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

Home , Arctocyon, Eurotamandua, Metacheiromys

AMERICAN MUSEUM Novitates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2761, pp. 1-3 1, figs. 1-12, tables 1-6 May 31, 1983

Eutherian Tarsals from the Late Paleocene of Brazil

RICHARD L. CIFELLI'

ABSTRACT Disassociated eutherian proximal tarsals (as- kokenia parayirunhor, and a new form; astragali tragalus, calcaneum) from Riochican (late Paleo- and calcanea of "condylarth" aspect are referred cene) fissure fills near Sao Jose de Itaborai, Rio de to Lamegoia conodonta, Victorlemoinea prototyp- Janeiro, Brazil, are described and, where feasible, ica, and Ernestokokenia protocenica. The fact that are assigned to dental species from that locality, some dentally primitive taxa (including a sup- based on predicted morphology, relative size, and posed congener of a dental and tarsal condylarth) relative abundance. Two cingulate xenarthrans are bear the diagnostic litoptern ankle specializations, present; one is probably a dasypodid, whereas the whereas others, including an advanced and den- other may pertain to the Glyptodontidae. Ankle tally litoptern-like form do not, heightens the specializations of Carodnia vierai are unlike those problem of distinguishing the two groups as cur- of astrapotheres and Dinocerata, but similar to rently recognized, and indicates that the funda- those of Pyrotherium and suggest reference to the mental specializations of the Litoptema are post- Pyrotheria; Tetragonostylops apthomasi is pedally cranial, not dental. Victorlemoinea, heretofore a primitive astrapothere. Notoungulate tarsals considered a macraucheniid litoptern but proba- from Itaborai, referred to Camargomendesia pris- bly pertaining to the new family, Sparnotherio- tina (Henricosborniidae) and Colbertia magellan- dontidae, shares derived pedal characters with the ica (?Oldfieldthomasiidae), are of generalized as- didolodonts; among the didolodonts, Ernestoko- pect within the order, and represent a primitive kenia protocenica, Lamegoia conodonta, and ?Di- morphotype from which later notoungulate fam- dolodus sp. seem to form a natural (monophyletic) ilies diverged. Litoptern tarsals are assigned to group definable by tarsal features. three species, Anisolambda prodromus, Ernesto- INTRODUCTION Little is known ofthe once-diverse but now 1948), the pre-Deseadan (early Oligocene; largely extinct eutherians ofSouth America- Marshall et al., 1977) record of xenarthrans currently grouped into some six to nine or- consists only of isolated cingulate scutes. The ders-prior to the ?early Eocene Casamayor- early record of fossil mammals in South an Land Mammal Age. With the exception America is almost exclusively Patagonian, of one partial skeleton of Utaetus (Simpson, and with the exception of a small ?Late Cre-

' Student, Department of Vertebrate Paleontology, American Museum of Natural History.

Copyright © American Museum of Natural History 1983 ISSN 0003-0082 / Price $3.35 2 AMERICAN MUSEUM NOVITATES NO. 2761 taceous faunule from Laguna Umayo, Peru diversity of well-preserved but (with one ex- (Grambast et al., 1967) begins with the ception) unassociated postcranial remains. In Riochican Land Mammal Age, presumed to the present paper I describe the eutherian be late Paleocene (Simpson, 1935a; Marshall, proximal tarsal bones (astragalus, calca- Hoffstetter, and Pascual, in press). The Pa- neum) and discuss their significance with re- tagonian Riochican localities, grouped into spect to the phylogeny of the various groups several faunal zones (Simpson, 1935a, 1948), involved. The American Museum ofNatural have produced a modest diversity of ungu- History has a small collection ofItaborai tar- lates represented largely by isolated teeth. sals, but the bulk of this study is based on The fauna from Sao Jose de Itaborai, Brazil, the large collections of the Divisao Nacional therefore assumes a position of singular de Producao Mineral. prominence in any treatment of early South Abbreviations for institutions cited in the American fossil mammals, because it con- text: AMNH, Department of Vertebrate Pa- tains the most diverse and best represented leontology, American Museum of Natural assemblage of Riochican vertebrates of any History, N.Y.; ACM, Pratt Museum, Am- fauna on the continent. herst College, Amherst, Mass.; DNPM and The fossils derive from predominantly DGM, Divisao Nacional de Producao Mi- argillaceous marls which fill fissures and un- neral, Rio de Janeiro, Brazil; MACN, Museo derground channels in a limestone, the Itabo- Argentino de Ciencas Naturales "Bernardino rai Formation, of probable Late Cretaceous Rivadavia," Buenos Aires, Argentina; or early Paleocene age (Francisco and Souza MNRJ, Departamento de Geologia e Miner- Cunha, 1978). The limestone occurs in a alogia, Museu Nacional de Rio de Janeiro, small, local tectonic basin in a Precambrian Brazil. gneissic and granitic basement rock intruded by Early Cretaceous basalts, about 25 km. northeast of Niteroi, State of Rio de Janeiro. ACKNOWLEDGMENTS The fossils have been recovered during the I am grateful to Drs. M. Coombs (Amherst course of commercial limestone quarrying College), D. Campos (Divisao Nacional de over the past 35 years. Producao Mineral), F. de Souza Cunha (Mu- The mammals of Itaborai have been de- seu Nacional de Rio de Janeiro), J. Bonaparte scribed largely by Paula Couto (1952a, 1952b, and M. Soria (Museo Argentino de Ciencas 1952c, 1952d, 1954) with additions, emen- Naturales "Bernardino Rivadavia"), and R. dations, and refiguring (1961, 1962, 1963, 1970, Pascual and M. Bond (Museo de La Plata) 1978a, 1978b, 1978c, 1979). There is some for permission to study specimens in their question as to the precise age of the fauna, care. Drs. M. McKenna, F. Szalay, L. Radin- and it is possible that faunules from the var- sky, and G. Simpson provided me with various fissures may not be isochronous. Al- ious advice or comments on an earlier draft though the fauna is obviously at ecologic of this paper, and their assistance is warmly variance with those known from Patagonia appreciated. Figures 4-12 were drawn by Mr. (Paula Couto, 1970), Paula Couto (1952a) H. Galiano, and I thank Mr. C. Tarka, Ms. has assigned the Itaborai assemblage a Rio- L. Meeker, and Dr. R. Tedford for advice in chican age, based on the presence of Car- the preparation of the illustrations. I grate- odnia, which occurs in Patagonia only in the fully acknowledge financial support for this earliest faunal zone of that age (Simpson, research, provided in part by the Theodore 1935a). The other genera in common, ?Nem- Roosevelt Memorial Fund, the Walker John- olestes (see Marshall, 1978), Patene, Ernes- son Fund, The Society of the Sigma Xi, and tokokenia, Asmithwoodwardia, Anisolamb- the Department ofGeological Sciences ofCo- da, and Victorlemoinea, occur in the lumbia University. Casamayoran or Casamayoran and Riochi- can of Patagonia. In addition to the well-represented denti- METHODOLOGY tions described and figured by Paula Couto, The Itaborai tarsals are ofparticular inter- the collections from Itaborai include a wide est but present special problems because of 1983 CIFELLI: EUTHERIAN TARSALS 3

A B

cal mt V lig Im abd dig V fib fac-- / m quad pl

per tub , fib col lig

ascal lig m ex dig brev- ex ret D C ast for

ect fac

FIG. 1. Right calcaneum (A, B) and astragalus (C, D) of Arctocyonferox (AMNH 16543), in plantar (B, D) and dorsal (A, C) views, illustrating terminology used in this paper. Dot pattern indicates areas of origin/insertion of muscles and ligaments as noted in the text. Abbreviations: ant ascal fac, anterior astragalocalcaneal facet; ascal lig, attachment of astragalocalcaneal ligament; cal mt V lig, attachment ofcalcaneal metatarsal V ligament; cub fac, cuboid facet; dig fl gr, groove for deep digital flexor tendon(s); ex ret, attachment of extensor retinaculum; ect fac, ectal facet; fib col lig, attachment of fibular (lateral) collateral ligament; fib fac, fibular facet; inart sul, interarticular sulcus; inf ast for, inferior astragalar foramen; mcl fac, facet for the medial collateral ligaments; med mal fac, facet for medial malleolus of tibia; med col lig, attachment of medial (tibial) collateral ligament; m abd dig V, attachment of abductor digiti quinti muscle; m ex dig brev, attachment of extensor digitorum brevis muscle; m fl dig sup, attachment of flexor digitorum superficialis muscle; m quad pl, attachment ofquadratus plantae muscle; nav fac, navicular facet; per tub, peroneal tubercle; pl lig, attachment of plantar ligaments (to the cuboid and cuneiforms) and tarsal fibrocartilage; sup ast for, superior astragalar foramen; sus, calcaneal sustentaculum; sus fac, sustentacular facet; tib tr, tibial trochlea; tub, tuber calcis. their occurrence. The fauna is well known associated postcranial remains are unknown. and well represented, and confusing elements There can be no predicted morphology for such as rodents and primates (which appear these forms. The establishment ofreliable as- in the Deseadan, early Oligocene) are absent. sociations is of paramount importance. I Yet, except for Carodnia, all tarsals lack as- have employed three interdependent meth- sociation and a number of members of the ods in making the associations proposed be- fauna (Asmithwoodwardia scotti, a new un- low; morphology, relative size, and relative named form, Ernestokokenia parayirunhor, abundance. Associations of astragali to cal- E. protocenica, Lamegoia conodonta, and canea were based on morphology, relative Tetragonostylops apthomasi) pertain to den- size and abundance, and "fit." tal groups, families, or even orders for which Work in progress strongly supports the hy- 4 AMERICAN MUSEUM NOVITATES NO. 2761

TABLE 1 Szalay and Decker (1974) and is given in fig- Genera of South American Ungulates with Ankles ure 1. Known by Associationa Relevant South American mammals for which the tarsus is known by association and Litopterna available for study are listed in table 1 (many Theosodon (Macraucheniidae) other far more specialized species were stud- Proterotherium (Proterotheriidae) ied but are not listed). Despite the fact that Protheosodon (?Proterotheriidae)b many available species pertain to later, pre- Astrapotheria sumably more derived groups, consideration Parastrapotherium (Astrapotheriidae) ofstructural features shared by diverse mem- Astrapotherium (Astrapotheriidae) bers of a higher category (i.e., orders Xenar- Pyrotheria thra, Notoungulata, Litopterna, Astrapo- Pyrotherium (Pyrotheriidae) theria, and Pyrotheria) in comparison to a primitive eutherian morphotype allows the Notoungulata distinction of shared, derived characters pe- Boreastylops (Notostylopidae) culiar to that higher category, thus permitting Thomashuxleya (Isotemnidae) Adinotherium (Toxodontidae) preliminary allocation of isolated tarsals. Scarrittia (Leontiniidae) To aid in association of ungulate tarsals by Rhynchippus (Notohippidae) relative size, comparison was made to a num- Homalodotherium (Homalodotheriidae) ber ofprimitive ungulates for which the ankle Protypotherium (Interatheriidae) is known by direct association. Ten species Pachyrukhos (Hegetotheriidae) ofcondylarths were selected for this purpose, a The ankle of these is known by direct association. largely on the basis of availability, but also b Simpson (1948, p. 1 7) doubted the association of because they represent a similar level ofden- hind limb to mandible made by Loomis (1914, p. 41), tal and pedal organization to the species in and guessed that the postcranial elements might pertain question. Molar size, highly correlated with to a notoungulate. The ankle in question (ACM 3001) body size (Kay, 1975; Gingerich, 1977), was is definitely that ofa litoptern, probably a proterotheriid, approximated by area (maximum length by and I therefore consider Loomis's association to be cor- maximum width) of M2, and was compared rect. to maximum astragalar length. The logarithmic linear regression offigure 2 indicates that log(M2 area)= -.77 + 1.55 log(astragalar pothesis forwarded by Szalay (Szalay and length), with the parameters a and b of the Decker, 1974; Szalay, 1977; see also Schaef- equation Y = a + bX estimated mathemati- fer, 1947; Kielan-Jaworowska, 1977; Nova- cally. (A sample ofnine fossil tapiroids yield- cek, 1980) that tarsals assigned to the Maas- ed a log linear regression with similar slope trichtian genus Protungulatum present with but slightly greater Y-intercept.) The more few exceptions the morphology which would questionable Itaborai tarsal phena were com- be predicted for a primitive eutherian mor- pared to dental species from the quarry and photype. Among ungulates, this type of tar- associated with them using the condylarth sus, with some minor modifications and vari- regression as a standard of comparison (fig. ations, has a wide distribution among 3). The sample size is small with a large mar- primitive members of families generally gin ofvariability (see tables 2 and 3), and the grouped in the order Condylarthra, including Itaborai collections include a wide variety of the Arctocyonidae, Phenacodontidae, an- growth stages with a very high proportion of isonchine Periptychidae, and mioclaenine juveniles. Nonetheless, this approach does Hyopsodontidae. This morphotype, as ex- set broad limits on association by size and is emplified by Arctocyon, is used as the basis more objective than assessment "by eye." for comparison in the descriptions given be- With one exception, the fossils from Ita- low, and reference should be made to the borai lack precise collecting information. references cited above concerning specific That exception is fortunately a large and im- features of the primitive eutherian tarsus. portant sample collected by D. Campos and The terminology employed is modified after L. Price in 1968, and is known to be from a 1983 CIFELLI: EUTHERIAN TARSALS 5

0.2 0.2 L. conodonta (no astragaii)

0 0 V prototypica 9 T apthomasi 1 17 -0.2 -0.2 A. fissidens d e 0 -

E. protocenica c -0.4 -0.4

-0.6 0c -0.6 E. parayirunhor

-0.8 -0.8

a new species 0D -2/ 0 0.2 0.4 0.6 0 0.2 0.4 0.6 FIG. 3. Logarithmic linear regression of M2 FIG. 2. Logarithmic linear regression of M2 area (ordinate) versus astragalar length (abscissa) area (ordinate) versus astragalar length (abscissa) of various Itaborai ungulates, with associations as of 10 condylarth species. a, Choeroclaenus turgi- hypothesized and discussed in the text, using the dunculus; b, Anisonchus sectorius; c, Chriacus gal- regression-line of figure 2 (log(M2 area) = -.77 + linae-, d, Hyopsodus walcottianus; e, Meniscothe- 1.55 log(astragalar length)) as a standard of ref- rium chamense; f, Tetraclaenodon puercensis; g, erence. Phenacodus wortmani; h, Periptychus rhabdodon; i, Arctocyon ferox; j, Phenacodus primaevus. ciduous teeth are present and the locus of certain isolated cheek teeth is difficult to de- termine. Such samples are pooled and con- single fissure. The remainder of the speci- sidered in more general terms. The relative mens, although now stored at the DNPM as abundances of dental and pedal remains are they were collected, were picked up by work- given in tables 4 and 5. Far more reliable ers and are of uncertain provenience within relative abundance comparisons would be the quarry. These collections (all at the made on the basis of minimum number of DNPM; the AMNH material was omitted individual counts, as commonly used in ar- from this part of the study) were pooled into cheological faunal analysis (e.g., Chaplin, five additional lots, based on matrix, color, 1971), but this was not possible for reasons and preservation of the bone, and date col- given above. lected. The astragali and calcanea from each Pending an urgently needed revision ofthe lot were sorted into morphological and size Itaborai fauna, soon to be undertaken in Bra- groups, counted, and measured. The dental zil, such emendations and additions as seem remains from each lot were then sorted to necessary are given below. species and where feasible were counted. Most of the dental material consists of iso- RESULTS lated cheek teeth, and it was not always possible to sort some ofthe closely similar forms NOTES ON THE FAUNA: A revised list of the (i.e., among the notoungulates), as many de- eutherians from Itaborai and their distribu- 6 AMERICAN MUSEUM NOVITATES NO. 2761

TABLE 2 Measurements (in Millimeters) of Itaborai Eutherian M2s Length Width Species OR X OR X N Condylarthra Asmithwoodwardia scotti 3.5 3.5 2.9-3.1 3.0 3 Ernestokokenia protocenica 8.1 8.1 6.9 6.9 1 Lamegoia conodonta 14.4 14.4 12.0 12.0 1 Victorlemoinea prototypica 11.4 11.4 7.2 7.2 1 Litopterna Unnamed 3.8-4.5 4.1 3.2-3.8 3.6 18 Ernestokokenia parayirunhor? 4.5-5.7 5.0 4.0-4.5 4.2 16 Anisolambda prodromusa 9.4-9.5 9.5 6.0 6.0 2 Notoungulata Itaboraitherium atavum 5.1 5.1 3.1 3.1 1 Camargomendesia pristina 4.8-5.6 5.2 3.8-4.7 4.2 12 Colbertia magellanicaa 5.9-8.0 6.9 4.1-6.0 5.0 17 Astrapotheria Tetragonostylops apthomasia 10.2-11.9 11.1 6.7-8.0 7.3 5 ?Pyrotheria Carodnia vieraia 28.8-33.5 31.2 27.2-28.2 27.7 2 aFrom Paula Conto, 1952a, 1952b.

tions among the various lots are given in table be added to this as yet hypothetical compos- 4. Asmithwoodwardia scotti, dentally the ite. The result is a sharply distinct form with most primitive known eutherian from South an enormously enlarged P4 (reminiscent of America, was at first referred by Paula Couto periptychid condylarths) and the molars (1952a) to the Hyopsodontidae, but later greatly enlarged from first to third, but this (1978b) transferred to the Didolodontidae, does not seem improbable and in fact rep- following arguments presented by McKenna resents enhancement of specialized trends (1956). It is the smallest ungulate from the known in other didolodonts such as Ernes- quarry and is extremely rare; aside from the tokokenia protocenica. Lamegoia is a rare described skull and mandible and five ad- form and is also found only in lots 2 and 3. ditional specimens in lot 2, only one other Victorlemoinea prototypica, heretofore con- specimen, from lot 3, is referred to this sidered a macraucheniid litoptern (Paula species, and that with some doubt. Ernes- Couto, 1952a, 1978b; see also Simpson, tokokenia protocenica, a medium-sized di- 1948), is tentatively excluded from that fam- dolodont, also has a limited distribution, oc- ily and order, based on pedal evidence pre- curring only in lots 2 and 3, where it is sented below. It is a rather rare form also and moderately abundant. Paula Couto (1952a) was found in lots ?1, 2, and 3. considered the isolated specimens MNRJ An undescribed species, dentally similar to 1841V and 1842V, a right P4 and left P4, species generally grouped in the Didolodon- respectively, as probably representing a new tidae but here recognized as litoptern because and distinctive member of the Didolodon- of tarsal specializations, is added to the fau- tidae. Comparison with other didolodonts, nal list. Its full description and detailed com- both from Patagonia and Itaborai, suggests parison with didolodonts and primitive li- that these pertain to Lamegoia conodonta topterns are in progress. Except for Paula Couto, 1952a. Additional isolated Asmithwoodwardia scotti, this is the smallest cheek teeth among the DNPM material may ungulate from Itaborai and is the single most 1983 CIFELLI: EUTHERIAN TARSALS 7 abundant non-notoungulate species. It may TABLE 3 seem paradoxical that a species heretofore Lengths (in Millimeters) of Eutherian Tarsals unrecorded should be represented by so from Itaborai many specimens, but except for two dubious Astragalus teeth it is known only from the 1968 Cam- pos/Price material, and thus was collected Species OR X N after Paula Couto's initial studies and de- Condylarthra scriptions. Dentally, it is rather primitive and Asmithwoodwardia scotti most comparable to Ernestokokenia para- Ernestokokenia protocenica 17.3-17.8 17.6 2 yirunhor in this fauna. Both have enlarged Lamegoia conodonta lower canines, but the new form is sharply Victorlemoinea prototypica 19.5-28.0 22.9 6 distinct in that the premolars are less molar- Litoptema ized (P4 and P4 lack the metacone and para- Unnamed 5.5-8.2 7.3 37 conid, respectively; P3 is a simple, unicuspid Ernestokokenia parayirunhor 11.0-12.2 11.6 15 tooth) yet are inflated, bringing to mind a Anisolambda prodromus 14.3-19.3 16.5 6 primitive mioclaenine hyopsodontid such as Notoungulata Ellipsodon. Ernestokokenia parayirunhor, Itaboraitherium atavum traditionally placed in the Didolodontidae Camargomendesia pristina 10.0-12.2 10.9 11 (Paula Couto, 1952a) but here considered li- Colbertia magellanica 11.0-17.2 13.9 50 toptern because ofthe tarsus, is questionably present in the Campos/Price collection, mod- Astrapotheria erately abundant in lots 2 and 3, and very Tetragonostylops apthomasi 18.7-23.3 21.0 5 abundant in lots 4 and 5. ?Pyrotheria Itaboraitherium atavum, a notoungulate Carodnia vierai 51.0 51.0 1 referred originally by Paula Couto (1954) to Xenarthra the Patagonian notostylopid genus Homa- Dasypodidae, indet. 11.2 11.2 1 lostylops but later (Paula Couto, 1970) placed ?Glyptodontidae, indet. 9.2 9.2 1 in the Oldfieldthomasiidae under the new ge- neric name, is very rare and is found only in Calcaneum lot 2 and questionably in lot 3. As currently Condylarthra conceived, the species Colbertia magellanica Asmithwoodwardia scotti is enormously variable in size and morphol- Ernestokokenia protocenica 38.0 38.0 1 ogy (see Paula Couto, 1952b). Isolated and Lamegoia conodonta worn teeth may be confused with Camar- Victorlemoinea prototypica 43.2-44.0 43.6 2 gomendesia pristina (a generally smaller and Litopterna more primitive species) where both occur be- Unnamed 13.3-17.1 15.2 26 cause, as noted above, ofthe difficulty ofdis- Ernestokokenia parayirunhor 21.1-23.0 22.0 6 tinguishing deciduous from adult teeth and Anisolambda prodromus of determining tooth locus within the jaw. Notoungulata Colbertia is found in all lots except 1 and ?4, Itaboraitherium atavum and is present in great (uncounted) numbers Camargomendesia pristina 20.4-22.9 21.7 11 in lot 3, where Camargomendesia pristina Colbertia magellanica 25.7-33.8 29.5 22 probably occurs also. This last species has Astrapotheria recently been considered by Paula Couto Tetragonostylops apthomasi - - - (1979) to belong to Othnielmarshia Ame- ghino, but the relatively high crowns, reduc- ?Pyrotheria tion of upper molar ectoloph folds, well-de- Carodnia vierai veloped lower molar entolophid (particularly on M3), and strongly marked upper molar pattern with at least second crista and cro- ognized as valid. Nonetheless, it shows no chet, and first crista variably present, seem specializations characteristic of other, more to oppose reference to the Patagonian genus. derived early notoungulate families, and is Camargomendesia is thus provisionally rec- therefore retained in the Henricosborniidae. 8 AMERICAN MUSEUM NOVITATES NO. 2761

TABLE 4 Relative Abundances of Eutherian Dentitions Lot 1 Lot 2 Lot 3 Lot 4 Lot 5 Lot 6 Species No. % No. % No. % No. % No. % No. % Asmithwoodwardia scotti - - 6 2 ?1 0 Unnamed 344 58 Ernestokokenia parayirunhor ?2 0 15 4 41 12 62 78 32 86 - Anisolambda prodromus 34 6 14 4 12 4 3 4 2 5 2 22 Ernestokokenia protocenica - - 46 14 11 3 ------Lamegoia conodonta - - 5 2 5 2 Victorlemoinea prototypica ?1 0 18 5 5 2 Itaboraitherium atavum - - ?10 3 a - ?2 3 Camargomendesia pristina 185 31 ?2 1 a _ 9 11 Colbertia magellanica - - 198 59 a _ - -J 3 8 1 11 Tetragonostylops apthomasi 24 4 ?70 21 213 65 4 5 - - 6 67 Carodnia vierai - - 20 6 46 14 - - a Lot 3 contains an enormous number ofisolated notoungulate teeth which time did not permit sorting or counting.

Tetragonostylops apthomasi pertains to the so distinct that even fragments are readily Trigonostylopidae, a family given ordinal recognized.) status by Simpson (1967), but here more con- The order Xenarthra is known at Itaborai servatively retained in the Astrapotheria. It only from four dasypodid plates collected by is found in all lots except 5, and is occasion- Rosendo Pascual and described by Scillato ally (lots 2, 3) very abundant. Carodnia vi- Yane (1976) as Prostegotherium aff. astrifer. erai, placed in a new and monotypic order, No additional plates were found in the col- Xenungulata, by Paula Couto (1952a) occurs lection, but several postcranial specimens, in- in modest numbers in lots 2 and 3. (The cluding astragali of a dasypodid and a ?glyp- counts for Tetragonostylops and Carodnia todontid were encountered. It cannot be are relatively inflated because their teeth are determined whether the dasypodid astragalus

TABLE 5 Relative Abundances of Eutherian Tarsals Lot I Lot 2 Lot 3 Lot 4 Lot S Lot 6 Species No. % No. % No. % No. % No. % No. %

Asmithwoodwardia scotti - - Unnamed 155 75 Ernestokokenia parayirunhor ?5 2 - - 18 10 7 37 1 50 2 9 Anisolambda prodromus - - - - 9 5 - - - - 1 4 Ernestokokenia protocenica - - 4 24 2 1 Lamegoia conodonta - - - - 1 1 Victorlemoinea prototypica 2 1 3 18 2 1 Itaboraitherium atavum - - - - 8 5 Camargomendesia pristina 40 19 - - j141 82 1 1 58 Colbertia magellanica - - J 1 50 11 48 Tetragonostylops apthomasi 4 2 6 35 ------11 48 Carodnia vierai - - 4 24 Dasypodidae, indet. ------1 5 Glyptodontidae, indet. - - 1 4 1983 CIFELLI: EUTHERIAN TARSALS 9

TABLE 6 Lengths (in Millimeters) of Figured Specimens Species/Specimen Figure Length Carodnia vierai Calcaneum, DGM 336M 3 91.7a Astragalus, DGM 336M 3 51.2 Tetragonostylops apthomasi Calcaneum, AMNH 55384 4 36.8 Astragalus, AMNH 55385 4 23.1 Ernestokokenia parayirunhor Calcaneum, AMNH 55396 5 21.2 Astragalus, AMNH 55394 5 12.2 Ernestokokenia protocenica Calcaneum, AMNH 55390 7 37.8 Astragalus, DNPM LE443 6 17.3 ? Victorlemoinea prototypica Calcaneum, DGM 890M 7 43.5 cm Astragalus, AMNH 55393 6 27.9 FIG. 4. Left astragalus and right calcaneum ?Lamegoia conodonta (DGM 336M) of Carodnia vierai, partially re- Calcaneium, AMNH 55389 7 45.2 stored. Left column, calcaneum in dorsal (A) and distal (B) views; right column, astragalus in plantar Colbertia magellanica (C), dorsal (D), and distal (E) views. Measurements Calcaneum, DNPM LE446B 8 27.2 given in table 3C. Astragalus, AMNH 55373 8 16.9 Dasypodidae, indet. Astragalus, DNPM LE449B 9 10.1 the astragalus, so that rotation at the crurotarsal joint would have been more strictly ?Glyptodontidae, indet. orthal with respect to the midtarsal joint than Astragalus, DNPM LE449A 9 9.0 it is primitively. The superior astragalar fo- a After restoration. ramen is absent, but a deep notch at the back of the tibial trochlea is continuous with the posteriorly expanded interarticular sulcus on pertains to the species for which plates have the inferior side of the bone, suggesting the been described. course of the nerve(s) or vessel(s) which tra- versed this region. There is no separate TARSAL DESCRIPTIONS groove for the digital flexor tendons. Both lateral and medial borders ofthe tibial troch- Carodnia vierai (?Pyrotheria) lea are sharply defined. The articular facet for Figure 4 the fibula is at a high angle to the tibial troch- DGM 335M includes a left astragalus and lea and extends the entire anteroposterior the partial skeleton DGM 336M includes a length of the body, as in Arctocyon, but is left astragalus and right and left calcanea of outwardly convex, unlike the condition in this species. The astragalus was described that genus. Anteriorly, the fibular shelf is ab- briefly by Paula Couto (1952a). The body is sent and the facet occupies the dorsal half of subquadrate and is much deeper medially the lateral body wall. Posteriorly, the fibular than laterally. The tibial trochlea presents a facet extends to the base of the body and is nearly flat surface mediolaterally and is therefore continuous with the posterolateral broadly convex anteroposteriorly. Unlike portion ofthe ectal facet. Although direct evi- Arctocyon, the tibial trochlea is subparallel dence is lacking (this region is not well pre- with rather than oblique to the long axis of served on the calcaneum), this implies that a 10 AMERICAN MUSEUM NOVITATES NO. 2761 modified, proboscidean type of calcaneofib- have been distal to that ofthe calcaneum and ular contact was present, at least posteriorly. cuboid, so that astragalocuboid contact, if In graviportal astrapotheres and uintatheres present, would have been oriented at a high where calcaneofibular contact is known to be angle to the head of the astragalus. Since no lacking, the fibular facet on the astragalus such facet is present, it seems probable that does not extend so clearly to the base of the an astragalocuboid articulationr was absent bone joining the ectal facet. The medial wall (serial tarsus). On the ventral side of the as- of the astragalus also presents a high angle to tragalus, the sustentacular and ectal facets are the tibial trochlea. The head and posterior elongate, broad, parallel, and nearly flat. portion of the astragalar body are extended They are separated by a deep interarticular medially. This, together with the fact that the sulcus. The sustentacular facet extends pos- low, unextended tibial trochlea describes a teriorly from the navicular facet, with which relatively narrow arc, indicates that the range it shares broad contact, and is somewhat con- of extension and flexion at the crurotarsal cave at its posterior extremity, where artic- joint was somewhat limited. ulation on the calcaneum continued some- The astragalar neck is extremely short and what onto the posterior border of the broad, extending over most of the width of sustentacular shelf. The posteromedial angle the body. The distal surface is tremendously of the astragalus is expanded into a promi- expanded dorsoventrally, especially on the nent knob, probably for the attachment of medial side. The navicular facet occupies hypertrophied medial collateral ligaments, nearly the entire distal surface of the head which extend to the distal end of the tibia as and is very flat. The slight curvature of con- well as the distal tarsals, and thus aid in lat- vexity forms a very low angle with respect to eral stability at the crurotarsal joint. the tibial trochlea, and this, together with the The poorly preserved calcaneum is short flat, non-oblique and subparallel astragalo- and robust. Little detail is discernible in the calcaneal facets, indicates that oblique rota- articular region on the dorsal surface of the tion at the midtarsal joint and therefore ca- body. The ectal prominence is nearly flat, on pability of eversion and inversion were a level with that of the sustentaculum, and limited or absent. A small facet for the short laterally slopes abruptly inferiad. This prob- medial collateral ligaments is well defined, ably corresponds to the modified fibular fac- and forms a sharp angle with the navicular et, as noted above. The sustentacular facet is facet on the medial side of the head.2 A cu- flat and does not extend onto the calcaneal boid facet appears to be lacking from the as- neck to reach the cuboid facet. The neck of tragalar head (there is a slight change in cur- the calcaneum (that portion distal to the as- vature medially on the head of DGM 336M, tragalocuboid articulations) is relatively but the head is broken and somewhat dis- short. On the lateral surface ofthe calcaneum torted.3 Articulation with the calcaneum in- the peroneal tubercle, to which portions of dicates the astragalonavicular articulation to the lateral collateral ligaments and quadratus plantae muscle attached, is well marked. A ridge extending posteroinferiorly from this 2 This is the "spring ligament" (see Camp and Smith, indicates the attachment of the calcaneal 1942) facet of Szalay and Decker, 1974. When well de- metatarsal V ligament and possibly the ab- fined (with a correspondingly well-developed area for ductor digiti quinti muscle. The distal end of origin of the short medial collateral ligaments on the the calcaneum bears a relatively small, dor- astragalus), it is often associated with a meniscus or soventrally compressed, somewhat medioin- sesamoidal ossification in the ligament. This ossification feriorly facing cuboid facet. Inferior and is often termed the "tibiale" (Matthew, 1937), but as somewhat medial to the cuboid facet is a Schaeffer (1947) noted, homology with the reptilian ele- large, rugose prominence. This would have ment has not been established and is improbable. 3 Paula Couto (1 952a, p. 386) supposed that a cuboid given origin to plantar ligaments attaching to facet was present on the astragalar head of Carodnia, the cuboid, cuneiforms, and tarsal fibrocar- but the feature he describes is medially placed and must tilage, which forms the deep surface of the be for the medial collateral ligaments. The cuboid bor- canalis tarsi containing the digital flexor ten- ders the astragalus on its lateral side. dons. 1983 CIFELLI: EUTHERIAN TARSALS 11

Tetragonostylops apthomasi C (Astrapotheria) A Figure 5

These astragali and calcanea are assigned to Tetragonostylops apthomasi because they I! A are of appropriate size (fig. 3) and because they are abundant where teeth of this species D are abundant (tables 4 and 5). Tetragono- stylops tarsals are conspicuously absent from lot 3, where numerous dental remains of it occur. Perhaps the five AMNH specimens were culled from this lot, as preservation and B coloration of the bone is similar. The astragalar body is low and subquadrate. The tibial trochlea, very slightly concave E in its center, has a relatively low anteroposterior curvature, indicating a modest arc of I flexion-extension at the crurotarsal joint. A cm prominent superior astragalar foramen is FIG. 5. Left astragalus (AMNH 55385) and present and a widening groove extends pos- calcaneum (AMNH 55384) of Tetragonostylops teromedially from it, dividing the tibial apthomasi. Left column, calcaneum in dorsal (A) trochlea and terminating in a well-marked and distal (B) views; right column, astragalus in groove for tendons of the deep digital flexor plantar (C), dorsal (D), and distal (E) views. Mea- muscle. A groove for these flexor tendons is surements given in table 6. seen also on the medial side of the calcaneal body, inferior and just proximal to the sustentaculum. The fibular border of the tibial trochlea is sharply defined and is somewhat row mediolaterally, and bears three distinct higher than the medial border. The facet for facets. The medial facet, for the medial col- the fibula, not quite vertical, occupies the en- lateral ligaments, is large and convexity along tire lateral side of the astragalar body and its major axis is vertical. The navicular facet slopes onto a very prominent fibular shelf. is rather flat and its major convexity (implied Juvenile astragali (determined by their small- direction of movement) forms a relatively er size, incomplete fusion ofthe head to neck, low angle with respect to the tibial trochlea. and more porous bone) indicate that devel- The navicular facet is continuous laterally opment of this shelf occurs ontogenetically, with the cuboid facet, which is large and well as it is little marked in young individuals. defined (alternating tarsus), and inferiorly The tibial trochlea extends up the neck ofthe with the sustentacular facet. The sustentac- astragalus. The medial crest of the tibial ular facet is large; posteriorly, it is nearly on trochlea is ill-defined and slopes medially the same plane as the digital flexor groove, into an extensive facet for the medial mal- so that, in contrast with Arctocyon, there was leolus of the tibia, indicating an increased little "overhang" and consequent articula- weight supporting area for the tibia. Inferior tion with the posterior margin of the calca- to the medial malleolar facet, the body ofthe neal sustentaculum. The ectal facet is broad, astragalus projects mediad. From below, this moderately concave, and faces more strictly projection is seen as a rugose, sweeping crest inferiorly than in Arctocyon, in which it faces that extends nearly the entire length of the somewhat laterally. Anteriorly, a secondary astragalus. It is evident that, as in Carodnia, facet continuous with the ectal is developed the short medial collateral ligaments were on the anterior face of the fibular shelf and very well developed. lateral wall of the astragalar body. This un- The astragalar neck is short; the head is usual articulation, seen also in Astraponotus, dorsoventrally expanded but somewhat nar- corresponds to a facet on the dorsal surface 12 AMERICAN MUSEUM NOVITATES NO. 2761 of the calcaneal neck, adjoining the cuboid though certain aspects ofthe morphology are facet. This locking mechanism would appear strikingly similar to artiodactyls, rodents, to have limited the possibility of pedal in- and rabbits. version at the midtarsal joint. Comparison of proterotheriid and mac- No complete calcanea are available, but raucheniid tarsals known by direct associa- judged from AMNH 55384, which has the tion with a hypothesized primitive morpho- dorsal portion of the tuber preserved, it was type as exemplified by Arctocyon indicates short and very robust. The astragalocalcaneal the following features of the astragalus and articulations are placed relatively far for- calcaneum to be derived for the Litopterna: ward. The most obvious feature of the cal- astragalar body spool-like, lacking the supe- caneum is the enormous development of the rior astragalar foramen, with salient, sub- peroneal tubercle into a massive, rugose, in- equal tibial and fibular crests, and with troch- feriorly directed shelf. The short fibular col- lea extending far posteroinferiorly (digital lateral ligaments and the mm. quadratus flexor groove not distinct); astragalar head plantae and extensor digitorum brevis (tarsal more or less semicylindrical, with major axis head ofthe flexor digitorum profundus) must subparallel to that of the tibial trochlea; na- have been strongly developed, indicating a vicular facet extending onto superior and in- reduced capacity for pedal inversion but ferior surfaces ofthe head, cuboid and medial powerful flexion capability ofthe ungual pha- collateral ligament facets lacking; astragalar langes. Anterolaterally on the peroneal shelf sustentacular facet anteroposteriorly elon- a well-marked groove for the m. abductor gate, ectal facet deeply concave and with pro- digiti quinti is present. nounced outward orientation; neck of cal- The calcaneal ectal facet is mildly convex caneum elongate, with strongly oblique and short. Unlike Arctocyon, it is rather ver- calcaneocuboid facet; sustentacular facet of tically oriented with respect to the body of calcaneum anteroposteriorly concave and the calcaneum; distally it faces nearly directly following same curvature as anterior part of anteriad, whereas its proximal portion faces ectal facet. nearly mediad. It is continuous with the well- Although this composite morphology is defined fibular facet, which is oriented at based primarily on Santacrucian (early Mio- about 45 degrees to the body of the calca- cene) forms, it is shared by several Itaborai neum and is strongly convex about both ma- species. The Itaborai litoptern tarsals are all jor and minor axes. The calcaneal sustentac- very similar and fall into three or perhaps ulum is short and bears a facet which is more size categories. The smallest tarsals are oriented anterodorsally. The sustentacular entirely restricted to the 1968 collection, facet achieves broad contact with the cuboid where they are found in great abundance (ta- facet. This latter is nearly vertical and has ble 5). This distribution coincides with that only a slight mediolateral concavity and me- of dental remains of the unnamed condy- dial orientation, again indicating a poor abil- larth-like species, also extremely abundant in ity to invert and evert the tarsus. The inferior lot 1 and not surely found elsewhere. Figure part of the tuber calcis, where the superficial 3 indicates further that these tarsals are of digital flexor takes partial insertion, appears appropriate size for the dental material and, to have been strongly projecting. The distal moreover, that they are too small to pertain inferior protuberance for origin of calcaneo- to any other known dental taxon from Ita- tarsal and tarsal fibrocartilage ligaments is borai, except for the very small and rare very poorly developed. Asmithwoodwardia scotti, which is restricted to lot 2 and possibly lot 3. Slightly larger litoptem tarsals, found in most samples but LITOPTERNA particularly abundant in lots 3 and 4, are of Figure 6 appropriate distribution and size only for Er- As with other advanced ungulate orders, nestokokenia parayirunhor, which is dentally litoptem proximal tarsals are readily recog- very close to the above species. nized as such and cannot be confused with There is some uncertainty as to the referral those of any other order of mammals, al- of the larger litoptem tarsals. These are very 1 983 CIFELLI: EUTHERIAN TARSALS 13

A C D

1 4P.

.: aI i. .. /1 !i ' ;_ f.. E

1.8

B *i.J

.. F

FIG. 6. Right calcaneum (AMNH 55396) and astragalus (AMNH 55394) of Ernestokokenia parayirunhor. Calcaneum shown in dorsal (A), distal (B), and lateral (C) views; astragalus shown in plantar (D), dorsal (E), and distal (F) views. Measurements given in table 6.

rare, generally fragmentary, and quite vari- log ofthe observed size range in each sample able in size (table 3). They could represent was calculated and divided by the sample either more than one species or one variable mean (an otherwise preferable basis of com- species, with the sample including juveniles. parison, standard deviation, was not used be- To assess these possibilities, the size range of cause the sample of large astragali is very the large tarsals was compared to that of the small). The natural log of OR/X in the small smallest litoptern, which may be assumed to species including juveniles is .163, and that be a single species and for which a large sam- of the large astragali is .130; it is therefore ple is available. To obtain a comparable mea- possible that only one species is represented, surement ofsize, astragalar length ofthe frag- the sample including immature individuals. mentary large specimens was estimated by As shown in figure 3, the log mean astragalar multiplication of neck width or body width length ofthis sample (including presumedju- by constants averaged from complete speci- veniles), .2175, is most appropriate for Er- mens. (Isometry is assumed: logarithmic nestokokenia protocenica, but the log mean regression is impractical because ofthe small of the two largest astragali is appropriate for number of complete specimens. The differ- either E. protocenica or Anisolambda pro- ence between the two methods is relatively dromus. These tarsals are referred to A. pro- slight unless the size discrepancies are large.) dromus on the following grounds: To account for differences in size, the natural (1) They are ofappropriate size, ifthe sam- 14 AMERICAN MUSEUM NOVITATES NO. 2761 ple represents one species and the fact that neck is relatively long. The head, relatively some specimens represent juveniles is taken narrower in the smaller forms than in A. into account. prodromus, presents nearly in a straight line (2) They pertain to a litoptern, and An- with respect to the astragalar body and neck, isolambda dentally appears to belong within rather than obliquely as in Arctocyon. Cuboid that group, whereas E. protocenica may or and medial collateral ligament facets are lack- may not. ing. The shape ofthe navicular facet describes (3) E. protocenica is absent from lot 6, an acute cone (i.e., is subcylindrical), with the where Anisolambda and a large astragalus oc- base or large diameter placed laterally, and cur. is considerably expanded onto the neck, both (4) "Condylarth" tarsals, averaging some- dorsally (lateral) and inferiorly (medial). The what smaller than the largest ofthese, are also axis of rotation appears to be nearly parallel known and are ofappropriate size for E. pro- to that at the crurotarsal joint. On the dorsal tocenica. surface of the neck a variably developed Each of these points considered individ- ridge, lacking in the smallest species, is pres- ually is open to question. Nonetheless, refer- ent proximal to the navicular facet. The pol- ral of the largest litoptem tarsals to one ished surface of this ridge indicates that the species, A. prodromus, does accord best with navicular contacted it during hyperextension the evidence at hand. at the midtarsal joint. The astragalar susten- The known litoptern tarsals from Itaborai tacular facet is elongate and anteriorly curves are thus assigned to three species: the unde- up the lateral side of the neck to the head. scribed form, Ernestokokenia parayirunhor, The ectal facet, well separated from the sus- and Anisolambdaprodromus. As with the no- tentacular by the interarticular sulcus, is very toungulates, the tarsals are virtually identical oblique and deeply concave. except for size, and the described and figured The astragalocalcaneal facets on the cal- morphology is that of E. parayirunhor, with caneum are located rather far posteriorly variations of the other forms as noted. (most notably in the small species), so that The astragalar body is spool-shaped, with the neck and tuber are ofapproximately equal a deep tibial trochlea sloping upward medi- length. The ectal facet is strongly convex; an- ally and laterally into salient, parallel, sub- teriorly, it faces mediodistally whereas pos- equal crests. The superior astragalar foramen teriorly it faces medioproximally. The ectal is absent. The medial crest, curved into a facet variably extends also onto the adjacent slightly tighter arc than the fibular, is inter- part of the calcaneal body. The fibular facet rupted anteriorly by a clearly defined groove is well developed, somewhat oblique to the coursing obliquely across it. The biological long axis ofthe calcaneum, and strongly con- significance ofthis structure is unknown (pos- vex anteroposteriorly but relatively flat trans- sibly it marks the attachment of an extensor versely. It extends anteriorly to the base of retinaculum); it is absent in the Santacrucian the ectal prominence at its junction with the genus Thoatherium and faint to absent in the neck, unlike Arctocyon. The neck anterior to Deseadan genus Protheosodon. The medial the ectal protuberance is relatively flat and is wall of the astragalar body is flat, and the defined medially by a salient ridge, to which proximal and distal ends of the medial mal- the extensor retinaculum, extensor digitorum leolar facet flare inward inferiorly, clearly de- brevis muscle, and astragalocalcaneal liga- fining the extent of the articular surface. The ment may have attached. The "beak" faintly lateral wall ofthe astragalar body, also nearly developed in Arctocyon and prominent in the vertical, is occupied by a broad fibular facet. South American condylarths described below The fibular shelf is prominent and is distinc- is lacking. Medially, the sustentacular shelf tive in being nearly horizontal. Posteriorly, and facet are somewhat extended anteropos- the articular surface for the tibia extends in- teriorly. The sustentacular facet is angled and feriorly to a considerable distance on both convex, so that it follows nearly the same medial and lateral crests, and the distinction curvature as the anterior part ofthe ectal fac- between it and the groove for the digital flexor et, although this is not nearly so well shown tendons is difficult to make. The astragalar as in later litoptems. A narrow articular strip 1 983 CIFELLI: EUTHERIAN TARSALS 15 runs anterodorsally from the sustentacular A D facet to the dorsomedial angle of the cuboid facet at the end ofthe calcaneum. The cuboid facet is L-shaped, very markedly oblique, and Ile. concave, facing anteromedially. It is bor- 4. dered laterally by a moderately developed peroneal tubercle, which distally carries an ill-defined groove for the abductor digiti quinti muscle and which continues proxi- mally to about the level ofthe posterior mar- E gin ofthe ectal protuberance. On the inferior B surface of the neck, the prominence for the calcaneotarsal ligaments is weakly developed and continues proximomedially as a ridge II along the anteroinferior margin of the sustentaculum.

Ernestokokenia protocenica (Condylarthra) F Figures 7, 8 C This species is dentally rather rare, occur- .' ring only in lots 2 and 3, and is moderately ;, abundant only in the latter ofthese. As noted .!,I elsewhere, it is possible that tarsals here !'. ascribed to Anisolambda prodromus actually cm cm pertain to E. protocenica. There is at present no means to resolve this problem definitive- FIG. 7. Right astragalus of ? Victorlemoinea ly, but the tarsals described below are attrib- prototypica (AMNH 55393; A, B, C) and left as- uted to E. protocenica because, as with the tragalus of Ernestokokenia protocenica (DNPM LE443: D, E, F), in plantar (A, D), dorsal (B, E), teeth, they are rarer than remains of A. pro- and distal (C, F) views. Measurements given in dromus and are slightly smaller than tarsals table 6. ascribed to that species, and because the dentition of E. protocenica is rather primitive, whereas that of A. prodromus is more de- monstrably litoptern in character (Paula malleolar facet extends far anteriorly onto the Couto, 1952a: see also Simpson, 1948). It neck and nearly to the head of the astragalus, should be noted also that these two tarsal where it curves abruptly mediad. The lateral morphs are appropriate in size and mor- wall of the astragalar body, also vertical, is phology only for the two mentioned species covered by an extensive fibular facet that ter- among Itaborai dental taxa. The possibility minates anteroinferiorly in a well-developed remains that one or both of the tarsal cate- fibular shelf. The head is relatively somewhat gories pertain(s) to (a) species unknown den- narrower transversely but is deeper than in tally and therefore that tarsals of either or Arctocyon, and bears a navicular facet that is both ofthese species are also as yet unknown, obliquely oriented, as in that genus. Medial but this does seem really unlikely. collateral ligament and cuboid facets are ab- The astragalar body is relatively deep and sent. A small supplementary facet continuous bears a moderately grooved tibial trochlea with that for the navicular but not with the with well-defined medial and lateral crests. sustentacular is present, and in life contacted The astragalar foramen persits in one smaller the dorsolateral neck of the calcaneum, as and presumably juvenile specimen, but is articulation of the two tarsals demonstrates. filled with cancellous bone and virtually The sustentacular facet is expanded distally obliterated in a larger specimen. The medial and is broadly continuous with the navicular 16 AMERICAN MUSEUM NOVITATES NO. 2761

c A B Aopl;-4

V.. -m,i .

. 1' ( K-xL

B C

cm 4s cm .1". h.oo FIG. 9. Right calcaneum of ? Victorlemoinea FIG. 8. Left calcaneum ofErnestokokenia pro- prototypica (DGM 890M) in lateral (A), dorsal (B), tocenica (AMNH 55390) in dorsal (A), distal (B), and distal (C) views. Measurements given in and lateral (C) views. Measurements given in table 6. table 6. facet. The interarticular sulcus is deep and for the lateral side of the astragalar head de- the ectal facet is approximately as in Arcto- scends from the apex of the prominence cyon in size and orientation. Posteriorly, the down its medial surface. The cuboid facet.is groove for the digital flexor tendons is well dorsoventrally elongate and somewhat con- marked and is somewhat offset from that of cave in that direction. Its major axis, follow- the posteroinferior margin ofthe tibial troch- ing the curvature ofconcavity, is oblique; the lea, as in Arctocyon. facet is transversely narrowed because, unlike The ectal protuberance on the dorsal sur- Arctocyon, the articular surface does not ex- face of the calcaneum is prominent and is tend to the inferomedial angle of the calca- situated farther posteriorly on the body than neal head. Inferior to the cuboid facet, the in Arctocyon. The articular surface for the tubercle for attachment of the calcaneotarsal fibula is well developed and strongly convex and tarsal fibrocartilage ligaments is mod- anteroposteriorly but relatively flat trans- estly developed. On the distolateral surface versely; an extensive articular surface is also of the calcaneum the peroneal tubercle is ro- developed on the lateral surface of the pro- bust but not large; the crest extending proxi- tuberance. The calcaneal ectal facet is broad mally from it terminates at the base of the and obliquely oriented with respect to the ectal prominence rather than passing inferior tuber, about as in Arctocyon. The sustentac- to it as in Arctocyon. ulum bears a rather small, ovoid facet which does not attain contact with the cuboid facet ? Victorlemoinea prototypica at the end of the calcaneum. At its anterior (Condylarthra) extremity, the calcaneum presents a very sa- Figures 7, 9 lient dorsally projecting beak. The dorsal sur- In addition to the tarsals described above, face of this prominence is rugose, and prob- the Itaborai collections include an enigmatic ably gave origin to the distal extensor assortment of relatively large ungulate as- retinaculum and the anterior astragalocal- tragali and calcanea which cannot be assigned caneal ligament. A narrow articular surface to known species without some (not great) 1983 CIFELLI: EUTHERIAN TARSALS 17 doubt. The remaining dental species for whichtarsalshavenotbeen assigned, Victorle- A moinea prototypica and Lamegoia conodonta, are large and correspondingly rare ungulates. There are no known litoptern tarsals, even those of a juvenile, large enough to pertain to either species. These large tarsals are rare, poorly preserved, and quite variable in size and morphology, and in the present lack of a large sample they cannot even be sorted with absolute confidence into phena representing the corresponding biological species: the two morphs described here are valid and quite distinct from each other, but it is possible that another species, unknown dentally, B is included in the sample of tarsals referred cm to Victorlemoinea prototypica. It does seem virtually certain, however, that one, or more likely both, of the species V. prototypica and L. conodonta are included in the two categories presented. FIG. 10. Left calcaneum of ?Lamegoia cono- The astragali grouped under the species V. donta (AMNH 55389) in dorsal (A), distal (B), and prototypica vary enormously in size and mor- lateral (C) views. Measurements given in table 6. phology; nonetheless, in all basic and im- portant features they are similar, and from the series available, most differences appear bears three facets. The medial collateral lig- to be size related-presumably reflecting on- ament facet is very broad, dorsoventrally and togenetic changes in proportions-or are due proximodistally convex, and meets the na- to differences in preservation. The observed vicular facet at a relatively high angle. The size range (adjusted) in a sample of six as- navicular facet, restricted to the median por- tragali, calculated as a natural log, is .163, tion of the head, is oriented much more ver- exactly that ofthe smallest Itaborai litoptem tically with respect to the astragalar body species, and it is therefore reasonable that a than in Arctocyon. The cuboid facet, occu- single species is represented. The log mean pying the inferolateral portion of the head, is length ofthese six specimens is somewhat less broadly continuous posteroinferiorly with than the predicted value for V. prototypica the sustentacular facet. Part of this facet was (fig. 3), but the largest is very close to that for contact with the neck and medial side of value. the calcanealdneck, but articulation of appro- The astragalar body is ofcomparable depth priately sized astragali and calcanea indicates to that of Arctocyon. The tibial trochlea is that astragalocuboid contact was also well de- very shallow, is posteriorly interrupted by an veloped (alternating tarsus). The sustentac- unreduced superior astragalar foramen, and ular facet is anteroposteriorly elongate and, is bordered by sharp lateral and rounded me- as noted, achieves very broad contact with dial crests. The medial wall is relatively ver- the navicular facet. The astragalar ectal facet tical, and the facet for the medial malleolus is somewhat transversely broader than in of the tibia is broad and extends well up the Arctocyon, but its moderate convexity and neck (nearly to the head in juvenile speci- outward presentation are as in that form. The mens) of the astragalus. The fibular facet on ectal and sustentacular facets are separated the lateral side of the astragalus is extensive by a deep interarticular sulcus. The groove and vertical, and a prominent fibular shelf is for the digital flexor tendons at the poste- developed anteroinferiorly. The neck is mod- roinferior margin of the astragalus is broad erately long. The head is transversely narrow and sharply distinct. but deep, both medially and laterally, and The ectal protuberance of the calcaneum 18 AMERICAN MUSEUM NOVITATES NO. 2761 is similar to that of Arctocyon; the calcaneal somedially, as in Arctocyon. The astragalar ectal facet is broadly convex anteroposte- sustentacular facet is small and rounded, and riorly and faces anteromedially. The fibular does not attain contact anteriorly with the facet is transversely broad and well devel- articular surface(s) ofthe head. A cuboid facet oped. The sustentacular shelf is relatively is probably lacking. deep; the sustentacular facet extends ante- The calcaneum certainly pertains to a ju- riorly to achieve broad contact with the cal- venile because the epiphysis of the tuber calcaneal cuboid contact. As in Ernestokokenia cis is unfused and lacking; nonetheless, the protocenica, the anterodorsal portion of the specimen is comparable in size to adult cal- calcaneal neck is prolonged dorsally into a canea referred above to ? Victorlemoinea pro- distinct protuberance; the cuboid facet, ex- totypica. The bone is porous and eroded, as tending up the distal face of this eminence, is common in immature tarsals, but in known is also very similar to that of E. protocenica features it is similar to that ofE. protocenica. except that it extends medially also, forming The calcaneal ectal facet bears a similar ori- an inverted V-shape. The tuberosity inferior entation and placement to that of Arctocyon, to the cuboid facet, for attachment of the but is somewhat less convex; the region of calcaneotarsal and tarsal fibrocartilage liga- the fibular facet is eroded, but contact be- ments, is modestly developed. The peroneal tween these two elements appears to have tubercle is rather enlarged and rugose, sug- been present. The sustentaculum is moderate gesting well-developed lateral collateral lig- in depth, unlike that ascribed to ? V. proto- aments and the quadratus plantae muscle. typica, and bears a relatively small, round facet which does not extend to the cuboid ?Lamegoia conodonta articulation (reflected in the isolated nature (Condylarthra) of the astragalar sustentacular facet, as de- Figure 10 scribed above). The dorsal surface ofthe calcaneal neck, as in E. protocenica and ? V. pro- Tarsals ascribed to Lamegoia conodonta totypica, is prolonged into a conspicuous are the rarest among Itaborai ungulates: one beak (a supplementary facet for the astragalus astragalus from DNPM lot 3 and one calca- is present on its medial border); the cuboid neum in the AMNH collections are referred facet occupies its distal surface and is thus here. The astragalus (not figured) is very dorsoventrally expanded. As in E. protocen- poorly preserved and no measurements could ica, this articular surface does not occupy the be reliably estimated, although it is as large inferomedial angle of the calcaneal head. In- as the largest referred to ? Victorlemoinea pro- feriorly, the protuberance for attachment of totypica, above. This is less than the expected calcaneotarsal and tarsal fibrocartilage liga- size ofL. conodonta (fig. 2B), unless the spec- ments is modestly developed. The peroneal imen pertains to a juvenile. In preserved tubercle on the distolateral corner of the cal- morphology, it is very similar to that of E. caneum is robust but modestly developed protocenica and to an unassociated specimen and extends posterosuperiorly to join the in the AMNH collection of ?Didolodus from base of the fibular side of the ectal protuber- the Casamayoran ofChubut (Colhue-Huapi). ance, as in E. protocenica but unlike Arcto- The body bears a moderately deep tibial cyon and all other forms considered here. trochlea with well-defined and subequal but rounded medial and lateral crests. It cannot be determined whether an astragalar foramen NOTOUNGULATA was present. As far as can be determined the Figure 11 medial and lateral walls of the body are ver- Isolated notoungulate astragali from the tical; medially, the facet for the medial mal- Early Tertiary of South America are readily leolus of the tibia is well developed and ex- recognized as such but are not easily identi- tends anteriorly onto the neck where it flares fied as to genus or even family. Judged from sharply mediad, as in E. protocenica. The comparison of representatives of all notoun- head is broad and deep, being nearly the size gulate families (for which the tarsus is known) of the body, and is somewhat rotated dor- with a hypothetical primitive ungulate mor- 1983 CIFELLI: EUTHERIAN TARSALS 19 photype as exemplified by Protungulatum or Arctocyon, the notoungulate proximal tarsus A C a appears to be distinctive in having the following combination of features: astragalar neck relatively long and constricted, with an oblique dorsal crest and bearing a small, sub- ..L ,l spherical head; medial protuberance on the j4 ', '. astragalar body; sulcus leading posterolater- I+: ally from the superior astragalar foramen, in- terrupting continuity of the lateral surface of I the tibial trochlea with the groove for the flexor tendons; and well-developed contact of astragalar sustentacular and navicular facets (corresponding to a well-developed contact of sustentacular with cuboid facets on B the calcaneum). Such a type, with few mod- cm ifications, is present in the Pleistocene ty- pothere Mesotherium. LJ: All notoungulate tarsals from Itaborai are extremely similar, frustrating attempts to sort them morphologically. The abundant and sole notoungulate species from the 1968 col- D F lection, Camargomendesia pristina, is distinct pedally in size and morphology from A Colbertia magellanica oflot 2 when the samples are compared as a whole, but structural variations and overlapping size range (table G 3) preclude identification of isolated tarsals in the common size range where both species occur. Colbertia magellanica is extremely E variable in size (see table 2 and Paula Couto, 1 952b) and may later be found to include two species (it is possible also that it is a single, sexually dimorphic species, since plots of dental and pedal measurements indicate two size peaks). Pending revision, it seems more FIG. 1 1. Left calcaneum (DNPM LE446B) and conservative to consider the known notoun- right astragalus (AMNH 55373) of Colbertia ma- gulate astragali and calcanea from Itaborai as gellanica. Calcaneum shown in dorsal (A), distal representing two species, Camargomendesia (B), and lateral (C) views: astragalus shown in dor- pristina and Colbertia magellanica. No sal (D), distal (E), posterior (F), and plantar (G) known foot bones may be assigned confi- views. Measurements given in table 6. dently to the very small Itaboraitherium atavum because ofsmall samples ofthat species, apparent broad size range in these notoun- posteriorly and with a shallow median con- gulates generally, and difficulty of distin- cavity, presents a slight angle to the long axis guishingjuveniles. The described and figured of the neck; the lateral border is higher than morphology is that referred to Colbertia ma- the medial and is moderately sharp. The lat- gellanica, but is also applicable to Camar- eral wall of the body is vertical and bears a gomendesia pristina, with differences as not- fibular shelf; posteriorly, the fibular facet ed. curves dorsad and reaches to the posterior In dorsal profile and in depth, the astraga- margin of the lateral tibial trochlea, subad- lar body is comparable to that of Arctocyon. jacent to the superior astragalar foramen. Be- The tibial trochlea, mildly convex antero- tween this foramen and the groove for the 20 AMERICAN MUSEUM NOVITATES NO. 2761 digital flexor tendons, a sulcus courses lat- facet is strongly convex along both major and erally to the fibular wall of the astragalar minor axes; the major axis has an orientation body. This evidently housed the nerve(s) or similar to that of Arctocyon, that is, oblique vessel(s) which passed through the astragalar to the axis of the crurotarsal joint. At its lat- foramen, and protected them from defor- eral margin, the navicular facet is broadly mation by the weight of the tibia during ex- continuous inferiorly with the astragalar sus- treme flexion or by the flexor tendons during tentacular facet, a continuity also well shown extreme extension (dorsiflexion) at the cru- on the calcaneum. The astragalocalcaneal rotarsal joint, and is unique to notoungulates facets do not appear otherwise to depart from among the groups studied. The medial border the hypothesized primitive condition. The of the tibial trochlea is not sharp, and the ectal facet is obliquely oriented, has moderate articular surface slopes medially to a mod- curvature and faces somewhat outward, and erate extent. Inferior to this, the body projects is separated from the sustentacular facet by medially to a well-marked protuberance a deep interarticular sulcus. The sustentac- which is approximately at the level of the ular facet occupies the inferior surface of the middle of the tibial trochlea. This is analo- astragalar neck, is slightly concave along that gous to the condition in Tetragonostylops and axis, and does not continue posteriorly to Carodnia, as described elsewhere, and as in the prominence marking the digital flexor those forms suggests well-developed medial groove, which is very distinct. short collateral ligaments, which would re- The calcaneum is rather similar to that of strict eversive movement at the crurotarsal Arctocyon. The ectal facet is broad, gently joint. The astragalar neck is constricted and convex, and laterally borders an obliquely moderately long (longer in Camargomen- oriented fibular facet. The sustentaculum, desia pristina), and bears a salient oblique whose anterior margin is at the level of the crest for attachment ofthe astragalocalcaneal posterior part of the cuboid facet, bears an ligament. The head, relatively smaller in C. oval facet which extends anterolaterally up pristina, is more expanded dorsoventrally on the neck of the calcaneum. A protuberance, its medial side but less broad transversely variable in development, marks the lateral than in Arctocyon, giving it a more spherical termination of the sustentacular facet, and appearance. The medial collateral ligament probably indicates the origin of the extensor facet is large but not sharply distinct from the digitorum brevis muscle and the extensor navicular facet; there is no facet for the cu- retinaculum. The cuboid facet is shallower, boid. Articulation with appropriately sized broader, and more obliquely oriented with calcanea indicates also that astragalocuboid respect to the long axis ofthe calcaneum than contact would have been slight, or more in Arctocyon. The inferior protuberance for probably lacking, so that the tarsus would origin of the calcaneocuboid and calcaneo- have been functionally serial.4 The navicular cuneiform ligaments is moderately developed. On the medial side of the body, a faint 4 Among primitive feet, the distinction between a "se- groove and ridge marking the course of the rial" and "alternating" tarsus seems to be in many cases deep digital flexor tendon(s) is visible, just slight and arbitrary; with isolated specimens it is often posteroinferior to the sustentaculum. The difficult to distinguish the two conditions. Early workers peroneal tubercle, on the fibular side of the (e.g., Cope, 1884; Osborn, 1889), who considered the calcaneum, is rather variable but generally serial condition to be primitive based on its presence in moderate in development. Phenacodus (see Matthew, 1897; Simpson, 1937), employed this distinction in the erection of higher classi- XENARTHRA fications. In this sense, where strongly modified higher The only previously described xenarthran groups ofmammals are concerned, it seems to be a useful material from Itaborai consists of four in- character and to reflect significant differences among adaptively similar animals, as shown by Osbom, 1889. complete plates referred by Scillato Yane Among ponderous, graviportal ungulates ("Amblypo- (1976) to the Dasypodidae as Prostegothe- da"), for instance, arsinoitheres and proboscideans have rium aff. astrifer Ameghino. The earliest pre- a completely serial tarsus, while that of uintatheres and viously known tarsae material also pertain- astrapotheres is very strongly alternating. ing to a dasypodid, Utaetus buccatus, derives 19831CIFELLI: EUTHERIAN TARSALS 21 from the Casamayoran of Patagonia and was basic dasypodid pattern is largely preserved described by Simpson, 1948. in the tamandua (Vermilingua), which also There is no genral agreement as to the possesses a distinctive pilosan characteristic: grouping of the major types of South Amer- the navicular facet on the astragalus, medially ican edentates (Xenarthra). Simpson (1945) and inferiorly convex, laterally presents a followed Flower (1882) in uniting sloths and bowl-like concavity.5 Anteaters possess a anteaters (Pilosa, sensu lato) to the exclusion number ofother derived features in common of armadillos and allies (Cingulata), but Pat- with sloths (Flower, 1882), and if such a re- terson and Pascual (1972) considered the ant- lationship is accepted, as it is here, common eaters (Vermilingua) as a separate group of derivation from an armadilloid type is im- equal rank to the other two, and not partic- plied. ularly allied with either. Comparison of as- Two astragali referable to the Xenarthra tragali of representatives of each of these are known from the DNPM Itaborai collec- major groups with that of a primitive mor- tions. They are clearly distinct from each oth- photype such as Protungulatum or Arctocyon er, but neither can be referred confidently to indicates the cingulate tarsus to be the most a known species or genus because of a con- primitive of these major groups (see below), tinuing lack of knowledge about early Xe- and that of the Dasypodidae to be primitive narthra. It seems likely that at least one per- within the Cingulata. As represented by the tains to the Dasypodidae, although reference living Dasypus novemcinctus, the astragalar to this family is reallyfaut de mieux because, body is low, broad, and anteroposteriorly as noted above, this design is considered shortened; the tibial trochlea is broad and primitive for the order. The other astragalus, lacks the superior astragalar foramen (prob- possibly also dasypodid, resembles glypto- ably a derived condition; see below). The me- dontids in certain features and is listed in that dial trochlear border is defined by a relatively family with some doubt. sharp crest and the medial wall of the astragalar body is vertical; the fibular crest, also DASYPODIDAE, GENUS AND SPECIES well marked but less so than the medial crest, UNDETERMINED is anteroposteriorly more elongate than the Figure 12 latter. The fibular wall is also vertical and a The body of this left astragalus is low, fibular shelf is variably developed among transversely broad, and anteroposteriorly modem species. The groove for the digital shortened. A vestige ofthe superior astragalar flexor tendons is not distinct from the pos- foramen, filled with cancellous bone, is pres- teroinferior margin ofthe tibial trochlea. The ent posterolaterally on the tibial trochlea, astragalar neck is transversely broad but shal- which is moderately deep. As in Utaetus, the low; the head, also shallow, bears a single medial crest is not so sharply defined as in facet (navicular) which in orientation and Dasypus or other known members of the shape is similar to that of Arctocyon. The family. Both medial and lateral walls of the sustentacular facet is relatively small and astragalar body are vertical, and a small fib- obliquely oriented with respect to the astragalar neck (probably derived); laterally it is separated by a very shallow interarticular 5 This unusual feature is present also in Manis, whose sulcus from the ectal facet, which is deeply astragalus is similar in other respects to that of Pro- concave and transversely broad. Many of schizmotherium, a megalonychid from the late Oligocene these features were indicated by Simpson of Patagonia (among sloths, the astragalus of megalo- (1931, p. 356) as characterizing the primitive nychids appears to be the most primitive). Thus, ifman- ids are related to the Xenarthra, as workers prior to about xernarthran astragalus. 1900 generally believed, then they would belong specif- The variations in the pilosan (sensu stricto) ically with the sloths. However, as detailed further below, astragalus are legion and extreme, and it is the astragalus of early fossil manids and palaeanodonts not immediately evident that megalonychid, lacks this specialization, being armadilloid in appearance mylodontid, or megatheriid types are derived (Matthew, 1918; Simpson, 1931). Emry (1970) has ar- from the primitive xenarthran type as ex- gued that palaeanodonts are pholidotans and are closely emplified by the Dasypodidae. However, the related to true manids. 22 AMERICAN MUSEUM NOVITATES NO. 2761 ular shelf protrudes from the anteroinferior size and configuration to that of Utaetus and angle ofthe lateral wall. As in cingulates gen- Arctocyon, not reduced as in modem Dasy- erally, the neck is relatively wide, short, and pus. The digital flexor groove is not distinct shallow, and a well-defined crest occupies its from the posteroinferior margin of the tibial dorsodistal extremity which, as in living da- trochlea. sypodids, would have given rise to the astragalonavicular ligament. The head is rela- DISCUSSION tively narrower transversely than in Utaetus and living forms, but the conformation ofthe Characters of the tarsus have been widely navicular facet is approximately the same. recognized as having phylogenetic signifi- Unlike Utaetus, which is more similar to cance since at least the time of Cope (e.g., Protungulatum in this respect, the astragalar 1884), but wide recognition does not imply sustentacular facet is somewhat elongate, ex- wide usage. This is somewhat of an irony tending posteromedially to the angle of the because most of the existing structure of or- astragalar body and contacting the groove for dinal and supraordinal ungulate classification the digital flexor tendons. This groove is dates to Cope's time, and is based largely on broad and poorly defined, not being distinct cheiridial morphology. Simpson's (1945) su- dorsally from the posterior margin ofthe tib- perorder Paenungulata, of particular interest ial trochlea. The astragalar ectal facet is sim- here, including the orders Pantodonta, Dino- ilar to known dasypodids in being broad and certa, Pyrotheria, Proboscidea, Embrithop- sharply concave, but the oda, Hyracoidea, and Sirenia, is essentially interarticular sulcus Osborn's (1898) Amblypoda, minus the pe- separating it from the sustentacular facet is riptychid condylarths (Simpson, 1937), and slightly better developed. including the "subungulates" generally con- ?GLYPTODONTIDAE, GENUS AND SPECIES sidered to be of African origin.6 Although a UNDETERMINED review of the Paenungulata (minus the Pan- Figure 12 todonta, including Deltatherium, which ap- This astragalus is quite distinct morpho- pear dentally not to belong here but, with the logically from the above and is also much Tillodontia, seem to represent a separate Late smaller. The tibial trochlea, which lacks the Cretaceous Cimolestes-like derivate; and mi- superior astragalar foramen, is very deep and nus the Hyracoidea, which may belong here its lateral borders are strong and condyloid. but do not fit in pedally) cannot be attempted This and the fact the medial tibial articular here, it is relevant to note that paenungulates, surface does not extend anteriorly past the plus the Astrapotheria (including trigono- middle of the astragalar body-describing a stylopoids) and Carodnia, share a basically V-shape-gives the astragalus a glyptodontid similar "amblypod" tarsal pattern: the as- appearance in dorsal aspect, similar to that tragalar body bears a low tibial trochlea with of the Santacrucian Propalaeohoplophorus. a broad anteroposterior curvature and little The and walls the are or no median concavity; the tibial trochlea medial lateral of body generally extends onto the neck of the as- both vertical; anterolaterally a fibular pro- tragalus which is short and robust (or lacking tuberance is faintly developed, medially, the entirely), and bears an expanded, flattened protuberance for attachment of the abductor head; the ectal and sustentacular facets are hallucis brevis muscle and the medial collat- flattened, elongate, and subparallel; and the eral ligaments, found in glyptodontids, is astragalar body is expanded medially. Large, lacking. The astragalar neck is shallow and of broad transversely, but notably less so than specialized members these groups tend to in Utaetus and other known dasypodids. The have feet which confusingly resemble each ectal facet is broad and rather deeply con- more cave, as in dasypodids and unlike the 6 These groups, and others under consideration here, derived condition seen in advanced glypto- were reshuffled among mammalian higher categories by dontids, in which it is planar and depressed McKenna, 1975, who erected the Mirorder Meridun- inferiorly to the level of the digital flexor gulata to contain all of South America's native ungulates groove. The sustentacular facet is similar in as a monophyletic unit. 1 983 CIFELLI: EUTHERIAN TARSALS 23 other, although most can be distinguished by diagnostic contrasting specializations. In a A C general way, the amblypod tarsus seems to be associated with a graviportal habitus (Os- born, 1929). Functional interpretation is hampered by the near absence of living -( forms, but from primitive representatives of .) several of the groups in question (Tetrago- nostylops, Astrapotheria; Phenacolophus, B Embrithopoda; Mongolotherium, Dinocera- D ta) it seems evident that these basic modifications ofthe foot are not due strictly to graviportal adaptations alone, since the genera 10 cited are relatively small, but perhaps are as- W.' I sociated with pentadactyl digitigrady, which is shared by all amblypods in the sense enjoined here. The non-Neotropical ambly- cm cm pods, whether an artificial assemblage or a group, fall into two categories FIG. 12. Left astragalus ofDasypodidae, genus monophyletic and species undetermined (DNPM LE449B; A, B) characterized by different specializations of and right astragalus of?Glyptodontidae, genus and the tarsus: (1) Embrithopoda-Proboscidea; in species undetermined (DNPM LE449A; C, D) in which the calcaneofibular contact is modified plantar (A, C) and dorsal (B, D) views. Measure- and the astragalar head reduces and loses ments given in table 6. contact with the cuboid (serial tarsus), and (2) Dinocerata; in which calcaneofibular contact is lost, the medial malleolar facet of the astragalus is well developed, and the astrag- traglar medial malleolar facet, are lacking. alocuboid contact is expanded, so that the Carodnia is not primitive but differently spe- calcaneum virtually loses its weight-bearing cialized in these respects. Comparison with function (alternating tarsus). Of the South Pyrotherium, which has one of the most de- American amblypods, Carodnia and the Pyr- rived ankles ofany known amblypod, is ham- otheria are similar to the first group, whereas pered because the tarsals are not well known the Astrapotheria is similar to the second. in that genus. The astragalus of Pyrotherium Carodnia and its synonym, Ctalecarodnia, was mentioned and figured several times by once considered a possible pyrothere (Simp- Ameghino (e.g., 1897) and was described by son, 1945), was placed in its own order, Xen- Gaudry (1909) in a posthumous publication.7 ungulata, by Paula Couto, who remarked The cuboid facet, as in Carodnia, is lacking "Carodnia is so distinct from all the known and the navicular facet is flat, not curving groups of ungulates that a detailed compar- medially as in astrapotheres. There are no ative study is not necessary to disclose its features which could be considered synapo- affinities" (Paula Couto, 1952, p. 386). Later morphous with notoungulates (such a special workers (e.g., Patterson and Pascual, 1972; relationship was proposed by Patterson, Simpson, 1978) have upheld this view, which 1977). Simpson (1935b) also noted a strong merely emphasizes the autapomorphous dental resemblance between Carodnia and characters of the genus. As noted above, the Holarctic Dinocerata, or uintatheres, and many advanced conditions ofthe ankle cited various workers have commented on a spe- in the description are common among am- cial relationship between these two groups blypods. It is significant, however, that as- trapotherian specializations, such as the 7The original specimen cannot now be found in the strongly developed astragalar cuboid facet, Ameghino Collection, but AMNH 11727 is a good cast medial extension ofthe navicular/medial col- of it. AMNH 11727 includes also an amorphous and lateral ligament facets on the astragalar head, enigmatic cast of a supposed calcaneum, but this spec- and strong anteroposteriorly extended as- imen adds no knowledge. 24 AMERICAN MUSEUM NOVITATES NO. 2761

(Wheeler, 1961; McKenna, 1981). Compar- facet,8 which is prominent in Tetragonosty- ison oftarsals of Carodnia to those ofa prim- lops and astrapotheres but lacking in Car- itive uintathere such as Mongolotherium or odnia and Pyrotherium. It is possible that the Bathyopsis reveals the same suite ofcontrast- specimen belongs to the Casamayoran pying specializations as regards the astrapo- rothere, Carolozittelia. theres. The proximal part of the ankle is known Traditionally, the poorly known trigono- in Parastrapotherium and Astrapotherium stylopoids have been considered close allies (Scott, 1937), and isolated astragali from the of the Astrapotheria. However, Simpson Mustersan may be referred to Astraponotus (1967) placed the family Trigonostylopidae with little doubt. In all forms, as in Tetra- in a new order, Trigonostylopoidea, stating gonostylops, the medial malleolar facet is well that "a common ancestor (i.e., with the as- developed and medially expanded, possibly trapotheres) would almost have to be on the assuming a weight-bearing function; the as- evolutionary level of the Condylarthra" tragalocuboid contact is expanded and the (1967. p. 196). This view was recently chal- navicular facet extends medially; the calca- lenged by Carbajal et al. (1977), based on a neal peroneal tubercle (unknown in Astra- consideration of the Casamayoran genus Al- ponotus) is enlarged and calcaneofibular con- bertogaudrya, referred by Simpson (1967) to tact is lacking (unlike the condition in the Trigonostylopidae but considered by Tetragonostylops). Because Deseadan and them to be an astrapotheriid. later astrapotheres are gigantic, strongly Ameghino (1904) described and figured modified graviportal forms, comparison of several isolated astragali from the Casama- Tetragonostylops with Astraponotus is in- yoran (Notostilopense in his usage) of Pata- structive. In the latter genus, the astragalar gonia as Trigonostylops "minimus," T. wort- fibular shelf and medial collateral ligament mani, and the Trigonostylopidae, indet. facets (primitive characters) lacking in later Restudy of these specimens indicates them astrapotheres are present and, significantly, all to be notoungulate, that of T. wortmani the unusual secondary astragalar ectal facet (Ameghino, 1904, fig. 8) possibly pertaining (for articulation with the dorsal astragalar to an isotemnid. In the same paper, Ame- neck) is present. This and the derived features ghino figured an astragalus he assigned to Al- noted above strongly suggest that the Itaborai bertogaudrya unica. This astragalus, noted in species, and by implication the family Tri- the original description of the family Alber- gonostylopidae, is closely related to the As- togaudryidae but not of the genus (Ameghi- trapotheria. In many respects, Tetragono- no, 1901), is contained in the genoholotype stylops is pedally more primitive than later specimen, MACN 12000. Simpson (1967, p. astrapotheres, and in others (e.g., the enor- 231) has noted that this probably was not mously developed peroneal shelf), it is clearly associated with dental remains of that spec- autapomorphous. As noted above, the As- imen. The astragalus is poorly preserved and trapotheria present tarsal specializations enigmatic as to affinities. The tibial trochlea similar to those of the Dinocerata although is low, with little curvature, and extends onto in one feature, loss ofcalcaneofibular contact, the shortened astragalar neck; a vertical me- this seems to be convergent (such a contact dial malleolar facet is present, and what re- is present in Tetragonostylops). If the paen- mains of the astragalocalcaneal facets indi- ungulates, in the sense enjoined here repre- cates them to have been elongate, flat, and sent a monophyletic group, then two major subparallel. This is reminiscent of astrapo- clades are suggested: Astrapotheria-Dino- theres (although more advanced in all re- cerata and Pyrotheria (including Carodnia)- spects than the Mustersan Astraponotus), but Proboscidea-Embrithopoda. This is, how- the head seems to have lacked the cuboid ever, frankly speculative at this juncture, since it is only the vague "amblypod" tarsal features cited above, coupled with generalized trends (such as graviportal adaptation 8 This cannot be determined with certainty, since some and gigantism) which unites this assemblage. of the astragalar head is missing. The South American condylarths and li- 1983 CIFELLI: EUTHERIAN TARSALS 25 topterns have long been considered closely idae, probably lacked the litoptern special- related: Scott, 1913, grouped the Didolodon- izations of the tarsus, although this will not tidae ("Didolodidae") within the Litopterna, be certain until an articulated skeleton is and Simpson (1948, p. 1 19) stated that "the found. Dentally, there are some molar spe- litopterns appear to represent the direct fur- cializations (e.g., posthypocone cingulum- ther evolution of the Condylarthra, continu- hypocone-metaconule crest in the upper ing in South America after they became ex- molars) of Victorlemoinea suggestive of the tinct in the rest of the world." Because most Macraucheniidae, but the genus is aberrantly of the described species from the early Ter- specialized in some respects and far more tiary are known only by isolated teeth, it is derived than the most primitive undoubted difficult to make extensive comparisons and members of the family, the Cramaucheni- erect dental diagnoses separating these two inae, which are known primarily from the groups. A reevaluation of the interrelation- Colhuehuapian (Soria, 1981) and Santacru- ships ofthese early forms is deferred, pending cian (Scott, 1910) but also by a few specimens completion ofwork in progress. It is evident, in the Deseadan. It seems likely that Victor- however, that the problem of distinguishing lemoinea is not a litoptern, but a precociously litopterns from didolodonts is intensified by specialized relative that was convergent upon the discovery that Ernestokokenia parayi- them. Its only clear relative is Phoradiadus runhor and a new undescribed form, dentally (Simpson, Minoprio, and Patterson, 1962) among the most primitive of all known eu- from the peculiar Divisadero Largo (?early therians from South America, bear the un- Deseadan) fauna from near Mendoza, Ar- mistakable litoptern ankle specializations, gentina (Soria, 1981, and personal observ.). whereas other didolodonts (including a sup- Possibly both are referable to the Sparno- posed congener of Ernestokokenia parayi- theriodontidae, which is based on a litoptern- runhor, E. protocenica) and a dental litop- like mandible, ofvery advanced aspect, from tern, Victorlemoinea prototypica, do not. the Casamayoran ofPatagonia (Soria, 1980). This suggests that, as in artiodactyls (Schaef- Tarsals of? Victorlemoinea prototypica and fer, 1948), the primary specializations of the the remaining condylarth-like species, Er- Litopterna were pedal not dental. These spe- nestokokenia protocenica and ?Lamegoia cializations, including modifications for conodonta, seem to form a natural group strictly orthal joint movement and for rapid when compared to other South American flexion of the crurotarsal joint, as seen in the forms and with various North American con- relatively posterior position of the astraga- dylarths. In all of them, the nearly vertical localcaneal facets on the calcaenum, plus the facet for the medial malleolus of the tibia is very small body size also suggest that the well developed and expanded anteriorly, ex- most primitive litopterns may have been sal- tending well onto the neck of the astragalus, tatory rather than cursorial in their mode of and the calcaneal cuboid facet is dorsally ex- locomotion. Perhaps these adaptations, panded into a prominent beak. These fea- which are similar in many respects for the tures are not known among the litopterns, but two locomotor modes, were in a sense pre- the litoptern tarsus is so modified that it is adaptive to the later radiations of large, cur- difficult to say with conviction that such char- sorial litopterns. acters were not present in a common ancestor The Itaborai litoptem tarsals, pertaining to of the two groups. Ernestokokenia protocen- Anisolambda prodromus, Ernestokokenia ica and ?Lamegoia conodonta are especially parayirunhor, and the unnamed form, are alike in a number ofderived similarities: the closely similar and resemble most those peroneal tubercle of the calcaneum extends ascribed to Protheosodon coniferus by Loo- posterodorsally to terminate at the base of mis, 1914 (see comments, table 1). The Santa the ectal protuberance, the astragalar head is Cruz proterotheres and macraucheniids medially and transversely expanded and have, respectively, distinct ankles which may lacks a medial collateral ligament facet and both be derived from this primitive pattern. cuboid contact, the astragalar tibial trochlea Victorlemoinea, considered by Simpson is deepened, and the anterior part ofthe me- (1948) to be a member of the Macraucheni- dial malleolar facet flares sharply mediad. A 26 AMERICAN MUSEUM NOVITATES NO. 2761

similar astragalus, tentatively assigned to Di- Hegetotheriidae among the Typotheria may dolodus, is known from the Casamayoran of be characterized and grouped by tarsal spe- Patagonia. These tarsals also lack the pre- cializations away from this generalized type. sumably autapomorphous features of ? Vic- As with the litopterns and South American torlemoinea, such as the anteroposteriorly condylarths, there seems to be no North expanded sustentacular facet (with a corre- American ungulate which presents pedal spe- sponding facet on the calcaneum), thickened cializations strongly indicative of notoun- calcaneal sustentaculum, and a large rugose gulate relationships. The tarsus of the peroneal tubercle. supposed Holarctic notoungulates, the The primitive members of the great order Arctostylopidae, is unknown. The most com- Notoungulata possessed a rather generalized parable North American ungulate tarsals per- tarsus, although the combination of derived tain to the Hyopsodontidae. In Promioclae- astragalocalcaneal conditions noted above nus aequidens, AMNH 823, the calcaneal are distinctive and do make isolated tarsals sustentacular facet extends anteriorly to readily recognizable. Colbertia magellanica achieve broad contact with the cuboid facet, is currently placed in the Oldfieldthomasi- the astragalar body has a median projection idae (Paula Couto, 1 952b) which, as generally (also true of Hyopsodus walcottianus, conceived, is a wastebasket family including AMNH 14654), and the lateral side of the generally primitive typotherian (including tibial trochlea is interrupted posteriorly by a both typotheres and hegetotheres) notoun- sulcus extending laterally from the superior gulates above the notioprogonian grade astragalar foramen (possibly but not surely (Simpson, 1948). Because of the absence of also true of H. walcottianus). These derived typically typotherian specializations (cristid features are characteristic of notoungulates, obliqua of lower molars with a narrow an- but the auditory region ofHyopsodus (Cifelli, terior attachment to the metalophid: upper 1982) is far more primitive than that of any molars with crochet and second crista joined, known notoungulate (see Patterson, 1934; first crista joined to the protoloph directly or Simpson, 1948, and Pascual, Vucetich, and by an antecrochet), Colbertia and some ofthe Fernandez, 1978). other oldfieldthomasiids, such as Max- The two known xenarthran astragali from schlosseria, Brachystephanos, and Xenoste- Itaborai are distinct from each other and phanos might with equal validity be placed surely represent different genera and proba- in the Isotemnidae. Simpson (1967) indicat- bly families. Both are assigned to the Cin- ed the accessory trigonid cuspule, which I gulata, although as noted this is based on the interpret as a paraconid, to be one of the absence of typical derived pilosan characters distinguishing features of isotemnids and it rather than the presence of distinctively cin- is found in the genera listed, but it has a still gulate features. The strongly constricted tib- wider distribution and is probably primitive ial trochlea of DNPM LE449a, with salient, for the Notoungulata as a whole. The tarsus condyloid, medial and lateral crests is (are) of Colbertia does not differ greatly from that derived condition(s) similar to glyptodon- of Camargomendesia pristina, which per- tids, suggesting that this family may have tains to a truly primitive family, the Henri- been distinct, at least pedally, by the Rioch- cosborniidae (Paula Couto, 1 978c, 1979; see ican; this specimen does, however, lack other Simpson, 1948, and Pascual, Vucetich, and distinctive and presumably later derived fea- Fernandez, 1978) or from that of Boreasty- tures of the glyptodont tarsus, such as a flat- lops (Notostylopidae), ?Pleurostylodon (Iso- tened sustentacular facet. The other astrag- temnidae), or Trachytherus (Mesotheriidae). alus is generally similar to that of Utaetus, This broad distribution among the major but differs from that genus and other dasy- groups of Notoungulata indicates that it is a podids in the (vestigial) presence of the su- generalized pattern for the order. Various no- perior astragalar foramen. It is curious that toungulate families, including the Homalo- dasypodid plates, generally such a conspic- dotheriidae, Leontiniidae, Notohippidae, and uous element of South American mammal the Toxodontidae among the Toxodonta and faunas, are entirely lacking in the extensive the Interatheriidae (see Stirton, 1 953) and the DNPM Itaborai collections, whereas a cur- 1 983 CIFELLI: EUTHERIAN TARSALS 27 sory survey revealed astragali of two cingu- mode of tooth reduction; (11) broad, flat late types as well as a number of other post- transverse processes on the caudal vertebrae, cranial remains. Considering the condition of and (13) distally elongated, medially turned the bone and the abundant herpetofauna, it pectoral crest on the humerus. On this basis is highly unlikely that this is an artifact of the two groups may be tentatively allied. Pos- collecting or preservational bias. sible contradictions to this arrangement in- Matthew and Granger (1918), following a clude palaeanodont-xenarthran resem- suggestion by Osborn (1904), placed their blances in the scapula (12; secondary scapular then new North American suborder Palaeano- spine, coracoscapular notch) and dorsolum- donta in the Edentata, suggesting a close bar vertebrae (9; deepened notches for nerve relationship to both xenarthrans and pholi- exits). Derived similarities of the Pholidota/ dotans but perhaps closer to the latter. In a Palaeanodonta and Xenarthra with respect to more extensive study, Simpson (1931) indi- other mammals, dismissed by Emry but pos- cated palaeanodonts to be closer to the Xe- sibly evidence of a common unique ancestor narthra. This arrangement, which he fol- within the Eutheria are mostly features func- lowed in 1945, was widely accepted until tionally related to the eating of colonial in- Emry (1970) presented detailed comparisons sects and powerful development ofthe limbs: arguing that palaeanodonts be referred to the general reduction and simplification of the Pholidota and that "there are no features of dentition, mandible, and mandibular gle- either the palaeanodonts or xenarthrans to noid; proportions of the limbs, digits, and indicate that they are more closely related to claws; everted and strongly crested ilia, and each other than either is to the Insectivora" contact ofthe ischium with caudal vertebrae. (Emry, 1970, p. 503). This statement is based The evidence afforded by the proximal tar- in part on Emry's unwarranted assumption sals is equivocal, partially because of uncer- that any similarities between the two groups tain association of skeletal materials. Simp- are necessarily convergent because palaeano- son (1931, p. 356) noted that, whereas the donts were contemporaneous with far more astragalus of Metacheiromys is distinctively derived true xenarthrans, and may be incor- xenarthran in appearance, that of Palaeano- rect. Possession of a more primitive condi- don is more primitive. Emry (1970, p. 500), tion by the palaeanodonts does not exclude however, considered astragali of these two that condition from being synapomorphous genera to be virtually identical and equally with xenarthrans and/or pholidotans, "armadilloid." In fact, there are two distinct through a common ancestor. Comparison astragalar types among specimens currently must be made to other mammals to deter- assigned to Palaeanodon; the narrow-bodied mine the primitiveness ofthe character with- morph, exemplified by AMNH 15137 (Mat- in the Eutheria in general. As Emry observed, thew and Granger, 1918, fig. 67), and the many of the palaeanodont-xenarthran simi- Metacheiromys-like morph, exemplified by larities noted by Simpson (1931) seem to be AMNH 14733. It can be argued that the latter primitive; this applies also to a number of exhibits a number ofderived xenarthran fea- palaeanodont-pholidotan similarities, from tures-most importantly a concave astraga- the same list, discussed by Emry. Such equiv- lar ectal facet, anteroposteriorly short but ocal characters, or those for which morpho- transversely wide body with a strongly cline polarity is unclear, include (2) propor- curved, deep tibial trochlea, and loss of a tions of the skull; (3) distribution of cranial distinct groove for the digital flexor ten- foramina; (5) auditory region; (6) brain; (8), dons-the former, however, is far more cervical vertebrae; ?(9) dorsolumbar verte- primitive and is vaguely xenarthran only in brae; (10) sacrum; (14) radius and ulna; (15) the somewhat shortened neck and transverse- carpus; (18) femur; (19) crus; and (20) as- ly narrowed head. Since the more primitive tragalus (numbers in parentheses are those astragalus is associated with partial skeletons used by Simpson, 1931, p. 368ff; Emry, 1970, including the type and a paratype ofPalaeano- p. 488ff). Palaeanodonts do appear to resem- don ignavus, I assume that it pertains to that ble manids but not xenarthrans, however, in species and the more Metacheiromys-like as- several specialized characters, such as (1) tragali (all from another locality) perhaps be- 28 AMERICAN MUSEUM NOVITATES NO. 2761 long to an early, unrecognized species of the it is possibly a primitive member of the pa- latter genus. Thus, either Metacheiromys is laeanodont-pholidotan complex. more closely related to the Xenarthra than is Storch (1981) has described a remarkably Palaeanodon, or both belong to a group in complete skeleton from the middle Eocene which the astragalus independently con- of Germany as Eurotamandua joresi, refer- verged on the hypothesized primitive xenar- ring it to the Myrmecophagidae. The speci- thran condition. Numerous derived condi- men is strikingly anteater-like, but as figured tions shared by these genera, described in seems to lack a number of myrmecophagid, detail by Matthew and Granger (1918) and pilosan, and xenarthran characters (see En- Simpson (1931) make this hypothesis the gelmann, 1978). The posterior dorsal and an- more probable. terior lumbar vertebrae (Storch, 1981, fig. 8) The situation is complicated further by each have posteriorly extended transverse consideration of the pholidotan ankle. As processes, contacting the centrum of the fol- noted above, the astragalus of Manis resem- lowing vertebrae. This is an uncommon (al- bles that of primitive (e.g., megalonychid) though not unparalleled) condition among sloths, particularly in the peculiar concavo- mammals, but is rather unlike the condition convex astragalonavicular articulation. Ifthe in the South American edentates. In those association ofMiocene-tarsals from Germany forms, the postzygapophysis is completely (described by Helbing, 1938) to Necromanis locked between the prezygapophysis and Filhol, 1884 (= Teutomanis Ameghino, 1905) transverse process of the preceding vertebra. is correct (Emry, 1970), then these pilosan- Ischiocaudal contact seems to be present in manid similarities were almost surely ac- Eurotamandua (Storch, 1981, fig. 7), but it quired independently. Necromanis appears does not appear to differ significantly from to lack the astragalar pivot (but see Szalay, the condition in Manis. As in Xenarthra (and 1977, p. 349) and is more primitive in certain many other mammals) a secondary scapular respects (presence of a prominent superior spine is present and teeth are lacking, as in astragalar foramen posteriorly dividing the anteaters. The astragalus is not well shown, tibial trochlea, astragalar body not trans- but ofother details in which pilosans are spe- versely expanded and with a median projec- cialized, Eurotamandua lacks the scapular tion) than the palaeanodont-Manis assem- fenestra and reduction ofthe zygomatic arch. blage considered as a whole. An astragalus The complete lack of teeth is a similarity to nearly identical with that of Necromanis is vermilinguans, but the muzzle is not clearly known from a partial ?pholidotan skeleton so elongate nor are the ribs flattened and from the Ulan Gochu beds of the Shara Mu- overlapping as in the South American forms. run region, Mongolia. Perhaps Eurotamandua also pertains to the Mention should be made of several recent pholidotan-palaeanodont assemblage which, discoveries of possible non-South American together with the aberrant and poorly under- xenarthrans, although the ankle is not well stood epoicotheres, may represent a signifi- known in these forms. Ernanodon antelios, cant but hitherto largely undocumented ra- from the ?late Paleocene of Guangdong, diation of North American-Old World south China, has been implicated as a prim- edentates. The balance ofevidence presented itive edentate (Ding, 1979). The posterior above suggests that, in essential agreement dorsal vertebrae are described as showing with Matthew (Matthew and Granger, 1918), "incipient" xenarthrous articulations, but these forms together with the Xenarthra may they are not unambiguously xenarthrous, and form a remotely related but monophyletic similarly incipient types are found among group of primitive Eutheria. other mammals such as didelphids and palaeanodonts. The ischium ofErnanodon may have contacted but probably did not fuse LITERATURE CITED with the caudal vertebrae; the skeleton is otherwise remarkably primitive, but the manid- Ameghino, F. like pectoral crest ofthe humerus and general 1897. Les mammif-eres cretaces de l'Argen- "edentate" trend ofthe dentition suggest that tine. Deuxieme contribution a la con- 1983 CIFELLI: EUTHERIAN TARSALS 29

naissance de la faune mammalogique Acad. Brasil. Cien., vol. 50, pp. 381- des couches a Pyrotherium. Bol. Inst. 416. Geogr. Argentina, vol. 18, pp. 406-521. Gaudry, A. 1901. Notices preliminaires sur les ongules 1909. Fossiles de Patagonie: Le Pyrotherium. nouveaux des terrains cretaces de Pa- An. Paleontol. (Paris), vol. 4, pp. 1-28. tagonie. Bol. Acad. Nac. Cien. Cordoba, Gingerich, P. D. vol. 16, pp. 349-429. 1977. Correlation of tooth size and body size 1904. La perforacion astragaliana en los ma- in living hominoid primates, with a note mlferos no es un caracter originaria- on relative brain size in Aegyptopithecus mente primitivo. An. Mus. Nac. Buenos and Proconsul. Amer. Jour. Phys. An- Aires, ser. 3, vol. 4, pp. 349-460. throp., vol. 47, pp. 395-398. Camp, C. L., and N. Smith Grambast, L., L. Martinez, M. Mattauer, and L. 1942. Phylogeny and function of the digital Thaler ligaments ofthe horse. Mem. Univ. Cal- 1967. Perutherium altiplanense, nov. gen., ifornia, Berkeley, vol. 13, pp. 69-124. nov. sp., premier mammifere meso- Carbajal, E., R. Pascual, R. Pinedo, J. A. Salfity, zoique d'Amerique du Sud. C.R. Acad. and M. G. Vucetich Sci., ser. D, vol. 264, pp. 707-710. 1977. Un nuevo mamifero de la Formacion Helbing, H. Lumbrera (Grupo Salta) de la Comarca 1938. Nachweis manisartiger Saugetiere im de Carahuasi (Salta, Argentina). Edad y stratifizierten europaiischen Oligocaen. correlaciones. Publ. Mus. Munic. Cien. Eclog. Geol. Helvetiae, vol. 31, pp. 293- Natur. Mar del Plata "Lorenzo Scaglia," 303. vol. 2, pp. 148-163. Kay, R. F. Chaplin, R. E. 1975. The functional adaptations of primate 1971. The Study of Animal Bones from Ar- molar teeth. Amer. Jour. Phys. An- chaeological Sites. London, Seminar throp., vol. 43, pp. 195-216. Press, pp. 1-170. Kielan-Jaworowska, Z. Cifelli, R. L. 1977. Evolution of therian mammals in the 1982. The petrosal structure of Hyopsodus Late Cretaceous of Asia. Part II. Post- with respect to that of some other un- cranial skeleton in Kennalestes and gulates, and its phylogenetic implica- Asioryctes. Palaeontol. Polon., vol. 37, tions. Jour. Paleontology, vol. 56, pp. pp. 65-83. 795-805. Loomis, F. B. Cope, E. D. 1914. The Deseado Formation of Patagonia. 1884. The Vertebrata of the Tertiary forma- Concord, The Rumford Press, pp. 1- tions ofthe west. Repts. U.S. Geol. Sur- 232. vey Territ., vol. 3, pp. 1-1003. McKenna, M. C. Ding, S. 1956. Survival ofprimitive notoungulates and 1979. A new edentate from the Paleocene of condylarths into the Miocene ofColom- Guangdong. Vert. Palasiatica, vol. 17, bia. Amer. Jour. Sci., vol. 254, pp. 736- pp. 57-64. 743. Emry, R. J. 1975. Toward a phylogenetic classification of 1970. A North American Oligocene pangolin the Mammalia. In Luckett, W. P., and and other additions to the Pholidota. F. S. Szalay, eds., Phylogeny of the Pri- Bull. Amer. Mus. Nat. Hist., vol. 142, mates. New York, Plenum Press, pp. pp. 455-510. 21-46. Engelmann, G. F. 1981. Early History and biogeography of 1978. The logic of phylogenetic analysis and South America's extinct land mam- the phylogeny of the Xenarthra (Mam- mals. In Ciochon, R. L., and A. B. Chi- malia). Ph.D. dissertation, Columbia arelli, eds., Evolutionary Biology of the Univ. New World Monkeys and Continental Flower, W. H. Drift. New York, Plenum Press, pp. 43- 1882. On the mutual affinities of the animals 77. composing the Order Edentata. Proc. Marshall, L. G. Zool. Soc. London, pp. 358-367. 1978. Evolution of the Borhyaenidae, extinct Francisco, B. H. R., and F. L. de Souza Cunha South American predaceous marsupi- 1978. Geologia e estratigrafia de bacia de Sao als. Univ. California Publ. Geol. Sci., Jose, Municipio de Itaborai, RJ. An. vol. 117, pp. 1-89. 30 AMERICAN MUSEUM NOVITATES NO. 2761

Marshall, L. G., R. Pascual, G. Curtis, and R. 1977. A primitive pyrothere (Mammalia, No- Drake toungulata) from the early Tertiary of 1977. South American geochronology; radio- northwestern Venezuela. Fieldiana metric time scale for middle to late Ter- (Geol.), vol 33, pp. 397-422. tiary mammal bearing horizons in Pat- Patterson, B., and R. Pascual agonia. Science, vol. 195, pp. 1325- 1972. The fossil mammal fauna of South 1328. America. In Keast, A., F. C. Erk, and Marshall, L. G., R. Hoffstetter, and R. Pascual B. Glass, eds., Evolution, Mammals, [In press] Geochronology of the continental and Southern Continents. Abany, State mammal bearing Tertiary of South Univ. New York Press, pp. 247-309. America. In Woodburne, M. O., ed., Paula Couto, C. de Paleontology as a Discipline in Geo- 1952a. Fossil mammals from the beginning of chronology. the Cenozoic in Brazil. Condylarthra, Matthew, W. D. Litopterna, Xenungulata, and Astra- 1897. A revision of the Puerco fauna. Bull. potheria. Bull. Amer. Mus. Nat. Hist., Amer. Mus. Nat. Hist., vol. 9, pp. 59- vol. 99, pp. 355-394. 110. 1952b. Fossil mammals from the beginning of 1937. Paleocene faunas ofthe San Juan Basin, the Cenozoic in Brazil. Notoungulata. New Mexico. Trans. Amer. Phil. Soc., Amer. Mus. Novitates, no. 1568, pp. 1- new series, vol. 30, pp. 1-5 10. 16. Matthew, W. D., and W. Granger 1952c. Fossil mammals from the beginning of 1918. A revision ofthe lower Eocene Wasatch the Cenozoic in Brazil. Marsupialia: and Wind River faunas. Part 5. Insec- Polydolopidae and Borhyaenidae. Ibid., tivora (continued), Glires, Edentata. no.1559.pp.1-27. Bull. Amer. Mus. Nat. Hist., vol. 34, 1952d. Fossil mammals from the beginning of pp. 429-483. the Cenozoic in Brazil. Marsupialia: Di- Novacek, M. J. delphidae. Ibid., no. 1567, pp. 1-26. 1980. Cranioskeletal features in tupaiids and 1954. On a notostylopid from the Paleocene selected Eutheria as phylogenetic evi- of Itaborai, Brazil. Ibid., no. 1693, pp. dence. In Luckett, W. P., ed., Compar- 1-5. ative Biology and Evolutionary Rela- 1961. Marsupiais f6sseis do Paleoceno do tionships of Tree Shrews. New York, Brasil. An. Acad. Brasil. Cien., vol. 33, Plenum Press, pp. 35-93. pp. 321-333. Osborn, H. F. 1962. Didelfideos f6siles del Paleoceno de 1889. The Mammalia of the Uinta Forma- Brasil. Rev. Mus. Argentino Cien. Na- tion. Part IV. The evolution of the un- tur. "Bernardino Rivadavia" Zool., vol. gulate foot. Trans. Amer. Phil. Soc., 8, pp. 135-166. new series, vol. 16, pp. 530-569. 1963. Um Trigonostylopidae do Paleoceno do 1898. Evolution of the Amblypoda. Part I. Brasil. An. Acad. Brasil. Cien., vol. 35, Taligrada and Pantodonta. Bull. Amer. pp. 339-351. Mus. Nat. Hist., vol. 10, pp. 169-218. 1904. An armadillo from the middle Eocene 1970. Novo notoungulado no Riochiquense (Bridger) of North America. Ibid., vol. de Itaborai. Iheringia, vol. 3, pp. 77-86. 20, pp. 163-165. 1978a. Ungulados f6sseis do Riochiquense de 1929. The titanotheres of ancient Wyoming, Itaborai, RJ, Brasil. I-Xenungulata. Dakota, and Nebraska. U.S. Geol. Surv. An. Acad. Brasil. Cien., vol. 50, pp. Monograph 55 (2 vols.), pp. 1-953. 203-207. Pascual, R., M. G. Vucetich, and J. Fernandez 1978b. Ungulados f6sseis do Riochiquense de 1978. Los primeros mamiferos (Notoungula- Itaborai, RJ, Brasil. II-Condylarthra e ta, Henricosborniidae) de la Formaci6n Litopterna. Ibid., vol. 50, pp. 209-218. Mealla (Grupo Salta, Subgrupo Santa 1978c. Ungulados f6sseis do Riochiquense de Barbara). Sus implicancias filogeneti- Itaborai, RJ, Brasil. III-Notoungulata, cas, taxon6micas, y cronologicas. Trigonostylopoidea. Ibid., vol. 50, pp. Ameghiniana, vol. 15, pp. 366-390. 219-226. Patterson, B. 1979. Ungulados f6sseis do Riochiquense de 1936. The internal structure ofthe ear in some Itaborai, RJ, Brasil. IV-Ratifica,co notoungulates. Geol. Ser. Field Mus. sobre os Notoungulata. Ibid., vol. 51, Nat. Hist., vol. 6, pp. 199-227. pp. 345-348. 1983 CIFELLI: EUTHERIAN TARSALS 31

Schaeffer, B. 1978. Early mammals in South America: fact, 1947. Notes on the origin and function of the controversy, and mystery. Proc. Amer. artiodactyl tarsus. Amer. Mus. Novi- Phil. Soc., vol. 122, pp. 318-328. tates, no. 1356, pp. 1-24. Simpson, G. G., J. L. Minoprio, and B. Patterson 1948. The origin of a mammalian ordinal 1962. The mammalian fauna of the Divisa- character. Evolution, vol. 2, pp. 164- dero Largo Formation, Mendoza, Ar- 175. gentina. Bull. Mus. Comp. Zool., vol. Scillato Yane, G. J. 127, pp. 239-293. 1976. Sobre un Dasypodidae (Mammalia, Xe- Soria, M. F. narthra) de Edad Riochiquense (Paleo- 1980. Una nueva y problematica forma de un- ceno superior) de Itaborai, Brasil. An. gulado del Casamayorense. Actas II Acad. Brasil. Cien., vol. 48, pp. 527- Congr. Argentina Paleontol. Bioestrat., 530. vol. 2, pp. 193-203. Scott, W. B. 1981. Los Litopterna del Colhuehuapense 1910. The Litopterna of the Santa Cruz beds. (Oligoceno tardio) de la Argentina. Rev. Repts. Princeton Univ. Expeditions Mus. Argentino Cien. Natur. "Bernar- Patagonia, vol. VII (Palaeontol. IV), pp. dino Rivadavia," vol. 3, pp. 1-54. 1-156. Stirton, R. A. 1913. A History of Land Mammals in the 1953. A new genus of interatheres from the Western Hemisphere. New York, Mac- Miocene ofColombia. Univ. California millan Co., pp. 1-693. Publ. Geol. Sci., vol. 29, pp. 265-349. 1937. The Astrapotheria. Proc. Amer. Phil. Storch, G. Soc., vol. 77, pp. 309-393. 1981. Eurotamandua joresi, ein Myrmeco- Simpson, G. G. phagidae aus dem Eozain der "Grube 1931. Metacheiromys and the Edentata. Bull. Messel" bei Darmstadt (Mammalia, Amer. Mus. Nat. Hist., vol. 59, pp. 295- Xenarthra). Senckenbergiana Lethaea, 381. vol. 61, pp. 247-289. 1935a. Occurrence and relationships ofthe Rio Szalay, F. S. Chico fauna of Patagonia. Amer. Mus. 1977. Phylogenetic relationships and a clas- Novitates, no. 818, pp. 1-21. sification of eutherian mammals. In 1935b. Descriptions ofthe oldest known South Hecht, M. K., P. C. Goody, and B. M. American mammals, from the Rio Chi- Hecht, eds., Major Patterns in Verte- co Formation. Ibid., no. 793, pp. 1-25. brate Evolution. New York, Plenum 1936. Skeletal remains and restoration of Press, pp. 315-374. Eocene Entelonychia from Patagonia. Szalay, F. S., and R. L. Decker Ibid., no. 826, pp. 1-12. 1974. Origin, evolution, and function of the 1937. The Fort Union of the Crazy Mountain tarsus in Late Cretaceous eutherians Field, Montana, and its mammalian and Paleocene primates. In Jenkins, F. faunas. Bull. U.S. Natl. Mus., vol. 169, A. Jr., ed., Primate Locomotion. New pp. 1-287. York, Academic Press, pp. 223-259. 1945. The principles of classification and a Vucetich, M. G. classification of mammals. Bull. Amer. 1980. Un nuevo Notostylopidae (Mammalia, Mus. Nat. Hist., vol. 85, pp. 1-350. Notoungulata) proveniente de la For- 1948. The beginning of the Age of Mammals macion Lumbrera (Grupo Salta) del in South America. Part 1. Ibid., vol. 91, noroeste Argentino. Ameghiniana, vol. pp. 1-232. 17, pp. 363-372. 1967. The beginning of the Age of Mammals Wheeler, W. H. in South America. Part 2. Ibid., vol. 1961. Revision of the uintatheres. Bull. Pea- 137, pp. 1-259. body Mus. Nat. Hist., vol. 14, pp. 1- 93.