New Specimens of Sparassodonta (Mammalia, Metatheria) From

Home , Abanico Formation, Alphadon, Australidelphia, Barbourofelidae, Barinya, Borhyaenidae

NEW SPECIMENS OF SPARASSODONTA (MAMMALIA, METATHERIA) FROM

CHILE AND BOLIVIA

RUSSELL K. ENGELMAN

Submitted in partial fulfillment of the requirements for the degree of

Master of Science

Department of Biology

CASE WESTERN RESERVE UNIVERSITY

January, 2019 CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of

Russell K. Engelman

candidate for the degree of Master of Science*.

Committee Chair

Hillel J. Chiel

Committee Member

Darin A. Croft

Committee Member

Scott W. Simpson

Committee Member

Michael F. Benard

Date of Defense

July 20, 2018

*We also certify that written approval has been obtained for any proprietary material

contained therein.

TABLE OF CONTENTS

NEW SPECIMENS OF SPARASSODONTA (MAMMALIA, METATHERIA) FROM

CHILE AND BOLIVIA ...... i

TABLE OF CONTENTS ...... iii

LIST OF TABLES ...... vi

LIST OF FIGURES ...... vii

ACKNOWLEDGEMENTS ...... 1

LIST OF ABBREVIATIONS ...... 4

ABSTRACT ...... 7

1. INTRODUCTION ...... 8

2. AUSTRALOGALE LEPTOGNATHUS GEN. ET SP. NOV., A SECOND SPECIES

OF SMALL SPARASSODONT (MAMMALIA, METATHERIA) FROM THE

MIDDLE MIOCENE LOCALITY OF QUEBRADA HONDA, BOLIVIA ...... 13

2.1 INTRODUCTION ...... 13

2.2 MATERIALS AND METHODS ...... 15

2.3 SYSTEMATIC PALEONTOLOGY ...... 17

2.4 DESCRIPTION ...... 19

2.5 DISCUSSION ...... 30

2.5.1 The p3 locus in UATF-V-001900 ...... 30

2.5.2 Systematic position of Australogale leptognathus ...... 34

2.5.3 Reduction of p3 in Metatherians ...... 42

iii

2.5.4 Paleobiology of Australogale leptognathus ...... 44

2.5.5 The Sparassodont Fauna of Quebrada Honda and the Laventan SALMA . 47

3. A LATE EOCENE SPARASSODONT (MAMMALIA, METATHERIA,

SPARASSODONTA) FROM THE LOCALITY OF LOS HELADOS (CENTRAL

CHILE) ...... 51

3.1 INTRODUCTION ...... 51

3.2 MATERIALS AND METHODS ...... 53

3.3 SYSTEMATIC PALEONTOLOGY ...... 53

3.4 DESCRIPTION ...... 55

3.5 DISCUSSION ...... 60

4. EOMAKHAIRA MOLOSSUS, A NEW BORHYAENOID SPARASSODONT

(MAMMALIA, METATHERIA, SPARASSODONTA) FROM THE EARLY

OLIGOCENE (?TINGUIRIRICAN) CACHAPOAL FAUNA, ANDEAN MAIN

RANGE, CENTRAL CHILE ...... 64

4.1 INTRODUCTION ...... 64

4.2 MATERIALS AND METHODS ...... 68

4.3 SYSTEMATIC PALEONTOLOGY ...... 70

4.4 DESCRIPTION ...... 73

4.4.1 Cranium...... 77

4.4.2 Mandible ...... 86

4.4.3 Dentition ...... 92

4.5 PHYLOGENETIC ANALYSIS ...... 118

4.6 DISCUSSION ...... 127

4.6.1 Paleobiology of SGOPV 3490 ...... 127

4.6.2 Carnassial Rotation in Sparassodonts ...... 135

4.6.3 Proborhyaenidae and the Origin of Thylacosmilids ...... 140

5. FUTURE DIRECTIONS ...... 151

SUPPLEMENTARY INFORMATION ...... 155

LITERATURE CITED ...... 278

LIST OF TABLES

Table 2.1. Mandibular depth of UATF-V-001900 compared to other small Neogene sparassodonts...... 20

Table 2.2. Measurements of the dentition and mandibular ramus of UATF-V-001900. .. 22

Table 2.3. Proportions of m1-2 in Australogale leptognathus gen. et. sp. nov. compared to other small sparassodonts and the slightly larger Sipalocyon gracilis...... 23

Table 2.4. Comparison of features among Australogale leptognathus gen. et sp. nov. and other very small to medium-sized South American tritubercular metatherians...... 37

Table 3.1. Anteroposterior lengths (in mm) of the lower dentition of SGOPV 6200...... 59

Table 3.2. Lower molar row anteroposterior length (Lm1-4), estimated body mass, and relative dentary depth at m4 (m4D) of SGOPV 6200 and species of Pharsophorus and

Plesiofelis...... 61

Table 4.1. Measurements of the holotype of Eomakhaira molossus (SGOPV 3490) in mm. Greatest dorsoventral height of maxilla was measured from the alveolar border of P3 to the dorsal border of the maxilla...... 79

Table 4.2. Dental measurements of the holotype of Eomakhaira molossus (SGOPV 3490) in mm...... 93

Table 4.3. Upper canine proportions of SGOPV 3490 compared to other sparassodonts, focusing on canine shape (ratio of anteroposterior length/labiolingual width) and relative canine size...... 94

Table 4.4. Morphometric values of the dentition used to infer dietary habits in

Eomakhaira molossus...... 133

Table 4.5. Angle of inward canting of the posterior upper molars in sparassodonts...... 137

LIST OF FIGURES

Figure 2.1. Temporal and geographical location of Quebrada Honda...... 14

Figure 2.2. UATF-V-001900, Australogale leptognathus gen. et sp. nov. (holotype), right

dentary fragment in A, labial; B, occlusal; and C, lingual views...... 19

Figure 2.3. Photograph (A) and line drawing (B) of m1-2 of UATF-V-001900 in occlusal

view...... 25

Figure 2.4. Line drawings of m1-2 of (A) Australogale leptognathus, UATF-V-001900;

(B) the small sparassodont Pseudonotictis pusillus, MLP 11-26, and (C) the didelphoid

Hesperocynus dolgopolae, FMNH P14469 (modified from Abello et al., 2015), scaled to

the same anteroposterior length, showing the greater resemblance of Australogale to

small sparassodonts than didelphoids...... 28

Figure 2.5. X-ray (A) and micro-CT (B) images of UATF-V-001900, showing the lack of a replacement tooth below p3 (second tooth from right) and the fully formed anterior root of m3 (at left)...... 31

Figure 2.6. Strict consensus tree of 8 most parsimonious trees showing the relationship of

Australogale leptognathus among sparassodonts and other metatherians...... 36

Figure 3.1. Map of Chile (left) and central Chile (inset box) showing the location of Los

Helados (LH, at arrow) and other selected Paleogene Chilean fossil mammal localities of

the Abanico Formation...... 53

Figure 3.2. Photo of SGOPV 6200 showing orientation of dentaries as they were

discovered...... 56

Figure 3.3. Left (A, C, E) and right (B, D, F) dentaries of the SGOPV 6200 in lingual view. Photographs of original specimen (A–B), cast of original specimen (C–D) with

vii natural casts and areas of preserved bone highlighted, and negatives of the natural casts

(E–F). Scale = 1 cm ...... 57

Figure 4.1. Middle Eocene-Oligocene South American Land Mammal “Ages”

(SALMAs)...... 66

Figure 4.2. Location of Cachapoal as well as the similar aged (likely coeval) locality of

Tinguirrica in central Chile...... 67

Figure 4.3. Photographs (A, C) and CT segmentation (B, D) of the holotype of

Eomakhaira molossus, a partial cranium of a senescent individual preserving the rostrum

and the anterior portion of the mandible (SGOPV 3490) in right (A-B) and left (C-D) lateral views...... 75

Figure 4.4. Photograph of the right postcanine dentition of SGOPV 3490, showing the

extremely worn morphology of the dentition...... 76

Figure 4.5. Relative height of the maxilla and depth of the dentary (measured at m3-4

embrasure) in sparassodonts, scaled by the length of the lower molar row (Lm1-4)...... 78

Figure 4.6. Cranium of the holotype of Eomakhaira molossus (SGOPV 3490) in palatal

view...... 81

Figure 4.7. Posterior palate of the holotype of Eomakhaira molossus (SGOPV 3490) in

oblique anteroventral view, showing the paired palatine tubercles and the broken border

of the minor palatine foramen...... 83

Figure 4.8. Left orbital region of the holotype of Eomakhaira molossus (SGOPV 3490) in oblique posterolateral view...... 85

Figure 4.9. Line drawings of the horizontal rami of Callistoe vincei (A) modified from

Babot et al. (2002), and Arctodictis sinclairi (B) modified from Forasiepi (2009),

viii

showing the curved horizontal ramus in Callistoe and the flat horizontal ramus with

“chin” in Arctodictis...... 87

Figure 4.10. Mandible of the holotype of Eomakhaira molossus (SGOPV 3490)...... 88

Figure 4.11. Symphyseal region of the holotype of Eomakhaira molossus (SGOPV 3490)

in anteroventral view, showing the smooth medial border on the anteriorly displaced left

dentary...... 90

Figure 4.12. Photograph (A) and CT images (B-C) of the right upper canine of the

holotype of Eomakhaira molossus (SGOPV 3490)...... 97

Figure 4.13. Comparison of borhyaenoid sparassodont canines. A, Borhyaena tuberata

(MACN-A 6203), showing the plesiomorphic condition of the canines in borhyaenoid

sparassodonts; B, Arminiheringia auceta (MACN-A 10972), showing the typical condition of the canines in proborhyaenids; C, Thylacosmilus atrox (MLP 35-X-4-1), showing the condition of the canines in thylacosmilids. Note the short height of the canine in Borhyaena (alveolar border marked by the white dotted line, specimen is rotated to better show root morphology) related to breakage and heavy wear in this individual, and the presence of enamel despite this heavy wear. Abbreviations: long. grooves, longitudinal grooves on canine root; med. keel, labial median keel; med. sulcus, labial median sulcus...... 100

Figure 4.14. Morphology of the left lower canine root in the holotype of Eomakhaira

molossus (SGOPV 3490)...... 102

Figure 4.15. Oblique lateral slice of SGOPV 3490 (parallel to long axis of right tooth

row), showing the position of the lower molars and the depth of the horizontal ramus. 106

Figure 4.16. Cross-section of the M4 of SGOPV 3490 in (A) oblique occlusal and (B) posterior view, showing the presence of three roots on this tooth...... 113

Figure 4.17. Results of the phylogenetic analysis under equal weights, showing the strict consensus of the four most parsimonious trees (MPTs)...... 121

Figure 4.18. Results of the phylogenetic analysis under implied weights showing the single recovered most parsimonious tree (MPT)...... 124

Figure 4.19. Size comparison among Paleogene proborhyaenids...... 129

Figure 4.20. Left posterior upper dentition (P3-M4) of the holotype of Arminiheringia auceta (MACN-A 10970/10972) in oblique ventral view...... 136

Figure 4.21. Temporal duration of the major lineages of mammalian saber-toothed carnivores with representative skulls of each clade at left...... 147

Figure 0.1. Anterior rostra of sparassodonts and the recently extinct thylacine, showing the presence of anteroventral maxillary foramina (amf) in the former and their absence in the latter...... 230

Figure 0.2. Palatal morphology in metatherians...... 232

ACKNOWLEDGEMENTS

The author would like to thank numerous individuals who provided help and advice

in producing this work. I thank D. Croft for helpful comments and advice regarding the

text of the thesis as a whole and A. Catena for advice in how to organize the thesis.

For Chapter 2, I thank R. Beck, A. Forasiepi, F. Goin, and N. Zimicz for helpful

information on the anatomy of other small extinct marsupials; R. Voss and K.

Travouillon for discussing premolars of extant didelphoids and peramelemorphians,

respectively; R. McCord (Arizona Museum of Natural History) for loans from the Larry

Marshall Marsupial Dentition Collection; R. Wherley and G. Svenson (CMNH) for

generating high-resolution photographs of UATF-V-001900 and casts of Pseudonotictis pusillus; L. Jellma and Y. Haile-Selassie (CMNH) for facilitating x-ray images of UATF-

V-001900; D. Su (Case Western Reserve University) for facilitating CT scanning of

UATF-V-001900; N. Gardner for assistance with the phylogenetic analysis and

Templeton test; the members of the 2013 field crew at Quebrada Honda (F. Carlini, P.

Carlini, A. Catena, D. Croft, M. Ciancio, and N. Drew) that discovered UATF-V-001900; and D. Croft, A. Forasiepi, and an anonymous reviewer who provided comments and criticism that greatly improved this manuscript. This research was funded by the National

Science Foundation (EAR 0958733 and EAR 1423058 to D. Croft).

Chapter 2 is reprinted by permission from Spinger Nature: Journal of Mammalian

Evolution (“Australogale leptognathus, gen. et sp. nov., a Second Species of Small

Sparassodont (Mammalia: Metatheria) from the Middle Miocene Locality of Quebrada

Honda, Bolivia”, R.K. Engelman, F. Anaya, and D.A. Croft), (2018), advance online publication, 27 June 2018 (doi: 10.1007/s10914-018-9443-z.)

For Chapter 3, I thank N. Wong for photographs of SGOPV 6200; A. Balcarcel for

preparation, molding, and casting of the specimen; D. Croft and P. Gans for information

on the geochronological context of Los Helados, E. Ruigómez for information on the

sparassodont specimens from La Gran Hondonada; R. Charrier and the Museo Nacional

de Historia Natural and the Consejo de Monumentos Nacionales, Santiago, Chile for

supporting research in the Abanico Formation, and D. Croft, J. Flynn, and A. Wyss for

comments and criticism that helped improve the manuscript. This research was supported

by funding from the National Science Foundation [DEB-9317943, DEB-0317014, and

DEB-0513476 to J. Flynn; DEB-9020213 and DEB-9318126 to A.Wyss] and the Frick

Fund, Division of Paleontology, AMNH.

For Chapter 4, I wish to thank the members of the 1996 field team who originally

discovered SGOPV 3490, as well as M. Brown for preparation of the specimen. I thank

A. Isch and Z.-X. Luo of the University of Chicago for their help in CT scanning SGOPV

3490, and N. Gard, C. Holliday, and A. Isch for helpful discussions regarding

segmentation and processing of CT data, A. Forasiepi, F. Goin, and C. Suarez for useful

discussions and additional information regarding the anatomy of sparassodonts and other

metatherians, M. Sánchez-Villagra for additional information on specimens of

Herpetotherium, R. Beck for advice on the phylogenetic analysis, R. Wherley and G.

Svenson (CMNH) for generating high-resolution photographs of SGOPV 3490 seen in

Figure 4.4, J. Galkin, A. Marcato, R. Voss, and E. Westwig (AMNH), R. Muelheim and

T. Matson (CMNH), A. Rountrey and C. Thompson (UMMZ), D. Lunde (USNM), and

D. Brinkman, M. Fox, and C. Norris (YPM-VPPU) for access to specimens in their care, and R. McCord (Arizona Museum of Natural History) for loans from the Larry Marshall

Marsupial Dentition Collection, and D. Croft for helpful comments and criticism that improved this manuscript. This research was supported by funding from the National

Science Foundation [DEB-0513476 to J. Flynn] and the Frick Fund, AMNH.

LIST OF ABBREVIATIONS

Anatomical Abbreviations— Upper and lower incisors, canines, premolars, and molars

are designated by I/i, C/c, P/p, and M/m, respectively.

Institutional Abbreviations

AC; Beneski Museum of Natural History, Amherst, USA

AMNH, American Museum of Natural History, New York, USA

CORD-PZ, Museo de Paleontología, Facultad de Ciencias Exactas, Físicas y

Naturales de la Universidad Nacional de Córdoba, Córdoba, Argentina

CM, Carnegie Museum of Natural History, Pittsburgh, USA

CMNH, Cleveland Museum of Natural History, Cleveland, USA

DGM, Divasão de Geologia e Mineralogia do Departamento Nacional da Produão

Mineral, Rio de Janeiro, Brazil

FMNH, The Field Museum, Chicago, USA

IGM, Museo Geológico Nacional, Servicio Geológico Colombiano (formerly

INGEOMINAS), Bogotá, Colombia

MACN-A, Museo Argentino de Ciencias Naturales “Bernardino Rivadavia”, Buenos

Aires, Argentina, Ameghino collection

MACN-PV, Museo Argentino de Ciencias Naturales “Bernardino Rivadavia”,

Buenos Aires, Argentina, vertebrate paleontology collection

MB.Ma, Museum für Naturkunde Berlin, Berlin, Germany

MLP, Museo de La Plata, La Plata, Argentina

MMH, Museo de Monte Hermoso, Monte Hermoso, Argentina

MMP, Museo Municipal de Ciencias Naturales de Mar del Plata, Mar del Plata,

Argentina

MNHC, Museo de Historia Natural de Cochabamba, Cochabamba, Bolivia

MNHN, Museum National d'Histoire Naturelle, Paris, France

MNHN-BOL, Museo Nacional de Historia Natural, La Paz, Bolivia

MNRJ, Museu Nacional e Universidade Federal do Rio de Janeiro, Rio de Janeiro,

Brazil

MPEF-PV, Museo Paleontológico Edigio Ferugulio, Trelew, Argentina, vertebrate paleontology collection

MPM-PV, Museo Regional Provincial Padre M. Jesús M. Río Gallegos, Santa Cruz,

Argentina, vertebrate paleontology collection

OMNH, Sam Noble Museum of Natural History, Norman, USA

PIMUZ, Paläontologisches Institut und Museum Zürich, Zurich, Switzerland

PIN, Institute of Paleontology, Russian Academy of Sciences, Moscow, Russia

PVL, Paleontología Vertebrados Lillo, Tucumán, Argentina;

SGOPV, Museo Nacional de Historia Natural, Santiago, Chile, vertebrate paleontology collections

UALVP, University of Alberta Laboratory for Vertebrate Paleontology, Edmonton,

Canada

UATF-V, Universidad Autónoma Tomás Frías, Potosí, Bolivia, vertebrate paleontology collection

UCMP, University of California Museum of Paleontology, Berkeley, California,

USA

UF, Florida Museum of Natural History, University of Florida, Gainesville, USA

UM, University of Michigan Museum of Natural History, Ann Arbor, USA

UMMZ, University of Michigan Museum of Zoology, Ann Arbor, USA

UNPSJB PV, Universidad Nacional de La Patagonia ‘San Juan Bosco’, Comodoro

Rivadavia, Argentina, vertebrate paleontology collection

USNM, National Museum of Natural History, Washington D.C., USA

YPFB Pal, Yacimientos Petrolíferos Fiscales de Bolivia in the Centro de Tecnología

Petrolera, Santa Cruz, Bolivia, paleontology collection

YPM-VPPU, Princeton University Collection, Yale Peabody Museum, New Haven,

USA

New Specimens of Sparassodonta (Mammalia, Metatheria) from Chile and Bolivia

RUSSELL K. ENGELMAN

ABSTRACT

The Sparassodonta were a group of carnivorous metatherians that were one of the dominant groups of terrestrial predators in South America during the continent’s

Cenozoic isolation. However, the fossil record of Sparassodonta has been largely biased towards sites in Argentina, particularly from the Southern Cone. Here, I describe three new species of sparassodonts, each represented by a single specimen, from Bolivia and

Chile. The first species, from the late middle Miocene of Quebrada Honda, Bolivia, represents one of the smallest sparassodonts known. The second species, from the late

Eocene of central Chile, is an early borhyaenoid and indicates that the diversity and disparity of borhyaenoids prior to the Oligocene was higher than previously appreciated.

The third species, from the early Oligocene of central Chile, is the one of the first early

Oligocene sparassodonts known from more than a single tooth and may represent an early thylacosmilid.

1. INTRODUCTION

In 1887, Florentino Ameghino described several fossils that had been collected earlier that year by his brother Carlos in Patagonia. This was modern science’s first encounter with the Sparassodonta, an extinct order of metatherian carnivores unlike anything alive today in the northern hemisphere. Indeed, in the first few decades after their discovery, there was significant debate over the affinities of sparassodonts, with various authors arguing for affinities with placental carnivorans, “creodonts”, thylacinids and other dasyuromorphians, or didelphimorphians (Ameghino, 1887; Sinclair, 1906;

Matthew, 1907; Wood, 1924; Cabrera, 1927; Simpson, 1941). Although sparassodonts are now universally regarded as metatherians either basal to all living forms or more closely related to New World groups than Australian forms, this debate still continues to this day in some form, as no consensus on the relationships of sparassodonts to other mammals has been established (Marshall and Kielan-Jaworowska, 1992; Szalay, 1994;

Suarez et al., 2016; Wilson et al., 2016; Carneiro, 2018). Nevertheless, in recent years great strides have been made in understanding the paleobiology and evolutionary history of sparassodonts, including locomotor habits (Argot, 2003a, b; Argot, 2004b; Argot,

2004c; Argot and Babot, 2011; Ercoli et al., 2012), dietary habits and guild structure

(Blanco et al., 2011; Ercoli et al., 2014; Forasiepi et al., 2015; Echarri et al., 2017; Croft et al., 2018), ontogeny (Forasiepi and Sánchez-Villagra, 2014; Engelman et al., 2015), and broader evolutionary trends (Prevosti et al., 2013; López-Aguirre et al., 2017; Croft et al., 2018).

However, in spite of these important discoveries, sparassodont paleontology still suffers from some major shortcomings. First, sparassodonts are extremely rare in the

fossil record. A survey of the literature suggests that only ~550 sparassodont specimens

have ever been described or alluded to in the literature (Engelman pers. obs.), a specimen

count roughly comparable to the number of specimens reported for African hominins

(Engelman pers. obs.; Nelson and Jurmain, 1982), despite the much greater time frame and ecological breadth of the former group. Sparassodonts are also rare within fossil mammal communities themselves, with sparassodont specimens being an order of magnitude rarer than would be expected based on comparisons with fossil carnivore faunas on other continents (Croft, 2006; Vizcaíno et al., 2010; Engelman et al., 2015).

There are several likely reasons for this, including methods such as screen-washing likely to sample small species only recently being implemented at many South American fossil localities (Goin et al., 2007), ecological partitioning of the terrestrial carnivore guild in

South America between sparassodonts and other predators such as sebecid crocodyliformes (Argot, 2004a; Argot and Babot, 2011), cariamiform birds (Argot,

2004a; Degrange et al., 2012; Ercoli et al., 2014), and certain armadillos (Vizcaíno and

De Iuliis, 2003; Gaudin and Croft, 2015) prior to the late Miocene, and an early adoption of hypercarnivory and subsequent “macroevolutionary ratchet” (Forasiepi, 2009; Solé and Ladevèze, 2017) combined with developmental constraints on a multi-carnassial postcanine tooth row (Werdelin, 1987) reducing the opportunities for sparassodonts to move into omnivorous niches and therefore lowering niche partitioning and species packing (Croft et al., 2018).

The sparassodont fossil record also exhibits a heavy temporal bias. Most sparassodont fossils come from a rather narrow interval of time, from the late Oligocene

(Deseadan SALMA) to late early Miocene (Santacrucian SALMA), with the fauna from

the late early Miocene of the Southern Cone being the best characterized in South

America (Sinclair, 1906; Argot, 2004a; Ercoli et al., 2012; Ercoli et al., 2014). By

contrast, other intervals, such as the late Eocene-early Oligocene and the middle-early

late Miocene have produced few sparassodont fossils and the evolutionary history of the

group during this time is relatively poorly known. The rarity of sparassodonts from the

late Eocene-early Oligocene and late middle-early late Miocene is of particular concern,

as these periods represent two major periods of climatic change in South America (the

Bisagra Patagonica and the period of global cooling following the Middle Miocene

Climatic Optimum) that resulted in large-scale turnover of the South American metatherian fauna (Goin et al., 2010; Zimicz, 2014; Engelman et al., 2017).

Sparassodonts appear to have reacted evolutionarily to both of these events (Forasiepi et al., 2007; Goin et al., 2010; Forasiepi et al., 2015; Croft et al., 2018), though the details of exactly how are currently unclear due to the poor fossil record.

Another significant issue in sparassodont paleontology involves location and geography. Most sparassodont fossils and a majority of sparassodont species (69% at the time of Forasiepi, 2009) come from Argentina, particularly the high-latitude provinces of

Chubut (including the impressive sequence of localities known from the Sarmiento

Formation; Vacan, Barrancan, Mustersan, ‘Canteran’, Deseadan, and Colhuehuapian

SALMAs) and Santa Cruz (including the fossil-rich Santa Cruz Formation; Sinclair,

1906; Prevosti et al., 2012) in Patagonia, though important fossils have also been found in northwestern Argentina, Mendoza province, and the Pampean region (Riggs, 1934;

Pascual and Bocchino, 1963; Babot et al., 2002; Forasiepi et al., 2015). The fossil record of sparassodonts outside Argentina is rather sparse, with only three localities (La Venta in

Colombia and Salla and Quebrada Honda in Bolivia) producing well-preserved sparassodont remains. Sparassodonts have also been documented from Venezuela

(Suarez et al., 2016), Chile (Marshall, 1990; Flynn et al., 2002), Brazil (Paula Couto,

1952; Ribeiro et al., 2010; Sedor et al., 2017), Peru (Goin and Candela, 2004), and

Uruguay (Mones and Ubilla, 1978), as well as other localities in Bolivia (Villarroel and

Marshall, 1983; Oiso, 1991; Croft et al., 2016), but these specimens are generally rather fragmentary and/or have not been fully described (for a more comprehensive review of sparassodont biogeography, see reviews in Forasiepi, 2009; Prevosti and Forasiepi,

2018).

As a result, this means that almost any well-preserved sparassodont specimen, particularly one from outside the Southern Cone or from a time period other than the late

Oligocene-early Miocene, is likely to provide important information about this group’s evolutionary history. In this paper I describe three new species of sparassodonts, all of which are represented by a single specimen and come from localities outside Argentina.

Chapter 2 describes a new species of small-bodied (~840 g) sparassodont from the late middle Miocene of Quebrada Honda, Bolivia. This taxon was preliminarily mentioned by

Croft et al. (2013), but is fully described here. This specimen is clearly morphologically distinct from the other small sparassodont described from this locality (Engelman and

Croft, 2014), despite being similar in size, and together with this species represents the first example of morphological niche partitioning between similar sized sparassodont species. Chapters 3 and 4 describe new sparassodonts from Chile, the first from outside the late early/earliest middle Miocene and the first from the fossil-rich Abanico

Formation. Chapter 2 describes a new sparassodont from the new late Eocene

(?Mustersan) locality of Los Helados, representing a Cladosictis-sized borhyaenoid with similarities to the much larger late Eocene-Oligocene genera Plesiofelis and

Pharsophorus. Chapter 3 describes a partial rostrum from the early Oligocene

(?Tinguirirican) Cachapoal Fauna, that despite its poor preservation shows similarities to the Thylacosmilidae, the origin of which is currently a major issue in sparassodont paleontology.

2. AUSTRALOGALE LEPTOGNATHUS GEN. ET SP. NOV., A SECOND SPECIES

OF SMALL SPARASSODONT (MAMMALIA, METATHERIA) FROM THE

MIDDLE MIOCENE LOCALITY OF QUEBRADA HONDA, BOLIVIA

2.1 INTRODUCTION

The Sparassodonta were a group of carnivorous metatherian mammals that occupied most predatory mammal niches in South America for much of the Cenozoic.

These animals were essentially the only group of predatory terrestrial mammals in South

America after the early Eocene (Woodburne et al., 2014) until the appearance of immigrant procyonid eutherians and autochthonous predatory didelphoids during the late

Miocene (Huayquerian South American Land Mammal ‘Age’ or SALMA; Simpson,

1974; Butler et al., 1984; Forasiepi et al., 2009). Many large sparassodonts (> 10 kg), as well as some medium-sized taxa (e.g., Sipalocyon and Cladosictis), are known from partial skeletons and relatively well-preserved remains (e.g., Sinclair, 1906; Riggs, 1934;

Marshall, 1977b; Babot et al., 2002; Argot, 2003b; Forasiepi et al., 2006; Forasiepi, 2009;

Argot and Babot, 2011). However, the same cannot be said of the smallest (≤ 1.5 kg) members of this group, most of which are known only from a single fragmentary and/or poorly preserved specimen. As a result, very little can presently be said about the anatomy and morphological diversity of the smallest sparassodonts, even though these animals clearly occupied ecological niches that were quite distinct from those of their much larger relatives.

The greatest number of well-preserved sparassodont remains comes from the late early Miocene (Burdigalian Age; Santacrucian SALMA) Santa Cruz Formation in

southern Argentina (Sinclair, 1906; Prevosti et al., 2012). Among sites that post-date

Santa Cruz, one of the few that preserves multiple species of sparassodonts is the late middle Miocene locality of Quebrada Honda in southern Bolivia (Figure 2.1). Prior to this writing, two species of sparassodonts had been described from this locality. The first,

Acyon myctoderos, was a relatively large (~15 kg) hathliacynid that is represented by several specimens including a nearly complete skull, a partially articulated incomplete skeleton, and partial upper and lower dentitions from a juvenile individual (Forasiepi et al., 2006; Engelman et al., 2015). A second, very small (< 1 kg) unnamed species is represented by a single specimen (UF 27881), a partial rostrum (Engelman and Croft,

2014). Phylogenetic analyses clearly demonstrate that this latter species does not pertain to the Hathliacynidae (Engelman and Croft, 2014; Forasiepi et al., 2015; Suarez et al.,

2016), despite its small size, making it by far the smallest Neogene sparassodont that does not pertain to that group.

Figure 2.1. Temporal and geographical location of Quebrada Honda. A, geological age of the Quebrada Honda faunal assemblage, with the age of the fossil-bearing rocks at Quebrada Honda (roughly 13.2–12.3 Ma, Laventan SALMA) denoted in dark grey; B, location of Quebrada Honda in South America. Modified from Croft et al. (2009), Croft et al. (2013), and Engelman et al. (2017)

In this contribution, I describe a new sparassodont from Quebrada Honda, the third species identified from this locality and the second one of very small size. This new species exhibits a unique combination of features that preclude its referral to any species previously described from Quebrada Honda or any other locality. Additionally, the nearly pristine dentition of the holotype specimen, which contrasts with that of many other small-bodied Miocene and Pliocene sparassodonts, provides an unparalleled view of the dental morphology of a small Neogene sparassodont and suggests that this species occupied a trophic niche distinct from that of UF 27881.

2.2 MATERIALS AND METHODS

Measurements of the specimen described in this study (UATF-V-001900) were made with digital calipers to the nearest 0.01 mm or taken from high-resolution photographs using Adobe Acrobat XI Pro. Measurements were double-checked using both methods to ensure the highest degree of accuracy. Additional data on sparassodonts and other metatherians were taken from either the literature or direct observations of specimens or casts. A complete list of specimens used for comparison be found in

Appendix 1.

In order to determine the identity of the tooth at the p3 locus, I analyzed this specimen using X-ray and micro-CT imaging. For micro-CT imaging, the specimen was scanned using an Inveon microPET/CT scanner at the Case Center for Imaging Research at Case Western Reserve University. The source-to-center distance was 262.440 mm, the source-to-detector distance was 335.390 mm, and the effective voxel size was 98.995255

μm. Digital analysis of the data set was performed using Dragonfly 2.0 (Object Research

Systems Inc., 2017).

The system of SALMAs used in this paper follows Flynn and Swisher (1995),

with modifications by Cione and Tonni (2005), Croft (2007), and Croft et al. (2009).

To test the phylogenetic affinities of the taxon described in this paper, I performed

a phylogenetic analysis by adding it (Appendix 2) to the matrix of Suarez et al. (2016) in

TNT 1.1 (Goloboff et al., 2008). The analysis was performed using the “New Technology

search” option using sectorial search, ratchet, tree drift, and tree fuse options under

default parameters, finding the minimum length 1000 times and then analyzing the

recovered trees under tree bisection reconnection branch swapping. In order to evaluate

the alternate hypothesis that this specimen represents a member of the Didelphoidea, I

performed a Templeton test following the methodology of Templeton (1983) and Larson

(1994) using the script created by Schmidt-Lebuhn (2016).

In this paper, “juvenile” is used to refer to individuals in which the adult dentition

has not entirely erupted (barring taxa like Thylacosmilus in which deciduous teeth are

never replaced; Goin and Pascual, 1987; Forasiepi and Sánchez-Villagra, 2014). I also differentiate between adults and “young adults” and use the latter term to refer to individuals in which all postcanine teeth have erupted but that exhibit non-dental juvenile features (e.g., retromolar space absent or not fully developed, alveolar margin retracted and interradicular process not developed on P/p3 and M/m4, immature bone texture).

I divide sparassodonts into five body size categories following Croft et al. (2018)

and refer to them as “very small” (≤ 1.5 kg), “small” (1.5-7 kg), “medium”, (7-21.5 kg),

“large” (21.5-50 kg), and “very large” (> 50 kg). Special attention is also drawn to the

16 hathliacynids Acyon, Cladosictis, and Chasicostylus, here termed “large hathliacynids”, which are significantly larger than other members of this family and are thought to represent a monophyletic group (see discussion in Engelman et al., 2015).

2.3 SYSTEMATIC PALEONTOLOGY

MAMMALIA Linnaeus, 1758

METATHERIA Huxley, 1880

SPARASSODONTA Ameghino, 1894

cf. HATHLIACYNIDAE Ameghino, 1894

AUSTRALOGALE, gen. nov.

Etymology—From the Latin prefix Australo-, meaning southern, and the Greek suffix – gale, meaning weasel.

Type species—Australogale leptognathus sp. nov.

Diagnosis—As for type and only species.

Australogale leptognathus, sp. nov.

Figures 2.2-2.4; Tables 2.1-2.2

Etymology—From the Greek leptos, meaning slender or thin, and gnathos, meaning jaw, referring to the gracile morphology of the lower jaw. Gender is masculine.

Holotype—UATF-V-001900, a right dentary fragment preserving most of p2, complete p3-m2, and the partial trigonid of m3.

Locality and Horizon—Unnamed formation of the Honda Group, Quebrada Honda,

Bolivia. Locality B-07-22, Río Rosario Local Area, basal red beds, equivalent to lower

levels of Quebrada Honda Local Area (i.e., below Unit 9 tuff of MacFadden and Wolff,

1981) based on correlation of MacFadden et al. (1990).

Age—Late middle Miocene, Serravallian age, Laventan SALMA. The reddish-orange

beds near the base of the section in the Río Rosario Local Area correlate to zone R1 of

the paleomagnetic section at the Quebrada Honda Local Area of MacFadden et al.

(1990:fig. 6). There, zone R1 underlies a 40K/40Ar date on sanidine of 12.83 ± 0.07 Ma

and correlates to Polarity Chron C5Ar.3r of the GPTS (MacFadden et al., 1990), which

spans 13.032–12.887 (Ogg, 2012).

Diagnosis—Differs from all other sparassodonts in having a p3 that is smaller than p2.

Differs from all sparassodonts except Hondadelphys fieldsi in having a well-developed precingulid on p3. Differs from all hathliacynids in that the medial end of the pre-

entocristid of m2 terminates labial to the base of the trigonid and does not form a

complete talonid basin. Differs from all hathliacynids except Pseudonotictis pusillus in

having a conical (as opposed to labiolingually compressed) entoconid on m1-2 and no

pre-entocristid on m1. Differs from all hathliacynids except Borhyaenidium spp. in

lacking a posterobasal heel on p2 (at least among taxa for which p2 is known) and in

having an m1 that is comparable in length to (rather than significantly shorter than) m2.

Differs from all very small hathliacynids except Borhyaenidium riggsi and Pseudonotictis

pusillus in having a long, relatively narrow m1 (L/W ratio > 2.6). Compared to other very

small sparassodonts and Sipalocyon, p2 of Australogale leptognathus is labiolingually

wider (L/W ratio = 2.82) than in Borhyaenidium spp. and Pseudonotictis pusillus but narrower than in Sipalocyon gracilis and Notocynus hermosicus.

2.4 DESCRIPTION

The dentary of UATF-V-001900 is extremely slender, with a maximum height of

5.11 mm and a maximum width of 3.34 mm (Figure 2.2). This gracile morphology is

apparent even when compared to other very small sparassodonts, such as Borhyaenidium,

Pseudonotictis, and Perathereutes (Table 2.1).

Figure 2.2. UATF-V-001900, Australogale leptognathus gen. et sp. nov. (holotype), right dentary fragment in A, labial; B, occlusal; and C, lingual views.

Table 2.1. Mandibular depth of UATF-V-001900 compared to other small Neogene sparassodonts. Measurements of other taxa were taken firsthand or are from Marshall (1981) or Villarroel and Marshall (1983). Abbreviations: D, depth of mandibular ramus below tooth; L, length of tooth

Taxon Specimen m1D m2D Lm2 m1D/Lm2 m2D/Lm2 Australogale leptognathus UATF-V-001900 4.81 5.03 4.69 1.03 1.07 gen. et sp. nov. MNHN-BOL-V- Borhyaenidium altiplanicus - 10.5 6.19 - 1.70 011889 Borhyaenidium musteloides MLP 57-X-10-153 9.90 10.16 5.54 1.79 1.83

Borhyaenidium riggsi FMNH P14409 9.41 10.20 5.90 1.59 1.73

Notictis ortizi MACN-A 3996 7.30 - 4.63 1.58 -

Notocynus hermosicus MLP 11-91 9.83 10.60 6.44 1.53 1.65 Perathereutes pungens MACN-A 684 8.66 8.91 5.01 1.73 1.78 Pseudonotictis pusillus MLP 11-26 7.57 7.91 4.62 1.64 1.71

The dentary is nearly uniform in height and slightly curved along its ventral edge but is

slightly shallower towards the anterior end of the jaw. The labial surface of this specimen

is somewhat convex, whereas the lingual side of the dentary is nearly flat. Only a single

mental foramen is preserved in UATF-V-001900, located directly below m1. By contrast, most other sparassodonts have two or more foramina between p2 and m2 (i.e., the portion of the ramus preserved in UATF-V-001900), including the hathliacynids Acyon spp.,

Borhyaenidium altiplanicus, Cladosictis spp., Sipalocyon spp., Notictis ortizi, Notogale mitis, and Perathereutes pungens. In this respect, UATF-V-001900 resembles the very small hathliacynids Pseudonotictis pusillus, Borhyaenidium riggsi, and B. musteloides, the slightly larger basal sparassodont Hondadelphys, as well as most living didelphids

(Didelphoidea, order Didelphimorphia) and dasyurids (Dasyuromorphia) (Voss and

Jansa, 2009), which have only one mental foramen in this region and two mental foramina in total (the additional mental foramen being located anterior to p2). Only the posterior edge of the mandibular symphysis is preserved in UATF-V-001900, but it

20 shows that this structure extended to a point below the posterior root of p2, similar to

Borhyaenidium (Pascual and Bocchino, 1963). In other hathliacynids, the mandibular symphysis usually extends below the anterior root of p3. The mandibular symphysis appears to have been unfused in Australogale, but not enough of its surface is preserved in UATF-V-001900 to determine whether it was smooth or rugose.

The distal border of the posterior alveolus of p1 is partially preserved and shows that p1 was separated from p2 by a small (0.46 mm) diastema. A second small (0.58 mm) diastema is present between p2 and p3. The p2 of Australogale is proportionally narrower than in Acyon myctoderos, Sipalocyon gracilis, Notocynus hermosicus, and Cladosictis patagonica but wider than Pseudonotictis pusillus, Notogale mitis, Acyon “herrerae”, and species of Borhyaenidium, comparable to the basal sparassodont Hondadelphys fieldsi and the hathliacynid Cladosictis centralis (Table 2.2). Data from other very small to medium-sized sparassodonts show that the length/width ratio of p2 tends to be consistent within a species (see Appendix 3), particularly within hathliacynids; future phylogenetic analyses may find this to be a taxonomically useful feature. Perhaps the most notable feature of Australogale is that p3 is much smaller than p2 in all dimensions

(Table 2.2; Figure 2.2), a feature common in didelphoids (Voss and Jansa, 2009) and dasyuromorphians (Wroe, 1997a, b; Muirhead and Wroe, 1998), but never previously reported in a sparassodont. The p3 of UATF-V-001900 is relatively much smaller than the other teeth in the tooth row, being roughly 65% the length of m2 (Table 2.2), smaller than in most didelphoids (with the exception of the late Miocene Zygolestes paranensis;

Goin, 1997a) and more similar to many dasyuromorphians (Tate, 1947).

Table 2.2. Measurements of the dentition and mandibular ramus of UATF-V-001900. All measurements in mm. Abbreviations: L, greatest length of tooth; W, greatest width of tooth; L(tr), length of trigonid; L(ta), length of talonid; L(pc), length of paracristid; W(tr), width of trigonid; W(ta), width of talonid; D, depth of mandibular ramus below tooth; B, breadth of mandibular ramus below tooth

Measurement p2 p3 m1 m2 m3 L 4.12 3.07 4.61 4.69 - W 1.46 1.38 1.63 2.26 - L(tr) - - 3.24 3.46 - L(ta) - - 1.37 1.23 - L(pc) - - 2.34 2.30 - W(tr) - - 1.63 2.26 2.38 W(ta) - - 1.57 1.77 - D 4.37 4.78 4.81 5.03 5.11 B 3.08 2.74 2.88 3.12 3.34

Both p2 and p3 have anterior cuspules or precingulids, with the cuspules on the

anterior tooth being larger and more distinct. The anterior cuspule of p2 is larger than in

the p2 of Pseudonotictis pusillus, Borhyaenidium musteloides, and Sipalocyon gracilis

and forms a more distinct cingulum. Borhyaenidium altiplanicus appears to have a larger

anterior cuspule on p2 than the three aforementioned taxa (Villarroel and Marshall, 1983)

and may more closely resemble the condition in Australogale, but specimens of B.

altiplanicus could not be directly observed for comparison. The anterior cuspule of p3 is

larger than in most hathliacynids but smaller than in didelphoids and the deciduous

premolars of most sparassodonts, about the same size as in Hondadelphys (UCMP

39251). The posterior faces of p2-3 of Australogale lack well-developed posterobasal

heels, cusps, or cingulids. The morphology of the distal end of p2 in Australogale differs from the condition seen in most didelphoids and Hondadelphys (in which the heel is cuspidate), and hathliacynids like Notocynus, Pseudonotictis, and Sipalocyon (in which a posterobasal heel is present but does not form a distinct cusp), but resembles that of

Borhyaenidium musteloides and Borhyaenidium altiplanicus (in which the posterobasal

heel is absent on p2). By contrast, most sparassodonts and didelphoids have a

posterobasal heel or cusp on p3 as well as dp3. The absence of these structures is clearly

not due to damage or wear given that the teeth of UATF-V-001900 are very well preserved. The anterior and posterior faces of p3 are nearly symmetrical in curvature, similar to hathliacynids but unlike didelphoids and Hondadelphys, though the main cusp

is directed slightly posteriorly, unlike the dp3 of other sparassodonts where this cusp is

vertical. Similarly, while the apex of the main cusp is not preserved on p2, the preserved

extent of this cusp more closely resembles sparassodonts like Hondadelphys and

hathliacynids than didelphoids.

Table 2.3. Proportions of m1-2 in Australogale leptognathus gen. et. sp. nov. compared to other small sparassodonts and the slightly larger Sipalocyon gracilis. Raw measurements can be found in Supplementary Table S2.2. Abbreviations: Lm1, length of m1; Wm1, width of m1; Lm2, length of m2; Wm2, width of m2

Taxon Specimen Lm1/Wm1 Lm1/Lm2 Wm1/Wm2 Australogale leptognathus UATF-V-001900 2.83 0.98 0.72 Borhyaenidium altiplanicus MNHN-BOL-V-011889 2.51 0.93 0.80 Borhyaenidium musteloides MLP 57-X-10-153 2.43 1.05 0.81 Borhyaenidium riggsi FMNH P14407 2.79 0.97 0.80 Notocynus hermosicus MLP 11-91 2.38 0.84 0.79 Notictis ortizi MACN-A 3996 2.66 0.93 0.80 Perathereutes pungens MACN-A 684 2.34 0.95 0.83 Pseudonotictis pusillus MLP 11-26 2.74 0.97 0.78 Sipalocyon gracilis MACN-A 647 2.54 0.90 0.78 Sipalocyon gracilis (S. MACN-A 686 2.32 0.92 0.77 “obusta”) Sipalocyon gracilis MACN-A 691 2.16 0.91 0.82 Sipalocyon gracilis MACN-A 5938 2.40 0.90 0.82 Sipalocyon gracilis MLP 11-7 2.38 0.94 0.79 Sipalocyon gracilis YPM-VPPU 15373 2.17 0.90 0.81

The m1 of Australogale is elongate and similar in length to m2 (Table 2.3; Figure

2.3). The only other sparassodonts with a similarly elongate m1 are Borhyaenidium riggsi

and the large hathliacynids Chasicostylus and Acyon. Comparisons of dental proportions

with other sparassodonts suggest that the comparatively elongate m1 of Australogale

leptognathus is due to this tooth being comparatively narrower labiolingually as well as

proportionally long relative to m2 (Table 2.3). The trigonid of UATF-V-001900 is composed of only two cusps, the paraconid and the protoconid; the metaconid is absent.

The protoconid is taller than the paraconid in m1-2, and both cusps increase in height distally in the molar row. The protoconid is roughly triangular in section, with a flat posterior face. The paracristid is long and oriented at a very acute angle relative to the tooth row, with the paracristid of m1 nearly parallel to the tooth row. The paracristid of m1 has a small notch, partially obliterated by wear, whereas the paracristid notch of m2–

3 is deeper and more prominent. The anterolabial cingulids of m1-2 are very small and almost absent, whereas the anterolabial cingulid on m3 is slightly larger. However, even in this tooth, the anterolabial cingulid is ridge-like and small and is only present on the anterior half of the paraconid; this resembles the condition in hathliacynids and differs from that of carnivorous didelphoids in which it forms a shelf that extends to the protoconid. The anteromedial edge of the anterolabial cingulum and the anterolateral face of the paraconid on m2-3 form a hypoconulid notch. This notch is more evident at the boundary between m2 and m3 due to the larger anterolabial cingulid on m3. There is no notch on m1 for the posterior end of p3, but the base of the paraconid on m1 is rounded and “swollen”, as has been described for some other sparassodonts (Villarroel and

Marshall, 1982; Forasiepi et al., 2006). By contrast, didelphoids other than

Sparassocynus typically have a notch on m1.

Although the talonids of UATF-V-001900 are nearly unworn, few detailed comparisons with other very small sparassodont taxa are possible because the majority of

these taxa are only known from a single specimen with few well-preserved teeth. This

meager sample also precludes well-supported assessments of interspecific (as opposed to

interspecific) variation. As a result, comparisons between UATF-V-001900 and these taxa are not exhaustive but made on a case-by-case basis based on available specimens.

Figure 2.3. Photograph (A) and line drawing (B) of m1-2 of UATF-V-001900 in occlusal view. Abbreviations: co, cristid obliqua; pec, pre-entocristid; ent, entoconid; hyp, hypoconid; hcl, hypoconulid; pa, paraconid; pr, protoconid.

The m1–2 talonids of UATF-V-001900 are basined and composed of three cusps: the hypoconid, the entoconid, and the hypoconulid, which differ in relative heights between the two teeth. In m1, the hypoconid is taller than the entoconid, whereas on m2, the entoconid is the tallest talonid cusp. The hypoconulid is small on both molars, almost nonexistent on m1, and slightly smaller than the entoconid on m2. In most larger

25 sparassodonts in which the entoconid is present (i.e., Lycopsis, Stylocynus, Cladosictis,

Acyon), the hypoconid is taller than the entoconid. However, in Pseudonotictis pusillus, the entoconid of m2-3 is taller than the hypoconid, and in Notocynus hermosicus, Notictis ortizi, and Hondadelphys fieldsi, the entoconid of m3 (a tooth not preserved in UATF-V-

001900) is significantly taller than the hypoconid. The talonids of m2 are not preserved in

Notocynus hermosicus or Notictis ortizi, and the relative heights of the entoconid and hypoconid are not clear in the only specimen of Hondadelphys with m1-2 that could be observed (UCMP 39251). In the specimens of Sipalocyon that could be examined, the hypoconid is always taller than the entoconid on m1, and the entoconid is always taller than the hypoconid on m3, but on m2, the height of the entoconid is variable, either taller than (MACN-A 647; MACN-A 5938), shorter than (YPM-VPPU 15373), or subequal to

(MACN-A 686, MLP 11-7) the hypoconid. The condition of the talonid cusps in

Perathereutes and Borhyaenidium could not be determined based on available specimens.

In didelphoids, the entoconid and hypoconid are nearly subequal in height except in a few taxa (e.g., Monodelphis, Hesperocynus, Lutreolina; Forasiepi et al., 2009; Voss and

Jansa, 2009) in which the entoconid is much lower. Although the entoconid is higher than the hypoconid on m2 in UATF-V-001900, the base of the hypoconid is much larger on m1-2, similar to the condition in other sparassodonts and didelphoids.

The entoconid of Australogale is unusual in its shape and the morphology of the pre-entocristid. The entoconids of m1-2 in UATF-V-001900 are conical, whereas in most sparassodonts (e.g., Patene, Hondadelphys, Sipalocyon, Cladosictis, Acyon, Notogale, most borhyaenoids) and many didelphoids (e.g., many didelphines, Monodelphis,

Thylatheridium, all species of sparassocynids except Sparassocynus maimarai), they are

26 labiolingually compressed on all teeth. However, in the holotype of Pseudonotictis pusillus (MLP 11-26), the entoconids of m1-2 resemble those of Australogale (Figure

2.4), though the entoconid of m3 is labiolingually compressed. The entoconid and hypoconulid of UATF-V-001900 are “twinned”, as in most didelphoids (except

Caluromysiops) and most non-borhyaenoid sparassodonts. There is no pre-entocristid on m1. This is unusual for sparassodonts but does also occur in Pseudonotictis pusillus. By contrast, the pre-entocristid is present on m2 in UATF-V-001900 but is short and terminates mesially lingual to the base of the trigonid. A pre-entocristid that ends lingual to the base of the trigonid is present in Hondadelphys fieldsi and several borhyaenoids

(borhyaenids, Plesiofelis, Pharsophorus, and Prothylacynus) but is not present in hathliacynids for which this character has been previously examined (e.g., Acyon,

Cladosictis, Sipalocyon, Sallacyon). No pre-entocristid appears to be present on m2 in

Pseudonotictis pusillus. A pre-entocristid is present on m3 in this taxon, and although this crest does not end lingual to the trigonid, it does end almost at the very lingual edge of the trigonid. A similar condition may be present in the m3 of Notictis ortizi (MACN-A

3996), but the entoconid is not clearly distinguishable from the cristid in the only known specimen of this taxon. In Notocynus hermosicus (MLP 11-91), which lacks a complete m1-2 but preserves m3 (see Marshall, 1981), the pre-entocristid of m3 is mostly lingual to the base of the trigonid but curves labially at its anterior end to form a closed talonid basin. A closed talonid basin is not present in the m1-2 of Australogale or Pseudonotictis.

Figure 2.4. Line drawings of m1-2 of (A) Australogale leptognathus, UATF-V-001900; (B) the small sparassodont Pseudonotictis pusillus, MLP 11-26, and (C) the didelphoid Hesperocynus dolgopolae, FMNH P14469 (modified from Abello et al., 2015), scaled to the same anteroposterior length, showing the greater resemblance of Australogale to small sparassodonts than didelphoids.

The cristid obliqua of UATF-V-001900 is oriented towards the base of the protoconid in m1-2 and shows a small notch on both teeth. This may be due to the low amount of wear on the molars. It is difficult to determine if notches in the cristid obliqua are present in many didelphoids and sparassodonts due to strong wear on this crest. These notches can be observed in the didelphoids Thylophorops (MMP 354-S) and the sparassodonts Lycopsis longirostrus (UCMP 38061), Pharsophorus lacerans (Forasiepi,

2009; Forasiepi et al., 2015), cf. Nemolestes (AMNH 29433), Sipalocyon gracilis (S.

“obusta”, MACN-A 686), Acyon myctoderos (UATF-V-000926), Cladosictis centralis 28

(MNHN Col. 5), and Acrocyon sectorius (MLP 11-70). It is rather noteworthy that many of the specimens that preserve a notch in the cristid obliqua (UCMP 38061, MACN-A

686, UATF-V-001900, UATF-V-000926, MLP 11-70) are considered to represent

juvenile or young adult individuals (Marshall, 1978b; Marshall, 1981; Forasiepi and

Sánchez-Villagra, 2014; Engelman et al., 2015). There is some indication that the

holotypes of Pseudonotictis pusillus (MLP 11-26) and Notocynus hermosicus (MLP 11-

91) may also have notches in the cristid obliqua of their lower molars, but this is not

entirely clear due to wear on the hypoconid and cristid obliqua. There is no labial

postcingulid, in contrast to the Oligocene hathliacynids Notogale (Patterson and

Marshall, 1978) and Sallacyon (Villarroel and Marshall, 1982) but resembling all

didelphoids, Hondadelphys, and the hathliacynids Borhyaenidium, Notictis, Notocynus,

Sipalocyon, Pseudonotictis, and Perathereutes.

The thickness of the enamel can be observed on the broken occlusal surface of p2

and on the preserved posterior end of m3, approximately where the base of the

protoconid would be on the complete tooth. The enamel is of uniform thickness on p2

and is approximately 0.08 mm thick. The enamel of m3, on the other hand, is not

uniformly thick across the tooth. The enamel is thicker closer to the paracristid notch

(0.155 mm), thinner further basally on the tooth (0.13 mm) and only 0.06 mm thick near

the base of the labial side and on the lingual side of the tooth. The thickness of the

enamel near the paracristid notch is unusual, given that in the much larger hathliacynids

Acyon and Cladosictis, the molar enamel is only 0.1 mm thick (Koenigswald and Goin,

2000; Engelman et al., 2015). This variability in enamel thickness suggests that tooth

enamel in hathliacynids may be variable within individual teeth as well as among

different teeth in the tooth row.

2.5 DISCUSSION

2.5.1 The p3 locus in UATF-V-001900

Perhaps the most notable feature of UATF-V-001900 is that its p3 is much smaller than p2, a characteristic that has not been described in any other sparassodont.

Although the small size of this tooth might suggest that it represents a dp3 rather than p3, several lines of evidence suggest this is unlikely to be the case. The tooth at the p3 locus of A. leptognathus is much more premolariform than the dp3 of other sparassodonts

(Forasiepi and Sánchez-Villagra, 2014; Engelman et al., 2015) and differs from the deciduous teeth of these taxa in lacking anterior or posterior cuspules (posterior cusps being present in all taxa for which dp3 is known, whereas the anterior cusp is small in

Acyon and Cladosictis) and well-developed crests on the main cusp. This tooth is also relatively larger than the dp3 of other sparassodonts (65% the length of m2, as opposed to

54-61% in other sparassodonts; Appendix 5), though allometry of the dp3 relative to other teeth is not well known in sparassodonts. The tooth also shows very little wear, whereas the dp3 of other sparassodonts in which m3 is fully erupted shows prominent wear facets. CT and X-ray images of UATF-V-001900 indicate that the roots of the tooth at the p3 locus in this taxon are fully formed and that no replacement tooth is present

(Figure 2.5). Additionally, m3 was fully erupted by the time the animal died, as the cemento-enamel junction of this tooth is level with that of m1–2, the paraconid is higher than that of m1–2, and the anterior root of m3 is fully developed. In sparassodont

specimens that retain dp3 and show a similar degree of molar eruption to UATF-V-

001900 (i.e., with m3 fully erupted or at least in the process of eruption), the roots of dp3 are at least partially resorbed, and the crown of p3 has calcified (e.g., YPM-VPPU 15097,

Cladosictis patagonica; Forasiepi and Sánchez-Villagra, 2014; UATF-V-000926, Acyon myctoderos, Engelman et al., 2015; UCMP 39251, Hondadelphys fieldsi, Engelman, pers. obs.). Indeed, in YPM-VPPU 15097, in which m3 is only partially erupted, the crown of p3 has already calcified. Similarly, in the extant Monodelphis domestica (Didelphidae) and Sminthopsis virginiae (Dasyuridae), by the time m3 has fully erupted, the crown of p3 has begun to calcify beneath dp3 (Cifelli et al., 1996; Luckett and Wooley, 1996). The stagodontid Didelphodon vorax shows a similar pattern, with p3 calcifying as m3 erupts

(Clemens, 1966).

Figure 2.5. X-ray (A) and micro-CT (B) images of UATF-V-001900, showing the lack of a replacement tooth below p3 (second tooth from right) and the fully formed anterior root of m3 (at left).

The roots of p3 in Australogale are small compared to those of other

sparassodonts. Gracile roots have been suggested to be an indicator of deciduous teeth

(Zack, 2012; Forasiepi and Sánchez-Villagra, 2014) but it is also possible the small roots could be due to the small size of the tooth itself, as the roots do not resemble the condition seen in unreplaced dp3s. In unreplaced dp3s, the anterior root is shorter than the posterior one due to greater resorption of the anterior root by the developing crypt of p3. In UATF-V-00190, the anterior and posterior roots are similar in length (Figure 2.5) with the anterior actually being slightly longer than the posterior.

Another possibility, though less likely, is that this tooth represents a dp3 and that p3 was either not replaced in Australogale leptognathus or replaced late in life (long after the eruption of m4). Some mammals do not replace their deciduous teeth as part of their normal life history, such as octodontoid rodents, in which dP/p4 are retained in many lineages (Vucetich et al., 2010; Arnal and Vucetich, 2015). Only a few metatherians, living or extinct, have been proposed to retain the deciduous premolars into adulthood.

One is the Australian numbat, Myrmecobiius fasciatus (Tate, 1951; but see Bensley,

1903), but in this taxon, the presumed dp3 is retained alongside p3 as a supernumerary tooth and its retention is probably related to this taxon’s myrmecophagous diet (Charles et al., 2013). The very small tooth at the third premolar locus in the late Miocene didelphid Zygolestes paranensis has been suggested to be a retained dp3, but this is not certain (Goin, 1997a; Goin et al., 2000). The thylacosmilids Thylacosmilus and

Patagosmilus retain dP3 into adulthood (Goin and Pascual, 1987; Forasiepi and Carlini,

2010; Forasiepi and Sánchez-Villagra, 2014) but appear to replace dp3 (at least in

Thylacosmilus, the lower dentition is unknown for Patagosmilus; Forasiepi and Sánchez-

Villagra, 2014). A few individuals of sexually mature didelphoids have been reported to retain deciduous premolars but represent cases where the onset of sexual maturity was

accelerated with respect to dental age rather than a prolonged retention of the DP/p3 or

evolutionary loss of P/p3 (Atramentowicz, 1986; Julien-Laferrière and Atramentowicz,

1990; Díaz and Flores, 2008). Although it is theoretically possible that UATF-V-001900 had a pattern of tooth replacement unlike that of any other sparassodont, this is not the most parsimonious interpretation in the absence of additional evidence, particularly given the relative conservatism in eruption patterns of p3 and m3-4 in sparassodonts (as opposed to the variation seen in the patterns of the upper dentition relative to the lower dentition; Forasiepi and Sánchez-Villagra, 2014; Engelman et al., 2015).

Pathological retention of the dp3 and loss of the developing p3 is possible but even less likely. Although agenesis of the permanent premolar is a common cause of retained deciduous premolars in humans, the loss of the developing permanent premolar is often caused by infection or physical trauma, and in most cases (80%), there is still some degree of deciduous tooth root resorption even though the permanent tooth is absent (Ith-Hansen and Kjær, 2000; Aktan et al., 2012). UATF-V-001900 shows no signs of pathology, and the tooth at the p3 locus shows no signs of root resorption.

Furthermore, there is not enough space in the dentary for a p3 equal to or greater in size than p2, as the distance between p2 and m1 (3.65 mm) is much smaller than the length of p2 (4.12 mm). In UATF-V-000926, a juvenile of Acyon myctoderos that is at least the same ontogenetic age as UATF-V-001900, the distance between p2 and m1 is greater than the length of p3 despite p3 still being in the process of erupting.

Nevertheless, the individual represented by UATF-V-001900 was probably not fully grown when it died, as there is still a distinct, functional paraconid on m1. In other sparassodonts, the m1 paraconid is rapidly lost with wear and is only distinct from the precingulum in juvenile or very young adult specimens (Engelman et al., 2015). These features (fully erupted p3 and m3, unworn m1 trigonid) suggest that UATF-V-001900 pertains to a late juvenile or young adult individual, near dental maturity but young enough that its teeth show little wear. It is possible that the gracile morphology of the dentary and limited posterior extension of the mandibular symphysis in this specimen are also partially related to its ontogenetic status, as the depth of the dentary is known to vary throughout ontogeny in sparassodonts (Engelman et al., 2015) as well as in other mammals and can even vary significantly between adult individuals of the same species

(Marshall, 1981; Prevosti et al., 2012). However, if the potential maximum depth of the dentary is extrapolated from UATF-V-000926 (which represents an individual of the hathliacynid Acyon myctoderos that is at least the same ontogenetic age as UATF-V-

001900 based on molar eruption), UATF-V-001900 still has a more gracile dentary than many other very small hathliacynids (Appendix 6).

2.5.2 Systematic position of Australogale leptognathus

The phylogenetic analysis of UATF-V-001900 resulted in eight most-parsimonious trees

(MPTs) of 1035 steps, a consistency index (CI) of 0.368, and a retention index (RI) of

0.685. In all eight trees, Australogale is recovered as a sparassodont (Figure 2.6).

Resolution within the Sparassodonta is poor, primarily because of the limited number of characters that can be coded for Australogale. Constraining Australogale to be a

didelphoid requires an additional nine steps (CI = 0.365, RI = 0.680). According to the

Templeton test (Appendix 7), the unconstrained topology (in which Australogale is recovered as a sparassodont) is a statistically significantly better explanation of the data than the topology in which Australogale is constrained to be a didelphoid (p-value <

0.025). Running the analysis under implied weights (K = 3) produced a single MPT with a best score of 111.76 that recovers Australogale within Hathliacynidae as a sister taxon to Sipalocyon (Appendix 8).

The overall morphology of UATF-V-001900 more closely resembles sparassodonts (specifically the hathliacynids Pseudonotictis and Borhyaenidium) than

any other group of Neogene metatherians (Table 2.4). Only three groups of tritubercular metatherians are known to have been present in South America during the late middle

Miocene: Didelphoidea, Sparassodonta, and Microbiotheria. Other groups of tritubercular

South American metatherians are restricted to the Paleogene (typically Eocene or older) and have not been recorded even from low-latitude localities (Goin, 1997b). UATF-V-

001900 lacks posteriorly salient hypoconids, one of the three synapomorphies of the lower dentition in microbiotherians identified by Goin and Abello (2013). The other two lower molar synapomorphies of Microbiotheria defined by Goin and Abello (2013), reduced or absent anterolabial and posterior cingulids, do occur in UATF-V-001900, but these features are widely distributed in metatherians. Indeed, didelphoids as a whole lack posterior cingulids (Voss and Jansa, 2009) and most sparassodonts (including

Hondadelphys and most Miocene hathliacynids) lack posterolabial cingulids and have reduced anterolabial cingulids.

Figure 2.6. Strict consensus tree of 8 most parsimonious trees showing the relationship of Australogale leptognathus among sparassodonts and other metatherians. Values to the upper left of each node are Bootstrap values and values to the lower left are Bremer supports. Important clades of Neogene South American metatherians (Didelphoidea, Microbiotheria, and Sparassodonta) are denoted in bold.

Table 2.4. Comparison of features among Australogale leptognathus gen. et sp. nov. and other very small to medium-sized South American tritubercular metatherians.

Character Australogale Didelphoidea Hondadelphys Hathliacynidae Microbiotheria Variable, typically Anterior face more Variable, typically Anterior face more Shape of main cusp on p2-3 Symmetrical anterior face more convex convex symmetrical convex Anteroposterior crests on Conical With sharp crests With sharp crests Conical With sharp crests? main cusp of p2-3 Typically present, absent Posterior cingulids on p2-3 Absent Absent Absent Present in didelphines p2/p3 size p2 > p3 Variable p2 < p3 p2 < p3 p2 < p3 Anterolabial cingulid Reduced Well-developed Reduced Reduced Reduced m1-2 metaconid Absent ≥ Paraconid height < Paraconid height Absent ≥ Paraconid height m1-2 paraconid > Hypoconid ≥ Hypoconid > Hypoconid > Hypoconid = Hypoconid m1-2 trigonid/talonid width Trigonid > talonid Trigonid ≤ talonid Trigonid ≤ talonid Trigonid > talonid Trigonid ≤ talonid Hypoconid/protoconid 29% >35% >35% 25-35% >35% height Middle of buccal Posterolabial corner of Posterolabial corner of Middle of buccal Posterolabial corner of Position of hypoconid margin of talonid talonid talonid margin of talonid talonid* Hypoconid posteriorly No No No No Yes salient? Variable, deep in many Hypoflexid Shallow Shallow Shallow Shallow species Morphology/position of Not posteriorly Variable, typically Not posteriorly Not posteriorly Not posteriorly hypoconulid protruding posteriorly protruding protruding protruding protruding Absent in most Posterolabial cingulid Absent Absent Absent Absent Miocene taxa

Distinguishing UATF-V-001900 from didelphoids is more difficult, as there are

no dental features that unequivocally differentiate didelphoids from other metatherians

(Voss and Jansa, 2009). Similarly, there are no currently recognized dental features that unambiguously differentiate hathliacynids from other sparassodonts (Forasiepi, 2009).

Nevertheless, there are several features of UATF-V-001900 that are more similar to hathliacynids than to didelphoids (Table 2.4; Figure 2.4). Many of these features also differentiate UATF-V-001900 from Hondadelphys, a small (~4 kg; Prevosti et al., 2013)

sparassodont from the coeval (MacFadden et al., 1990; Flynn et al., 1997) locality of La

Venta, Colombia, that exhibits many plesiomorphic dental features relative to other

members of the group (Marshall, 1976b; Goin, 1997b).

The lower premolars of Australogale are more sparassodont-like than didelphoid-

like. In most didelphoids, the main cusp of p2-3 is typically asymmetrical in lateral view,

with an anterior face that is more convex than the posterior face, and the base of this cusp

almost always extends to the anterior end of the tooth. By contrast, in the premolars of

hathliacynids (except for the p3 of Borhyaenidium altiplanicus and Pseudonotictis

pusillus), this cusp does not reach the anterior end of the tooth and the curvatures of the

anterior and posterior faces of the main cusp are about the same. Hondadelphys shows an

unusual combination of features; like didelphoids, the p2-3 of Hondadelphys have

anterior faces that are more convex than the posterior faces, but the base of the main cusp

does not extend to the anterior end of the tooth. The premolars of Australogale more

closely resemble the condition seen in hathliacynids than didelphoids or Hondadelphys;

the main cusp of p3 is symmetrical and does not reach the anterior end of the tooth.

Similarly, although the tip of p2 is not preserved in Australogale, its main cusp does not

extend to the anterior end of the tooth, and the curvature of the preserved portion of the

main cusp more closely resembles small hathliacynids than didelphoids or Hondadelphys.

The lower premolars of didelphoids also tend to have well-developed posterior

cuspules and cingulids (except in didelphines) as well as well-developed anterior and posterior crests, even in species in which the anterior and posterior faces of the main cusp of p2-3 are more symmetrical (e.g. caluromyids and sparassocynids). This is not the case in hathliacynid sparassodonts, in which the protoconids of p2-3 have little to no development of crests. Hathliacynids also have relatively less developed posterior cuspules compared to didelphoids (a posterobasal heel is present typically not cuspate, except in Borhyaenidium) and usually lack cingulids. Once again, Hondadelphys shows an unusual combination of features, with similarities to both groups. The p2-3 of

Hondadelphys have well developed anterior and posterior crests and posterior cusps

(based on UCMP 39251) but no posterior cingulids. Australogale resembles hathliacynids and differs from didelphoids and Hondadelphys in lacking well-developed crests or posterior cusps on the premolars and, unlike most didelphoids, has no posterior cingulids.

One major difference between Australogale and didelphoids (that is more similar to sparassodonts) is the morphology of the anterolabial cingulid on the lower molars.

Well-developed, shelf-like anterolabial cingulids are present in all known didelphoids,

including sparassocynids (see Abello et al., 2015, fig. 3), Hyperdidelphys (Goin and

Pardiñas, 1996), Zygolestes (Reig, 1957; Goin et al., 2000), and all extant taxa (Voss and

Jansa, 2009). By contrast, in most sparassodonts, including Hondadelphys and all

hathliacynids, the anterolabial cingulid is smaller and narrower, being restricted to the

39 anteriormost portion of the paraconid. This latter condition is also present in

Australogale.

Another feature that differentiates Australogale from didelphoids that is present in many sparassodonts is the absence of a metaconid on m1–2. All didelphoids, even those most specialized for hypercarnivory (Hyperdidelphys inexpectata and Sparassocynus bahiai; see Zimicz, 2014), still retain a metaconid on m1–2 (Reig and Simpson, 1972;

Goin and Pardiñas, 1996). Indeed, in most didelphoids, the metaconid is either taller than or subequal to the paraconid. In Hondadelphys, a metaconid is present but lower than the paraconid and protoconid (Marshall, 1976b). By contrast, the m1 metaconid is absent in all hathliacynids and borhyaenoids, and the m2 metaconid is absent in all hathliacynids and many borhyaenoids (e.g. Lycopsis, Prothylacynus, proborhyaenids). The only didelphoid that may lack a metaconid on any of its lower molars is Zygolestes paranensis, in which the metaconid may be absent on m4 (Reig, 1957; Goin, 1997a).

The disparity in height between the trigonid and talonid in Australogale

(measured as the percent difference in height between the protoconid and hypoconid; see

Forasiepi, 2009) is more similar to hathliacynids than didelphoids or Hondadelphys. In didelphoids, including extinct taxa such as Hesperocynus, Sparassocynus and Zygolestes, as well as the sparassodont Hondadelphys, the hypoconid is typically > 35% the height of the protoconid. In hathliacynid sparassodonts, the hypoconid is between 25–35% the height of the protoconid. In Australogle, the hypoconid is only about 29% the height of the protoconid, resembling the condition in hathliacynid sparassodonts.

Australogale also differs from didelphoids in the size and shape of the talonid. In most didelphoids, even carnivorous forms like sparassocynids, the talonid is typically

40 subequal to or wider than the trigonid (especially on m1) and the talonid basin is usually wider than long. By contrast, the talonid basin is narrower than the trigonid in hathliacynids, and the talonid basin is subequal in length and width. In Pseudonotictis pusillus, the talonid of m2 is narrower than the trigonid. By contrast, the talonid of m1 in

P. pusillus is subequal in width to the trigonid at its base, but its talonid basin is narrower than the trigonid. In Hondadelphys, the talonid is subequal to wider than the trigonid, but the talonid basin is subequal in length and width. The talonid of Australogale is subequal in length and width and the talonid of m2 is narrower than the trigonid. The talonid of m1 is subequal in width to the trigonid at its base, but the talonid basin is narrower than the trigonid, resembling the condition in Pseudonotictis.

Other aspects of the talonid of Australogale are more similar to sparassodonts

(especially hathliacynid sparassodonts) than didelphoids. The hypoconid is located in the middle of the buccal margin of the talonid, as in hathliacynids; in Hondadelphys and didelphoids, this cusp is at the posterolingual corner of the talonid. The hypoflexid is shallow, similar to Hondadelphys and hathliacynids but unlike many didelphoids, where it is very deep. Finally, the hypoconulid is not very distinct from the posterior border of the talonid in occlusal view. This condition is also present in hathliacynid sparassodonts but contrasts with the condition in most didelphoids, in which the hypoconulid is posteriorly salient and the posthypocristid is transverse to the long axis of the dentary, making it appear as though the hypoconulid is protruding from the posterior border of the talonid. However, this is not universal for didelphoids, as Glironia and caluromyids have a condition more closely resembling that of sparassodonts. In Hondadelphys, the postcristid is oriented more transversely than in other sparassodonts, but at the same time,

41 the hypoconulid is positioned more anteriorly than in didelphoids, making the posterior border of the talonid in this taxon more similar to other sparassodonts than didelphoids.

Although it is not possible to conclusively resolve the position of Australogale within Sparassodonta, a few other morphological features are suggestive of this taxon’s phylogenetic affinities. As mentioned above, Australogale differs from most sparassodonts with the notable exception of Pseudonotictis pusillus in lacking a pre- entocristid on m1 and having a conical entoconid on m1-2. Australogale also resembles members of the genus Borhyaenidium in the absence of a posterobasal heel on p2 and having a mandibular symphysis that only extends to the level of p2 (but see below).

Australogale resembles Pseudonotictis and Borhyaenidium in having only one mental foramen posterior to p2 (though this feature is also present in Hondadelphys), relatively sectorial premolars (Appendix 3), and a long, narrow m1 (Table 2.3). Together, these features suggest that Australogale may eventually be found to be closely related to these two genera.

2.5.3 Reduction of p3 in Metatherians

As mentioned above, a p3 that is much smaller than p2 is a derived trait commonly seen in dasyurids and didelphoids, though the relative disparity in size between p2 and p3 varies within these other clades, with some species having a p3 that is larger than p2. Other metatherian taxa in which p2 is larger than p3 (see Appendix 9 for references) include Glasbius intricatus, Monodelphopsis travassosi, the peradectid

Didelphidectes pumilis, the basalmost thylacinids Muribacinus and Badjcinus, the dasyuromorphian Mayigriphus, the herpetotheriid Copedelphys titanelix, at least some

peramelemorphians (Travouillon et al., pers. comm. January 2015), and possibly the

numbat (Myrmecobius fasciatus), if the second postcanine tooth represents p2 and one of the smaller posterior teeth represents p3. By contrast, p2 is smaller than or similar in size to p3 in a much wider range of metatherians (see Supplementary Appendix 9 for references) including deltatheroidans, Kokopellia, Asiatherium, Alphadon sensu stricto, pediomyids, stagodontids, other peradectids and herpetotheriids, microbiotheres,

Caroloameghinia, pucadelphyids (Pucadelphys and Andinodelphys), Szalinia, the early australidelphian Djarthia, sparassodonts (with the exception of Australogale), and several metatherians from Itaboraí whose higher level systematic placement is uncertain. This distribution of character states suggests that a reduced p3 is a derived character state in didelphoids and dasyurids that was later independently reversed in some lineages. Given that this feature evolved independently in two major radiations of carnivorous Cenozoic metatherians (didelphoids and dasyurids), it is unsurprising that this trait would also evolve within the Sparassodonta.

Interestingly, many reversals in p3 size appear to correlate with an increased degree of carnivory, including in derived thylacinids, sparassocynids, and certain genera of thylamyines (Lestodelphys and Thylamys) and marmosines (Thylatheridium and

Monodelphis). A direct correlation between the relative size of p3 and the degree of carnivory seems to have evolved in parallel in marmosines and thylamyines: p3 is larger than p2 in the most carnivorous taxa (Lestodelphys, Thylatheridium), smaller than p2 in more omnivorous taxa (Gracilinanus, Marmosa), and equal in size to p2 in taxa with an intermediate diet (Thylamys, Monodelphis) (see Vieira and Astúa de Moraes, 2003;

Zimicz, 2014 for dietary information). Additionally, the condition in Badjcinus and

Muribacinus suggests that a p2 that is larger than p3 may be plesiomorphic for

dasyuromorphians as a whole and was only later reversed in derived thylacinids.

Although a relatively small p3 has apparently evolved multiple times in living and

extinct marsupials, the exact reasons why this has occurred have not been studied in

detail. In extant Australian carnivorous marsupials (dasyuromorphians), a small p3 has

been suggested to correlate with a short rostrum (brevirostry) and increased bite force or

mechanical advantage at the canines (Thomas, 1887; Bensley, 1903; Archer, 1976; Wroe,

1997a), though this has not been tested biomechanically. It also is not clear why p3

would be reduced in dasyuromorphians rather than p1, as typically occurs in other

mammals (e.g., carnivorans). Regardless, the idea that a small p3 is an adaptation for

brevirostry in Australogale leptognathus is at odds with its relatively long rostrum (based

on the premolar region) with diastemata between p1-2 and p2-3. A similar condition (i.e.,

a p3 that is smaller than p2 and a relatively long rostrum with diastemata) is present in many other taxa such as Badjcinus and many extant didelphoids. This demonstrates that

an evolutionary reduction in the size of p3 is not always correlated with brevirostry in

metatherians and, perhaps, that the selective forces responsible for the evolution of this

feature in Australogale differed from those in dasyuromorphians.

2.5.4 Paleobiology of Australogale leptognathus

Australogale leptognathus is estimated to have weighed approximately 840 g

based on the anteroposterior length of m2 and the regression equations of Gordon (2003)

(Appendix 10). Body mass could only be determined based on the length of m2 for

several reasons. First, the lower molars of most sparassodonts are labiolingually narrow

44 compared to didelphoids and dasyurids, in part due to the absence of a metaconid, which would bias any estimates based on the width of these teeth (see also Fortelius, 1990).

Additionally, as noted above, the m1 of UATF-V-001900 deviates from the typical morphology seen in didelphoids, dasyurids, and most other very small sparassodonts, being proportionally longer than would be expected based on these taxa (nearly as long as m2), which would cause estimates based on m1 to be biased. This is supported by the fact that mass estimates based on m1 are nearly 33% larger than those based on m2

(Appendix 10). By contrast, the size of the m2 compared to the remaining teeth

(particularly p2 and m3) suggest more typical proportions for this tooth. A body mass estimate of 840 g would make Australogale leptognathus one of the smallest known sparassodonts, along with taxa like Notictis ortizi, Pseudonotictis pusillus, Pseudonotictis chubutensis, the specimen of Patene from the Lumbrera Formation (PVL 2618), and UF

27881 (Engelman and Croft, 2014; Zimicz, 2014).

Despite being one of the smallest known sparassodonts, Australogale leptognathus was much larger than most extant didelphoids. With the exception of didelphines (i.e., Didelphis, Metachirus, and close relatives) and caluromyids such as

Caluromys, the body mass of most didelphoids is < 100 g (Gordon, 2003; Voss and

Jansa, 2009). Even among larger extant didelphoids, many species have a body mass much smaller than that estimated for A. leptognathus, such as Philander quica (Macedo et al., 2007; Vieira and Almeida Cunha, 2008), Caluromys philander, and Metachirus nudicaudatus (Stallings, 1989; Voss et al., 2001). Australogale is also larger than all currently known pre-late Miocene didelphoids (Marshall, 1976b; Goin, 1997b; Goin et al., 2007; Antoine et al., 2013; Goin and Abello, 2013), the largest of which likely had a

45 body mass of no more than 450-500 g (Appendix 10). However, Australogale is smaller than many of the extinct predatory opossum taxa known from later in the Cenozoic (late

Miocene-Pliocene) such as Hyperdidelphys and Thylophorops (Zimicz, 2014). This adds further support to the idea of size-based ecological separation between didelphoids and sparassodonts throughout the late Cenozoic, with geologically older small sparassodonts occupying body-size niches that would later be filled by predatory didelphoids

(Engelman and Croft, 2014).

In recent years, there has been great progress in inferring the paleobiology of sparassodonts and other South American faunivorous metatherians through quantitative methods of the dentition (e.g., Prevosti et al., 2012; Zimicz, 2014). Unfortunately, almost all of these metrics are based on m4, which is not preserved in UATF-V-001900. A few morphological features of Australogale leptognathus provide some limited insight into this animal’s paleobiology, even though it is not possible to make extensive comparisons with other taxa. The premolars (p2-3) and m1 of this taxon are highly sectorial, even relative to other small sparassodonts. The talonids of m1-2 are comparable in size to those of Pseudonotictis and Borhyaenidium but not as large as in Perathereutes or

Sipalocyon, suggesting a primarily faunivorous diet more closely resembling that of the former genera. However, Australogale also has a relatively shallow dentary compared to most small hathliacynids, and the mandibular symphysis is smaller (extending only to p2 rather than p3 as in most small hathliacynids), though it is possible these features may be exaggerated by the young age of the holotype. Additionally, the presence of diastemata between the premolars suggests that Australogale had a relatively long snout, probably shorter than Borhyaenidium but longer than taxa such as Notictis based on comparisons

46 of p2-m2 in these taxa. This suggests that Australogale was likely feeding on prey items much smaller than itself, such as small (mouse-sized) mammals, birds, reptiles, and potentially even large invertebrates.

2.5.5 The Sparassodont Fauna of Quebrada Honda and the Laventan SALMA

Perhaps the most significant aspect of the discovery of Australogale leptognathus is that it increases the both the taxonomic and ecomorphological diversity of sparassodonts at Quebrada Honda. Very small (< 1.5 kg) sparassodonts are extremely rare in the fossil record; all post-Santacrucian (late early Miocene) species are known from a single published specimen. Moreover, prior to this study, no post-Santacrucian locality had produced more than a single very small species (Zimicz, 2014). Multiple very small species were almost certainly present at other post-early Miocene localities based on analogy with modern carnivorous marsupial faunas (e.g., Tasmania; Jones and

Barmuta, 1998) and direct evidence from fossil localities such as Santa Cruz (Ercoli et al., 2014; but see Croft 2013), but Quebrada Honda is the only post-Santacrucian site where the presence of more than one species of very small sparassodont (viz.,

Australogale leptognathus and UF 27881) has been documented directly rather than inferred.

Functional interpretations of the craniodental morphologies of Australogale leptognathus and UF 27881 suggest that these animals varied significantly in their dietary habits, providing a likely explanation for their coexistence despite their similar size (ca.

840 g and 940 g, respectively; Appendix 10). UF 27881 has a short, robust rostrum with large canines and no diastemata between the premolars, features that have been

interpreted as adaptations for hunting relatively large prey (Engelman and Croft, 2014).

By contrast, Australogale is characterized by a relatively longer snout (the length of p2-3

in UATF-V-001900 is longer than the entire premolar row of UF 27881), a gracile

dentary, and sectorial premolars separated by diastemata, features that suggest a diet of

relatively small prey (as noted above). Niche segregation based on dietary adaptations

rather than body size contrasts with what is observed in the Santa Cruz Formation, the

only other South American fossil assemblage where more than very small sparassodont is

known from well-preserved remains. In this fauna, the two very small (< 2 kg) species

(Perathereutes pungens and Pseudonotictis pusillus) are very similar to one another in

both size and dietary adaptations. Although Pseudonotictis and Perathereutes do differ

from one another slightly in craniodental morphology, including dentary depth, talonid

size, and premolar proportions (Marshall, 1981; Prevosti et al., 2012), these are rather subtle differences compared to those that distinguish Australogale from UF 27881.

Marsupial carnivore niche differentiation at Quebrada Honda also differs from the pattern seen in the modern Tasmanian dasyuromorphian community, where Dasyurus maculatus,

D. viverrinus, and Sarcophilus harrisii show significant differences in size in addition to craniodental morphology and locomotor habit (Jones and Barmuta, 1998; Jones, 2003).

Australogale is also noteworthy from a biogeographic standpoint, as it reinforces the distinctiveness of the Quebrada Honda’s mammal fauna relative to that of the coeval site of La Venta. As of this writing, no more than three or four genera and perhaps only a single species are shared between these two sites (Croft, 2007, 2016; Catena et al., 2017;

McGrath et al., 2018), and they have no sparassodont genera or even families in common

(Engelman et al., 2015). As yet another Quebrada Honda metatherian that is distinct from

those of La Venta, Australogale further illustrates the strong faunal provinciality present

in South America during the Miocene (Flynn and Wyss, 1998; Croft et al., 2004, 2007;

Croft et al., 2009; Catena et al., 2017). The absence of very small sparassodonts from La

Venta is striking, as a variety of other small (< 500 g) marsupials including didelphoids

(Marshall, 1976b; Goin, 1997b), paucituberculatans (Dumont and Bown, 1997), and

microbiotheres (Goin, 1997b) have been described from this site, due in large part to

extensive use of screen-washing techniques to collect very small specimens (Goin,

1997b). Therefore, all other things being equal, one would expect very small mammals to

be much better represented at La Venta than at sites such as Quebrada Honda where

deposits suitable for screen-washing are not present.

The fossil record shows a gradual decline in sparassodont disparity and diversity

beginning in the late Miocene, at least in the Southern Cone, where the record is most

complete (see Forasiepi et al., 2007; Prevosti et al., 2013; Zimicz, 2014; Croft et al.,

2018; Prevosti and Forasiepi, 2018). Sparassodonts were most diverse (in terms of number of species) and morphologically disparate during the late Oligocene and early

Miocene (Prevosti et al., 2012; Ercoli et al., 2014; Croft et al., 2018). Larger members of the Hathliacynidae (i.e., species like Cladosictis, > 3.5 kg) are last recorded in the early late Miocene (Chasicoan SALMA), borhyaenids and other non-thylacosmilid borhyaenoids are last recorded in the latest Miocene (Huayquerian SALMA), and small hathliacynids and thylacosmilids are last recorded in the “middle” Pliocene

(Chapadmalalan SALMA; see Prevosti et al., 2013; Esteban et al., 2014). The middle

Miocene thus represents a critical point in the evolution of sparassodonts, linking high middle Cenozoic diversity with their later decline. The three sparassodonts presently

known from Quebrada Honda (Australogale, UF 27881, and Acyon myctoderos) and the five documented at La Venta (Hondadelphys fieldsi, Lycopsis longirostrus, Dukecynus magnus, Anachylsictis gracilis, and an unnamed but clearly morphologically distinct taxon represented by IGM 251108 that may represent a second species of thylacosmilid;

Marshall, 1976b, 1977b; Goin, 1997b) demonstrate that late middle Miocene sparassodonts spanned a broad range of morphologies, comparable to those of early

Miocene sparassodonts. This observation is supported quantitatively if one uses the methods of Croft et al. (2018) to compare Laventan sparassodont disparity (with

Pseudonotictis as a proxy for Australogale, which cannot be coded for most characters) to Santacrucian disparity sensu stricto (i.e., excluding Patagosmilus, which has not yet

been recorded from a Santacrucian site and was treated as a range-through taxon by Croft

et al. 2018); Laventan sparassodont disparity is actually higher (1.160) than Santacrucian

disparity (0.746). Sparassodont diversity during the Laventan SALMA is also relatively

high (at least 8 spp.), comparable to that of the late Oligocene (Deseadan SALMA; see

Forasiepi et al., 2015; Croft et al., 2018) and lower than only the late early Miocene

(Colhuehuapian and Santacrucian SALMAs; see Croft et al., 2018, fig. 1; Prevosti and

Forasiepi, 2018, tab. 5.1). This strongly argues against a decline in sparassodont diversity prior to the late Miocene, a pattern reminiscent of that recently documented for paucituberculatan marsupials (Engelman et al., 2017). Additional sampling of late

Miocene-Pliocene sites, particularly in the northern two-thirds of South America, is necessary to further clarify the pattern of decline in sparassodont diversity.

3. A LATE EOCENE SPARASSODONT (MAMMALIA, METATHERIA,

SPARASSODONTA) FROM THE LOCALITY OF LOS HELADOS (CENTRAL

CHILE)

3.1 INTRODUCTION

Sparassodonts are extinct metatherian mammals that dominated many terrestrial carnivore niches in South America from the Paleocene (Tiupampan or Peligran South

American Land Mammal "Age", or SALMA, depending on whether the Tiupampan

Allqokirus australis is a sparassodont; Marshall and Muizon, 1988; Forasiepi and

Rougier, 2009) to the middle Pliocene (Chapadmalalan SALMA; Goin and Pascual,

1987; Prevosti et al., 2013). Sparassodonts have been recovered across the continent, from the La Guajira Peninsula of Colombia in the north (Suarez et al., 2016), to fossil beds of Santa Cruz Province, Argentina in the south (Sinclair, 1906; Prevosti et al.,

2012). The record of sparassodonts is heavily geographically biased, however, with most fossils coming from Argentina (Forasiepi, 2009) and a few localities in Bolivia

(Villarroel and Marshall, 1982; Petter and Hoffstetter, 1983; Forasiepi et al., 2006;

Engelman and Croft, 2014; Engelman et al., 2015) and Colombia (Marshall, 1977b;

Goin, 1997b).

Only a few fossils of sparassodonts have ever been discovered from Chile. Most

Chilean sparassodont fossils come from the earliest middle Miocene locality of Alto Río

Cisnes in the Río Frias Formation (Figure 3.1). Marshall (1990) identified four species of sparassodonts from this locality, the hathliacynids Sipalocyon gracilis and Cladosictis patagonica and the borhyaenoids Prothylacynus patagonicus and Borhyaena tuberata.

This sparassodont fauna is overall very similar to slightly older sparassodont faunas from the Santa Cruz Formation (Marshall, 1990; Prevosti et al., 2012). More recently, Flynn et al. (2002) and Flynn et al. (2008) reported Cladosictis and cf. Sipalocyon in faunal lists of late early Miocene levels at Pampa Castillo in southern Chile (~47° S) and Laguna del

Laja in south-central Chile (~37.5° S), respectively. Notably, all of these occurrences span a relatively narrow temporal interval (late early Miocene-earliest middle Miocene;

Flynn and Swisher, 1995; Flynn et al., 2002; Flynn et al., 2008), and all appear to represent taxa previously known from the Santa Cruz Formation of Argentina.

Here, I describe a sparassodont from the locality of Los Helados in the Abanico

Formation, near Estero Los Helados in the Tinguiririca River drainage in central Chile

(Figure 3.1). Although the Abanico Formation is best known for producing early

Oligocene mammal fossils (e.g., Flynn et al., 2003b; Hitz et al., 2006; Croft et al., 2008b;

Bertrand et al., 2012; Bradham et al., 2015, and references therein), its assemblages span a wide range of ages, from Eocene to Miocene (Flynn et al., 2003a; Hitz et al., 2006;

Croft et al., 2008a; Flynn et al., 2012). The specimen described here is late Eocene in age, making it one of the few late Eocene sparassodonts known. Along with the specimen from Chapter 4, this specimen is also the first sparassodont from to be formally described from the Abanico Formation, a group whose previous absence from the Abanico

Formation has been somewhat surprising given the formation’s overall richness in mammal fossils (Croft, 2006).

Figure 3.1. Map of Chile (left) and central Chile (inset box) showing the location of Los Helados (LH, at arrow) and other selected Paleogene Chilean fossil mammal localities of the Abanico Formation. Gray area in inset box represents Abanico Formation outcrops. Abbreviations: Az, Azufre (middle Eocene?); Cp, Cachapoal (early Oligocene); LQ, Los Queñes (late Eocene? and late Oligocene?); Tn, Tinguiririca (early Oligocene); Tp, Tapado (middle Eocene?). 3.2 MATERIALS AND METHODS

The system of South American Land Mammal Ages (or SALMAs) used in this paper follows Flynn and Swisher (1995), as modified by Flynn et al. (2003b), Ré et al.

(2010), Woodburne et al. (2014), and Krause et al. (2017).

3.3 SYSTEMATIC PALEONTOLOGY

MAMMALIA Linnaeus, 1758

METATHERIA Huxley, 1880

SPARASSODONTA Ameghino, 1894

BORHYAENOIDEA Simpson, 1930

Gen. et sp. indet.

Figures 3.2-3.3; tables 3.1-3.2

Specimen—SGOPV 6200, fragments of the symphyseal region of the mandible preserving the left and right lower canines and the roots of left p1-3, along with the

natural casts of the lingual face of the dentaries, including the crowns of p1-m4 on both

sides.

Locality—Los Helados, central Chile. SGOPV 6200 derives from volcaniclastic

sediments of the Abanico (= Coya-Machalí) Formation in the greater Tinguiririca River

drainage (~35° S), in the Andean Main Range of central Chile, roughly 20 km west of the

border with Argentina. It was recovered from ~20° west-dipping strata near the crest of

the divide between the Azufre and Los Helados drainages.

Age—Late Eocene, Mustersan SALMA. The geology of Los Helados has not been formally described but is considered to be similar in age to the locality of Azufre, which is also considered to be of late Eocene age (Flynn et al., 2012). Unpublished 40Ar/39Ar

dates from igneous rocks bracketing the fossil-producing layers indicate that the fossils

from Los Helados are probably between 37-36 Ma old (D. Croft and P. Gans, pers.

comm. 2017), at least partially overlapping the ~38-37 Ma range estimated for Mustersan

SALMA faunas at Gran Barranca based on radioisotopic dating of the El Rosado and Bed

10 tuffs (Bond and Deschamps, 2010; Madden et al., 2010; Ré et al., 2010; Dunn et al.,

2013). Referral of Los Helados to the Musteran SALMA is also supported by preliminary identifications of the associated fauna, which includes polydolopid metatherians, dasypodid xenarthrans, the archaeohyracid notoungulate Pseudhyrax, “notopithecine” interatheriids, and toxodontian notoungulates that are less hypsodont than those typical of early Oligocene (Tinguirirican) sites.

3.4 DESCRIPTION

SGOPV 6200 preserves part of the symphyseal region of the mandible, as well as natural casts of the left and right dentaries. The left dentary measures approximately 11.0 cm in anteroposterior length (the posterior edge cannot be identified on the right side).

The two dentaries are separated at the symphysis and splayed apart, with the labial side of each dentary facing the observer (though little bone is preserved) and the natural mold primarily preserving the lingual surface (Figure 3.2). The symphysis apparently was not fused in SGOPV 6200 (contrary to the condition in Prothylacynus and proborhyaenids), as the two halves of the mandible became separated prior to burial. Since the two dentaries are splayed out, with their lingual faces directed into the slab, the extent of the mandibular symphysis and whether its surface is smooth or rugose cannot be determined.

Nevertheless, judging from the smooth surfaces of the casts (Figure 3.3e, f), the symphysis did not reach the p3/m1 embrasure and probably extended only slightly posterior to p2 (contrary to the condition in proborhyaenids and some borhyaenids). A structure that may represent part of the mandibular canal is preserved on the left dentary, suggesting the anteriormost mental foramen opened at or anterior to the level of p2.

Figure 3.2. Photo of SGOPV 6200 showing orientation of dentaries as they were discovered. Both dentaries preserve the base of the coronoid process, which is separated from m4 by a space of ~4 mm, but neither preserves enough of the coronoid process to determine the angle between its anterior border and the tooth row. The opening of the mandibular canal is infilled with matrix on the right dentary, indicating that the mandibular foramen was positioned near the anteroposterior midpoint of the coronoid process. A small portion of the coronoid notch may be preserved on the left dentary, which suggests that the mandibular condyle was probably in line with the tooth row, a feature typical of carnivorous mammals including most sparassodonts.

The ventral border of the dentary posterior to m4 is noteworthy in being nearly straight in lateral view. This is an uncommon feature in sparassodonts, having previously been documented only in Pharsophorus lacerans, Prothylacynus patagonicus,

Proborhyaena gigantea, and Thylacosmilus atrox. By contrast, this region is strongly curved in lateral view in most other sparassodonts, including Pharsophorus tenax, and

Plesiofelis schlosseri which previously has been suggested to be closely related to

Pharsophorus (Cabrera, 1927; Simpson, 1948; Marshall, 1978b). Although the ventral

border of the dentary of the holotype of Pharsophorus lacerans (MACN-A 52-391) is

straight, in a second specimen tentatively referred to this taxon, MPEF-PV 4190 (Goin et al., 2010), this border is curved posterior to m4, calling this specimen’s identification into question. This finding, along with the reinterpretation of “Pharsophorus” antiquus (now

Australohyaena antiquua; Forasiepi et al., 2015) as a borhyaenid (rather than a basal borhyaenoid), highlights the need for a future revision of Pharsophorus and other non- proborhyaenid borhyaenoids (i.e., Plesiofelis).

Figure 3.3. Left (A, C, E) and right (B, D, F) dentaries of the SGOPV 6200 in lingual view. Photographs of original specimen (A–B), cast of original specimen (C–D) with natural casts and areas of preserved bone highlighted, and negatives of the natural casts (E–F). Scale = 1 cm

The only observable remnant of the incisors in SGOPV 6200 is a small fragment of a left lower incisor (Figure 3.3b). This incisor is appressed to the base of the left

canine, suggesting it represents either i2 or i3 (both of which contact the lower canine in

many sparassodonts). The lower canines are large and single-rooted. The base of the right

lower canine shows that the canine roots were closed in adults of this species.

Small diastemata separate the canine and p1, p1 and p2, and possibly p2 and p3.

The three premolars are simple, triangular in profile, and similar in shape. Whether p1 is

oriented obliquely to the tooth row, as it is in most borhyaenoids, cannot be determined.

The premolars, best seen in casts of the right p1-3 (Figure 3.3e, f), gradually increase in

size posteriorly (Table 3.1). Small posterobasal heels are present on p1-3, but the preservation of these teeth only as partial molds makes it difficult to compare the size of these heels to those of other sparassodonts. The main cusp of p3 is asymmetric in lateral view; its anterior face is convex, whereas its posterior face is less curved and only slightly concave. This tooth lacks an anterobasal cuspule or shelf, and its main cusp is strongly curved posteriorly. The p3 of SGOPV 6200 is posteriorly canted relative to the rest of the tooth row. The right p3 appears more canted than the left. A posteriorly-canted p3 is a common feature in borhyaenoid sparassodonts, occurring in the basal borhyaenoids Plesiofelis schlosseri and Pharsophorus lacerans, and the borhyaenids

Australohyaena antiquua, Arctodictis sinclairi, and Borhyaena macrodonta (Marshall,

1978b; Forasiepi et al., 2015) in addition to SGOPV 6200. The roots of p3 are much larger, more massive, and more divergent than those of p1-2, but the roots of this tooth are not as bulbous as in species with inferred bone-cracking habits like borhyaenids and proborhyaenids. The premolars as a whole are notably more gracile and much less robust than in Pharsophorus, Plesiofelis, and borhyaenids.

Table 3.1. Anteroposterior lengths (in mm) of the lower dentition of SGOPV 6200. * = estimated measurement.

Side p1 p2 p3 m1 m2 m3 m4 p1-3 m1-4 Left — — 7.60 6.22 8.33 9.29 9.51 15.24* 32.15 Right 4.45 4.84 6.54 — 8.59 9.25 9.67 19.17 33.36

Like the premolars, the molars of SGOPV 6200 are preserved solely as natural molds, limiting us to a lingual view (best seen from the “positive” cast created by infilling the molds—Figure 3.3e, f). As in nearly all sparassodonts, the paraconid and protoconid are the largest and tallest cusps of the molars, each increasing in height from m1 to m4. Well-developed crests occur on the anterior and posterior edges of the protocone. Casts of the lower molars show a small metaconid on m2-4 that is larger on m2-3 than on m4, best observed on the left m3 (Figure 3.3e). Metaconids are variably present on m2-4 among borhyaenoids, being absent in Lycopsis, Pseudothylacynus,

Prothylacynus, Angelocabrerus, proborhyaenids, and thylacosmilids but present in borhyaenids and the Eocene-Oligocene borhyaenoids Plesiofelis and Pharsophorus, though the metaconid of m4 is absent in Arctodictis and some individuals of Borhyaena

(Forasiepi et al., 2015). The left m1 (the natural mold of the right m1 does not preserve the crown) shows no evidence of a metaconid, suggesting a metaconid was absent on m1 in SGOPV 6200, as is the case for all borhyaenoid and hathliacynid sparassodonts. The natural molds of the molar talonids are not well preserved, limiting what can be said about their morphology. The cast of the right m4 suggests that its talonid consisted of a simple “heel” without a basin. A heel-like m4 talonid is a derived attribute of borhyaenoids among sparassodonts, suggesting that SGOPV 6200 pertains to a member of this clade.

3.5 DISCUSSION

SGOPV 6200 resembles the late Eocene basal borhyaenoid Plesiofelis schlosseri

and the Oligocene Pharsophorus spp. in several respects, including: a posteriorly inclined

p3 with an anterior edge that is more curved than the posterior edge, presence of a

metaconid on m2-4, and a straight posteroventral margin of the dentary in lateral view

(shared with Pharsophorus lacerans but absent in Plesiofelis and possibly Pharsophorus tenax). Some of these features are also present in some borhyaenids (specifically the posteriorly inclined p3 and the metaconid on m2-4), but SGOPV 6200 differs from all borhyaenids in lacking bulbous roots on its premolars and having a straight posteroventral margin of the dentary in lateral view.

SGOPV 6200 also differs from Plesiofelis and Pharsophorus in several respects.

First, this specimen pertains to an animal is much smaller than any species of Plesiofelis or Pharsophorus. Both Plesiofelis and Pharsophorus are large-bodied taxa, with the smallest species (Pharsophorus tenax) estimated to have weighed about 20 kg (Table

3.2). By contrast, SGOPV 6200 represents a much smaller taxon, with a molar row length only about 55-65% that of Plesiofelis and Pharsophorus lacerans and about 70% that of

Pharsophorus tenax, suggesting an animal about the size of the hathliacynid Cladosictis patagonica (approximately 6.5 kg; Table 3.2). In addition, the dentary of SGOPV 6200 is

much shallower – in both absolute and relative terms – than that of Plesiofelis or

Pharsophorus (Table 3.2). SGOPV 6200 further differs from these taxa in having

diastemata between the lower premolars and differs from Pharsophorus lacerans in that

p1-3 increase gradually in size, rather than p1 being significantly smaller than p2-3 (the

condition is unknown in Plesiofelis and Pharsophorus tenax).

Table 3.2. Lower molar row anteroposterior length (Lm1-4), estimated body mass, and relative dentary depth at m4 (m4D) of SGOPV 6200 and species of Pharsophorus and Plesiofelis. Body mass is estimated using Lm1-4 and the regression equation of Myers (2001) for dasyuromorphians. Body mass and dentary depth for SGOPV 6200 estimated using the left dentary, as the ventral border of the dentary and Lm1-4 are better preserved on this side. The body masses of Pharsophorus lacerans and Plesiofelis schlosseri represent extrapolations from the modern dasyuromorphian comparative dataset used by Myers (2001), but the estimated masses for SGOPV 6200 and Pharsophorus tenax are within the range of variation of the sample used to generate the equation and thus likely to be more accurate.

Specimen Taxon Lm1-4 Mass (kg) m4D m4D/Lm1-4 SGOPV 6200 Borhyaenoidea indet. 33.4 6.5 21.9 0.66 (left) MACN-A 52-391 Pharsophorus lacerans 56.0 34.3 42.5 0.76 MACN-A 11653 Pharsophorus lacerans 53.0 28.7 — — Pharsophorus cf. P. MPEF-PV 4190 56.8 35.9 42.9 0.75 lacerans AC 3004 Pharsophorus tenax 48.0 20.9 34.6 0.72 MLP 11-114 Plesiofelis schlosseri 58.6 39.6 45.7 0.78

SGOPV 6200 is about the same size as the late Eocene Procladosictis anomala

(Croft et al., 2018), which is known only from a highly unusual upper dentition, precluding direct comparison; however, P. anomala has been suggested to be a hathliacynid or basal sparassodont rather than a borhyaenoid (Marshall, 1981; Forasiepi,

2009). In summary, SGOPV 6200 probably represents a borhyaenoid closely related to

Plesiofelis schlosseri and Pharsophorus spp., but differs from these taxa in enough features (smaller size, shallower dentary) that it most likely represents a new taxon.

However, because SGOPV 6200 is a natural cast that lacks many of the regions commonly used to diagnose borhyaenoid species (i.e., the talonids), I refrain from naming a new taxon here. Many Eocene sparassodont specimens consist solely of molar teeth (e.g., some of the Mustersan specimens listed below), and if SGOPV 6200 were

61 designated as the holotype of a new species it would be extremely difficult to assign specimens to any given taxon (the trigonid, which is preserved in SGOPV 6200, is highly conservative in sparassodonts and aside from the presence of the metaconid does not preserve any phylogenetically significant information).

SGOPV 6200 is also important because of its late Eocene (Mustersan) age. Only a handful of Mustersan sparassodont specimens have previously been described: the holotype of Procladosictis anomala from Gran Barranca (Marshall, 1981), three specimens from the locality of Cerro del Humo (including the holotype of the borhyaenoid Plesiofelis schlosseri; Simpson, 1948; Marshall, 1978b), four specimens from La Gran Hondonada (E. Ruigómez; personal comm.), which have been referred to

Procladosictis and Plesiofelis but not described (Cladera et al., 2004, table 1), and an isolated lower molar of a proborhyaenid assigned to Callistoe sp. from Antofagasta de la

Sierra (Goin et al., 1998; Powell et al., 2011).

Indeed, the pre-late Oligocene record of Sparassodonta is remarkably poor, with most occurrences representing basal sparassodonts (e.g., Patene), or proborhyaenids, a grouping of exclusively Paleogene sparassodonts only distantly related to most Miocene forms (except possibly thylacosmilids; see Babot, 2005; Forasiepi et al., 2015; Chapter 4 of this thesis). If proborhyaenids are deeply nested within Borhyaenoidea, as consistently recovered by many studies (Muizon, 1999; Babot et al., 2002; Forasiepi, 2009; Engelman and Croft, 2014; Forasiepi et al., 2015; Suarez et al., 2016), then the major lineages of late Cenozoic sparassodonts (e.g., hathliacynids, borhyaenids) must have originated by the middle Eocene (Vacan subage of the Casamayoran SALMA) based on the earliest widely-accepted occurrence of a proborhyaenid (Babot et al., 2002; Powell et al., 2011),

62 or potentially even the early Eocene based on possible proborhyaenid remains from the

Las Violetas Fauna (Gelfo et al., 2010; Krause et al., 2017). An early or middle Eocene divergence of major sparassodont clades is supported by a recent report of a possible borhyaenid from the middle Eocene locality of La Barda (Lorente et al., 2016). SGOPV

6200 indicates that the morphological diversity of Eocene sparassodonts was greater than previously thought, suggesting that non-proborhyaenid borhyaenoids were far more diverse and morphologically disparate in Eocene faunas than currently indicated by the fossil record. This observation is compatible with the long ghost lineages inferred for many Neogene taxa and further highlights the need for additional sampling of Eocene localities from throughout South America.

4. EOMAKHAIRA MOLOSSUS, A NEW BORHYAENOID SPARASSODONT

(MAMMALIA, METATHERIA, SPARASSODONTA) FROM THE EARLY

OLIGOCENE (?TINGUIRIRICAN) CACHAPOAL FAUNA, ANDEAN MAIN

RANGE, CENTRAL CHILE

4.1 INTRODUCTION

The Sparassodonta were a group of carnivorous metatherian mammals that played important roles in the ecosystems of Cenozoic South America, occupying many of the ecological niches filled by placental carnivorans and “creodonts” on other continents.

Sparassodonts exhibited a great deal of morphological diversity, ranging from forms < 1 kg in body mass that occupied roughly the same ecological niche as mustelids (e.g., see

Chapter 2) to forms over 100 kg that converged morphologically with saber-toothed cats

(Wroe et al., 2013; Croft et al., 2018; Prevosti and Forasiepi, 2018). Nevertheless, despite their importance in Cenozoic South American ecosystems, the fossil record of sparassodonts prior to the late Oligocene (Deseadan South American Land Mammal

‘Age’ or SALMA) is very poor. Although phylogenetic analyses have suggested most sparassodont lineages had already diverged from one another by the middle Eocene at the latest (Forasiepi, 2009; Lorente et al., 2016), most major groups of Neogene sparassodonts including borhyaenids, thylacosmilids, and hathliacynids have no pre-

Deseadan record (but see Lorente et al., 2016) and several intervals of the Paleogene have little to no sparassodont record at all (Prevosti and Forasiepi, 2018).

The early Oligocene is a particularly poorly known period in the history of sparassodonts. Indeed, several authors have remarked on the near-absence of

sparassodont remains from the Tinguirirican (López-Aguirre et al., 2017; Croft et al.,

2018; Prevosti and Forasiepi, 2018), the only SALMA currently recognized in the early

Oligocene (Figure 4.1). The poor Tinguirirican record of sparassodonts is almost

certainly due to the limited number of fossil localities known from this interval, which

include the localities of Tinguririca and Cachapoal in the Abanico Formation of central

Chile (Flynn et al., 2003b; Hitz et al., 2006; Croft et al., 2008b; Bertrand et al., 2012;

Bradham et al., 2015, and references therein) and the localities of La Cancha (Goin et al.,

2010) and Cañadon Blanco in Chubut Province, Argentina (the location of the last of these is now lost; Wyss et al., 1994; Reguero et al., 2003). Two other localities, La

Cantera (Goin et al., 2010) and Barrancas Blancas (Dozo et al., 2014), are also thought to be early Oligocene in age, but are slightly younger than the Tinguirirican faunas (La

Cantera based on radiometric dates, Barrancas Blancas based on biochronology; Ré et al.,

2010; Dozo et al., 2014) and are often assigned to an informal “Canteran” interval. The only Tinguirirican locality from which sparassodonts have been described is La Cancha, which has produced a fragmentary upper molar of an extremely small Pseudonotictis- sized sparassodont and an isolated premolar of unknown position (Goin et al., 2010). The only other record of sparassodonts from the early Oligocene comes from La Cantera, where a left dentary tentatively assigned to the borhyaenoid Pharsophorus lacerans has been described (Goin et al., 2010).

Figure 4.1. Middle Eocene-Oligocene South American Land Mammal “Ages” (SALMAs). The La Cantera Fauna is represented as a thin bar as this local fauna is thought to represent a very short interval of geologic time (< 150 ka; Dunn et al., 2013). Figure modified from Croft et al. (2008b) based on data in Ré et al. (2010), Dunn et al. (2013), and Krause et al. (2017). In particular, it is surprising that sparassodonts have not been previously described from the early Oligocene localities of the Abanico Formation (e.g.,

Tinguiririca, Cachapoal), the geological group from which most early Oligocene fossils are found (Figure 4.2). Many other groups of mammals are known from the early

Oligocene localities of the Abanico Formation, including polydolopid, rosendolopid, and argyrolagoid polydolopimorphian metatherians (Flynn and Wyss, 1999, 2004; Goin et al.,

2010); dasypodid cingulates (Carlini et al., 2009); the sloth Pseudoglyptodon (McKenna et al., 2006); chinchillid and dasyproctid rodents (Bertrand et al., 2012); an indaleciid

(Wyss et al., 1994); and notostylopid (Bradham et al., 2015), interatheriid (Hitz et al.,

2006), “archaeohyracid” (Croft et al., 2003; Reguero et al., 2003), leontiniid (Croft et al.,

2008b), “notohippid” (Wyss et al., 1994; Croft et al., 2008b), and homalodotheriid

(Bradham et al., 2015) notoungulates.

Figure 4.2. Location of Cachapoal as well as the similar aged (likely coeval) locality of Tinguirrica in central Chile. Gray area in inset box represents outcrops of the Abanico Formation. Here, I describe a new genus and species of sparassodont from the early

Oligocene locality of Cachapoal in the Abanico Formation, which is thought to be

Tinguirirican in age (Hitz et al., 2006; Croft et al., 2008b; Flynn et al., 2012). This taxon represents one of the first sparassodonts to be described from the Abanico Formation

(along with the specimen described in Chapter 3 from the late Eocene locality of Los

Helados), and represents the first Tinguirirican sparassodont known from more complete remains than an isolated tooth. Additionally, despite being represented by a senescent

67 individual, this species also shows an unusual combination of characters reminiscent of thylacosmilids, a group whose origins are currently a major issue in sparassodont systematics, and may prove to be related to this group.

4.2 MATERIALS AND METHODS

Details of SGOPV 3490, the specimen described in this paper, are difficult to distinguish in its natural state and much of the specimen’s anatomy is hidden by very dense matrix. In order to better elucidate the morphology of the specimen, SGOPV 3490 was CT scanned at the PaleoCT facility at the University of Chicago using a μCT scanner

(GE Phoenix v/tome/x 240kv/180kv scanner). The specimen was scanned at 180 kV and

150 μA with 0.5 mm Cu beam filter, producing a scan of 2021 slices with a voxel size of

0.058 mm. Because of extreme beam hardening and poor differentiation between rock and bone the specimen could not be segmented through automated thresholding and had to be segmented manually. Manual segmentation was performed in Amira 5.3.3 and visualization of the specimen was performed in Avizo 8.0. Additional examination and measurement of the CT scan was performed in Dragonfly 2.0 (Object Research Systems

Inc., 2017).

Additional data on sparassodonts and other mammals was taken either from direct observation by one of the authors or from the primary literature. A complete list of specimens and references used for comparison can be found in Appendix 11.

Many comparisons of SGOPV 3490 involve dimensions of the canines. To standardize terminology throughout this paper, unless otherwise noted, “length” refers to the anteroposterior length of the tooth, “width” refers to its labiolingual width at the

alveolar border, and “height” refers to the extent of the canine above/below the alveolar

border of the tooth.

Certain groups of sparassodonts (proborhyaenids, thylacosmilids) are unusual in

that they have open-rooted, ever-growing canines (as opposed to incisors or molars), a

feature only otherwise seen in bandicoots among metatherians (Aplin et al., 2010; K.

Travouillon, pers. comm.). Several classifications have been proposed to quantify the diversity of high-crowned teeth in mammals (e.g., Simpson, 1970a; Mones, 1982;

Koenigswald, 2011). However, these schemes show a confusing degree of overlap, to the point where which classification scheme is used in the literature mostly depends on personal preference (Billet et al., 2008; Koenigswald, 2011; Ciancio et al., 2014).

Generally, when scientists describe hypsodonty in mammals, they are concerned with defining three major parameters: (1) how tall are the teeth (2) are the teeth ever-growing or rooted, and (3) how are the teeth formed and the occlusal surface maintained. Each of the three classification schemes proposed above primarily focuses on a different yet equally important aspect of high-crowned teeth in mammals (i.e., Simpson, 1970a;

Mones, 1982; Koenigswald, 2011, respectively). In this paper, hypsodonty is used to refer

to teeth that are tall but are not necessarily ever-growing, whereas hypselodonty is used to

refer to animals in which the teeth continue to grow throughout the animal’s lifespan.

Within hypselodonty, I distinguish between protohypselodonty, in which the teeth are

functionally hypselodont for most of the animal’s adult life but the roots close in

extremely old individuals (e.g., the notoungulates Trachytherus and Nesodon; Billet et

al., 2008; Cassini et al., 2012) and euhypselodonty, in which open roots are maintained

throughout the animal’s entire lifespan. The way in which the teeth develop/are

69 maintained (e.g., enamel band hypsodonty, dentine hypsodonty) follows Koenigswald

(2011).

Saber teeth have evolved numerous times in carnivorous mammals, most famously in the felid subfamily Machairodontinae but also in barbourofelids, nimravids,

“creodonts”, and the sparassodont family Thylacosmilidae. In order to avoid confusion the term “sabertooth” in this paper is used to refer to all carnivorous mammals with large,

“saber-like” upper canines, whereas saber-toothed members of the Felidae are specifically referred to as “machairodontines” or “saber-toothed cats”.

The system of South American Land Mammal Ages or SALMAs used in this paper primarily follows Flynn and Swisher (1995), with modifications by Flynn et al.

(2003b), Croft et al. (2009), Ré et al. (2010), Woodburne et al. (2014), Dunn et al.

(2013), and Krause et al. (2017).

4.3 SYSTEMATIC PALEONTOLOGY

MAMMALIA Linnaeus, 1754

METATHERIA Huxley, 1880

SPARASSODONTA Ameghino, 1894

BORHYAENOIDEA Simpson, 1930

PROBORHYAENIDAE Ameghino, 1897

aff. THYLACOSMILIDAE (Riggs, 1933)

EOMAKHAIRA gen. nov.

Etymology—From the Greek prefix “Eo-”, meaning dawn, and “makhaira”, a type of short sword or large knife, in reference to the labiolingually narrow canines of this species and the fact that it may represent a close relative or early member of the

Thylacosmilidae.

Type species—Eomakhaira molossus sp. nov.

Diagnosis—As for type and only species.

Eomakhaira molossus, sp. nov.

Figures 4.3-4.4, 4.6-4.8, 4.10-4.12, 4.14-4.16; Tables 4.1-4.2

Holotype—SGOPV 3490, a partial rostrum of a senescent individual preserving the right

maxilla with C-P3, the alveoli and partial roots of M1-2, and part of M3; the left upper

dentition with C-P3, the anterior root of M1, and M3-4; the left and right horizontal rami

of the mandible with the lower canines and most of the postcanine dentition, with parts of

the coronoid process also preserved; the entire left nasal and small parts of the right; parts

of the palatine; and parts of the inner wall of the orbit (better preserved on the left side

than the right).

Diagnosis—Small borhyaenoid sparassodont, smaller than all other known species

except for Fredszalaya hunteri, SGOPV 6200 from Los Helados (see Chapter 3), and the

possible plesiomorphic thylacosmilid from La Venta (IGM 251108). Differs from all

borhyaenoid sparassodonts except thylacosmilids in the absence of longitudinal striations

on the canine roots, having a labial median keel and posterior ridge on the upper canines,

having a P3 significantly longer than p3 (except possibly Proborhyaena), having an

anteroposteriorly shorter mandibular symphysis and having a shorter p1-3 relative to m1-

4. Differs from all borhyaenoids except Australohyaena antiquua and Lycopsis spp. in

having three roots on M4. Differs from all proborhyaenids (in addition to the features

listed above) in lacking a labial median canine sulcus, having a relatively deeper maxilla

and shallower dentary, an unfused mandibular symphysis, and relatively narrower

canines (except possibly Proborhyaena). Differs from borhyaenids (in addition to the

features listed above) in lacking a metaconid on m2-4, having a deep ectoflexus on M3, straight postcanine tooth row, relatively narrower canines, and possibly in having open- rooted canines. Differs from all thylacosmilids in having three premolars, no mandibular flange, replacement of P3 (status of dP3 is unknown for Anachlysictis), and a deeper dentary. P3 labiolingually wider than in Prothylacynus, but labiolingually narrower than in Pharsophorus, Fredszalaya, borhyaenids, and Callistoe. The p3 is labiolingually wider

than in Prothylacynus or Pharsophorus, but narrower than in borhyaenids, Plesiofelis,

Proborhyaena, and Arminiheringia. Differs from Fredszalaya hunteri in having a much

less bulbous P3, palatal process of maxilla > 1.5, and absence of well-developed stylar cusp B and protocone on upper molars. Differs from SGOPV 6200 in being larger, having a much deeper dentary, lacking diastemata between the lower premolars, having a p3 with bulbous roots, and lacking a metaconid on m2-4. Differs from the possible thylacosmilid from La Venta in being larger, having no mandibular flange, a deeper dentary, and a labiolingually narrower canine.

Type Locality—Cachapoal Local Fauna, Abanico Formation, Central Chile

Age—Early Oligocene, ?Tinguirirican SALMA. The geology and fauna of Cachapoal have not been fully described (Croft et al., 2008a), and as such it is difficult to precisely

correlate it with other localities in the Abanico Formation. The Cachapoal fauna is

thought to be at least 29.3 ± 0.1 million years old, based on the dating of a volcanic tuff

that is either contemporaneous with or located just above the fossil-producing layers

located in a neighboring drainage (Charrier et al., 1997; Flynn and Wyss, 2004). Based on the shared presence of several taxa, including the polydolopid Kramadolops (Flynn and Wyss, 2004), the “archaeohyracids” Protarchaeohyrax and Archaeotypotherium

(Croft et al., 2008b) and the interathere Johnbell hatcheri (Hitz et al., 2006), the

Cachapoal fauna is probably similar in age to the Tinguiririca Fauna (31.5 Ma; Flynn et al., 2003b).

Etymology—From the Greek molossus, a term used to refer to short-snouted, robust- skulled dog breeds such as mastiffs and bulldogs, which in turn refers to the short, robust snout of this species. Gender is masculine.

4.4 DESCRIPTION

SGOPV 3490 is a partial rostrum (Figure 4.3) preserving parts of the right maxilla and left and right horizontal rami of the mandible along with parts of the right upper dentition and fragments of several other bones of the rostrum (e.g., left nasal). Compared to some other fossils described from the Abanico Formation, SGOPV 3490 is not particularly well preserved. The specimen has been heavily altered through diagenesis and the matrix is very dense, making it difficult to distinguish rock from bone even through CT imagery. Many parts of the specimen have been damaged or destroyed during preservation, leaving the remaining material isolated but in near-life position. For example, the left lower molar row is preserved essentially in-situ, although the alveolar

bone surrounding them is not preserved (primarily on the labial side, some is present on

the lingual side). This type of preservation is common in the Abanico Formation (Croft,

pers. comm.), and is thought to be due to the preservation of these specimens by volcaniclastic lahars, destroying thinner, more fragile parts of the skeleton while leaving other elements (particularly resistant material such as teeth and petrosals) “floating” in life position (McKenna et al., 2006). The specimen shows clear signs of crushing and distortion, particularly on the left side where elements of the skull have been displaced anteriorly and show signs of breakage. By contrast, the right side of the specimen is nearly undistorted, as the teeth are in near-occlusal position and the maxilla and mandible do not show signs of crushing.

The correct orientation of this specimen such that it is in proper anatomical view is also difficult to determine. Many of the features typically use to orient a specimen

(e.g., the alveolar line of the postcanine tooth row or the ventral edge of the dentary), do not appear to be accurate indicators of orientation in SGOPV 3490. Orienting the skull based on these features typically results in physically impossible orientations of other parts of the skull in ways that cannot be accounted for by distortion (see below). The orientation of the specimen used in this paper was chosen because it results in the fragments of the palate being nearly parallel to the frontal plane and the roots of several postcanine teeth (P2-3, m1-3) nearly vertical, as would be expected if the skull was in life position. This position also results in an orientation of the canines and an angle of inclination of the lower molar row that is within the realm of variation seen in other sparassodonts (e.g., Arctodictis, Arminiheringia, Australohyaena, Callistoe, some individuals of Thylacosmilus). Conclusively determining the proper orientation of the

rostrum in Eomakhaira will require more complete specimens of this taxon in which the

orientation of the rostrum can be more objectively determined.

Figure 4.3. Photographs (A, C) and CT segmentation (B, D) of the holotype of Eomakhaira molossus, a partial cranium of a senescent individual preserving the rostrum and the anterior portion of the mandible (SGOPV 3490) in right (A-B) and left (C-D) lateral views. For the CT segmentation, the cranium (maxilla, palate, nasal, etc.) is shown in purple, teeth are shown in yellow, and the dentary is shown in green. Anterior is to the right in A–B and to the left in C–D.

Figure 4.4. Photograph of the right postcanine dentition of SGOPV 3490, showing the extremely worn morphology of the dentition.

Compounding these issues in describing SGOPV 3490 is this specimen’s

ontogenetic status. Based on several features, SGOPV 3490 likely represents a highly

senescent individual. The canines are extremely blunt, even compared to many other

sparassodonts, with the apices of both the upper and lower canines nearly rounded. Most

of the postcanine dentition is also heavily worn, the main cusp of the left p2 is nearly

worn flat and the occlusal morphology of the right M3 is virtually obliterated. The P3 has

worn to the point that the roots of the teeth are functioning in occlusion and the occlusal

faces of m1 and the trigonid of m2 are virtually flat due to wear (Figure 4.4). The obliteration of the occlusal morphology of m2 classifies this specimen as a senile individual under the system of individual dental age stages of Anders et al. (2011). Even

M4, which is typically the last tooth to erupt in sparassodonts (Forasiepi and Sánchez-

Villagra, 2014; Engelman et al., 2015) and shows the least amount of wear in many specimens, exhibits well-developed wear facets in SGOPV 3490. The extreme wear seen on the dentition clearly happened in the animal’s lifetime, rather than as breakage and abrasion after the animal died, as the posterior face of P3 and the trigonid of m1 are almost in occlusion and their wear facets match perfectly. The senescent nature of the specimen has destroyed most of the specimen's dental morphology and makes it difficult to describe features of SGOPV 3490 (or at least determine whether a particular feature of this specimen is a real morphological feature of adults of this taxon or age-related) even when these elements are well preserved.

4.4.1 Cranium

The maxilla of Eomakhaira molossus is strikingly deep. Indeed, a comparison of maxillary proportions in sparassodonts (Table 4.1, Figure 4.5) found that the maxilla of

Eomakhaira is relatively deeper than in most other sparassodonts, including the proborhyaenids Arminiheringia and Callistoe and the robust-skulled borhyaenid

Arctodictis sinclairi. Only Australohyaena antiquua, Arctodictis munizi, and

Thylacosmilus atrox had deeper maxillae among the taxa analyzed. The maximum height of the maxilla in Paraborhyaena could not be determined with certainty but based on the estimates of Petter and Hoffstetter (1983) it seems to be more similar to the Eocene proborhyaenids than Eomakhaira. A small portion of the dorsal border of the infraorbital foramen is preserved and indicates that this structure opened dorsal to P3/M1 embrasure, as in Patagosmilus, Thylacosmilus, and Australohyaena but unlike the Eocene taxa

Callistoe and Arminiheringia, where the infraorbital foramen opens above or anterior to the anterior root of P3; Borhyaena or cf. Proborhyaena (MLP 79-XIII-18-1), in which the foramen opens slightly more anteriorly over the posterior root of P3; or in species of

Arctodictis, where the foramen is more posterior (above or posterior to the posterior root of M1; Forasiepi, 2009).

Figure 4.5. Relative height of the maxilla and depth of the dentary (measured at m3-4 embrasure) in sparassodonts, scaled by the length of the lower molar row (Lm1-4). Several taxa are denoted with reconstructions of the skull for orientation. Skull reconstructions of Arctodictis sinclairi, Acyon myctoderos, Callistoe vincei, and Thylacosmilus atrox modified from Forasiepi (2009), Forasiepi et al. (2006), Babot et al. (2002), and Riggs (1934), respectively. For raw data see Appendix 12. Although the alveolar border of the maxilla posterior to the infraorbital foramen is fragmentary, what parts are preserved suggest that Eomakhaira lacked maxillary

“cheeks”. Maxillary “cheeks”, which are lateral protrusion of the maxillae beyond the tooth row posterior to the infraorbital foramen best seen in ventral view, are a highly variable feature within sparassodonts, present in borhyaenids, Prothylacynus, and many hathliacynids but absent in most species of Lycopsis (except L. torresi), Acyon,

Patagosmilus, and MLP 79-XII-18-1. Contrary to previous studies, maxillary “cheeks” appear to be absent in Paraborhyaena and Thylacosmilus based on observations of the

skulls of these taxa (Engelman personal obs., see also Petter and Hoffstetter, 1983; Goin

and Pascual, 1987). The state in Arminiheringia could not be determined based on

available information.

Greatest dorsoventral height of maxilla 42.8 Greatest width of nasals (estimated as twice greatest width 24.8 of right nasal) Width of nasals at the level of the canines 6.24 Maximum width of palate between canines 22.9 Approximate width of palate at the level of the infraorbital 28.0 foramen Approximate maximum width of palate (at level of M3) 44.2 Length of C-M3 (approximate) ~49 Length of P1-3 20.3 (left), 19.5 (right) Length of M1-3 (approximate) ~26.7 Length of c-m4 (right) Length of p1-3 16.4 (left) Length of m1-4 37.3 (right) Length of symphysis 20.2 Depth of dentary below p3 21.2 (left), 21.6 (right) Depth of dentary below m3 30.3 (right) Estimated greatest depth of dentary below m4 31.8 (right)

Based on the left nasal, the nasals of Eomakhaira formed an angle of roughly

100° with the frontals in dorsal view, greater than the acute naso-frontal suture seen in

Callistoe but narrower than the broad naso-frontal suture of Paraborhyaena,

Patagosmilus, Pharsophorus, and borhyaenids. There is no internasal projection of the

frontals (i.e., the naso-frontal suture is V-shaped, not W-shaped). As in Callistoe, it

appears part of the nasals may have extended onto the lateral surface of the snout, but this

cannot be determined for certain without better preserved remains. The anterior half of

the nasals are relatively narrow, approximately only 25% of the greatest width of the

nasals overall. This disparity in width between the anterior and posterior nasals is slightly

greater than is typical for borhyaenoids, which on average have nasals whose anterior

portion is about 30% or so the greatest width of the nasals (Appendix 13). The only

exception is Callistoe, in which the anterior and posterior ends of the nasals are much

more similar in width. The hathliacynids Sipalocyon and Acyon show less disparity in

width between the anterior and posterior nasals than any borhyaenoid (except Callistoe,

see above), though the proportions in Cladosictis more closely resemble those of

borhyaenoids. Interestingly, the sparassodont UF 27881, originally described as a basal

sparassodont (Engelman and Croft, 2014) but recovered as a borhyaenoid in later

analyses (Forasiepi et al., 2015; Suarez et al., 2016), does not resemble borhyaenoids in

its nasal proportions, instead more closely resembling a specimen tentatively assigned to

the basal sparassodont Hondadelphys (IGM 250364; Goin, 1997b).

Although the nasals of borhyaenoids are relatively conservative in overall

proportions, they vary from one another in terms of their absolute size (using the greatest width of the nasals as a proxy for overall nasal size), being proportionally larger in taxa like Arctodictis and Australohyaena and relatively smaller in taxa like Prothylacynus and

Acrocyon. It is not clear if this is a simple allometric relationship, as the largest sparassodonts examined tend to have the proportionally largest nasals (i.e., Arctodictis,

Arminiheringia, and Australohyaena, but notably not Paraborhyaena). Patagosmilus and

Callistoe also have relatively small nasals for borhyaenoids, more similar to hathliacynids

in relative size, but again it is not clear if this is due to allometry.

Figure 4.6. Cranium of the holotype of Eomakhaira molossus (SGOPV 3490) in palatal view. Anterior is to the right. Lacrimal fragment in teal, palatine in blue, all other bones of the cranium (i.e., maxilla) in purple. Abbreviations: mpps, medial postpalatine spine; ptub, palatine tubercles; pp, palatal pit; dental abbreviations as in Materials and Methods. The palate of SGOPV 3490 is fragmented and patchily preserved, but enough of the palate is present that it is possible to determine that Eomakhaira lacked maxillopalatine fenestrae, as in all other sparassodonts (Figure 4.6). The length/width ratio of the palatal process of the maxilla in ventral view is greater than 1.5 even after accounting for distortion, similar to most sparassodonts (with the exception of thylacosmilids and the borhyaenids Australohyaena and Arctodictis). At the posterior end of the palate in ventral view are two long, low ridges that extend across the entire palatine bone (Figure 4.7). These structures are termed the palatal ridges here, due to uncertainties regarding their homology. In many metatherians, including the stem marsupials

Didelphodon (Wilson et al., 2016), Pucadelphys (Marshall and Muizon, 1995), and

Andinodelphys (Muizon et al., 1997) and several groups of crown marsupials, the posterior edge of the palate has a prominent postpalatine torus, which often gives the posterior palate a very straight or angular appearance. However, other metatherians, including deltatheroidans (Forasiepi, 2009; Bi et al., 2015), Mayulestes (Muizon, 1998), basal didelphoids (caluromyids and Glironia; Voss and Jansa, 2009), and some dasyuromorphians (Barinya, Thylacinus; Wroe, 1999; Engelman pers. obs.) lack a palatine torus. Among metatherians, sparassodonts are notable in lacking a palatine torus.

In most sparassodonts (e.g., UF 27881, Cladosictis, Sipalocyon, Borhyaena,

Australohyaena, Thylacosmilus) the area of the palatine surrounding the choanae is slightly thickened, but does not form a palatine torus. In species of Arctodictis, this feature is even more pronounced (Forasiepi et al., 2004; Forasiepi, 2009), but this thickening of bone still follows the borders of the choanae. By contrast, proborhyaenid sparassodonts have ridges at the back of the palate that do not follow the borders of the choanae (in these taxa the ridges are distinct from the anterior borders of the choanae and the medial borders of the choanae are not thickened). In Callistoe these ridges of bone are much more prominent than the thickened choanal border seen in other sparassodonts, but are not completely straight. By contrast, the palatal ridges of Arminiheringia and

Eomakhaira are nearly straight. However, these structures do not form a complete palatine torus, as they are paired structures that do not come in contact at the midline line like the palatine tori of other metatherians and do not constrict the choanae (i.e., the choanae still exhibit the classic “double-arch” pattern typical of sparassodonts). The condition in the two largest species of proborhyaenids, Proborhyaena and

Paraborhyaena, is unknown. It is possible that these palatal ridges might be homologous

to the thickened borders of the choanae seen in other sparassodonts or palatine tori of

other metatherians, but this remains to be determined.

Figure 4.7. Posterior palate of the holotype of Eomakhaira molossus (SGOPV 3490) in oblique anteroventral view, showing the paired palatine tubercles and the broken border of the minor palatine foramen. Abbreviations: mpf, minor palatine foramen; ptub, palatine tubercles. Just lateral to the palatine ridges are a pair of small foramina. These foramina have been slightly damaged due to the anterior displacement of the palatine and based on their position likely represent the minor palatine foramen (Figure 4.7). As in most metatherians, including many sparassodonts (including UF 27881, Cladosictis,

Arctodictis, Callistoe, Arminiheringia, Patagosmilus, and some but not all specimens of

Thylacosmilus), the minor palatine foramen is located between the maxilla and the

83 palatine. In Callistoe and Arminiheringia, because of the large palatal ridges at the posterior end of the palate, the minor palatine foramina are under the palatal ridges. The condition in SGOPV 3490 is intermediate in some respects between the condition seen in

Callistoe/Arminiheringia and Patagosmilus; the minor palatine foramen is very close to the palatine tubercles, but the foramen itself is much more laterally positioned than in the former taxa and is close to the upper molar row as in Patagosmilus. In terms of size, the minor palatine foramen of Eomakhaira is relatively large and more similar in size to

Patagosmilus than Arminiheringia or Callistoe. Accounting for deformation and the borders of the minor palatine foramen, the horizontal process of the palatine does not appear to have extended posterior to M4, similar to the condition in Borhyaena,

Patagosmilus, and some specimens of Prothylacynus.

In addition to the rostral regions of the skull SGOPV 3490 also preserves a small portion of the orbital region, primarily on the left side. The extent of the orbital region on the left side consists of part of the ascending process of the palatine posteriorly and several isolated plates of bone separated by distinct gaps and holes that form part of the orbital wall anteriorly (Figure 4.8). Based on their position and morphology these bone fragments likely represent part of the orbital process of the lacrimal and the anterior part of the ascending process of the palatine. No clear suture is visible between the palatine and the maxilla in lateral view. The gaps between these plates of bone, based on their location and size, were likely once foramina whose boundaries were artificially enlarged by the destruction of the fragile bone surrounding the rim of the foramina and the sutures between bones due the specimen’s preservation in a lahar flow.

Figure 4.8. Left orbital region of the holotype of Eomakhaira molossus (SGOPV 3490) in oblique posterolateral view. Fragment of the facial process of the lacrimal in teal, palatine in blue, all other bones of the cranium in purple. Abbreviations: laf, lacrimal foramen, mpf, minor palatine foramen; spf, sphenopalatine foramen. Two distinct openings are visible between the fragments of the anterior orbit. The more posterior, larger opening, which opens approximately dorsal to M4, compares well to the sphenopalatine foramen. Compared to other sparassodonts for which the position of the sphenopalatine foramen could be observed, this foramen in Eomakhaira is slightly located more anterior than in Australohyaena antiquua and Arctodictis sinclairi (though more similar to the former), but more posterior than in Patagosmilus goini, in which the sphenopalatine foramen opens much further anteriorly over M2 (Forasiepi and Carlini,

2010). According to Forasiepi and Carlini (2010) the location of the sphenorbital foramen in most sparassodonts, including the thylacosmilid Thylacosmilus, more closely

resembles the condition in Australohyaena and Arctodictis than Patagosmilus, which is

considered an autapomorphy of the genus.

The anteriormost of the fragments of the orbital wall appears to represent a small

portion of the orbital process of the lacrimal. At the anterior end of this fragment is a

small, partially preserved canal that appears to correspond with the lacrimal foramen.

Based on the position of this foramen and the surrounding elements, the lacrimal foramen

appears to have opened inside the orbit. This compares well with what is known in all

other sparassodonts, in which there is a single lacrimal foramen (except in Lycopsis

padillai and Arminiheringia, where there are two, and Callistoe, which is polymorphic in this feature; Suarez et al., 2016) that opens within the orbit. Of course it is not possible to say for certain whether one or two foramina were present due to the limited preservation of the lacrimal.

SGOPV 3490 also preserves parts of the jugal on both sides of the skull. On the left side, the jugal is only represented by a fragment of “floating bone” near the orbital region. Part of the jugal is also likely preserved on the right side based on the extent of bone preserved (i.e., representing parts of the rostrum formed by the anterior jugal in other metatherians) and the presence of remnants of a marrow cavity dorsal to the upper molars, but the specimen is too damaged to determine the shape and location of the maxillo-jugal suture.

4.4.2 Mandible

The mandible of Eomakhaira is robust and deep (Figure 4.10). However, the dentary of this taxon is relatively shallower than in the proborhyaenids Callistoe and

Arminiheringia (Figure 4.5, Appendix 27), though it is deeper than in Proborhyaena and

Paraborhyaena. It is possible that the depth of the dentary could be overestimated due to breakage, but by exactly how much is unclear. Like Callistoe, Arminiheringia, and

Pharsophorus tenax, the horizontal ramus is shallower anteriorly and greatly increases in depth posteriorly in the tooth row, as opposed to the borhyaenids Australohyaena and

Arctodictis, the proborhyaenids Proborhyaena and Paraborhyaena, and the basal borhyaenoid Pharsophorus lacerans where the ventral border of the horizontal ramus of the dentary is flat in lateral view and almost uniform in depth across the jaw and the mandible has a distinct “chin” (Figure 4.9). Indeed, the shape of the horizontal ramus in

Eomakhaira shows a particular resemblance to that of Arminiheringia (MACN-A 10970).

The deepest point of the mandible is estimated to have been below m4 based on the curvature of the horizontal ramus, similar to that of most other sparassodonts.

Figure 4.9. Line drawings of the horizontal rami of Callistoe vincei (A) modified from Babot et al. (2002), and Arctodictis sinclairi (B) modified from Forasiepi (2009), showing the curved horizontal ramus in Callistoe and the flat horizontal ramus with “chin” in Arctodictis. Figures not to scale.

Figure 4.10. Mandible of the holotype of Eomakhaira molossus (SGOPV 3490). A, C, E; Right dentary in labial (A), lingual (C), and occlusal (E) view. B, D; left dentary in labial (B) and lingual (D) view. Abbreviations: cor. process, coronoid process of dentary; ling. sulcus, lingual sulcus of the lower canine; men. foramen, mental foramen; post. symp., posterior border of the mandibular symphysis. Anterior is to the right in A, D, and E and to the left in B, C. The mandibular symphysis of SGOPV 3490 appears to have been unfused based on several lines of evidence. First, the two dentaries are displaced anteroposteriorly relative to one another, with the left being positioned slightly anterior relative to the right, something that would not be possible if the symphysis is fused. The anterior portion of the right hemimandible is missing but the left preserves a straight ventral edge, rather

than a jagged break as would be expected if the symphysis was fused and broken

postmortem. The medial face of the left dentary also shows a small amount of smooth

surface bone texture, which would not be present if the symphyseal surface was

obliterated by fusion. Finally, in sparassodonts with fused rami, when the dentary does

break it tends to break immediately posterior to the mandibular symphysis (e.g., MACN-

A 706, Prothylacynus patagonicus; MLP 85-VII-3-1, Arctodictis sinclairi; UATF-V-

000129, Paraborhyaena boliviana) at the point where the dentary is least reinforced, or

produce a much more uneven break rather than separating exactly at the reinforced, fused

symphysis. Although the mandibular symphysis of sparassodonts is known to become

increasingly rugose or even fuse with age, the advanced ontogenetic age of SGOPV 3490

(see below) indicates that the absence of a fused symphysis in this taxon is not due to its

ontogenetic status. Although it is not possible to tell how rugose the symphyseal surface

was, it is clear that the symphysis was neither fully fused, as in proborhyaenids and

Arctodictis, nor characterized by well-developed interdigitating ridges, as in Borhyaena.

Based on the curvature of the medial face of the mandible which in most mammals

typically denotes the posterior edge of the mandibular symphysis, the symphysis in

SGOPV 3490 extended to the level p2/p3 embrasure, or at most just slightly below the anterior root of p3 (Figure 4.10). This is true regardless of the orientation of the cranium.

This is rather short compared to closely related sparassodonts. In most proborhyaenids, as well as the borhyaenids Arctodictis sinclairi and Arctodictis munizi, the symphysis typically extends to the level of the p3/m1 embrasure, and in Arminiheringia auceta it actually extends below the level of the molars (Babot et al., 2002; Zimicz, 2012). In

Australohyaena antiquua and Prothylacynus patagonicus, the symphysis extends to the

level of the posterior root of p3. In Borhyaena macrodonta, Borhyaena tuberata, and

Pharsophorus lacerans (including MPEF-PV 4190, the specimen of Pharsophorus cf. P. lacerans from La Cantera) the symphysis extends to the midpoint between the two roots of p3. In Plesiofelis and Pharsophorus tenax, the symphysis extends below the anterior root of p3, but it does not extend as far posteriorly as in SGOPV 3490. Goin et al. (2007) described the symphysis of Acrocyon riggsi as extending to the anterior root of p3, but the specimen these authors examined (MLP 85-II-1-1) could not be examined first-hand to compare its morphology with the other taxa examined. In Thylacosmilus atrox and

Anachlysictis gracilis, the symphysis is much shorter than in all of the aforementioned taxa, ending below the canines.

Figure 4.11. Symphyseal region of the holotype of Eomakhaira molossus (SGOPV 3490) in anteroventral view, showing the smooth medial border on the anteriorly displaced left dentary. No scale available.

The mandible of SGOPV 3490 has only two mental foramina, one beneath p2-3 and the other below the m1–2 embrasure. The opening of the second foramen is well- defined on both sides and points posteriorly, whereas the anterior foramen is not well-

90 preserved on either side of the mandible but its existence and position can be inferred by the bone fragments beneath the premolar row. The presence of only two mental foramina in Eomakhaira is rather unusual for a borhyaenoid sparassodont. In general, most mammals have only one or two mental foramina, with individual variants of some species having three (Voss and Jansa, 2009). By contrast, most hathliacynids, borhyaenoids, and the basal sparassodont Stylocynus have three or more mental foramina, with individuals of some species having as many as five or six (Acrocyon riggsi, Arctodictis sinclairi,

Australohyaena antiquua, Borhyaena tuberata, and Lycopsis longirostrus). In addition, in most of these species the anterior mental foramen is greatly enlarged compared to the posterior foramen/foramina. Although it is possible that SGOPV 3490 had more mental foramina than the two preserved here, the location of the preserved bone fragments of the dentary and the distribution of mental foramina in other sparassodonts (typically between p2 and m1/2) make this unlikely to be the case. Because the opening of the anterior mental foramen is not well preserved on either side of the mandible, it is not possible to determine if this foramen was much larger than the posterior mental foramina as in these other forms.

A significant portion of the coronoid process is preserved on the left side of the skull, indicating that this process was well-developed and tall. However, due to crushing and distortion of the left side of the skull, it is not possible to make observations on the shape of the masseteric fossa or calculate the angle between the anterior border of the coronoid process and the tooth row based on the left side of the specimen. A small portion of the ascending ramus is also preserved on the right side of the specimen, which unlike the left does not appear to have undergone significant taphonomic deformation.

Assuming there is no deformation of the right side of the skull, this suggests the anterior

border of the coronoid process formed an angle of approximately 110-113° to the tooth row, which is within the range of variation typically seen in sparassodonts (Forasiepi,

2009)

4.4.3 Dentition

The only potentially preserved part of the precanine dentition is a small cylindrical tooth fragment resembling part of an incisor appressed to the lingual side of the right canine, in a similar position to the i3 of other sparassodonts. If this fragment does represent part of a lower incisor, then the lower incisors of Eomakhaira molossus

would be proportionally smaller than those of Arminiheringia auceta, Arctodictis

sinclairi, and potentially even Australohyaena antiquua and Paraborhyaena boliviana

(scaling by both p3 and m4), though the incisors are not as small as in MLP 77-VI-13-1,

a specimen assigned to Arctodictis sinclairi that has been noted to have relatively small

teeth compared to other specimens of this taxon (Goin et al., 2007).

Perhaps the most immediately noticeable feature of the holotype of Eomakhaira is

its large, robust upper and lower canines (Table 4.2). The canines of Eomakhaira are

disproportionately large compared to most sparassodonts, only comparable to

proborhyaenids, thylacosmilids, and the borhyaenids Australohyaena, Acrocyon, and

Arctodictis in terms of relative size (Table 3), all of which (except for Acrocyon) have

been noted to have relatively large canines compared to other sparassodonts (Engelman

and Croft, 2014). It is not clear why Paraborhyaena boliviana is recorded as having

relatively small canines compared to other proborhyaenids. This may be an actual feature

of the species, or it may be due to the fact that the specimen examined (MNHN SAL 51)

has a highly damaged M3. Both the upper and lower canines of Eomakhaira have a well-

developed median sulcus on their lingual side, which makes the canines appear somewhat

reniform in cross-section (Figure 4.10d, Figure 4.12b, c). By contrast, there is no sulcus on the labial side of the lower canine or the exposed lingual surface of the upper canine.

A very shallow lingual sulcus is present on the basal third of the lingual surface of the upper canine, but is not present on any exposed portion of this tooth.

Table 4.2. Dental measurements of the holotype of Eomakhaira molossus (SGOPV 3490) in mm. Measurements are reported to the nearest tenth of a millimeter due to uncertainties in measurement. * = estimated measurement.

Upper Dentition Lower Dentition Left Right Left Right Length 11.4 11.7 Length 11.0 — C c Width 6.6 6.9 Width 5.8 — Length — 5.2 Length 4.1 — P1 p1 Width — 2.6 Width 2.5 — Length 6.2 — Length 6.6 6.5 P2 p2 Width 3.2 — Width 4.3 4.3 Length 8.6 9.1 Length 7.5 7.6 P3 p3 Width 4.9 5.0 Width 4.2 4.1 Length — * Length — 7.9 M1 m1 Width — — Width — 4.2 Length — * Length — 9.0 M2 m2 Width — — Width — 5.3 Length 7.9 — Length — 10.2 M3 m3 Width 5.1* — Width — 5.5 Length 4.1 — Length — 12.0 M4 m4 Width 7.1 — Width — 6.3

Table 4.3. Upper canine proportions of SGOPV 3490 compared to other sparassodonts, focusing on canine shape (ratio of anteroposterior length/labiolingual width) and relative canine size. * = juvenile individuals. For raw measurements and how relative canine size was calculated see Appendix 12.

L/W Relative Taxon Specimen Family Ratio Size Eomakhaira molossus SGOPV 3490 Proborhyaenidae 1.71 1.12 Arminiheringia auceta MACN-A 10972 Proborhyaenidae 1.48 1.04 Callistoe vincei PVL 4187 Proborhyaenidae 1.38 1.28 cf. Proborhyaena MLP 79-XII-18-1 Proborhyaenidae 1.75 — Paraborhyaena boliviana MNHN SAL 51 Proborhyaenidae 1.45 0.85 Proborhyaenidae indet. MHNT-VT-1400/1401 Proborhyaenidae 1.45 0.64 Thylacosmilus atrox MLP 35-X-4-1 Thylacosmilidae 2.45 1.13 Thylacosmilus atrox FMNH P14531 Thylacosmilidae 2.50 1.11 Thylacosmilus atrox MMP 1470 Thylacosmilidae 2.63 0.84 ?Thylacosmilidae sp. nov. IGM 251108 ?Thylacosmilidae 1.56 — cf. Dukecynus sp.* UCMP 32950 Basal Borhyaenoidea 1.41 — cf. Pharsophorus* AMNH 29591 Basal Borhyaenoidea 1.55 — Lycopsis longirostrus* UCMP 38061 Basal Borhyaenoidea 1.27 0.51 Pharsophorus lacerans MNHN SAL 96 Basal Borhyaenoidea 1.35 — Prothylacynus patagonicus MACN 11453 Basal Borhyaenoidea 1.45 0.75 Prothylacynus patagonicus* MACN-A 5931 Basal Borhyaenoidea 1.26 0.67 Hondadelphys fieldsi UCMP 37960 Basal Sparassodonta 1.81 0.80 Acrocyon riggsi FMNH P13433 Borhyaenidae 1.36 1.00 Arctodictis munizi MLP 11-65 Borhyaenidae 1.60 1.29 Arctodictis munizi CORD-PZ 1210-1/5 Borhyaenidae 1.46 1.27 Arctodictis sinclairi MLP 85-VII-3-1 Borhyaenidae 1.30 1.16 Australohyaena antiquua UNPSJB-PV 113 Borhyaenidae 1.34 1.12

Australohyaena antiquua FMNH P13633 Borhyaenidae 1.42 — Borhyaena macrodonta MACN 52-390 Borhyaenidae 1.51 0.92 Borhyaena tuberata MACN 6203-6265 Borhyaenidae 1.47 0.85 Borhyaena tuberata MACN 5780 Borhyaenidae 1.52 — Borhyaena tuberata YPM-VPPU 15701 Borhyaenidae 1.22 0.99 Borhyaena tuberata YPM-VPPU 15120 Borhyaenidae 1.45 0.86 Acyon myctoderos MNHN-Bol-V-003668 Hathliacynidae 1.42 0.60 Borhyaenidium riggsi FMNH P14409 Hathliacynidae 1.43 0.46 Cladosictis centralis MACN 11639 Hathliacynidae 1.64 0.84 Cladosictis patagonica MACN 5927 Hathliacynidae 1.49 0.91 Cladosictis patagonica MACN 6280-6285 Hathliacynidae 1.42 0.62 Cladosictis patagonica AMNH 9134 Hathliacynidae 1.52 0.72 Cladosictis patagonica YPM-VPPU 15046 Hathliacynidae 1.50 0.71 Cladosictis patagonica YPM-VPPU 15170 Hathliacynidae 1.64 0.83 Cladosictis patagonica YPM-VPPU 15702 Hathliacynidae 1.43 0.95 Notogale mitis YPM-VPPU 21871 Hathliacynidae 1.47 — Sipalocyon externa MACN-A 52-383 Hathliacynidae 1.44 0.72 Sipalocyon gracilis MACN-A 692 Hathliacynidae 1.51 0.79 Sipalocyon gracilis YPM-VPPU 15373 Hathliacynidae 1.48 0.61 Sipalocyon gracilis AMNH 9254 Hathliacynidae 1.55 0.80 Sipalocyon gracilis YPM-VPPU 15029 Hathliacynidae 1.45 0.71 Sipalocyon gracilis YPM-VPPU 15154 Hathliacynidae 1.43 0.91 Sparassodonta gen. et sp. nov. UF 27881 incertae sedis 1.22 0.90 Sparassodonta indet. MUSM 1649 incertae sedis 1.26 —

The presence of median canine sulci were considered to be a distinguishing

feature of proborhyaenids or a clade of proborhyaenids + thylacosmilids by Babot et al.

(2002). However, observations show that median canine sulci are much more widely

distributed in sparassodonts than previously noted, also being present in the borhyaenids

Australohyaena antiquua, Arctodictis sinclairi, Arctodictis munizi, and Borhyaena

macrodonta, the basal borhyaenoid Pharsophorus lacerans (as seen in the holotype

MACN-A 32-391, YPM-VPPU 20551, and MPEF-PV 4190, a specimen assigned to

Pharsophorus cf. P. lacerans from La Cantera; Patterson and Marshall, 1978; Goin et al.,

2010), and an indeterminate sparassodont from the Fitzcarrald Arch (Tejada-Lara et al.,

2015). Sinclair (1930) and Marshall (1978b) reported lingual sulci in Acrocyon riggsi, but

I could not observe these in this taxon firsthand. In addition, in Thylacosmilus, a species previously considered to lack sulci on the upper canines, a completely extracted upper canine of one specimen (FMNH P14531) appears to show a shallow median sulcus at the extreme base of the lingual side of the canine, even though the extra-alveolar portions of the tooth are completely flat. The lower canines of Thylacosmilus have median sulci on both their labial and lingual faces (Goin and Pascual, 1987). In all of these cases, the sulci are either less prominent on (e.g., proborhyaenids, Arctodictis munizi) or are outright absent from the labial surface (all other taxa).

The surface of the canines are smooth (Figure 4.12a), contrary to the condition seen in most borhyaenoid sparassodonts in which the roots of the canines are covered in a series of small longitudinal grooves, which may extend close to the apex of the canine in forms such as Arminiheringia and Proborhyaena (Figure 4.13a, b). This is a feature

which has been observed in nearly all borhyaenoid taxa, with the exception of

thylacosmilids (Figure 4.13c) and possibly Lycopsis viverensis.

Figure 4.12. Photograph (A) and CT images (B-C) of the right upper canine of the holotype of Eomakhaira molossus (SGOPV 3490). A, Right upper canine of SGOPV 3490 in lateral view, showing the bluntness of the canine apex, and the absence of enamel, longitudinal ridges, and labial median canine sulci. B, transverse section of the canine just dorsal to the exposure of the upper canine (the actual point at the level of exposure is obscured by a crack), showing the posterior ridge of this tooth. The anterior root of P1 (aP1) is denoted to show that ridge is not an artifact of postmortem damage. C, transverse section of the canine slightly apical to the point of exposure of the upper canine, showing the median labial keel and the lingual median sulcus of this tooth. Approximate location of sections in B-C denoted by arrows on A. Anterior is to the right in all images and lingual is to the right in B-C. Abbreviations: aP1, anterior root of P1; mk, labial medial keel; ms, lingual median sulcus; pr, posterior ridge. Scale = 1 cm in A and 5 mm in B-C. The upper canines of are not completely ovate in cross-section. The lingual side is slightly flatter than the labial side, but not to the degree seen in species like

Thylacosmilus atrox (Riggs, 1934). The labial surface of the right upper canine actually shows a slight keel (Figure 4.12c), less pronounced than but similar to what has been

described in Thylacosmilus (Riggs, 1934; Turnbull, 1978), though this feature is not present on the left upper canine. There is a small carina on the posterior edge of the tooth, which is slightly easier to see in the left canine than the right (Figure 4.12b). This carina is not present on the lower canines, and does not appear to be present in the upper canines of Arminiheringia auceta, MLP 79-XII-18-1, or any borhyaenid that could be observed.

By contrast, the upper canines of thylacosmilids, like most saber-toothed mammals, are dominated by well-defined carinae that give the tooth a knife-like appearance. In

Thylacosmilus both the anterior and posterior faces of the upper canines have well- defined edges (though the posterior is much sharper than the anterior; Riggs, 1934;

Turnbull, 1978; Goin and Pascual, 1987), whereas in Patagosmilus the upper canine has a blunt anterior face and a sharp posterior ridge (Forasiepi and Carlini, 2010), more similar to the condition in Eomakhaira.

The upper canines of Eomakhaira are relatively mediolaterally compressed compared to other borhyaenoid sparassodonts (Table 4.3). The exact length/width ratio of the palate is highly dependent on the orientation of the specimen, ranging from 1.50-1.77 based how the cranium is oriented. However, the lower part of this range is likely an underestimate as it requires a more oblique orientation of the palate and tooth roots compared to the orientations that produce the higher estimates, with the actual L/W ratio of the canines probably between 1.68-1.71. Another issue is the labial median keel on the right canine, which makes the canines appear much wider based on simple measurements of maximum anteroposterior length/mediolateral width than if the keel was absent and the teeth were more ovate. Regardless of the orientation used, the upper canines of

Eomakhaira are more mediolaterally compressed than most other borhyaenoids,

including Pharsophorus, Prothylacynus, the borhyaenids Acrocyon, Arctodictis,

Australohyaena, and Borhyaena, the possible thylacosmilid IGM 251108, the

proborhyaenids Arminiheringia, Paraborhyaena, Callistoe, and the indeterminate

proborhyaenid from the Tremembe Formation (Couto-Ribeiro, 2010), but less than in

Patagosmilus, Thylacosmilus, or cf. Proborhyaena (MLP 79-XII-18-1). In terms of non- sparassodont carnivores, the canine proportions of Eomakhaira molossus are comparable to the machaeroidine creodont Machaeroides eothen (Gazin, 1946), the nimravids

Dinictis felina and Nimravus brachyops (Barrett, 2016) and the machairodontine felid

Pseudaelurus quadridentatus (Antón et al., 2012). The similarity to the latter three is rather noteworthy, as these taxa are considered to be sabertooths yet show some of the least amount of specialization for a sabertooth lifestyle within their respective clades

(Meachen-Samuels, 2012; Antón, 2013).

The state of the roots of the canines is difficult to determine. Based on CT- imaging the canine roots of SGOPV 3490 were closed at the time of death, but several features suggest that they may have been open in adult individuals of Eomakhaira and their closure is related to the advanced ontogenetic age of the holotype, as has been suggested for the proborhyaenid Proborhyaena (Bond and Pascual, 1983; but see Babot et al., 2002). First, the canines of SGOPV 3490 lack any trace of enamel, being entirely composed of dentine. This differs from what is seen in sparassodonts with non- hypselodont canines (e.g., borhyaenids), where the apex of the canine retains some traces enamel on non-occluding surfaces (such as the lateral side of the upper canines) even in old individuals. Although it is possible that the enamel could be completely lost on one canine due to wear and post-mortem damage the chances of the enamel being lost on all

99 four canines is less likely. This is especially true on the right side of the specimen, where enamel is absent on both the upper and lower canines but present on the right P1 and p2 less than a few millimeters away. This is very similar to what is seen in proborhyaenids

(which have open-rooted canines), where enamel is typically absent on the canines but present on the postcanine teeth. The absence of enamel in and of itself is not indicative of open-rooted canines, but when combined with other evidence it is highly suggestive.

Figure 4.13. Comparison of borhyaenoid sparassodont canines. A, Borhyaena tuberata (MACN-A 6203), showing the plesiomorphic condition of the canines in borhyaenoid sparassodonts; B, Arminiheringia auceta (MACN-A 10972), showing the typical condition of the canines in proborhyaenids; C, Thylacosmilus atrox (MLP 35-X-4-1), showing the condition of the canines in thylacosmilids. Note the short height of the canine in Borhyaena (alveolar border marked by the white dotted line, specimen is rotated to better show root morphology) related to breakage and heavy wear in this individual, and the presence of enamel despite this heavy wear. Abbreviations: long. grooves, longitudinal grooves on canine root; med. keel, labial median keel; med. sulcus, labial median sulcus.

The upper canines of SGOPV 3490 are also relatively tall, especially given their extreme degree of wear. By contrast, in specimens of many other sparassodonts that show

100

heavy tooth wear and do not have open-rooted canines in adulthood (e.g., borhyaenids) the canines are shorter due to breakage and/or wear (Figure 4.13a). Indeed, in terms of relative canine height, the exposed portion of the upper canine in SGOPV 3490 is significantly longer than that of other sparassodonts in which the upper canine is worn or broken in vivo (e.g., MACN-PV 14453, Prothylacynus patagonicus; MACN-A 5780 and

MACN-A 6203, Borhyaena tuberata), only comparable to specimens in which the canine shows little apical wear (especially among borhyaenoids) and to a lesser extent those where the canine is known to be hypselodont (i.e., Callistoe and Arminiheringia, which have much longer crowns than any of the other taxa examined). Indeed, the worn canines of SGOPV 3490 are actually longer than those of some individuals in which the canines show little apical wear (e.g., FMNH P13433, Acrocyon riggsi; YPM-VPPU 15373,

Sipalocyon gracilis; MNHN-Bol-V-003668, Acyon myctoderos; UCMP 38061, Lycopsis longirostrus, though it is possible the upper canine of this last individual is not fully erupted, see (Forasiepi and Sánchez-Villagra, 2014). Although the canines of SGOPV

3490 are blunt, they show no signs of in vivo breakage, suggesting their rounded apices are simply due to extreme wear.

Finally, on the lower left canine, the canine with the best preserved root, there is some evidence that the root of this tooth has not completely closed. Although the pulp cavity throughout most of the tooth is very narrow, the base of the canine has a flared end and is characterized by a deep depression connected to the pulp cavity (Figure 4.14). The absence of the basal tip of the canine does not appear to be due to damage, as the base of the alveolus is surrounded by bone and the edges of the depression are smooth and rounded rather than jagged as would be expected if the basal tip of the canine broke off or

101

was destroyed by the heat of the lahar flow. Given the advanced ontogenetic age of the

animal, this supports the idea that the roots of the canines had been open during much of

the adult life of the animal and had only recently closed. It is not clear if the same is true

of the upper canines and right lower canine, the right lower canine is shattered and the

bases of the upper canines cannot be easily distinguished from the surrounding matrix and bone and it cannot be unequivocally determined if the roots were open, closed, or in the process of closing. Comparisons with proborhyaenids are difficult as although CT scans have been performed on Proborhyaena, Arminiheringia, Callistoe, and

Thylacosmilus (Goin and Pascual, 1987; Babot et al., 2002), complete datasets have never

been published. Given the conflicting observations in SGOPV 3490 and the obfuscating

effects of ontogeny, the state of the canine roots in normal (i.e., non-senescent adult

individuals of this taxon) cannot be determined with certainty.

Figure 4.14. Morphology of the left lower canine root in the holotype of Eomakhaira molossus (SGOPV 3490). A, CT reconstruction of the root of the left lower canine of SGOPV 3490 in oblique lateral view, showing the prominent depression on the proximal end of the canine and its connection to the pulp cavity. B, Oblique lateral slice of SGOPV 3490 showing the flared, smooth distal end of the left lower canine. The canine roots of Eomakhaira appear to be less curved and less extensive than

in other sparassodonts. This is more pronounced in the lower canines than the upper ones.

102

In Eomakhaira, the roots of the upper canines end somewhere above P3. By contrast, in

Callistoe and Arminiheringia the cross-section of the canine roots is still relatively large above P3, suggesting the canines extended slightly more posteriorly in these taxa (Babot

et al., 2002). A similar condition is suggested by the description of a specimen of

Australohyaena (UNPSJB PV 113; Forasiepi et al., 2015). The preserved portion of the

lower canines are nearly subvertical, similar to what has been described for most

borhyaenoids with the notable exception of Arminiheringia, in which the lower canines

are procumbent. The roots of the lower canines are much shorter and more vertical than

the upper canines, ending either at the level of the p1-2 embrasure or at most ending

posteriorly below the anterior root of p2. This is different in from what is seen in

Callistoe, Arminiheringia, Proborhyaena, and Arctodictis in which the roots of the lower

canines extend posteriorly at least to the level of p3 (Babot et al., 2002; Forasiepi, 2009), if not farther (in Arminiheringia they extend below the molar row; Zimicz, 2012), and in

Callistoe and Arctodictis are more curved than in Eomakhaira.

Eomakhaira exhibits the normal metatherian postcanine dental formula of three premolars and four molars. The alveolar margin of the dentary is higher on the lingual side than the labial side, a feature which has been noted in some other sparassodonts

(Forasiepi et al., 2015). As all in other sparassodonts, the premolars increase in size from

P/p1-3. However, the relative size of the premolars differs between the upper and lower tooth row. In the upper dentition, P3 is very large and P1-2 are much closer in size. By contrast, in the lower dentition, p2-3 are more similar in size and p1 is distinctly smaller.

However, in both the upper and lower dentitions, P/p1 are very small. The base of the lower premolar row and the border of the dentary at the alveoli descend sharply in height

103 from p1-3. This is a common feature in sparassodonts, also present in Prothylacynus patagonicus, Pharsophorus cf. lacerans (MPEF-PV 4170), Borhyaena macrodonta,

Arctodictis sinclairi, Australohyaena antiquua, Callistoe vincei, and Proborhyaena gigantea. The premolar row is relatively short compared to other sparassodonts

(Appendix 15). The only sparassodont that comes close to Eomakhaira in this regard is

Paraborhyaena, in which the premolar row also appears to be relatively short but potentially not as short as in this taxon.

The right P1 and both the left and right p1 are oriented obliquely to the tooth row at an angle of about 35°, but the left P1 is oriented nearly parallel to the tooth row. Given that three of the four premolars are oriented obliquely to the tooth row and the left maxilla is not as well-preserved as the right maxilla or left and right dentaries, it seems more likely that the orientation of the left P1 is due to postmortem deformation and the orientation of the other three P/p1 more closely reflects the orientation of these teeth in life. The idea that the orientation of the left P1 is not natural is supported by the fact that variation in the orientation of P/p1 to this degree in a single individual has otherwise never been documented in sparassodonts. The crown and roots of P/p2 of Eomakhaira are oriented in-line with the rest of the tooth row, as in Pharsophorus (though possibly not

MPEF-PV 4170 from La Cantera, which more closely resembles Arctodictis sinclairi in this respect), Callistoe, Arminiheringia auceta, and Borhyaena, rather than oblique to the tooth row as in Proborhyaena, Paraborhyaena, Arminiheringia contigua (MACN-A

10317) Arctodictis, Acrocyon, and Australohyaena.

The postcanine tooth row of Eomakhaira is straight and P3 is positioned parallel to the long axis of the tooth row, as in Pharsophorus, Borhyaena, and all proborhyaenids.

104

By contrast, in Arctodictis and Australohyaena the postcanine tooth row is not straight and P3 is oriented obliquely in the jaw (this is more pronounced in Arctodictis munizi and

Australohyaena antiquua than Arctodictis sinclairi). Acrocyon riggsi appears to show an

intermediate state between the two conditions; in the holotype (FMNH P13433) the P3 is

in line with the rest of the upper postcanine tooth row but the crown is slightly oblique.

However, Sinclair (1930) expressed doubts that this arrangement of the tooth row was

natural, suggesting that it may instead be an artifact of the heavily restored nature of the

holotype skull. FMNH P13433 is rather unusual in that the postcanine tooth rows are

parallel to the canine rather than diverging posteriorly to form a more triangular palate, a

feature which is seen in every sparassodont for which the midline suture is known except

Hondadelphys, which supports this interpretation.

The P/p3 are robust teeth, with broad, bulbous roots. However, these teeth are

proportionally narrower labiolingually than those of borhyaenids (Appendix 16), which

tend to be much more bulbous (average L/W ratio of 1.43, with the taxa that more closely

resemble SGOPV 3490 in other respects like Arctodictis and Australohyaena towards the

lower end of the spectrum), as well as species of Pharsophorus (P. lacerans, YPM-VPPU

20551; P. tenax, AC 3192). Compared to proborhyaenids, the P3 of Eomakhaira is more

sectorial than Callistoe vincei, but less so than MLP 79-XII-18-1, comparable to specimens of Arminiheringia. The p3 of Eomakhaira is also proportionally narrower labiolingually than most borhyaenids, the basal borhyaenoid Plesiofelis, and the

proborhyaenids Proborhyaena and Arminiheringia, comparable to the proborhyaenids

Paraborhyaena and Callistoe, but less sectorial than the basal borhyaenoids

Prothylacynus or Pharsophorus.

105

Figure 4.15. Oblique lateral slice of SGOPV 3490 (parallel to long axis of right tooth row), showing the position of the lower molars and the depth of the horizontal ramus. Note how the worn surfaces of P3 and M1 closely match one another. Abbreviations: post. root, posterior root. The P3 of Eomakhaira is longer than the p3 (Appendix 16), being nearly 13 to 19 percent longer than the lower tooth (depending on whether the left or the right P3 is used for comparison). This is unusual, as in most other sparassodonts that resemble

Eomakhaira (i.e., borhyaenids, proborhyaenids, and Pharsophorus, the P3 is typically

106

about the same length as p3 (with P3 ranging from 93% to 108% the length of p3). The

only borhyaenoid that may resemble Eomahaira in this respect is MLP 79-XII-18-1, a

specimen assigned to Proborhyaena by Bond and Pascual (1983) but considered to not be

a proborhyaenid by Babot et al. (2002) (referred to as cf. Proborhyaena in this work).

Based on its roots, the P3 of this specimen is massive, nearly 30 mm in length, much

longer than the p3 of the largest known specimen of Proborhyaena gigantea (AMNH

29576, in which it is ~24 mm long), despite MLP 79-XII-18-1 pertaining to a

comparatively smaller animal based on upper molar lengths. If MLP 79-XII-18-1 really does pertain to Proborhyaena, it implies that the P3 of this taxon was much longer than p3, though as mentioned above the assignment of this specimen to this taxon is controversial. Thylacosmilids (specifically Thylacosmilus, the only genus in this family for which the upper and lower dentition is known) also have a tooth at the P3 locus that is much longer than p3, but as this tooth represents dP3 rather than P3 (Goin and Pascual,

1987; Forasiepi and Carlini, 2010; Forasiepi and Sánchez-Villagra, 2014) comparisons with other sparassodonts may be uninformative in this respect.

Based on the orientation of the rostrum, p3 appears to have been inclined posteriorly (Figure 4.15). This is a common feature in borhyaenoid sparassodonts, occurring in the basal borhyaenoids Plesiofelis schlosseri and Pharsophorus lacerans, the borhyaenids Australohyaena antiquua, Arctodictis sinclairi, and Borhyaena macrodonta

(Marshall, 1978b; Forasiepi et al., 2015), the specimen from Los Helados described in

Chapter 3 (SGOPV 6200), and possibly the proborhyaenid Proborhyaena gigantea

(Engelman, pers. obs.) in addition to Eomakhaira. Close-ups of the left p3 show that the enamel is distributed much more apically on this tooth than in p2 or m1-2, differing from

107

what is seen in other sparassodonts (i.e., Pharsophorus, borhyaenids) but somewhat

resembling what is seen in specimens of Proborhyaena (AMNH 29576) and Callistoe

(PVL 4187). It is not clear if this is a natural feature of this specimen or is simply an

artifact of preservation.

Many borhyaenoid sparassodonts (specifically borhyaenids, proborhyaenids, and

closely related taxa) are characterized by lower postcanine teeth with extremely robust,

“bulbous” roots. The extent of this condition ranges from taxa in which bulbous roots

only being present on p3 (e.g., Borhyaena) to taxa in which bulbous roots are present on all premolars and some of the molars (e.g., Australohyaena antiquua and most proborhyaenids). The roots of p3 in Eomakhaira are extremely robust (Figure 4.15), to

the point that the anterior and posterior roots are almost appressed and there is little

interradicular space between them. However, it is difficult to compare the condition in

this specimen to other sparassodonts, because previous studies have typically defined the

roots as “bulbous” if they are wider than the crown in occlusal view (see Forasiepi,

2009), something which is not possible for SGOPV 3490 because the crown of p3 is not

completely preserved. Nevertheless, the overall morphology of p3 in SGOPV 3490 more

closely resembles sparassodonts in which the roots of p3 are bulbous than those in which

they are not. The condition of p3 in Eomakhaira resembles what is seen in the p2-m2 of

Arctodictis sinclairi, in which the roots of p2-m2 are so bulbous that there is virtually no interradicular space between the anterior and posterior alveoli of these teeth (Forasiepi,

2009: fig. 19). The crowns of the right p2 and m1 are preserved and on both of these teeth the roots are not wider than the crown, suggesting they were not bulbous. Nevertheless, the roots of the right p2 are still comparatively more robust than the roots of the molars.

108

The morphology of the upper molars is best preserved on the right M3-4, though the left M3 is also somewhat well-preserved. However, as mentioned above, these teeth are all highly worn. The M3 is so worn that it is essentially pyramidal in shape, with only a single, poorly-distinguished cusp. Based on comparisons with other sparassodonts, this cusp most likely represents the metacone. This worn cusp is located close to the edge of the tooth, suggesting the stylar shelf was extremely narrow. There is no evidence of an anterolabial cingulum, a paracone, or any stylar cusps, but it is not possible to tell for certain whether these structures were truly absent or merely obliterated by wear or damage to the teeth. There is a deep and very prominent ectoflexus on M3. The ectoflexus of Eomakhaira is considered “deep” following the criteria of Davis (2007) and

Williamson et al. (2012), in which the ectoflexus is considered “deep” if it is >10% and

“shallow” if it is ≤ 10% the labiolingual width of the tooth. Among short-snouted borhyaenoids, deep ectoflexi on M3 are present in Prothylacynus patagonicus,

Proborhyaena gigantea, and Callistoe vincei (Appendix 17), and Callistoe also has a deep ectoflexus on M2. By contrast, the upper molar ectoflexi are shallow or absent in

Pharsophorus tenax, all borhyaenids (Borhyaena spp., Arctodictis spp., Acrocyon riggsi, and Australohyaena antiquua) and Thylacosmilus atrox. Interestingly, the M3 ectoflexus of Patagosmilus goini is deeper than in borhyaenids and Pharsophorus tenax, almost at the threshold between “deep” and “shallow” (Appendix 17), in contrast to Thylacosmilus atrox and the indeterminate Colhuehuapian thylacosmilid described by Goin et al. (2007)

in which the ectoflexus is extremely shallow or absent. This raises the question as to

whether the loss of the ectoflexus is convergent in thylacosmilids, borhyaenids, and P. tenax. The roots of M3 are extremely splayed. A similar condition was observed in

109

Proborhyaena, Paraborhyaena, and MLP 79-XII-1-1, but it is possible this condition is present in additional taxa and was simply not able to be observed (especially in the absence of CT imaging).

Interestingly, despite the great ontogenetic age of the specimen, there is no evidence of carnassial rotation (sensu Mellett, 1969) in SGOPV 3490, as seen in the proborhyaenid Arminiheringia auceta (Marshall, 1978b). Carnassial rotation produces a

highly distinctive wear pattern on the dentition, resulting in completely flat posterolingual faces of M1-3 (extending from the metastyle to the protocone and sometimes exposing

the pulp cavity in ontogenetically older individuals, as in MACN-A 10970/10972) and anterolingual face of M4. These distinctive wear faces are not seen in SGOPV 3490. The lingual surface of the left M3 where the protocone should be is unusually flat, but the wear does not resemble what is seen in Arminiheringia or Hyaenodon. The flat lingual surface of the left M3 is obliquely oriented to the postmetacrista in SGOPV 3490, rather than parallel to it due to medial rotation of the tooth to maintain occlusion, and what is preserved of the postmetacrista does not suggest wear any different from that seen in most sparassodonts. The condition of the right M3 is hard to determine, but it looks like a matching surface is not present on this tooth the protocone may be better preserved.

Given the poor preservation of the upper molars in general (see above) this feature on the left M3 is probably better interpreted as damage sustained in vivo, during preservation, or as a result of preparation.

Nevertheless, the orientation of the right M3 and left M3-4 suggests the crowns of these teeth were canted medially in life. It is clear that this rotation is not an artifact of taphonomic distortion as the crowns are canted in complementary but opposing directions

110

on the left and right side (i.e., the crowns of both the left and right M3 are canted

medially). Similarly, the lower molar row is slightly oblique, such that the molars are

canted labially. This feature has been documented in some other sparassodonts, such as

Australohyaena, and observations show it to be present in some other taxa such as

Arminiheringia. This feature may be related to the occluso-lingual orientation of the upper molars present in many sparassodonts.

The left M4 is the best preserved of the upper molars. The morphology of this tooth is highly simplified with only two major cusps, the paracone and the stylar cusp B.

These cusps are linked by a well-developed preparacrista, which is so well-developed that the two cusps are not even distinct from the crest (though this may be due to exacerbation by wear). There is no anterior cingulum on M4. Similarly, there is no cingulum on the labial edge of the molar, as in some other sparassodonts with simplified

M4s such as Prothylacynus patagonicus (MACN-A 707). The metacone is absent. There is no vestigial preparacrista posterior to the paracone, as has been described for the M4 of some sparassodonts like Patagosmilus (Forasiepi and Carlini, 2010) and Arctodictis

(Forasiepi, 2009). The protocone is best preserved on this tooth and is highly reduced; barely even a small swelling of enamel on the lingual side of the paracone. This is different from what is seen in taxa such as Callistoe or Patagosmilus, where it is forms a

vestigial yet distinct cusp (Babot et al., 2002; Forasiepi and Carlini, 2010), or Borhyaena macrodonta (MACN-A 32-390) or Pharsophorus tenax (AC 3192), where it is vestigial yet retains a small basin. Unfortunately, the M4 of the largest proborhyaenids,

Paraborhyaena and Proborhyaena, is either unknown or was not available for comparison.

111

Despite its highly simplified morphology, CT imaging of the M4 of SGOPV 3490

indicates this tooth has three roots (Figure 4.16). Basally, there are three distinct roots

located labially, lingually, and posteriorly, but further apically the lingual and posterior

roots merge together until only two roots are distinct at the level of the crown. By

contrast, most sparassodonts with highly simplified ("linear") M4s are considered to have

only two roots on these teeth. These include nearly all short-snouted borhyaenoids for which the M4 is known, including Acrocyon riggsi, Arctodictis spp., Borhyaena spp.,

Callistoe vincei, Paraborhyaena boliviana, Patagosmilus goini, Pharsophorus tenax,

Prothylacynus patagonicus, and Thylacosmilus atrox. The only short-snouted borhyaenoid to definitely have three roots on M4 is Australohyaena antiquua, which is notable as phylogenetic analyses have recovered this taxon as deeply nested within a clade formed of taxa two-rooted M4s (Forasiepi et al., 2015). This raises the question if some of the other sparassodont taxa identified as having two roots on M4 really have two roots on this tooth, as the presence of a third root in M4 in SGOPV 3490 was only identified through CT imagery. Externally, without the aid of CT imagery, SGOPV 3490 appears to have two roots on M4, the same way in which this condition is ascertained in other forms.

112

Figure 4.16. Cross-section of the M4 of SGOPV 3490 in (A) oblique occlusal and (B) posterior view, showing the presence of three roots on this tooth. The lower molars of Eomakhaira do not appear to be strongly imbricated, at least

not to the degree seen in some other sparassodonts such as the borhyaenids

Australohyaena, Acrocyon, and Arctodictis or the thylacosmilid Thylacosmilus. The m3-4

are slightly angled relative to the tooth row, comparable to the degree of imbrication seen

in Arminiheringia and Proborhyaena, though moreso than in Paraborhyaena,

Pharsophorus (specifically the holotype MACN-A 52-391), and possibly Borhyaena

(based on MACN-A 52-366, assigned to Borhyaena macrodonta). The alveolar line of the lower molars ascends anteroventrally-posterodorsally at a significant angle to the frontal plane. Although at first glance this would seem to be an artifact of preservation,

especially given the absence of most of the alveolar bone, several factors suggest that this

reflects a real, rather than taphonomic, feature. First, rotating the specimen such that the

alveolar line is parallel to the frontal plane results in an impossible orientation of the

cranium, resulting in features such as the nasals pointing anterodorsally. Second, P3/m1

are preserved in near-occlusal position on the right side of the specimen, and unlike the

lower molars the alveolus of the right P3 is partially preserved. Finally, an anteroventral-

113 posterodorsal inclination of the tooth row is seen in a few other sparassodonts, including

Arminiheringia auceta, Paraborhyaena boliviana, Arctodictis sinclairi, Australohyaena antiquua, and possibly Acrocyon riggsi (Goin et al., 2007; fig. 8a). In these taxa, the alveolar line of the lower molars is oriented at an angle of about 8° to the frontal plane, comparable to what is seen when SGOPV 3490 is oriented in the position I consider to be closest to life position.

Little can be said about the two anteriormost molars (m1-2). As mentioned above the occlusal morphology of the right m1-2 has been obliterated by wear, whereas on the left side m1 is missing its crown and m2 is completely gone, possibly due to greater distortion on the left side of the skull. The only anatomical feature that can be determined in the anterior molars is that the posterior lobe of the crown was not lower than the anterior, as it is (to variable degrees) in all borhyaenids, Prothylacynus, Plesiofelis,

Pharsophorus, Arminiheringia, Proborhyaena, and Thylacosmilus. In this respect

Eomakhaira resembles Callistoe, which is the only other short-snouted borhyaenoid to lack this feature. The m3 is not much better, but it preserves slightly more morphology than the anterior teeth.

The m4 (specifically, the right m4) is the best preserved of the lower molars, and is the only lower molar in which the morphology of the crown has not been mostly obliterated by wear. Nevertheless, the shape of the paracristid on m4 shows that this tooth was also highly worn, comparable to the degree of wear seen in the holotypes of

Angelocabrerus daptes (MMP 967M; Simpson, 1970b) or a specimen assigned to

Pharsophorus cf. P. lacerans (MPEF-PV 4190; Goin et al., 2010). As in all borhyaenoid sparassodonts, the m4 of Eomakhaira is characterized by two main cusps; a tall

114

protoconid and a slightly shorter paraconid. The morphology of m4 more closely

resembles Proborhyaena or Paraborhyaena than any other sparassodont, with a

posteriorly salient protoconid located at the posterior end of the tooth and the talonid

being virtually absent. This contrasts with the anterior molars (m1-3), which show evidence of very small, if worn, talonids. This cusp is smaller than in Callistoe or

Arminiheringia, about the same size as in Proborhyaena or Paraborhyaena, but larger than in Arctodictis, in which the talonid is completely absent but the protoconid is not positioned at the posterior end of the tooth.

The metaconid is clearly absent on m4. Determining the presence or absence of the metaconid on m2-3 is more difficult due to the worn and “exploded” nature of the teeth, but it appears to be absent on at least m3. In borhyaenoid sparassodonts with a small metaconid (e.g., borhyaenids, Pharsophorus) there is often a small, low protuberance at the posterolingual corner of the tooth representing a small or worn metaconid (see figures 10-11 in Forasiepi et al., 2015). In SGOPV 3490, on the other hand, the posterolingual surface is smooth and the base of the protoconid extends to the lingual margin of the tooth; there is no distinct cusp that would suggest the presence of a metaconid. In fact, the posterior margin of m3 is very similar to the holotype of

Arminiheringia auceta (MACN-A 10970), being a simply dorsolingually-ventrolabially oriented ridge with no well-developed cusps.

The anterior root of the posterior lower molars (primarily m3-4) is much larger and more robust than the posterior one. This condition is seen in some other sparassodonts (see Appendix 18), including Borhyaena tuberata, Arctodictis sinclairi,

Acrocyon sectorius, Proborhyaena gigantea, Thylacosmilus atrox, and the proborhyaenid

115

from Antofagasta de la Sierra (based on direct observations of specimens via X-rays or isolated teeth), and may also be present in Arminiheringia auceta and Paraborhyaena boliviana (based on what can be observed of the roots at the alveolus). However, this condition is not present in the specimens of Australohyaena antiquua, Acyon myctoderos,

Cladosictis patagonica, Hondadelphys fieldsi, Notogale mitis, Pharsophorus lacerans

(based on direct observation), and possibly Stylocynus paranensis (based on inference)

observed. An anterior molar root that is much larger and more robust than the posterior

may prove to be an important phylogenetic or paleoecological character in sparassodonts, but a comprehensive survey of this feature is beyond the scope of this study. The degree of development of this feature also seems to vary between taxa: in Arctodictis sinclairi only m3-4 have unequally sized roots, whereas in Proborhyaena and Paraborhyaena this

feature is present on m2-4. In Eomakhaira, the disparity in size is present on m2-4 but is

much less pronounced on m2 (where the two roots are almost the same size) than on m3-

4. Interestingly, this feature is not very pronounced in the extant Sarcophilus harrissii.

The anterior root of m4 is more robust than the posterior one, but the two roots do not differ in their length and on m2-3 the roots are similar in height and robustness (Fiani,

2015). This does not appear to be related to bone-crushing habits as a similar condition also occurs in the carnassials of several non-durophagous carnivorans including felids and barbourofelids (Tseng et al., 2010; Kupczik and Stynder, 2012).

It is not clear whether a labial postcingulid was present in Eomakhaira. There is a small depression on the labial side of m1 (Figure 4.4), but given the wear on this tooth it is not clear if this is part of a labial cingulid that extends to the posterior end of the tooth or merely a depression on the posterolabial face of m1. Unfortunately, the m4, which is

116

the best preserved tooth in this specimen, sometimes lack labial postcingulids in other

sparassodonts even when the cingulid is preserved on m1-3 (e.g., Arctodictis), though in other species there is a postlabial cingulid on m4 (e.g., Pharsophorus lacerans, MACN-A

52-391). The presence of a labial postcingulid would not be unexpected given its distribution in sparassodonts. Labial postcingulids are absent in the basal borhyaenoids

Lycopsis, Pseudothylacynus, and Prothylacynus, but are present in cf. Nemolestes

(AMNH 29433; Forasiepi et al., 2015) Plesiofelis, Pharsophorus (both P. lacerans and P. tenax), all borhyaenids (Acrocyon spp., Australohyaena antiquua, Arctodictis spp., and

Borhyaena spp.), and the proborhyaenid Callistoe (Forasiepi et al., 2015). The presence or absence of a posterolabial cingulid could not be determined on the specimens of

Arminiheringia, Proborhyaena, or Paraborhyaena observed. The condition in thylacosmilids is not entirely clear. Photographs of the holotype of Anachlysictis gracilis in Goin (1997b) appear to show a posterolabial cingulid, whereas pictures of

Thylacosmilus (Riggs, 1934; Marshall, 1976a) show a structure that could be a posterolabial cingulid, but this is not clear. Interestingly, most phylogenies of sparassodonts (Forasiepi, 2009; Engelman and Croft, 2014; Forasiepi et al., 2015; Suarez et al., 2016) imply no less than six independent losses of the posterolabial cingulid (in

Hondadelphys, Stylocynus, at least twice in hathliacynids, in Lycopsis, and in

Prothylacynus), not counting taxa known from lower dentitions that have not been considered in phylogenetic analyses (i.e., Pseudothylacynus, Dukecynus), or else repeated

loss and reacquisition of the labial postcingulid within Sparassodonta. The evolutionary

lability of the postlabial cingulid in sparassodonts has important implications for broader

studies of phylogenetic relationships within Metatheria, in which the absence of the labial

117

postcingulid has been considered a synapomorphy of crown-group marsupials (reversed

in dasyuromorphians and Djarthia; Voss and Jansa, 2009; Maga and Beck, 2017), but an

in-depth discussion of this is beyond the scope of this study.

4.5 PHYLOGENETIC ANALYSIS

In order to determine the phylogenetic affinities of Eomakhaira, I performed a

phylogenetic analysis using a modified version (Appendix 20) of the most recent phylogenetic matrix of sparassodont interrelationships (Suarez et al., 2016) with several additional or revised characters and codings. The proborhyaenid Proborhyaena gigantea was also added to the analysis in order to better test relationships within the

Proborhyaenidae. The specimen MLP 79-XII-18-1 was not included in the coding of P. gigantea due to uncertainty in its taxonomic identity (see Babot et al., 2002). A complete list of changes from Suarez et al. (2016) can be found in Appendix 22. This matrix was compiled in Mesquite (Maddison and Maddison, 2008) and analyzed in TNT 1.1

(Goloboff et al., 2008) under both equal and implied weights. The use of implied weighting in phylogenetics has been criticized (e.g., Congreve and Lamsdell, 2016;

Madzia and Cau, 2017; but see Goloboff et al., 2017), but in this study implied weighting was primarily used as a means to test hypotheses. That is, if a topology was recovered in both the equal and implied weights tree, it can be considered to be more robustly supported than if the topology was recovered via only one method.

In contrast to previous analysis, which used Deltatheroida (Forasiepi, 2009) or

Deltatheroides cretacicus (Engelman and Croft, 2014; Forasiepi et al., 2015; Suarez et al., 2016) as the outgroup, I rooted the phylogenetic analysis on Holoclemensia texana

118 for several reasons. First, due to the constraints of TNT, only a single taxon can be defined as an outgroup. Constraining Deltatheroides as the outgroup means that the monophyly of Deltatheroida, which is generally agreed upon by most authors and phylogenetic analyses (e.g., Bi et al., 2015; Rougier et al., 2015), can never be tested, possibly affecting character state reconstructions. Secondly, several recent phylogenetic analyses of Metatheria and Theria have recovered Holoclemensia as a metatherian basal to the split between Deltatheroida and Marsupialiformes (Luo et al., 2003; Bi et al., 2015;

Rougier et al., 2015; Wilson et al., 2016; Carneiro, 2018) or as a basal therian or even eutherian (Vullo et al., 2009; Averianov et al., 2010; Carneiro and Oliveira, 2017). The taxonomic history of Holoclemensia is complex (see discussion in Beck, In press) but no study has ever unambiguously recovered Holoclemensia as a member of

Marsupialiformes. Although Holoclemensia is only known from dental remains

(Slaughter, 1968; Davis and Cifelli, 2011) placing it as the outgroup is probably more accurate than constraining it to be a member of Marsupialiformes, which may affect character state polarity. Constraining Holoclemensia or Deltatheroides to be the outgroup taxon produced no change in the arrangement of taxa within Marsupialiformes.

As in previous version of this matrix (Forasiepi, 2009; Engelman and Croft, 2014;

Forasiepi et al., 2015; Suarez et al., 2016), the thylacinid Thylacinus cynocephalus was coded in this matrix but Thylacinus was excluded from the analysis. Running the matrix with T. cynocephalus results in Thylacinus being recovered as the sister taxon to

Sparassodonta and Sparassodonta being recovered within Dasyuromorphia. However, this result is almost certainly spurious. Examination of the matrix shows many of the characters that purportedly support a sister group relationship between Thylacinus and

119

sparassodonts, such as the absence of a palatine torus (21[1]), a paracanine fossa formed

solely by the premaxilla (6[1]), a connate paracone and metacone (147[0]), and a

metaconid that is absent or smaller than the paraconid (180[1], 181[2], and 182[0]) are

apomorphies of Thylacinus or Thylacinus and its close relatives within Thylacinidae

(Murray and Megirian, 2000; Wroe and Musser, 2001; Yates, 2014), rather than

symplesiomorphies of the group. Furthermore, many of these features are specializations

related to carnivory, which are considered to be highly homoplastic in mammals (Polly,

1996; Muizon and Lange-Badré, 1997; Solé and Ladevèze, 2017). The recovery of a

Thylacinus + Sparassodonta clade is more likely due to the absence of morphologically

plesiomorphic thylacinids such as Nimbacinus and non-autapomorphic characters in this

matrix that link Thylacinus to other Australian taxa. A close relationship between

sparassodonts and thylacinids is not supported by tarsal (Szalay, 1994) and genetic

(Thomas et al., 1989; Krajewski et al., 1992; Miller et al., 2009) evidence, as well as the

fact that both genetic and total evidence phylogenies suggest that dasyurids and thylacinids last shared a common ancestor in the Oligocene (Kealy and Beck, 2017, and references therein) whereas several sparassodonts in this analysis come from the Eocene.

Tree analyses were performed in TNT using the “New Technology search” option using sectorial search, ratchet, tree drift, and tree fuse options under default parameters, finding the minimum length 1000 times and then analyzing the recovered trees under tree bisection reconnection branch swapping. The equal weights analyses produced four most parsimonious trees (MPTs) with a length of 1076 steps, a consistency index of 0.365, and a retention index of 0.689. A strict consensus of the MPTs under equal weights is shown in Figure 4.17. The implied weights analysis produced a single MPT with a best score of

120

116.96656, shown in Figure 4.18. In general, the overall topologies of the trees were similar with only minor differences between the equal weights and implied weights analysis. As a result, the results of both are discussed together below.

Figure 4.17. Results of the phylogenetic analysis under equal weights, showing the strict consensus of the four most parsimonious trees (MPTs). Numbers to the upper left of each node represent Bremer supports, numbers to the lower left represent bootstrap values. The overall topology of both the equal weights and the implied weights tree are very similar to that of previous analyses, with the notable exception of recovering a monophyletic Deltatheroida (which, as noted above, is related to not constraining a

121 deltatheroidan as an outgroup). Perhaps most importantly, Sparassodonta was not recovered as one of the most basal branches within Metatheria nor as a sister taxon to

Mayulestes, Pucadelphys, and Andinodelphys, contra Wilson et al. (2016) or Cohen (In press), in spite of the fact that this analysis includes data from the specimens of

Didelphodon vorax described by Wilson et al. (2016). Instead, sparassodonts are recovered crownward of Kokopellia, Asiatherium, Pediomyidae, and Alphadon. The three

Tiupampa taxa were recovered crownward of Sparassodonta as the sister taxon to

Peradectidae. In this respect, it is interesting to note that Goin (2003) suggested that may be closely related to peradectids and Carneiro (2018) recovered a close relationship between the Tiupampa taxa and Peradectidae. Herpetotherium was recovered in different positions between the equal weights and implied weights, either as the sister taxon to

Australidelphia (equal weights) or as basal to crown Marsupialia but more closely related to crown marsupials than sparassodonts and most other metatherian taxa (e.g.,

Pediomyidae, Alphadon, Asiatherium). Both positions have been recovered by previous phylogenetic analyses (Forasiepi, 2009; Horovitz et al., 2009; Beck, 2012; Engelman and

Croft, 2014).

In both the equal and implied weights analyses stagodontids were recovered as the sister taxon to Sparassodonta. Marshall et al. (1990) suggested a close relationship between stagodontids and sparassodonts, but this arrangement was criticized by later authors (Marshall and Kielan-Jaworowska, 1992; Fox and Naylor, 1995; Muizon, 1998).

Engelman and Croft (2014) also recovered a relationship between stagodontids and sparassodonts, but most phylogenetic analyses including members of both groups have not (Rougier et al., 1998; Forasiepi, 2009; Forasiepi et al., 2015; Suarez et al., 2016;

122

Carneiro and Oliveira, 2017; Carneiro, 2018). However, there is some reason to believe that the association between stagodontids and sparassodonts recovered in this study is spurious. Many of the derived characters that link stagodontids and sparassodonts in this analysis, including a lateral process of the premaxilla that extends beyond the anterior border of the canine (8[1]), premaxilla-nasal contact posterior to the canine (12[1]), nasals that do not extend over the narial opening (13[1]), well-developed sagittal crest

(74[0]), and a metaconid lower than the paraconid (182[0]), are also seen in dasyuromorphians (or thylacinids among dasyuromorphians, as is the case for a metaconid lower than the paraconid; Murray and Megirian, 2000; Yates, 2014) and may be convergent adaptations for carnivory. Recovering a topology where stagodontids and sparassodonts are not sister taxa takes three additional steps. The recovered association between stagodontids and sparassodonts may also be due to the fact that several taxa in this analysis such as Kokopellia, pediomyids, and Alphadon are primarily known from dental remains, especially since most of the characters that support this relationship are cranial (especially related to the premaxillae, which is not preserved in these taxa nor in

Asiatherium or peradectids).

The relationships between members of the Hathliacynidae in this analysis are slightly better resolved than in previous analyses. There has been little consensus as to the relationships of the member of this group both between phylogenetic analyses

(Muizon, 1999; Forasiepi et al., 2006; Forasiepi, 2009; Suarez et al., 2016) and even within a single study (Engelman and Croft, 2014; Forasiepi et al., 2015), to the point that the poor resolution of phylogenetic relationships within the Hathliacynidae have been noted by previous authors (Forasiepi et al., 2015). Two major groups of hathliacynids are

123 recovered in this analysis: a clade of the large-bodied hathliacynids Cladosictis and

Acyon and a polytomy between the three smaller-bodied taxa Sallacyon, Notogale, and

Sipalocyon. In the implied weight analysis this polytomy is resolved with Notogale and

Sipalocyon being recovered as sister taxa.

Figure 4.18. Results of the phylogenetic analysis under implied weights showing the single recovered most parsimonious tree (MPT). Numbers to the lower left represent bootstrap values. One of the primary causes of instability within the Hathliacynidae in this and previous analysis is the hathliacynid Sallacyon hoffstetteri. The position of this taxon has

124

been noted to be highly unstable in previous analyses, and in some studies the position of

this taxon is the only thing that differs between the topology of hathliacynids in the

recovered MPTs (Engelman and Croft, 2014). This instability is mostly due to the large

number of character states missing for this taxon. Despite being known from a specimen

preserving most of the cranium posterior to P3 (MNHN SAL 92), this specimen has

almost never been figured or examined in detail since its initial description by Petter and

Hoffstetter (1983) (and even the figures and photos in this publication do not show

enough anatomical detail to evaluate many features in this taxon) and so many features

that may be present in this specimen cannot be evaluated, including several features of

the posterior rostrum, dentition, and braincase. Muizon (1999) described the braincase of

the specimen, but primarily focused on the auditory region. If Sallacyon is removed from

the analysis, the topology of Hathliacynidae in this study is the same as that of Forasiepi

et al. (2006), Forasiepi (2009), Engelman and Croft (2014), and the implied weights

analysis of Forasiepi et al. (2015). A re-examination of MNHN SAL 92, as well as the

description of additional specimens from Salla which preserve parts of the rostrum

anterior to P3 and appear to show unusual features such as a cuspidate metacone on M4

(Anaya Daza et al., 2010; Engelman pers. obs.) would be of great help in elucidating the

phylogenetic position of this early hathliacynid.

Eomakhaira molossus was recovered as a member of the Proborhyaenidae in both analyses. However, rather than recovering a monophyletic Proborhyaenidae, proborhyaenids were found to be paraphyletic relative to Thylacosmilidae in both the equal and implied weights tree, with Eomakhaira being recovered as the sister group of undoubted thylacosmilids (Patagosmilus, Proborhyaena). This relationship is not due to

125

the presence of Eomakhaira, as removing this taxon from the matrix and re-running the analysis recovered the same arrangement. The clade of Eomakhaira + Thylacosmilidae was found to be sister to either Proborhyaena gigantea or a clade of Paraborhyaena +

Proborhyaena. Interestingly, despite Paraborhyaena, Proborhyaena, and the clade of

Eomakhaira being recovered as a polytomy in the equal weights analysis, Paraborhyaena

and Proborhyaena were recovered as sister taxa with relatively high support in the

implied weights analysis. Callistoe vincei was recovered as the most basal member of the

Proborhyaenidae (including thylacosmilids).

Although Eomakhaira is most parsimoniously recovered as the sister taxon to thylacosmilids within Proborhyaenidae in this study, it should be noted that this specimen could only be coded for a limited number of characters (45/321 characters; ~14%, not

counting multistate characters) and codings for some of these characters may be affected

by preservation and ontogenetic status. Therefore, it is possible that future studies could

recover Eomakhaira elsewhere within Proborhyaenidae, as a borhyaenid, or a

borhyaenoid outside the clade of (Borhyaenidae + (Proborhyaenidae +

Thylacosmilidae)). Constraining Eomakhaira to not be the sister taxon of

Thylacosmilidae requires only a single additional step and results in Eomakhaira being

recovered as the most basal proborhyaenid or the sister taxon of Callistoe. Constraining

Eomakhaira to be outside the clade of Proborhyaenidae + Thylacosmilidae requires only

two additional steps, placing Eomakhaira as sister to the clade of Borhyaenidae +

(Proborhyaenidae + Thylacosmilidae), though the placement of Eomakhaira outside this

clade is based on a single character, an unfused mandibular symphysis, which may not

represent the ancestral condition for the clade of Borhyaenidae + (Proborhyaenidae +

126

Thylacosmilidae) as the symphysis of Borhyaena is typically unfused in most individuals

(possibly related to ontogeny). Constraining Eomakhaira as a borhyaenid requires three

additional steps and recovers Eomakhaira as the most basal borhyaenid. Constraining

Proborhyaenidae and Thylacosmilidae to be monophyletic (leaving Eomakhaira as a

floating taxon) requires four additional steps and recovers a variety of topologies for

Proborhyaenidae, Thylacosmilidae, Borhyaenidae, and Eomakhaira, with Eomakhaira

usually being recovered as the most basal thylacosmilid or proborhyaenid. Constraining a

relationship of Thylacosmilidae + Borhyaenidae (leaving Eomakhara as a floating taxon)

requires five extra steps and recovers Eomakhaira as the basalmost proborhyaenid. This

instability, as well as other anatomical observations such as the apparent presence of a

precingulid and talonid with an entoconid in the thylacosmilid Anachlysictis gracilis

(Goin, 1997b) and the report of a metaconid in the potential thylacosmilid from La Venta

(Goin, 2003) described by Goin (1997b) suggests that the results of this phylogenetic analysis should be interpreted with some caution. For this reason I defer the decision of whether or not Eomakhaira should be assigned to the Thylacosmilidae until better preserved material is discovered.

4.6 DISCUSSION

4.6.1 Paleobiology of SGOPV 3490

4.6.1.1 Body Size

In terms of paleobiology, Eomakhaira molossus is notable for its small size. The body mass of Eomakhaira was estimated using two criteria: the length of the lower molar row (Lm1-4) via the regression equation of Myers (2001) and the length of m3 (which

127

was considered to be the best predictor variable of body mass by Zimicz [2012] and

Forasiepi et al. [2015]) via the regression equation of Gordon (2003). In both cases the

regression equations used were based on dasyuromorphians, the extant metatherian group

that most closely resembles borhyaenoid sparassodonts in terms of their dental

morphology. Although the use of these equations has been suggested to produce

inaccurate mass estimates in some cases because most sparassodonts are far larger than

any living carnivorous marsupial (Forasiepi et al., 2015), SGOPV 3490 is small enough

that it is not an outlier relative to the extant species examined. Both regression equations

produce body mass estimates of only about 9.3 kg for Eomakhaira molossus, comparable

in size to a male Tasmanian devil (Sarcophilus harrisii; Rose et al., 2016). SGOPV 3490

also compares well to the skull of a Tasmanian devil in terms of general size, though the

skull of SGOPV 3490 is deeper and narrower than that of Sarcophilus.

Eomakhaira is much smaller than the four other currently recognized genera of

proborhyaenids (Callistoe, Arminiheringia, Proborhyaena, and Paraborhyaena,

Appendix 19) and only 40% the size of the next smallest proborhyaenid, Callistoe (~23

kg; Argot and Babot, 2011). The presence of such a small proborhyaenid sparassodont in

the Oligocene is quite unexpected, given that this was also the period when members of

the Proborhyaenidae reached their largest sizes (> 50 kg in Paraborhyaena and

Proborhyaena, though both of these taxa are only known from the late Oligocene). There are two possible explanations for this phenomenon. First, it is possible that Eomakhaira as well as Proborhyaena and Paraborhyaena represent part of an Oligocene radiation of proborhyaenids, of which gigantism and small size/incipient sabertooth features were but two evolutionary experiments (the latter of which may have proved more successful).

128

The Eocene-Oligocene transition was a period of great turnover in South American metatherian faunas, likely due to climatic change, involving the extinction of most of the previously dominant groups and radiation of the surviving forms (Goin et al., 2010; Goin et al., 2016), including the first appearance of hathliacynids and borhyaenids (Petter and

Hoffstetter, 1983; Forasiepi et al., 2015). It is possible that proborhyaenids reacted to these environmental changes in the same way as these other sparassodont groups.

Figure 4.19. Size comparison among Paleogene proborhyaenids. From largest to smallest, Proborhyaena gigantea (in blue), the largest known proborhyaenid (scaled after AMNH 29576, the largest specimen of this taxon); Callistoe vincei (in green), the smallest named proborhyaenid prior to this study (scaled after the holotype specimen, PVL 4187); Eomakhaira molossus (in red), scaled after SGOPV 3490 (silhouette modified from Proborhyaena). Proborhyaena and Callistoe sillhouetes by Zimices and Steven Traver, respectively, from PhyloPic. Alternatively, the Oligocene proborhyaenids could represent the remnants of a much older radiation, dating back to the Eocene, of which only a few lineages such as those that gave rise to Eomakhaira, Proborhyaena, and Paraborhyaena surviving into later times.

At present, the fossil record of sparassodonts is too incomplete to favor either hypothesis.

129

Either way, this suggests that there was a significantly greater amount of

ecomorphological diversity within the Proborhyaenidae than previously recognized that

existed as late as the early Oligocene. This is further supported by the proborhyaenid

from the late Oligocene (Deseadan SALMA) locality of Taubaté (Couto-Ribeiro, 2010),

which has not been included in any phylogenetic analysis but is much larger than SGOPV

3490 and much smaller than Paraborhyaena, approximately the same size as

Arminiheringia auceta (Appendix 19).

One interesting question raised by the small size of Eomakhaira is whether

another specimen, MLP 88-V-10-4, an isolated m4 from the locality of Antofagasta de la

Sierra in northwestern Argentina, belongs to this taxon or a closely related form. This

specimen was originally assigned to Arminiheringia by Goin et al. (1998). However, later

authors (Babot et al., 2002; Babot, 2005; Powell et al., 2011) referred this specimen to the

genus Callistoe, based on its smaller size and reduced talonid. Although MLP 88-V-10-4 is slightly larger than the holotype of Eomakhaira (m4 length 13.95 mm in MLP 88-V-

10-4 versus 12.0 mm in SGOPV 3490), it is much closer in size to this specimen than the

holotype of Callistoe (m4 length 17 mm; Babot et al., 2002). Additionally, MLP 88-V-

10-4 resembles SGOPV 3490 in that the protoconid is positioned at the posterior end of

the tooth and the talonid is virtually absent, in contrast to Callistoe where the protoconid

is positioned slightly more anterior on the tooth and there is a small posterior talonid

cuspule.

The small size of Eomakhaira relative to other proborhyaenids suggests there may

have been a possible dwarfing event in the evolutionary history of this taxon. SGOPV

3490 is not only small compared to proborhyaenids but to most thylacosmilids as well

130

(Appendix 19). The only members of the thylacosmilid-proborhyaenid clade to be comparable to SGOPV 3490 in size are the Colhuehuapian thylacosmilid from Gran

Barranca (MLP 92-X-10-6, M3 length 7.5 mm versus 7.6 mm in SGOPV 3490; Goin et al., 2007), and the possible thylacosmilid from La Venta (IGM 251108, m4 length 8.7 mm versus 12.0 mm in SGOPV 3490; Goin, 1997b), which is significantly smaller. If

MLP 88-V-10-4 from Antofagasta de la Sierra also belongs to this taxon or a closely related form it would further support the idea of a dwarfing event in the evolutionary history of this taxon, as it would imply a decrease in size from the late Eocene to the early Oligocene across the Eocene-Oligocene transition. A dwarfing event in the late

Eocene might explain how thylacosmilids might have arisen from within proborhyaenids,

as large carnivorous mammals usually exhibit very short temporal durations (< 10 Ma)

and tend to be characterized by turnover events and replacement by other groups (Van

Valkenburgh, 1999; Van Valkenburgh et al., 2004) rather one large carnivore lineage

giving rise to another large carnivore lineage.

4.6.1.2 Dietary Habits

Several lineages of sparassodonts, including borhyaenids and proborhyaenids,

have been suggested to have had the ability to break bones at the level of p3, based on

morphological features such as an interlocking or fused mandibular symphysis, deep

dentaries, bulbous premolars with long roots, and microfractures in the tooth enamel

(Ercoli et al., 2014) as well as morphometric analyses of the teeth and jaws (Blanco et al.,

2011; Forasiepi et al., 2015). At first glance, Eomakhaira appears to be another robust

bone-cracking form based on its deep dentary, but other features suggest that this species

was not strongly specialized for bone-cracking. It has been suggested that in deep-jawed

131

carnivorous mammals the forms most specialized for bone cracking exhibit almost no

difference in dentary depth between the primary carnassial and the bone-cracking

premolar (Palmqvist et al., 2011; Forasiepi et al., 2015). In sparassodonts, this was noted

specifically in Australohyaena (Forasiepi et al., 2015), but also occurs in Proborhyaena,

Paraborhyaena, and Arctodictis. By contrast, in Eomakhaira the dentary under p3 is

much shallower than under m4 (Table 4.1). The symphysis of Eomakhaira is also much

less extensive than in bone-cracking forms. In most sparassodonts with inferred bone-

cracking habits the symphysis completely underlies the main bone-cracking premolar,

often extending to the level of the p3/m1 embrasure to or even the posterior root of m1 in

Arminiheringia auceta (Babot et al., 2002; Zimicz, 2012). In Eomakhaira the symphysis is more anterior and ends approximately at the level of the p2/p3 embrasure. The roof of the skull is also not vaulted at the level of the primary bone-cracking teeth in Eomakhaira as is typically seen in other bone-cracking mammals (Werdelin, 1989), including the

vaulted nasals of Australohyaena (Forasiepi et al., 2015).

Zimicz (2012; 2014) and Forasiepi et al. (2015) used several metrics modified

from those originally used Van Valkenburgh (1989) to examine dietary habits in

carnivorous metatherians (primarily degree of carnivory and specializations for

durophagy/bone-cracking). Several of these metrics can be applied to the holotype of

Eomakhaira (Table 4.4). The premolar shape (width of p3/length of p3), which measures

the robustness of the last lower premolar, is 0.54 in Eomakhaira, below the threshold

between bone-crackers and non-durophagous forms (bone-crackers > 0.58). Relative

premolar size (width of p3/cube root of body mass in kg), which measures the robustness

of the last lower premolar relative to body size, is 1.32, which is also lower than the

132

threshold between bone-cracking and non-durophagous (bone-crackers > 2.6). The

relative premolar length (length of p3/length of m4) is 0.63, below the threshold of

hypercarnivores (hypercarnivores > 0.7, non-hypercarnivores < 0.7). However, it is

noteworthy that many sparassodonts that otherwise show strong adaptations for

hypercarnivory (Callistoe, Arminiheringia, Australohyaena) also fall below this

threshold, calling into question whether this parameter can be taken at face value when

evaluating metatherians.

Table 4.4. Morphometric values of the dentition used to infer dietary habits in Eomakhaira molossus. Methodology for calculating these parameters and critical values for dietary categories based on Van Valkenburgh (1989), Zimicz (2012), and Forasiepi et al. (2015). Parameter Value Critical Values Bone-crackers > 0.58, Premolar shape (width of p3/length of p3) 0.54 other carnivores < 0.58 Relative premolar size (width of p3/cube Bone-crackers > 2.6, 1.32 root of body mass in kg) other carnivores < 2.6 Relative premolar length (length of 0.63 Hypercarnivores > 0.7, p3/length of m4) other carnivores < 0.7 "Cat-like" hypercarnivores > 0.9, Relative trigonid length (length of m4 bone-cracking 0.91 trigonid/length of m4) hypercarnivores 0.8-0.9, other carnivores < 0.8 (but see text) Relative grinding area (square root of grinding area on m4/length of trigonid on ~0 Hypercarnivores < 0.48, m4 other carnivores > 0.48

The relative trigonid length (length of m4 trigonid/total length of m4) is 0.91,

within the range of specialized hypercarnivores and comparable to some of the most

specialized carnivores within Sparassodonta, including the borhyaenoid Angelocabrerus, the borhyaenids Australohyaena, Arctodictis, and Acrocyon, the proborhyaenids

Proborhyaena and Paraborhyaena, and the thylacosmilid Thylacosmilus (Zimicz, 2012;

133

Forasiepi et al., 2015; Croft et al., 2018). Zimicz (2012) and Forasiepi et al. (2015) considered a relative trigonid length of >0.9 to be indicative of cat-like hypercarnivory and a relative trigonid length between 0.8 and 0.9 to be indicative of hypercarnivory with bone-crushing specializations. However, only three extant bone-crushing taxa were included in this comparative dataset, the striped, brown, and spotted hyenas. Brown and striped hyenas are unusual among extant large hypercarnivores in that they have a small but functional grinding area on their carnassial molar (Ewer, 1954) whereas Crocuta is more similar to felids (and have a relative trigonid length > 0.9; Van Valkenburgh, 1989).

Given the habits of extant brown and striped hyenas, this may be related to the significant amount of fruit in the diet of these species, which is not present in the diet of spotted hyenas (Kruuk, 1976; Owens and Owens, 1978; Mills, 2015). There is little reason to think that trigonid length would be correlated with bone-cracking habits in hyaenids, as hyenas primarily break bones with their premolar teeth, rather than their molars as in some other carnivorans (e.g., extant canids; Werdelin, 1989). Additionally, the relative trigonid length of the extant hypercarnivorous canids Cuon, Lycaon, Speothos is only

0.72-0.74 (Van Valkenburgh, 1989). The fact that all living bone-cracking carnivorans

have relative trigonid lengths of 0.8-0.9 is more likely due to historical contingency and

the small number of living bone-cracking carnivorans (N = 3) than any functional reason, especially since this value is within the range of variation seen in hypercarnivores in general.

The final parameter, relative grinding area, has been measured using two methods in sparassodonts: using only m4 (Prevosti et al., 2013) and using the entire lower molar row (Croft et al., 2018). Although using the entire molar row is probably the more

134

accurate method of assessing dietary habits (see discussion in Croft et al., 2018), m1-3

are too heavily worn in the holotype of Eomakhaira to calculate RGA in this manner. The

m4 of Eomakhaira most closely resembles the condition in taxa like Proborhyaena,

Paraborhyaena, Arctodictis, and Thylacosmilus, which Prevosti et al. (2013) considered

to lack a functional talonid and therefore coded RGA as 0. Together, these features

suggest Eomakhaira was a hypercarnivore but was not specialized for bone-crushing.

4.6.2 Carnassial Rotation in Sparassodonts

Mellett (1969) described an unusual condition in the hyaenodontan Hyaenodon in

which the crowns of the upper molars rotate throughout ontogeny so that they cant increasingly medially (termed “carnassial rotation” by this author). Although Mellett

(1969) primarily discussed carnassial rotation in relation to Hyaenodon, he also mention the condition was present in the hyaenodontan Hemipsalodon, the oxyaenid Patriofelis, and “an unnamed Pliocene marsupial sabertooth” (almost certainly Thylacosmilus atrox, given that no other well-preserved thylacosmilid remains were known at the time).

However Goin and Pascual (1987) did not notice any carnassial rotation in the specimens of Thylacosmilus they observed. Marshall (1978b) reported the occurrence of carnassial rotation in several other sparassodonts, most prominently in Arminiheringia auceta but to a lesser degree in specimens of Acrocyon, Arctodictis, and Borhyaena. Bond and Pascual

(1983) also described carnassial rotation as being present in the senescent specimen they assigned to Proborhyaena (MLP 79-XII-18-1). Carnassial rotation in Hyaenodon and

Arminiheringia produces a very distinct form of wear (Figure 4.20), with the entire posterolingual face the upper molars forming a sharp, flat edge with vertical wear facets

135 roughly parallel to the main shearing blade of the tooth (typically the postmetacrista of upper molars) In the holotype of Arminiheringia (MACN-A 10970/10972) this condition is particularly pronounced to the point that the pulp cavities of M1-3 were exposed in vivo in this animal.

Figure 4.20. Left posterior upper dentition (P3-M4) of the holotype of Arminiheringia auceta (MACN-A 10970/10972) in oblique ventral view. Note the well-developed wear facets extending from the protocone to the metastylar corner of the tooth roughly parallel to the main shearing crest of the unworn molars (postmetacrista of M1-3 and preparacrista of M4) due to carnassial rotation. Photo by Darin Croft.

However, our observations have found that upper molars that are canted medially are present in a much wider range of sparassodonts than previously thought. In addition to SGOPV 3490, medially canted molars (here defined as when the angle between the base of the crown of M3-4 and the palate is > 30°, though in all cases the angle was >

35°) were observed in the hathliacynid sparassodonts Acyon myctoderos, Cladosictis patagonica, Sipalocyon gracilis, and the borhyaenoids Acrocyon riggsi, Arctodictis sinclairi, Prothylacynus patagonicus, Pharsophorus tenax, and Arminiheringia sp. (Table

4.5). By contrast, in the sparassodonts Patene simpsoni, Patene coluapiensis,

Hondadelphys fieldsi, and UF 27881 the molars are much less canted (< 35°, generally less than 20°), though the base of the crown of the posterior molars in these taxa was not close to level with the palate, as in didelphoids (Table 4.5). 136

Table 4.5. Angle of inward canting of the posterior upper molars in sparassodonts. This angle is measured as the angle between the base of the crown of M3–4 in posterior view and the floor of the palate. Angle in SGOPV 3490 is approximate given crushing of the skull. Comparative data from Didelphis and Dasyurus from Macrini (2005a, 2005b)

Taxon Group Specimen Angle Eomakhaira molossus Proborhyaenidae SGOPV 3490 ~42° Arminiheringia sp. Proborhyaenidae MLP 82-V-1-1 43.89° Acrocyon riggsi Borhyaenidae FMNH P13433 48.58° Arctodictis sinclairi Borhyaenidae MLP 85-VII-3-1 49.07° Basal Pharsophorus tenax AC 3192 35.87° Borhyaenoidea Prothylacynus Basal MACN-A 5931 45.54° patagonicus Borhyaenoidea Prothylacynus Basal MACN-A 706 38.55° patagonicus Borhyaenoidea Acyon myctoderos Hathliacynidae UF 26933 40.21° Cladosictis patagonica Hathliacynidae MACN-A 5950 38.45° Cladosictis patagonica Hathliacynidae MACN-A 5927 44.69° Sipalocyon gracilis Hathliacynidae MACN-A 692 46.72° YPM-VPPU Sipalocyon gracilis Hathliacynidae 38.27° 15373 Sparassodonta UF 27881 UF 27881 24.61° incertae sedis Hondadelphys fieldsi Basal Sparassodonta UCMP 37960 14.57° Patene coluapiensis Basal Sparassodonta AMNH 28448 15.66° Patene simpsoni Basal Sparassodonta MNRJ 1331-V 17.76° Didelphis virginiana Didelphidae TMM M-2517 8.3° Dasyurus hallucatus Dasyuridae TMM M-6921 24.0°

This angle could not be quantified in Australohyaena, Borhyaena, cf. Proborhyaena,

Paraborhyaena, or Callistoe due to lack of relevant views for measuring, but observations of the dentitions of these taxa suggest that the posterior upper molars were also strongly canted inwards as in their close relatives Arminiheringia, Acrocyon, and

Arctodictis. UF 27881 (which is missing the crowns of M2-3, but the alveolar bone

surrounding the base of the crown is preserved) is particularly interesting in this respect,

as it shows a greater degree of medial canting than Patene or Hondadelphys, but still

137

significantly lower than other sparassodonts. No undoubted members of the group

Hathliacynidae + Borhyaenoidea that could be observed first-hand were noted to not exhibit pronounced medial canting of the upper molars, suggesting this may be a useful phylogenetic feature to characterize this group. However, a systematic assessment of carnassial rotation in sparassodonts and other metatherian taxa could not be performed due to the limited number of specimens that could be observed.

Our observations also show a distinction between “carnassial rotation” (inward

rotation of the teeth over the course of the animal’s lifespan) as defined by Mellett (1969)

and molars that are simply canted medially. Two of the specimens in which medial

canting of the posterior molars could be examined, MACN-A 5931 (Prothylacynus

patagonicus) and MLP 82-V-1-1 (Arminiheringia sp.) are subadult specimens, with

M/m4 still erupting (Forasiepi and Sánchez-Villagra, 2014). Nevertheless, the angle of

the occlusal faces relative to the palate are comparable to that seen in adult sparassodonts,

and in Prothylacynus the molars of the subadult specimen are actually more canted than

the adult (Table 4.5). Most sparassodonts (including Acrocyon, Arctodictis, and

Borhyaena) also do not exhibit the distinctive wear pattern seen in the teeth of

Arminiheringia, even in older specimens. However, the subadult specimen of

Arminiheringia, despite being ontogenetically young, already shows wear facets similar

to the holotype of Arminiheringia on the posterolingual faces of M1-2 (Forasiepi and

Sánchez-Villagra, 2014: fig 2b-c). The condition in cf. Proborhyaena (MLP 79-XII-18-1)

is harder to determine. The wear on the upper molars of this specimen does not seem to

be identical to that in Arminiheringia, but is the most similar to the condition in

Arminiheringia out of all other sparassodonts. These observations suggests that the upper

138

molars of most sparassodonts did not rotate throughout life but actually erupted in canted

position, and that Arminiheringia and possibly Proborhyaena are the only currently described sparassodonts to show carnassial rotation similar to what Mellett (1969) described in Hyaenodon.

It is not clear why medially canted upper molars are so widespread among sparassodonts. Mellett (1969) suggested that the carnassial rotation seen in Hyaenodon was a necessary adaptation in carnivorous mammals with multiple shearing teeth, anisognathus lower jaws, and fused mandibular symphyses because wear on the cheek teeth could not be compensated for by lateral motion of the lower jaw as in similar forms with unfused symphyses. However, in sparassodonts, inwardly canted upper molars are present in many taxa with ligamentous symphyses (e.g., Acyon, Cladosictis, Sipalocyon) that would have allowed for some degree of flexibility at the symphysis. Furthermore, as noted above, the upper molars are medially canted in immature specimens of

Arminiheringia sp. and Prothylacynus patagonicus and the angle of displacement of the molars in these specimens is comparable to that observed in specimens of adult sparassodonts (including other specimens of the same taxon). This indicates that the medial canting of the upper molars was not something that developed gradually over the animal’s lifespan, as would be expected if the function of medially canted carnassials in these taxa was to maintain precise occlusion as the primary shearing teeth wore down, but rather was present as soon as M3-4 erupted. Finally, many sparassodonts also show a lateral canting the lower molars (e.g., Australohyaena antiquua, see Forasiepi et al.,

2015), which is also present to a lesser degree in Dasyurus (Macrini, 2005a), whereas no such canting is present in Hyaenodon (Mellett, 1969)

139

4.6.3 Proborhyaenidae and the Origin of Thylacosmilids

4.6.3.1 Eomakhaira molossus and the Evolution of Thylacosmilidae

Ever since the first well-preserved specimens of thylacosmilids were described in the 1930s (Riggs, 1933, 1934), scientists have struggled to understand how these animals developed their highly specialized morphology and how they were related to other sparassodonts. Scott (1937) noted similarities that might link Thylacosmilus to the

Eocene proborhyaenid Arminiheringia, but other authors were skeptical of this idea given the great disparity in morphology and geologic time between the two taxa and suggested these similarities might be due to convergent evolution (Simpson, 1948; Marshall,

1976a). Part of this issue was due to the fact that earlier workers were primarily dealing with Thylacosmilus atrox, the geologically youngest and most specialized member of the

Thylacosmilidae. It wasn’t until the 1990s that geologically older, more plesiomorphic thylacosmilids were described that might provide insights into the origin of this group.

Goin (1997b) described the thylacosmilid Anachlysictis gracilis and a second, unnamed taxon potentially belonging to this family (IGM 251108) from the middle Miocene locality of La Venta, Colombia, both of which exhibit a less specialized morphology than the Mio-Pliocene Thylacosmilus. More recently, Goin et al. (2007) and Forasiepi and

Carlini (2010) described even older thylacosmilids from the early Miocene

(Colhuehuapian) and early middle Miocene (Colloncuran) of Patagonia, respectively.

These early specimens are still relatively morphologically specialized, more similar to

Thylacosmilus than either La Venta taxon, indicating a much older (Oligocene) origin for the Thylacosmilidae (Goin et al., 2007). However, until now, no specimens have been

140

described that could represent Oligocene thylacosmilids or link thylacosmilids with any

other group of sparassodonts.

Perhaps the most important result of the phylogenetic analysis in this chapter is

the position of Eomakhaira molossus as the sister taxon to the thylacosmilids

Patagosmilus and Thylacosmilus and within a paraphyletic Proborhyaenidae. A close

relationship between proborhyaenids and thylacosmilids is not a completely novel result;

Marshall et al. (1990), Muizon (1999) and Babot et al. (2002) found proborhyaenids and thylacosmilids to be sister groups, whereas Babot (2005), the equal weights analysis of

Suarez et al. (2016), and the implied weight analysis of Forasiepi et al. (2015) recovered

Thylacosmilidae within a paraphyletic Proborhyaenidae. Indeed Babot (2005) recovered

Proborhyaena gigantea as the sister taxon to Thylacosmilidae, a position partially supported by the results of this study. Additionally, as mentioned above, a relationship between thylacosmilids and proborhyaenids was suggested as early as Scott (1937).

However, this is the first phylogenetic analysis of Sparassodonta to unambiguously recover Thylacosmilidae within a paraphyletic Proborhyaenidae both under equal and implied weights.

Additionally, Eomakhaira represents a morphological intermediate between the morphology typically seen in proborhyaenids and that typically seen in thylacosmilids, with similarities to both groups. Prior to this study, even though some analyses suggested a relationship between proborhyaenids and thylacosmilids, there were no taxa that spanned the rather large morphological disparity between the two groups. Proborhyaenids in general tend to be morphologically conservative, with the four previously named genera looking fairly similar to one another and no taxon showing any similarity to non-

141 proborhyaenid sparassodonts (except possibly Proborhyaena, see below).

Thylacosmilids, on the other hand, are characterized by numerous autapomorphies and other features which based on character state reconstruction are thought to be reversals to the ancestral condition from a derived one related to sabertooth adaptations (Riggs, 1934;

Goin, 1997b; Babot et al., 2002; Forasiepi and Carlini, 2010; Forasiepi and Sánchez-

Villagra, 2014). Eomakhaira resembles proborhyaenids in having lingual sulci on the upper canines, three premolars with no retention of dP3 (both features plesiomorphic for sparassodonts in general), no mandibular flange (plesiomorphic for sparassodonts), and possibly open-rooted lower canines.

On the other hand, in many other respects Eomakhaira resembles thylacosmilids.

The canines of Eomakhaira have no longitudinal grooves (the loss of which is almost unique to thylacosmilids among borhyaenoids), are relatively labiolingually narrow compared to other sparassodonts, and the upper canines lack a labial median sulcus but have a well-developed median keel. The maxilla is much deeper than in proborhyaenids, more like borhyaenids and thylacosmilids, and compared to the Eocene proborhyaenids

Callistoe and Arminiheringia (though possibly not the Oligocene Paraborhyaena and

Proborhyaena) the dentary is shallower and the infraorbital foramen is more posteriorly located, over the posterior root of P3. The symphysis of Eomakhaira is also more similar to thylacosmilids than proborhyaenids. In most borhyaenoid sparassodonts (especially proborhyaenids), the symphysis is extensive, and often fused in adults. By contrast, in

Eomakhaira the symphysis is much less extensive, ending at the p2/3 embrasure or at most the anterior root of p3, and is unfused. Additionally, in Eomakhaira the P3 is much longer than p3 and the premolar row is relatively short. This somewhat resembles what is

142 seen in thylacosmilids, in which the dP3 is much longer than p3 and the premolar row is relatively short (though Thylacosmilus exhibits a reorganization of the postcanine dentition related to recentering the molar teeth under the area of maximal bite force;

Goin, 1997b; Forasiepi and Carlini, 2010)

Interestingly, some of these features may be shared with Proborhyaena gigantea, especially if MLP 79-XII-18-1 pertains to this species, including a relatively shallower dentary, more posteriorly positioned infraorbital foramen, labiolingually narrow upper canines, and a (d)P3 that is much larger than p3. This is especially noteworthy as most of these characters were either not included in this analysis or were not coded for P. gigantea due to being determined at least in part by MLP 79-XII-18-1, whose assignment to Proborhyaena has been debated (Bond and Pascual, 1983; Babot et al., 2002; Forasiepi and Sánchez-Villagra, 2014), and suggests that Proborhyaena might be the closest relative to thylacosmilids among the four other proborhyaenid taxa.

4.6.3.2 Evolution of Saber Teeth in Sparassodonta

Despite having no incontrovertible living representatives, saber teeth appear to have been a successful morphotype in mammals, having independently evolved in

“creodonts” (Machaeroidinae), nimravids, barbourofelids, felids (Machairodontinae), and sparassodonts (Antón, 2013). However, the distribution of saber teeth is very asymmetrical among mammals. Despite evolving at least four separate times in carnivorous placentals, saber-toothed canines only evolved once in metatherians, in

Proborhyaenidae + Thylacosmilidae. No other groups of metatherians aside from thylacosmilids are known to have developed saber teeth, despite the large number of metatherian lineages that have adopted predatory habits, including deltatheroidans,

143 stagodontids, pucadelphyids, the marsupialiform Anatoliadelphys, other groups of sparassodonts (i.e., hathliacynids), didelphoids, dasyurids, thylacinids, and thylacoleonids

(suggestions that the extant didelphid Monodelphis dimidiata represents an extant sabertooth predator have not been supported by later analyses; Blanco et al., 2013;

Chemisquy and Prevosti, 2014).

One potential reason for the rarity of metatherian sabertooth lineages is the way in which these distinctive upper canines are thought to have been used. Many researchers have noted is that saber teeth are a highly precise weapon which requires some degree of experience to use correctly (Emerson and Radinsky, 1980; Akersten, 1985; Antón and

Galobart, 1999; Wheeler, 2011). An inaccurate bite can greatly reduce the efficiency of the saber-tooth feeding apparatus (Wheeler, 2011) and the tall, labiolingually narrow shape of these teeth means they can easily break if exposed to sudden, unpredictable loads (Van Valkenburgh and Ruff, 1987), such as those produced by struggling prey, and leave the animal unable to hunt. This evolutionary dilemma may be the reason behind an unusual ontogenetic feature of most placental sabertooths: prolonged retention and large size of the deciduous canines. In two groups of placental sabertooths, the Nimravidae and the Barbouofelidae, the deciduous canines are often as large as the adult teeth and are replaced relatively late in ontogeny (Bryant, 1988; Wysocki et al., 2015). This condition is much less pronounced in machairodontine felids, though the deciduous canines in taxa like Smilodon are still much larger and replaced later in ontogeny than the deciduous canines of non-sabertoothed felids and are still capable of performing a similar function to the adult teeth (Wysocki et al., 2015). This prolonged use of the deciduous canines and late eruption of the replacement teeth meant that if a young, inexperienced placental

144

sabertooth broke its canines, it had a second chance to survive once the adult canines

erupt. However, unlike placental mammals, marsupials do not have any deciduous

canines and only have one generation of teeth at the canine locus. Therefore, the ever-

growing (hypselodont) upper canines of thylacosmilids may represent a different way to

achieve the same function: if a young thylacosmilid broke its upper canines, it could

survive off of carrion and other food sources until the canine regrows to an appropriate

size (though the growth rate of thylacosmilid canines is not known, and no thylacosmilid

canine that was clearly broken in vivo has been described).

Interestingly, the results of the phylogenetic analysis in this paper suggest that

open-rooted canines were not a novel innovation of thylacosmilids, but a plesiomorphic

feature inherited from non-sabertoothed ancestors (i.e., proborhyaenids). Therefore,

because they lack deciduous canines, ever-growing canines may be a prerequisite for

evolving saber-teeth in metatherians. This may be one reason why despite evolving four

separate times in carnivorous placentals, saber-toothed canines only evolved once in

metatherians. A similar analogy between the late-erupting permanent canines of

thylacosmilids and the ever-growing canines of thylacosmilids in lieu of the absence of a

deciduous canine was noted by Marshall (1976a), though this author did not note the

rarity of sabertoothed metatherian lineages or the potential constraint of requiring ever-

growing canines as a prerequisite exaptation for saber teeth.

The development of open rooted canines in proborhyaenids and thylacosmilids is probably the result of paedomorphosis, as marsupial canines have been noted to remain open for much longer in ontogenetic history than placentals (Jones, 2003; Chemisquy and

Prevosti, 2014), and in juvenile non-proborhyaenid, non-thylacosmilid sparassodonts the

145 canine roots have been noted to remain open until relatively late in ontogeny (Forasiepi and Sánchez-Villagra, 2014; Engelman et al., 2015). This is supported by the presence of closed roots in senescent specimens of proborhyaenids including SGOPV 3490 and MLP

79-XII-18-1 (Bond and Pascual, 1983), which suggest that canine hypselodonty was achieved by delaying root closure until extremely late in the animal's life. Such evolutionary transitions from closed rooted to fully hypselodonty via postponement of root formation have been observed in the teeth of other mammals, such as notoungulates

(Madden, 2015). Interestingly, paedomorphosis also seems to be involved in other aspects of the evolution of thylacosmilids, such as the lack of replacement of dP3

(Forasiepi and Sánchez-Villagra, 2014) and possibly the unfused mandibular symphysis in these animals, and may have played a significant role in shaping the unusual anatomy of later members of this group.

It is also possible that other features typical of sabertoothed mammals in thylacosmilids, such as adaptations for a wide gape (Emerson and Radinsky, 1980; Slater and Van Valkenburgh, 2008; Antón, 2013), may have also been exaptations from non- sabertoothed proborhyaenids. In Callistoe vincei, the combined upper length of the upper and lower canines is comparable to the length of the upper canines of Thylacosmilus

(Powell et al., 2011), suggesting that Callistoe would have had to open its jaws to a comparable degree to Thylacosmilus in order to achieve clearance between the canines

(Powell et al., 2011; Wroe et al., 2013). However, jaw function and gape in Callistoe and other sparassodonts has not been investigated in detail.

Another unusual feature of saber-toothed sparassodonts compared to their placental counterparts is their longevity. Although it is possible that the similarities

146

between Eomakhaira and thylacosmilids are as a result of convergent evolution, similar

to what has been observed between nimravids, barbourofelids, and saber-toothed cats

(Morlo et al., 2004), the phylogenetic evidence currently suggests that all saber-toothed and sabertooth-like sparassodonts represent a monophyletic clade. If Eomakhaira represents a form analogous to Nimravus or Pseudaelurus among placental saber-tooth clades, then thylacosmilids would have a biochron extending from the early Pliocene (3

Ma) to the early Oligocene (≤ 29.3 Ma), resulting in a minimum temporal range of 26 Ma

(Figure 4.21).

Figure 4.21. Temporal duration of the major lineages of mammalian saber-toothed carnivores with representative skulls of each clade at left. Metatherian lineages in grey, placental lineages in black. Representative skulls, from top to bottom: Thylacosmilus atrox (Thylacosmilidae); Machaeroides eothen (Machaeroidinae), Hoplophoneus primaevus (Nimravidae); Barbourofelis fricki (Barbourofelidae); and Smilodon fatalis (Machairodontinae). All skull images from Antón (2013). By contrast, placental sabertooth clades have been far less fortunate in the lengths of their respective dynasties (Figure 4.21). Saber-toothed “creodonts” (Machaeroidinae) have an age range of approximately 11.3 Ma (52.8-41.5 Ma; Dawson et al., 1986;

Robinson et al., 2004; Kelly et al., 2012; Tomiya, 2013; Zack, in press), though this is a very rough estimate given the poor fossil record of the clade (being known from at most

147

six species and less than ten specimens). Nimravids are known from the mid-late Eocene

(~42 Ma; Suyin et al., 1977; Averianov et al., 2016) to the end of the Oligocene (23 Ma;

Bryant, 1996; Peigné, 2003), a temporal range of 19 Ma. Barbourofelids first appear in the early Miocene (20-19 Ma; Morales et al., 2001; Morlo et al., 2004) and last appear in

the late middle Miocene (6 Ma; Tedford et al., 2004); a range of 13 to 14 Ma. Finally, the

machairodontine felids (here defined as the first appearance of Pseudaelurus sensu

stricto) diverged from other felids as early as the middle Miocene (16 Ma)(Werdelin et

al., 2010; Robles et al., 2013) and disappeared in the end-Pleistocene extinctions (< 1

Ma), resulting in a temporal range of roughly 17 Ma.

Even if Eomakhaira turns out to not be a close relative of thylacosmilids, thylacosmilids would still likely have a temporal range extending back into the

Oligocene, given the presence of an upper molar that closely resembles the morphologically specialized thylacosmilid Patagosmilus from early Miocene

Colhuehuapian SALMA (20.2-20.0 Ma; Goin et al., 2007; Ré et al., 2010). If one assumes the Colhuehuapian specimen represents the absolute earliest occurrence of thylacosmilids (with no ghost lineages), this produces a temporal range of at minimum 17

Ma, still greater than that of barbourofelids and comparable to that of machairodontines.

Accounting for ghost lineages (as less specialized thylacosmilid lineages such as

Anachlysictis are considered to have already diverged before this time), would easily produce a temporal range longer than that of nimravids (19 Ma) and likely extend back into the Oligocene.

One potential explanation for relative longevity of thylacosmilids relative to placental sabertooths is the lack of potential replacing competitors due to South

148

America’s Cenozoic isolation (Patterson and Pascual, 1968; Simpson, 1980; Croft et al.,

2018). Sparassodonts were essentially the only group of carnivorous mammals in South

America until the beginning of the Great American Biotic Interchange (Prevosti and

Forasiepi, 2018) and given the potential developmental constraints on developing a sabertooth dentition in metatherians it seems unlikely that a competing group of sabertooths would potentially arise from within the Sparassodonta. However, there are several problems with this hypothesis. First, there is currently no evidence that the extinction of thylacosmilids was due to competitive exclusion. Although the idea that thylacosmilids were outcompeted by machairodontines used to be popular in the literature (e.g., Marshall, 1977a; Simpson, 1980), this idea is no longer supported by most authors given that at least 1.2 Ma separates the last record of thylacosmilids and the first record of machairodontines in South America (Forasiepi et al., 2007; Prevosti et al.,

2013) and the low number of sparassodont specimens in the “middle” Pliocene

Chapadmalalan SALMA relative the late Miocene Huayquerian SALMA (mostly

Thylacosmilus in both cases) despite similar collecting efforts and preservational biases suggesting a decline in abundance prior to their last appearance (Prevosti and Forasiepi,

2018).

Additionally, competitive replacement does not appear to have been a factor in the extinction of other groups of sabertooths. In the case of nimravids and barbourofelids, on most continents nimravids never even encountered their potential replacements, creating periods in which there were no sabertooth carnivorous mammals such as North

America’s “cat gap” (Hunt and Joeckel, 1988). Similarly the youngest records of machaeroidine “creodonts”, which are only known from North America, are at most 41.5

149

Ma (and are potentially older), whereas nimravids first appear on this continent 36.5 Ma

(Averianov et al., 2016; Barrett, 2016). Despite the poor fossil record of Machaeroidinae,

this is a significant temporal gap (5 Ma) between the two groups. Barbourofelids and

saber-toothed cats did encounter one another, and so some authors have suggested that

the appearance of the latter may be related to the extinction of the former (Geraads and

Güleç, 1997; Antón, 2013). However, the coexistence of barbourofelids and

machairodontine felids in Eurasia and North America for several million years has led

other authors to suggest that barbourofelids and saber-toothed cats did not compete with

one another, but that the more morphologically specialized barbourofelids were more

prone to extinction (Morlo, 2006). This is a common pattern in sabertoothed carnivores,

which show relatively high rates of extinction and turnover potentially correlated with their extreme morphological and ecological specializations (Naples et al., 2011; Piras et al., 2018). In this respect, the temporal longevity of thylacosmilids is even more noteworthy, especially given that they survived several major faunal/climatic changes in

South America such as the Bisagra Patagonica (Goin et al., 2010), the middle Miocene climatic optimum (Croft et al., 2016), and the expansion of grasslands during the late

Miocene (Pascual and Ortiz Jaureguizar, 1990). Far from being inferior “knock-offs” of placental sabertooths, as they are commonly regarded outside the specialist literature, thylacosmilids were highly successful predators that in many ways were able to outperform their placental counterparts.

150

5. FUTURE DIRECTIONS

The three sparassodont specimens described in this paper each expand on our knowledge of the morphological diversity and evolution of this extinct mammal group.

The unusual morphology represented by Australogale leptognathus expands the known morphological diversity of very small sparassodonts, and represents one of the few very small sparassodonts in which the talonid morphology can be described. It also shows that similar-sized sympatric sparassodonts were able to avoid competition by morphological niche partitioning, whereas previous studies of metatherian carnivore guilds found niche partitioning to be heavily driven at least in part by body size (Jones and Barmuta, 1998;

Ercoli et al., 2014). SGOPV 6200 represents one of the few non-proborhyaenid Eocene borhyaenoids, which increases the known diversity of Eocene sparassodonts and suggests these animals (especially non-proborhyaenid borhyaenoids) were much more diverse during the Eocene than previously thought. Eomakhaira exhibits a combination of traits completely unlike any previously described sparassodont (somewhat labiolingually narrow canines, deep maxilla, small size, etc.) and shows some similarities to the thylacosmilids, specifically suggesting a link between the proborhyaenids and thylacosmilids. These sparassodonts all come from localities outside the traditional famous collecting localities in Argentina (specifically Patagonia, the pampean region, and northwest Argentina), highlighting the important contributions extra-Argentine localities can make to our understanding of the evolutionary history of South America.

It is difficult to suggest future directions based on the research presented here.

Generally, only a single sparassodont fossil is collected over the course of a field season, even at highly productive localities (Engelman, pers. obs.), due to their extreme rarity in

151

the fossil record. Therefore, advances in sparassodont paleontology tend to be due more

to the discovery of new specimens than re-examination of old material, and new

expeditions to the field cannot reliably be expected to produce new material.

Nevertheless, a few avenues of future research present themselves from the results of

these studies.

The fact that the position of Australogale was not well-resolved in the

phylogenetic analysis, mostly due to the absence of morphologically similar

hathliacynids such as Psueudonotictis due to the limited material known for them and

accessibility of specimens, suggests the need for a re-examination, redescription, and rediagnosis of most very small sparassodont taxa. In particular, redescription of

Borhyaendium musteloides, the only very small Neogene sparassodont to be represented by cranial remains and one of only two (along with Pseudonotictis pusillus; Argot,

2003b) with associated postcrania might prove very useful. The holotype and only known skull of this species (MLP 57-X-10-153) is heavily restored with wax, obscuring many important anatomical details of the cranium. CT-scanning of the holotype of B. musteloides might be one way to resolve this issue, as this method had proven useful in restoring historical specimens whose morphology has been distorted or obscured with

wax or plaster by previous preparators (Grohé et al., 2015)

With SGOPV 6200, the discovery of a distinct Eocene borhyaenoid strongly

highlights the need for a revision and reexamination of Eocene sparassodont taxa. Only

two non-proborhyaenid borhyaenoids (Plesiofelis schlosseri and Angelocabrerus daptes)

have previously been described, the latter of which was named based on a specimen

which is now lost (Prevosti and Forasiepi, 2018). The current state of knowledge of

152

Eocene sparassodonts is heavily dependant on specimens listed in faunal lists, book chapters, and conference abstracts but have not been formally published or fully described (Simpson, 1948; Cladera et al., 2004; Babot, 2005; Babot and García-López,

2010; Gelfo et al., 2010; Powell et al., 2011; Forasiepi and Sánchez-Villagra, 2014;

Lorente et al., 2016). At the time of this writing many of these specimens are currently on loan to other researchers and are not available for examination, but some of these specimens are not. For example, the specimens from La Gran Hondonada (Cladera et al.,

2004) do not appear to be on loan and are in need of description, having been only assigned to Plesiofelis or Procladosictis, the latter of which is dubious given the genus is only represented by an maxillary fragment and the specimens from La Gran Hondonada all represent dentary fragments and Procladosictis has been a wastebasket taxon in the past (Marshall, 1981). Additionally, SGOPV 6200 suggests a need for the revision of

Pharsophorus and Pharsophorus-like sparassodonts (i.e., Plesiofelis, which is thought to be closely related to Pharsophorus and was even regarded as synonymous by some authors; Simpson, 1948). Several specimens have been assigned to several species of

Pharsophorus, but these specimens differ from one another in morphology and it is not clear if they all represent a monophyletic group (Marshall, 1978b; Patterson and

Marshall, 1978; Goin et al., 2010; Zimicz, 2012). Indeed, one species assigned to

Pharsophorus, “P.” antiquus, was recently re-examined with the discovery of new material and found to be a borhyaenid (Australohyaena; Forasiepi et al., 2015) SGOPV

6200 exhibits features similar to some specimens of Pharsophorus (canted p3, flat ventral border of dentary in lingual view), but not others (e.g., P. tenax, the La Cantera

153

Pharsophorus). Additionally, SGOPV 6200 clearly does not belong to Pharsophorus, as

it much smaller and more gracile than any specimen assigned to this genus.

Although the description of Eomakhaira provides a tantalizing possible look into

the origins of thylacosmilids, knowledge of the species and its anatomy is hampered by

the poor preservation of the senescent holotype. More and better preserved material of

this species will be needed in order to fully understand its evolutionary relationships and

paleoecology. It is even possible that some of the features of Eomakhaira may have to be reinterpreted in the light of new material given the poor preservation of the holotype. For example, given that a metaconid has been reported in the possible basal thylacosmilid from La Venta (IGM 251108), it is possible, though unlikely, that a metaconid was present in m2-3 in Eomakhaira and absent on m4 (as in Arctodictis and Australohyaena)

but merely not observable due to damage during preservation. However, at this time,

there are no undescribed specimens from the Abanico Formation that could represent a

second specimen of Eomakhaira. Perhaps the best way to build on the results of the study of SGOPV 3490 would be a full description of the middle Miocene Anachlysictis gracilis and IGM 251108, to give a better idea of the morphology of the most plesiomorphic undoubted thylacosmilids and to finally be able to examine these species within a

phylogenetic context. Such a study is already in preparation as of this writing (Suarez,

pers. comm., 2018).

154

SUPPLEMENTARY INFORMATION

TABLE OF CONTENTS

Appendix 1. List of comparative materials used in the study of UATF-V-001900...... 158

Appendix 2. Codings for Australogale leptognathus based on the characters of Suarez et al. (2016)...... 159

Appendix 3. Relative proportions of p2 in Australogale leptognathus, Hondadelphys

fieldsi, and hathliacynid sparassodonts...... 160

Appendix 4. Raw molar measurements used to calculate relative proportions of m1-2 in

Table 2.3...... 161

Appendix 5. Relative size of the tooth at the p3 locus in UATF-V-001900 compared to

the dp3 of other sparassodonts scaled by m2...... 162

Appendix 6. Estimated maximum potential depth of the dentary of Australogale

assuming the holotype represents an incompletely grown individual based on

comparisons with juvenile and adult specimens of the hathliacynid Acyon myctoderos.

...... 163

Appendix 7. Methodology and results of the Templeton test...... 164

Appendix 8. Most parsimonious tree recovered under implied weights, showing the

recovered position of Australogale (in bold) within the Hathliacynidae ...... 165

Appendix 9. Premolar proportions in non-sparassodont, non-didelphoid, non-dasyurid metatherians...... 166

Appendix 10. Body mass estimates of UATF-V-001900 (holotype of Australogale leptognathus), UF 27881, and two specimens of pre-late Miocene didelphoids (MLP 77-

VI-13-26 and IGM 184600)...... 168

Appendix 11. List of comparative material examined in the study of SGOPV 3490. .... 169 155

Appendix 12. Relative proportions of the maxilla and the mandible in sparassodonts used

in Figure 4.5...... 171

Appendix 13. Nasal proportions of sparassodonts...... 172

Appendix 14. Raw measurements and references for upper canine data presented in Table

4.3...... 174

Appendix 15. Relative length of the premolar row (p1-3) relative to the molar row (m1-4) in various species of short-snouted borhyaenoids...... 177

Appendix 16. Proportions of the third upper and lower premolars (P/p3) in short-snouted

borhyaenoids...... 179

Appendix 17. Percent labiolingual depth of the ectoflexus of M3 in selected short-

snouted borhyaenoids...... 183

Appendix 18. List of sparassodont specimens for which the state of the roots of the lower

molars could be observed...... 184

Appendix 19. Length of M3 and m4 in various sparassodonts, showing the variation in

size in members of the Proborhyaenidae and Thylacosmilidae...... 185

Appendix 20. Character-state matrix used in the phylogenetic analysis of Eomakhaira

molossus...... 187

Appendix 21. List and description of 321 characters used in the phylogenetic analysis of

SGOPV 3490, modified from Suarez et al. (2016)...... 203

Appendix 22. New and changed character states from Suarez et al. (2016) ...... 229

Appendix 23. List of comparative specimens and references used to code new and revised

characters in the phylogenetic matrix...... 267

156

Appendix 24. List of sparassodont specimens where paired anteroventral maxillary

foramina have been observed...... 269

Appendix 25. List of metatherian taxa in which the palatal canine foramina were absent.

...... 270

Appendix 26. Measurements of the angle of the tooth rows relative to the midline of the skull...... 273

Appendix 27. Recoding of dentary depth values for metatherians in the phylogenetic

matrix ...... 275

157

Appendix 1. List of comparative materials used in the study of UATF-V-001900. In addition, extant didelphoid morphology was assessed based on Voss and Jansa (2009). The specimens column lists specimens or casts examined firsthand, whereas the references column refers to observations from the primary literature Taxon Group Specimens References Acrocyon sectorius Sparassodonta MLP 11-70 Marshall 1978 Acyon myctoderos Sparassodonta UATF-V-000926 Forasiepi et al., 2006 Villarroel and Marshall, Borhyaenidium altiplanicus Sparassodonta - 1983 Borhyaenidium musteloides Sparassodonta MLP 57-X-10-153 Marshall, 1981 Borhyaenidium riggsi Sparassodonta FMNH P14409 Marshall, 1981 Cladosictis centralis Sparassodonta MNHN Col. 5 Marshall, 1981 MACN-A 674; MACN- Cladosictis patagonica Sparassodonta Marshall, 1981 A 5927 Didelphis brachyodonta Didelphoidea MMP 604-M (cast) - Forasiepi et al., 2009; Hesperocynus dolgopolae Didelphoidea - Abello et al., 2015 UCMP 37960, UCMP Hondadelphys fieldsi Sparassodonta Marshall, 1976 39251 MACN-A 1615 (cast); Hyperdidelphys inexpectata Didelphoidea - MACN-A 11654 (cast) Lutreolina sp. Didelphoidea FMNH P14487 (cast) - Lutreolina biforata Didelphoidea MACN-A 7592 (cast) - Myrmecobius fasciatus Myrmecobiidae USNM 83707 - Marshall, 1981; Notogale mitis Sparassodonta - Villarroel and Marshall 1983; Forasiepi, 2009 Notictis ortizi Sparassodonta MACN-A 3996 (cast) Marshall, 1981 Notocynus hermosicus Sparassodonta MLP 11-91 Marshall, 1981 Perathereutes pungens Sparassodonta MACN-A 684 (cast) Marshall, 1981 Plesiofelis schlosseri Sparassodonta MLP 11-114 - Pseudonotictis pusillus Sparassodonta MLP 11-26 Marshall, 1981 MACN-A 647; MACN- Marshall, 1981; Sipalocyon gracilis Sparassodonta A 5938; MLP 11-7; Forasiepi, 2009 YPM-VPPU 15373 Sipalocyon gracilis (S. Sparassodonta MACN-A 686 - “obusta”) Sparassocynus bahiai Didelphoidea MLP 11-119 Abello et al., 2015 Reig et al. 1987, Abello Sparassocynus derivatus Didelphoidea MACN-PV 17909 et al., 2015 Sparassocynus maimarai Didelphoidea - Abello et al., 2015 “Sparassocynus” Didelphoidea - Abello et al., 2015 heterotopicus Thylatheridium cristatum Didelphoidea MACN-A 6443 (cast) - MMP 354-S (cast); Thylophorops chapadmalensis Didelphoidea - MMP 767-M (cast) Zygolestes paranensis Didelphoidea MACN-A 8889 (cast) Reig, 1957; Goin 1997b Zygolestes tatei Didelphoidea - Goin et al., 2000

158

Appendix 2. Codings for Australogale leptognathus based on the characters of Suarez et al. (2016). No changes were made to the scorings of other taxa relative to Suarez et al. (2016). Abbreviations: “?”, missing or unknown data; a, 0/1 (uncertainty); “-” inapplicable characters

?????????? ?????????? ?????????? ?????????? ?????????? ?????????? ?????????? ?????????? ?????????? ????????02 ?????????? ???????0?? ?00?1?011? ?????????? ?????????? ????????0? 1??1111110 011?100a11 ?01-110??? ?????????? ?????????? ?????????? ?????????? ?????????? ?????????? ?????????? ?????????? ?????????? ?????????? ?????????? ???????

159

Appendix 3. Relative proportions of p2 in Australogale leptognathus, Hondadelphys fieldsi, and hathliacynid sparassodonts. Measurements of hathliacynids are from Patterson and Marshall (1978), Marshall (1981), Villarroel and Marshall (1982), Villarroel and Marshall (1983), Forasiepi et al. (2006), Engelman et al. (2015), and direct observation. Measurements of Hondadelphys are from Marshall (1976b) and (Goin, 1997b). All observations are written as “average ± standard deviation (minimum- maximum)”.

Taxon No. of Observations (Specimens) p2 Length/Width Ratio Australogale leptognathus gen. et sp. nov. 1 2.82 Acyon herrerae 2 (1) 3.42 ± 0.04 (3.39-3.44) Acyon myctoderos 2 (2) 2.34 ± 0.00 (2.34) Borhyaenidium altiplanicus 1 3.13 Borhyaenidium musteloides 1 3.12 Cladosictis centralis 2 (1) 2.79 ± 0.05 (2.75-2.83) Cladosictis patagonica 14 (12) 2.55 ± 0.15 (2.31-2.79) Hondadelphys fieldsi 4 (4) 2.76 ± 0.17 (2.52-2.90) Notocynus hermosicus 1 2.5 Notogale mitis 2 (2) 2.98 ± 0.09 (2.92-3.04) Pseudonotictis pusillus 1 3.08 Sipalocyon gracilis 12 (10) 2.65 ± 0.09 (2.50-2.76)

160

Appendix 4. Raw molar measurements used to calculate relative proportions of m1-2 in Table 2.3. All measurements were taken directly from the specimens or casts of the specimens, with the exception of Borhyaenidium altiplanicus, which was measured from figures in Villarroel and Marshall (1983). * = estimated measurement.

Specimen Taxon Lm1 Wm1 Lm2 Wm2 UATF-V-001900 Australogale leptognathus 4.61 1.63 4.69 2.26 MNHN-BOL-V-011889 Borhyaenidium altiplanicus 5.78 2.30 6.19 2.89 MLP 57-X-10-153 Borhyaenidium musteloides 5.63* 2.32 5.37 2.87 FMNH P14407 Borhyaenidium riggsi 5.71 2.05 5.90 2.56 MACN-A 3996 Notictis ortizi 4.30 1.62 4.63 2.04 MLP 11-91 Notocynus hermosicus 5.39 2.27* 6.44* 2.85* MACN-A 684 Perathereutes pungens 4.62 1.98 4.86 2.38 MLP 11-26 Pseudonotictis pusillus 4.49 1.64 4.62 2.11 MACN-A 5938 Sipalocyon gracilis 5.31 2.22 5.87 2.91 MACN-A 647 Sipalocyon gracilis 5.28 2.08 5.88 2.66 Sipalocyon gracilis (S. MACN-A 686 5.89 2.53 6.40 3.08 “obusta”) MACN-A 691 Sipalocyon gracilis 5.91 2.74 6.49 3.35 MLP 11-7 Sipalocyon gracilis 5.81 2.44 6.18 3.09 YPM-VPPU 15373 Sipalocyon gracilis 5.53 2.55 6.12 3.15

161

Appendix 5. Relative size of the tooth at the p3 locus in UATF-V-001900 compared to the dp3 of other sparassodonts scaled by m2. Scaling was done by m2 rather than m1 because of the unusually long m1 of Australogale leptognathus. The only specimen of Prothylacynus patagonicus that preserves dp3, MACN-A 642, does not preserve m2, but based on the length of m1 and proportions of the lower molars of Prothylacynus (based on measurements in (based on measurements in Marshall, 1979) the relative size of dp3 in this taxon appears to be similar to that of Acyon and Borhyaena. Abbreviations: L, length.

Taxon Specimen Ldp/p3 Lm2 % length Reference of m2 Australogale UATF-V- 3.07 4.69 65% Present Study leptognathus 001900 Cladosictis YPM-VPPU 4.63 7.59 61% Yale Peabody Museum of patagonica 15097 Natural History Online Catalog Cladosictis MPM-PV 3646 5.26 8.52 62% Forasiepi, pers. comm. patagonica 2015 Acyon myctoderos UATF-V- 5.68 10.52 54% Engelman et al. (2015) 000926 Borhyaena tuberata MPM-PV 3554 7.34 13.47 54% Forasiepi, pers. comm. 2015

162

Appendix 6. Estimated maximum potential depth of the dentary of Australogale assuming the holotype represents an incompletely grown individual based on comparisons with juvenile and adult specimens of the hathliacynid Acyon myctoderos. The juvenile specimen of A. myctoderos (UATF-V-000926) is based on the stage of molar eruption (m3 fully erupted in both specimens). Measurements of A. myctoderos are from Engelman et al. (2015). * = Estimated or extrapolated measurement. Abbreviations: D, depth of mandibular ramus below tooth; L, length of tooth

m1D/ m2D/ Taxon Specimen m1D m2D Lm2 Lm2 Lm2 Australogale leptognathus UATF-V-001900 4.81 5.03 4.69 1.03 1.07 (actual) Australogale leptognathus UATF-V-001900 5.66* 5.77* 4.69 1.20* 1.23* (estimated) Acyon myctoderos (juvenile) UATF-V-000926 14.51 16.02 10.52 1.38 1.52 Acyon myctoderos (adult) MNHN-Bol-V- 17.07 18.42 10.05 1.70 1.83 003668

163

Appendix 7. Methodology and results of the Templeton test. Following the methodology of Templeton (1983) and Larson (1994), one tree was chosen at random from the pool of most parsimonious trees (MPTs) and compared to the single MPT recovered when Australogale, Didelphis, and Monodelphis were constrained to form a clade. These trees were imported into a single data file, with the first representing the overall best unconstrained tree and the second representing the alternate phylogenetic hypothesis to test against following the script of Schmidt-Lebuhn (2016). Because all MPTs are the same length, it should be irrelevant which MPT is chosen to use in the Templeton test, nevertheless, the test was performed with several different MPTs to double-check the results: the results did not differ in significance. The results of the Templeton test are as follows:

Templeton test of first tree in memory against second tree Rank Difference Rankscore 1 -1.000000 -7000000 2 -1.000000 -7000000 3 1 7 4 1 7 5 1 7 6 1 7 7 1 7 8 1 7 9 1 7 10 1 7 11 1 7 12 -1.000000 -7000000 13 1 7 14 2 14 Sum of negative ranks 21 Number of non-zero values 14 Critical value is 25 for 5 percent, 21 for 2.5 percent, and 15 for 1 percent Significant at the 2.5 percent level

164

Appendix 8. Most parsimonious tree recovered under implied weights, showing the recovered position of Australogale (in bold) within the Hathliacynidae

165

Appendix 9. Premolar proportions in non-sparassodont, non-didelphoid, non-dasyurid metatherians. All observations are based on specimens for which the crowns of p2-3 are preserved in a single individual, or specimens where the alveoli for one tooth are clearly larger than the other. Taxa for which p2 and p3 are nearly equal in size are denoted by an asterisk

Taxa in which p2 > p3 Taxon Classification Reference Glasbius intricatus Glasbiidae Clemens, 1966 Monodelphopsis travassosi Marsupialiformes incertae sedis Marshall, 1987 Didelphidectes pumilis Peradectidae Korth, 1994 Copedelphys titanelix Herpetotheriidae Korth, 1994 Badjcinus turnbulli Thylacinidae Muirhead and Wroe, 1998 Muribacinus gadiyuli Thylacinidae Wroe, 1996 Mayigriphus orbus Dasyuromorphia incertae sedis Wroe, 1998 Myrmecobius fasciatus Myrmecobiidae Engelman pers. obs.

Taxa in which p2 ≤ p3 Taxon Classification Reference Lotheridium mengi Deltatheriidae (Deltatheroida) Bi et al., 2015 Deltatheridium pretrituberculare Deltatheriidae (Deltatheroida) Rougier et al., 1998 Kokopellia juddi Marsupialiformes incertae sedis Cifelli and Muizon, 1997 Asiatherium reshetovi Marsupialiformes incertae sedis Szalay and Trofimov, 1996 Alphadon halleyi Marsupialiformes incertae sedis Montellano, 1988 Alphadon marshi Marsupialiformes incertae sedis Clemens, 1966 Alphadon wilsoni Marsupialiformes incertae sedis Lillegraven, 1969 Pediomyidae Pediomyidae Clemens, 1966; Lillegraven, 1969 Eodelphis spp. Stagodontidae Scott and Fox, 2015 Didelphodon spp. Stagodontidae Fox and Naylor, 2006; Wilson et al. 2016 Herpetotheriidae (except C. Herpetotheriidae Korth, 1994; Horovitz, 2008 titanelix) Microbiotheriidae Microbiotheriidae Marshall, 1982a Pucadelphys andinus Pucadelphyidae Marshall and Muizon, 1995 Andinodelphys cochabambensis Pucadelphyidae Muizon et al., 1997 Szalinia gracilis Marsupialiformes incertae sedis Muizon and Cifelli, 2001 Peradectes elegans* Peradectidae Simpson, 1935 ?Peradectes elegans Peradectidae Fox, 1983 Djarthia murgonensis Australidelphia incertae sedis Godthelp et al., 1999; Beck et al., 2008 Bobbschaefferia fluminensis Polydolopimorphia incertae sedis Marshall, 1987 Caroloameghinia mater* Caroloameghiniidae Marshall, 1982b Didelphopsis cabrerai Sternbergiidae Marshall, 1987 Eobrasilia coutoi Marsupialiformes incertae sedis Marshall, 1987 Gaylordia spp. Marsupialiformes incertae sedis Marshall, 1987 Guggenheimia brasiliensis* Protodidelphidae Marshall, 1987 Itaboraidelphys camposi* Sternbergiidae Marshall, 1987

166

Marmosopsis juradoi* Marsupialiformes incertae sedis Marshall, 1987 Minisculodelphys minimus* Jaskhadelphyidae Marshall, 1987 Mirandatherium alipioi Marsupialiformes incertae sedis Marshall, 1987 Protodidelphis vanzolinii Protodidelphidae Marshall, 1987; Oliveira and Goin, 2011 Protodidelphis mastodontoides Protodidelphidae Marshall, 1987; Oliveira and Goin, 2011 Sternbergia itaboraisensis Sternbergiidae Marshall, 1987

167

Appendix 10. Body mass estimates of UATF-V-001900 (holotype of Australogale leptognathus), UF 27881, and two specimens of pre-late Miocene didelphoids (MLP 77- VI-13-26 and IGM 184600). Measurements of MLP 77-VI-13-26 and IGM 184600 taken from Goin and Abello (2013) and figures in Goin (1997b), respectively. Goin (1997b) does not attempt to identify the position of the isolated lower molar figured in this paper (IGM 184600), though based on its morphology the tooth clearly does not represent m4 (and most likely pertains to m2-3 based on comparisons with other didelphoids) so body mass estimates were restricted to m1-3.

Taxon Specimen Measurement Equation Mass (g) Australogale UATF-V- Gordon (2003), didelphids only 954 leptognathus 001900 Gordon (2003), dasyurids only 1192 m1 Gordon (2003), didelphids + dasyurids 1111 Average 1086 Gordon (2003), didelphids only 861 Gordon (2003), dasyurids only 819 m2 Gordon (2003), didelphids + dasyurids 840 Average 839 Sparassodonta UF 27881 LM1-4, from Myers (2001), dasyuromorphians 940 gen. et sp. nov. alveoli Didelphoidea MLP 77- LM1 Gordon (2003), didelphids only 461 indet. VI-13-26 WM1 Gordon (2003), didelphids only 312 LM2 Gordon (2003), didelphids only 447 LM3 Gordon (2003), didelphids only 391 - Average 403 ?Didelphinae IGM Lm?1 Gordon (2003), didelphids only 532 indet. 184600 Lm?2 Gordon (2003), didelphids only 443 Lm?3 Gordon (2003), didelphids only 373

168

Appendix 11. List of comparative material examined in the study of SGOPV 3490. Specimen refers to material that could be observed directly, whereas references refer to data and observation taken from the primary literature.

Taxon Specimens References Acrocyon riggsi FMNH P13433 Goin et al., 2007 Acrocyon sectorius Marshall, 1978 Anachlysictis gracilis — Goin, 1997b Arctodictis sinclairi MLP 77-VI-13-1, MLP 85- Forasiepi, 2009 VII-3-1 Arctodictis munizi — Forasiepi et al., 2004 Arminiheringia auceta MACN-A 10970/10972 — (holotype) Arminiheringia contigua MACN-A 10317 (cast) — Arminiheringia sp. — Zimicz, 2012; Forasiepi and Sanchez-Villagra, 2014; Forasiepi personal comm. Australohyaena antiquua MACN-A 52-322; FMNH Forasiepi et al., 2015 P13633 Borhyaena macrodonta MACN-A 52-366 Borhyaena tuberata MACN-A 6203-6265 Sinclair, 1906; Cabrera 1927; Forasiepi, 2009 Callistoe vincei — Babot et al., 2002 Paraborhyaena boliviana UATF-V-000129 Petter and Hoffstetter, 1983 Patagosmilus goini — Forasiepi and Carlini, 2010; Forasiepi personal. comm. Pharsophorus lacerans MACN-A 52-391 (holotype), Patterson and Marshall, YPM-VPPU/MNHN SAL 1978; Petter and Hoffstetter, 1983 Pharsophorus tenax AC 3004 (cast), AC 3192 Marshall, 1978 (cast) Pharsophorus cf. P. MPEF-PV 4190 Goin et al., 2010; Zimicz, lacerans 2012 Plesiofelis schlosseri MLP 11-114 (holotype) — Proborhyaena gigantea AMNH 29576, MACN-A 52-382 cf. Proborhyaena MLP 79-XII-18-1 Bond and Pascual, 1983 Proborhyaenidae indet. — Couto Ribeiro, 2010

169

Prothylacynus patagonicus Forasiepi, 2009 Sarcophilus harrissii CMNH 18915 — Thylacosmilidae? gen. et — Goin, 1997b sp. nov. (IGM 251108) Thylacosmilidae indet. — Goin et al., 2007

170

Appendix 12. Relative proportions of the maxilla and the mandible in sparassodonts used in Figure 4.5. Maxilla height measured from the dorsalmost point of maxilla to the alveolar border. Mandible height measured at m3/4 embrasure.

Maxilla Dentary Relative Relative Taxon Specimen Family m1-4 Height Depth (m4) Maxilla Height Dentary Depth Eomakhaira molossus SGO-PV 3490 Proborhyaenidae 37.3 42.8 31.8 1.15 0.85 Basal Lycopsis longirostrus UCMP 38061 58 44.81 30.29 0.77 0.52 Borhyaenoidea Prothylacynus Basal YPM-VPPU 15700 47.98 38.08 32.69 0.79 0.68 patagonicus Borhyaenoidea Arctodictis munizi CORD-PZ 1210-1/2 Borhyaenidae 51.3 74.18 56.13 1.45 1.09 Arctodictis sinclairi MLP 85-VII-3-1 Borhyaenidae 45 46.39 41.15 1.03 0.91 Australohyaena UNPSJB PV 113 Borhyaenidae 61 74.9 58.79 1.23 0.96 antiquua Borhyaena tuberata YPM-VPPU 15701 Borhyaenidae 54.43 46.12 36.67 0.85 0.67 Acyon myctoderos MNHN-Bol-V-003668 Hathliacynidae 41.1 27.08 18.94 0.66 0.46 Cladosictis patagonica MACN-A 5927 Hathliacynidae 30.7 24.27 25.85 0.79 0.84 Cladosictis patagonica YPM-VPPU 15170 Hathliacynidae 33.13 21.64 16.45 0.65 0.50 Sipalocyon gracilis YPM-VPPU 15373 Hathliacynidae 29.94 14.94 10.45 0.50 0.35 Arminiheringia auceta MACN-A 10970/10972 Proborhyaenidae 53.34 44.7 50.8 0.84 0.95 Callistoe vincei PVL 4187 Proborhyaenidae 51.05 45.91 45.95 0.90 0.90 Thylacosmilus atrox MMP 1443 Thylacosmilidae 56.39 97.03 42.16 1.72 0.75 Thylacosmilus atrox FMNH P145312 Thylacosmilidae 62.65 99.48 41.06 1.59 0.66

1 Estimated based on extent of lacrimal and frontal 2 Based on Plate I, in which the mandible has been partially restored after FMNH P14344 171

Appendix 13. Nasal proportions of sparassodonts. Relative width of anterior nasals to posterior nasals (Ratio) calculated by dividing width of anterior nasals (ANa) by greatest width of nasals posteriorly (PNa). Relative size of nasals calculated by dividing greatest width of nasals posteriorly by length of M3 (LM3) as a scaling measurement. Similar results were achieved when M1-3 was used as a scaling measurement. Thylacosmilus could not be included in this analysis as although it has been noted to have narrow anterior nasals the anterior portion of the nasals are enclosed inside the skull beneath the enlarged alveoli for the upper canines. All nasal measurements were calculated based on one side of the specimen in order to compare with taxa for which the nasals are bilaterally incomplete.

PNa/ Taxon Specimen Family ANa PNa Ratio LM3 LM3 Reference Eomakhaira SGO-PV 3490 Proborhyaenidae 3.12 12.4 0.25 7.9 1.57 Present Study molossus Acrocyon riggsi FMNH P13433 Borhyaenidae 7.42 23.02 0.32 12.75 1.81 Engelman pers. obs. CORD-PZ 1210- Arctodictis munizi Borhyaenidae 11.83 35.36 0.33 16 2.21 Forasiepi et al. (2004) 1/2 Arctodictis MLP 85-VII-3-1 Borhyaenidae 8.48 27.65 0.31 12.8 2.17 Forasiepi (2009) sinclairi Australohyaena UNPSJB-PV 113 Borhyaenidae 13.39 44.89 0.30 16.65 2.70 Forasiepi et al. (2015) antiqua Borhyaena MPM-PV 3625 Borhyaenidae 10.49 34.15 0.31 21.27 1.61 Ercoli et al. (2014) tuberata Borhyaena YPM-VPPU 15120 Borhyaenidae 6.54 23.94 0.27 14.22 1.68 Sinclair (1906) tuberata Borhyaena YPM-VPPU 15701 Borhyaenidae 9.97 29.84 0.33 15.01 1.99 Sinclair (1906) tuberata Pharsophorus MNHN SAL 96/ Patterson and Marshall Basal Borhyaenoidea 8.25 25.85 0.32 — — lacerans YPM-VPPU 20551 (1978)

172

Prothylacynus MACN-A 5931 Basal Borhyaenoidea 4.53 18.83 0.24 13.1 1.44 Engelman pers. obs. patagonicus (juv.) Prothylacynus Marshall (1979), MACN-A 14453 Basal Borhyaenoidea 6.97 22.66 0.31 12.5 1.81 patagonicus Engelman pers. obs. Callistoe vincei PVL 4187 Proborhyaenidae 7.81 15.15 0.52 13.3 1.14 Babot et al. (2002) Paraborhyaena Petter and Hoffstetter MNHN SAL 51 Proborhyaenidae — 39.73 — 24 1.66 boliviana (1983) Arminiheringia Simpson (1948), MACN-A 10970 Proborhyaenidae 9.14 33.53 0.27 14 2.39 auceta Engelman pers. obs. Forasiepi and Carlini Patagosmilus MLP 07-VII-1-1 Thylacosmilidae 5.03 16.35 0.31 13 1.26 (2010), Forasiepi pers. goini comm. MNHN-Bol-V- Acyon myctoderos Hathliacynidae 5.92 16.01 0.37 10.85 1.48 Forasiepi et al. (2006) 003668 Cladosictis MACN-A 5927 Hathliacynidae 3.94 12.54 0.31 8.85 1.42 Engelman pers. obs. patagonica Cladosictis YPM-VPPU 15702 Hathliacynidae 3.69 12.70 0.29 8.92 1.42 Sinclair (1906) patagonica Sipalocyon AMNH 9254 Hathliacynidae 3.25 9.45 0.34 6.35 1.49 Sinclair (1906) gracilis Sallacyon Petter and Hoffstetter MNHN SAL 92 Hathliacynidae — 9.14 — 6.1 1.50 hoffstetteri (1983) ?Hondadelphys IGM 250364 Basal Sparassodonta 3.37 7.55 0.45 6.5 1.16 Goin (1997b) sp. Sparassodonta Sparassodonta incertae Engelman and Croft UF 27881 2.70 6.54 0.41 4.4 1.49 gen. et sp. nov. sedis (2014)

173

Appendix 14. Raw measurements and references for upper canine data presented in Table 4.3. Abbreviations: LC, greatest anteroposterior length of canine; WC, greatest labiolingual width of canine perpendicular to LC; LM3, length of M3. Canine shape (L/W ratio) was calculated by dividing LC by WC. Relative canine size was determined by calculating the average diameter of the canine in order to account for differences in shape between sparassodonts (e.g., the extremely labiolingually narrow canines of thylacosmilids) by treating LC and WC as the major and minor axes of an ellipse and using the formula LC WC = 2 2 2 2 𝑑𝑑 where d equals the average diameter of the canine tooth,𝜋𝜋 �then� scaling𝜋𝜋 � this� measurement� � by dividing by LM3. LM3 was chosen as the scaling variable as it was the measurement that could be compared in the greatest number of specimens, though similar results with minor differences in rank order were obtained when using M1-3 as the scaling variable. * = juvenile individuals (individuals in which all of the teeth have not fully erupted). Juvenile individuals are highlighted as they are expected to have canines that appear relatively wider labiolingually at the exposed base of the canine due to the late eruption and apexification of the canine in sparassodonts (Chemisquy and Prevosti, 2014; Engelman and Croft, 2014; Forasiepi and Sánchez-Villagra, 2014).

L/W Relative Taxon Specimen Family LC WC LM3 Ratio Size Reference Eomakhaira molossus SGOPV 3490 Proborhyaenidae 11.6 6.8 7.9 1.71 1.12 Present Study Arminiheringia auceta MACN-A 10972 Proborhyaenidae 19.6 13.2 15.5 1.48 1.04 Engelman and Croft (2014) Callistoe vincei PVL 4187 Proborhyaenidae 20 14.5 13.3 1.38 1.28 Engelman and Croft (2014) cf. Proborhyaena MLP 79-XII-18-1 Proborhyaenidae 29 16.6 — 1.75 — Bond and Pascual (1983)

Paraborhyaena boliviana MNHN SAL 51 Proborhyaenidae 24.6 17 24 1.45 0.85 Petter and Hoffstetter (1983)

MHNT-VT-1400/ Proborhyaenidae indet. Proborhyaenidae 14.8 10.2 19.3 1.45 0.64 Couto-Ribeiro (2010) 1401 Thylacosmilus atrox MLP 35-X-4-1 Thylacosmilidae 28.4 11.6 16 2.45 1.13 Engelman and Croft (2014) Thylacosmilus atrox FMNH P14531 Thylacosmilidae 35 14 20 2.50 1.11 Engelman and Croft (2014)

174

Thylacosmilus atrox MMP 1470 Thylacosmilidae 21 8 15.5 2.63 0.84 Engelman and Croft (2014) IGM 251108 IGM 251108 ?Thylacosmilidae 9.2 5.9 — 1.56 — Goin (1997b) cf. Dukecynus sp.* UCMP 32950 Basal Borhyaenoidea 19 13.5 — 1.41 — Marshall (1978b) cf. Pharsophorus* AMNH 29591 Basal Borhyaenoidea 12.85 8.3 — 1.55 — Engelman, pers. obs. Lycopsis longirostrus* UCMP 38061 Basal Borhyaenoidea 10 7.9 17.5 1.27 0.51 Engelman and Croft (2014) MNHN SAL 96/ Pharsophorus lacerans Basal Borhyaenoidea 13.8 10.2 — 1.35 — Petter and Hoffstetter (1983) YPM-VPPU 20551

Prothylacynus patagonicus MACN 11453 Basal Borhyaenoidea 11.2 7.75 12.35 1.45 0.75 Engelman and Croft (2014)

Prothylacynus patagonicus* MACN-A 5931 Basal Borhyaenoidea 9.2 7.3 12.3 1.26 0.67 Engelman and Croft (2014)

Hondadelphys fieldsi UCMP 37960 Basal Sparassodonta 8.5 4.7 7.9 1.81 0.80 Engelman and Croft (2014) Acrocyon riggsi FMNH P13433 Borhyaenidae 15 11 12.8 1.36 1.00 Engelman and Croft (2014) Arctodictis munizi MLP 11-65 Borhyaenidae 28 17.5 17.2 1.60 1.29 Engelman and Croft (2014) Arctodictis munizi CORD-PZ 1210-1/5 Borhyaenidae 24.57 16.82 16 1.46 1.27 Engelman and Croft (2014) Arctodictis sinclairi MLP 85-VII-3-1 Borhyaenidae 18.15 14 13.7 1.30 1.16 Engelman and Croft (2014)

Australohyaena antiquua UNPSJB-PV 113 Borhyaenidae 21.55 16.05 16.65 1.34 1.12 Forasiepi et al. (2015)

Australohyaena antiquua FMNH P13633 Borhyaenidae 23.3 16.4 — 1.42 — Forasiepi et al. (2015)

Borhyaena macrodonta MACN 52-390 Borhyaenidae 17.7 11.7 15.6 1.51 0.92 Marshall (1978b) Borhyaena tuberata MACN 6203-6265 Borhyaenidae 14.8 10.1 14.4 1.47 0.85 Engelman and Croft (2014) Borhyaena tuberata MACN 5780 Borhyaenidae 15.5 10.2 — 1.52 — Engelman and Croft (2014) Borhyaena tuberata YPM-VPPU 15701 Borhyaenidae 16.5 13.5 15 1.22 0.99 Sinclair (1906) Borhyaena tuberata YPM-VPPU 15120 Borhyaenidae 14.5 10 14 1.45 0.86 Sinclair (1906)

175

MNHN-Bol-V- Acyon myctoderos Hathliacynidae 8.95 6.3 12.5 1.42 0.60 Engelman and Croft (2014) 003668 Borhyaenidium riggsi FMNH P14409 Hathliacynidae 4 2.8 7.3 1.43 0.46 Engelman and Croft (2014) Cladosictis centralis MACN 11639 Hathliacynidae 9 5.5 8.4 1.64 0.84 Engelman and Croft (2014) Cladosictis patagonica MACN 5927 Hathliacynidae 10 6.7 9 1.49 0.91 Engelman and Croft (2014) Cladosictis patagonica MACN 6280-6285 Hathliacynidae 6.34 4.46 8.6 1.42 0.62 Engelman, pers. obs. Cladosictis patagonica AMNH 9134 Hathliacynidae 7.54 4.96 8.5 1.52 0.72 Engelman, pers. obs. Cladosictis patagonica YPM-VPPU 15046 Hathliacynidae 7.5 5 8.6 1.50 0.71 Sinclair (1906) Cladosictis patagonica YPM-VPPU 15170 Hathliacynidae 9 5.5 8.5 1.64 0.83 Sinclair (1906) Cladosictis patagonica YPM-VPPU 15702 Hathliacynidae 10 7 8.8 1.43 0.95 Sinclair (1906) Notogale mitis YPM-VPPU 21871 Hathliacynidae 7.03 4.78 — 1.47 — Engelman, pers. obs. Sipalocyon externa MACN-A 52-383 Hathliacynidae 5.6 3.9 6.5 1.44 0.72 Engelman and Croft (2014) Sipalocyon gracilis MACN-A 692 Hathliacynidae 5.92 3.93 6.14 1.51 0.79 Engelman, pers. obs. Sipalocyon gracilis YPM-VPPU 15373 Hathliacynidae 4.5 3.05 6.03 1.48 0.61 Engelman, pers. obs. Sipalocyon gracilis AMNH 9254 Hathliacynidae 6.38 4.11 6.38 1.55 0.80 Engelman, pers. obs. Sipalocyon gracilis YPM-VPPU 15029 Hathliacynidae 5.8 4 6.8 1.45 0.71 Sinclair (1906) Sipalocyon gracilis YPM-VPPU 15154 Hathliacynidae 6 4.2 5.5 1.43 0.91 Sinclair (1906) Sparassodonta gen. et sp. UF 27881 incertae sedis 4.4 3.6 4.44 1.22 0.90 Engelman and Croft (2014) nov. Sparassodonta indet. MUSM 1649 incertae sedis 11.22 8.9 — 1.26 — Tejada-Lara et al. (2015)

176

Appendix 15. Relative length of the premolar row (p1-3) relative to the molar row (m1-4) in various species of short-snouted borhyaenoids. * = premolar row lengths that were estimated by subtracting the reported lengths of m1-4 from p1-m4. Specimens of thylacosmilids could not be compared because thylacosmilids have only two premolars and Thylacosmilus exhibits a reorganization of the molar row related to the unusual craniodental anatomy of this taxon (Goin, 1997b).

Lp1-3/ Taxon Specimen Lp1-3 Lm1-4 Lm1-4 Reference Eomakhaira molossus SGOPV 3490 16.4 37.3 0.44 Present Study Arctodictis munizi MACN 5915 32* 67 0.48 Marshall (1978b) Arctodictis munizi MACN 5919 34.4* 64 0.54 Marshall (1978b) Arctodictis munizi MLP 11-65 38.9* 60.6 0.64 Marshall (1978b) Arctodictis munizi MLP 11-85 37.5* 63 0.60 Marshall (1978b) MLP 85-VII- Arctodictis sinclairi 24.51 44.68 0.55 Present Study 3-1 MACN-A Arminiheringia auceta 40.09 58.07 0.69 Present Study 10970 UNPSJB-PV Forasiepi et al. Australohyaena antiquua 37* 61 0.61 113 (2015) MACN-A 52- Borhyaena macrodonta 36.64 57.15 0.64 Present Study 366 YPM-VPPU Borhyaena tuberata 33.68 50.78 0.66 Sinclair (1906) 15701 Borhyaena tuberata MACN 9342 32.5* 50.5 0.64 Marshall (1978b) Borhyaena tuberata MACN 9345 32.35* 51.5 0.63 Marshall (1978b) Babot et al. Callistoe vincei PVL 4187 35.69 51.7 0.69 (2002) UATF-V- Paraborhyaena boliviana ~33.0 86.81 0.383 Present Study 000129 MNHN SAL Petter and Paraborhyaena boliviana 37.7 82.1 0.46 51 Hoffstetter (1983) MPEF-PV Pharsophorus cf. lacerans 32.77 58.44 0.56 Present Study 4170 MACN-A 52- Pharsophorus lacerans 34.9 55.93 0.62 Present Study 391 Pharsophorus lacerans MACN 11653 35* 56 0.63 Marshall (1978b) Proborhyaena gigantea AMNH 29576 61.88 109.71 0.56 Present Study

3 Measurement uncertain. The region of the dentary between the lower canine and anterior root of p3 is not well-preserved in this specimen. 177

MACN-A 52- Proborhyaena gigantea 48.35 88.02 0.55 Present Study 382 Prothylacynus patagonicus MACN-A 706 26.05 47.74 0.55 Present Study YPM-VPPU Prothylacynus patagonicus 26.91 48.78 0.55 Sinclair (1906) 15700 Prothylacynus patagonicus MLP 11-38 24.7* 47.3 0.52 Marshall (1979) MACN 52- Pseudothylacynus rectus 27.4* 41.4 0.66 Marshall (1979) 369

178

Appendix 16. Proportions of the third upper and lower premolars (P/p3) in short-snouted borhyaenoids. * = Size of P3 relative to p3 using the left P3 of SGOPV 3490 calculated using the right p3, as the left p3 is not preserved enough to give an accurate measurement.

LP3/ Lp3/ LP3/ Taxon Specimen Family LP3 WP3 Lp3 Wp3 WP3 Wp3 Lp3 Reference SGO-PV 3490 Eomakhaira molossus Proborhyaenidae 8.58 4.9 — — 1.75 — 1.13* Present Study (left) SGO-PV 3490 Eomakhaira molossus Proborhyaenidae 9.06 5.2 7.57 4.09 1.74 1.85 1.20 Present Study (right) Basal Shockey and Fredszalaya hunteri UF 172501 8.8 6.4 — — 1.38 — — Borhyaenoidea Anaya (2008) Pharsophorus cf. P. MPEF-PV Basal Goin et al. — — 12 8.85 — 1.36 — lacerans 4190 Borhyaenoidea (2010) YPM-VPPU Basal Patterson and Pharsophorus lacerans 10.4 7.7 — — 1.35 — — 20551 Borhyaenoidea Marshall (1978) MACN-A 52- Basal Patterson and Pharsophorus lacerans — — 14.6 7.3 — 2.00 — 391 Borhyaenoidea Marshall (1978) MACN-A Basal Patterson and Pharsophorus lacerans — — 11.9 6.2 — 1.92 — 11653 Borhyaenoidea Marshall (1978) Basal Patterson and Pharsophorus tenax AC 3192/3004 11.4 6.9 12.25 6.08 1.65 2.01 0.93 Borhyaenoidea Marshall (1978) Basal Patterson and Plesiofelis schlosseri MLP 11-114 — — 11.00 6.50 — 1.69 — Borhyaenoidea Marshall (1978) Prothylacynus Basal MACN 11453 8.60 4.60 — — 1.87 — — Marshall (1979) patagonicus Borhyaenoidea Prothylacynus Basal MACN 11453 8.60 4.40 — — 1.95 — — Marshall (1979) patagonicus Borhyaenoidea

179

Prothylacynus YPM-VPPU Basal 9.98 5.93 10.09 4.88 1.68 2.07 0.99 Sinclair (1906) patagonicus 15700 Borhyaenoidea FMNH Marshall Acrocyon riggsi Borhyaenidae 10.10 7.40 — — 1.36 — — P13233 (1978b) FMNH Marshall Acrocyon riggsi Borhyaenidae 10.00 7.50 — — 1.33 — — P13233 (1978b) MACN 9364- Marshall Acrocyon sectorius Borhyaenidae 10.20 6.50 — — 1.57 — — 9385 (1978b) MACN 9364- Marshall Acrocyon sectorius Borhyaenidae 10.30 6.20 — — 1.66 — — 9385 (1978b) MACN 5918- Marshall Arctodictis munizi Borhyaenidae 14.10 10.50 15.00 10.30 1.34 1.46 0.94 5921 (1978b) Marshall Arctodictis munizi MLP 11-65 Borhyaenidae 14.00 9.50 13.00 8.00 1.47 1.63 1.08 (1978b) Marshall Arctodictis munizi MLP 11-65 Borhyaenidae 13.50 9.00 — — 1.50 — — (1978b) Marshall Arctodictis munizi MLP 11-85 Borhyaenidae 15.00 8.50 1.76 0.00 (1978b) CORD-PZ Forasiepi et al. Arctodictis munizi Borhyaenidae 13 9.5 1.37 1210-1/5 (2004) FMNH Marshall Arctodictis sinclairi Borhyaenidae — — 11.00 6.70 — 1.64 — P13526 (1978b) MLP 82-V-2- Goin et al. Arctodictis sinclairi Borhyaenidae 10.00 7.70 — — 1.30 — — 116 (2007) MLP 85-VII- Arctodictis sinclairi Borhyaenidae 11 8.5 10.2 6.5 1.29 1.57 1.08 Forasiepi (2009) 3-1 Australohyaena UNPSJB-PV Forasiepi et al. Borhyaenidae 12.4 9.1 12.6 8.2 1.36 1.54 0.98 antiqua 113 (2015) MACN-A 52- Marshall Borhyaena macrodonta Borhyaenidae 12.5 8.1 12.2 6.4 1.54 1.91 1.02 390 (1978b)

180

MACN-A 52- Marshall Borhyaena macrodonta Borhyaenidae 12.6 8 — — 1.58 — — 390 (1978b) MPEF-PV Marshall Borhyaena macrodonta Borhyaenidae 11.00 8.22 — — 1.34 — — 1467 (1978b) MACN 6203- Marshall Borhyaena tuberata Borhyaenidae 12.20 9.30 11.60 7.10 1.31 1.63 1.05 6265 (1978b) MACN 6203- Marshall Borhyaena tuberata Borhyaenidae 12.60 9.50 12.00 7.30 1.33 1.64 1.05 6265 (1978b) MACN 9341- Marshall Borhyaena tuberata Borhyaenidae 11.70 8.60 — — 1.36 — — 9342 (1978b) MACN 9341- Marshall Borhyaena tuberata Borhyaenidae 12.00 8.30 11.30 6.50 1.45 1.74 1.06 9342 (1978b) MACN 9344- Marshall Borhyaena tuberata Borhyaenidae 11.40 7.70 11.00 6.60 1.48 1.67 1.04 9349 (1978b) MACN 9344- Marshall Borhyaena tuberata Borhyaenidae 11.30 7.60 10.90 6.50 1.49 1.68 1.04 9349 (1978b) MACN Arminiheringia auceta 10970/ 10972 Proborhyaenidae N/A N/A 10.60 6.60 — 1.61 — Croft pers. obs. (left) MACN Arminiheringia auceta 10970/ 10972 Proborhyaenidae 11.50 6.50 11.30 7.10 1.77 1.59 1.02 Croft pers. obs. (right) Arminiheringia MACN-A Engelman pers. Proborhyaenidae — — 8.98 5.45 — 1.65 — contigua 10317 obs. Engelman pers. Arminiheringia sp. MLP Proborhyaenidae 11.58 8.71 — — 1.33 — — obs. Forasiepi and MLP 82-V-1- Arminiheringia sp. Proborhyaenidae 13.99 7.54 — — 1.86 — — Sánchez- 1 Villagra (2014) Babot et al. Callistoe vincei PVL 4187 Proborhyaenidae 8.5 6.6 8.5 4.5 1.29 1.89 1.00 (2002)

181

Babot et al. Callistoe vincei PVL 4207 Proborhyaenidae 8.0 5.5 — — 1.45 — — (2002) MLP 79-XII- Bond and cf. Proborhyaena Proborhyaenidae 32.5 14.99 — — 2.17 — — 18-1 Pascual (1983) Petter and Paraborhyaena MNHN SAL Proborhyaenidae — — 21.13 11.37 — 1.86 — Hoffstetter boliviana 51 (1983) Proborhyaena cf. MNHN-DP Mones and Proborhyaenidae — — 16.30 10.65 — 1.53 — gigantea 720 Ubilla (1978) MACN-A 52- Patterson and Proborhyaena gigantea Proborhyaenidae — — 20.5 11.5 — 1.78 — 382 Marshall (1978) Patterson and Proborhyaena gigantea AMNH 29576 Proborhyaenidae — — 22.5 14.5 — 1.55 — Marshall (1978)

182

Appendix 17. Percent labiolingual depth of the ectoflexus of M3 in selected short- snouted borhyaenoids.

M3 M3 % Ectoflexus ectoflexus width Taxon Specimen Family Width depth (mm) (mm) Eomakhaira SGO-PV 3490 Proborhyaenidae 11.17% 0.21 1.88 molossus Pharsophorus Basal AC 3192 9.07% 0.90 9.93 tenax Borhyaenoidea Prothylacynus Basal MACN-A 707 5.72% 0.53 9.27 patagonicus Borhyaenoidea Prothylacynus Basal MACN-A 707 7.92% 0.53 6.69 patagonicus Borhyaenoidea Prothylacynus Basal MACN-A 5931 9.23% 0.94 10.24 patagonicus Borhyaenoidea Acrocyon riggsi FMNH P13433 Borhyaenidae 8.62% 0.79 9.17 Borhyaena MACN-A 52- Borhyaenidae 0.00% 0 10.48 macrodonta 390 Borhyaena MACN-A 6203 Borhyaenidae 0.00% 0 11.38 tuberata Australohyaena FMNH P13800 Borhyaenidae 5.44% 0.78 14.33 antiquua Arminiheringia MACN-A 10972 Proborhyaenidae 9.68% 1.57 16.22 auceta Arminiheringia MACN-A 10972 Proborhyaenidae 10.80% 1.74 16.11 auceta Proborhyaena AMNH 29576 Proborhyaenidae 12.21% 2.57 21.04 gigantea Patagosmilus MLP 07-VII-1-1 Thylacosmilidae 10.13% 1.23 12.15 goini Thylacosmilidae MLP 92-X-10-6 Thylacosmilidae 5.12% 0.55 10.75 indet.

183

Appendix 18. List of sparassodont specimens for which the state of the roots of the lower molars could be observed. State of the molar roots was determined either by direct observation from isolated teeth or X-rays/CT scans or inferred based on the apparent size of the roots at the alveolar line.

Taxon Specimen Family Method Anterior root of m3-4 much larger than posterior Eomakhaira molossus SGO-PV 3490 Proborhyaenidae Direct observation Acrocyon sectorius YPM-VPPU Borhyaenidae Direct observation 15210 Arctodictis sinclairi MLP 85-VII-3- Borhyaenidae Direct observation 1 Arminiheringia auceta MACN-A- Proborhyaenidae Inferred 109790 Borhyaena tuberata MACN-A 6203 Borhyaenidae Direct observation Paraborhyaena UATF-V- Proborhyaenidae Inferred boliviana 000129 Proborhyaena gigantea MNHN-DP Proborhyaenidae Direct observation 720 Proborhyaenidae indet. MLP 88-V-10- Proborhyaenidae Direct observation 4 Thylacosmilus atrox —4 Thylacosmilidae Direct observation Anterior and posterior roots of m3-4 similar in size Acyon myctoderos UATF-V- Hathliacynidae Direct observation 000926 Australohyaena antiqua UNPSJB-PV Borhyaenidae Direct observation 113 Cladosictis patagonica MACN-A 674 Hathliacynidae Direct observation Hondadelphys fieldsi UCMP 37960 Basal Sparassodonta Inferred (tooth partially displaced from alveoli) Notogale mitis AC 3060 Hathliacynidae Direct observation Pharsophorus lacerans MACN 11-652 Basal Borhyaenoidea Direct observation Stylocynus paranensis MLP 11-94 Basal Sparassodonta Inferred

4 Based on comments in Goin and Pascual (1987). 184

Appendix 19. Length of M3 and m4 in various sparassodonts, showing the variation in size in members of the Proborhyaenidae and Thylacosmilidae. Specimens of Arminiheringia cultrata and A. contigua with M3 or m4 could not be measured.

Taxon Specimen Age and Locality LM3 Lm4 Reference Callistoe vincei PVL 4187 Middle Eocene; Lumbrera Formation 13.3 17.0 Babot et al. (2002) Callistoe vincei PVL 4207 Middle Eocene; Lumbrera Formation 13.0 — Babot et al. (2002) Arminiheringia cultrata MACN-A 10317 Middle Eocene, Gran Barranca Member — 19.3 Engelman, pers. obs. Arminiheringia auceta MACN-A 10324 Middle Eocene, Gran Barranca Member — 20.55 Engelman, pers. obs. MACN-A 10970/ Arminiheringia auceta Middle Eocene, Gran Barranca Member 17.2 21.0 Croft, pers. obs. 10972

Proborhyaenidae aff. MLP 88-V-10-4 Late Eocene, Antofagasta de la Sierra — 13.95 Goin et al. (1998) Eomakhaira molossus Eomakhaira molossus SGOPV 3490 Early Oligocene; Cachapoal 7.9 12.0 Present Study Proborhyaena gigantea AMNH 29576 Late Oligocene, Rinconada de los Lopez 29.6 35.6 Patterson and Marshall (1978) Proborhyaena gigantea MACN-A 52-382 Late Oligocene, uncertain provenance — ~31.5 Patterson and Marshall (1978) Proborhyaena gigantea MNHN-DP 720 Late Oligocene, Fray Bentos Formation — 28.8 Mones and Ubilla (1978) Paraborhyaena boliviana MNHN SAL 51 Late Oligocene, Salla ~24 ~25 Petter and Hoffstetter (1983) Paraborhyaena boliviana UATF-V-000129 Late Oligocene, Salla — 30.26 Engelman pers. obs. Proborhyaenidae sp. nov.? MHNT-VT-1401 Late Oligocene, Tremembé 19.3 — Couto-Ribeiro (2010) Thylacosmilidae indet. MLP 96-X-10-6 Early Miocene, Colhue Huapi Member 7.5 — Goin et al. (2007) Patagosmilus goini MLP 07-VII-1-1 Early Middle Miocene, Collón Curá Formation 13 — Forasiepi pers. comm. ?Thylacosmilidae indet. IGM 251108 Late Middle Miocene, La Venta — 8.7 Goin (1997b) Anachlysictis gracilis IGM 184247 Late Middle Miocene, La Venta — 13.4 Goin (1997b) Thylacosmilus atrox MACN-A 5892 Late Miocene, Ituzaingó Formation — 12.4 Marshall (1976a) Thylacosmilus atrox FMNH P14344 Late Miocene, Andalgalá Formation — 16.4 Marshall (1976a) Thylacosmilus atrox FMNH P14531 Late Miocene, Corral Quemado Formation 20.0 — Goin and Pascual (1987) Thylacosmilus atrox FMNH P14474 Late Miocene, Corral Quemado Formation 16.0 — (Goin and Pascual, 1987) Thylacosmilus atrox MMP 1470 Late Miocene, Cerro Azul Formation 15.5 13.6 (Goin and Pascual, 1987) 185

Thylacosmilus atrox MACN-A 7916 Early Pliocene, Monte Hermoso — 12.8 Marshall (1976a) Thylacosmilus atrox MACN-A 6639 "Middle" Pliocene, Chapadmalal Formation — 13.5 Marshall (1976a) Thylacosmilus atrox MMP 1443 "Middle" Pliocene, Chapadmalal Formation 15.5 15.4 Goin and Pascual (1987) Thylacosmilus atrox MMP 1484 "Middle" Pliocene, Chapadmalal Formation — 11.2 Goin and Pascual (1987)

186

Appendix 20. Character-state matrix used in the phylogenetic analysis of Eomakhaira molossus. ? = unknown, - = inapplicable character. Polymorphisms: a = 0 &1, b = 1 & 2, c = 2 &3, d = 0&2, e = 1&3, f= 0&1&2. Uncertainties: g = 0/1, h = 1/2, i= 2/3/4.

Holoclemensia texana

???????????????????????????????????????????????????????????

??????????????????????0?1000100?210000110000212110?0??01100

1000210000000000?000000????????????????????????????????????

???????????????????????????????????????????????????????????

??????????????????????????

Deltatheroides cretacicus

?????0?00????????b00002????????????????????????????????????

???????????????????????????????????????????0???????????????

????00001?0??0????0???1000000-0021000000100001002?0????00??

0???00??0????0?0???--?0????????????????????????????????????

???????????????????????????????????????????????????????????

??????????????????????????

Deltatheridium pretrituberculare

?00??0000000000000000a200d0000?00aa?0??????00???00?00?-????

??0??0????????1?????000000000-000?0?01000020001100110201100

00000000000000000001??1000000-0121000000100001000000?000001

000000000001-0000000000????????????????????????????????????

187

???????????000???????????????????????????????????????00????

00-??000?00???????????????

Didelphodon vorax

000000?111?11???0210100101120??????0?0?01101-1?00??110????1

?????00???0???0???00110000000001011?110110210?01111??211?10

00000110111010a1100???00111010003201101000a201001010?102001

22000110010010011002101????????????????????????????????????

??????????????????????????????????????????00????0????????1?

0????0????100?????????????

Eodelphis browni

?????????????????11????10??????????0????0100a?-00???b??????

????????????????????1????????????????a0010211a02?1????11100

??00010?11101001??0???0?111010003201101000a201000110?102001

2200011001001001100a101????????????????????????????????????

???????????????????????????????????????????????0???0???????

??????????????????????????

Kokopellia juddi

???????????????????????????????????????????????????????????

???????????????????????????????????????0?0211?1???????00000

??0?010??00??0?000????0??00010??320100?10001a1000100?001100

22001010000011000001001????????????????????????????????????

188

???????????????????????????????????????????????????????????

??????????????????????????

Pediomyidae

?????????????????b00??01001?????????????????????????b??????

?????0??????????????0100000001?1001?0a001021a?0??a???????00

????000101000020000???0?111a1000420101100013122a2000?00100a

2200201001101001a002101????????????????????????????????????

??????????????????????????????????????????00????0????????0?

0????0????101?????????????

Alphadon spp.

??????????????????1???010??????????????????????????????????

???????????????????????????????????????01021aa11?1????0??00

??00000??1000000000???0?110a10003201001100a1010110000001a01

2200201001a0a001100a001????????????????????????????????????

???????????????????????????????????????????????????????????

??????????????????????????

Asiatherium reshetovi

0?0?0????????10?010???0?0012???1?11?????0000?0-?0?11b?0???1

??????????????20?????????0?10???0??????00021??1201???????00

??0?000000000000000???001000101032010111101112201010?001101

22001010011001010002001??????????????????1??????0????0??000

189

0??00??00000?00?00???00000??0000000?0???0??????????????????

??????0???????????????????

Mayulestes ferox

00000000000101000a00100000120?0101001100000110-011001100000

0110010001000010??0000101010010100001000?021??0011000100100

0000010011000010?0021?001210101031010011000101001010?001101

1200101000100011000200100?0010000??????0000???1000000?11100

0001100000010000???0010011-01000?0000000100010000???0110000

00-0000101101???????0?????

Pucadelphys andinus

000000000?00010100001000?0120??100000020002110-0110011-0000

00000100?1001010?0000a10100001010000110000210a0011010100100

00000a001000001010021?001111110042010011010112011010?001101

2200211000100101000200100?00110000000?100111000?00?10001100

000000000000001?????11012001000010001?0010001001000?0100000

00-0000111101???1???0????1

Andinodelphys cochabambensis

100000000??00111000?10?000120?h011?010?00??01?-0111001-0000

0100?1???1???010???00a101000010101001?000021??0??100??00100

00?000001000?01??00???00111111003201001100011101101??001001

2200211001100001000200100?001?00???0???0?????00?00?1????10?

190

?0?0?10??00??01??????100?1-?1000??0???????0?10010???0?100??

??????01??1????0101?0?????

Herpetotherium fugax

110000001??a?a110b101001001200b1001?01000101?0-01?112000??0

?1000100011?1120??00011000101?0100001?10102101211100??00010

0000000100000010000???001211110031aa001010021101f110000100b

220020100210a1011002101???????????????????????????00??01100

00?0110a0011??1??????10?????000??000??11?010000101??0??1000

?0-0?0001110????1?????????

Peradectidae

100?00000??001a10a101?1??01200?a1a0?1?20?????????11010????0

??0???????????10?????11000020?010?01??0010f1??????????00110

0?00000000000010a00???001h011a0032a?10110001110a1aa0?001102

22002010021010011002101????????????????????????????????????

???????????????????????????????????????????????????????????

??????????????????????????

Didelphis albiventris

1110000a100a0111001010010012a020110010a000211000111120000aa

0a0a010000a0a000100001100001110100001000102101a011000000100

00000001a00001-0a002000012111100321--01011021101a0000001001

220020100110111110020001a11a11110101b0a210001001011000ab100

191

000a1aa011000001110110001011000010000001121000010100a001101

0110001-11a1110100011010a1

Monodelphis spp.

01000001a00a01110110100100121020110010a00020100011112000?aa

0a00010001a000a01000011000011101100011a010b1020a11000000100

000000001a000a11a0020000121111a0211--01011020101a000000100b

220020100210101110020001a100a11001000012100a0001000100a100a

0001010001000011110?10001011a000?000000111100001010?0000001

0110001-??01010100010010??

Dromiciops gliroides

000000000000010112100021011210b111000011002000--11a1200??21

01000001111111211000011000111111100011101021022211010100000

000000001000000000020?0001111011421--1011003022001000001100

22002011111001011002100a01000-1?0110b?0a1010111101010012000

1000001000100010021?020010100000?00001011210000101001020100

0100011-111011000000101001

Dasyurus spp.

0000100001011a11121001a10d1201211a0010100021000011112001?21

11100100111a10100100011000121011010011a010b10201111a0211110

000010?0000?---0-00?0?000211111032000010110201d021a00001012 b2002010-210110110021aa110a01110010aa0a100011011010110a0110

192

11001b1111010110021?11002010000a000001011211100001000010001

01111101a11010101110010000

Sminthopsis crassicaudata

0000100011021101121011110b110020111000110020000011112000?20

0100010a011011200100011000021111011011101011021211100201100

??0000?000000a1010020?000211111032a10010100201d021110001001

22002010021001010002101100000110011010211001000101200002001

10000011000010000???100020101020?00100011211110101000000011

0111111-110000101?1?010000

Thylacinus cynocephalus

121001011a021111031011a1021101100a0110110001000001111001?01

1a10010010a01000010001000000110111?0111010e1012110111211110

0000000000010000000?0?1002010-10201—110110411202110001—

12-11200111201-00121--11110110101011012001011?0001010010020

002110021111010111?211020121-0a0100102110?11111101010012200

0100-1110001001000111-0-0100

Patene simpsoni

???????????????????????????????????????????????????????????

???????????????????????????????????????010a10????0?????1?10

????010??0000010?10???001200100021010110100111010110?0?1012

1100101000001111100110a????????????????????????????????????

193

???????????????????????????????????????????????????????????

??????????????????????????

Sallacyon hoffstetteri

??101????????0101?0??1?0??111???001????00?01?0?10?0?110?1?1

1??0??0???????0???010??00?10???????0??????c????????????????

??????0????????????????00200001?200?1?10110300100100???-12-

1?20011?1101011100-101?????????????????????????????????????

???????????????????????????????????????????????????????????

?????????????????????????

Notogale mitis

???1??0110?h??????11??h?1??????????????0?????011??1111001?1

100?1101??????0??0010??00010?????0?01?????3???????11?211110

0?000001010??0?0?????????20000??201--1?01103?01001?000?-12-

1120011?100?01?1??-101?????????????????????????????????????

???????????????????????????????????????????????????????????

?????????????????????????

Acyon myctoderos

111111111??11000011111b01?11???0001?????000100110?011?????1

1??11??0??01?000?0?10??0001?1????????1101031011110111221110

00000001000100000102011002000020200a1110110311b00110001--12

-1120011110a1011100-10010???1a???????1?????????????????????

194

???????????????????????????????????????????????????????????

??????????????????????????

Cladosictis patagonica

111111a111011a100a0111b011110010001010200001001101a11100111

11011?0a00a10000?0010??0?01???0????0?11a10310aa110111221110

0000000100010000010??11002000-2120011110110311b00110001--12

-1120011110a10111a0-10a101?0111010?12??101a00?0111?11101100 a??102a?101010000??1111121-1001??11?111110?100?1111??201000

1100000000001?10??0?0??000

Sipalocyon spp.

01011101100110000a0111e0111100100010a0100001001101111100111

10011?0?0?110000?0010110001010?10?0011101021000010111211110

000000010a010000010???1002000-b020011100110210100110?01--12

-112001111001011100-1001011010??1????0??011????11??0???1?00

10???2???0???00?0??0?1???1-???11?1?????11011001011110200110

11000000101010?0??0001?000

UF 27881

?0???1????00?0000a0111301a1?00?1001????????????????????????

??????????????h??????????????????????????????????????????10

????00001?0???????0???000?0?????1?1--?????0i???????????????

???????????????????????????????????????????????????????????

195

???????????????????????????????????????????????????????????

??????????????????????????

Hondadelphys fieldsi

01?1??????????1?1111?01?1?1????1001??0?0000100110?011?01??1

1?0101000?0?0000??01011001?01001????111010010??????????1110

??0?000100010??0??02??00121110003101111010031120011101?0011

22?00011?11011011012100101??11??1????0?????????1???????????

??????????????00???????????1?11????????????????????????????

??????????????????????????

Prothylacynus patagonicus

001001011?111000110111b011a10011001011211001001001101000001

0101011010110000??110??111101?0?0??0111101e1001110111221110

00100100110100101102111002000-11101--210110311000111?01--12

-11200a11101---0201-10010110011010???1?101000?01?1?11111000

0a01100000110000???1112121-11111011011111011000011110??????

???????????01010??????1000

Stylocynus paranensis

?????????????????g1????????????????????????????????????????

???????????????????????????????????????1?03????????????1110

????000??101?010110???0??21010??31???1?01103b?2101100010012

21000011110010111002100????????????????????????????????????

196

???????????????????????????????????????????????????????????

??????????????????????????

Borhyaena tuberata

0010011111101000110111b011010011001011211011011001001a01000

0110011100110000??110111011011010?0111111a31011110221221110

0010010011110001010??11002000-21001--1101104201001111010012

1-1a001a12-1---02a1---100110001010?120????1???0111?011?a?10

1011021?11?11100?1?1111121-1112????0??111??????????1?22???0

110000100000??????????0?11

Arctodictis munizi

001001111?10100012010110111???010010?1211011011001001?11?00

0110011110110000??110?????????0????1?1121131011110221221110

??1a021011111002110???1002000-21001--1?0110420100112?010012

1-11001112-1---0201---1????????????????????????????????0???

?01???????????????????????????????????1?1??????????????????

??????????????????????????

Arctodictis sinclairi

00100111111110001201011012110001001011211011011001001011?00

01100111101?0000??110111011011010?0111121131011110221221110

001a021011110001110???1012000-21001--?101104201001121010012

1-11001112-1---0201---10011000??10012021111?000111101101010

197

101112111001110001?1112121-????101?010111011001011110221000

11000000000010000000010011

Pharsophorus lacerans

?0???11????1101011????201??????00010?10110?10???0??????????

??????????????0?????011101101??10?0?11?100310?????1??221110

??1a0100110?1010?00???1??20?0???h?????????????????110010012

111000111201---020a21a1????????????????????????????????????

?????????????????????????????????????????011?000111?0??????

??????????????????????????

Thylacosmilus atrox

0110110110011-1111110-b01111???1001110110011001101001?01?01

0000011110a10001??110???01?01???0001111a0031010b103?-2-2-11

100a1-0---1?-??b1?1???1112010-210-1--0101204212001111?1--12

-1120012?121-?0021----?101101110110120?a011?0?????????00010

10111111110???00???1?????1?0001?011201111011000011111001110

1110000011001?0???????01??

Paraborhyaena boliviana

?110?111??0??000??11112????????????????0????01110??011?10??

0??0??1??01?0001??1101?1??1????????0?1111131??????1??212110

11110110111??0?2??0???100????-????????????????????1?1??????

???20??2???????????????????????????????????????????????????

198

???????????????????????????????????????????????????????????

??????????????????????????

Callistoe vincei

111?111111121110101?11201111?0110aa011111011101101001111?00

0101111110?10000??1101??00101101?000111211310?1210110201110

11110100110?0022?10???1002000-2100???1101?040110?110??1--??

??1200?1???????????-??10011?001?11??2??1?11???01?1201111?00

0?1102111?00?10001?0??????????1101??1?11101?101?11110???010

???????????1?00?0?0?011?0?

Patagosmilus goini

0g????0??????010111?0-30110101?0001??111???????????????????

??????????????0??????????????????????1???????????????????11

1?0?1-0--?0?-?????1???1102000-21001--0101203201001?????????

???????????????????????????????????????????????????????????

??????????????????????????

Australohyaena antiquua

001??1111111100011010120111100?0001?11??111101?001001011??0

010?0?11?00??000??110??10??0??01?????1121131012110221221110

0010010011111022110???1002000-10001--1101101201001120010012

1-1000111221---0201---1????????????????????????????????????

199

???????????????????????????????????????????????????????????

??????????????????????????

Lycopsis longirostrus

111111????????10?01????01?1??0??001??1??000?00110?101?10001

01010111??01?000??1101?101101?0?0??1111010310a1110?????1110

001000010001000010021?1002000-2020011110110311b00110?01--12

-112001111001010200-11010??????0???????1??????0?1????1?1000

1011021?11000100?1001?1121-0001???0?101?10????????????1????

??????1?10????1???????1?01

Lycopsis padillai

????????????????-a1???2??????????00????????????????????????

???????????????????????????????????????????????????????????

??????????????????????1012000-10200111101203101001?????????

???????????????????????????????????????????????????????????

??????????????????????????

Lycopsis torresi

??????????????????0???2????????????????????????????????????

????????????????????????????????????????10?10a1??????????10

00100?0???01?0ga100???1?12000-1020011110120310100110??1--12

-112001112001010200-1?0????????????????????????????????????

200

???????????????????????????????????????????????????????????

??????????????????????????

Lycopsis viverensis

?????????????????01?11?0101????????????????????????????????

???????????????????????????????????????0?031?????????????10

??0?010?00010???1?0???1012000-h?200111101203101001???01--12

-112001112001010200-100????????????????????????????????????

???????????????????????????????????????????????????????????

??????????????????????????

Proborhyaena gigantea

???????????????????????????????????????????????????????????

???????????????????????????????????????101310?1??0???????1?

?111011??111?022?1????1??2000???00???0?012041000g11?1?1??12

??120012???????????????????????????????????????????????????

???????????????????????????????????????????????????????????

??????????????????????????

Eomakhaira molossus

?g???????????0???h1?11h0?101???????????????????????????????

???????????????????????????????????????2?031?????????????1?

?10a010011???0?1?10????01?????2000??????1?0?g???????1?????2

???200???2????????????g????????????????????????????????????

201

???????????????????????????????????????????????????????????

??????????????????????????

202

Appendix 21. List and description of 321 characters used in the phylogenetic analysis of SGOPV 3490, modified from Suarez et al. (2016). * = Ordered character. 1. Length of the skull 0 Short (Less than twice width at level of zygomatic arch) 1 Long (Greater than twice width at level of zygomatic arch) 2. Length of rostrum* 0 Less than 1/3 total length of skull 1 Between 1/3 and 1/2 total length of skull 2 More than 1/2 total length of skull 3. Width of braincase versus maximum postorbital width 0 Braincase wider than maximum postorbital width 1 Braincase narrower than maximum postorbital width 4. Dimensions of braincase 0 As wide as long, or slightly wider than long 1 Much wider than long 5. Level of the palate relative to the basicranium 0 Palate lower than basicranium 1 Palate and basicranium at the same level 6. Paracanine fossa 0 Formed by both maxilla and premaxilla 1 Formed solely by premaxilla 7. Precanine notch 0 Absent 1 Present 8. Lateral palatal process of premaxilla 0 Anterior to or just reaches anterior border of canine alveolus 1 Posterior to anterior border of canine alveolus 9. Posterior border of incisive foramen 0 Anterior to or just reaches anterior border of canine alveolus 1 Posterior to anterior border of canine alveolus 10. Position of medial palatal process of premaxilla 0 Horizontal 1 Inclined dorsally, forming an incisive fossa 11. Dorsal process of premaxilla in narial platform 0 Absent 1 Present 12. Posteriormost point of premaxilla-nasal contact* 0 Anterior or at the level of the canine 1 Posterior to the canine 2 Posterior to p2

203

13. Anterior extent of nasals 0 Protrude anteriorly, obscuring the nasal opening in dorsal view 1 Retracted posteriorly, exposing the narial opening in dorsal view 14. Shape of naso-frontal suture 0 Open W-shape or posteriorly convex 1 Acute W or V-shaped 15. Postorbital processes 0 Absent or indistinct 1 Well-developed 16. Fronto-maxillary or naso-lacrimal contact 0 Naso-lacrimal contact 1 Fronto-maxillary contact 17. Angle of maxillo-jugal contact 0 More than 140 degrees 1 Between 95 and 140 degrees 18. Location of the infraorbital foramen* 0 Anterior or dorsal to the anterior root of P3 1 Dorsal to the posterior root of P3 2 Dorsal to M1 3 Posterior to M1 19. Flaring of maxillary "cheeks" behind infraorbital foramen 0 Present 1 Absent 20. Large foramen at anteroventral end of maxilla medial to canines 0 Absent 1 Present 21. Palatal length/width ratio 0 Lesser than or equal to 1.5 1 Greater than 1.5 22. Shape of the palate 0 Rectangular (molar rows near parallel) 1 Triangular (wider posteriorly) 23. Number of palatal pits* 0 None 1 One (between M3-M4) 2 Two (between M2-M3 and M3-M4) 3 Three (one between each pair of molars) 24. Maxillopalatine fenestrae 0 Absent 1 Present

204

25. Major palatine foramen 0 One pair opening in maxilla, palatine, or maxillo-palatine suture 1 Many small foramina on the surface of the maxilla 26. Minor palatine foramen* 0 Large 1 Small 2 Incomplete or absent 27. Posterior extent of palatines 0 Extend to the level of the last molar 1 Extend beyond the level of the last molar 28. Posterior end of palatines 0 Concave posteriorly (single-arched) 1 Concave posteriorly (double-arched) 2 Straight due to presence of a palatine torus 29. Palatine reaches level of infraorbital canal 0 Present 1 Absent 30. Position of sphenorbital foramen 0 Posterior to the level of the posterior border of lacrimal 1 Anterior or at the level of the posterior border of lacrimal 31. Development of pterygoids* 0 Well-developed and expanded on medial side, with midline contact 1 Well developed and expanded on medial side, but no midline contact 2 Reduced, not expanded on medial side 32. Anterior extent of lacrimal 0 Restricted to orbit 1 Extending onto rostrum 33. Lacrimal tubercle 0 Present 1 Absent 34. Position of lacrimal foramina 0 Within orbit 1 Exposed on face 35. Number of lacrimal foramina 0 Two 1 One 36. Glenoid process of jugal 0 With articular facet 1 Without articular facet 37. Orbital crest

205

0 Absent 1 Present 38. Interparietal 0 Present 1 Absent (or fused with parietal) 39. Shape of fronto-parietal suture 0 Formed by posterior wedge of frontals 1 Straight 2 Formed by anterior wedge of parietals 40. Parietal-alisphenoid or fronto-squamosal contact 0 Parietal-alisphenoid 1 Fronto-squamosal 41. Width of glenoid cavity 0 Less than twice anteroposterior length 1 More than twice anteroposterior length 42. Distinct preglenoid process of squamosal 0 Absent 1 Present 43. Morphology of postglenoid process* 0 Wide and low 1 Wider than high (as high as half the width) 2 As wide as high 44. Location of postglenoid foramen 0 Behind postglenoid process 1 Medial to postglenoid process 45. Suprameatal foramen 0 Above suprameatal crest 1 Below suprameatal crest 46. External acoustic meatus 0 Longer than wide 1 Wider than long 47. Paracondylar process of exoccipital and post-tympanic process of squamosal 0 Paracondylar process larger 1 Both processes similar in length 48. Orientation of the post-tympanic and/or paracondylar processes 0 Ventrally projecting 1 Anteroventrally projecting 49. Alisphenoid glenoid process 0 Absent 1 Present

206

50. Optic foramen and sphenorbital fissure 0 Separate 1 Joined 51. Transverse foramen 0 Absent 1 Present 52. Tympanic process of alisphenoid 0 Absent 1 Present 53. Hypotympanic sinus 0 Absent 1 Formed by squamosal, petrosal, and alisphenoid 2 Formed by alisphenoid and petrosal 54. Medial process of the squamosal 0 Absent 1 Present 55. Concave process of alisphenoid contributing to antero-dorsal portion of hypotympanic sinus 0 Present 1 Absent 56. Extra sinuses posterior to the hypotympanic sinus 0 Absent 1 Present 57. Pneumatization of squamosal 0 Absent 1 Present 58. Eustachian foramen 0 No impression 1 Notch on the alisphenoid 2 Foramen on petrosal 59. Composition of foramen ovale 0 Between petrosal and alisphenoid 1 On alisphenoid 60. Secondary foramen ovale 0 Absent 1 Present 61. Foramen for the greater petrosal nerve 0 Distinct notch or foramen 1 Without distinct notch or foramen 62. Position of carotid foramen

207

0 Anterior to the basisphenoid-basoccipital suture 1 At the level of the basisphenoid-basoccipital suture 63. Hypoglossal foramina 0 Two or more 1 One 64. Groove between hypoglossal foramina and foramen for inferior petrosal sinus 0 Shallow or absent 1 Well-defined with prominent lateral borders 65. Size of jugular foramen relative to fenestra vestibuli 0 Subequal 1 Larger 66. Jugular fossa 0 Absent 1 Present 67. Median keel in basioccipital 0 Absent 1 Present 68. Median crest of basisphenoid/presphenoid (sphenoid crest) 0 Present 1 Absent 69. Dorsal margin of the foramen magnum 0 Formed only by exoccipitals 1 Formed by both exoccipitals and supraoccipital 70. Mastoid foramen or other emissary foramina in the occiput 0 Present 1 Absent 71. Connection between condylar articular facets in ventral view 0 Absent 1 Present 72. Inclination of the major axis of the condyle in posterior view 0 Inclined (less than 55 degrees) 1 Vertical to subvertical (between 90 and 55 degrees) 73. Supraoccipital in posterior view 0 Concave 1 Convex or flat 74. Sagittal crest* 0 Prominently developed (extending to frontals) 1 Weakly developed (not extending to frontals) 2 Absent 75. Position of nuchal crest

208

0 At or posterior to the level of the condyles 1 Anterior to the condyles 76. Morphology of the stapes 0 Columelliform (not perforated by stapedial foramen) 1 Bicrurate (perforated by stapedial foramen) 77. Ectotympanic shape 0 Ring-shaped 1 Expanded 78. Position of petrosal 0 At the level of the ventral margin of the braincase 1 Dorsal to the ventral level of the braincase 79. Mastoid portion of the petrosal 0 Contributes to the occipital shield 1 Excluded from the occipital shield 80. Petrosal-squamosal fusion 0 Absent 1 Present 81. Cavum epiptericum 0 Floored by petrosal and alisphenoid 1 Floored primarily or exclusively by alisphenoid 82. Internal acoustic meatus 0 Deep with thick prefacial commissure 1 Shallow with thin prefacial commissure 83. Subarcuate fossa 0 Deep 1 Shallow 84. Deep sulcus for carotid artery on anterior end of promontorium 0 Absent 1 Present 85. Epitympanic wing of petrosal 0 Present 1 Absent 86. Prootic canal 0 Present 1 Absent 87. Rostral tympanic process of petrosal* 0 Absent or low ridge 1 Tall ridge restricted to the anterior half of the promontorium 2 Tall ridge reaching anterior half of promontorium 88. Paroccipital process of petrosal

209

0 Distinct process 1 Indistinct or absent 89. Position of hiatus fallopii 0 At a distance from the anterior edge of the petrosal 1 On the anterior edge of the petrosal 90. Stylomastoid foramen 0 Absent 1 Present 91. Floor of cavum supracochleare 0 Absent 1 Present 92. Stapedial ratio 0 Rounded, less than 1.8 1 Elliptical, more than 1.8 93. Contribution of squamosal to epitympanic recess 0 Small 1 Extensive 94. Fossa incudis 0 Continuous with epitympanic recess 1 Separated from the epitympanic recess 95. 'Petrosal crest' (sensu Muizon, 1999) 0 Present 1 Absent 96. Stapedial fossa 0 Twice the size of fenestra vestibuli 1 Small and shallow 97. Foramina for temporal rami 0 On parietal or squamosal 1 Absent 98. Post-temporal canal or notch 0 Present 1 Absent 99. Shape of dentary (depth below m3/m4 embrasure/length m1-4)* 0 Shallow (less than 0.6) 1 Intermediate (between 0.6 and 0.8) 2 Deep (greater than 0.8) 100. Ventral margin of jaw behind m4 0 Straight 1 Curved 101. Mandibular symphysis

210

0 Unfused 1 Fused 102. Posteriormost mental foramen* 0 Below p3 1 At p3/m1 embrasure 2 Below m1 3 Posterior to m1 103. Retromolar space 0 Absent 1 Present 104. Labial mandibular foramen inside masseteric fossa 0 Absent 1 Present 105. Shape of the angular process* 0 Shelf-like (ASL/AL > 0.81) 1 Intermediate (0.72 < ASL/AL < 0.81) 2 Rod-like (ASL/AL < 0.72) 106. Angle between anterior border of coronoid process and tooth row* 0 Between 95 and 105 degrees 1 Between 106 and 125 degrees 2 Greater than 126 degrees 107. Position of the mandibular foramen* 0 Posterior to the mid-point of the coronoid process 1 At the mid-point of the coronoid process 2 Anterior to the mid-point of the coronoid process 108. Morphology of mandibular condyle 0 Subspherical 1 Cylindrical 109. Position of mandibular condyle relative to tooth row 0 Below or at level of tooth row 1 Above level of tooth row 110. Number of upper incisors* 0 Five 1 Four 2 Three 3 Two or fewer 111. Shape of first upper incisor (serially homologous I1)* 0 Enlarged 1 Subequal to or smaller than remaining incisors 2 Absent

211

112. Size of I4 versus I3 0 I4 subequal to I3 1 I4 larger 113. Size of I5 versus I4* 0 I5 subequal to I4 1 I5 smaller than I4 2 I5 absent 114. Shape of upper incisor arcade* 0 Parabolic 1 Slightly anteriorly convex 2 Transverse 115. Number of lower incisors* 0 Four 1 Three 2 Two or less 116. Staggered lower incisor (serially homologous i3) 0 Absent 1 Present 117. Size of canines 0 Relatively small 1 Enlarged 118. Shape of upper canines 0 Not saber-like 1 Saber-like 119. Roots of upper canines 0 Closed in adults 1 Open 120. Roots of lower canines 0 Closed in adults 1 Open 121. Surface of the roots of the canines 0 Smooth 1 With small grooves and ridges 122. Prominent median sulci on labial and lingual faces of canines 0 Absent 1 Present 123. Number of premolars 0 Three 1 Two or less 124. Orientation of P/p1 relative to tooth row*

212

0 Parallel to tooth row (less than 19 degrees) 1 Obliquely oriented to tooth row (20 degrees or more) 2 Transversely oriented to tooth row 125. Orientation of P/p2 relative to tooth row 0 Parallel to tooth row 1 Oblique 126. Diastema anterior to P1 0 Absent 1 Present 127. Diastema posterior to P1 0 Present 1 Absent 128. Diastema posterior to p1 0 Present 1 Absent 129. Shape of premolars 0 Uninflated 1 Inflated, with apical wear strongly developed 130. Protoconid of p1 0 Anteroposteriorly aligned with or anterior to anterior root 1 Posterior to anterior root 131. Cusp on the posterior heel of P3 0 Absent or vestigial 1 Well-developed 132. Size of p2 0 Smaller than p3 1 Larger than p3 133. Change in height of lower premolars 0 Increase gradually in height 1 Abrupt change in size between p1 and p2-3 2 Abrupt change in size between p1-2 and p3 134. Roots of lower premolars* 0 Flat (as wide as crown) 1 Bulbous on only one premolar 2 Bulbous on all premolars and some molars 135. Precingulid or cingulid cusp on p2 0 Absent 1 Present 136. Symmetry of main cusp on p3 0 Anterior edge of cusp more convex than posterior edge

213

1 Both edges similar in curvature 137. Replacement of dP3 0 dP3 is replaced 1 dP3 is not replaced 138. Timing of eruption between dP/p3 and M/m3-4 0 p3 erupts before m3 1 p3 and m3 erupt almost simultaneously 2 p3 erupts almost simultaneously with or after m4 139. Timing of eruption between M3-4 and m4 0 M3 and m4 erupt simultaneously 1 M/m4 erupt simultaneously 140. Morphology of dp3 0 With trigonid and talonid 1 With a main cusp and smaller accessory cusps 141. Size of molars increasing posteriorly 0 Moderate posterior increase in size 1 Marked posterior increase in size 142. Shape of upper molar row 0 Straight 1 Bowed 143. Width of M4 relative to M3 0 Narrower than M3 1 Subequal to wider than M3 144. Size of metacone relative to paracone (based on M2 when possible)* 0 Slightly smaller 1 Subequal to slightly larger 2 Larger 145. Position of the metacone relative to paracone (based on M2 when possible) 0 Approximately at the same level 1 Lingual 146. Shape of paracone and metacone 0 Conical 1 Subtriangular with a flat labial face 147. Bases of paracone and metacone 0 Adjoined 1 Separate 148. Centrocrista 0 Straight 1 V-shaped 149. Metacone on M4*

214

0 Present and distinct 1 Extremely reduced 2 Absent 150. Number of roots on M4 0 Three 1 Two or less 151. Size of protocone* 0 Vestigial or absent 1 Small and without basin 2 Small with basin 3 Somewhat expanded anteroposteriorly 4 Greatly expanded anteroposteriorly 152. Height of protocone* 0 Less than 60% of para/metacone height 1 Between 60 to 80% para/metacone height 2 Greater than or equal to 80% of para/metacone height 153. Paraconule and metaconule 0 Present 1 Absent 154. Wing-like cristae associated with para- and metaconules 0 Absent 1 Present 155. Relative position of para- and metaconule (based on M2 when possible) 0 At or lingual to the midpoint between protocone and para/metacone 1 Closer to the paracone or metacone 156. Orientation of the preparacrista (based on M2 when possible) 0 Nearly perpendicular to long axis of tooth 1 Oriented anterobucally to long axis of tooth 2 Absent 157. Lengths of preparacrista on M3 and M4 0 M4 preparacrista shorter 1 M4 preparacrista subequal or longer than M3 preparacrista 158. Postmetacrista (based on M3 if possible) 0 Strongly developed (longer than preparacrista) 1 Weakly developed (shorter than preparacrista) 159. Orientation of postmetacrista (based on M3 if possible) 0 Nearly perpendicular to tooth row 1 Oblique to tooth row 160. Preparacingulum (based on M3 if possible)* 0 Expanded

215

1 Short 2 Vestigial to Absent 161. Postcingulum 0 Absent or weakly developed 1 Present 162. Stylar shelf (based on M3 if possible)* 0 Uniform in width, 50% or more of total transverse width 1 Uniform in width, but less than 50% of total transverse width 2 Slightly reduced labial to paracone 3 Strongly reduced labial to paracone 4 Vestigial to absent 163. Deep ectoflexus on upper molars* 0 On M2 and M3 1 On M3 only 2 Strongly reduced or absent 164. Stylar cusp A* 0 Absent 1 Smaller than StB 2 Large, subequal to StB 165. Stylar cusp B* 0 Large 1 Small or forming an ectocingulum 2 Vestigial or absent 166. Stylar cusp C 0 Absent 1 Present 167. Stylar cusp D* 0 Absent 1 Present, smaller than stylar cusp B 2 Present, larger than stylar cusp B 168. Stylar cusp E 0 Present and distinct 1 Indistinct or absent 169. Size of m4 0 m4 subequal or smaller tham m3 1 m4 larger than m3 170. Posterior lobe of the crown lower than anterior lobe* 0 Absent 1 Present only on m1-2 and slightly developed 2 Present on m1-3 and strongly developed

216

171. Roots of lower molars (based on m3-4 when possible) 0 Both roots similar in size 1 Anterior root much larger than posterior root 172. Talonid of m4 relative to m3 0 Talonid of m4 reduced and narrower than m3 1 Talonid of m4 similar to m3 173. Alignment of the main cusps of m1 0 Reverse triangle acute 1 Single longitudinal row 174. Trigonid configuration posterior to m1* 0 Open, with paraconid anterolingual 1 Acute, with paraconid more posteriorly placed 2 Anteroposteriorly compressed 175. Metaconid position 0 Aligned with paraconid 1 Metaconid at extreme lingual margin of tooth 176. Orientation of postprotocristid/metacristid 0 Transverse to lower jaw 1 Parallel or oblique to lower jaw 177. Trigonid versus talonid length (m1-m3)* 0 Trigonid smaller than talonid 1 Trigonid subequal to talonid 2 Trigonid larger than talonid 178. Trigonid versus talonid width (m1-m3)* 0 Very narrow (subequal to the base of the metaconid or protoconid) 1 Narrow (but wider than the base of the metaconid or protoconid) 2 Subequal to wider than the trigonid 179. hypoconid versus protoconid height (based on m2-3)* 0 hypoconid/protoconid height radio less than 20% 1 hypoconid/protoconid height radio between 25-35% 2 hypoconid/protoconid height radio between 40-60% 180. Metaconid on m1 0 Present 1 Absent 181. Metaconid on m2-4* 0 Present on m2-4 1 Absent on m4 2 Absent on m2-4 182. Paraconid height relative to metaconid (m2-4)* 0 Taller

217

1 Subequal 2 Lower 183. Height of protoconid 0 Tallest cusp of the trigonid 1 Subequal to metconid or paraconid 184. Mesiolingual vertical crest of the paraconid 0 Rounded 1 Forming a keel 185. Precingulid* 0 Well-developed, extending from the protoconid to paraconid basins 1 Reduced, extended only on the base of the paraconid 2 Absent 186. Paraconid of m1 0 Distinct 1 Low and confluent with cingulum 187. Length versus width of talonid basin (based on m2 when possible)* 0 Longer than wide 1 Subequal 2 Wider than long 188. Location and presence of hypoconid 0 Appoximately at the middle of the buccal margin of the talonid 1 At the posterobuccal corner of the tooth 2 Absent 189. Presence of the entoconid 0 Present 1 Vestigial or absent 190. Shape of the entoconid 0 Conical 1 Labio-lingually compressed 191. Height of entoconid 0 Smaller than the hypoconid 1 Subequal to larger than the hypoconid 192. Location of entoconid 0 At the posterolingual corner of the tooth 1 Between the metaconid and posterior tooth margin 193. Position of hypoconulid 0 In posteromedial position 1 Lingually placed and twinned with entoconid 194. Hypoconulid of m4* 0 Tall

218

1 Short 2 Absent 195. Pre-entocristid 0 Present 1 Absent 196. Direction of the pre-entocristid 0 To the base of the trigonid 1 Lingual to the trigonid 197. Cristid obliqua* 0 Lingual to the carnassial notch 1 To the carnassial notch 2 Labial to the carnassial notch 198. Lower molar hypoflexid 0 Deep (40-50% of talonid width) 1 Shallow or absent 199. Carnassial notch in cristid obliqua 0 Absent 1 Present 200. Labial postcingulid 0 Absent 1 Present 201. Atlas intervertebral foramen 0 Absent 1 Present 202. Atlas transverse foramen 0 Absent 1 Present 203. Ventral foramen on transverse process of axis 0 Absent 1 Present 204. Posterior extent of atlantal transverse processes 0 Anterior or just reaches caudal facets for axis 1 Extend caudally beyond level of caudal facets for axis 205. Anterior extent of atlantal transverse processes 0 Does not reach level of atlantal foramen or groove 1 Extends anterior beyond atlantal foramen or groove 206. Shape of cranial facets 0 Only concave 1 Dorsal edge curved 207. Atlas and intercentrum

219

0 Unfused 1 Fused 208. Axis transverse foramen 0 Absent, represented by a notch 1 Present, enclosed 209. Axis posterior spinous process extension 0 Extends beyond the level of the postzygapophyses 1 Extends to the level of the postzygapophyses 210. C3-C4 ventral sagittal process 0 Absent 1 Present 211. C5 transverse process heads overlap transversally 0 present 1 absent 212. C5 and T1 body length 0 C5 subequal or longer than T1 1 C5 shorter than T1 213. C6 spinous process 0 Protuberance 1 Lamina 214. C7 transverse foramen 0 Absent 1 Represented by a notch 2 Complete foramen 215. Shape of anterior face of C7 centrum 0 Circular to ovoid 1 Rectangular to trapezoidal 216. Position of tallest spinous process of thoracic vertebrae 0 On T1 1 On T2 2 On T3 217. Anticlinal vertebra 0 On lumbar 1 On thoracic 2 No anticlinal vertebra 218. Foramen on dorsal arch of last lumbar vertebra 0 Present 1 Absent 219. Metapophyses in third lumbar vertebra anterior to last 0 Absent or weak

220

1 Present 220. Ventral median keel on lumbar vertebra 0 Absent 1 Present 221. Auricular process of sacrum 0 Developed on two sacral vertebrae 1 Developed on one sacral vertebra 222. Size of sacral spinous process 0 Shorter than last lumbar 1 Taller than last lumbar 223. Length of the tail 0 Shorter than twice the length of the precaudal vertebral column 1 Greater than twice the length of the precaudal vertebral column 224. Angle between scapular spine and dorsal border of scapula 0 Acute or almost straight (between 80 and 95 degrees) 1 Obtuse (between 100 and 110 degrees) 225. Coracoid process 0 Large (extends beyond medial border of glenoid cavity) 1 Small (just reaches medial border of glenoid cavity) 226. Ventral extension of acromion process 0 Extends ventrally below glenoid cavity 1 Does not extend ventrally below glenoid cavity 227. Width of infraspinous fossa 0 Less than 1/4 its length 1 More than 1/4 its length 228. Width of the acromion process at the level of the neck* 0 Wider than infraspinous fossa 1 Subequal 2 Narrower than infraspinous fossa 229. Infraspinous/supraspinous fossa width at the level of the neck 0 Supraspinous fossa subequal or wider 1 Supraspinous fossa narrower 230. Scapular notch 0 More than 130 degrees 1 Between 90 and 130 degrees 231. Clavicle 0 Present 1 Absent 232. Medial process for teres major 0 Absent

221

1 Present 233. Tricipital line of humerus* 0 Absent 1 Ridge or crest 2 Massive crest continuous with deltopectoral crest 234. Capitulum for radius on humerus 0 Spherical 1 Cylindrical 235. Entepicondylar foramen 0 Present 1 Absent 236. Olecranon fossa or foramen 0 Large fossa 1 Foramen 237. Laminar supinator crest/ectepicondylar crest* 0 Large 1 Intermediate 2 Absent 238. Greater tuberosity height relative to humeral head height 0 Greater tuberosity subequal or lower in height to humeral head 1 Greater tuberosity is higher 239. Development of greater tuberosity in proximal view 0 Small (less than half the anteroposterior length of head) 1 Large (greater than or equal to half the anteroposterior length of head) 240. Extension of the deltoid crest 0 Restricted to proximal half of humerus 1 Reaches distal half of humerus 241. End of deltoid crest 0 Merging with diaphysis 1 Forming a distinct angle or process 242. Relative heights of trochlea and capitulum in lateral view* 0 Longer proximal extension of capitulum 1 Subequal 2 Longer proximal extension of trochlea 243. Humerus medial epicondyle size 0 Large 1 Small 244. Humerus distal end size 0 Large 1 Small

222

245. Lateral extension of capitulum 0 Rounded 1 Straight 246. Depth of intercondylar notch in posterior view 0 Wide and relatively shallow concave 1 Narrower and concave posteriorly 247. Curvature of the posterior border of the humeral shaft 0 Curved 1 Straight 248. Medial development of the ulnar anconeal process 0 Does not protrude beyond medial border of olecranon process 1 Medially protruding 249. Medial curvature of the ulna 0 Present 1 Absent 250. Posterior border of the ulna 0 Anteriorly curved 1 Straight or posteriorly curved 251. Shape of articular facet for humerus 0 Anteroposteriorly compressed 1 Circular 252. Distal shaft of radius 0 Oval (wider than long) 1 Rounded (almost as wide as long) 253. Prepollex 0 Absent 1 Present 254. Distolateral process of scaphoid* 0 Absent 1 Present, does not separate lunate from magnum 2 Present, separates lunate from magnum 255. Number of plantar tubercles (distal heads) on trapezium 0 Two 1 One 256. Angle between transverse axis of proximal and distal epiphyses of metacarpal I 0 Absent 1 Present 257. Orientation of ilium relative to ischium 0 Prominent dorsally 1 Aligned with ischium

223

258. Tuberosity for rectus femoris muscle 0 Absent 1 Protuberance 2 Depression 259. Length of iliac neck* 0 Longer than 15% total pelvis length 1 Between 6 and 15% total pelvis length 2 Less than 6% total pelvis length 260. Greater sciatic notch 0 Greater than 120 degrees 1 Between 90 and 115 degrees 261. Iliac and gluteous fossa 0 No fossa 1 Two fossa subequal in size 2 Gluteous fossa larger 262. Epipubic bones 0 Present 1 Absent 263. Proximal size of epipubic bones 0 Short 1 Long 264. Torsion between proximal and distal epiphyses of femur 0 Present 1 Absent 265. Relative heights of greater trochanter and femoral head 0 Greater trochanter lower or equal in height to femoral head 1 Greater trochanter higher than femoral head 266. Lesser trochanter of femur 0 Present 1 Vestigial or absent 267. Femoral condyles* 0 Lateral condyle wider than medial condyle 1 Subequal 2 Medial condyle wider than lateral condyle 268. Ossified patella 0 Absent 1 Present 269. Parafibula 0 Present 1 Absent

224

270. Femoro-fibular articulation 0 Present 1 Absent 271. Tibia length relative to femur length 0 Tibia subequal to or longer than femur 1 Tibia shorter than femur 272. Proximal dimensions of tibia* 0 Larger mediolaterally than anteroposteriorly 1 Subequal 2 Larger anteroposteriorly than mediolaterally 273. Tibia shape 0 Sigmoid 1 Straight 274. Torsion between proximal and distal epiphyses of tibia 0 Present 1 Absent 275. Type of distal articulation of tibia 0 Spiral 1 Sagittal 276. Posterior shelf of tibia 0 Present but does not extend posteriorly beyond the medial astragalotibial facet 1 Present and extends posteriorly beyond the medial astragalotibial facet 277. Distal malleolus of tibia 0 Indistinct or absent 1 Distinct 278. Angle between medial and lateral astragalotibial facets* 0 90 degrees 1 Intermediate 2 180 degrees 279. Astragalonavicular facet extends onto ventromedial side of head 0 Absent 1 Present 280. Width and height of navicular facet in distal view 0 Transversely wider 1 Dorsoventrally wider 281. Visibility of medial plantar tuberosity in dorsal view 0 Not visible 1 Visible 282. Angle between lateral tibial and fibulal facets 0 No angle

225

1 With angle 283. Medial extent of sustentacular facet 0 Does not reach the medial edge of neck 1 Reaches the medial edge of neck 284. Astragalar canal 0 Present 1 Absent 285. Width of astragalar neck 0 Neck wider than head 1 Neck narrower or as wide as head 286. Major orientation of posterior astragalocalcaneal facet 0 Anteromedial-posterolateral 1 Posteromedial-anterolateral 287. Malleolar shelf of astragalus 0 Absent 1 Present 288. Astragalo-distal tuber 0 Absent 1 Present 289. Connection between astragalonavicular facet and sustentacular facet 0 Present 1 Absent 290. Longest dimension of sustentacular facet 0 Anteromedial-posterolateral 1 Sagittally longer 2 Transversely longer 291. Orientation of the calcaneoastragalar facet* 0 Medial 1 Intermediate 2 Dorsal 292. Calcaneal peroneal tubercle 0 Protuberance 1 Crest-like 293. Position of peroneal tubercle 0 Anterior, non-protruding 1 At a distance from the anterior end of the calcaneus 294. Calcaneal peroneal groove for the peroneous longus 0 Indistinct or weakly developed 1 Distinct, deep separation 295. Position of sustentaculum

226

0 Reaches anterior end of calcaneus 1 Subterminal 296. Outline of sustentacular process 0 Triangular or rounded 1 Rectangular 297. Mesiolateral orientation of sustentacular facet 0 Medial 1 Dorsal 298. Anteroposterior orientation of sustentacular facet 0 Dorsal 1 45 degrees dorsoanteriorly 299. Sustentacular facet morphology 0 Slightly concave or flat 1 Posteriorly convex 300. Secondary distal calcaneoastragalar facet 0 Absent 1 Present 301. Sustentacular and posterior calcaneoastragalar facets 0 Separate 1 Merged 302. Calcaneal facet for fibula 0 Present 1 Absent 303. Orientation of calcaneal facet for fibula 0 Dorsal 1 Lateral 304. Length of the tuber calci 0 Longer than the body 1 Shorter than the body 305. Medial curvature of the tuber calci 0 Present 1 Absent 306. Ventral curvature of the tuber calci 0 Present 1 Absent 307. Proximal calcaneocuboid facet 0 Absent 1 Present 308. Angle between proximal and distal areas of calcaneocuboid facet 0 No angle

227

1 Oblique calcaneocuboid facet 309. Spatial relationship between navicular and entocuneiform 0 Entocuneiform anterior to navicular 1 Entocuneiform extends proximally medial to the distal area of the navicular 310. Angle between navicular and distal metatarsal facets of ectocuneiform 0 Oblique 1 Parallel to the distal facet 311. Prehallux 0 Absent 1 Present 312. Mt I length relative to Mt III 0 Greater than or equal to than 50% the length of Mt III 1 Less than 50% the length of Mt III or Mt I absent 313. Metatarsal V proximal process 0 Does not extend ventral to cuboid 1 Extends ventral to cuboid 314. Proximal ends of metatarsal II and III 0 Subequal in length 1 Mt II extends more proximally than Mt III 315. Ridge on proximal articular facet of metatarsal I 0 Absent 1 Present 316. Mt III thickness relative to that of Mt IV 0 Mt III thicker or subequal to Mt IV 1 Mt III thinner 317. Mt III thickness relative to that of Mt I 0 Mt I thicker than Mt III 1 Mt III thicker than Mt I 318. Median keel on palmar/plantar surface of metapodials 0 Sharp 1 Blunt 319. Foot ungual phalanx of digit IV in proximal view 0 Larger dorsoventrally than mediolaterally 1 Larger mediolaterally than dorsoventrally 320. Groove on dorsal surface of tip of ungual phalanges 0 Absent 1 Present 321. Dorsal border of ungual phalanges 0 Forming a crest-like border 1 Rounded

228

Appendix 22. New and changed character states from Suarez et al. (2016)

New or Revised Characters and Character States

Character 20 – Large foramen at anteroventral end of maxilla medial to canines

(anteroventral maxillary foramina)

The anterior palate of sparassodonts is characterized by a pair of large, paired,

anteriorly opening foramina that are positioned medial to the canines and typically

slightly posterolateral to the incisive foramina. These foramina can be readily

distinguished from the other small foramina that cover the maxilla (which are thought to be homologous with the major palatine foramen; Forasiepi, 2009) in that the latter are

smaller, do not open anteriorly, and are irregularly distributed across the palate and not

bilaterally symmetrical. These large, paired, anteriorly opening foramina have been

repeatedly noted in descriptions of sparassodonts (e.g., Forasiepi et al., 2006; Forasiepi,

2009; Engelman and Croft, 2014) but have never been named or examined in a

phylogenetic context. Here, I name these foramina anteroventral maxillary foramina.

Anteroventral maxillary foramina appear to be a diagnostic feature of the

Sparassodonta, as they are consistently present in every member of the Sparassodonta for

which this region is known (Figure 0.1; Appendix 24), and are consistently absent in

most non-sparassodont metatherian for which this character could be evaluated

(Appendix 25). No sparassodonts for which the area of the palate around the canine could

be examined clearly lacked anteroventral maxillary foramina. There are a few

metatherians that exhibit a condition that somewhat resembles that in sparassodonts. In

the phalangerids Ailurops ursinus (USNM 217576) and Trichosurus vulpecula (CMNH

18950), there are small paired foramina medial to the numerical first premolar and in

229

Vombatus ursinus (CMNH 18946), there are a pair of small paired foramina posterior to

the incisive foramina, but these are much smaller than the foramina in sparassodonts and

it is unclear if these are homologous.

Figure 0.1. Anterior rostra of sparassodonts and the recently extinct thylacine, showing the presence of anteroventral maxillary foramina (amf) in the former and their absence in the latter. A, Thylacinus cynocephalus (CMNH 18916, Thylacinidae); B, Sparassodonta gen. et sp. nov. (UF 27881, Sparassodonta incertae sedis), C, Notogale mitis (YPM- VPPU 21871, Hathliacynidae), and D, Arctodictis sinclairi (MLP 85-VII-3-1, Borhyaenidae). Abbreviations: amf, anteroventral maxillary foramen; pf, premaxillary foramina. Scale (upper right) = 10 mm. The dasyurid Dasyuroides byrnei and the thylacinid Thylacinus cynocephalus

consistently exhibit paired accessory foramina on the palate, but in these taxa the paired

foramina are smaller, do not open anteriorly, and are located in the premaxillae (Figure

0.1). The numbat, Myrmecobius fasciatus seems to consistently have paired foramina

located in its paracanine fossa (CMNH 18914, USNM 83707), but it is not clear if these

foramina are homologous to those of sparassodonts. Indeed, small nutrient foramina are

present at the ventral border of the premaxilla-maxilla suture, suggesting the paracanine

foramina in numbats may be related to these foramina. In Didelphopsis cabrerai (MNRJ

1429-V), there is a small, anteriorly-opening foramen present at the level of the P2/3 embrasure, but this foramen differs from that of sparassodonts in being much smaller and narrower.

230

Anteroventral maxillary foramina are present in Sarcophilus harrissii. However,

the presence of these foramina was variable between individuals, being present in some

specimens (USNM 582023; USNM 582024), but absent in others (USNM 142598,

CMNH 18915). The foramina were always bilaterally present or absent, no specimens

showed only a single foramen. This seems to be a condition unique to Sarcophilus

harrissii among extant dasyuromorphians, as no anteroventral maxillary foramina could be observed in several individuals of Dasyurus viverrinus, Dasyurus maculatus,

Thylacinus cynocephalus, or one the one individual of D. albopunctatus I was able to

observe (USNM 521036). Finally, in some didelphine opossums (e.g., Lutreolina,

Didelphis, Philander) some individuals exhibit small foramina slightly posterior to the

canines (e.g, Voss et al., 2018, fig. 12), but these foramina are often asymmetrical,

irregularly placed, and are not bilaterally symmetrical or open anteriorly as in

sparassodonts. It is uncertain how these foramina compare to the anteroventral maxillary

foramina in sparassodonts.

It is not clear what anatomical structures passed through these foramina,

especially since sparassodonts are extinct and it is not possible to simply dissect a living

specimen. The most likely possibility is that it transmitted portions of the major palatine

artery and nerve, especially given the reduction of the major palatine foramen in

sparassodonts. This is supported by the fact that some didelphine opossums exhibit a

trough running from the maxillopalatine fenestrae (the normal opening that transmits

these structures in marsupials) to the approximate location of the anteroventral maxillary

foramina in sparassodonts (e.g. MACN-A 17781, Hyperdidelphys inexpectata). However, it is also possible that it could have carried a branch of the greater palatine artery or

231

nasopalatine nerve, which normally pass through the incisive foramina, given how close

the anteroventral maxillary foramina are to the major palatine foramen. The presence of

anteroventral maxillary foramina and the absence of a single, large major palatine

foramen are not completely correlated, given that in other mammals a single, large major

palatine foramen is absent without anteroventral maxillary foramina being present

(Gaudin and Wible, 2004).

Character 22 – Shape of the palate

The overall shape of the palate in ventral view varies significantly in non- diprotodont metatherians. In most non-diprotodont metatherians, the molar rows are nearly parallel to the long axis of the skull, making it appear rectangular in shape in ventral view (Figure 0.2a). However, in some specialized carnivorous forms, including sparassodonts, as well as dasyuromorphians (except Myrmecobius) and the didelphimorphian Hesperocynus (Forasiepi et al., 2009), the molar rows diverge posteriorly, causing the palate to appear triangular in shape (Figure 0.2b).

Figure 0.2. Palatal morphology in metatherians. A, Didelphis virginiana (CMNH 21719) showing the “rectangular” palatal morphology between the molars; B, Thylacinus cynocephalus (CMNH 18916) showing the “triangular palatal morphology between the molars; and C, Thylacosmilus atrox (FMNH P14531, modified from Riggs, 1934) showing the bowed morphology of the molar rows.

232

This difference is easiest to see in taxa with well-developed protocones like

didelphids and hathliacynids. In taxa with rectangular palates, like didelphids, the

medialmost extent of the molars, represented by an imaginary line drawn through roughly

the medial edge of the protocones, is nearly parallel to the midline suture, whereas in taxa

with triangular palates this line is oriented at a distinct angle. However, because this

cannot be used in taxa such as borhyaenids, in which the protocone is greatly reduced and

there is significant carnassial rotation, this character was measured by measuring the

angle between the greatest length of the upper molar row and the midline suture (Figure

0.3). The palate is considered rectangular when this angle is ≤ 10°, and triangular when it is > 10°.

Figure 0.3. Method of determining codings for Character 22, “shape of palate”, demonstrating using a skull of the borhyaenid Borhyaena tuberata (modified from a picture of YPM-VPPU 15120 in Engelman and Croft, 2014). Angle θ refers to the angle between the molar row and the midline of the palate. This character is similar to but distinct from Character 21 (palatal length/width ratio) which refers to the proportions of the palate, rather than its overall shape, and is

233

more affected by the overall length and width of the palate (particularly the rostrum) than

its shape. This can be seen in taxa like Prothylacynus, which has a palatal length/width

ratio > 1.5 but a triangular palate, or Dromiciops, which has a palatal length/width ratio <

1.5 but has a rectangular palate. In Patagosmilus and Thylacosmilus, this character is coded as “?” because the molar row is not straight but curved.

Character 99 – Shape of the dentary (depth below m3–4 embrasure/length of m1–4)

In previous analyses, this character was coded based on the depth below m3-4 divided by the total length of the dentary, but has been changed here for several reasons.

For one, although several taxa in this analysis (Kokopellia, Mayulestes, Stylocynus) are known from much of the horizontal ramus the posterior portion including the coronoid and mandibular processes and the mandibular condyle is unknown, making it impossible to code these taxa based on the previous criteria. Additionally, calculating the depth of the dentary based on m1–4 rather than the total length of the dentary avoids confounding taxa that have shallow jaws with those that have deeper but elongated jaws due to diastemata between the premolar row (e.g., Cladosictis, Acyon, Thylacinus). A list of revised codings can be found in Appendix 27.

Character 117 – Size of canines

Character 118 – Shape of upper canines

These characters were separated (originally character 115 in Forasiepi, 2009) to avoid confounding canine size with canine shape. In Patagosmilus and Thylacosmilus, the canines are both relatively large in addition to the upper canines being saber-like.

SGO-PV 3490 is coded as “?” for this character, as the canine L/W ratios for this taxon

234

show it being on the borderline between the two states, comparable to some very

unspecialized sabertoothed placentals like the nimravids Dinictis and Nimravus.

Character 119 – Roots of upper canines

Character 120 – Roots of lower canines

In previous versions of this matrix (Forasiepi, 2009; Engelman and Croft, 2014;

Forasiepi et al., 2015), as well as other studies of sparassodont interrelationships (Babot et al., 2002), the state of the roots of the upper and lower canines were coded as a single character. However, the way this character was formulated meant that in previous analyses the presence of open roots on the upper canines and open roots on the lower canines were treated as morphologically unrelated states, and therefore the open roots of the upper canines were not counted as a potentially shared feature between proborhyaenids and thylacosmilids. As a result, this character has been divided into the two characters here.

It is also worth noting that although the roots of the lower canines are closed in

Thylacosmilus, their morphology suggests this state may represent a secondary reversal

from a proborhyaenid-like condition (Goin and Pascual, 1987).

Character 122 – Prominent median sulci on labial and lingual faces of canines

References: Babot et al. (2002, character 33).

Proborhyaenids are often diagnosed by the presence of median sulci on their

canines, making the canine appear “figure 8-shaped” in cross-section (Babot et al., 2002).

These are different from the small longitudinal grooves and ridges that characterize the

canines roots of borhyaenoids (character 121), which cover the entire canine. However,

observations by the authors have found that median canine sulci are much more widely

235

distributed in sparassodonts than previously thought, occurring in Pharsophorus,

Arctodictis . Generally, in non-proborhyaenid sparassodonts that have canine sulci

(except Arctodictis) there is a lingual sulcus but no labial one. This agrees with the

observations of Babot et al. (2002), who noticed that the labial sulcus was generally

shallower than the lingual in proborhyaenids.

Thylacosmilus is coded as polymorphic for this character because although the

upper canines of Thylacosmilus lack labial and lingual sulci, these features are still

present on the lower canines (Goin and Pascual, 1987). Additionally, an extracted upper

canine of Thylacosmilus shows what may be a vestigial median sulcus near its basal end.

Similarly, Patagosmilus is coded as “?” because the lower dentition is unknown and it is possible this taxon shows a similar lower canine to Thylacosmilus.

Character 125 – Orientation of P/p2

In several sparassodonts, including Arctodictis spp., Paraborhyaena boliviana,

and Proborhyaena gigantea, both P/p1 and P/p2 are oriented obliquely to the tooth row.

Character 130 – Protoconid of p1

In most metatherians, the apex of the p1 protoconid is positioned very far

anteriorly on the tooth, either dorsal to or even anterior to the anterior root of this tooth.

By contrast, in most sparassodonts (except P. simpsoni among taxa examined) the apex

of the protoconid of p1 is more posteriorly located. This condition is not unique to

sparassodonts among metatherians, also occurring in the sparassocynid Hesperocynus

(Forasiepi et al., 2009), peramelemorphians (Echymipera, Perameles) and some dasyuromorphians (Barinya, Thylacinus, possibly Myrmecobius if the sequentially first premolar is homologous to p1 in this taxon), among taxa examined. This character was

236

coded as “?” for Dasyurus, given the uncertain homology of the missing premolar locus

(possibly p2; Luckett, 1993). If the anterior premolar in this taxon is homologous with

p1, then the apex of the protoconid is also posteriorly positioned.

This character is different from (e.g, Rougier et al., 1998, character 11; Luo et al.,

2003, character 142; Ladevèze and Muizon, 2007, character 12; 2010, character 70), in

that this character primarily looks at the morphology of p1 rather than the procumbency

of P1. These two characters are not directly correlated, given that the P1 is not

procumbent in taxa for which the p1 protoconid is anteriorly positioned such as

Asiatherium, Herpetotherium, and Dromiciops. This character can also be coded for taxa

in which p1 is obliquely oriented relative to the tooth row (e.g, stagodontids and some

borhyaenoid sparassodonts). In the stagodontids Didelphodon (Clemens, 1966) and

Eodelphis (Scott and Fox, 2015), the apex of the p1 protoconid is located directly over the anterior root, whereas in borhyaenoids the apex of the p1 protoconid is positioned more posteriorly, closer to the midpoint of the tooth.

The coding of this character in Patene simpsoni is based on PVL 2618, a specimen from the late middle Eocene of northwestern Argentina (Goin et al., 1986;

Chornogubsky et al., 2018). However, as mentioned by Engelman and Croft (2014), PVL

2618 may not belong to P. simpsoni and needs to be reevaluated given that it is 20% smaller and nearly 10 million years younger than the Brazilian holotype of P. simpsoni from Itaboraí (Woodburne et al., 2014; Chornogubsky et al., 2018). Codings were used from PVL 2618 to maintain continuity with previous studies, in which this specimen was coded as part of P. simpsoni.

Character 137 – Replacement of dP3

237

In most metatherians, dP3 is replaced. However, in the thylacosmilids

Patagosmilus and Thylacosmilus, dP3 is not replaced and instead the deciduous tooth is retained throughout life (Goin and Pascual, 1987; Forasiepi and Carlini, 2010). In these

taxa, the tooth at the P3 locus can be identified as dP3 based on having three gracile roots

instead of two robust ones (Forasiepi and Sánchez-Villagra, 2014), making it possible to

identify whether dP3 was replaced even in taxa for which only the alveolus of P3 is

known (e.g., Stylocynus).

Character 138 – Timing of eruption between p3 and m3-4

Character 139 – Timing of eruption between M3-4 and m4

Character 138 is derived in part from Character 130 of previous versions of this

matrix. However, it has been reworded and recoded to focus solely on the lower dentition

given the variation in the eruption sequences of the upper dentition in sparassodonts

(Forasiepi and Sánchez-Villagra, 2014; Engelman et al., 2015). In addition, an extra

character was added given variation in the eruption of M3-4 relative to m4. In the groups

of extant marsupials examined (didelphids, microbiotheres, dasyurids), M3 and m4 erupt

in synchrony, whereas in Mayulestes and Pucadelphys M/m4 erupt in synchrony (Cifelli

and Muizon, 1998b). The borhyaenoids Arminiheringia, Lycopsis, and Prothylacynus

also exhibit a synchronous eruption of M/m4 (Forasiepi and Sánchez-Villagra, 2014), but

in Acyon, the only hathliacynid for which the eruption sequence of the postcanine teeth is

known, M3 and m4 erupt synchronously (Engelman et al., 2015).

Character 142 – Shape of upper molar row

In most sparassodonts, as well as most metatherians, the upper molar row is

straight (Figure 0.2). However, in thylacosmilids (primarily Patagosmilus goini and

238

Thylacosmilus atrox, but to a lesser degree also Anachlysictis gracilis), the molar row is distinctly bowed (Goin, 1997b; Forasiepi and Carlini, 2010)

Character 159 – Orientation of postmetacrista (based on M3 if possible)

Notes: Character 55 of Williamson et al. (2012)

In most Cenozoic metatherians (e.g., Herpetotherium, Sparassodonta,

Didelphidae, Dromiciops, Dasyuromorphia), as well as some Mesozoic groups such as deltatheroidans, the postmetacrista are oriented at an oblique angle to the tooth row. By contrast, in other Mesozoic metatherians (e.g., pediomyids, stagodontids, Kokopellia), the postmetacrista are nearly perpendicular to the tooth row. Postmetacrista were coded as

“oblique” if they were oriented at an angle of more than 10° to the centrocrista (or the imaginary line between the apices of the paracone and metacone in species where the centrocrista is discontinuous or absent), and “nearly perpendicular” if the angle between these two features was less than 10°. In species where the paracone is located slightly lingual to the metacone, this difference was not enough to affect the coding of the character.

Character 164 – Stylar Cusp A

Additional state “0” (absent) to this character. This character represents a clear morphocline and so has been coded as ordered.

Character 165 – Stylar Cusp B

The states of this character represent a clear morphocline, and so this character has been recoded as ordered.

Character 167 – Stylar Cusp D

239

The states of this character represent a clear morphocline, and so this character has been recoded as ordered.

Character 177 – Trigonid versus talonid length

This character represents a clear morphocline, so it has been reordered and coded as ordered.

Character 181 – Metaconid on m2–4

Based on observations in several groups of carnivorous marsupials, the loss of the metaconid in the posterior molars (m2-4) appears to occur in a stepwise fashion, starting from the m4 and then making its way up to the tooth row to m2-3. In sparassodonts, metaconids are present on m2-4 in the Oligocene borhyaenoids Pharsophorus and

Australohyaena, variably present on m4 in the borhyaenid Borhyaena, and then lost on m4 (only present on m2-3) in specimens of the early Miocene Arctodictis (Forasiepi et al., 2015). Similarly, in several species of thylamyine opossums, the metaconid is much smaller on m4 than in the anterior molars, and may even be lost in the m4 of Zygolestes paranensis (Goin, 1997a). Therefore, due to this variability, I added an additional character state to this character, reflecting the presence or loss of the metaconid on m4.

This character follows a logical sequence and was thus coded as ordered. Although it is likely that a further intermediate state was present (metaconid on m2, absent on m3-4), none of the taxa examined exhibited this state. The loss of the metaconid of m1 (character

180) appears to be decoupled from that of m2-4 (Forasiepi et al., 2015), likely because the m1 of metatherians appears to be homologous to the dp4 of eutherians (O'Leary et al.,

2013).

Character 312 – MtI length relative to MtIII

240

One major difference between dasyuromorphians and sparassodonts is that the pes

of most dasyuromorphians has a very short MtI (absent in some taxa like Thylacinus and

Myrmecobius), which in most dasyuromorphians is clawless. In Herpetotherium

(Horovitz et al., 2008a), Pucadelphys (Argot, 2002; Muizon and Argot, 2003),

Andinodelphys (Muizon and Argot, 2003), as well as extant caenolestids and peramelemorphians (Szalay, 1994), MtI is also very short (less than 50% the length of

MtIII). Didelphids and Dromiciops have a well-developed MtI that is at least half the length of MtIII and has well-developed articular surfaces (Szalay, 1994).

The MtI is only known for a few sparassodonts, but in these species MtI is generally long compared to the other metatarsals, unlike the condition in dasyuromorphians. In the sparassodonts Arctodictis sinclairi (Forasiepi, 2009) and

Callistoe vincei (Argot and Babot, 2011 ) the MtI is slightly greater than 50% the length of MtIII. The MtI is also known in Sipalocyon gracilis (Argot, 2003b) and Lycopsis longirostrus (Argot, 2004b), but only the proximal and distal ends are preserved, respectively, and their length relative to MtIII cannot be determined.

Taxon-Specific Character Changes

Codings for new or revised characters are listed above unless otherwise noted.

Deltatheroides cretacicus

- Character 8 (lateral palatal process of premaxilla) coded as “0” (anterior or just

reaches anterior border of canine alveolus)

- Character 9 (posterior border of incisive foramen) coded as “0” (anterior or just

reaches anterior border of canine alveolus)

241

- Character 18 (location of the infraorbital foramen) coded as “1&2” (dorsal to

posterior root of P3 or dorsal to M1)

- Character 21 (palatal length/width ratio) coded as “0” (less than 1.5)

- Character 123 (number of premolars) coded as “0” (three)

- Character 126 (diastema anterior to P1) coded as “0” (absent)

- Character 127 (diastema posterior to P1) coded as “0” (absent)

Characters coded following Rougier et al. (2004)

Deltatheridium pretrituberculare

- Character 90 (stylomastoid foramen) coded as “0” (absent) based on Beck (2012)

Kokopellia juddi

- Character 154 (wing-like cristae associated with para- and metaconules) coded as

“1” (present) based on Cifelli and Muizon (1997)

Asiatherium reshetovi

- Character 23 (number of palatal pits) coded as “0” (zero) after Szalay and

Trofimov (1996).

- Character 24 (maxillopalatine fenestrae) coded as “?”

Comments: The palate of the holotype of Asiatherium reshetovi (PIN 3907) is highly

damaged, and as a result it is not possible to determine if the maxillopalatine fenestrae

were present in this taxon.

- Character 74 (sagittal crest) coded as “2” (absent).

The holotype of A. reshetovi preserves no sagittal crest on its frontals or the preserved

extent of the parietals.

242

- Character 85 (epitympanic wing of petrosal) coded as “0” (present) based on Beck

(2012).

Alphadon spp.

For many years, almost all dentally unspecialized Cretaceous North American metatherians were assigned to the genus Alphadon (e.g., Clemens, 1966; Lillegraven,

1969). However, more recent studies have recognized that Alphadon sensu lato is a wastebasket taxon and the taxa previously assigned to this genus do not form a monophyletic group (Cifelli, 1990; Johanson et al., 2003; Williamson et al., 2012). Some taxa previously assigned to Alphadon have been reassigned to their own genera

(Eoalphadon, Protalphadon, Turgidodon, Varalphadon) whereas others have not been recovered as part of a monophyletic Alphadon in phylogenetic analyses (e.g., “Alphadon” halleyi) or have not been evaluated in a phylogenetic context to test if they belong to

Alphadon (e.g., “Alphadon” eatoni). Forasiepi (2009) coded a composite terminal taxon

for Alphadon using several taxa that are no longer referred to Alphadon sensu stricto or whose assignment to Alphadon is uncertain, including Turgidodon russelli, Protalphadon lulli, “Alphadon” halleyi, and “Alphadon” eatoni. As a result, I re-evaluated the codings of Alphadon spp. in this matrix using the previously published literature (Clemens, 1966;

Lillegraven, 1969) to ensure the codings for this terminal taxon are based only on specimens that belong to the core clade of Alphadon (A. marshi, the type species; A. sahnii, and A. wilsoni) recovered by Williamson et al. (2012).

- Character 115 (number of lower incisors) coded as “?”

- Character 116 (staggered lower incisor) coded as “?”

243

The morphology of the lower incisor row is unknown in undoubted species of Alphadon.

Previous iterations of this matrix used codings from Rougier et al. (1998), who coded

Alphadon as having four lower incisors and no “staggered” incisor. However, it is not

clear what specimens these authors used to code these characters. Three lower incisor

alveoli are known in a specimen of Protalphadon lulli (Clemens, 1966). Cifelli and

Muizon (1998a) suggest that three incisors were present in “Alphadon” eatoni (and at

least three are in the holotype), but this is not certain. The lower incisors have not been

described for any other species of Alphadon sensu lato (Williamson et al., 2012).

- Character 131 (cusp on the posterior heel of P3) coded as “0” (absent or vestigial)

Comments: The posterior cusp of P3 in a specimen of A. marshi (UALVP 2389;

Lillegraven, 1969) is comparable in development to that of Pucadelphys and

Asiatherium, and so has also been coded “0” here.

- Character 133 (change in height of the lower premolars) coded as “0” (increase

gradually in height)

Comments: In specimens assigned to the species Alphadon marshi (UCMP 50299 and

50300; Clemens, 1966), the premolars gradually increase in size from p1–3. It is not clear

where the original character coding of “2” (abrupt change in size between p1-2 and p3)

came from, but it may be based on UCMP 46882, which has since been assigned to the

genus Protalphadon (Cifelli, 1990). As a result, I code this character as “0” here.

- Character 138 (timing or eruption between p3 and m3–4) and 149 (morphology of

dp3) coded as “?”

Comments: These characters appear to be coded based on Alphadon eatoni, which is the

only member of this genus that is known from deciduous dentition is preserved in situ

244

(Cifelli and Muizon, 1998a, b). However, given that A. eatoni has never been tested to determine if it really does belong to Alphadon sensu stricto, I code these characters as “?” for the time being.

- Character 155 (relative position of para- and metaconule) coded as “0” (at or

lingual to the midpoint between paraconule and metaconule) based on Clemens

(1966) and Williamson et al. (2012).

Eodelphis browni

A review of previously published literature has found that AMNH 14169, the specimen examined by Forasiepi (2009) to provide character state information for Eodelphis cutleri, is actually the holotype of the stagodontid E. browni (Matthew, 1916). As a result, the name of this terminal taxon has been changed and additional information has been added from the primary literature.

− Character 18 (location of the infraorbital foramen) coded as “1” (dorsal to the

posterior root of P3)

− Character 19 (flaring of maxillary “cheeks” behind infraorbital foramen) coded as

“1” (absent)

− Character 43 (morphology of postglenoid process) coded as “0” (wide and low)

− Character 127 (diastema posterior to P1) coded as “1” (absent)

− Character 131 (cusp on the posterior heel of P3) coded as “1” (well-developed)

Comments: Coded based on Scott and Fox (2015).

Didelphodon vorax

− Character 1 (length of the skull) coded as “0” (short, less than twice width of skull

at level of zygomatic arches)

245

− Character 2 (length of rostrum) coded as “0” (less than 1/3 length of skull)

− Character 3 (width of braincase versus maximum postorbital width) coded as “0”

(braincase wider than maximum postorbital width)

− Character 4 (dimensions of braincase) coded as “0” (as wide as long, or slightly

wider than long)

− Character 5 (level of palate relative to the basicranium) coded as “1” (palate lower

than basicranium)

− Character 6 (paracanine fossa) coded as “0” (formed by both maxilla and

premaxilla)

− Character 8 (lateral palatal process of premaxilla) coded as “1” (posterior to

anterior border of canine alveolus)

− Character 9 (posterior border of incisive foramen) coded as “1” (posterior to

anterior border of canine alveolus)

− Character 12 (posteriormost point of premaxilla-nasal contact) coded as “0”

(posterior to the canine)

− Character 13 (anterior extent of nasals) coded as “1” (retracted posteriorly,

exposing the narial opening in dorsal view)

− Character 17 (angle of maxilla-jugal contact) coded as “0” (more than 140

degrees)

− Character 18 (location of the infraorbital foramen) coded as “2” (dorsal to M1)

− Character 21 (palatal length/width ratio) coded as “1” (greater than 1.5)

− Character 26 (minor palatine foramen) coded as “1” (small)

246

− Character 27 (posterior extent of palatines) coded as “1” (extend beyond the level

of the last molar)

− Character 28 (posterior end of palatines) coded as “2” (straight due to presence of

palatine torus)

− Character 29 (palatine reaches level of infraorbital canal) coded as “1” (present)

− Character 38 (interparietal) coded as “0” (present)

− Character 40 (parietal-alisphenoid or fronto-squamosal contact) coded as “0”

(parietal-alisphenoid contact)

− Character 41 (width of glenoid cavity) coded as “1” (more than twice

anteroposterior width)

− Character 52 (tympanic process of alisphenoid) coded as “1” (present)

− Character 59 (composition of foramen ovale) coded as “1” (on alisphenoid)

− Character 74 (sagittal crest) coded as “0” (extending to frontals)

− Character 110 (number of upper incisors) coded as “1” (four)

− Character 113 (size of I5 versus I4) coded as “2” (I5 absent)

− Character 119 (roots of upper canines) coded as “0” (closed in adults)

− Character 120 (roots of lower canines) coded as “0” (closed in adults)

− Character 121 (surface of the roots of the canines) coded as “0” (smooth

− Character 126 (diastema anterior to P1) coded as “0” (absent)

− Character 127 (diastema posterior to P1) coded as “1” (absent)

Comments: Modified based on Wilson et al. (2016).

Mayulestes ferox

- Character 98 (posttemporal notch/foramen) coded as “1” (present)

247

According to Muizon (1998), the mastoid portion of the petrosal is disarticulated in the

holotype of M. ferox, but it preserves a notch for a posttemporal notch/foramen.

- Character 132 (size of p2) coded as “0” (smaller than p3)

According to Ladevèze and Muizon (2007), the p2 of Mayulestes is slightly smaller than

p3.

- Character 133 (change in height of lower premolars) coded as “1” (abrupt change

in size between p1 and p2–3)

Comments: In a subadult specimen of Mayulestes ferox (MNHC 8267; Cifelli and

Muizon, 1998b), the p2 is seen to be relatively large, comparable to p3, whereas the p1 is

highly reduced in size, as in the holotype of M. ferox (Muizon, 1998).

- Character 105 (angle between anterior border of coronoid process and tooth row)

coded as “0” (between 95 and 105 degrees)

- Character 107 (position of the mandibular foramen) coded as “0” (posterior to the

midpoint of the coronoid process

- Character 108 (morphology of mandibular condyle) coded as “1” (cylindrical)

- Character 109 (position of mandibular condyle relative to tooth row) coded as “1”

(above level of tooth row)

Cifelli and Muizon (1998b) figure an otherwise undescribed subadult specimen of

Mayulestes ferox (MNHC 8267) that preserves many aspects of the dentary and lower dentition that are otherwise unknown in the holotype of M. ferox, including the

morphology of the posterior part of the dentary.

Pucadelphys andinus

248

- Character 107 (position of the mandibular foramen) coded as “0” (posterior to the

midpoint of the coronoid process.

In all specimens of Pucadelphys for which the position of the mandibular foramen could

be observed (YPFB Pal 6108; MNHC 8266), the foramen was located posterior to the

midpoint of the coronoid process.

- Character 111 (shape of I1) coded as “1” (subequal to smaller than remaining

incisors)

According to Ladevèze and Muizon (2007), an undescribed specimen of Pucadelphys

indicates that I1 was not enlarged relative to the other incisors.

- Character 124 (orientation of P/p1 relative to tooth row) changed to “0&1” based

on Ladevèze et al. (2011b)

As shown in Ladevèze et al. (2011b, fig. 4), in some specimens of P. andinus the first

premolar is oriented obliquely to the tooth row (MNHC 8381), whereas in other

individuals it is straighter (MNHC 8266). Additionally, in some specimens p1 is in line

with the lower tooth row while P1 is obliquely oriented (MNHC 8381).

- Character 53 (hypotympanic sinus) coded as “1” (formed by squamosal,

alisphenoid, and petrosal)

According to Beck et al. (2014), a hypotympanic sinus is present in P. andinus and has a squamosal contribution.

- Character 321 (dorsal border of ungual phalanges) coded as “1” (rounded) based

on Argot (2002)

Andinodelphys cochabambensis

249

- Character 15 (postorbital processes) coded as “1” (well-developed) based on

Muizon et al. (1997).

- Character 17 (angle of maxilla-jugal suture) coded as “0” (more than 140 degrees)

based on Muizon et al. (1997).

- Character 172 (talonid of m4 relative to m3) coded as “0” (talonid of m4 reduced

and narrower than m3)

Comments: Coded from Ladevèze and Muizon (2007)

Herpetotherium fugax

- Character 6 (paracanine fossa) coded “0” (formed by premaxilla and maxilla)

- Character 7 (precanine notch) coded “0” (absent)

- Character 9 (posterior border of incisive foramen) coded as “1” (posterior to the

anterior border of canine alveolus)

- Character 12 (posteriormost point of premaxillo-nasal contact) coded as “0&1”

anterior or at the level of the canine and posterior to the canine)

- Character 18 (location of the infraorbital foramen) coded “1&2” (dorsal to

posterior root of P3 and dorsal to M1)

- Character 111 (shape of first upper incisor) coded as “1” (enlarged)

- Character 259 (length of iliac neck) coded as “0” (longer than 15% pelvis length)

- Character 267 (width of femoral condyles) coded as “0” (lateral condyle wider

than medial condyle)

- Character 294 (calcaneal peroneal groove for the peroneus longus) coded as “0’

(indistinct or weakly developed)

- Character 304 (length of the tuber calci) coded as “0” (shorter than the body)

250

- Character 305 (medial curvature of the tuber calci) coded as “0” (absent)

- Character 306 (ventral curvature of the tuber calci) coded as “0” (absent)

These characters were coded based on Sánchez-Villagra et al. (2007), Horovitz et al.

(2008a), and observations of PIMUZ 2613, MB.Ma.50672, and MB.Ma.50672.

- Character 29 (palatine reaches level of infraorbital canal) coded as “1” (present)

Coded based on personal observations of several specimens (AMNH ) in the American

Museum of Natural History.

- Character 38 (interparietal) coded as “1” (absent or fused with parietal)

According to Voss and Jansa (2009, p. 88), the interparietal is absent in Herpetotherium

Mimoperadectes houdei

− Character 32 (anterior extent of lacrimal) coded as “1” (extending onto orbit)

− Character 37 (orbital crest) coded as “1” (present)

− Character 50 (optic foramen and sphenorbital fissure) coded as “1” (joined)

− Character 68 (median crest of basisphenoid) coded as “1” (absent)

− Character 119 (roots of upper canines) coded as “0” (closed)

These codings are based on observations of the holotype of M. houdei in Horovitz et al.

(2009).

− Character 53 (hypotympanic sinus) coded as “1” (formed by squamosal, petrosal,

and alisphenoid)

Comments: The holotype of M. houdei has been observed to have a squamosal

contribution to the hypotympanic sinus (Voss in Beck et al., 2014).

Dromiciops gliroides

- Character 86 (prootic canal) coded as “1” (absent) based on Beck (2012)

251

Dasyurus spp.

- Character 88 (parocciptial process of petrosal) coded as “1” (indistinct to absent)

based on Beck (2012)

- Character 304 (length of the tuber calci) coded as “0&1” (longer than body and

shorter than body)

Measurements from Bassarova et al. (2008) show that individuals of Dasyurus are

polymorphic for this character, even within a single species (D. maculatus). Therefore, this character has been recoded as polymorphic.

Patene simpsoni

- Character 100 (ventral margin of jaw behind m4) coded as “1” (curved)

- Character 102 (posteriormost mental foramen) coded as “0&1” (below p3 and

below p3/m1 embrasure

- Character 103 (retromolar space) coded as “1” (present)

- Character 104 (labial mandibular foramen inside masseteric fossa) coded as “0”

(absent)

Comments: Codings for these characters based on MNRJ 1351-V

- Character 193 (position of hypoconulid) coded as “1” (lingually placed and

twinned with entoconid”

Comments: Coded based on DGM 798-M.

Hondadelphys fieldsi

− Character 1 (length of the skull) coded as “0” (short, less than twice width of skull

at level of zygomatic arches)

252

− Character 2 (length of rostrum) coded as “1” (between 1/2 and 1/3 total length of

skull)

Comments: The holotype of Hondadelphys fieldsi, UCMP 37960, consists of a partial skeleton from a single individual that includes much of the skull, including the left and right maxillae, left dentary, and the basicranium (including the glenoid fossa). Based on these elements, it is possible to estimate the total length of the skull, the width of the skull at the level of the zygomatic arch, and the length of the rostrum (preorbital length of skull). The estimated total length of the skull of Hondadelphys is about 122 mm (based on the total length of the dentary plus the anteroposterior length of the basicranium posterior to the glenoid fossa), and the width of the skull is approximately 32.74 mm based on the distance from the lateral edge of the zygomatic arch to the midpoint of the skull. This means the length of the skull is less than twice its width across the zygomatic arches, and therefore “short” by the definition of Forasiepi (2009). The lacrimal is not preserved, but the facial process of the maxilla suggests the anterior border of the orbit extended to about the level of the talonid of m3, if not slightly more posterior. This roughly agrees with what is observed in a specimen tentatively assigned to Hondadelphys

(IGM 25034) by Goin (1997b). This would make the rostrum of Hondadelphys about

39% the length of the skull, categorizing the specimen as state “1” (rostral length between 1/2 and 1/3 total skull length), similar to many sparassodonts. Increasing the estimated length of the rostrum to account for a greater contribution by the lacrimal does not greatly affect the estimated proportion.

- Character 25 (major palatine foramen) coded as “1” (many small foramina on the

maxilla)

253

- Character 27 (posterior extent of palatines) coded as “1” (extend beyond the level

of the last molar)

- Character 154 (wing-like cristae associated with para- and metaconules) coded as

“0&1” (present).

- Character 178 (trigonid versus talonid width) coded as “2” (subequal to wider

than trigonid)

Comments: These characters were coded based on observations of the holotype (UCMP

37960). In the case of the major palatine foramen (character 25), the preserved regions of the palatal processes of the maxilla are covered in numerous small foramina, as in other sparassodonts.

- Character 138 (timing of eruption between p3 and m3-4) coded as “2” (p3 erupts

simultaneously with or after m4)

Comments: In UCMP 39251, p3 is in the process of erupting whereas m3 has already

erupted based on the presence of an intraradicular process at this locus, indicating that p3

erupted after m3 was already erupted, as in other sparassodonts (Forasiepi and Sánchez-

Villagra, 2014).

- Character 114 (shape of upper incisor arcade) coded as “?”.

Comments: The upper incisor row is unknown in Hondadelphys, and only the alveoli and

roots of the lower incisors are known (Goin, 1997b). Although it is possible that the shape of upper incisor arcade is as it was coded in previous versions of this matrix, in the absence of more direct evidence this character is better coded as “?”.

Stylocynus paranensis

- Character 18 (location of the infraorbital foramen) coded as “0/1”

254

Comments: Observations of a specimen of S. paranensis (MACN-PV 13203) shows that

the infraorbital foramen in this taxon opens roughly over the middle of P3.

- Character 19 (flaring of maxillary “cheeks” coded as “1” (absent)

Comments: Coded based on observations of MACN-A 5893 and MACN-PV 13203.

- Character 114 (shape of upper incisor arcade) coded as “?”.

Comments: The upper incisor row is unknown in Stylocynus, and only the alveoli of the

lower incisors are known (Marshall, 1979). Although it is possible that the shape of upper incisor arcade is as it was coded in previous versions of this matrix, in the absence of more direct evidence this character is better coded as “?”.

UF 27881

- Character 30 (position of sphenorbital foramen) coded as “0” (posterior to the

posterior border of the lacrimal) based on observations of UF 27881.

- Character 141 (size of molars increasing posteriorly) coded as “0” (moderate

posterior increase)

Comments: As noted by Engelman and Croft (2014), the M3 of UF 27881 appears

similar in size to M1, and does not exhibit the pronounced increase in size seen in other

sparassodonts like Borhyaena.

- Character 8 (lateral palatal process of premaxilla) coded as “?”

- Charater 9 (posterior border of incisive foramen) coded as “?”

Comments: Additional observations of UF 27881 and comparisons with the condition in

other sparassodonts suggest that the state of these characters cannot be determined in this

specimen.

Notogale mitis

255

- Character 7 (precanine notch) coded as “0” (absent)

- Character 8 (lateral process of premaxilla) coded as “1” (posterior to anterior

border of canine alveolus)

- Character 9 (posterior border of the incisive foramina) coded as “1” (posterior to

the anterior border of the canine alveolus)

- Character 10 (position of medial palatal process of the premaxilla) coded as “0”

(horizontal)

- Character 12 (posteriormost point of premaxilla-nasal contact) coded as “1/2”

(posterior to the canine/posterior to p2)

- Character 25 (major palatine foramina) coded as “1” (many small foramina

opening on the surface of the maxilla)

- Character 111 (shape of I1) coded as “1” (subequal to smaller than other incisors)

- Character 113 (size of I5 versus I4) coded as “2” (I5 absent)

- Character 114 (shape of upper incisor arcade) coded as “1” (slightly anteriorly

convex)

- Character 121 (surface of the roots of the canines) coded as “0” (smooth)

Comments: These characters were all coded based on personal observations of YPM-

VPPU 21871.

- Character 19 (flaring of maxillary “cheeks” behing infraorbital foramen) coded as

“1” (absent)

Comments: No expansion of the maxillary “cheeks” is present in AC 3117

- Character 23 (number of palatal pits) coded as “1/2” (one or two)

256

Comments: At least one palatal pit is present between M3–4 of AC 3117, and a second pit may have been present between M2–3.

Sallacyon hoffstetteri

- Character 5 (level of the palate relative to the basicranium) coded as “1” (palate

and basicranium at the same level)

- Character 34 (position of lacrimal foramina coded as “0” (within orbit)

- Character 35 (number of lacrimal foramina) coded as “1” (exposed on face)

- Character 149 (metacone on M4) coded as “1” (extremely reduced)

- Character 169 (size of m4) coded as “1” (m4 larger than m3)

- Character 194 (hypoconulid of m4) coded as “1” (short)

Comments: These characters coded based on (Petter and Hoffstetter, 1983)

- Character 29 (palatine reaches level of infraorbital canal) coded as “1” (absent)

based on Petter and Hoffstetter (1983)

Comments: According to Petter and Hoffstetter (1983), the infraorbital canal of MNHN

SAL 92 opens at the level of the lacrimal foramen, which is approximately over M2. By contrast, the palatines only extend anteriorly to the level of M3. Therefore, this character is coded as “absent”.

- Character 60 (secondary foramen ovale) coded as “1” (present)

Comments: This character is coded following Ladevèze and Muizon (2007)

- Character 44 (location of postglenoid foramen) coded as “1” (medial to

postglenoid process)

Comments: Coded following observations of MNHN SAL 92 in Muizon (1994) and

Muizon (1999)

257

− Character 119 (roots of upper canines) coded as “?”

− Character 120 (roots of lower canines) coded as “?”

Comments: The canines are unknown for Sallacyon.

Sipalocyon spp.

- Character 149 (metacone on M4) coded as “1&2” (absent or present but indistinct

from the cingulum)

Comments: Observations of specimens of S. gracilis show that the metacone of M4 is

absent in some specimens (YPM-VPPU 15373) and present in others (MACN-A 692).

Cladosictis patagonica

- Character 106 (angle between anterior border of coronoid process and tooth row)

coded as “0&1” (between 95 and 105 degrees and 106 and 125 degrees) based on

observations of MACN-A 5927.

- Character 165 (Stylar cusp B) coded as “0&1” (small or forming an ectocingulum

and vestigial or absent)

Observation of some specimens of C. patagonica (e.g., MACN-A 5980) show the occasional presence of a small StB that sometimes forms an ectocingulum.

- Character 189 (presence of entoconid) coded as “0&1” (absent and present)

Observations of specimens assigned to C. patagonica indicate that the entoconid is present in most individuals, but is absent in others (e.g., MACN-A 674).

Acyon myctoderos

- Character 9 (posterior border of incisive foramen) coded as “1” (posterior to

anterior border of canine alveolus)

258

- Character 12 (posteriormost point of premaxilla-nasal contact) coded as “1”

(posterior to the canine)

- Character 21 (palatal length/width ratio) coded as “1” (greater than 1.5)

- Charater 34 (position of lacrimal foramina) coded as “0” (within orbit)

- Character 35 (number of lacrimal foramina) coded as “1” (one)

- Character 44 (location of postglenoid foramen) coded as “1” (medial to

postglenoid process)

- Character 47 (paracondylar process of exoccipital and post-tympanic process of

squamosal) coded as “1” (both processes similar in length)

- Character 106 (angle between anterior border of coronoid process and tooth row)

coded as “1” (between 106 and 125 degrees)

- Character 107 (position of the mandibular foramen) coded as “1” (at the midpoint

of the coronoid process)

- Character 119 (roots of upper canines) coded as “0” (closed in adults)

- Character 120 (roots of lower canines) coded as “0” (closed in adults)

- Character 206 (shape of cranial facets) coded as “1” (dorsal edge curved)

Comments: These characters were coded based Forasiepi et al. (2006).

- Character 23 (number of palatal pits) coded as “1&2” (one and two)

- Character 154 (wing-like cristae associated with para- and metaconules) coded as

“0&1” (present and absent).

Comments: Changes in these characters were coded following recognition of individual

variation in new specimens of this species (UF 26921–26941 and UATF-V-000926;

Engelman et al., 2015).

259

- Character 138 (timing of eruption between dp3 and m3–4) coded as “2” (p3 and

m4 erupt almost simultaneously)

- Character 139 (timing of eruption between M3–4 and m4) coded as “0” (M3 and

m4 erupt simultaneously)

- Character 140 (morphology of dp3) coded as “1” (with a main cusp and smaller

accessory cusps)

- Character 148 (shape of centrocrista) coded as “0” (straight)

- Character 165 (stylar cusp B) coded as “0&1” (absent or small or forming an

ectocingulum)

Comments: These characters are coded based on the juvenile specimen described by

Engelman et al. (2015)

- Character 189 (presence of entoconid) coded as “0&1” (absent and present)

- Character 190 (shape of the entoconid) coded as “1” (labio-lingually compressed)

- Character 191 (height of the entoconid) coded as “0” (smaller than the hypoconid)

- Character 192 (location of entoconid) coded as “1” (between metaconid and

posterior tooth margin)

- Character 195 (pre-entocristid) coded as “0&1” (present and absent)

- Character 196 (direction of pre-entocristid) coded as “0” (to the base of the

trigonid)

Comments: Observations of specimens assigned to A. myctoderos indicate that the

entoconid is present in some specimens (UATF-V-000926), but is absent in others (e.g.,

MNHN-Bol-V-003668). When present, the morphology of this cusp is similar to other hathliacynids.

260

Lycopsis longirostrus

− Character 114 (shape of upper incisor row) coded as “?”

Comments: The premaxilla, upper incisor row, and lower incisor row are all unknown for the holotype (Marshall, 1977b) and only other known specimen (Goin, 1997b) of L. longirostrus, and as a result the shape of the incisor row in this taxon cannot be evaluated.

Lycopsis torresi

- Character 133 (change in height of lower premolars) coded as “0/1” (increase

gradually in height/abrupt change in size between p1 and p2-3) as this character

was coded on a single specimen for which only the right premolar row is

complete (MLP 11-113) in Suarez et al. (2016) and therefore represents an

uncertainty rather than a polymorphism.

Lycopsis viverensis

- Character 21 (palatal length/width ratio) coded as “1” (greater than 1.5)

Comments: The holotype of L. viverensis (MMH 87-6-1) preserves most of the postcanine portion of the maxilla including the midline suture (Forasiepi et al., 2003).

Based on this it is possible to extimate the proportions of the palatal process of the maxillary to some extent (the width is known, but the length is not). The ratio of the greatest preserved length of the maxilla versus the estimated bilateral width of the palate in this specimen is much greater than 1.5, and would have been even higher if the maxilla anterior to P1 was preserved. Therefore this character can be safely coded as “1” (greater than 1.5).

- Character 133 (change in height of lower premolars) coded as “?”

261

- Character 134 (roots of lower premolars) coded as “?”

Comments: These characters were changed to “?” because p3 is unknown in L. viverensis

(Forasiepi et al., 2003)

- Character 149 (metacone on M4) changed to “1/2” (extremely reduced or absent)

as Forasiepi et al. (2003) note that there is a small structure on this tooth that

could correspond to a vestigial metacone.

Prothylacynus patagonicus

− Character 5 (level of palate relative to basicranium) coded as “0” (palate lower

than basicranium) based on observation of MACN-PV 14453

Pharsophorus lacerans

− Character 18 (location of the infraorbital foramen) was coded as “0” (anterior or

dorsal to anterior root of P3)

Comments: In MNHN SAL 96/YPM-VPPU 20551 (the only specimen that can be

confidently assigned to Pharsophorus that preserves the infraorbital foramen), the

infraorbital foramen is located over the anterior root of P3.

− Character 15 (postorbital processes) changed from “0” (absent or indistinct) to

“1” (well-developed) based on observations of MNHN SAL 96/YPM-VPPU

20551

− Character 100 (ventral margin of the jaw below m4) was coded as “0” (straight),

based on observations of the holotype of this species

Comments: The ventral margin of the jaw of a specimen tentatively assigned to P.

lacerans (MPEF-PV 4190; Goin et al., 2010) appears to be curved, but this specimen has not been studied in detail to determine if it belongs to P. lacerans (see main text).

262

− Character 145 (position of the metacone relative to the paracone) coded as “0”

(approximately at the same level).

− Character 147 (bases of paracone and metacone) coded as “0” (adjoined)

Comments: Coded based on Patterson and Marshall (1978)

Borhyaena tuberata

− Character 5 (level of palate relative to basicranium) coded as “0” (palate lower

than basicranium) based on observation of MACN-A 5780 and observations of

specimens in Sinclair (1906)

Australohyaena antiquua

− Character 1 (length of the skull) coded as “1” (short, less than twice width of skull

at level of zygomatic arches)

Comments: Although the only known skull of Australohyaena (UNPSJB-PV 113) does

not preserve a complete zygomatic arch, the preserved width of this skull across the

zygomatic arches is nearly 228 mm. The skull of the same specimen is only estimated to

have been about 300 mm long, making the width of the zygoma easily more than half the

length of the skull. In order for the skull to be considered “long”, the glenoid region of

this specimen would need to be nearly 70 mm shorter than it actually is.

− Character 178 (trigonid versus talonid width) coded as “1” (narrow, but wider

than the base of the metaconid or protoconid)

Arctodictis sinclairi

− Character 5 (level of palate relative to basicranium) coded as “0” (palate lower

than basicranium) based on observation of MLP 85-VII-3-1

263

− Character 313 (metatarsal V proximal process) coded as “0” (does not extend

ventral to cuboid)

Comments: Observations of the pes of MLP 85-VII-3-1 (Forasiepi, 2009, fig 46) show that the proximal process of MtV does not extend proximally.

Arctodictis munizi

− Character 5 (level of palate relative to basicranium) coded as “0” (palate lower

than basicranium) based on observation of CORD-PZ 1210-1/2 in Forasiepi et al.

(2004)

Callistoe vincei

− Character 5 (level of palate relative to basicranium) coded as “1” (palate and

basicranium at same level)

− Character 38 (interparietal) coded as “1” (absent or fused with the parietal)

− Character 40 (parietal-alisphenoid or fronto-squamosal contact) coded as “1”

(fronto-squamosal contact)

− Character 165 (stylar cusp B) coded as “1” (small or forming and ectocingulum)

− Character 167 (stylar cusp D) coded as “?”, as Babot et al. (2002) suggest the

posterolabial cingulum on the upper molars could represent a reduced stylar cusp

D, a condition that is seen in some other sparassodonts (e.g., MPEF-PV 4345;

Goin et al., 2010)

Comments: These characters coded based on Babot et al. (2002).

− Character 34 (position of lacrimal foramina) coded as “0&1” (two and one)

− Character 35 (number of lacrimal foramina) coded as “0&1” (two and one)

264

Comments: The number and position of the lacrimal foramina differ between different specimens of C. vincei (Forasiepi et al., 2015). Therefore, these characters are polymorphic, rather than uncertainties.

Paraborhyaena boliviana

− Character 19 (flaring of maxillary “cheeks”) coded as “1” (absent)

− Character 21 (length/width ratio of palatal process of maxilla in ventral view)

coded as “1” (greater than 1.5)

Comments: Coded based on Petter and Hoffstetter (1983)

− Character 185 (precingulid) coded as “2” (absent)

Comments: Coded based on observations of UATF-V-000129

Thylacosmilus atrox

− Character 5 (level of palate relative to basicranium) coded as “1” (palate and

basicranium at same level)

− Character 19 (flaring of maxillary “cheeks”) coded as “1” (absent)

Comments: Coded based on observations of FMNH P14531

− Character 111 (shape of I1) coded as “?” (present).

Comments: Although it is clear Thylacosmilus has a reduced compliment of incisors compared to other sparassodonts, the homologies of the retained incisors are uncertain.

Upper incisors were clearly present based on wear facets (Churcher, 1985), but no specimen of Thylacosmilus preserves the upper incisors. It is possible that one of the lost upper incisors in Thylacosmilus corresponds to I1, based on arguments for incisor homology in borhyaenids (Forasiepi, 2009), but it is also possible that the lost upper incisors represent different loci, especially given the unusual upper incisor arcade of

265

Paraborhyaena (Petter and Hoffstetter, 1983) where there are two pairs of incisors

(possibly I1-2 and I3-4) that are mesiodistally aligned with one another. Until a specimen of Thylacosmilus with the upper incisors is discovered, their number and possible homologies remain uncertain.

266

Appendix 23. List of comparative specimens and references used to code new and revised characters in the phylogenetic matrix.

Taxon Specimens References Deltatheroides cretacicius — Rougier et al. (2004) Deltatheridium — Rougier et al. (1998) pretrituberculare Holoclemensia texana — Davis and Cifelli (2011) Didelphodon vorax — Wilson et al. (2016) Eodelphis browni AMNH 14169 Scott and Fox (2015) Kokopellia juddi — Cifelli and Muizon (1997) Pediomyidae — Davis (2007) Alphadon spp. — Clemens (1966) Asiatherium reshetovi — Trofimov and Szalay (1994), Szalay and Trofimov (1996) Mayulestes ferox — Cifelli and Muizon (1998b), Muizon (1998), Babot et al. (2002) Pucadelphys andinus — Cifelli and Muizon (1998b), Babot et al. (2002) Andinodelphys — Babot et al. (2002) cochabambensis Herpetotherium fugax AMNH 22304 Fox (1983), Horovitz et al. (2008b) Peradectidae — Horovitz et al. (2009) Didelphis albiventris UMMZ 125453 Macrini (2005b) Monodelphis spp. — Macrini (2001), Wible (2003) Dromiciops gliroides — Marshall (1978a), Gosselin- Ildari (2006) Dasyurus spp. CMNH 18912 Macrini (2005a) Sminthopsis crassicaudata — Archer (1981), Macrini (2009) Thylacinus cynocephalus CMNH 18916 Engelman, pers. obs. Patene simpsoni MNRJ 1331-V Marshall (1981), Goin et al. (1986) Sallacyon hoffstetteri — Petter and Hoffstetter (1983) Notogale mitis AC 3117; YPM-VPPU Marshall (1981) 21871

267

Acyon myctoderos UF 26921, UF 26933, Forasiepi et al. (2006), UATF-V-000926 Engelman et al. (2015) Cladosictis patagonica MACN-A 5927, MACN- Sinclair (1906), Marshall A 5950 (1981) Sipalocyon spp. AMNH 9254, MACN-A Sinclair (1906), Marshall 691 (1981) UF 27881 UF 27881 Engelman and Croft (2014) Hondadelphys fieldsi UCMP 37960 Suarez pers. comm. Prothylacynus patagonicus MACN-A 707, MACN-A Sinclair (1906), Marshall 5931 (1979), Forasiepi and Sánchez-Villagra (2014) Stylocynus paranensis MACN-A 5893, MACN- — PV 13203, MLP 11-94 Borhyaena tuberata MACN-A 6203-6265 Sinclair (1906), Forasiepi et al. (2015) Arctodictis munizi — Forasiepi et al. (2004) Arctodictis sinclairi MLP 85-VII-3-1 Forasiepi (2009), Forasiepi et al. (2015) Pharsophorus lacerans MACN-A 52-391 Thylacosmilus atrox FMNH P14531 Goin and Pascual (1987), Babot et al. (2002) Paraborhyaena boliviana UATF-V-000129 Petter and Hoffstetter (1983), Babot et al. (2002) Callistoe vincei — Babot et al. (2002) Patagosmilus goini — Forasiepi and Carlini (2010) Australohyaena antiquua — Forasiepi et al. (2015) Lycopsis longirostrus UCMP 38061 Forasiepi and Sánchez- Villagra (2014) Lycopsis padillai — Suarez et al. (2016) Lycopsis torresi MLP 11-113 — Lycopsis viverensis — Forasiepi et al. (2003) Proborhyaena gigantea AMNH 29576, MACN-A Mones and Ubilla (1978), 52-382 Patterson and Marshall (1978) Eomakhaira molossus SGO-PV 3490 —

268

Appendix 24. List of sparassodont specimens where paired anteroventral maxillary foramina have been observed.

Taxon Specimen Reference Sparassodonta incertae sedis Hondadelphys fieldsi UCMP 37960 Personal observation Sparassodonta gen. et sp. Engelman and Croft UF 27881 nov. (2014) Hathliacynidae Notogale mitis YPM-VPPU 21871 Personal observation cf. Sallacyon UATF-V-000165 Anaya-Daza et al. 2010 AMNH 9254, MACN-A 692, Sinclair 1906; Personal Sipalocyon gracilis YPM-VPPU 15029, YPM-VPPU observation 15154, YPM-VPPU 15373 Cladosictis centralis MACN-A 11639 Personal observation MACN-A 5927; MACN-A 5950; MACN-A 6280; YPM-VPPU Sinclair 1906; Personal Cladosictis patagonica 15170, YPM-VPPU 15046, observation YPM-VPPU 15702 Acyon myctoderos MNHN-Bol-V-003668 Forasiepi et al. (2006) Basal Borhyaenoidea MACN-A 5931; MACN-PV Prothylacynus patagonicus Personal observation 14453 Marshall (1978b); cf. Pharsophorus AMNH 29591 Personal observation Borhyaenidae Australohyaena antiquua UNPSJB-PV 113 Forasiepi et al. (2015) Arctodictis sinclairi MLP 85-VII-3-1 Forasiepi (2009) Arctodictis munizi CORD-PZ 1210-1/2 Forasiepi et al. 2004 MACN-A 6203-6265; YPM- Borhyaena tuberata VPPU 15120, YPM-VPPU 15701 Sinclair 1906 Proborhyaenidae Petter and Hoffstetter Paraborhyaena boliviana MNHN SAL 51 (1983) Thylacosmilidae FMNH P14531, MLP 35-X-4-1 Riggs (1934); Personal Thylacosmilus atrox (= FMNH P14474) observation Only specimens where anterior maxillary foramina could unequivocally be observed are listed. These foramina may be present in the holotype of Callistoe vincei (PVL 4187; Babot et al., 2002) but this could not be determined with certainty.

269

Appendix 25. List of metatherian taxa in which the palatal canine foramina were absent.

Higher Taxonomic Taxon Specimens References Group Deltatheroida Sulestes karakshi — Averianov et al. (2010) Deltatheroides — Rougier et al. cretacicus (2004) Deltatheridium — Rougier et al. pretrituberculare (1998) Pediomyidae Protolambda hatcheri — Clemens (1966) Stagodontidae Didelphodon vorax — Wilson et al. (2016) Eodelphis browni — Scott and Fox (2015) Marsupialiformes Didelphopsis cabrerai MNRJ 1429-V — incertae sedis Eobrasilia coutoi — Simpson (1947) “Gurlin Tsav skull” — Szalay and Trofimov (1996) Peradectidae Mimoperadectes — Horovitz et al. houdei (2009) Pucadelphyidae Pucadelphys andinus — Marshall and Muizon (1995), Ladevèze et al. (2011b) Mayulestidae Mayulestes ferox — Muizon (1998) Herpetotheriidae Herpetotherium fugax AMNH 22304 — Polydolopimorphia Epidolops ameghinoi — Beck (2016) Groeberia minioprioi — Pascual et al. (1994) Kramadolops — Flynn and Wyss mckennai (2004) Didelphimorphia Glironia venusta AMNH 71394 Voss and Jansa (2009) Caluromysiops irrupta AMNH 208151 Voss and Jansa (2009) Caluromys derbianus USNM 335001 Caluromys lanatus AMNH 13199 Hyladelphys AMNH 267338 Voss et al. kallinowskii (2001), Voss and

270

Jansa (2009) Gracilinanus emiliae — Voss et al. (2001) Marmosa murina CMNH 18878 — Monodelphis D. Croft Wible (2003) domestica personal collection Metachirus AMNH 97320, Voss and Jansa nudicaudatus USNM 546190 (2009) Didelphis albiventris UMMZ 125453 — Didelphis virginiana CMNH 21719 — Hyperdidelphys — Goin and inexpectata Pardiñas (1996) Hyperdidelphys — Goin and parvula Pardiñas (1996) Thylophorops — Goin and perplana Pardiñas (1996) Thylophorops — Goin and chapadmalensis Pardiñas (1996) Paucituberculata Caenolestes USNM 513429 — fuliginosus Palaeothentes lemoinei — Forasiepi et al. (2014) Microbiotheria Microbiotherium — Sinclair (1906) tehuelchum Dromiciops gliroides FMNH 127448 — Notoryctemorphia Notoryctes typhlops CMNH 19326 O'Leary et al. (2013) Peramelemorphia Galadi speciosus — Travouillon et al. (2010) Galadi amplus — Travouillon et al. (2013) Peroryctes broadbenti — Aplin et al. (2010) Echymipera kalubu AMNH — 194713, USNM 277441 Perameles nasuta CMNH 18882 — Macrotis lagotis CMNH 18881 — Dasyuromorphia Sminthopsis USNM 218463 — fuliginosus

271

Phascogale calura AMNH 199687 — Dasyuroides byrnei AMNH 257603 — Dasyurus viverrinus CMNH 18913 — Dasyurus maculatus CMNH 18912 — Sarcophilus harrissii CMNH 18915 — Barinya wangala — Godthelp et al. (1999) Badjcinus turnbulli — Muirhead and Wroe (1998) Mutpuracinus — Murray and archibaldi Megirian (2006) Nimbacinus dicksoni — Wroe and Musser (2001) Thylacinus CMNH 18916 — cynocephalus Myrmecobius fasciatus Malleodectes mirabilis — Arena et al. (2011) Diprotodontia Acrobates pygmaeus USNM 588356 — Cercartetus concinnus USNM 218466 — Petauroides volans USNM 277373 — Pseudocheirops FMNH 128365 — cupreus Pseudocheirus USNM 237694 — peregrinus Phascolarctos cinereus CMNH 18955 — Hypsiprymnodon USNM 283443 — moschatus Bettongia lesueur CMNH 18928 — Macropus rufogriseus CMNH 18924 —

272

Appendix 26. Measurements of the angle of the tooth rows relative to the midline of the skull. Angle 1: Angle between midline suture and medial border of the protocones (M1-4 if possible, M1-3 if the protocone of M4 is reduced). Angle 2: Angle between midline suture and greatest length of M1-4, this measurement was used to code Character 22.

Taxon Specimen Reference Angle 1 Angle 2 Deltatheroides cretacicus AMNH 21700 Rougier et al. 3.1 6.7 (2004) Deltatheridium PSS-MAE 133 Rougier et al. 6.8 10.05 pretrituberculare (1998) Didelphodon vorax NDGS 431 Wilson et al. 0.3 6.9 (2016) Pucadelphys andinus MNHC 8266 Ladevèze et 7.9 9.0 al. (2011a) Andinodelphys MNHC 8264 Muizon et al. 2.6 8.55 cochabambensis (1997) Mayulestes ferox MNHC 1249 Muizon 1.1 3.55 (1998) Herpetotherium fugax UA 8572 Fox (1983) 6.15 7.6 Glironia venusta AMNH Pers. obs. 7.3 7.3 Caluromys derbianus Pers. obs. 6.2 3.5 Didelphis virginiana CMNH 21719 Pers. obs. 4.35 6.4 Didelphis albiventris UMMZ 125453 Pers. obs. 3.05 7.95 Monodelphis brevicaudata CM 52729 Wible (2003) 5.25 6.4 Dromiciops gliroides FMNH 127448 Pers. obs. 9.5 6.45 Dasyurus maculatus CMNH 18912 Pers. obs. 9.8 10.75 Sminthopsis fuliginosus USNM 218463 Pers. obs. 9.95 10.06 Thylacinus cynocephalus CMNH 18916 Pers. obs. 9.85 14.85 Hondadelphys fieldsi UCMP 37960 Pers. obs. 4.5 5.5 UF 27881 UF 27881 Pers. obs. 19 20.7 Acyon myctoderos MNHN-Bol-V- Forasiepi et al. 14 15.3 003668 (2006) Cladosictis patagonica MACN-A 5950 Pers. obs. 13.8 15.45 Cladosictis patagonica MACN-A 5927 Pers. obs. 17.5 23.5 Sallacyon hoffstetteri Sipalocyon gracilis AMNH 9254 Pers. obs. 10.6 12.15 Sipalocyon gracilis Pers. obs. Lycopsis viverensis MMH 87-6-1 Forasiepi et al. 12.3 13.5 (2003) Prothylacynus patagonicus MACN-PV Pers. obs. 13.95 19.2

273

Australohyaena antiquua UNPSJB PV 113 Forasiepi et al. 12.75 13.5 (2015) Arctodictis sinclairi MLP 85-VII-3-1 Pers. obs. 11.9 14.6 Borhyaena tuberata Callistoe vincei PVL 4187 Babot et al. 13.75 14.2 (2002) Eomakhaira molossus SGOPV 3490 Pers. obs. — 24.55

5 Estimated due to crushing to skull. Palate is clearly triangular in ventral view but exact angle is difficult to determine. 274

Appendix 27. Recoding of dentary depth values for metatherians in the phylogenetic matrix Measurements taken from both personal observation of specimens as well as references listed in the Reference column. Critical values based on the codings in previous iterations of this matrix (e.g., Forasiepi, 2009) and the data collected here are Shallow: < 0.6; Intermediate: 0.6-0.8; and Deep: > 0.8. Depth under m3 also shown for proborhyaenids and Arctodictis sinclairi to show how the holotype of Eomakhaira molossus (SGOPV 3490) compares to these taxa.

m3D/ m4D/ Coding in Previous Taxon Specimen Lm1-4 m3D Lm1-4 m4D Lm1-4 Reference Versions Deltatheridium Kielan-Jaworowska ZPAL MgM-I/91 9.21 — — 3.68 0.40 Shallow pretrituberculare (1975) Cifelli and Muizon Kokopellia juddi OMNH 26361 8.61 — — 3.67 0.43 — (1997) Szalay and Trofimov Asiatherium reshetovi PIN 3907 7.3 — — 2.8 0.38 Shallow (1996) Alphadon marshi UCMP 50299 9.00 — — 3.67 0.41 Clemens (1966) Shallow Alphadon marshi UCMP 50300 8.75 — — 3.17 0.36 Clemens (1966) Shallow Pediomys elegans UCMP 46883 7.75 — — 2.88 0.37 Clemens (1966) Shallow Eodelphis browni AMNH 14169 18.49 — — 10.44 0.56 Scott and Fox (2015) Shallow Didelphodon vorax LACM 15433 ~25 ~19 0.76 Clemens (1968) Intermediate 6.49- 2.31- 0.34- Pucadelphys andinus Multiple, see reference — — Ladevèze et al. (2011b) Shallow 7.40 3.90 0.57 Andinodelphys MNHC 8264 11.28 — — 4.68 0.42 Muizon et al. (1997) Shallow cochabambensis Mayulestes ferox MHNC 1249 13.25 6.27 0.47 6.81 0.51 Muizon (1998) — Mimoperadectes labrus UM 66144 13.07 — — 6.6 0.5 Engelman pers. obs. Shallow Herpetotherium fugax PIMUZ 2613 7.74 — — 2.46 0.31 Horovitz et al. (2008a) Shallow Didelphis albiventris USNM 536826 17.87 — — 11.40 0.64 Engelman pers. obs. Shallow Monodelphis CM 52729 9.1 3.6 0.40 Wible (2003) Shallow brevicaudata Dromiciops gliroides AMNH 92147 5.80 — — 2.61 0.26 Marshall (1978a) Shallow

275

Dasyurus maculatus CMNH 19821 23.08 — — 10.78 0.47 Engelman pers. obs. Shallow Thylacinus cynocephalus CMNH 50.65 — — 27.49 0.54 Engelman pers. obs. Shallow Hondadelphys fieldsi UCMP 37960 31 — — 17 0.55 Marshall (1976b) Shallow Marshall (1979), Stylocynus paranensis MLP 11-94 37 — — 56 0.66 — Engelman pers. obs. Acyon myctoderos MNHN-Bol-V-003668 41.10 — — 20.08 0.49 Forasiepi et al. (2006) Shallow Shallow and Cladosictis patagonica YPM-VPPU 15170 30.73 — — 21.34 0.33 Sinclair (1906) Intermediate Cladosictis patagonica MACN-A 9360 Shallow and 29.5 — — 14 0.47 Marshall (1981) Intermediate Sipalocyon gracilis MACN-A 691 26 — — 13.2 0.51 Marshall (1981) Shallow Sipalocyon gracilis MACN-A 5938 22.7 — — 11.8 0.52 Marshall (1981) Shallow Sipalocyon gracilis MACN-A 5940 22.8 — — 11.6 0.51 Marshall (1981) Shallow Sipalocyon gracilis MACN-A 5964 24.9 — — 14.6 0.59 Marshall (1981) Shallow Sipalocyon gracilis MACN-A 5965 25 — — 13.3 0.53 Marshall (1981) Shallow Sipalocyon gracilis YPM-VPPU 15029 26.16 — — 13.46 0.51 Sinclair (1906) Shallow Lycopsis longirostrus UCMP 38061 58 — — 30.30 0.52 Marshall (1977b) Shallow Lycopsis longirostrus IGM 250974 60.2 — — 30.9 0.51 Goin (1997b) Shallow Prothylacynus Marshall (1979), MACN-A 706 47.5 — — 31 0.65 Intermediate patagonicus Engelman pers. obs. Prothylacynus YPM-VPPU 15700 47.5 — — 31.5 0.66 Sinclair (1906) Intermediate patagonicus Patterson and Marshall Pharsophorus lacerans MACN-A 52-391 55.60 — — 37.31 0.67 — (1978) Borhyaena tuberata MACN-A 9342 34.5 — — 50.5 0.68 Marshall (1978b) Intermediate Borhyaena tuberata MACN-A 9345 31 — — 51.5 0.60 Marshall (1978b) Intermediate Arctodictis sinclairi MLP 85-VII-3-1 46.36 37.28 0.80 41.15 0.91 Forasiepi (2009) Deep Arctodictis munizi MLP 11-85 61.50 — — 44.00 0.72 Forasiepi et al. (2004) Deep Arctodictis munizi MACN-A 5915-17 62.70 — — 51.00 0.81 Forasiepi et al. (2004) Deep Arctodictis munizi MACN-A 5918-21 64.50 — — 49.90 0.77 Forasiepi et al. (2004) Deep Arctodictis munizi MLP 11-65 59.00 — — 50.00 0.85 Forasiepi et al. (2004) Deep

276

Thylacosmilus atrox FMNH P14344 52.64 — — 28.67 0.54 Riggs (1934) Intermediate Goin and Pascual Thylacosmilus atrox MMP 1443 56.38 — — 43.69 0.77 Intermediate (1987) Callistoe vincei PVL 4197 51.05 41.69 0.82 45.95 0.90 Babot et al. (2002) Deep Arminiheringia auceta MACN-A 10790 53.34 46.36 0.87 50.80 0.95 Engelman, pers. obs. —

Paraborhyaena boliviana UATF-V-000129 86.81 61.71 0.71 68.98 0.79 Engelman, pers. obs. Deep

Proborhyaena gigantea AMNH 25976 109.71 78.98 0.72 81.53 0.74 Engelman, pers. obs. — Eomakhaira molossus SGOPV 3490 37.3 29.4 0.79 31.8* 0.85 Present Study —

277

LITERATURE CITED

Abello, M. A., M. Reyes, de los, A. M. Candela, F. Pujos, D. Voglino, and B. M. Quispe.

2015. Description of a new species of Sparassocynus (Marsupialia: Didelphoidea:

Sparassocynidae) from the late Miocene of Jujuy (Argentina) and taxonomic

review of Sparassocynus heterotopicus from the Pliocene of Bolivia. Zootaxa

3937:146-170.

Akersten, W. A. 1985. Canine function in Smilodon (Mammalia, Felidae,

Machairodontinae). Contributions in Science, Natural History Museum of Los

Angeles County 356:1-22.

Aktan, A. M., İ. Kara, İ. Şener, C. Bereket, S. Çelik, M. Kırtay, M. E. Çiftçi, and N.

Arıcı. 2012. An evaluation of factors associated with persistent primary teeth.

European Journal of Orthodontics 34:208-212.

Ameghino, F. 1887. Enumeración sistemática de las especies de mamíferos fósiles

coleccionados por C. Ameghino en los terrenos eocenos de la Patagonia austral.

Boletín del Museo de La Plata 1:1-24.

Ameghino, F. 1894. Enumération synoptique des espécies de mamífères fossiles des

formations éocénes de Patagonia. Boletín de la Academia de Ciencias de Córdoba

13:259–452.

Ameghino, F. 1897. Les Mammifères crétacés de l'Argentine. Deuxième contribution à la

connaissance de la faune mammalogique des couches à Pyrotherium. Boletín del

Instituto Geográfico Argentino 18:406-521.

Anaya Daza, F., B. Shockey, and D. A. Croft. 2010. Sparassodonts of Salla: Species

richness and new taxa of Carnivorous marsupials from the late Oligocene of

278

Bolivia. Journal of Vertebrate Paleontology, SVP Program and Abstracts Book,

2010:53A.

Anders, U., W. von Koenigswald, I. Ruf, and B. H. Smith. 2011. Generalized individual

dental age stages for fossil and extant placental mammals. Paläontologische

Zeitschrift 85:321-339.

Antoine, P.-O., M. Roddaz, S. Brichau, J. Tejada-Lara, R. Salas-Gismondi, A.

Altamirano, M. Louterbach, L. Lambs, T. Otto, and S. Brusset. 2013. Middle

Miocene vertebrates from the Amazonian Madre de Dios Subandean Zone, Perú.

Journal of South American Earth Sciences 42:91-102.

Antón, M. 2013. Sabertooth. 243 pp. Indiana University Press, Bloomington.

Antón, M., and À. Galobart. 1999. Neck function and predatory behavior in the scimitar

toothed cat Homotherium latidens (Owen). Journal of Vertebrate Paleontology

19:771-784.

Antón, M., M. J. Salesa, and G. Siliceo. 2012. Machairodont adaptations and affinities of

the Holarctic late Miocene homotherin Machairodus (Mammalia, Carnivora,

Felidae): The case of Machairodus catocopis Cope, 1887. Journal of Vertebrate

Paleontology 33:1202-1213.

Aplin, K. P., K. M. Helgen, and D. P. Lunde. 2010. A review of Peroryctes broadbenti,

the giant bandicoot of Papua New Guinea. American Museum Novitates 3696:1-

41.

Archer, M. 1976. The dasyurid dentition and its relationships to that of didelphids,

thylacinids, borhyaenids (Marsupicarnivora) and peramelids (Peramelina:

Marsupialia). Australian Journal of Zoology 39:1-34.

279

Archer, M. 1981. Systematic revision of the marsupial dasyurid genus Sminthopsis

Thomas. Bulletin of the American Museum of Natural History 168:1-223.

Arena, D. A., M. Archer, H. Godthelp, S. J. Hand, and S. Hocknull. 2011. Hammer-

toothed ‘marsupial skinks'; from the Australian Cenozoic. Proceedings of the

Royal Society B: Biological Sciences 278:3529-3533.

Argot, C. 2002. Functional-adaptive analysis of the hindlimb anatomy of extant

marsupials and the paleobiology of the paleocene marsupials Mayulestes ferox

and Pucadelphys andinus. Journal of Morphology 253:76-108.

Argot, C. 2003a. Functional adaptations of the postcranial skeleton of two Miocene

borhyaenoids (Mammalia, Metatheria), Borhyaena and Prothylacinus, from South

America. Palaeontology 46:1213-1267.

Argot, C. 2003b. Postcranial functional adaptations in the South American Miocene

borhyaenoids (Mammalia, Metatheria): Cladosictis, Pseudonotictis and

Sipalocyon. Alcheringa: An Australasian Journal of Palaeontology 27:303-356.

Argot, C. 2004a. Evolution of South American mammalian predators (Borhyaenoidea):

anatomical and palaeobiological implications. Zoological Journal of the Linnean

Society 140:487-521.

Argot, C. 2004b. Functional-adaptive analysis of the postcranial skeleton of a Laventan

borhyaenoid, Lycopsis longirostris (Marsupialia, Mammalia). Journal of

Vertebrate Paleontology 24:689-708.

Argot, C. 2004c. Functional-adaptive features and palaeobiologic implications of the

postcranial skeleton of the late Miocene sabretooth borhyaenoid Thylacosmilus

280

atrox (Metatheria). Alcheringa: An Australasian Journal of Palaeontology 28:229-

266.

Argot, C., and J. Babot. 2011. Postcranial morphology, functional adaptations and

palaeobiology of Callistoe vincei, a predaceous metatherian from the Eocene of

Salta, north-western Argentina. Palaeontology 54:447-480.

Arnal, M., and M. G. Vucetich. 2015. Main radiation events in Pan-Octodontoidea

(Rodentia, Caviomorpha). Zoological Journal of the Linnean Society 175:587-

606.

Atramentowicz, M. 1986. Dynamique de population chez trois marsupiaux didelphides

de Guyane. Biotropica 18:136-149.

Averianov, A., E. Obraztsova, I. Danilov, P. Skutschas, and J. Jin. 2016. First nimravid

skull from Asia. Scientific Reports 6:25812.

Averianov, A. O., J. D. Archibald, and E. G. Ekdale. 2010. New material of the Late

Cretaceous deltatheroidan mammal Sulestes from Uzbekistan and phylogenetic

reassessment of the metatherian-eutherian dichotomy. Journal of Systematic

Palaeontology 8:301-330.

Babot, J., and D. A. García-López. 2010. A new borhyaenoid (Mammalia, Metatheria,

Sparassodonta) from the middle Eocene of Salta Province, Argentina: In X

Congreso Argentino de Paleontología y Bioestratigrafía y VII Congreso

Latinoamericano de Paleontología.

Babot, M. J. 2005. Los Borhyaenoidea (Mammalia, Metatheria) del Terciario inferior del

Noroeste argentino. Aspectos filogenéticos, paleobiológicos y bioestratigráficos.,

pp. 454. Universidad Nacional de Tucumán.

281

Babot, M. J., J. E. Powell, and C. Muizon, de. 2002. Callistoe vincei, a new

Proborhyaenidae (Borhyaenoidea, Metatheria, Mammalia) from the early Eocene

of Argentina. Geobios 35:615-629.

Barrett, P. Z. 2016. Taxonomic and systematic revisions to the North American

Nimravidae (Mammalia, Carnivora). PeerJ 4:e1658.

Bassarova, M., C. M. Janis, and M. Archer. 2008. The calcaneum: On the heels of

marsupial locomotion. Journal of Mammalian Evolution 16:1-23.

Beck, R. 2012. An ‘ameridelphian’ marsupial from the early Eocene of Australia

supports a complex model of Southern Hemisphere marsupial biogeography.

Naturwissenschaften 99:715-729.

Beck, R. M. D. 2016. The Skull of Epidolops ameghinoi from the Early Eocene Itaboraí

Fauna, Southeastern Brazil, and the Affinities of the Extinct Marsupialiform

Order Polydolopimorphia. Journal of Mammalian Evolution:1-42.

Beck, R. M. D. In press. Current understanding of the phylogeny of Metatheria: A

review; pp. in F. J. Goin, and A. M. Forasiepi (eds.), New World marsupials and

their extinct relatives: 100 million years of evolution. Springer.

Beck, R. M. D., H. Godthelp, V. Weisbecker, M. Archer, and S. J. Hand. 2008.

Australia's oldest marsupial fossils and their biogeographical implications. PLoS

ONE 3:e1858.

Beck, R. M. D., K. J. Travouillon, K. P. Aplin, H. Godthelp, and M. Archer. 2014. The

osteology and systematics of the enigmatic Australian Oligo-Miocene metatherian

Yalkaparidon (Yalkaparidontidae; Yalkaparidontia; ?Australidelphia;

Marsupialia). Journal of Mammalian Evolution 21:127-172.

282

Bensley, B. A. 1903. On the relationships of the Australian Marsupialia; with remarks on

the relationships of marsupials in general. Transactions of the Linnean Society of

London, Zoology 9:83-217.

Bertrand, O. C., J. J. Flynn, D. A. Croft, and A. R. Wyss. 2012. Two new taxa

(Caviomorpha, Rodentia) from the early Oligocene Tinguiririca Fauna (Chile).

American Museum Novitates 3750:1-36.

Bi, S., X. Jin, S. Li, and T. Du. 2015. A new Cretaceous Metatherian mammal from

Henan, China. PeerJ 3:e896.

Billet, G., C. d. Muizon, and B. Mamani Quispe. 2008. Late Oligocene mesotheriids

(Mammalia, Notoungulata) from Salla and Lacayani (Bolivia): implications for

basal mesotheriid phylogeny and distribution. Zoological Journal of the Linnean

Society 152:153-200.

Blanco, R. E., W. W. Jones, and G. A. Grinspan. 2011. Fossil marsupial predators of

South America (Marsupialia, Borhyaenoidea): bite mechanics and

palaeobiological implications. Alcheringa: An Australasian Journal of

Palaeontology 35:377-387.

Blanco, R. E., W. W. Jones, and N. Milne. 2013. Is the extant southern short‐tailed

opossum a pigmy sabretooth predator? Journal of Zoology 291:100-110.

Bond, M., and C. M. Deschamps. 2010. The Mustersan age at Gran Barranca: a review;

pp. 255-263 in R. H. Madden, A. A. Carlini, M. G. Vucetich, and R. F. Kay (eds.),

The Paleontology of Gran Barranca. Evolution and Environmental Change

through the Middle Cenozoic of Patagonia. Cambridge University Press,

Cambridge.

283

Bond, M., and R. Pascual. 1983. Nuevos y elocuentes restos craneanos de Proborhyaena

gigantea Ameghino, 1897 (Marsupialia, Borhyaenidae, Proborhyaeninae) de la

Edad Deseadense. Un ejemplo de coevolución. Ameghiniana 20:47-60.

Bradham, J., J. J. Flynn, D. A. Croft, and A. R. Wyss. 2015. New notoungulates

(Notostylopidae and basal toxodontians) from the early Oligocene Tinguiririca

Fauna of the Andean Main Range, Central Chile. American Museum Novitates

3841:1-24.

Bryant, H. N. 1988. Delayed eruption of the deciduous upper canine in the sabertoothed

carnivore Barbourofelis lovei (Carnivora, Nimravidae). Journal of Vertebrate

Paleontology 8:295-306.

Bryant, H. N. 1996. Nimravidae; pp. 453-475 in D. R. Prothero, and R. J. Emry (eds.),

The Terrestrial Eocene-Oligocene Transition in North America. Cambridge

University Press, Cambridge.

Butler, R. F., L. G. Marshall, R. E. Drake, and G. H. Curtis. 1984. Magnetic polarity

stratigraphy and 40K-40Ar dating of late Miocene and early Pliocene continental

deposits, Catamarca Province, NW Argentina. The Journal of Geology 92:623-

636.

Cabrera, Á. 1927. Datos para el conocimiento de los dasiuroideos fósiles argentinos.

Revista del Museo de La Plata 30:271–315.

Carlini, A. A., M. R. Ciancio, J. J. Flynn, G. J. Scillato-Yané, and A. R. Wyss. 2009. The

phylogenetic and biostratigraphic significance of new armadillos (Mammalia,

Xenarthra, Dasypodidae, Euphractinae) from the Tinguirirican (Early Oligocene)

of Chile. Journal of Systematic Palaeontology 7:489-503.

284

Carneiro, L. M. 2018. A new species of Varalphadon (Mammalia, Metatheria,

Sparassodonta) from the upper Cenomanian of southern Utah, North America:

Phylogenetic and biogeographic insights. Cretaceous Research 84:88-96.

Carneiro, L. M., and É. V. Oliveira. 2017. Systematic affinities of the extinct metatherian

Eobrasilia coutoi Simpson, 1947, a South American early Eocene Stagodontidae:

Implications for "Eobrasiliinae". Revista Brasileira de Paleontologia 20:355-372.

Cassini, G. H., D. A. Flores, and S. F. Vizcaíno. 2012. Postnatal ontogenetic scaling of

nesodontine (Notoungulata, Toxodontidae) cranial morphology. Acta Zoologica

93:249-259.

Catena, A. M., D. I. Hembree, B. Z. Saylor, F. Anaya, and D. A. Croft. 2017. Paleosol

and ichnofossil evidence for significant Neotropical habitat variation during the

late middle Miocene (Serravallian). Palaeogeography, Palaeoclimatology,

Palaeoecology 487:381-398.

Charles, C., F. Solé, H. G. Rodrigues, and L. Viriot. 2013. Under pressure? Dental

adaptations to termitophagy and vermivory among mammals. Evolution 67:1792-

1804.

Charrier, R., J. J. Flynn, A. R. Wyss, L. Zapata, and C. C. Swisher, III. 1997.

Antecedentes bio y cronoestratigráficos de la Formación Coya-Machalí–Abanico,

entre los Ríos Maipo y Teno (33° 55’ y 35° 10’ L.S.), Cordillera Principal, Chile

Central. VIII Congreso Geológico Chileno, Actas I:465-469.

Chemisquy, M. A., and F. J. Prevosti. 2014. It takes more than large canines to be a

sabretooth predator. Mastozoología Neotropical 21:27-36.

285

Chornogubsky, L., A. N. Zimicz, F. J. Goin, J. C. Fernicola, P. Payrola, and M. Cárdenas.

2018. New Palaeogene metatherians from the Quebrada de Los Colorados

Formation at Los Cardones National Park (Salta Province, Argentina). Journal of

Systematic Palaeontology:1-17.

Churcher, C. S. 1985. Dental functional morphology in the marsupial sabre-tooth

Thylacosmilus atrox (Thylacosmilidae) compared to that of felid sabre-tooths.

Australian Mammalogy 8:201-220.

Ciancio, M. R., E. C. Vieytes, and A. A. Carlini. 2014. When xenarthrans had enamel:

insights on the evolution of their hypsodonty and paleontological support for

independent evolution in armadillos. Naturwissenschaften 101:715-725.

Cifelli, R. L. 1990. Cretaceous mammals of southern Utah. I. Marsupials from the

Kaiparowits Formation (Judithian). Journal of Vertebrate Paleontology 10:295-

319.

Cifelli, R. L., and C. Muizon, de. 1997. Dentition and jaw of Kokopellia juddi, a

primitive marsupial or near-marsupial from the medial Cretaceous of Utah.

Journal of Mammalian Evolution 4:241-258.

Cifelli, R. L., and C. D. Muizon. 1998a. Marsupial mammal from the Upper Cretaceous

North Horn Formation, central Utah. Journal of Paleontology 72:532-537.

Cifelli, R. L., and C. D. Muizon. 1998b. Tooth eruption and replacement patterns in early

marsupials. Comptes Rendus de l'Académie des Sciences. Series IIA. Earth and

Planetary Science 326:215-220.

286

Cifelli, R. L., T. B. Rowe, W. P. Luckett, J. Banta, R. Reyes, and R. I. Howes. 1996.

Fossil evidence for the origin of the marsupial pattern of tooth replacement.

Nature 379:715-718.

Cione, A. L., and E. P. Tonni. 2005: Bioestratigrafía basada en mamíferos del Cenozoico

superior de la provincia de Buenos Aires, Argentina. Paper presented at the

Geología y Recursos Minerales de la Provincia de Buenos Aires XVI Congresos

Geológica, La Plata, 2005.

Cladera, G., E. Ruigomez, E. O. Jaureguizar, M. Bond, and G. M. López. 2004.

Tafonomía de la Gran Hondonada (Formación Sarmiento, Edad-mamífero

Mustersense, Eoceno Medio) Chubut, Argentina. Ameghiniana 41:315–330.

Clemens, W. A., Jr. 1966. Fossil Mammals of the Type Lance Formation Wyoming Part

II. Marsupialia. University of California Publications in Geological Sciences

62:1-122.

Clemens, W. A. J. 1968. A mandible of Didelphodon vorax (Mammalia: Marsupialia).

Contributions in Science, Natural History Museum of Los Angeles County 133:1-

11.

Cohen, J. E. In press. Earliest divergence of stagodontid (Mammalia: Marsupialiformes)

feeding strategies from the Late Cretaceous (Turonian) of North America. Journal

of Mammalian Evolution:1-13.

Congreve, C. R., and J. C. Lamsdell. 2016. Implied weighting and its utility in

palaeontological datasets: a study using modelled phylogenetic matrices.

Palaeontology 59:447-462.

287

Couto-Ribeiro, G. d. 2010. Avaliação morfológica, taxonômica e cronológica dos

mamíferos fósseis da Formação Tremembé (Bacia de Taubate), Estado de São

Paulo, Brasil, pp. 112. Universidade de São Paulo, São Paulo.

Croft, D. A. 2006. Do marsupials make good predators? Insights from predator-prey

diversity ratios. Evolutionary Ecology Research 8:1193-1214.

Croft, D. A. 2007. The middle Miocene (Laventan) Quebrada Honda Fauna, southern

Bolivia, and a description of its notoungulates. Palaeontology 50:277-303.

Croft, D. A. 2013. What constitutes a fossil mammal community in the early Miocene

Santa Cruz Formation? Journal of Vertebrate Paleontology 33:401-409.

Croft, D. A. 2016. Horned Armadillos and Rafting Monkeys: the Fascinating Fossil

Mammals of South America. 320 pp. Indiana University Press, Bloomington.

Croft, D. A., F. Anaya, D. Auerbach, C. Garzione, and B. J. MacFadden. 2009. New data

on Miocene Neotropical provinciality from Cerdas, Bolivia. Journal of

Mammalian Evolution 16:175-198.

Croft, D. A., F. Anaya, A. M. Catena, M. R. Ciancio, and R. K. Engelman. 2013. New

species, local faunas, and paleoenvironmental data for the middle Miocene

Quebrada Honda Fauna, Bolivia. Journal of Vertebrate Paleontology, SVP

Program and Abstracts Book 2013:109.

Croft, D. A., M. Bond, J. J. Flynn, M. A. Reguero, and A. R. Wyss. 2003. Large

archaeohyracids (Typotheria, Notoungulata) from central Chile and Patagonia

including a revision of Archaeotypotherium. Fieldiana: Geology (New Series)

49:1-38.

288

Croft, D. A., A. A. Carlini, M. R. Ciancio, D. Brandoni, N. E. Drew, R. K. Engelman,

and F. Anaya. 2016. New mammal faunal data from Cerdas, Bolivia, a middle-

latitude Neotropical site that chronicles the end of the Middle Miocene Climatic

Optimum in South America. Journal of Vertebrate Paleontology 36:e1163574.

Croft, D. A., R. Charrier, J. J. Flynn, and A. W. Wyss. 2008a. Recent additions to

knowledge of Tertiary mammals from the Chilean Andes, pp. 91-96.

Croft, D. A., R. K. Engelman, T. Dolgushina, and G. Wesley. 2018. Diversity and

disparity of sparassodonts (Metatheria) reveal non-analogue nature of ancient

South American mammalian carnivore guilds. Proceedings of the Royal Society

B: Biological Sciences 285:1-9.

Croft, D. A., J. J. Flynn, and A. R. Wyss. 2004. Notoungulata and Litopterna of the early

Miocene Chucal Fauna, northern Chile. Fieldiana Geology, new series 50:1-52.

Croft, D. A., J. J. Flynn, and A. R. Wyss. 2007. A new basal glyptodontid and other

Xenarthra of the early Miocene Chucal Fauna, northern Chile. Journal of

Vertebrate Paleontology 27:781-797.

Croft, D. A., J. J. Flynn, and A. R. Wyss. 2008b. The Tinguiririca Fauna of Chile and the

early stages of "modernization" of South American mammal faunas. Arquivos do

Museu Nacional, Rio de Janeiro 66:191-211.

Davis, B. M. 2007. A revision of "pediomyid" marsupials from the Late Cretaceous of

North America. Acta Palaeontologica Polonica 52:217-256.

Davis, B. M., and R. L. Cifelli. 2011. Reappraisal of the Tribosphenidan Mammals from

the Trinity Group (Aptian—Albian) of Texas and Oklahoma. Acta

Palaeontologica Polonica 56:441-462.

289

Dawson, M. R., R. K. Stucky, L. Krishtalka, and C. C. Black. 1986. Machaeroides

simpsoni, new species, oldest known sabertooth creodont (Mammalia), of the Lost

Cabin Eocene. Rocky Mountain Geology 24:177-182.

Degrange, F. J., J. I. Noriega, and J. I. Areta. 2012. Diversity and paleobiology of

Santacrucian birds; pp. 138-155 in S. F. Vizcaíno, R. F. Kay, and M. S. Bargo

(eds.), Early Miocene Paleobiology in Patagonia: High-Latitude

Paleocommunities of the Santa Cruz Formation. Cambridge University Press,

Cambridge.

Díaz, M. M., and D. A. Flores. 2008. Early reproductive onset in four species of

Didelphimorphia in the Peruvian Amazon. Mammalia 72:126-130.

Dozo, M. T., M. Ciancio, P. J. Bouza, and G. Martínez. 2014. New association of

Paleogene Mammals in Eastern of Patagonia (Chubut Province, Argentina):

biochronological and paleobiogeographical implications. Andean Geology

41:224-247.

Dumont, E. R., and T. M. Bown. 1997. New caenolestoid marsupials; pp. 207-212 in R.

F. Kay, R. H. Madden, R. L. Cifelli, and J. J. Flynn (eds.), Vertebrate

Paleontology in the Neotropics: The Miocene Fauna of La Venta, Colombia.

Smithsonian Institution Press, Washington, D.C.

Dunn, R. E., R. H. Madden, M. J. Kohn, M. D. Schmitz, C. A. E. Strömberg, A. A.

Carlini, G. H. Ré, and J. Crowley. 2013. A new chronology for middle Eocene–

early Miocene South American Land Mammal Ages. Geological Society of

America Bulletin 125:539-555.

290

Echarri, S., M. D. Ercoli, M. A. Chemisquy, G. Turazzini, and F. J. Prevosti. 2017.

Mandible morphology and diet of the South American extinct metatherian

predators (Mammalia, Metatheria, Sparassodonta). Earth and Environmental

Science Transactions of the Royal Society of Edinburgh 106:277-288.

Emerson, S. B., and L. Radinsky. 1980. Functional analysis of sabertooth cranial

morphology. Paleobiology 6:295-312.

Engelman, R. K., F. Anaya, and D. A. Croft. 2015. New specimens of Acyon myctoderos

(Metatheria, Sparassodonta) from Quebrada Honda, Bolivia. Ameghiniana

52:204-225.

Engelman, R. K., F. Anaya, and D. A. Croft. 2017. New palaeothentid marsupials

(Paucituberculata) from the middle Miocene of Quebrada Honda, Bolivia, and

their implications for the palaeoecology, decline and extinction of the

Palaeothentoidea. Journal of Systematic Palaeontology 15:787-820.

Engelman, R. K., and D. A. Croft. 2014. A new species of small-bodied sparassodont

(Mammalia, Metatheria) from the middle Miocene locality of Quebrada Honda,

Bolivia. Journal of Vertebrate Paleontology 34:672–688.

Ercoli, M. D., F. J. Prevosti, and A. Álvarez. 2012. Form and function within a

phylogenetic framework: locomotory habits of extant predators and some

Miocene Sparassodonta (Metatheria). Zoological Journal of the Linnean Society

165:224-251.

Ercoli, M. D., F. J. Prevosti, and A. M. Forasiepi. 2014. The structure of the mammalian

predator guild in the Santa Cruz Formation (late early Miocene). Journal of

Mammalian Evolution 21:369-381.

291

Esteban, G., N. Nasif, and S. M. Georgieff. 2014. Cronobioestratigrafía del Mioceno

tardío – Plioceno temprano, Puerta de Corral Quemado y Villavil, provincia de

Catamarca, Argentina. Acta Geológica Lilloana 26:165-192.

Ewer, R. F. 1954. Some adaptive features in the dentition of hyaenas. Annals and

Magazine of Natural History 7:188-194.

Fiani, N. 2015. Dental Radiology; pp. 205-222 in L. Vogeinest, and G. Allan (eds.),

Radiology of Australian Mammals. CSIRO Publishing, Collingwood.

Flynn, J. J., R. Charrier, D. A. Croft, P. B. Gans, T. M. Herriott, J. A. Wertheim, and A.

R. Wyss. 2008. Chronologic implications of new Miocene mammals from the

Cura-Mallín and Trapa Trapa formations, Laguna del Laja area, south central

Chile. Journal of South American Earth Sciences 26:412-423.

Flynn, J. J., R. Charrier, D. A. Croft, R. H. Hitz, and A. R. Wyss. 2003a: The Abanico

Formation of the Chilean Andes: An exceptional Eocene-Miocene record of South

American mammal evolution. Paper presented at the 63rd Annual Meeting of the

Society of Vertebrate Paleontology, St. Paul, Minnesota, 2003a.

Flynn, J. J., R. Charrier, D. A. Croft, and A. R. Wyss. 2012. Cenozoic Andean Faunas:

Shedding New Light on South American Mammal Evolution, Biogeography,

Environments, and Tectonics; pp. 51-75 in B. D. Patterson, and L. P. Costa (eds.),

Bones, Clones, and Biomes. The History and Geography of Recent Neotropical

Mammals. University of Chicago Press, Chicago.

Flynn, J. J., J. Guerrero, and C. C. Swisher, III. 1997. Geochronology of the Honda

Group; pp. 44-59 in R. F. Kay, R. H. Madden, R. L. Cifelli, and J. J. Flynn (eds.),

292

Vertebrate Paleontology in the Neotropics: The Miocene Fauna of La Venta,

Colombia. Smithsonian Institution Press, Washington, DC.

Flynn, J. J., M. J. Novacek, H. E. Dodson, D. Frassinetti, C. McKenna, M. A. Norell, K.

E. Sears, C. C. Swisher, III, and A. R. Wyss. 2002. A new fossil mammal

assemblage from the southern Chilean Andes: implications for geology,

geochronology, and tectonics. Journal of South American Earth Sciences 15:285-

302.

Flynn, J. J., and C. C. Swisher, III. 1995. Cenozoic South American Land Mammal Ages:

correlation to global geochronologies; pp. 317-333 in W. A. Berggren, D. V.

Kent, M.-P. Aubry, and J. Hardenbol (eds.), Geochronology, Time Scales, and

Global Stratigraphic Correlation. SEPM (Society for Sedimentary Geology)

Special Publication 54.

Flynn, J. J., and A. R. Wyss. 1998. Recent advances in South American mammalian

paleontology. Trends in Ecology and Evolution 13:449-454.

Flynn, J. J., and A. R. Wyss. 1999. New marsupials from the Eocene-Oligocene transition

of the Andean Main Range, Chile. Journal of Vertebrate Paleontology 19:533-

549.

Flynn, J. J., and A. R. Wyss. 2004. A polydolopine marsupial skull from the Cachapoal

Valley, Andean Main Range, Chile. Bulletin of the American Museum of Natural

History 285:80-92.

Flynn, J. J., A. R. Wyss, D. A. Croft, and R. Charrier. 2003b. The Tinguiririca Fauna,

Chile: biochronology, paleoecology, biogeography, and a new earliest Oligocene

293

South American Land Mammal "Age". Palaeogeography, Palaeoclimatology,

Palaeoecology 195:229-259.

Forasiepi, A., F. J. Goin, and A. A. Tauber. 2004. Las especies de Arctodictis Mercerat,

1891 (Metatheria, Borhyaenidae), grandes carnívoros del Mioceno del América

del Sur. Revista Española de Paleontología 19:1-22.

Forasiepi, A. M. 2009. Osteology of Arctodictis sinclairi (Mammalia, Metatheria,

Sparassodonta) and phylogeny of Cenozoic metatherian carnivores from South

America. Monografías del Museo Argentino de Ciencias Naturales 6:1–174.

Forasiepi, A. M., M. J. Babot, and N. Zimicz. 2015. Australohyaena antiqua (Mammalia,

Metatheria, Sparassodonta), a large predator from the late Oligocene of Patagonia.

Journal of Systematic Palaeontology 13:505-523.

Forasiepi, A. M., and A. A. Carlini. 2010. A new thylacosmilid (Mammalia, Metatheria,

Sparassodonta) from the Miocene of Patagonia, Argentina. Zootaxa 2552:55-68.

Forasiepi, A. M., F. Goin, and A. G. Martinelli. 2009. Contribution to the knowledge of

the Sparassocynidae (Mammalia, Metatheria, Didelphoidea), with comments on

the age of the Aisol Formation (Neogene), Mendoza Province, Argentina. Journal

of Vertebrate Paleontology 29:1252-1263.

Forasiepi, A. M., F. J. Goin, and V. di Martino. 2003. Una nueva especie de Lycopsis

(Metatheria, Prothylacyninae) de la Formación Arroyo Chasicó (Mioceno Tardío)

de la provincia de Buenos Aires. Ameghiniana 40:249-253.

Forasiepi, A. M., A. G. Martinelli, and F. J. Goin. 2007. Revisión taxonómica de

Parahyaenodon argentinus Ameghino y sus implicancias en el conocimiento de

294

los grandes mamíferos carnívoros del Mio-Plioceno de América de Sur.

Ameghiniana 44:143-159.

Forasiepi, A. M., and G. W. Rougier. 2009. Additional data on early Paleocene

metatherians (Mammalia) from Punta Peligro (Salamanca Formation, Argentina):

comments based on petrosal morphology. Journal of Zoological Systematics and

Evolutionary Research 47:391-398.

Forasiepi, A. M., and M. R. Sánchez-Villagra. 2014. Heterochrony, dental ontogenetic

diversity, and the circumvention of constraints in marsupial mammals and extinct

relatives. Paleobiology 40:222-237.

Forasiepi, A. M., M. R. Sánchez-Villagra, F. J. Goin, M. Takai, N. Shigehara, and R. F.

Kay. 2006. A new species of Hathliacynidae (Metatheria, Sparassodonta) from

the middle Miocene of Quebrada Honda, Bolivia. Journal of Vertebrate

Paleontology 26:670-684.

Forasiepi, A. M., M. R. Sánchez-Villagra, T. Schmelzle, S. Ladevèze, and R. F. Kay.

2014. An exceptionally well-preserved skeleton of Palaeothentes from the Early

Miocene of Patagonia, Argentina: new insights into the anatomy of extinct

paucituberculatan marsupials. Swiss Journal of Palaeontology.

Fortelius, M. 1990. Problems with using fossil teeth to estimate body sizes of extinct

mammals; pp. 207-227 in J. Damuth, and B. J. MacFadden (eds.), Body Size in

Mammalian Paleobiology: Estimation and Biological Implications. Cambridge

University Press, Cambridge.

Fox, R. C. 1983. Notes on the North American Tertiary marsupials Herpetotherium and

Peradectes. Canadian Journal of Earth Sciences 20:1565-1578.

295

Fox, R. C., and B. G. Naylor. 1995. The relationships of the Stagodontidae, primitive

North American Late Cretaceous mammals; pp. 247-250 in A. Sun, and Y. Wang

(eds.), Sixth Symposium on Mesozoic Terrestrial Ecosystems and Biota. China

Ocean Press, Beijing.

Fox, R. C., and B. G. Naylor. 2006. Stagodontid marsupials from the Late Cretaceous of

Canada and their systematic and functional implications. Acta Palaeontologica

Polonica 51:13-36.

Gaudin, T. J., and D. A. Croft. 2015. Paleogene Xenarthra and the evolution of South

American mammals. Journal of Mammalogy 96:622-634.

Gaudin, T. J., and J. R. Wible. 2004. On the cranial osteology of the yellow armadillo

Euphractus sexcinctus (Dasypodidae, Xenarthra, Placentalia). Annals of Carnegie

Museum 73:117-196.

Gazin, C. L. 1946. Machaeroides eothen Matthew, the saber-tooth creodont of the

Bridger Eocene. Proceedings of the United Stated National Museum 96:335-347.

Gelfo, J. N., L. Chornogubsky, G. M. López, F. J. Goin, and M. Ciancio. 2010:

Biochronological relationships of the mammal fauna from the Paleogene of Las

Violetas, Chubut Province, Argentina. Paper presented at the X Congreso

Argentino de Paleontología y Bioestratigrafía-VII Congreso Latinoamericano de

Paleontología, La Plata, 2010.

Geraads, D., and E. Güleç. 1997. Relationships of Barbourofelis piveteaui (Ozansoy,

1965), a late Miocene nimravid (Carnivora, Mammalia) from central Turkey.

Journal of Vertebrate Paleontology 17:370-375.

296

Godthelp, H., S. Wroe, and M. Archer. 1999. A new marsupial from the early Eocene

Tingamarra local fauna of Murgon, southeastern Australia: a prototypical

Australian marsupial? Journal of Mammalian Evolution 6:289-313.

Goin, F. J. 1997a. Sobre la edad y afinidades de Zygolestes paranensis Ameghino, 1898

(Marsupialia: Didelphidae: Marmosinae). Neotrópica 43:15-19.

Goin, F. J. 1997b. New clues for understanding Neogene marsupial radiations; pp. 187-

206 in R. F. Kay, R. H. Madden, R. L. Cifelli, and J. J. Flynn (eds.), Vertebrate

Paleontology in the Neotropics: the Miocene Fauna of La Venta, Colombia.

Smithsonian Institution Press, Washington, DC.

Goin, F. J. 2003. Early marsupial radiations in South America; pp. 30-42 in M. Jones, C.

Dickman, and M. Archer (eds.), Predators with Pouches: the Biology of

Marsupial Carnivores. CSIRO Publishing, Collingwood.

Goin, F. J., A. Abello, E. Bellosi, R. Kay, R. Madden, and A. Carlini. 2007. Los

Metatheria sudamericanos de comienzos del Neógeno (Mioceno Temprano, Edad-

mamífero Colhuehuapense). Parte I: Introducción, Didelphimorphia y

Sparassodonta. Ameghiniana 44:29-71.

Goin, F. J., and M. A. Abello. 2013. Los Metatheria sudamericanos de comienzos del

Neógeno (Mioceno temprano, edad-mamífero Colhuehuapense. Parte II:

Microbiotheria y Polydolopimorphia. Ameghiniana 50:51-78.

Goin, F. J., M. A. Abello, and L. Chornogubsky. 2010. Middle Tertiary marsupials from

central Patagonia (early Oligocene of Gran Barranca): understanding South

America's Grande Coupure; pp. 69–105 in R. H. Madden, A. A. Carlini, M. G.

Vucetich, and R. F. Kay (eds.), The Paleontology of Gran Barranca. Evolution

297

and Environmental Change through the Middle Cenozoic of Patagonia.

Cambridge University Press, Cambridge.

Goin, F. J., A. Candela, and G. López. 1998. Middle Eocene marsupials from

Antofagasta de la Sierra, northwestern Argentina. Geobios 31:75-85.

Goin, F. J., and A. M. Candela. 2004. New Paleogene marsupials from the Amazon Basin

of eastern Perú. Science Series, Los Angeles Museum of Natural History 40:15–

60.

Goin, F. J., C. Montalvo, and G. Visconti. 2000. Los marsupiales (Mammalia) del

Mioceno superior de la Formación Cerro Azul (Provincia de La Pampa,

Argentina). Estudios Geológicos 56: 101-126.

Goin, F. J., R. M. Palma, R. Pascual, and J. E. Powell. 1986. Persistencia de un primitivo

Borhyaenidae (Mammalia, Marsupialia) en el Eoceno temprano de Salta (Fm.

Lumbrera, Argentina). Aspectos geologicos y paleoambientales relacionados.

Ameghiniana 23:47–56.

Goin, F. J., and U. F. J. Pardiñas. 1996. Revision de las especies del genero

Hyperdidelphys Ameghino, 1904 (Mammalia, Marsupialia, Didelphidae). Su

significacion filogenetica, estratigrafica y adaptativa en el Neogeno del Cono Sur

Sudamericano. Estudios Geológicos 52:327-359.

Goin, F. J., and R. Pascual. 1987. News on the biology and taxonomy of the marsupials

Thylacosmilidae (late Tertiary of Argentina). Anales de la Academia Nacional de

Ciencias Exactas, Físicas y Naturales de Buenos Aires 39:219-246.

298

Goin, F. J., M. O. Woodburne, A. N. Zimicz, G. M. Martin, and L. Chornogubsky. 2016.

A Brief History of South American Metatherians: Evolutionary Contexts and

Intercontinental Dispersals. 237 pp. Springer, New York.

Goloboff, P. A., J. A. Farris, and K. C. Nixon. 2008. TNT, a free program for

phylogenetic analysis. Cladistics 24:774-786.

Goloboff, P. A., A. Torres, and J. S. Arias. Weighted parsimony outperforms other

methods of phylogenetic inference under models appropriate for morphology.

Cladistics:n/a-n/a.

Gordon, C. L. 2003. A first look at estimating body size in dentally conservative

marsupials. Journal of Mammalian Evolution 10:1-21.

Gosselin-Ildari, A. 2006. Dromiciops gliroides, Vol. 2018. Digital Morphology.

Grohé, C., G. E. Rössner, and M. Spaulding. 2015. Reality or fantasy? Assessing the

condition of fossil specimens with CT data: In Journal of Vertebrate

Paleontology, SVP Program and Abstracts Book, 2015, A. MacKenzie, E.

Maxwell, and J. Miller-Camp eds, pp. 136, Dallas, Texas.

Hitz, R., J. J. Flynn, and A. R. Wyss. 2006. New basal Interatheriidae (Typotheria,

Notoungulata, Mammalia) from the Paleogene of central Chile. American

Museum Novitates 3520:1-32.

Horovitz, I., S. Ladevèze, C. Argot, T. E. Macrini, T. Martin, J. J. Hooker, C. Kurz, C.

Muizon, de, and M. R. Sánchez-Villagra. 2008a. The anatomy of Herpetotherium

cf. fugax Cope, 1873, a metatherian from the Oligocene of North America.

Palaeontographica Abteilung A 284:109-141.

299

Horovitz, I., S. Ladevèze, C. Argot, T. E. Macrini, T. Martin, J. J. Hooker, C. Kurz, C.

Muizon, de, and M. R. Sánchez-Villagra. 2008b. Herpetotherium cf. fugax, Vol.

2018. Digital Morphology.

Horovitz, I., T. Martin, J. Bloch, S. Ladevèze, C. Kurz, and M. R. Sánchez-Villagra.

2009. Cranial anatomy of the earliest marsupials and the origin of opossums.

PLoS ONE 4:e8278.

Hunt, R. M. J., and R. M. Joeckel. 1988. Mammalian biozones in nonmarine rocks of the

North American continental interior: biostratigraphic resolution within the "cat

gap.". Rocky Mountain Section, Geological Society of America Abstracts with

Program 20:421.

Huxley, T. H. 1880. On the application of the laws of evolution to the arrangement of the

Vertebrata, and more particularly of the Mammalia. Proceedings of the Zoological

Society of London 43:649-662.

Ith-Hansen, K., and I. Kjær. 2000. Persistence of deciduous molars in subjects with

agenesis of the second premolars. European Journal of Orthodontics 22:239-243.

Johanson, Z., P. Ahlberg, and A. Ritchie. 2003. The braincase and palate of the

tetrapodomorph sarcopterygian Mandageria fairfaxi: Morphological variability

near the fish–tetrapod transition. Palaeontology 46:271-293.

Jones, M. E. 2003. Convergence in ecomorphology and guild structure among marsupial

and placental carnivores; pp. 285-296 in M. Jones, C. Dickman, and M. Archer

(eds.), Predators with Pouches: The Biology of Carnivorous Marsupials. CSIRO

Publishing, Collingwood.

300

Jones, M. E., and L. A. Barmuta. 1998. Diet overlap and relative abundance of sympatric

dasyurid carnivores: a hypothesis of competition. Journal of Animal Ecology

67:410-421.

Julien-Laferrière, D., and M. Atramentowicz. 1990. Feeding and reproduction of three

didelphid marsupials in two Neotropical forests (French Guiana). Biotropica

22:404-415.

Kealy, S., and R. Beck. 2017. Total evidence phylogeny and evolutionary timescale for

Australian faunivorous marsupials (Dasyuromorphia). BMC Evol Biol 17:240.

Kelly, T. S., P. C. Murphey, and S. L. Walsh. 2012. New records of small mammals from

the middle Eocene Duchesne River Formation, Utah, and their implications for

the Uintan-Duchesnean North American Land Mammal Age transition.

Paludicola 8:208-251.

Kielan-Jaworowska, Z. 1975. Evolution of the therian mammals in the late Cretaceous of

Asia. Part I. Deltatheridiidae. Palaeontologica Polonica 33:103-131.

Koenigswald, W., von. 2011. Diversity of hypsodont teeth in mammalian dentitions –

construction and classification. Palaeontographica Abteilung A 294:63-94.

Koenigswald, W., von, and F. J. Goin. 2000. Enamel differentiation in South American

marsupials and a comparison of placental and marsupial enamel.

Palaeontographica Abteilung A 255:129-168.

Korth, W. W. 1994. Middle Tertiary marsupials (Mammalia) from North America.

Journal of Paleontology 68:376-397.

Krajewski, C., A. C. Driskell, P. R. Baverstock, and M. J. Braun. 1992. Phylogenetic

relationships of the thylacine (Mammalia: Thylacinidae) among dasyuroid

301

marsupials : evidence from cytochrome b DNA sequences. Proceedings of the

Royal Society of London. Series B: Biological Sciences 250:19.

Krause, D. W., W. C. Clyde, M. Ibañez-Mejía, M. D. Schmitz, T. Barnum, E. S. Bellosi,

and P. Wilf. 2017. New age constraints for early Paleogene strata of central

Patagonia, Argentina: Implications for the timing of South American Land

Mammal Ages. GSA Bulletin 129:886-903.

Kruuk, H. 1976. Feeding and social behavior of the striped hyaena (Hyaena vulgaris

Desmarest). East African Wildlife Journal 14:91-111.

Kupczik, K., and D. D. Stynder. 2012. Tooth root morphology as an indicator for dietary

specialization in carnivores (Mammalia: Carnivora). Biological Journal of the

Linnean Society 105:456-471.

Ladevèze, S., C. Muizon, de, R. M. D. Beck, D. Germain, and R. Cespedes-Paz. 2011a.

Earliest evidence of mammalian social behaviour in the basal Tertiary of Bolivia.

Nature 474:83-86.

Ladevèze, S., and C. d. Muizon. 2007. The auditory region of early Paleocene

Pucadelphydae (Mammalia, Metatheria) from Tiupampa, Bolivia, with

phylogenetic implications. Palaeontology 50:1123-1154.

Ladevèze, S., and C. d. Muizon. 2010. Evidence of early evolution of Australidelphia

(Metatheria, Mammalia) in South America: phylogenetic relationships of the

metatherians from the late Palaeocene of Itaboraí (Brazil) based on teeth and

petrosal bones. Zoological Journal of the Linnean Society 159:746-784.

302

Ladevèze, S., C. d. Muizon, R. M. D. Beck, D. Germain, and R. Cespedes-Paz. 2011b.

Earliest evidence of mammalian social behaviour in the basal Tertiary of Bolivia.

Nature 474:83-86.

Larson, A. 1994. The comparison of morphological and molecular data in phylogenetic

systematics; pp. 371-390 in B. S. B. Schierwater, B. Streit, G. P. Wagner, and R.

DeSalle (eds.), Molecular Ecology and Evolution: Approaches and Applications.

Birkhäuser, Basel.

Lillegraven, J. A. 1969. Latest Cretaceous mammals of upper part of Edmonton

Formation of Alberta, Canada, and review of marsupial-placental dichotomy in

mammalian evolution. The University of Kansas Paleontological Contributions

50:1-122.

Linnaeus, C. 1758. Systema Naturae per Regna Tria Naturae, Secundum Classes,

Ordines, Genera, Species, cum Characteribus, Differentiis, Synonymis, Locis. 824

pp. Laurentiii Salvii, Stockholm.

López-Aguirre, C., M. Archer, S. J. Hand, and S. W. Laffan. 2017. Extinction of South

American sparassodontans (Metatheria): environmental fluctuations or complex

ecological processes? Palaeontology 60:91-115.

Lorente, M., L. Chornogubsky, and F. J. Goin. 2016. Presencia de un posible boriénido

(Mammalia, Metatheria) en el Eoceno temprano medio de la localidad de La

Barda (Provincia de Chubut, Argentina): In XXX Jornadas Argentinas de

Paleontología de Vertebrados, pp. 84, Buenos Aires, Argentina.

303

Luckett, W. P., and P. A. Wooley. 1996. Ontogeny and homology of the dentition in

dasyurid marsupials: development in Sminthopsis virginiae. Journal of

Mammalian Evolution 3:327-364.

Luo, Z.-X., Q. Ji, J. R. Wible, and C.-X. Yuan. 2003. An Early Cretaceous Tribosphenic

Mammal and Metatherian Evolution. Science 302:1934-1940.

Macedo, J., D. Loretto, and M. S. C. Mello. 2007. História natural dos mamíferos de uma

área perturbada do Parque Nacional da Serra dos Órgãos; pp. 165-181 in C.

Cronemberguer, and E. V. Castro (eds.), Ciência e Conservação na Serra dos

Órgãos. IBAMA, Brasilia.

MacFadden, B. J., F. Anaya, H. Perez, C. W. Naeser, P. K. Zeitler, and K. E. Campbell,

Jr. 1990. Late Cenozoic paleomagnetism and chronology of Andean basins of

Bolivia: evidence for possible oroclinal bending. Journal of Geology 98:541-555.

MacFadden, B. J., and R. G. Wolff. 1981. Geological investigations of Late Cenozoic

vertebrate-bearing deposits in southern Bolivia. Anais do II Congresso Latino-

Americano de Paleontología 2:765-778.

Macrini, T. E. 2001. Monodelphis domestica, Vol. 2018. Digital Morphology.

Macrini, T. E. 2005a. Dasyurus hallucatus, Vol. 2018. Digital Morphology.

Macrini, T. E. 2005b. Didelphis virginiana, Vol. 2018. Digital Morphology.

Macrini, T. E. 2009. Sminthopsis crassicaudata, Vol. 2018. Digital Morphology.

Madden, R. H. 2015. Hypsodonty in mammals: Evolution, geomorphology and the role

of earth surface processes. Cambridge University Press, Cambridge.

Madden, R. H., R. F. Kay, M. a. G. Vucetich, and A. A. Carlini. 2010. Gran Barranca: a

23-million-year record of middle Cenozoic faunal evolution in Patagonia; pp.

304

423-439 in R. H. Madden, A. A. Carlini, M. G. Vucetich, and R. F. Kay (eds.),

The Paleontology of Gran Barranca. Evolution and Environmental Change

through the Middle Cenozoic of Patagonia. Cambridge University Press,

Cambridge.

Maddison, W. P., and D. R. Maddison. 2008. Mesquite: a modular system for

evolutionary analysis, Vol. 2.75.

Madzia, D., and A. Cau. 2017. Inferring ‘weak spots’ in phylogenetic trees: application to

mosasauroid nomenclature. PeerJ 5:e3782.

Maga, A. M., and R. M. D. Beck. 2017. Skeleton of an unusual, cat-sized marsupial

relative (Metatheria: Marsupialiformes) from the middle Eocene (Lutetian: 44-43

million years ago) of Turkey. PLoS ONE 12:e0181712.

Marshall, L. G. 1976a. Evolution of the Thylacosmilidae, extinct saber-tooth marsupials

of South America. Paleobios 23:1-30.

Marshall, L. G. 1976b. New didelphine marsupials from the La Venta Fauna (Miocene)

of Colombia, South America. Journal of Paleontology 50:402-418.

Marshall, L. G. 1977a. Evolution of the carnivorous adaptive zone in South America; pp.

709-721 in M. K. Hecht, P. C. Goody, and B. M. Hecht (eds.), Major Patterns in

Vertebrate Evolution. Plenum Press, New York.

Marshall, L. G. 1977b. New species of Lycopsis (Borhyaenidae: Marsupialia) from La

Venta Fauna (late Miocene) of Colombia, South America. Journal of

Paleontology 51:633–642.

Marshall, L. G. 1978a. Dromiciops australis. Mammalian Species 99:1-5.

305

Marshall, L. G. 1978b. Evolution of the Borhyaenidae, extinct South American

predaceous marsupials. University of California Publications in Geological

Sciences 117:1–89.

Marshall, L. G. 1979. Review of the Prothylacyninae, an extinct subfamily of South

American "dog-like" marsupials. Fieldiana Geology, new series 3:1-49.

Marshall, L. G. 1981. Review of the Hathlyacyninae, an extinct subfamily of South

American "dog-like" marsupials. Fieldiana Geology, new series 7:1–120.

Marshall, L. G. 1982a. A new genus of Caroloameghiniinae (Marsupialia: Didelphoidea:

Didelphidae) from the Paleocene of Brazil. Journal of Mammalogy 63:709-716.

Marshall, L. G. 1982b. Systematics of the South American marsupial family

Microbiotheriidae. Fieldiana Geology, new series 10:1-75.

Marshall, L. G. 1987. Systematics of Itaboraian (middle Paleocene) age “opossum-like”

marsupials from the limestone Quarry at São José de Itaboraí, Brazil; pp. 91-160

in M. Archer (ed.), Possums and Opossums: Studies in Evolution, Vol. I. Surrey

Beatty and Sons in association with the Royal Zoological Society of New South

Wales, Chipping Norton.

Marshall, L. G. 1990. Fossil Marsupialia from the type Friasian Land Mammal Age

(Miocene), Alto Rio Cisnes, Aisen, Chile. Revista Geológica de Chile 17:19-55.

Marshall, L. G., J. A. Case, and M. O. Woodburne. 1990. Phylogenetic relationships of

the families of marsupials; pp. 433-505 in H. H. Genoways (ed.), Current

Mammalogy. Volume 2. Plenum Press, New York.

Marshall, L. G., and Z. Kielan-Jaworowska. 1992. Relationships of the dog-like

marsupials, deltatheroidans and early tribosphenic mammals. Lethaia 25:361-374.

306

Marshall, L. G., and C. Muizon, de. 1995. Part II. The skull; pp. 21-90 in C. d. Muizon

(ed.), Pucadelphys andinus (Marsupialia, Mammalia) from the early Paleocene of

Bolivia. Mémoires du Muséum national d’Histoire naturelle, Paris.

Marshall, L. G., and C. d. Muizon. 1988. The dawn of the Age of Mammals in South

America. National Geographic Research 4:23–55.

Matthew, W. D. 1907. The Relationships of the ‘Sparassodonta’. Geological Magazine

4:531-535.

Matthew, W. D. 1916. A marsupial from the Belly River Cretaceous. With critical

observations upon the affinities of the Cretaceous mammals. Bulletin of the

American Museum of Natural History 37:477-500.

McGrath, A. J., F. Anaya, and D. A. Croft. 2018. Two new macraucheniids (Mammalia:

Litopterna) from the late middle Miocene (Laventan South American Land

Mammal Age) of Quebrada Honda, Bolivia. Journal of Vertebrate

Paleontology:e1461632.

McKenna, M. C., A. R. Wyss, and J. J. Flynn. 2006. Paleogene pseudoglyptodont

xenarthrans from central Chile and Argentine Patagonia. American Museum

Novitates 3536:1-18.

Meachen-Samuels, J. A. 2012. Morphological convergence of the prey-killing arsenal of

sabertooth predators. Paleobiology 38:1-14.

Mellett, J. S. 1969. Carnassial rotation in a fossil carnivore. American Midland Naturalist

82:287-289.

Miller, W., D. I. Drautz, J. E. Janecka, A. M. Lesk, A. Ratan, L. P. Tomsho, M. Packard,

Y. Zhang, L. R. McClellan, J. Qi, F. Zhao, M. T. P. Gilbert, L. Dalén, J. L.

307

Arsuaga, P. G. P. Ericson, D. H. Huson, K. M. Helgen, W. J. Murphy, A.

Götherström, and S. C. Schuster. 2009. The mitochondrial genome sequence of

the Tasmanian tiger (Thylacinus cynocephalus). Genome Research 19:213-220.

Mills, M. G. L. 2015. Living near the edge: a review of the ecological relationships

between large carnivores in the arid Kalahari. African Journal of Wildlife

Research 45:127-137.

Mones, A. 1982. An equivocal nomenclature: What means hypsodonty? Paläontologische

Zeitschrift 56:107-111.

Mones, A., and M. Ubilla. 1978. La edad Deseadense (Oligoceno inferior) de la

Formación Fray Bentos y su contenido paleontologico, con especial referencia a

la presencia de Proborhyaena cf. gigantea Ameghino (Marsupialia:

Borhyaenidae) en el Uruguay. Nota preliminar. Comunicaciones Paleontologicas

del Museo de Historia Natural de Montevideo 7:151-158.

Montellano, M. 1988. Alphadon halleyi (Didelphidae, Marsupialia) from the Two

Medicine Formation (Late Cretaceous, Judithian) of Montana. Journal of

Vertebrate Paleontology 8:378-382.

Morales, J., M. J. Salesa, M. Pickford, and D. Soria. 2001. A new tribe, new genus and

two new species of Barbourofelinae (Felidae, Carnivora, Mammalia) from the

Early Miocene of East Africa and Spain. Earth and Environmental Science

Transactions of the Royal Society of Edinburgh 92:97-102.

Morlo, M. 2006. New remains of Barbourofelidae (Mammalia, Carnivora) from the

Miocene of Southern Germany: implications for the history of barbourofelid

migrations. Beiträge zur Paläontologie 30:339-346.

308

Morlo, M., S. Peigné, and D. Nagel. 2004. A new species of Prosansanosmilus:

implications for the systematic relationships of the family Barbourofelidae new

rank (Carnivora, Mammalia). Zoological Journal of the Linnean Society 140:43-

61.

Muirhead, J., and S. Wroe. 1998. A new genus and species, Badjcinus turnbulli

(Thylacinidae: Marsupialia), from the late Oligocene of Riversleigh, northern

Australia, and an investigation of thylacinid phylogeny. Journal of Vertebrate

Paleontology 18:612-626.

Muizon, C., de. 1994. A new carnivorous marsupial from the Palaeocene of Bolivia and

the problem of marsupial monophyly. Nature 370:208-211.

Muizon, C., de. 1998. Mayulestes ferox, a borhyaenoid (Metatheria, Mammalia) from the

early Palaeocene of Bolivia. phylogenetic and paleobiologic implications.

Geodiversitas 20:19-142.

Muizon, C., de, and R. L. Cifelli. 2001. A new basal "didelphoid" (Marsupialia,

Mammalia) from the early Paleocene of Tiupampa (Bolivia). Journal of

Vertebrate Paleontology 21:87-97.

Muizon, C., de, R. L. Cifelli, and R. Céspedes Paz. 1997. The origin of the dog-like

borhyaenoid marsupials of South America. Nature 389:486-489.

Muizon, C. d. 1999. Marsupials skulls from the Deseadan (late Oligocene) of Bolivia and

phylogenetic analysis of the Borhyaenoidea (Marsupialia, Mammalia). Geobios

32:483-509.

Muizon, C. d., and C. Argot. 2003. Comparative anatomy of the Tiupampa

didelphimorphs; an approach to locomotory habits of early marsupials; pp. 43-62

309

in M. Jones, C. Dickman, and M. Archer (eds.), Predators with Pouches: the

Biology of Marsupial Carnivores. CSIRO Publishing, Collingwood.

Muizon, C. d., and B. Lange-Badré. 1997. Carnivorous dental adaptations in tribosphenic

mammals and phylogenetic reconstruction. Lethaia 30:353-366.

Murray, P., and D. Megirian. 2000. Two new genera and three new species of

Thylacinidae (Marsupialia) from the Miocene of the Northern Territory,

Australia. The Beagle, Records of the Museums and Art Galleries of the Northern

Territory 16:145–162.

Murray, P. F., and D. Megirian. 2006. Cranial morphology of the Miocene thylacinid

Mutpuracinus archibaldi (Thylacinidae, Marsupialia) and relationships within the

Dasyuromorphia. Alcheringa: An Australasian Journal of Palaeontology 30:229-

276.

Myers, T. J. 2001. Prediction of marsupial body mass. Australian Journal of Zoology

49:99-118.

Naples, V. L., L. D. Martin, and J. P. Babiarz. 2011. The Other Saber-Tooths: Scimitar-

Tooth Cats of the Western Hemisphere. John Hopkins University Press,

Baltimore.

Nelson, H., and R. Jurmain. 1982. Introduction to Physical Anthropology. West

Publishing Co., St. Paul.

O'Leary, M. A., J. I. Bloch, J. J. Flynn, T. J. Gaudin, A. Giallombardo, N. P. Giannini, S.

L. Goldberg, B. P. Kraatz, Z.-X. Luo, J. Meng, X. Ni, M. J. Novacek, F. A. Perini,

Z. S. Randall, G. W. Rougier, E. J. Sargis, M. T. Silcox, N. B. Simmons, M.

Spaulding, P. M. Velazco, M. Weksler, J. R. Wible, and A. L. Cirranello. 2013.

310

The placental mammal ancestor and the post–K-Pg radiation of placentals.

Science 339:662-667.

Object Research Systems Inc. 2017. Dragonfly, Vol. 2.0. Object Research Systems

(ORS) Inc., Montreal.

Ogg, J. G. 2012. Geomagnetic polarity time scale; pp. 85-113 in F. M. Gradstein, J. G.

Ogg, M. Schmitz, and G. Ogg (eds.), The Geologic Time Scale. Elsevier, New

York.

Oiso, Y. 1991. New land mammal locality of middle Miocene (Colloncuran) age from

Nazareno, southern Bolivia; pp. 653-672 in R. Suárez-Soruco (ed.), Fósiles y

Facies de Bolivia—Vol. I: Vertebrados. Yacimientos Petrolíferos Fiscales

Bolivianos, Santa Cruz, Bolivia.

Oliveira, É. V., and F. J. Goin. 2011. A reassessment of bunodont metatherians from the

Paleogene of Itaboraí (Brazil): Systematics and age of the Itaboraian SALMA.

Revista Brasileira de Paleontologia 14:105-136.

Owens, M. J., and D. D. Owens. 1978. Feeding ecology and its influence on social

organization in Brown hyenas (Hyaena brunnea, Thunberg) of the Central

Kalahari Desert. African Journal of Ecology 16.

Palmqvist, P., B. Martínez-Navarro, J. A. Pérez-Claros, V. Torregrosa, B. Figueirido, J.

M. Jiménez-Arenas, M. Patrocinio Espigares, S. Ros-Montoya, and M. De Renzi.

2011. The giant hyena Pachycrocuta brevirostris: Modelling the bone-cracking

behavior of an extinct carnivore. Quaternary International 243:61-79.

Pascual, R., and A. Bocchino. 1963. Un nuevo Borhyaeninae (Marsupialia) del Plioceno

medio de Hidalgo (La Pampa). Ameghiniana 3:97-107.

311

Pascual, R., F. J. Goin, and A. A. Carlini. 1994. New data on the Groeberiidae: Unique

late Eocene-early Oligocene South American marsupials. Journal of Vertebrate

Paleontology 14:247-259.

Pascual, R., and E. Ortiz Jaureguizar. 1990. Evolving climates and mammal faunas in

Cenozoic South America. Journal of Human Evolution 19:23-60.

Patterson, B., and L. G. Marshall. 1978. The Deseadan, early Oligocene, Marsupialia of

South America. Fieldiana Geology 41:37-100.

Patterson, B., and R. Pascual. 1968. The fossil mammal fauna of South America.

Quarterly Review of Biology 43:409-451.

Paula Couto, C. d. 1952. Fossil mammals from the beginning of the Cenozoic in Brazil.

Marsupialia: Didelphidae. American Museum Novitates 1567:1–26.

Peigné, S. 2003. Systematic review of European Nimravinae (Mammalia, Carnivora,

Nimravidae) and the phylogenetic relationships of Palaeogene Nimravidae.

Zoologica Scripta 32:199-229.

Petter, G., and R. Hoffstetter. 1983. Les marsupiaux du Déséadien (Oligocène inférieur)

de Salla (Bolivie). Annales de Paléontologie 69:175-234.

Piras, P., D. Silvestro, F. Carotenuto, S. Castiglione, A. Kotsakis, L. Maiorino, M.

Melchionna, A. Mondanaro, G. Sansalone, C. Serio, V. A. Vero, and P. Raia.

2018. Evolution of the sabertooth mandible: A deadly ecomorphological

specialization. Palaeogeography, Palaeoclimatology, Palaeoecology 496:166-174.

Polly, P. D. 1996. The skeleton of Gazinocyon vulpeculus gen. et comb. nov. and the

cladistic relationships of Hyaenodontidae (Eutheria, Mammalia). Journal of

Vertebrate Palentology 16:303-319.

312

Powell, J. E., M. J. Babot, D. A. García-López, M. V. Deraco, and C. Herrera. 2011.

Eocene vertebrates of northwestern Argentina: annotated list; pp. 349–370 in J. A.

Salfity, and R. A. Marquillas (eds.), Cenozoic Geology of the Central Andes of

Argentina. SCS Publisher, Salta.

Prevosti, F. J., A. Forasiepi, and N. Zimicz. 2013. The evolution of the Cenozoic

terrestrial mammal guild in South America: competition or replacement? Journal

of Mammalian Evolution 20:3-21.

Prevosti, F. J., and A. M. Forasiepi. 2018. Evolution of South American Mammalian

Predators During the Cenozoic: Palaeobiogeographic and Palaeoenvironmental

Contengencies. 196 pp. Springer, Cham.

Prevosti, F. J., A. M. Forasiepi, M. D. Ercoli, and G. F. Turazzini. 2012. Paleoecology of

the mammalian carnivores (Metatheria, Sparassodonta) of the Santa Cruz

Formation (late early Miocene); pp. 173–193 in S. F. Vizcaíno, R. F. Kay, and M.

S. Bargo (eds.), Early Miocene Paleobiology in Patagonia: High-Latitude

Paleocommunities of the Santa Cruz Formation. Cambridge University Press,

Cambridge.

Ré, G. H., E. S. Bellosi, M. Heizler, J. F. Vilas, R. H. Madden, A. A. Carlini, R. F. Kay,

and M. G. Vucetich. 2010. A geochronology for the Sarmiento Formation at Gran

Barranca; pp. 46-58 in R. H. Madden, A. A. Carlini, M. G. Vucetich, and R. F.

Kay (eds.), The Paleontology of Gran Barranca. Evolution and Environmental

Change through the Middle Cenozoic of Patagonia. Cambridge University Press,

Cambridge.

313

Reguero, M. A., D. A. Croft, J. J. Flynn, and A. R. Wyss. 2003. Small archaeohyracids

from Chubut Province, Argentina and central Chile: implications for trans-

Andean temporal correlation. Fieldiana: Geology (New Series) 48:1-17.

Reig, O. A. 1957. Sobre la posición sistemática de "Zygolestes paranensis" Amegh. y de

"Zygolestes entrerrianus" Amegh., con una reconsideración de la edad y

correlación del "Mesopotamiense". Holmbergia 5:209-226.

Reig, O. A., J. A. W. Kirsch, and L. G. Marshall. 1987. Systematic relationships of the

living and Neocenozoic American "opossum-like" marsupials (suborder

Didelphimorphia), with comments on the classification of these and of the

Cretaceous and Paleogene New World and European metatherians; pp. 1-89 in M.

Archer (ed.), Possums and Opossums: Studies in Evolution, Vol. I. Surrey Beatty

and Sons in association with the Royal Zoological Society of New South Wales,

Chipping Norton.

Reig, O. A., and G. G. Simpson. 1972. Sparassocynus (Marsupialia, Didelphidae), a

peculiar mammal from the late Cenozoic of Argentina. Journal of Zoology

167:511-539.

Ribeiro, A. M., G. M. López, and M. Bond. 2010. The Leontiniidae (Mammalia,

Notoungulata) from the Sarmiento Formation at Gran Barranca, Chubut Province,

Argentina; pp. 170-181 in R. H. Madden, A. A. Carlini, M. G. Vucetich, and R. F.

Kay (eds.), The Paleontology of Gran Barranca: Evolution and Environmental

Change through the Middle Cenozoic of Patagonia. Cambridge University Press,

Cambridge, UK.

314

Riggs, E. S. 1933. Preliminary description of a new marsupial saber-tooth from the

Pliocene of Argentina. Geological Series, Field Museum of Natural History 6:61-

66.

Riggs, E. S. 1934. A new marsupial saber-tooth from the Pliocene of Argentina and its

relationships to other South American predacious marsupials. Transactions of the

American Philosophical Society 24:1-32.

Robinson, P., G. F. Gunnell, S. L. Walsh, W. C. Clyde, J. E. Storer, R. K. Stucky, D. J.

Froehlich, I. Ferrusquia-Villafranca, and M. C. McKenna. 2004. Wasatchian

through Duchesnean Biochronology; pp. 106-155 in M. O. Woodburne (ed.), Late

Cretaceous and Cenozoic Mammals of North America. Columbia University

Press, New York.

Robles, J. M., J. Madurell-Malapeira, J. Abella, C. Rotgers, R. Carmona, S. Almécija, J.

Balaguer, and D. M. Alba. 2013. New Pseudaelurus and Styriofelis remains

(Carnivora: Felidae) from the middle Miocene of Abocador de Can Mata (Vallès-

Penedès Basin). Comptes Rendus Palevol 12:101-113.

Rose, R. K., D. A. Pemberton, N. J. Mooney, and M. E. Jones. 2016. Sarcophilus

harrisii. Mammalian Species 49:1-17.

Rougier, G. W., B. M. Davis, and M. J. Novacek. 2015. A deltatheroidan mammal from

the Upper Cretaceous Baynshiree Formation, eastern Mongolia. Cretaceous

Research 52, Part A:167-177.

Rougier, G. W., J. R. Wible, and M. J. Novacek. 1998. Implications of Deltatheridium

specimens for early marsupial history. Nature 396:459-463.

315

Rougier, G. W., J. R. Wible, and M. J. Novacek. 2004. New specimen of Deltatheroides

cretacicus (Metatheria, Deltatheroida) from the Late Cretaceous of Mongolia).

Bulletin of Carnegie Museum of Natural History:245-266.

Sánchez-Villagra, M., S. Ladevèze, I. Horovitz, C. Argot, J. J. Hooker, T. E. Macrini, T.

Martin, S. Moore-Fay, C. Muizon, de, T. Schmelzle, and R. J. Asher. 2007.

Exceptionally preserved North American Paleogene metatherians; adaptations and

discovery of a major gap in the opossum fossil record. Biology Letters 3:318-322.

Schmidt-Lebuhn, A. N. 2016. TNT script for the Templeton test.

Scott, C. S., and R. C. Fox. 2015. Review of Stagodontidae (Mammalia, Marsupialia)

from the Judithian (Late Cretaceous) Belly River Group of southeastern Alberta,

Canada. Canadian Journal of Earth Sciences 52:682-695.

Scott, W. B. 1937. A history of land mammals in the Western Hemisphere. 786 pp. The

Macmillan Company, New York.

Sedor, F. A., É. V. Oliveira, D. D. Silva, L. A. Fernandes, R. F. Cunha, A. M. Ribeiro,

and E. V. Dias. 2017. A New South American Paleogene Land Mammal Fauna,

Guabirotuba Formation (Southern Brazil). Journal of Mammalian Evolution

24:39-55.

Shockey, B. J., and F. Anaya. 2008. Postcranial osteology of mammals from Salla,

Bolivia (late Oligocene): form, function, and phylogenetic implications; pp. 135-

157 in E. J. Sargis, and M. Dagosto (eds.), Mammalian Evolutionary Morphology:

A Tribute to Frederick S. Szalay. Springer, New York.

Simpson, G. G. 1930. Post-Mesozoic Marsupialia; pp. 1-87 in J. F. Pompeckj (ed.),

Fossilium Catalogus. I: Animalia. W. Junk, Berlin.

316

Simpson, G. G. 1935. The Tiffany fauna, upper Paleocene. I.-Multituberculata,

Marsupialia, Insectivora, and ?Chiroptera. American Museum Novitates 795:1-

19.

Simpson, G. G. 1941. The affinities of the Borhyaenidae. American Museum Novitates

1118:1-6.

Simpson, G. G. 1947. A new Eocene marsupial from Brazil. American Museum

Novitates 1357:1-7.

Simpson, G. G. 1948. The beginning of the age of mammals in South America. Part I.

Bulletin of the American Museum of Natural History 91:1–232.

Simpson, G. G. 1970a. Additions to knowledge of the Groeberia (Mammalia,

Marsupialia) from the mid-Cenozoic of Argentina. Breviora:1-17.

Simpson, G. G. 1970b. Mammals from the early Cenozoic of Chubut, Argentina.

Breviora 360:1–13.

Simpson, G. G. 1974. Notes on Didelphidae (Mammalia, Marsupialia) from the

Huayquerian (Pliocene) of Argentina. American Museum Novitates 2559:1-15.

Simpson, G. G. 1980. Splendid Isolation: the Curious History of South American

Mammals. 266 pp. Yale University Press, New Haven, Connecticut.

Sinclair, W. J. 1906. Mammalia of the Santa Cruz Beds. Marsupialia; pp. 333-460 in W.

B. Scott (ed.), Reports of the Princeton University Expeditions to Patagonia,

1896-1899. Princeton University, E. Schweizerbart'sche Verlagshandlung (E.

Nägele), Stuttgart.

Sinclair, W. J. 1930. New carnivorous Marsupialia from the Deseado Formation of

Patagonia. Field Museum of Natural History, Geological Memoirs 1:35-39.

317

Slater, G. J., and B. Van Valkenburgh. 2008. Long in the tooth: evolution of sabertooth

cat cranial shape. Paleobiology 34:403-419.

Slaughter, B. H. 1968. Earliest known marsupials. Science 162:254-255.

Solé, F., and S. Ladevèze. 2017. Evolution of the hypercarnivorous dentition in mammals

(Metatheria, Eutheria) and its bearing on the development of tribosphenic molars.

Evolution & Development 19:56-68.

Stallings, J. R. 1989. Small mammal inventories in an eastern Brazilian park. Bulletin of

the Florida State Museum, Biological Sciences 34:153-200.

Suarez, C., A. M. Forasiepi, F. J. Goin, and C. Jaramillo. 2016. Insights into the

Neotropics prior to the Great American Biotic Interchange: new evidence of

mammalian predators from the Miocene of northern Colombia. Journal of

Vertebrate Paleontology 36:e1029581.

Suyin, D., Z. Jiajian, Z. Yuping, and T. Yongsheng. 1977. The age and characteristic of

the Liuniu and the Dongjun faunas, Bose Basin of Guangxi. Vertebrata

PalAsiatica 15:35-45.

Szalay, F. S. 1994. Evolutionary History of the Marsupials and an Analysis of

Osteological Characters. Cambridge University Press, New York.

Szalay, F. S., and B. A. Trofimov. 1996. The Mongolian Late Cretaceous Asiatherium,

and the early phylogeny and paleobiogeography of Metatheria. Journal of

Vertebrate Paleontology 16:474-509.

Tate, G. H. H. 1947. On the anatomy and classification of the Dasyuridae (Marsupialia).

Bulletin of the American Museum of Natural History 88:97-156.

318

Tate, G. H. H. 1951. The banded anteater, Myrmecobius Waterhouse (Marsupialia).

American Museum Novitates 1521:1-8.

Tedford, R. H., L. B. Albright, III, A. D. Barnosky, I. Ferrusquia-Villafranca, R. M.

Hunt, Jr., J. E. Storer, C. C. Swisher, III, M. R. Voorhies, S. D. Webb, and D. P.

Whistler. 2004. Mammalian biochronology of the Arikareean through

Hemphillian interval (late Oligocene through early Pliocene epochs); pp. 169-231

in M. O. Woodburne (ed.), Late Cretaceous and Cenozoic Mammals of North

America: Biostratigraphy and Geochronology. Columbia University Press, New

York.

Tejada-Lara, J. V., R. Salas-Gismondi, F. Pujos, P. Baby, M. Benammi, S. Brusset, D. De

Franceschi, N. Espurt, M. Urbina, and P.-O. Antoine. 2015. Life in proto-

Amazonia: Middle Miocene mammals from the Fitzcarrald Arch (Peruvian

Amazonia). Palaeontology 58:341-378.

Templeton, A. R. 1983. Phylogenetic inference from restriction endonuclease cleavage

site maps with particular reference to the evolution of humans and the apes.

Evolution 37:221-244.

Thomas, O. 1887. On the homologies and succession of the teeth in the Dasyuridae with

an attempt to trace the history of the evolution of mammalian teeth in general.

Philosophical Transactions of the Royal Society of London 178:443-462.

Thomas, R. H., W. Schaffner, A. C. Wilson, and S. Pääbo. 1989. DNA phylogeny of the

extinct marsupial wolf. Nature 340:465.

319

Tomiya, S. 2013. New carnivoraforms (Mammalia) from the middle Eocene of

California, USA, and comments on the taxonomic status of 'Miacis' gracilis.

Palaeontologia Electronica 16:1-29.

Travouillon, K. J., Y. Gurovich, M. Archer, S. J. Hand, and J. Muirhead. 2013. The genus

Galadi: three new bandicoots (Marsupialia, Peramelemorphia) from Riversleigh’s

Miocene deposits, northwestern Queensland, Australia. Journal of Vertebrate

Paleontology 33:153-168.

Travouillon, K. J., Y. Gurovich, R. M. D. Beck, and J. Muirhead. 2010. An exceptionally

well-preserved short-snouted bandicoot (Marsupialia; Peramelemorphia) from

Riversleigh's Oligo-Miocene deposits, northwestern Queensland, Australia.

Journal of Vertebrate Paleontology 30:1528-1546.

Trofimov, B. A., and F. S. Szalay. 1994. New Cretaceous marsupial from Mongolia and

the early radiation of Metatheria. Proceedings of the National Academy of

Sciences 91:12569-12573.

Tseng, Z. J., G. T. Takeuchi, and X. Wang. 2010. Discovery of the upper dentition of

Barbourofelis whitfordi (Nimravidae, Carnivora) and an evaluation of the genus

in California. Journal of Vertebrate Paleontology 30:244-254.

Turnbull, W. D. 1978. Another look at dental specialization in the extinct sabre-toothed

marsupial, Thylacosmilus, compared with its placenta! counterparts; pp. 399-413

in P. M. Butler, and K. A. Joysey (eds.), Development, Function and Evolution of

Teeth. Academic Press, New York.

320

Van Valkenburgh, B. 1989. Carnivore dental adaptations and diet: A study of trophic

diversity within guilds; pp. 410-436 in J. L. Gittleman (ed.), Carnivore Behavior,

Ecology, and Evolution. Springer US.

Van Valkenburgh, B. 1999. Major patterns in the history of carnivorous mammals.

Annual Review of Earth and Planetary Sciences 27:463-493.

Van Valkenburgh, B., and C. B. Ruff. 1987. Canine tooth strength and killing behavior in

large carnivores. Journal of the Zoological Society of London 212:379-397.

Van Valkenburgh, B., X. Wang, and J. Damuth. 2004. Cope's Rule, Hypercarnivory, and

Extinction in North American Canids. Science 306:101.

Vieira, E. V., and D. Astúa de Moraes. 2003. Carnivory and insectivory in Neotropical

marsupials; pp. 271-284 in M. Jones, C. Dickman, and M. Archer (eds.), Predators

with Pouches: The Biology of Carnivorous Marsupials. CSIRO Publishing,

Collingwood.

Vieira, M. V., and A. Almeida Cunha, de. 2008. Scaling body mass and use of space in

three species of marsupials in the Atlantic Forest of Brazil. Austral Ecology

33:872-879.

Villarroel, C., and L. G. Marshall. 1982. Geology of the Deseadan (early Oligocene) age

"Estratos Salla" in the Salla-Luribay Basin, Bolivia, with description of new

Marsupialia. Geobios, mémoire spécial 6:201–211.

Villarroel, C., and L. G. Marshall. 1983. Two new late Tertiary marsupials

(Hathlyacyninae and Sparassocyninae) from the Bolivian Altiplano. Journal of

Paleontology 57:1061-1066.

321

Vizcaíno, S. F., M. S. Bargo, R. F. Kay, R. A. Fariña, M. Di Giacomo, J. M. G. Perry, F.

J. Prevosti, N. Toledo, G. H. Cassini, and J. C. Fernicola. 2010. A baseline

paleoecological study for the Santa Cruz Formation (late-early Miocene) at the

Atlantic coast of Patagonia, Argentina. Palaeogeography, Palaeoclimatology,

Palaeoecology 292:507–519.

Vizcaíno, S. F., and G. De Iuliis. 2003. Evidence for advanced carnivory in fossil

armadillos (Mammalia: Xenarthra: Dasypodidae). Paleobiology 29:123-138.

Voss, R. S., J. F. Díaz-Nieto, and S. A. Jansa. 2018. A revision of Philander

(Marsupialia: Didelphidae), part 1: P. quica, P. canus, and a new species from

Amazonia. American Museum Novitates 3891:1-70.

Voss, R. S., and S. A. Jansa. 2009. Phylogenetic relationships and classification of

didelphid marsupials, an extant radiation of New World metatherian mammals.

Bulletin of the American Museum of Natural History 322:1-177.

Voss, R. S., D. P. Lunde, and N. B. Simmons. 2001. The mammals of Paracou, French

Guiana: a neotropical lowland rainforest fauna. Part 2. Nonvolant species.

Bulletin of the American Museum of Natural History 262:1-236.

Vucetich, M. G., A. G. Kramarz, and A. M. Candela. 2010. Colhuehuapian rodents from

Gran Barranca and other Patagonian localities: the state of the art; pp. 206-219 in

R. H. Madden, A. A. Carlini, M. G. Vucetich, and R. F. Kay (eds.), The

Paleontology of Gran Barranca. Evolution and Environmental Change through the

Middle Cenozoic of Patagonia. Cambridge University Press, Cambridge.

Vullo, R., E. Gheerbrant, C. Muizon, de, and D. Néraudeau. 2009. The oldest modern

therian mammal from Europe and its bearing on stem marsupial

322

paleobiogeography. Proceedings of the National Academy of Sciences

106:19910-19915.

Werdelin, L. 1987. Jaw geometry and molar morphology in marsupial carnivores;

analysis of a constraint and its macroevolutionary consequences. Paleobiology

13:342-350.

Werdelin, L. 1989. Constraint and adaptation in the bone-cracking canid Osteoborus

(Mammalia: Canidae). Paleobiology 15:387-401.

Werdelin, L., N. Yamaguchi, W. E. Johnson, and S. J. O'Brien. 2010. Phylogeny and

evolution of cats (Felidae); pp. 59-82 in D. Macdonald, and A. Loveridge (eds.),

Biology and Conservation of Wild Felids. Oxford University Press, Oxford.

Wheeler, H. T. 2011. Experimental Paleontology of the Scimitar-tooth and Dirk-tooth

Killing Bites; pp. 19-34 in V. L. Naples, L. D. Martin, and J. P. Babiarz (eds.),

The Other Saber-Tooths: Scimitar-Tooth Cats of the Western Hemisphere. John

Hopkins University Press, Baltimore.

Wible, J. R. 2003. On the cranial osteology of the short-tailed opossum Monodelphis

brevicaudata (Didelphidae, Marsupialia). Annals of Carnegie Museum 72:137-

202.

Williamson, T. E., S. L. Brusatte, T. D. Carr, A. Weil, and B. R. Standhardt. 2012. The

phylogeny and evolution of Cretaceous–Palaeogene metatherians: cladistic

analysis and description of new early Palaeocene specimens from the Nacimiento

Formation, New Mexico. Journal of Systematic Palaeontology 10:625-651.

323

Wilson, G. P., E. G. Ekdale, J. W. Hoganson, J. J. Calede, and A. Vander Linden. 2016.

A large carnivorous mammal from the Late Cretaceous and the North American

origin of marsupials. Nature Communications 7:13734.

Wood, H. E. 1924. The position of the "sparassodonts" with notes on the relationships

and history of the Marsupialia. Bulletin of the American Museum of Natural

History 51:77-101.

Woodburne, M. O., F. J. Goin, M. Bond, A. A. Carlini, J. N. Gelfo, G. M. López, A.

Iglesias, and A. N. Zimicz. 2014. Paleogene land mammal faunas of South

America; a response to global climatic changes and indigenous floral diversity.

Journal of Mammalian Evolution 21:1-73.

Wroe, S. 1996. Muribacinus gadiyuli, (Thylacinidae; Marsupialia), a very plesiomorphic

thylacinid from the Miocene of Riversleigh, northwestern Queensland, and the

problem of paraphyly for the Dasyuridae (Marsupialia). Journal of Paleontology

70:1032-1044.

Wroe, S. 1997a. Mayigriphus orbus gen. et sp. nov., a Miocene dasyuromorphian from

Riversleigh, Northwestern Queensland. Memoirs of the Queensland Museum

41:439-448.

Wroe, S. 1997b. A reexamination of proposed morphology-based synapomorphies for the

families of Dasyuromorphia (Marsupialia). I. Dasyuridae. Journal of Mammalian

Evolution 4:19-52.

Wroe, S. 1999. The geologically oldest dasyurid, from the Miocene of Riversleigh, north-

west Queensland. Palaeontology 42:501-527.

324

Wroe, S., U. Chamoli, W. C. H. Parr, P. Clausen, R. Ridgely, and L. Witmer. 2013.

Comparative biomechanical modeling of metatherian and placental saber-tooths:

A different kind of bite for an extreme pouched predator. PLoS ONE 8:e66888.

Wroe, S., and A. Musser. 2001. The skull of Nimbacinus dicksoni (Thylacinidae:

Marsupialia). Australian Journal of Zoology 49:487-514.

Wysocki, M. A., R. S. Feranec, Z. J. Tseng, and C. S. Bjornsson. 2015. Using a novel

absolute ontogenetic age determination technique to calculate the timing of tooth

eruption in the saber-toothed cat, Smilodon fatalis. PLoS ONE 10:e0129847.

Wyss, A. R., J. J. Flynn, M. A. Norell, C. C. Swisher, III, M. J. Novacek, M. C.

McKenna, and R. Charrier. 1994. Paleogene mammals from the Andes of central

Chile: a preliminary taxonomic, biostratigraphic, and geochronologic assessment.

American Museum Novitates 3098:1-31.

Yale Peabody Museum of Natural History. 2016. YPM-VPPU 015097, Vol. 2017. Yale

Peabody Museum of Natural History,.

Yates, A. M. 2014. New craniodental remains of Thylacinus potens (Dasyuromorphia:

Thylacinidae), a carnivorous marsupial from the late Miocene Alcoota Local

Fauna of central Australia. PeerJ 2:e547.

Zack, S. P. 2012. Deciduous dentition of Didymictis (Carnivoramorpha: Viverravidae):

implications for the first appearance of “Creodonta”. Journal of Mammalogy

93:808-817.

Zack, S. P. in press. A skeleton of a Uintan machaeroidine ‘creodont’ and the phylogeny

of carnivorous eutherian mammals. Journal of Systematic Palaeontology:1-37.

325

Zimicz, A. N. 2012. Ecomorfología de los marsupiales paleógenos de América del Sur,

Vol. PhD. Universidad Nacional de La Plata, La Plata.

Zimicz, N. 2014. Avoiding competition: the ecological history of late Cenozoic

metatherian carnivores in South America. Journal of Mammalian Evolution

21:383–393.

326