FOSSIL MAMMALS FROM THE BEGINNING OF THE CENOZOIC IN BRAZIL FOSSIL MAMMALS FROM THE BEGINNING OF THE CENOZOIC IN BRAZIL 0 CONDYLARTHRA, LITOPTERNA, XENUNGULATA, AND ASTRAPOTHERIA CARLOS DE PAULA COUTO Museu Nacional, Rio de Janeiro BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY VOLUME 99 : ARTICLE 6 NEW YORK : 1952 BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY Volume 99, article 6, pages 355-394, text figures 1-3, plates 32-43, tables 1-18 Issued August 8, 1952 Price: $1.00 a copy INTRODUCTION THIS PAPER IS THE FOURTH preliminary re- miles northeast of Niteroi, the capital of the port on the fossil mammals from the Paleocene state, and on approximately the same latitude of Sio Jose de Itaborai, State of Rio de as the city of Rio de Janeiro. This quarry is Janeiro, Brazil. The final report, which will under the management of the Companhia include detailed descriptions of all the mam- Nacional de Cimento Portland (Maui), malian groups represented in Sio Jos6 de which has given generous cooperation in the Itaborai, will be published by the Museu work of collecting, and to which I present Nacional, Rio de Janeiro, in collaboration my thanks for this help. FIG. 1. Map of part of the State of Rio de Janeiro, Brazil, showing the locality of the limestone quarry of Sao Jose de Itaborai (circle near the center). The latitudes are referred to the meridian of the city of Rio de Janeiro. Adapted from Price and Paula Couto, 1950. with the Divisao de Geologia e Mineralogia A note and bibliography on the geology of do Departamento Nacional da Produglo this deposit were given by the author in an Mineral. Completion of the final report will earlier paper (1949). take considerable time, and for this reason I The fossil vertebrates came from channels have decided to present a series ofpreliminary and underground caves produced in the reports on the results of my investigations of limestone by pluvial and underground water. the different groups involved. As they are definitely more primitive than All the material here described has been the corresponding groups in the Casamayoran collected in a limestone quarry at SAo Jos6 (Notostylops beds-lower Eocene) of Pata- de Itaborai, State of Rio de Janeiro, some gonia, and as one of the genera here consid- 359 360 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY VOL. 99 ered (Carodnia) is known elsewhere only in Mrs. Rachel Husband Nichols in the Osborn the Rio Chican (upper Paleocene) of Pata- Library, from the staff of the Library, and gonia, it seems to me that they must be from the Division of Photography during the considered as of Paleocene age, synchronous preparation of this paper. The photographs with the Rio Chican. reproduced as plates 32 to 37 and 40 to 43 The limestone is perhaps older than the were especially taken for this work by the fossiliferous marl which fills the channels Division of Photography. and underground caves. At any event it is Messrs. Carl Sorensen and Walter Soren- not more recent than the Rio Chican. Its sen, and William Fish, of the Laboratory of age, and consequently the age of its inverte- Vertebrate Paleontology of the American brate fauna as described by Maury (1935), Museum are deserving of special acknowledg- is possibly lower Paleocene or upper Cre- ments for the careful preparation of, respec- taceous, at least in part (some of the verte- tively, the material of Carodnia vieirai, new brate specimens were found in the white and species, and of the complete mandible of Tri- gray limestone). gonostylops apthomasi here described. These studies were greatly aided by the To Dr. George Gaylord Simpson I am opportunity of direct comparison between deeply indebted for valuable suggestions and the Paleocene remains from Sao Jos6 de constructive criticism during the preparation Itaborai and the specimens in the rich collec- of this paper and for the loan of his unpub- tion of the Patagonian and North American lished manuscript on the Patagonian lower lower Tertiary groups in the American Eocene astrapotheres, which has given impor- Museum of Natural History. This was made tant information about the Casamayoran possible by the generous support given me forms and the whole group in general. My by the John Simon Guggenheim Memorial opinion on the Patagonian species of Trigo- Foundation, in New York, to which I offer nostylops is based mostly on Simpson's direct my best acknowledgments. I also wish to observations on the type specimens in the thank the General Secretary of that founda- Argentinean museums, but in part on my own tion, Mr. Henry Allen Moe, for the kind direct observations on the same specimens. attention and courtesy which I have invari- In an earlier paper (Price and Paula Couto, ably received from him. 1950) I selected T. wortmani as the best-de- To Dr. Matias G. de Oliveira Roxo, former fined species of this genus. director, and Dr. Alberto Ribeiro Lamego, The following abbreviations are used to present director, of the Divisao de Geologia e designate the locations of the specimens: Mineralogia do Departamento Nacional da Produgio Mineral, Rio de Janeiro, I am A.M.N.H., the American Museum of Natural to study the History very grateful for permission D.G.M., Divisao de Geologia e Mineralogia do important material of the collection of that Departamento Nacional da Produgao Mineral institution. (formerly Servigo Geol6gico do Brasil), To the American Museum of Natural His- Brazil tory I am grateful for the solicitous assist- M.A.C.N., Museo Argentino de Ciencias Nat- ance that I have received from its Depart- urales "Bernardino Rivadavia," Buenos Aires ment of Geology and Paleontology, from M.N.R.J., Museu Nacional do Rio de Janeiro SYSTEMATICS THIS STUDY is concerned with the Condy- Eocene (Casamayor) of that region. Two new larthra, Litopterna, Xenungulata, new order, species, closely related to their Patagonian and Astrapotheria. congeneric species, are present in the Paleo- The most surprising result of the study of cene of S-ao Jos6 de Itaborai. the condylarths of Sao Jos6 de Itaborai was The known Patagonian remains are merely the revelation that the genus Asmithwood- isolated upper and lower molars. Among the wardia Ameghino is not a member of the specimens that were collected in Sao Jose de Didolodontidae (the condylarth family re- Itaborai are an almost complete left upper stricted to South America) but a legitimate jaw, with p8-4 and M'-8 present, and a median representative of the North American Hyop- fragment of a left lower jaw, with M1_3 sodontidae in the South American Paleocene, present, as well as several upper and lower a discovery made possible because the new isolated premolars and molars. material of this genus is the most complete These more complete remains enable me to yet found. add to the diagnosis of this genus, which was This fact is not really surprising, since the given by Simpson (1948, p. 110) on the basis idea that the South American Cenozoic of the material collected in Patagonia. mammals arose, at least for the most part, DIAGNOSIS: Upper molars wider than long; from North American primitive groups that no mesostyle; accessory cusps poorly de- immigrated to South America during or veloped; continuous external cingulum, some- immediately after Cretaceous times is gener- times interrupted at the base of the paracone; ally accepted by most of the specialists, large hypocone in M'-2, not present in M3. Pa including the author. On the contrary, it is longer than wide, triangular, similar to pl-2 strong evidence in favor of this opinion. in Didolodus. P4 as in Didolodus. Lower Another important result of this work was molars usually without cuspule on hypo- the discovery that the genus Carodnia conid-metaconid crest. Lower C strong, Simpson, first described from a single tooth slightly procumbent, with sharp, pointed, from the Patagonian Paleocene, does not and turned-up tip. PI small, compressed belong in any of the known mammalian anteroposteriorly, and single-rooted. P2.8 orders. It is a very strange and aberrant form, simple, two-rooted. P4 molariform, with high here considered as the first known representa- trigonid, low and short talonid, strong proto- tive of a new order (Xenungulata). conid and metaconid, and small paraconid; The Order Astrapotheria is represented in low and more or less equally developed Sao Jose de Itaborai by a species of the hypoconid and entoconid. Teeth generally of genus Trigonostylops Ameghino, a species simple construction, with smooth enamel which in its general characters appears to be and bunoid cusps. more primitive than the species of Patagonia DIscussIoN: Ernestokokenia is very simi- (Casamayor, lower Eocene). It also fur- lar to Didolodus, but differs from it in p8-4, nished almost certain evidence of the descent which are very different in shape, in the of this genus from the most primitive complete absence of the mesostyle in its condylarthran stock in South America. This upper molars, and in the lack of a cuspule in may also explain the slight resemblance the hypoconid-metaconid crest. between this group and the Litopterna. Ernestokokenia protocenica, new species ORDER CONDYLARTHRA COPE, 1881 Plate 32, figures 1, 2 FAMILY DIDOLODONTIDAE SCOTT, 1913 HOLOTYPE: M.N.R.J. No. 1430-V. Frag- ERNESTOKOKENIA AmEGHINO, 1901 ment of left maxilla with P3-4 and M1-8 This genus was established by Ameghino present. Collected by Carlos de Paula Couto, for species based on remains collected in the 1948. lower Tertiary formations of Patagonia. PARATYPE: M.N.R.J. No. 1431-V. Part of The species now known are from the left lower jaw with M1.. present.