
 
 The group is traceable from Palmeira in that quarter southwest- 
 ward, making a sweeping curve through Parana that passes 60 km. 
 to the west of Tibagy and then turns east of south through Iraty to 
 Lauro Muller and thence to the Atlantic coast. Possibly overlapped 
 just a little to the northeast of Porto Alegre by higher divisions, the 
 Iraty has been proved by occasional outcrops and by borings along 
 the southern side of the Jacuhy River, running nearly due westward, 
 passing to the north of Sao Sepe and thence through Sao Gabriel, 
 after which its exact course is not known with certainty, but it is 
 exposed on the Uruguayan border between Paso Maria Isabel and 
 Paso Minarano, and turning southwestward stretches along the base 
 
 72 White (1908), pp. 181-191; Oliveira (1918), p. 16. 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 
 
 79 
 
 of the Cerro Largo and, curving through Fraile Muerto, strikes west- 
 northwestward toward the Rio Negro. It was cut in the boring at 
 Tacuarembo beneath the Triassics and is known to reappear on the 
 farther side of the basin at Villarrica in Paraguay, where fossil bones 
 occur, while Millward has reported the extension of this and the 
 Estrada Nova in the south of Matto Grosso and Goyaz, apparently 
 from the foot of the Sierra de Aquidauana northeastward past Rio 
 Bonito to Rio Verde, but this quarter is as yet only slightly known. 
 
 Fossils —Important is the frequency with which the singular 
 little reptile Mesosaurus has been found at various points along the 
 known huge extent of outcrop within certain layers in the black 
 shales. In places the shales are just crowded with skeletons and in¬ 
 deed, not infrequently, such remains are discovered in bore-hole 
 cores. Along with it occurs the very closely allied but generally 
 larger form Stereo sternum, though curiously it is stated that the latter 
 is practically confined to the limestone layers. Some splendid show 
 specimens of both these reptiles, and of Stereosternum embedded in 
 white limestone, are displayed in the State Museum at Sao Paulo. 
 
 It is essential to observe that long ago Lydekker, and recently 
 that high authority Von Huene, 73 have refused to recognize the gen¬ 
 eric status of Stereo sternum, both regarding it as a species of Meso¬ 
 saurus, and for the present we shall therefore consider them as 
 Mesosaurus tumidus and M. brasiliensis respectively. Von Huene, 
 reviewing the known South African species, has furthermore placed 
 them all in the one form, M. tenuidens, with the opinion that the 
 latter is probably distinct from the two species represented in Brazil. 
 Noteworthy is the fact that this unique creature—clearly a free- 
 swimming reptile—has as yet been found only in the Parana basin 
 and not elsewhere in South America, although strata of equivalent 
 age are certainly represented elsewhere in that continent, while 
 across the Atlantic it has been recorded at intervals from the 
 Kaokoveld in Southwest Africa down to the south of the Karroo, 
 though not to the east or northeast (see Fig. 7, on which the approx¬ 
 imate limits are indicated). 
 
 Walther has mentioned the finding on the Rio Jaguarao in north¬ 
 eastern Uruguay of impressions resembling those described by White 
 from the Bonito of Santa Catherina under the name of Hastimima, 
 originally thought to be a plant, but later discovered to belong to an 
 eurypterid. Doctor Pacheco has informed me of the occurrence in 
 Sao Paulo State of Schizodus (?), while Broili in a letter to Walther 
 has identified certain fish-scales from the boring at Isla de Zapata near 
 Melo as belonging to a ganoid. Not improbably the fossil wood 
 Dadoxylon pedroi Zeill. comes from the Iraty of Sao Paulo. Derby has 
 furthermore remarked upon the association in Sao Paulo State—an 
 
 73 Von Huene (1925), pp. 118-120. 
 
80 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 important one—of Psaronius with Stereosternum and Lycopodiopsis 
 derbyi , and the latter has been discovered even above the Iraty 
 group. 74 
 
 Just as in the Cape Province with the shales of the White band, 
 so the bituminous nature of the Iraty shales has led to much pros¬ 
 pecting for coal, while their petroliferous odor when broken has 
 incidentally been responsible for the drilling of several deep bore¬ 
 holes in search of oil; in both countries, it might be remarked, the 
 shales are often sapropelitic in character and will furnish oil on 
 distillation. 
 
 The exclusive presence of Mesosaurus and the outstanding litho¬ 
 logical similarities, down to minutiae, render the correlation of the 
 Iraty group with the White band of the Dwyka series absolutely assured , 
 although a distance of 6,300 km. separates their nearest outcrops, 
 whereas these two zones are traceable not more than 1,000 km. 
 away from the Atlantic shore in either case. 
 
 (8) The Estrada Nova Group 
 
 The Estrada Nova group, or, as it has been called in Sao Paulo, 
 the Corumbatahy group, follows the Iraty with seeming conformity 
 from just south of Palmeira in that State to Tacuarembo in Uruguay, 
 an assemblage that seldom has good exposures, crumbling readily 
 and so releasing the cherty concretions that form so characteristic 
 a feature of it. 75 
 
 The beds are arenaceous shales—gray, dark greenish or reddish, 
 pink, lavender or lilac—always well-laminated, with intercalated 
 bands of relatively pure sandstone very rich in siliceous nodules, and 
 nodules of chert, together with some white limestone. While most 
 of the bright coloration is certainly due to weathering, some of it is 
 original, as borings have proved, and the superficial pink, heliotrope, 
 or lilac tints help materially to distinguish this group from the lower 
 division, although proving a source of trouble when attempting to 
 discriminate it from the succeeding deeply colored strata. Taken 
 by I. C. White as the top of this group, because it was overlain by the 
 Triassic Rio do Rasto series, is the Rocinha limestone, a band attain¬ 
 ing the maximum thickness of 3 meters, though only rarely exposed, 
 and known along the Rio Rocinha in Santa Catherina, at Therezina 
 in Parana, and Fartura in Sao Paulo, but which is supposed to be 
 represented at other points by cherty beds with fossil mollusca. 76 
 The thickness of the group is taken as 150 meters, but the inclusion 
 of certain higher beds in Parana that can be placed in the Estrada 
 Nova would bring up the value in that State to fully 300 meters. 
 
 The Estrada Nova is important because of its fossiliferous charac- 
 
 74 White (1908), p. 371. 
 
 75 Ibid., pp. 191-195. 
 
 76 Oliveira (1918), p. 17. 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 
 
 81 
 
 ter, indicative of deposition in an estuary with occasional incursions 
 of the ocean. At Rio Claro, east of Marechal Mallet Station in 
 Parana, mollusca have been found in the cherts and siliceous oolites 
 (Zone i) and also at a slightly higher level in a calcareous sand¬ 
 stone (Zone 2), these two fossiliferous horizons having been recog¬ 
 nized at several other points to the west of the railway between 
 Marechal Mallet and Iraty, these occurrences having been detailed by 
 Oliveira and certain of the forms described by Holdhaus. 77 (See below.) 
 
 The sequence that I was able to examine between Marechal 
 Mallet and Rio Claro is a rather important one. In the first bore-hole, 
 situated 4 km. to the west of Rio Claro, there are fully 200 meters 
 of beds—bluish-green shales with white calcareous and sandy well- 
 banded layers—overlying the Iraty group. The soft fossiliferous 
 sandstone with “Solenomorpha ” etc., taken by Oliveira as the top of 
 the Estrada Nova, crops out about 40 meters higher still, yet above 
 the latter the rising ground on the road to Marechal Mallet displays 
 for a vertical distance of 40 meters similar strata, except for the 
 absence of cherts. Above this the beds become more arenaceous, yet 
 not until nearly 75 meters above the “ Solenomorpha ” band do strata 
 make their appearance, having strong red coloration, from which 
 point red and purple sandstones and mudstones are visible, followed 
 by buff fine-grained sandstones and greenish mudstones and shales, 
 with intrusions of dolerite up to the summit of the Serra do Tigre. 
 
 The upper portion of these 300 meters of beds overlying the Iraty 
 could preferably be included in the Estrada Nova, despite the 
 absence of cherts, the higher strata with bright-colored shales and 
 mudstones being allotted to the Rio de Rasto. 
 
 In Rio Grande do Sul this group has nowhere been mapped, but 
 in Uruguay, entering from Brazil, it is known to run southward 
 down the Canada de Ibanez past Melo and a little to the north of 
 Fraile Muerto and thence northwestward across the Rio Negro and 
 up the Tacuarembo Valley, covering a wide area to the south and 
 east of the town of that name. It is also known to surround much of 
 the granite inlier of the Department of Riviera. 
 
 The only place where I was able to study these beds in Uruguay 
 was in the shallow basin crossed by the railway between Fraile 
 Muerto and Melo, where, proceeding from the former station, the 
 laminated olive and buff clayey shales with ribs of sandstone and 
 limestone belonging to the Iraty pass up at kilometer 390 into 
 greenish, lavender, and then into red, carmine, and white laminated 
 beds, with thin layers of white cherty rock, and are hence very like 
 the Corumbatahy group of Sao Paulo. Between kilometers 392 and 
 393 some red sandstones appear, and, though the well-laminated 
 character of the shales persists for a short distance, the beds give way 
 
 77 Holdhaus (1919). 
 
82 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 to a group of less well bedded red and variegated clays and mud¬ 
 stones and red sandstones somewhat evenly stratified, sometimes 
 cross-bedded, as at 395. At kilometer 400, just before reaching 
 Banado de Medina, a thick sandstone of this kind rests on red clays, 
 and such arenaceous beds extend almost to Melo and occupy a con¬ 
 siderable area to the north and northwest. All the rocks from about 
 kilometer 394 onward can, I think, be taken as representing the Rio 
 do Rasto group. The bore-hole at Buena Vista, 20 km. north of 
 Melo, reveals a similar sequence with about 245 meters of strata 
 ascribable to the Estrada Nova, while Walther states that to the 
 south of Melo siliceous concretions have weathered out from these 
 beds in great numbers. 
 
 In Paraguay, according to Beder, 78 there are at Villarrica, in addi¬ 
 tion to the undoubted Iraty cherts, some with oolitic structure and 
 also some higher reddish friable sandstones that have yielded species 
 of “ Solenomorpha .” According to Leme, 79 Millward has found in the 
 south of Goyaz similar beds with plant fossils and lamellibranchs, 
 presumably belonging to this group and covering a wide area round 
 about Rio Bonito and Jatahy and thence southwestward towards 
 Aquidauana on the railway to Porto Esperanga. 
 
 ( 9 ) The Fossils of the Estrada Nova 
 
 The stratigraphical position of the Estrada Nova is at the 
 moment involved, as this so-called “group” would appear to include 
 beds of rather different ages. Of the lamellibranchs that occur at 
 a number of points in Parana, Holdhaus 80 has described several 
 species of Solenomorpha and one of Sanguinolites , maintaining a 
 Permian age for the formation. Dr. Reed, who has determined the 
 mollusca collected by me from Zones 1 and 2 at Rio Claro, doubts 
 Holdhaus’s identifications, having discovered in these two horizons 
 only Trias sic forms, such as Pachycardia, Anodontophora , Myophoria , 
 and Schafhautlia, as detailed on page 150. 
 
 Although the sequence has hitherto been regarded as an 
 unbroken one, a stratigraphical hiatus must he present , though the 
 precise plane thereof has not yet been located in the field. Any 
 beds below this surface of discontinuity must lie conformably upon 
 the Iraty and hence belong to the Permian; those above it being of 
 Triassic age, must perforce be associated with the overlying Rio 
 do Rasto and are presumably conformable therewith. Future 
 work will no doubt bring about a better re-arrangement of the 
 Estrada Nova. 
 
 Fish scales and bones have been discovered, also silicified wood has 
 been determined by D. White as Dadoxylon nummularium, a species 
 in which no annual rings can be made out, though Walther has found 
 
 78 Beder (1923), p. 11. 79 Leme (1924), p. 179. 80 Holdhaus (1919). 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 83 
 
 other kinds in Uruguay in which the rings are perfectly well defined. 
 At kilometer post 124 on the railway, 17 km. south of Iraty, I 
 obtained a fragment of Glossopteris , while, from near the top of the 
 group in the area between Marechal Mallet and Roxo Roiz, Zeiller 81 
 has determined Glossopteris browniana , G. angustifolia, Txniopteris 
 possibly T.feddeni, Pecopteris sp., Cladophlebis sp., and equisetaceous 
 stems. A frond in the Geological Survey Museum, Rio, labeled 
 “Glossopteris” belongs probably to the Triassic genus Sagenopteris. 
 
 (10) The Triassic Overlap 
 
 Before proceeding to the description of the higher groups, it will 
 be proper to discuss this particular question because of the light 
 which it sheds on the succession, more particularly in Parana. 
 
 Evidence, palaeontological as well as stratigraphical, has gradu¬ 
 ally been accumulating, as can be found on perusing the writings of 
 Walther, Baker, and others, to show that a break must exist within 
 the Parana basin between beds that are certainly Triassic and those 
 that are definitely Permian or older, thus tending to bring the strati¬ 
 graphy of this part of the continent into harmony with that of western 
 Argentina. Future work will have to be directed toward the mapping 
 of the “ disconformity, ” but the gap, so far as can be made out, 
 seemingly represents the whole of the uppermost Permian and perhaps 
 much of the lower Triassic. 
 
 In the north of Sao Paulo the Triassic Botucatu sandstone reposes 
 on the ancient floor of schists near the Agua Suja diamond mine and 
 at Franca, while farther south, at Monte Santo the Pyramboia—the 
 equivalent of the Rio do Rasto—rests directly upon the Itarare 
 tillite. West of Mococa, on the border of Minas Geraes, Doctor 
 Pacheco informs me that the lower groups have suffered some flexing, 
 which action may have been due to inter-Triassic earth-movements. 
 At Palmeira the Pyramboia rests on the Iraty, and farther south on 
 the Corumbatahy or Estrada Nova, which relationship, so far as 
 can be gathered, thenceforward holds good into Santa Catherina, 
 though the latest geological map of Parana (1925; scale 1:1,000,000) 
 by Doctor Oliveira makes the Rio do Rasto transgress across the 
 Itarare and Passa Dois in the vicinity of Thomazina. In Santa 
 Catherina, I. C. White observed the contact with the Rocinha lime¬ 
 stone to be an unconformity over the short length exposed. 
 
 The Rio do Rasto extends on the coast from Morro Conventos 
 (lat. 2 9 0 ) to Conceiao do Arroio and thence westward to the north of 
 Porto Alegre, but according to information from Doctors Rego and 
 Lofgren this formation also builds the large hill to the east of that 
 town and stretches past Vumiao into the Serra Geral ( not the great 
 range of that name in Santa Catherina), immediately to the east. 
 
 81 Holdhaus (1919), p. 30. 
 
84 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 It is not improbable, judging from the height to which the granite 
 rises at Porto Alegre, that the Rio do Rasto overlaps across the 
 Bonito group of the Gravatahy coal-field to come to rest on the gran¬ 
 ite in the south. Proceeding up the Jacuhy Valley, with its exten¬ 
 sive alluvial plains, shales regarded as Iraty were observed by Doctor 
 Rego, 3 km. west of Rio Pardo, the beds round about belonging to 
 the Rio do Rasto. The deep boring at Ferreira, however, close to 
 Cachoeira, which I saw in progress, proved the Rio do Rasto, down 
 to 73 meters at least, with the Estrada Nova (including a 30-meter 
 sheet of dolerite) beneath and the top of the Iraty at 260 meters. 
 
 While the Tuber ao and Iraty make a sweeping curve around the 
 granite axis in the country between Sao Sepe and Bage, followed 
 farther to the north by the Rio do Rasto in regular fashion, Wal- 
 ther’s 82 observation of an area at Cagapava made by the Triassic 
 resting directly on the crystallines is of great importance. Between 
 Cagapava and the Seibal Copper Mine the sediments are only from 
 6 to 8 meters thick, with a red basal conglomerate, and are covered 
 by highly vesicular amygdaloids, presumably the equivalents of the 
 Serra Geral eruptives. At Sao Sebastiao, on the railway to Bage, 
 another outlier occurs. 
 
 Over much of Uruguay, just as in Brazil, the Rio do Rasto has 
 not yet been carefully delimited from the top of the rather lithologic¬ 
 ally similar Estrada Nova, and a pseudo-conformity probably exists, 
 as, for example, in the Department of Cerro Largo, the younger group 
 being distinguishable by the predominance of red sandstones, often 
 massive and cross-bedded. But in the important section described 
 by Walther 83 in the Cerro Mirinaque east of Tacuarembo, 80 meters 
 of these arenaceous red beds rest directly on the granite, whereas at 
 Tacuarembo, where such strata compose both the plain and the 
 neighboring spurs, the rocks cut in the deep boring embrace the 
 whole of the Iraty and Tuber ao. Examination of the core leads me 
 to believe that the bright cream, red, and brown sandstones, gritty 
 at their base, are here 126 meters thick and rest upon 24 meters only 
 of maroon and heliotrope mudstones that can be ascribed to the 
 Estrada Nova and these in turn upon the Iraty, etc. The overlap is 
 also discovered far to the south, nearly 100 km. away from the main 
 basin, where, on the Barrija Negro in the Minas district, Walther 84 
 has recorded, resting upon the crystal lines, gray, yellow, and dark- 
 red ripple-marked sandstones, coarse conglomerates, etc., overlain 
 by vesicular basalt. Boring No. 51 in the soil-covered region at 
 Santa Rosa on the Melo Railway, 55 km. from Montevideo, proved 
 red and chocolate-colored amygdaloid down to 78 meters at least, 
 and, after piercing some softer rock beneath, penetrated a red con- 
 
 82 Walther (1912) and (1912a). 
 
 83 Ibid. (1919), p. 112. 
 
 84 Ibid. (1912), p. 401. 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 85 
 
 glomerate with boulders of gneiss, grits, quartzites, etc., down to in 
 meters. The red gritty sandstones at Sauce, however, appear to me 
 to be probably of Cretaceous or Tertiary age. 
 
 From the record of the San Cristobal bore-hole 85 in Santa Fe, 
 Argentina, the Rio do Rasto apparently rests on the Bonito at about 
 1,140 meters, the Iraty not being identifiable, while that at Galeguay 
 in Entre Rios east of Rosario shows beneath 463 meters of Neogene 
 36 meters of volcanics (fine-grained porphyrites) resting directly 
 upon amphibolite without any Botucatu. According to Leme, the 
 transgression is again to be found far to the northwest, where, on the 
 west side of the basin at Aquidauana, the Botucatu followed by vol¬ 
 canics reposes upon schists. 
 
 These various and isolated observations, though scattered over 
 an enormous area, collectively justify our concluding that the Rio do 
 Rasto must in a number of places, chiefly be it noted on the margins 
 of the basin, overlap and be unconformable to the lower groups, thus 
 forming a parallel to the Karroo system. Within the basin, however, 
 the molluscan horizons (Triassic) would prove that this plane of 
 overlapping is there situated somewhere within the limits of the 
 Estrada Nova as correctly defined, presumably nearer its base, 
 these marine beds being elsewhere overstepped by the continental 
 Rio do Rasto. The tracing out of this break will be an important 
 investigation for the future. The circumstances in the Parana basin 
 would seem to form an almost exact parallel to those in the north of 
 the Orange Free State in South Africa, where, by the dying out of 
 the Molteno beds, two maroon and mainly argillaceous formations 
 are brought together, the upper Triassic Red beds coming to rest 
 directly upon the lower Triassic upper Beaufort beds in that area, 
 but in the central Transvaal upon the lower Permian Ecca beds 
 generally. 
 
 Attention can also be drawn to the transgression of the Trias- 
 Rhastic (Stage IV) in Argentina across Stages III, II, and even I of 
 the Paganzo, as detailed earlier, 86 of which Stage II displays many 
 points of lithological similarity with the Passa Dois series. In the 
 Parana basin, however, unlike the pre-Cordilleran region, the older 
 groups were only slightly warped and eroded during the early 
 Triassic. 
 
 (11) The Sao Bento Series—Triassic 
 
 In the Parana basin this is divided into: 
 
 3. Serra Geral eruptives, up to 500 meters in thickness 
 
 2. Botucatu sandstone_o to 100 meters in thickness 
 
 1. Rio do Rasto beds ... 100 to 400 meters in thickness usually 
 
 This arrangement, it will be observed, differs slightly from the 
 scheme proposed by I. C. White. 
 
 85 Hausen (1919), p. 35. 
 
 86 Section F (5). 
 
86 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 The placing of the lowest of these devisions in the Permian, as has 
 been done by Oliveira , 87 has nothing to support it, since the mollus- 
 can remains found beneath it belong to undoubtedly Triassic forms. 
 
 The beds again with vertebrate fossils while overlapping the 
 older groups in certain places, pass without a break into the Botu¬ 
 catu, where that division is recognizable, overlain in turn by and 
 intimately associated with the basic lavas or “traps. ” 
 
 The Triassic occupies a huge region—the greater part of the 
 basin in fact—for which the volcanic group is mainly responsible 
 areally. In places it is unconformably covered by patches of the 
 Bauru (Cretaceous) beds, while to the west the succession passes 
 beneath the Neogene of the lower Parana River, though whether 
 some faulting has been responsible for its lower position in that direc¬ 
 tion is not yet assured. 
 
 (12) The Triassic Continental Sediments 
 
 The Rio do Rasto , 88 also called the Pyramboia group in Sao 
 Paulo, is so closely associated with the Botucatu (or Sao Bento) 
 sandstone, that the two groups must be described together. Appar¬ 
 ently only in the south of Parana and in Santa Catherina, according 
 to the account of White, is the upper division well differentiated and 
 clearly distinguishable from the low r er; in Rio Grande do Sul it is in 
 places not clearly separable, while over much of Uruguay the paler 
 color of the top of the sedimentary formation is largely a secondary 
 character. The term Botucatu, furthermore, seems to have been also 
 used by the Geological Survey of Brazil for the sedimentary inter¬ 
 calations between the lavas. 
 
 The friability of the lower red sandstones and mudstones, 
 coupled with the hardness of the overlying Botucatu or the traps, 
 has been responsible for an escarpment traceable nearly right around 
 the basin, this feature being bordered by many picturesque outliers, 
 from table-topped to conical. To a wonderful degree the rocks and 
 scenery recall the “Basalt-Cave sandstone—Red beds” topography 
 around Basutoland or in the Kaokoveld and parts of southern Rho¬ 
 desia. The development in Santa Catherina is from all accounts 
 most like that of Basutoland and the features seen elsewhere like 
 those in Rhodesia. 
 
 The Rio do Rasto or Pyramboia consists of friable red, often 
 vivid, sandstones with occasional white blotches often false-bedded 
 on a large scale; layers of fine-grained pink, red, purplish, and occa¬ 
 sionally greenish clays are also present, but there is not the well- 
 marked lamination and alteration of shales and thin sandstones 
 characteristic of the Estrada Nova, while the diagonal bedding in 
 the sandstones is also most conspicuous. 
 
 87 Oliveira (1918), p. 11. 
 
 88 White (1908), pp. 197-211. 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 87 
 
 At Taquara, north of Porto Alegre, where the upper part of the 
 brown-red sandstones was of uniform character through a thickness 
 of fully 75 meters, I found the false-bedding most pronounced, the 
 planes often dipping at angles up to 40° from the horizontal in 
 opposite directions in surprising fashion through individual thick¬ 
 nesses in some places of 12 meters or more of rock. The sandstone is 
 quarried for paving slabs (Plate XII, B ), being built up of alternate 
 laminae of fine-grained rather clayey material and coarser, friable 
 “millet seed” grains in strictly parallel fashion and not in lenses; 
 mica is practically absent, but iron ores are abundant and a cal¬ 
 careous matrix is indicated. In the coarser bands not only the grains 
 of quartz, but even those of feldspar (which are quite fresh), are 
 well rounded, and up to a millimeter in diameter, and there can be 
 no doubt that this formation is of eolian origin. The same was found 
 to be the case about Sant’ Anna and Riviera, on the border of Uru¬ 
 guay, the smaller grains tending to be subrounded. Doctor Pacheco 
 informs me that the sand grains in the Pyramboia of Sao Paulo are 
 well rounded, each coated with a thin pellicle of iron oxide. 
 
 In the basal part of the group, as seen along the railway between 
 Porto Alegre and Santa Maria and again between Rivera and Tacua- 
 rembo in Uruguay the sandstones become fine-grained and pass 
 into mudstones of uniform nature and wonderfully brilliant hue 
 wherever freshly exposed, while irregular calcareous patches are 
 common. Crumbling readily, they give rise to sandy or loamy soils 
 of wide distribution, from which the red tint has been removed by 
 bleaching. What were considered to be the basal beds were cut in 
 the Tacuarembo boring at 126 meters and were coarser and gritty. 
 Some curious veins or reefs, within which the sandstones have been 
 partially or wholly silicified, are frequent in northern Uruguay, as 
 described by Walther; from examination of two of these near Paso 
 del Cerro it seems not improbable that they have been produced by 
 heated waters descending through fissures from the lavas above. 
 
 In the northeastern part of Paraguay the red sandstones inter- 
 bedded with and capped by basalts occur at Bella Vista and build 
 the Sierra de Amambay, while they extend northeastward from 
 Aquidauana into Goyaz and onward, with a large outlying area to the 
 northwest in the Chapada. 
 
 The thickness of the group must be considerable; at least 400 
 meters between Taquara and Santa Maria, but less than half that in 
 Parana, Santa Catherina, and Uruguay, and according to Baker 89 
 under 100 meters in the north. In the San Cristobal bore-hole it is 
 perhaps 240 meters thick. 
 
 The Botucatu sandstone is a little coarser and paler, generally 
 creamy or gray, though sometimes terra-cotta or red, and then hard 
 
 89 Baker (1923), p. 68. 
 
88 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 to distinguish from the underlying Rio do Rasto, particularly as it 
 tends to display the same conspicuous false-bedding. Pacheco states 
 that in Sao Paulo the grains are usually angular, but at Ityrapina 
 I found the sandstones, with which basalt flows are interbedded, to 
 be pinkish, gritty though friable, with quartz grains up to 3 mm. in 
 diameter, many of them oval or spherical, while the bulk were sub- 
 angular and not angular. Here the group was over 200 meters thick. 
 
 In Parana and Santa Catherina the sandstone, according to 
 White, 90 is very massive and makes long lines of yellow cliffs sinking 
 in a southerly direction until they come down almost to sea-level 
 behind Torres. The deep bore-hole No. 3, situated 20 km. inland 
 from that place, proved, according to Doctor Lofgren, 122 meters of 
 Botucatu overlying 145 meters of Rio do Rasto at least, but, curi¬ 
 ously to the southwest, neither at Taquara nor at Santa Maria, is 
 the upper member represented, though it is reported that a pale 
 sandstone underlies the lavas in the ragged escarpment between 
 these two towns. 
 
 Traced from Rivera to Tacuarembo, the top of the Rio do Rasto 
 immediately underlying the basalts develops a harder facies, paler 
 in exposures, and thus resembles the Botucatu superficially, but is 
 actually red when fractured, and moreover shows the same degree of 
 false-bedding as in the Rio do Rasto below. This feature can be 
 .ascribed mainly to the indurating action of the basalts, for the top is 
 often brecciated. This material commonly makes scarps and some¬ 
 times caps isolated hills, as in the symmetrical Tres Cerros east of 
 Tacuarembo or in the similarly styled elevations to the north of 
 Uruguay ana in Argentina mentioned by Hausen . 91 
 
 Whether the Botucatu is merely a slightly different facies of the 
 uppermost part of the Rio do Rasto, or whether it is really a separate 
 formation only partially developed or missing in certain portions of 
 the basin, thus allowing the basalts to repose directly upon the lower 
 division, is not quite clear, and only detailed mapping could settle 
 this point, which is of no great importance, fortunately. 
 
 There can hardly be any doubt, however, as to the predomi¬ 
 nantly eolian origin of the Sao Bento sediments, which is all the more 
 compelling when comparison is made with the obviously equivalent 
 Cave and Forest sandstone (Stormberg) of the Cape and Southern 
 Rhodesia, with their underlying red beds. The first named is similar 
 lithologically to the Botucatu, displaying features indicative of a 
 loess-like origin, while the second, as shown by “millet-seed” grains 
 and strong false-bedding, is of eolian origin, the sediments in both 
 cases being followed by effusion of basalts in which certain inter¬ 
 bedded sandstones occur. 
 
 90 White (1908), pp. 211-217. 
 
 91 Hausen (1919), PI. VI. 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 
 
 89 
 
 (a) Fossils —Under such circumstances the life of the period 
 would have been scanty and the poverty of fossils would thus find 
 an explanation. Silicified wood is abundant at many points in the 
 Rio do Rasto, both low down as well as at the top, for example at 
 Rivera; such is probably mainly referable to Dadoxylon. 
 
 Of Filicales we have only the record of Cladophlebis , determined 
 by Kurtz 92 as C. denticulcita Brongt. from the base of the Tres Cerros 
 near Tacuarembo, but which after an examination I feel sure is a 
 distinct and probably new species, though closely related to the 
 foregoing, which is a Rhaetic-Liassic form. 
 
 From a point a little to the south of Santa Maria da Bocca do 
 Monte the remains were obtained of Scaphonyx fischeri S. Woodw., 93 
 a dinosaur allied to Euskelesaurus from the upper Triassic Red beds of 
 South Africa, below which latter horizon dinosaurs have not yet been 
 discovered. Leme 94 mentions the finding of a dinosaur allied to 
 Scaphonyx at Cambembe, 50 km. to the north of the Chapada in 
 Matto Grosso, while Von Ihering has recorded a dinosaur tooth in 
 the sandstone of Sao Jose do Rio Preto, 450 km. to the northwest of 
 Sao Paulo. Baker has noted the presence of bones near the coast to 
 the northeast of Torres. 
 
 A layer in the core from the bore-hole at Ferreira in Rio Grande 
 do Sul, cutting the lower part of the Rio do Rasto, yielded me valves 
 of a small Estheria almost circular in outline and certainly distinct 
 from the Cyzicus ( Estheria) draperi so common in Argentina. 
 
 It is important to note that as yet no forms ascribable to the 
 Permian have been found in the sediments of the Sao Bento series 
 which must be considered as of Trias-Rhxtic age. 
 
 (13) The Serra Geral Eruptives or Traps and the Associated Sediments 
 
 Crowning the Botucatu, or the Rio do Rasto, where the former 
 can not definitely be differentiated therefrom, comes a great capping 
 of lavas, basaltic to andesitic in composition, while in certain locali¬ 
 ties sediment must have been in course of deposition during inter¬ 
 vals of quiescence to show up now as intercalations between the 
 lavas, though only in the lower part of the volcanic group. Elsewhere, 
 as in western Parana and Santa Catherina and in parts of Uruguay, 
 only volcanic material is to be found. 
 
 Attention must be drawn to Woodworth’s account 95 dealing with 
 the trap plateaus of Santa Catherina, and particularly to the valu¬ 
 able summary by C. L. Baker entitled ‘ ‘ The Lava Fields of Parana. ” 96 
 As shown in Baker’s map, the eruptives of the Parana basin form a 
 complete elongated area, 1,300 km. long, but of rather variable 
 width, 500 km. across at the broadest, narrowing toward the south- 
 
 92 Walther (1919), p. 107, footnote. 94 Leme (1924), p. 189. 
 
 93 White (1908), pp. 217-225. 95 Woodworth (1912), pp. 91-99. 
 
 96 Baker (1923); the scale on the map is incorrectly given. 
 
90 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 west and occupying a tract of fully 800,000 square kilometers in ex¬ 
 tent, including certain stretches concealed beneath the Bauru beds. 
 Excepting on the west, the lavas generally terminate in erosional es¬ 
 carpments ; they frequently give rise to terraced features and are 
 responsible for the multitude of rapids and cataracts on the rivers, out¬ 
 standing among which are the magnificent falls of the Iguassu 97 with 
 a height of 60 meters and those of La Guayra on the upper Parana. 
 
 While the contact between the volcanics and the sediments 
 beneath is as a rule a perfectly even surface, attention must be drawn 
 to certain irregularities first studied in the town of Taquara, north 
 of Porto Alegre in Brazil, and seen again close to the railway near 
 Paso del Cerro and in the Cuchilla de Tambores io km. south of 
 Tacuarembo in Uruguay. These disclosed the fact that the basalts 
 in these three spots were resting upon and filling up shallow hollows 
 in an originally uneven surface of sandstone, which depressions in 
 every case trended a little south of west, a feature worthy of record, 
 since in South Africa several cases are known of a similar kind along 
 the Basalt-Cave sandstone contact; these apparently represent 
 primitive drainage lines over the sandy surface in existence at the 
 commencement of the period of volcanicity. The section shown in 
 Figure 6 may therefore be not an uncommon one. 
 
 Flows earlier than those usually seen were found deep down in 
 the Botucatu sandstone at Ityrapina in Sao Paulo, where the deep 
 boring gave the following downward section, starting at the base of 
 the main traps: sandstone, 85 meters; basaltic flow, 15 meters; sand¬ 
 stone, 77 meters; basaltic flow, 3 meters; sandstone becoming redder, 
 more clayey, and calcareous, and hence transitional to the Pyramboia, 
 48 meters. In South Africa such a similar case has been recorded 
 from the Maclear district of the Cape Province, where local eruptions 
 antedated the dominant period of volcanicity. 
 
 Thin sandstone intercalations are recorded by Baker generally, by 
 Walther in Uruguay, and by Sobral and Hausen in Misiones, and are 
 almost always medium to fine grained—sometimes finer and coarser 
 banded—red, red-brown, terra-cotta, or even purplish colored, 
 poorly or cross bedded sandstones or not unusually intensely hard, 
 glassy quartzites breaking with a conchoidal fracture and so making 
 very stony outcrops. While probably much of this induration was 
 effected by heat from the overlying lava, the depth through which 
 such hardening extends rather suggests the action of hot solutions 
 carrying silica. On the other hand, it should not be overlooked that 
 they might possibly have acquired some of this exceptional character 
 through original silicification prior to the outpouring of the succeed¬ 
 ing lava-flow, for it is well known that under an arid climate sands 
 may be converted into “surface quartzites” or “silcretes” under 
 
 87 See Hausen (1919), PI. IV. 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 
 
 91 
 
 M—I 
 
 O 
 
 8.3 
 
 P ^ 
 
 3 <D 
 
 m’C 
 
 • »H >-k 
 
 ■4^ O 
 a3 Td 
 
 I * 
 
 § s 
 
 rt a) 
 
 d a 
 
 ° J2 
 fcuO 'So 
 d w> 
 •43 0:3 
 
 w 
 
 CD cC 
 
 *H .» 
 
 s 
 
 M a3 
 O O 
 
 03 ft 
 o 2 
 
 ’o 
 
 .a 3 
 
 $ ^ 
 o3 CD 
 
 •r- 1 t-i 
 
 >< o3 
 
 H - 
 
 ►o 
 
 5 h w 
 Pi - 
 
 © 
 
 M 
 
 d W 
 •g o3 
 
 g.d 
 
 2 o 
 ^ 3 
 
 CQ 
 
 Ifl 
 
 M C /3 
 
 w £ 
 
 •S'S 
 
 <D 
 
 <v > 
 
 *rj .!h 
 
 2 crt 
 
 C t /5 
 
 c3 o3 
 
 o 6 
 
 _ <U 
 O u 
 O 
 
 (* g 
 d « 
 
 Vh Pi 
 d +> 
 d 
 
 CT I '~' 
 03 ~ 
 
 ^ to 
 rj V 
 M d 
 d 0 
 O 1« 
 
 ^ d 
 -d w 
 
 |s 
 
 a; cS 
 
 Pt 
 a « 
 ,rH o 
 
 So 
 
 M VrH 
 
 Pi 
 O w 
 
 M 
 
 <d 
 
 a s 
 
 .2 a) 
 
 O'? 
 <D 0) 
 CO to 
 
 I 'd 
 
 vO 
 
 6 
 p—I 
 
 pH 
 
 0) 
 
 Pi 
 
 purely natural conditions, as in the Kala¬ 
 hari in South Africa, and that the Botu- 
 catu was apparently accumulated under 
 such desert conditions. This is suggestive 
 because of the very similar silicified char¬ 
 acter beneath the basalts of the Triassic 
 Forest sandstone of Rhodesia, also a for¬ 
 mation of eolian and presumably of desert 
 origin. A noteworthy type of rock exten¬ 
 sively used as ballast on the railways at 
 Santa Maria is a coarse gray grit, which 
 from inquiry I afterward found came from 
 Barras, on the railway close to Passo 
 Fundo, apparently from an intercalation 
 relatively high up in the volcanics, though 
 of this I can not be sure. It is remarkable 
 in that the samples picked up contained 
 tiny “dreikanter. ” Conglomerates have 
 been reported to occur in the Botucatu, 
 and these are manifestly worthy of de¬ 
 tailed study in view of the information 
 they might yield regarding the geographi¬ 
 cal and climatic conditions in Brazil at 
 the close of the Triassic epoch. 
 
 Above Taquara, on the road to Morro 
 Fortaleza, not far above the base of the 
 lava group, there crops out an agglomer¬ 
 ate bed (Fig. 6, j), with a fine, pinkish, 
 sandy matrix in which was set an abund¬ 
 ance of subangular to rounded fragments 
 of different kinds of lavas, which bed 
 passes upward into red cross-bedded 
 friable sandstone exactly similar to those 
 of the Rfo do Rasto seen in the town below, 
 and ending in a sharply defined ‘ ‘ baked 
 junction against the flow overlying; the 
 band totals 8 meters in thickness. Ten 
 meters higher up occur thin sandstone 
 veins running vertically down into cracks 
 in the underlying basalt from a higher 
 thin intercalation of red sandstone; clearly, 
 just as in South Africa, eolian conditions 
 seem to have lingered on while the earliest 
 lavas were being erupted. 
 
92 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 Some of the thin intercalated sandstones and quartzites in the 
 basin are reported to contain fragmental igneous matter, and indeed, 
 in the Territory of Misiones in northeastern Argentina, such layers 
 may be tufaceous in their character. 97 Such bands have been passed 
 through in certain of the borings, for example in those at Posadas, 
 Solari, and Mercedes (Corrientes). It is worthy of note that tufa¬ 
 ceous layers should also have been found at a lower horizon (in the 
 Rio do Rasto) in the bore-hole at San Cristobal (Santa Fe) still 
 farther to the west, where the volcanics were struck beneath the 
 Neogene at 73 5 meters, the Triassic sediments being cut at 900 meters, 
 underlain by possibly Bonito beds from about 1,141 to 1,384 meters. 
 A rather similar succession was intersected between 860 and 1600 
 meters in the boring at Tostado, a little to the northeast, which helps 
 to link up the strata of the Parana basin with the Paganzo of the 
 Cordoba and La Rioja districts. In Laguna Paiva No. 2, a little to 
 the north of Santa Fe, the amygdaloids were struck at 602 meters 
 and not bottomed at 638 meters. 
 
 (14) The Volcanics 
 
 As various observers have pointed out, many of the flows are 
 relatively coarse-grained, compact rocks difficult to distinguish from 
 intrusive sheets, but as a rule such types, when satisfactorily exposed, 
 display vesicular structure at the top, if not at the very base. Though 
 not recorded by previous workers, I found “pipe amygdaloids” 
 developed at the bottom of the flows in several places and at Taquara 
 also “vesicle cylinders” and “bubble trains” set vertically in non- 
 vesicular basalt, just as can be instanced from the Stormberg series. 98 
 
 Some of the lavas are highly vesicular and even cavernous, Brazil 
 and Uruguay being noted for the large quartz-filled geodes and hydro- 
 lytes that have become weathered out in the red loamy soil; others 
 are very dense, in which case there is sometimes a columnar structure. 
 In color they range from black to dark green, purple, brown, red, or 
 chocolate. Native copper is occasionally to be found in them. The 
 thickness of the individual flows varies from 0.5 meter upward, the 
 general value being probably anything up to 50 meters or thereabouts, 
 though Woodworth has spoken of sheets from 95 up to perhaps 190 
 meters thick in the eastern part of Rio Grande do Sul. There are no 
 references to the occurrence of thin layers of red bole or earth between 
 successive effusions, and in this respect the volcanics are like those 
 of South Africa, for which the evidence points to the rapid welling out 
 of flow after flow without any lengthy time interval, probably under 
 a dry climate such as would have hindered the rapid weathering of 
 the lava crusts. 
 
 As regards the thickness of the group, Baker 99 states that in 
 different parts of the basin figures of from 250 to 500 meters have 
 97 Hausen (1919). 98 Du Toit (1907). 99 Baker (1923), p. 72. 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 93 
 
 been measured, although much rock must have been removed by 
 erosion from over this great region. In one of the borings at Curuzu 
 Cuatia, on the lower Parana River, between Corrientes and Concor¬ 
 dia, the base of the volcanics (from which intercalated sediments 
 were absent) was not passed through at 486 meters even. In this 
 respect, therefore, the group is comparable with the Stormberg 
 basalts, wherein a thickness of over 1,000 meters has been recorded 
 in Basutoland and 400 meters in the Kaokoveld. 
 
 (15) Petrology 
 
 The bulk of the flows are “melaphyres” in which olivine is 
 absent or else in small individuals, usually much altered to serpentine, 
 or, as in Uruguay according to Walther, to iddingsite, while the 
 structure is in general ophitic to subophitic, features that exactly 
 duplicate those of the Stormberg volcanics. Hausen observed no 
 porphyritic structures in Misiones, but Walther mentions types 
 approaching the porphyrites in Uruguay, while Baker refers to 
 augite-porphyrite and Branner records the presence of such types in 
 Rio Grande do Sul and Hussak in Santa Catherina and Sao Paulo. 
 This is instructive because, although porphyritic basalts are unrep¬ 
 resented in the Union of South Africa, they appear occasionally up 
 the Zambezi-Linyanti Valley and are well marked in the Kaokoveld 
 of Southwest Africa. Andesitic varieties are sometimes present in 
 Brazil, as at Santa Maria. 
 
 Doubtless the lavas issued mainly from fissures, now occupied 
 by narrow vertical dikes of basaltic or doleritic rock, but at Santa 
 Maria what was taken to be a volcanic neck was observed in the face 
 of the spur north of the town, facing Pedreira and overlooking the 
 railway to Passo Fundo. It cuts the hard interbedded sandstone 
 capping the ridge and is filled with a pale scoriaceous rock, banded in 
 nearly vertical fashion parallel to the wall of the pipe. 
 
 Regarding the age of the Serra Geral eruptives, it can hardly be 
 doubted that, like those in western Argentina and Patagonia, they 
 are Rhaeto-Liassic, though direct evidence thereon is wanting. 
 
 (16) The Intrusive Dolerites or Diabases 
 
 Penetrating the strata of the Santa Catherina system, but more 
 commonly its lower half, are these well-known extensive sheets and 
 dikes of basic rock that are manifestly closely allied to the Serra 
 Geral eruptives, though, so far as can be gathered, of slightly more 
 basic composition. Numerous references will be found in the geolog¬ 
 ical literature to such sills, but no comprehensive account thereof. 
 Baker has indicated on his small-scale map the areas outside the 
 basin proper where they have been observed by him, and he notes 
 that they also penetrate the basement formations in Paraguay and 
 the southern part of Matto Grosso. 
 
 They are common in the Itarare, Bonito, and Iraty beds, but the 
 
94 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 impression that I obtained from journeys between Porto Alegre 
 through Santa Maria and Rivera to Montevideo, where excellent 
 sections are furnished by numerous railway cuttings, was that such 
 intrusions were distinctly scarce in the upper part of the system and 
 then generally dike-like, just as is the case in the Karroo. Further, 
 it may be affirmed that the relative proportion of igneous rock is 
 much less than in the dolerite-riddled portion of the Union of 
 South Africa. Their precise relationship to the eruptives has not yet 
 been worked out, but arguing from analogy from South Africa 1 , they 
 would most probably have been injected into the framework of South 
 America during or probably just after the eruptive phase, and hence 
 probably date from the beginning of the Jurassic. 
 
 They display no unusual petrological characters, ranging from 
 dolerites without olivine to olivine-dolerites, and having subophitic 
 or ophitic structures. A typical sample from a bore-hole near Rio 
 Claro in Parana possessed a specific gravity of 3.02. ' 
 
 SECTION L. THE NORTHEASTERN REGION OF BRAZIL 
 
 Mainly to Lisboa, 2 Crandall, 3 and Small 4 is due our knowledge of 
 the quite considerable area of Permo-Triassic in Maranhao, Piauhy, 
 and the northeastern part of Goyaz. Quite unjustified is its southerly 
 extension, as shown on Branner’s geological map of Brazil, so as to 
 join up with the strata of the Parana basin, the formations filling the 
 broad gap being either pre-Carboniferous or post-Triassic, though 
 probably the latter may conceal some Gondwana beds. 
 
 The strata are lying nearly flat, resting on crystallines, but in the 
 east, according to Small, unconformably on the thick Serra Grande 
 series, described elsewhere (possibly of early Palaeozoic age), and are 
 overlapped to the north by the Cretaceous and the coastal Tertiaries. 
 
 The lowest beds, of Permian age, make a broad, sweeping horse¬ 
 shoe, extending from the north of the Pamahyba River southward 
 through Piauhy, then westward to the Rio Tocantins and down the 
 latter to the borders of Para. The higher horizons lie within the 
 curve, the Triassic appearing toward the north and covering a 
 limited area about the headwaters of the Mearim River in Maranhao. 
 
 Called by Lisboa the ‘ ‘ Pamahyba series ” and divided by him into 
 several groups totaling a few hundred meters in thickness, so far as 
 one can gather, the beds consist successively of gray calcareous shales 
 and gray to white sandstones, a thick red sandstone having purple 
 spots and with pisolitic limestone, a gray sandstone followed by green 
 and chocolate shales, and limestones with cherty bands. Glacial 
 signs are absent at the base. Psaronius associated with gymno- 
 spermous wood has been found in many localities in this wide region 
 throughout the full thickness of strata, together with the cast of a 
 stem referred to Sigillaria in Piauhy. Since the first-named occurs, 
 
 1 Du Toit (1920). 2 Lisboa (1914). 3 Crandall (1910). 4 Small (1914). 
 
GONDWANA SYSTEM IN BRAZIL, URUGUAY, AND PARAGUAY 95 
 
 according to Derby, in Sao Paulo State associated with Stereosternum 
 and Lycopodiopsis, such would suggest that the Parnahyba series 
 probably represents horizons from about the Iraty upward and hence 
 mainly of Permian age, the beds being all of fresh-water origin, 
 apparently. 
 
 The succeeding “Mearim series” consists of a few hundred 
 meters of white, splotched, and red sandstones overlain by flow’s of 
 basic amygdaloidal lavas, which are also intercalated in the red 
 sandstones, the strata building table-topped hills, well represented 
 between Grajahu and Pastos Bons, while they and the underlying 
 beds are, furthermore, cut by intrusive dolerite in the form of sheets 
 and dikes. 
 
 There can be no doubt that this upper series, older than the 
 Cretaceous and so like the Sao Bento of southern Brazil, is of Triassic 
 age, support for which is found in the discovery of scales of Semio- 
 notus near Floriano on the Parnahyba River. 
 
 Attention might furthermore be directed to the general parallel¬ 
 ism in facies with the ‘ ‘ Karroo ” succession of the Low'er as contrasted 
 with the Upper Congo basin in Africa, namely, in the absence of 
 glacials, the shaly and often slightly calcareous character of the 
 Lualaba beds with fish and crustacean remains, and the succeeding 
 yellow or pink friable Lubilache beds, though volcanic rocks are not 
 known to overlie the latter. 
 
 SECTION M. EASTERN BRAZIL 
 
 In the coastal region of Bahia and Sergipe, Branner 5 has described 
 under the name of the “Estancia beds” certain strata, which he 
 placed in the Permian, that occupy the region from Bahia nortlrward 
 to the Rio Sao Francisco, though much concealed by Cretaceous and 
 Tertiary formations. Subsequently Soper 6 has removed a part of this 
 to the Cretaceous and has furthermore stated that the Estancia beds 
 pass down, apparently wdthout a break, into a thick series of rocks 
 that rest unconformably upon an uneven surface of crystallines and 
 proterozoic (?) quartzites. It totals no less than 3,700 meters, the 
 beds being tilted and sometimes considerably disturbed, and even 
 faulted, the foldings trending generally a little east of north, but 
 along other axes as w^ell, overlain discordantly by gently inclined 
 Cretaceous and by horizontal Tertiaries. 
 
 In the low r er part are seen shales, slates, and gray to pink lime¬ 
 stones, followed by a great mass of green, blue, dark or dark red, often 
 micaceous, shales, succeeded by red sandstones and by blue lime¬ 
 stones. The finding of Alethopteris branneri White in a dark shale 
 beneath red Cretaceous sandstone at Aracy, and according to Leme 7 
 of Psaronius near to Conselho, would suggest an upper Carbonifer¬ 
 ous or more likely a lowrer Permian age for the upper part of the series; 
 
 6 Branner (1913). 6 Soper (1914), pp. 25-37. 7 Leme (1924), p. 181. 
 
96 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 consequently the basal portion may extend well back into the 
 Carboniferous. 
 
 Comparisons may be instituted with the gently tilted formations 
 in the western part of the Belgian Congo, particularly the middle 
 portion of the Kundulungu, called by Cornet the “Mpioka series, ” 
 with its red coloration. 
 
 It is just possible that the horizontal limestones and shales of the 
 Salitre Valley in Bahia, extending over wide areas to the southwest 
 thereof, resting unconformably upon much older formations, and 
 tentatively placed by Branner 8 in the Permian, may belong to the 
 Estancia series, but palaeontological data are essential before any 
 further advances can be made. In the coastal zone to the south of 
 Bahia near Cannavieiras beds occur on the Rio Pardo and Rio 
 Jequitinhonha that have yielded plants referred by Hartt to Astero- 
 phyllites (?) scutigera, suggested by D. White as perhaps a Phyllo- 
 theca. 9 
 
 Nevertheless, the following important points have been estab¬ 
 lished: (i) The presence of some undoubted late Palaeozoic continen¬ 
 tal beds in the coastal region; (2) their partially calcareous nature; 
 and (3) their disturbance by early mesozoic movements not alto¬ 
 gether coinciding in trend with the ancient “Brasilides, ” the signifi¬ 
 cance of which will be remarked on in Chapter VI. 
 
 SECTION N. THE CRETACEOUS OF THE INTERIOR 
 
 Brief reference only will be made to the continental Bauru beds 
 of Sao Paulo State that rest horizontally and unconformably on the 
 Triassic traps and are probably of Cretaceous age. Their precise 
 distribution being not very well known, these beds are not indicated 
 on the accompanying geological map, but it might be remarked that 
 they occupy wide areas, forming much of the higher ground border¬ 
 ing the Parnahyba, Rio Grande, Tiete, and Paranapanema, and 
 extend down the valley of the Parana to Guayra, while beyond at 
 San Ignacio near Posadas the sandstones have been brought down to 
 river-level by warping. The transgressive base is indicated by the 
 fact that they rest on the traps between Bauru and Agudos, but on 
 the Triassic sandstone at Botucatu. They are composed chiefly of 
 massive and often false-bedded white to red sandstones with a cal- 
 careo-argillaceous cement. Farther north they form a broad covering 
 along and west of the Rio Sao Francisco, known here as the “Pira- 
 pora sandstone, ” and appear to connect up with the slightly younger 
 but marine Cretaceous of rather similar character of Piauhy and 
 Ceara. These beds rest very largely upon a peneplain cut across the 
 underlying formations of Brazil, and thus reveal the very consider¬ 
 able denudation of the continent during the Jurassic epoch. 
 
 8 Branner (1913), p. 626; (1919) p. 237; Crandall (1919), p. 244. 9 White (1908), p. 423. 
 
CHAPTER VI 
 
 GEOLOGICAL HISTORY OF THE AFRO-AMERICAN 
 
 LAND MASS 
 
 That these two continents were intimately connected during 
 several geological epochs will, I venture, be acknowledged after the 
 perusal of the preceding pages, though the manner of such union 
 would admittedly be speculative. In preparing this review an 
 attempt was made first of all to write the historical account, irrespec¬ 
 tive of any hypothesis as to the manner of such union or of the 
 ultimate mode of separation of the land-masses though it became evi¬ 
 dent, as the data were assembled, that they pointed very definitely in 
 the direction of the displacement hypothesis, and that they could 
 most satisfactorily be interpreted in the light of that brilliant 
 conception. 
 
 It was finally decided to treat the subject from this viewpoint 
 rather than from any other, but only in such a way as to bring out 
 particular features the significance of which would otherwise be 
 overlooked; the hypothesis itself could then be discussed indepen¬ 
 dently and without prejudicing this particular chapter. 
 
 THE SILURO-DEVONIAN 
 
 Knowledge concerning the early limits of the Afro-American 
 mass only becomes definite during the late Silurian, when the ocean 
 must have covered the central Sahara, western Africa, the lower 
 portion of the Amazon Valley, parts of Bolivia, Jachal, San Juan, and 
 Mendoza, but not the Nevado de Famatina. The faunas up to 
 the very close of that epoch were, it is essential to note, boreal in 
 their aspect. 
 
 Attention has more especially to be focussed on far distant 
 Bolivia, because of the light which its splendidly developed lower 
 Palaeozoics shed on the precise age of the unfossiliferous sandstone 
 formations of Parana and the Cape, customarily ascribed to the 
 lower Devonian. The valuable accounts of the stratigraphy of this 
 country by d’Orbigny, 10 Steinmann, 11 Ulrich, 12 Knod, 13 and others, 
 supplemented recently by Kozlowski 14 and Bonarelli, 15 clear up 
 many points, although showing the need for further investigation. 
 
 Following Knod and Kozlowski, the normal Bolivian succession, 
 resting usually on the Ordovician, is divisible into: (i) The lower or 
 Ida sandstone, about 170 meters thick, with unfossiliferous lower 
 
 10 D’Orbigny (1842). 12 Ulrich (1893). 14 Kozlowski (1923). 
 
 11 Steinmann (1904). 13 Knod (1908). 15 Bonarelli (1921), pp. 39-56. 
 
 97 
 
98 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 half (a greenish gray or whitish quartzose sandstone overlain by 
 gray calcareous sandstone and calcareous flagstone; (2) the Conu- 
 laria and Crinoid beds some 200 meters thick (dark shales and sandy 
 shales with calcareous nodules, like those of Parana); (3) the upper 
 or Huamampampa sandstone up to 250 meters thick (grayish and 
 unfossiliferous). These 700 to 800 meters of beds are generally 
 agreed upon as representing the lower Devonian alone (Oriskany), 
 but Kozlowski has stated that around Lake Titicaca there is 
 (4) the Sicasica formation (quartzites of a warm color, micaceous 
 brown or white shales with sandy layers and black siliceous nodules), 
 with a fauna showing middle Devonian forms together with a few 
 lower Devonian types followed by unfossiliferous sandstones and 
 shales—Onondaga-Hamilt on. 
 
 Now, both in Parana and the Cape the sequence is almost identi¬ 
 cal, the barren Furnas and Table Mountain sandstones being respec¬ 
 tively followed by the dark fossiliferous Pont a Grossa and Bokkeveld 
 shales, which are unquestionably equivalent to the Conularia and 
 Crinoid shales of Bolivia, as borne out by their particular faunas. 
 
 On the other hand, Kozlowski 16 and Bonarelli 17 have independ¬ 
 ently pointed out that in parts of Bolivia this important system 
 actually includes in its basal section strata that belong to the upper¬ 
 most Silurian, and which, although of no great thickness, are wide¬ 
 spread to the north and southeast of Sucre, being characterized by 
 the presence of Clarkeia antisiensis (Liorhynchus bodenbenderi ) and 
 several other forms not found associated with the typical Devonian 
 faunas. At Tarabuco, east-southeast of Sucre, the succession resting 
 upon the Ordovician consists of 40 meters of barren white quartzite, 
 30 meters of yellow to red sandstones with Clarkeia , overlain by some 
 hundreds of meters of yellowish sandy shales, in the upper part of 
 which fossils characteristic of the Icla shales occur; at Finca Con¬ 
 st ancia to the north of Sucre a rather similar order can be seen. That 
 is to say, where the normally unfossiliferous basal group of whitish 
 quartzitic sandstones has developed a marine facies, the latter now 
 carries not Devonian but Silurian forms, wherefore the presumption 
 is strong that the basal part of the equivalent Furnas and Table 
 Mountain sandstone is in each case of uppermost Silurian age. It is 
 significant that the only molluscan remains yet found in the Cape 
 formation—unfortunately not identifiable—should occur at the 
 very base of the sandstone. 
 
 A parallel is forthcoming from Jachal, where Keidel, 18 studying 
 the unbroken succession in the Lomas de los Piojos, has recorded the 
 zone with Clarkeia antisiensis and Monograptus priodon, composed of 
 
 “Kozlowski (1923), pp. 8-10. 
 
 “Bonarelli (1921), pp. 46-53. 
 
 18 Keidel (1921), pp. 32-37, 51-54; Clarke (1913), P- 332. 
 
GEOLOGICAL HISTORY OF THE AFRO-AMERICAN LAND MASS 99 
 
 green and dark micaceous shales, which are followed by 800 meters of 
 “Devonian,” the basal 500 meters being unfossiliferous sandy 
 shales, greenish and bluish gray sandstones with dark sandstone ribs, 
 followed by similar gray and green beds, but with the typical forms 
 of the “austral” Devonian of Bolivia, Parana, Falklands, and Cape. 
 
 THE DEVONIAN OSCILLATIONS 
 
 These illuminating sections in Bolivia prove, therefore, that the 
 great “Devonian” transgression really began just prior to the close 
 of the Silurian. At that period Central Africa had been land probably 
 as far northward almost as Nigeria and southward right to the Cape, 
 while most of eastern and southern Brazil, Uruguay, southeastern 
 Bolivia, and northern, central, and eastern Argentina were bound 
 to that continent. 
 
 Along the margins of this land-mass were then accumulated the 
 siliceous deltaic deposits of Parana, Matto Grosso, Sierra de la 
 Ventana, the Falklands, and the Cape (from Clan william in the 
 west to Zululand in the east), while, as the ocean advanced and the 
 area of this section of Gondwanaland decreased, the delta fringe 
 crept inward and marine muds and sands with a lower Devonian 
 fauna of austral type came to be laid down on them. In Africa, the 
 Gold Coast region was ultimately submerged, and in the opposite 
 direction the southern third of the Cape and the whole of the Falk¬ 
 land Islands; whether the Cordoba-San Luiz region remained as 
 an island is unknown, but certainly Jachal and Salta were inundated, 
 though probably not the Famatina, Puna de Atacama, or the center 
 of La Rioja. 
 
 A cool climate is expressly indicated by the glacial zone of the 
 Table Mountain series, and indirectly by the dark gray, blue, and 
 green coloring of the marine strata, by the absence of limestones, and 
 scarcity of corals in them; everywhere a shallow ocean has to be 
 postulated, which would explain the widespread transgression of its 
 waters in certain territories. 
 
 THE DEVONIAN FAUNAS 
 
 With the lower Devonian, for reasons not yet understood, came 
 strong faunal differentiations. The boreal element still predominates 
 in the Sahara, 19 as shown by species from Rhineland and North 
 America, but an austral bias is signaled by such forms as Homolo- 
 notus herscheli , Leptocoelia flabellites, Spirifer rousseauxi , etc. The 
 Maecuru (Amazon) assemblage, while showing alliances with the 
 European, displays the North American—Oriskany-Onondaga— 
 facies, but the austral element is nevertheless quite marked in certain 
 
 19 Haug (1905); Lemoine (1913), p. 19. 
 
100 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 groups, particularly among the trilobites, with peculiarities among 
 the brachiopods, e.g., the Spirifers, which are more marked still in 
 the Bolivian and west Argentine areas. The fauna of Matto Grosso 
 is not far removed from that of Bolivia, but the west Argentine one 
 again stands apparently closer to that of the Falklands than to 
 Parana. Clarke 20 has commented on the fact that the Falkland fauna 
 is much closer allied to the Bokkeveld of the Cape than to the South 
 American assemblages, peculiar only so long as one disregards the 
 doctrines of the displacement hypothesis, for under the latter these 
 islands would have been nearer geographically to South Africa than 
 to Brazil. We know that some high ground must have lain in early 
 Devonian times to the west and northwest of the Cape (sufficient to 
 support local glaciers), while the sources of the siliceous formations of 
 Sierra de la Ventana and Parana probably lay to the east rather than 
 the west. Consequently the Parana region might well have occupied 
 an embayment with a promontory intervening between the Brazilian 
 and Falkland shore-lines. 
 
 Our knowledge concerning much of the periphery of Gondwana- 
 land ceases with the lower Devonian, but during the mid-Devonian 
 the ocean persisted in the Amazon region and in northern Bolivia. 
 The Brazilian Erere fauna shows increased affinities with the North 
 American, in addition to forms from the Maecuru, but it also has 
 alliances with the Sicasica fauna of Bolivia through such forms as 
 Tropidoleptus carinatus and Homolonotus dekayi , typically Hamilton 
 species, the latter represented also in Gold Coast Colony. 21 In Ahnet 
 (Sahara) the mid-Devonian carries Agoniatites vanuxemi , characteris¬ 
 tic of the Marcellus stage of New York, while the upper Devonian of 
 Tassili shows affinities with North America. Faunas of late Devonian 
 age are not yet known in South America outside of the Amazon 
 Valley. 
 
 y THE CARBONIFEROUS EPOCH 
 
 The oscillating shallow seas then withdrew—probably the conti¬ 
 nent rose isostatically—for in South Africa, the Falklands, and the 
 Sierra de la Ventana area siliceous estuarine and fluviatile beds of 
 presumably upper Devonian-lower Carboniferous age (derived in 
 the first-named from the erosion of the old Devonian land-mass) 
 succeed the marines with but a few plant remains. At Paracas on 
 the coast of Peru are the only other known plant-bearing strata, 
 probably of middle Carboniferous age. 
 
 Only in southern Peru, a little north of Lake Titicaca, 22 are marine 
 beds represented in South America, of undoubted lower Carbonifer¬ 
 ous (Avonian) age, though present in Gold Coast Colony and the 
 Sahara, and it would appear that the ocean then retreated to beyond 
 
 20 Clarke (1913), pp. 13, 56. 21 Kitson (1925). 22 Douglas (1920). 
 
GEOLOGICAL HISTORY OF THE AFRO-AMERICAN LAND MASS 101 
 
 its limits during the Silurian, thus presenting an increased area of 
 land to erosion. A certain amount of folding occurred in the pre- 
 Andine regions about this time, seemingly with the injection of 
 granites in Peru and Bolivia. 
 
 With the upper Carboniferous came a return of the ocean, which 
 entered the lower Amazon region and occupied a large part of Peru 
 and Bolivia, but the block between the Puna de Atacama and Potosi 
 remained land, around which a gulf of the sea must have crept to 
 penetrate into the western part of San Juan and parts of Chile on 
 about the same parallel of latitude; all the country to the northeast, 
 east, and southeast of San Juan probably remained unsubmerged. 
 
 The upper Carboniferous occurs on either side of the Amazon 
 River with wide extension—a group of limestones, shales, and sand¬ 
 stones, perhaps 600 meters thick, with an abundant fauna, including 
 Productus cor a , P. semireticulatus, Athyris subtilita, Spirifer earnera- 
 tus , etc., allied to the upper Carboniferous of the United States and 
 Urals and showing resemblances to the Carboniferous of Bolivia and 
 Peru on the one hand and of the Sahara on the other. Most curiously, 
 as Doctor Reed has well pointed out, the equivalent fauna from 
 Barreal is quite distinct from those recorded elsewhere in North or 
 South America, the forms being nearly all either closely allied to or 
 comparable with established species from the lower Carboniferous of 
 Europe or the middle or upper Carboniferous of Russia or parts of 
 Asia. Probably, therefore, the western part of San Juan was not in 
 direct communication with the ocean to the north. The fauna, it 
 might be noted, displays no affinities with those of eastern or west¬ 
 ern Australia. 
 
 THE CARBONIFEROUS GLACIATION 
 
 Time is revealing more and more the number of places where the 
 / Carboniferous glaciers passed into a shallow ocean and, melting, 
 I dropped their burden, for example in Tasmania, New South Wales, 
 Western Australia, the Salt Range in India, the San Juan district in 
 Argentina, and Parana in Brazil. It has furthermore become evi¬ 
 dent that over wide areas the ice must have traveled across a gently 
 undulating surface nearly at, if not below, sea-level, as is proved by the 
 shales with marine mollusca in Parana and Southwest Africa, by the 
 Dwyka shales of South Africa, Bonito coal measures of Brazil, etc. 
 The general low relief of the ice-ridden territory is indeed one of the 
 most unexpected peculiarities of this Palaeozoic glaciation. 
 
 I have adduced evidence to show that in South Africa 23 the glaci¬ 
 ation apparently started at several centers, in each case over some¬ 
 what higher ground, but that the ice-sheets at the maximum stage of 
 glaciation seem to have coalesced, to discharge southward into a 
 
 23 Du Toit (1921). 
 
102 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 body of probably fresh water. An important center was situated in 
 Southwest Africa, but it is not at all unlikely that a similar feature 
 lay to the northwest or even west of the Cape (from which direction 
 had indeed come the local glaciers of the Devonian), suggestive being 
 the “boulder pavement” at Eland’s Vlei on the Doom River, with 
 its groovings directed nearly west-east. 
 
 In South America the evidence consistently favors an ice center 
 situated out in the present Atlantic, the significance of which will be 
 appreciated from the study of Figure 7. The position allotted to the 
 Falkland Islands in this geographical scheme is determined approxi¬ 
 mately by a series of rather pretty comparisons: (1) The folding is 
 along two axes, one directed west-northwest and the other at right 
 angles thereto, but of fairly open character as a rule, with dips else¬ 
 where generally low over wide areas, and thus comparable with the 
 Ceres-Clanwillian region, and neither with the Caledon-Riversdale 
 sections nor with Argentina, where the first-mentioned set of foldings 
 is dominant and intense; (2) the lower Devonian sandstones are 
 gritty and pebbly, as in the Cape-Clanwilliam region or in Parana, 
 and not quartzitic, as in Mossel Bay or in the Sierra de la Ventana, 
 moreover, the glacial horizon of the western Cape may, according to 
 Baker, just possibly be represented in West Island; (3) the fauna of 
 the Devonian shales in the Falklands is very closely related to that 
 of the Bokkeveld, less so to that of Parana; (4) the Devono-Carbonif- 
 erous quartzites of the Falklands are identical with the Witteberg 
 series of the Cape and not closely comparable with the greywacke 
 phase of the Sierra de la Ventana; (5) the strongly cleaved tillite of 
 the islands is slightly disconformable to these quartzites, as in Clan- 
 william, whereas the glacials of the Rio Sauce Grande are conform¬ 
 able to the strata below, just as in the south of the Cape (Worcester); 
 
 (6) the tillite is of great thickness, with some intercalated shales, and 
 hence like that of the southern Karroo and the Rio Sauce Grande; 
 
 (7) the Lafonian sandstone, just above the tillite, has no lithological 
 representative in the Cape, but finds its equivalent in the Itarare 
 series of Uruguay and Brazil; (8) the absence of the “Mesosaurus 
 zone” links the islands with Argentina; (9) the great development 
 of Lafonian sediments devoid of coals recalls the facies of the Sierra 
 de Pillahuinco and the Karroo of the southern Cape rather than that 
 of the northern Karroo or of Parana; (10) the upper Lafonian beds 
 show lithological resemblances with the upper Karroo and not the 
 Triassic of South America; (11) the few dolerite intrusions suggest 
 again the Cape-Clanwilliam region. 
 
 If we admit the displacement hypothesis, these affinities would 
 collectively assign the Falklands to a position along the northern 
 edge of the broad belt of foldings that linked the Cape with Argen¬ 
 tina, but would place them somewhat nearer to Africa. Despite the 
 
GEOLOGICAL HISTORY OF THE AFRO-AMERICAN LAND MASS 103 
 
 fact that these islands rise from the Patagonian coastal shelf, and 
 that the late Mesozoic foldings (Patagonides) strike southeastward 
 near the Gulf of San Jorge, and therefore toward them, as remarked 
 by Windhausen, 24 the Falklands, if we ignore the mere tectonic 
 parallelism, display not the slightest affinities with Patagonia, where 
 the late Triassic volcanics and sediments rest upon the crystallines, 
 with no representatives of the Palaeozoics. It must accordingly be 
 concluded that this outpost of Gondwanaland belongs to the north¬ 
 east rather than to the west and actually affords very striking sup¬ 
 port to the displacement hypothesis. The record by Schwarz of 
 granite, probably from volcanic ejectamenta, in Tristan da Cunha is 
 an indication of crystalline plutonics underlying part of the South 
 Atlantic basin. 
 
 The north-south directed groovings beneath the tillite in the 
 Falklands would lend a modicum of support to such an hypothetical 
 ice-center situated out in the Atlantic, though impinging on Brazil, 
 as indicated by the common absence of the glacials in the southern 
 corner of that country, by the fluvio-glacials in Santa Catherina, 
 Uruguay, and San Juan, and the marine glacials in Parana and near 
 Barreal in San Juan. The conditions in Goyaz and Matto Grosso are 
 not yet known, while it is uncertain that ice formed at this date in 
 Bolivia, since the glacials of the Bermejo series may perhaps be of 
 Permian age. 
 
 In 1921 , 25 1 ventured to point out that the evidence in South Africa 
 favored the deduction that the several centers had become glaciated 
 in turn, the earliest on the west, the latest on the east. Both in New 
 South Wales (Seaham area) and in San Juan Rhacopteris and Car diop¬ 
 ter is are present between the lowermost glacials, while such is not the 
 case in Brazil, South Africa, India, Victoria, or Tasmania. Although 
 the palaeontological evidence on all these points is not as definite as 
 could be desired, it would appear that glaciation commenced in the 
 middle or early upper Carboniferous in the Southern Hemisphere in 
 the Seaham area, extended to Argentina, then to South Africa, to the 
 west of Australia, Tasmania, and India during the upper Carbonifer¬ 
 ous, and recurred during the early Permian in New South Wales 
 and perhaps still later in Bolivia and the eastern Congo. 
 
 Under the displacement hypothesis, not only would all the areas 
 affected be brought together in the most simple manner, within an 
 oval perhaps no larger than the African continent, but the non¬ 
 synchronism of glaciation in the several parts would find their 
 explanation as the natural outcome of an extensive active ice-cap, 
 the margins of which progressively advanced in one direction and 
 retreated in the other, so that with time the sheet had as a whole 
 migrated over the surface, glaciating various sections in turn, exactly 
 
 24 Windhausen (1921a), p. 6. 26 Du Toit (1921). 
 
104 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 as seems to have been the case with the Pleistocene Ice Age in the 
 Northern Hemisphere. This view would perhaps meet Coleman’s 
 well-founded criticism of the excessive size of the single ice-cap under 
 this supposition. 26 
 
 The causes of this intense refrigeration, lasting through about 
 an entire geological epoch, are still largely speculative, but some 
 additional suggestions may be thrown out. I had tentatively 
 advanced the hypothesis of “eccentric ice-capping” to explain the 
 growth and migration of ice centers, but a second aspect, not yet 
 brought forward, is a possible variation in the crustal heat of the 
 earth locally, all explanations having hitherto sought for such causes 
 outside the lithosphere. Admitting the probability of Joly’s theory 
 of the connection between radioactivity and diastrophism, it would 
 follow that periods of liquefaction of the subcrust and the abnormally 
 rapid escape of internal heat by volcanism and hypabyssal injection 
 would be succeeded by intervals of quiescence, endothermic reac¬ 
 tions, and reduced heat transference. A certain section of the crust 
 might thereupon come to have over, a period an exceptionally low 
 geothermic gradient, under which refrigeration would at least be 
 facilitated, provided that climatic and other well-recognized predis¬ 
 posing agencies were concurrently present. That the entire crust 
 of the globe should have been simultaneously affected in any such 
 manner is, however, almost unthinkable. 
 
 CARBO-PERMIAN SEDIMENTATION 
 
 During the closing stages of glaciation, the upper Carboniferous 
 ocean, with its characteristic Spirifer and Productus fauna, occupied 
 the Amazon region, Peru, and Bolivia, and penetrated into the 
 western part of the San Juan. In some extraordinary fashion, the 
 remarkable Eurydesma fauna of the Salt Range and Hunter River 
 area makes its appearance in Southwest Africa, though by what 
 channel this specialized lamellibranch form reached this spot is 
 unknown, but presumably via the Cape, for a route across Africa 
 from Mozambique, while not yet proved impossible, is rather unlikely. 
 
 In southern Brazil, Uruguay, Paraguay, and western Argentina 
 deltaic deposits were laid down, the “Glossopteris flora” flourished, 
 and the coals of Brazil were formed; the areas marginal thereto 
 received estuarine, lacustrine, or “basin” deposits, for example the 
 Cape, Southwest Africa, and eastern Argentina. A deepening of the 
 waters then drowned these deltaic or shallow-water deposits, except 
 in western Argentina, Bolivia, and Africa north and east of the Cape, 
 in which directions the ground presumably stood higher, and within 
 this immense area of open water the black, sulphureted, and often 
 limy muds of the Iraty-upper Dwyka shales were, as suggested by 
 
 26 Coleman (1926), p. 260. 
 
GEOLOGICAL HISTORY OF THE AFRO-AMERICAN LAND MASS 105 
 
 the writer, deposited in an estuary or land-locked sea like the Gulf of 
 Bothnia. 27 The deeper, sulphureted waters were unfavorable to 
 bathyal life and almost the only forms were fishes, the swimming 
 Mesosaurus , and Crustacea. The development of bedded limestones 
 in Sao Paulo indicates communication in that direction with the 
 northern Permo-Carboniferous ocean—a broad region of marine 
 deposition in the Triassic after an intervening period of elevation 
 and erosion. 
 
 In Natal, Transvaal, and Rhodesia 28 deltaic coal measures were 
 laid down on an extensive scale, into which crept a few relics of the 
 northern flora, introduced presumably by way of Brazil, though fail¬ 
 ing to reach Rhodesia or India. In the west and south of the Cape, in 
 the Falklands, and in eastern Argentina the deposits were of the kind 
 called “basin” by the writer for want of a better term—a monoto¬ 
 nous succession of fine sands and silts, often ill-differentiated, with 
 occasional “marls” and calcareous nodules, laid down on interior 
 plains or in very wide and shallow depressions subject to flooding and 
 desiccation like those of Turkestan. The conditions were, however, 
 favorable for vertebrate life, as shown by the wonderful reptilian and 
 amphibian fauna of the Beaufort beds of southern Africa, the forms 
 appearing earliest in the southwestern comer of the Karroo and 
 spreading thereafter to the northeast, through East Africa to Mada¬ 
 gascar, India, and even Russia, accompanied by members of the 
 Glossopteris flora. 
 
 THE PERMO-TRIASS1C MOVEMENTS 
 
 Hitherto crustal movements had been mild and generally in the 
 vertical direction over this section of Gondwanaland, but at the 
 close of the Permian a lengthy arc of compression (the Gondwanides) 
 came into being, extending east-west through the south of the Cape, 
 the Sierra de la Ventana, Sierra Pintada, and northward to the 
 Puna de Atacama, the strata therein becoming crumpled and in places 
 overturned to the north and northeast. That these may have been 
 “coastal” ranges is suggested by the presence of marine Triassic and 
 Liassic beds at a few localities outside the arc, whereas just inside the 
 area inclosed by the latter—a ‘ ‘ foreland ’ ’•—only continental deposits 
 have been found. During the early Triassic these border ranges 
 were, however, actively eroded, so that in the later part of that epoch 
 sediments and volcanics were being deposited unconformably upon 
 the tilted Devono-Permian succession, both in the Cape and Argen¬ 
 tina, to the south of the fold-belt in places upon the pre-Devonian 
 basement, and to the north and northeast thereof upon only slightly 
 disturbed Permian, and hence without marked discordance as a rule. 
 
 27 Du Toit (1921). 
 
 28 The conditions are more fully discussed by Du Toit (1926), Chapter XXI. 
 
106 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 THE ARIDITY OF THE CLOSING TRIASS1C 
 
 The climate, which presumably had generally been warm and 
 dry, appears to have finally become more and more parched, until 
 toward the close of the Triassic the material being deposited over 
 wide areas in Brazil, Uruguay, Paraguay, and the southern half of 
 the African continent consisted of fine-grained, cream, pink, or red 
 sands, sometimes with well-rounded and polished grains and with . 
 marked false-bedding, stuff which must largely have been of desert 
 origin, e.g., the Botucatu and Rio de Rasto in the Parana Basin, the 
 Forest-Bushveld-Cave sandstone of the Union, Southern and North¬ 
 ern Rhodesia, the Lubilache of the Congo basin, and probably the 
 Adigrat beds of Abyssinia, covering areas totaling several millions of 
 square kilometers. 
 
 Lithologically these various deposits in no small degree resemble 
 the Quaternary Pampean loess of Argentina, with its prevalent pink¬ 
 ish, sandy material with calcareous concretions, largely windbome 
 in its origin and of great thickness in places; in other respects the 
 Tertiary-Quaternary Kalahari beds of Bechuanaland, with fluviatile 
 and dune sands overlying in places marls, clays, and some con¬ 
 glomerates. The fact that the conditions in Europe during the 
 closing part of the Triassic were largely desert would almost suggest 
 a period of universal desiccation, although animal life could by no 
 means have been scanty. Fluviatile and lacustrine phases have 
 yielded the well-known Thinnfeldia flora (together with Estheria ), 
 significantly on the margins of this gigantic region, i.e., western 
 Argentina, Chile, Cape Province, peninsular India, Tonkin, New 
 South Wales, Queensland, and Tasmania. 
 
 THE MESOZOIC ERUPTIONS 
 
 Outstanding too, is the enormous scale of the basic effusions at 
 the close of the Triassic, under which vast territories in the two con¬ 
 tinents were flooded with basalts, in places to depths of over 1,000 
 meters, i.e., the Parana basin, the limited region in Maranhao, the 
 much-dissected relics in western Argentina, together with extensive 
 outpourings, mainly of acid character, in Neuquen and Patagonia, 
 the Stormberg lavas of the Union of South Africa, Lebombo, South¬ 
 ern Rhodesia, lower Zambezi Valley, and Kaokoveld. In practically 
 every case thin sedimentary intercalations are to be found in the 
 basal part of the volcanic groups, certain of which indicate the per¬ 
 sistence of arid conditions. 
 
 Conspicuous too are the accompanying sills and dikes of dolerite 
 (diabase) that, avoiding the folded region, ramify through the 
 Gondwana beds or penetrate the basement in the Karroo region, the 
 Zambezi Valley, Tasmania, Antarctica, the Parana basin, and possi¬ 
 bly Guiana, and that were injected during or immediately after the 
 eruptive phase, probably during the Liassic. 
 
GEOLOGICAL HISTORY OF THE AFRO-AMERICAN LAND MASS 107 
 
 THE JURASSIC TRANSGRESSION 
 
 The upward transfer of these enormous volumes of magma led 
 to instability along the borders of Gondwanaland and to the trans¬ 
 gression of the Liassic and Jurassic ocean across its margins, and the 
 breaking-up of that continent seemingly dates from this period. We 
 find fossiliferous marine beds with dolomite and gypsum in Tunis 
 and Abyssinia, a wide development of Jurassic through British East 
 Africa, Tanganyika Territory, and Madagascar, while on the oppo¬ 
 site side of the land-mass about the border of Chile and Argentina 
 the Liassic is characterized by eruptive matter—rhyolite, dacite, 
 and spilite—while marine beds of this epoch covering many stages 
 were laid down along an enormous stretch of the Andine region. 
 
 THE CRETACEO-TERTIARY 
 
 It is a striking fact that no early marine Cretaceous should 
 fringe the coasts either in the section between Bahia Blanca and 
 Bahia or that between Cape Agulhas and Angola, suggesting, indeed, 
 that the two continents may have still been united at the commence¬ 
 ment of that epoch, although Madagascar and India must almost 
 certainly have parted from Africa by that time. The early Cre¬ 
 taceous ocean nevertheless assuredly trespassed over the southern 
 extremity of Africa, and also covered parts of Neuquen and the 
 Andine region. 
 
 The mid-Cretaceous, so far as can be elucidated the Turonian, 
 saw a great renewal of orogenic movement in Africa, directed east- 
 west over the wide belt traversed by the earlier Cape foldings, with 
 considerable marginal fracturing both there and in the Natal- 
 Zululand coastal area. Evidence of synchronous crustal movements 
 has been discovered in Neuquen, with arc-like flexures bridging the 
 gap between the Gondwanides and the (posthumous) Andine fold¬ 
 ings, while correlated flexures strike southward through Chubut, 
 curving eastward to the mouth of the Rio Deseado, disturbances 
 that have been called the “ Patagonides ” by Windhausen. 29 Then 
 came a period of extensive erosion, after which the later Cretaceous 
 ocean transgressed into those regions, entering by an arm situated 
 between the Sierras of Buenos Aires and the Patagonian platform, 
 as far inland as the Sierra Pintada. 
 
 The weight of evidence would ascribe the base-leveling in the 
 Sierra de la Vent ana and the accumulation of the ‘ ‘ high-level gravels 
 to the well-established Eogene transgression that followed the first 
 phase of the Andine movements, while the upheaval thereof would 
 synchronize with the second phase and, preceding the “Parana” 
 ingression, date from the early Pliocene. The correspondence with 
 the southern Cape is hence remarkably close. 
 
 28 Windhausen (1924), p. 202. 
 
108 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 A wealth of information will be found in the writings of Wind- 
 hausen on the palaeogeography of northern Patagonia, the outstand¬ 
 ing item being the fact that the Cretaceo-Tertiary (San Jorge) fauna 
 of the Rio Negro and Chubut is of South Atlantic and Antarctic 
 facies and does not show North Atlantic affinities. 
 
 The little that is to be gleaned from Eastern Brazil is confirma¬ 
 tory. The late Tertiary lacustrine deposits stand at a relatively high 
 level, e.g. y Curitiba, Sao Paulo, and near Diamantina, and point to a 
 comparatively recent uplift, while the physiography of the Parahyba 
 Valley between Sao Paulo and Rio de Janeiro strongly suggests that 
 the lengthy crystalline chain of the Serra da Itajuba running parallel 
 to the coast is a fault scarp with downthrow on the ocean side. In 
 South Africa, too, the general movement of the continent has been 
 in an upward direction to the amount of several hundred meters. 
 
 THE ANDINE MOVEMENTS 
 
 Under the displacement hypothesis the South American conti¬ 
 nent is viewed as having proceeded to drift westward during the 
 latter part of the Mesozoic, crumpling up in its path the marine sedi¬ 
 ments bordering the western side of the Brazilian “shield, ” a process 
 that led to the extravasation of magma and the injection of hypabys- 
 sal and plutonic matter into that belt of extreme compression and of 
 upheaval. The Tertiary foldings and overthrustings, directed usually 
 toward the east, that characterize the Cordillera and extend from 
 Venezuela to Terra del Fuego, find their explanation in logical fashion 
 under this hypothesis, the eruptive matter being regarded not as the 
 cause of such movements, but as the product of such compression 
 upon the zone underlying this belt of crumpling, an action that has, 
 moreover, not yet ceased. 
 
 Hindered by the birth of these western ranges, the drainage of 
 the continent would have been reversed, while the spasmodic ele¬ 
 vation of the chains would have caused oscillations in the level of the 
 slowing earth-block and encroachings and retreats of the waters of 
 the Atlantic. In the broadening gap between Africa and South Amer¬ 
 ica the presumably basaltic ocean-floor with its accumulating mantle 
 of pelagic deposits would normally have been in tension, and found¬ 
 ering of sections thereof is a possibility that can be invoked to explain 
 the volcanicity of the South Atlantic basin, as exemplified by its 
 many volcanic islands. 
 
 Much could be written on the Tertiary history of this vast ocean 
 basin having an intimate and vital bearing on the subject under 
 review, but enough has, I venture, been set down to impress the 
 reader with the exceptionally illuminating and outstanding character 
 of this stimulating theory of earth structure, undoubtedly the most 
 fascinating of geological hypotheses to-day. 
 
CHAPTER VII 
 
 BEARING ON THE DISPLACEMENT HYPOTHESIS 
 
 While the general geological resemblance between those portions 
 of the two continents that face the South Atlantic basin has long 
 been perceived, the outcome of these present studies is essentially to 
 emphasize this geological parallelism, which is nothing less than 
 extraordinary, considering the enormous stretches of ocean parting 
 these two land-masses. 
 
 Such points of resemblance have now become so numerous as 
 collectively almost to exceed the bounds of coincidence, while they 
 are, moreover, confined not to one limited region nor to one epoch, 
 but implicate vast territories in the respective land-masses and 
 embrace times ranging from pre-Devonian almost to the Tertiary. 
 Furthermore, these so-called “coincidences” are of a stratigraphical, 
 lithological, palaeontological, tectonic, volcanic, and climatic nature. 
 
 Of prime importance , moreover , is that evidence obtainable from the 
 study of the phasal variations displayed by particular formations when 
 traced within their respective continents. 
 
 In illustration, let us consider the case of two equivalent forma¬ 
 tions, the one in South America beginning on or near the Atlantic 
 coast at A and extending westward to A' and the other in Africa 
 starting similarly near the coast at B and stretching eastward to B'. 
 Then it can be affirmed that more than one such instance can be 
 designated, where the change of facies in the distance A A' or BB' is 
 greater than that found in AB, although the full width of the Atlantic 
 intervenes between A and B. In other words, these particular forma¬ 
 tions along the two opposed shores tend to resemble one another 
 more closely than either one or both of their actual and visible 
 extensions within the respective continents. With the multiplication 
 of such examples, drawn from more than one geological epoch, such 
 a singular relationship can no longer be regarded as wholly fortuitous 
 and a definite explanation therefore has accordingly to be sought. 
 An analysis, moreover, shows that this unexpected tendency is 
 equally marked, whether the formations involved be marine, deltaic, 
 continental, glacial, eolian, or volcanic. 
 
 If, on the other hand, the two land-masses are pictured as having 
 been moved closer together, as in Figure 7, a great number of obser¬ 
 vations and deductions are now found to be brought into apparent 
 harmony, and these possible “coincidences” are disposed of in the 
 simplest fashion. 
 
 This is precisely what the displacement hypothesis effects, there¬ 
 by providing a simple explanation of many otherwise puzzling 
 
 109 
 
110 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 observations. The fact that many eminent scientists have cast doubt 
 upon its geophysical possibility should not be permitted to cloud the 
 issue any more than that the existence of former “land bridges’' 
 should be denied because of cogent objections based upon the doc¬ 
 trine of isostasy. 
 
 It is not proposed to discuss here the physical basis of that 
 hypothesis, nor is it desired to deal with this problem as a whole, 
 such as has been done by Wegener. 30 The intention is merely to set 
 forth some of the data regarding Africa and South America and to 
 state the conclusions to be drawn therefrom, that are distinctly 
 awkward of explanation under the current and orthodox view of 
 “land bridges,” but which, on the contrary, appreciably favor the 
 “hypothesis of continental disruption.” Incidentally, some few 
 observations will be made having a bearing upon other parts of 
 Gondwanaland, that in turn suggest lines of future research in those 
 particular countries. 
 
 Of prime importance is the extraordinarily close correspondence 
 in the outlines of the opposed shores of the two continents, as has 
 been pointed out and discussed by others long before Wegener, and 
 which is particularly marked when comparison is made not with 
 maps, but on the face of a terrestrial globe. Next is the presence of the 
 central Atlantic rise beneath the ocean, with its surprisingly sym¬ 
 metrical position nearly midway between the Old World and the New. 
 
 Interpreted mathematically , the great regularity of these three 
 features, extending through the entire length of the South Atlantic, 
 would betoken an enormously high probability that such features had 
 owed their origin to one and the same set of tectonic forces at a 
 relatively late geological period. Upon this rise are, furthermore, 
 aligned certain of the volcanic islands of the southern Atlantic. This 
 otherwise profitable subject must, however, be relinquished in favor 
 of the more momentous geological aspect. 
 
 Commencing in the south, the following relationships can be 
 considered as more or less established: 
 
 * (I) The section south of Bahia Blanca and that below the Zuurberg 
 
 in the Uitenhage district show several points of agreement in that: 
 (a) The upper Triassic (i) is predominantly of volcanic origin, (2) 
 rests discordantly upon Permian or older beds that are affected by 
 Permo-Triassic movements, (3) is influenced by mid-Cretaceous 
 disturbances following more or less the older tectonic axes, and (4) 
 is overlain by marine Cretaceous and Tertiary, the former of which 
 contains variegated clays and is in places gypsiferous. In the Cape 
 the volcanics are, however, basic, in Argentina largely acid, though 
 in Mendoza and San Juan basaltic flows are represented. ( b ) The 
 lower Cretaceous in northwestern Neuquen on the southern side of 
 
 30 Wegener (1924). 
 
BEARING ON THE DISPLACEMENT HYPOTHESIS 
 
 111 
 
 the “ Gondwanides ” contains a typical “Uitenhage” invertebrate 
 fauna, this being the only part of South America where such a 
 faunal assemblage has yet been found, and like that in the Sundays 
 River Valley has been tilted before late Cretaceous times. ( c ) There 
 are great developments of nearly flat marine Tertiaries from Eocene 
 to Pliocene in age, in which most strikingly an “Atlantic” fauna is 
 not at all or only partially represented. 
 
 (II) The ranges north of Bahia Blanca undoubtedly correspond 
 with the Cape Fold ranges, in each case showing: (a) Intense fold¬ 
 ing, with overturning toward the north or northeast, strata up to the 
 Permian being involved. ( b ) The quartzites of the Sierra de la Ven- 
 tana correspond lithologically with the Table Mountain sandstone 
 of the region about Mossel Bay and George, (c) The fossiliferous 
 Devonian is like that of the latter region, but the equivalent of the 
 Witteberg series is composed of greywackes instead of hard white 
 quartzites. ( d ) The thick, folded and cleaved glacials in Argentina 
 duplicate the characters of the Dwyka tillite and are conformable 
 with the beds below, as in the Cape south of latitude 33° S. (e) In 
 Argentina, dark shales overlie the glacials, as in the Cape, and pass 
 up into beds like the Ecca of the southwestern Karroo, only some¬ 
 what more quartzitic, but in the first-named, just as in Pondoland, 
 the horizon with Mesosaurus is not recognizable. (/) Dolerite intru¬ 
 sions are absent in each case, (g) Consolidated ferruginous “gravels” 
 rest on bevels cut across the folded beds on the northern flanks of the 
 Sierra de la Vent ana and duplicate in most extraordinary fashion 
 the early Tertiary “high-level gravels” of the south of the Cape, 
 having similarly been raised and dissected at a subsequent date. 
 
 ( h ) Traced northwestward from Sierra de la Vent ana into San Juan 
 and Jachal, the lower Devonian quartzites have become darker and 
 less like the Table Mountain sandstone, the fossiliferous slates 
 greener, the equivalent of the Witteberg is apparently represented by 
 grits and greywackes, the glacials now rest unconformably on the 
 older divisions and show characters very different from those in 
 the Cape, at the same time becoming compound; tfie plants 
 include northern Carboniferous forms, while the succeeding strata, 
 the equivalent of the Ecca, become quite unlike those of the west¬ 
 ern Cape. 
 
 (III) To the northeast of the Sierra de la Ventana the “Gond- 
 wanides” are fading out, and at Olavarria and Tandil the pre- 
 Devonian rocks are only slightly affected by them, though a folding 
 transverse thereto, trending northeastward, can be made out, thus 
 paralleling the conditions between Clan william and Namaqualand. 
 Closer comparison is impossible, since the Pampean of the lower 
 ground round about conceals these older groups. 
 
 (IV) Comparing the region stretching from Uruguay to Minas 
 
112 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 in Brazil with that between Clan william and the Kaokoveld, we find: 
 (a) The almost horizontal Furnas sandstone of Parana is similar to 
 the equivalent Table Mountain sandstone of Clan william and Van 
 Rhynsdorp. (b) Each is thinner and softer than in the south and each 
 carries small quartz pebbles, (c) Each is succeeded by the marine 
 Devonian shales with calcareous nodules—Ponta Grossa shales and 
 Bokkeveld beds—with closely allied austral faunas, (d) The soft 
 Tibagy sandstone with Spirifers duplicates the “ Fossiliferous sand¬ 
 stone” of the Bokkeveld of Ceres and Clan william, (e) The base of the 
 Itarare glacials is unconformable and transgressive, exactly as is the 
 case with the Dwyka in Calvinia, which stratigraphical relation is 
 maintained toward the north. (/) There is irregularity in the charac¬ 
 ter and thickness of the glacials in each country, owing to the 
 uneven nature of the floor, etc. (g) The glacials are absent at the 
 base of the Gondwana system in Rio Grande do Sul and in Santa 
 Catherina, just as in parts of the Transvaal. Qi) A southeasterly 
 source is presumed for the Brazilian glacials, a northerly, north¬ 
 easterly, and perhaps northwesterly origin for those in the western 
 Cape and in Southwest Africa, (i) The Iraty shales are identical 
 lithologically and palaeontologically with the White band of the 
 Dwyka, each containing the reptile Mesosaurus, not known in other 
 parts of the world. The South African formation is fully developed 
 from Robertson across to Grahamstown and northward to Mariental, 
 but is not represented in Natal, the Transvaal or about Palapye; the 
 South American extends from Uruguay to Sao Paulo and westward 
 into Paraguay, but is unknown in the bore-holes in Argentina west 
 of the Parana River or in Bolivia. Thus the basin within which this 
 uncommon type of originally organic sulphureted mud was deposited 
 is fairly well defined and must have formed an estuary opening 
 probably to the north, where the facies becomes calcareous, a con¬ 
 clusion to which the presence of a marine phase in the glacials in 
 Parana and Southwest Africa would give support, (j) An important 
 difference is to be found through the Brazilian Bonito or “coal 
 measures ” underlying this horizon, whereas the Natal and Transvaal 
 Ecca “coal measures” succeed the latter, but I should like to record 
 that in the western Karroo I have found a zone of gray fissile 
 sandstones such as is unrepresented farther to the east or northeast, 
 in corresponding position in the upper Dwyka shales. Presumably 
 the Bonito is a deltaic phase that “invaded” the basin in which 
 the “post-glacial” deep-water shales were being laid down, (k) The 
 Estrada Nova, which shows resemblance to stages II and III of the 
 Paganzo, but not to the presumably equivalent horizon of Bolivia, 
 strongly recalls the Ecca of southwest Africa, which again is widely 
 different from that of the Cape, Transvaal, and Rhodesia. Silicified 
 wood is common in both countries, but marine mollusca of this age 
 
BEARING ON THE DISPLACEMENT HYPOTHESIS 
 
 113 
 
 have not yet been found in South Africa. (/) The higher Permian 
 beds are apparently absent from the Parana basin and from South¬ 
 west Africa, which has seemingly been the result of the positive 
 earth movements at the close of the Palaeozoic, (m) The Sao Bento 
 series is pseudo-conformable to the underlying strata in the Parana 
 basin, just as the Stormberg series rests without angular break upon 
 the upper Beaufort series in the north of the Orange Free State or 
 upon the Ecca in the central Transvaal, (i) Each series transgresses 
 across the lower Gondwanas, i.e., along the margins of the Parana 
 basin and in the region lying to the north and northeast of the Cape 
 Province; (2) strata with the “ Thinnfeldia flora” are developed 
 toward the base, almost solely in the south, i.e., in the Cape-Basuto- 
 land region and in western Argentina and the fresh-water crustacean 
 Cyzicus (. Estheria ) draperi is common to both these areas; (3) the 
 brilliantly colored Rio do Rasto parallels the Stormberg Red beds and 
 Bushveld marls and like the latter formations contains dinosaurian 
 remains belonging to allied forms; (4) the Botucatu and uppermost 
 portion of the Rio do Rasto agree lithologically with the Cave- 
 Bushveld-Forest sandstone, all of which have been formed largely 
 by eolian agency under an arid climate; silicification of the sand¬ 
 stones is common; (5) the succeeding and widespread volcanics are 
 interbedded in their basal portions with thin sandstone bands or 
 tufaceous sandstones; (6) the only places where the volcanics form 
 the actual coast are near Torres and in the Kaokoveld respectively ; 
 (7) the lavas are basalts with but little olivine or else basaltic andes¬ 
 ites, but, while ophitic and subophitic structures are the rule in the 
 Cape, a porphyritic structure is occasional in southern Rhodesia and 
 is dominant both in the Kaokoveld and in the Parana basin, while 
 tuffs are rare. They show bands of pipe-amygdaloid and were evi¬ 
 dently poured out in rapid succession, apparently under a dry climate 
 and presumably from fissures, for except in the Basutoland region 
 volcanoes are practically unknown; (8) these effusions terminate the 
 succession in both countries, (m) Widespread injections of dolerite 
 characterize the flat-lying Karroo of the Cape, Natal, and southwest 
 Africa and also the Parana basin, but are not found in the south of 
 the Cape, northern Transvaal, or Southern Rhodesia generally, nor in 
 southern or western Argentina. ( n ) Marine Jurassic and Cretaceous 
 beds are absent from along the opposite coasts. (0) Kimberlite and 
 melilite-basalt of presumably late Cretaceous age pierce the strata 
 in the form of pipes both in South Africa (for example, in Namaqua- 
 land and the Gibeon district) and in eastern Brazil (for example, in 
 the Serra das Lages 100 km. from Rio de Janeiro, 31 and at Patos, in 
 the western half of Minas, where a neck penetrates Triassic sand¬ 
 stone). (p) The Cretaceous of east-central Brazil is largely a conti- 
 
 31 Rimann (1915). 
 
114 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 nental deposit covering a huge region in the interior, overlapping the 
 Sao Bento volcanics and paralleling in certain ways the extensive 
 Kalahari beds of southwest Africa, (q) The detrital diamonds of south¬ 
 west Africa that have been derived from Tertiary marine beds near 
 Luderitz are quite unlike those won from the kimberlite pipes of 
 South Africa, but show crystallographic and physical points of 
 resemblance with the gems obtained in eastern Brazil from gravels 
 or from pipes, such as the Boa Vista Mine, piercing the Espinago 
 quartzites and apparently of pre-Gondwana age. (r) The pre-Devo¬ 
 nian beds consist of quartzites, limestones, and slates folded along 
 axes roughly parallel to the respective coasts and injected in places 
 by granitic bodies, i.e., the Bambuhy and Assunguy series, at inter¬ 
 vals from Bahia to Uruguay, and the Nama system from Luderitz 
 to Cape Agulhas. (5) Brouwer has called attention to the presence 
 of several centers of alkaline rocks, both plutonic and effusive, on 
 either side of the Atlantic, for example, in Southwest Africa the 
 important group of foyaites of the Granit Berg, etc., and the phono- 
 lites of the Klinghardt Mountains near Luderitz, and in Brazil the 
 well-known Poqos de Caldas north of Sao Paulo, the various similar 
 occurrences in the State of Rio de Janeiro, such as the Serra de 
 Gericino, Serra de Tingua, Cabo Frio, etc. 
 
 (V) Proceeding farther north we find: (a) Opposed to the dis¬ 
 turbed area of Cretaceous and Tertiary of Angola and Loanda are 
 those of Bahia and Sergipe, in which their equivalents, with similar 
 faunal elements of the Atlantic and Mediterranean types, display 
 somewhat less tilting, (b) Inland therefrom in Brazil are the folded 
 Permian and probably Carboniferous Estancia beds, with some 
 resemblances to the Kundulungu of the lower Congo region, while 
 beneath the latter lies a great calcareo-dolomitic formation, which 
 may perhaps be compared with the ancient limestones of Bahia. 
 
 (VI) Farther north is found a belt of Cretaceo-Eocene along the 
 coast from Natal westward that rises up to form the plateau along 
 the boundary of Ceara with Piauhy and is seen again in the northern 
 part of Maranhao. These strata can be compared with the beds of 
 approximately similar age in the coastal portions of Gold Coast, 
 Dahomey, and Cameroons, extending up the valleys of the Niger and 
 Benue and occupying a wide area to the north. In Nigeria the 
 Cretaceous is somewhat folded and faulted. Reference might be 
 made to the well-known volcanic line of the Cameroons that extends 
 southwest ward into the ocean, in view of the presence off the comer 
 of Brazil of the phonolite island of Fernando Noronha. 
 
 (VII) The Gondwana outlier of Maranhao and Piauhy consti¬ 
 tutes a fairly close parallel with the development known as the Lubi- 
 lache in the western part of the Congo Basin in ( a ) the absence of the 
 glacial group, (b) the absence of coals, (c) the equivalence apparently 
 
BEARING ON THE DISPLACEMENT HYPOTHESIS 
 
 115 
 
 of the upper red sandstones of Brazil with the cream and red friable 
 sandstones in the Congo. Mesozoic volcanics are, however, not yet 
 known in west-central Africa, whereas the Brazilian strata are over- 
 lain by basalts and penetrated by dolerites. 
 
 (VIII) The Silurian and Devonian strike south-south west ward 
 through the Sahara and southwestward through West Africa, with 
 isolated patches of the latter system in Gold Coast Colony, and are 
 apparently not much disturbed generally, a condition that can be 
 paralleled with the corresponding systems forming the syncline of 
 the lower Amazon Valley resting on the Archaean granite, with trend 
 oblique to the coast-line, the more so in that the Devonian in each 
 case includes, in addition to European elements, some North Ameri¬ 
 can forms, and also displays certain important affinities with the 
 austral fauna. Presumably land lay to the south, and along that 
 Devonian coast migration was possible toward Europe in the one 
 direction and toward Amazonas, Matto Grosso, and Bolivia in the 
 other. A point that might be noted in view of the diamondiferous 
 deposits of the Guianas is the frequency with which that gem occurs 
 in the coastal belt of West Africa, e.g., in Liberia and Gold Coast. 
 
 While the above can be taken as merely summarizing the evidence 
 on the subject, it will be clear to all acquainted with the literature 
 that many other points of similarity or else of analogy could readily 
 be discovered in the geology of these two land-masses. Sufficient, I 
 venture to think, has already been set down to bring out the aston¬ 
 ishing geological agreement between more or less comparable sections 
 of the respective coastal zones, from which it will be conceded that 
 the evidence so far would distinctly appear to favor the displacement 
 hypothesis rather than orthodox ideas. It is nevertheless highly 
 desirable that more impartially minded persons should criticize the 
 data here given and judge whether those amazing resemblances have 
 been correctly interpreted or whether some more rational explana¬ 
 tion for them could not be formulated. 
 
 Accepting provisionally this hypothesis , it will next be instructive 
 to attempt some graphic representation thereof. 
 
 Regarding the various possible dispositions of the continental 
 masses, it can first of all be remarked that actual contiguity of the 
 opposed shore-lines can most definitely be ruled out, and, secondly, 
 that even apposition of the borders of the continental shelves, as 
 favored by Wegener, may perhaps hardly be warranted, for utilizing 
 the line of reasoning based on phasal variation , the differences actu¬ 
 ally noticed between the various facies of the equivalent formations 
 where they come closest together are of such a degree as to demand 
 a fairly wide gap. One of the order , perhaps of from 400 to 800 km., 
 would indeed seem to he needed , if all the observed phenomena are to 
 he satisfactorily accounted for. 
 
116 
 
 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 - Cretaceous and 
 l Eocene 
 
 
 ✓ 
 
 . Condwana 
 
 
 «*—•— 
 
 Silurian , Devonian' 
 and Carboniferous 
 
 Late Eozoic and 
 e a r/y Pa la e oz oic ■ 
 
 • Gondwanides and 
 Cape fold in 3s 
 
 v 
 
 v V 
 
 V V 
 
 Granitic basement 
 
 PosPTriassic uplifts 
 
 Fig. 7—Suggested continental restoration under the Displacement Hypothesis. 
 
 - Limit of Mesosaurus 
 
 —.- And me fold mss 
 
 - BrasiUdes and 
 
 Post-Nama foldings 
 
BEARING ON THE DISPLACEMENT HYPOTHESIS 
 
 117 
 
 Such is schematically represented in Figure 7, which is tentatively 
 suggested as perhaps best meeting the needs of the case, though the 
 intervening space is actually a little less than the phasal variations 
 would rightly demand. 
 
 Further critical comparative studies should enable a better ori¬ 
 entation to be made, the diagram advanced being admittedly only 
 a first approximation. Granted, too, that such crustal slipping could 
 have taken place, it must not be overlooked that some regional dis¬ 
 tortion might have resulted during the drifting apart of the floating 
 continental blocks, an action inferred from the inflection of the 
 Andine foldings in the neighborhood of Cape Horn. 
 
 It will be noticed that in this figure the Falkland Islands have, 
 following the discussion in Chapter VI, been moved up into a posi¬ 
 tion between Cape Town and La Plata. Only now can the remark¬ 
 able correspondence of the various fold-lines, and even of the more 
 moderate archings, of different ages be properly appreciated. The 
 rupture of the Afro-American mass is also more than hinted at in the 
 strict parallelism of the zigzagging of the two coast-lines near Rio 
 de Janeiro and Angola respectively. It is surprising, too, though not 
 deliberately arranged, that the space between the two shores should 
 correspond so closely in its two boundaries with the plan of the 
 “central ridge” of the South Atlantic. These two sections of coast¬ 
 line, it should furthermore be noted, are composed of crystalline 
 rocks, whereas both to the north and south considerable stretches 
 can be found where late Tertiary and Quaternary marine sediments 
 fringe the shores and consequently tend to obscure the original out¬ 
 lines of the supposedly fractured masses. 
 
 Further elaboration is hardly necessary, and the diagram is 
 accordingly left to the criticism of the reader. 
 
CHAPTER VIII 
 
 CONCLUSIONS 
 
 Throughout this work the intention has been to detail the geo¬ 
 logical peculiarities of the opposed sections of the two land-masses, 
 to point out their resemblances and differences, and to draw such 
 legitimate conclusions in regard to the geographical, climatic, and 
 other circumstances in each country as the facts would appear to the 
 writer to warrant, though irrespective of any particular theory of 
 continental union during the past, for that they were so united is 
 indicated by a wealth of evidence of a practically indisputable 
 character. 
 
 But, having proceeded thus far, one could scarcely avoid making 
 use of these data and indulging in a certain amount of speculation as 
 to their possible manner of union during the later Palaeozoic and 
 early Mesozoic eras. Such an analysis, it can be affirmed, does not 
 favor the notion of one or more relatively narrow connecting links or 
 “land bridges” lasting down into the early Mesozoic, but on the 
 contrary supports the presumption of some continuous land area 
 embodying those sections of Gondwanaland that are now represented 
 in the two continents. It furthermore strongly favors the admittedly 
 revolutionary idea that geographically these two portions were 
 appreciably closer in the past; indeed, the evidence is, I think, of 
 sufficient weight to warrant such a viewpoint being adopted, as a 
 working hypothesis at any rate. Upon studying the remnants of 
 Gondwanaland in the other continents—India, Australia, etc.—we 
 find the facts and the deductions therefrom in similar accord, and 
 the impression becomes decidedly strengthened that the displace¬ 
 ment hypothesis, if not an actual explanation of the phenomena, 
 would at least seem to contain more than a germ of the truth, 
 despite its revolutionary and heterodox nature and apparent lack 
 of agreement with geophysical considerations. 
 
 Nevertheless, as already remarked, geological evidence almost 
 entirely must decide the probability of this hypothesis, for those 
 arguments based upon zoo-distribution are incompetent to do so, 
 being as a rule equally, though more clumsily, explicable under the 
 orthodox views involving lengthy land connections afterward sub¬ 
 merged by the oceans. 
 
 The hypothesis sets forth to explain in simple and logical fashion 
 a host of problems—geological, palaeontological, tectonic, climatic, 
 and biological—that have hitherto proved difficult or impossible of 
 satisfactory solution, and is incidentally serving to stimulate geo- 
 
 118 
 
CONCLUSIONS 
 
 119 
 
 logical thought in fashion almost unparalleled in the history of 
 that science. For example: 
 
 (1) It condenses all the sections of Gondwanaland with their 
 markedly similar formations and geological histories within the 
 compass of one moderately sized continent , instead of demanding a 
 land-mass very much larger than Eurasia. 
 
 (2) It brings into close association away from the equator all the 
 areas glaciated in the “Permo-Carboniferous” and thus succeeds in 
 eliminating an outstanding difficulty in necessitating somewhere or 
 other a refrigeration in the subtropics or tropics, a trouble inherent 
 under any other hypothesis, even in those postulating a wide move¬ 
 ment of the South Pole. 
 
 (3) It can explain the apparent non-synchronism of glaciation in 
 the several sections and many other observed peculiarities as pri¬ 
 marily due to the progressive migration of the main ice center or 
 centers. 
 
 (4) It simplifies the problem of the distribution of the ‘ ‘ Glossop- 
 teris ” and “ Thinnfeldia ” floras. 
 
 (5) It brings together the wide areas over which arid conditions 
 seem generally to have prevailed at the close of the Triassic. In 
 making a review of this problem nothing is more conspicuous than 
 the tendency under the orthodox viewpoint of overlooking the enor¬ 
 mous improbability that over widely separated parts of the globe, 
 situated in different latitudes and longitudes and under totally 
 differing oceanic environments (and therefore of atmospheric dis¬ 
 tributions) there should have coexisted huge areas throughout the 
 entirety of which the climate should consistently have at any one 
 epoch been either frigid or pluvial or arid. 
 
 (6) It assembles the earlier Mesozoic lava-fields. 
 
 (7) It would explain Gondwanaland as a well-defined and almost 
 isolated unit down to about the end of the Jurassic epoch, after which 
 the continental mass is viewed as proceeding to break up and the 
 sectors to begin drifting apart, with the Cretaceous ocean progres¬ 
 sively penetrating between the blocks and generally coming to over¬ 
 lap their borders in the fully established marine transgressions of 
 the late Mesozoic. 
 
 In the hypothesis as formulated by Wegener, the New World is 
 regarded as having parted from the Old up the full length of the 
 Atlantic from south to north at quite a late stage in the Tertiary, * 
 against which supposition many objections could be raised. Else¬ 
 where 32 I have tentatively suggested the radically different concep¬ 
 tion of two such parent continents, that to the south being Gond¬ 
 wanaland, which, by impinging on each other, gave rise to the enor¬ 
 mously extended orogenic structures made by the Atlas, Alps, 
 
 32 Du Toit (1921a). 
 
120 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 Carpathians, Caucasus, Himalayas, etc., or, as Suess has called 
 them, the “Posthumous Altaides. ” Concentrating attention upon 
 the southern continent, it can be pointed out that one of the deter¬ 
 mining factors in the conjectured breaking up of the mass may have 
 been the development around the continental “shield” of the 
 peripheral trough of subsidence or “fossa, ” within which sedimenta¬ 
 tion had proceeded through lengthy periods during Palaeozoic, Meso¬ 
 zoic, and even Cainozoic times, whereby an engirdling zone of crustal 
 weakness would have been produced favoring the centrifugal separa¬ 
 tion postulated. 
 
 These encircling oceanic deposits would thereafter have been 
 pushed up into folds in advance of the drifting crustal sectors, a 
 process that would ultimately have served to arrest their outward 
 movement, for, in conspicuous fashion, excluding “Angaraland, ” all 
 the known portions of Gondwanaland now lie within a practically 
 closed orogenic ring constituted by the Atlas, south European, 
 Iranian, and Himalayan folds on the north, the Malay, Polynesia, 
 and New Zealand crumplings on the east, the west Antarctica belt 
 on the soutlp and the Andine flexures on the west, curving eastward 
 into Venezuela. Along certain sections of this “orogen, ” to use 
 Kober’s term, intense compression and overthrusting then took 
 place, accompanied by considerable magmatic activity. 
 
 Such a view would, moreover, serve to explain the temporary 
 reunion of certain of the continents at some stage in the Tertiary, as 
 the facts of zoographical distribution would apparently demand. A 
 rather similar process is conceived to have operated in the Northern 
 Hemisphere, where the continental fragments lie within a great oval 
 ring of mainly Tertiary foldings, from Mexico to Alaska, Japan, the 
 Himalayas, and onward as before. These two orogenic belts are 
 pressed together in the section extending from the West Indies to 
 the East Indies, in which two regions distorted arcs with marked 
 inflections have originated from their supposed mutual interference. 
 As will be observed from a map, these two orogenic “fossae, ” when 
 rolled out into one plane, come to resemble a figure 8, while the 
 remainder of the surface of the globe constitutes the immense 
 basin of the Pacific, along the margins of which the two fold-systems 
 have crowded. 
 
 SOME GENERAL REMARKS 
 
 These pages have of necessity dealt in very summary fashion 
 with this fascinating subject, more particularly as related to those 
 territories beyond the respective opposed regions known personally 
 to the writer. It is therefore to be hoped that scientists acquainted 
 with those particular outside regions may be induced to set down 
 their observations on those areas, supporting or else refuting the pre- 
 
 ! 
 
BIBLIOGRAPHY 
 
 121 
 
 sumptions here put forward on behalf of this hypothesis, since a 
 strict and impartial criticism is indeed required if this riddle in early 
 history is to be deciphered. The cooperation of geologists, palaeon¬ 
 tologists, zoologists, botanists, and physicists is urgently needed, the 
 discussion advanced here dealing admittedly with only a very 
 limited, albeit extremely important, section of the globe, from which, 
 under the eclectic hypothesis of crustal instability, corroborative or 
 destructive evidence might reasonably be expected. By virtue of 
 their enormous lengths of opposed coast-line and extraordinary geo¬ 
 logical parallelisms, the two most favored continents from which evi¬ 
 dence is to be drawn would appear to be Africa and South America, 
 and, should the details herein set forth appear to be worthy of 
 serious consideration, it is to be hoped that more detailed investiga¬ 
 tions may shortly be instituted elsewhere on the two sides of the 
 Atlantic for the purpose of clearing up some of the many crucial 
 questions that must be regarded as sub judice. 
 
 Even if the facts collected should cause the displacement hypoth¬ 
 esis to be discarded, the researches made through its stimulus would 
 have served a most useful purpose in establishing data for the fuller 
 comprehension of this problem of the past history of our planet. 
 
 BIBLIOGRAPHY 
 
 Aguirre, E. 
 
 1879. La geologfa de la Sierra Baya. Anales. Sociedad Cientijica Argentina, 
 VIII, pp. 34 - 45 - 
 
 1891. La Sierra de la Ventana. Anales. Sociedad Cientijica Argentina, 
 XXXII, pp. 20-33. 
 
 Andersson, J. G. 
 
 1907. Contributions to the Geology of the Falkland Islands. Wiss. Ergebn. 
 Schwed. Siidpolar-Exp. 1901-3. Bd. 3, Lief. 2. Stockholm. 
 
 Arocena, E. Terra. 
 
 1926. Nota sobre el piso de Itarare y los sedimentos marinos del Rinc6n de 
 Alonso. Revista de la Asociacion Politecnica, Montevideo, pp. 1-20. 
 
 Av£-Lallemant, G. 
 
 1890. El Paramillo de Uspallata. Anales . Sociedad Cientijica Argentina. 
 Baker, C. L. 
 
 1923. The Lava Field of the Parand Basin, South America. Journal oj 
 Geology, XXXI, pp. 66-79. 
 
 Baker, H. A. 
 
 1923 ? Final Report on Geological Investigations in the Falkland Islands. 
 London. 
 
 Beder, R. 
 
 1923. Sobre un hallazgo de fosiles permicos en Villarrica. Boletin. Academia 
 
 Nacional Ciencias. Cordoba, XXVII, pp. 9-12. 
 
 Beder, R. and A. Windhausen. 
 
 1918. Sobre la Presencia del Devdnico en la parte media de la Republica 
 del Paraguay. Boletin. Academia Nacional Ciencias. Cordoba, 
 tom. XXIII, pp. 255-262. 
 
 Berry, E. W. 
 
 1922. Carboniferous Plants from Peru. Amer. Jour. Sci., Ill, p. 189. 
 
 1924. Mesozoic Plants from Patagonia. Amer. Jour. Sci., VII, p. 473. 
 
122 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 Bodenbender, W. 
 
 1894. Devdno y Gondwana en la Republica Argentina. Bol. Acad. National 
 Ciencias. Cordoba, XV, pp. 201-255. 
 
 1896. Sobre le edad de algunas formaciones carboniferas de la Republica 
 Argentina. Revista del Museo de la Plata, VII, p. 129. {•*-> 
 
 1902. Contribucion al conocimiento de la Precordillera de San Juan y 
 Mendoza y de las sierras de la Republica Argentina. Boletin. Acad. 
 National Ciencias . Cordoba, XVII, p. 203. 
 
 1905. La Sierra de Cordoba. Anales. Minist. Agricultura Argentina . Tom. 
 I, num. 2. 
 
 1911. Constitucidn geoldgica de la parte meridional de La Rioja y Regiones 
 Limitrofes. Bol. Acad. National Ciencias. Cordoba , XIX, pt. I, 
 
 pp. 1-220. 
 
 1912. Ibid. Anales. Minist. Agricultura Argentina, tom. VII, num. 3. 
 
 1916. El Nevado de Famatina. Bol. Acad. National Ciencias. Cordoba , XXI, 
 pp. 100-182. 
 
 Bonarelli, G. 
 
 1914. La estructura geol6gica y los yacimientos petroKferos del Distrito 
 Minero de Oran. Minist. Agricultura Argentina. Bol. num. 9, serie B. 
 1921. Tercera contribucidn al conocimiento geoldgico de las Regiones 
 Petroliferas subandinas del norte. Anales. Minist. Agricultura 
 Argentina, XV, num. I. 
 
 Brackebusch, L. 
 
 1891. Reisen in den Cordilleren der argentinischen Republik. Verhandl. 
 
 Gesell. Erdkunde Berlin, pp. 53-79. 
 
 1892. Mapa geologico del interior de la Republica Argentina (1 in 1,000,000). 
 
 Gotha. 
 
 Branner, J. C. 
 
 1910. 
 
 1913- 
 
 I9i5- 
 
 1919. 
 
 Brouwer, H. A. 
 
 1921. 
 
 Clarke, J. M. 
 
 1890. 
 
 1899. 
 
 1900. 
 
 1913- 
 
 Coleman, A. P. 
 
 1918. 
 
 1926. 
 
 Crandall, R. 
 
 1910. 
 
 1919. 
 
 Delhaes, G. 
 
 1913. 
 
 The Tombador escarpment, in the State of Bahia, Brazil. Amer. 
 
 Journ. Science, XXX, pp. 335-343. 
 
 The Estancia beds of Bahia, Sergipe, and Alagoas, Brazil. Amer. 
 
 Journ. Science, XXXV, pp. 619-632. 
 
 Geologia Elementar (Segunda edicao). Rio de Janeiro. 
 
 Outlines of the Geology of Brazil, to accompany the Geologic Map of 
 Brazil (also in Portuguese). Bull. Geolog. Soc. America, XXX, pp. 
 189-338. 
 
 De alkaligesteenten van de Serra do Gericino ten Noordwesten van 
 Rio de Janeiro en de overeenkomst der eruptiefgesteenten van 
 Brazilie en Zuid-Afrika. Konink. Akadem. Wetensch. Amsterdam, 
 XXIX, pp. 1005-1020. 
 
 The Trilobites of the Sandstone of Ererd and Maecuru, State of Pard, 
 Brazil. Archiv. Museu. Nation., Rio de Janeiro, IX, pp. 1-58. 
 Molluscos devonianos do Estado do Pard, Brazil (Portuguese and 
 English). Archiv. Museu. Nation., Rio de Janeiro, X, pp. 49-174. 
 The Palaeozoic Faunas of Para, Brazil. Albany, New York. 
 
 Fosseis Devonianos do Parand. Monographias do Serviqo Geologico e 
 Mineralogico do Brasil (Portuguese and English), vol. I. 
 
 Permo-Carboniferous Glacial Deposits of South America. Journ. 
 Geology, XXVI, p. 310. 
 
 Ice Ages, Recent and Ancient. The Macmillan Co., London. 
 
 Geographia, geologia, .... norte do Brazil, Ceara, etc. Publicacao 
 4, Inspectoria de Obras Contra as Seccas. Rio de Janeiro. 
 
 Notes on the Diamond Region of Bahia, Brazil. Econom. Geology , 
 XIV, no. 3, p. 220. 
 
 Sobre la presencia del rdtico en la costa patagonica. Minist. Agri¬ 
 cultura Argentina, Bol. num. I, serie B. 
 
BIBLIOGRAPHY 
 
 123 
 
 Delhaye, F., and M. Sluys. 
 
 1921. Les Calcaires du Bas-Congo, Bruxelles, pp. 1-79. 
 
 Derby, O. A. 
 
 1887. On Nepheline Rocks in Brazil. Quart. Journ. Geol. Soc. London, 
 
 XLII, pp. 457 - 473 . 
 
 1891. Ibid., XLVII, pp. 251-265. 
 
 1894. The Amazonian Upper Carboniferous Fauna. Journ. Geol., II, pp. 
 480-501. 
 
 1906. The Serra do Espinago. Journ. Geol., XIV, p. 374. 
 
 Dixey, F. 
 
 1925. The Geology of Sierra Leone. Quart. Journ. Geol. Soc. London, 
 LXXXI, pp. 195-220. 
 
 Douglas, J. D. 
 
 1914. Geological Sections through the Andes of Peru and Bolivia. Quart. 
 
 Journ. Geol. Soc. London, LXX, pp. 1-51. 
 
 1920. Geological Sections through the Andes of Peru and Bolivia, ibid., 
 LXXVI, p.41. 
 
 Draper, D. 
 
 1911. The Diamond-bearing Deposits of Bagagem and Agua Suja (Brazil). 
 Trans. Geol. Soc. South Africa, XIV, pp. 8-19. 
 
 1920. The High-level Diamond-bearing Deposits of Brazil. Revista de 
 
 Sciencias, Rio de Janeiro, IV. num. 1, pp. 1-11; num. 2, pp. 33-45. 
 
 1921. The High-level Diamond-bearing Breccias of Diamantina, Brazil. 
 
 Trans. Geol. Soc. S. Africa, XXIII, pp. 43-51. 
 
 Du Toit, A. L. 
 
 1907. Pipe Amygdaloids. Geolog. Magazine, IV, p. 13. 
 
 1920. The Karroo Dolerites of South Africa. Trans. Geol. Soc. S. Africa, 
 
 XXIII, pp. 1-42, 
 
 1921. The Carboniferous Glaciation of South Africa. Trans. Geol. Soc. S. 
 
 Africa, XXIV, pp. 188-217. 
 
 1921a. Land Connections between the other Continents and South Africa. 
 
 S. African Journ. Sci., XVIII, pp. 120-140. 
 
 1926. The Geology of South Africa. Edinburgh. 
 
 Evans, J. W. 
 
 1894. The Geology of Matto Grosso. Quart. Journ. Geol . Soc. London, L, 
 pp. 85-103. 
 
 Ferraz, L. C. 
 
 1921. ExcursSes scientificas no Estado de S. Catherina. Annaes de Escola 
 de Minas de Ouro Preto, num. 17, p. 1. 
 
 Fourmarier, P. 
 
 1924. Carte Geologique du Congo Beige. Revue Universelle des Mines. 
 7 e serie, tom. IV, no. 4, p. 182. Liege. 
 
 Geinitz, H. B. 
 
 1876. Ueber Rhatische Pflanzen und Tierreste in den argentinischen Pro- 
 vinzen La Rioja, San Juan, und Mendoza. Cassel. 
 
 Gothan. W. 
 
 1925. Sobre restos de plantos fbsiles procedentes de la Patagonia, con un 
 apendice; Plantas reticas de Marayes (Pirov. de San Juan). Bol. 
 Acad. Nacion. Ciencias. Cordoba, XXVII, pp. 197-212. 
 
 Gregory, J. W. 
 
 1916. Contributions to the Geology of Benguella. Trans. Roy. Soc. Edin¬ 
 burgh, LI, pt. Ill, no. 13, pp. 495-536. 
 
 Groeber, P. 
 
 1918. Estratigraphia del Dogger en la Republica Argentina. Minist. Agri- 
 cultura Argentina, Bol. 18, serie B. 
 
 1918a. Edad y extensi6n de las estructuras de la Cordillera entre San Juan 
 y Nahuel-Huapi. Physis. Buenos Aires, IV, pp. 208-240. 
 
 1922. Pdrmico y Tridsico en la costa de Chile. Physis. Buenos Aires, V, pp. 
 3I5-3I9- 
 
124 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 Guillemain, C. 
 
 1911. Zur Geologie Uruguays. Zeits. Deutsch. Geol. Gesell. Berlin, 63, 
 
 Monatsber., p. 203. 
 
 1912. Beitrage zur Geologie Uruguays. Neu. Jahrb. f. Min., Beil. Bd. 
 
 xxyi, p. 208. 
 
 Halle, T. G. 
 
 1912. On the Geological Structure and History of the Falkland Islands. 
 Bull. Geolog. Inst. Univ. Upsala, XI, pp. 115-229. 
 
 Harder, E. C. and R. T. Chamberlin. 
 
 1915. The Geology of Central Minas Geraes, Brazil. Journ. Geology, XXIII, 
 no. 4, pp. 341-378; no. 5, pp. 385-424. 
 
 Documents scientifiques de la Mission Saharienne, VIII, Paleonto- 
 logie, Paris, pp. 751-832. 
 
 On the Lithology and Geological Structure of the Sierra de Umango 
 Area, Province of La Rioja, Argentine Republic. Acta Academica 
 Aboensis, Mathematica et Physica, I, Abo. 
 
 Contribucion al estudio de la Petrographla del Territorio Nacional de 
 Misiones. Minist. Agricult. Argentina, Bol. 21, serie B. 
 
 La Sierra de la Ventana. Revista Museo. La Plata, III, pp. 3-11. 
 Contribucion el estudio de la geologla de la Provincia de Buenos Aires. 
 
 Revista Museo. La Plata, VII, pp. 477-489. 
 
 Contribuciones al conocimiento de la geologia de la Provincia de 
 Buenos Aires. Publicaciones de la Universidad de la Plata, num. 1. 
 Beitrage zur Geologie der argentinischen Provinz Buenos Aires. 
 Petermann's Geograph. Mitt., L, Heft IV. 
 
 Heald, K. C., and K. F. Mather. 
 
 1922. Reconnaissance of the Eastern Andes between Cochabamba and 
 
 Santa Cruz, Bolivia. Bull. Geol. Soc. America, XXXIII, pp. 553-570. 
 
 Holdhaus, K. 
 
 1919. Sobre alguns Lamellibranchios Fosseis do Sul do Brasil. Servigo. Geol. 
 e. Mineral, do Brasil, Monograph II. 
 
 Huene, F. von. 
 
 1925. Fossilfiihrende Karrooschichten im nordlichen Sudwestafrika. Neu. 
 Jahrb. f. Min., Beil. Bd. LII, pp. 115-122. 
 
 Jaworski, E. 
 
 1923. (On the Marine Triassic in South.America.) Neues Jahrb. Min. Geol., 
 
 Beil. Bd. XLVII, pp. 93-200, t. IV-VI. 
 
 Katzer, F. 
 
 1903. Grundzuge der Geologie des unteren Amazonas. Leipsig, pp. 1-296. 
 
 Keidel, J. 
 
 1916. La Geologia de las Sierras de la Provincia de Buenos Aires y sus 
 
 relaciones con las montanas de Sud Africa y los Andes. Anales. 
 Minist. Agricult. Argentina, XI, num. 3. 
 
 1917-18. Ueber das patagonische Tafelland, etc. Zeits. Deutsch. Wissen. 
 
 Vereins Buenos Aires, Jahrg. 1917, Heft. 5 and 6; Jahrg. 1918, 
 Heft 1,3, etc. 
 
 1921. Observaciones geoldgicas en la Precordillera de San Juan y Mendoza. 
 
 Anales. Minist. Agricult. Argentina, XV, num. 2, Monographfa II. 
 
 1922. Sobre la distribucidn de los Depositos Glaciares del Permico conocidos 
 
 en la Argentina. Bol. Acad. Nacion. Ciencias. Cordoba, XXV, 
 pp. 239-368. 
 
 1925. Sobre la estructura tectdnica de las Capas Petrollferas en el territorio del 
 Neuquen. Minist. Agricult. Argent. Public, num. 8, secci6n Geologia. 
 Kitson, A. E., and Davies A. Morley. 
 
 1925. Devonian Rocks at Accra, Gold Coast, West Africa. Compte-Rendu, 
 XIII e Congres. Geol. Internt., 1922., p. 945. 
 
 Haug, E. 
 
 1905 - 
 
 Hausen, H. 
 
 1921. 
 
 Hausen, J. 
 
 1919. 
 
 Hauthal, R. 
 
 1892. 
 
 1896. 
 
 1901. 
 
 1904. 
 
BIBLIOGRAPHY 
 
 125 
 
 Knod, R. 
 
 1908. Beitrage zur Geologie und Palaeontologie von Sudamerika (von G. 
 
 Kober, L. 
 
 1921. 
 
 Kozlowski, R. 
 
 Steinmann). Devonische Faunen Boliviens. Neu. Jahrb. f. Min., 
 Beil. Bd. XXV, pp. 493-600. 
 
 Der Bau der Erde. Berlin. 
 
 1923. Faune Dtivonienne de Bolivie. Annal. de Paleont., XII, fasc. 1 and 2. 
 
 Kurtz, F. 
 
 1895. Sobre la existencia del Gondwana Inferior en la Republica Argentina. 
 
 Revista Museo. La Plata, VI, pp. 125-139. 
 
 1895a. Contribuciones a la Palaeophytologia Argentina, Ibid., pp. 117-124. 
 1921-22. Atlas de Plantas Fdsiles de la Republica Argentina. Adas Acad. 
 Nacion. Ciencias. Cordoba, VII, pp. 129-153. 
 
 Lecointre, G., and P. Lemoine. 
 
 1925. Sur les graptolites Gothlandiens de la Guinee frangaise. Comptes 
 
 Leme, A. B. P. 
 
 Rendus, Paris, pp. 387-388. 
 
 1924. Evolugao da Estructura da Terra geologia do Brasil. Museu Nacional, 
 Rio de Janeiro . 
 
 Lemoine, P. 
 
 1913. Afrique Occidentale. Handbuch Regionl. Geol., Heidelberg, Bd. VII, 
 Abt. 6A, Heft 14. 
 
 Lisboa, M. A. R. 
 
 1914. 
 
 The Permian Geology of Northern Brazil. Amer. Journ. Science, 
 XXXVII, pp. 425-443. 
 
 Lundqvist, G. 
 
 1919. 
 
 Fossile Pflanzen der Glossopteris- flora aus Brasilien. Kungl. Svenska. 
 Vetens. Akad. Handl., Bd. 60, No. 3, pp. 1-36. 
 
 Mather, K. F. 
 
 1922. 
 
 Front Ranges of the Andes between Santa Cruz, Bolivia, and Embar- 
 
 cacion, Argentina. Bull. Geol. Soc. America, XXXIII, pp. 703-764. 
 
 Meyer, H. L. F. 
 
 1914. 
 
 Carbonfaunen aus Bolivia und Peru. Neu. Jahrb. f. Min., Beil. Bd. 
 XXXVII, pp. 590-652. 
 
 Nagera, J. J. 
 
 1919. 
 
 La Sierra Baya. Anales. Minist. Agricult. Argentina, XIV, num. 1, 
 Monografia II. 
 
 1919a. Nota geologica sobre el Cerro San Augustin, Balcarce. Minist. Agri¬ 
 cult. Argentina, bol. 22, serie B. 
 
 Oliveira, E. P. de 
 
 1918. 
 
 Regioes carboniferas dos Estados do Sul. Servigo Geologico e Minera- 
 logico do Brasil. Rio de Janeiro. 
 
 1920. 
 
 Rochas petroliferas do Brasil. Servigo Geologico e Mineralogico do 
 Brasil, boletim 1. 
 
 1925- 
 
 Mappa Geologico do Estado do Parana, (scale 1 in 1,000,000). Semi go 
 Geologico e Mineralogico do Brasil. 
 
 Orbigny, A. d\ 
 1842. 
 
 Voyage dans l’Amdrique meridionale (1826-1833). Paris and Strass- 
 burg, 1835-1847. 
 
 Pastor^, F. 
 
 1915* 
 
 Estudio geoldgico y petrogrdfico de la Sierra del Morro. Anales. 
 Minist. Agricult. Argentina, XI, num. 2. 
 
 Penck, W. 
 
 1920. 
 
 Rassmuss, J. 
 
 1918. 
 
 Der Siidrand der Puna de Atacama (NW. Argentinien). Leipzig. 
 
 Investigation de la estructura tectdnica de la Cuenca imbrifera del 
 Rio de la Rioja. Minist. Agricult. Argentina, bol. 17, serie B. 
 
126 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 Rassmuss, J.— Continued. 
 
 1921. Observaciones geologicas en Salta. Minist. Agricult. Argentina, bol. 
 
 4 » P* 2 5 - 
 
 1922. Apuntes geoldgicos sobre los hallazgos de Carbdn al sur del Lago 
 Nahuel-Huapi. Minist. Agricult. Arge?itina, bol. 28, serie B. 
 
 1922a. La Cuenca de Marayes. Ibid., bol. 32, serie B. 
 
 Reed, F. R. C. 
 
 1925. Revision of the Fauna of the Bokkeveld Beds. Annals S. African 
 Museum , XXII, pt. I, no. 2, pp. 27-225. 
 
 Rimann, E. 
 
 1915. Ueber Kimberlit und Alnoit in Brasilien. Tscher. Min. u. Petr. Mitt. 
 XXXIII, p. 244. 
 
 1917. A Kimberlita no Brasil. Annaes. Escola. Minas. Ouro Preto. no. 15. 
 
 1918. Estudio geologico de la Sierra Chica. Bol. Acad. Nacion. Ciencias. 
 
 Cordoba, XXIII, 2,pp. 129-202. 
 
 Rogers, A. W. and A. L. Du Toit. 
 
 1909. The Geology of Cape Colony. London. 
 
 Schiller, W., and H. Keidel. 
 
 1907. Ueber den Bau der argentinischen Anden. Sitz. Berich. K. K. Akad. 
 Wissen. Wien., Math-natur. Kl., CXVI, sect. I, pp. 649-674. 
 
 Seward, A. C. 
 
 1922. On Carboniferous Plants from Peru. Quart. Journ. Geol. Soc. London 
 
 LXXVIII, p. 278. 
 
 Seward, A. C., and J. Walton. 
 
 1923. On a Collection of Fossil Plants from the Falkland Islands. Quart. 
 
 Journ. Geol. Soc. London, LXXIX, pp. 313-332. 
 
 SlEMIRADSKI, I. VON. 
 
 1893. Eine Forschungsreise in Patagonien. Petermann’s Mitt., Bd. 39, pp. 
 49-62. 
 
 Small, H. L. 
 
 1914. Geologia e supprimento d’agua subterranea no Piauhy e parte do 
 Ceara. Publicaqao No. 32, Inspectoria Federal de Obras contra as 
 Seccas, Rio de Janeiro. 
 
 Sopper (Soper), R. 
 
 1914. Geologia e supprimento d’agua subterranea em Sergipe e no 
 nordeste da Bahia. Publicaqdo No. 34, Inspectoria de Obras contra 
 as Seccas, Rio de Janeiro. 
 
 Stappenbeck, R. 
 
 1910. La Precordillera de San Juan y Mendoza. Anales. Minist. Agricult. 
 Argentina, IV, num. 3, pp. 1-187. 
 
 1913. Apuntes Hidrogeologicos sobre el Sud-Este de la Provincia de Men- 
 doze. Minist. Agricult. Argentina, bol. 6, serie B. 
 
 1917. Geologia de la Falda Oriental de la Cordillera del Plata. Anales. 
 
 Minist. Agricult. Argentina, XII, num. 1. 
 
 1921. Estudios geoldgicos e hidrologicos en la Zona Subandina de las Pro- 
 vincias de Salta y Tucuman. Anales. Minist. Agricult. Argentina, 
 XIV, num. 5. 
 
 1927. Ueber T ransgressionen und Regressionen des meers und Gebirgsbildung 
 in Sudamerika. Neues Jahrb. Min. Geol., Beil. Bd. LVIII B, pp. 
 453 - 496 . 
 
 Steinmann, G., H. Hoek, and A. von Bistram. 
 
 1904. Zur Geologie des sudostlichen Boliviens. Centralb. f. Min. Geol. u. 
 Pal., Stuttgart, V, 5, pp. 1-4. 
 
 Steinmann, G. and Solms-Laubach. 
 
 Das Auftreten und die Flora der Rhatischen Kohlenschichten von La 
 Temera (Chile). Neu. Jahrb. f. Min., Beil. Bd. XII, pt. 3, pp. 581— 
 609. 
 
BIBLIOGRAPHY 
 
 127 
 
 Stelzner, A. 
 
 1878. Beitrage zur Geologie und Palasontologie der argentinischen Repub- 
 lik. Cassel. 
 
 Sussmilch, C. A., and T. W. E. David. 
 
 1920. Sequence, Glaciation, and Correlation of the Carboniferous Rocks of 
 the Hunter River District, New South Wales. Proc. Roy. Soc. 
 
 SZAJNOCHA, L. 
 
 N. S. Wales, LIII, pp. 246-338. 
 
 1888. Ueber fossile Pflanzenreste aus Cacheuta in der argentinischen 
 
 Taylor, F. B. 
 
 Republik. Wien. 
 
 1910. Bearing of the Tertiary Mountain Belt in the Origin of the Earth’s 
 
 Thomas, I. 
 
 1905. 
 
 Plan. Bull. Geol. Soc. America, 21, pp. 179-226. 
 
 Neue Beitrage zur Kenntniss der devonischen Fauna Argentiniens. 
 Zeits. Deutsch. Geol. Gesell. Berlin, 57, pp. 233-290. 
 
 Ulrich, A. 
 
 
 1893. Palaeozoische Versteinerungen aus Bolivien. Neu. Jahrb.f. Min., Beil. 
 
 Valentin, J. 
 
 1894. 
 
 Bd. VIII, pp. 1-116. 
 
 Rdpido estudio sombre las sierras de los partidos de Olavarria y del 
 Azul. Revista Museo La Plata, VI, pp. 1-24. 
 
 Walther, K. 
 
 1912. 
 
 Ueber Transgressionen der oberen “ Gondwana-Formation ” in 
 Sudbrasilien und Uruguay, Centralb. f.M.G.P., no. 13, pp. 398-405. 
 
 1912a. Zur Geologie der Gegend von Seibal im Staate Rio Grande do Sul 
 und ihrer Kupfererzlagerstatten. Zeitsch. f. Prakt. Geologie, XX, 
 
 1919. 
 
 pt. 10, pp. 404-414. 
 
 Lineas fundamentales de la Estructura Geolbgica de la Republica del 
 Uruguay. Montevideo. 
 
 1924. 
 
 Los Resultados de las Perforaciones practicadas en el departamento 
 de Cerro Largo en busca de Carbon de Piedra. Instituto de Geologia 
 y Perforaciones, bol. 6. 
 
 1924a. Estudios Geomorfologicos y Geologicos. Montevideo. 
 
 Wegener, A. 
 
 1915. 
 
 1924. 
 
 White, D. 
 
 1907. 
 
 Die Enstehung der Kontinente und Ozeane. Brunswick. 
 
 The Origin of Continents and Oceans. London. 
 
 Permo-Carboniferous Climatic Changes in South America. Journ. 
 Geology, XV, pp. 615-633. 
 
 1908. 
 
 Fossil Flora of the Coal Measures of Brazil. Final Report of the Bra¬ 
 zilian Coal Commision, Rio de Janeiro, Part III, pp. 337-617. 
 
 White, I. C. 
 
 1908. 
 
 Final Report of the Brazilian Coal Commission (Portuguese and 
 English), Rio de Janeiro, pp. 1-336. 
 
 WlCHMANN, R. 
 
 1918. 
 
 Geologia e Hidrologfa de Bahia Blanca y sus alredores. Anales. 
 Minist. Agricult. Argentina, XIII, num. 1. 
 
 1919. 
 
 Contribucion a la Geologia de la Region comprendida entre el Rio 
 Negro y Arroyo Valcheta. Anales. Minist. Agricult. Argentina, 
 XIII, num 4. 
 
 1922. 
 
 Observaciones Geologicas en el Gran Bajo de San Julian. Minist. 
 Agricult. Argentina, bol. 30, serie B. 
 
 Windhausen, A. 
 
 1916. 
 
 Los yacimientos petrollferos de la Zona Andina. Minister. Agricult. 
 Argentina, bol. 15, serie B. 
 
 1918. 
 
 The Problem of the Cretaceous-Tertiary Boundary in South America, 
 Amer. Journ. Science, XLIV, pp. 1-53. 
 
128 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 Windhausen, A.— Continued. 
 
 1918a. Rasgos de la historia geologica de la planicie costanera en la Patagonia 
 septentrional. Bol. Acad. Nacion. Ciencias, Cordoba, XXIII, pp. 
 319-364. 
 
 1921. Informe sobre un viaje de reconocimiento geoldgico. Chubut. Minist. 
 
 Agricult. Argent., bol. 24, serie B. 
 
 1921a. Ensayo de una clasificacidn de los Elementos de Estructura en el 
 subsuelo de la Patagonia. Bol. Acad. Nacion. Ciencias, Cordoba, 
 XXV, pp. 125-139. 
 
 1922. Estudios geoldgicos en el valle superior del Rio Negro. Minist. 
 
 Agricult. Argentina, bol. 29, serie B. 
 
 1924. Lineas generales de la constitucidn geoldgica de la Region situada al 
 oeste del Golfo de San Jorge. Bol. Acad. Nacion. Ciencias, Cordoba, 
 XXVII, pp. 167-320. 
 
 Woodward, A. Smith. 
 
 1908. On some reptilian bones from the State of Rio Grande do Sul. Final 
 Rept. Brazilian Coal Commission, pp. 202-206; 230-231. 
 
 Woodworth, J. B. 
 
 1912. Geological Expedition to Brazil and Chile. Bull. Museum. Compar. 
 Zoology, Harvard Coll., LVl, no. 1, geol. ser. X, 1. 
 
 Zeiller, R. 
 
 1895. Note sur la flore fossile des gisements houillers de Rio Grande do Sul 
 
 (Brezil meridional). Bull. Soc. Geol. de France, XXIII, p. 601. 
 
 1896. Remarques sur la flore fossile de TAltai. . . . Republique Argentine 
 
 Bull. Soc. Geol. France, XXIV, p. 466. 
 
APPENDIX 
 
 UPPER CARBONIFEROUS FOSSILS FROM ARGENTINA 
 
 By 
 
 F. R. Cowper Reed, M.A., Sc.D., F.G.S. 
 Sedgwick Museum, Cambridge , England 
 
 With four plates 
 
 129 
 

 
GENERAL REMARKS 
 
 The fossils submitted to me by Doctor Du Toit from the Que- 
 brada del Salto, Barreal, San Juan, Argentina, are mostly preserved 
 in a tough purplish sandy mudstone rather than shale, with an 
 irregular subconchoidal fracture, and, though often fragmentary and 
 imperfect, are occasionally in a good state of preservation. 
 
 R. Stappenbeck, 1 in 1911, recorded the following fossils from the 
 same locality and apparently from the upper of the two fossiliferous 
 horizons herein discussed, which he referred to the stage of Spirifer 
 supramosquensis, a Russian species which Nikitin regarded as 
 characteristic of the Gshelian: 
 
 Spirifer supramosquensis Nik. 
 Productus cor a D’Orb. 
 Productus lineatus Waag. 
 Productus pustulatus Keys. 
 Orthis sp. 
 
 Chonetes sp. 
 
 Platystoma sp. 
 
 Pleurotomaria sp. 
 
 Euomphalus sp. aff. parvus Waag. 
 Orthoceras sp. 
 
 Cyathocrinus sp. 
 
 Bryozoa. 
 
 No description or figures of the species were given, and there 
 has not been, to the best of my knowledge, any later account of 
 the fauna. 
 
 In the succeeding descriptions the figures 7281 to 7376 represent 
 the catalogue numbers of the specimens, which are preserved in the 
 South African Museum at Cape Town. 
 
 DESCRIPTION 
 
 Fenestella aff. perelegans Meek 
 
 Most of the specimens of Fenestella in this collection are poorly preserved 
 and fragmentary and only in one case is the poriferous face well preserved. 
 In this specimen (7324) the branches are subparallel, the dissepiments 
 slender and equidistant, the fenestrules rectangular and oblong, being about 
 ip2 times as long as wide and 9 to 10 occurring in a distance of 10 mm.; 
 there are 4 to 5 rather widely spaced small cells to each fenestrule, one of the 
 cells usually lying at the base of the dissepiments, and the low median 
 carina bears a row of much smaller pores (or hollow tubercles) numbering 
 7 to 9 to each fenestrule. The reverse face is rather coarsely striated longi¬ 
 tudinally. The poriferous keel occurs in many species of the genus and was 
 figured by McCoy 2 in F. formosa from the Carboniferous limestone of 
 Ireland. But it is on upper Carboniferous horizons that this type of Fenestella 
 is most common, and we may especially compare our form with F. perelegans 
 Meek 3 of the coal measures of Nebraska, but also recorded by Waagen and 
 
 1 Stappenbeck, Geol. Aufbau d. Vorkordillere, Geol. u. Palaeont. Abhandl. N. F., 
 Bd. IX, Hefts, I9ii,pp. 31,32. 
 
 2 McCoy, Syn. Carb. Foss. Ireland, 1844, p. 201, PI. XXIX, Fig. 2. 
 
 3 Meek and Hayden, Final Rept. Nebraska, 1871, p. 153, PI. VII, Figs. 3 to 3d. 
 
 131 
 
132 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 Pichl 4 from the Producius limestones of the Salt Range. There are also 
 several allied Russian species from the upper Carboniferous, such as F. 
 saraneana Stuck. 5 and F. eichwaldi Stuck., 6 and the author 7 has described a 
 form allied to the latter species from the upper Carboniferous of Chitral. 
 F. laosensis Mansuy, 8 from the Permo-Carboniferous of Indo-China, may 
 also be compared. 
 
 Fenestella sp. ind. 
 
 Among the fragments of zoaria of Fenestella occurring in the collection 
 there seems to be another species (7306, 7307) characterized by its finer 
 mesh-work, 12 to 16 fenestrules occurring in a distance of 10 mm. The 
 reverse face is longitudinally striated, but the poriferous face is unknown 
 and the affinities of the species are therefore uncertain. 
 
 Polypora aff. biseriata Ulrich 
 
 A species of Polypora is represented by several poor and fragmentary 
 specimens (7286, 7285, 7311, 7295, 7298). The one (7286) which has the 
 celluliferous face of the zoarium best preserved shows also portions of the 
 reverse face. The zoarium consists of subparallel, thick, straight branches, 
 frequently bifurcating and bearing two rows of rather large alternating cells, 
 with an additional one usually at the base of each dissepiment. Traces of 
 a median line of small nodes can be occasionally detected. The reverse face is 
 ornamented with 8 to 10 rather coarse longitudinal thread-like lines, and 
 is flattened for most of its length, but seems to swell up in places into low, 
 broad nodes. The dissepiments are short and thickened at their base and 
 form oval fenestrules of subequal size longer than wide but narrower than 
 the branches; about 7 occur in a length of 10 mm. 
 
 The relations of this imperfectly known species seem to be with P. 
 biseriata Ulrich 9 and P. varsoviensis Prout, 10 both from the Warsaw Forma¬ 
 tion, Illinois. 
 
 Orbiculoidea saltensis sp. nov. (Plate XIII, Fig. 1) 
 
 Shell subcircular. Upper valve low, conical, with excentral apex situ¬ 
 ated at about one-third of the diameter from posterior margin. Surface 
 ornamented with about fifteen coarse rounded subequidistant concentric 
 liras, mostly bearing small, low, closely placed, hollow tubercles, the whole 
 crossed by very delicate radial strias. Lower valve (incompletely known) 
 flattened or gently concave, with hollow, low, conical apex and a small 
 foramen at the end of a partly inclosed broad groove to posterior margin. 
 
 Dimensions — I (7335): Diameter, 20 mm.; height, c. 8 mm. II (7340): 
 Diameter, c. 27 mm.; height, c. 9 mm. 
 
 Remarks —There are three specimens (7335, 7340, 7310) of this new 
 species, one of which (7335) has the lower valve showing the posterior 
 groove, though in a poor state of preservation. The other two are less perfect 
 and only show the upper valve. The peculiar tuberculation of the concentric 
 rugae seems to be an original and definite character, but it is only well seen 
 
 4 Waagen and Pichl, Salt-Range Foss. (Palseont. Ind., Ser. XII, 1885), p. 777; PI. 
 
 LXXXVII, Figs. 1 to 3. 
 
 6 Stuckenberg, Mem. Com. Geol. Russ., vol. X, no. 3, 1895, p. 231, t. XX, Fig. 16. 
 
 6 Ibid., p. 234, t. XXII, Fig. 1. 
 
 7 Reed, Upper Carb. Foss. Chitral (Palasont. Ind., n.s., vol. VI, mem. 4, 1925), p. 19, 
 pi. II, Figs. 6, 6 a, 7. 
 
 8 Mansuy,Mem. Serv. Geol. Indo-Chine, vol. II, fasc. 4,1913, p. 23, PI. I, Figs. 2 a to c . 
 
 9 Ulrich, Geol. Surv. Illinois, vol. 8, 1890, p. 592, pi. LX, Figs. 4, 4 a and 4 b. 
 
 10 Ibid., p. 593, PI. LX, Figs. 2, 2 a, and 2 b. 
 
UPPER CARBONIFEROUS FOSSILS FROM ARGENTINA 
 
 133 
 
 in two specimens, the third (7310) being much crushed and distorted. The 
 relations of this species seem to be with Orbiculoidea missouriensis (Shu- 
 mard) 11 of the upper Carboniferous rather than with 0. nitida Phill. 
 
 Productus lineatus Waagen (Plate XIII, Fig. 9) 
 
 This species, 12 which is included by many authors in Productus cor a 
 D’Orb., is represented by one small complete specimen of a pedicle valve 
 (7365) and several fragments of larger shells. The synonymy of this species 
 has frequently been discussed, 13 but to the author 14 Waagen’s type seems at 
 any rate to mark a distinct variety of the form known as Productus cora in 
 the lower Carboniferous of Europe. 
 
 Productus curvirostris Scliellwien var. nov. barrealensis (Plate XIII, Figs. 4 a, 4 b) 
 
 Shell with strongly convex narrow, elongated pedicle-valve, having 
 prominent incurved acutely pointed beak projecting behind and over short 
 straight hinge-line. Ears small, triangular, gently convex, well marked 
 off from body by impressed line. Surface of valve ornamented with few 
 regularly distributed spinose tubercles, semirecumbent, widely spaced, and 
 arranged in a roughly quincunx order, with weak concentric rugae becoming 
 more marked on the flanks and with fine, irregular, undulating concentric 
 lamellose striae over the whole valve. 
 
 Dimensions —(7315): Length, 16 mm.; width (cardinal), 10 mm.; 
 height, c. 8 mm. 
 
 Remarks —This shell, which is represented by one well-preserved speci¬ 
 men of a pedicle-valve (7315) and by a fragment of another may be regarded 
 as a variety of Productus curvirostris Schellw. 15 of the Permo-Carboniferous 
 of Italy. Mansuy 16 has also recorded this species from Indo-China. It is 
 undoubtedly allied to Pr. acuteatus Mart., but is sufficiently distinct, as 
 Schellwien and others have pointed out. 
 
 Productus cf. juresanensis Tschernyschew 
 
 A fragment of a pedicle-valve of a species of Productus (7281) is too 
 imperfect for a satisfactory determination, but it appears to possess the 
 ornamentation of Pr. juresanensis Tschern., 17 and it may be compared with 
 it rather than with Pr. nebrascensis Owen, 18 which is closely allied and is 
 also an upper Carboniferous species. Pr. scabriculus Mart, belongs to the 
 same group. 
 
 Productus (Marginifera ) cf. echinatus Waagen 
 
 One imperfect pedicle valve (7364) appears to possess the external char¬ 
 acters of Pr. (. Marginifera) echinata Waagen, 19 as figured by Waagen from 
 
 u Schuchert, Bull. 87 U. S. Geol. Surv., 1897, p. 280 (for synonymy and references). 
 
 12 Waagen, Salt Range Foss. I (Palaeont. Ind.), p. 673, PI. LXVI, Figs. 1, 2; PI. 
 LXVII, Fig. 3. 
 
 13 Hayasaka, Science Rept. Tohoku Imper. Univ., ser. 2, Geol., 1922, pp. 86-93, 
 
 14 Reed, Up. Carb. Foss. Chitral, etc., Palaeont. Ind., vol. VI, mem. no. 4, 1925, p. 28, 
 PI. Ill, Figs. 12 to 14a. 
 
 15 Schellwien, Palaeontographica, XXXIX, 1902, p. 26, t. Ill, Figs. 12 to 14. Gortani, 
 Palaeont. Ital., vol. XII, 1906, p. 23, t. II, Figs. 1 to 3; id., Boll. Soc. Geol. Ital., vol. XXV, 
 1906, fasc. 2, p. 14, Fig. 7. 
 
 16 Mansuy, Mem. Serv. G60I. Indo-Chine, vol. II, fasc. 4, 1913, p. 32, PI. II, Figs. 
 9a toe; ibid., vol. Ill, fasc. 3, 1914, p. 16, PI. II, Figs, ioatoc. 
 
 17 Tschernyschew, Mem. Com. G60I. Russ., vol. XVI, no. 2,1902, p. 620, t. XLVII, 
 Figs. 1, 2; t. LIII, Fig. 4. 
 
 18 Schuchert, Bull. 87 U. S. Geol. Surv., 1897, p. 327 (for synonymy and references). 
 
 14 Waagen, op. cit., p. 727, PI. LXXVIII, Fig. 1. 
 
134 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 the lower Productus limestone of the Salt Range. The thin, recumbent 
 spines on the surface, becoming more upstanding or tubercular posteriorly, 
 and the transverse swollen shape of the valve are clearly preserved; the 
 broad short beak, hollowed umbonal slopes, and hinge-line can also be 
 distinguished. It had been labeled Strophalosia cf. indica Waag., but is 
 totally different from that species. 
 
 Productus ( Marginifera ) spinulo-costatus Abich var. nov. peregrina (Plate XIII, Fig. 2) 
 
 There is one fairly perfect pedicle-valve (7374) which might be regarded 
 as a variety of Productus (Marg.) ornatus Waag., 20 but seems more allied to 
 Pr. (Marg.) spinulo-costatus Abich, 21 of the Djulfa beds, and if better known 
 we might consider it to be a distinct species. The shell is transversely semi¬ 
 elliptical, with a straight hinge-line, equal to, or nearly equal to, its width; 
 the pedicle-valve is strongly swollen, with a weak, broad median depression 
 down the middle; the beak is broad, rounded, obtuse, swollen, and incurved 
 and overhangs the hinge-line. The ears are small, triangular, depressed, 
 and sharply marked off from the umbonal slopes, which descend steeply to 
 them and are somewhat excavated. The surface of the valve is ornamented 
 with weak, low, rounded, indistinct ribs, somewhat discontinuous, being 
 mainly due to radially elongated recumbent stout spines arranged in a 
 roughly quincunx fashion, but widely spaced; there are about eight such 
 faint ribs on each side of the median depression, which bears only one short 
 similar rib arising at about half the length of the valve; three to five small spi- 
 nose tubercles form an irregular line along the junction of the ears with the 
 body, and two to three more occur on the ears. Fine concentric striae cross 
 the valve, with a few stronger growth-lines becoming more marked on the 
 flanks. 
 
 From typical examples of Pr. (Marg.) ornatus it differs by its more 
 elongated recumbent spine bases, and thus more resembles the Armenian 
 species, Pr. (Marg.) spinulo-costatus , which Freeh 22 and Hayasaka 23 figure 
 also from China. Pr. walcottianus Girty 24 may also be compared. 
 
 Chonetes cf. pseudovariolata Nikitin 
 
 Two small specimens of Chonetes in rather a poor state of preservation 
 seem to be comparable to Ch. pseudovariolata Nikitin, 25 of the Gshelian stage 
 of Russia, rather than with Ch. lissarensis Diener, 26 of the Zewan beds of 
 Kashmir; but such small and imperfectly known shells are almost impossible 
 to determine specifically. 
 
 Chonetes granulifer Owen (Plate XIII, Fig. 6) 
 
 One large specimen of a species of Chonetes (7319) shows the pedicle- 
 valve with the shell broken away from most of the middle part and the 
 outer surface only preserved on a small portion near the anterior margin. 
 It is transversely semi-elliptical in shape, with the cardinal angles slightly 
 
 20 Waagen. op. cit., p. 721, PI. LXXVII, Figs. 5 and 6. 
 
 21 Freeh and Arthaber, Palaeoz. in Hocharmenien, Beitr. z. Palaeont. u. Geol. Oesterr. 
 Ung., Bd. XII, 1900, pp. 262-264, t. XX, Figs. 5, 7,8, and 9; Diener, Palaeont. Ind., n.s., 
 vol. V, mem. 2, 1915, p. 82, PI. VIII, Figs. 13, a to d, PI. IX, Figs. 1 and 2. 
 
 22 Freeh in Richthofen’s “China,” Bd. V, 1911, p. 175, PI. XXVII, Fig. 2. 
 
 23 Hayasaka, Science Rept. Tohoku Imper. Univ. Ser. 2, Geol., 1922, p. 99, PI. IV, 
 Figs. 15 and 16. 
 
 24 Girty, Prof. Paper 58, U. S. Geol. Surv., 1908, p. 269, PI. XXI, Figs. 27, a and b. 
 
 “Nikitin, Mem. Com. Geol. Russ., vol. V, no. 5, 1890, p. 27, PL II, Figs. 1 to 4; 
 Loczy, Beschr. Palaeont. Stratigr. Result. Reise Graf. Bela Szechenyi in Ostasien 
 (Budapest, 1898), p. 73, t. Ill, Figs. 8 to 13. 
 
 26 Diener, Himal. Foss. (Palaeont. Ind. Ser. XV), vol. I, pt. 4, 1897, p. 19, PI. II, 
 Figs. 4 and 6; ibid., n.s., vol. V, mem. 2, 1915, p. 83, PI. IX, Fig. 3. 
 
UPPER CARBONIFEROUS FOSSILS FROM ARGENTINA 
 
 135 
 
 pointed but not produced into ears. The cardinal margin shows seven to 
 eight small, obliquely directed spine-bases, the largest being at the cardi¬ 
 nal angle. The body is gently convex, and the valve is not definitely flat¬ 
 tened or depressed in the middle, a sinus being completely wanting. The 
 surface is covered with numerous fine, granulated, equal radial lines, fifty to 
 sixty occurring on the margin in a space of io mm., apparently increasing by 
 bifurcation and intercalation, and in all there must be at least 200 such lines 
 around the whole margin of the valve. The interior shows a low, narrow 
 median ridge extending from the beak for about one-third the length of the 
 valve, and separating weakly impressed flabellate diductor scars, on each 
 side and in front of which, to a distance of about three-fourths the radial 
 diameter of the valve, there are coarse, rather closely set papillae, somewhat 
 irregularly arranged in radial lines; outside them there is a marginal band 
 of smaller, more regular, closer, and more numerous papilla (in both cases, 
 of course, shown on the internal cast as punctae). 
 
 This shell in some respects seems to resemble Ch. variolata D’Orb., 27 
 and especially its variety baroghilensis Reed 28 from the upper Carboniferous 
 of Chitral, in its internal character, but it has no definite median flattening 
 or sinus. Many of the shells referred to Ch. granulifer Owen 29 bear a closer 
 resemblance to our shell, though this specific name has been used in some¬ 
 what a wide sense and still requires a detailed diagnosis and strict limita¬ 
 tion, but it is an indication of an upper Carboniferous horizon. 
 
 Camarophoria cf. mutabilis Tschemyschew? 
 
 Two crushed and imperfect valves (7327, 7352) of a rhynchonelloid 
 shell having three to four ribs in the sinus and on the fold and four to five on 
 each side seem to be comparable with the variable species Camarophoria 
 mutabilis Tschem., 30 typically occurring on the Schwagerina horizon of 
 Russia and also recorded from Indo-China by Mansuy. 31 
 
 Spirifer cf. supramosquensis Nikitin (Plate XIV, Fig. 1) 
 
 Stappenbeck 32 has already recorded Spirifer supramosquensis from this 
 locality in Argentina. It is characteristic of the Gshelian, according to> 
 Nikitin. 33 There is one specimen (7316) in the present collection which, 
 seems indistinguishable from this species, and especially resembles the 
 figure of the shell from the Urals referred by De Verneuil 34 to Sp. crassus 
 De Ron., but considered by Nikitin to be identical with Sp. supramosquen¬ 
 sis. The shells which Hayasaka 35 and Mansuy 36 respectively figure as Sp. 
 nikitini Tschern. from the upper Carboniferous of Manchuria (in associ¬ 
 ation with Sp. supramosquensis) and from the Productus limestone of Indo- 
 China bear a considerable resemblance, and Hayasaka remarks on its like¬ 
 ness to the American Sp. mexicanus Shum. of the upper Carboniferous. 
 
 27 D’Orbigny, Voyage dans l’Amerique Merid. Ill, pt. 4, 1842, p. 49, PI. IV, Figs. 10 
 and 11. Kozlowski, Brach. Carb. Super. Bolivie, Ann. Pateont. IX, 1914, p. 55, PI. 
 VIII, Fig. 16. 
 
 28 Reed, Up. Carb. Foss. Chitral (Pal. Ind., n.s., vol. VI, mem. 4, 1925), p. 40, PI. Ill, 
 Figs. 1 to 4. 
 
 29 Greene, Journ. Geol., vol. XVI, 1908, pp. 654-663, Pis. I-IV; Girty, Bull. 544, U. S. 
 Geol. Surv., 1915, p. 59, PI. VI, Figs. 12 to 133. 
 
 30 Tschemyschew, op. cit., p. 491, t. XLV, Figs. 1 to 15. 
 
 31 Mansuy, op. cit., vol. II, fasc. 4, p. 90, PI. IX, Figs. 14a to c. 
 
 32 Stappenbeck, Geol. Aufb. d. Vorkordillere (Geol. u. Palaeont. Abh. N. F., Bd. IX, 
 Heft 5, 1911), pp. 31-32. 
 
 33 Nikitin, Mem. Com. G60I. Russ., vol. V, no. 5,1890, pp. 66, 165, t. Ill, Figs. 1 to 3. 
 
 34 De Verneuil, Geol. Russ. II, Palaeont., p. 165, t. VI, Fig. 2. v - 
 
 33 Hayasaka, op. cit., 1922, p. 125, PI. VI, Figs. 10 to 13. 
 
 36 Mansuy, Mem. Serv. Geol. Indo-Chine, vol. 2, fasc. 4,1913, p. 66, PI. VI, Fig. 4. 
 
136 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 Gortani 37 would regard Sp. fritschi Schellw. as only a variety of Sp. supra- 
 mosquensis , and there are many closely allied species from the upper Carbon¬ 
 iferous of Russia, such as Sp. ussensis Stuck., 38 which are difficult to separate. 
 
 Spirifer mexicanus Shumard var. nov. neotropica (Plate XIV, Figs. 3, 4) 
 
 There are two specimens (7325, 7366) of a species of Spirifer belonging 
 to the Sp. mexicanus group which much resemble the varieties of Sp. 
 mexicanus from Texas, figured by Girty, 39 and are distinguished by the 
 more definitely demarcated but smooth channel in the median sinus of the 
 pedicle-valve. This median channel in the sinus of one of our specimens 
 is deep, narrow, devoid of ribs, and preserves an almost uniform width, 
 but in the other (7366) is more open. The shell is longer, more oval, and 
 narrower in shape than Sp. supramosquensis , being longer than wide, with 
 a shorter hinge-line and smaller beak. The interior of part of the valve is 
 seen in one specimen and shows rather coarse pitting toward the lateral 
 margins on each side of the large median muscular area. There are seven 
 to ten low, rounded, flattened, simple ribs of subequal size on each lateral 
 lobe, and on the rounded slopes of the sinus there are three or four nar¬ 
 rower equal similar ribs, and on the upper edge two much sharper ones. 
 The cardinal ribs are very faint. Very fine radial lines and delicate con¬ 
 centric striation ornament the whole surface of the shell. 
 
 Dimensions —Length, c. 25 mm.; width, c. 30 mm. 
 
 Remarks —In spite of its more quadrate shape and simple ribs it seems 
 probable that this shell is more allied to Sp. mexicanus than to Sp. nikitini 
 Tschern., 40 but it is certainly a new variety. 
 
 Spirifer wynnei Waagen var. nov. argentina (Plate XIV, Figs, ya, 7 b) 
 
 Shell transversely subelliptical, about twice as wide as long; hinge-line 
 less than width of shell; cardinal angles rounded. Pedicle-valve convex, 
 deeper than opposite valve; beak prominent, moderately high, somewhat 
 incurved; hinge-area triangular, gently concave, inclined steeply to plane 
 of valve, well-defined from lateral lobes by sharp edges; surface of valve 
 crossed by shallow, broad, rounded median sinus, not sharply marked off 
 from lateral lobes, increasing in width and depth anteriorly, with floor pro¬ 
 duced with broad, rounded, projecting tongue arching steeply upward, so 
 as to be nearly at right angles to plane of valve; lateral lobes rounded, gently 
 to strongly convex from side to side. Sinus holding nine to ten subequal, 
 narrow, low, rounded ribs, which mostly divide at about half their length, 
 making fifteen to eighteen on the anterior margin. Lateral lobes with ten to 
 •twelve similar, rather larger ribs, which bifurcate once at about half their 
 length and sometimes again nearer margin, or rarely trifurcate, giving an 
 appearance of fasciculation and numbering twenty-four to twenty-eight of 
 subequal size on the margin, those near the cardinal edge becoming smaller 
 and fainter. Brachial valve less convex and shallower than pedicle-valve, 
 with low, broad, rounded median fold increasing in width anteriorly, not 
 marked off from lateral lobes; beak small, slightly elevated, scarcely 
 incurved; hinge-area narrow, with sharply defined edges, lying in plane of 
 valve. Fold with five to seven narrow, low, rounded ribs, arising from the 
 beak, and each bifurcating at about one-third to one-half its length, and 
 
 37 Gortani, Palseont. Ital., vol. XII, 1906, p. 26, t. II, Figs. 15 to 18. 
 
 38 Stuckenberg, Mem. Com. G60I. Russ., n.s., Liv. 23,1905, pp. 44, 123, t. Ill, Fig. 17; 
 t. V, Figs. 3 and 4; t. VII, Fig. 11. 
 
 39 Girty, Prof. Paper 58, U. S. Geol. Surv., 1908, p. 360, PI. XIII, Figs. 1 to 9. 
 
 40 Tschemyschew, op. cit., p. 542, t. XIII, Fig. 2. 
 
UPPER CARBONIFEROUS FOSSILS FROM ARGENTINA 
 
 137 
 
 •some dividing a second time so as to form sixteen to eighteen ribs on the 
 front margin. Lateral lobes with five to six similar but rather larger ribs on 
 each side, which bifurcate soon after their origin and mostly again at about 
 half their length or near the margin, so as to make about twenty to twenty- 
 four in all on the margin, with a decided tendency to fasciculation. In both 
 valves the ribs on the fold and in the sinus tend to be smaller than those 
 on the lateral lobes. 
 
 Dimensions —(7309): Length of brachial valve, 31 mm.; width of bra¬ 
 chial valve, 50 to 54 mm.; depth of conjoint valves, 26 mm. 
 
 Remarks —This form, which is represented by five imperfect specimens 
 in different states of preservation, is undoubtedly closely allied to Sp. 
 wynnei Waagen, 41 but it can hardly be considered identical with the Salt 
 Range type-examples, though it may be regarded as a variety. The chief 
 differences are in the tendency to fasciculation of the ribs and in the fact 
 that they are smaller on the fold and sinus than on the lateral lobes. The 
 rounded character and low elevation of the fold on the brachial valve and 
 the sharply demarcated hinge-area on the pedicle-valve are features which 
 distinguish it from another Salt Range species, Sp. oldhamianus Waag. ; 42 
 the number of ribs is also less. The lower Carboniferous species Sp. crassus 
 De Kon. 43 is perhaps allied to our shell, but it has a less marked median fold 
 and sinus, and no tendency to fasciculation of the ribs. Some of the shells 
 which from time to time have been referred to Sp. striatus Mart, show 
 a marked resemblance in the ribbing, but differ by not possessing rounded 
 cardinal angles, which is also one of the chief distinctions of our shell from 
 the true Sp. cameratus Morton, though Derby 44 attributed some Brazilian 
 shells of this shape to this species. The fine-ribbed specimens from the British 
 lower Carboniferous which Davidson 45 at first referred to Sp. duplicicosta 
 Phill. but subsequently 46 transferred to Sp. striatus are almost inseparable 
 from our shells, but the present author does not agree with Davidson’s views 
 -about the species. Sp. volgensis Stuck. 47 from the Russian upper Carbonif¬ 
 erous may also be compared. Diener 48 has figured and described examples 
 of Sp. wynnei from Chitichun, which much resemble our specimen. The 
 species which Nikitin 49 described from the Gshelian of Russia as Sp. post- 
 striatus was referred in 1889 by Tschemyschew to Sp. wynnei (but subse¬ 
 quently in 1902 to Sp. cameratus ); it differs from our shell by having a more 
 triangular shape and longer hinge-area, the latter extending to the cardinal 
 -angles instead of being considerably less than the width of the shell. The 
 ribbing, however, of Sp. poststriatus is more like that of our shell than is that 
 of Sp. wynnei , being subfasciculate. Undoubtedly our shell belongs to the 
 same group as that containing Sp. striatus , Sp. duplicicosta , Sp. cameratus , 
 Sp. condor, Sp. fasciger, Sp. ravana, and Sp. marcoui , all of which are 
 recorded by Tschemyschew 50 from the upper Carboniferous of Russia. 
 Hayasaka, 51 who describes Sp. wynnei from the upper Carboniferous of 
 
 41 Waagen, Salt Range Foss., I, p. 517, PI. XLIV, Figs. 6 and 7 
 
 42 Ibid., I, p.518, Pl.XLVI. 
 
 43 De Koninck, Faune Calc. Belg., VI (Ann. Mus. R. Belg., XIV, 1887), p. 137, 
 PI. XXIII, Figs. 3 to 9. 
 
 44 Derby, Bull. Cornell Univ., 1874, p. 12, PI. I, Fig. 9 (non cet.). 
 
 45 Davidson, Mon. Brit. Foss., Brach. II, p. 21, PI. Ill, Figs. 7 to 9. 
 
 * 6 Ibid., p. 221. 
 
 47 Stuckenberg, Mem. Com. Geol. Russ, n.s., Livr. 23, 1905, pp. 40, 122, t. Ill, Fig. 19. 
 
 48 Diener, Himal. Foss., vol. 1, pt. 3 (Palaeont. Ind. Ser. XV.) 1897, p. 44, PI. VII, 
 Figs. 1 to 4. 
 
 49 Nikitin, op. cit., 1890, pp. 64, 164, t. II, Figs. 16 to 19. 
 
 50 Tschemyschew, op. cit., 1902, pp. 133-146. 
 
 61 Hayasaka, op. cit., 1922, p. 'I27, PI. VI, Figs. 15 and 16. 
 
138 GEOLOGICAL COMPARISON OP SOUTH AMERICA WITH SOUTH AFRICA 
 
 Manchuria, notes points of resemblance between it and Sp. volgensis Stuck. 
 It is said also to occur in the “lower Permian” of Australia, according to 
 Freeh, and in the upper Carboniferous of California, and Schellwien 52- 
 figures it from the Trogkofel beds of the C-arnic Alps. 
 
 Spirifer saltensis sp. nov. (Plate XIV, Figs. 8a, 8b) 
 
 Shell transversely elliptical, much compressed; cardinal angles strongly 
 rounded; cardinal line less than width of shell. Pedicle-valve gently convex, 
 most so posteriorly; beak small, pointed, elevated, incurved, with concave 
 shoulders; hinge area small, triangular, nearly in plane of valves, with 
 sharply defined edges. Surface of valve with weak, broad, shallow, but 
 well-defined, rounded sinus holding four to five narrow, low, rounded, 
 simple, equal ribs; lateral lobes covered with eight to nine much broader, 
 low, rounded, flattened ribs, becoming narrower toward the cardinal angles, 
 and each dividing near the margin into two, or usually three, much smaller 
 ribs, giving a subfasciculate appearance. Brachial valve less convex than 
 pedicle-valve, with a low, scarcely raised, but definite broad median fold, 
 increasing rather rapidly in width and carrying a median weak, narrow, 
 simple rib, and on each side of it a stronger but narrow, rounded, bifurcated 
 rib, the division taking place at about the third of its length; lateral lobes 
 with six to seven much broader, low, flattened ribs dividing into three 
 smaller ones near the margin, and clearly forming fascicles; toward the 
 cardinal margin all the ribs become very weak and faint. 
 
 Dimensions —(7296): Length, 20 mm.; width, 29 mm.; thickness, 10mm. 
 
 Remarks —The relations of this shell, of which there is only one speci¬ 
 men which is fortunately nearly perfect, are undoubtedly with the earner a- 
 tus group of species; the fasciculation of the ribs is a distinctive feature. 
 The rounded elliptical shape of the shell and short hinge-line are unlike 
 Sp. fasciger or its allies, although the ribbing especially recalls some forms 
 attributed to Sp. musakheylensis Dav ., 63 which is often regarded as only a 
 variety of Sp. fasciger. 5 * But the elliptical shape, rounded cardinal angles, 
 and short hinge-line are found combined with a similar division and sub- 
 fasciculation of the ribs in some shells referred by Davidson 55 to Sp. dupli - 
 cicosta Phill. from the British lower Carboniferous, and to Sp. cameratus 
 Morton from Brazil , 56 of which species our shell might even be considered 
 a variety. 
 
 Spirifer barrealensis sp. nov. (Plate XIV, Fig. 2) 
 
 Shell transversely subtriangular. Pedicle-valve convex, with deep, well- 
 defined median, subangular sinus, increasing slowly in width to front margin; 
 hinge-line long, equal to width of valve; cardinal angles acute; beak high, 
 prominent, acutely pointed, incurved; hinge-area high, triangular, concave, 
 nearly in plane of valve. Surface of shell marked with five to six low, rounded, 
 simple, weak ribs on each lateral lobe, becoming fainter and almost obsolete 
 toward the cardinal angles; sinus with lateral slopes bearing a faint, low 
 narrower rib on each side, arising at about half the length of the valve by the 
 unequal division of the larger rib forming the edge of the sinus, the floor of 
 
 62 Schellwien, Fauna d. Trogkofelsch., Abh. k. k. geol. Reichsanst., XVI, 1900, p. 75, 
 t. X, Figs. 5 and 6. 
 
 63 Diener, Himal. Foss., vol. I, pt. 4, p. 35, PI. Ill, Figs. 3 and 4; PI. IV, Figs. I and 
 2; PI. V, Fig. 1. 
 
 64 Reed, Upper Carb. Foss. Chitral, Palaeont. Ind., n.s., vol. VI, mem. 4, 1925, p. 42 
 and references. 
 
 65 Davidson, op. cit., PI. II, Fig. 11 (non cet.). 
 
 66 Derby, op. cit., p. 12, PI. I, Figs. 3, 14 (non cet.). 
 
UPPER CARBONIFEROUS FOSSILS FROM ARGENTINA 
 
 139 
 
 the sinus being narrow, concave, and the width of a rib. Whole surface of 
 valve covered with closely placed concentric lamellose striae and growth¬ 
 lines, and with very delicate radial striae. Brachial valve unknown. 
 
 Dimensions —(7371): Length, c. 17.5 mm.; width, c. 32.0 mm. 
 
 Remarks —The affinities of this species seem to be with Sp. aff. boon - 
 ensis Swallow, described by Girty 57 from Oklahoma. The one specimen in 
 the present collection is crushed and imperfect, only the central portion 
 and left lateral lobe of the pedicle-valve being preserved. Sp. keilhavii 
 Von Buch 58 from the upper Carboniferous of the Urals, etc., is less like it. 
 
 Spirifer aff. rajah Salter (Plate XIV, Fig. 5) 
 
 There is one imperfect specimen (7317) of a pedicle-valve of a species of 
 Spirifer which may be a variety of Sp. barrealensis, but it differs in being 
 subquadrate in shape, in possessing more numerous and narrower ribs on 
 the lateral folds, and by having on the steeply inclined, flattened slopes of 
 the sinus two to three very faint and narrower ribs; the floor of the sinus 
 is flattened, of nearly uniform width, and rather wider than one of the 
 inner ribs on the lateral lobes, just as in Sp. barrealensis. The minute 
 ornamentation is identical. In our specimen the anterior part of the shell 
 is sharply bent under the posterior part, but the sinus, with its short pro¬ 
 jecting tongue, and the concentric lamellae near the edge, are well pre¬ 
 served, and it is in this region that the ribbing is most clearly shown. The 
 cardinal angles are obtusely angular, and the shell much resembles certain 
 of the specimens figured by Broili 59 from Timor, which he refers to the 
 Himalayan species Sp. rajah Salter. Sp. samarensis Stuck, is also allied. 
 
 Spirifer (Martinia ) cf. simensis Tschernyschew (Plate XIV, Fig. 6) 
 
 There are two specimens of pedicle-valves in the collection (7338, 7363) 
 which seem comparable to, if not identical with, the shell figured and 
 described by Tschernyschew 60 from the Schwagerina horizon in the Urals 
 under the name Martinia simensis. Mansuy 61 has recorded this species 
 from Tonkin. 
 
 Dimensions —(7363): Length, c. 15 mm.; width, c. 15 mm. 
 
 Spirifer (Martiniopsis ) cf. aschensis Tschernyschew 
 
 An imperfect pedicle-valve (7361) which shows the umbo, the median 
 channel-like sinus, and the faint ribs on the lateral lobes, may be compared 
 with Tschernyschew’s Martiniopsis aschensis 62 rather than with any of the 
 other closely allied species of the same subgenus described by the same 
 author from the Schwagerina horizon in Russia. 
 
 Spiriferina zewanensis Diener (Plate XIII, Fig. 3) 
 
 There is one nearly perfect but somewhat compressed specimen (7312) of 
 a species of Spiriferina having the pedicle-valve in a good state of preserva¬ 
 tion and not crushed; the brachial valve is somewhat flattened by pressure, 
 but it shows all essential characters, and the ornamentation of the surface 
 
 67 Girty, Bull. 544, U. S. Geol. Surv., 1915, p. 91, PI. XI, Fig. 5. 
 
 68 Tschernyschew, op. cit., p. 527, t. XL, Figs. 1 to 4. 
 
 69 Broili, Palasont. Timor, VII, No. 12, 1916, p. 34, t. CXIX, Figs. 1 to 11, t. CXX, 
 Figs. 1 to 6. 
 
 60 Tschernyschew, op. cit., p. 569, t. L, Figs. 1 to 3. 
 
 61 Mansuy, op. cit., vol. V, fasc. 4, 1916, p. 33, PL V, Fig. 10. 
 
 62 Tschernyschew, op. cit., p. 557, t. L, Fig. 4. 
 
140 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 of both valves is beautifully preserved. In the pedicle-valve there are four 
 (and a weaker fifth) strong angular folds on each side of the median sinus, 
 which is broad, subangular, and rather rapidly increases in width anteriorly; 
 the pair of folds bordering it are the strongest of the series, the others suc¬ 
 cessively decreasing in size toward the cardinal region, a smooth triangular 
 space devoid of ribs being left along the cardinal line. The beak is small, 
 acute, elevated and slightly incurved, but not high, and the hinge-area is 
 low and narrow. The brachial valve has a narrow, prominent median fold, 
 rounded rather than subangular and very slowly increasing in width anteri¬ 
 orly ; there are only three strong subangular folds on each side and traces of 
 a fourth one. The cardinal angles of both valves are broken off, but seem to 
 have been acute or subacute. The surface of both valves is covered with a 
 series of strong, equidistant, regular, thick imbricating lamellae about twenty 
 to twenty-five in number, and the shell substance is coarsely punctate. A 
 very delicate radial striation is also visible with a lens on the surface. 
 
 Dimensions —Length, 14 mm.; width (estim.), 28 mm. 
 
 Remarks —This shell is much like Sp. octoplicata Sow. as redefined 
 by North, especially the mut. 8 Vaughan, 63 but the species from the Zewan 
 beds of Kashmir which Diener 64 first described as Sp. cf. kentuckiensis 
 Shum., but subsequently as Sp. zewanensis sp. nov., seems to be indistin¬ 
 guishable from our specimen. 
 
 Reticularia notica sp. nov. (PI. XIII, Figs. 8a, 8b, 10) 
 
 Shell subcircular. Pedicle valve moderately convex; beak high, promi¬ 
 nent, pointed, incurved, with umbonal slopes hollow; cardinal area concave, 
 triangular, steeply inclined; hinge-line short, about half width of shell. Sur¬ 
 face of valve marked with continuous median shallow sinus, scarcely increas¬ 
 ing in width anteriorly. Ornamentation of shell consisting of numerous 
 closely placed subequidistant concentric thin lamellae with their edges pro¬ 
 vided with minute equidistant, rather widely spaced, small, short, simple, 
 recumbent spines, not crossing the lamellae but forming narrow fringes. 
 Brachial valve unknown. 
 
 Dimensions —Length, 15 to 16 mm.; width, 14 to 15 mm. 
 
 Remarks —This species, of which there is one broken example (7326) of 
 a pedicle valve having the surface beautifully preserved and showing the 
 ornamentation, and three other decorticated and poor specimens (7376, 
 7345, 7372) can not be referred to Ret. lineata Mart., because of the median 
 sinus and the character of the marginal spinose fringes. But as often remarked, 
 R. lineata has been made to include a variety of forms. It is more like Ret. 
 setigera Hall and Whitfield 65 in the ornamentation and in the median sinus 
 and shape we may also compare R. orientalis Mansuy, 66 and Ret. rostrata 
 Kut. 67 R. waageni Loczy, 68 from southern China and especially Squamularia 
 guadelupensis Shum. 69 from Texas, may also be compared. 
 
 63 North, Quart. Journ. Geol. Soc., vol. LXXVI, pt. 2, 1920, p. 217, PI. XIII, Fig. 9. 
 
 64 Diener, Himal. Foss., vol. I, pt. 2, 1899, p. 61, PI. V, Figs. 11 and 12; id. Anthrac. 
 Foss. Kashmir, etc. (Palasont. Ind., vol. VI, mem. 2, 1915), p. 90, PI. IX, Fig. 12. 
 
 65 Hall and Whitfield, Explor. 40th Parallel, p. 270, PI. VI, Figs. 17 and 18. 
 
 66 Mansuy, op. oil., vol. II, fasc. 4, p. 81, PI. IX, Fig. 2. 
 
 67 Tschernyschew, op. cit., p. 575, t. XX, Figs. 14 to 18. 
 
 68 Loczy, Beschr. Wiss. Ergeb. Pal. Stratigr. Result. Reise des Grafen Bela Szechenyi. 
 (Budapest, 1898), p. no, t. IV, Figs. 1 and 2. 
 
 69 Girty, op. cit., 1908, pp. 367 to 369, PI. XIV, Figs. 2 and 3. 
 
UPPER CARBONIFEROUS FOSSILS FROM ARGENTINA 
 
 141 
 
 Ambocoelia plano-convexa Shumard 
 
 This well-known brachiopod, which occurs in the upper Carboniferous 
 of America, 70 Asia, 71 and Europe 72 and was recorded by Derby 73 from the 
 coal measures of Brazil and by Kozlowski 74 from the upper Carboniferous of 
 Bolivia, is represented by one pedicle-valve in the collection, measuring 
 about 7.5 mm. in width and about 5 mm. in length. 
 
 Athyris ( Spirigerella) ? sp. 
 
 There is one fragmentary specimen of a broadly subcircular pedicle- 
 valve with a weak anterior, broad, median sinus, and strong concentric 
 growth-ridges which seems to belong to some species of Athyris or Seminula 
 and may perhaps be compared with Spirigerella grandis Waag. 75 and the 
 allied species from the Productus limestones. We may also compare some of 
 the shells from the upper Carboniferous of Bolivia figured by Kozlowski 76 
 under the name Seminula argentea (Shepard), for the genus of our specimen 
 is uncertain. 
 
 Hemiptychina cf. sublxvis Waagen (Plate XIII, Fig. 5) 
 
 A small oval terebratuloid shell (7300 showing only the pedicle-valve 
 has the external appearance and shape of Hemiptychina sublcevis Waag., 77 
 but its identification is somewhat uncertain. It measures about 6.5 mm. 
 in length. 
 
 Pseudamusium stappenbecki sp. nov. (Plate XIII, Figs. 11 and 13) 
 
 Shell obliquely subelliptical, longer than high, transverse, bilaterally 
 asymmetrical, having an oblique axis and anterior swing; gently biconvex, 
 compressed. Right valve with body gently convex, more or less flattened; 
 beak pointed, with its edges meeting at a right angle or more than a right 
 angle, situated excentrally at about one-third the length of the valve from 
 posterior end; anterior end of valve rounded, projecting in front, strongly 
 arched forward above, curving back sharply below into inferior margin, which 
 makes a wider, gentler curve and passes obliquely up into shorter, more 
 semicircular posterior margin; junction of body above with posterior ear 
 straight; junction of body with anterior ear slightly concave. Anterior ear 
 large, narrow, long, acutely triangular, depressed, sharply marked off from 
 body, rounded in front, without byssal slit; surface marked with four to five 
 broad, low radial folds. Posterior ear small, short, depressed, smooth, weakly 
 but distinctly marked off from body, with obtuse posterior angle, higher 
 than long, extending a short distance down posterior margin of body. 
 Surface of body marked with very faint, low, wide, flattened radial ribs 
 of subsequal size, separated by incised lines and only distinct toward the 
 margins; a few closely placed fine raised radial lines are present on anterior 
 and posterior upper margins of body near junctions with ears; strong con¬ 
 centric stride and growth-ridges at unequal distances apart cover the whole 
 surface of both valves. Left valve similar to right valve, but with the fine 
 radial lines extending over more of the surface of the body. 
 
 Dimensions —Length (oblique), 25 mm.; height, 22 mm. 
 
 70 Girty, Bull. 544 U. S. Geol. Surv. 1915, p. 94 and refs. 
 
 71 Mansuy, op. cit., vol. II, fasc. 4, 1913, p. 82, PI. IX, Fig. 4. 
 
 72 Tschemyschew, op. cit., p. 575, t. XX, Fig. 1. 
 
 73 Derby, Bull. Cornell Univ., 1874, p. 19, PI. VIII, Figs. 12, 16, 18, PI. IX, Fig. 7. 
 
 74 Kozlowski, op. cit., 1914, p. 76, Text-Fig. 19, PI. I, Fig. 5, PI. X, Figs. 1 to 14. 
 
 75 Waagen, op. cit., p. 461, PI. XXVI, Figs. 1 to 7 (especially Figs. 3 and 4). 
 
 76 Kozlowski, op. cit., 1914, p. 79, PI. XI, Figs. 1 to 46. 
 
 77 Waagen, op. cit., p. 364, PI. XXVII, Figs. 1 to 3; Diener, Palasont. Ind., ser. XV, 
 Vol. I, PI. V, 1903, p. 40, PI. II, Figs. 3 a to d. 
 
142 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 Remarks —This species is represented by several fairly perfect speci¬ 
 mens, but the radial ribbing is only seen in those having the shell more or 
 less preserved, most of the specimens being casts in which the concentric 
 lines and ridges are more conspicuous. 
 
 The right valves (7287, 7290, 7303) are more complete than any of the 
 left valves (7301?, 7302, 7291, 7289, 7284), but there seems to be little or no 
 difference in their characters. The generic reference of this shell is to 
 Pseudamusium as defined by Wheelton Hind, and it seems allied especially 
 to Ps. anisotum Phill. 78 and to Pecten sibiricus De Vern., 79 which Hind con¬ 
 siders to be probably identical. 
 
 Diener 80 has figured an unnamed species of t( Aviculo pecten” from the 
 Zewan beds which seems rather to resemble the Argentine shell. The well- 
 known Russian species Av. sericeus De Vern . 81 is less oblique and less trans¬ 
 verse, and the same differences are noticeable in the Amazonian upper 
 Carboniferous shell which Katzer 82 has figured as Euchondria neglecta 
 Geinitz. But Herrick’s species Crenipecten foerstii zz of the Pottsville fauna 
 closely resembles our species in character. 
 
 Pseudamusium cf. ellipticum (Phillips) (Plate XIII, Fig. 12) 
 
 There seems to be another species of Pseudamusium in the collection 
 which is only represented by one good example of a left valve (7304), and 
 it differs from the foregoing in its shape and apparently in its ornamenta¬ 
 tion. It is not oblique and is higher than long, being a bilaterally symmetri¬ 
 cal oval shell like a Pecten; the sides of the beak which is central in position, 
 meet at about 75 0 ; the anterior ear is shorter broader, and more equilater- 
 ally triangular than in Ps. stappenbecki and shows concentric lines instead of 
 radial ribs. The long, straight shoulders of the body descend nearly to the 
 middle of the valve. The surface is gently convex, and though the shell is 
 mostly missing, there is a regular fine radial lineation visible over the whole 
 surface, passing into broad, flattened ribs near the margins, which are more 
 distinct than in the previous species, and the concentric ridges and striae 
 are less well developed. 
 
 This shell seems to resemble Ps.fibrillosum (Salt.) of the Coal Measures 
 of England, as described by Wheelton Hind , 84 but still more Ps. ellipticum 
 Phill . 85 of the Carboniferous limestone of Great Britain and Ireland. 
 
 Dimensions —(7304): Height, 25 mm.; length, 19 mm. 
 
 Pseudamusium cf. fibrillosum (Salter) 
 
 There is one left valve (7324) of a species of Pseudamusium which seems 
 more like Ps. fibrillosum (Salt.) 85 as described by Wheelton Hind than 
 Ps. anisotum, Ps. ellipticum, or Ps. redesdalense Wh. Hind. It has the up¬ 
 right oval shape and vertical axis of Ps. ellipticum and Ps. fibrillosum, but 
 the finer ornamentation of the latter without any wide, flat ribs. 
 
 Aviculopecten barrealensis sp. nov. (Plate XIII, Fig. 14) 
 
 Shell obliquely oval, gently convex. Body subtriangular, rounded; 
 beak subanterior, small, with sides meeting at about 90°. Anterior margin 
 of shell short, strongly arched, sweeping down and back obliquely into well- 
 
 78 Wheelton Hind, Brit. Carb. Lamellibr., vol. II, 1903, p. 104, PI. XXI, Figs. 13 to 20. 
 
 79 De Verneuil, Geol. Russ., II, Palaeont., p. 329, PI. XXI. Fig. 7. 
 
 80 Diener, Himal. Foss., vol. I, pt. 2, 1899 (Palseont. Ind.), p. 16, PI. I, Fig. 3. 
 
 81 Stuckenberg, Mem. Com. Geol. Russ., vol. XVI, no. 1, 1898, p. 200, t. I, Fig. 22. 
 
 82 Katzer, Grundziige d. Geol. d. unter. Amazonasgebietes (Leipzig, 1903), p. 185, 
 t. VIII, Fig. 6. 
 
 83 Morningstar, Geol. Surv. Ohio, Ser. 4, Bull. 25, 1922, p. 230, PI. XIII, Figs. 7, 8. 
 
 84 Wheelton Hind, op. cit., p. 106, PI. XVI, Figs. 16 to 22. 
 
 85 Ibid., p. 103, PI. XX, Figs n to 18. 
 
UPPER CARBONIFEROUS FOSSILS FROM ARGENTINA 
 
 143 
 
 arched inferior margin which curves up sharply behind with short rounded 
 projecting posterior end. Ears well marked off from body, flattened 
 depressed; anterior ear rather large, triangular, projecting, with rounded 
 anterior margin; [posterior ear larger (?), triangular, with acute (?) posterior 
 angle]. Surface of valve covered with fine, regular, straight, mostly equi¬ 
 distant, threadlike, raised radiating ribs, of equal strength, thirty to thirty- 
 three in number, less than half being primary and continuous from beak to 
 margin, the rest arising by intercalation mostly at about one-fourth the 
 length of the others; interspaces wide, flat, all of them as well as the ribs 
 being crossed by rather closer, finer, regular, threadlike raised lines which 
 are equidistant in the lower half of the shell, but become more and more 
 crowded and numerous toward the beak. Anterior ear with concentric 
 growth-lines. 
 
 Dimensions —(7362): Length (oblique), 11.5 mm.; height, 9.0 mm. 
 
 Remarks —Only one specimen (7362) occurs in the collection, but it is 
 a nearly perfect left valve. The surface of this shell is divided up into 
 numerous oblong cancellce by the crossing of the radial and concentric 
 lines, as in A.forbesii McCoy, 86 which it also much resembles in shape, but 
 this Argentine species differs by the ribs being of equal strength, and the 
 obliquity of the shell is more marked and the body narrower and not subor- 
 bicular. Ps. radialis (Phill.) as figured by Waagen 87 from the Salt Range 
 has a somewhat similar ornamentation. 
 
 Nuculana (Leda ) cf. bellistriata Stevens (Plate XIII, Fig. 7) 
 
 One small right valve of a species of Nuculana occurs in the collection 
 (7339)* but it is only an internal cast and does not show the dentition. It 
 seems somewhat like N. bellistriata Stevens, 88 from the Pennsylvanian of 
 the United States, but is shorter and blunter behind. Probably it is a 
 new species. 
 
 Dimensions —Length, 10 mm.; height, 7 mm. 
 
 Euomphalus subcircutaris Mansuy (Plate XV, Figs. 1,2,4) 
 
 There is a species of Euomphalus rather abundantly represented in the 
 collection and including several well-preserved specimens which seems 
 absolutely indistinguishable from Mansuy’s E. subcircularis 89 from the 
 Carboniferous of Chouei Tang, Yunnan. Apparently this is the shell which 
 Stappenbeck mentions from Barreal as resembling E. parvus Waag., 90 but 
 as being much larger. 
 
 The specimens average about 25 to 30 mm. in diameter, and some 
 of them (7322) are well preserved and show both the apical and basal sur¬ 
 faces. In all respects they agree with Mansuy’s figures and description, 
 which need not be repeated. 
 
 Pleurotomaria ad vena sp. nov. (Plate XV, Fig. 5) 
 
 Shell low, conical, heliciform, height and diameter being about equal, 
 basal whorl occupying about half the height; apical angle about 130°; 
 whorls rounded, five to six in number, rapidly decreasing in size to apex. 
 Basal whorl large, rounded below, more or less flattened above, rising at 
 an angle of 45 0 to 6o° from the periphery, which is obtuse; slit-band promi- 
 
 86 Wheelton Hind, op. cit., p. 83, PI. XVIII, Figs. 3 to 7. 
 
 87 Waagen, op. cit., p. 280, PI. XXIII, Fig. 5. 
 
 88 Girty, Bull. 544, U. S.Geol. Surv., 1915, p. 122, PI. XIV, Figs. 1 to 9; Jakowlew, 
 Mem. Com. Geol. Russ., n.s., Livr. 4, 1903, p. 10. 
 
 89 Mansuy, op. cit., vol. I, fasc. 2, 1912, p. 105, PI. XIX, Figs. 3a to d. 
 
 90 Waagen, op. cit., p. 89, pi. IX, Fig. 2. 
 
144 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 nent, situated slightly above the periphery, narrow, concave, deep, with 
 sharp raised edges. Upper whorls rounded. Sutures not deeply sunken, 
 with slit-band on upper whorls situated on or slightly above them. Base of 
 shell convex, rounded; umbilicus small, closed, depressed; lip of mouth 
 somewhat thickened. Shell ornamented with revolving liras crossed by finer 
 spiral lines which at their intersection form small pustules or nodules giving 
 a moniliform character to the lirae; on apical surface of each whorl above the 
 slit-band there are five to seven of such equidistant revolving liras of equal 
 strength, closely nodulated, mostly with finer, similarly nodulated liras 
 between them; below the slit-band on the basal whorl are more numerous 
 and more closely placed finer nodulated liras, with the closely placed spiral 
 lines bending back to meet the slit-band at an angle of about 20°. 
 
 Dimensions —Height, c. 30 mm.; basal diameter, c. 35 mm. 
 
 Remarks —There is one specimen (7355) somewhat crushed but showing- 
 well the basal whorl, slit-band, and ornamentation. Another (7360) shows 
 the spire and upper whorls better. The species seems to be allied to PL 
 punjabica Waag. 91 of the upper Productus limestone of the Salt Range, 
 but differs in the upper whorls being more rounded, in the slit-band being 
 sutural, and in the more marked nodulation of the revolving liras. We may 
 perhaps also compare PI. timorensis Wanner 92 from the Permo-Carbonifer¬ 
 ous of Timor and Ptychomphalus venustus Mansuy 93 from the correspond¬ 
 ing beds of Indo-China, especially with regard to the shape of the shell and 
 its ornamentation. 
 
 Pleurotomaria argentina sp. nov. (Plate XV, Figs. 3a, 3 b, 3c) 
 
 Shell conical, subtrochiform; apical angle 75 0 ; whorls five to six in 
 number, bluntly angulated at periphery, slowly decreasing in size. Basal 
 whorl large, fully half the total height of the shell, with swollen, convex 
 rounded base, subangular periphery bearing a prominent, well-defined, 
 sunken, narrow slit-band having sharply raised edges, and with a somewhat 
 flattened apical face sloping up to suture-line. Upper whorls with promi¬ 
 nent suture-band situated a little distance above suture-line, which is not 
 sunken. Surface of whorls marked with regular revolving lirae, numbering 
 six to seven on apical face, equidistant and bearing regular sharp, equi¬ 
 distant, small spinose tubercles; below the slit-band on the upper whorls 
 are one to three similar lirae. Lower surface of basal whorl with several 
 similar equidistant lirae immediately below slit-band, but nearer umbilicus 
 they become finer and less distinct or obsolete, fine spiral growth-lines 
 being here specially developed. Umbilicus sunken, closed. Mouth large, 
 subcircular. 
 
 Dimensions —(7334): Height, c. 27 mm.; diameter, c. 24 mm. 
 
 Remarks -—This species is closely allied to PI. ( Phanerotrema) gray - 
 villensis Norvv. and Pratt, as described by Girty 94 from the Pennsylvanian of 
 Oklahoma, differing chiefly in the absence of strong spiral lines crossing the 
 revolving lirae, and of the transverse ridges below the suture-lines on the 
 apical surface of the whorls. 
 
 Pleurotomaria barrealensis sp. nov. (Plate XVI, Figs. in, 16) 
 
 Shell subturbinate, composed of four to five rounded whorls, rather 
 rapidly decreasing in size to apex; apical angle about 8o°. Basal whorl 
 
 91 Waagen, op. cit., p. 115, PI. XI, Figs. 3 and 4. 
 
 92 Wanner, Palaeont. Timor, Lief. XI, Abh. XVIII, 1922, p. 20, t. 151, Figs. 9 a and 6. 
 
 93 Mansuy, op. cit., vol. II, fasc. 4, 1913, p. 102, PI. XI, Fig. 3. 
 
 94 Girty, Bull. 544 U. S. Geol. Surv., 1915, p. 149, PI. XXIII, Figs. 2 to 8. 
 
UPPER CARBONIFEROUS FOSSILS FROM ARGENTINA 
 
 145 
 
 large, rounded-subquadrate in section, being somewhat flattened on outer 
 face; slit-band narrow, inconspicuous, situated at junction of outer face 
 with convex apical face; inferior surface of basal whorl rounded, but not 
 strongly convex, somewhat flattened; umbilicus sunken, small; mouth 
 transversely oval, wider than high, with thickened inner lip. Upper whorls 
 with more or less flattened vertical outer face below the slit-band, which is 
 situated in about the middle of each whorl. Inferior surface of basal whorl 
 crossed by strong subequidistant, sublamellose, spiral growth-lines and by 
 numerous rather closely placed, subequidistant, thick, thread-like, smooth 
 revolving liras, not tuberculated; outer face of basal whorl ornamented with 
 similar more widely spaced revolving liras; apical face (not well preserved) 
 ornamented with similar revolving liras. 
 
 Dimensions —Height, c. 55 mm.; diameter at base, c. 55 mm. 
 
 Remarks —There is only one specimen (7323) of this shell, but it is 
 nearly complete, though the surface ornamentation is destroyed, except on 
 the basal whorl, and the slit-band is nowhere distinct. We may compare this 
 species with PI. orientalis Roemer, as figured by Fliegel 95 from the upper 
 Carboniferous of Padang, Sumatra, and with PL cf. punjabica Waag. as 
 figured by Diener 96 from the Productus shales of the Himalayas. 
 
 Metoptoma ? sp. 
 
 One much-crushed subcircular shell (7318) may possibly be referred to 
 the genus Metoptoma , for apparently it had a low capuliform shape with an 
 elevated submarginal apex scarcely overhanging the posterior margin. The 
 surface of the shell seems to have been gently convex, very obliquely coni¬ 
 cal, and there are traces of twx> to three small, short marginal radial folds 
 in the median line on the anterior edge. On one part of the central pre- 
 umbonal region regular fine, equidistant, incised radial lines are visible, and 
 a few concentric growth-strias and ridges sweep around the margins and pass 
 beneath the umbo. 
 
 Dimensions —Diameter, 22 to 25 mm.; height, + 4 mm. 
 
 Orthoceras sp. 
 
 One small fragment of a subcylindrical shell with a circular cross-section 
 is probably referable to a species of Orthoceras. The specimen is only 10.5 
 mm. in length and shows five to six horizontal septa in that distance. The 
 siphuncle is central and circular and measures one-sixth to one-seventh of 
 the diameter. The walls of the cameras seem to be unusually thick. The 
 external surface of the shell seems to be smooth, except for very delicate 
 concentric striae. The affinities of this fragment are not clearly determinable, 
 owing to the poorness of the specimen. Another fragment (7354) of appar¬ 
 ently the same species consists of one broken, isolated camera; the septa 
 are seen to be deeply concave and the siphuncle apparently expands in the 
 upper part of each chamber. If the thickness of the walls of the chambers 
 as seen in the other specimen is due to a marginal infilling we may per¬ 
 haps compare Pseudorthoceras knoxense Girty 97 from the Pennsylvanian of 
 America, which also agrees in its slow rate of tapering, central siphuncle, 
 distance apart of the septa, and circular section. The rate of tapering, 
 central siphuncle, circular section, and distance apart of the septa in our 
 specimens also suggest a comparison with Orth, punjabiense Waag. 98 from 
 the middle Productus limestone of Musakheyl. 
 
 95 Fliegel, Palaeontographica, Bd. XLVIII, Lief. 3, 1901, p. 113, t. VIII. Figs. 7 and 8. 
 
 96 Diener, Palaeont. Ind., ser. XV, Vol. I, pt. 5, 1903, p. 100, PI. V, Figs. 1 to 3. 
 
 97 Girty, Bull. 544 U. S. Geol. Surv.. p. 227, PI. XXVII, Figs. 1 to 6 
 
 98 Waagen, op. cit p. 71, PI. VI. Fig. 11. 
 
146 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 COMPOSITION AND CORRELATION OF THE FAUNA 
 
 The following is a list of the species in the present collection 
 which have been identified and described above: 
 
 * Fenestella aft. perelegans Meek. 
 
 sp. ind. 
 
 Polypora aff. biseriata Ulr. 
 
 Orbiculoidea saltensis sp. nov. 
 
 *t Productus lineatus Waag. 
 
 * curvirostris Schellw. var. nov. barrealensis. 
 
 * cf. juresanensis Tschern. 
 
 * (Marginifera ) cf. echinata Waag. 
 
 * ( Marginifera) spinulo-costatus Abich var. nov. peregrina. 
 
 *Chonetes cf. pseudovariolata Nik. 
 
 granulifer Owen. 
 
 *Camarophoria cf. mutabilis Tschern. 
 
 *Spirifer cf. supramosquensis Nik. 
 
 * wynnei Waag. var. nov. argentina. 
 saltensis sp. nov. 
 
 barrealensis sp. nov. 
 
 mexicanus Shum. var. nov. neotropica. 
 
 * aff. rajah Salt. 
 
 * (Martinia) cf. simensis Tschern. 
 
 * ( Martiniopsis ) cf. aschensis Tschern. 
 
 *Spiriferina zewanensis Diener. 
 
 Reticularia notica sp. nov. 
 
 *\Ambocoelia plano-convexa Shum. 
 
 Athyris (Spirigerella ) ? sp. 
 
 *Hemiptychina cf. sublcevis Waag. 
 
 Pseudamusium stappenbecki sp. nov. 
 
 cf. ellipticum (Phill.). 
 cf. fibrillosum (Salt.). 
 
 Aviculopecten barrealensis sp. nov. 
 
 Nuculana (Leda ) cf. bellistriata Stevens. 
 
 Euomphalus subcircularis Mansuy. 
 
 Pleurotomaria advena sp. nov. 
 
 argentina sp. nov. 
 barrealensis sp. nov. 
 
 Metoptoma ? sp. 
 
 Orthoceras sp. 
 
 With regard to the two horizons in the shales from which Doctor 
 Du Toit’s specimens were collected the following lists indicate the 
 distribution of the species: 
 
 Lower horizon (7281-7313). 
 
 Polypora aff. biseriata. 
 
 Fenestella sp. ind. 
 
 Productus cl. juresanensis. 
 
 Spirifer wynnei var. argentina. 
 saltensis. 
 
 Spiriferina zewanensis. 
 
 Hemiptychina cf. sublcevis. 
 
 Pseudamusium stappenbecki. 
 
 cf. ellipticum. 
 
UPPER CARBONIFEROUS FOSSILS FROM ARGENTINA 
 
 147 
 
 Upper horizon (7314-7376). 
 
 Fenestella aff. perelegans. 
 
 Orbiculoidea saltensis. 
 
 Productus lineatus. 
 
 curvirostris var. barrealensis. 
 
 ( Marginifera) cf. echinata. 
 
 ( Marginifera) spinulo-costatus var. peregrina. 
 
 Chonetes cf. pseudovariolata. 
 granulifer Owen. 
 
 Camarophoria cf. mulabilis. 
 
 Spirifer cf. supramosquensis. 
 barrealensis. 
 
 mexicanus var. neotropica. 
 
 {Martinia) cf. simensis. 
 
 ( Martinia ) cf. aschensis. 
 
 Reticularia notica. 
 
 Pseudamusium cf.fibrillosum. 
 
 Aviculopecten barrealensis. 
 
 Nuculana cf. bellistriata. 
 
 Euomphalus sub circular is. 
 
 Pleurotomaria advena. 
 
 argentina. 
 
 barrealensis. 
 
 Metoptoma ? sp. 
 
 Orthoceras sp. 
 
 The species marked with an asterisk (*) in the above list occur 
 either in the Carboniferous of Russia or parts of Asia, though several 
 are represented here by varieties. In many cases we have not been 
 able to feel sure that the other species are identical, owing to the 
 paucity or poor preservation of the specimens in the present col¬ 
 lection, so that they can only be compared with already described 
 forms. A few are, however, identified without hesitation; such are 
 Productus lineatus , Chonetes granulifer, Amboccelia planoconvexa, Spiri- 
 jerina zewanensis, and Euomphalus sub circular is. 
 
 The species marked with a dagger (f) have been recorded by 
 previous geologists from South America, but these are very few, and 
 the remainder show little affinity. The fauna, in fact, of this Argen¬ 
 tine locality is quite distinct from that of Brazil or Bolivia, and does 
 not show any striking resemblance to any occurring in North 
 America, except perhaps to the Guadelupian type. 
 
 Derby," in 1894, gave a list of the Amazonian upper Carbonifer¬ 
 ous fauna, but without any figures or descriptions. Many had been 
 described in his earlier paper on the Brazilian Carboniferous fossils 
 from Itaituba. The most complete list of the upper Carboniferous 
 fossils from the Amazon region, with descriptions and figures of 
 several of the species, was given by Katzer, 100 in 1903. Among the 
 species occurring in this area and the present collection from Barreal 
 
 99 Derby, Journal of Geology, vol. II, 1894, p. 480. 
 
 100 Katzer, Grundzuge d. Geol. d. unter Amazonasgebietes (Leipzig, 1903), pp. 142- 
 188, 263-268, t. IV-VIII. 
 
148 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 are: Productus lineatus and Amboccelia plano-convexa, while probably 
 some of the shells referred to Sp. condor by Katzer ( e.g., op. cit., t. 
 V, Fig. i) should be identified with Sp. ef. supramosquensis , and Pr. 
 nebraskensis may be the same as the shell compared with Pr. juresa- 
 nensis. But the Amazonian fauna, which is a large one, consists 
 mainly of species not recognized at Barreal. Katzer (op. cit., 
 pp. 246-253) has discussed the relations of this fauna with that of 
 other regions and has referred to the literature on the subject. 
 
 The relations of the Argentine fauna to that of the upper Car¬ 
 boniferous of Bolivia, of which the brachiopods alone have been 
 adequately studied, is not close, as is seen by Kozlowski’s 101 memoir. 
 But Amboccelia plano-convexa and Productus lineatus ( = cora (pars) 
 Kozl.) are species which the two faunas have in common. 
 
 Meyer, 102 in 1914, described marine upper Carboniferous faunas 
 from Peru and Bolivia and discussed their occurrence generally in 
 South America, referring to Stappenbeck’s discovery of such at 
 Barreal, but without giving any further particulars. Of the Argen¬ 
 tine species, Productus cora (including Pr. lineatus ) alone is men¬ 
 tioned as occurring in Bolivia, but in Peru we note, in addition to this 
 species, Amboccelia plano-convexa. Reviewing the evidence, Meyer 
 assigns all the faunas to the upper Carboniferous and not to the 
 Permian. Douglas 103 in 1914 briefly described a similar fauna of the 
 Titicaca district, ascribing it to the upper Carboniferous or Permo- 
 Carboniferous, and he recognized that a few forms were related 
 to “Permian” forms from the Salt Range. None of the Argentine 
 species appear to occur. In a subsequent paper 104 he records the 
 occurrence of lower Carboniferous fossiliferous beds in southern 
 Peru, but the fauna is quite distinct from that of Barreal. 
 
 There is not any very close resemblance with the fauna of any 
 part of the Productus limestones of Asia, though a few species are 
 apparently identical and several are allied or comparable. Hayasaka , 105 
 however, has described a small fauna of brachiopods from the 
 upper Carboniferous of the coal-bearing series of Hon-kei-ko, 
 Manchuria, which includes the following species occurring at 
 Barreal: Camarophoria cf. mutabilis, Spirifer supramosquensis, Sp. 
 wynnei, and some other allied species of Spirifer. 
 
 If we accept Holland’s correlation of the Productus limestones 106 
 we only find Spiriferina zewanensis of the Zewan beds as a Permian 
 form, though Pleurotomaria advena is more allied to an upper Produc- 
 
 101 Kozlowski, Brach. Carb. Super. Bolivie, Annales de Paleont. IX, 1914, pp. 1 to 100 
 Pis. I to XI. 
 
 102 Meyer, Neues Jahrb. f. Mineral, etc., Beil. Bd. XXXVII, 1914, pp. 590-651, 
 t XIII XIV 
 
 103 Douglas, Quart. Journ. Geol. Soc., vol. LXX, 1914, pp. 30-37, Pis. VIII, IX. 
 
 1M Ibid., vol. LXXVI, 1920, p. 41. 
 
 105 Hayasaka, op. cit., 1922, pp. 117-137. 
 
 106 Holland, Mem. Geol. Surv. India, vol. LI, pt. 1, 1926, p. 22. 
 
UPPER CARBONIFEROUS FOSSILS FROM ARGENTINA 
 
 149 
 
 tus limestone species than to any other. Productus lineatus and 
 Amboccelia planoconvexa, though occurring in the Zewan beds, are 
 abundant also on lower horizons and in the so-called upper Carbonif¬ 
 erous of Russia and America. Chonetes granulifer is typically a 
 Pennsylvanian species; and the species Sp. supramosquensis is char¬ 
 acteristic of the Gshelian, which occurs immediately above the 
 Moscovian and is therefore well below the Permian. If we look at 
 the rest of the species from Barreal, we observe that nearly all are 
 closely allied to or comparable with established species from either 
 the lower Carboniferous of Europe or the middle or upper Car¬ 
 boniferous of Russia or North America, and that none are limited 
 to undisputed Permian beds. Thus among the new forms, Spirifer 
 saltensis shows close affinities with some examples of Sp. duplicicosta 
 of the lower Carboniferous; Sp. barrealensis is especially allied 
 to Sp. boonensis; Reticularia notica is related to R. setigera of the 
 lower Carboniferous; Pseudamusium stappenbecki has affinities with 
 upper Carboniferous species; Pleurotomaria argentina closely resem¬ 
 bles a species from the Pennsylvanian of the United States, and 
 PI. barrealensis is like an upper Carboniferous form from Sumatra. 
 
 Thus the weight of evidence from the fossils leads us without 
 hesitation to refer this fauna from Barreal to the Carboniferous , and 
 it seems that it should not be placed at the top of that formation, but 
 probably toward the base of the upper division. It is interesting to note 
 that Holland (op. cit.) places the glacial boulder bed of the Salt 
 Range in the Carboniferous, for the tillite of Barreal is stated to pass 
 up into and even include some of the beds from which the present 
 collection of fossils was obtained. 
 
150 GEOLOGICAL COMPARISON OF SOUTH AMERICA WITH SOUTH AFRICA 
 
 TRIASSIC FOSSILS FROM RIO CLARO, PARANA, BRAZIL 
 
 The following species have been determined provisionally from the 
 material forwarded by Doctor Du Toit, but it is intended to publish full 
 descriptions and figures elsewhere at some later date. The fossils from 
 the upper zone have presented difficulties in their identification owing to 
 their occurrence in the form of casts. 
 
 Lower Horizon—Zone i : 
 
 Pachycardia aff. rugosa Hauer. 
 
 Anodontophora aff. trapezoidalis Mansuy. 
 
 Trigonodus aff. rablensis (Gredler). 
 
 Radiolaria (not determinable). 
 
 Upper Horizon—Zone 2 : 
 
 Pachy car dia neotropica sp. nov. 
 
 Myophoria (Myophoriopis ) aff. carinata Bittn. 
 
 aff. lineata Miinst. 
 
 Megalodus ? neolropicus sp. nov. cf. triqueter (Wulf). 
 
 Gonodon ( Schafhautlia ) paranaense sp. nov. 
 
 Modiola aff. subcarinata Bittn. 
 
 Solenomorpha in the sense used by Holdhaus is not represented in the 
 collection, nor are there any species common to the two zones. Permian 
 forms are absent, the fossils from both horizons displaying undoubted 
 Triassic characters suggestive of the upper rather than the lower Trias. 
 
INDEX 
 
 Authorities Cited in the Text 
 
 Aguirre, 19, 21 
 
 Arocena, 3, 4 
 
 Ave-Lallemant, 47, 48 
 
 Baker, C. L., 61, 83, 87, 89, 90, 92-93 
 
 Baker, H. A., 11, 26 
 
 Beder, 60, 77 
 
 Berry, 40, 55 
 
 Bodenbender, 28 , 35 ~ 36 , 38-39, 42 - 44 , 
 46-47 
 
 Bonarelli, 56, 58, 98 
 
 Brackebusch, 43 
 
 Branner, 62-64, 7 °, 93 , 95 - 96 
 
 Brouwer, 114 
 
 Chamberlin, 65 
 
 Clarke, 11, 60-61, 100 
 
 Coleman, 23, 31, 67, 70-71, 74 , 104 
 
 Crandall, 64-65, 94 
 
 Daly, 2 
 
 Derby, 61, 63-64, 94, 147 
 Douglas, 57, 148 
 Draper, 65, 67 
 Du Toit, 94, 101, 103, 119 
 Evans, 61 
 Ferraz, 59, 74 
 Fourmarier, 9 
 Gothan, 55 
 Gregory, 9 
 Groeber, 27, 51 
 Guillemain, 67, 73 
 Halle, 11-12 
 Harder, 65 
 Haug, 11 
 Hausen, H., 35 
 Hausen, J., 88, 90, 93 
 Hauthal, 19, 23 
 Heald, 55-57 
 Holdhaus, 81-82 
 Holland, 148-149 
 Huene, 79 
 Katzer, 147-148 
 
 Keidel, 14, 18-30, 32-35, 37, 42, 47, 49, 
 51,98 
 Knod, 97 
 
 Kozlowski, 97-98, 148 
 
 Kurtz, 36, 38-41, 89 
 
 Leme, 59, 61, 63-64, 82, 85, 89, 95 
 
 Lemoine, 10 
 
 Lisbda, 94 
 
 Lundqvist, 77 
 
 Mather, 55-57 
 
 Meyer, 148 
 
 Nagera, 5 
 
 Oliveira, 4, 67, 70, 72, 75, 81,83-84 
 Orbigny, 56, 62 
 Pacheco, 71, 76, 79, 83, 87 
 Pastore, 5 
 
 Penck, 36-38, 43 , 35 
 Rassmuss, 51-52 
 Reed, 5, 35, 60, 82, 101 
 Rego, 83-84 
 Rimann, 53, 70 
 Rogers, 27 
 Schiller, 19, 20 
 Seward, 11-12, 40 
 Sobral, 4, 90 
 Solms-Laubach, 50 
 Soper, 95 
 
 Stappenbeck, 28, 32-33, 35 , 38, 43, 47 ~ 49 > 
 
 131 
 
 Stelzner, 47 
 Taylor, 1 
 
 Walther, 3-4, 62, 66-67, 69, 73, 76, 78, 
 82-84, 87, 90, 93 
 Wegener, 1, no, 115, 119 
 White, D., 77-78, 82, 96 
 White, I. C., 66, 72, 80, 83-84, 86 
 Windhausen, 27, 32, 51, 60, 103, 107-108 
 Woodworth, 70-74, 89, 92 
 Wright, 2 
 Zeiller, 77, 82 
 
 Fossils Described in Appendix 
 
 Ambocoelia plano-convexa, 141 
 Athyris, 141 
 
 Aviculopecten barrealensis, 142 
 Camarophoria cf. mutabilis, 135 
 Chonetes granulifer, 134 
 
 cf. pseudovariolata, 134 
 Euomphalus subcircularis, 143 
 Fenestella aff. perelegans, 131 
 Hemiptychina cf. sublaevis, 141 
 Metoptoma, 145 
 Nuculana, 143 
 Orbiculoidea saltensis, 132 
 Orthoceras, 145 
 Pleurotomaria advena, 143 
 
 argentina, 144 
 barrealensis, 144 
 Polypora aff. biseriata, 132 
 
 Productus curvirostris, 133 
 
 cf. juresanensis, 133 
 lineatus, 133 
 
 (marginifera) cf. echinata, 133 
 spinulo-costatus, 
 134 
 
 Pseudamusium cf ellipticum, 142 
 cf. fibrillosum, 142 
 stappenbecki, 141 
 Reticularia notica, 140 
 Spirifer barrealensis, 138; cf. aschensis, 139 
 cf. mexicanus, 136 
 saltensis, 138 
 cf. simensis, 139 
 supramosquensis, 135 
 wynnei, 136 
 
 Spiriferina zewanensis, 139 
 
 151 
 
152 
 
 INDEX 
 
 General 
 
 Abaet6, 64, 70 
 Accra, 10-11 
 Adiantites, 36, 40 
 
 Afro-Americanlandmass, 1, 97,99,107,109 
 rupturing of, 108, 117, 119 
 Agoniatites, 100 
 Agua Suja, 70, 83 
 Alethopteris, 95 
 Alhuampa, 46, 58, 77 
 Alkaline rocks, 114 
 Amazon syncline, 13, 15 
 
 valley, Carboniferous, 57, 101, 104, 
 H 7 -I 49 
 
 valley, Devonian, 61, 99, 115 
 Ambocoelia, 147-149 
 Amygdaloid, 84, 92, 95 
 Anaplophora, 45 
 
 Andes, 14, 45, 108. See Cordillera 
 Andesite, 48, 89 
 
 Andine foldings, 14, 28, 42, 107-108, 117, 
 120 
 
 Angola, 9, 16, 107, 114, 117 
 Annularia, 32, 77 
 Anodontophora, 82, 150 
 Araucaria, 55 
 Archaean: 
 
 Africa, 6 
 Brazil, 13, 115 
 Archaeopteris, 36, 40 
 Areado sandstone, 70 
 Argentina: 
 
 Carboniferous, 29-41 
 compared with Australia, 34, 39-41, 
 54, 103, 106 
 
 Devonian, 22, 33, 35, 55, 98, in 
 Gondwana, 22-58 
 
 glacials, 14, 22-25, 28, 30-36, 41, 
 
 hi 
 
 Ordovician, 19, 28, 31-32, 56 
 Silurian, 18, 28, 97-98 
 Arthrophycus, 19 
 Assunguy series, 59, 63, 114 
 Asterocalamites, 36-37 
 Asuncion, 60, 77 
 Athyris, 101 
 Atlantic basin: 
 
 central rise, no 
 origin, 108, 119 
 volcanicity, 103, 108, no 
 Atlantic coasts, 15-16, 18, 107, no, 117 
 comparison, 1516 
 Atrypa, 19 
 
 Austral Devonian faunas, 6, 12, 16, 22, 99 
 Australia, comparison with Argentina, 41, 
 . 54 
 
 Avicula, 45 
 
 Bahia, 13, 16, 64-65, 95~96, 107, 114 
 
 Bahia Blanca, 14-15, 19, 27, 107, 110-111 
 
 Baiera, 47-48, 50 
 
 Bajo de Velis, 37, 39, 43~44 
 
 Balcarce, 19 
 
 Bambuhy series, 63-64, 70, 114 
 Barreal, 34 ~ 35 , 4 7 , ioi, 131, 148-149 
 Basaltic lavas: 
 
 Argentina, 27, 48, no 
 Brazil, 14, 89-94, Io6 , 113 
 
 Basaltic lavas —continued 
 South Africa, 8, 106, 113 
 Basement rocks, 6, 13-15, 64-65, 115 
 Bauru beds, 61, 86, 96 
 Beaufort series, 8, 105 
 age, 8 
 flora, 8, 13 
 
 Bello Horizonte, 63-64 
 Benguella, 9, 15 
 Bermejo series, 55-58 
 age, 57 
 
 glacials, 58, 103 
 Bituminous shales, 48, 50, 69 
 B6a Vista, 65, 114 
 Bokkeveld beds, 6, 12, 112 
 Argentina, 20, 22, in 
 Bolivia, 98 
 Brazil, 59, 60, 112 
 Falklands, 12 
 fauna, 100 
 Bolivia: 
 
 Carboniferous fauna, 56-57, 147-149 
 Carboniferous flora, 40 
 Permo-Triassic, 55-58 
 Siluro-Devonian, 97-98 
 Bonito beds: 
 age, 78 
 
 Argentina, 76, 85, 92 
 Brazil, 75-76, 112 
 coals, 75-76 
 flora, 77 
 Paraguay, 77 
 Uruguay, 69, 76 
 
 Boreal, Devonian faunas, 16, 61, 99, 100 
 Botrychiopsis, 36, 40 
 Botucatu beds, 83, 85-87, 90, 106, 113 
 interbedded lavas, 90 
 Boulder: 
 
 Beds. See Glacials 
 Pavements, 24, 34, 102 
 Brasilides, 14, 18, 62, 96 
 Brazil : 
 
 Carboniferous marine, 13,72, 101,147 
 Cretaceous, 61, 86, 96 
 Devonian, 13-^15, 60-61, 99, 100, 112 
 Permo-Triassic, 75-95 
 Pre-Cambrian, 15, 64-65 
 Triassic marine, 81-82, 150 
 Brazilian Coal Commission, 66, 75 
 Bushveld sandstone, 8, 106, 113 
 Cagapava, 63, 84 
 Cacheuta, 50, 54 
 Calamites, 40 
 
 Calchaquenos, 28, 31, 35, 50 
 Calvinia, 7 
 Camarophoria, 135 
 Cameroons, 10-n, 16, 114 
 Campinas, 70, 74, 76 
 Cape, geology, 6-9 
 Cape Corrientes, 18 
 Cape Fold-ranges, 7, 14, 19, in 
 Cape system, 6, 12 
 Capetown, 15, 62 
 Caraga quartzite, 64 
 Carboniferous: 
 
 Amazon, 13, 101, 147 
 
INDEX 
 
 153 
 
 Carboniferous —conti nued 
 Argentina, 29-41 
 
 fossils described, 131-149 
 affinities, 147-148 
 horizons, 146-147 
 Asia, 147-149 
 Bolivia, 56-57, 148 
 Chile, 51 
 
 Eastern Brazil, 13, 96 
 flora, 29, 37, 39-41, 100 
 glacials, Falklands, 12 
 
 South Africa, 7, 19, 101-102 
 South America, 15, 30-35, 72, 100 
 conformity below, 7, 23 
 glaciation, 101-104 
 movements, 6, 100 
 ocean, 100-101, 104 
 Peru, 100, 148 
 Russia, 147-149 
 South Africa, 7 
 Western Africa, 11 
 Cardiopteris, 29, 36-38, 40, 103 
 Carpinterfa, 31, 36, 40 
 Cave sandstone, 8, 88, 90, 106, 113 
 Ceara, 13, 16, 114 
 Ceres, 7, 102 
 Cerro Bold, 37, 39, 45-46 
 Bonete, 23, 25 
 Largo, 63, 67, 73, 76, 81, 84 
 Morado, 46 
 Pelado, 32, 37-38 
 Challao, 48 
 Chapada, 61, 87 
 Chemnitzia, 56 
 Cherts, 80-82 
 Chile, Carboniferous, 51 
 Liassic, 107 
 Rhaetic 50 
 Triassic, 50-51 
 Chiropteris, 54 
 Chubut, 51, 107-108 
 Cladophlebis, 47, 49, 50, 54, 89 
 Clanwilliam, 15, 60, 102, m-112 
 Clarkeia, 98 
 Coal measures, 8 
 Coals: 
 
 Bonito, 75-76 
 Chile, 50 
 
 Paganzo, 32, 38, 46-47, 53 
 Uruguay, 76 
 Cochabamba, 56 
 Congo, 9, 16 
 
 Congo basin, 9, 95, 106, 114 
 glacials, 10, 58 
 
 Continental disruption. See Displacement 
 Hypothesis 
 Copiapo, 50 
 
 Cordaites, 36-37, 39-41, 77 
 Cordillera, 32, 50-51, 56, 108 
 Cdrdoba, 45, 53 
 Corumbatahy beds, 80-81 
 Cretaceous: 
 
 Argentina, 27, 47-48, 52, 55, 58, no 
 Brazil, 14, 16, 61, 9=5-96, 113-114 
 Mediterranean facies, 16 
 Movements, 10, 51, 107, no 
 Ocean, 107, 119 
 Patagonia, 14, 52 
 
 Cretaceous —continued 
 
 South Africa, 9, 52, in 
 Western Africa, 10, 16, 114 
 Culm flora, 32 
 Curitiba, 71, 108 
 Cuyaba, 61 
 
 Cyzicus, 47, 50, 52, 54, 113 
 Dadoxylon, 13, 47, 75, 79, 82 
 Damuda series, 8 
 Devonian: 
 
 Amazon, 13, 15, 61, 100, 115 
 Argentina, 22, 33, 35, 55, 98, hi 
 B olivia, 55, 57, 61, 97-98, 100 
 Brazil, 14—15, 59-61, 100, 112 
 change facies, 22 
 Climate, 99, 100 
 Falklands, 12, 100 
 faunas, 99, 100 
 
 austral, 6, 12, 16, 22, 99 
 boreal, 16, 61, 99, 100 
 North American, 99, 100, 145 
 ocean, 99, 100, 115 
 oscillations, 99 
 Sahara, n, 16, 99, 100, 115 
 South Africa, 6, 12, 16, 100, 111-112 
 Uruguay, 60, 62 
 Western Africa, n, 15, 100, 115 
 Diamantina, 64-65, 108 
 Diamonds, 59, 61, 64-65, 114-115 
 Discina, 72 
 
 Displacement Hypothesis, 1,2, 17, 41, 97, 
 108-110, 118, 121 
 arguments for, 103, 109, 115 
 graphic representation, 116 
 Dinosaurs, 89, 113 
 Dogger, 51 
 Dolerites: 
 
 age, 15, 94 - 106 
 Brazil, 81, 93 ~ 95 , 106, 113 
 Falklands, 13, 102 
 South Africa, 9, 94, 106, 113 
 Dolomite: 
 
 Argentina, 18 
 Brazil, 65, 114 
 Congo, 10 
 South Africa, 6 
 Uruguay, 63, 67 
 Drakensberg volcanics, 8 
 Durazno, 74 
 Dwyka series, 7 
 age, 7 
 
 glacials, 7, 112 
 
 in Argentina, 22-25, 30-36, in 
 in Brazil, 70-72, 74-75 
 in Falklands, 12 
 in Uruguay, 67-69, 73-74 
 shales, 7, 16 
 
 White band, 7, 78, 80, 104-105 
 Brazil, 80 
 
 Earth movements. See Foldings 
 Ecca series, 7, 8, 77, 112 
 age, 8, 77 
 facies, 7, 8 
 flora, 8, 13 
 
 in Argentina, 21, 26, in 
 Eocene, 10, 16 
 Eogene, 107 
 Equisetites, 57 
 
154 
 
 INDEX 
 
 Eremopteris, 40 
 Espinago quartzites, 64-65 
 Estancia beds, 64, 95, 114 
 Estheria, 47-52, 54 “ 55 , 89, 113 
 Estrada Nova group, 80-84 
 age, 82, 150 
 break in, 82, 85 
 fossils, 81-82, 150 
 Eurydesma, 104 
 Falkland Islands, 11-13, 99, 102 
 Devonian, 100, 102 
 geodesic position, 13, 16, 102, 117 
 glacials, 12, 102 
 tectonics, 11, 102 
 Famatina, 38, 44-45, 99 
 Favosites, 63 
 Fenestella, 34, 56, 132 
 Flora: 
 
 Bonito, 77-78 
 Carboniferous Peru, 100 
 Ecca, 8, 77 
 Lafonian, 12-13 
 Liassic, 55 
 Molteno, 54 
 Paganzo, 37-41 
 Trias-Rhaetic, 49, 50, 54 
 Foldings: 
 
 Andine, 14, 28, 42, 107-108,117, 120 
 Brasilides, 14, 18, 62, 96 
 Cape, 7, 14, 19, 107 
 Cretaceous, 10, 51, 107, no 
 Gondwanides, 18, 26-27, 46, 52, 62, 
 105, 107, hi 
 H ercynian, 27 
 
 Permo-Triassic, 18, 27, 46, 105 
 Tertiary, 14, 28, 42, 120 
 Forest sandstone, 18, 88, 91, 106, 113 
 Fraile Muerto, 63, 67-69, 73, 79, 81 
 Furnas sandstone, 59-61, 71, 98, 112 
 Gangamopteris, 12, 32, 36-38 
 flora, 7, 16, 77 
 Glacials: 
 
 Argentina, 14,22-25,28,30-36,41,111 
 
 age, 35, 39, 4L 7§, 103, 149 
 Bolivia, 58 
 
 Brazil, 14, 70-72, 74~75 
 Congo, 10 
 Falklands, 12, 102 
 marine, 33-35, 72, 104 
 Uruguay, 67-69, 73-74 
 Glaciation, Carboniferous, 101-104, 119 
 causes, 104 
 centers, 101-104, 112 
 striated surfaces below, 7, 12, 30, 33, 
 35, 68, 103 
 
 Glossopteris, 12, 36-38, 40-41, 49, 77, 82 
 flora, 12, 14-16, 37, 39, 41, 44, 77, 
 104-105, 119 
 
 Gold coast, 10, 15, 99, 100, 114-115 
 Gondwana beds: 
 
 Argentina, 22-58 
 Brazil, 13-14, 70-72, 74-96 
 Falklands, 12 
 
 Flora. See Glossopteris Flora 
 South Africa, 7, 8 
 Uruguay,67-69, 73-76 
 Gondwanaland, 1,2, 14, 78-84, 87-93, 99 , 
 100, 103, 105, 118-120 
 
 Gondwanaland, breaking up, 107-108 
 glaciation, 103-104 
 
 Gondwanides, 18, 26-27, 46, 52, 62, 105,, 
 107, in 
 
 Goyaz, 61, 74, 79, 82, 87, 94 
 Grao Mogul, 64-65 
 Gravatahy, 72, 83 
 Guandacol, 35 
 Halobia, 51 
 Hastimima, 79 
 Hercynian folds, 27 
 High-level gravels, 19, 26-27, io 7 , 111 
 age, 27, 107 
 Hilario, 47 
 Holcostephanus, 52 
 Homalonotus, 99, 100 
 Huasco, 50 
 Interglacial zones, 25 
 Ipswich series, 52, 54 
 Iratf, 78, 82 
 
 group, 16, 78-80, 84, 95, 104, 112 
 fossils, 79 
 limestones, 78 
 Paraguay, 79, 82 
 South Africa, 80, 112 
 Uruguay, 69, 78 
 Itabirite, 64 
 Itarare series, 69, 70-74 
 banded shales, 73 
 base, 68-69, 71-72, 112 
 center of glaciation, 74-75 
 fossils, 72 
 
 source of erratics, 74 
 uneven floor, 71, 73 
 Jachal, 35, 98-99, in 
 Jaguariahyva, 60, 70 
 Jujuy, 43, 56 
 Jurassic, 46, 55, 107 
 erosion, 96 
 transgression, 107 
 Kalahari, 16, 91, 114 
 Kaokoveld, 7, 15, 106, 112 
 Karroo, 15 
 basin, 66 
 system, 7-9 
 
 Argentina, 23-26, 42, 54 
 
 Brazil and Uruguay, 71, 78-80,, 
 85, 88-89, 92-93, 101-106, 
 m-113 
 
 Falklands, 12-13, 16 
 Kimberlite, 113 
 Kuibis, 6 
 
 Kundulungu, 9, 10, 96, 114 
 Kuttung series, 40-41 
 Lafonian system, 12, 102 
 Lages, 72 
 
 Lake Titicaca, 40, 56-57, 98, 100 
 Land-bridges, 1, no, 118 
 La Rioja, 42, 45, 99 
 La Ternera, 50, 55 
 Lavras series, 64-65 
 Lebombo, 52, 106 
 Leoncito Encima, 33 
 Lepidodendron, 36-37, 39, 40, 77 
 Lepidophloios, 77 
 Leptoccelia, 35, 99 
 Liassic: 
 
 Argentina, 51, 54 
 
INDEX 
 
 155 
 
 Liassic —continued 
 
 Argentina flora, 54 
 volcanics, 51, 107 
 Chile, 50 
 
 Dolerites, 15, 94, 106 
 Limestones: 
 
 Argentina, 18-19 
 Brazil, 61, 114 
 Congo, 10 
 South Africa, 6 
 Uruguay, 62-63, 67 
 Lingula, 57, 60, 72 
 Lualababeds, 9, 95 
 Lubilache beds, 9, 95, 106, 114 
 Liideritz, 6, 15, 62, 114 
 Lycopodiopsis, 77, 80, 95 
 Maldonado, 63 
 Mandiyuti conglomerate, 57 
 Maranhao, 13-14, 16, 94, 106, 114 
 Marattiopsis, 49 
 Marayes, 47, 52 
 Marechal Mallet, 81-82 
 Mariental, 7, 72, 112 
 Materal formation, 57 
 Matto Grosso, 14-15, 61, 79, 89, 99, 100 
 Mearim series, 95 
 Megalodus, 82, 150 
 Melania, 56 
 Melaphyre, 45, 5 1 , 9 2 
 Melilite basalt, 113 
 Melo, 63, 67, 76, 79, 81-82 
 Mendoza, 45, 48-49, 52, no 
 Mesosaurus, 7, 16, 79, 105, 112 
 Minas, 63 
 
 Minas Geraes, 13, 64 
 Minas series, 64-65 
 Miocene elevation, 27 
 Molteno beds, 8 
 age, 8, 54 
 flora, 8, 54 
 Monograptus, 98 
 Montevideo, 62, 84 
 Moscovian, 35, 149 
 Myophoria, 82, 150 
 Nama system, 6, 18, 62-63, 114 
 Neocomian: 
 
 Argentina, 16, 52 
 Cape, 16, 52 
 
 Neogene, 14, 16, 43, 66, 85-86 
 Neuquen, 16, 52, 106-107, no 
 Neuropteridium, 36-37, 77 
 Nevado de Famatina, 38, 44-45, 99 
 New South Wales, 34, 39, 77, 103-104 
 Nigeria, 10, 99, 114 
 Noeggerathiopsis, 37 
 Northern Carboniferous flora, 38, 40 
 Olavarria, 18, hi 
 O quita formation, 57 
 Ordn, 43 
 Orbicuioidea, 60 
 Ordovician: 
 
 Argentina, 19, 28, 31-32, 56 
 Bolivia, 97 
 Brazil, 13-14, 62-63 
 Uruguay, 62-63, 68 
 Orleans conglomerate, 72, 77 
 Ouro Preto, 64-65 
 Pachycardia, 82, 150 
 
 Paganzo system, 14, 28-54, 112-113 
 coals, 32, 38, 46-47, 53 
 description, 43-54, 58 
 distribution, 42-43, 46 
 flora, 37-41 
 folding, 31, 46 
 marine facies, 34-35 
 name, 41 
 
 overlapping, 39, 43 
 stages, 43 > 
 
 unconformity below, 28, 30, 33, 35, 
 
 38, 46 
 
 volcanics, 45-46 
 Palermo shales, 75-76 
 Palmeira, 78, 80, 83 
 
 Pampean formation, 16, 18-19, 27, 106, 
 
 in 
 
 ranges, 52-53 
 Paraguay: 
 
 Bonito, 77 
 Devonian, 60 
 Estrada Nova, 82 
 Gondwana, 60 
 Iraty, 79, 82 
 Triassic, 87 
 Paramillo, 47-48 
 Parana, 14, 59, 61, 63, 70, 83, 87 
 Basin, 66-94, io 4 
 Devonian, 15, 59, 60 
 Parnahyba, 13 
 series, 94 
 
 Passa Dois series, 78-82, 85 
 composite character, 82, 85 
 in Argentina, 85 
 mollusca, 82, 150 
 Patagonia, Cretaceous, 27, 55 
 Patagonides, 103, 107 
 Pennsylvanian, 40-41 
 Peralta, 25 
 Permian: 
 
 Argentina, 14, 25-26, 43-45 
 Bolivia, 57 
 Brazil, 13, 94-96 
 movements, 10, 18, 27, 46, 105 
 South Africa, 8, 112 
 Pernambuco, 13 
 Peru, 40-41, 100 
 
 Carboniferous, 100, 148 
 Petroleum, 50, 58, 80 
 Phacops, 61 
 Phoenicopsis, 47, 54 
 Phyllotheca, 37, 46, 57-58, 77 
 Piauhy, 13, 16, 94, 114 
 Pillahuincd beds, 19, 25-26, 31 
 Pipe amygdaloid, 48, 92, 113 
 Piracicaba, 71 
 Pirapora sandstone, 96 
 Plants, fossil. See Flora 
 Pliocene, 50, 107 
 Polypora, 34 
 
 Ponta Grossa, 59, 60, 70-72, 75 
 shales, 59, 60-61, 98, 112 
 Porto Alegre, 72, 75, 78, 83, 86-87 
 Porto Uniao, 71 
 Potosi, 56, 101 
 Potrerillos, 49, 50 
 Pre-Cambrian, 
 
 Brazil, 15, 64-65 
 
156 
 
 INDEX 
 
 Pre- Cambri an —continued 
 Cape, 6, 15 
 Southwest Africa, 6 
 Pre-Cordillera, 14, 27, 33, 41, 46 
 Productus, 34, 51, 56, 101, 104, 147-149 
 limestones, 148 
 Psaronius, 80, 94-95 
 Pseudamusium, 34 
 Pseudomonotis, 51 
 Pyramboia group, 83, 86-87 
 Quaternary, 28, 43, 117 
 Quartz prophyry, 51-52 
 Queensland, 40-41, 54 
 Radioactivity, 104 
 Red beds, 8, 85, 89, 113 
 Reptilian fossils, 8 
 Rhacopteris, 29, 36-38, 103 
 Rhaetic: 
 
 Argentina, 42, 45-55 
 Brazil, 14, 89 
 Chile, 50 
 Floras, 50, 54 
 South Africa, 8, 54 
 Rhipidopsis, 37 
 Rhodesia, 8, 88, 106 
 Rhyolite, 52, 107 
 Rinconada, 36 
 Rio Claro, 81-82, 94 
 Rio do Rasto group, 81-89, 9U JI 3 
 origin, 87 
 
 overlapping, 83-85 
 
 Rio Grande do Sul, 66, 72, 75, 81, 86, 92 
 Rio Sauce Grande conglomerate, 24 
 Rio (River): 
 
 Atuel, 27, 51, 54 
 Choapa, 51 
 Chubut, 51 
 de Jejenes, 31, 36-37 
 de la Mina, 28, 31, 38 
 Grande, 28-32, 36-37 
 la Plata, 15 
 
 Negro (Argentina), 108 
 Negro (Brazil), 70-72 
 Negro (Uruguay), 60 
 Neuquen, 51 
 Niger, 11, 114 
 Parnahyba, 94-96 
 Sao Francisco, 13, 15, 63 
 Sauce Grande, 19, 20, 22-23, 2 7 > 102 
 Rivera, 87, 89 
 
 Roches moutonn£es, 7, 30, 69, 74 
 Rocinha limestone, 80, 83 
 Sagenopteris, 83 
 Sahara, 11, 16, 57, 99, 100 
 Saharides, 10 
 Saldan, 53 
 Salta, 55, 58, 99 
 Salt Range, 104, 148 
 Samaropsis, 78 
 San Cristobal, 76, 85, 87, 92 
 San Jorge formation, 16, 108 
 San Juan, 28, 41-43, 101, 103-104, 110-11 
 Sanogasta, 44 
 San Rafael, 27, 32 
 Sao Bento series, 85-93 
 age, 89 
 area, 89 
 fossils, 89 
 
 Sao Bento series —continued 
 origin, 88 
 
 Sao Jeronymo, 72, 75 
 
 Sao Paulo, 63, 71, 74, 76, 78-79, 81, 83, 
 94, 105, 108, 114 
 
 Santa Catherina, 71, 80, 83, 86-89, 93 
 system, 14, 42, 66-94 
 
 Argentina, 76, 85, 92-93 
 classification, 67 
 overlapping, 72-73, 83 
 Santa Cruz: 
 
 Bolivia, 55, 57 
 Patagonia, 51, 55 
 Santa Maria, 87, 89, 91, 93 
 Scaphonyx, 89 
 Schizodus, 79 
 Schizoneura, 46, 58 
 Semionotus, 48-49, 54, 95 
 Serra Geral eruptives, 84, 89-93 
 age, 93 
 petrology, 93 
 thickness, 92 
 Serra Grande series, 94 
 Serrinha, 59, 70 
 
 Sierra Chica de Zonda, 28-32, 35-36, 38-39 
 de Cerro Blanco, 46-47 
 de Cordoba, 43, 46, 53 
 de la Huerta, 52 
 del Alto, 55-56 
 
 de la Ventana, 19-27, 99, 102, ill 
 folding, 20-21, 25, 102, hi 
 shale band, 21 
 quartzites, 21, 62, 102, 111 
 tillite, 22-25, 102, hi 
 de los Llanos, 37, 39, 44 
 de Mandiyuti, 55, 57 
 de Paganzo, 46 
 
 de Pillahuinco, 19, 20, 23-25, 102 
 
 de Sanogasta, 43 
 
 de Tandil, 18, 63 
 
 de Umango, 35, 43 
 
 deVelazco, 37, 39 
 
 de Vilgo, 43 
 
 Leone, 11, 15 
 
 Pintada, 22, 27, 32, 51, 105, 107 
 Valle Fertil, 36, 43, 46 
 Villa Unidn, 35, 43, 46 
 Sigillaria, 32, 77, 95 
 Silurian, Amazon, 13, 15, 115 
 Argentina, 18, 28, 97-98 
 Bolivia, 100 
 ocean, 97-99 
 
 Western Africa, 11, 15, 115 
 Solenomorpha, 81, 150 
 South Africa compared with South 
 America, 15-16, 19, iio-m 
 South America, geology, 13-14 
 South Atlantic, 108 
 Southern Flora. See Glossopteris flora 
 Southwest Africa, 6, 7, 79, 93, 102, 104, 
 112-114 
 
 Sphenophyllum, 77 
 Sphenopteris, 36, 40-41 
 Spirifer, 34-35, 56, 60, 63, 99, 101, 104, 
 1 47-149 
 
 Spiriferina, 34, 56 
 Stenopteris, 54 
 Stephanian, 39 
 
INDEX 
 
 157 
 
 Stereosternum, 79, 80, 95 
 Stormberg series, 8, 42, 88, 113 
 age, 8, 54 
 divisions, 8 
 transgression, 8, 113 
 volcanics, 8, 52, 106 
 Stromatopora, 19 
 Sucre, 98 
 
 Table Mountain sandstone, 6, 12, 20, 
 in-112 
 age, 98 
 
 Argentina, 21, 111 
 Bolivia, 98 
 Falklands, 12 
 glacials, 6 
 
 Tacuaremb6, 74, 76, 79, 81, 84, 87, 89, 90 
 Tasniopteris, 47-48, 50, 54, 82 
 Taquara, 86-87, 90-92 
 Tatuhy group, 76 
 Tertiaries: 
 
 Argentina, 28, 52, 108, ill 
 Brazil, 13, 95, 108 
 South Africa, 9, hi 
 W estern Africa, 10 
 Tertiary foldings, 28, 107-108, 120 
 high-level gravels 19,26-27,107 
 Thinnfeldia, 8, 46-48 
 
 flora, 16, 45, 53, 106, 113, 119 
 Tibagy, 59, 60, 70, 75 
 
 sandstone, 22, 60, 112 
 Tillite: 
 
 Argentina, 14-15, 19, 22-25, 3 °~ 3 U 
 33-36, 44 
 Bolivia, 58 
 Brazil, 70-71, 74 
 South Africa, 7 
 Uruguay, 68, 73-74 
 See Glacials 
 Tinogasta, 44 
 Togoland, 10, 16 
 Tombador series, 64 
 Tontal series, 33 
 Tornquist, 20 
 Torres, 13, 66, 87, 89 
 Tostado, 76, 92 
 Trapiche, 37 
 Traps, 86, 89 
 Trias-Rhaetic: 
 
 Argentina, 47-56, 5 8 
 
 Trias-Rhaetic —continued 
 
 Brazil, 13-14, 61, 67, 80-93 
 Chile, 50 
 Patagonia, 51, 55 
 South Africa, 8, no 
 Uruguay, 80-91 
 Triassic: 
 
 Argentina, 45-46, 92 
 Bolivia, 56 
 
 Brazil, 13-14, 61, 67, 80-91, 95 
 climate, 87-91, 106, 113, 119 
 Congo, 9 
 
 Marine, 45, 50, 81-82, 150 
 tuffs, 51-52 
 
 unconformity below, 8, 14-15, 27, 32, 
 42, 46-47, 50, 67, 74, 83-85, 105, 
 no 
 
 volcanics, 27, 51-52, 85, 94, 119 
 Trigonia, 52 
 Tropidoleptus, 60 
 Tuberao series, 75-77, 84 
 Tucumdn, 43, 55 
 Uitenhage, 52, no 
 fauna, 16, 52 
 
 Argentina, 16, 52, in 
 Uralian, 35, 40, 56 
 Uruguay: 
 
 Devonian, 60, 62 
 Gondwana: 
 coals, 76 
 
 glacials, 67-69, 73-74 
 volcanics, 84, 90-93 
 Ordovician, 62-63, 68 
 Uspallata, 32, 49, 51 
 Volcanics: 
 
 Liassic, 51, 107 
 Paganzo, 45-46 
 Serra Geral, 89-93 
 
 Argentina, 85, 92-93 
 Stormberg, 8, 93-94 
 Triassic, 48, 51-52, 85, 90-94, 119 
 Wegener’s Hypothesis. See Displacement 
 Hypothesis 
 Western Africa, 10-11 
 Westphalian, 39, 40 
 White band, 7, 78, 80, 112 
 Witteberg series, 7, 12, 102 
 Argentina, 20, in 
 Falklands, 12 
 
PLATES 
 
 I. Folded Devonian quartzites, Sierra de la Ventana. 
 
 II. A. Shale band in Devonian quartzites, Sierra de la Ventana; B. Cleavage in til- 
 lite, Sierra de la Ventana. 
 
 III. A. Tertiary “high-level gravels,’' Sierra de la Ventana; B. Lower glacial zone, 
 
 Sierra Chica de Zonda, San Juan. 
 
 IV. A. Glaciated surface beneath lowest glacial zone, San Juan; B. Lowest glacial 
 
 zone, Sierra Chica de Zonda, San Juan. 
 
 V. A . Second glacial zone, Sierra Chica de Zonda, San Juan; B. Second glacial zone, 
 Sierra Chica de Zonda, San Juan. 
 
 VI. Second glacial zone, Sierra Chica de Zonda, San Juan. 
 
 VII. A. Second glacial zone, Sierra Chica de Zonda, San Juan; B. Dome in Paganzo 
 beds, Carpinterla, San Juan. 
 
 VIII. A. Dome in Carboniferous beds, Barreal, San Juan; B. Paganzo beds, Bajo de 
 Velis, San Luiz. 
 
 IX. Triassic beds, Barreal, San Juan. 
 
 X. A. Tree-trunks in Triassic beds, Paramillo de Uspallata; B. Triassic beds, 
 Marayes, San Juan. 
 
 XI. A. Triassic conglomerates, Saldan, Cordoba; B. Triassic conglomerates, Capilla 
 del Monte, Cordoba. 
 
 XII. A. Devonian Furnas sandstone, Ponta Grossa, Parana; B. Rio do Rasto sand¬ 
 
 stones, Taquara, Porto Alegre. 
 
 XIII. Carboniferous fossils, Barreal, Argentina. 
 
 XIV. Carboniferous fossils, Barreal, Argentina. 
 
 XV. Carboniferous fossils, Barreal, Argentina. 
 
 XVI. Carboniferous fossils, Barreal, Argentina. 
 
 m l!BRMtf M THt 
 
 27 1928 
 
 UNlVttlSMY 01 tU'N°* s 
 
 158 
 

 N 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE I 
 
 Folded Devonian quartzites north of the Cerro Naposta Grande, looking east 
 
 From a photograph by Dr. R. Beder 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE II 
 
 A—Shale band in quartzites of the Sierra de la 
 Ventana gorge below the Tres Picos 
 
 B—Cleavage in tillite, railway cutting a little northeast of Sierra de la Ventana station 
 

 
 
 
 
 
 
 THE LIBRARY 
 
 OF THE 
 
 UHIVERSITY OF ILLINOIS 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE III 
 
 A—Tertiary “ high-level 
 
 gravels" on folded Devonian beds, west side of the Arroyo San 
 Bernardino, Sierra de la Ventana 
 
 B—Lower glacial zone, immediately to the north of the Rio Grande, Sierra Chica de Zonda, 
 
 San Juan 
 

Carnegie Inst. Wash. Pub. 381 
 
 PLATE IV 
 
 A—Glaciated surface of Ordovician limestone beneath lowest glacial zone, a little north of 
 the Rio Grande, Sierra Chica de Zonda, San Juan; the hammer indicates the direction 
 of the glacial striae 
 
 B—The lowest glacial zone, resting on tilted Palaeozoics, followed by plant-bearing shales 
 crowned with sandstones, a little north of the Rio Grande, Sierra Chica de Zonda, 
 San Juan 
 
THE LIBRARY 
 
 OF THE 
 
 HMIVERSITY OF ILL! 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE V 
 
 A—Second glacial zone, a little north of the Rio Grande, Sierra Chica de Zonda, San Juan. 
 The dark tillite forms the central peak and is sandwiched between well-stratified beds; 
 the third glacial zone forms the point on the extreme left 
 
 B—Second glacial zone, Arroyo de Jejenes, San Juan. Below are sandstones, followed by 
 the glacials with lens of sandy material and overlain by sandstones which form the 
 summit of the cliff 
 
THE LIBRARY 
 
 OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
The second glacial zone, Arroyo de Jejenes, San 
 Juan, showing lenticular band of sandy material 
 
 Carnegie Inst. Wash. Pub. 381 
 
 PLATE VI 
 

 THE LIBRARY 
 OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
 
 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE VII 
 
 A—The second glacial zone, Arroyo de Jejenes, San Juan, showing practically 
 
 unstratified tillite 
 
 B—Dome in Paganzo beds, followed by tilted Calchaquenos and overlain by horizontal 
 
 Quaternary gravels; west of Carpinteria, San Juan 
 
THE LIBRARY 
 OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE VIII 
 
 A—Dome in Carboniferous beds near Quebrada del Salto, Barreal, San Juan, showing in 
 foreground boulder shales and fossiliferous shales, followed by sandstones forming 
 the hills in the background 
 
 B—Plant-bearing Paganzo beds, Bajo de Velis, Sierra de San Luiz. The range in the back¬ 
 ground is made of crystalline rocks 
 

 > • 
 
 
 
 
 
 THE LIBnARY 
 
 OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
 
 
 
 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE IX 
 
 A—Triassic plant-bearing strata of Stage IV, Quebrada del Jarillal, Barreal; 
 
 on extreme right beds of Stage V 
 
 B—Triassic beds of Stage IV, overlain by the red conglomeratic strata of Stage V 
 
 Quebrada del Salto, Barreal, San Juan 
 

 
 THE LIBRARY 
 
 OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE X 
 
 A—Tree trunks standing erect in Triassic volcanic ash, Paramillo de Uspallata 
 
 B—Tilted sandstones and conglomerates of the coal-bearing section of Stage IV, Carrizal 
 
 near Marayes, Sierra de la Huerta, San Juan 
 
THE LIBRAUY 
 OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE XI 
 
 B—Triassic conglomeratic sandstones, Los Terrones, near Capilla del Monte 
 
 Sierra de Cordoba 
 

 THE LIBfMRy 
 OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
 
 
 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE XII 
 
 A—Devonian Furnas sandstone, showing current-bedding. East of Ponta Grossa, Parana 
 
 B—Rio do Rasto " millet-seed " sandstones, municipal quarry, Taquara, near Porto Alegre. 
 The marked dip from left to right is due to false bedding, the formation being actually 
 horizontal 
 
THE LIBRARY 
 
 OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE XIII 
 
 Carboniferous Fossils, Barreal, Argentina 
 
 (i) Orbiculoidea saltensis sp. nov. ( 7335 ), upper valve. X 1.5. (2) Productus ( Marginifera ) 
 
 spinulo-costatus Abich var. nov. peregrina (7374), pedicle-valve. X 1.25. (3) Spiriferina zewanensis 
 Diener (7312), pedicle-valve. X 1.5. (4a, 46) Productus curvirostris Schellwien var. nov. barrealensis 
 (7315), pedicle-valve, a, top view; b, front view. X 2. (5) Hemiptychina cf. sublcevis Waagen (7300), 
 
 pedicle-valve. X 3 - ( 6 ) Choneles granulifer Owen (7319), pedicle-valve (shell partly wanting). X 1.5. 
 
 (7) Nuculana (Leda ) cf. bellistriata Stevens (7339), internal cast. X 2. (8a, 8 b) Reticularia notica sp. nov. 
 
 (7326); a, pedicle-valve. X 1.5; b, portion of shell. X 8. (9) Productus lineatus Waagen (7365), side view, 
 
 pedicle-valve. X 2. (10) Reticularia notica sp. nov. (7445), pedicle-valve. X 1.5. (11) Pseudamusium 
 
 stappenbecki sp. nov. (7291), left valve. X 1.5. (12) Pseudamusium cf. elliplicum (Phillips) (7304), left 
 
 valve. X 1.5. (13) Pseudamusium stappenbecki sp. nov. (7290), right valve. X 1.5. (14) Aviculopecten 
 
 barrealensis sp. nov. (7362), left valve. X 4. 
 
THE LIBRARY 
 OF THE 
 
 UNIVERSITY' OF ILLINOIS 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE XIV 
 
 7b 8b 
 
 Carboniferous Fossils, Barreal, Argentina 
 
 (i) Spirifer of. supramosquensis Nikitin (7316), pedicle-valve. X 1.25. (2) Spirifer barrealensis 
 
 sp. nov. (7371), pedicle-valve. X 1.5. (3) Spirifer mexicanus Shumard var. nov. neolropica (7366), pedicle- 
 valve. X 1.25. (4) Spirifer mexicanus Shumard var. nov. neotropica (7375), pedicle-valve. X 1.25. (5) 
 
 Spirifer aff. rajah Salter (7317), pedicle-valve (anterior part bent underneath). X 1.5. (6) Spirifer 
 
 ( Martinia ) cf. simensis Tschernyschew (7363), pedicle-valve. X 1.5. (7 a, 7b) Spirifer Wynnei Waagen var. 
 nov. argenlina (7309), a brachial valve, X 1.25; b front view. X 1.25. (8a, 8 b) Spirifer saltensis sp. nov. 
 
 (7296); a, pedicle-valve. X 1.5; b, brachial valve. X i.S- 
 

 
 
 
 THE LIBRARY 
 OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
 
 
 
 
 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE XV 
 
 3 c 
 
 Carboniferous Fossils, Barreal, Argentina 
 
 (i) Euomphalus subcircularis Mansuy (7328), base. X 1. (2) Euomphalus subcircularis (7344) 
 
 apical surface. X 1.25. (3a, 3 b, 3c) Pleurotomaria argentina sp. nov. (7334); a, b, side-views. X 1.5; c, por¬ 
 tion of base. X 6. (4) Euomphalus subcircularis (7322), apical surface. X 1.25. (5) Pleurotomaria 
 
 advena sp. nov. (7355)- X i-5- 
 
THE LIBRARY 
 OF THE 
 
 UNIVERSITY OF ILUttOIS 
 
 
Carnegie Inst. Wash. Pub. 381 
 
 PLATE XVI 
 
 Carboniferous Fossils, Barreal, Argentina 
 
 Pleurotomaria barrealensis sp. nov. (7323); a, side view. X 1.5; b, base, X 1.5. 
 

 \ 
 
 
 
 
 
 THE LIBRARY 
 OF THE 
 
 UNIVERSITY OF ILLINOIS 
 
 
 
 

 
 
 
 
 
 
 
CARNEGIE INSTITUTION OF WASHINGTON 
 
 PUB. No. 381