Novitates PUBLISHED by the AMERICAN MUSEUM of NATURAL HISTORY CENTRAL PARK WEST at 79TH STREET, NEW YORK, N.Y

AMERICAN MUSEUM Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2820, pp. 1-26, figs. 1-10, tables 1-9 June 26, 1985

Biostratigraphy of the Casamayoran, Early Eocene, of Patagonia

RICHARD L. CIFELLII

ABSTRACT The Casamayoran, presumed to pertain to the leontologic evidence suggests that these differences early Eocene, is the most diverse and best known are temporal and not ecological in nature. Faunal of South America's early to middle Tertiary land composition of the two localities is compared mammal ages. Because the composition of local qualitatively and statistically with respect to the faunas and the stratigraphic position of assem- magnitude ofdifference seen in preceding and suc- blages at any given locality have not been evalu- ceeding South American land mammal ages and ated, however, the Casamyoran has remained a to roughly equivalent North American land mam- dimensionless age, despite its seminal importance mal ages. Two new subdivisions of the Casama- to understanding the early radiations ofmammals yoran, the Barrancan and the Vacan, are defined on the South American continent. on this basis. In addition, a composite stratigraph- Preliminary analysis offield data and collections ic plot oftaxa from the younger ofthe two localities obtained by G. G. Simpson in 1930-1931 suggests (the Gran Barranca) produced a number of non- a faunal succession in the Casamayoran of Pata- overlapping range zones within the Barrancan; sta- gonia. Large, well documented samples are avail- tistical analysis indicates that the probability these able from two localities with relatively wide strati- observations are due to sampling error is small in graphic distributions of Casamayoran mammals: most, but not all, cases. The other Patagonian local Caniadon Vaca and the Gran Barranca south of faunas are briefly considered with respect to this Lago Colhue Huapi, central Chubut. The com- scheme, and an hypothesis oftheir relative ages is positions ofthese two local faunas differ markedly, presented. and lithostratigraphic, biostratigraphic, and pa-

INTRODUCTION It is now widely accepted that South Amer- Tertiary-insular, or nearly so, with respect ica was an island continent for most of the to its biota. The Tertiary record of South

1 Curator of Paleontology, Department ofGeology, Museum ofNorthern Arizona, Flagstaff, Arizona 86001; Post- doctoral Fellow, Division of Mammals, National Museum ofNatural History, Smithsonian Institution, Washington, D.C. 20560.

Copyright © American Museum of Natural History 1985 ISSN 0003-0082 / Price $1.70 America's mammalian faunas is fairly well local faunas or to the superposition of assem- known and, as a consequence, provides an blages from given localities, the Casa- ideal "natural experiment" for the study of mayoran has remained an age without tem- historical biogeography and evolutionary poral dimension. principles in general. Of particular interest The collections made in central Patagonia are the earliest faunas, those from the begin- by the Scamtt Patagonian Expeditions (1 930- ning of the Age of Mammals, because they 1931, 1933-1934), under the direction of are critical to understanding the phylogeny, George Gaylord Simpson, are of paramount adaptive radiations, and biogeographic de- importance in this regard. Today, more than ployment of mammals on the South Amer- 50 years later, they include the largest col- ican continent. Of these early assemblages, lections of Casamayoran mammals and are the Casamayoran is by far the most diverse the only ones for which precise stratigraphic and well known. provenience data are available. In the present The great Argentine paleontologist, Flo- paper I present the stratigraphic data for large, rentino Ameghino, who first described the well documented samples from two localities "Notostylops" (Casamayoran) fauna, be- with relatively wide stratigraphic distribu- lieved it to include dinosaurs and thus con- tions of Casamayoran mammals and consid- sidered it to be Late Cretaceous in age. This er the composition of other Patagonian local view has long since been shown to be incor- faunas of Casamayoran age. rect (Simpson, 1932) but, because the mammalian fauna is so dissimilar to those from anywhere else in the world, the age of the Casamayoran is not well established in either I am deeply grateful to the late Dr. George Gaylord Simpson for his advice and for per- relative or radiometric terms. Simpson (1 940) considered the age of Patagonian marine for- mission to use his fieldnotes of the Scanitt Patagonian Expeditions. All original field ob- mations and placed the Casamayoran in the early Eo~ene.~Welcome new evidence, in the servations and descriptions presented herein form of radiometric dates and a magnetic were made by him and are directly abstracted polarity sequence, is now at hand for part of from these notes, now deposited in the ar- the underlying Rio Chico Formation, fixing chives of the Department of Vertebrate Pa- its age as 6 1 to 56 million years (Marshall et leontology, American Museum of Natural al., 198 1). Such data are not yet available for History. This paper is therefore largely the rocks enclosing a Casamayoran fauna and the results of Simpson's efforts, although I of Riochican-Casamayoran boundary is placed, course assume full responsibility for any mis- interpretations I may have introduced. I ded- by convention, at the Paleocene boundary (Marshall, 1982). icate this paper to his memory. I thank Dr. Despite nearly a century of effort, South Malcolm C. McKenna of the Department of America's early land mammal history is so Vertebrate Paleontology for access to the collections and notes of the Scamtt Expeditions. inadequately documented that only very gen- Helpful review comments on an earlier draft eral observations can be made. A critical missing element is a precise chronology, either of this paper were provided by Drs. Mc- Kenna, L. G. Marshall, D. Savage, and R. H. absolute or relative. The stratigraphic relationships of the faunas have not yet been Tedford, and their advice is warmly appre- ciated. Partial support for this research was evaluated and, because little attention has been provided by the Undergraduate-Graduate paid to variation between the composition of Research Program administered by the By suggesting that the arctostylopid Palaeostylops, American Museum of Natural History and of the Gashato fauna, Mongolia, was Paleocene in age supported by the Greenwall Foundation. and ancestral to both the Wasatchian Arctostylops and the notoungulatesof South America, Matthew and Gran- HISTORICAL AND GEOLOGICAL ger (1925) had already implied that the earliest Nearctic BACKGROUND forms (i.e., those of the Casamayor fauna; the Riochican was not then known) were of latest Paleocene or early The first mammals to be described from Eocene age. what is now known as the Casamayoran were 1985 CIFELLI: THE CASAMAYORAN 3

FIG. 1. Central Patagonia index map. Localities: 1, Gran Barranca south of Lago Colhue Huapi (black rectangular area; shown in greater detail on fig. 2); 2, Caniadon Vaca; 3, Caniado Hondo; 4, Caiiadon Lobo. Tapera de L6pez, not on the map, is approximately 30 km northeast ofLaguna Palacios; the exact location of several other (Ameghino) localities is uncertain. collected by Carlos Ameghino at the Gran three or even four distinct faunas were in- Barranca south of Lago Colhue Huapi, cen- cluded. The French collector Andre Tour- tral Chubut, in 1895-1896, and were pub- nouer, guided by Carlos Ameghino, made a lished by Florentino Ameghino (1897) in his small collection of Casamayoran mammals second work devoted to the Pyrotherium (De- at Cafnadon Lobo, near Punta Casamayor, seadan) fauna. Carlos later noted (Ameghino, and "Casamayoran" has since been adopted 1913-1936, vol. 21, p. 105) that the fossils as the geographic term for Ameghino's Noto- in question pertained to a much earlier, pre- stylops fauna (Simpson, 1933). Pyrotherium, fauna and, following Carlos' In addition to the Gran Barranca and suggestion, Florentino named this the No- Cafnadon Lobo, a number of other localities tostylops fauna after that characteristic and in Patagonia have yielded Casamayoran abundant genus (Ameghino, 1899). Floren- mammals; these have been reviewed by tino Ameghino subsequently (1902) distin- Simpson (1948, 1967a) and Marshall et al. guished two subdivisions of the Notostylops (1983). Some of the localities under consid- fauna and later (1903) added a "basal" part. eration here are shown in figure 1; these and In his great treatise of 1906, Ameghino aban- others for which data are available are in- doned these somewhat hypothetical subdi- cluded in the Casamayoran local fauna lists visions but noted that a large span of time of table 1. was probably represented and that perhaps Mammalian fossils ofCasamayoran age in 4 AMERICAN MUSEUM NOVITATES NO. 2820

TABLE 1 Composition of Casamayoran Local Faunasa

Localitiesb Species 1 2 3 4 5 6 7 8 9 10 11 12 Coona pattersoni Simpson, 1938 - - X ------Coona gaudryi Simpson, 1964 X ------Caroloameghinia mater Ameghino, 1901 X C. tenuis Amegbino, 1901 ...... X Arminiheringia auceta Ameghino, 1902 ?X - - ?X Patene coluapiensis Simpson, 1935 X Polydolops thomasi Ameghino, 1897 X X .X- - P. serra Ameghino, 1902 X P. princeps (Ameghino, 1902) X P. primulus (Ameghino, 1902) X P. borcurhor Simpson, 1948 - X Amphidolops serrula Ameghino, 1902 X - - - X Eudolops tetragonus Ameghino, 1897 X - - - X E. caroliameghinoi (Ameghino, 1903) - - X UtaetusbuccatusAmeghino, 1902 X - - - - ?X U. lenis (Ameghino, 1902) X U. deustus Ameghino, 1902 X "Pseudostegotherium" chubutanum Ameghino, 1902 X Prostegotherium notostylopianum Ameghino, 1902 X Adiantoides magnus Cifelli and Soria, 1983 X Sparnotheriodon epsilonoides Soria, 1980 X Didolodus multicuspis Ameghino, 1897 X D. latigonus (Ameghino, 1902) X D. minor Simpson, 1948 X Paulogervaisia inusta Ameghino, 1901 X P. porca (Ameghino, 1901) X P. mamma (Ameghino, 1901) X Proectocion argentinus Ameghino, 1904 X P. precisus Ameghino, 1904 X Enneoconus parvidens Ameghino, 1901 X Asmithwoodwardia subtrigona Ameghino, 1901 X Ernestokokenia nitida Ameghino, 1901 X E. patagonica (Ameghino, 1901) - - X Victorlemoinea labyrinthica Ameghino, 1901 - - - X V. emarginata Ameghino, 1901 - - - X V. sp. indet. - X Ernestohaeckelia acutidens Ameghino, 1901 . ..X- Anisolambdafissidens Ameghino, 1901 X X A. adunca (Ameghino, 1901) X A. amel Simpson, 1948 - X Guilielmofloweria plicata Ameghino, 1901 X Henricosbornia lophodontaAmeghino, 1901 ?X X - - X - - H. ampla (Ameghino, 1904) - - - X - - Othnielmarshia lacunifera Ameghino, 1901 - X - - X - - Peripantostylops minutus (Ameghino, 1901) - X - - X - - Notostylops murinus Ameghino, 1897 X - -X - - - N. pendens (Ameghino, 1901) - X - - - - X - - - - ?X N. appressus (Ameghino, 1902) - X N. pigafettai Simpson, 1948 - - - X - - Homalostylops parvus (Ameghino, 1897) X Homalostylops sp. nov.? - X Edvardotrouessartia sola Ameghino, 1901 - X . . . .X 1985 CIFELLI: THE CASAMAYORAN 5

TABLE 1-(Continued) Localitiesb Species 1 2 3 4 5 6 7 8 9 10 11 12 Maxschlosseria praeterita Ameghino, 1901 x M. minima (Ameghino, 1897) x M. rusticula (Ameghino, 1901) x M. consumata (Ameghino, 1901) - X - x Oldfieldthomasia debilitata (Ameghino, 1901) x x 0. parvidens Ameghino, 1901 x 0. sp. indet. x - - Ultrapithecus rutilans Ameghino, 1901 Acoelodus oppositus Ameghino, 1897 x Ameghino, 1902 x "A." proclivus x Paginula parca Ameghino, 1901 x x Archaeopithecus rogeri Ameghino, 1897 x Acropithecus rigidus (Ameghino, 1901) x- X- ?X- x _X Notopithecus adapinus Ameghino, 1897 - x x x Antepithecus brachystephanus Ameghino, 1901 x - - Transpithecus obtentus Ameghino, 1901 Eohyrax rusticus Ameghino, 1901 x - - x x - - E. isotemnoides Ameghino, 1904 xr E. praerusticus Ameghino, 1902 x - - Pleurostylodon modicus Ameghino, 1897 x -X x x P. similis Ameghino, 1901 ?P. recticrista (Ameghino, 1904) x - - x Anisotemnus distentus (Ameghino, 1901) x x x x Acoelohyrax complicatissimus (Ameghino, 1904) x Isotemnus primitivus Ameghino, 1897 x x I. latidens (Ameghino, 1901) xx x Thomashuxleya rostrata Ameghino, 1901 x Thomashuxleya externa Ameghino, 1901 x x Scaglia kraglievichorum Simpson, 1957 -xx Trigonostylops wortmani Ameghino, 1897 X ?X x x Albertogaudrya unica Ameghino, 1901 x -- x x x A. sp. indet. x - Carolozittelia tapiroides Ameghino, 1901 x Florentinoameghinia mystica Simpson, 1932 - x - a Data from Simpson, 1948, 1964, 1967a, 1967b, and personal observations. b Localities: 1, Cohu6 Huapi (Gran Barranca); 2, Cafiad6n Vaca (excluding faunule 6); 3, Cafiad6n Hondo; 4, Colhu6 Huapi norte; 5, Cerro Blanco; 6, Rio Chico oeste; 7, Rio Chico este; 8, Cerro del Humo; 9, Cerro Negro; 10, Pico Salamanca; 11, Cafiad6n Lobo; 12, Tapera de L6pez.

Patagonia lie at the base ofa series ofmainly rine Monte Leon ("Patagonia") Formation, pyroclastic sediments, the "tobas de Sar- respectively. Largely loessic in origin (Spal- miento" (Sarmiento tuffs) ofFeruglio (1938). letti and Mazzoni, 1977), the "Sarmiento The Sarmiento tuffs, or "Sarmiento group" group" is nonetheless highly varied in li- (a term which has been used both formally thology and complex in internal structure and informally), also includes mammal-bear- (Simpson, 1940). Despite the long span of ing beds of Mustersan, Deseadan, and Col- time represented, it includes many hiatuses huehuapian age (all four faunas are found in and is relatively thin, perhaps 150 to 190 m superposition at the Gran Barranca south of thick (Feruglio, 1949). Simpson (1940) ap- Lago Colhue Huapi) and are bounded below plied the geographic South American land by the terrestrial sandstones/claystones ofthe mammal age names (see Simpson, 1933) to Rio Chico Formation and above by the ma- the rock units containing the respective 6 AMERICAN MUSEUM NOVITATES NO. 2820 mammal faunas, based in part on well marked thology and bed thickness preclude direct erosional planes between them. The term correlation of strata between their sections. "Casamayor" was proposed as a stage, how- Andreis (1977) proposed separate litho- ever, and has never been defined as a rock stratigraphic names for "Sarmiento group" unit. (See Savage, 1962; Pascual et al., 1965; sediments ofCasamayoran and Deseadan age and Simpson, 1971, for discussions regarding in the Caniadon Hondo region (see also Simp- the confusion of lithostratigraphic with son, 1935; Schaeffer, 1947). Andreis pro- chronostratigraphic and geochronologic posed the name Caniadon Hondo Formation terms.) for the Casamayoran beds, which in this area Based on detailed analysis of a 114.5-m overlie the Visser Member of the Rio Chico thick section measured at km 163 ofthe rail- Formation (Andreis et al., 1975) with angular way between Comodoro Rivadavia and Sar- unconformity. He referred to the unconform- miento (central Chubut, fig. 2), Spalletti and ably overlying Deseadan sediments as the El Mazzoni (1977) recognized three sedimen- Sol Formation. Andreis (1977) proposed these tary cycles in beds ofthe "Sarmiento group." names as replacements for those adopted by They believed that these cycles correspond, Simpson (1940) in the beliefthat names such respectively, to the Casamayoran, Muster- as "Deseado Formation" and "Casamayor san, and Deseadan land mammal ages. The Formation" confuse lithostratigraphic with sedimentary cycles are characterized by vary- chronostratigraphic terms (sensu Hedberg, ing percentages ofbentonites, tuffs, "homeo- 1976). Unlike the earlier terminology, the conglomerates," and paleosols; they were lithostratigraphic units proposed by Andreis correlated to the respective land mammal ages (1977) are defined as such, but their status as according to apparently correlative environ- replacements (as well as the status and cir- mental changes in faunal composition (Pas- cumscription ofthe names used by Simpson) cual and Odreman Rivas, 1971). Based on remains unclear or dubious. The areal extent study of a nearby section (8.5 km north of ofthe "El Sol Formation" was not given; the km 162 on the Comodoro Rivadavia to Sar- Cafiadon Hondo Formation" was defined as miento railway), Spalletti and Mazzoni later a strictly local entity, thinning to the south (1979) concluded that the "Sarmiento group" and not occurring to the west ofthe Rio Chi- constitutes a lithologic entity and defined it co. Thus, if the latter is to be recognized as as a formation with three members. Accord- valid, new names will have to be proposed ing to these authors, the lowest, or Gran Bar- for correlative strata in the region ofCainadon ranca Member, conformably overlies the Rio Vaca and probably elsewhere as well. In light Chico Formation, lacks vertical cyclicity, and of these confusions and unclear definitions, contains mammalian fossils ofCasamayoran I use the term "Sarmiento group," in an in- age. An erosional contact separates the Gran formal sense, to refer to all sediments of Barranca Member from the overlying Puesto Casamayoran to Colhuehuapian age in cen- Almendra Member. The Puesto Almendra tral Patagonia. Member consists ofalternating tuffs and conglomerates, contains mammalian fossils of Deseadan age near its base, and includes a CASAMAYORAN LOCAL FAUNAS basalt radiometrically dated at 35 ma. The AND BIOSTRATIGRAPHY uppermost, or Colhuehuapi Member, con- cordantly overlies the Puesto Almendra Simpson (1948, 1967b, and references Member and contains fossils of Colhuehu- therein) revised the mammals from the Casa- apian age. Spalletti and Mazzoni (1979) in- mayoran ofPatagonia and published locality dicated that the Gran Barranca Member data for such of the Ameghino type speci- probably corresponds to the lower and mid- mens as that information was available dle cycles of their earlier (1977) section and (1967a). Table 1 presents a revised list of that the lower part of the Puesto Almendra mammal species from the Casamayoran of Member may correspond to the upper cycle, Patagonia, based largely on these publica- but they stressed that lateral changes in li- tions but with some unpublished additions 1985 CIFELLI: THE CASAMAYORAN 7 and emendations (Savage and Russell, 1983, GRAN BARRANCA pp. 86-88, list a composite Casamayoran One ofthe most important fossil mammal fauna). None ofthese is definitely known from localities in the world, the Gran Barranca the Riochican, but a number of them belong south of Lago Colhue Huapi, Chubut (fig. 1) to genera also known from the Ernestoko- was discovered by Carlos Ameghino in 1895. kenia chaishoer zone (?transitional Riochi- Here, the four mammal faunas of the "Sar- can-Casamayoran) of the Bajo de la Palan- miento group"-Casamayoran, Mustersan, gana upper sandstone and the Kibenikhoria Deseadan, and Colhuehuapian-lie together Riochican zone ofthe Ca-nadon Hondo sand- in a single superposed sequence. Detailed sec- stone (Simpson, 1 935).3 There are no species tions through the "Sarmiento group" sedi- and, with a few dubious exceptions, no genera ments in this vicinity were described by Spal- known from the Casamayoran or Mustersan letti and Mazzoni (1977, 1979). Fossil and later faunas. These exceptions include mammals collected in the region by the Scar- the virtually indistinguishable large isotem- ritt Expeditions were keyed to measured sec- nids Thomashuxleya and Periphragnis, from tions; those profiles incorporating significant the Casamayoran and Mustersan, respective- Casamayoran assemblages are shown in fig- ly,4 and the aberrant ?didolodontoid Adian- ures 3-5 and their approximate locations are toides, now known by a species from the given in figure 2. The only direct lithologic Casamayoran of Caiiadon Vaca in addition correlation between all sections is a marker to the genotype from the Divisaderan (Cifelli horizon traced throughout5 by Simpson and Soria, 1983). ("marker bed" of figures 3-5). This is a tuff, Most of Carlos Ameghino's Casamayoran several meters in thickness, and usually mas- localities and a number ofothers were visited sive and white with a pink hue in a few ex- and collected by the Scarritt Patagonian Ex- posures. It almost always forms a prominent peditions. The collections so made are large salient, and many fossils ofCasamayoran age and, in most cases, well documented strati- were collected from it. This tuff may be graphically, and therefore offer an unparal- equivalent to level "P11" of Spalletti and leled opportunity to determine the contents, Mazzoni (1977, p. 265) and perhaps also to superposition, and age relationships of the the "primer nivel fosilifero" of Spalletti and various Casamayoran local faunas. The two Mazzoni (1979, p. 273). In all of the profiles largest samples were collected at the Gran which include the lower parts of the strati- Barranca, south of Lago Colhue Huapi, and graphic column, about 45 m of alternating at Cafnadon Vaca, west ofthe Rio Chico, cen- tuffs and bentonitic clays (McCartney, 1933), tral Chubut (fig. 1). Because these samples containing Casamayoran fossils, intervene are large, diverse, rather dissimilar, and span between the marker tuff and partially indu- relatively great stratigraphic intervals, initial rated tuffs and clays ofthe "Sarmiento group," comparison is limited to the two faunas. the "argiles fissilaires" of Ameghino (Mc- Cartney, 1934). Fossils have not been found in these diagenetically altered sediments. Al- 3Cabrera (1936; cited from Simpson, 1948, p. 126 though no continuous section is exposed, and 1967a, p. 71, and not seen by me) assigned some Riochican specimens to Casamayoran species; these Simpson estimated these opalized beds to identifications were considered doubtful by Simpson in range from 35 to 45 m in thickness; they lie the works cited ( have not seen the materials). Pascual in apparent concordance on the alternating (in Marshall et al., 1983, and personal commun.) reports sandstones and detrital claystones of the Rio the presence of typically Casamayoran species in the Chico Formation. Fossils were not recovered ?transitional Riochican-Casamayoran Bajo de la Palan- from the Rio Chico Formation in this vicin- gana upper sandstone fauna. ity by the Scarritt Expeditions, but Riochican 4 Simpson, 1967b, p. 163, comments: "Periphragnis mammals have since been discovered here is so closely similar to Thomashuxleya that the generic diagnosis [ofPeriphragnis] is not clear-cut and ifthe two were of the same age they would perhaps be considered s This horizon has been relocated by subsequent work- generically identical." ers (Marshall, personal commun.). 8 AMERICAN MUSEUM NOVITATES NO. 2820

FIG. 2. Details of black rectangular area on figure 1, showing approximate locations of sections measured at the Gran Barranca, south of Lago Colhue Huapi (figs. 3-5) by G. G. Simpson. Contours represent meters above sea level.

(Pascual and Bond, 1981). Above the marker m. The stratigraphically highest Casamayor- horizon, about 20 more meters of variously an mammals are registered on profile V (fig. colored alternating tuffs and bentonitic clay- 5), about 20 m above the base of the marker stones containing Casamayoran fossils occur. bed and in an indurated, manganese-bearing At this point in profiles I and II (and other pink tuff. This tuff may be unconformable measured sections not included here) a basal with underlying beds; Simpson observed it conglomerate and channel series begins; to grade laterally into the channel series noted traced eastward, this grades into an impure above. (The few specimens from the site in tuffwhich may or may not be unconformable question, no. 15, are somewhat rolled and with overlying strata (profile V). may have been reworked.) With the excep- Casamayoran mammals from the Gran tion of Periphragnis exauctus, the composi- Barranca were recovered from numerous ho- tion of this highest assemblage is Casama- rizons in the "Sarmiento group," from about yoran. To the west, on profile I (fig. 3), a 10 m above the silicified tuffs and claystones typically Mustersan assemblage (site 4), in- to (and including) the base of the channel cluding Distylophorus alouatinus, Periphrag- series, and were found in great abundance in nis exauctus, Rhyphodon sp., and Astrapono- the marker horizon. The maximum recorded tus sp., occurs in the base ofthe channel series vertical distribution of Casamayoran fossils and at a level slightly higher than that of site at the Gran Barranca is therefore about 60 15; Mustersan fossils (pertaining to the no- 1985 CIFELLI: THE CASAMAYORAN 9

section I GRAN BARRANCA rest of section omitted ______...... broadlybedded variegated fine -sands, tuffs, and clays, all - variable. Abundant fossils. c top of channel beds clay and sand section 11 Highly variable channel beds. top of exposure _____ Hard beds vesicular and .-...-soft channel beds ; ___ /weathering org., partly x-bedded C. ?disconformity <-^ --1 hard nodular bed fine clay, grn. above lava ------=-= gry., ylw. tuff and gry. to ylw. elsewhere x-bedded gravel ar[nd sand. Basal Ilocal lava flow pebbles include ve,sicular lava fine wht. sand 1/ (transition) - unconformity soft x-bedded sand and gravel u a local unconformity, small relief mottled clay, grading upward ==--=--=-=---massve_wht.msiewh.tftuf from pnk. to gry. to ylw. ______====tn--=wc= =_ === _ ======_ =- .. local hard org. channel bed .______ylw. clay and tuff _ _======- - as e ht t f massive gry. tuff _ ylw. clay - hard org. channel beds ---=-=---=----|_ipr uf n. gry. and pale ylw. clay, -----== mssvetuf it some silicified patches ______=------yl.tfacosc/pthe,tffpb or ylw. Channel series. ---=-==--=- ipretffrai Clay ball conglomerate ------trm gy.to yl. gry. to ylw. clay, shrlotc silicified inconspicuously banded les below B. ------'(rnio) wht. tuff impure buff, tongry gry. clay, grading upward to ylw. ig upward _- __ -. /tuffand to pnk. banded gry. and ylw. tuff, clay tuffaceous clay ------_= -whtchstuffpebay / hard gry. band (marker bed) \ ?unarpconfomtyc massive wht. tuff--- tuff - unconformity 643, _ tfissile ylw. clay 5o =._ gry. to grn. clay and sandy clay (transition) btfine ylw. clay clay with Mn bands bottom of exposurie - (transition) pnk., buff, gry., ylw. clays. Horiz. banding. Many gypsum seams. bottom of exposure FIG. 3. Measured sections I and II at the Gran Barranca south of Lago Colhue Huapi, redrawn after fieldnotes of G. G. Simpson. Numbers at left indicate Mustersan and Casamayoran fossil horizons or sites listed in table 2. See figure 2 for approximate locations ofsections. ?D, faunule ofprobable Deseadan age; C, faunule of Colhuehuapian age. tohippid genus Eomorphippus; sites 16 and conformity nearly 20 m higher than site 4. 17) are recorded also in section V, where they The channel series is thus taken here to rep- were collected both above and below an un- resent a lithologic change and, probably, hia- 10 AMERICAN MUSEUM NOVITATES NO. 2820

GRAN BARRANCA section III top of exposure section IV tuffaceous clays C(rest of section omitted) (transition) gry. impure tuff gry. impure tuff with Mn spots ' -- f-, near bottom, many bones--- ylw. clay sharp contact (marker bed) pale pnk. or gry. tuff and tuffaceous clays with clay, few fossils Mn spots near top

ylw. to pnk. tuffs and tuffaceous clays, sparse bones 12. impure gry. tuff, many fossils -local hard pink band - pale gry., pnk., and ylw. tuff and clay _ gry. to ylw. massive tuff

- gry. tuffaceous clay

ylw. tuff gypsum bands gry. tuffaceous clay, at least one brightly colored mottled bed gry. to buff tuffaceous clay with gypsum veins -variegated, soft, bright clays bedded gypsum and upper bed of gry., silicified tuff gypsiferous clay with limonite concretions, capping partially metamorphosed tuffs and clays ______- -soft gypsiferous clay

-=-=-=== ------thick series of partly - metamorphosed and silicified tufts and clays FIG. 4. Measured sections III and IV at the Gran Barranca south of Lago Colhue Huapi, redrawn after fieldnotes of G. G. Simpson. Numbers at left indicate Casamayoran fossil horizons or sites listed in table 2. See figure 2 for approximate locations of sections.

tus, coinciding with the distribution of Casa- base of the channel beds described in the mayoran and Mustersan fossils at the Gran present paper. The lower part of the Puesto Barranca. Fossils have unfortunately not been Almendra Member, consisting largely ofcon- recovered from areas which may have had glomerates, was considered by these authors more continuous deposition or less erosion to be early Oligocene in age because it in- during the time represented by this hiatus. cludes a basalt flow, possibly correlative to Approximately 20 m ofsediments, the same that at Cerro Blanco dated by Marshall et al. distance as that between the marker horizon (1977) at 35 ma. and because it contains a and the channel series, also occur between Deseadan fauna. Sediments ofMustersan age Spalletti and Mazzoni's (1979, fig. 2) "primer were therefore probably not deposited (or nivel fosilifero" and the unconformity sep- were removed prior to renewed deposition arating their Gran Barranca and Puesto Al- during the Deseadan) in the area studied by mendra members of the "Sarmiento For- Spalletti and Mazzoni (1979). In another sec- mation." It is thus possible that the "primer tion, these authors had previously (1977) re- nivel fosilifero" is a correlate of the marker corded Casamayoran fossils from a level P1 1 horizon and the Gran Barranca-Puesto Al- (possibly equivalent to the marker horizon, mendra unconformity is a correlate of the as noted above); a marked unconformity sep- 1985 CIFELLI: THE CASAMAYORAN I 1

GRAN BARRAN section V section VI --(rest of section omitted) (rest of section omitted) 17 \irreg. bedded pnk. tuff, limonite and Mn concretions,conglomerate 4 = \ ~~~lava 16: of tuff pebbles in lower part o*'-:4.;~~~--.' unconformity - unconformity 0;°oX.;9 0 O Channel beds. Tuffs with many strong vo,*>,;°s,s|^O-@_ o x-bedded sands O..and. tuff pebbles

unconformity _- _- _- -_-_____-______-______-_ 0.°,.-^.w~_ _ _ -_ gry. tuff, in places grn. arLnd with bright ylw. patches ______- ______-_- _- -_ -_ -_ =- =- =- -= = -= -______-_ _ _ __- _ _ _ -______- _ _ _ -_- _ _ _ __- _ _ _ -_======_=-_ _ _ __- _ _ _ massive, impure, faintly bedded ______-= = ==-_ = -tuff with Mn, especially in lower part =------=--- ipr n.tf _ ---__ =-_ I = impure pnk. tuffs n r c. lays 0 = = = gipry. tolpn. tuffsadgn - - _= ==-_ = _ g y.to p k.tu ------. 21 -___-____-__-_- _ _ _ __- _ _ _ -=______--= -= -= -_ , -_-______--= -_--= -_--= -_-= -_-_ -----=--- iprewt-adyw.tf -______-_ _ _ __- _ _ _ (marker bed) -______-_ _ _ __- _ _ _ -______- _ _ _ = =_ _ _ ==_ l~~~~~~~~~~~~~~~~~~~~~~---=-===- ==---- - m uewh.adyw u ______- -_- -_- -_ -_ -______- -_- -_- -_- -_ -_

______--- hard band -impure tuft and fine ylw. clay | -_-==-==-= =-= -======- - wht. tuftf = tUff 2C _- ==_- _- _- _ gry. -fine ylw. and gin. clay, __======- some tuft ------= - ylw. to pnk. tuff

*-impure, irreg. ylw. tuft |/ 1 _---=------impure ylw. tuff *-gry. to pnk. tuft, partly indurated and concretionary|

-ipr__-tuff-ylw to brn -impure ylw. tuft e------cnrtoar r.tf 1E _-----= = imue l. rbuftuf --=-=-=--- massive wht. tuff =---- =-----/tiksre fvreae -_-=- -- -=-- -iieoeeioay patl -_ ------_ -/n J, - - -=--=-=--- -sharp contact impuetamworphosedtuffs ,ified an lht _ fine ylw. clay = Q_ tufmadetamophsebrg --(rest of section omitted) FIG. 5. Measured sections V and VI at the Gran Barranca south of Lago Colhue Huapi, redrawn after fieldnotes of G. G. Simpson. Numbers at left indicate Mustersan and Casamayoran fossil horizons or sites listed in table 2. See figure 2 for approximate locations of sections.

arates their horizons P14 and P15 approxi- unconformity and channel beds recognized mately 20 m above level P11. This is an ap- here, although it should be noted that Spal- propriate level to correspond to the letti and Mazzoni (1977) did not observe con- I I

00 l 111111111 ll

Bm11111111111,I,,,,,,,,, t3T X XR,iiX X XR 1111..... I1N- x , , , , , , x , , , x , , x , x X I I, X 11- x , , , , , , x , , x x x , , x , , x , , x ,

|~~~~~~~~~~~~~ X

011_0VwI XI I bX X X~IXI II IIXIIgX II X1 I 1 X1 1I I1 I8III1

-J

C ~~ ~ ~ r Zr a4 ni~~~~~~~~~~~~~~~~~~~~

t- Ii It < 3 Id 9tI! e d E i I cz44mCtURNittitZtutC WNC44mCoNCCc

12 1985 CIFELLI: THE CASAMAYORAN 13

glomerates at this point in their section. If these correlations are correct, the distribution

I offaunas presented herein corroborates Spal- 0 letti and Mazzoni's (1977) association ofsed- c w imentary cycles to the Casamayoran, Mus- tersan, and Deseadan land mammal ages, 00 respectively. 0 The composition of the Gran Barranca assemblages is given in table 2, and the strati- I X graphic ranges of identified mammals from the Gran Barranca, taken from profiles I-VI, g are combined in figure 6. Although local vari- C.) I lii iI ations in lithology and bed thicknesses pre- U, clude exact correlation, the thicknesses of 00 strata between the marker horizon and the channel series unconformity and the diagenetically altered tuffs and claystones are approximately the same in all sections where 410,0 they are recorded. Therefore, there is no rea- son to believe that the relative vertical position of the various assemblages would be more o d different were the sections lithologically I- correlatable. X: ,,0 Many species are known by only one or a 00 I I U few specimens. Ofthe more abundant species 0.- I I (those assumed to be most likely to be sam- N pled at any given horizon if they were orig- mH I inally present), Notostylops murinus, Old- fieldthomasia debilitata, Notopithecus adapinus, Pleurostylodon modicus, and Trigono- U0 stylops ?wortmani are not known from levels above the marker bed, while Homa- CU C lostylops parvus, Ultrapithecus rutilans, and Antepithecus brachystephanus are not known iiQ from below that horizon. These distributions 0 may be significant because most ofthe species listed are relatively abundant where they occur. Nonetheless, the assemblages are un- ll l l~~~41 evenly distributed throughout the section and o, sampling is poor for large parts of it, espe- ~~41U 3 cially above the marker horizon. The distri- 0o bution ofspecimens in the combined section is shown in figure 7. About 68 percent of F iC t R , identified materials derived from or near the marker tuff (sites 1, 2, 3, 5, 6, 7, 1 1, 13, 21); 4 24 percent came from scattered levels below GM that stratigraphic interval (sites 9, 10, 12, 18, 19, 20); and the remainder from above it (sites 8, 14, 15). The possibility that species absences from the upper or lower part of the section are the result of sampling inadequa- cies may be evaluated by determining the 14 AMERICAN MUSEUM NOVITATES NO. 2820

30 20 10 0 10 20

METERS FROM BASE OF MARKER BED Caroloameghinia mater Patene coluapiensis Polydolops thomasi 2 |l|i11l4lq |Amphidolops serrula Amphidolops sp.I II ~~~~~~~Utaetus buccatusI Utaetus sp. 3 Ernestokokenia nitida ||@p11 ,, ?Henricosbornia lophodonta n murinus 18 I i I | | Notostylops I?O 11111 Notostylops sp. 3 Homalostylops parvus 5 Homalostylops sp. Oldfieldthomasia debilitata 26 IT t zitS |I Ultrapithecus rutilans 15 Notopithecus adapinus 12 N. a. adapinus 6 N. a. reduncus 7 Antepithecus brachystephanus 32 Transpithecus ?obtentus Eohyrax isotemnoides 3 |_0 + Pleurostylodon modicus 14 Pleurostylodon sp. Anisotemnus distentus Anisotemnus sp. I ,, Isotemnus primitivus Thomashuxleya rostrata 3 Trigonostylops ?wortmani 8 O|| ||| Albertogaudrya unica I sp.

FIG.6.AbdsDistylophorus alouatinus a t BarashofLooPeriphragnis exauctus -V 2

3-5).Thebaseofthemarkerhorizon, traced throughout,Eomorphippus obscurus t c ?E. pascuali |

18 19 20 '12 9 ah, 10,l,1,5,21'2 3.11 6 7,13 14 8 15 4 RELATIVE~~~~~~~~~~~~~~~~~AlbertogaudryaSITE LEVEL TOTAL NUMBER OF SPECIMENS/ FIG. 6. Abstracted stratigraphic distributions ofCasamayoran and Mustersan mammals at the Gran Barranca south ofLago Colhu'e Huapil, combining distributional clata from measured sections I-VI (figs. 3-5). The base of the marker horizon, traced throughout, is the lithostratigraphic datum for this composite. Numbers at right denote fossil horizons or sites listed in table 2 and indicated on figures 3-5 (Mustersan faunules nos. 16 and 17 are omitted). Open circles: three or fewer specimens of that species recovered from a given horizon; closed circles, more than three. Light horizontal lines: less than a total of five specimens placed stratigraphically for that species; heavy lines, five or more specimens. Other taxa known from the Gran Barranca locality region cannot be placed precisely in this stratigraphic column. 1985 CIFELLI: THE CASAMAYORAN 15

125- p (predicted abundance value for sample in which that species is absent) 100. co = UE PNx; 0 75 * Q a where n = number of specimens pertaining 0 (base of marker bed) to that species, N = the total number ofspec- 50. imens from the marker horizon (sites 1, 2, 3,

25. 5, 6, 7, 1, 13, 21), and Nx = the total number of specimens from above (sites 8, 14, 15) or r-m below (sites 9, 10, 12, 18, 19, 20) the marker 0 6 12 18 24 30 36 42 48 54 60 66 on the dis- section horizon. (The value used depends meters of tribution of the species in question.) The FIG. 7. Histogram showing stratigraphic dis- probability that a given species absence is the tribution with respect to relative abundance offos- result of sampling inadequacy (i.e., the prob- recovered by the Scarritt sil mammal specimens ability that it was present but no specimens Patagonian Expeditions at the Gran Barranca. As marker horizon is were found) is determined by using a normal with figure 6, the base of the binomial distribution: used as a lithostratigraphic datum plane; numbers approximation' to the at left indicate total number of specimens re- X-PNx covered from each 6-m interval. (N + Nx)P(1 - P)' variable (in relative abundance of each species where it where X is the binomial random is found and determining the probability that this case, equal to zero because it is absence its nonrepresentation elsewhere is due to data which are being evaluated), and Z is the chance. Using the binomial distribution, standard normal random variable (X population parameters are determined as follows: The probabilities of species absences due to chance are given in table 3. These prob- P (abundance of occurrence for a given abilities are not great for those species outside species) the lower part ofthe section (where sampling n is good), especially for otherwise abundant N + Nx species such as Antepithecus brachystephanus. The higher values for those absent from S (standard deviation) levels above the marker horizon reflects the = v(N + NX)P(1- Pj smaller sample size from that portion of the

TABLE 3 Distributional Significance of the More Abundant Gran Barranca Speciesa,b Species n P s p Z A Notostylops murinus 18 .099 4.02 1.58 -.39 .3483 Oldfieldthomasia debilitata 26 .144 4.72 2.30 -.49 .3121 Notopithecus adapinus 25 .138 4.64 2.21 -.48 .3156 Pleurostylodon modicus 14 .077 3.59 1.23 -.34 .3669 Trigonostylops ?wortmani 8 .044 2.76 0.70 -.25 .4013 Homalostylops parvus 5 .033 2.19 1.55 -.71 .2389 Ultrapithecus rutilans 15 .100 3.67 4.70 -1.28 .1003 Antepithecus brachystephanus 32 .213 5.01 10.01 -2.00 .0228 a N (number of specimens from marker horizon localities, nos. 1, 2, 3, 5, 6, 7, 11, 13, 21) = 134; N, (number of specimens from the upper localities, nos. 8, 14, 15) = 16; N2 (number of specimens from the lower localities, nos. 9, 10, 12, 18, 19, 20) = 47. b n = no. of specimens; P = probability of occurrence; s = standard deviation; p = predicted value of occurrence; Z = value of transformed normal random variable; A = area of normal curve for Z c 0. 16 AMERICAN MUSEUM NOVITATES NO. 2820

TABLE 4 Distributional Significance of Notopithecus adapinus Subspeciesa Subspecies n P s p Z A Notopithecus adapinus adapinus 6 .045 2.39 1.58 -.66 .2546 Notopithecus adapinus reduncus 5 .200 2.37 24.60 -10.40 .0000 a See table 3 for abbreviations. N (marker horizon localities) = 134 - 11 Notopithecus adapinus specimens = 123; N1 (number of specimens from localities 12, 20) = 35.

section, and are greatest for the rarer forms 8). The lowest fossils in place are about 5 m such as Trigonostylops ?wortmani. above the Rio Chico contact on section II; The taxonomy and relative stratigraphic three additional fossil levels on section II are positions of the subspecies of Notopithecus at about 12, 18, and 86 m, respectively, above adapinus have been discussed by Simpson the Rio Chico contact. The lowest three of (1967b, pp. 78-80). Specimens recognizable these horizons (sites 3, 4, 5) are highly pro- as belonging to both subspecies, N. a. adapi- ductive; the small assemblage from site 6 is nus and N. a. reduncus, have not been found of great interest because it includes species at the same horizon. Rather, N. a. adapinus otherwise unknown from Caiiadon Vaca. Two is restricted to localities at the level of the fossil horizons are recorded on section I. As- marker horizon, whereas N. a. reduncus oc- semblage 1, probably a lateral equivalent to curs only at two levels (sites 12, 20) well be- site 4 on section II, produced numerous iso- low that horizon (fig. 6). As before, the prob- lated teeth; assemblage 2, about 41 m higher, abilities that these observed distributions are includes but a few specimens. The known the result of chance nonrepresentation may Casamayoran of Ca-nadon Vaca thus occu- be determined by a normal approximation to pies some 81 m of section, with most spec- the binomial distribution (table 4). The gra- imens deriving from 5 to 19 m above the Rio cile-jawed (?a primitive condition) N. a. re- Chico contact. duncus, rather abundant at sites 12 and 20, The known fauna from Cafiadon Vaca is was almost certainly never present at the given in table 5. Some 30 species of fossil marker horizon; there is an approximately 25 mammals are included; with the exception of percent chance that the absence of N. a. site 6, the individual assemblages are highly adapinus at localities 12 and 20 is a result of similar in composition. Assemblage 1 in- sampling error. cludes nine species not known from elsewhere but it also has the most species, and this difference probably reflects the fact that CARAD6N VACA it is better sampled than the others. Large samples of Casamayoran mammals Excluding site 6, the fauna ofCainadon Vaca were recovered by the first Scarritt Patagoni- is strikingly dissimilar to that of the Gran an Expedition also at Caniadon Vaca, west of Barranca (tables 1, 2, 5). Of a total of 416 the Rio Chico and approximately 60 km identified species collected at known hori- northeast ofthe localities south ofLago Col- zons from either locality by the Scarritt Ex- hue Huapi (see fig. 1). "Sarmiento group" peditions, only three or four are common to rocks, consisting of variously colored alter- both. The shared species include Polydolops nating tuffs and bentonitic claystones, are ex- thomasi, Isotemnus primitivus, and (with posed on the north side ofthe cainadon. Con- some doubt) Henricosbornia lophodonta and glomeritic sandstone and an impure tuffoccur Anisolambda fissidens. Henricosbornia loph- at the base of the "Sarmiento group"; these odonta, abundant at Caniadon Vaca, is known overlie with a sharp but planar contact the gray shales (with no observable bedding) and red sandstones of the Rio Chico Formation. 6Including the Ameghino Collection, which includes Casamayoran fossils from Cainadon Vaca a number of specimens of uncertain exact provenience are registered on two measured sections (fig. from the Gran Barranca, the total is about 60. 1985 CIFELLI: THE CASAMAYORAN 17

CANADON VACA CANADON VACA section I section 11 -top of exposure - ylw. clay and wht. tuff partly indurated wht. ylw. clays and tuffs, and gry. tuff some bone frags. -pale ylw. clay

____ fine wht. tuff or irreg. bedded, often nodular tuffaceous clay light tuff and ylw. clay ylw. or grn. clay ylw. clays and soft tuffs, - impure wht. tuff, much Mn little Mn _ ylw. clay ylw. impure tuff and clay, -- irreg. indurated wht. tuff 2 some Mn pale ylw. clay ylw. clay, many Mn concretions _---_ ~wht. tuff ylw. clay, some Mn impure ylw. tuff and clay, some Mn concretions -=--_ clays and impure tuffs, -,ylw. tuffaceous clay, - wht. or pale colored, 0 weathering gry. weathering pnk. or ylw. wht. tuff \ylw. tuffaceous clay, weathering gry. __massive or faintly bedded =\wht. to ylw. tuffaceous clay ylw. tuff t. and nodular tuff -, irreg. wht. or ylw. impure tuft - massive impure ylw. tuff, massive tuff grading into tuffaceous clay ylw. --hard, white, vesicular tuft, alt. ylw. tuffaceous clays and limonite concretions _ ylw., wht., or grn. -impure ylw. tuft impure hard tuffs -_ thin-bedded impure tuff, sand, gravel massive wht. to buff tuff wht. tuft with Mn -impure ylw. concretionary tuffs gry. shale (Rio Chico Fm.) -brn. or grn. tuffaceous clays, some gypsum and Mn -massive wht. tuff :7 - bottom of exposure

FIG. 8. Measured sections I and II at Cafiadon Vaca, west ofthe Rio Chico (see fig. 1), redrawn after fieldnotes of G. G. Simpson. Section I was measured on the southern rim of Pampa Pelada in the upper part of Cafiadon Vaca; section II was taken about 5 km east of section I. Numbers at left indicate Casamayoran fossil sites or horizons listed in table 2. by a single M2 (AMNH 28726) from the specimens to Homalostylops parvus (a species marker horizon at the Gran Barranca. This modestly abundant in the upper horizons at tooth is somewhat larger than those of the the Gran Barranca). These specimens present Cainadon Vaca sample. The presence of An- characters somewhat more primitive than isolambdafissidens at both localities depends those ofthe Gran Barranca sample and prob- on hypothetical synonymies which I have ably, but not surely, represent a separate yet presented elsewhere (Cifelli, 1983). Simpson closely allied species. (1948, p. 212) referred two Cainadon Vaca Fallaw (1979) has shown that the least 18 AMERICAN MUSEUM NOVITATES NO. 2820

TABLE 5 Composition of Cafiad6n Vaca Faunulesa 1 2 3 4 5 6 Polydolops thomasi X X - P. borcurhor X - - Prostegotherium sp. indet. X - - Didolodus minor X - - Asmithwoodwardia subtrigona X - - Victorlemoinea sp. indet. X - - Anisolambdafissidens X - - A. amel X - - Henricosbornia lophodonta X - X X Othnielmarshia lacunifera X - X Peripantostylops minutus X - Notostylops murinus - - - - - X N. pendens X - X X X N. appressus - - - - X N. sp. indet. - X Homalostylops sp. nov.? X Maxschlosseria consumata X M. minuta - - X Ultrapithecus cf. rutilans - - - - - X Acropithecus rigidus X - ?X X Notopithecus adapinus - - - - - X Eohyrax praerusticus - - - - - X Pleurostylodon similis X - X X X Isotemnus primitivus ?X - X Thomashuxleya externa X - X X X Trignostylops wortmani ?X - - ?X ?X Albertogaudrya sp. indet. - - - - X a Numbers refer to sites or levels in measured sections offigure 7. Data from the collections ofthe Scarritt Patagonian Expeditions only; occurrence not known by precise stratigraphic horizon ignored. See table 1 for author and date of species. biased7 binary coefficient offaunal similarity 61 and 19, respectively. These figures are is that proposed by Simpson (1960); this may comparable to those of some successive land be calculated as follows: mammal ages; by comparison, a similarly

C computed index for Gidley and Scarritt quar- x 100, ries, Montana (Torrejonian and Tiffanian N North American land mammal ages, respec- where C is the number of taxa common to tively) yielded a species value of38 (Simpson, both faunas and Nis the total number oftaxa 1960, p. 310; see discussion below). Assem- in the less diverse of the two faunas. Indices blage 6, the stratigraphically highest of the of taxonomic resemblance for genera and Cainadon Vaca sample, is of special interest species from Ca?nadon Vaca and the Gran in this regard because the included taxa (No- Barranca, based on the data of table 1, are tostylops ?murinus, Eohyrax praerusticus, Notopithecus adapinus, and Ultrapithecus cf. rutilans) are typical ofthe Gran Barranca and 7Raup and Crick (1979) have shown that this coeffi- are otherwise unknown from Cainadon Vaca. cient is biased and have developed a probabilistic ap- The faunal dissimilarity between these two proach to assessing faunal similarity, using Monte Carlo localities is reflected in a comparison of the computer simulations. Further, more detailed analysis of the present data is clearly warranted and will be pre- relative representation of ungulate families sented elsewhere. (fig. 9). The Cainadon Vaca fauna is domi- 1985 CIFELLI: THE CASAMAYORAN 19 nated by archaic families (Henricosborni- idae, Isotemnidae); the Gran Barranca by 301 CANADON VACA more advanced forms (Oldfieldthomasiidae, Interatheriidae). The relative primitiveness 20- of the Cafnad6n Vaca fauna is reflected also in comparison of sister taxa (closely related species or genus pairs) of the two localities. 10- Homalostylops sp. from Caniadon Vaca differs from H. parvus, the Gran Barranca species, in being smaller; the cheek teeth are o I I lr-]I t lower crowned, and the median lower molar a))X O O coX ' o trigonid cusp is more salient. Notostylops *n0 a) c-i XcD murinus, from the Gran Barranca, is ad- CD~ ~ ~ ~ ~ -cc o vanced with respect to the Ca-nadon Vaca Z o cC 0 species (N. pendens, N. appressus) in that the 00 CD 0-co -cCZ co Cu ~ 4U upper premolar series is enlarged, with an Z 0 < < -C CO a. internal groove on P3 4 (Simpson, 1948, pp. 191-201). Maxschlosseria, an archaic old- COLHUE HUAPI fieldthomasiid (see Bond, 1981) represented 30 at Ca-nad6n Vaca by M. consumata, differs from its close relative Ultrapithecus (repre- 20 sented at Colhue Huapi by U. rutilans) in having lower crowned cheek teeth and relatively smaller posterior upper premolars, with 10 a lesser development of the postcingulum. Archaeopithecus, known from the Gran Bar- ranca but not represented in the AMNH col- 0 1 1 1 1 1 1I 1 1 F-I1 1 lections, is more derived than Acropithecus FIG. 9. Comparison of the relative represen- (from Ca-nad6n Vaca) in having more trans- tation ofungulate families between the combined verse upper premolars (particularly Pl-2) and faunas of the Gran Barranca and Cainadon Vaca. in the reduction of ectoloph folds on the up- Numbers at left refer to total of specimens refer- per cheek teeth (Simpson, 1967b, p. 63). The able to a given family. character polarities ofother sister taxa ofthe two localities are not yet well understood. It seems unlikely that these faunal differ- miento group," beginning several meters ences are ecological in nature. Considered as above the contact with the Rio Chico For- wholes, each fauna extends through a rela- mation. No measured section at the Gran tively great stratigraphic section, implying Barranca includes this contact; however, sampling over a long span oftime and, in the the stratigraphically lowest fossil mam- case ofthe Gran Barranca at least, the fossils mals from south of Lago Colhue Huapi derive from geographically widely distribut- probably derive from at least 40 m above ed places. The two districts are rather close the Rio Chico contact. geographically, and there is no geologic evi- 2. The uppermost assemblage (site 6) from dence to suggest the presence ofdifferent cli- Cafiadon Vaca, which is considerably matic regimes or zoogeographic barriers be- higher than the other sites from that lo- tween them. That the faunal differences cality, shares all its identifiable taxa with between Ca-nadon Vaca and the Gran Bar- the Gran Barranca and none with Caina- ranca are temporal and not ecological is fur- don Vaca sites 1-5. ther suggested by lithostratigraphic, biostrati- 3. The fauna of Cainadon Vaca is more sim- graphic, and paleontologic evidence: ilar to that of the Rio Chico Formation than is that of the Gran Barranca (table 1. The levels offossil concentration at Cafia- 6). The archaic nature ofthe Cafiad6n Vaca don Vaca lie near the base of the "Sar- fauna is also reflected in the relative rep- 20 AMERICAN MUSEUM NOVITATES NO. 2820

RIOCHICAN CASAMAYORAN MUSTERSAN -.1 ______

Gran Barranca

Caiiadon Vaca

Colhue Huap(Norte

Caniadon Lobo

BA IACA Tapera de Lopez eal I' lt Rio Chico Oeste

Rio Chico Este

VACAN

FIG. 10. Hypothesized relative age relationships of the better represented Casamayoran local faunas ofPatagonia, using the distributions at the Gran Barranca and Caiiadon Vaca as standards forcomparison. See text for discussion.

resentation ofungulate families and in the before, are given in table 6, and an hypothesis fact that it includes the more primitive oftheir relative age relationships is shown in members of sister taxa, where they can be figure 10. The values for these indices cor- identified. roborate the distinction ofthe Gran Barranca and Ca-nadon Vaca assemblages because they OTHER PATAGONIAN LOCALITIES show a negative correlation of similarity to the two faunas. The faunas from Colhue Hua- It is of interest to consider the composi- pi norte, Ca-nadon Lobo, and Tapera de L6- tions of other Casamayoran local faunas in pez are highly similar to that of the Gran the context ofthe sequences discussed above, Barranca. The exact provenience ofthe spec- although they are (for the most part) not well imens from Colhue Huapi norte (all are in represented or understood stratigraphically. the Ameghino Collection, currently housed Ofthose listed in table 1, only five additional in the Museo Argentino de Ciencas Natu- localities have yielded assemblages including rales, Buenos Aires) is uncertain. However, five or more identified mammal taxa. Four the presence of species found in the lower ofthe five (Tapera de L6pez, Rio Chico este, horizons at the Gran Barranca (Notopithecus Rio Chico oeste, Colhue Huapi norte) are adapinus, Trigonostylops wortmani, Pleuro- regional designations and are not known to stylodon modicus, Thomashuxleya rostrata) be true faunal assemblages; even the exact and the absence of species such as Homalo- location of some of the Ameghino localities stylops parvus, Uhrapithecus rutilans, and is unclear (see Simpson, 1948, 1967a; Mar- Antepithecus brachystephanus suggest corre- shall et al., 1983). lation with the lower Gran Barranca assem- Indices of faunal similarity for these lo- blage. calities and for the Riochican, calculated as Caiiadon Lobo (=Caniadon Toumouer), the 1985 CIFELLI: THE CASAMAYORAN 21

TABLE 6 indicates probable correlation with the mark- Simpson's Coefficient of Faunal Similarity for er bed. Casamayoran Local Faunas and the Riochican Casamayoran mammals were collected near the Tapera de L6pez, Chubut, by the second X Scarritt Patagonian Expedition (1933-1934) 0 N~~ and were described by Simpson (1948, 0 a 1 967b). Work in progress indicates that these mammals came from several districts and that, because the geology of the region is not yet well understood and because much ofthe region is covered with vegetation, the strati- Gran Barranca graphic relationships of these areas cannot Cafiad6n Vaca 19 now be determined. Two species (Isotemnus Colhue Huapi norte 88 0 ?primitivus and Polydolops thomasi) are Rio Chico oeste 35 18 25 shared with both Cainadon Vaca and the Gran Rio Chico este 14 57 0 14 Barranca; several other species are shared only Cafiad6n Lobo 80 0 20 20 20 with the latter locality but, with the exception Tapera de L6pez 86 29 14 29 0 0 of Notopithecus adapinus, they are not very Riochican 29 47 13 21 40 0 29 diagnostic for purposes of chronologic correlation because of their rareness. Marsupial specimens were recovered in unusual abundance from the Tapera de L6pez. type Casamayoran (Simpson, 1933), is The correlation of Ameghino's localities sparsely fossiliferous and has yielded a small along the Rio Chico is less clear. Their precise but important fauna. Carlos Ameghino ap- location is not known, and probably one, or parently collected only one (unidentifiable) both, represents an area rather than a circum- specimen here (Simpson, 1967a, p. 68); the scribed fossil locality. Many species from west Scarritt Expeditions recovered a single iden- of the Rio Chico are represented by single tifiable specimen (Caroloameghinia mater, specimens, holotypes, and are unknown else- Simpson, 1948, p. 38) at Ca-nadon Lobo. The where; they are thus ofno help in relative age main collection from this locality is that made determination. The presence of Oldfieldtho- by Andr6 Tournouer in 1903, described by masia debilitata, Trigonostylops wortmani, Gaudry (1906) and by Simpson (1964, Anisotemnus distentus, and Notopithecus 1967b). The "Sarmiento group" is repre- adapinus suggests an age roughly equivalent sented here by about 69 m ofalternating tuffs to that ofthe lower part ofthe Gran Barranca and claystones, occasionally with manganese section; nonetheless, the large isotemnids nodules, overlying an unknown thickness of Pleurostylodon similis and Thomashuxleya opalized tuffs and clays. The Monte Leon externa are common to Cafiadon Vaca but Formation overlies the "Sarmiento group" not to the Gran Barranca. Perhaps the Rio at Cainadon Lobo. Simpson discovered un- Chico oeste locality represents an assemblage identifiable mammal bones approximately 25 from different stratigraphic horizons, or per- m above the silicified tuffs and clays on the haps it represents a local fauna intermediate north side ofCainadon Lobo and about 21 m in age between well known Ca-nadon Vaca above the same on the south side ofthe cafna- and Gran Barranca assemblages. don. (The single identifiable specimen col- The fauna from east of the Rio Chico lected by the Scarritt Expeditions was found strongly resembles that of Caiiadon Vaca in on the surface and its exact provenience is that it includes Maxschlosseria spp., Notosty- uncertain.) Ofthe modest fauna, Trigonosty- lops pendens, Edvardotrouessartia sola, Ac- lops wortmani, Oldfieldthomasia sp., Utaetus ropithecus rigidus, Thomashuxleya externa, buccatus, and Antepithecus brachystephanus and especially the henricosborniids Henri- are Gran Barranca forms; the presence ofthese cosbornia lophodonta, Othnielmarshia la- taxa best known from the marker horizon, cunifera, and Peripantostylops minutus. Con- but commonly found above and below it, tradictions to correlation with Cainadon Vaca 22 AMERICAN MUSEUM NOVITATES NO. 2820

TABLE 7 TABLE 8 Generic List of Riochican Mammals from Simpson's Coefficient of Faunal Similarity (Gen- Patagonia era) for North American Early Tertiary Land Mammal Agesa Seumadia Othnielmarshia Polydolops Peripantostylops Ages Index Gashternia Kibenikhoria Puercan-Torrejonian 30 Ernestokokenia Maxschlosseria Torrejonian-Tiffanian 44 Victorlemoinea Seudenius Tiffanian-Clarkforkian 53 Wainka Isotemnus Clarkforkian-Wasatchian 63 Anisolambda Shecenia Henricosbornia Carodnia a Data from Savage and Russell, 1983. Highly dubious occurrences were ignored. Wasatchian includes Gray Bull, Lysite, and Lost Cabin faunas; all are composites.

are the rare species Paginula parca (otherwise recorded from Cerro Negro and the Cafnadon can into three zones, named after taxa typical Lobo local fauna, ofprobable Gran Barranca of each: Ernestokokenia chaishoer, the age), and Anisotemnus distentus, a species youngest, including mammals from the Bajo shared with the Gran Barranca fauna. de la Palangana upper sandstone; Kibeni- khoria, intermediate in age, including the DISCUSSION Cafnadon Hondo sandstone fauna; and Ca- CASAMAYORAN LAND MAMMAL rodnia, the oldest, including lower Bajo de la AGE BOUNDARIES Palangana fossils and those from Cerro Re- dondo. Simpson (1935a) noted that, with the The problem ofdefining land mammal ages exception of the few known taxa from the in South America is particularly acute be- Carodnia zone, the fauna considered as a cause, except for the appearance of rodents whole is similar to that of the Casamayoran and primates in the ?late Eocene and the great and that distinction between the faunas is interAmerican interchange (which began at somewhat arbitrary with respect to interme- the close ofthe Miocene), no immigrants are diate assemblages. Of the 17 genera recog- available for use as "datum planes." It is to nized here from the Riochican of Patagonia, be expected that as transitional faunas be- nine are also known from the Casamayoran. come known, the distinctions between land The index of faunal similarity between the mammal ages will become less evident; many two land mammal ages, 53, is comparable to of the classic faunal distinctions in South indices for the Paleocene and early Eocene America's Tertiary record are based on ob- land mammal ages of North America (table served hiatuses in the record. 8). Of Casamayoran local faunas, Cainadon A generic list ofRiochican mammals from Vaca and Rio Chico este are most similar to Patagonia is given in table 7 (a composite the Riochican (see table 6). species list is given by Savage and Russell, With the possible exception of Periphrag- 1983, p. 46). Transpithecus and Notopithecus, nis and several other (highly dubious) occur- typical Casamayoran genera, were included rences (see Simpson, 1948, 1967b), no genera in the Riochican faunal list of Simpson are known from both the Mustersan and (1967b, p. 249); I here more conservatively Casamayoran. Pascual (1965) referred a local refer them to Notopithecinae, indet.8 Simp- fauna that he collected at Paso de los Indios, son (1935a) divided the then known Riochi- Provincia del Chubut, to the Casamayoran, but noted that it is not typical and in some respects is more similar to Mustersan assem- These and other Casamayoran taxa (both genera and species) are recorded in the Riochican by Pascual, as blages. He did not supply a faunal list, but noted above. Precise boundary circumscription of the the basis for this assessment seems to be: (1) Casamayoran must await detailed study ofrelevant Mus- the relatively advanced morphology of the tersan and Riochican assemblages, especially that ofthe isotemnid notoungulates and ofOxybunothe- Bajo de la Palangana upper sandstone. rium praecursor, and (2) the absence of No- 1985 CIFELLI: THE CASAMAYORAN 23 tostylops and of henricosborniid notoungu- Vaca and because the difference is recogniz- lates, and the rarity of Homalostylops. able in other local faunas of Casamayoran Oxybunotherium, based on lower molars, is age in Patagonia, two distinct Casamayoran a probable synonym ofProectocion, based on subages may be recognized. For these I pro- upper cheek teeth (Cifelli, 1983). The exact pose the names Barrancan and Vacan. The proveniences ofthe holotypes ofthe dubious- index of generic similarity between the two ly distinct P. argentinus and P. precisus are subages, 61, is approximately equal to that not known, but they were collected in the for the Clarkforkian and Wasatchian land Casamayoran beds south ofLago Colhue Hu- mammal ages ofNorth America (63; see table api (Simpson, 1948, pp. 108-109); a speci- 8). Riochican mammals are now known from men surely conspecific with the holotype of the Rio Chico Formation at Ca-nadon Vaca "Oxybunotherium praecursor," AMNH (M. F. Soria, personal commun.) and, be- 28769, was collected by the first Scarritt Pata- cause a Barrancan assemblage is also known gonian Expedition in Valle Hermoso (south- from that locality (site 6, section II; fig. 8), east of the main barranca). The associated the Vacan is thus known to be bounded by fauna includes Notopithecus adapinus and preceding and succeeding faunas in strati- Trigonostylops ?wortmani; correlation with graphic superposition. Vacan mammals are the lower portion of the Gran Barranca sec- not yet known from the Gran Barranca, but tion therefore seems most probable. The ab- the Barrancan fauna is there bounded above sence of henricosbomiid notoungulates at by the Mustersan. As identified by the first Paso de los Indios is thus unsurprising; the appearance of abundant and characteristic significance of other absences (such as No- taxa, these subages may be defined as follows: tostylops) cannot be evaluated without a complete faunal list and relative abundance Vacan Barrancan data. Anisolambda amel Utaetus Nearly all of the genera and species hith- Othnielmarshia lacuni- Didolodus multicuspis erto referred to the Casamayoran are there- fera Oldfieldthomasia debi- fore characteristic ofit, but the present study Peripantostylops minu- litata tus Homalostylops parvus indicates that many of them are rare or are Notostylops pendens Oldfieldthomasia debi- confined to given stratigraphic intervals with- Notostylops appressus litata in it; except for taxonomic revisions incor- Maxschlosseria consu- Ultrapithecus porated here, these results are in agreement mata Acoelodus with the faunal list given by Pascual and Acropithecus Paginula Odreman Rivas (1971). Notostylops, for Thomashuxleya exter- Notopithecus which the fauna was originally named, is re- na Eohyrax isotemnoides stricted to and spans the entire Casamayoran Thomashuxleya rostra- recognized here. Judged by their presence in ta the Gran Barranca, Cainad6n Vaca, and other Antepithecus brachys- Patagonian local faunas, it seems likely that tephanus Polydolops thomasi, Homalostylops, Eohy- rax, Isotemnus primitivus, Thomashuxleya, Superpositional (and therefore probably age and Trigonostylops have nearly the same related) faunal change is evident within the range. They may be taken as index taxa for Barrancan. Some of the observed distribu- the Casamayoran Land Mammal Age, and tion patterns may reflect sampling inadequa- their first appearance distinguishes the Casa- cies; nonetheless, an "early" and a "late" Bar- mayoran from the Riochican. rancan may be informally distinguished by the presence or absence9 ofcharacteristic and SUBDIVISIONS OF THE CASAMAYORAN abundant species: The foregoing analysis indicates the faunas of the Gran Barranca and Caiiadon Vaca to 9 All ofthese species occur together in the marker ho- be markedly different. Because this difference rizon. The subspecies Notopithecus adapinus reduncus results from the greater antiquity ofCafiadon is totally restricted to the early Barrancan. 24 AMERICAN MUSEUM NOVITATES NO. 2820

TABLE 9 Ameghino's Subdivision of the Casamayoran Ameghino, 1902 Simpson, 1967aa Notostylops superior Notopithecidae Utaetus lenisb Archaeopithecidae Machlydotherium sparsumb Henricosbornidae [sic] Meteutatus percarinatusb Carolozittelia Utaetus buccatus Albertogaudrya U. deustus Thomashuxleya Prostegotherium astrifer Didolodus P. notostylopianum Euprotogonia [Ernestokokenia]c Pseudostegotherium chubutanum Prohyracotherium [Henricosbornia] Paulogervaisia mamma Lophiodonticulusb Paulogervaisia porca Eochalicotherium [Isotemnus]c Paulogervaisia inusta Isotemnus "Acoelodus" priclivus Selenoconus [Henricosbornia]c Maxschlosseria minima Nephacodus [Didolodus]c Eohyrax rusticus Paulogervaisia Pleurostylodon modicus Trigonostylops Isotemnus colhuehuapensisb Notostylops Anisotemnus distentus Acelodus [sic] Pleurostylodon crassiramus" Oldfieldthomasia Albertogaudrya unica Miolania argentinad Dinosaurs Notostylops inferior Caroloameghinia Utaetus buccatus Maxschlosseria Maxschlosseria praeterita Ernestokokenia Notopithecus adapinus Amilnedwarsia [sic] Rutimeyeria Polydolops Ideodelphisb Argyrolestes peralestinus Nemolestes spalacotherinus Dinosaurs a After revisions by Simpson, not the name originally published. b Status dubious or uncertain. c Names in brackets follow revisions by Simpson, 1948, 1967b. d Not a mammal, but a homed turtle.

Early Barrancan Late Barrancan both Barrancan and Vacan age and therefore Notopithecus adapinus Homalostylops parvus may partly occlude the gap between the type Oldfieldthomasia debi- Ultrapithecus rutilans faunas (see fig. 10). litata Ameghino (1902) distinguished two sub- Pleurostylodon modi- Antepithecus brachy- divisions of what is now known as the Casa- cus stephanus mayoran fauna, a "Notostilopense superior" Notostylops murinus and a "Notostilopense inferior," and later Most of the other well known Casamayor- (1903) added a "basal" part (table 9). an localities of Patagonia are comparable to Ameghino's lists for the superior and inferior the Gran Barranca in content; Ameghino's Notostylops faunas are compared to desig- localities east and west of the Rio Chico in- nations he left with his collections or trans- clude (in different proportions) mammals of mitted orally to Simpson by Carlos Ameghi- 1985 CIFELLI: THE CASAMAYORAN 25 no (Simpson, 1967a; see discussion of the Bond, M. "basal" Notostylops therein). The basis for 1981. Un nuevo Oldfieldthomasiidae (Mam- these subdivisions is not now evident and malia, Notoungulata) del Eoceno infe- may have been in part hypothetical. In sev- rior (Fm. Lumbrera, Grupo Salta) del NW Argentino. An. II Congr. Latino- eral instances (such as the assignment of Old- Americano Paleontol., Porto Alegre, vol. fieldthomasia to the "Notostilopense supe- 2, pp. 521-536. rior" and of Maxschlosseria to the Cabrera, A. "Notostilopense inferior") Ameghino's re- 1936. Estado actual de la cuestion del limite ferrals are in accord with those ofthe present cretaceo-terciario en la Argentina. Mus. study, but the faunal lists are otherwise so La Plata, Obra del Cincuentenario, vol. contradictory that his zonation cannot be said 2, pp. 1-22. to have an adequate basis. Cifelli, R. L. 1983. The origin and phylogeny of the South LITERATURE CITED American Condylarthra and early Ter- tiary Litopterna (Mammalia). Am. Mus. Ameghino, F. Novitates, no. 2772, pp. 1-51. 1897. Mammiferes cretaces de l'Argentine. Cifelli, R. L., and M. F. Soria Deuxieme contribution a la connais- 1983. Systematics of the Adianthidae (Litop- sance de la faune mammalogique des terna, Mammalia). Ibid., no. 2771, pp. couches a Pyrotherium. Bol. Inst. Geogr. 1-25. Argentina, vol. 18, pp. 406-521. Fallaw, W. C. 1899. Sinopsis geologicopaleontologica. Su- 1979. A test of the Simpson Coefficient and plemento. Adiciones y correlaciones. La other binary coefficients of faunal sim- Plata, pp. 1-13. [Seen and cited as a ilarity. Jour. Paleontol., vol. 53, pp. separate; also published in Ameghino, 1029-1034. 1913-1936, vol. 12, pp. 763-767.] FeXruglio, E. 1902. Cuadro sinoptico de las formaciones se- 1938. Nomenclatura estratigrafica de la Pat- dimentarias terciarias y cretacicas de la agonia y Tierra del Fuego. Bol. Inf. Pe- Argentina, en relacion con el desarrollo trol., Y.P.F., Buenos Aires, no. 171, pp. y descendencia de los mamiferos. An. 54-67. Mus. Nac. Hist. Nat., vol. 8, pp. 1-12. 1949. Descripcion geologica de la Patagonia. 1903. L'age des formations sedimentaires de Minst. Ind. Com. de la Nacion, Div. Patagonie. An. Soc. Cientif. Argentina, Gen. Y.P.F., Buenos Aires, vol. 2, pp. vols. 50, 54, pp. 3-231. 1-349. 1906. Les formations sedimentaires du cre- Gaudry, A. tace superieur et du tertiare de Patagon- 1906. Fossiles de Patagonie. Etude sur une ie, avec un parallele entre leurs faunes portion du monde antarctique. An. Pa- mammalogiques et celles de l'ancien leontol., vol. 1, pp. 101-143. continent. An. Mus. Nac. Buenos Aires, Hedberg, A. D. (ED.) vol. 15 (ser. 3a, vol. 8), pp. 1-568. 1976. Inernational Stratigraphic Guide. New 1913-1936. Obras completas y corresponden- York, Wiley, pp. 1-200. cia cientifica de Florentino Ameghino, McCartney, G. C. A. Torcelli, ed. Impresiones Oficiales, 1933. The bentonites and closely related rocks La Plata, vols. 1-24. of Patagonia. Am. Mus. Novitates no. Andreis, R. R. 630, pp. 1-16. 1977. Geologia del area de Caniadon Hondo, 1934. The "argiles fissilaires", a series ofopal- Depto. Escalante, Provincia del Chu- bearing rocks of Patagonia. Am. Mus. but, Repuiblica Argentina. Mus. La Pla- Novitates, no. 687, pp. 1-8. ta, Obra del Centenario, vol. 4, pp. 77- Marshall, L. G. 102. 1982. Calibration of the Age of Mammals in Andreis, R. R., M. M. Mazzoni, and L. A. Spalletti South America. Gebios, mem. spec., no. 1975. Estudio estratigrafico y paleoambiental 6, pp. 427-437. de las sedimentitas terciarias entre Pico Marshall, L. G., R. F. Butler, R. E. Drake, and Salamanca y Bahia Bustamante, Pro- G. H. Curtis vincia de Chubut, Repiiblica Argentina. 1981. Calibration of the beginning of the Age Rev. Asoc. Geol. Argentina, vol. 30(1), ofMammals in Patagonia. Science, vol. pp. 85-103. 212, pp. 43-45. 26 AMERICAN MUSEUM NOVITATES NO. 2820

Marshall, L. G., R. Hoffstetter, and R. Pascual Schaeffer, B. 1983. Geochronology of the continental 1947. An Eocene serranid from Patagonia. mammal bearing Tertiary of South Am. Mus. Novitates, no. 1331, pp. 1-9. America. Palaeovertebrata, Mem. Ex- Simpson, G. G. traordinaire, pp. 1-93. 1932. The supposed association of dinosaurs Marshall, L. G., R. Pascual, G. H. Curtis, and with mammals of Tertiary type in Pat- R. E. Drake agonia. Ibid., no. 566, pp. 1-21. 1977. South American geochronology: radio- 1933. Stratigraphic nomenclature of the early metric time scale for middle to late Ter- Tertiary ofcentral Patagonia. Ibid., no. tiary mammal-bearing horizons in Pat- 644,pp.1-13. agonia. Science, vol. 195, pp. 1325- 1935. Occurrence and relationships ofthe Rio 1328. Chico fauna ofPatagonia. Ibid., no.818, Matthew, W. D., and W. Granger pp. 1-21. 1925. Fauna and correlation of the Gashato 1940. Review of the mammal-bearing Ter- Formation of Mongolia. Am. Mus. tiary ofSouth America. Proc. Am. Phil. Novitates, no. 18, pp. 1-12. Soc., vol. 83(5), pp. 649-709. Pascual, R. 1948. The beginning of the Age of Mammals 1965. Un nuevo Condylarthra (Mammalia) de in South America. Part 1. Bull. Am. edad Casamayorense de Paso de los In- Mus. Nat. Hist., vol. 91(1), pp. 1-232. dios (Chubut, Argentina). Breves con- 1960. Notes on the measurement offaunal re- sideraciones sobre la edad Casamayor- semblance. Am. J. Sci., vol. 258A, pp. ense. Ameghininana, vol. 4, pp. 57-65. 300-311. Pascual, R., and M. Bond 1964. Los mamiferos casamayorenses de la 1981. Epidolopinae subfam. nov. de los Poly- coleccion Tournouer. Rev. Mus. Argen- dolopidae (Marsupialia, Polydolopo- tino Cien. Nat. "Bernardino Rivada- idea). An. II Congr. Latino-Americano via," Paleontol., vol. 1(1), pp. 1-21. Paleontol., Porto Alegre, vol.2, pp. 479- 1967a. The Ameghinos' localities for early Ce- 488. nozoic mammals in Patagonia. Bull. Pascual, R., and 0. E. Odreman Rivas Mus. Comp. Zool., vol. 136(4), pp. 63- 1971. Evolucion de las comunidades de los 76. vertebrados del terciario Argentino. Los 1967b. The beginning of the Age of Mammals aspectos paleozoogeograficos y paleo- in South America. Part 2. Bull. Am. chimaticos relacionados. Ameghiniana, Mus. Nat. Hist. vol. 137, pp. 1-259. vol. 8(3-4), pp. 372-412. 1971. Clasificacion, terminologia y nomencla- Pascual, R., E. J. Ortega Hinojosa, D. Gondar, tura provinciales para el Cenozoico ma- and E. Tonni malifero. Rev. Asoc. Geol. Argentina, 1965. Las edades del Cenozoico mamalifero vol. 26(3), pp. 281-297. de la Argentina, con especial atencion a Spalletti, L. A., and M. M. Mazzoni aquellas del territorio bonaerense. An. 1977. Sedimentologia del Grupo Sarmiento en Com. Invest. Cien. Buenos Aires, vol. un perfil ubicado al sudeste del Lago 6, pp. 165-193. Colhue Huapi, Provincia de Chubut. Raup, D. M., and R. E. Crick Obra del Centenario, Mus. La Plata, vol. 1979. Measurement offaunal similarity in pa- 4, pp. 261-283. leontology. Jour. Paleontol., vol. 53, pp. 1979. Estratigrafia de la Formacion Sarmien- 1213-1227. to en la barranca sur del Lago Colhue Savage, D. E. Huapi, Provincia del Chubut. Rev. Asoc. 1962. Cenozoic geochronology of the fossil Geol. Argentina, vol. 34(4), pp. 271- mammals of the western hemisphere. 281. Rev. Mus. Argentino Cien. Nat. "Ber- nardino Rivadavia," Cien Zool., vol. 8(4), pp. 53-67. Savage, D. E., and D. E. Russell 1983. Mammalian Paleofaunas of the World. Reading, Mass., Addison-Wesley Publ. Co., pp. 1-432.