MEMOIRES DU MUSEUM NATIONAL D’HISTOIRE NATURELLE

Redacteur en chef ( Editor-in-Chief) : Jean-Lou Justine

Redacteurs ( Editors ) : Jean-Marie Betsch, Philippe Bouchet, Christian Erard & Jean-Lou Justine

Assistante de redaction (Copy editor) : Bernadette Charles

Adresse (Address)

Memoires du Museum national d'Histoire naturelle

57, rue Cuvier

F-75005 Paris

Tel. : [33] (1) 40 79 34 37

Fax. : [33] (1) 40 79 38 08

e-mail : memoires@mnhn.fr

Les Memoires du Museum national d'Histoire

naturelle publient des travaux originaux majeurs,

tels que des monographies ou des volumes a auteurs

multiples. Les auteurs sont invites, pour toutes les

questions editoriales. a prendre contact avec le

directeur de la publication. Les manuscrits peuvent

etre en frangais ou en anglais.

Vente en France :

Museum national d’Histoire naturelle

Service des Publications Scientifiques

Diffusion : Delphine Henry

57, rue Cuvier

F-75005 Paris

Tel. : [33] (1) 40 79 37 00

Fax : [33] (1) 40 79 38 40

Parution et prix irreguliers. Les ordres perma¬

nents d'achat et les commandes de volumes separes

sont regus par le Service des Publications Scientifi¬

ques, Diffusion (pour la France et les dom-tom

uniquement), ou par Universal Book Services. Ca¬

talogue sur demande. Une liste des derniers titres

parus figure en page 3 de couverture.

The Memoires du Museum national d'Histoire

naturelle publishes major original contributions, such

as monographs or multi-authored volumes. Prospec¬

tive authors should contact the Editor-in-Chief. Ma¬

nuscripts in French or English will be considered.

Sales Office :

Universal Book Services

Dr W. Backhuys

p.o. Box 321

NL-2300 ah Leiden

Tel. : [31] (71) 517 02 08

Fax : [31] (71) 517 18 56

Volumes are published at irregular intervals, and

at irregular prices. Standing orders and orders for

single volumes should be directed to the Service des

Publications Scientifiques du Museum , (France and

dom-tom only) or Universal Book Services. Price list

and catalogues are available on request. Recently

published memoirs are listed on page 3 of the cover.

Printed on acid-free paper

Imprime sur papier non acide

B n. b . 1 ,^' h .t que Cen * r ale Museum

3 3001 00125988 5

Source: MNHN, Paris

Pucadelphys andinus (Marsupialia, Mammalia)

from the early Paleocene of Bolivia

bibi

DUi

as*

ISBN : 2-85653-223-3

ISSN : 1243-4442

Photocopies :

Les Memoires du Museum adherent au Centre Fran^ais d'Exploitation du Droit de Copie

(CFC), 3, me Hautefeuille, 75006 Paris. Le CFC est membre de l'lnternational Federation of

Reproduction Rights Organisations (IFRRO). Aux Etats-Unis d'Amerique, contacter le

Photocopies:

The Memoires du Museum adhere to the Centre Frangais d'Exploitation du Droit de Copie

(CFC), 3, rue Hautefeuille, 75006 Paris. The CFC is a member of International Federation of

Reproduction Rights Organisations (IFRRO). In USA, contact the Copyright Clearance Center,

27, Congress Street, Salem, Massachusetts 01970.

MEMOIRES DU MUSEUM NATIONAL D'HISTOIRE NATURELLE

TOME 165

PALfiONTOLOGIE

Pucadelphys andinus (Marsupialia, Mammalia)

from the early Paleocene of Bolivia

Larry G. MARSHALL*, Christian DE MUIZON**, Denise SlGOGNEAU-RUSSELL***

* Institute of Human Origins, 1288, 9 lh Street, Berkeley

California 94710, USA

** Institut Frangais d’Etudes Andines, URA 12 CNRS

Cas. 18-1217, Lima 18, Peru

*** Museum national d'Histoire naturelle

Laboratoire de Paleontologie, URA 12 CNRS

8, rue Buffon, F-75005 Paris, France

EDITIONS

DU MUSEUM

PARIS

1995

Source: MNHN, Paris

Source: MNHN , Paris

CONTENTS

SOMMAIRE

Part I: The locality of Tiupampa: age, taphonomy and mammal fauna

(L. G. Marshall, C. de Muizon & D. Sigogneau-Russell)

Pages

Age . 13

Taphonomy . 13

Mammalian Faunal list . 17

Part II: The skull (L. G. Marshall & C. de Muizon)

Abstract . 2 *

Resume . 22

Extended French Summary . 22

Introduction . 22

Systematic Paleontology . 28

Description . 31

Upper dentition . 31

Lower dentition .

Dentary .

Skull . 37

General Skull Structure. 37

Nasal . 39

Premaxilla . 39

Maxilla . 42

Palatine . 42

Lacrimal . 4 5

Jugal . 47

Frontal . 47

Parietal . 49

Postparietal . 50

Orbitosphcnoid . 50

Alisphenoid . 50

Basisphenoid .

Squamosal . 52

Basioccipital . 54

Exoccipital . 54

Supraoccipital . 56

Petromastoid . 56

Ec to tympanic ... ^4

Discussion . ^ 4

Dental comparison . ^5

Character Analysis . 67

Dentition . 67

Dentary . 68

Bones of the Skull . 69

Foramina of the skull . 71

Ear Region . 73

Conclusions . 82

Acknowledgments . 84

Source

LARRY G. MARSHALL. CHRISTIAN DE MUIZON & DENISE SIGOGNEAU-RUSSELL

References . 84

Measurements . 90

Part III : The postcranial skeleton (L. G. Marshall & D. Sigogneau-Russell)

Abstract .

Resume .

Extended French summary .

Introduction .

Systematic Paleontology.

Description .

General Features .

Axial skeleton .

Cervical vertebrae .

Allas .

Axis .

Cervicals 3 to 7

Thoracic vertebrae .

Ribs .

Lumbar vertebrae .

Sacral vertebrae .

Caudal vertebrae .

Discussion .

Appendicular skeleton .

Shoulder girdle .

Scapula .

Clavicle .

Interclavicle ....

Forelimb .

Humerus .

Ulna .

Radius .

Manus .

Discussion .

Pelvic girdle .

Ilium .

Ischium .

Pubis .

Os marsupium .

Hindlimb .

Femur .

Tibia .

Fibula .

Pcs .

Discussion.

Habits .

Locomotion .

Way of life .

Conclusions .

Acknowledgments .

References .

Appendix: measurements .

101

104

106

108

I 19

1 19

I 19

122

125

127

130

132

135

137

139

145

148

149

150

151

152

153

156

Source: MNHN, Paris

ABSTRACTS

Marshall, L. G., Muizon. C. de & Sigogneau-Russell. D., 1995.— Pan I: The locality of Tiupampa: age. taphonomy and

mammalian fauna. In : C. de Muizon (ed.). Pucadelphysandinus (Marsupialia, Mammalia) from the early Palcocene of Bolivia.

Mem. Mus. natn. Hist, nat .. 165 : 11-20. Paris ISBN : 2-85653-223-3.

Published November 2T' 1 1995.

The specimens of Pucadelphys andinus studied in this volume were collected in the early Palcocene beds of the Santa Lucia

Formation at Tiupampa (Bolivia). In this first part, a shon presentation is given of the locality of Tiupampa, of the horizon from

which the skeletons of Pucadelphys andinus were collected and of their age. We also provide an updated list of the mammalian

fauna yielded by the site, and present comments on the taphonomy and paleoenvironmcnt of the oldest known undoubted marsupial

skeletons.

Marshall, L. G.. Muizon C. de. 1995.— Part II: The skull In: C. de Muizon (ed.), Pucadelphys andinus (Marsupialia,

Mammalia) from the early Paleocene of Bolivia. Mem. Mus. natn. Hist, nat., 165 : 21-90. Paris ISBN : 2-85653-223-3.

Published November 2?" 1995.

One of the earliest known complete dentitions and skulls of a metatherian mammal are represented by Pucadelphys andinus

Marshall & Muizon (1988) from the Early Paleocene age Santa Lucfa Formation at Tiupampa (type locality of the Tiupampan

Land Mammal Age) in southcentral Bolivia. A detailed analysis of the dentition, dentary and skull reveals that the vast majority

of states in Pucadelphys can be regarded as mammalian and/or tribosphenid plesiomorphies, while metatherian synapomorphies

include: cheek tooth formula of P3/3 +M4/4; prootic canal reduced in length and width, and does not open endocranially; prootic

sinus continues onto squamosal side of petromastoid within a deep sulcus; presence of sphenoparietal emissary vein which

occupies deep sulcus on squamosal side of petromastoid, is continuous with sulcus for prootic sinus, and exits skull through

postglenoid foramen; length/width ratio of fenestra vestibuli 1.4; and absence of stapedial artery. Pucadelphys lacks an ossified

alisphenoid bulla, but has what is interpreted to be a small anterior lamina of the petromastoid and a large foramen ovale (exit of

V3) which opens between the anterior lamina and alisphenoid. The presence of an anterior lamina of the petromastoid fused to

the pars petrosa associated to the presence of an anteroposteriorly expanded alisphenoid in Pucadelphys seems to contradict the

hypothesis of Presley & Steel (1976) and Presley (1981 ) on the evolution of the tribosphenid middle ear and lateral wall of the

skull based on ontogenetic studies. This is the First record of an anterior lamina in a tribosphenid fossil mammal. The molar structure

of Pucadelphys is indistinguishable from members of the family Didelphidae. and it is placed in this family, within the Order

Didelphimorphia. Pucadelphys represents an intriguing example of mosaic evolution and illustrates that we have much to learn

about early metatherian and tribosphenid evolution.

Marshall, L. G. & Sigogneau-Russell, D.. 1 995.— Part III: Postcranial skeleton. In: C. df. Muizon (ed.), Pucadelphys andinus

(Marsupialia, Mammalia) from the early Paleocene of Bolivia. Mem. Mus. natn. Hist, not., 165 : 91-164. Paris ISBN : 2-85653-

223-3.

Published November 27°' 1995.

I he earliest and most complete articulated skeletons ot fossil metatherians yet known are represented by four specimens of

Pucadelphys andinus from the Santa Lucfa Formation (early Paleocene) at Tiupampa in southcentral Bolivia. Two sets of what

arc interpreted to be male-female pairs were found in a three dimensional, life-like, snout-rump position in burrow-nests that were

apparently dug in a bank along a meandering river. The animals apparently died as the result of a flood which entrapped them in

their burrows and filled the latter with water and sediment. A detailed comparative study of the postcranial bones reveals that the

vast majority of character states in Pucadelphys are regarded as mammalian, tribosphcnic or metatherian plesiomorphies (e.g. atlas

not perforated by transverse canal and with a persisting suture between ossified intercentrum and atlantal arch; absence of

transverse canal on axis, with possible un fused rib; absence of enclosed transverse canal on CV7; robust fibula; presence of ossified

os marsupium; etc.). Character states of uncertain polarity include the presence of only one vertebra articulating with the ilium

(fulcral vertebra), and a long non-prehensilc tail. The tarsus has a bicontact upper ankle joint (UAJ) as in living Didelphidae;

moreover the calcaneum shows a partially plantar orientation of the cuboid facet which can be interpreted as foreshadowing the

specialisation of later Didelphidae; the situation then is more advanced than in the “plesiomorphic metatherian morphotype" of

Szalay (1982 a. b; 1984); the only characters of the latter persisting in Pucadelphys are the large peroneal process and the

“remarkably broad transverse dimensions from peroneal process to the medial margin of the sustentaculum”. Collectively these

characters support the view, based on the study of the skull and dentition (Marshall & Muizon. this volume), that Pucadelphys

represents the plesiomorphic taxon within the family Didelphidae. Functional considerations of the skeletons suggest that

Pucadelphys was essentially terrestrial, quite agile, and possessed limited bounding and digging capabilities.

Source: MNHN , Paris

Part I: The locality of Tiupampa: age, taphonomy and

mammalian fauna

Larry G. Marshall *, Christian de Muizon**

& Denise Sigogneau-Russell***

^Institute of Human Origins, 1288, 9th Street, Berkeley

California 94710, USA

**Institul Fran^ais cTEtudes Andines, URA 12 CNRS

Cas. 18-1217. Lima 18, Peru

***Museum national d'Histoire naturelle

Laboratoire dc Paleontologie, URA 12 CNRS

8 rue Buffon, F-75005 Paris, France

ABSTRACT

The specimens of Pueadelphys andinus studied in this volume were collected in the early Paleocene beds of the Santa Lucfa

Formation at Tiupampa (Bolivia). In this first part, a short presentation is given of the locality of Tiupampa, of the horizon from

which the skeletons of Pueadelphys andinus were collected and of their age. We also provide an updated list of the mammalian

fauna yielded by the site, and present comments on the taphonomy and paleocn vironment of the oldest known undoubted marsupial

skeletons.

RESUME

Premiere partie : ie gisement de Tiupampa : age, taphonomie et faune de mammiferes

Les specimens dc Pueadelphys andinus qui font l’objet de ce memoire ont ete r£coltes dans des niveaux du Paleocene ancien

de la Formation Santa Lucfa a Tiupampa (Boli vie). Cette premiere panic offre une breve presentation de la localite de Tiupampa.

dc f horizon d’oii proviennent les squelettes et de leur age. Ellc contient aussi une liste mise h jour dc la faune mammalienne

d’accompagnement, ainsi que des informations sur la taphonomie et Ie paleoenvironnement de ces squelettes de marsupiaux. qui

sont les plus anciens actuellcment connus.

Marshall, L.G., Muizon, C. de& Sigogneau-Russell, D., 1995.— Part I: The locality of Tiupampa: age. taphonomy and

mammalian fauna. In: Muizon, C. de (ed.), Pueadelphys andinus (Marsupialia, Mammalia) from the early Paleocene of Bolivia.

Mem. Mus. natn. Hist, nat., 165 : 11-20. Paris ISBN : 2-85653-223-3.

LARRY G. MARSHALL. CHRISTIAN de MUIZON & DENISE SIGOGNEAU-RUSSELL

Fig. 1. — Map of southcentral Bolivia showing location of Tiupampa (top) and detail map of Tiupampa area showing fossil

vertebrate sites (bottom). The skeletons of Pucadelphys andinus are from site 1 (“the quarry").

FtG. /. - Carte de la partie suddu centre de la Bolivie montranl la situation de Tiupampa (en haul), et carte detaillee de la region

de Tiupampa montrant les locallies d vertebresfossiles (en bas). Les squelettes de Pucadelphys andinus proviennent de

la localite I ("the quarry").

Source: MNHN, Paris

PUCADELPHYS ANDINUS: THE LOCALITY OF TIUPAMPA

The locality of Tiupampa, recognized as a promising site as early as 1980 by R. Hoffstteter,

C. de Muizon and P. Taquet, yielded its first mammalian remains (a marsupial) during a second

campaign (1982) organized in particular by L. G. Marshall and Ch. of. Muizon. It was in the

course of a 4th campaign (1985) that the skeletons of Pucadelphys andinus presented here were

found.

LOCALITY. HORIZON, AND AGE

These specimens were recovered from the Santa Lucia Formation (sensu Gayet et al., 1992;

Muizon, 1992; Muizon & Brito. 1993) at Tiupampa, located about 95km southeast of Cochabamba

(65°35'W, 18°02’S),Mizque Province, Department of Cochabamba, southcentral Bolivia (Fig. l,top).

Details of the topography are shown on Caita Nacional Boliviana San Vicente Quadrangle, Hoja 6439

1, serie H731,1/50,000, 1968. All specimens are from site 1 (“the quarry") (Fig. 1, bottom).

The rich and taxonomically diverse vertebrate fauna discovered at Tiupampa in 1982 was

initially believed to be from the “Late Cretaceous” age El Molino Formation (Marshall et al.,

1983; 1985; Muizon et al., 1983, 1984; Marshall & Muizon. 1988). However, recent detailed

studies of the biostratigraphy (Gayet et al., 1992) and stratigraphy (Sempere, 1994) of the El

Molino and overlying Santa Lucia Formations reveal the following: 1) uncontested Late Creta¬

ceous age “guide" fossils (i.e. dinosaurs, some selachians and fish) occur only in the lower and

basal middle members of the El Molino Formation, which collectively span early and late

Maastrichtian times; 2) the upper member of the El Molino Formation is Early Paleocenc

(Danian); 3) the K/T boundary lies in the upper part of the middle member of the El Molino

Formation; but. another interpretation is given by Jaillard et al. (1993: 650) who definitely

consider the El Molino Formation as a lateral equivalent of the Upper Vilquechico Formation from

Peru, which they ascribed to a latest Campanian-late Maastrichtian interval (Jaillard etal., 1993:

655). The mammal-bearing level at Tiupampa, which has yielded the type fauna of the Tiupampian

Land Mammal age, has been assigned to the Santa Lucia Formation (Gayet et al.. 1992; Muizon,

1992; and Muizon & Brito, 1993) and has been given an early Paleocene age (Ortiz Jaureguizar

& Pascual, 1989; Marshall, 1989; Bonaparte et al., 1993; Muizon & Brito, 1993). Finally,

studies in progress on the geochronology of the El Molino and Santa Lucia Formations (Sempere

& Marshall, in prep.; Sempere et al., in prep.) would suggest an age slightly younger in the

Paleocene than stated before. Pending formal publication of these results, the Tiupampa land

mammal fauna and the Tiupampian land-Mammal age will be referred here to the early Paleocene

following Van Valen (1988), Muizon (1992), Bonaparte et al. (1993), and Muizon & Brito

(1993).

Stratigraphic sections showing the location of the fossil level in the Tiupampa section are

provided by Muizon et al. (1983, fig. I) and Marshall et al. (1985, fig. 4, right).

TAPHONOMY

The Santa Lucia Formation at Tiupampa was deposited in channels of meandering rivers on

a flat alluvial plain. The water-lain sediments surrounding the skeletons are a well-sorted, weakly

LARRY G. MARSHALL. CHRISTIAN nr, MUIZON & DENISE SIGOGNEAU-RUSSELL

Source; MNHN, Paris

PUCADELPHYS ANDINUS: THE LOCALITY OFTIUPAMPA

consolidated, fine-grained quartzose-sandstone with sparse feldspar grains, some hematite and

traces of red clay between the predominantly quartz grains. The red color comes from the iron

dioxide. The presence of several taxa of crocodiles attests to a warm, probably subtropical,

climate.

The specimens were collected from sediment blocks that had been transported about 2km

from site I (Fig. 1, bottom) to a screen-washing location; they were first observed by P.-Y.

Gagnier. The blocks containing the skeletal remains were prepared at the MNHN by C. de Muizon

and L. G. Marshall. This revealed that four partial skeletons were represented and that these

occurred in pairs (6105 and 6106, 6110 and 6111, respectively; see below).

The individuals of each pair were apparently in a “snuggle” position and facing in opposite

directions (Figs 3 and 4). We interpret this relationship to indicate that these animal-pairs

(probably male and female, see below) were in burrow-nests and sleeping (or resting) in a snout-

rump position typical of many living didelphids (see below). We further believe that these animals

probably died as the result of a flood that entrapped them in their burrows and filled the latter with

water and sediment. The “died-in-a-burrow” hypothesis is supported by the fact that frogs (some

represented by complete skeletons, Baez, 1992) are abundant throughout “the quarry” fossil level,

and that these apparently lived in a pond or oxbow adjacent to a bank that was used as a burrow-

nest site; this interpretation would also explain the exceptionnally high number of fossils in this

level, which may have died as a result of a single carastrophic flood (Sempere & Marshall, in

prep.).

There was apparently little time lapse between death and the onset of fossilization (i.e.

permanent entombment in the sediment). This is evidenced by the fact that the majority of bones

are articulated, suggesting little or no postmortem dismemberment. Nevertheless, some rib

fragments in both specimen-pairs are situated “out-of-context” of the main skeletons, no distal

phalanges or claws are articulated with the four partial pes and only five caudal vertebrae posterior

to C9 are preserved. These features are interpreted to indicate that some movement of sediment

Fig. 2. — Pucadelphys andinus. A. specimen-pair YPFB Pal 6105. holotype. (top) and YPFB Pal 6106 (bottom) photographed

at initial stage of preparation; 6105 faces right; 6106 faces left. X3/4. B, line-drawing of same specimen-pair after

preparation (skull of YPFB Pal 6105 detached). Bones of YPFB Pal 6105 arc stippled and bones of YPFB Pal 6106 are

left white to facilitate identification and association. X 1,5. In Roman letters, unpaired bones and paired bones of the right

side; in italics, paired bones of the left side; cross-hatched, unaltributed bones. Abbreviations: A. astragalus; AT. atlas;

AX. axis; C, caudal vertebra; CA, calcaneum; CL. clavicle; Cl), cuboid; CV. cervical vertebra; F. femur; FI. fibula; H.

humerus; IL. ilium; IN. interclavicle; IS. ischium; L, lumbar vertebra; MA. os marsupium; MT. metatarsal; NA.

navicular; PU. pubis; R. rib; RA, radius; S. sacral vertebra; SC. scapula; T, thoracic vertebra; TI, tibia; IJ, ulna.

Fig. 2. Pucadelphys andinus. A, couple de specimens (YPFB Pal 6/05. holotype. en ham) et YPFB Pal 6106 (en bas)

photographies ciu debut de la preparation: YPFB Pal 6/05 est sur le cote droit. 6/06 sur le cote gauche. X 3/4. B. dess in

an trail des tnemes specimens apres preparation (le crane de YPFB Pal 6105 a ete retire). X 1.5. Les os de YPFB Pal 6105

soul point i I les et ceux de YPFB Pal 6106 ont ete laisses en blanc pourfaciliter I ‘identification et les associations. En lettres

romaines, os impairs el os pairs du cote droit: en italic/ues, os pairs du cote gauche: en hachures, os non attribues.

Abreviations: A. cistragale: AT. atlas: AX. axis. C. vertebre caudate: CA, calcaneum: CL, clavicule; CV. cuboide: CV.

vertebrecervicale: F. femur: FI, fibula: 11. humerus: IL. ilion; IN, interclavicule; IS. ischium; L. vertebre lombaire; MA.

os marsupial: MT. metatarsien; NA, naviculaire: PU, pubis; R, cote; RA, radius; S, vertebre sacree; SC. scapula; T.

vertebre thoracique; TI. tibia; V, cubitus.

LARRY G. MARSHALL. CHRISTIAN de MUIZON & DENISE SIGOGNEAU-RUSSELL

Source: MNHN, Paris

PUCADELPHYS ANDINUS: THE LOCALITY OF TIUPAMPA

around the corpses occurred, causing minor displacement and/or loss of some fragile extremity

bone elements. Diagenetic processes were, however, exceedingly weak as evidenced by the facts

that the sediments are poorly consolidated, the bones show no or only minor evidence of

deformation, and the skulls and skeletons are preserved in a three dimensional, “life-like” state.

MAMMALIAN FAUNAL LIST

The Locality of Tiupampa has yielded an abundant mammalian fauna. An updated list

(modified from Muizon, 1992 and Muizon & Brito, 1993) is given below:

Class Mammalia

Infra-class Metatheria

Order Sparassodonta

Family Hathliacynidae

Allqokirus australis Marshall & Muizon, 1988

Family Mayulestidae

Mayulestes ferox Muizon, 1994

Order Peradectia

Family Peradectidae

Peradectes cf. austrinum

Roberthoffstetteria nationalgeographica

Marshall, Muizon & Sige, 1983

Order Microbiotheria

Family Microbiotheriidae

Khasia cordillerensis Marshall & Muizon, 1988

Fig. 3. — Pucadelphys andinus. A, specimen-pair YPFB Pal 6110 (bottom) and YFPB Pal 6 111 (top) photographed at initial stage

of preparation. YPFB Pal 6110 faces right; YPFB Pal 6111 faces left. X 3/4. B. line drawing of same specimen-pair after

preparation (skull of YPFB Pal 61 lOdeiached). X 1.5. Bones of YPFB Pal 6111 are stippled and bones of YPFB Pal 6110

are left while to facilitate identification and association. In Roman letters, unpaired and paired bones of the right side; in

italics, paired bones of the left side; cross-hatched, unattributed bones. In the part of YPFB Pal 6110 corresponding to Fig.

4. bones are drawn reversed. For abbreviations, see caption to Fig. 2.

Fig. 3. Pucadelphys andinus. A. couple cle specimens YPFB Pal 6110 (en has) el YPFB Pal 6111 (en haul) photographies au

debut de la preparation. YPFB Pul6IIOest sur le cote droit: YPFB Pal 6111 sur le cote gauche. X 3/4: B. dessin au trait

des memes specimens aprespreparation (le crane de YPFB Pal 6110a ete retire). X 1.5. Les os de YPFB Pal 6111 sont

pointilles et ceux de YPFB Pal 6110 out ete laisses en blanc pour faciliter I identification et les associations. En leltres

romaines. os impairs et os pairs du cote droit: en italiques. os pairs du cote gauche ; en hachures, os non attribues. Sur la

partie de YPFB Pal 6110 correspondant a la Fig. 4. le dessin a ete renverse. Pour les abreviations. voir legeiule Fig. 2.

LARRY G. MARSHALL, CHRISTIAN dr MUIZON & DENISE SIGOGNEAU-RUSSELL

Order Didelphi morphia

Family Didelphidae

Pucadelphys andinus Marshall & Muizon, 1988

Incadelphys antiquus Marshall & Muizon. 1988

Mizquedelphyspilpinensis Marshall & Muizon, 1988

Andinodelphys cochabambensis Marshall & Muizon. 1988

Tiulordiafloresi Marshall & Muizon, 1988

Family Jaskhadelphydae

Jaskhadelphys minutus Marshall & Muizon, 1988

Infra-class Eutheria

Order Lcptictida

Family Palaeoryctidae?

cf. Cimolestes sp.

Family indet.

Gen. and sp. indet.

Order Pantodonta

Family Alcidedorbignyidae

Alcidedorbignya inopinata Muizon & Marshall, 1987

Order Condylarthra

Family Mioclaenidae

Molinodus suarezi Muizon & Marshall. 1987

Tiuclaenus minutus Muizon & Marshall,1987

Tiuclaenus sp. nov. 1

Tiuclaenus sp. nov.2

Pucanodus gagnieri Muizon & Marshall, 1991

Mioclaenidae nov. gen., nov. sp.

Family Mioclaenidae or Didolodontidae

Andinodus boliviensis Muizon & Marshall, 1987

Family ?Periptychidae

aff. IMimatuta

Order Condylarthra incertae sedis

Family Kollpaniidae

Kollpania tiupampina Marshall & Muizon, 1988

Order Notoungulata

Family Henricosborniidae or Oldfiedthomasiidae

ABBREVIATIONS OF INSTITUTIONS

MNHN. Laboratoire de Paleontologic, Museum national d'Histoire naturelle, Paris, France.

YPFB Pal. Paleontology collection of Yacimientos Petrolfferos Fiscales de Bolivia in the Centro

de Tecnologfa Petrolera, Santa Cruz, Bolivia.

Source: MNHN . Paris

PUCADELPHYS ANDINUS: THE LOCALITY OF TIUPAMPA

Fig. 4. — Pucadelphysandinus. YPFB Pal 6110.

reversed photo (X 3) of accessory block,

to show its continuity with main block in

Fig. 3 (see explanation Fig. 26).

Fig. 4. — Pucadelphys andinus. YPFB Pal 6110,

photo ren versee (X3) chi bloc accessoire,

pour montrer la continuite avec le bloc

principal de la Fig. 3 (voir explication

Fig . 26).

REFERENCES

Baez, A.M., 1992. — A new Early Paleocene Neobatrachian frog from the Santa Lucfa Formation in South Central Bolivia, with

comments on the Cretaceous and Early Tertiary batrachofaunas of South America. In: Fossilesy Facies de Bolivia (R.

Suarez, ed.), Revista Tecnica de YPFB, Dec. 1991.12 (2-3): 529-540, Santa Cruz.

Bonaparte, J. F.. Van Valen. L. M.. Kramartz, A., 1993. — La fauna local de Punta Peligro, Paleoccno inferior, de la Provincia

del Chubut, Patagonia, Argentina. Evolutionary monographs. 14: 1-61.

Gayet. M., Marshall, L. G.. & Sempere, T.. 1992. — The vertebrate-bearing Late Cretaceous-Palcoccnc of Bolivia and its

stratigraphic context: a review. Revista Tecnica de Yacimientos PetroUferos biscales de Bolivia , Cochabamba,

December 1991. 12(3-4): 393-433.

J aillard, E.. Cappetta. H., Ellenberger. P., Feist. M.. Grambast-Fessard, N., Lefranc, J.-P. & Sige, B., 1993. — Scdimemology,

paleontology, biostratigraphy and correlation of the late Cretaceous Vilquechico Group of southern Peru. Cretaceous

Research , 14 : 623-661.

Marshall, L. G., 1989. — The K-T boundary in South America: on which side is Tiupampa? National Geographic Research, 5

(3): 268-270.

Marshall. L. G.. & Muizon C. de, 1988. — The dawn of the age of mammals in South America. National Geographic Research,

4 (1): 23-55.

Marshall. L. C... Muizon, C. de. Gayet, M.. Lavenu, A.. & Sige. B.. 1985. — The “Rosetta Stone” for mammalian evolution in

South America. National Geographic Research, 1(2): 247-288.

LARRY G. MARSHALL. CHRISTIAN de MUIZON & DENISE SIGOGNEAU-RUSSELL

Marshall, L.G., Muizon, C. de,&Sig6. B., 1983. — Late Cretaceous mammals (Marsupialia) from Bolivia. Geobios , 16 (6): 739-

745, Lyon.

Muizon, C. de, 1992. — La fauna de mamfferos deTiupampa (Palcoccno inferior, Formacfon Santa Lucia) Bolivia. Revista tecnica

de Yacimientos Petrolferos Fiscales de Bolivia . December 1991. 12 (3-4): 575-624.

Muizon, C. de, 1994. — A new carnivorous marsupial from the Paleocene of Bolivia and the problem of marsupial monophyly.

Nature, 370: 208-211.

Muizon, C. de,& Brito, I. M., 1993. — Le bassin calcairede Sao Jose deltaborai (Rio de Janeiro. Bresil): ses relations fauniques

avee le site de Tiupampa (Cochabamba, Bolivie). Annates de Paleontologies 79 (3): 233-268.

Muizon, C. de, Gayet, M., Lavenu. A., Marshall, L. G., Sige. B., & Villarroel, C.. 1983. — Late Cretaceous vertebrates

including mammals from Tiupampa, southcentral Bolivia. Geobios, 16: 747- 753.

Muizon, C. de. Marshall, L. G.. & Sige, B., 1984. — The mammal fauna from the El Molino Formation (Late Cretaceous-

Maestrichtian) at Tiupampa, southcentral Bolivia. Bulletin du Museum national d'Histoire naturelle , Paris, Serie 6,

Section C. 4: 327-351.

Ortiz Jaureguizar, E. & Pascual, R.. 1989. — South American land-mammal faunas during the Cretaceous-Tertiary transition:

evolutionary biogeography. Contribution to the symposium: Cretaceous in Latin America. Part A, Events and

sedimentological record. Buenos Aires: 231-252.

Sempere, T.. 1994. — Kimmeridgian? to Paleocene tectonic evolution of Bolivia. In: Cretaceous tectonics of the Andes. J. A.

Salfity (ed.) Vieweg: 167-212, Wiesbaden.

Sempere, T. & M arshall. L.G., in prep. — Location of the Tiupampa vertebrates in the Late Cretaceous-Paleocene stratigraphy

of Bolivia: resolution of a paleontological debate.

Sempere T.S., Marshall L.G.. Butler, R.L., Sharp W. & Swisher C. in prep. — Chronostratigraphy of the Maastrichtian-

Paleoccnc of Bolivia and a chronology of Paleocene land ammmal faunas of South America.

Van Valen, L.. 1988. — Paleocene dinosaurs or Cretaceous ungulates in South America. Evolutionary Monographs. 10: 1-7

Source: MNHN , Paris

Part II: The skull

Larry G. Marshall * & Christian de Muizon**

institute of Human Origins, 1288, 9th Street, Berkeley

California, 94710, USA

**Institut Fran^ais d’Etudes Andines, URA 12 CNRS

Cas. 18-1217, Lima 18, Peru

ABSTRACT

One of the earliest known complete dentitions and skulls of a metatherian mammal are represented by Pucadelphys andinus

Marshall & Muizon, (1 988) from the early Paleocene age Santa Lucia Formation at Tiupampa (type locality of the Tiupampan

Land Mammal Age) in southcentral Bolivia. A detailed analysis of the dentition, dentary and skull reveals that the vast majority

of states in Pucadelphys can be regarded as mammalian and/or tribosphenid plesiomorphies. while metatherian synapomorphies

include: cheek tooth formula of P3/3+ M4/4; prootic canal reduced in length and width, and does not open endocranially: prootic

sinus continues onto squamosal side of petromastoid within a deep sulcus; presence of sphenoparietal emissary vein which

occupies deep sulcus on squamosal side of petromastoid, is continuous with sulcus for prootic sinus, and exits skull through

postglenoid foramen; length/width ratio of fenestra vestibuli 1.4; and absence of stapedial artery. Pucadelphys lacks an ossified

alisphenoid bulla, but has what is interpreted to be a small anterior lamina of the petromastoid and a large foramen ovale (exit of

V3) which opens between the anterior lamina and alisphenoid. The presence of an anterior lamina of the petromastoid fused to

the pars petrosa associated to the presence of an anteroposteriorly expanded alisphenoid in Pucadelphys seems to contradict the

hypothesis of Presley & Steel (1976) and Presley (1 98 1) on the evolution of the tribosphenid middle ear and lateral wall of the

skul 1 based on ontogenetic studies. This is the first record of an anterior lamina in a tribosphenid fossil mammal. The molar structure

of Pucadelphys is indistinguishable from members of the family Didelphidac, and it is placed in this family, within the Order

Didelphimorphia. Pucadelphys represents a classic example of mosaic evolution and illustrates that we have much to learn about

early metatherian and tribosphenid evolution.

Marshall, L. G.. Muizon C. de, 1995.— Part II: The skull. In: Muizon, C. de (ed.), Pucadelphys andinus (Marsupialia,

Mammalia) from the early Paleocene of Bolivia. Mem. Mus. natn. Hist, nat ., 165 : 21-90. Paris ISBN : 2-85653-223-3.

LARRY G. MARSHALL & CHRISTIAN de MUIZON

RESUME

Deuxieme partie : le crane

Lcs cranes cl dentitions de Pitcadelphyscmdimts Marshal!. & Muizon, ( 1988) decriis dans ce travail sont parmi les plus anciens

connus chez lcs metatheriens. Us provienneni dti Paleoccne inferieur de la Formation Santa Lucia a Tiupampa (locality-type de

I'agede mammiferescontinentaux.Tiupampien)en Boliviecentrale. Une analyse detailleede la dentition, du dentaire etdu crane

revele que la grande majority des etats de Pucadelphys peuvent etre considers comme des plesiomorphies au sein de mammifercs

et/ou des tribosphenides. Les synapomorphies de metatherien sont : dents jugales au nombre de P 3/3 ct M 4/4: canal prootique

reduit cn longueur et en largeur. ne s’ouvrant pas a Pinterieurdu crane; sinus prootique se prolongeant sur la face squamosale du

petromastoVde dans un profond sillon; presence d'unc veine emissairc sphenoparietale occupant un profond sillon sur la face

squamosale du petromastoVde. cn continuity avec le sillon pour le sinus prootique et sortant du crane a travers le foramen

postglenoide; rapport longueur/largeur de la fenetre vestibulaire de 1.4; absence d’art&re stapediale. Pucadelphys ne possede pas

de processus tympanique de LalisphenoVde mais possede une petite lame anterieure du petromastoVde et un grand foramen ovale

(sortie du V3) qui s’ouvrc entre la lame anterieure et LalisphenoVde. La presence, chez Pucadelphys , d'unc lame anterieure du

petromastoVde fusionnec a la pars petrosa et associee a la presence d'un alisphenoVde dilate antero-posterieurement, scmble

contredire L hypothese de Presley & Steel (1 98 1) et Presley (1981) sur revolution de Lorcillc moyenneet de la paroi cranienne

des tribosphenides. fondyesiirdesetudcsontogcnctiques. II s'agildu premiersignalementd'une lame anterieure du petromastoVde

chez un mammilere tribosphenidc fossilc. La structure des molaires de Pucadelphys cst indiflerenciablcdccclle des membres de

la famille des Didelphidae ct ce genre est range dans cettc famille au sein de I'ordrcdes Didelphimorphia. Pucadelphys represente

un exemple classiquc devolution en mosaVque ct i I lustre bien que nous avons encore beaucoup a apprendre sur les debuts de

revolution des meiathcricns ct des iribosphynides.

RESUME DEVELOPPE

Les os craniens et basicranicns de mammi teres sont des elements ires i mportanls dans les reconstructions phy logenetiqucs.

Toutefois. ties pen de cranes de mammifercs tribosphenides sont connus pour le mcsozoVque et le Paleoccne inferieur et aucun

crane de metatherien incomes'* n’est connu au mesozoVque. Plusieurs squelettes et cranes d’un marsupial Didelphidae

{Pucadelphys andinus) ont etc decouverts dans le Paleoccne inferieur de la Formation Santa Lucia (Bolivie) ct constituent, avec

cclui de Mayuleslesferox (un borhyaenoide provenant du memegisement), lcs plus anciens cranes complets connus ratlaches avec

security a ce groupe de mammifercs.

Dans ce travail nous presenlons une description detaillcc des cranes de Pucadelphys andinus Marshall & Muizon

(1988). nous discutons scs principaux caracteres craniens et les analysons dans une optique orientee vers la comprehension de la

phylogenie des tribosphenides (Marsupiaux et Placentaires). Sont deems successivemcnt les dents superieures, les dents

inferieurcs. le dentaire et le crane.

Famille Didelphidae

Pucadelphys andinus Marshall & Muizon. 1988

DESCRIPTION

La formule dentaire de Pucadelphys andinus est I 5/4. Cl/1. P 3/3. M4/4. Les incisivcs superieures sont disposees en serie

continue et sont jointives. Elies sont croissantcsdc 11 a 14 et I5eslrcduite. Les prcmolairessontbiradiculeeset presentent une forte

augmentation de taille de PI a P3. Elies ont une couronne simple constituee d'unc grandecuspidc triangulairc et d'un minuscule

talon postericur. La P3 cst haute et sa couronne descend nettement plus bas que le plan occlusal des molaires. Les molaires

superieures sont triangulaircs ct presentent un paracone plus petit que le mctacone (pour Ml-M3), une ccntrocrista en V ct un

protocone relati vement gros avec un bord posterieur renfle. La metacrista cst petite mais legcrement plus grande que la paracrista.

La plateforme stylairc est dcveloppee ct possede des styles bien marques. Les incisivcs inferieures sont de taille croissante de i I

a i4 et sont jointives. Les premo I ai res inferieures sont semblables aux superieures. Lcs molaires inferieures possedent un trigonide

eleve plus large que long. Le protoconide est legcrement plus grand que le metaconidc. Icquel est nettement plus devcloppe que

Source: MNHN, Paris

PUCADELPHYS ANDINUS: THE SKULL.

le paraconide. La protocristide, droite et tranchante, est orientce transversalemem. Le talonidc possede un bassin profond el des

cuspidcs bien marquees, entoconide el hypoconulide etant distinciemeni accolds. Le deniaire presente une branche horizontal

longue el gracile, une fosse massdtdrine peu profonde, un vastc processus coronoi'dcet un condyle situe au-dessus du plan occlusal.

Le crane possede un museau court el obtus. Les arcades zygomatiques sont grandes cl monirent une orbite en communication avec

la fosse tempo rale. 11 possede une legere Crete sagittale parietale et une crete lambdoi'de marquee. Les nasaux sont allonges et

presentent un forte dilatation de lcur extremitd posterieurc. Les premaxillaires sont gracilcs el sont performs ventralement par des

foramens incisifs allongds, lesquels affectentdgalement les maxillaires. L'arcdentaireest presque semi-circulaire. Les maxillaires

possedent uneouverturedu canal infraorbitaire situee au-dessus de P3. Lepalatin possede un fort torus postpalatin situe en arricre

de M5. En vue vcntrale, maxillaires et palatins ne presentent pas de foramens palatins. Le lacrimal forme le bord anterieur de

1’orbite: il est grand et s'etend largemcnt sur la face dorsale du rostre. Le jugal est un os gracilc, sigmoi'de qui participc a

I’elaboration de la cavitd glenoVde. Le frontal occupe environ le tiers du crane en vue dorsale et laterale. II porte parfois un petit

processus supraorbitaire el participe a la formation du foramen ethmoi'dicn. Le parietal est une vaste plaque osseuse qui conslitue

la majeure partie du toil et des parois de la cavitd cranienne. L'orbitosphdnoi'de est un os petit qui possdde un grand foramen

sphdnorbitaire situe a sajonction avec le frontal et le palatin. L’alisphenoide forme la region anterieurc et laterale de la bo?te

cranienne. Cet os participc il la formation de trois foramens. Le foramen ovale est dnormc et se situe entre le bord posterieur de

Palisphenoidc et le bord anterieur du pelromastoidc, medialemcnt & la cavitd glenoVde. Antcro-medialemcnt. le foramen

entocarotidien est borde latdralement par I’alisphdnoide et medialement par le basisphdnoide. Un foramen rotundum bien marque

s'observe antcrieuremcni. II n’existe pas de processus tympanique de 1'alispheno'ide. Le squamosal presente une cavitd glenoidc

profonde et allongee transversalemem. munie d'un grand processus postgldnoide et d'un petit processus preglenotde. Le

squamosal posscdc quatre foramens sur sa face externe : 1) le foramen postzygomatique, sur le bord postcrieur de la racine

posterieurc de I’arcade zygomatique, dorsalement & I'apex du processus postgldnoide, 2) le foramen subsquamosal qui s’ouvre

dorsalement au mdat auditif externe, 3) le foramen postgldnoide. le plus grand, dont l’ouverture est situee sur le flanc posterieur

de la racine postdrieure de I'arcade zygomatique, 4) le foramen posttemporal qui s’ouvre entre le squamosal et la pars mastoidea

du petromastoi'de. Ventralement. Ic squamosal participe a la formation du bord lateral du foramen ovale. Le pelromastoidc est un

os complexe forme de la pars petrosa et la pars mastoidea. Cette derniere affleure largemcnt sur la face postdrieure du crane entre

I'exoccipital, le squamosal et le supraoccipital. Ventralement, la pars mastoidea participe a la formation de la moitic laterale du

petit processus exoccipital et, lateralement, forme un processus mastoide relativement grand. Le bord ventral de la pars mastoidea

forme une levredirigeeantdro-ventralementquiconstitue le processus tympanique caudal (sensu Wible, 1990). II forme une partie

du plancher du sinus mastoide dpitympanique. Une cchancrurc stylomastoidienne bien marquee s’observe lateralement au

processus mastoide; elle donne passage au nerf facial et a la veine laterale de la tete. Le bord lateral du processus mastoide s’Ctend

antero-dorsalcment, separant le bord lateral du sinus mastoide dpitympanique de la fossa incudis. situde dans la region postdrieure

du reccssus dpitympanique. La pars petrosa presente un grand promontoirc lisse et piriforme dont I’apex est dirige antdro-

medialement. II est borde antero-mcdialemcnt par un sinus marque correspondant au passage de la carotide interne se dirigeant

vers le foramen carotidicn; lateralement au sinus carotidicn se trouve la fosse du muscle tensor tympani. Le bord medial du

promontoire presente un sillon interne marque pour le passage du sinus pdtreux infdrieur. Sur la face laterale du promontoire. la

l enetre ovale se situe a la hauteur du bord postero-dorsal du meat auditif externe; son rapport longucur/largeur est de 1.4. Antdro-

lateralemcnt & la fenetreovale,dans le sillon qui borde latdralement lcpromontoire.se trouve I’ouverturcpostdrieurcdu canal facial

secondaire. Latdralement a cette ouverture. se trouve un petit foramen. I’ouverture mediale du canal prootique. Lc rdeessus

epitympanique est une petite depression allongee situee latdralement au canal facial; son extremitd posldrieure est la fossa incudis.

Sur la face dorsale ou cdrdbelleuse du petromastoVde s’observe une grande fossa subarcuatact le meat auditif interne. Quatre sillons

entourent le corps du petromastoi'de sur sa face cdrdbellaire : I) ventro-medialement, le sillon pour le sinus pdtreux inldrieur. 2)

postdro-ventralement. un sillon vertical pour le sinus juguiaire. 3) postero-dorsalement une petite poche reprdsente le sillon pour

le sinus sigmoi'de. 4) le long du bord dorso-latdral se trouve un sillon allonge pour le sinus prootique. Antero-latdralement & la crista

petrosa, dans une aile osseuse se projetant antero-dorsalement. sc trouve une depression pour le lobe temporal du cervcau et lc

ganglion trigdmine. Chez les autres Didelphidac Ic ganglion repose sur I’alisphdnoide et il n’existe pas de projection osseuse. Cette

aile antdro-laterale du petromastoi’de contribue au plancher et au bord interne de la paroi cranienne. Elle reprdsente apparemment

une rclique de la lame antdrieure du pdtreux (sensu Crompton & Jenkins. 1979). Sur la face vcntrale du petromastoi'de. un contact

irrdgulier s’observe entre le bord antdro-lateral du promontoire et la plateforme osseuse qui conslitue le bord posterieur du foramen

ovate. Cette portion reprdsente aussi probablemcnt une partie de la lame anterieurc du pdtromastoi'de. Sur la lace laterale de I os.

on observe un sillon profond et large qui recevait lc sinus prootique: a son extremitd ventralc. un petit loramen correspond au

dcbouche lateral du canal prootique.

LARRY G. MARSHALL & CHRISTIAN de MUIZON

DISCUSSION

La comparison des cranes de Pucadelphys est problematic du fait dc Pabsence de cranes de marsupiaux fossiles du Tertiaire

ancien ou du Cretace qui soient raisonnablcment bien connus. Asiatherium du Cretace supcricur dc Mongolic (Szalay, 1993) est

probablement un marsupial mais sa description detaillee est en cours et, de surcroit. cette espece est loin dc faire Punanimite quant

a son attribution aux metathericns. Mayulestesferox Muizon (1994) du mcme age et du meme gisement que Pucadelphys , est connu

par un crane remarquablement complet mais sa description detaillee est cn cours. En revanche il existe de nombreux restes

dentaircs dc marsupiaux dans le Cretace supcrieur d'Amerique du Nord et dans Ic Palcoccnc d’Amerique du Sud. La discussion

comprcndra done une premiere partie qui comparer les dents de Pucadelphys aux formes voisincs d'Ameriques du Nord et du

Sud. et une seconde partie qui passera cn revue les principaux caracteres craniens de Pucadelphys afin d’en determiner la polarity

et qui tentera d’analyser les modalites dc revolution du crane chez les Theria et plus precisement chez les Tribosphcnida.

COMPARAISON DENTAIRE.— Parmi les Didelphidae fossiles d'Amerique du Sud. Sternbergia itaboraiensis duPa\eocene

“moyen” d'ltaboraf (Brcsil) est Pespece qui se rapproche le plus de Pucadelphys andinus par la taillc et la structure de ses dents

a tel point que la premiere constituc un ancetre morphologique ideal pour la seconde. Pucadelphys et Sternbergia , en plus de leur

structure dentairc presque identique. possedent la meme morphologic du protoconc renfle posterieurement. et le meme important

developpement des styles B et C. Les molaires de Sternbergia different toutefois de cclles de Pucadelphys par leur bassin du

trigone, moins profond. leur paracone. plus reduit, leurcentrocrista formant un V plus marque. leur plateforme stylairc. plus petite,

leur talonidc. plus large, leur trigonide. plus bas. leur hypoconulide et entoconide. plus grands et par la presence d’un cingulum

labial.

Deux autres Didelphidae d’ Itaboraf presentent d’ importantes ressemblances structurales avec Pucadelphys ; ce sont: Mar mo sop sis

juradoi et Itaboraidelphys camposi. Bien que dans ces cas la resscmblance ne soit pas aussi frappante que pour Sternbergia ,

Pucadelphys pourrait egalcment representer un bon ancetre morphologique pour ces deux genres.

Pucadelphys montre egalement d’ interessantes ressemblances avec plusieursesp£ces d'Alphadon et de Protalphadon d’ Am6rique

du Nord, genres classes parmi les Peradectidae par Marshall et al. (1989). Les trois genres presentent des structures dentaires

similaires mais Pucadelphys differe des deux formes nord-americaines par la possession d’une centrocrista en V. d’un paracone plus

petit que le metacone (ils sont sub-egaux chez Alphadon et Protalphadon) et d’un protocone a flanc postericur renfle, trois caracteres

de Didelphidae. En fait, Alphadon tnarshi represente un bon ancetre structural pour les Didelphidae (Clemens, 1966) et done pour

Pucadelphys . Cette interpretation confirme Phypothese dc Clemens (1966) et de Crochet (1980) qui pensent que les Didelphidae

pourraient avoir leur origine parmi les Peradectidae nord-americains. II est d'ailleurs interessantdeconstater que certains specimens

d 'Alphadon marshi ont une centrocrista en V et un paracone plus petit que le mdtacone (Cifelli, 1990 : 315).

ANALYSE DE CARACTERES. — Sont passes en revue et discutes ci-dessous les etats de caracteres qui paraissenl importants

pour la comprehension de revolution des mammiferes. Ils seront presentes dans Pordre suivant: dentition, denlaire, os du crane,

foramens craniens. et region auditive.

CARACTERES DENTAIRES

Nombre d'incisive s. — Pucadelphys presentc Petal plesiomorphc qui est la possession de 5/4 incisives.

Structure des incisives superieures. — Pucadelphys est plcsiomorphe par la morphologie conique de ses incisives qui

sont d’une hauteur subegale.

Structure des incisives inferieures. — Tous les Didelphidae actuels ont une i3 dont la racine est ddcalee en quinconce

vers Parriere. Cette disposition est absente chez Pucadelphys qui retient done Petat plesiomorphc pour ce caractere.

Nombre de molaires et premolaires. — Pucadelphys presentc P3/3 et M4/4, ce qui constitue Petat plesiomorphc de ce

caractere au sein des marsupiaux.

Structure des molaires et premolaires. — Avec une centrocrista en V, un paracone plus petit que le metacone et un

protocone renfl£ postiSrieurement, Pucadelphys est en parfait accord avec la structure dentairc observee chez les Didelphidae.

DENTAIRE

Processus angulaire inflechi. — Pucadelphys , qui possedc une telle structure, presente Petat plesiomorphe de ce

caracfere au sein des Tribosphcnida.

Si lion mylohyoidien. — La presence d’un sillon mylohyoYdicn chez Pucadelphys est une plesiomorphie au sein des

mam mi leres.

Foramen mandibulaire labial. — L'absence de ce type de foramen chez Pucadelphys represente Petat apomorphe de

ce caractere puisqu'il existe chez la pluparl des mammifcrcs primitifs comme Kielantherium . Prokennalestes , Oetlestes, et

Zalambdalestes. Sa presence chez Microbiolherium gallegoense. un Microbiotheriidae du Miocene infcricur d’Argentine,

constituc probablement une reversion.

PUCADELPHYSANDINUS : THE SKULL

OS DU CRANE

Orbite grande el confluente avec la fosse temporale. — Pucadelphys , qui presente la condition observce sur tous les

cranes de theriens du Cretace. est plesiomorphe pour ce caractere.

Lacrimal. — Un lacrimal possedant une grande aile faciale est un etat plesiomorphe present chez Pucadelphys.

Contact nasal-lacrimal. — Un contact maxillaire-frontal est fetat plesiomorphe de ce caractere qui sc rencontre chez

les cynodontes, les tritylodontes, les multitubercules, Morganttcodon. Vincelestes, et Deltatheridium. II existe egalement chez les

Borhyaenoidea (probablcment un etat plesiomorphe)el chez le vombatoide Wynyardia( probablementunereversion). Pucadelphys,

avec un net contact nasal-lacrimal, presente done I’etat derive de ce caractere.

Contact alisphenoide-parietal. — Pucadelphys presente un large contact entre falisphenoide et le parietal, ce qui

constitue fetal plesiomorphe de ce caractere.

Fosses palatines. — L’ absence de fosses palatines chez Pucadelphys est un etat plesiomorphe qui existe egalement chez

les cynodontes. les monotremes, les multitubercules, Vincelestes , Deltatheridium, et chez de nombreux metatheriens.

Processus preglenotde du frontal. — Cette piesiomorphic probable des Tribosphcnida est presente chez Pucadelphys.

FORAMENS DU CRANE

Canal transverse. — Cette structure est absente chez Morganucodon , les monotremes. les multitubercules, les

Deltatheroida. les Borhyaenoidea, certains Didelphidae, certains Dasyuridae et Pucadelphys, et constitue fetat plesiomorphe au

sein des mammiferes. Un canal transverse est present chez tous les autres marsupiaux.

Foramen ovale. — Le terme de foramen ovale est ici utilise pour designer le passage de la branche mandibulaire du

trijumeau ( V3) independamment des os qu’il traverse. Le terme de foramen pseudovale utilise par plusieurs auteurs (MacIntyre,

1967; Archer, 1976a) est extremcment confus car souvent utilise avec des sens differents.

Foramen subsquamosal. — La presence de ce foramen chez Pucadelphys est un etat plesiomorphe retrouve chez tous

les metatheriens et chez les eutheriens cretaces, Asioryctes et Kennalestes.

Foramen postzygomatique. — La presence de ce foramen chez Pucadelphys est une piesiomorphic qui sc retrouve chez

tous les marsupiaux.

REGION AUDITIVE

Bulle tympanique ossifiee. — Pucadelphys ne possede pas de processus tympanique de I’alisphenoide. L'absence de

bulle tympanique ossifiee represente apparemment Petat plesiomorphe chez les mammiferes (Novacek, 1977). L’abscnce de

processus tympanique de I’alisphenoide est une piesiomorphic chez Pucadelphys (et non une reversion, commc cela lul suppose

par Marshall & Kielan-Jaworowska, 1992) et P etude du crane de Mayulestes, qui ne possede pas non plus de processus

tympanique de Palisphdnoide, suggere que ce caractere est apparu plusieurs fois au cours de revolution des marsupiaux (Muizon,

1994).

Orientation de Pectotympanique et de la membrane tympanique. — Les avis divergent sur la position primitive de

Pectotympanique. Pour Novacek (1977), la position primitive est subhorizontale car ccttc disposition est presente chez les

monotremes et chez Lambdopsalis et parce que Pectotympnique est subhorizontal dans les stadcs prdcoces du devcloppement

ontogeniquede tous les mammiferes. Pour Kielan-Jaworowska (1981) Pectotympanique est primitivcment oriente a 45° car e’est

la position de Pangulaire (= ectotympanique) chez les cynodontes et chez Morganucodon. De plus, Pectoympanique des

eutheriens Asioryctes et Kennalestes, qui fut retrouve in situ sur certains specimens, presente une position a 45° de Phorizontale.

Bien que Pectotympanique de Pucadelphys soit inconnu. la position de la fenetre ovale laisse supposer qu’il avait une position

subverticale comme chez tous les autres marsupiaux. Quel que soit Petat plesiomorphe de ce caractere, la condition de Pucadelphys

serait done dcrivee.

Sinus auditif. — L'absence de sinus auditifs chez Pucadelphys , qui sont en general abondants et vastes chez la plupart

des autres marsupiaux, est un etat plesiomorphe au sein du groupe.

Contribution de la pars mastoidea a /' occiput. — Cet etat, present chez Pucadelphys ainsi que chez de nombreux autres

mammiferes, est apparemment une piesiomorphic pour les Tribosphcnida.

Mastoide et processus paroccipital. — L’absence de processus paroccipital el la petite taille du processus mastoide,

comme cela sc rencontre chez Pucadelphys, representent apparemment un etat primiti I pour les mammiferes.

Forme de la fenetre ovale et de la base du stapes. — Avec un rapport longueur/largeur de la fenetre ovale de 1,4,

Pucadelphys est voisin de la condition plesiomorphe pour les marsupiaux.

Anere stapediale. — L’artere stapediale est presente chez les embryons de tous les mammiferes actuels mais seulement

chez les adultes d'Ornithorhynchus et d’eulheriens. Son parcours, marque par un profond sillon sur le promontoire, a etc mis en

evidence chez les multitubercules et chez de nombreux eutheriens fossiles. L'artcre stapediale disparatt chez les marsupiaux

adultes, lesquels ne presentent pas de sillon sur le promontoire et sont caract6rises par cctte synapomorphic. La presence d’une

artfcre stapediale chez I’adulte est une piesiomorphic chez les mammiferes et rabsence de sillon du promontoire (perte de f artcre

stapediale) chez Pucadelphys serait indicateur d’un dtat apomorphe au sein des mammiferes mais plesiomorphe au sein des

marsupiaux.

LARRY G. MARSHALL & CHRISTIAN de MUIZON

Coins de la carolide interne. — Lc cours de ce vaisseau chez Pucadelpliys esl situc sur le bord medial du promontoire.

Ceiic disposition est egalement presente chez tous les autres marsupiaux. chez les monotremcs. chez les eutheriens du Crctacd

supcricurd' Asie el chez certains eutheriens aciucls (rongeurs, lagomorphes. ongulcs) el constituc I'etal plesiomorphe (Presley.

1979). L'etat apomorphe. ou la carotidc internees! laterale au promontoire. existechez tous les autres eutheriens et a du apparaitre

plusieurs I'ois au cours de Involution. Pucadelpliys retient done Petal plesiomorphe pour ce caractcre.

Sillon pour le nerf facial. — La presence d'un sillon pour le nerf facial, chez Pucadelpliys. est unc plesiomorphie au sem

des mammiferes.

Sinus et foramen pitreux infirieurs . — La presence de ccs structures che/. Pucadelpliys constituc une plesiomorphie au

sein des mammi feres.

Lame antcrieure. — Le petromastoide de Pucadelpliys possede unc lame antcro-laferale qui participe a la formation de

la paroi interne de la boTte cranienne. A la hauteur de cette lame, la paroi externe du crane est formee par I’alisphenoide et le

squamosal. Mcdialemcnt cette lame possede une fosse qui recevait tres probablemcnt le ganglion trigdmine. Chez des mammi feres

primitifs comme Morganucodon. Haldanodon , les multilubercules, Vincelestese t les monotremcs, le petromastoide presente une

grande lame anferieure qui participe seule a la formation de la paroi laterale du crane. Elle s’article anferieurement avec le bord

posterieur de I'alisphenoide et sa face mediale presente une fosse qui recevait lc ganglion trigeminc. Chez les marsupiaux et les

placentaires actuels. il n’existe pas de lame anierieure du petromastoide et, seul, I’alisphenoide participe a la construction de la

paroi laterale du crane. Wible (1990) a cepcndant note qu’il existait parfois chez certains marsupiaux, une structure extrememenl

reduite qui pourrait representer unc relique de la lame antcrieure des mammiferes primitits. La structure laminaire observee chez

Pucadelpliys , laquellc possede sur sa face mediale la meme fosse pour lc ganglion trigemine que cclle observee chez les

mammiferes non-tribospheniques cites plus haul et qui occupe la meme position que chez ces formes, est ici consider^ comme

homologue de la lame antcrieure. La presence chez Pucadelpliys d' unc lame antcrieure encore bien developpce, quoique nettement

plus reduite que chez les mammi feres non tribospheniques, semble demontrer que la reduction et la disparition de la lame anierieure

du pelromasto'jde esl une synapomorphie des Tribosphcnida, comme 1’ avail mentionne Wible (1990)- Pucadelpliys rcpresentcune

etape intermediaire dans cette reduction; chez ce genre, la lame antcrieure est expulsec de la face externe de la paroi laterale du

crane par r apparition d'une articulation du squamosal avec I’alisphenoide mais reste toujours presente. bien que reduite, sur la

face interne de la paroi laterale du crane. A ce stadc. la paroi laterale du crane, dans la region anierieure du petromastoide, est done

une double lame dont la portion laterale est formee par le squamosal et I'alisphenoide et la portion mediale par la lame antcrieure

du petromastoide. Lc stade suivant (marsupiaux et placentaires actuels) voil la disparition de la lame antcrieure. Pucadelpliys est

done la demonstration paleontologique de la polarite d' un caractcre important de Involution de la paroi cranienne des mammi feres

et confirirte les hypotheses de Wible (1990) et de Hopson & Rougilr (1993).

CONCLUSIONS

Le crane dc Pucadelpliys represent un cas classiquc devolution en mosaique. Pucadelpliys possede de nombreux caracferes

plesiomorphes de marsupiaux. de Tribosphcnida, de theriens et dc Mammiferes. II possede neanmoins des caracferes apomorphes

qui tendent a le rapprocher de la famille des Didclphidae. La presence d’une lame anierieure du petromastoide est un caracterc

plesiomorphe pour un marsupial qui tendrait a en faire le groupe frerc de tous les autres marsupiaux et done h le placer dans un

taxon nouveau; en fait Pucadelpliys presente deja I'apomorphie des Tribosphcnida consistant en une reduction de la lame

antcrieure du petromastoide. unc structure qui a pu disparaitre plusieurs fois au cours de revolution de ce groupe. L’absence de

processus tympanique de I'alisphenoide (structure longtemps consideree comme une synapomorphie dcs marsupiaux) tendrait

egalement a en faire le groupe frere plesiomorphe pour ce caracterc dc tous les autres marsupiaux ou a envisager une reversion.

En fait, retude de Mayidestes ferox, un borhyaenoide du meme gisement que Pucadelpliys et ne possedant pas de processus

tympanique de I’alisphenoide. a montre que cette structure ctait sans doute apparue plusieurs fois au cours de revolution des

marsupiaux et ne constituc pas un obstacle a la classification dc Pucadelpliys parmi les Didelphidae.

Les seuls caracferes derives, au sein des marsupiaux. que possede Pucadelpliys sont ccux dc la structure de ses molaircs

superieures qui lc rangent sans equivoque au sein des Didelphidae.

PUCADELPHYS ANDINUS: THE SKULL

INTRODUCTION

Skulls and basicranial bones of mammals are very important in phylogenetic studies, yet

these elements are extremely rare in rocks of Cretaceous and earliest Paleocene Age. Because of

a dearth of such specimens, little is known of the early cranial evolution of Tribosphenida

(metatherians, eutherians and their common ancestors with tribosphenic dentitions; sensu

McKenna, 1975). In fact, the character states of skull features in the direct common ancestor of

metatherians and eutherians are unknown, yet inferences about these states have been made based

on study of a few late Cretaceous specimens and of Cenozoic and living taxa (Kemp, 1982; 1983).

To date, there are only two non-tribosphenid therians for which cranial material is known.

One is a eupantothere from the late Jurassic of Portugal, Henkelotherium guimorotae which is

represented by a very partial skull associated with a nearly complete skeleton (Henkel & Krebs,

1977; Krebs, 1987; Krebs, 1991), and the other is the eupantothere Vincelestes neuquenianus

from the early Cretaceous of Argentina, which is known from nearly complete skulls and skeletons

which have not yet been fully described (Bonaparte & Rougier, 1987; Rougier & Bonaparte,

1988; Rougier et ai, 1992; Hopson & Rougier, 1993).

Eutherians are the best known of late Cretaceous tribosphenids, and complete or partial

skulls of Asioryctes, Barunlestes, Kennalestes and Zalambdalestes have been described from the

?late Santonian and/or Campanian of Asia (Kielan-Jaworowska, 1981; 1984; Kielan-Jaworowska

& Trofimov, 1980).

In sharp contrast, basicranial remains of metatherians from the late Cretaceous are

extremely rare. Dcltatheroida, the possible plesjomorphic sister-group of marsupials (Kielan-

Jaworowska & Nessov, 1990; Marshall & Kielan-Jaworowska, 1992) are known by several

partial skulls and by one almost complete skull (e.g. Anonymous, 1983), although the basicranial

region has not been described. Specimens of metatherians which have been described include a

right ventrolateral corner of a skull of Eodelphis brpwni (Stagodontidae) from the Belly River

Formation of Alberta (Matthew, 1916); a left ventrolateral corner of a skull of Didelphodon vorax

(Stagodontidae) from the Lance Formation of Wyoming (Clemens, 1966; a second ear region

described by Clemens as D. vorax is a multituberculate; see Wible, 1990); a nearly complete left

petromastoid from the Hell Creek Formation of Montana which Archibald (1979) described as

either Pediomys hatched, P.florencae or Alphadon rhais ter but which Wible (1990) calls Petrosal

Type D (possibly A. rhaisterf, and ten petromastoids from the late Cretaceous ol North America

which Wible (1990) refers to as Petrosal Type A (one specimen, probably Pediomys hatched or

P.florencae), B (one specimen, possibly P. hatched or P.florencae), and C (eight specimens,

indeterminate). Recently Trofimov & Szalay (1993) gave a short presentation of a new' marsupial

from the late Cretaceous of Mongolia. The specimen is an almost complete skeleton of a

Monodelphis scalops -sized animal which has a marsupial dental formula and marsupial features

in its dental morphology. In the early Paleocene of the Santa Lucia Formation (Bolivia) another

specimen of early Metatheria is a partial skeleton of a new borhyaenoid. Mayulestes ferox,

Muizon, 1994, which includes a very complete skull. As mentioned in its preliminary description,

the skull of Mayulestes bears several features that question the traditional definition of Metatherians.

LARRY G. MARSHALL & CHRISTIAN de MUIZON

In this part of the volume we describe nearly complete skulls of the metathcrian Pucadelphys

andinus Marshall & Muizon (1988) from the early Palcocene age Santa Lucia Formation (sensu

Gayet et al. 1992) at Tiupampa in southcentral Bolivia (Fig. 1). With a nearly complete

undescribed skull of a deltatheroid from the late Cretaceous of Asia (Anonymous, 1983; Kielan-

Jaworowska & Nessov, 1990), the marsupial skeleton mentioned above (Trofimov & Szalay,

1993), and the skeleton of Mayulestes, these are among the earliest complete skulls of metatherians

(or possible metatherians) yet known. As shown below, these skulls reveal numerous character

states which have important bearing on the phylogenetics of Metatheria, in particular, and on

Tribosphenida, in general. The postcranial skeleton is described in the part III of this volume

(Marshall & Sigogneau-Russell, 1995).

Abbrevations of Institutions. — MNHN, Institut de Paleontologie, Museum National d’Histoire

Naturelle, Paris; YPFB Pal, paleontology collection of Yacimientos Petrolfferos Fiscales de

Bolivia in the Centro de Tecnologt'a Petrolera, Santa Cruz, Bolivia.

SYSTEMATIC PALEONTOLOGY

Serial designation for teeth follows Luckett (1992) contra Archer (1978), Hershkovitz

(1982), Marshall & Muizon (1988) and Muizon (1992) (i.e. premolars are PI, P2, P3; deciduous

tooth is dP3; permanent molars are Ml, M2, M3, M4); terminology for molar structure is shown

in Figure 5; usage of Metatheria follows Marshall el al. (1989), [although we admit that the

Deltatheroida could possibly be included in this taxon as recommended by Kielan-Jaworowska

& Nessov (1990) and Marshall & Kielan-Jaworowska (1992); usage of Theria follows Kielan-

Jaworowska et al. (1987), and Tribosphenida follows McKenna (1975); suprageneric ranks of

Metatheria follows Marshall el al. (1989). All measurements are in millimeters (mm). Abbre¬

viations of teeth are as follows: c, lower canine; C. upper canine; i, lower incisor; I, upper incisor;

m, lower molar; M, upper molar; p, lower premolar; P, upper premolar.

Legion TRIBOSPHENIDA McKenna, 1975

Infraclass METATHERIA Huxley, 1880

Order DIDELPHIMORPHIA (Gill, 1872) Marshall el al. 1989

Family DIDELPHIDAE Gray, 1821

Genus PUCADELPHYS Marshall & Muizon, 1988

Type-Species. — Pucadelphys andinus Marshall & Muizon, 1988

Diagnosis. — Same as for type and only known species.

PUCADELPHYS ANDINUS: THE SKULL

Pucadelphys andinus Marshall & Muizon, 1988

Holotype. — YPFB Pal 6105, a nearly complete skull (missing part of left zygomatic arch)

and associated dentaries with alveoli of 12-5, and II and C1-M4 present on right side; alveoli of

11 and 13-5, crowns of 12 and Cl, roots of PI, P2 missing, posterior half of P3, and Ml-4 present

on left side; right dentary missing ventromedial edge of horizontal ramus with alveoli of il and

i4, root of i2, complete i3, complete c 1 -m4 (p3 not fully erupted); posterior part of left dentary

stylar

shelf talon

i-il-1

paracingulum paracone

\ paracrista / paraconule

stylar cusp A '

stylar cusp "B" s

stylar cusp "C" v ^

ectoflexus—..

stylar cusp D--

stylar cusp "E--

,,protocone

—preprotocrista

'-postprotocrista

metaconule

metacone

centrocrista

metacrista

RIGHT UPPER MOLAR

precingulid-^^

03 paracristid—

protoconid^

hypoflexid--

§ cristid obliqua--

00 hypoconid

postcristid

,paraconid

^prefossid

^metaconid

--protocristid

_ —entoflexid

— entocristid

'-entoconid

'-postfossid

'hypoconulid

postcingulid

LEFT LOWER MOLAR

Fig. 5. —Terminology for molar structure of a generalized marsupial (after Marshall, 1987).

Fig. 5. — Terminologie de la structure des molaires d'un marsupial primitif (d’apres Marshall, 1987).

LARRY G. MARSHALL & CHRISTIAN de MUIZON

with ill 1-4: and associated articulated partial skeleton of neck and thoracic regions (figured by

Marshall & Muizon, 1988. Figs 1. 3A).

Hypodigm. — The holotype, YPFB Pal 6105; YPFB Pal 6107. greater part of skull and

articulated dentaries with root of right Cl and complete right P1-M4 and p2- m4. complete left

CI-M4 and c 1 -m4, and palatal and basicranial part of skull crushed slightly internally; YPFB Pal

6108. greater part of dorsoventrally crushed skull with associated dentaries. with complete Cl-

M4, alveoli of i 1-4 and cl-m4 present (missing lips of p3 and m3 protoconid) on right side, and

root of Cl. complete P1-M4, root of cl, alveoli of pi, p2-3 present (missing all but bases of

crowns) and m2-4 present (m2 missing tips of para-and metaconid. m4 missing tip of paraconid)

on left side; YPFB Pal 6109. rostral part of skull with right C1-P3 and trigon of M2, left P1-M4

(figured by Marshall & Muizon, 1988, fig. 3B), greater part of left dentary with bases of c 1 and

p 1, and p2-m4 complete (figured by Marshall & Muizon, 1988, fig. 3C); YPFB Pal 6110, greater

part of skull and articulated dentaries with most of dentition (missing only 11-5, left P1 and tip of

left p2) and part of articulated thoracic region of skeleton (figured by Marshall & Muizon, 1988,

fig. 2); YPFB Pal 6470. anterior part of skull (dorsoventrally crushed) with left 12 complete,

alveoli of 11 and 13-5, complete C1-M4, right C1-M4 complete, basicranium with complete right

petromastoid and four articulated vertebrae, anterior half of right dentary with cl and p2-m2

complete, and alveoli of p 1, and greater part of left dentary with i 1 complete, alveoli of i2-4, c 1

complete, roots of pi, and p2-m4 present (trigonid of m4 missing); YPFB Pal 6471, premaxilla

with right C1 -P1. fragment of left maxilla with M2-4 complete and attached zygomatic arch with

glenoid fossa, exoccipital region of skull with three articulated vertebrae, anterior ends of left and

right dentaries with c 1-pl in each, and posterior part of left dentary with m3-4; YPFB Pal 6472,

right maxilla and premaxilla with complete C1-M4. complete right dentary with bases of i 1 -4, and

complete cl-m4; YPFB Pal 6473, left maxilla with complete Ml-4, greater part of left dentary

with root of i I, base of i2, complete i3-4. c 1 missing tip of crown, and virtually complete pi-m4;

YPFB Pal 6474, greater part of left dentary of juvenile with three incisors and alveolus of a fourth,

c I erupting. p2 and m2 complete, and trigonid of erupting m3; YPFB Pal 6475, right maxilla with

M1 -4 (M2 missing antcrolabial cornerof crown); YPFB Pal 6476, right maxilla with M2-4; YPFB

Pal 6477, right maxilla with M2-3 and parastyle of M4; YPFB Pal 6478. right maxilla with M2-

3; YPFB Pal 6479, anterior part of right dentary with cl-p3 complete and broken base of ml;

YPFB Pal 6480. fragment of a right dentary with posterior half of p3, complete m 1, and trigonid

of m2; YPFB Pal 6481, fragment of right dentary with m3-4 (trigonids broken on both); YPFB

Pal 6482. fragment of right dentary with ml missing trigonid and complete m2; YPFB Pal 6483

(=MNHN Vil 118), an isolated left m2 missing lip of paraconid (figured by Muizon et al. 1984,

fig. 8); YPFB Pal 6484 (=MNHN Vil 127), an isolated left m4 missing tip of metaconid and

posterior edge of entoconid (figured by Muizon el al ., 1984. fig. 9); YPFB Pal 6485. fragment ol

left maxilla with broken roots of Cl-Pl and complete P2-M4.

Source:

PUCADELPHYS ANDINUS: THE SKULL

Horizon and Locality. — The specimens of Pucadelphys andinus described here were

recovered from the Santa Lucia Formation (sensu Gayet et al., 1992) at Tiupampa, located about

95 km southeast of Cochabamba (see Marshall et al., part I of this volume).

Agf.. Tiupampan Land Mammal Age (sensu Ortiz Jaureguizar & Pascual, 1989;

Marshall, 1989), early Paleocene (Van Valen, 1988; Ga yet etai, 1992; Bonaparte et al ., 1993 ;

Muizon & Brito, 1993; Marshall et al., part I of this volume).

Diagnosis (skull morphology only). — 15/4 Cl/1 P3/3 M4/4; 11-5 in continuous series (no

diastem); i3 not staggered; cheek teeth structure as in Didelphidae: paracone smaller in size in

occlusal view and distinctly lower in lateral view than metacone; centrocrista V-shaped with apex

of V pointing labially toward stylar cusp C; stylar shelf well developed; stylar cusp B highest of

stylar cusps; stylar cusps A, C and D subequal in size and only slightly smaller than stylar cusp

B; parastyle spur-like, extends anteriorly and overhangs posterolabial edge of preceding tooth;

remnant ol mylohyoid groove; large medially inflected angular process; rostrum short; secondary

palate solid (no evidence of vacuities) and extends posteriorly to point behind M4; nasals narrow

anteriorly, much expanded posteriorly with W-shaped contact with frontals; small frontal-maxilla

contact; lacrimal with large facial wing; lacrimal canal large, opens within orbit; orbit confluent

with temporal fossa; jugal participates in formation of preglenoid process; glenoid fossa situated

posteriorly opposite anterior half of promontorium; subsquamosal foramen present; tympanic

area uncovered, no evidence of ossified auditory bulla or of ossified ectotympanic; tympanic

membrane apparently oriented obliquely at 45° or more from horizontal; no auditory sinuses;

piomontorium large, teardrop-shaped, inflated ventrally; sulci for facial nerve, sigmoid sinus,

inferior petrosal sinus, prootic sinus, and sphenoparietal emissary vein; internal carotid artery

medial in position; no evidence of transverse canal; prootic canal for lateral head vein (vena capitis

lateralis) present; petromastoid with anteroanterior laminar structure which is regarded here as

homologous to the anterior lamina of the non-tribosphenid mammals; foramen ovale (exit for V3)

located posteriorly opposite anterior part of glenoid fossa, rimmed anteriorly by alisphenoid and

posteriorly by anterior lamina of petromastoid; external acoustic meatus very small.

DESCRIPTION

Anatomical features are described in the following order; upper dentition, lower dentition,

dentary and skull .

11-5 (Figs 10 and 12). — The right 11 is preserved in YPFB Pal 6105 and the left 12 in YPFB

Pal 6470; both teeth have small rounded crowns. As shown by the alveoli of 11 -5 preserved on the

right and left side of YPFB Pal 6105 and left side of YPFB Pal 6470, a slight size increase occurs

from II to 14 (13 and 14 are subequal in size) and 15 is smaller than 14, being similar in size to I].

The edges of the alveoli of 11-4 are on the same level, while that of 15 is situated more dorsally

in the anteriormost edge of a large fossa in the posterior part of the premaxilla which accomodates

LARRY G. MARSHALL & CHRISTIAN de MUIZON

the tip of cl. r 1 -5 occur in a continuous sequence with no spaces between them or between the left

and right Il's. A distinct diastema (1.4 mm on left side of YPFB Pal 6105) separates 15 from

Cl.

Cl (Figs 10 and 12).—This tooth is large, pointed and arcs ventroposteriorly such that the

tip of the crown is about ventral to the posterior edge of the crown base. In cross section C1 is ovoid,

being longer than wide (2.2 mm x 1.2 mm at the base of right Cl in YPFB Pal 6105). Cl is

consistently larger than c 1 in specimens where both are preserved. Cl is either in direct contact

with PI or separated from it by a very small diastema.

P1 -3 (Figs 6 and 7). — These three teeth are all double-rooted; an increase in crown length,

width and height occurs from PI to P3. PI is considerably smaller than P2-3 (Table 1); in occlusal

view it is ovoid, a low ill-defined heel occurs posterobasally, and the crown tip lies ventral to the

posterior edge of the anterior root. P2 is slightly smaller than P3 (Table 1); the crown tip lies ventral

to the anterior edge of the posterior root, at about the same height as the occlusal surfaces of M1 -

4. A distinct cusp occurs posterobasally and a very small cuspule occurs anterobasally. Weak basal

cingula usually occur along labial and lingual edges of crown posteriorly, and there is a distinct

posterolingual swelling of the crown. P3 is the largest of upper premolars; the crown tip lies

medially below the center of the tooth and is considerably higher than the occlusal surfaces of M1 -

4. There is a distinct posterobasal cusp, an anterobasal cuspule (which is larger than that on P2),

a distinct posterobasal cingulum labially, and a weaker basal cingulum lingually which extends

from the anterobasal cuspule to the posterobasal cusp.

Ml-4 (Figs 6 and 7). — In average length M1=M2>M3>M4 and in average width

M 1<M2<M3>M4 (Table 1). The protocone is large, broad anteroposteriorly, and inflated basally,

particularly posterolingually. The preprotocrista is slightly shorter and more rectilinear than the

postprotocrista which has a distinct posterolingual bow on M1 -3. The trigon is distinctly basined

on Ml-3, less so on M4. The para- and metacone are well developed on Ml-3. The paracone is

slightly smaller in occlusal view and lower in lateral view than the larger and higher metacone on

M1 -3 (on M4 the paracone is very large and high and the metacone, very reduced to absent). The

metaconule is slightly larger than the paraconule on M2-3, subequal on M1. The centrocrista is

V-shaped (dilambdodont; sensu Crochet, 1980) with apex of V pointing labially toward stylar

cusp C. The para- and metacone are weakly V-shaped (not connate) with their apex pointing

lingually toward the para- and metaconule, respectively. The paracone, the centrocrista and the

metacone form a weak W-shaped structure in occlusal view. The paracrista is low, short,

rectilinear, and unites the paracone with a large stylar cusp B on Ml-4. The metacrista is low,

slightly longer than the paracrista on Ml-3 (particularly on Ml-2) and with a posterior bow (on

M2-M3). The stylar shelf is well developed on Ml-3 with an anteroposteriorly elongate basin

separating para- and metacone from the row of large stylar cups. The stylar cusp B is the largest

of stylar cusps; the stylar cusps A, C and D are also well developed, subequal in size and only

slightly smaller than the stylar cusp B on M1-2. On M3 the stylar cusp A is subequal to B, while

C and D are more reduced; on M4 stylar cusps A and B are fused into a large parastylc (other stylar

cusps are extremely reduced or absent). There is no distinct evidence of a stylar cusp EonMI-

4. A prominent parastyle, formed by spur-like stylar cusps A and B, extends anteriorly and

Source .

PUCADELPHYS AN DIN US: THE SKULL

Fig. 6. — Pucadelphys andinus. SE.V1 photos of left P3-M4 (A, labial: B. occlusal: C\ lingual views) of YPFB Pal 6470; and left

p2-m4 (D, labial; E, occlusal; F. lingual views) of YPFB Pal 6109. X 15.

Fig. 6. — Pucadelphys andinus. Photos MEB de P3-M4 gauches (A, vue labiale; B. vue occlusale; C. vue linguale ) d'YPFB Pal

6470; et p2-m4 gaudies (I), vue labiale; E, vue occlusale; F, vue linguale) d'YPFB Pal 6109. X 15.

Source: MNHN, Paris

LARRY G. MARSHALL & CHRISTIAN de MUIZON

i -1-1_i

m m

Fig. 7. Pucadelphys andinus. A-C. left P3-M4 (A, labial; B. occlusal; C. lingual views), based on YPFB Pal 6109 and 6470; D-

F. left p3-m4 (D. labial; E. occlusal; F, lingual views), based on YPFB Pal 6109 and 6473.

he. 7. —Pucadelphys andinus. A-C. P3-M4gauches (A, vue labiate; B, vue occlusale; C, vue linguale), d'apres YPFB Pal 6109

ei 6470; D-h. p3-m4 gauches (I), vue labiale; E, vue occlusale; F, vue linguale), d'apres YPFB Pal 6/09 el 6473.

Source: MNHN, Paris

PUCADELPHYS ANDINUS: THE SKULL

overlaps the postcrolabial edge (i.e. metastylar region) of the preceding tooth. A distinct

paracingulum extends along anterolabial edge of crown from stylar cusp A; it becomes very

narrow below the tip of the paracone (it disappears on M4) and continues lingually to disappear

below the paraconule. The ectoflexus is absent or very shallow on M1 and increases in depth from

M 1 to M3. M2 differs from M1 in being slightly wider, having a slightly deeper ectoflexus, and

a longer paracrista relative to metacrista. M3 differs from Ml-2 in having a distinct ectoflexus,

an enormous spur-like parastyle on which stylar cusps A and B are large and subequal in size,

smaller stylar cusps C and D, a metastyle area reduced in size relative to the more prominent

parastyle area, para- and metacrista more subequal in length, and a crown less bulbous and

narrower anteroposteriorly (i.e. more transversely elongate).

i 1 -4 (Fig. 8). — A left i 1 is preserved in YPFB Pal 6470, a right i3 in YPFB Pal 6105, a right

i4 in YPFB Pal 6472, and left i3-4 in YPFB Pal 6473. The i 1-4 have simple ovoid crowns and

increase slightly in size from i 1 to i4; i4 is in direct contact with cl.

c 1 (Fig. 8). — This tooth is large, pointed, arcs in anterodorsal direction. It is ovoid in cross-

section, longer than wide (1.5 mm x 1.0 mm in right YPFB Pal 6105). The axis of the tooth is set

at a slight oblique angle relative to the cheek tooth series (i.e. anterior edge labial, posterior edge

more lingual), cl is consistently smaller than Cl.

pi-3 (Figs 6 and 7). — All three teeth are two-rooted, pi is very small relative to p2-3, and

is separated from cl and p2 by small (0.5 mm) diastems. It is ovoid in occlusal view with the

highest point of the crown set above the anterior edge of the anterior root. p2-3 are subequal in

length and width (Table 1) although p2 tends to be slightly lower in height; both are distinctly

premolariform with well developed posterobasal heels. The crown tips arc about same level as

protoconid tips of ml-4. The p2 differs from p3 in having the principal cusp narrower

anteroposteriorly and inclined more anteriorly so that the tip occurs above the middle of the

anterior root (on p3 the anteroposteriorly broader primary cusp is more dorsally directed so that

tip occurs dorsal to posterior edge of anterior root); p2-3 occur in close succession without spaces

separating them or p3 from m 1.

m 1 -4 (Figs 6 and 7). — In average length m 1 <m2=m3>m4, in average width m 1 <m2=m3>m4

(Table 1). The labial side of teeth is higher than the lingual side, and the trigonid is well elevated

above the talonid. The trigonid of in 1 is relatively long (paraconid set anterolabially), while those

of m2-4 are foreshortened anteroposteriorly and distinctly wider than long (paraconid set more

lingually). The protoconid is the highest of the trigonid cusps and slightly higher than the

metaconid. The paraconid is smaller than the metaconid and inclined anteriorly. Protoconid and

metaconid are united by a straight trenchant protocristid. The talonid is broad and distinctly

basined on ml-4, being wider than the trigonid on ml-3, and narrower on m4. The hypoconid is

the dominant talonid cusp on ml-3; it is V-shaped in appearance, and considerably larger than

entoconid and hypoconulid in occlusal view. In lateral view of in 1-3 all three talonid cups are

subequal in height. A short cristid obliqua contacts the trigonid at the posteromedial surface of the

protoconid, labial to the protocristid notch. The hypoflexid is well developed on ml-4. Entoconid

and hypoconid are subequal in size in occlusal view on ml-3, and the hypoconulid is "twinned”

with the entoconid (i.e. set closer to entoconid than to hypoconulid). On m4 entoconid and

LARRY G. MARSHALL & CHRISTIAN de MUIZON

Source: MNHN , Paris

PUCADELPHYS ANDINUS: THE SKULL

hypoconulid are fused basally, and the hypoconid is distinctly elevated above the much lower

entoconid and hypoconid. A precingulid is well developed on ml-4; a postcingulid (i.e. shelf

extending ventrolabially from tip of hypoconulid across posterior surface of hypoconid) is well

developed on'm 1 -3 but absent on m4.

Dkntary (Figs 8 and 9). — The horizontal ramus is long and slender. The ventral border of

the bone is relatively flat between p 1 and m3, bending dorsally anterior to p2 and posterior to m3.

The deepest point occurs below m3-4, with ,a gentle decrease in depth from m3 to pi.

The articular surface of the symphysis is slightly rugose and covered by numerous vascular

foramina; a well-developed symphyseal ligament was apparently present. The long axis of the

symphyseal surface lies at an angle of about 40° to the horizontal axis of the dentary and extends

posteriorly to below p2. There arc typically two large mental foramina, one below pi and another

below m 1. A groove on the medial surface of the dentary, about 1/3 of the way above the ventral

edge, extends from the anterior edge of the angular process to below ml, becoming shallower

anteriorly. This represents the mylohyoid groove (Bensley, 1902) and marks the course of a

neurovascular bundle which probably included the mylohyoid artery and nerve. The angular

process is very large and inflected medially; it is situated far posteriorly, more or less below the

condyloid process. The mandibular foramen is situated dorsal to the anterior edge of the angular

process and below the middle of the coronoid process. The masseteric fossa is shallow with a

posteroventral border (between the condyloid process and the point directly ventral to the middle

of the coronoid process) developed into a broad laterally directed shelf. In dorsal and ventral

views, the posterior part of the dentary has a broad V-shape due to the large medially inflected

angular process and the large laterally directed expansion of the posteroventral part of the

masseteric border. The condyloid process is wide transversely; the portion projecting lateral to the

vertical plane of the dentary is about three times longer (transversally) than that projecting

medially. It occurs dorsal to the horizontal plane through ml-4, at about the same level as the tip

of c 1. Its articular surface is broadly convex dorsoposteriorly. The coronoid process of the

ascending ramus is broad anteroposteriorly, thin transversely with the anteroposterior axis in the

same vertical plane as the labial sides of ml -4. Its upper margin is rounded (convex dorsally) and

its posterior border is markedly concave posteriorly.

General Skull Structure (Figs 10-14 and 16). — There are five partial or nearly complete

skulls. The most complete and least distorted is the type YPFB Pal 6105 (Figs 10 and 18) which

is the basis for the reconstructions in Figs 10, 15 and 17 along with details incorporated from other

Fig. 8. — Pucadelphys andinus. Right dentary of YPFB Pal 6108 (A. lateral; B. dorsal; C, medial views). Stereophotos X 2.5.

Fig. 8. — Pucadelphys andinus. Deniaire droit d'YPFB Pal 6108 (A, vue laterale; B. vue dorsale; C. vue mediate). Photos

stereoscopiques X 2.5.

LARRY G. MARSHALL & CHRISTIAN de MUIZON

specimens as indicated. In YPFB Pal 6105 the nasals and the left part of the rostrum anterior to

the infraorbital canal are slightly crushed ventrally, but the remainder of the skull is virtually

undistorted. Specimens YPFB Pal 6107 (Fig. 14) and 6110 (Fig. 13) are crushed slightly

transversely, particularly in the basicranial and palatal areas. This crushing is most extreme in

YPFB Pal 61 10 where the basioccipital is missing, the exoccipitals are in near contact and the

Fig. 9. — Pucadelphys andinus. Reconstruction of right dentary of YPFB Pal 6108 (A, lateral; B. dorsal; C. medial views).

Abbreviations: ap. angular process of dentary: corpd, coronoid process of dentary; cpd, condyloid process of dentary;

DEN(ar). ascending ramus of dentary; DEN(hr), horizontal ramus of deniary; masf, masseteric fossa; menf, mental

foramen; mf, mandibular foramen (inferior alveolar foramen); mg. mylohyoid groove (=meckelian groove); psmf,

posterior shelf of masseteric fossa; sd, symphysis of dentary.

Fig. 9. Pucadclphysandinus. Reconstitution du dentaire droitd'YPFBPal6I08(A, vuelaterale; B. vuedorsale; C, vue mediate).

Abreviations: ap. procesus angulaire du dentdire; corpd. processus coronoide du dentaire; cpd. condyle articulaire du

dentaire; DEN(ar), tranche ascendante du dentaire; DEN (hr), tranche horizon tale du dentaire; masf fosse masseterine;

menf foramen mentonnier; mf foramen manditulaire (=foramen alveolaire inferieur); mg. sillon mylohyoidien (= si lion

de Meckel); psmf plateforme posterieure de la fosse masseterine; sd, symphyse du dentaire.

Source: MNHN. Paris

PUCADF.LPHYS ANDINUS: THE SKULL

anterior ends of the pars petrosa of the petromastoids are in direct contact. The fourth skull. YPFB

Pal 6108 (Fig. 16), is crushed dorsoventrally and the dorsal surface is displaced to the left relative

to the ventral surface. The rostral portion of the fifth specimen, YPFB Pal 6109 (Fig. 11), shows

virtually no distortion on the right side, although the nasals are slightly depressed anteriorly and

the posterior part of the left maxilla is separated from the rest of the skull.

In YPFB Pal 6105 the ratio of the condylar-premaxilla length (28.5 mm) and the breadth

between the outer edges of the zygomatic arches (approx. 19.5 mm) is about 0.68. The ratio

betweenthe condylar-canine length (25.0 mm) and the broadest point of the braincase measured

between the outer edges of the mastoid processes (11.0 mm) is 0.44.

The most distinct aspect of the skull in dorsal view is the short rostrum and overall bilobed

appearance, with a broad area between the lacrimals and another between the posterior bases of

the large zygomatic arches (Fig. 12). The narrowest point of the skull, excluding the nasal region,

is just anterior to the frontal-parietal suture, the site of the interorbital constriction. The orbit is

broadly confluent with the temporal fossa.

A weak sagittal crest occurs in all specimens, but is best developed in YPFB Pal 6105 where

it begins on the posterior part of the frontals; in the other specimens the sagittal crest occurs only

on the parietals and begins at the frontal-parietal suture. In all specimens the sagittal crest becomes

higher and more prominent posteriorly. A dorsoposteriorly directed lambdoidal crest of the

postparietal is well developed in all specimens. In ventral view (Fig. 12) the palate, including the

dentition, and the basicranium are similar in size and shape, being broadly triangular with their

apices directed anteriorly.

In lateral view (Fig. 12) the recess for the external acoustic meatus is small, broadly U-

shaped, and bounded posteriorly by a small posttympanic process of the squamosal; it is situated

immediately ventral to the subsquamosal foramen and dorsoposteriorly to the postglenoid

foramen. The glenoid fossa is situated far posteriorly, opposite the anterior extremity of the

promontorium of the pars petrosa. The foramen ovale is also situated far posteriorly at the

posteroventral corner of the alisphenoid opposite the anterior part of the glenoid fossa; it is rimmed

anteriorly by the alisphenoid and posteriorly by the anterior lamina of the petromastoid; its lateral

margin touches the squamosal suture. The occipital plate slopes upwards and slightly forwards

from the occipital condyles.

Nasal (Figs 11 and 12). — Anteriorly the nasals are elongate and subequal in breadth. At a

point approximately dorsal to P2 they flare sharply laterally and posteriorly toward the frontal-

maxilla suture, where their sutures with the frontals abruptly bend medially and, before meeting,

make a sharp bend anteriorly. The nasal-frontal suture thus has a distinct W-shape, with the base

of the W pointing posteriorly and the top anteriorly. The anterior-most edges of the nasals are

broken in all specimens, but as suggested in YPFB Pal 6105 (right side) and 6110 in which they

are most complete, they extended somewhat beyond their point of contact with the premaxilla and

were apparently slightly longer medially.

Premaxilla (Fig. 12). — In dorsal and lateral views the premaxilla is sharply V-shaped. The

ascending process is directed posteriorly between the nasal and maxilla and its posterior extremity

LARRY G. MARSHALL & CHRISTIAN de MUIZON

Source: MNHN. Paris

PUCADELPHYS ANDINUS: THE SKULL

Fig. 10. —Pucadelphys andinus. Skull of YPFB Pal 6105 (hololype) (A, dorsal; B, ventral; C, occipital; D, right lateral; E. left

lateral views). Stereophotos X 2.5.

Fig. 10. Pucadelphys andinus. Crane d'YPFB Pal 6105 {hololype) (A, vue dorsale; B. vue vent rale: C. vue occipitale; D. vue

late rale droite : F, vue late rale gauche). Photos stereoscop iques X 2.5.

Source: MNHN, Paris

LARRY G. MARSHALL & CHRISTIAN ni MUIZON

is dorsal to the middle of C1. The ascending process of the premaxilla forms the dorsolateral edge

of the nares and its lateral surface is slightly convex. The anterior edges of the two premaxillae

have a distinctly rounded (sharply convex) shape when viewed dorsally. In ventral view the

premaxilla-maxilla suture extends posteromedially from the anterior-most edge of the maxilla

(lateral to 14), across the fossa which receives cl. Medially to the anterior mid-point of Cl, the

suture bends anteromedially and reaches the lateral edge of the incisive foramen in its posterior

quarter. On the medial edge of the incisive foramen, the premaxilla-maxilla suture runs

posteromedially from the posterior quarter of the foramen to the median plan of the skull. The

paired incisive foramina are well developed and elongate anteroposteriorly, extending from a

point medial to 13, anteriorly and to a point medial to the anterior edge of Cl, posteriorly. The

anterior 75% of the bar separating the incisive foramina medially is formed by the premaxilla; the

posterior 25% is formed by the maxilla.

Maxilla (Figs 12 and 15). — A small wing of the anterolateral edge of the maxilla overlaps

the posterolateral edge of the premaxilla to a point labial to 15 and forms the lateral rim of the fossa

between 15 and Cl (Fig. 12). Posterodorsally, the maxilla is wedged between the nasal and

lacrimal, and has a small, but clear, contact with the frontal. From the posterior edge of the frontal-

maxilla.suture, the maxilla-lacrimal suture runs in an anterior direction, then antcroventrally along

Fig. 11 — Pucadelphys andinus. Rosiral pari of skull of YPFB Pal 6109 (A, dorsal: B. right lateral views). Stereophotos X 2.5.

Fig. I1. — Pucadelphys andinus. Portion rostrate du crane d'YPFB Pat 6109 (A. rue dorsate: B. vue tolerate droite). Photos

siereoscopiques X 2.5.

Source: MNHN. Paris

PUCADELPHYS ANDINUS: THE SKUI.L

the anterior rim ol the orbit, contacting the jugal at the lower edge of the orbit dorsal to the posterior

edge of M2. From that point, as seen in lateral and ventral views, the maxilla-jugal suture extends

in a straight line posteroventrally.

The anterior (rostral) opening of the infraorbital canal is large, slightly elongate dorsoven-

trally (the ventral edge is slightly broader than the dorsal edge), and opens dorsal to P3. The

posterior (orbital) opening is also large, slightly elongate transversely. The dorsal rim is formed

by the lacrimal, the ventral and lateral rims by the maxilla, and the ventromedial rim by a sliver

of the palatine which extends anteriorly between the lacrimal and maxilla into the infraorbital

canal, as is clearly seen on left side of YPFB Pal 6105 (Fig. 15). The infraorbital canal transmits

the infraorbital branch of V2, a branch of the infraorbital artery to the mesial side of maxilla and

premaxilla, and a small vein in Recent metatherians (Archer, 1976a).

Within the orbit the maxilla forms the floor and, along its medial edge, has a broad contact

with the palatine. The large postpalatine foramen is bordered by the maxilla laterally and by the

palatine dorsally, medially and ventrally (Fig. 15).

In ventral view(Fig. 12),the posteromedial edgeofthemaxillaextendsanteromedially from

the lateral side of the postpalatine foramen, roughly parallel to the lingual edges of M3-4, and

apparently has a broad transverse contact with the palatine medial to the embrasure between M2

and M3. This area of the palate is broken in all specimens and the suture itself is not visible;

however, the relatively complete palate in YPFB Pal 6105 shows no vacuities in the palatal portion

of the maxilla.

Palatine (Figs 10 and 12). — The palatine contribution to the secondary palate is preserved

only in YPFB Pal 6105 where it is nearly complete. The posterior edge of the palatine is a

transversely thickened bar, the postpalatine torus, which extends below the level of the palate and

which it encloses posteriorly. This torus lies posterior to the M4's and its maximum thickness is

located laterally, posterior to the protocone’s of the M4’s. A large ovoid postpalatine foramen

(which probably transmits the descending palatine vein, palatine artery and palatine branch of

cranial nerve V) occurs at the posterolateral corner of the palatine (as seen only on the left side

of YPFB Pal 6105). This foramen is clearly seen in anterior and posterior views, while in ventral

view only the antcriormost edge is visible since the foramen is tucked above the postpalatine torus

and opens vertically. The posterior surface of the postpalatinc torus is ornamented by three spines:

a lateral one on either side at the medial edge of the postpalatine foramen and a medial one at the

contact of the two palatines. These spines give the ventral surface of the choanal orifice a broad

m-shaped appearance. Immediately anterior to the torus and situated medially on each palatine is

a pair of shallow parallel sulci (palatine sulci) which run anteriorly and probably extended to a

middle palatine foramen as occurs in living Didetphis, although this foramen is not preserved in

specimens of Puccidelphys. The anterior extent of the palatine is not fully preserved and the

presence or absence of vacuities cannot be securely established, although if vacuities were present

they were small. Given the relatively complete palate of YPFB Pal 6105, and the apparent absence

of vacuities, the palate is shown as solid in the reconstruction (Fig. 12).

Part of the orbital or laterovertical process of the palatine is preserved in YPFB Pal 6105 (left

side), 6108 (left side) and 6110 (right side) (Fig. 15). A large elliptical sphenopalatine foramen,

which transmits the sphenopalatine artery and nerve into the naso-pharyngeal space above the

LARRY G. MARSHALL & CHRISTIAN de MU1ZON

sc PA

Source: MNHN. Paris

PUCADELPHYS AND IN US: THE SKI'LL

palate, has a posteriorly directed opening within the vertical surface of the palatine at the

posteromedial edge of the orbit just dorsomedial to the postpalatine foramen.

The palatine makes a small contribution to the medial edge of the orbital floor, extending

medially in sutural contact with the maxilla from the lateral side of the postpalatine foramen, then

almost directly anteriorly into the ventromedial corner of the infraorbital canal (Fig. 15). Its

contact with the lacrimal is a broad suture extending vcntrally from the anterior point of contact

with the frontal to the lower edge of the infraorbital canal, where it turns sharply anterodorsally

into that canal. Dorsally the palatine has a broad and nearly horizontal contact with the frontal; this

suture extends posteriorly to a point above the sphenopalatine foramen where it bends

ventroposteriorly to a point just posterodorsal to that foramen, then extends posteriorly along the

ventrolateral edge of the orbitosphenoid, and has a small contact posteriorly with the anterior edge

of the basisphenoid (as seen in YPFB Pal 6105, left side) (Fig. 15).

Lacrimal (Figs 12 and 15). — This bone forms the anterior rim of the orbit. In dorsolateral

view, the large facial wing of the lacrimal has adistinct quarter-moon shape. Viewed posterolaterally

(Fig. 15), the lacrimal has a rectangular shape (long axis is dorsoventral) and contacts the maxilla

anterodorsally, anteriorly and ventrolateral ly, the jugal anteroventrally, the frontal dorsoposteriorly,

and the palatine ventroposteriorly and ventromedially. The large elliptical shaped lacrimal

Fig. 12. — Pucadelphys andinus. Reconstruction ofthe skull (A, dorsal: B, right lateral; C. ventral views) based primarily on YPFB

Pal 6 1 05 (holotype). Abbreviations: AS, alisphenoid; BO, basioccipital; BS, basisphenoid; dal', dorsal atlantal facet; doc,

dorsal occipital condyle; ef. eniocarolid foramen (=anterior carotid foramen, carotid canal); EO, cxoccipital; eop.

cxoccipital process; elf. ethmoid foramen; flc. fossa for lower canine: fm, foramen magnum; fo. foramen ovale; FR,

frontal; fr, foramen rotundum: gf. glenoid fossa: if, incisive foramen; ifc. infraorbital canal; JU. jugal: js. jugular sulcus;

LA. lacrimal: Ic. lambdoidal crest (=nuchal crest): If. lacrimal foramen; lpps. lateral postpalatine spine: mp, mastoid

process; mpf, middle palatine foramen; nipps. medial postpalatine spine; MX. maxilla; NA. nasal; oof, optic-orbital

foramen; opf, supraorbital process of frontal; PA. parietal; pgf. postglenoid foramen: PL. palatine: PM(pm), pars

mastoidea of petromastoid (=mastoid s.s.); PM(pp), pars petrosa of petromastoid (=petrosal s.s.): PMX. premaxilla;

pogps, postglenoid process of squamosal: PP. postparictal; ppf. postpalatine foramen; ppt. postpalatine torus: prgpj

preglenoid process of jugal; prgps, preglenoid process of squamosal; ps. palatine sulcus; ptc. pterygoid canal; ptf,

posttemporal foramen: ptp. posttympanic process; pzf. postzygomalic foramen; sc, sagittal crest: SO. supraoccipital; sof.

supraorbital foramen; SQ, squamosal: ssf, subsquamosal foramen (=suprameatal. postsquamosal).

Fig. 12. — Pucadelphys andinus. Reconstitution du crcine (A, vue dorsale; B, vue lateral e droite; C, vue vent rale) principalement

d'apres YPFB Pat 6105 (holotype). Abreviations: AS, alisphenoide; BO. basioccipital; BS, basisphenoide; daffacette

atlantale dorsale; doc. condyle occipital dorsal; ef. foramen entocarotidien (=foramen carotidien anterienr, canal

carotidien): FO, cxoccipital; eop. processus cxoccipital; etf foramen ethmoidien: flc. fosse, pour la canine inferieure;

fin, foramen magnum; fo. foramen ovate; FR. frontal; fir. foramen rotundum; gf.fosse glenoide; if foramen incisif; ifc,

canal infraorbitaire; JU, jugal; js. sillon jugulaire; LA, lacrimal; Ic. Crete lambdoide (=crete nuchale); If. foramen

lacrimal; lpps, epine postpalatine laterale; mp. processus mastoide; mpf foramen palatin moyen; mpps, epine

postpalatine mediale; MX. maxillaire; NA. nasal; oof. foramen optico-orbitaire; opf, processus supra-orbitaire du

frontal; PA. parietal: pgf. foramen postglenoide ; PL. palatin: PM(pm). pars mastoidea du petreux (=mastoide s.s.);

PM(pp). pars petrosa du petromastotde (=petreux s.s.); PMX. premaxillaire; pogps. processus postglenoide du

squamosal; PP. postparictal; ppf. foramen postpalatin: ppt. torus postpalatin; prgpj processus preglenoide du jugal;

prgps, processus preglenoide du squamosal; ps, sillon palatin: ptc. canal pterygoide; ptf, foramen post temporal; ptp.

processus posttvmpanique; pzf. foramen postzygomalique; sc, crete sagittate; SO, supraoccipital: soj. foramen

supraorbitaire; SQ, squamosal; ssf foramen subsquamosal (=suprameatal. postsquamosal).

LARRY G. MARSHALL & CHRISTIAN de MUIZON

Fig. 13. — Pucadelphysandinus. A-C, skull with attached mandible ofYPFB Pal 6110 (A, dorsal; B, ventral; C. right lateral views).

Stereophotos X 2.5.

Fig. 13. — Pucadelphys andinus. A-C, crane el mandibule en connexion d’YPPFB Pal 6110 (A.vue dorsale; B, vue ventrale; C,

cue laterale droile). Photos stereoscopiques X 2.5.

Source: MNHN, Paris

PUCADELPHYS ANDINUS: THE SKULL

foramen, which transmits the naso-lacrimal duct, opens within the orbit and has two distinct canals

of subequal size situated one above the other along a dorsorncdial-ventrolateral axis (Fig. 15).

Jugal (Fig. 12). — Complete jugals are preserved in YPFB Pal 6105 (right side) and 6471 (left

side). This is a large bone which contributes to the formation of about 60% of the zygomatic arch.

Anteriorly it has a broad diagonal contact with the maxilla, a small contact with the lacrimal

anterodorsally-anteromedially, and it forms the ventral edge of the orbital rim. From its contact

with the maxilla it arcs in a poslerolaterodorsal direction, being dorsoventrally deep and

transversely narrow. At its dorsal and lateral greatest extent, which is approximately lateral to the

interorbital constriction, the jugal makes contact with the zygomatic process of the squamosal

dorsally and continues ventrally along it to the anterolateral edge of the glenoid fossa where it

forms a large preglenoid process. In lateral view, the ventroposterior spine of the jugal and dorsal

zygomatic process of the squamosal have an elongate diagonal contact (Fig. 12). A shallow sulcus

occurs along the ventrolateral edge of the jugal contribution to the zygomatic arch marking the

point of insertion of the masseter muscle (Fig. 12). The greater part of the zygomatic arch,

including the jugal and squamosal contributions, is subequal in breadth and depth.

Frontal (Figs 12 and 15). — This bone covers all of the medial 30% of the dorsal area and

most of the medial 30% of the lateral area of the cranium. Anteriorly it has a distinct W-shaped

contact with the nasals (Fig. 12), and lateral to this is a small distinct contact with the maxilla. The

lacrimal contact is broad and nearly horizontal dorsally, and broad but nearly vertical laterally.

There is a broad nearly horizontal contact with the dorsal edge of the orbital part of the palatine,

and a small wedge of the frontal extends ventrally and separates the palatine and orbitosphenoid

LARRY G. MARSHALL & CHRISTIAN de MUIZON

Source: MNHN. Paris

PUCADELPHYS ANDINUS: THE SKULL

Fig. 14 .—Pucadelphys andimis. Skull with attached mandible of YPFB Pal 6107 (A, dorsal: B. ventral: C. posterior: D. left lateral

views). Stereophotos X 2.5.

PtG. 14. Pucadelphys andinus. Crane et mandibttle en connexion d'YPFB Pal 6107 (A . vue dorsale; /?. vue ventrale; C. vue

paste tie tire; /). vue I ate rale gauche). Photos stereoscopiques X 2.5.

(Fig. 15). The frontal contacts the small orbital contribution of the orbitosphenoid along its

anterior and dorsal edges. More posteriorly, the frontal has a broad arcuate contact with the

alisphenoid. ventral to the interorbital constriction on the anterior surface of the cranium. The

contact with the parietal extends from the lateral surface of the anterior part of the cranium in an

almost straight line dorso-dorsoantcriorly, just posterior to the interorbital constriction. A weak

supraorbital process is sometimes present (as in YPFB Pal 6105 and 6108) on the anterolateral

surface of the frontal, and a small anteriorly opening supraorbital foramen consistently occurs just

ventral to this process within the dorsal edge of the orbit. A well developed oval-shaped ethmoid

foramen opens anteriorly at the frontal-alisphenoid-orbitosphenoid contact. The anteromedial

and dorsal wall of this foramen are formed by the frontal, the posterolateral wall by the

alisphenoid, the anteroventral wall by the orbitosphenoid. and the posteroventral wall by the

basisphenoid as seen in YPFB Pal 6105 (right side) and 6108 (lelt side).

Parietal (Fig. 12). — The large plate-like parietals cover the greater part of the dorsal and

dorsolateral surfaces of the cranium. They have a broad ventroposteriorly directed contact with

the frontal, a broad continuous contact with the dorsal parts of the alisphenoid and squamosal

ventrally, and a broad convex contact with the postparietal posteriorly. The outer walls of the

parietals are distinctly convex, reflecting the presence of well developed cerebral hemispheres,

and are ornamented dorsomedially by irregular scars lor attachment ot the temporalis muscle.

LARRY G. MARSHALL & CHRIS TIAN de MUIZON

These scars are seen to some degree in all specimens, but are most prominent in YPFB Pal 6105

in which a distinct scar-region on the left parietal extends from the frontal-parietal suture directed

toward, but not reaching, the postparietal-squamosal suture. On the right parietal of YPFB Pal

6105 and in all other specimens these scars are less distinct and are concentrated on the

dorsomedial area of this bone.

Postpariktal (Figs 12 and 17). — This bone is tightly fused with the parietal and the suture

is often difficult to identify. Nevertheless, careful study shows that the contact with the parietal

is broadly concave and extends ventrally where it meets the dorsoposterior end of the squamosal,

then turns sharply posteriorly having a nearly horizontal suture with the squamosal. A distinct

suture separates the postparietal from the underlying supraoccipital (Fig. 17). The postparietal

overhangs the entire dorsal edge of the supraoccipital so that the lambdoidal crest is formed

entirely by the posterior edge of the postparietal. Viewed posteriorly the postparietal also overlaps

the dorsolateral edge of the pars mastoidea of the petromastoids (Fig. 17).

Orbitosphenoid (Fig. 15). — Parts of this bone are preserved only in YPFB Pal 6105 and 6108.

Viewed laterally this bone makes a small contribution to the posteroventral base of the orbital

region. It contacts the frontal anteriorly and dorsally, the palatine anteroventrally and the

basisphenoid posteroventrally. In ventral view it forms the posteromedial roof of the choanal

orifice, the posterolateral parts of the roof are formed by small wings of the palatine. Within the

orbit (Fig. 15) a large ovoid, anteriorly opening optic-orbital foramen (which transmits cranial

nerves II, III, IV, V1, VI. the ophthalmic artery, and a vein which drains the eye to the cavernous

sinus) has the entire medial wall formed by the orbitosphenoid, the anteroventral edges by the

palatine, the anterior and dorsal edges by the frontal, and apparently the alisphenoid posterolaterally

although the alisphenoid is broken in this area in all specimens. The posterior side of the

orbitosphenoid has a distinct convex saddle-shaped surface marking the confluence of the left and

right optic-orbital foramina; the posterodorsal part has the form of laterally projecting horns which

separate the upper edge of the optic-orbital foramen from the lower edge of the more dorsally

situated ethmoid foramen.

Alisphenoid (Figs 12. 15. and 19). — This bone forms most of the anterior ventrolateral parts

of the ventral surface of the cranium. Anterodorsally it has a broad horizontal contact with the

frontal (Fig. 15); posterodorsally a broad horizontal contact with the parietal; posteriorly a broad

convex contact with the squamosal, which continues ventrally toward the base of the zygomatic

process, passes anteriorly around it, continues on the ventral surface of the skull and reaches the

posterolateral corner of the foramen ovale (which transmits V3, and possibly the major transverse

sinus of the external jugular vein; see below) (Figs 12 and 19). The alisphenoid forms the anterior

rim of the foramen ovale and, at its contact with the petromastoid, lateral to the position of the small

medial lacerate foramen (which is located at the junction of the basisphenoid-basioccipital-

petromastoid-alisphenoid; sensu MacIntyre, 1967) it extends anteriorly in tight sutural contact

with the basisphenoid. bisects the entocarotid foramen (alisphenoid forms lateral rim, basisphenoid

Source . MNHN, Paris

PUCADELPHYS ANDINUS : THE SKULL

medial rim), which transmits the internal carotid artery and a small vein from the inferior petrosal

sinus, and continues anteriorly to the anterior edge of the basisphenoid. The anteroventral extent

of the alisphenoid is not preserved in any specimen, but it probably formed the posterolateral rim

of the optic-orbital foramen as in Didelphis , and based on the size and breakage surface

anteroventrally there was clearly a wing-like structure which extended antcrolatcrally toward the

lateral edge of the choanal orifice (Fig. 15). On the anteroventral surface, immediately lateral and

parallel to the basisphenoid contact, is a sulcus marking the position of the pterygoid canal (for

the pterygoid or vidian nerve; sec Jollie, 1962: 44) or sutural contact with the pterygoid (seen in

YPFB Pal 6105). This bone is not well preserved in any specimen. The pterygoid canal can be

followed posteriorly to the anterolateral edge of the entocarotid foramen. Immediately lateral and

slightly dorsal to this canal is a large ovoid, anteriorly opening foramen rotundum (seen only in

YPFB Pal 6108) which transmitted the maxillary branch of the trigeminal nerve, V2 (Jollie, 1962:

44) (Figs 12 and 15).

Fig. 15. — Pucadelphys andinus. Detail of left orbital region based primarily on YPFB Pal 6105 (holotype). Abbreviations: AS.

alisphenoid; BS. basisphenoid; co. choanal orifice; etf. ethmoid foramen; fr. foramen rotundum; FR. frontal; ifc.

infraorbital canal; LA. lacrimal; If. lacrimal foramen; mpps. medial postpalatine spine; MX. maxilla; NA, nasal; oof,

optic-orbital foramen; opf, supraorbital process of frontal; OS, orbitosphenoid; PA. parietal; PL, palatine; ppf.

postpalatinc foramen; sof, supraorbital foramen; sphf, sphenopalatine loramen.

Fig. 15. — Pucadelphys andinus. Detail de la region orbitaire gauche principalement d'apr'es YPFB Pal 6105 (holotype).

Abreviations: AS, alisphenoide; BS, basisphenoide; co. orifice de la clioane; etf foramen ethmoide; fr. foramen

rotundum; FR, frontal; ifc . canal infraorbitaire; LA, lacrimal: If foramen lacrimal: mpps, epine postpalatine mediate;

MX, maxillaire; A I A, nasal; oof foramen optico-orbitaire; opf processus supraorbi tai re dufrontal; OS. orbitosphenoide;

PA, parietal; PL. palatin: ppf foramen postpalatin; sof foramen supraorbitaire; sphf foramen sphenopalatin.

LARRY G. MARSHALL & CHRISTIAN df. MUIZON

Basisphenoid (Figs 12 and 15). — As seen in YPFB Pa! 6105 and 6108, this is an anteroposteriorly

elongate bone which is broadest posteriorly along its contact with the basioccipital and narrows

anteriorly to its contact with the orbitosphenoid; sutures with the basioccipital and orbitosphenoid

are distinct. Laterally it has broad contacts with the alisphenoid: these bones arc tightly fused and

the suture is, in places, difficult to identify. The posterior surface of the basisphenoid is convex

ventral ly, while anterior to the entocarotid foramen it is flat. The anteriorend of this bone in YPFB

Pal 6105 is broken transversely, a feature which appears to represent natural breakage but which

may mark sutural contact between the basisphenoid (s.s.) and a smaller presphenoid. However,

this bone is quite similar to that in Didelphis and appears to be the basisphenoid which is simply

lacking its anterior end.

Squamosal (Figs 12,17 and 19). — This bone has a broad concave contact with the alisphenoid

anteriorly; a broad and nearly horizontal contact with the parietal dorsoanteriorly and dorsomcdial ly,

and with the postparietal dorsoposteriorly; an irregular vertical contact with the lateral surface of

the pars mastoidea of the petromastoid posteriorly; and ventrally its contact with the petromastoid

extends anteromedially from the lateral base of the mastoid process, passes along the lateral edge

of the fossa incudis and cpitympanic recess, and joins the short transversely oriented squamosal

suture at the posterolateral corner of the foramen ovale.

In ventral view the glenoid fossa is deep and transversely elongate; a large postglenoid

process of the squamosal occurs along its posterior edge and attains its greatest ventral extension

medially; a much lower, but still distinct, preglenoid process of the squamosal occurs along the

anteromedial edge of the fossa and a large preglenoid process of the posterior edge of the jugal

borders the fossa anterolaterally.

Three foramina perforate the squamosal externally. The first and smallest is a postzygomatic

foramen on the posterobasal surface of the zygomatic arch dorsal to the deepest point of the

postglenoid process (Fig. 17). This foramen apparently transmits a vein from the squamosal root

of the zygomatic arch to the sphenoparietal emissary vein (Archer. 1976a). It is usually single,

transversely elongate, and opens laterally; however, it is sometimes double (as in left side of YPFB

Pal 6105) with a second smaller opening set dorsomedially to the primary one.

The second foramen is the subsquamosal which is large, oval, and opens posteroventrally

at the base of the zygomatic arch just dorsal to the external acoustic meatus (Fig. 12). It is most

cleat ly seen in posterolateral view, while in dorsal view it is obscured by a posterior swelling of

the squamosal along its upper edge and a similar but smaller swelling occurs along its lower edge.

Thus, this foramen opens into a shallow postzygomatic depression in the squamosal. Externally

the foramen opens within the squamosal, while internally (as seen in right side of YPFB Pal 6110)

the medial side of the canal that it transmits is formed by the petromastoid and the lateral side by

the squamosal (Fig. 19). This foramen transmits an artery from the postglenoid foramen onto the

parietal area of the cranium which supplies the temporalis muscle, and a vein from the parietal area

to the sphenoparietal emissary vein, whichexits through the postglenoid foramen (Archer, 1976a;

WlBLE, 1990).

Source MNHN, Paris

PUCADELPHYS ANDINUS: THE SKULL

The third and largest is the postglenoid foramen which is ovoid in shape and opens

ventrolaterally along the posteromedial edge of the postglenoid process (Figs 12 and 19). The

course of the vessels this foramen transmits continues as a groove beyond the opening toward the

ventromedial edge of the postglenoid process. The external opening occurs solely within the

squamosal, while internally (as seen on the right side of YPFB Pal 6110; Fig. 19) at a level

corresponding to the dorsal rim of the external acoustic meatus, the medial side of this canal is

formed by the petromastoid and the lateral side by the squamosal. This foramen transmits the

Fig. 16. - Pucaddphys andinus. Skull of YPFB Pal 6108 (A. dorsal; B. ventral views). Stereophotos. X 2.5.

Fig. 16. — Pucadelphys andinus. Crane d'YPFB Pal 6108 {A. rue dorsale; II. rue rentrale). Photos slereoscopiques, X 2.5.

LARRY G. MARSHALL & CHRISTIAN de MUIZON

sphenoparietal emissary (which externally becomes the postglenoid) vein from the prootic sinus

(Wible, 1990), and the postglenoid artery (Archer, 1976a; Wible, 1990).

A fourth but smaller foramen, the posttemporal, opens between the squamosal and pars

mastoidea of the petromastoid directly posterior to the subsquamosal foramen and dorsal to the

mastoid process. This foramen transmits the arteria diploetica magna and vena diploetica magna

(Wible, 1990) which pass through a canal bordered by the petromastoid medially and squamosal

laterally (Fig. 21).

Basioccipital (Figs 12,17,19 and 21). — As seen in YPFB Pal 6105,6107 and 6470, this bone

has the shape of an arrow-head base and the sutures with adjacent elements are well defined. The

basioccipital has a broad transverse linear contact with the basisphenoid anteriorly between the

left and right medial lacerate foramina. From these foramina it extends posterolaterally in contact

with the medial surface of the pars petrosa of the petromastoid to the anteromedial corner of the

posterior lacerate foramen, where it bends sharply medially forming a small but distinct convex

rim over the inferior petrosal foramen (basioccipital forms ventral and medial rims, exoccipital

forms ventromedial rim. and pars petrosa of petromastoid forms lateral rim) which transmits the

inferior petrosal sinus. On the medial side of the foramen for the inferior petrosal sinus the

basioccipital contacts the exoccipital and the suture arcs posteromedially to a point on the lateral

side of the ventral occipital condyle. The posterior edge of the basioccipital forms the ventral rim

of the loramen magnum where a pair of well developed ventral occipital condyles are separated

by a distinct U-shaped intercondylar fossa. The foramen magnum transmits most of the sigmoid

sinus of the transverse venous system and the spinal root of nerve XI out of cranium (Dom el aL,

1970; Wible, 1990) and the vertebral arteries into cranium (Archer, 1976a).

The ventral surface of the basioccipital has a low medial keel extending anteriorly from the

middle of the intercondylar fossa to the basisphenoid suture; the lateral edges of the basioccipital

are elevated surfaces extending parallel to the petromastoid suture and they mark the floor of the

inferior petrosal sinus; between each of these lateral elevations and the medial keel is a broad fossa

for the rectus capitis muscle; the posterolateral edge of the basioccipital, between the elevated

ventral occipital condyles and the ventral rim of the inferior petrosal foramen, is distinctly convex

dorsally.

Exoccipital (Figs 12, 17, 19 and 21). — Viewed ventrally this bone contacts the basioccipital

medially, it loims the posterodorsal rim of the inferior petrosal foramen anteromedially; and the

medial and posterior rim of the posterior lacerate foramen (which transmits nerves IX, X, XI and

probably a small branch of the sigmoid sinus to the internal jugular vein) anteriorly; it also makes

a small contact with the pars petrosa of the petromastoid via a septum between the inferior petrosal

and posterior lacerate foramina; and laterally it contacts the pars mastoidea of the petromastoid.

Ventially, in the large fossa just anterior to the elevated prominence of the lower rim of the

dorsal occipital condyle, are typically three small subequal-sized condyloid foramina (Fig. 19)

which apparently transmit branches of cranial nerve XII as occurs in some specimens of Didelphis

Source MNHN, Paris

PUCADELPHYS ANDINUS: THE SKULL

( Jollie, 1 962: 55) but may also transmit branches of the sigmoid sinus to the internal jugular vein

(Archer, 1976a. fig. 2B). Of these three foramina the posterior-most passes under the medial part

of the dorsal occipital condyle; a second occurs anterolaterally between the first and posterior

lacerate foramen; and the third occurs anterior to the first and opens either totally within the

exoccipital just lateral to the basioccipital suture or within the suture itself.

Viewed posteriorly (Fig. 17) the exoccipital has a broad arcuate (convex) contact with the

pars mastoidea of the petromastoid which extends from the center of the exoccipital process

ventrally to the junction with the supraoccipital dorsally, where it bends sharply medially having

a slight arcuate (convex) contact with the supraoccipital to the dorsal edge of the dorsal atlantal

facet; medially, it forms the lateral rim of the posteriorly facing and transversely ovoid foramen

magnum. On its posterior dorsomedial surface, the exoccipital has a prominent dorsal atlantal

facet at the point of contact with the supraoccipital and a large ovoid dorsal occipital condyle

occurs ventromedially. The exoccipital contributes to the formation of the medial half of the very

small exoccipital process ventrolatcrally. The exoccipital process is delimited medially by a

trough in the exoccipital separating it from the dorsal occipital condyle, and another trough

laterally in the petromastoid separating it from the larger mastoid process.

Pi Gi 17 . Pucadelphvs andinus. Reconstruction of the posterior view of the skull and Zygomatic arches, based primarily on YPFB

Pal 6105 (holotype). Abbreviations: BO. basioccipital; daf. dorsal atlantal facet; doc. dorsal occipital condyle; EO.

exoccipital; cop! exoccipital process; f. foramen; fm. foramen magnum; icf. intercondylar fossa; JU. jugal; lc. lambdoidal

crest (=nuchal crest); mfo. mastoid foramen; nip. mastoid process; pgf. postglcnoid foramen; PM(pm). pars mastoidea

of petromastoid (=mastoid s.s.); pogps. postglenoid process of squamosal; PP. postparietal: prgpj prcglenoid process ol

jugal; ptf, posttemporal foramen; p/.f. postzygomatic foramen: SO. supraoccipital: SQ. squamosal; voc. ventral occipital

condyle.

f, G 17 ' _ Pucadelphys andinus. Reconstitution de la vueposterieure du crane et desarcades zygomatiquesprincipalement d apres

YPFB Pal 6105 (holotype). Abreviations: BO. basioccipital; daffacette atlantale dorsale; doc. condyle occipital dorsal;

EO. exoccipital: eop. processus exoccipital:f foramen: fin. foramen magnum: icf. fosse intercondxlienne; JU. jugal: lc.

crete lambdoide (=crete nuchale): mfo. foramen mastoide: mp. processus mastoide:pgf. processuspostglenoide ; PM(pm).

pars mastoidea du petromastoide I =mastoide s.s.): pogps. processus postglenoide du squamosal; PP. postparietal; prgpj

processus preglenoide du jugal; ptf. foramen posttemporal; pzf. foramen poslzygomatique; SO. supraoccipital; SQ.

squamosal; voc. condyle occipital ventral.

LARRY G. MARSHALL & CHRISTIAN de MUIZON

In YPFB Pal 6105 the distance between the outer edges of the dorsal occipital condyles is

8.0 mm and 4.5 mm between the inner edges: the maximum breadth of the foramen magnum,

dorsal to the dorsal occipital condyles, is 5.0 mm and the maximum depth is 3.0 mm.

Supraoccii’ital (Figs 12 and 17). — This bone occupies most of the dorsal surface of the

occipital plate and forms the dorsomedial rim of the foramen magnum. It has a broad contact with

the exoccipital ventrolaterally, a broad contact with the pars mastoidea of the petromastoid

laterally, and a broad arcuate contact with the postparietal dorsally. As noted above, the contact

with the latter occurs ventral to the lambdoidal crest which is formed entirely by the postparietal.

The occipital area is hidden in dorsal view by the postparietal extension of the lambdoidal crest.

There is no distinct vertical medial crest, but in YPFB Pal 6105 the posterior surface is ornamented

with small ridges, scars and shallow pits marking areas of insertion of the nuchal muscle (these

features are not seen in other specimens). There are only three small foramina in the supraoccipital:

one (left and right) dorsomedial to the point of contact of the petromastoid and exoccipital; and

a single foramen dorsomedially. between the supraoccipital and postparietal below the lambdoidal

crest at its point of contact with the sagittal crest.

Petromastoid (Figs 12, 17, 19 and 21). — This is the most complex bone of the skull. For

purposes of description it can be divided into two principal components: the pars mastoidea which

contributes to the formation of the lateral part of the occipital region and houses the subarcuate

lossa on the cerebellar side: and the pars petrosa which houses the inner ear in a teardrop shaped

promontorium on the ventral side between the basioccipital and squamosal, and the internal

acoustic meatus on the cerebellar side (MacIntyre, 1972).

Fig. 18. — Pucadelpliys cindinus. Ventral view of basicranium of YPFB 6105 (holotypc). Stereophotos, X 5.

FlG - 18 ~ p "cadelphys andinus. Vue venlrale de la base du crane d’YPFB 6105 (holotype). Photos Slereoscopiques, X 5

Source: MNHN, Paris

PUCADELPHYS ANDINUS: THE SKULL

The pars mastoidea is best seen in posterior view (Fig. 19) where it occupies the lateral area

of the occipital region and contacts theexoccipital ventromedially, the supraoccipital dorsomedially,

the ventrolateral edge of the postparietal dorsally, and the squamosal along most of its lateral

surface. A liny mastoid foramen typically occurs on the posterior surface of the pars mastoidea

just lateral to the supraoccipital-cxoccipital-petromastoid juncture (Fig. 17), and internally opens

along the posterior side of the sigmoid sinus (Fig. 21). A small posttemporal foramen (sensu

Wible, 1990) occurs on the lateral surface of the occiput between the squamosal and the pars

mastoidea (Figs 12 and 21; see below). A large ovoid surface facing ventrolaterally is ornamented

with muscle scars, and was probably the major area for insertion of muscles which move the head

and neck (Figs 12 and 17).

Viewed ventrally (Fig. 19) the pars mastoidea contributes posteromedially to the formation

of the lateral half of the very small exoccipital process and it forms all of the larger mastoid process

laterally. The anteroventral rim at its medial-most edge is united with the pars petrosa by a narrow

bridge of bone on the lateral side of the posterior lacerate foramen. Lateral to this bridge, the

anteroventral rim of the pars mastoidea forms an anteriorly directed lip which represents the

caudal tympanic process of the petromastoid (sensu Wible, 1990). It underhangs a fossa located

between it and the posterior end of the promontorium of the pars petrosa which is called the

mastoid epitympanic sinus (sensu Archer, 1976a: 314). Lateral to the caudal tympanic process

of the petromastoid, is the well developed mastoid process, and along its lateral and medial

surfaces anteriorly is the stylomastoid notch, through which the facial nerve and lateral head vein

leave the middle ear (Wible, 1990). The lateral extremity of the anteroventral rim of the pars

mastoidea extends anterodorsally, up to the medial border of the fossa incudis at the posteromedial

extremity of the epitympanic recess, lateral to the fenestra vestibuli.

The pars petrosa, viewed ventrally, is dominated by a large teardrop-shaped promontorium

(=pars cochlearis) which is broadest posteriorly and tapers anteromedially to a point at the junction

ofthealisphenoid-basisphenoid-basioccipital. In lateral view the ventral border of the promontorium

is at a level slightly dorsal to the ventral edge of the dorsal occipital condyles. The ventral surface

of the promontorium is nearly smooth and the minor topographic variations largely reflect turns

of the cochlear duct.

Anteriorly, a broad shallow depression is located between the promontorium and the large

foramen ovale; it begins at the anteromedial edge of the large ovoid bulge of the promontorium,

extends anterior and parallel to the basioccipital suture, crosses onto the alisphenoid just lateral

to the medial lacerate foramen, and passes on the alisphenoid, to the posterior edge ot the

entocarotid foramen. This depression marks the route ot passage ot the internal carotid aitery (i.e.

sulcus for internal carotid artery) on its way to the entocarotid foramen; the passage of the internal

carotid artery is thus situated medially (sensu Presley, 1979). The large tossa lateral to this

depression on the anterolateral surface of the promontorium is for insertion of the tensor tympani

muscle. The medial edge of the promontorium parallel and adjacent to the basioccipital suture is

elevated and rounded, a feature produced by the internal sulcus for the interior petiosal sinus

which exits the basicranium through the inferior petrosal foramen at the posteromedial edge ot the

promontorium.

LARRY G. MARSHALL & CHRISTIAN de MUIZON

O__2

Source: MNHN. Paris

PUCADELPHYS ANDINUS: THE SKULL

Two openings occur on the posterior surface of the promontorium. The first and smallest

is the fenestra cochleae which is situated on the posterolateral surface of the promontory

anterolateral to the lateral edge of the posterior lacerate foramen and opens laterally; it occurs at

the ventral edge of a shallow depression, the fossula fenestra cochleae. Viewed posteriorly, this

fenestra occurs immediately anterior to the small exoccipital process at the exoccipital-pars

inastoidea suture, and its ventral opening is at about the same level as the medial part of the ventral

rim of the pars inastoidea (i.e. the lowest point of the trough separating the exoccipital and mastoid

processes). The second and larger opening is the fenestra vestibuli which accommodates the

footplate of the stapes and annular ligament; it has an ovoid (anteroposteriorly elongate) shape,

a length/width ratio of 1.4 based on two specimens, and occurs dorsally on the lateral surface of

the promontorium within a shallow depression, the vestibular fossula. In lateral view, this fenestra

is seen at the posterodorsal edge of the external acoustic meatus. It thus lies anterolateral and a little

dorsal to the fenestra cochleae, and the two fenestra are separated by a broadly rounded and

vertically oriented swelling of the posterolateral corner of the promontorium.

Fig. 19. — Piiccidelphys andinus. A. reconstruction of the right side of the basicranium in ventral view, based on YPFB Pal 6105

(holotype) and 6110; B, right ear region in ventral view (YPFB Pal 6110). The foramen ovale is restored. Abbreviations:

ac, aquaeductus cochleae: al. anterior lamina; AS. alisphenoid; BO. basioccipital; BS, basisphenoid; cf, condyloid

foramen; cpmv, crista promontorii medioventralis; ctpp. caudal tympanic process of pars mastoidcaof petromastoid; doc.

dorsal occipital condyle: cam. external acoustic meatus; ef, cntocarolid foramen (=antcrior carotid foramen, carotid

canal); EO, exoccipital: cop. exoccipital process; er. epitympanic recess; f. foramen; fc. fenestra cochleae (=rotunda); fcs.

floor of caVum supracochleare; ffc. fossula fenestra cochleae; fi, fossa incudis; fo, foramen ovale; frcin. fossa for rectus

capitis muscle; fs, facial sulcus; fstni. fossa for stapedial muscle; fv. fenestra vestibuli (=ovalis); gf, glenoid fossa; icf,

intercondylar fossa; ipf. inferior petrosal foramen (internal jugular foramen); lape. lateral aperture of prootic canal; It.

lateral trough; mes, mastoid epitympanic sinus of epitympanic recess; mil', median lacerate foramen; mp, mastoid process;

pc. prootic canal; pgf, postglenoid foramen; plf. posterior lacerate foramen (=jugular foramen); PM(pm), pars mastoidea

of petromastoid (=mastoid s.s.); PM(pp). pars petrosa of petromastoid (=petrosal s.s.); pogps. postglenoid process of

squamosal; pr. promontorium of pars petrosa of petromastoid; ptp. posttympanic process; pzf. postzygomalic foramen;

sff. secondary facial foramen; sica, sulcus for internal carotid artery; sips, sulcus for inferior petrosal sinus; smn,

stylomastoid notch: SQ. squamosal; ssf. subsquamosal foramen (=supramcatal. postsquamosal); tape, tympanic aperture

of prootic canal; ttf. tensor tympani fossa; vf. vestibular fossula; voc, ventral occipital condyle.

Fig. 19. — Pucadelphys andinus. A, Reconstitution du cote droit de la base du crane en vue ventrale d'apres YPFB Pal 6105

(holotype) et 6110; B, vue ventrale de la region auditive droite (YPFB Pal 6110). Le foramen ovale est reconstitue.

Abreviations: ac, aqueduc cochlea ire; al, lame anterieure; AS, alisphenoide; BO, basioccipital; BS, basisphenoide; cf.

foramen condylien; cpmv. crista promontorii medioventralis; ctpp, processus tympanique caudal de la pars mastoidea

du petromastoide; doc, condyle occipital dorsal; earn, meat acoustique externe; ef, foramen entocarotidien (=foramen

carotidien anterieur, canal carotidien); FO. exoccipital; cop, processus exoccipital; er. recessus epitympanique; f

foramen; fc. fenestra cochleae (=rotunda); fcs, plancher du cavum supracochleare; ffc. fossula fenestra cochleae; Ji,

fossa incudis; fo, foramen ovale; from. fosse pour le muscle rectus capitis; fs. sillon facial; fstm, fosse pour le muscle

stapedial; fv, fenestra vestibuli (=ovalis); gf, fosse glenoide; icf, fosse intercondylienne; ipf, foramen petreux inferieur

(=foramenjugulaire interne); tape, ouverture late rale ducanal prootique; It. depression laterale; mes. sinus epitympanique

mastoide du recessus epitympanique; mlf trou dechire moyen; mp, processus mastoide; pc, canal prootique; pgf, foramen

postglenoide; plf trou dechire poslerieur(=foramen jugulaire); FM(pm), pars mastoidea du petromastoide (=mastoide

s.s.); PM(pp), pars petrosa du petromastoide (=petreux s.s.); pogps, processus postglenoide du squamosal; pr.

promontoire de la pars petrosa du petromastoide; ptp, processus posttympanique; pzf. foramen postzygomatique; sff.

foramen facial secondaire; sica. sillon pour la carotide interne; sips, sillon pour le sinus pretreux inferieur; smn.

echancrure stylomastoidienne; SQ. squamosal; ssf foramen subsquamosal (=suprameatal. postsquamosal); tape,

ouverture tympanique du canal prootique; ttf, fosse pourle muscle tensor tympani; vf. fossette vestibulaire; voc, condyle

occipital ventral.

LARRY G. MARSHALL & CHRISTIAN dc MUIZON

A large ovoid secondary facial foramen, which is the opening of the facial nerve (VII) canal

into the middle ear, occurs just anterior to the dorsal edge of the fenestra vestibuli. The secondary

facial foramen opens posteriorly and a sulcus for the facial nerve (sulcus facialis) occurs

immediately posterior to this foramen and separates the fenestra vestibuli from the epitympanic

recess and the lateral trough (see below); these features are clearly seen in YPFB Pal 6110

(Fig. 19).

The epitypanic recess is “the extension of the middle ear cavity that lies dorsal to the

tympanic membrane and contains the mallear-incudal articulation” (Wible, 1990: 188), (see also

Van der Klaauw, 1931: 73: and Archer. 1976a: 226). It is also called the fossa capitis mallei. The

posterior extremity of the epitympanic recess is a deep and narrow pit where the ligament of the

crus breve of the incus attaches: it is the fossa incudis or fossa crus breve incudis. In Pucadelphys,

the epitympanic recess is an oblique, slightly concave depression in the roof of the tympanic

cavity, approximately three times longer than wide, and bordered laterally and medially by thin

bony rims. It is limited anteriorly by the squamosal at the anterior edge of the postglenoid foramen.

The lateral rim of the epitympanic recess forms the medial wall of the postglenoid foramen (this

is clearly seen on the right side of YPFB Pal 6110).The fossa incudis is a small pit located at the

posteromedial extremity of the epitypanic recess.

The lateral trough is a well developed anteroposteriorly elongate slit that opens medially just

lateral to the secondary facial foramen. It is formed by a depression in the petromastoid

overhanging the medial rim of the epitympanic recess as clearly seen on the right side of YPFB

Pal 6110 (Fig. 19). In the posterior end of the lateral trough is a small foramen which represents

the tympanic aperture of the prootic canal which connects the lateral head vein (vena capitis

lateralis; Archer, 1976a: 302; Wible, 1990) to the prootic sinus (Fig. 19). The prootic canal has

a lateral opening within the petromastoid (lateral aperture of prootic canal) into the postglcnoid

canal (as seen on left side of YPFB Pal 6110) where the lateral head vein joins with the prootic

sinus at its juncture with the sphenoparietal emissary vein (Fig. 21).

A small fossa for the origin of the stapedius muscle (fossa stapedius or fossa muscularis

minor) occurs on the medial side of the lateral part of the stylomastoid notch and dorsal to it, just

posterolateral to the fenestra vestibuli, posterior to the sulcus for the facial nerve. The fossa is

shallow, dorsoventrally elongated, and faces medially. It is backed on to the fossa incudis

(Fig. 19).

A tiny anteriorly directed opening foramen occurs on the lateral side of the fossa for the

tensor tympani muscle, adjacent to the promontorium between the secondary facial foramen and

the foramen ovale (Fig. 19). This foramen leads into a passage which joins with the posterior end

of the facial canal within the petromastoid.

Two tiny openings occur in the posteromedial wall of the petromastoid, within the posterior

lacerate foramen, which connect with the vestibular cavity (Figs 19, 21). The first is the

aquaeductus cochleae which transmits the perilymphatic duct and a vein; it is situated in a vertical

sulcus just dorsal to the outer opening of the posterior lacerate foramen within the jugular sulcus,

on the medial side of the bridge joining the pars mastoidea and pars petrosa. Viewed posteriorly

this opening is at about the same level as the dorsal rim of the fenestra cochleae. The second

Source:

PUCADELPHYS ANDINUS: THE SKULL

opening is the aquaeductus vestibuli which transmits a vein and the endolymphatic duct. It is a

small vertical slit deep within the posterior lacerate foramen, dorsal to the aquaeductus cochleae,

on the anterior side of the jugular sulcus at a level corresponding to the loweredge of the subarcuate

fossa posterior to the crus commune (Fig. 21); it opens posteriorly on the back surface of the broad

crista separating the subarcuate fossa and jugular sulcus.

The cerebellar surface of the petromastoid consists principally of a large broadly ovoid

central body oriented dorsolaterally-ventromedially, with two large openings (Fig. 21). The first

opening, the subarcuate fossa, is the more dorsoposterior of the two and is a deep spheroidal pocket

which opens medially; it houses the paraflocculus of the cerebellum. The second opening, the

internal acoustic meatus, is more ventromedially located and is separated from the subarcuate

fossa by a broad horizontally oriented septum of which the posteromedial part is the crus

commune. The internal acoustic meatus is a broad shallow opening containing two deeper pits

separated by a low transverse septum. The larger of these pits is the foramen acusticum inferius

which is set posteromedial to the other and serves for transmitting two branches of the

vestibulocochlear nerve (VIII); it has a fusiform shape and opens dorsomedially. The smaller of

the pits is the foramen acusticum supcrius which is set anterolateral to the first and is a vertical

slit which opens posteromedially and a little ventrally; it transmits the facial (VII) and branches

of the vestibulocochlear (VIII) nerves. The ventral rim of the internal acoustic meatus opens onto

a large anteromedially projecting platform, which is rounded and thickened anteriorly to form a

divide which separates, on the one hand a broad medially directed sulcus leading from the foramen

acusticum inferius on its posterior side, and on the other hand a smaller more anteriorly directed

and dorsally situated sulcus running from the foramen acusticum superius on its anterior side.

Fig. 20. — Pucadelphys andinus. View of the cerebellar (dorsomedial) surface of petromastoid of YPFB Pal 6470. Stercophotos. X 5.

Fig. 20. — Pucadelphys andinus. Vue de la surface cerebellaire (dorso-mediale) du peiromastoide de YPFB Pal 6470. Photos

stereoscopiques, X 5.

LARRY G. MARSHALL & CHRISTIAN de MUIZON

Four principal sulci surround the central body of the petromastoid on its cerebellar surface

(Fig. 21): 1) Along the ventromedial edge is an elongate sulcus for the inferior petrosal sinus which

is bordered dorsally by the crista promontorii medioventralis. This sinus connects “the venous

sinus (sinus cavernosus) encasing the pituitary gland and optic chiasma with the internal jugular

vein just before it emerges from the inferior petrosal foramen” (MacIntyre, 1972: 291). 2)

Posteroventrally, is a shallow and nearly vertical sulcus for the jugular sinus. 3) Posterodorsally

is a small pocket representing the sulcus for the sigmoid sinus. A tiny medially opening foramen

in the deepest part of this sinus extends laterally and apparently connects with the canal from the

small foramen in the dorsoposterior end of the sulcus for the prootic sinus (sec below). There is

also the cerebellar opening of the mastoid foramen for the occipital emissary vein. 4) Along the

dorsolateral edge is an elongate sulcus for the prootic sinus, a primary tributary of the lateral head

vein, which is walled by the squamosal laterally and by a crest of the petromastoid medially.

Anterolaterally to the crista petrosa is a large depression in an antcrdorsally projecting wing

of the petromastoid for part of the temporal lobe of the cerebrum and the trigeminal (=gasserian

or semilunar) ganglion of the trigeminal nerve (V). In other didelphids the trigeminal ganglion lays

on the alisphenoid and no anterodorsally projecting wing is observed. At the anteromedial edge

of this depression along the lateral surface of the crista petrosa is a small foramen which opens

anteromedially into a groove (seen clearly in YPFB Pal 6107 and 6470); this is the hiatus Fallopii

which transmits the greater petrosal nerve, a branch of the facial nerve (VII). The anterolateral

wing makes an extensive contribution to the floor of the middle cranial fossa (Fig. 21) and the

internal side of the lateral wall of the braincase (the lateral side is formed by the squamosal and

the alisphenoid; it apparently represents the relic of a reduced anterior lamina of the petrosal (sensu

Crompton & Jenkins, 1979: 71. figs 3-5B, C) (see below for justification and discussion). In

ventral view, an irregular contact is clearly seen between the anterolateral border of the

promontorium and the flat shelf which forms the posterior edge of the foramen ovale in YPFB Pal

6105 (Fig. 19). This portion probably represents part of the anterior lamina. The anterior lamina

is covered laterally by the squamosal and the alisphenoid. Its uncovered portion, visible in ventral

view of the skull, lies medial to the squamosal. It forms the posterior rim of the foramen ovale,

extending from the squamosal suture laterally across the anterior surface of the promontorium as

a broad uneven ledge to the alisphenoid contact lateral to the medial lacerate foramen.

On the lateral (squamosal) side of the petromastoid (Fig. 21) is a broad shallow anteroventrally

directed sulcus which transmitted the prootic sinus along its greater length and the sphenoparietal

emissary vein ventral ly. The lateral aperture of the prootic canal occurs at the anteroventral end

of this sulcus and this aperture approximates the boundary between the prootic sinus (dorsally)

and sphenoparietal emissary vein (ventrally). A small anteriorly directed opening foramen occurs

dorsally on the petromastoid, just posterior to the sulcus for the prootic sinus. This foramen

transmits a branch of the prootic sinus posteriorly within the pars mastoidea, and the canal

apparently unites medially with the small foramen in the dorsal end of the sigmoid sinus (Fig. 21).

A small shallow sulcus unites this foramen with the dorsoposterior edge of the sulcus for the

prootic sinus (Fig. 21).

Source: MNHN, Paris

PUCADELPHYS ANDINUS: THE SKULL

0_ 4 _2

m m

Fig. 21. — Pucadelphys andinus. Details of petromastoid (A, cerebellar or dorsomedial surface; B, squamosal or lateral surface),

based on YPFB Pal 6470. Abbreviations: ac, aquacductus cochleae; al. anterior lamina; av, aquaeductus veslibuli

(=endolymphatic foramen); BO, basioccipital; cc, crus commune; cf, condyloid foramen; cp, crista petrosa; cpmv. crista

promontorii medioventralis; EO. exoccipital; f. foramen; fai. foramen acusticum inferius (=area cochleae); fas. foramen

acusticum superius (=area facialis); fc, fenestra cochleae (=rotunda); ftg, fossa for trigeminal ganglion; fv, fenestra

vestibuli (=ovalis); hF. hiatus Eallopii; iam, internal acoustic meatus; ipf. inferior petrosal foramen (internal jugular

foramen); js. jugular sulcus; lape. lateral aperture of prootic canal; mfo, mastoid foramen; mlf. median lacerate foramen;

mp, mastoid process; plf, posterior lacerate foramen (=jugular foramen); PM, petromastoid; pr. promontorium of pars

petrosa of petromastoid; ptn. posttcmporal notch; saf. subarcuate fossa (=floccular, parafloccular fossa); sdv. sulcus for

diploetic vessels; sips, sulcus for inferior petrosal sinus; smn, stylomastoid notch; SO, supraoccipital; sps, sulcus for

prootic sinus; SQ, squamosal; ssev. sulcus for sphenoparietal emissary vein; ssn. subsquamosal notch; sss, sulcus for

sigmoid sinus; tape, tympanic aperture of prootic canal; ttf. tensor tympani fossa.

Fig. 21. — Pucadelphys andinus. Details da petromastoide (A, face cerebellaire on dorso-mediale; B, face squamosale oa

late rale), d 'apres YPFB Pal 6470. Abrevialions: ac. aquaeductus cochleae; al. lame anlerieure; av. aquaeductus vestibuli

(=foramen endolymphatique); BO. basioccipital; cc, crus commune; cf. foramen condylien; cp. crista petrosa; cpmv.

crista promontorii medioventralis; EO. exoccipital; f foramen; fai. foramen acusticum inferius (=area cochleae); fas.

foramen acusticum superius (=area facialis); fc. fenestra cochleae (=rotunda); ftg, fosse pour le ganglion trigeminal;

fv. fenestra vestibuli (=ovalis); hF. hiatus Fallopii; iam, meat auditif interne; ipf. foramen petreux inferieur (=foramen

jugulaire interne); js. sillon jugulaire; lape, ouverture laterale du canal prootique; mfo. foramen mastoide; mlf iron

declare moyen; mp, processus mastoide; plf. iron declare posterieur (=foramen jugulaire); PM, petromastoide; pr,

promontoire de la pars petrosa du petromastoide; ptn, echancrure posttemporale; saf fossa subarcuata (= fosse

flocculaire on parafloculaire); sdv. sillon pour les vaisseaux diploetiques; sips, sillon pour le sinus petreux inferieur; smn,

echancrure stylomastoidienne; SO. supraoccipital; sps. sillon pour le sinus prootique; SQ. squamosal; ssev. sillon pour

la veine emissaire sphenoparietale; ssn. echancrure subsquamosale; sss. sillon pour le sinus sigmoide; tape, ouverture

tympanique du canal prootique; ttf, fosse pour le muscle tensor tympani.

A less well defined sulcus extends posteriorly from the sulcus for the sphenoparietal

emissary vein along the ventral edge of the pars mastoidea, then bends dorsoposteriorly. This

sulcus apparently transmitted the arteria diploetica magna and vena diploetica magna (sensu

LARRY G. MARSHALL & CHRISTIAN de MUIZON

Wible, 1990) which passed through the posttemporal foramen at the posttemporal notch. The vena

diploetica magna united with the sphenoparietal emissary vein at a point level with the lateral

aperture of the prootic canal. The sphenoparietal emissary vein and arteria diploetica magna

passed ventrally and exited the skull through the postglenoid foramen (Wible, 1990).

Ectotympanic. — There is noevidence of an ossified ectotympanic (the bone which supports

the tympanic membrane) nor are there facets in the ear region which marked the site of attachment

of this bone.

DISCUSSION

The comparison of the skulls of Pucadelphys andinus is difficult in spite of the good

preservation of the remains. As a matter of fact, there is no described skull of fossil marsupials

from the Cretaceous, and the skull of Ma yulestesferox from the early Paleocene of Tiupampa, with

which P. andinus could reasonably be compared, is only known by a preliminary description

(Muizon, 1994) and its thorough study is under progress by one of us (C. M.). However, two

undescribed probable metatherian skulls from the late Cretaceous of Mongolia have been

mentioned in the litterature. Kielan-Jaworowska & Nessov (1990) referred to the Melatheria a

complete dcltatheroidan skull wcl 1 known as the Gurl in Tsav Skull and Trofimov & Szalay (1993)

mentioned the skeleton of a Monodelphis scalops- sized Asiadelphia from the Barun Goyot

Formation of Mongolia. Another specimen, known from the middle Paleocene of Sao Jose de

Itaboraf (Brazil), belongs to a polydolopid and consequently cannot be used for a comparison with

a didelphid; furthermore, its very poor state of preservation does not permit reasonable compari¬

sons. However, several well preserved skulls (or partial skulls) of primitive mammals are known

from the Jurassic and the Cretaceous. The most important are: morganucodontids [ Morganucodon ,

( Kermack. 1963; and Kermack etui, 1981)]. Gobiconodon (Jenkins & Schaff, 1988), Sinoconodon

(Crompton & Luo, 1993), docodonts [Haldanodon (Lillegraven & Krusat, 1991 )J, several

multituberculates, [among others: Kamptobaatar and Sloanbaatar (Kielan-Jaworowska, 1971),

Catopsbaatar and Chulsanbaatar (KiELAN-i aworowska, 1974), Catopsalis (Kielan-Jaworowska

& Sloan, 1974), Kryptobaatar and Tugrigbaatar (Kielan-Jaworowska & Dashzeveg, 1978),

Nemegtbaatar, (Kielan-Jaworowska et al. , 1986)]. eupantotheres [Vincelestes (Bonaparte &

Rougier. 1987; Rougier & Bonaparte, 1988; Rougier et al., 1992; and Hopson & Rougier,

1993)], deltatheroidans [Deltatheridium (Kielan-Jaworowska, 1975: a partial skull only) and the

Gurlin Tsav skull, (Kielan-Jaworowska & Nessov, 1990)], and eutherians /Aviorycto, Barunlestes,

Kennalestes and Zalambdalestes (Kielan-Jaworowska, 1981; 1984; Kielan-Jaworowska &

Trofimov. 1980)]. Although not directly related to Pucadelphys they are extremely useful for

determining the primitive features of the genus. The skulls of Tiupampa ( Pucadelphys and

Mayulestes) are the oldest known undoubted marsupial skulls and, in the future, they will represent

the best sample for the study of primitive marsupials. Pucadelphys is also one of the oldest known

didelphids and its teeth can be compared with the abundant dental remains from the “middle”

PUCADELPHYS ANDINUS: THE SKULL

Paleocene of Itaboraf which were revised by Marshall (1987). Consequently, the discussion will

be divided into two sections. The first section will compare the teeth of Pucadelphys to those of

the Paleocene didelphids of Itaboraf and of the late Cretaceous marsupial dental remains of North

America related to the Alphaclon group. The second section will consider the major cranial

characters of Pucadelphys and, by comparing them to the available skulls of Mesozoic mammals,

we will try to determine their phylogenetic state and polarity.

Dental Comparison. — Among all the fossil didelphids known in the lower Tertiary of South

America, Stembergia itaboraiensis from the middle Paleocene (Itaboraian) of Sao Jose de Itaboraf

(Brazil) is the closest to P. andiniis with respect to size as well as to tooth structure. Both species possess,

among other features, the same didelphid characters, which are the paracone smaller than the metacone

and the V-shaped centrocrista. The affinities of the two forms were noted since the first discoveries at

Tiupampa by Muizon et al. (1984) who described a tooth that they related to a "Stembergia- like”

marsupial and that must now be referred to P. andinus. Stembergia itaboraiensis was described by

Paula Couto (1970) on the basis of a mandible fragment with m4 and the talonid of m3. A revision

of S. itaboraiensis was provided by Marshall (1987) who included in the hypodigm of the species

several other specimens of the lower and upper dentition. With the exception of their dental structure,

which is almost identical, P. andinus and S. itaboraiensis share the same morphology of the protocone,

which is long anteroposteriorly, triangular-shaped, and inflated posteriorly, and the important

development of the styles B and C. However, the two species clearly differ in the different development

of several of their elements. The upper molars of Stembergia differ from those of Pucadelphys in their

shallower trigon basin, in their more reduced paracone, in their more deeply V-shaped centrocrista, in

their smaller conules, in their smaller stylar shelf with a shallower basin, in the reduction of the style

A which does not form a parastyle as large as in Pucadelphys , in their very reduced style D (while it

is often the same size as style C or larger in Pucadelphys). in the reduction of the metacone of their M4

which is fused to the posterolabial angle of the tooth, in the loss of the posterior part of the stylar shelf

basin of M4 and in the great anterolabial projection of the parastylar region of the stylar shelf of M4

with a larger paracrista. The lower molars of Stembergia differ from those of Pucadelphys in that the

talonids of m3 and m4 are always wider than the trigonids (in Pucadelphys the talonid of these teeth

is sometimes as wide as the trigonid but is often narrower), in the anteroposteriorly longer trigonid, in

the lower and stouter trigonid cusps, in that the metaconid and paraconid arc more separated one from

the other, a condition which more widely opens the trigonid basin lingually, in the hypoconid which

has a better marked V-shaped morphology, in the larger hypoconulid and entoconid and in the presence

of well developed anterior and labial cingula (they are absent in Pucadelphys). Most of the dental

characters of Pucadelphys are more primitive than those of Stembergia and, as already expressed by

Marshall & Muizon (1988) and Muizon (1992), the former represents a good morphological ancestor

for the latter, as far as the teeth are concerned.

Two other marsupials from Itaboraf share structural similarities with Pucadelphys andinus ; they

are Marmosopsis juradoi Paula Couto (1962) and Itaboraidelphys camposi Marshall &

Muizon (1984).

LARRY G. MARSHALL & CHRISTIAN de MUIZON

M.juradoi is known by several upper and lower jaws and its hypodigm has been revised by

Marshall (1987). It differs from P. andinus in its smaller size, its upper molars proportionally

slightly narrower and longer and its molar cusps sharper and less bulky. In comparison to those

of Pucadelphys, the upper molars of Marmosopsis differ in having a higher metacone relative to

the paracone, a smaller protocone, a shorter and straight preprotocrista, smaller conules, a

shallower stylar shelf basin, a much higher metacrista which is straight in occlusal and anterior

views (it is concave in Pucadelphys), a slightly shorter paracrista, reduced styles C. D, and E. The

lower molars of Marmosopsis have a higher trigonid (relatively to the talonid) with a smaller and

lower paraconid, a shorter talonid with a smaller and shallower basin, a crest-1 ike entoconid which

links the hypoconulid to the posterior edge of the metaconid (in Pucadephys and Stembergia a

distinct entoconid is present) and the loss of a distinct hypoconid on m4. Marmosopsis d i ffers from

Stembergia and resembles Pucadelphys in having a relatively little marked V-shaped centrocrista

(a primitive feature); however, it differs from both genera in having a very large and straight

metacrista, in the crest-like entoconid, and in the reduction of the talonid of the m4 which has no

hypoconid. Marshall & Muizon (1988) and Muizon (1992) have suggested that Mizquedelphys

from the early Paleocene of Tiupampa could represent a possible morphological ancestor for

Marmosopsis, as far as the teeth are concerned.

The molars of Itaboraidelphys camposi are also structurally similar to those of Pucadelphys

andinus but differ from them in their size, approximately 50% larger, in their more robust

morphology and in their lower and stouter cusps. The upper molars of Itaboraidelphys differ from

those of Pucadelphys in being transversally narrower, in the more pronounced V-shaped

centrocrista, in the larger conules, in the smaller style C which is clearly closer in size to style B

than D and is sometimes twinned, and in the larger and straighter metacrista in occlusal view. The

lower molars of Itaboraidelphys differ from those of Pucadelphys in their higher trigonid (relative

to the talonid) which is more open lingually, in the shallower trigonid and talonid basins, in the

stronger anterocingulid, in the shorter talonid and in the smaller hypoconulid. Marshall &

Muizon (1988) and Muizon (1992) have stated that Andinodelphys from the early Paleocene of

Tiupampa could represent a possible morphological ancestor for Itaboraidelphys, although

Marshall et al. (1989) regard it as a plesion of indeterminate familial attribution which could be

related to the origin of the Australian marsupials.

It is true that Marmosopsis and Itaboraidelphys differ more from Pucadelphys than does

Stembergia, which is the reason why we have suggested a closer relationship between the latter

two (Marshall & Muizon, 1988; Muizon, 1992). However, there is apparently no feature that

could exclude Pucadelphys from being a possible ancestor of Marmosopsis and Itaboraidelphys,

as the principal derived features of Pucadelphys are the V-shaped centrocrista and the larger and

higher metacone relatively to the paracone, two didelphid characters.

Pucadelphys also shows dental similarities with the various species of Alphadon and

Protalphadon of the late Cretaceous of North America, which were classified in the family

Peradectidae by Marshall et al. (1989). The three genera present a similar dental structure with

large styles A to D. but Pucadelphys differs from the other two genera in possessing a V-shaped

centrocrista (straight in Alphadon and Protalphadon), a paracone smaller than the metacone (they

Source: MNHN. Paris

PUCADELPHYS ANDINUS: THE SKULL

are subequal in size in Alphadon and Protalphadon) and an inflated posterior base of the

protocone, which are didelphid characters. Clemens (1966) has suggested that the Tertiary

didelphids were structurally similar to species like Alphadon lulli and Alphadon marshi. As a

matter of fact, Alphadon marslii represents a good structural ancestor for the dental morphology

of Pucadelphys andinus, an observation which confirms the hypothesis of Clemens (1966),

Crochet (1980) and Reig etal. (1987) which states that the Didelphidae may have originated from

the North American Peradectidae. It is noteworthy that some specimens of Alphadon marshi (from

the Maastrichtian of the Scollard Formation, Alberta, Canada) have a slightly V-shaped centrocrista

and a paracone smaller than the metacone (Cifelli, 1990: 315); the same has been observed by

Sahni (1972: 383 and fig. I4p) in a molar referred to Alphadon cf. rhaister from the Campanian

of Judith River Formation (Montana). However, when present, these features are generally less

developed than in the Didelphidae.

Character Analysis. — In this section we analyze character states in Pucadelphys which are

of potential importance in phylogenetic inference. This analysis is made in order to determine

which states are plesiomorphic or apomorphic for Metatheria in particular and Tribosphenida in

general. The relevance of these states for the phylogenetic relationships of Pucadelphys are

summarized in the conclusions below. The character states are discussed in the following order:

dentition, dentary, skull bones, skull foramina, and ear region (petromastoid).

Dentition

Number of incisors. — Hershkovitz (1982) has shown that the ancestral metatherian (and

probably eutherian) primitively had five upper and lower incisors, although the highest number

known is 5/4 which resulted from the loss of il. Therefore, so far, the plesiomorphic number of

incisors for Tribosphenida and Metatheria is 5/4 (II, 12,13,14,15/i2, i3, i4, i5) (Marshall, 1979;

Hershkovitz, 1982; Clemens&Lillegraven, 1986:71). Pucadelphys thus retains the plesiomorphic

state for Tribosphenida and Metatheria.

Structure of upper incisors. — The relative size and structure of metatherian upper incisors

are discussed by Takahashi (1974), Archer (1976b), and Reig et al. (1987). Metatheria were

apparently plesiomorphic in having 11 -5 conical in shape, while the spatulate shape as occurs in

Microbiotheriidae and Australian taxa is the derived state (Archer, 1976b; Reig et al., 1987).

Many Didelphidae (Takahashi, 1974) and Dasyuridae (Archer, 1976b) have the II hypsodont

and semiprocumbent relative to 12, and a small space (diastema) separated II and 12. Archer

(1976b) regarded both states as plesiomorphic for Metatheria, and Hershkovitz (1982) suggested

that these states evolved subsequent to loss of the i 1 and are associated with grooming. Although

both states occur in many generalized metatherians, some taxa have one state but not the other.

In addition, Dromiciops has 11-5 equally spaced and II is not hypsodont (see Reig et al.. 1987, fig.

49) as also occurs in some Dasyuridae (Archer, 1976b, Table 1). The states in Pucadelphys in

which 11-5 are conical, subequal in height and equally spaced (no diastema) are apparently

LARRY G. MARSHALL & CHRISTIAN de MUIZON

plesiomorphic for metatherians because it is possible to derive all the variations seen in

Didelphidae, Dasyuridae. Microbiotheriidae and other groups from a Pucadelphys- like ancestor.

The states in Pucadelphys probably also occurred in the ancestral tribosphenid.

Structure of lower incisors. — Hershkovjtz (1982) illustrated that all fossil metatherians

except Microbiotheriidae have what he termed a staggered i3 (i.e. the second incisor of the four

occurring in generalized metatherians) which has “a bony alveolar buttress on labial surface,

greater root exposure on lingual surface and medially staggered position of root and alveolus”

(p. 191). In addition, the staggered i3 is larger than the i2, i4 or i5. Given the nearly universal

occurrence of this staggered i3 in metatherians, Hershkovitz concluded that this state was

plesiomorphic for this group. He also demonstrated that this state occurs in an edentulous dentary

from the Albian of Texas (see his fig. 5) which places a minimum age for the metatherian-eutherian

dichotomy. However, the lower incisors in Pucadelphys (YPFB Pal 6107), Microbiotheriidae

(Marshall. 1982) and generalized eutherians (Hershkovitz, 1982) are arranged in a linear series

and have no staggered i3. This state was thus probably plesiomorphic for Tribosphenida and was

retained in eutherians and in a few metatherians ( Pucadelphys , Microbiotheriidae). The staggered

i3 is here regarded as a derived state which appears to be synapomorphic for metatherians except

Pucadelphys and Microbiotheriidae.

Number of premolars and molars. — The plesiomorphic number of permanent premolars and

molars in Tribosphenida was probably P5/5 and M4/4, while loss of one molar would account for the

plesiomorphic state in eutherians ol P5/5 and M3/3 and loss of two premolars for the plesiomorphic

state in metatherians of P3/3 and M4/4 (Dashzeveg & Kjelan-Jaworowska, 1984, pp. 225-226). The

loss of deciduous incisors, canines and premolars (only the P3/3 has a predecessor, the HDP 3/3) is also

a derived state for metatherians (Archer et ah, 1985). Pucadelphys thus retains the plesiomorphic

metatherian dental formula of P3/3 and M4/4.

Structure of premolars and molars. — The structure of the cheek teeth in generalized

metatherians is discussed by Archer (1976b), Reig et al. (1987), Marshall (1987) and Marshall

et al. (1989). Based on these studies, Pucadelphys has a cheek tooth structure which agrees

perfectly with members of the family Didelphidae (i.e. paracone smaller and lower than metacone;

V-shaped (dilambdodont) centrocrista; well developed stylar shelf with large stylar cusps A, B,

C and D: posteriorly expanded protoconal base; trigonid shorter than talonid; cristid obliqua

contacts posterior wall of trigonid labial to protocristid notch; entoconid tall and spire-like;

hypoconulid lower than entoconid; well developed pre- and postcingulids). The cheek tooth

character states of Didelphidae relative to other metatherians are listed in Marshall et al (1989

figs 1,2).

Dent ary

Inflected angular process. — The angular process of the dentary is strongly inflected

medially in all metatherians except the living Tarsipes where it is absent, and in Phascolarctos

and Mynnecobius where it is weakly developed. The angular process is inflected medially in most

late Cretaceous eutherians from Asia (i.e. Kennalestes , Asioiyctes and Baninlestes; Kielan-

PUCADELPHYS ANDINUS: THE SKULL

Jaworowska, 1981.1984; Kielan-Jaworowska el al., 1979) and in some species of Gypsonictops

and Cimolestes from North America, in Didymoconidae from the late Eocene-middle Oligocene

of Asia, and in various groups of living and fossil rodents (Marshall, 1979 and references

therein). One is also present in the early Cretaceous age pantothere Vincelestes from South

America (Bonaparte & Rougier, 1987) and in some triconodonts, symmetrodonts, and

Multitubcrculata (Miao, 1988: 878). The distribution of an inflected angular process suggests that

it is a plesiomorphic state for Tribosphenida. Pucadelphys is thus plesiomorphous in this feature.

Mylohyoid groove. — The mylohyoid groove marks the course of a neurovascular bundle

which includes the mylohyoid artery and nerve, and presumably in early taxa the persistent

Meckel’s cartilage (Krebs, 1971 ). It is present in Jurassic dryolestids (Krebs, 1971); in docodonts,

triconodonts and eupantotheres (Kielan-Jaworowska, 1981:6 1); in the early Cretaceous triconodont

Gohiconodon ostromi (Jenkins & Sciiaff, 1988:5); in the early therian Kielantherium (Dashzeveg

& Kielan-J aworowska, 1984: 221); in the Cretaceous eutherians Prokennalestes and Kennalestes

(Kielan-Jaworowska, 1981: 61 and Kielan-Jaworowska & Dashzeveg, 1989); in some living

adult eutherians [insectivorans, edentates, cetaceans; (Bensley, 1902)]: and in embryos of

monotremes (Watson, 1916). It is absent in multituberculates, some dryolestids (i.e. Crusafontia ;

Krebs, 1971), paurodontids ( Amphitherium , Peramus); and some early eutherians (. Asioryctes',

Kielan-Jaworowska. 1 981 :6 1 ). It is absent in the Deltatheroida (Kielan-Jaworowska & Nessov,

1990), but present in Pucadelphys , embryos of some living marsupial taxa and, sporadically, in

some adult specimens of Didelpltis , Trichosurus, Phalanger, Perameles and Petauroides (Bensley,

1902). The presence of a mylohyoid groove is thus the plesiomorphic state for Mammalia, and

Pucadelphys retains the primitive condition. The presence or absence of a mylohyoid groove

appears (contra Archer el al., 1985, fig. 3, point 14) to be of dubious value in phylogenetic

inference.

Labial mandibular foramen. — Kielan-Jaworowska & Dashzeveg (1989) named labial

mandibular formen a small foramen on labial side of the dentary at the base of the coronoid process

present in many primitive mammals. This small foramen occurs in Kielantherium (aegialodontid)

from the early Cretaceous of Asia, in Prokennalestes (otlestid) from the Aptian-Albian of Asia,

in Otlestes (otlestid) from the late Cenomanian of Uzbekistan and in Zalambdalestes

(zalambdalestid) from the Campanian of Mongolia (Dashzeveg & Kielan-Jaworowska, 1984;

and Kielan-Jaworowska & Dashzeveg, 1989), and in the microbiothere metatherian

Microbiotherium gallegosense from the early Miocene ot Argentina (Marshall, 1982). Dashzeveg

& Kielan-Jaworowska (1984) and Kielan-Jaworowska & Dashzeveg (1989) regard the presence

of this foramen as a therian plesiomorphy. There is no trace of this foramen in Pucadelphys.

Bones or Skull

There are numerous states in Pucadelphys which, as shown by previous workers,

apparently represent Tribosphenida plesiomorphies. These include: nasals flared posteiioily

(Gregory, 1920: 139); alisphenoid and squamosal have major component in lateral wall of

braincase (Kermack etal., 1981: 135; Kemp, 1982:305. 1983 : 372); glenoid fossa set posteriorly

LARRY G. MARSHALL & CHRISTIAN de MUIZON

opposite anterior half of promontorium (Kielan-Jaworowska, 1981: 65); occipital plate slopes

upwards and slightly forwards from condyles (Kielan-Jaworowska, 1981: 65); and lambdoidal

crest sharp, sagittal crest weak or absent (Kielan-Jaworowska et al., 1979: 227).

Orbit large, confluent with temporal fossa. — This state occurs in Vincelestes (Bonaparte &

Rougier, 1987), Deltatheroida (Kielan-Jaworowska, 1975), Zalambdalestes (Kielan -Jaworowska,

1984: 113), Asioryctes and Kennalestes (Kielan-Jaworowska, 1981: 56), Pucadelphys, and in

generalized Cenozoic and living metatherians and eutherians. This is apparently the plesiomorphic

state for Theria.

Lacrimal. — The lacrimal has a large facial wing in cynodonts, Morganucodon (Kermack

etal, 1981), Vircelestes (Bonaparte & Rougier, 1987), Deltatheridium (Kielan-Jaworowska,

1975), Pucadelphys and in generalized living and fossil metatherians (Gregory, 1920). In

Vincelestes, Deltatheridium , Pucadelphys and in generalized living and fossil metatherians the

lacrimal forms a prominent antorbital rim and the lacrimal foramen occurs within the orbital rim.

I he structure ol the lacrimal and position ot the lacrimal foramen in Pucadelphys apparently

represents the plesiomorphic state for Tribosphenida.

Nasal-lacrimal contact. — A broad nasal-lacrimal contact occurs in cynodonts,

tritylodontids, multituberculates, AforganwcodowiKERMACK eta/., 1981), Vincelestes ( Bonaparte

& Rougier, 1987, fig. 2A), Deltatheridium (Kielan-Jaworowska, 1975: 122, fig. 3A), and among

metatherians only in South American Borhyaenoidea (Marshall, 1978; Marshall etal., 1989)

and in the Australian Wynyardia (Gregory, 1920). A broad nasal-lacrimal contact is the

plesiomorphic state for mammals (Gregory, 1920). Living and other fossil metatherians (including

Pucadelphys) are derived in having no nasal-lacrimal suture but a clear frontal-maxilla contact

(Gregory, 1920; Marshall & Kielan-Jaworowska, 1992). An intermediate condition within

metatherians occurs in some Didelphidae ( Didelphis , Chironectes) which have only a narrow and

variable frontal-maxi I la contact (Gregory, 1920:139). Most eutherians are also derived in having

a broad frontal-maxilla contact (Novacek, 1986: 27).

Contact of alisphenoid and parietal. — A broad contact between the alisphenoid and

parietal on the outside of the skull occurs in the early eutherians Asioryctes and Kennalestes

(Kielan-Jaworowska, 1981), in Pucadelphys, and in many Cenozoic and living metatherians (i.e.

all Didelphidae, Myrmecobiidae, most Dasyuridae, some Borhyaenoidea; Archer, 1976a: 309).

There is no contact in Thylacinidae, Peramelidae, Vombalidae, some Phascolarctidae, some

Borhyaenoidea, and some Dasyuridae (Archer, 1976a: 309). A broad alisphenoid-parietal

contact is regarded as the plesiomorphic state in Tribosphenida.

Palate. The bony palate is solid (without vacuities) in cynodonts, monotremes, some

multituberculates ( Kamptobaatar ), the eupantothere Vincelestes (Bonaparte & Rougier, 1987),

the late Cretaceous eutherians Zalambdalestes (Kielan-Jaworowska, 1984: 108), most living and

Cenozoic eutherians, the deltatheroidan Deltatheridium (Kielan-Jaworowska, 1975: 122), and

many metatherians (Borhyaenoidea, Sparassocynus, Caluromys, Dasycercus, Dasyuroides,

species groups ol Antechinus and Sminthopsis, Myrmecobius, Dactylopsila, Petaurus and

Dactylonax) (Marshall, 1979 and references therein). In contrast, the palate is perforated by

large vacuities, often two pairs, in some multituberculates (i.e. Sloanbaatar, Nemegtbaatar,

Source . MNHN, Paris

PUCADELPHYS ANDINUS: THE SKULL

Bulganbaatar and Ptilodus), some eutherians (rabbits, some rodents, macroscelidids, hedgehogs,

Carpolestes), some Deltatheroida (Kielan-Jaworowska & Nessov, 1990), and many living and fossil

metatherians (Marshall, 1979; Reig etal., 1987). Because of the wide occurrence of palatal vacuities

among metatherians, their presence has generally been regarded as plesiomorphic for this group.

However, when vacuities are present they are generally quite variable within families and even within

species. Ontogenetic studies of metatherian embryos have shown that the palatal plates of the maxilla

and palatine bones are at first solid and develop vacuities by gradual resoiption of bone. These and other

observations favor the view that a solid palate was the plesiomorphic state for mammals (Marshall,

1979 and references therein). Pucadelphys is thus plesiomorphic for this state.

Preglenoid process of jugal. — The jugal is deep, long and extends posteriorly to form a

preglenoid process along the anterolateral edge of the glenoid fossa in the late Cretaceous

metatherian Didelphodon vorax (Clemens, 1966: 72), in Cenozoic and living metatherians, in

Deltatheroida (Kielan-Jaworowska & Nessov, 1990), in the early Cretaceous eupantothere

Vincelestes (Bonapartc & Rougier, 1987), and in some living eutherians (elephants, some

hyraxes, some rodents; Marshall, 1979). In some late Cretaceous eutherians from Asia (Asioryctes,

Kennalestes, Barunlestes) the jugal is long, extends to the anterior edge of the glenoid fossa, but

does not form a preglenoid process (Kielan-Jaworowska, 1981; 1984). The presence of a

preglenoid process of the jugal is clearly a plesiomorphic state for metatherians, and the presence

of this slate in Vincelestes and some eutherians suggests that it is plesiomorphic for Theria as well.

Pucadelphys thus retains the plesiomorphic metatherian (and therian) state.

Foramina of Skull

Pucadelphys retains two states which are regarded by some workers as plesiomorphic for

Tribosphenida: V2 exits skull through foramen rotundum in alisphenoid (Kermack et al ., 1981:

135), and presence of a foramen lacerum medium (MacIntyre, 1967; Archer, 1976a). However,

a foramen rotundum does not occur in Kennalestes or Asioryctes, and Kielan-Jaworowska (1981:

65) regarded a “foramen rotundum confluent with sphenorbital fissure” as the plesiomorphic

therian state. The earliest known foramen rotundum in a eutherian occurs in the ?middle

Campanian age Barunlestes (Kielan-Jaworow'sk3 & Trofimov, 1980).

Transverse canal. — There is no transverse canal in Morganucodon (Kermack et al.,

1981), monotremes (Watson, 1916), multituberculates (Kielan-Jaworowska et al, 1986),

Deltatheroida (Kielan-Jaworowska & Nessov, 1990). Borhyaenoidea (except for a possible

vestige in Lycopsis; Marshall, 1977: 415), Pucadelphys, some didclphids (i.e. Caluromys) and

some dasyurids (some Planigale) (Archer, 1976a) or in late Cretaceous eutherians from Asia

(Asioryctes,Barunlestes,Kennalestes,Zalambdalestes', Kielan-Jaworowska, 1981:1984; Kielan-

Jaworowska & Trofimov, 1980). Among metatherians a transverse canal occurs in thylacinids,

myrmecobiids, peramelids, most didelphids and most dasyurids (Archer, 1976a).

Archer (1976a) regarded a transverse canal as a plesiomorphic state for metatherians, yet

the distribution of this feature suggests that its absence was the plesiomorphic state for mammals.

If this polarity is true, then a transverse canal evolved multiple times within metatherians, a

LARRY G. MARSHALL & CHRISTIAN de MUIZON

position supported by data in Archer (1976a: 307). For example, when present, the transerse

canal, which transmits a major venous sinus that drains the base of the brain around the pituitary,

varies considerably in size and morphology; “in dasyurids, thylacinids, some didelphids (e.g.

Marmosa and Monodelphis) the canal passes not only into endocranium but also transversely

through the basisphenoid”, sometimes “via more than one canal”. In these taxa the transverse

venous sinus “may enter the endocranium via the sulci for the cntocarotid canals”, while in other

taxa (e.g. peramelids and other didelphids) “both ends of this canal lead anteriorly for a

considerable distance before communicating transversely through the basisphenoid”.

Archer (1976a: 307) stressed that taxa which lack this canal have a venous drainage pattern

considerably different from those taxa in which it is present. In some species of Planigale, for

example, the “venous sinus passes transversely via the anterior edges of the enormous foramina

pseudovale [sic, see below] which are so large that they incise the aiisphenoid in the normal

position of the transverse canal”. Given the arrangement in some species of Planigale, we

speculate that in metatherians like Pucadelphys which lack this canal, part of the venous drainage

around the pituitary may leave the skull through the large foramen ovale. Pucadelphys, which

lacks a transverse canal, retains the plesiomorphic metatherian (and mammalian) state.

Foramen ovale. — This term is used (sensu Kielan-Jaworowska et al., 1986: 584) to

designate the foramen which transmits the mandibular branch of the trigeminal nerve (V3). Within

this definition is included the foramen “pseudovale” of earlier workers, a term which Kielan-

Jaworowska et al. (1986: 584) “rejected as prejudicial to possible homology of anterior lamina

of petrosal with part of mammalian aiisphenoid (Presley, 1981; Kemp, 1983)”. This move took

into consideration the fact that the shape of this foramen and detailed relations with surrounding

bones vary within mammals, and that MacIntyre's (1967) use of “pseudovale” is not consistent

regarding these relationships since the foramen may either be totally surrounded by the anterior

lamina or be in the junction between two or more bony elements. Because of this, “pseudovale”

was regarded as not being a clear-cut alternative to ovale and its use implies a false impression of

precision of definition (Presley, personal communication, 1989). Thus, the term ovale is used in

preference for homology regarding transmittal of the V3 and gives no weight to the bones

surrounding the V3 in different mammal groups.

A great deal of confusion regarding the term ovale and “pseudovale"has resulted from the

studies of MacIntyre (1967) and Archer (1976a). MacIntyre, for example, recognized a

"pseudovale” in some living eutherians (i.e. some perissodactyls, artiodactyls, rodents) which

formed as a result ol the disappearance of a true foramen ovale (which was surrounded entirely

by the aiisphenoid) by its union with the foramen lacerum medium, which primitively in mammals

is located at the junction of the alisphenoid-basisphenoid-basioccipital-petrosal (MacIntyrc,

1967: 834). In this sense, MacIntyre’s “ pseudovale” is derived from an ancestral (plesiomorphic)

state in which a true foramen ovale was surrounded by the aiisphenoid which in turn is derived

from yet another ancestral state in which a “pseudovale” was located between the anterior lamina

of the petrosal and the aiisphenoid. MacIntyre’s “pseudovale” is thus used in two significantly

di I ferenl non homologous contexts: true "pseudovale” (as in cynodonts, tritylodonts, monotremes)

- true ovale (a derived state in generalized tribosphenids) - and “pseudovale” (=true ovale +

foramen lacerum medium), (another derived state, in some living eutherians).

Source: MNHN, Paris

PUCADELPHYS ANDINUS: THE SKULL

Archer (1976a) recognized a foramen ovale and “pseudovale” in the taxa he studied, and

in some cases identified both on the same side of the same specimen (i.e. most Didelphidae: 307).

He defined (p. 219) his “pseudovale” as: “Transmits cranial nerves. Sometimes carries very small

arterial or venous anastomoses linking internal cranial and external cranial vessels and sometimes

near its posterior edge, carries small branch of internal carotid artery into eustachian canal”. In his

Plate 1 where he labels this foramen on specimens, what he calls the “pseudovale” is in reality the

medial lacerate foramen as evidenced by its position at the alisphenoid-basioccipi tal- basisphenoid-

petrosal junction (MacIntyre, 1967: 834). In other instances, Archer uses “pseudovale” sensu

MacIntyre (1967) in part, meaning the foramen formed by union of the medial lacerate foramen

and ovale. Archer thus uses “pseudovale” in reference to two different nonhomologous foramina.

However, Archer's use of foramen ovale appears to be correct (i.e. foramen in alisphenoid which

transmits V3) and it is interesting that he regards (p. 307) its absence as a dasyurid plesiomorphy.

Because Archer used the presence, absence and structural variation of his “pseudovale” and ovale

in phylogenetic inference, his conclusions based on these features need reconsideration.

Subsquamosal foramen. —This foramen occurs in Pucadelphys, in all metatherian groups

studied by Archer (1976a: 314) and in the late Cretaceous eutherians Asioryctes and Kennalestes

(Kielan-Jaworowska,1981 : 61). Archer (1976a) noted that this foramen is variably developed: in

some specimens it occurs on one side of the skull but not the other, while in other specimens it lies

adjacent to the postglenoid foramen and is separated from it by only a splint of bone. The presence of

this foramen represents a plesiomorphic state within Theria (Kielan-Javvorowska, 1981: 65).

Postzygomatic foramen. — This foramen occurs in Pucadelphys and in all metatherian

groups studied by Archer (1976a: 314). It is a plesiomorphic feature within Metatheria.

Ear Region

Pucadelphys retains the following states which are plesiomorphic for Tribosphenida: a

triossicular middle ear mechanism (Miao & Lillegraven, 1986; Kemp, 1983); an inflated

promontorium to house the spiral cochlea (MacIntyre, 1972; Prothero, 1983: 1044); a fully

coiled cochlea (380°) (Miao & Lillegraven. 1986; Wible, 1990); a fossa incudis with prominent

lateral wall formed by the squamosal (Wible, 1990); a true aqueductus cochleae (Kermack etai,

1981: 137; Wible, 1990); the internal acoustic meatus is a broad shallow pit with foramina which

transmit a branch of the facial nerve (VII) and two branches of the acoustic nerve (VIII) (Prothero,

1983: 1041); the presence of a posttemporal foramen (Wible, 1990); and the presence of a lateral

head vein (absence of this vein is a eutherian synapomorphy; Marshall, 1979; Wible, 1990). It

also has three states which are metatherian synapomorphies: the prootic canal is reduced in length

and width, and does not open endocranially (absence of this canal is a eutherian synapomorphy;

Wible. 1990); the prootic sinus continues onto the squamosal side of the petromastoid within a

deep sulcus (Wible, 1990); and the presence, on the squamosal side of the petromastoid, of a

sphenoparietal emissary vein which occupies a deep sulcus continuous with the sulcus for the

prootic sinus, and exits skull through the postglenoid foramen (Wible, 1990).

LARRY G. MARSHALL & CHRISTIAN de MUIZON

Ossified auditory bulla. — In Pucadelphys there is no evidence of an ossified auditory

bulla. This same state occurs in monotremes, soricids, some talpids, and some fossil eutherians

(i.e. Asioryctes, Kennalestes, Palaeoryctes) and apparently represents the plesiomorphic mam¬

malian state (Novacek. 1977: 141). “A condition similar to that in the monotreme Tachyglossus ,

where the tympanic cavity was bordered ventrally by a thin connective tissue (dense fibrous and/

or areolar) membrane located between the ventromedial surface of the nearly horizontal tympanic

ring and the ventral surface of the petrosal, was the most primitive condition in monotremes,

marsupials, and placentals” (Novacek, 1977: 144) (however, see below for discussion on the

orientation of the typmanic ring).

Based on an authoritative study of fossil and living metathcrians, Archer (1976a: 310)

concluded that, among living taxa, didelphids were most plesiomorphic in aspects of their bullar

morphology. Reig et al. (1987: 22) suggested that, among living didelphids, Metachirus and

Philander were the most generalized. In these taxa the ossified bulla is incomplete: a small

tympanic process of the alisphenoid occurs anteriorly, a small tympanic process of the pars petrosa

of the petromastoid occurs posteriorly, and much of the middle ear in macerated skulls is open

ventrally. This same architecture was regarded as the plesiomorphic metatherian bullar state by

Novacek (1977) and apparently occurred in the late Cretaceous stagodontid Didelphodon vorax

(Clemens, 1966: 55). There are thus two features (both absent in Pucadelphys) which warrant

special consideration in the evolution of the metatherian bulla: the ossified tympanic process of

the alisphenoid and the tympanic process of the pars petrosa.

An ossified tympanic process of the alisphenoid occurs in the vast major of metathcrians and

is generally regarded as a synapomorphy of that group (Kielan-Jaworowska & Nf.ssov, 1990),

although its phylogenentic value has been recently questioned by Muizon (1994). This feature is,

however, absent in borhyaenoids [with the exception of Cladosictis and Sipalocyon (Muizon,

1994)], in vombatids (Patterson, 1965), and in what appears to be a late Cretaceous metatherian

that was figured but not discussed by MacIntyre (1967, fig. 3). Kielan-Jaworowska (1981: 61)

suggested that “the marsupial alisphenoid bulla originated from the enlarged quadrate ramus of

the alisphenoid of early therians” and that in late Cretaceous eutherians ( Asioryctes and probably

Kennalestes) “a strongly inflated quadrate ramus of the alisphenoid occupies the same position

as the alisphenoid bulla in marsupials and is probably homologous to it”. However, in eutherians

the alisphenoid rarely contributes to the formation of the ossified bulla and in those taxa where

it is developed (i.e. macroscelidids, some insectivores; Novacek, 1977; Segall, 1970) it has a

different shape and occurs in adifferent location than in metatherians (Kielan-Jaworowska, 1981:

61-62). Furthermore, in insectivores “the tympanic process of the basisphenoid takes the place of

the tympanic process of the alisphenoid of marsupials” (Segall, 1970: 169). These observations

clearly suggest that the ossified tympanic wing of the alisphenoid in metatherians and eutherians

are nonhomologous and evolved independently in each group. This view is supported by the fact

that an ossified alisphenoid bulla is absent in monotremes (W atson, 1916), in the early Cretaceous

eupantothere Vincelestes (Bonaparte & Rougier, 1987; Rougier & Bonaparte, 1988), in

Pucadelphys , in Mayulestes (Muizon, 1994) and in some late Cretaceous eutherians (Kielan-

Jaworowska, 1981; 1984). Furthermore, Muizon (1994) has shown that the absence of tympanic

PUCADELPHYS ANDINUS: THE SKULL

Process of the alisphenoid in Mayulestes and in most other borhyaenoids is a symplesiomorphy

within the superfamily [not a reversal as stated by others (Marshall & Kielan-Jaworowska,

1992)] and that its presence in Cladosictis and Sipalocyon is a synapomorphy of this clade. The

absence of a tympanic process of the alisphenoid in the didelphid Pucadelphys reinforces

Muizon’s statement that this structure must have appeared several times during marsupial

evolution and should not be used to diagnose this group of mammals.

The tympanic process of the pars petrosa (sensu Reig et al. , 1987; =rostral tympanic process

of the petrosal. Wible, 1990; ectotympanic process of the periotic, Archer, 1976a: 230) is

“intimately related" to the posteroventral end of the ossified ectotympanic; these bones are

typically in contact but never fused (Archer, 1976a: 230). A tympanic process of the pars petrosa

is absent and the promontorium is smooth in Sinoconodon, morganucodontids, triconodontids,

Vincelestes, many metatherians (Dcltatheroida, Pucadelphys , most Borhyaenoidea; living

Phalangeridae, Vombatidae, and possibly Phascolarctidae), and many eutherians (late Cretaceous

taxa from Asia, many Cenozoic and living goups) (Wible, 1990). A low ridge which may represent

an incipient state of this structure occurs in the metatherian Didelphodon vorax and in Petrosal

Types A-D of Wible (1990). This feature is well developed in living metatherians (except those

mentioned above), in some eutherians (euprimates, erinaceomorphs), possibly in some

multituberculates, and monotremes (although in this group it is not clearly homologous with that

in metatherians and eutherians; Wible, 1990). The distribution of this structure among various

groups thus suggests that mammals primitively had a smooth promontorium and lacked a

tympanic process of the pars petrosa. Wible (1990) concluded that the absence of this feature was

a plesiomorphic state for Eutheria, but was uncommital about the plesiomorphic state in

Metatheria, suggesting only “that ridges and processes have been added to and lost from the

promontorium a number of times in marsupials”. We believe that metatherians (as in Pucadelphys)

primitively lacked this structure and that it evolved independently in various lineages, possibly

in association with the development and enlargement of an ossified auditory bulla.

Orientation of ectotympanic and tympanic membrane. — In living metatherians and

eutherians (except soricids and some talpids) the ossified ectotympanic and tympanic membrane

are oriented in a nearly vertical plane, and this state has been regarded as a therian synapomorphy

(Kemp, 1983:375). However, this view is not corroborated by fossil or ontogenetic evidence which

suggests two other possibilities. First, Novacek (1977) suggested that the ectotympanic was

primitively subhorizontal in position. This possibility is supported by the horizontal position of

the ectotympanic in the living monotreme Tachyglossus (Novacek, 1977: 144, Fig. 6), in the late

Paleocene multituberculate Lambdopsalis (Miao & Lillegraven, 1986), and the knowledge that

the ectotympanic is oriented horizontally in the early ontogeny in all mammals and remains in this

position in monotremes but rotates to an inclined position in most therians (de Beer, 1937) which

thus represents a derived state (Rowe, 1988). Second, Kielan-Jaworowska (1981) suggested that

the ectotympanic was primitively inclined at an angle of 45° from horizontal. This possibility is

based on the observations that in cynodonts and some early mammals with a double jaw joint (i.e.

Morganucodon) the angular (=ectotympanic) is sharply inclined (Kielan-Jaworowska, 1981:

38). Support for this possibility came with the discovery of the late Cretaceous eutherians

LARRY G. MARSHALL & CHRISTIAN de MUIZON

Asioryctes and Kennalestes in which the in situ ectotympanic was released from the lower jaw but

still retains its primitive anterior position opposite the posterior part of the dentary and is roughly

parallel to it at 45° from horizontal. If this does indeed represent the primitive therian state, then

both the subhorizontal state in Tcichyglossus and the near vertical state in some living metatherians

and eutherians are derived conditions.

Yet, the same data base has been interpreted by some workers to support both possibilities

(i.e. subhorizontal is primitive and inclined at 45° is primitive), suggesting that differences

between the two are relative and in part semantic. The best example is Novacek’s (1977: 145)

paper in which he argues that “the last common ancestor of marsupials and placentals... and the

most primitive members of each group... had a simple tympanic ring only slightly inclined to the

horizontal plane of the skull” which resembled the monotreme-like state. In his fig. 7a he

illustrates this plesiomorphic state with the ectotympanic inclined at 35° from horizontal, while

on p. 141 he notes that "the simple ring-shaped ectotympanic lies at a low angle (less than 50°)

to the horizontal plane of the skull”, a state seen in monotremes, soricids, some talpids, and

possibly in Asioryctes and Palaeoryctes.

In Pucadelphys there is no evidence of an ossified ectotympanic, nor are there facts to

demonstrate that one was present. Its absence is attributed to destruction during fossilization. The

tympanic membrane was apparently subvertical or slightly obliquely oriented at an angle much

larger than 45° from horizontal, because this is the general orientation of the fenestra vestibuli (see

YPFB Pal 6110) and the ectotympanic parallels it in Recent mammals (J. Wible, written

communication, 1990). Therefore, whatever the primitive condition is (either sub horizontal or at

45°) the inferred subvertical position of the ectotympanic of Pucadelphys would represent a

derived condition.

Auditory sinuses. — “Auditory sinuses... are cavities within or between bones in the

auditory region of the skull (other than the epitympanic recess) which communicate directly with

the epitympanic recess” (Archer, 1976a: 226). Such sinuses are absent in Morganucodon

(Kermack et al. , 1981), monotremes (Watson, 1916), Vincelestes (Bonaparte & Rougier, 1987),

late Cretaceous eutherians ( Asioryctes, Barunlestes , Kennalestes, Zalamhdalestes ; Kielan-

Jaworowska, 1981; 1984; Kielan-Jaworowska & Trofimov, 1980), and generalized living

eutherians (Van Kampen, 1905; Novacek, 1977). Among Cenozoic and living metatherians,

auditory sinuses are primitively absent, although a small alisphenoid hypotympanic sinus appears

in all groups studied by Archer (1976a) except some borhyaenoids. The absence of auditory

sinuses in Pucadelphys thus represents the plesiomorphic state for mammals.

Pars mastoidea contribution to occiput. — The pars mastoidea has an extensive exposure

on the occiput in monotremes ( Tachyglossus ), in late Cretaceous eutherians ( Asioryctes,

Kennalestes-, Kielan-Jaworowska, 1981, fig. 2), and among metatherians in Didelphodon vorax,

Eodelphis browni. Petrosal Types A and B from the late Cretaceous (Wible, 1 990), in Mayulestes

(Muizon, 1994) and in many taxa studied by Archer (1976a). An extensive contribution of the pars

mastoidea to the occiput also occurs in Pucadelphys, and this apparently represents the plesiomorphic

state for Tribosphenida.

Source . MNHN, Paris

PUCADELPHYS ANDINUS: THE SKULL

Mastoid and paroccipital processes. — These processes are either absent or only

incipiently developed in Morganucodon , monotremes, Vincelestes (Bonaparte & Rougier,

1987), in late Cretaceous Asiatic eutherians ( Asioryctes , Barunlestes,Kennalestes, Zalambdalestes',

Kielan-Jaworowska, 1981; 1984; Kielan-Jaworowska & Trofimov, 1980), and in most gener¬

alized metatherians ( Mayulestes; Myrmecobiidae; most Dasyuridae; Caenolestidae;

Microbiotheriidae; some Didelphidae, Lestodelphys, Monodelphys) (Archer, 1976a; Reig et

al. , 1987). A well developed mastoid and/or paroccipital process does occur in various late

Cretaceous (Wible, 1990), Cenozoic and Recent metatherian groups (Archer, 1976a; Reig eta!.,

1987), where they are derived and apparently evolved independently in various lineages (Archer,

1976a). The absence of a paroccipital process and small size of mastoid process as occur in

Pucadelphys , apparently represent or approximate the plesiomorphic state for mammals.

Shape of fenestra vestihuli and stapedial footplate. — The ratio of length to width of the

fenestra vestibuli approximates or represents that of the stapedial footplate. In the monotreme

Tachyglossus this ratio is 1.0 (Segall, 1970: 203, fig. 26); in Morganucodon 1.1, triconodontids

and Bug Creek multituberculates 1.3, Vincelestes 1.0 (Wible, 1990); and the late Paleocene

multituberculate Lambdopsalis bulla 1.0 (Miao & Lillegraven, 1986). Among metatherians,

living Didelphidae have ratios ranging from 1.3 - 1.5 (Segall, 1970), Didelphodon vorax 1.4

(Archibald, 1979), late Cretaceous taxa studied by Wible (1990) range from 1.3-1.6, while living

Dromiciops and Macropus have 2. 1 (Segall, 1970). Bug Creek eutherians range from 2.0 - 2.4

(Archibald, 1979) and living eutherians 1.8 - 2.9 (Segall, 1970). These data suggest that: a

circular fenestra vestibuli with a ratio of 1.0 is plesiomorphic for Mammalia (Archibald, 1979;

Segall, 1970; Prothero, 1983; Miao & Lillegraven, 1986; Wible, 1990); a slightly ovoid

fenestra with a ratio of 1.3 is a synapomorphy of metatherians and possibly Tribosphenida;

Pucadelphys with a ratio of 1.4 approximates the plesiomorphic state for metatherians; and an

elliptical fenestra with a ratio of 1.8 is a synapomorphy of eutherians (Wible, 1990).

Stapedial artery’. — The stapedial artery, a branch of the internal carotid artery, occurs in

embryos of living monotremes, metatherians and eutherians, but only in adults of Omithorhynchus and

some eutherians (Presley, 1979). As evidenced by a groove which marks the course of this artery on

the pars petrosa of monotremes and some eutherians, it is believed to have been present in

multituberculates (Kielan-Jaworowska etal. . 1986), late Cretaceous eutherians from Asia {Barunlestes,

Zalambdalestes-, Kielan-Jaworowska. 1984: 108) and North America (Wible, 1990) and various

Cenozoic eutherians (Presley, 1979). In metatherians, a stapedial artery is absent in adults of all living

taxa(TANDLER, 1899; Presley, 1979) and the groove is absent in fossil taxa (Archer, 1976a; Archibald,

1979; Clemens, 1966; Wible, 1990). The presence of a stapedial artery in adults is thus a plesiomorphic

state for mammals (Presley, 1979: 241, fig. 3), while the absence of this artery in adult metatherians

represents a synapomorphy of that group (Wible, 1990). Pucadelphys , which lacks a groove tor this

artery, has the plesiomorphic state for metatherians.

Course of internal carotid artery’. — The course of the internal carotid artery, which is

derived from the embryonic dorsal aorta, has two basic states in mammals which arise by a

process of differential growth affecting the relative positions of the dorsal aorta and cochlear

promontory” ( Presley, 1979: 238). Wible (1986) cautions that the position of the internal carotid

LARRY G. MARSHALL & CHRISTIAN de MUIZON

artery with regard to the cochlear promontory cannot be inferred from the location of the

entocarotid foramen alone. The probable plesiomorphic state for mammals (Archibald, 1979;

Presley, 1979) is called the medial internal carotid artery (MICA) where the vessel passes medial

to the cochlear promontory on its way to the entocarotid foramen and leaves little or no trace on

the bone. This state occurs in monotremes, all metatherians (MacIntyre, 1972: 291), in late

Cretaceous Asian eutherians (Kielan-Jaworowska, 1981:58-59; 1984: 115: Kielan-Jaworowska

& Trofimov, 1980), and in some living eutherians [e.g. rodents, rabbits, ungulates (Presley, 1979:

Kielan-Jaworowska. 1981: 57)]. The apomorphic state is called the promontory internal carotid

artery (PICA) where the vessel passes upon or lateral to the cochlear promontory, often in a sulcus.

This state occurs in some primates, insectivores and some carnivorans, and may have evolved

multiple times (Presley, 1979). Pucadelphys thus retains the plesiomorphic MICA state, where

the course of the internal carotid is marked by a shallow sulcus.

In addition, the presence of a large entocarotid foramen between the alisphenoid and

basisphenoid which transmits the internal carotid as in Pucadelphys appears to be the plesiomorphic

state in metatherians (Archer, 1976a), eutherians (Kielan-Jaworowska, 1981), and their

tribosphenid ancestor.

Sulcus for facial nerve. — A sulcus for the facial nerve on the petromastoid has been

regarded as a synapomorphy of metatherians and eutherians (i.e. MacIntyre, 1972); yet, this

feature also occurs in morganucodontids, triconodontids, multituberculates, monotremes and the

eupantothere Vincelestes', it is therefore a plesiomorphy for mammals ( Wible. 1990). Pucadelphys

retains the plesiomorphic state.

Inferior petrosal sinus and foramen. — The inferior petrosal sinus joins posteriorly with

the internal jugular vein and exits the skull through the inferior petrosal (internal jugular)

foramen (Archibald, 1979). The term internal jugular foramen (=canal) was proposed by Archer

(1976a) to replace the posterior carotid foramen of Gregory (1910: 233) and Patterson (1965:

2) which they erroneously believed transmitted a branch of the internal carotid artery; this error

was perpetuated by Clemens (1966: 73) and Marshall (1977, 1978).

A sulcus for the inferior petrosal sinus occurs in Morganucodon , monotremes, and in many

metatherians and eutherians, and apparently represents a plesiomorphy (nectherian synapomorphy

of MacIntyre, 1972) within mammals. An inferior petrosal foramen occurs nearly universally in

metatherians (Archer, 1976a: 309) but is apparently absent in some (microbiotheres, Acrobates,

possibly Phascolarctos ; Paterson, 1965: 6), and is widely distributed among fossil and living

eutherians (Patterson, 1965). Its presence apparently represents the plesiomorphic stale for

Tribosphenida. Pucadelphys retains both an inferior petrosal sinus and foramen.

Anterior lamina. — The anterior lamina is a bony blade located at the anterodorsal border

of the petrosal in several primitive mammals: Adelobasileus , Haldanodon, Megazostrodon,

Morganucodon, Sinoconodon, Trioracodon , multituberculates, Vincelestes , and monotremes

(see Wible & Hopson, 1993 for original references). The anterior lamina is perforated by a

foramen (foramen ovale) for the V3 (mandibular branch of the trigeminal nerve), in Morganucodon,

Sinoconodon, Adelobasileus, most multituberculates, Vincelestes, and Ornithorhynchus. However,

there is a single trigeminal foramen between the anterior lamina and alisphenoid in Megazostrodon

PUCADELPHYS ANDINUS: THE SKULL

and Haldanodon (Wible & Hopson, 1993). The V2 (maxillary branch of the trigeminal nerve) is

also enclosed in the anterior lamina in Morganucodon, Sinoconodon, Adelobasileus, and

Vincelestes.

In the non-tribosphenic taxa mentioned above “the anterior lamina contributes to the lateral

wall of the cavum epiptericum, the extradural space within which the trigeminal ganglion lies”

(Wible & Hopson, 1993: 47). Kermack et al. (1981: 119), in their authoritative study of the skull

anatomy of Morganucodon, state that the trigeminal ganglion (= semilunar ganglion) lies in a deep

pocket of the medial side of the anterior lamina, lateral to the medial opening of the aqueductus

Fallopii for the facial (VII) cranial nerve. A similar condition is observed in Trioracodon

(Kermack, 1963: 87), in multituberculates (Kielan-Jaworowska et al., 1986), in Haldanodon

(Lillec.raven & Krusat, 1991:97) and in Omithorhynchus (Zeller, 1989a, b). The medial aspect

of the anterior lamina of Vincelestes has not been described. In the recent didelphid Monodelphis

domestica, the trigeminal ganglion lies anterolateral to the hiatus Fallopii in a shallow fossa of the

alisphenoid and the anterior border of the petrosal (Maier, 1987; Hopson & Rougier, 1993). In

Pucadelphys, the fossa observed on the medial side of the lamina of the pars petrosa anterolateral

to the hiatus Fallopii. is interpreted here as the Gasserian fossa or fossa for the trigeminal ganglion.

Since this ganglion is located in the anterior lamina of the petrosal in Morganucodon , Trioracodon ,

Haldanodon , multituberculates, and Omithorhynchus, the anterior wing of the pars petrosa of

Pucadelphys is interpreted here as a structure homologous to the anterior lamina of these early

mammals. Furthermore, the anterior lamina of Pucadelphys occupies the same position as the

anterior lamina in early cynodonts (Kemp, 1983: 375), morganucodonts (Kermack et al., 1981),

triconodonts (Kermack, 1963; Crompton & Jenkins, 1979, fig. 3-5 A-C), and Vincelestes (Hopson

& Rougier, 1993). The anterior lamina of Pucadelphys is greatly reduced when compared to that

of Morganucodon, Sinoconodon , multituberculates and Vincelestes but it is larger than in any

other marsupials. However, this feature does not constitute a true novelty for the group since

Wible (1990: 200) and Wible & Hopson (1993: 60), stated that the great reduction or absence of

the anterior lamina of the petrosal is a synapomorphy of the marsupials and cutherians, therefore

implicitly accepting the presence of a reduced anterior lamina in some Tribosphenida.

The presence of an anterior lamina in Pucadelphys confirms Wible’s assertion that “the

anterior lamina...may not be entirely absent in marsupials” (Wible, 1990: 200). As explained by

this author, “the prootic canal runs through the anterior lamina of the petrosal in Sinoconodon,

morganucodontids, triconodontids, multituberculates, and Vincelestes”. The occurrence of a

prootic canal in some marsupials seems to confirm that the structure it passes through may

represent a reduced anterior lamina (in peramelids, dasyurids and some didelphids). However,

Wible & Hopson (1994) have recently shown that the prootic canal of marsupials is probably not

homologous to that in morganucodonts, triconodonts, multituberculates, and Vincelestes. If they

are correct, therefore, the prootic canal should not be used for homologies of the anterior lamina.

The anterior lamina of Pucadelphys differs from that of the non-tribosphenic mammals

mentioned above, because it does not participate in the closure of the lateral wall of the skull; it

is covered laterally by the squamosal and the alisphenoid which articulate together and replace the

LARRY G. MARSHALL & CHRISTIAN di: MUIZON

anterior lamina in that function. At the level of the anterior lamina of Pucadelphys, the lateral wall

of the skull is thus duplicated; it is formed by the alisphenoid and the squamosal laterally and by

the petrosal (anterior lamina) medially. This condition demonstrates that the anterior lamina was

expelled from the lateral wall of the skull before it totally regressed. As a consequence of the

reduction of the anterior lamina, in marsupials, the mandibular branch of the trigeminal nerve (V3)

exits the skull (via the foramen ovale) between the alisphenoid and the anterior border of the

petrosal (this condition is very clear in the early diverging borhyaenoid marsupial Mayulestes).

In Pucadelphys , however, the anterolateral border of the petrosal (the posterior edge of the

foramen ovale) probably retains part of the anterior lamina; therefore, the foramen ovale of

Pucadelphys is located between the alisphenoid and the anterior lamina of the petrosal. In

eutherians the foramen ovale is located in the alisphenoid.

In morganucodonts, triconodonts, Sinoconodon and Adelohasileus the anterior lamina is

large and the alisphenoid is relatively short anteroposteriorly but tall dorsoventrally (Kermack,

1963; and Kermack etal, 1981; Lucas & Luo, 1993; Crompton & Luo, 1993). Monotremes have

a large anterior lamina posterodorsal to a reduced (anteroposteriorly and dorsoventrally) alisphenoid

and posteroventral to the orbitosphenoid. In most multituberculates the anterior lamina is large,

posterodorsal (or posterior) to a large orbitosphenoid and posterodorsal to the alisphenoid

(Kielan-Jaworowska, 1971; 1974: Kemp, 1988). The alisphenoid of multituberculates is gener¬

ally small (Kielan-Jaworowska, 1971; Kemp, 1982), but in the multituberculate Lambdopsalis it

is large, totally separates the anterior lamina from the orbitosphenoid, and has tribosphenid

proportions (Miao, 1988). Large anterior lamina and alisphenoid also occur in Vincelestes

(Rougier & Bonaparte, 1988; Hopson & Rougier, 1993).

The homologies of the anterior lamina and alisphenoid in monotremes and in the other non-

tribosphenic mammals mentioned above have been discussed by Presley & Steel (1976),

Presley (1980: 1981), and Hopson & Rougier (1993). The status of the anterior lamina of

monotremes has been abundantly studied in order to elucidate whether or not it is homologous to

the anterior lamina of the other non-tribosphenic taxa mentioned above. The anterior lamina in

living monotremes is an intramembranous ossification called the “lamina obturans” which fuses

secondarily to the pars petrosa (Vandebroek, 1964; Wible, 1990). Presley (1981) has confirmed

that the "lamina obturans”, in living monotremes, is an independent center of intramembranous

ossification within the sphenobturator membrane. This author also states that “there seem to be

no fundamental difference between therians and monotremes in the early development of this

bone. Difference arise only when it fuses with its neighbours. ... Cynodonts may have had an

anterior part contributing to the broad blade of the epipterygoid and a posterior forming the smaller

anterior process of the petrosal”. The proportions of these two parts of the lamina obturans may

have varied since in triconodonts (Kermack, 1964) and morganucodonts (Kermack etal., 1981)

the anterior lamina of the petrosal is large and the epipterygoid blade small. In multituberculates,

given the extreme reduction of the alisphenoid (epipterygoid) the lamina obturans may have

contributed extensively to the petrosal, approaching the monotreme condition (Presley, 1981).

These studies suggest that the anterior lamina fused to the pars petrosa as in early cynodonts,

morganucodonts and triconodonts is derived from the posterior part of the “lamina obturans”.

Source MNHN. Paris

PUCADELPHYS ANDINUS: THE SKULL

which apparently was an ossification independent from the pars petrosa; while the anterior lamina

in monotremes, multituberculates represents some combination of posterior and anterior part of

the “lamina obturans”. Miao (1988) has suggested that the “lamina obturans” has an apparent

plasticity for alliance within itself and with adjacent bones, and this could explain the within group

variability of the anterior lamina in multituberculates. Considering that the blade of the alisphenoid

and the anterior lamina of the petrosal are probably equivalent, Presley (1981) concluded that a

simple change in the fusion of these elements during development could produce a therian pattern

of the lateral wall of the braincase. This was ratified by Kemp (1983:374), who noted: “the sidewall

of the non-therian braincase is apparently homologous [s.l.] with the sidewall of the modern

therian braincase, and the two groups differ only in the relatively trivial matter of which bone the

anterior lamina [s.l.] finally fuses with towards the end of ontogeny”.

Recently. Hopson & Rougier (1993) have questioned the interpretation of Presley & Steel

(1976), Presley (1981), and Kemp (1983). Presley (1981) and Zeller (1989 a. b) have demon¬

strated that in Oniithorhynchus the lamina obturans (anterior lamina) ossifies within the spheno-

obturator membrane, entirely separate from the otic capsule and only in later development

becomes synostosed to it (Hopson & Rougier, 1993:278). The lamina obturans of Omithorhynchus

begins to ossify in the dorsal part of the spheno-oblurator membrane, quite distant from the ala

temporalis (the endochondral portion of the alisphenoid) and only late in ontogeny it expands

anteroventrally to contact the small alisphenoid (Hopson & Rougier. 1993: 282). Maier (1987)

described the development of the alisphenoid of Monodelphis domesticci. He demonstrated that

the intramembranous portion of this bone is initiated adjacent to the ala temporalis and only at a

fairly late ontogenetic stage expands posteriorly to contact the otic capsule. Hopson & Rougier

(1993: 283-284) therefore conclude that: “even though both elements form in the spheno-

obturator membrane, the developmental evidence indicates that they are distinct, therefore non-

homologous, ossifications”. The monotremc lamina obturans is presumably homologous with the

anterior lamina (Hopson & Rougier, 1993:284) and therefore, contrary to Presley & Steel (1976)

and Presley (1981), the anterior lamina is not homologous with the intramembranous portion

(formed within the spheno-obturator membrane) of the alisphenoid. Furthermore, the alisphenoid

of Vincelestes is similar to that of modern therians in its great anteroposterior expansion and its

position relative to the otic capsule. It is noteworthy that in Oniithorhynchus the trigeminal

ganglion lies in a fossa of the medial side of the lamina obturans (Zeller, 1989b). Therefore, the

homology of the monotreme lamina obturans with the anterior lamina (which we accept here)

reinforces the interpretation of the anterior lamina of Pucadelphys, since the trigeminal ganglion

appears to lie constantly in the same structure, the small fossa on the medial side of the anterior

lamina lateral to the hiatus Fallopii.

The occurrence of a reduced anterior lamina in a marsupial is by no mean a surprise since

the evolution of the therian ( Vincelestes + marsupials + placentals) lateral wall of the braincase

is characterized by the development of an anteroposteriorly elongated alisphenoid (observed in

all therians) and the reduction and loss of the anterior lamina of the petrosal in the Tribosphenida.

The presence, in Pucadelphys, of an anterior lamina, large for a marsupial but reduced when

compared to that of Vincelestes (the sister group of the Tribosphenida), demonstrates that the

LARRY G. MARSHALL & CHRISTIAN de MUIZON

morphology of the lateral wall of the braincase of Pucadelphys is intermediate between that of

Vincelestes and that observed in the other marsupials. Therefore, it may approximate the condition

of the ancestral Tribosphenida. The morphology of the lateral wall of the braincase of Pucadelphys

corroborates Wible (1990) and Wible & Hopson (1993) synapomorphy of the Tribosphenida

(marsupials + placentals): reduction or absence (interpreted here as loss) of the anterior lamina of

the petrosal. It is also in agreement with the interpretation of Vincelestes as the sister group of

marsupials + placentals (Tribosphenida) (Row, 1993; and Wible & Hopson, 1993). Furthermore,

it corroborates the interpretation of Hopson & Rougier (1993) who have refuted the hypotheses

ot Presley & Steel (1976) and Presley (1981) on the homologies of the elements of the lateral

wall of the mammalian braincase.

CONCLUSIONS

The skull of Pucadelphys represents a classic example of mosaic evolution. It has a

commonly observed combination of plesiomorphic and apomorphic states which, collectively,

complicates its placement in recent phylogenies (i.e. Marshall etal., 1989; Marshall & Kielan-

Jaworowska, 1992) although recent discoveries have shed some light to its interpretation

(Muizon, 1994). However that may be, it illustrates that we still have much to learn about early

metatherian and tribosphenid evolution.

The vast majority of structures in the skull and dentary of Pucadelphys represents

plesiomorphic states for Tribosphenida and/or Mammalia. A few features represent metatherian

synapomorphies (i.e. cheek tooth formula P3/3 and M4/4; prootic canal reduced in length and

width, and does not open endocranially; prootic sinus continues onto squamosal side of

petromastoid within a deep sulcus; presence of sphenoparietal emissary vein which occupies a

deep sulcus on squamosal side of petromastoid, is continuous with sulcus for prootic sinus, and

exits skull through postglenoid foramen; length/width ratio of fenestra vestibuli; absence of

stapedial artery) which securely demonstrate that Pucadelphys is a member of this group, as

presently defined.

The remaining structures suggest contradictory affinities within Metatheria. The derived molar

features agree perfectly with members of the Didelphidae (sensu Marshall et al, 1989) and for this

reason Pucadelphys has been assigned to this family (Marshall. & Muizon, 1988). However,

Pucadelphys retains the plesiomorphic state of a non-staggered i3, while all other metatherians (except

Microbiotheriidae) have the derived state of a staggered i3. It is also plesiomorphic in having 11-5

conical, equally spaced and II not hypsodont, while most other metatherians are derived in either

having spatulate 11-5 (a synapomorphy of Australidelphia; Marshalls al. , 1989) or II hypsodont and

separated from 12 (many Didelphidae and Dasyuridae).

The plesiomorphic mammal state of a clear nasal-lacrimal contact occurs among metatherians

only in Borhyaenoidea, Deltatheroida (if they actually are metatherians) and the Australian

Wynyardia , while all other known taxa (including Pucadelphys) have the derived state of a frontal-

maxilla contact.

Source: MNHN . Paris

PUCADELPHYS ANDINUS: THE SKULL

Pucadelphys lacks an ossified alisphenoid bulla, a feature present and traditionally regarded as

a synapomorphy (Kielan-Jaworowska & NeSsov, 1990) in all other metatherians. The Deltatheroida,

which apparently possess a tympanic process of the alisphenoid, are classified by Kielan-Jaworowska

& Nessov (1990) within the metatherians. The absence of a tympanic process of the alisphenoid in some

Borhyaenoidea has been inteipreted as representing a derived state from an ancestor which had an

alisphenoid bulla (Archer, 1976a; Marshall & Kielan-Jaworowska, 1992). However, the morphol¬

ogy of the skull of MayulestesJ'erox, a borhyaenoid from the early Paleocene of Tiupampa indicates

that the lack of a tympanic process of the alisphenoid is the plesiomorph condition for the borhyaenoids

(Muizon, 1994). Furthermore, the absence of a tympanic process of the alisphenoid in Pucadelphys ,

an early Paleocene didelphoid may indicate its independent acquisition in the other members of the

superfamily. It is noteworthy that a tympanic process is probably present in the late Cretaceous

stagodontid, Didelphodon (Clemens, 1966) and in a probable marsupial (Asiadelphia) from the Barun

Goyot Formation (middle Campanian) of Mongolia (Trofimov & Szalay, 1993). This reinforces the

idea that the tympanic process of the alisphenoid appeared several times independently within

marsupials (Muizon, 1994) and should not be used to diagnose this group, which, therefore, partially

questions the inclusion of the Deltatheroida within the Metatheria (Kielan-Jaworowska &

Nessov, 1990).

The presence of an anterior lamina fused to the petromastoid, and a foramen ovale (exit of

V3) which opens between the anterior lamina and alisphenoid represent the most peculiar

structures in the skull of Pucadelphys. These features have not been reported as such in other

tribosphenids, although their presence has been predicted in the tribosphenid ancestor because

these states occur in non-tribosphenid mammals. Comparisons of Pucadelphys with Vincelestes

(the sister group of the Tribosphenida) and with the other marsupials indicates that it represents

a morphological stage of the structure of the lateral wall of the braincase, intermediate between

the primitive ( Vincelestes ) and modern (tribosphenid) therians and therefore allows a better

understanding of character states and evolution. It is probable that Pucadelphys represents or

approximates the plesiomorphic state in the morphology of the lateral wall of the braincase of early

Tribosphenida.

Pucadelphys thus appears to represent the paleontological confirmation of the hypothesis

of Hopson & Rougier (1994) who have refuted that of Presley & Steel (1976) and Presley (1981)

for the evolution of the tribosphenid ear region. If the skull of Pucadelphys was of early-late

Cretaceous age and had a generalized tribosphenid molar structure, then it would probably be

celebrated as the long sought “missing-link” and be readily accommodated and accepted into the

current scheme of early tribosphenid evolution. However, Pucadelphys is a metatherian, has a

derived didelphid molar structure, and is early Paleocene in age. Thus the predicted anterior

lamina and foramen ovale bordered by it and the alisphenoid occur in a dentally derived and

geologically recent metatherian.

Considering the major differences existing between Pucadelphys and the other didelphids

(especially concerning the ear region and the lateral wall of the braincase), one could be tempted

to classify it in a new family. However, as mentioned above, all the cranial features of Pucadelphys

are plesiomorphic for marsupials and the only clearly derived features within marsupials concern

LARRY G. MARSHALL & CHRISTIAN de MUIZON

the dental morphology and relate Pucadelphys to the Didelphidae. We regard Pucadelphys as a

primitive Didelphidae, member of the order Didelphimorphia (sensa Marshall et al , 1989)

because its molar structure (which is currently the foundation of metatherian systematics) is

indistinguishable from that family. We thus give preference to molar structure in classifying

Pucadelphys within Metatheria.

ACKNOWLEDGEMENTS

Aspect of this study were made possible by three grants (2467- 82. 2908-84, 3381-86) from the National Geographic

Society; an “Action Spdcifique” (N° 1337) from the MNHN, Paris; a grant from the Gordon Barbour Fund. Department of

Geological and Geophysical Sciences. Princeton University; two grants from the National Science Foundation (EAR-8804423.

INT-8814059); a Collaborative Research Grant (86/0013) from the North Atlantic Treaty Organization (NATO) obtained by C.

de Muizon forL. G. M. and two “Profcsseur Associe" appointments at the Museum National d'Histoire Naturelle(MNHN), Paris,

France obtained by C. de Muizon for L. G. M. Fossil collecting was initially carried-out under auspices of the Instituto Boliviano

de Biologfade Altura (IBB A). La Pazand the Office de la Recherche Scientifiqucct Technique d'Outre-Mer (Mission ORSTOM),

Bolivia; then in collaboration with the Centro de Tecnologfa Petrolera (CTP) of the Yacimientos Petrolfferos y Fi scales de Bolivia

(YPFB). Santa Cruz, Bolivia; and currently in collaboration with the Asociacion Boliviana de Palcontologfa and the Fundacion

Para Las Ciencias in Cochabamba, Bolivia. All specimens described in this study were collected between 1982-1987 and are the

property of CTP. YPFB. Special thanks to M. Suarez R.. R. Suarez S.. C. Molina M., J. Lobo B.. P.-Y. Gagnier, and R. CSspedes

for their collaboration and logistic support. For help on various aspects of this work we thank J. A. Case, M. C. McKenna, K. A.

Kermack, Z. Kielan-Jaworowska, R. Presley. D. E. Russell, B. Sige, D. Sigogneau, .1. R. Wible. M. O. Woodburne, and an

anonymous reviewer. D. Russell and D. Sigogneau provided casts for comparative study. The stereophotos were made by D.

Serrette, the line drawing are by E. Liebman, the SEM’s are by C. Weber-Chancoc-ne.

REFERENCES

Anonymous. 1983. — Paleontological Institute of the USSR Academy of Sciences, Vnestorgizdat. Isdatelstov, N 3496: 1-35.

Archer. M., 1976a. — The basicranial region of marsupicamivores (Marsupialia), relationships of carnivorous marsupials, and

affinities of the insectivorous marsupial peramelids. Zoological Journal of the Linnean Society of London, 59:217-322.

Archer, M.. 1976b.—The dasyurid dentition and its relationships tothaiofdidelphids.thylacinids.borhyacnids(Marsupicarnivora)

and peramelids (Peramelina: Marsupialia). Australian Journal of Zoology, Supplementary Scries. 39: 1-34.

Archer, M.. 1978. — The nature of the molar-premolar boundary in marsupials and reinterpretation of the homology of marsupial

cheekteeth. Memoirs of the Queensland Museum, 18: 157-164.

Archer, M.. Flannery. T. F.. Ritchie, A. & Molnar. R. E.. 1985. — First Mesozoic mammal from Australia: an early Cretaceous

monotreme. Nature, 318: 363-366.

Archibald, J.D., 1979.—Oldest known eutherian stapes and a marsupial petrosal bone from the late Cretaceous of North America.

Nature, 281: 669-670.

Beer. G. de, 1937. — Development of the Vertebrate Skull. London, Oxford University Press, 552 p.

Bensley. B. A.. 1902. — On the identification of the Meckelian and mylohyoid grooves in the jaws of Mesozoic and Recent

mammals. University of Toronto Studies, Biological Series, 3: 75-82.

Bonaparte, J. F., & Rougier. G., 1987. — MamiTeros del Crctacico Inferior de Patagonia. IV Congreso Latinoamericano de

Paleontologia . Santa Cruz, Bolivia, 1: 343-359.

Source MNHN . Paris

PUCADELPHYS ANDINUS: THE SKULL

Bonaparte ,J. F., van Valon. L. M., & Kramartz, A., 1993. — La fauna local dc Punta Peligro, Paleoceno inferior.de la Provincia

del Chubut, Patagonia. Argentina. Evolutionary monographs. 14: 1-61.

Cifelli, R. L.. 1990. — Cretaceous mammals of southern Utah. I. Marsupials from the Kaiparowits Formation (Judithian). Journal

of Vertebrate Paleontology . 10 (3): 295-319.

Clemens, W. A., 1966. — Fossil mammals of the type Lance Formation. Wyoming. Part II. Marsupialia. University of California

Publications in Geological Sciences, 62: 1-122.

Crochet, J. Y., 1980. — Les Marsupiaux du Tertiaire d'Europe. Editions dc la Fondation Singer-Polignac. Paris. 279 p.

Crompton, A. W., & Jenkins, F. A., 1979. — Origin of mammals, pp. 59-73, In: Mesozoic Mammals: The First Two-Thirds of

Mammalian History,]. A. Lillegraven.Z. Kielan-Jaworowskaand W. A. Clemens (eds). University of California Press,

Berkeley.

Crompton. A. W., & Luo. Z.. 1993. — Relationships of early mammals. Sinoconodon. Morganucodon oehleri, and Dinnetherium,

pp. 30-44 hr. M animal Phytogeny. Mesozoic differenciation , Multituberculates, Monotremes. early Therians and

Marsupials. F. S. Szalay, M. J. Novacek and M. C. McKenna (eds). Springer-Vcrlag, New York.

Dashzeveg, D., & Kielan-Jaworowska, Z.. 1984. — The lower jaw of an aegialodontid mammal from the early Cretaceous of

Mongolia. Zoological Journal of the Linnean Society of London, 82: 217- 227.

Dom. R.. Fisher, B. L., & Martin. G. F., 1970. — The venous system of the head and neck of the opossum ( Didelphis virginiana).

Journal of Morphology, 132: 487-496.

Gayet, M.. Marshall, L. G.. & Sempere, T., 1992. — The vertebrate-bearing late Cretaceous-Paleocene of Bolivia and its

stratigraphic context: a review. Revista Tecnica de Yacimientos Petrolferos Fiscales dc Bolivia, Cochabamba, 12 (3-

4, December 1991): 393-433.

Goodrich, E. S., 1958. — Studies on the structure and Development of Vertebrates. Dover, London.

Gregory, W. K., 1910. — The orders of mammals. Bulletin of the American Museum of Natural History . 27: 1-524.

Gregory. W. K.. 1920. — Studies in comparative myology and osteology; N IV. A review of the evolution of the lacrymal bone

of vertebrates with special references to that of mammals. Bulletin of the American Museum of Natural History, 42 (2):

95-263.

Henkel, S., & Krebs. B.. 1977. — Der ertse Fund eines Saugetier-Skelcttes aus der Jura-Zeit. Umschau in Wissenschaft und

Technik, 7: 217-218.

Hershkovitz, 1982. —The staggered marsupial lower third incisor (13). Geobios, Memoire Special. 6: 191-200.

Hopson. J. A.. & Rougier. G. W., 1993. — Braincasc structure in the oldest skull of a therian mammal: implications for mammalian

systematic and cranial evolution. American Journal of Science, 293: 268-299.

Jenkins, F. A.. & Schafe. C. R.. 1988. — The early Cretaceous mammal Gobiconodon (Mammalia. Triconodonta) from the

Cloverly Formation in Montana. Journal of Vertebrate Paleontology. 8 (1): 1-24.

Jollie. M.. 1962. — Chordate Morphology. Reinhold Publishing Corporation. New York.

Kemp. T. S.. 1982. — Mammal-like Reptiles and the Origin of Mammals. Academic Press, New York. 363 pp.

Kemp. T. S., 1983. — The relationships of mammals. Zoological Journal of the Linnean Society of London, 77: 353-384.

Kemp. T. S., 1988. — A note on the Mesozoic mammals, and the origin of therians. pp. 23-29. In: The Phytogeny and Classification

of the Tetrapods , M. J. Benton (ed.). Volume 2: Mammals, Clarendon Press, Oxford.

Kermack, K. A.. 1963. — The cranial structure of the triconodonts. Philosophical Transactions of the Royal Society of London,

B246: 83-103.

Kermack, K. A.. & Kielan-Jaworowska. Z.. 1971. — Therian and non-therian mammals, pp. 103-115, In: Early Mammals, D.

M. Kermack and K. A. Kermack (eds.). Zoological Journal of the Linnean Society of London, Supplement 1, Number

50.

LARRY G. MARSHALL & CHRISTIAN de MUIZON

Kermack. K. A., Mussett, F., & Rigney, H. W., 1981. — The skull of Morganucodon. Zoological Journal of the Linnean Society

of London, 71: 1-158.

Kielan-Jaworonvska, Z.. 1971. — Skull structure and affinities of the Multituberculata. Results of the Polish-Mongolian

Paleontological Expeditions. Part III. Palaeontologia Polonica, 25: 5-41.

Kielan-Janvorowska. Z.. 1974. — Multituberculate succession in the late Cretaceous of the Gobi Desert (Mongolia). Results of

the Polish-Mongolian Paleontological expeditions. Part V. Palaeontologia Polonica, 30: 23-44.

Kielan-Jaworowska.Z.. 1975. — Evolution of the thcrian mammals in the late Cretaceous of Asia. Part I. Dellatheridiidae. Results

of the Polish-Mongolian Paleontological Expeditions. Part VI. Palaeontologia Polonica, 33: 103-132.

Kielan-Jaworonvska. Z., 1981. — Evolution of the therian mammals in the late Cretaceous of Asia. Part IV. Skull structure of

Kennalestes and Asioryctes. Results of the Polish-Mongolian Paleontological Expeditions. Part IX. Palaeontologia

Polonica, 42: 25-78.

Kielan-Jaworonvska. Z., 1984. — Evolution of the therian mammals in the late Cretaceous of Asia. Part V. Skull structure in

Zalambdalestidae. Results of the Polish-Mongolian Paleontological Expeditions. Part X. Palaeontologia Polonica, 46:

107-117.

Kielan-Janvorowska. Z.. & Dashzeveg. D.. 1989. — Eutherian mammals from the early Cretaceous of Mongolia. Zoologica

Scripta , 18 (2): 347-355.

Kielan-Jaworonvska. Z., & Nessov. L. A.. 1990. — On the metatherian nature of the Deltatheroida. a sister group of the

Marsupialia. Lethaia, 23: 1-10.

Kielan-Janvorowska. Z.. & Trofimov. B. A., 1980. — Cranial morphology of the Cretaceous eutherian Barunlestes. Acta

Palaeontologica Polonica, 25 (2): 167-185.

Kielan-Jaworonvska, Z., Bown, t. m., & Lillegraven, J. A.. 1979. — Eutheria. pp. 221-258. In: Mesozoic Mammals: The First

Two-Thirds of Mammalian History. J. A. Lillegraven, Z. Kiclan-JaNvoroNvska and W. A. Clemens (eds). University of

California Press, Berkeley.

Kielan-Janvorowska, Z.. Crompton. A. W., & Jenkins. F. A.. 1987. — The origin of egg-laying mammals. Nature, 326: 871 -873.

Kielan-Jaworonvska, Z.. Presley. R., & Poplin, C, 1986. — The cranial vascular system in lacniolabidoid multituberculate

mammals. Philosophical Transactions of the Royal Society of London. 313 (1164): 525-602.

Krebs. B.. 1971. — Evolution of the mandible and lower dentition in dryolestids (Pantotheria, Mammalia), pp. 89-102. In: Early

Mammals , D. M. Kermack and K. A. Kermack (eds). Zoological Journal ofthe Linnean Society’of London, Supplement

I. Number 50.

Krebs. B.. 1987. — The skeleton of a Jurassic cupantothere and the arboreal origin of modern mammals, pp. 132-137. hr.Mesozoic

Terrestrial Ecosystems, P. J. Currie and E. II. Koster (eds). Occasional Paper of the Tyrrell Museum of Paleontology.

Number 3.

Krebs. B.. 1991. — Das Skeletl von Henkelotherium guimarotae gen. et sp. nov. (Eupantotheria. Mammalia)ausdem oberen Jura

von Portugal. Berliner Geowissenschaftliche Ahhandlungen. ser. A. 133: 1-121.

Lillegraven, J. A., & Krusat. G.. 1991. — Cranio-mandibular anatomy of Haldanodon exspectatus (Docodonta. Mammalia)

from the late Jurassic of Portugal and its implications to the evolution of mammalian characters. University of Wyoming

Contribution to Geology, 28: 39-138.

Lucas. S. G.. Luo, Z.. 1993. — Adelobasileus from the upper Triasic of West Texas: the oldest known mammal. Journal of

Vertebrate Paleontology, 13 (3): 309-334.

Luckett, P.. 1993. — An ontogenetic assessment of dental homologies in therian mammals, pp. 182-204. In: Mammal Phytogeny,

Mesozoic differentiation, Multituberculates, Monotremes, early Therians and Marsupials. F. S. Szalay, M. J. Novacek

and M. C. McKenna (eds), Springer-Vcrlag. New York.

Source: MNHN. Paris

PUCADELPHYS ANDINUS: THE SKULL

MacIntyre. G. T.. 1967. — Foramen pscudovale and quasi mammals. Evolution, 21: 834-841.

MacIntyre, G. T.. 1972. —The trisulcale petrosal pattern of mammals, pp. 275-303, In: Evolutionary Biology, T. Dobzhansky,

M. K. Hecht and W. C. Steere (eds.). Appleton-Century-Crofts, New York. Volume 6.

Maier. W., 1987. — The ontogenetic development of the orbitotemporal region in the skull of M onodelphis domestica

(Didelphidae, Marsupialia) and the problem of the mammalian alisphenoid, pp. 71-90. In: Morphogenesis of the

Mammalian Skull. H. J. Kuhn and U. Zeller (eds). Mammalia Depicta , Heft 13, Beihefte zur Zeitschrifl fur

Saugetierkundc, Verlag Paul Parey Spitalerstrasse, Hamburg.

Marshall. L. G.. 1976. — New didclphinc marsupials from the La Venta fauna (Miocene) of Colombia, South America. Journal

of Paleontology, 50: 402-418.

Marshall. L. G., 1977. — A new species of Lycopsis (Borhyacnidae: Marsupialia) from the La Venta fauna (late Miocene) of

Colombia. South America. Journal of Paleontology, 51: 633-642.

Marshall, L. G.. 1978. — Evolution of the Borhyacnidae, extinct South American predaceous marsupials. University of

California Publications in Geological Sciences, 117: 1-89.

Marshall, L. G.. 1979. — Evolution of metatherian and eutherian (mammalian) characters: a review based on cladistic

methodology. Zoological Journal of the Linnean Society of London, 66: 369-410.

Marshall. L. G.. 1982. — 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 Sao Jose de Itaborai, Brazil, pp. 91 -160. In: Possums and Opossums. M. Archer (ed.). Surrey Beatty & Sons. Sydney.

Marshall, L. G.. 1989. — The K-T boundary in South America: on which side is Tiupampa? National Geographic Research, 5

(3): 268-270.

Marshall. L. G.. & Kielan-Jaworowska. Z.. 1992. — Relationships of the dog-like marsupials, deltatheroidans and early

tribosphenic mammals. Lethaia, 25: 361-374.

Marshall. L. G., & Muizon, C. de, 1984. — Un nouveau marsupial didelphide (Itaboraidelphys camposi nov. gen., nov. sp.) du

Paleocene moyen (Itaboraien) de Sao Jose de Itaborai (Brcsil). Comptes rendus de PAcademic des Sciences, Paris.

299: 1297-1300.

Marshall, L. G.. & Muizon, C. de. 1988. — The dawn of the age of mammals in South America. National Geographic Research.

4(1): 23-55.

Marshall. L. G., & Sigogneau-Russoll. D.. 1995. — Part III: the postcranial skeleton. In: C. de Muizon (ed.), Pucadelphys

andinus (Marsupialia. Mammalia) from the early Paleocene of Bolivia. Memoires du Museum national d'Histoire

nature He. 165: 91-164. Paris.

Marshall. L. G., Case. J. A.. & Woodburne. M. O.. 1989. — Phylogenetic relationships of the families of marsupials. Current

Mammalogy. 2: 433-502.

Marshall, L. G.. Muizon. C. de. Gayet, M.. Lavenu, A., & Sige, B.. 1985. — The “Rosetta Stone" for mammalian evolution in

South America. National Geographic Research, 1 (2): 247-288.

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, 35: 477-500.

McKenna, M. C, 1975. —Toward a phylogenetic classification of the Mammalia, pp. 21-46. In: Phytogeny of the Primates . W.

P. Luckett and F. S. Szalay (eds). Plenum Press, New York.

Miao, D.. 1988. — Skull morphology of Lamb dop sal is bulla (Mammalia. Multituberculata) and its implications to mammalian

evolution. Contributions to Geology. University of Wyoming. Special Paper. 4: 1-104.

Miao. D.. & Lillegraven, J. A.. 1986. — Discovery of three ear ossicles in a multituberculate mammal. National Geographic

Research, 2 (4): 500-507.

LARRY G. MARSHALL & CHRISTIAN de MUIZON

Muizon. C. de. 1992. — La fauna de mamiTeros dcTiupampa (PaIcoceno inferior. Formacfon Sania Lucia) Bolivia. Revista tecnica

de Yaciminentos Petroliferos Fiscales de Bolivia , 12 (3-4, December 1991): 575-624.

Muizon. C. de. 1994. — A new carnivorous marsupial from the Paleocene of Bolivia and the problem of marsupial monophyly.

Nature , 370: 208-211.

Muizon. C. de. & Brito. I. M.. 1993. — Lc bassin calcaircdc Sao Josede Itaboraf (Rio de Janeiro. Bresil): ses relations fauniques

avec le site de Tiupampa (Cochabamba. Bolivic). Annales de Paleontologie. 79 (3): 233-268.

Muizon. C. de. Gayet. M.. Lavenu. A.. Marshall. L. G.. Sige, B.. & Villaroel.C.. 1983. — Late Cretaceous vertebrates including

mammals from Tiupampa. southcentral Bolivia. Geobios, 16: 747- 753.

Muizon. C. de. Marshall, L. G.. & Sige, B.. 1984. —The mammal fauna from the El Molino Formation (late Cretaceous-

Maestrichtian) at Tiupampa. southcentral Bolivia. Bulletin du Museum national d'Histoire nature lie, Paris, Serie 6.

Section C. 4: 327-351.

Novacek, M. J.. 1977. — Aspects of the problem of variation, origin, and evolution of the eutherian auditory bulla. Mammal

Review, 7: 131-149.

Novacek, M. J.. 1986. —The skull of leptictid insectivorans and the higher-level classification of eutherian mammals. Bulletin

of the American Museum of Natural History, 183 (1): 1-111.

Novacek, M. J., Wyss, A. R.. & McKenna, M. C., 1988. — The major groups of eutherian mammals, pp. 31 -71. In: The Phytogeny

and Classification of the Tetrapods , M. J. Benton (cd.). Volume 2: Mammals. Clarendon Press, Oxford.

Ortiz Jaureouizar, E.. & Pacual, R.. 1989. — South American land-mammal faunas during the Cretaceous-Tertiary transition:

evolutionary biogeography. Contribution to the symposium, Cretaceous in Latin America, Part A Events and

sedimentological record , Buenos Aires: 231-252.

Patterson. B.. 1965. — The auditory region of the borhyaenid marsupial Cladosictis. Breviora, 217: 1-9.

Paula couto, C. de. 1962. — Didelfideos fosiles del Paleoceno de Brasil. Revista del Museo Argentino de Ciencias naturales

"Bernardino Rivadavia ”, Ciencias Zoologicas, 112: 135-166.

Presley. R.. 1979. — The primitive course of the internal carotid artery in mammals. Acta Anatomica. 103: 238-244.

Presley, R., 1980. — The braincase in Recent and Mesozoic therapsids. Memoiresde la Societe Geologique de France, n s. 139:

159-162.

Presley. R.. 1981. — Alisphenoid equivalents in placentals, marsupials, monotremes and fossils. Nature, 294: 668-670.

PRESLEY, R.. & Steel. F. L. D., 1976. — On the homology of the alisphenoid. Journal of Anatomy, 121: 441-459.

Prothero. D. R.. 1983. — The oldest mammalian petrosals from North America. Journal of Paleontology. 57 (5): 1040-1046.

Reig, O. A.. Kirsch. J. A., & Marshall. L. G., 1987. — Systematic relationships of the living and Neocenozoic opossum-like

marsupials (suborder Didclphi morphia) with comments on the classification of these and the Cretaceous and Paleogene

New World and European metatherians. pp. 1 -89. In: Possums and Opossums, M. Archer (cd.), Surrey Beatty & Sons,

Sydney.

Rougher, G.. & Bonaparte. J. F., 1988. — La pared lateral del craneo de Vi n celestes neuque n ian us (Mammalia, Eupantotheria)

y su importancia en el estudio de los mamiTeros Mesozoicos. V JornadasArgentinas de Paleontologfa de Vertebrados,

Resumenes, La Plata (16-19 de Mayo, 1988): 14.

Rougher, G.. Wible, J. R.. & Hopson, J. A., 1992 — Reconstruction of the cranial vessels in the early Cretaceous mammal

Vincelesles neuquenianus: implications for the evolution of the mammalian cranial vascular system. Journal of

Vertebrate Paleontology, 12 (2): 188-216.

Rowe, T.. 1988. — Definition, diagnosis, and origin of Mammalia. Journal of Vertebrate Paleontology, 8 (3): 241-264.

Rowe, T., 1993. — Phylogenetic systematic^ and the early history of mammals, pp. 129-145. In: Mammal Phytogeny. Mesozoic

differenciation. Multituberculates, Monotremes, early Therians and Marsupials, F. S. Szalay. M. J. Novacek and M.

C. McKenna (eds). Springer-Verlag, New York.

Source: MNHN, Paris

PUCADELPHYS ANDINUS: THE SKULL

Sahni, A., 1972. — The vertebrate fauna of the Judith River Formation, Montana. Bulletin of the American Museum of Natural

History . 147: 321-412.

Segall, W.. 1969a. — The middle ear region of Dromiciops . Acta Anatomica, 72: 489-501.

Segall. W., 1969b. — The auditory ossicles (malleus, incus) and their relationships to the tympanic in marsupials. Acta

Anatomica, 73: 176-18 1 .

Segall, W„ 1970. — Morphological parallelisms of the bulla and auditory ossicles in some insecti vores and marsupials. Fieldiana

Zoology. 51: 169-205.

Tandler, J.. 1899. — Zur vergleichenden Anatomie der Kopfarterien bei den Mammalia. Denkschriften der Kaiserlichen

Akademie der Wissenschaften, 69: 677-784.

Takahashi. 1974. — Variagao morfologica de incisivos cm Didclfideos (Marsupialia-Didclphinae). Anais Academia Brasileira

CienciaSy Rio de Janeiro. 46 (3-4): 413-416.

Trofimov, B. A.. & Szalay. F. S.. 1993. — New group of Asiatic marsupials (order Asiadelphia) from the late Cretaceous of

Mongolia. Journal of Vertebrate Paleontology. 13 (3): 13A. Abstracts (supplement to Number 3).

Vandebroek, G., 1964. — Rcchcrches sur Porigine des mammiferes. Annates de la Societe Royale de Belgique . 94: 117-160.

Van der Klaauvv. C. J.. 1913. —The auditory bulla in some fossil mammals. Bulletin of the American Museum of Natural History.

62: 1-352.

Van Kampen, P. N.. 1905. — Die Tympanalgegend des Saugetierschadels. Morphologisches Jahrbuch, 34: 321-722.

Van Valen, L., 1988. — Paleocene dinosaurs or Cretaceous ungulates in South America. Evolutionary Monographs, 10: 1 -79.

Watson, D. M. S., 1916. — The monotreme skull: a contribution to mammalian morphogenesis. Philosophical Transactions of

the Royal Society of London. B207: 311 -374.

Wible ,J. R., 1986. — Transformations in the extracranial course of the internal carotid artery in mammalian phylogeny. Journal

of Vertebrate Paleontology. 6: 313-325.

Wible. J. R., 1990. — Petrosals of late Cretaceous marsupials from North America, and a cladistic analysis of the petrosal in thcrian

mammals. Journal of Vertebrate Paleontology. 10(2): 183-205.

Wible, J. R.. & Hopson, j. a., 1993. — Basicranial evidence for early mammal phylogeny, pp. 45- 62, in Mammal Phylogeny.

Mesozoic differencial ion, Multituberculates, Monotremes. early Therians and Marsupials. F. S. Szalay. M. J. Novacek

and M. C. McKenna (eds). Springer-Verlag, New York.

Wible. J. R., & Hopson, J. A., 1994. — Homologies of the prootic canal in mammals and non-mammalian cynodonts. Journal of

Vertebrate Paleontology , 14 (3): 52A, Abstracts (supplement to Number 3).

Zeller, U., 1989a. — Die Entwicklung und morphologic des Schadels von Omithorhynchus anatinus (Mammalia: Prototheria:

Monotremata). Senckenbergischen Naturforschenden Gesellschaft Abhandlungen, 545: 1-188.

Zeller, U.. 1989b. — The braincase of Omithorhynchus, pp. 386-391. In: Trends in vertebrate morphology. H. Splechtna and

II. Hilgers (eds). G. Fisher. Stuttgart.

Source. MNHN. Paris

LARRY G. MARSHALL & CHRISTIAN de MUIZON

§ •?

-£• o

-3“ O

- Ofj

^ a

z; u.

^ ~

O >

C £

^ c

E o

</5

c °

ZJ —

E ^

2 E

~ =

£ -

i c

•—

“3 ^

— -Ci

03 C

C/5 C<-

>> *>

-C ^

—

o i

~ >

03 «U

S so

2- s

V; >

“I 5

2pO

-8 .§

x ’o

^ s-

I "8

1.1

— a

- *C

E- o

: i %

-T

-J

O'*

>/-)

8.88

0.28

3.15

11.30

0.25

*^r

i/S

S .

•Cl

*^r

l/S

o’

C-I

c-i

V»

r**

—

CM*

—

r 1 -

—

-4

■4

.—

U-j

—

-J

—

r-~

4.84

061

0.07

4.21

—

-3-

c-i

•

—

■5

■s

-J

.—

-4

_u:

■ ■

i—

1.89

0.08

—

1.09

0.08

7.34

“

-J

«/S

r»

«/“>

-T

•o,

-J

r~~

1.77

5.65

1.47

60’0

6.12

”

«n

ICY

«n

C**

r i

•O

»o

C l

-3-

*/“>

“

«CY

-4-

0.49

900

12.24

r—

■4

0.51

0.06

11.76

p»

-J

c *

O IX

o i x

Source: MNHN, Paris

Part III: Postcranial skeleton

Larry G. Marshall * & Denise Sigoc.neau-Russell**

* Institute of Human Origins, 1288, 9th Street, Berkeley

California 94710, USA

**Museum national d'Histoire naturelle

Laboratoire de Paleontologie, URA 12 CNRS

8 rue Buffon. F-75005 Paris, France

ABSTRACT

The earliest and most complete articulated skeletons of fossil metatherians yet known arc represented by four specimens of

Pucadelphys andinus from the Santa Lucia Formation (early Paleocene) at Tiupampa in southcentral Bolivia. Two sets of what

arc interpreted to be male-female pairs were found in a three dimensional, life-like, snout-rump position in burrow-nests that were

apparently dug in a bank along a meandering river. The animals probablydie d as the result of a flood which entrapped them in

their burrows and tilled the latter with water and sediment. A detailed comparative study of the postcranial bones reveals that the

vast majority of character states in Pucadelphys are regarded as mammalian, tribosphcnic ormetatherian plesiomorphies (e.g. atlas

not perforated by transverse canal and with a persisting suture between the ossified intercentrum and atlantal arch; absence of

transverse canal on axis, with possible unfused rib; absence of enclosed transverse canal onCV7; robust fibula; presence of ossified

os marsupium; etc.). Character states of uncertain polarity include the presence of only one vertebra articulating with the ilium

(fulcraI vertebra), and a long non-prehensilc tail. The tarsus has a bicontact upper ankle joint (UAJ) as in living Didclphidae;

moreover the calcancum shows a partially plantar orientation of the cuboid facet which can be interpreted as foreshadowing the

specialisation ol later Didelphidae; the situation then is more advanced than in the “plesiomorphic metatherian morphotype” of

Szalay (1982 a. b; 1984); the only characters of the latter persisting in Pucadelphys are the large peroneal process and the

“remarkably broad transverse dimensions from peroneal process to the medial margin of the sustentaculum". Collectively these

characters support the view, based on the study of the skull and dentition (Marshall & Muizon, 1995), that Pucadelphys represents

the plcsiomorphic taxon within the family Didelphidae. Functional considerations of the skeletons suggest that Pucadelphys was

essentially terrestrial, quite agile, and possessed limited bounding and digging capabilities.

Marshall. L. G. & Sigogneau-Russell, D.. 1995.— Part III: Postcranial skeleton. In: Muizon, C. de <ed.). Pucadelphys

andinus (Marsupialia, Mammalia) from the early Paleocene of Bolivia. Mem. Mus. natn. Hist . licit.. 165 : 91-164. Paris ISBN :

2-85653- 223-3.

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

RESUME

Troisieme partie : Ie squelette postcranien

Les squclettes les plus ancicns et les plus complets de metatheriens fossiles connus a ce jour sont representcs par quatre

specimens de Pucadelphys andinus , en provenance de la Formation Santa Lucia (Paleocene inferieur), a Tiupampa. dans le sud

de la Boli vie centrale. Deux ensembles de cc que nous avons interprete comme des couples male-femclle ont Cte conserves en trois

dimensions, en position de vie. et disposes tete-bechc dans ce qui scmble avoir ete la berge d’un meandre fluviatile. Ces animaux

sont apparemment morts a la suite d’une inondation qui les a pieges dans leur terrier en remplissant ce dernier d’eau et de boue.

L'etudc detaillee et comparee dcs os postcraniens a revelc que la grande majorile des caracteres de Pucadelphys peuvent etre

considers comme etant dans un etat plcsiomorphe pour les Mammiferes, les Tribosphenida et/ou les Metatheriens ( e.g . atlas

imperfore et gardant une suture entre l’intercentre ossifie et I’arche atlantale ; axis depourvu de canal transverse, mais possedant

une cote libre persistante : absence de veritable canal transverse sur CV7 ; fibula robuste ; presence d’un os marsupium

ossifie; etc.). L’existence d’une seule vertebre fulcrale (SI) et d'une longue queue non prehensile constituent, eux, des caracteres

de polarite incertaine. L’articulation du tarsc est de type «bicontact» comme chez les Didelphidac actuels ; en outre la facette

cuboi’de du calcanCum presente une orientation en partie plantaire qui peut etre interprets comme prCfigurant la specialisation

des Didelphidae ultericurs ; ce tarse est done plus specialise que celui du «plesiomorphic metatherian morphotype» de Szalay

(1982a.b ; 1984). ne le rappclant que par le grand processus peroneen du calcaneum et la largcur remarquable qui sCpare cc

processus et le bord medial du sustentaculum. Prise dans son ensemble, cette analyse du squelette soutient Popinion, basee sur

l'etudc du crane et de la denture, selon laquellc Pucadelphys represente le taxon plcsiomorphe a Pinterieur des Didelphidac. Les

considerations fonctionnelles resultant de Petude de ces squelettes suggerent que Pucadelphys etait un animal essentiellement

terrestre. plutot vif et capable, dans certaincs limites. de sauter aussi bien que de creuser.

RESUME DEVELOPPE

Les restes de squelettes postcraniens des mammiferes paleogenes sont extremcmenl rares et. qui plus est, difficiles a identifier

en raison de Pabsence dissociation avec les dents, organes sur Iesquels est essentiellement fondee la taxonomic de ces animaux.

Or les squelettes etudies dans cette troisieme partie sont non sculcment associes a dcs cranes munis de leur denture (voir Part.

II de ce volume), mais ils sont tr£s complets (95% des os sont represents) et tres bien conserves. Ils sont enfin les plus anciens

restes squeletliques connus de marsupiaux, cette position etant jusqu'ici tenue par des os tarsiens isoles du Paleocene superieur

d’ltaborai. Bresil.

L’etude detaillee de ces squelettes est done du plus haul interet. Les comparisons ont ete faites, d’une part avec les rares

squelettes connus de mammiferes mesozoiques: I'eothericn Eozostrodon, du Jurassiquc inferieur d’AIriquedu Sud. lethcricn non

tribosphenique Henkelotherium , du Jurassiquc superieur du Portugal, et les placentaires du CretacC inferieur de Mongolie ; avec

d'autre part les petits didclphides generalises actuels : Metachirus, Monodelphis, Mannosa et Didelphis , avec aussi Perameles.

La colonne cervicalc de Pucadelphys andinus est celle d’un petit didelphide actuel, si ce n’est que l’atlas montre la persistance

d'une suture entre i'intercentre et fare neural, et que Taxis est depourvu de canal transverse et garde des cotes axiales, deux

caracteres primitifs. La colonne thoracique est non moins gcneralisee dans son ensemble ; par contre la colonne lombaire est

interpretee commederivCe. en raison de Tallongement progressif du corps vertebral et des apophyses transverses, et de la hauteur

des epines neurales dirigees vers Tavant: une telle morphologie ne se retrouve pas chez les didclphides examines, mais bien chez

la forme fouisseuse Perameles. Dans ce contexte. le sacrum est considere comme specialise, avec deux vertebres donl une

seulement s’appuie sur T i lion. La longueur de la queue reste imprecise, mais les vertebres caudales conserves ne montrent aucune

specialisation prehensile, contrairement a cellcs d’ Henkelotherium ou de Didelphis.

La ceinture scapulaire est identique a celle dcs didelphides actuels les plus primitifs ; Thumerus montre une vaste surface

d’insertion pour les extenseurs, comme celui des petites formes terrestres ; Ic cubitus et le radius sont encore tres robustes. Les

os du carpe et de la main n’ont pas ete conserves.

La bassin, bien que massif, presente, comme le sacrum, des specialisations de type peramelide, avec une grande expansion

dorso-ventrale de Taile antericure de Tilion, un grand foramen obturateur, et des os marsupiaux reduits. Au contraire, les os de

la cuisse et de la jambe sont plus primitifs que ceux des petits didclphides actuels. L’astragalc n’est pas tres bien conserve, mais

il semblc que sa morphologie etait plus plcsiomorphe que celle de Didelphis. Quant au calcaneum, il est depourvu de facette

fibulaire ; il est done plus evoluc que celui defmissant le “morphotype metathcrien” de Szalay (1982a. b) et se rapproche du

“morphotype didelphide” ; il conserve pourtant un fort processus peroneen. interprete dans cc contexte particulier comme un

caractere plCsiomorphe. Un point interessant concerne Torientation de la facette cuboidienne, consideree comme annonQant la

condition des didelphides actuels. Le pied lui-meme parait avoir etc long et relativemcnt rigide.

Source MNHN , Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

Dc P ensemble de ces caracteres, les auteurs deduisent pour Pucadelphys andinus un mode de vie terrestre et non arboricole (une

adaptation considdrce comme primitive pour les Theria Tribosphenida. Krebs. 1991): I’astragale et le calcancum en particulier

nc montrent pas les caracteres relids a unc telle specialisation ; malheureusement la configuration de lajonction tibia-astragale n'a

pu etre precisee. nous ignorons done s’il y avait renversement du pied comme observe par Jenkins & Me Learn (1984) chez

quelques didelphidcs arboricoles.

Dans le detail, la brievetd des apophyses epineuses des vertebres cervicales suggere une bonne mobilite du cou. L’etroitesse

de I’espace separant radius et cubitus imposait une rotation limitee de I’avant-bras. Le ddveloppement des metapophyses dorsales,

la longueur des apophyses epineuses des vertebres lombaires et la largeur dc Pextremitd distale de I’humdrus constituent autant

de potentialites fouisseuses ; de memc, la longueur des epincs lombaires et la conformation du sacrum (possible mobilite de la

jonction sacro-iliaque, grand angle ilio-sacre, largeur et orientation de la surface iliaque destinee aux abducteurs et extenseurs de

la cuisse) devaient favoriser le fouissement, sans qu’une telle specialisation soit. chez Pucadelphys, poussee aussi loin que chez

Perameles. Par ailleurs les proportions relatives des membres superieuret inferieur sont compatibles avec une bonne agilite mais

ne correspondent pas a celles d'un animal coureur; enfin cettc espece sembie avoir ete capable de sauter. mais sans etre reellement

specialisee dans cette direction.

En ce qui concerne le mode de vie. cette forme devait etre nocturne comme la plupart des petits didelphidcs. dormant le jour

dans un nid-terrier ct cherchant sa nourriture durant la nuit. La connaissance des ma-urs des petits didelphides actuels (qui sont

habituellement solitaires) conduit h penser que ces quatre squelcttes associes deux a deux representent ceux de couples formes

durant la periode de reproduction ; les terriers etaient creuses dans la berge d'unc riviere, dont la crue subite a rempli ces nids de

sediments ayant favorise la conservation en position naturelle. Cette conclusion est corroborec par l’abondance des fossi les trouves

dans le gisement (comportant en particulier des squelcttes presque complcts de grenouilles).

Peu de caracteres du squelette permettent de preciser la position phylogenique de Pucadelphys andinus, si ce n'est la

configuration du tarse. Celle-ci, jointe aux conclusions concernant le crane et la denture, font considerer Pucadelphys comme le

taxon le plus pldsiomorphe a I'interieur des Didelphidae.

INTRODUCTION

Postcranial bones of mammals are very important in phylogenetic studies; yet these

elements are extremely rare in rocks of Cretaceous and Paleocene age, a time when the basic

branches of this class became established. Associated skeletons which provide functional and

phylogenetic information are even rarer, and those with associated dentitions and skulls are almost

non-existent. Because of a dearth of such specimens, little is known of the early postcranial

evolution of Tribosphenida (metatherians, eutherians and related forms with tribosphenic

dentitions; sensu McKenna, 1975). In fact, the postcranial character states in the direct ancestor(s)

of metatherians and eutherians are currently inferred from the study of all too few isolated

Cretaceous and Paleocene elements, of Eocene to Pleistocene fossils, and of living taxa.

There are, to date, only two non-tribosphenid therians (both eupantotheres) for which

associated postcranial material is known. One is Henkelotherium guimarotae from the Late

Jurassic of Portugal, described by Krebs (1987; 1991), and the other is Vincelestes neuquenianus

from the Early Cretaceous of Argentina, which is not yet described (see Bonaparte & Rougier,

1987 ; Rougier et al. , 1992 ; Wibble & Hopson, 1993).

Eutherians are the best known of Late Cretaceous tribosphenids, and partial skeletons of

Asioryctes, Zalambdalestes and Barunlestes have been described from the ? Late Santonian and/

or Campanian of Asia (Kielan-Jaworowska, 1977, 1978).

In contrast, postcranial remains of metatherians from the Late Cretaceous and Paleocene are

presently known only from isolated calcanea and astragali (Szalay, 1982a and b; 1984);

associated skeletons have not been reported. The earliest nearly complete skeleton of a metatherian

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Fig. 22. —The “sarigue fossile” from Montmartre, Cuvier collection MNHN 7905 (top) and 7904 (bottom): opposite halves of

same specimen, type of Peratherium cuvieri Fischer, 1829. X 1.

Pig. 22. — La sarigue fossile de Montmartre, collection Cuvier MNHN 7905 (haul) et 7904 (has): moities opposees du meme

specimen, type de Peratherium cuvieri Fischer, 1829. X 1.

Source: MNHN. Paris

PUCADELPHYS AND IN US: POSTCRANIAL SKELETON

SARIGUE fossile .

Fig. 23. — The “sarigue fossile” from Montmartre (after Cuvier 1804. PI. 19). type of Percitheriuni cuvieri (Fischer. 1829).

Fig. 23. — La sariguefossile de Montmartre (d'apres Cuvier 1804. Pi 19). type de Peratherium cuvieri ( Fischer. 1829).

Source MNHN. Paris

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

was long represented by the classic “sarigue fossile” collected from the Butte Montmartre in

northcentral Paris (Figs 22, 23). The fossil is from the “Gypse de Montmartre”, assigned to the

Late Eocene (i.e. Headonian Land Mammal Age, Savage et al., 1994). It was first described by

Cuvier (1804), named Didelphis cuvieri by Fischer (1829), and is now classified as Peratherium

cuvieri (Crochet, 1980).

Hence the major interest of the four nearly complete skeletons, two of which have articulated

skulls, of the metatherian Pucadelphys andinus, from the Early Paleocene Santa Lucia Formation

at Tiupampa in southcentral Bolivia. These are currently the earliest and most complete articulated

skeletons of metatherians yet known, and they provide the first opportunity to securely assess

aspects of the postcranial structure of a member of this group at the “Beginning of the Age of

Mammals”.

Unless otherwise specified, all numbers cited below, in the figure captions and in the

Appendix (i.e. 6105, 6106, 6110, 6111) pertain to YPFB.

SYSTEMATIC PALEONTOLOGY

The skeleton-pairs described in this study were collected with articulated skulls, one of

which (6105) was designated the type of Pucadelphys andinus Marshall & Muizon, 1988. A

detailed study of the dentitions and skulls of 6105 and 6110 (as we! 1 as numerous other specimens),

shows that both skeleton-pairs are referable to this species (Marshall & Muizon, 1995). The

systematic position of the skeletons is thus as follows:

Legion TRIBOSPHENIDA McKenna, 1975

Infraclass METATHERIA Huxley, 1880

Order DIDELPHIMORPHIA (Gill, 1872) Marshall et al., 1989

Family DIDELPHIDAE Gray, 1821

Genus PUCADELPHYS Marshall & Muizon, 1988

Pucadelphys andinus Marshall & Muizon, 1988

Diagnosis (postcranial skeleton morphology only). — Atlas not perforated by transverse

canal, and with a persisting suture between ossified intercentrum and atlantal arch; absence of

transverse canal on axis, with possible unfused rib; absence of enclosed transverse canal on CV7;

lumbar series specialized compared to that of other didelphids (gradual lengthening of vertebral

body and transverse processes, long anteriorly directed neural spines); single fulcral vertebra (SI);

specialized pelvis (ilium dorsoventrally expanded anteriorly, large obturator foramen, small

ossified os marsupium); possible movable sacro-iliac joint; long (±30 caudals) non-prehensile

tail; digging specializations of the humerus (no third distal articular surface, large areas for

extensors of forearm and carpus); robust fibula; calcaneum with bicontact upper ankle joint (U AJ)

(Szalay, 1982a, b), large peroneal process and remarkably broad transverse dimensions from

peroneal process to medial margin of sustentaculum.

Source; MNHN, Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

DESCRIPTION

Measurements (\n mm) of the individual bones of the specimens of Pucadelphys described

below are given in Tables 1 to 20 in the Appendix.

General Features. — Skeleton-pairs YPFB Pal 6105 and 6106 (Figs 2, 24A and 25A). The

animals are in a snout-rump position; 6105 faces to the right and 6106 to the left (referring to the

positions as in the photos). Part of the dorsal surface and the entire ventral surface of both

individuals are presently visible, and the latter was presumably on the floor of the burrow when

the animals died and were fossilized. Both individuals are adults as evidenced by the facts that the

skull of 6105 has a slightly worn adult dentition and the epiphyses, although still distinct, are all

firmly attached to the diaphyses. 6106 is slightly larger than 6105 (see Appendix).

On 6105, the seven cervical and eleven thoracic vertebrae are in a nearly straight line. The

left forelimb is extended posteriorly and parallel along the body. The left hindlimb (as seen on the

opposite surface of the block) has the femur on top of the thoracic region of 6106, the tibia and

fibula extend posteriorly over the thoracic vertebrae to T11, and the pes is on the right side of that

specimen. These features indicate that the pelvic region of 6105 was lying upon the upper thoracic,

neck and possibly head region of 6106.

On 6106. the body is in an arched position as shown by the arrangement of the articulated

T1 to C9 vertebrae. The right forelimb lies along side and nearly parallel to the body and, as shown

by the humerus, was extended posterolateral ly. The proximal part of the left forelimb is extended

anterolaterally under the posterior lumbar-pelvic region of 6105, with the ulna and radius flexed

sharply anteriorly along the posterolateral side of 6105. The right hindlimb is flexed anteriorly

under the abdominal area. The left hindlimb is extended laterally, with the tibia and fibula under

the thoracic region of 6105.

Skeleton-pairs YPFB Pal 6110 and 6111 (Figs 3,4 and 26). The animals are also in a snout-rump

position; 6110 faces to the right and 6111 to the left (referring to the positions as in the photos). The

dorsal surfaces of both individuals are visible. Both individuals are subadults as demonstrated by the

facts that the skull of 6110 has an unworn adult dentition and the epiphyses are unfused and often

separated from the diaphyses in both individuals. 6110 is slightly larger than 6111 (see Appendix) and

both individuals are notably smaller (average about 20%) than 6105 and 6106.

On 6110, the pelvic area is nearly horizontal. The tail bends sharply dorsolaterally to the

left and the end of the tail lies upon, and parallels, the thoracic vertebrae of 6111. In the posterior

thoracic region the body begins to twist to the left, with the cervicals and head completely on their

left side. The proximal end of the right forelimb is extended posterolaterally and the ulna is flexed

sharply anteriorly. The left forelimb extends parallel along the posterior side of the proximal part

of the right forelimb. The relationship of the forelimbs and cervical vertebrae clearly shows that

the animal was lying on its left shoulder. The right hindlimb was extended nearly perpendicular

to the body, while the left hindlimb was in a tightly anteriorly flexed position directly under the

body.

Source; MNHN, Paris

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Source: MNHN, Paris

I'UCADELPHYS AN DIN US: POSTCRANIAL SKELETON

6111 was lying on its right shoulder with the right forelimb extending nearly perpendicular

on the left side of the body (as seen by the position of the distal end of the humerus and proximal

ends of the ulna and radius). The left forelimb was flexed tightly against the thoracic region of the

right side of the body and the “elbow" abuts vertebrae T8 and T9. The proximal end of the right

hindlimb extends anteriorly under the body such that the femur parallels the lumbar vertebrae and

the distal end (tibia and fibula) extends nearly perpendicular on the left side of the body. The left

hindlimb is extended anterolaterally on the left side of the body, and the distal parts of the tibia

and fibula (and pes) lie upon the lumbar region of 6110.

Collectively, the four specimens include about 95% of the complete skeleton. The only

missing elements are the manus (see p. 127), the first metatarsal, some tarsals and phalanges of

the pes, and an estimated 17 posterior caudal vertebrae.

Axial Skeleton.

6105 7 cervicals (left half of atlas, axis, CV3-CV7)

9 thoracics (T1 -T9, fragments of T10-T11)

6106 13 thoracics (T1 -T13)

6 lumbars (L1-L6)

2 sacrals (SI-S2)

9 caudals (C1-C9)

6110 7 cervicals (atlas, axis, CV3-CV7)

8 thoracics (T6-T13)

6 lumbars (L1-L6)

2 sacrals (S1-S2)

9 caudals (C1-C5, C16?-C 17?, C207-C21 ?)

6111 7 thoracics (T7-T13)

6 lumbars (LI-L6)

2 sacrals (S1-S2)

9 caudals (C1-C9)

Cervical Vertebrae

Allas (Figs 25B, 27 and 28A, B; Table 1).— Two elements: 6105, left half (attached to

axis), ventral and lateral views; 6110, isolated left and right halves, nearly complete.

Fig. 24. — Pucadelphys cmdinus. Stereophotos. A, specimen-pair 6105-6106, partial view. X 1; B, 6106, right scapula, anterior

view; thoracic vertebrae and ribs, ventral view. X 3.

Fig. 24. Pucadelphys andinus. Stereophotos. A. coupledesspecimens 6105-6106, vueparlielle. X /; B. 6106 , omoplatedroite.

vue anterieure; vertebres thorciciques et cotes, vue ventmle. X 3.

100

LARRY G. MARSHALL & DENISE SIGOC.NEAU-RUSSELL

Source: MNHN, Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

101

The dorsal arch (da) is narrow ventrally and considerably broader dorsomedially. Laterally,

the dorsal arch joins the base of the ventral arch and projects as a flat, horizontal, ovoid transverse

process (6105), constricted at the base. In 6110, only the root of the transverse process is preserved

on both halves (b tp); but the lateral-most surface does not appear to be broken, suggesting that

the transverse process had a different center of ossification and fused with the root only in adults.

Anterior to the transverse process, the articular facet for the occipital condyle (If of) is wide and

distinctly concave; posterior to it, the articular facet for the atlas body (i.e. anterior part of the axis

sensu Into ) is smaller and only slightly concave (ax f). The two articular facets (for occipital and

atlas body) are not situated directly opposite each other anteroposteriorly: the occipital facet

slopes posteromedial ly at an angle of about 40°, while the atlas body facet is more transverse and

faces mostly posterolaterally at an angle of about 33°. As for the ventral component of the atlas

(intercentrum), what must be its lateral part appears to be suturally joined to the left atlas of 6105

and directed towards the ventral surface of the dens. Supporting this interpretation is the presence

of a facet on the posteroventromedial edge of the atlas arch of 6110. immediately adjacent to the

articular facet for the atlas body of the axis (Fig. 27C, si); we thus assume that the intercentrum

was ossified in the adult. The groove for the vertebral artery at the anterior base of the dorsal arch

is deep in the adult 6105 and shallower in the subadult 6110 (t s). Behind the transverse process,

a deeper sulcus (for “neuro-vascular structures” according to Jenkins & Parrington, 1976)

separates the latter from the axial facet. A very small canal is clearly visible on the right side of

6110 (t c?). the posterior opening of which, just behind the root of the transverse process, being

larger than the anterior one which is in front of the same process; whether this is homologous to

a transverse canal remains uncertain (see discussion p. 111). The atlantal canal for cranial nerve

I, situated on the medial side of the occipital facet, is not closed.

Axis (Figs 25,28C-E,29Aand38A;Table 1). — Twoelements:6105,ventral,medial and

left lateral views; 6110, all surfaces visible.

The suture between the atlantal and axial component is distinct in 6110 on the dorsal and

ventral surface of the body, while in 6105 it only appears as an elevated transverse ridge on the

ventral surface, but is still distinct on the dorsal surface. The atlas component of the axis body is

wider than the axial part. Both components have two (paired) distinct nutrient foramina on the

dorsal face of the body. The dens is oriented anterodorsally, is slightly flattened dorsoventrally,

and transversely convex bothdorsally and ventrally. A distinct medial keel extends longitudinally

along the ventral surface of the body and posteriorly forms a broad lip. Lateral to the medial keel

are two broad depressions which are bordered by the ventral root of the transverse processes (Fig.

28, tp v). The neural arch (6105) is very long dorsally relative to its ventral base; hence the anterior

Pig. 25. — Puccidelphys andinus. Stercophotos. A, specimen-pair 6105-6106. partial view. X I ; B. 6105, cervical vertebrae,

interclavicle and left clavicle, ventral view. X 3.

Fig. 25. — Pucadelphys andinus. Stereophotos. A. couple des specimens 6105-6106, vue poriielle. X I; II, 6/05, vertebres

cervicales, interdavicule et clavicule, gauche, vue vent rale . X 3.

102

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Source: MNHN . Pans

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

103

Fig. 27. — Pucadelphys andinus. 6110. atlas: A. right half, lateral view: B, left half, dorsolateral view: C, left half, mesial view.

X 13.5. Abbreviations: b tp. base of transverse process; d a. dorsal arch; If of. lateral face of occipital facet; si. surface

for intercentrum: t s. transverse sulcus.

Fig. 27. — Pucadelphys andinus. 6110. atlas: A, moitiedroite, vuelaterale: B. moitie gauche, vuedorso-laterale; C. moitie gauche,

vue mesiale. X 13.5. Abreviations: h tp, base da processus transverse; d a, arc dorsal; If of, face late rale de la facet te

occipitale; si, surface pour Pintercentre; t s, sillort transverse.

Fig. 26. — Pucadelphys andinus. Stereophotos. A. specimen-pair 6110-6111, partial view. X 1; B. the same, partial view. X 1;

C,6110, posterior vcrtebraeandhindlimb: this block appeared to represent the posterior part of 611 Oonly after preparation;

but the ventral surface of this block had been prepared, while on the main block the preparation was of the opposite surface;

in order to link the two blocks, it is necessary to reverse C and also turn it 90° (see Figs 3 and 4). X I.

Fig. 26. — Pucadelphys andinus. Stereophotos. A, couple des specimens 6110-6111, vuepartielle. X 1; B, le meme, vue partielle.

X1; C, 6110, vertebresposterieures et membreposterieur: ce bloc ne s 'est revele representer la panic posterieure de6l 10

qu 'apres preparation; mais c ’est la surface vent rale de ce bloc qui avail ete preparee. alors que sur le bloc principal la

preparation portait sur la face opposee; a fin de relie r les deux blocs, it a done ete necessaire de renverser C et de le tourner

de 90° (voir figs 3 et 4). X 1.

104

LARRY Ci. MARSHALL & DENISE SIGOGNEAU-RUSSELL

border is deeply concave and the atlanto-axial intervertebral space is large. The posterior border

of the neural arch is straight. The spine itself is not intact but was low, long and thin; anteriorly

it overhangs the vertebral body (dens not included); posteriorly it only slightly overhangs CV3.

The anteriorly facing atlantal facets (at 0 are oval, short and directed almost vertically. The

postzygapophysial facets face lateroventrally. The transverse process had two roots which, as seen

on 6105. did not unite laterally; as a result the transverse “canal” between the two roots remains

a wide sulcus (t s). The lateral extremity of the dorsal root on both sides of 6105 is uniformly

concave and does not appear to have been broken; this may have been the attachment surface for

a small cervical rib. a remnant of which may be represented by the bone fragment located along

the anterior surface of the dorsal root of CV3 of 6105.

Cervicals 3 to 7 (Figs 25, 28D-E and 29A; Table 2). — Two series: 6105, CV3-7; 6110,

CV3-7 (fragmentary).

The epiphyses on all elements remain distinct. C V3 is notably longer than CV4 and both are

slightly larger than CV5-CV7, which are also subequal in length. The intervertebral spaces narrow

from CV3 to C V7. The zygapophysial facets are only slightly inclined lateroventrally on CV3, and

become nearly horizontal on CV4 (not preserved on CV5 to CV7). The vertebral bodies are’low

and wide, and the spine was apparently short (only the base is preserved on CV3-CV5). The two

roots ol the transverse process of CV3 to CV6 unite laterally, enclosing the transverse canal (t c);

together they produce a thick posterior process, while the ventral root has also an anteroventral

projection or tubercle (t). There are no free cervical ribs.

( V3 has a distinct median ventral keel which broadens posteriorly and ends in a wide lip;

a nutrient foramen pierces the base of the keel. The lateral depressions arc well developed, but

shallower than those on the axis. The anterior projection of the ventral root of the transverse

process is reduced to a tubercle (Fig. 28E, t). The spinous process is nearly vertical, very short,

and the anterior border of the arch is closely appressed to the posterior border of the axial arch,

suggesting that there may have been some contact between the two arches.

Fic;, 2S. - Pucadelphys cmdinus. A-B. alias: A. 6110. left half, dorsolateral face: B, 6105. left half, ventral face C 6110 axis left

lateral face: D-E. 6105, cervical vertebrae CV2-CV7: D. right lateral face: E, ventral face. All drawn as preserved X 8

Abbreviations: at I. atlantal facet: ax f. axial facet: b n sp. base of neural spine: b t P , base of transverse process: d. dens;

d a. dorsal arch: il, inferior lamella of CV6: If ax f. lateral face of axial facet; If or, lateral face of occipital facet; n a. neural

arch: n s. neural spine; poz, post/.ygapophysis; prz. prezygapophysis, s. suture between atlantal and axial component- 1

tubercle, sec text: tc. transverse canal: t p. transverse process, tp-d. dorsal root of transverse process; tp-v. ventral root

ol transverse process; ts. transverse sulcus; v f, vascular foramen; v k. ventral keel.

I ia 28 ' ~ ^cadelphys andinus. A-B, atlas: A, 6110, moitie gauche, face dorso-laterale; B, 6105, moitiegauche face ventrale-

C, 6110, axis, face laterale gauche: D-E. 6105. vertebrescervicales CV2-CW7: D.face hue,ale droite: E face ventrale

Dessm en I eta, de conservation. X 8. Abreviations: atffacette a,Ian,ale: ax f. facette axiale: h n sp base de repine

neurale: h tp. base du processus transverse: d. dens: d a. arc dorsal: il, lamelle inferieure de CV6: Ifaxf. face laterale

de la facette ax,ale: If of, face laterale de la facette occipitale: n a, arc neural: n s, epine neurale: poz, pos'tzygapophyse:

prz, prezygapophyse: s, suture entre les composants atlantal cl axial: t. lubercule, voir texte; tc. canal transverse: t p.

processus transverse: tp-d. racine dorsale du processus transverse: tp-v, racine ventrale du processus transverse: ts, sillon

transverse; vj. foramen vasculaire; i* k, carene ventrale.

Source

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

105

poz

at f

tp-v

CV7 CV6

' \

CV5 CV4

\ t

~tp-v

CV3 TP-d axis(CV2)

Source MNHN, Paris

106

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

CV4 has a broad, low, median ventral thickening that widens considerably posteriorly (it

is nearly as wide as the posterior epiphysis); the lateral depressions are thus reduced and very

shallow. The anterior tubercle of the ventral root of the transverse process is quite prominent. The

neural arch is similar to that of CV3 and had probably some contact with that of CV2. The spine

itself was very short.

CV5 is flat ventrally. The anterior tubercle of the ventral root of the transverse process is

broken, but it was clearly longer than that of CV4; the dorsal root is covered by the inferior lamella

(lamina ventralis) of CV6 (Fig. 28 D-E, i 1). The left part of the dorsal arch is preserved (6105);

its anterior and posterior borders are straight.

Between the two inferior lamellae of CV6 (the anterior part of which corresponds to the

elongation of the above mentioned tubercle and the posterior part to the ventral component of the

posterior process), the vertebral body is slightly concave; the expansion of the dorsal root of the

transverse process is here separated from its ventral counterpart and is oriented more laterally than

posteriorly.

On CV7, the ventral surface of the centrum has a low longitudinal median ridge and very

shallow lateral depressions. The dorsal root of the transverse process is now completely laterally

oriented, the ventral root is reduced to a small spine, and the two roots arc not fused (i.e. the

“transverse canal” is open laterally).

Thoracic vertebrae (Figs 24B, 29B, 30, 32A and 47A, B; Table 3). — Present in all four

specimens: 6105, T1 to T9, ventral view (although T5 to T9 are represented only by ventral part

of centra); 6106. T1 to T13, ventral and both lateral views; 6110, dorsal views of ventral half of

T6 to T9, dorsal and both lateral views of T10 to T13; 6111, T12 and T13, dorsal and partial lateral

views.

The presence of a small transverse process on the 14th dorsal (i.e. LI) demonstrates that

there are 13 thoracic vertebrae.

As seen best in 6106, the vertebral body increases slightly in length from T1 to T13. T1

to T4 have a longitudinal median keel which is broad and low on Tl, higher and narrower on T2

to T4. The vertebral body is ventrally angular on T3 to T5, convex on T6 and T7, flat on T8 and

T9, and with a longitudinal median trough bordered by low broad lateral ridges on T10 to T13

(a feature progressively developed from T10 to T13).

The neural spine is broken on all exposed vertebrae, except Tl.Tl 1 and T12 of 6106, and

T9 and T10 of 6111 where it is low, narrow (it thickens at the top) and inclined slightly

posterodorsally; it arose on the posterior half of the arch and was transversely thin at the root. The

articular surfaces of the dorsal zygapophyses become more vertical posteriorly and are widely

separated.

On Tl and T2, the transverse processes which bear the parapophyses for the ribs are

ventrally situated opposite the anterior end of the body; they are flat and project quite far laterally,

flaring slightly after a constriction at the root. Beginning on T3 they arise more medially on the

body, and are also more dorsally situated due to the ventral angulation of the body; they become

shorter from T5 to T7, and on T8 (and those more posteriorly), they merely constitute a less and

less prominent facet on the lateral side of the body, to become undistinguishable on Tl 1-T12.

Source . MNHN, Paris

PUCADELPHYSANDINUS: POSTCRANIAL SKELETON

107

Fig. 29. Pucadelphysandinus. Stereophoios. A. 6105, cervical vertebrae, dorsal view. X 3; B. 6106. vertebraeT10-L4. ventral

view. X 3.

Pic. 29. — Pucadelphys andinus. Siereoplmios. A. 6105. verlebres ceivicales, vue dorsale. X3; B. 6106. veriebres T10-L4. vue

ventrale. X 3.

108

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

T8 is a transitional vertebra with the appearance of the first anapophysis (ap) and the first

metapophysis (mp). These apophyses become progressively longer from T9 to T13; they interlock

and produce a secondary articulation at least from T10 to T13. The costal facets are not clearly

discernible on the vertebral bodies.

Ribs (Fig. 24). — There are 13 pairs of ribs. The first nine pairs have tuberculae and

capitulae, while the last four pairs lack tuberculae. The distal ends of all ribs are broken, so it is

not possible to determine their exact length. The parts that are preserved consistently show a lack

of curvature along the ventral half, indicating the presence of a relatively deep chest cavity. They

are narrow, proximally compressed anteroposteriorly, with a faint longitudinal sulcus.

Lumbar vertebrae (Figs 29B. 30, 31.32 and 41; Table 4). — Complete series (LI to L6) in

three specimens: 6106, ventral, lateral and partial dorsal views; 6110. dorsal, ventral and lateral

views (L2 and L3 are damaged); 6111. dorsal, lateral and partial ventral views.

The vertebral body of the lumbars (especially of LI) is very much like that of the last

thoracics, only deeper; its length increases from LI to L4. There are usually two nutrient foramina

(Fig. 32B. n f) on the ventral body surface of all lumbars. but they are more irregularly situated

than those on the thoracics. The ventral surface of LI has a shallow median longitudinal

depression bordered by low broad lateral ridges, almost identical to the condition of T13; on L2

the two ridges are closer and on L3 they meet in a median longitudinal crest; L4 to L6 are gently

convex transversely. The anterior and posterior rims curve slightly ventrally, producing an

incipient “saddle-shape”, and the body is completely platycoelous, at least posteriorly.

The neural spine, as seen on 6106, is very narrow anteroposteriorly (Fig. 31); it is dorsally

directed on L1 and L2, and sharply inclined anteriorly on L3 to L6; it increases markedly in height

from LI to L4 (L4 to L6 are subequal). Due to bifurcation of the posterior base of the neural spine,

the latter is more anteriorly situated on L4 to L6 and a basal sulcus develops along its posterior

border.

The zygapophyses are more robust and clearly more inclined than on the thoracics, with

articular facets almost vertical. The meta- and anapophyses are longer than on the thoracics. as are

the “trous de conjugaison”. These metapophyses arc tightly interlocked with the zygapophyses

(Fig. 30C).

The narrow and thin transverse processes (tv p) increase sharply in size from LI (where they

are only incipiently developed) to L6; they are inserted anteriorly on the body, all anteriorly

directed, and project more and more ventrally from L2 to L6 where they make a sharp angle with

the body.

Sacral vertebrae (Figs 31, 32 and 41; Table 5). — Three specimens: 6106, complete in

ventral view (though slightly masked by underlying bones) and partly visible in dorsal view; 6110,

ventral view of SI and part of S2: 6111. dorsal view, but dorsal wall of vertebral body is broken

on posterior half of SI and all of S2 (thus is visible the dorsal half of the ventral vertebral body).

The sacrum consists of two vertebrae which together form a short and wide triangle

(“apex" points posteriorly), but only S1 contacts the ilium (fulcral vertebra). The vertebral body

Source:

PUCADELPHYS AND IN US: POSTCRANIAL SKELETON

109

of S2 is longer than that of SI, as well as that of Cl. The dorsal intervertebral foramen, elongate

between L6 and S1, is very small between S1 and S2, and S2 and C1. The suture between S1 and

S2 is distinct, and so is the ossification fissure between body and epiphyses.

Fig. 30. — Schematic representation of thoracic and anterior lumbar vertebrae, in ventrolateral view (anterior is to left): A,

Didelphis (MNHN A 3293 III 545).T6toL3; B. Metachirus nudicaudatus (MNHN 1988-68). T6 to L2; C. Pucadelphys

and in us (6106). T7 to L3. X 5/8. Abbreviations: ap. anapophysis; nip. metapophysis; po. postzygapophysis; r. rib

attachment; tvp, transverse process. The dots indicate the direction of the neural spines.

Fig. 30. — Representation schematique des vertebres thoraciques et lombaires anterieures. en vue ventro-laterale (Pavant est

a gauche): A. Didelphis (MNHN A 3293 III 545), T6ciL3: B, Metachirus nudicaudatus (MNHN 1988-68), T6d L2: C,

Pucadelphys andinus (6106), T7a L3. X 5/8. Abreviatipns: ap, anapophyse: nip. metapophyse; po. postzygapophyse: r,

point d'attache des cotes: tvp, processus transverse. Les pointilles indiquent la direction des epines neurales.

110

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

The ventral surface of the S1 vertebral body is almost flat; the sacral ribs (Fig.32, s r) flare

anterolaterally, and the ends bend sharply ventrally with the lowest point at about mid-length. S2

is gently transversely convex. The transverse process of S1 flares anterolaterally and fuses with

a posterior extension of the S2 transverse process, thus closing laterally the sacral foramen (s f).

Fig. 31. — Pucadetphys andinus. Stereophotos. A, 6106. vertebrae L6-SI-S2. C1-C2 and left side of pelvis, dorsal view. X 3; B.

6106, vertebrae C4-C7. dorsal view. X 3.

Pig. 31. — Pucadelphys andinus. Siereopliotos. A. 6106. vertebres L6-S1-S2. C1-C2 el cole gauche eh, bassin, vue dorsale. X

3; Ii. 6106, vertebres C4-C7, vue dorsale. X 3.

Source: MNHN, Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

111

The latter is subcircular, surrounded anteriorly and laterally by SI, and posteriorly by S2. Two

nutrient foramina occur on the ventral body surfaces of S1 and S2. The ilio-sacral angle is

estimated to have been around 15°.

Dorsally on SI. the prezygapophyses are expanded as two widely separated wings; on S2

they are smaller, slightly closer to each other and, above all, not so ventrally flared. Still dorsally

and behind the prezygapophyses, the base of the SI transverse process is marked by a thick ridge,

almost a step, oblique anterolaterally; lateral to it flares the thin sacral rib.

The base of the neural spine is preserved on SI and S2; on SI it extends along nearly the

entire length of the arch, with the main part located at the posterior end; the spine was probably

of moderate height and directed dorsally. On S2 the spine is very reduced and situated at the

extreme posterior end only. On both SI and S2 it is very thin transversely.

Caudal vertebrae (Figs 31,33 A and 34; Table 6a and b). — Represented in three specimens:

6106, Cl toC9, ventral view; 6110, Cl toC5, ventral view, C4 to C5, lateral view. Cl6?, ventral

view, Cl7?, lateral, dorsal and ventral views, C20? and C21?, lateral, dorsal and ventral views;

61 1 1. C1 to C9, dorsal view (C1 is fragmentary).

C1 to C5 are anterior caudals (/.<?. those with zygapophyses), C6 is transitional, and C7 is

the first posterior caudal. There were no haemal apophyses.

The lengths of the vertebral body on C1 to C4 are subequal in 6106 (and are all shorter than

S2); beginning with C5 there is a notable elongation. Ventrally the vertebral body of C1 to C3 has

a broad median longitudinal sulcus bounded laterally by low broad ridges; C4 and C5 have a low

broad median ridge which flares posteriorly (on C5 it is more rounded); the longitudinal median

lidge becomes narrower on C6 and C7, and more elevated on C8 and C9. The anterior and posterior

epiphyses remain distinct ventrally and dorsally.

The neural spine of C1 (6106) is very short and narrow, located at the posterior end of the

arch, and directed posterodorsally; on C2 none is detectable; C3 is not visible; there is apparently

no spine on C4 and C5; from C6 to C9 the fused zygapophyses form a very low but long keel. Pre-

and postzygapophyses are functional on C1 to C5, and prezygapophyses only on C6; they remain

oblique; C7 has only tubercles that do not contact the preceding vertebra; ana- and metapophyses

persist and interlock until C5 or C6.

’I he transverse processes (Figs. 31,33 t p) on C1 to C3 are broad, situated at mid-length of

the body, and are nearly perpendicular to the latter; on C4 and C5 they arise posteriorly on thebody

and flare posterolaterally; on C6 the body is narrow anteriorly and the transverse processes are

broad and “wing-like” along the posterior part of the body; on C7 these posterior transverse

processes are reduced, and there appear narrower, posterolaterally directed, anterior transverse

processes; on C8 and C9, these anterior components have the lateral edges bent ventrally and they

are slightly broader than the posterior components, which are reduced to lateral expansions.

The vertebral body of Cl6? to C21? consists of an elongate rod with greatly reduced

anterior and posterior transverse processes; it decreases in width (but not length) from C16 to C21.

Discussion

Comparisons with Eozostrodon-Megazostrodon, Henkelotherium, Asioryctes and

Barunlestes are based, respectively, on the works of Jenkins & Parrington (1976), Krebs (1991),

and Kielan-Jaworowska (1977, 1978). These taxa were chosen as out-group comparisons.

112

LARRY G. MARS! I ALL & DENISE SIGOGNEAU-RUSSELL

Among living didelphids, comparison was made principally with Metachirus nudicaudatus

(subadult, last molar not fully erupted, Laboratoire de Zoologie, Mammiferes et Oiseaux, MNHN,

specimen no 1988-68), Monodelphis domestica. Philander opossum, Didelphis virginiana (old

adult) and Perameles nasuta (this taxon is used because it shares numerous specializations of the

pelvic region with Pucadelphys as will be demonstrated below) (specimens from the collections

in (he Laboratoire d'Anatomie Comparee. MNHN, Paris, nos. 1967-330, A 3307, 1900-182 and

Fig. 32. — Pucadelphys and in us. 6106. thoracic, lumbar and sacral vertebrae: A. TI3 and LI-3 (ventral face); B, L5-6, and Sl-

2 (ventral face). X 5. All drawn as preserved. Abbreviations: ap. anapophysis; c, centrum; e. epiphysis; il s. iliac surface;

n f, nutrient foramen; s f. sacral foramen; s r, sacral rib; t p. transverse process.

hie. 32. — Pucadelphys andinus. 6106, vertebres thoraciques, lombaires et sacrees: A, TI3 et LI-3 (face vent rale); B, L5-6, et

SI-2 (face vent rale). X 5. Dessin en Petal de conservation. Abreviations: ap, anapophyse; c, centrum; e, epiphyse; il s.

surface iliaque; n f foramen nourricier; s f foramen sac re; s r, cote sacree; t p, processus transverse.

Source MNHN , Paris

PUCADELPHYS ANDINUS : POSTCRANIAL SKELETON

1 13

1880-1020, respectively); special emphasis was laid on Metachirus because this taxon is one of,

if not the, most primitive extant member of this group (Reig et al„ 1987). For these comparisons,

the works by Coues (1872), Slipjer (1946) and Tate (1933) were also used.

The first point to be noted is the vertebral formula: Pucadelphys has 7 cervicals. 13 thoracics,

6 lumbars, 2 sacrals and ± 30 caudals, which is similar to that of many living didelphids (e.g.

Didelphis, Metachirus, Philander).

Atlas. — No atlas is preserved in Henkelotherium. Sutural linkage of the ossified

intercentrum with the atlantal arch, as suggested in the above description of Pucadelphys. was also

observed in Barunlestes , but not Asioryctesox Eozostrodon. A persisting suture between these two

components makes the atlas of Pucadelphys less derived than in Metachirus , Monodelphis and

Didelphis , where synostosis is complete. The presence of a fully enclosed transverse canal (though

very narrow) is recorded in Barunlestes and considered as a derived state; this canal is not present

in Asioryctes , Didelphis , Monodelphis or Metachirus (which means that the arteria vertebralis

ran in a groove at the base of the transverse process); it thus seems doubtful that the liny canal

mentioned in Pucadelphys 6110 (Fig. 28, t c?) is homologous with the transverse canal, since the

arterial sulcus is clearly visible at the base of the transverse apophysis; besides, a tiny canal similar

to that in Pucadelphys is recorded in Caenolestes by Osgood (1921), who does not consider it as

homologous to the transverse canal. On the other hand, the non closure of the canal for cranial

nerve I as described in Pucadelphys is known only in three living marsupial genera (i.e.

Perameles, Marmosa and Monodelphis). Other similarities with Monodelphis and Metachirus

include the shallowness of the concavity of the axial facet and the shape of the transverse process.

In conclusion, only one character of the atlas of Pucadelphys (persisting suture between

ossified intercentrum and atlantal arch) is more primitive than in the atlas of the living non-

specialized didelphids.

Axis. — No axis is preserved in Henkelotherium. The large intervertebral space (between

atlas and axis) observed in Pucadelphys was also recorded in Eozostrodon and. based on the

regular anterior concavity of the axis, was probably present in Asioryctes and Barunlestes. Among

the examined didelphids, this space is longest in Didelphis, where the axis is very specialized

anteriorly (i.e. the spinous process is very deep with a ventrally protracted anteroventral edge).

The suture visible between the atlantal and axial components of the axis body (less

accentuated in the adult 6105 than in the subadult 6110) is typical of primitive mammals as well

as living adult marsupials: it is discernible in Metachirus (but only faintly visible in Perameles.

Monodelphis and Didelphis) and is also recorded in Zalamhdalestes and Asioryctes. The axial

part is shorter than the atlantal part (dens excluded) in these last two genera; it is longer in

Metachirus. Monodelphis. Perameles and Didelphis , as in Pucadelphys.

The axis spine of Pucadelphys is not as specialized as in Asioryctes or Zalamhdalestes. It

is also not as specialized as in Didelphis or Perameles , but is very similar to that of Metachirus

and Monodelphis-, as already mentioned, the anterior border of the axial arch in Pucadelphys is

indeed more similar in shape and orientation to that of Metachirus than to that of Didelphis , where

1 14

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Source: MNHN . Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

115

it is more sharply inclined ventrally. As for the posterior border, it slightly contacts the arch of CV3

as in Metachirus and Monodelphis, but not Philander.

The two roots of the transverse process, separated in Pucadelphys, are united in the living

marsupials as well as in Zalambdalestes; thus, the absence of a transverse canal in Pucadelphys

represents the plesiomorphic state. Finally, an unfused axial rib may have been present in

Pucadelphys-. among living marsupials a freely articulating cervical rib is only reported in

Perameles (Grasse, 1967: 608), Caenolestes and Phascogale (Osgood, 1921).

In conclusion, the axis shows two characters (absence of transverse canal -only the sulcus

is present-, unfused axial rib) less derived than in generalized living didelphids.

Cervicals 3-7. — The epiphyses of the cervical vertebral bodies are not distinct in

Henkelotherium or Eozostrodon , but remain so in Pucadelphys (and in the examined marsupials

except Didelphis). In contrast, the neural-centrum suture is open in Henkelotherium but not in

Pucadelphys. The shape of the vertebral body of the latter is similar to that of Metachirus. The

neural spines are rod-like in Eozostrodon, low in Henkelotherium and Metachirus, almost non

exislant in Perameles, Monodelphis and Philander (at least until CV5) but high in Didelphis where

they synostose with each other (a specialization of that genus); they were reduced to tubercles in

Asioryctes and zalambdalestids, and apparently low and separated in Pucadelphys.

An enclosed transverse canal is present in CV3-CV6 of Pucadelphys, but absent on CV7 as

in zalambdalestids, Metachirus, some specimens of Caenolestes and several other marsupials

(e.g. Perameles)-, this again represents the plesiomorphic state. There were no cervical ribs other

than axial in Pucadelphys (nor in Henkelotherium, but possibly some in Eozostrodon).

The presence of an inferior lamella on CV6 is typical of all therians known in that respect,

including Asioryctes and Barunlestes ; the situation is not known in Henkelotherium.

In conclusion, the cervical vertebrae in Pucadelphys resemble those of living generalized

didelphids.

In summary, primitive characters of the cervical vertebrae of Pucadelphys (i.e. absence of

a transverse canal on atlas, axis and CV7, possible axial rib, shape of the axis arch) are not

collectively found in any of the living didelphids examined.

Thoracic vertebrae. — In Henkelotherium the number of thoracics and even presacrals is

not known, and no certain thoracic vertebra is preserved.

The number of thoracics in Pucadelphys (13) is the same as in living didelphids and

Perameles. T8-T9 are transitional vertebrae in Pucadelphys (Fig. 30C) and Monodelphis, T7-8 in

Didelphis (Fig. 30A), T9 in Philander and T9-10 in Metachirus (Fig. 30B).

Fig. 33. — Pucadelphys cuidinus. Stercophotos. A. 6106, vertebrae C4 to C9, ventral view. X 3; B. 6105. left partial pes, dorsal

view. X 3.

Fig, 33. — Pucadelphys andinus. Stercophotos. A, 6106. vertebres C4-C9. vue ventrale. X 3; B, 6105, pied gauche partiel, vue

dorsale. X 3.

116

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Fig. 34.

a tp

C 16? __

C17?-

— Pucadelphys andinus. Caudal vertebrae: A, 6106, C4 to 8. ventral face; B, 6110, Cl 77, C18?. C20? and C21 ? ventral

lace. X 5. All drawn as preserved. Abbreviations: at p. anterior transverse process: e. epiphysis; c, centrum; t p transverse

process; p tp, posterior transverse process.

- Pucadelphys andinus. Vertebres caudales: A, 6106. C4-C8.face ventrale; li, 6110. C/7?. CIS?. C20?. el C21? (ace

vent rale. X 5. Dessins en Fetal de conservation. Abreviations; at p. processus transverse anterieur; e epiphvse • c

centrum; t p; processus transverse; t p. processus transverse posterieur. ’ ’

Source MNHN , Paris

PUCADELPHYS AND1NUS: POSTCRAN1AL SKELETON

117

In Asioryctes (only (he first thoracic is preserved) and Barunlestes (the last three preserved),

the transverse processes are short; there is no spine, only a tubercle; the vertebral body is wider

than long; and accessory processes (meta- and anapophyses) are also present.

In Didelphis, Perameles and Metachirus, the thoracic spines, longer in the anterior part of

the series than in Piicadelphys, become short and stouter posteriorly; in Monodelphis and

Philander, they are short and inclined more and more posteriorly, while in Didelphis they contact

each other on the posterior vertebrae, a specialized state relative to that of the other genera. In

Piicadelphys , there is no evidence of an anticlinal spine, no more than in living didelphids, but in

Perameles the spinal orientation reverses at Til. Finally, in Piicadelphys, Metachirus and

Perameles , the epiphyses are distinct, whereas they are hardly so in Didelphis or Monodelphis.

In conclusion, the thoracic series in Piicadelphys , though not completely preserved, is

comparable to that of the living didelphids.

Lumbar vertebrae. — The number of lumbars is not known for certain in Henkelotherium,

but is estimated to be at least six; there were possibly seven in Barunlestes , and there are six in

living didelphids as in Piicadelphys. In Piicadelphys, Metachirus, Monodelphis and Perameles,

as in Barunlestes, the lumbars show a lengthening of the body more posteriorly in the series than

in Henkelotherium, Philander or Didelphis. The epiphyses are distinct in the latter genera, not by

a suture, but by a thickening of the ends; they remain distinct by a suture in Piicadelphys,

Metachirus and Perameles.

In Piicadelphys, the lumbar spinous processes arc narrow, high and dorsally or anteriorly

directed as in Perameles-, in Metachirus, they are long, low and posteriorly oriented in the anterior

lumbars, then they become higher, straight and finally slightly anteriorly directed; this arrange¬

ment is intermediate between that in Piicadelphys on the one hand, and that in Didelphis or

Philander in the other, where they arc all low and posteriorly directed except on L6; the same is

true of Monodelphis, but the spines are even lower. The latter were long and low in Barunlestes,

at least on L3. In Henkelotherium the spinous processes are not preserved.

Accessory processes are recorded on the lumbars of Eozostrodoir, in Henkelotherium the

anapophyses seem to be well developed; in Metachirus and Piicadelphys they are better developed

than on the thoracics. They remain discrete in Monodelphis, where only metapophyses are

developed, but they are stout in Didelphis. Their development seems to be related to a reduction

of lateral flexion.

The short transverse processes of the lumbars of Henkelotherium do not seem to show an

elongation towards the sacrum. They are also short in Barunlestes "as in Didelphis ” (Kielan-

Jaworowska, 1977); those of Metachirus and Piicadelphys form a very similar gradual series,

being anterolaterally directed (as in Perameles), while those of Philander and Monodelphis are

more discrete and those of Didelphis are more robust and form a more rigid ensemble.

In conclusion, the lumbar vertebrae of Piicadelphys form a rather specialized series (within

didelphids at least), with a gradual lengthening of the body and transverse processes, and long

neural spines anteriorly directed; this situation is comparable to that of Perameles.

118

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Sacral vertebrae. — The sacrum is relatively much longer in Barunlestes than in

Pucadelphys , due principally to elongation of S2. But in both genera, it is made of two vertebrae

and there is only one fulcral vertebra. According to Kielan-Jaworowska (1978) this situation is

primitive for Eutherians (the mean number of sacral vertebrae being 3). However, Lessertisseur

& Saban (in Grasse, 1967: 624) state: “le nombre total de vertebres soudees en un sacrum est

variable et ne permet guere d'envisager une ligne claire d’evolution”. Besides, Henkelotherium

has two true sacrals (with two robust sacral ribs), Eozostrodon two to three; in addition, cynodonts

commonly had four to five (Jenkins, 1971), the Early Cretaceous triconodont Gobiconodon

probably had three (Jenkins & Schaff, 1988) and monotremes do have three. This would seem to

contradict Kielan-Jaworowska’s statement. Among didclphids, the sacrum is made of two

vertebrae; but whereas the two are almost equally apposed to the ilium in Monodelphis and

Didelphis, there is only one andahalf in Philander and only onein Metachirus (which incidentally

covers less than 50% of the auricular surface of the ilium as opposed to more than 50% in

Pucadelphys): moreover in Metachirus, S2, which has even a smaller contact with SI than in

Pucadelphys , has short transverse processes, shorter even than those of Cl: clearly, in this taxon

considered as primitive on other grounds, there is practically no differenciation of S2 towards a

sacral status. But Perameles, which shows numerous specializations in that area, has the same

sacral structure as Metachirus and Pucadelphys, with an even smaller auricular surface.

To conclude, if our interpretation of the lumbar vertebrae of Pucadelphys as derived is

correct, it would seem coherent to consider its configuration of the sacrum as also derived, at least

within didelphids.

Caudal vertebrae. — Comparison of the four distal caudals of 6110 with Metachirus

indicates that they apparently correspond to Cl6? + 07?, and C20? + C21?. Assuming that the

subsequent caudals were similar to those of the living didelphids (see below), the number of

caudals for Pucadelphys is estimated to be about 30. The number in some living didelphids is as

follows: Monodelphis, 18?: Didelphis, 26 to 29; Chironectes, 29 (Flower, 1885); Philander, 32;

and Metachirus, 32. Also, Pucadelphys resembles Monodelphis and Didelphis in having 5

anterior caudals, a transitional C6, and the posterior caudals begin with Cl: in Metachirus C5 is

transitional while in Philander it is C4.

In Didelphis, the caudal vertebrae have specializations associated with a prehensile tail

( Krause & Jenkins, 1983: 242): e.g. tail long, commonly twice or more the length of the precaudal

vertebral column; a median sulcus for abductor muscles and tendons crosses ventrally all the

vertebrae, zygapophyses are more vertical, transverse processes are broad and robust for muscle

attachments and present even in most distal caudals; moreover, haemal apophyses, that enclose

abductor tendons and muscles, are large and developed along nearly entire length of tail; finally,

sacral spinous processes are relatively well developed, commonly subequal to the height of the

spinous processes of posterior lumbar vertebrae. No such specializations except high sacral

spinous processes and a very slight ventral sulcus on caudals exist in Philander and Metachirus,

haemal apophyses in Monodelphis-, none of them were found in Pucadelphys. The caudal

vertebrae of Henkelotherium also show prehensile specializations: haemapophyses, ventral

sulcus, and mostly considerable elongation from C6.

Source . MNHN, Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

119

The “evolution” of the caudal transverse processes observed along the tail is different in the

examined didelphids: in Didelphis and Monodelphis the long transverse lamina divides into an

anterior and a posterior component persisting together for a distance; in Metachirus the wide

transverse process also divides, but the posterior component practically disappears on the vertebra

following this division; in Pucadelphys, the situation is more complex than in this genus (see

p. 109). As stated above, the transitional vertebra is C6 in Pucadelphys, Didelphis, Monodelphis

and Caenolestes, and C5 in Metachirus.

Spinous processes are totally absent on the caudals of Caenolestes and Monodelphis , very

faint in Metachirus and Didelphis , high on anterior caudals in Perameles , while rather low in

Pucadelphys.

In conclusion, there is no indication that the tail was prehensile in Pucadelphys.

Appendicular Skeleton.

Shoulder Girdle

Scapula (Figs 24B, 35B, 36 and 38 A; Table 7). — Four elements: 6105, left, external and

partial internal views; 6106, partial right, internal and external views - distal extremity of left,

internal view; 6110, distal fragment of left, internal view; 6111, proximal fragment of left,

external view.

Notable features are the straightness of the posterior border (Fig. 36. pb), the relative

anteroposterior narrowness of the scapular plate (s p), the slight anteroposterior convexity of the

internal face along the posterior border, the anteromedial protrusion of the large coracoid process

(c), and the depth of the supracoracoid incisure (s i). The plate itself is very thin and the

suprascapular border (s b) is not thickened. The spine (s), for which we have only partial views

(the best being on 6105), is very thin, relatively high and probably deflected; it decreases in height

proximal ly. The acromion (a) (as preserved only on 6106, right) is a thin triangular (apex directed

posteriorly) plate-like process, set perpendicular on the anterior part of the spine and parallel to

the scapular blade. Its anterior edge extends slightly beyond, but parallels, that of the glenoid

cavity (g c); its anterior-most projection lies nearly dorsal to the tuber scapulae (t s). The

supraspinous fossa (s 0 seems to have been slightly wider than the infraspinous fossa (i f)

(testifying to a robust supraspinous muscle which itself contributes to a good shoulder mobility

(Jenkins & Weijs, 1979: 408). The glenoid cavity is ovoid in shape, very shallow, and the tuber

scapulae high but not prominent (for insertion of the coraco-humeral ligament). The margin of the

glenoid fossa is only slightly thickened. What may be the scapulocoracoid suture is visible only

in 6106 (left).

Clavicle (Figs 25B, 26B, 35A, 37A and 38: Table 8). — Three elements: 6105, nearly

complete left, anterior view - sternal extremity of right, posterior view; 6110, almost complete

right, posterior view - incomplete left, anterior and dorsal views.

The clavicle has the form of a very elongate S. The scapular extremity (Fig. 37, sc e) is

slightly oblong and was probably thickened, but it is flattened as preserved; the sternal extremity

120

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Source: MNHN, Paris

PUCADELPHYS AND/NUS: POSTCRANIAL SKELETON

121

Fig. 36. — Pucadelphys andinus. Left scapula, reconstructed (6105 and 6106 combined): A. external face; B. internal face; C,

anterior transverse section; D. posterior transverse section. X 5. Abbreviations: a. acromion; a b. anterior border; c p.

coracoid process; cs s, coraco-scapular suture; g c. glenoid cavity; i f. infraspinous fossa; n, neck: p b, posterior border;

s. spine: s b, suprascapular border; s f. supraspinous fossa; s i, suprascapular incisure; s p. scapular plate; t s. tuber scapulae.

F-ig. 36. — Pucadelphys andinus. Omoplate gauche reconstitute (baste siir6105 el 6106): A. face externe; 8, face interne: C.

section transversale anttrieure; D, section transversaleposttrieure. X 5. Abrtviations: a. acromion; a b, hordanterieur:

cp, processus coracoidien; cs s, suture coraco-scapulaire; g c, cavitt gltnoide; if fosse sous-tpineuse; n, col; p b. bord

positrieur; s, tpine; s b. bordsupra-scapulaire; sf fosse sus-tpineuse; s i, incisure suprascapulaire; sp. lame scapulaire;

t s. tuber scapulae .

Fig. 35. Pucadelphys andinus. Slereophotos. A. 6105. left clavicle, scapula, humerus, ulna and radius, ventroposterior view.

X 3; B. 6106. distal extremity of left scapula, humerus, ulna and radius, posterolateral view. X 3.

hie. 35. — Pucadelphys andinus. Stereophotos. A, 6/05, clavicule gauche, omoplate, humerus, cubitus et radius, vue ventro-

posttrieure. X 3; 8. 6/06. ext remite distale de V omoplate gauche, humerus,cubitus et radius, vue posttro-lattrale. X 3.

122

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSEL.L

(st e) is much flatter and expanded, concave anteriorly (for the omosternal cartilage linking it to

the interclavicle) with a raised and thickened posterior border; it is situated at a right angle relative

to the shaft; its posterior face is relatively flat. The anterior face of the shaft shows a spiral-like

transverse convexity, while the posterior face is flatter.

Interclavicle (Figs 25B and 37B: Table 9). — Preserved only in 6105, very slightly

distorted, ventral and partial dorsal views.

The interclavicle is well ossified and relatively long; the anterior process (which supports

anteriorly the clavicle) and the posterior process (which supports the first rib laterally) are only

slightly unequal in breadth, but the anterior part is shorter than the posterior one. Ventrally it bears

a long and high keel which heightens distally, where the bone is as high as it is wide. The short

triangular arms face laterally and terminate with a “costal tubercle” (Fig. 37, c t). A broad

anteroposterior sulcus occurs at the base of the arms between the median ridge (m r) and costal

tubercle. Dorsally the bone seems to have been slightly concave.

Fig. 37. — Pucadelphys andinus. 6105: A. left clavicle and B, interclaviclc, ventral face. X 5. Drawn as preserved. Abbreviations:

a. transverse arm; am b. anteromesial border; c e. cephalic extremity; c s, caudal surface; c t. costal tubercle; g, groove;

m r. median ridge; pi b, posterolateral border; sc e. scapular extremity; st e, sternal extremity.

Fig. 37. — Pucadelphys andinus. 6105: A. clavicule gauche el B, interclavicule, face ventrale. X 5. Dessins en Fetal de

preservation. Abreviations: a: branche transversale; am b, bord antero-mesial; c e: ext remite cephalique; c s. surface

caudale; c t. tubercule costal: g. sillon: m r, cote mediane; pi b, bord postero-lateral: sc e. extremite scapulaire: st e.

extrernite sternale.

Forelimb

Humerus (Figs 35, 38 and 39; Table 10). — Six elements: 6105. nearly complete left,

anterior view - distal half of right, all views; 6106. most of right diaphysis, anterior view -

complete left, posterointernal view; 6110, nearly complete right, posterior view - nearly complete

left, internal view; 6111. distal half of left, posterointernal view - a small fragment of the distal

end of the right.

The humerus is stout. The head (Fig. 39, h) covers about 50% of the width of the proximal

end, and extends only slightly beyond the level of the tuberosities; it is transversely narrow and

Source:

PUCADELPHYSANDINUS: POSTCRANIAL SKELETON

123

Fig. 38. — Pucadelphys andinus. Stereophotos. A, 6110, axis, left lateral view; fragments of left scapula and humerus, internal

view: left clavicle, anterior view. X3; B. 6110, right humerus, posterior view, left ulna and radius, internal view; left

clavicle, posterior view, X 3.

/•/g. 38. Pucadelphys andinus. Stereophotos. A, 6110, axis, vue laterale gauche; fragments d'omoplate gauche et d'humerus,

vue interne; clavicule gauche, vue anterieure. X 3; B, 6110, humerus droit, vue posterieure; cubitus et radius gauches,

vue interne; clavicule gauche, vue posterieure. X 3.

124

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

curves sharply posteroventrally. The greater tuberosity (g t) is subequal in breadth, but is overall

larger, than the lesser tuberosity (1 1 ); it is separated from the head by a slight sulcus, and grades

distally into a long deltopectoral crest (dp c) which extends half the length of the diaphysis. The

large deltopectoral surface is triangular (apex points distally), planar, and somewhat turned down

internally; it bears only small rugosities. The lesser or internal tuberosity is not sharply demarcated

from the head (only a narrow sulcus is visible on the left side of 6106). A wide and deep bicipital

groove (b g) separates the two tuberosities anteriorly and continues distally between the deltoid

Fig. 39. — Pucadelphys cmdinus. Humerus: A. 6105, left, anterior face: B, 6106. left, posteromesial face; C.6110, right, posterior

face. X 5. All drawn as preserved. Abbreviations: b g, bicipital groove: c, capitulum; dp c. deltopectoral crest; ect,

ccicpicondvle; ent. entepicondyle; en f. entepicondylar foramen; g t. greater tuberosity; h, head; i a s. internal articular

sulcus; I t, lesser tuberosity; o f. olecranon fossa; s r. supinator ridge; t. trochlea.

Fig. 39. — Pucadelphys andinus. Humerus: A, 6/05, gauche, face anterieure; B, 6106, gauche, facepostero-mesiale; C, 6110,

droit, face poslerieure. X 5. Dessins en I 'etat de conservation. Abreviations: b g, si I Ion bicipital: c. capitulum: dp c, crete

delto-pectorale: ect, ectepicondyle; ent. entepicondyle: en f, foramen entepicondylien; g t, grande tuberosite; h. tete;

ias, sillon articulaire interne: 1 1 . petite tuberosite: o f fosse olecranienne; s r, crete du supinateur: t. trochlee.

Source MNHN. Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

125

crest and the posteroexternal border of the bone. Distal to the deltopectoral crest, the diaphysis is

narrow, subcylindrical in cross-section; but the bone rapidly flattens anteroposteriorly and flares

transversely; at that level, it is limited externally by a broad and thin supinator ridge (s r) that

extends along about 30% of the diaphysis. Internally this expanded part is limited by the

entepicondylar “crest” which consists of two ridges that border a pronounced sulcus, itself

perforated by an elliptical entepicondylar foramen (en f); the anterointernal ridge ascends

obliquely toward the deltopectoral crest. The suture between the proximal epiphysis and the

diaphysis remains distinct in adults.

The distal end is transversely wide, but the articular surfaces are low and occupy only a

narrow part of this width. The small ectepicondyle (ect) is not thickened anteroposteriorly; the

entepicondyle (ent) is ovoid, slightly convex, and protrudes internally. External to it is a broad

sulcus, sharply delimited posteriorly on its external side, less sharply anteriorly, and obliquely

oi iented (this is the internal articular sulcus of Osgood (1921 ), which in Caenolestes receives the

inner upper lip of the sigmoid cavity of the ulna). The capitulum (c) is wide anteriorly; posteriorly

it is reduced to a crest reaching the level of the base of the entepicondylar foramen. Anteriorly the

trochlea (t) and capitulum grade into one another without delimitation; posteriorly and above the

capitulum, the bone is depressed by a deep and short fossa which receives the olecranon during

flexion of the forearm (o f). The measurements effected on the humerus of 61 10 according to the

method proposed by Kielan-Jaworowska & Gambaryan (1994) indicate that there was practi¬

cally no twisting of the bone.

Ulna (Figs 35, 38B, 40A-D and 45C; Table 11). — Six elements: 6105, left, complete

except for distal epiphyses, anteroexternal view - isolated proximal half of right, internal, external

and posterior views; 6106, proximal half of left, internal view (olecranon incomplete): 6110, left,

internal view - right proximal part, external view; 6111, median fragment of left diaphysis -

isolated proximal half of right, anterointernal view.

In length, the ulna is equal to or slightly greater than the humerus. It is wider than the radius,

hence stout. The olecranon process is long, with a convex upper border (6111 right); it is slightly

inclined internally; there is no terminal tubercle. The anterior margin of the olecranon appears to

be slightly concave. Its internal face is depressed by a longitudinal sulcus, while its external face

is practically flat. The “beak” of the olecranon is directed externally, with no internal projection.

The greater sigmoid cavity (g s c) is narrow, high, oblique dorsoventrally and externointernally,

convex transversely on the internal part for the humerus, and flat on the external part, also

apparently for the humerus. The lesser sigmoid cavity (s s c) for the radius is apparently small and

only slightly excavated. There is a well developed coronoid apophysis (c a), but there is no internal

coronoid tubercle; this apophysis is strongly protruding internally, anteriorly underlain by a

depression with an orifice, and limited laterally by a strong ridge. The diaphysis is very

compressed laterally so as to form a ridge posteriorly; distally it becomes more ovoid, slightly

arched and slightly narrower. No distal extremity is preserved: the most complete specimen

(6105) did not have the distal epiphysis fused, and it was lost during fossilization.

Radius (Figs 35, 38B, 40D-F and45C; Table 12). — 6105, left, complete except for distal

126

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Source: MNHN, Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

127

epiphysis, anteroexternal view - isolated, proximal half of right, anterior, external and internal

views; 6106, left proximal third, internal view; 6110, left, internal view; 6111, fragment of left

diaphysis - isolated proximal half of right, anterointernal view.

The anteroposterior curvature of the bone is gentle, the lateral one is more pronounced

distal ly. The proximal articular surface (Fig. 40, p a s) is ovoid and compressed with a transverse

long axis, and noticeably concave. Under the thickened proximal border, there follows a neck (n),

itself immediately followed by a strong bicipital tuberosity (b t), rather highly situated. The

anterior face of the bone is convex proximally, becomes flat distal ly, the anterior ridge quickly

disappearing; the external face, wide proximally, becomes more convex distally, where a median

sulcus is discernible, presumably for the ulnar crest; the internal face narrows to a ridge distally.

The diaphysis is triangular in cross-section at mid-length. No distal extremity is preserved.

In each case, the ulna and radius are, as preserved, in contact along nearly their entire length:

only under the bicipital tuberosity is there a very narrow interosseous space. This testifies to a very

limited capacity of rotation. Proximally the bones are situated anteroposteriorly relative to each

other, but soon torsion brings the radius in an internal position relative to the ulna.

Manus. — Nothing is unquestionably preserved of the hand, but see cautionary remarks

under pes (p. 139).

Discussion

Scapula. — In Megazostrodon , there is no distinct supraspinous fossa. In Henkelotherium ,

the scapula remains more primitive than in Pucadelphys with a massive coracoid component, but

the supraspinous fossa is said to cover almost two-thirds of the external surface.

No scapula is preserved in Asioryctes. In Banmlestes , where the ventral part is preserved,

the extensive tuber scapulae and the coracoid process were developed in a way more reminiscent

of metatherians than of modern claviculate eutherians. In Pucadelphys. the tuber scapulae is

smaller than the coracoid process, as in Metachirus. In Banmlestes. the spine starts at about the

same level as in Pucadelphys, that is, slightly higher than in Didelphis. In the latter, the two fossae

arc subequal (as in Perameles), the spine is thicker than in Pucadelphys, the posterior border more

convex, the glenoid cavity flatter and more complex, and the internal face regularly concave; there

is no overhang of the coracoid process, and the plate itself is relatively thicker.

In Philander, the inlraspinous fossa remains very slightly larger than the supraspinous

Fig. 40. — Pucadelphys andinus. A-C, 6105, proximal pari of righl ulna: A. internal face; B, poslcrior face; C, external face. D.

isolated, proximal end of articulated right ulna and radius, anterior view; E- F. proximal end of left radius: E, 6105, internal

face; F, 6110. anteroexternal face. X 5. All drawn as preserved. Abbreviations: b t, bicipital tuberosity; c a. coronoid

apophysis; ext. external; I'e I, fossa for exterior ligament; g s c. greater sigmoid cavity; int, internal: n, neck; o b, olecranon

beak; o p. olecranon process; p a s. proximal articular surface; R, radius; s s c, lesser sigmoid cavity; U, ulna.

Fig. 40 .— Pucadelphys andinus. A-C. 6/05. partie proximate du cubitus droit: A, face interne; I), face posterieure; C.faceexterne;

D. specimen isole. extremite proximate du cubitus el du radius en articulation, vue anterieure; E-F, extremite proximate

du radius gauche: E. 6105, face interne; F. 6/lO,faceantero-externe.X5. Dessins en Fetal de conservation. Abreviations:

b t. tuberosite bicipitale; c a. apophyse coronoide; ext, externe;fe l, fosse pourle ligament externe; g s c. grande cavite

sigmoide; int. interne; n. col: o b. bee de Folecrane: o p, processus olecranien; p a s. surface articulaire proximate; R,

radius; s s c. petite cavite sigmoide; U. cubitus.

128

LARRY G. MARSHALL & DENISE SICOONEAU-RUSSELL

fossa: otherwise the bone is comparable to that of Metachirus; but in the latter as in Monodelphis,

it is the supraspinous fossa which is slightly larger than the infraspinous, the spine is very thin,

the posterior border straight, and the acromion not thickened, all character states as in Pucadelphys.

However, in Metachirus the tuber scapulae is lower, and the coracoid process less recurved

internally than in the other two genera; the coracoscapular suture remains distinct, at least in the

subadult specimen which we used for comparisons. Finally, the scapula of Monodelphis has a

deflected spine as in Pucadelphys.

It thus appears that Pucadelphys has an unspecialized scapula, combining all primitive

character states found isolated in living didelphids.

Clavicle. — The assignment of this bone to Eozostrodon is only tentative. The clavicle of

Henkelotherium , which must have been quite long, is rounder than that of Pucadelphys and not

widened sternally. No clavicle is preserved in the Mongolian eutherians.

The clavicle of Didelphis shows the same general features as that of Pucadelphys (but the

torsion is more accentuated relative at least to the preserved state in Pucadelphys , which may be

the result of some flattening during fossilization) and the interclavicularextremity more flattened;

also the scapular extremity does not make an angle on the long axis; finally it is relatively shorter

(1/2 of length of the scapula, compared to 2/3 in Pucadelphys).

In Metachirus , the clavicle is not as clearly S-shaped as in Pucadelphys, and there is no

posterior spiral convexity; but the sternal extremity is hollowed ventrally with a raised medial

border as in Pucadelphys. The clavicle of Monodelphis is more similar to that of Pucadelphys , but

its sternal extremity, though less angled relative to the shaft, is relatively wider; the same is true

of Metachirus. In Perameles, the clavicle is absent or rudimentary.

In conclusion, the clavicle of Pucadelphys is that of a typical non specialized didelphid.

Interclavicle. — The homologies of this bone have been widely disputed; it seems logical

to consider it as homologous to the episternum of monotremes and to the “manubrium” of

Caenolestes (Osgood. 1921). No interclavicle is known in Cretaceous therians, and, according to

Krliss (1991), none was developed in Henkelotherium. As for Eozostrodon, the same remarks

noted for the clavicle apply to the interclavicle.

In Didelphis, the interclavicle is shorter than in Pucadelphys and the arms divide it

longitudinally into two equal parts; it is also more concave dorsally, and the ventral keel

completely disappears distally. Similarly in Perameles, this keel becomes considerably lower

distally. In Metachirus the bone is very similar to that of Pucadelphys. In Monodelphis, the

anterior and posterior parts are also quite unequal as in the latter two genera, but there is practically

no ventral keel.

Overall, the interclavicle of Pucadelphys is typically didelphid.

Humerus. — In Henkelotherium, the bone is long and robust; a peculiarity is a strongly

marked muscular insertion on the mesial face of the shaft not found in Pucadelphys, which

presumably would represent the derived state for this character: in Eozostrodon the bump is even

more accentuated than in Henkelotherium. Also in the latter, the deltoid crest is robust but

Source: MNHN, Paris

PUCADELPHYSANDINUS: POSTC'RANIAL SKELETON

129

relatively shorter (1/3 of the bone against 1/2 in Pucadelphys). The entepicondyle is prominent.

Compared to Barunlestes, 6105 has a less curved diaphysis, the deltopecloral crest is more

prominent; there is the same groove between both tuberosities. But most notable is the difference

concerning the distal extremity, which is much wider in Pucadelphys-, also the entepicondyle is

better developed and there is no supratrochlear foramen; finally, the susepicondylian crest is

lacking in Barunlestes , and the supinator ridge is not as wide and sharp.

In Metachirus, the bone is relatively more slender than in Pucadelphys , the deltopcctoral

surface is as smooth and long but not deflected as in Pucadelphys, and the supinator ridge is

relatively shorter. There is no third distal articular surface; such a feature, described above in

Pucadelphys, is mentioned by Osgood (1921) for Caenolestes as being exceptional for marsup¬

ials, and he considers it as more indicative of digging than cursorial habits. The humerus of

Didelphis is more similar to that of Pucadelphys than to that of Metachirus: the proportions of

the bone are slightly different and the head is relatively wider, but dorsally the latter does not

extend beyond the tuberosities any more than in Pucadelphys-, the deltopectoral ridge extends over

half of the length of the bone and the deltopectoral surface is turned downward as in Pucadelphys ;

finally, a clear sulcus also separates the entepicondyle and trochlea, without however articulating

with the upper lip of the ulna. The situation is the same in Philander and Monodelphis, whose

humerus is very close to that of Pucadelphys in every aspect: proportions, triangular and deflected

deltopectoral surface, high supinator ridge, and distal width with a third articular surface. That of

Perameles also shows a deflected deltopectoral surface, but the bone is even stouter, the supinator

ridge less high and a supratrochlear foramen is present.

An interesting feature in the humerus of Pucadelphys is the great development of the areas

for the extensors of the forearm and the carpus, although it is relatively no greater than in terrestrial

didelphids such as Monodelphis or Metachirus. To be noted is the difference between the humeri

of these last two forms, both terrestrial, but with some saltatorial capabilities in Metachirus.

Ulna - Radius. — The ulna and radius of Henkelotherium are slightly longer than the

humerus; the proportions are not clear in Pucadelphys since the distal epiphyses of these two bones

are not preserved, but they appear to have been subequal; the two elements arc about equal in

Didelphis and Metachirus. In Henkelotherium. the ulna and radius were more slender and closely

appressed over the whole length; another notable difference is the stronger olecranon and the

deeper sigmoid incisure than in Pucadelphys. In Barunlestes the two bones have proximally an

anteroposterior position relative to each other, a position more typical of metatherians than of

eutherians. The olecranon is “slightly bent forwards” (Kielan-Jaworowska, 1978) like in

Pucadelphys, but it is shorter, notched anteriorly and terminated by a tubercle; the beak of the

olecranon is extended internally in a lip which is absent in Pucadelphys.

In Metachirus the ulna is more gracile and the olecranon shorter than in Pucadelphys, and

there is an even more developed upper internal lip above the greater sigmoid cavity, as in

Barunlestes-, also the lesser sigmoid cavity is more distinct and hollow. The same remark

concerning the upper lip applies to Didelphis-, on the contrary, in Monodelphis the internal lip is

130

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

reduced: it thus seems that, contrary to Osgood’s statement, there is no relationship between the

presence of an internal articular surface on the distal extremity of the humerus (present in

Monodelphis) and the presence of an internal upper lip on the ulna; the latter may also correspond

to an extension of the surface for the digit flexor muscles. In Monodelphis. the bone is again more

gracile than in Pucadelphys, but the olecranon is as high and the morphology very similar. The

olecranon is particularly high in Philander, but the bone is rather stout.

In Barunlestes, the proximal surface of the radius is much more complex than in Pucadelphys.

this time in a eutherian fashion. In Metachi ms, Monodelphis and Didelphis. this surface is similar

to that of Pucadelphys but more rounded. In the latter, the diameter of the radius relative to that

of the ulna comes closer to that of Metachirus and Monodelphis than to Didelphis, but the bones

are relatively more stout.

In Metachirus and Didelphis, the ulna and radius remain slightly apart for the whole length

in the latter and for the upper third in the former, Metachirus being thus closer in this respect to

Pucadelphys. The two bones remain somewhat more apart in Monodelphis. the radius being more

curved. Perameles is very specialized in this area and not comparable to the other forms examined

here.

In conclusion, the general stoutness of the ulna and radius in Pucadelphys c an be considered

as primitive, and no particular morphological specialization distinguishes these bones from those

of living terrestrial didelphids.

Pelvic Girdle (Figs 31 A. 41. 42 anil 47B; Table 13).

Three elements: 6106. complete, ventral view; 6110. left ilium (all views) and ischium

(external view) - part of right ilium and ischium (internal view); 6111. nearly complete, left and

right ilium and ischium, dorsal view.

The suture between the ilium and ischium is clear dorsally, that between the ischium and

pubis clear ventrally, but that between ilium and pubis is less distinct.

Ilium. —The ilium forms a little more than half the length of the pelvis, which makes it

relatively short; its main characteristics are the strong anterior eversion, the thinness and the

dorsoventral expansion of the anterior part. It is an elongated, spatula-shaped bone, with a deep

dorsovcntral iliac wing, and a long but not so deep iliac body. The iliac wing flares anterolaterally;

the dorsal border is acute, the ventral border even more so and the iliac crest which unites them

anteriorly is rounded and thin. The lateral (gluteal) surface is concave laterally in the longitudinal

direction; more laterally and dorsovcntrally, the wing is divided by an oblique crest into a concave

lateroventral (iliac) surface (Fig. 42, i s) and a convex lateral surface. The internal surface is

uniformly concave for vertebral muscles (anterior half) and for the sacral rib more distally

(auricular half); the entire iliac wing remains thin.

The much thicker iliac body is narrower dorsoventrally, but flares somewhat distally in front of

the acetabulum (a c); its ventral (iliac) face (i 0 is well delimited and rounded. The dorsolateral lace

is slightly concave, and the endopel vie face deeply concave. The dorsal border joins the iliac wing by

adeep notch, the great sciatic (g s n). Details of the acetabulum are not observable, the cavity being filled

in all cases with the femoral head; the majority of the cavity is ischiatic.

Ischium. — The ischium is nearly as long as the ilium, extending posteriorly from the ilium

Source: MNHN. Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

131

as a thick longitudinal branch which flattens distally. In dorsal view this branch is twisted, the

medial border becoming dorsolateral (long sciatic spine (s s) plus weakly concave sciatic notch)

and at the same time the medial face becomes dorsal. The ventral face of the same branch is slightly

convex immediately behind the acetabulum and the cotyloid notch; it becomes flat posteriorly on

Fig. 4!. — Pucadelphys andinus. Stereophotos. 6106, pelvis et L3-C6, vue ventrale; tibia, peroni el tarse droits, vue exteme. X 3.

132

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

both sides; the distal branch is wide and flat and makes a right angle with the longitudinal branch.

The ischiatic participation to the symphysis is not known since the suture between pubis and

ischium is not visible at that level. The ischiatic tuberosity (i t) is little pronounced. There is no

obturator tubercle.

Pubis. The two branches of the pubis make an obtuse angle one with the other; the acetabular

branch is relatively narrow and the longitudinal branch even narrower. There is only a weak pcctineus

tubercle (p t) at the ilio-pubic junction. The obturator foramen (o f) is obliquely elongated and the pubic

symphysis (s) relatively long. There is no pubic tubercle at the anterior junction of the two pubes.

Os marsupium .—The os marsupium (nearly complete left and right in 6106; proximal part

of right in 6111) is a long. thin, flat and relatively short bone that is broad basally, where it occupies

only about 40% of the anterior pubic border; it tapers proximally and bends slightly laterally

(Fig. 41; Table 14).

-IS

Fig. 42. — Pucadelphys cmdinus. 6106 and 6111 combined, left lateral face of pelvis, restored. X 5. Abbreviations: I, ilium; IS,

ischium; P. pubes. ac\ acetabulum; a cr, anterior crest; c b m, contact border for os marsupium; c n. cotyloid notch; g s

n. great sciatic notch; i f. iliac face; i f s. inferior posterior spine; i s, iliac surface; i t. ischiatic tuberosity; o f. obturator

foramen; s, symphysis; s f, sacral face: s s, sciatic spine; s s n, small sciatic notch.

Fig. 42. — Pucadelphys andinus. Face laterale gauche du bassin (reconstruction basee sur6i06 et 6111). X 5. Abreviations: /.

ilion; IS. ichion; P. pubis: ac. acetabulum; a cr. crete anterieure; c b m. bord de contact pour I 'os marsupial; c n, incisure

cotyloide: g s n. grande incisure sciatique; if, face iliaque; i f s. epine inferieure posterieure; i s. surface iliaque; i t.

tuberosite ischiatique; o f foramen obturateur; s. symphyse; sf face sacree; s s. epine sciatique; s s n. petite incisure

sciatique.

Fig. 43. — Pucadelphys andinus. Stcreophotos. A, 6106. right femur, posterior view; partial tibia and fibula, external view, X 3;

B. 6105. left femur, tibia and fibula, external view. X 3.

Fig. 43. — Pucadelphys andinus. Stereophotos. A. 6106. femur droit, vue posterieure; tibia et pe rones. vue externe. X 3; B, 6/05.

femur gauche, tibia et perone, vue externe. X 3.

Source MNHN. Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

133

Source MNHN, Paris

134

LARRY G. MARSHALL & DENISE S1G0GNEAU-RUSSELL

HlNDUMB

Femur (Figs 43, 44. 45C and 47A; Table 15). — Seven elements: 6105. left, missing

proximal end. anterior and external views; 6106, complete left, posteroexternal view - complete

right, posterior view; 6110. complete right, posterior and both lateral views - complete left,

anterior and partly posterior views: 6111. complete left, anterior, posterior and external views -

complete right, anterior and external views.

The femur is slightly longer than the humerus; it is practically straight, and greatly

expanded proximally.

The head (Fig. 44. h) is wide, hemispherical, and its base is circumscribed by a suture line.

A very shallow fovea capitis for the ligamentum teres is visible, in an exccntric position. The neck

is very short, making an angle of 40° with the diaphysis. The greater trochanter (g t) is slightly bent

anterointernally; it does not reach the level of the top of the head but only that of its sutural base;

it is broad and borders externally the deep trochanteric fossa (t f); a marked and short ridge delimits

the latter externally; but no clear intertrochanteric ridge delimits it internally, only a marked bump

(tc c). The lesser trochanter (11) is particularly large, lamellar and V-shaped with the apex directed

internally; a notch joins it with the base of the head. The third trochanter (t l) is represented by

a modest swelling on the external face of the diaphysis situated across from the distal edge of the

lesser trochanter.

On the anterior face, the bone is extremely concave under the crest of the greater

trochanter, then convex except for a sulcus along the edge of the external border. Further down,

the diaphysis is ovoid in cross-section, with an anterior flattening.

On the distal end, the condyles protrude posteriorly as usual; the external condyle (e c) is

wider transversely and the internal condyle (i c) longer anteroposteriorly; but they reach about the

same level distally. Anteriorly, the patellar fossa is high, wide and shallow. The external trochlear

crest is slightly sharper than the internal one. The susepicondylar tubercles (s c t) are modest but

distinct.

Patella. —There is no evidence of an ossified patella.

Tibia (Figs 41.43, 45. 46, 47A and 48; Table 16). — Six elements: 6105, complete left,

external view; 6106. complete right, external view - complete left (except for small part of

diaphysis), internal view; 6110, proximal half of left, external view; 6111, left, external view -

right, posterior view.

The tibia is slightly longer than the femur. The diaphysis is nearly straight (at least as

preserved), although in anterior view the tibial crest has a slight sigmoidal curve. On the proximal

surface, the interarticular spine is higher than the two lateral articular facets, of which the external

one (Fig. 46, e f) is the larger. The anterior, external and internal tuberosities are not very

prominent: a distinct fibular facet (f f) occurs on the posteroexternal surface, just below the

external articular facet which overlaps the shaft. Proximally, the diaphysis is triangular in cross

section (apex anterior), while distally it is flattened externointernally, but it remains relatively

wide anteroposteriorly. The external surface is flat proximally. then transversely concave distally;

the internal surface, barely convex proximally, becomes flat distally. Posteriorly and proximally,

Source:

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

135

Fig. 44. — Pucadelphys andinus. Femur, partly restored: A, anterior face of right (combined left and right 6106 and 6110); B,

posterior face of right (combined left and right 6106 and 6110): C, external face of right proximal end (6110+61 11); D.

external face of left distal end (left 6111 and right 6110 reversed); E, 6110, distal surface of right; F, 6106, internal view

of left distal end. X 5. Abbreviations: e c. external condyle; etc. external trochanteric crest; g t, greater trochanter; h. head:

i c, internal condyle; i n. intercondylar notch; i tc, intertrochanteric crest; I a, ligneaprc; 11. lesser trochanter; p f, popliteal

fossa; pa f. patellar fossa; s c t. sus-condylar tubercle; t c c, bump (voir texte); t f, trochanteric fossa; 11. third trochanter.

Fig. 44. Pucadelphys andinus. Femurpartiellement restaure: A, face anterieure du femur droit (reconstitution basee sur 6106

et6l 10); B.face posterieure du femur droit (basee sur 6106et6110); C, face externe de l 'ext remite proximo le droite (basee

sur 6110 et 6111): D. face externe de l 'extremite distale gauche (basee sur le specimen 611 1 gauche el le specimen 6110

droit re averse); E. 6110. surface distale du femur droit; F. 6106. vue interne de Fextremite distale gauche. X 5.

Abreviations: e c, condyle externe: e t c, crete trochanterienne externe; g t, grand trochanter; h, tele; i c, condyle interne;

i n; echancrure intercoixdylienne; itc. crete intertrochanterienne; l a. ligne apre; 1 1 . petit trochanter; p f fosse poplitee;

pa f fosse patellaire; s c t. tubercule sus-condylien: l c c. bosse ( voir texte); t f fosse trochanterienne: 11 . troisieme

trochanter.

136

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Source: MNHN, Paris

PUCADELPHYSANDINUS: POSTCRANIAL SKELETON

137

there is a wide and deep vertical depression; a faint “ligne oblique", which separates the popliteal

surface from that for the flexor digitorum, continues to the mid-length of the bone. Distally, the

bone widens somewhat; the internal malleolus (i m) projects slightly more distally than the

external malleolus (e m) of the fibula. The distal articular surface itself is not accessible.

Fibula (Figs 41.43,45.46A, 47A and 48; Table 17). — Six elements: 6105, complete left

(except for distal epiphysis), posterior and external views; 6106, complete right, posteroexternal

view - left, anteroextcrnal view; 6110. proximal half of left, external view; 6111, greater part of

left, anterointernal view - nearly complete right (lacking only distal epiphysis), internal view.

The fibula is slightly shorter and more slender than the tibia. The two bones are completely

separated. The fibula flattens somewhat proximally and distally, while remaining relatively

robust at mid-length; at that level, it is semi-circular in cross-section. Proximally the styloid

apophysis (Fig. 46. s a) is prominent; a distinct tibial facet occurs on the internal face making a

salient lip, and the bone is strongly concave posteroexternally for a short distance. At the distal

end, there is a broad internal facet for the astragalus (a f) and the external malleolus (e m) is well

developed.

Fig. 45. — Pucadelphys andinus. Stereophotos. A, 6106, left tibia and Fibula, tarsals and two metatarsals, internal view. X 3;

B, 6105, left tibia and fibula, lateral view. X 3; C, 6106, left femur, proximal part of tibia and fibula (top), internal

view; ulna and radius (bottom), anterior view, and right foot, dorsal view. X 3.

Fig. 45. — Pucadelphys andinus. Stereophotos. A, 6106. tibia et perone gauches, tarsiens el deuxmetatarsiens. vue interne. X 3;

B, 6/05, tibia et perone gauches. vue laterale. X 3; C, 6106, femur gauche, parlie proximate du tibia et dupe rone (en haul),

en vue interne; cubitus et radius (en bas), en vue anterieure, et pied droit, en vue dorsale. X 3.

138

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

SO -

Fig. 46. — Pucadelphysandiiuis. A. 6106. right tibia and fibula, anteroexternal face; B. 6106. right tibia, anterior face; C, 6106.

right proximal surface. X 5. All drawn as preserved. Abbreviations: a f, astragalar facet; e f, external facet; e m, external

malleolus; f c, femoral condyle; f f. fibular facet; i b, interosseous border; i f, internal facet; i g s. interglenoid spine; i ni.

internal malleolus; s a, styloid apophysis; t c, tibial crest.

Fig. 46. Pucadelphys andinus. A, 6106. tibia et fibula droits, face antero-externe; B, 6106. tibia droit, face anterieure; C, 6106,

tibia droit, surface proximate. X 5. Dessins en Fetal de conservation. Abreviations: a f, facette astragalienne: effacette

externe; e in. malleole externe; f c, condyle femoral; f f facette fibulaire; i b, bord interosseux; i J. facette interne;

i g s, epine interglenoide; i m, malleole interne; s a, apopliyse styloide; t c, Crete tibiale.

Source: MNHN , Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

139

Pcs

Elements of the pes that include some tarsals and associated metatarsals are represented

on the left and right sides of 6106. Reference of these elements to the pes is certain because the

tarsals remain articulated with the tibia and fibula. Two other element groupings (6105 left, 6110

left) are also cautiously assigned to the pes because of their possible relationships with tibia and

fibula, although in these cases they are not directly articulated with these bones. The elements

common to these two last groupings are indistinguishable, but the ?cuboids of 6110 are not

identical to those of 6106 and the ?metatarsals of 6105 are generally shorter and more robust than

in 6106. We attribute these differences to individual variation, but because the cuboid of the pes

and unciform of the manus arc very similar in didclphids, we are unable to state decisively that

the elements in 6105 and 6110 are not, in fact, of the manus.

Astragalus (Figs 41,45 A and C, and 48). — Three elements: 6106, left, dorsal view—right,

partial ventral view; 6111. left, ventroposterior view. All three specimens are displaced and

squeezed between the distal ends of the tibia and fibula on one hand, and the calcaneum on the

other, and thus arc distorted and give only partial views of the bone.

It appears that no neck separated the “head” from the “body": the former is only partly

visible. Dorsally, a continuous ridge (the medial trochlear crest) separates the inner malleolar

facet (Fig. 48, i m f), deep and elongated, from the libial or trochlear facet (t 0; the latter is shallow.

A ridge essentially parallel to the first but weaker and shorter (the lateral trochlear crest) isolates

a concave fibular facet (f f) on the external side of the astragalus. For the plantar face, and if one

interprets correctly 6111, the concave surface corresponding to the calcaneum (CaA of Szalay,

1982) is high and narrow, and limited distally by a condyle; a wide intcrosseus sulcus separates

the CaA from the distal condyle, where the surfaces for the sustentaculum tali, and for the

navicular more distally, are practically continuous one with the other.

Calcaneum (Figs 41.45A and C, 48, 49, 50 and 51; Table 18). — Five elements: 6105,

complete left, isolated, all views; 6106, nearly complete right, external view - nearly complete left,

ventral view; 6110, nearly complete right, isolated, all views; 6111, incomplete left, external view.

The following description is based on 6105 left (Fig. 49). The posterior half or heel accounts

for almost 50% of the total length of the calcaneum; it curves internally at the posterior end. It is

high dorsoventrally and narrow from side-to-side; the dorsal border is even narrower than the

ventral one. It terminates with a robust tuber calcanei (t c), the posterior face of which slopes

anteroventrally; the medial side is more bulbous than the lateral. The anterior half of the calcaneum

is subequal in length and breadth. The dorsal surface is dominated by a high, ovale tuberosity for

one of the two calcaneo-astragalar facets (CaA); medial to it is a broad shelf, the sustentaculum

tali (s t), supporting the small and ill-defined second calcaneo-astragalar facet; close to the CaA

is the narrow, anteroposteriorly elongate sulcus calcanei (s c). pierced by a tiny vascular foramen.

The long axis of the CaA is oblique relative to the axis of the heel (-45°); its surface itself is convex,

more so on the external side, and no clear fibular surface (CaFi) is discernible (unless the external

border of this facet is for the fibula?). Anterior to the CaA is a notable expansion, the distal process

140

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Fio. 47. — Pucadelphys andinus. Stereophotos. A. 6110. right femur, external view; half tibia and fibula and last thoracic vertebra,

ventral view. X 3; 13.6110. proximal part of left femur, left ischium, right cuboid and metatarsals II. 111. IV. dorsal view.

X 3.

Fig. 47. — Pucadelphys andinus. Stereophotos. A. 6110. femur droit, vue externe; demi-tibia et perone et dermeres vertebres

thoraciques. vue ventrale. X 3: IS. 6110. partie proximate du femur gauche, de I’ischion gauche, du cubotde droit et des

metaiarsiens II. III. IV. vue dor sale. X 3.

Source: MNHN , Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

141

(d p), which is slightly convex dorsally, a convexity limited internally by a strong concavity, but the

sulcus for the tendon of the peroneus brevis is poorly defined (gt pb). The concavity is limited externally

by a rounded rim corresponding to a well developed peroneus process (p p) (somewhat incomplete on

the specimen). The edge of the distal and peroneal processes limit dorsally the cuboidal facet (c f).

Fig. 48. Pucadelphys and inns. A. 6106. right pes. dorsal lace: B, 6106. Icfl pes. ventrolateral face; C. 6110. left metatarsal III.

proximal extremity, external face. X 5. Drawn as preserved. Abbreviations: A, astragalus; CA. calcaneum; CU, cuboid;

F. fibula; MT. metatarsal; NA. navicular; T. tibia; c, cavity for adjacent metatarsal; e m, external malleolus; f f, fibular

facet: i m, internal malleolus; i m f, inner malleolar facet; p p, peroneal process; r. ridge, see text; su, sustentacular condyle;

t f. trochlear facet.

F,a Pucadelphys ahdinus. A, 6106, pied droit, face dorsale; B, 6)06, pied gauche, face ventro-laterale; C,6110, metatarsien

HI gauche, extremite proximate, face externe. X 5. Dessins en l 'etat de conservation. Abreviations: A. astragale; CA.

calcaneum: CU, cuboide; F, perone: MT. metatarsien; NA. naviculaire; T. tibia: c, cavite destinee an metatarsien

adjacent: e m. malleole externe;fffacettefibulaire: i m. malleole interne; i mf facette malleolaire interne;pp, processus

peroneen; r. cote, voir texte; su. condyle sustentaculaire; t f, facette trochleaire.

142

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

In plantar (ventral) view, the peroneal process forms a triangular point anteriorly; it is quite

protruding and is bordered internally by a bump (b), the distal plantar tubercle; between the two is a

groove, presumably for the digit V abductor. The cuboidal facet is vaguely diamond-shaped and quite

shallow; it is inclined from anterior to posterior and external to internal, and nearly perpendicular to the

long axis of the bone.

The calcaneum 6110 is smaller than 6105, as is the entire specimen; it is also somewhat abraded.

Its main difference concerns the cuboidal facet, which is oriented more internally than distally.

Fig. 49.— Pucadelphys andinus. 6105, left calcaneum: A, external face; B. ventral face; C, dorsal face. X 15. Abbreviations: b.

bump (see text); CaA. astragalarcondyle; c f. cuboidal facet; d p. distolateral process; gt pb. groove for tendon of peroneus

brevis; h, heel; p p. peroneal process; s c. sulcus calcanei; s t, sustentaculum tali; t c, tuber calcanei.

Fig. 49.— Pucadelphys andinus. 6105 , calcaneum gauche: A, face externe: B.face venirale; C.face dor sale. X 15. Abreviaiions:

h; hosse (voir texte): CaA. condyle astragalien; cf facette cuboidienne; dp. processus disto-lateral; gt pb. sillon pour

le tendon du peroneus brevis: h. talon; p p. processus peroneen; s c. sulcus calcanei: s t. sustentaculum tali; t c, tuber

calcanei.

Source MNHN. Paris

PUCADELPHYS AND IN US: POSTCRANIAL SKELETON

143

Navicular (Figs 4, 41, 45C and 48). — One element: 6106, right, complete but partly

hidden by adjacent bones.

C uboid (Figs 45 A and C. 48 and 52: Table 19). — Two certain elements: 6106, right, dorsal

view—left, plantar view. Two uncertain elements: 6110 right, all views except ventral; and 6105

left, dorsal and external views.

The cuboids of 6106 show a triangular dorsal face (Fig. 52C), with a distal base articulating

with the whole surface of metatarsal IV and a small part of metatarsal V (su 1V-V). This dorsal

face is almost flat; its internal border is slightly indented for the cuneiform. Proximally a crest

separates the internal and external surfaces: the former is deep, flat and quadrangular. The external

face (Fig. 52B), articulated for more than the proximal half with the calcaneum (c s), and dorsally

with the navicular (n s) seems to be almost flat; the distal half of this external face bears two

tuberculae separated by a deep pit for the tendon of the peroneus longus muscle (t p I). Finally,

the ventral face seems to be narrow and convex from side-to-side.

The cuboid? of 6110 differs essentially in its proportions, being stouter, and also in the

calcancar facet, which is clearly convex proximodistally. The dorsoexternal bump is much more

protruding than the ventral one. These differences may be attributed to better preservation of the

bone in 6110. On the distal face, the facet for metatarsals IV and V forms a broad triangle; no relief

delimitates the respective facets for these two metatarsals; however, that for metatarsal V must

have been very narrow. A very small astragalar facet is discernible on the proximal corner.

Cuneiform IV? (Figs 5 and 52A, C; Table 19). — One element: 6110, left, dorsal and internal

views.

It is transversely narrow, short, concave and recurved mesially for contact with cuneiform

III. convex laterally for the cuboid. It seems to bear a proximal bump ventrally.

Fig. 50. — Pucadelpliys andinus. 6110. right calcaneum: A. plantar view; B. internal view; C. dorsal view; D. external view.

X 9. Drawn as preserved. Abbreviations: see caption to Fig. 49.

biG. 50. Pucadelphys andinus. 6110, calcaneum droit: A, vue plantaire; B. vue interne; C, me dorsale; D, me externe. X 9.

Dessins en Be tat de conservation. Abreviations: voir legende Fig. 49.

144

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Metatarsals (Figs 4B, 33B. 45A and C, 47B and 48; Table 20). — Represented in four

specimens: 6105?, left. II-V, dorsal view; 6106, right. Hl-V, dorsal view - left, IV and proximal half

of V, dorsal view; 6110?, left, II, III and distal pait of IV, dorsal view.

As mentioned above, the relative robustness of the left metapodials of 6105 compared to the

undisputable metatarsals of 6106 (though of lesser amplitude than the discrepancy observed between

the foot and hand of Didelphis) led us to suspect the reference of these metapodials to the manus rather

than pcs. Nevertheless, we cautiously choose to interpret this difference as individual variation.

In length, the middle metatarsals (II, III. IV) are about one third the length of the tibia-fibula;

metatarsals III and IV are subequal in length, II is slightly shorter and V is the shortest. On II to

IV, the distal end is broader than the proximal; V differs in having, proximaily, a broad laterally

directed flange (extending well beyond the cuboidal articulation) while its distal end is narrower.

The shape of these distal ends is similar for III and IV; but on II the mesial side is expanded and

on V the lateral side is expanded; this end is strongly convex with a deflection on each side, and

a depression dorsally and ventrally. On II to IV the diaphysis is subrounded in cross-section at mid-

Fig. 51.— Didelphis. Leftcalcaneum: A. plantar view (with cuboid); B, dorsal view. X 3. Abbreviations; Ca A, astragalar condyle;

cu, cuboid process; mtIV, surface for metatarsal IV; p p. peroneal process; s c, sulcus calcanei; st, sustentaculum tali.

Fig. 51. Didelphis. Calcarteum gauche: A. vue plantaire (avec cuboide); B, vue dorsale. X 3. Ahreviations: CaA, condyle

astragali en; cu, processus cuboidien: mtl V. surface pour le meiatarsien IV: p p. processus peroneen; s c, sulcus calcanei:

st, sustentaculum tali.

Source

PUCADELPHYS ANDINJS: POSTCRANIAL SKELETON

145

length. Internally and proximally each metatarsal is encased in a pit of the previous one (Fig. 48C.

c), thus making the pes a rather rigid unit. This arrangement and preservation of the digits II to V

suggest that digit I was, at least, partly divergent, hence possibly opposable.

Phalanges (Fig. 33B). — One proximal phalange in 6105 (left). It is about half the length

of the metatarsals. The proximal extremity is wider than the distal, which is convex with no lateral

deflection; only a pit borders this convexity on each side.

Fig. 52. — Pucadelphys andinus. 6110. left ?cuboid and '.’cuneiform IV: A. proximal face of both: B. external face of'.’cuboid:

C. dorsal lace of both; D. distal face of ?cuboid. X 9. All drawn as preserved. Abbreviations: c s. calcanear surface; CU.

cuboid; CUN, cuneiform; ext, external; int, internal; n s, navicular surface; suIII, surface for metatarsal 111; suIV. surface

tor metatarsals 1V-V; t p 1, pit for tendon of peroncus longus.

hiG. 52. - Pucadelphys andinus. 61/0. ? cuboide et ?cuneiforme IVgauches: A, face proximate des deux os; B.face exierne du

?cuboide; C, surface dorsale des deux os; D,face distale du ?cuboide.X9. Dessins en l 'etat de conservation. Abreviations:

c s. surface calcaneenne; CU, cuboide; CUN, cuneiforme; ext, exierne; int, interne; n s. surface naviculaire; suIII,

surface pour le metatarsien III; suIV, surface pour les metatarsiens IV- V; t p l, fosse pour le tendon du peroneus longus.

Discussion

Pelvis. — In Henkelotherium, the pelvis is very similar to that of Pucadelphys, with an

elongate ilium and symphysis, a clear incisura acetabuli, and a small participation of the pubis to

the acetabulum; however, in Henkelotherium, the ilium is relatively longer, apparently not as deep

and somewhat thicker, and no posterior iliac spine is developed. The ischium is more robust

distally, the obturator foramen relatively smaller, and the stout os marsupium occupies the whole

anterior border of the pubis; there is no ilio-pectineus tubercle, which we interpret as the primitive

state.

146

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

In Barunlestes, the main differences with Pucadelphys concern the iliac wing, which is

relatively longer and narrower, and the wider obturator foramen; the ilio-pectineus tubercle is

developed as well as a sciatic spine. There may have been an os marsupium in that genus, as

suggested by the presence of a concavity on the anterior margin of the pubis (Kielan-J aworowska,

1975).

In Metcichirus. the ilium is not as everted as in Pucadelphys ; it is also relatively longer, with

a thickened anterior border and a not so deep ventral wing. But the ischium and pubis are very

similar to those of Pucadelphys, with a similar limitation of the os marsupium to the medial half

of the anterior border of the pubis, and an open symphysis angle. Perameles also resembles

Pucadelphys in the high, thin, everted iliac wing, though it is relatively longer; but the ventral face

of the ilium is wider, and there is practically no incisure in the acetabulum; the os marsupium is

narrower than in Pucadelphys, being reduced to a rod; the distal ischiatic branch is also wider,

hence the obturator foramen is smaller and the ventral puboischiatic carina is not salient. This

carina is, on the contrary, very salient in Monodelphis, where the ilium is longer relative to the

ischio-pubic part than in Pucadelphys and thicker than in Metachirus. Finally, the pelvis of

Didelphis is more different, with a much longer and thicker ilium; the lateral ridge is very thick

and the ventral face is ventroexternal; it has a relatively longer sacral surface for the two sacral

vertebrae; the os marsupium is much wider distally. The pelvis of Philander is close to that of

Didelphis , with a lower ilium and a short puboischium, but the very short os marsupium occupies

only half of the anterior pubic border. The sutures between the three pelvic components are clear

only in the subadult specimen of Metachirus used for comparison.

We conclude that Pucadelphys has a peramelid-like ilium, the dorsoventral anterior

expansion of which we regard as a specialization, as well as the large size of the obturator foramen

and the smallness of the os marsupium. But the relative shortness of the ilium and the massiveness

of the ischio-pubic ensemble conform to what we infer to be the primitive state.

Femur. — In Henkelotherium, the femur is very similar to that of Pucadelphys-, but the

angle between the head and diaphysis is only 30° and there is no third trochanter. Barunlestes

shows more differences: angle 60°, neck longer, greater trochanter higher, lesser trochanter less

expanded proximodistally but more internally, internal condyle larger.

The femur of Metachirus is, as the humerus, more slender than that of Pucadelphys, with

also a longer neck; the tuberosity internal to the trochanteric fossa is much more prominent, the

greater trochanter reaches higher and there is hardly any third trochanter (the latter is said to be

rare in marsupials: only three genera are mentioned by Osgood. 1921); finally, the slope of the

distal border is quite steep, the medial condyle extending distally beyond that of the lateral

condyle. The same remarks apply to Didelphis, where these features are more accentuated. The

femur of Philander is robust and particularly wide distally; the neck and lesser trochanter are well

developed, and the greater trochanter reaches the top of the head. The femur of Monodelphis

resembles that of Pucadelphys in its proportions, its low greater trochanter and the quasi-absence

of a neck, but the intertrochanteric fossa is less deep, the lesser trochanter is not so well developed

though of the same shape and there is no third trochanter, no more than in Perameles-, in this genus

Source

PUCADELPHYS AND1NUS: POSTCRANIAL SKELETON

147

(he lesser trochanter is even more reduced, and shows again some specializations, particularly

distally, that arc probably related to fossorial adaptations. The fovea capitis occupies the same

position in Pucadelphys and Metachirus, but it is better delimited and deeper in the latter; it is also

taint in Monodelphis and poorly delimited, but situated higher on the head; it is eccentric in

Eozostrodon too, less so in Didelphis and Perameles, but clearly circular.

Overall, the femur of Pucadelphys appears somewhat more generalized than those of the

living didelphids.

Tibia-Fibula. — From the figurations, the tibia of Henkelotherium seems to have been more

curved than in Pucadelphys-. in any case this is true of Metachirus, where the bone narrows more

in the distal half. But in the latter genus, the general morphology is very close to that of

Pucadelphys , including detailsof the proximal surface and those of the internal face of the internal

malleolus. One difference concerns the medial proximal facet, which is narrower, longer and

deeper; these characters are even more accentuated in Didelphis and Perameles. The tibia of

Monodelphis is also sigmoidal, narrower anteropostcriorly, but otherwise again very similar to

that of Pucadelphys.

In Barunlestes the tibia and fibula are partly fused (an apomorphy found in some

euthenans), and the tibia is more curved and narrower distally than in Pucadelphys.

In Henkelotherium , the fibula remains quite thick as in Pucadelphys. In Metachirus (and

Monodelphis) the fibula is at the same time much more slender and more expanded proximally,

and the tibial articular surface is less salient medially and distally. The specialisations toward

fossorial adaptations observed on the femur of Perameles apply to the tibia-fibula. In all

didelphids the two bones are completely and widely separated.

The fibula of Pucadelphys is clearly not as specialized as in living didelphids, being

intermediate between those of the latter and that of Henkelotherium.

Astragalus. — The astragalus of Pucadelphys differs from that of Didelphis , where a short

neck sepaiates head and body: the condition in Pucadelphys , also present in Caenolestes, is

considered to be primitive by Osgood (1921). The configuration of the two tibial facets suggests

that the tibia of Pucadelphys was well locked with the astragalus so as to allow only slight lateral

movements. Moreover, in Didelphis the CaA is more crescent-like, the interosseous relatively

narrower and the condyle for the sustentaculum tali is more distinct. Dorsally the trochlear is even

shallower but wider.

Only a very partial astragalus is preserved in Henkelotherium. That of Barunlestes is very

different, with a distinct neck, and a medial ridge shorter than the lateral ridge. The astragalus of

Deccanolestes, a placental from the late Cretaceous of India, is of interest in our comparisons

because it has been shown by Godinot & Prasad (1994) to display arboreal specializations, in the

trochlea (lateral crest higher than the medial crest, a condition opposite to that of Pucadelphys),

in the angulation of the navicular facet relative to the sustentacular facet (again contrary to the

situation of Pucadelphys).

Calcaneum. The preserved part of the calcaneum of Henkelotherium is reduced to the

tuber calcis. That of Zalambdalestes lacks a peroneal tubercle and the cuboidal facet is

148

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

perpendicular to the long axis of the bone, both states being more derived than those of

Pucadelphvs. In Didelphis( Fig. 51), the calcaneum has a relatively shorter heel, the sustentacular

facet is clearly delimited (not in Pucadelphys) but there is no deep sulcus between it and the CaA;

the long axis of the CaA (which is relatively less salient) is not so oblique relative to that of the

heel. An important specialization of the calcaneum of didelphids is the plantar indentation which

extends the cuboidal facet, itself completely terminal; the distoventral orientation of the cuboidal

facet in Pucadelphys would thus seem to foreshadow the didelphid state.

The calcaneum of Pucadelphys is thus different from that of Szalay’s “metatherian

morphotype” (1982a and b); in particular, the apparent absence of a fibular facet (CaFi, Szalay,

1982) corresponds rather to his “didelphid morphotype tarsus”, with a bicontact upper ankle joint

(UAJ); however, this calcaneum resembles the plesiomorphic metatherian morphotype in having

a large peroneal process and “remarkably broad transverse dimensions from peroneal process to

the medial margin of the sustentaculum” (Szalay, 1982a: 626). The presence of a large peroneal

process is also mentioned in Deccanolestes (Prasad & Godinot, 1994) but is there interpreted as

“related to powerful movements of inversion and eversion", hence to a very mobile foot; this

character is indeed associated, in this genus, with other indicators of such a condition, for instance:

difference in the size of the astragalar and calcanear caA, well rounded astragalar navicular facet.

Cuboid. — The cuboid of Didelphis is indented distally, and the articular surface for

metatarsal V is better delimited. The greatest difference occurs ventrally, with the development

of the proximal longitudinal process directed towards the secondary facet on the calcaneum. The

intermeshing of cuboid and cuneiform III occurs in Didelphis and also in Philander.

Pes. — No tarsus of Metachirus, Philander or Monodelphis was available for comparison.

The pes of Perameles is very specialized for digging, with a narrow tarsus and metatarsus.

In Henkelotherium, the metatarsals represent about 1/3 of the length of the tibia-libula. In

Monodelphis, the metatarsals equal less than 1/4 the length of tibia-fibula and the digits are

interlocked proximally as in Pucadelphys. In Philander, the metatarsals are short (1 /6th the length

of tibia-fibula) and not linked to one another; II>III =IV>V. In Didelphis , the metatarsals,

independent from one another, are only 1 /5th the length of the tibia-fibula; III>II =IV>V.

Thus, the fool of Pucadelphys (where the metatarsals represent more than 1/3 the length of

the tibia-fibula) was relatively long and rigid.

The proximal phalanges are almost as long as the metatarsals in Henkelotherium and

Philander, slightly more than half as long in Monodelphis and Didelphis (and about half in

Pucadelphys).

HABITS

About 22 species of mammals are known from Tiupampa, of which 50% are eutherian and 50%

metatherian. As demonstrated by all skeletal and dental remains thus far collected, Pucadelphys

andinus is the most abundant species of mammal in this fauna (Marshall & Muizon, 1988).

Source .

PUCADELPHYSANDINUS: POSTCRANIAL SKELETON

149

Based on the adult specimens (6105 and 6106), Pucadelphys andinus measured about 135

mm from tip of snout to base of tail, and about 270 mm from tip of snout to tip of tail (assuming

the presence of ±30 caudal vertebrae). In size it compares with some living species of Mannosa

and with Lestodelphis halli (Marshall. 1977), and thus weighed about 50 grams.

Locomotion. — Many living didelphids are terrestrial ( e.g. Lutreolina, Lestodelphis.

Metachirus , Monodelphis , some Marmoset) and those that have arboreal capabilities (e.g. some

Mannosa. Philander. Didelphis) are often found on the ground. Perameles is specialized for

digging and lives in burrows. From the above description, what can be deduced about the

locomotion in Pucadelphys ?

Hildebrand (1961) concluded from his study of body proportions in didelphids that the

measurements of the various long bones did not yield any significant results: “the more arboreal

animals differ from the semi-arboreal and terrestrial animals in behavior patterns but not in

morphology". Also. Jenkins (1971) insisted justly on the importance of physiological factors in

locomotion : “even if biomechanical conditions are similar, there may be profound differences ..

in, for example, agility, in ratio or intensity of locomotor activity”; and in 1974, the same author

noted: “the differences between arboreality and terrestriality of tree shrews is at times only

behavioral". Finally. Barnett & Napier (1953) wrote that “there is remarkably little morphologic

difference between the form and mobility of the fibula in the ecologically distinct species

(arboreal, unspecialized terrestrial and aquatic)”. However, Grand (1983) remarked that,

whereas Metachirus and Monodelphis are both terrestrial, the hindlitnb elongation in the former

genus coincides with its agility and bounding locomotion; Monodelphis , with his subequal

anterior and posterior limbs, remains effectively slower.

Belore proposing locomotion capabilities of Pucadelphys , we shall summarize its skeletal

characteristics:

1. Short spinous processes of the cervicals, are compatible with good flexibility of the neck.

2. Large dorsal metapophyses, long lumbar spinous processes as well as broad distal end of

humerus are suggestive of digging abilities, while the same long and widely distant lumbar spines

favor leaping capabilities (Slijper, 1946; Gambaryan, pers. comm., 1992).

3. I he smallness of the space between ulna and radius suggests a very limited rotation of the

forearm; the relation between the two bones seems to indicate a slight anteroexternal proximal

superposition, a side-by-side position along the diaphyses and a slight anterointernal superposi¬

tion distally.

4. The pelvis usually reflects locomotory specializations: leaping or digging forms have recog¬

nizable features on the bone, but these are extreme cases; for less specialized forms, interpretation

is not so clear. Also, the shape of the pelvis not only depends on function but also on its relation

to the viscera, as indicated by Elftman (1929). In Pucadelphys we have noted a possible mobility

of the sacro-iliac joint (as in Perameles)-. the broadness of the sacrum and the large ilio-sacral angle

(both as in the digging form Perameles ); the wide, flat and laterally flaring surfaces for the erector

spinae and gluteus medius (abductor and extensor of the thigh), again as in Perameles-, similarly

the pronounced downward curvature of the transverse processes of lumbar vertebrae corresponds

to powerful erector spinae (Elftman, 1929). The propulsive part part of the pelvis (behind the iliac

150

LARRY G. MARSHALL & DENISE SIGOCNEAU-RUSSELI.

blade) is short, though not as short as in Perameles. The ischia are elongated to provide leverage

for the hamstrings and increase the power of the adductors as extensors of the thigh as in leaping

or digging forms; measurements for is/il ratio are: Perameles, 0.67; Pucadelphys, 0.53; Didelphis,

0.36', Metachirus, 0.47; Monodelphis, 0.37. All of these observations suggest digging capabilities

for Pucadelphys. However, the posterior border of the ischium is slightly inclined forward and

downward (while it is more vertical in Perameles): this can be interpreted as giving less strength

to the extensors; also there is no strong buttress on the acetabulum, anteriorly or posteriorly (such

a buttress is strong in digging lorrns), and finally the iliac muscle (which extends from the vcntial

surface of the ilium to the smaller trochanter and prevents the body from falling forward when

discing) is esteemed to have been modest in Pucadelphys. But it should be kept in mind that there

are many ways of digging and thus adaptive strategies can be different.

5. Relative limb proportions of Pucadelphys, with subequal fore- and hindlimbs, are intermediate

between those of Metachirus and Monodelphis, and testify to a good agility (Julien-Laferriere,

1991).

6. A tibia locked to the astragalus corresponds to a relatively rigid hindlimb (as opposed to an

arboreal form); the relatively strong fibula (as in generalized marsupials: Barnett & Napier.

1953 ) lias only a limited distal contact with the tibia. Finally, the well developed internal and

external malleoli usually indicate saltatorial capabilities (Barnett & Napier, 1953)

7. The astragalus and calcaneum of Pucadelphys do not show arboreal specializations such as

mentioned above (Godinot & Prasad, 1994), though the large peroneal process may indicate good

mobility. The pes is relatively long, and it is deduced from the preservation of 6106 that the 1st

digit (itself missing on the specimen) was at least slightly divergent; if not, one would assume that

metatarsal 1 would have been preserved along with II and III (the big toe is only slightly divergent

in Monodelphis, more so in Metachirus-, note that divergence does not necessarily mean

opposability, and simple opposabi lity does not mean arboreality). Moreover, the united metapodials

suggest a rigid pes: united digits and long metapodes are usually related to digitigrady, but the

latter normally accompanies at least a reduction of digit I, a situation unknown in Pucadelphys.

Finally, the configuration of the tibio-astragalus joint is not known; we thus ignore if it was such

as to allow reversal of the foot, as observed by Jenkins & McLearn (1984) in some arboreal

didelphids.

Based on the above, we prudently infer that Pucadelphys was essentially terrestrial, quite

agile but neither cursorial nor saltatorial; it appears to have been capable of bounding and also of

some digging, these two abilities being less developed than in Metachirus and Perameles

respectively.

Way of life. — Pucadelphys was probably nocturnal as are most living didelphids, sleeping

in a burrow-nest during the day and foraging for food at night. Most living didelphids are solitary

except during the breeding season (Walker. 1964); McManus (1970) stresses the poor social

behavior of Didelphis. However, it is not uncommon that one individual would share a burrow

with a congenere (or even a “foreigner”, Shirer & Fitch, 1970), especially in cold weather

(didelphids with their naked ears, tail and paws are ill adapted to cold weather: Fitch & Siiirf.r,

Source:

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

151

1970) and also depending on the population density (Crawley ,1973). Yet, there are no records

that two males would huddle together (McManus, 1970; Shirer & Fitch, 1970). These

observations suggest two interpretations for the specimen-pairs of Pucadelphys andinus : I) they

were male-female pairs that bonded for a restricted breeding season; or 2) that they were two

females that shared the same burrow. For females of Didelphis, Fitch & Shirer (1970) have

observed that, “from time to time, two individuals were staying together in the same den

simultaneously; on some occasions it seemed that the animals must have been side-by-side and

in actual contact”. A definitive choice between these two interpretations is not possible, as there

arc no features in either the skull or skeleton of didelphids which permit secure sexual

identification [although in small mammals, females are often larger than males; Wooller el al.,

1981; but in Trichosurus , males are only 1% larger than females (Crawley , 1973)]. Nevertheless,

the occurrence of two specimen-pairs (i.e., a repeated association) of Pucadelphys at Tiupampa

suggests the most natural and simplistic relationship (i.e. male-female pairs), rendering the first

interpretation as most likely.

Ground nests of all of the mentioned didelphids have been found in hollow logs, under rocks

or in burrows (Walker. 1964). Didelphis curls in a den when it is cool, limbs close to the body,

head under body; when very hot it lies on its back (McManus, 1970). Thus, the “died-in-a-burrow”

hypothesis (sec above), and the three dimensional life-like position of the fossilized specimens are

consistent with observed behavior and habits of some living didelphid taxa.

CONCLUSIONS

The vertebrate fauna from Tiupampa accumulated in channels of meandering rivers on a flat

alluvial plain. The presence of several groups of crocodiles attests to a warm, probably subtropical

climate. The fauna is from the middle member of the Santa Lucfa Formation which, based on a

detailed magnetostratigraphic study, would be between 59.5 and 59.0 Ma (Sempere etal ., in prep.).

The four specimens of Pucadelphys andinus described here represent the earliest and most

complete articulated skeletons of metatherians yet known. The two sets of what are interpreted

to be male-female pairs were found in a three dimensional, life-like, snout-rump position in

burrow-nests that were apparently dug in a river bank. The animals probably died as the result

of a flood which entrapped them in their burrows and filled the latter with water and sediment.

Functional considerations of the skeletons suggest that Pucadelphys was essentially terrestrial,

quite agile, and possessed limited bounding and digging capabilities. The vast majority of

character states in the skeleton of Pucadelphys are regarded as mammalian, tribosphenic and

metatherian plesiomorphies (i.e. atlas imperforate and with a persisting suture between ossified

intercentrum and atlantal arch; absence of transverse canal on axis, with possible unfused rib;

absence of enclosed transverse canal on CV7; strong fibula; presence of ossified os marsupium).

The plesiomorphic states of the cervical vertebrae are not collectively found in any living

didelphid examined.

Numerous other skeletal features of Pucadelphys also do not occur as a suite in any of the

didelphids examined. These include; 1) a single fulcral vertebra (S2); 2) a long, non-prehensile

152

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

tail; 3) specialized lumbar series (gradual lengthening of vertebral body and transverse processes,

and long anteriorly directed neural spines); 4) specialized pelvis (ilium dorsoventrally expanded

anteriorly, large obturator foramen, small os marsupium); 5) possible movable sacro-iliac joint;

and 6) digging specializations of the humerus [no third distal articular surface (Osgood. 1921),

large areas for extensors of forearm and carpus]. Except for feature 2 and a small os marsupium

(part of 4), the others are specializations reflecting the bounding and digging capabilities noted

above.

The only postcranial element that reflects the phylogenetic position of Pucadelphys

within Metatheria is the calcaneum. This bone has a bicontact upper ankle joint (UAJ) which is

diagnostic of the family Didelphidae; moreover, the partially plantar orientation of the cuboidal

facet can be interpreted as foreshadowing the state in living Didelphidae where the calcaneum has

a distinct ventral indentation of the cuboidal facet. The calcaneum of Pucadelphys also

approaches the "plesiomorphic metatherian morphotype” of Szalay (1982a, b; 1984) in having

a large peroneal process and remarkably broad transverse dimensions from peroneal process to

the medial margin of the sustentaculum. This combination of tarsal states supports the interpre-

tation based on the study of the skull, dentary and dentition (Marshall & Muizon, this volume)

that Pucadelphys represents the most plesiomorphic taxon within the family Didelphidae.

ACKNOWLEDGEMENTS

Aspects of this study were made possible by three grants (2467-82. 2908-84. 3381-86) from the National Geographic

Society; a grant from the Gordon Barbour Fund. Department of Geological and Geophysical Sciences. Princeton University; two

grants from the National Science Foundation (EAR-8804423. INT-8814059); a Collaborative Research Grant (no. 86/0013) from

die North Atlantic Treaty Organization (NATO, obtained by C. de Muizon for L. G. M.); five months as “Professeur Associe” for

L. G.M. at the Museum National d'Histoire Naturelle (MNHN). Paris; an “Action Specifiquc” (no. 1337) from the MNHN; and

a six month visiting professorship from the Ministere de la Recherche. France (1994) obtained for L.G.M. by Prof. Claude Babin

(Lyon). Fossil collecting was initially carried out under the auspices of the Instituto Boliviano de Biologiade Altura (IBB A). La

Pa/, and the Instilut Frangais de Recherche Scientifique pour le Devcloppement en Cooperation (Orstom); then in collaboration

with the Centro de Tecnologia Petrolera (CTP) of the Yacimentos Petroliferos Fiscales Bolivianos (YPFB). Santa Cruz. Bolivia;

and finally with the Associacion Boliviana de Paleontologia and the Fundacion Para Las Ciencias in Cochabamba. Bolivia. The

specimens described in this study are the property of the YPFB (CTP). Special thanks to M. Suarez R.. R. Suarez S.. C Molina

M. . J. Lobo B.. M. Gayet. P.-Y. Gagnier, R. Cespedes and C. de Muizon for their collaboration and/or for providing logistic

support. We are deeply thankful to Dr. P. P. Gambaryan for having examined our specimens and given useful suggestions, and

to Prof. Z. Kielan-Jaworowska who provided constructive remarks on an early version of the manuscript, and later accepted the

arduous task of refereeing the final version. Dr. F. Jenkins, who performed very thoroughly the same time-consuming toil, greatly

contributed to a supplement of rigor and precision in the text. The ink drawings are by E. Liebman (Chicago), except Figs 8. 18,

19 and all lettering, which are by Mine Pilard; the photographs are by D. Serrette and L. Merlette (stereos), except Figs 11 and

33. which are by C. Weber-Chancogne (SEM) (URA 12 CNRS for the last four technicians).

Source

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

153

REFERENCES

Barnett. CR.& Nabbr. JR - 1953.- The form and mobility of the fibula in me.atherian mammals. Journal of Anatomy. 53:

Bonaparte J. & Rougier G W 1987— Mam.Teros del Cretdc,co Inferior dc Patagonia. IV Congreso Latinoamericano de

haleontojogia, Santa Cruz, Bolivia, 1: 343-359.

Coues, E.. 1871- The osteology and myology of Didelpkys virginiana. Memoirs of the Boston Society of Natural History. 2:

Crawley. MC 1973—A live-trapping study of Australian brush-tailed possum. Trichosurus vulpecula (Kerr), in the Orogongo

Valley. Australian Journal of Zoology. 2: 75-90. b =

Crochet. J.-Y.. 1980— Les Marsupiaux du Teniaire d'Europe. Fondation Singer-Polignac. Paris: 279 p.

Cuvier. C.. 1804— Memoire sur lesquelet.e presque entier d’un petit quadruple du genre dcs sarigues. trouve dans la pierre

a platrc ties environs de Paris. Annales du Museum national d'Histoire naturelle , Paris. 5: 277-292, pi. 19.

Elftman. H_0 1929— Functional adaptations of the pelvis in marsupials. Bulletin of the American Museum of Natural Him

so: lo9-232.

Fischer, J.B.. 1829. — Synopsis Mammalium. Stuttgart. 62: 1-752.

Fitch. H.S.. & Shirer. H.W.. 1970— A radiotelemetric study of spatial relationships in the opossum. The American Naturalist.

o4: 1 /()-186.

Flower, W.H., 1885.— An Introduction to the Osteology of the Mammalia. McMillan: 1-383.

Gambaryan, P., 1974.— How mammals run: anatomical adaptations. Halsted Press. Jerusalem. XI + 367p.

Gidley. J.W.. 1919.- Significance of divergence of the first digit in the primitive mammalian foot. Journal of the Washington

Academy of Science . 9(10): 273-281.

Godinot, M.. & Prasad, G. V.R., 1994.— First Cretaceous arboreal eutherians discovered in India. Naturwissenc/iaften, 81: 79-81.

Grand. 1.1 1983.— Body weight: its relationship to tissue composition, segmental distribution of mass, and motor function. III.

I he Didelphidae of French Guiana. Australian Journal of Zoology, 31: 299-312.

Grasse, P.P.. 1967.— Trane de Zoologic : Anatomie, Systematique, Biologie. Mammiferes : teguments ct squelettc T XVT

I -1162. Masson, Paris.

Hildebrand, M., 1961.— Body proportions of didelphid (and some other) marsupials, with emphasis on variability American

Journal of Anatomy, 109: 239-249.

Jenkins, F. A.. 1971 — Limb posture and locomotion in the Virginia opossum (Didelphis marsupialis) and in other non-cursorial

mammals. Journal of Zoology, London. 165: 303-315.

Jenkins. F.A.. 1974— Tree shrew locomotion and primate arborcalism. In: Primate Locomotion (F.A. Jenkins, ed ) Academic

Press: 85-115.

Jenkins. F ‘^^ 2 ^ learn ’ D ’ l9 84.— Mechanisms of hind foot reversal in climbing mammals. Journal of Morphology. 182 (2):

Jenkins, F.A.. & Parrington, F.R.. 1976.—The postcranial skeletons of the Triassic mammals Eozostrodon . Megazostrodon and

Erythrotherium. Philosophical Transactions of the Royal Society of London. B. 273: 387-431.

Jenkins. F.A.. & Schaff. C.R.. 1988.— The Early Cretaceous mammal Gobiconodon (Mammalia, Triconodonta) from the

Cloverly Formation in Montana. Journal of Vertebrate Paleontology. 8(1): 1-24.

Jenkins, F.A.. & Weijs, W.A., 1979.— The functional anatomy of the shoulder in the Virginia opossum ( Didelphis virginiana)

Journal of Zoology. London, 188: 379-410.

Julien-Laferriere, D.. 1991.— Organisation du peuplement dc marsupiaux en Guyane frangaise. Revue d'Ecologie (Terre Vie),

46 : 125-144.

154

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Kielan-Jaworowska, Z., 1975.— Possible occurrence of marsupial bones in Cretaceous eutherian mammals. Nature. 255:

698-699.

Kielan-Jaworowska. Z., 1977.— Evolution of the therian mammals in the Late Cretaceous of Asia. Part II. Postcranial skeleton

in Kennalestes and Asioryctes. Palaeontologia Polonica. 37: 65-83.

Kielan-Jawo#wska. Z.. 1978.— Evolution of the therian mammals in the Late Cretaceous of Asia. Part III. Postcranial skeleton

in Zalambdalestidae. Palaeontologia Polonica, 38: 3-41.

Kielan-Jaworowska. Z., & Gambaryan. P.P.. 1994.— Postcranial anatomy and habits of Asian multituberculate mammals.

Fossils and Strata, 36: 1-92.

Krause. D.W.. & Jenkins. F.A.. Jr. 1983.-The postcranial skeleton of North American multituberculatcs. Bulletin of the Museum

of Comparative Zoology , Harvard. 150 (4): 199-246.

Krebs. B.. 1987.—The skeleton of a Jurassic cupantothere and the arboreal origin of modern mammals. In: Mesozoic Terrestrial

Ecosystems (P.J. Currie and E.H. Koster, cds). Occasional Papers of the Tyrrell Museum of Paleontology. (9):

132-137.

Krebs. B.. 1991.— Das Skelett von Henkelotheriumguimarotaegc n. et sp. nov. (Eupantotheria. Mammalia) ausdem Obercn Jura

von Portugal. BerlinergeowissenschaflicheAbhandlungen, A. 133: 110 p.. Berlin.

Mckenna. M.C.. 1975.— Toward a phylogenetic classification of the Mammalia. In: Phytogeny of the Primates (W.P. Luckett

and F.S. Szalay. eds.). Plenum Press. New York: 21-46.

Mcmanus, J.J.. 1970.— Behavior of captive opossum. Didelphis marsupialis virginianus. The American Naturalist, 84: 144-169.

Marshall. L.G., 1977.— Lestodelphis halli Mammalian Species Series, The American Socien of Mammalogists. 91: 1-4.

Marshall. L.G., Case, J.A., & Woodburne, M.O.. 1989.— Phylogenetic relationships of the families of marsupials. Current

Mammalogy. 2: 433-502.

Marshall. L.G.. & Muizon, C. de, 1988.—The dawn of the age of mammals in South America. National Geographic Research ,

4(1): 23-55.

Marshall, L.G.. & Muizon. C. de. 1995.— Part II: the skull. In: C. de Muizon (ed.). Pucadelphys andinus (Marsupialia,

Mammalia) from the early Paleoccne of Bolivia. Memoires du Museum national d 'Histoire naturelle , 165:21 -90, Paris.

Marshall, L.G., Sempere, T.S., Butler, R.F.. in prep.— Chronology of the mammal-bearing Paleocene of South America.

Journal of South American Earth Sciences.

Osgood. W.H., 1921.— A monographic study of the American marsupial Caenolestes. Field Museum of Natural History , Zoology

series. 14: 1-156.

Prasad. G.V.R., & Godinot. M.. 1994.— Eutherian tarsal bones from the Late Cretaceous of India. Journal of Paleontology, 68

(4): 892-902.

Reig. O.A.. Kirsch. J.A.W., & Marshall, L.G., 1987.— Systematic relationships of the living and Mesozoic opossum-like

marsupials (suborder Didclphimorphia), with comments on the classification of these and of the Cretaceous and

Paleocene New World and European metatherians. pp. 1 -89. In: Possums and Opossums (M. Archer, cd.). Royal

Zoological Society of New South Wales, Sydney, Australia.

Rougier. G.W.. Wible. J.R.. & Hopson. J.. 1992.— Reconstruction of the cranial vessels in the early Cretaceous mammal

Vincelestes neuquenianus: implications for the evolution of the mammalian cranial vascular system. Journal of

Vertebrate Paleontology . 12 (2): 188-216.

Sempere. T.S., Butler, R.L.. Marshall, L.G., Sharp. W., & Swisher, C.. in press.— Chronostratigraphy of the Maastrichtian-

Slratigraphy and chronology of late Cretaceous-early Paleocene strata in Bolivia and northwest Argentina. Bulletin of

the Geological Society of America.

S hirer, 11.W..& Fitch, S.H.. 1970.—Comparison from radio-tracking of movements and denning habits of the racoon, the striped

skunk and the opossum in north-eastern Kansas. Journal of Mammalogy. 51: 491-503.

Slijper, E.J.. 1946.— Comparative biologic-anatomical investigations on the vertebral column and spinal musculature of

mammals. Verhandelingen der Koninklijke Nederlandse Akademie van Wetenschappen. Apfling Natuurkunde. 42:

1-128. Amsterdam.

Source MNHN, Paris

PUCADELPHYS AND IN US: POSTCRANIAL SKELETON

155

Szalay, F.S.. 1982a.— A new appraisal of marsupial phylogeny and classification, pp. 621 -640. In: Carnivorous Marsupials (M.

Archer, ed.). Royal Zoological Society of New South Wales. Sydney, Australia.

Szalay, F.S., 1982b.— Phylogenetic relationships of the marsupials. Geobios. Memoire Special 6: 177-190.

Szalay. F.S.. 1984.— Arboreality: is it homologous in metatherian and eutherian mammals? Evolutionary Biology. 18 (6):

215-258.

Tate, G.H.H., 1933.— A systematic revision of the marsupial genus Marmosa. Bulletin of the American Museum of Natural

History, 66: 1-250.

Walker, E.P., 1964.— Mammals of the World. The Johns Hopkins Press, Baltimore, vol.l: 1-644.

Wible, J.R.. & Hopson, J.A.. 1993.— Basicranial evidence for early mammal phylogeny. hr. Mammal phylogeny, Szalay et al.

cds.: 45-62.

Wooller, R.D.. Renfree, M.B., Russell. E.M.. Dunning, A., Green, S.W.. & Duncan, P., 1981.— Seasonal changes in a

population of the nectar-feeding marsupial Tarsipes spencerae (Tarsipedidae). Journal of Zoology, London, 195:

267-279.

156

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

APPENDIX: measurements

In the following tables, all measurements (in mm) are of specimens ol Pucadelphys. The abbreviation ca.

is used in the sense of “estimated”. But it should be made clear that all measurements are to be taken w,th caution,

the figures obtained depending on the orientation and focus choosen by the operator.

Dans les tableaux suivants, Unites les mesures (en mm) sent cellos de Pucadelphys andtnus

ca indique qu 'il s ’agit d ’une estimation. Mats toutes ces mesures doivent etre considerees aveeprudence, les

obtenues dependant en partie de /'orientation el de la mise au point choistes par I operateur.

Table 1. — Measurements of atlas and axis.

tableau 1 . — Mesures de Tatlas el de I 'axis.

specimen

6105

6110

Atlas (CVI):

1.3

Width of axoidian facet

1.5

Mxm length of arch

2.0

2.4

Ventral length

0.9

1.2

Axis (CV2):

Length of neural arch

4.1

4.Z

Mxm length

5.4

4.7

Height of dens

0.7

0.5

1.1

Width of dens

1.2

Height of neural arch

ca. 3.9

Mxm width at atlantal articulation

5.0

4.4

Width of tv.proc. (dorsal root) as preserved

8.0

2.4

Posterior width of centrum

2.8

Table 2. — Measurements of cervical vertebrae 3 to 7.

Tableau 2. — Mesures des vertebres cervicales 3 a 7.

specimen

6105

CV 3

Centrum:

Length

2.4

Anterior width

2.4

Posterior width

2.6

CV 4

Centrum:

Length

2.0

Anterior width

2.8

Posterior width

2.6

CV 5

Centrum:

Length

1.8

CV 6

Centrum:

Length

1.7

CV7

Centrum:

Length

1.65

Source: MNHN, Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

157

Table 3. — Measurements of thoracic vertebrae.

Tableau 3. — Mesures des vertebres thoraciques.

specimen

6105

6106

6110

6111

Centrum:

Length

2.3

2.4

Anterior width

2.9

1.9

Posterior width

2.3

2.0

Width of transverse processes

8.4

6.8

Centrum:

Length

2.4

Anterior width

2.3

Posterior width

2.8

Width of transverse processes

ca. 5.0

Centrum:

Length

2.5

2.7

Width of transverse processes

ca. 5.0

Centrum:

Length

2.75

2.7

Centrum:

Length

2.6

2.9

Centrum:

Length

2.55

2.8

2.2

Anterior width

ca. 1.6

Posterior width

ca. 1.8

Centrum:

Length

2.0

3.0

ca. 2.2

Anterior width

ca.1.6

Posterior width

ca. 1.7

Centrum:

Length

2.6

3.0

2.2

Anterior width

ca. 2.0

Posterior width

ca. 2.0

Centrum:

Length

ca. 2.8

3.0

2.3

ca. 1.9

Anterior width

ca. 2.1

Posterior width

ca. 2.2

TIO

Centrum:

Length

3.1

2.3

ca. 1.9

Til

Centrum:

Length

3.1

2.4

ca. 2.0

TI2

Centrum:

Length

3.2

2.45

2.4

TI3

Centrum:

Length

3.6

2.65

2.5

Source: MNHN, Paris

158

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Table 4. — Measurements of lumbar vertebrae.

Tableau 4. — Mesures des vertebres lombaires.

specimen

6106

6110

6111

Centrum: Length

3.7

3.0

2.5

Anterior width

3.0

2.0

Posterior width

3.6

ca. 2.4

Width of transverse processes

4.0

Centrum: Length

4.5

3.0

ca. 2.9

Anterior width

2.9

Posterior width

2.8

Width of transverse processes

5.0

Centrum: Length

4.9

3.8

ca. 3.3

Anterior width

2.8

4.0

Posterior width

3.0

Width of transverse processes

5.6

Centrum: Length

5.3

3.8

3.3

Anterior width

2.8

Posterior width

3.1

Width of transverse processes

6.5

Length of spinous process

4.00

Centrum: Length

5.0

3.5

3.7

Anterior width

3.0

Posterior width

4.5

3.1

Width of transverse processes

6.9

Length of spinous process

4.20

Centrum: Length

4.3

3.0

ca. 2.9

Anterior width

Posterior width

3.9

2.5

Width of transverse processes

8.1

Length of spinous process

4.90

Table 5. — Measurements of sacral vertebrae.

Tableau 5. — Mesures des vertebres sacrees.

specimen

6106

6110

6)11

S 1 + S 2

Length

7.4

6.5

5.5

S 1

Centrum: Length

3.4

3.1

ca. 2.5

Anterior width

2.5

Posterior width

3.0

2.3

ca 1.9

Width of transverse processes between pelves

10.0

8.4

6.6

S 2

Centrum: Length

4.0

3.4

ca. 2.5

Anterior width

3.0

ca. 2.3

Posterior width

3.0

Sacral foramen:

(r)

Length

1.0

1.05

Width

1.0

Source: MNHN , Paris

PUCADELPHYS AND/NUS: POSTCRANIAL SKELETON

159

Table 6a. — Measurements of caudal vertebrae (Cl - C9).

Tableau 6a. — Me sure s des vert eb res caudales (Cl - C9).

specimen

6106

6110

61 i 1

C 1

Centrum: Length

3.7

ca. 2.3

Anterior width

2.5

Posterior width

2.6

C 2

Centrum: Length

3.7

ca. 2.3

Anterior width

2.5

2.2

Posterior width

2.7

ca. 2.25

Width of transverse processes

ca. 10.0

C 3

Centrum: Length

3.5

2.8

Anterior width

Posterior width

2.1

Width of transverse processes

ca. 9.8

ca. 5.5

C 4

Centrum: Length

3.5

ca. 2.6

2.5

Anterior width

1.9

Posterior width

3.0

1.7

2.4

Width of transverse processes

9.0

ca. 6.6

5.6

Centrum: Length

4.0

ca. 3.5

3.3

Anterior width

2.4

1.9

Posterior width

2.1

.2.2

Width of transverse processes

ca. 5.4

6.0

C 6

Centrum: Length

5.5

4.3

Anterior width

2.15

2.1

Posterior width

2.9

2.5

Width of transverse processes

ca. 7.0

4.4

Centrum: Length

7.0

5.0

Anterior width

2.7

2.2

Posterior width

3.0

2.2

Width of transverse processes:

4.1

4.8

posterior

ca. 6.0

C 8

Centrum: Length

7.7

6.9

Anterior width

2.6

2.5

Posterior width

2.1

2.3

Width of transverse processes:

5.8

posterior

4.6

C 9

Centrum: Length

8.0

Anterior width

2.4

2.2

Posterior width

2.3

Width of transverse processes:

4.6

posterior

4.0

Source: MNHN . Paris

160

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Table 6b. — Measurements of caudal vertebrae (C16?, Cl 7?, C20?, C21 ?).

Tableau 6b. — Mesures des vertebres caudales (C16?, Cl 7?, C20?, C2J?).

specimen

6110

C 16?

Centrum:

Length

ca. 6.5

Anterior width

1.9

Width of transverse processes: anterior

2.9

C 17?

Centrum:

Length

7.0

Anterior width

1.4

Posterior width

1.4

C 20?

Centrum:

Length

6.1

Anterior width

1.4

Posterior width

1.6

C 21?

Centrum:

Length

6.4

Anterior width

1.3

Posterior width

1.2

Table 7. —

- Measurements of scapula.

Tableau 7.

— Mesures de 1 'omoplate.

specimen

6105

6106

(1)

(r)

Length

ca. 15.0

ca. 18.0

Maximum proximal width

3.8

4.0

Width of glenoid fossa

ca. 2.0

Table 8. —

Measurements of clavicle.

Tablmu 8. -

- Mesures de la clavicule.

specimen

6105

6110

(1)

(r)

Length

10.5

Width of sternal end

2.0 ca. 1.8

Width of scapular end

1.1

1.2

Median width

0.9

0.8

0.7

Source: MNHN. Paris

PUCADELPHYS AND/NUS: POSTCRANIAL SKELETON

161

Table 9. — Measurements of interclavicle.

Tableau 9. — Mesures de Pinterclavicule.

specimen

6105

Total length

8.9

Length anterior to costal tuberculae

3.1

Length posterior to costal tuberculae

5.0

Width of end cephalic

1.2

caudal

1.4

Width between costal tuberculae

3.7

Table 10. — Measurements of humerus.

Tableau 10. — Mesures de l’humerus.

specimen

6105

6106

(1)

6110

0) (r)

Length

17.5

20.7

ca. 16.3 17.0

Transverse width proximal end

3.8

Ant.-post, width proximal end

3.7

3.75 3.4

Transverse width distal end

6.3

6.0

4.5

Table 11. —

Tableau 11.

Measurements of ulna.

— Mesures du cubitus.

specimen

6105

6106

6110

new specimen

(1)

(r)

Length of olecranon + sigmoid cavity

ca. 4.6

5.0

Length of sigmoid cavity

1.9

1.8

2.5

Depth of olecranon at mid-length

1.6

2.1

2.0

Depth of sigmoid cavity at mid-length

1.9

1.7

Table 12. — Measurements of radius.

Tableau 12. — Mesures du radius.

specimen

6105

6110

(I)

(1)

Length

ca. 14.0

Width of proximal articular surface

1.65

Minimum width of diaphysis

1.1

Source: MNHN, Paris

162

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Table 13. — Measurements of pelvis.

Tableau 13. — Mesures da bassin.

specimen

Length (ant. edge of ilium to post, edge of ischium)

Width between iliac wings (ant. width)

Width between lateral edges of ischium (post, width)

Length of pubic symphysis

Length of ilium

Length of ischium

Width of post, end of ischium

Obturator foramen Length

Width

6106

6110

(1)

6111

(1) (r)

24.0

18.0

17.0

ca. 10.0

13.0

ca. 10.0

10.0

ca. 10.0

10.0 ca. 9.0

ca.8.0

8.8 9.0

ca.4.5

4.5

6.8

4.2

Table 14 . — Measurements of os marsupium.

Tableau 14 . — Mesures de Vos marsupial.

specimen

6106

0) (r)

Length

9.3

Width of contact with pelvis

3.1 3.2

Table 15 . — Measurements of femur.

Tableau 15. — Mesures du femur.

specimen

6106

(1)

(r)

6110

(1)

(r)

(1)

6111

(r)

Length

25.0

24.5

ca. 20.0

21.0

19.3

19.0

Proximal transverse width

5.7

5.8

5.6

4.5

Distal transverse width

ca. 4.4

4.3

4.0

Distal ant.-post, width

3.6

3.5

ca.3.6

ca.3.8

3.1

Source: MNHN. Paris

PUCADELPHYS ANDINUS: POSTCRANIAL SKELETON

163

Table 16. — Measurements of tibia.

Tableau 16.— Mesures du tibia.

specimen

6105

(1)

6106

(1) (r)

6110

(1)

6111

(r)

Length

23.5

26.5

ca. 19.0

Proximal transverse width

3.7

3.8

4.0

Proximal ant.-post, width

3.3

3.4

3.5

ca. 3.1

Transverse distal width

3.3

ca. 3.1

2.2

Table 17. — Measurements of fibula.

Tableau 17. — Mesures du perone.

specimen

6105

(1)

6106

(I) (r)

6110

(1)

6111

(1)

Length

ca. 22.5

ca. 24.5

Proximal transverse width

ca. 3.0

2.6

Proximal ant.-post, width

1.9

Transverse distal width

2.3

2.4

2.0

Ant.-post, distal width

2.0

1.7

Table 18. — Measurements of calcaneuni.

Tableau 18. — Mesures du calcaneuni.

specimen

6106

(1) (r)

6110

(r)

6111

(1)

Length

5.3 5.6

4.2

ca. 4.0

Width of tuber

1.5 1.6

Distal width

4.0

Source: MNHN, Paris

164

LARRY G. MARSHALL & DENISE SIGOGNEAU-RUSSELL

Table 19. — Measurements of tarsals.

Tableau 19.— Mesures des tarsiens.

specimen

6105

(I)

6106

(r)

6110

(I)

Cuboid:

Length (dorsal length)

1.6

2.3

1.6

Distal width

1.65

1.5

Cuneiform IV:

Length (dorsal length)

Distal width

1.0

0.6

Table 20. —

Measurements of mctapodials.

Tableau 20. — Mesures des metapodes.

specimen

6105

6106

6110

(r)

(1)

Metapodial II:

ca. 6.5

Length

7.0

Proximal width

0.9

0.8

Distal width

1.5

ca. 1.25

Metapodial III:

Length

7.5

ca. 8.0

ca. 7.0

Proximal width

1.1

0.8

Distal width

1.5

1.4

Metapodial IV:

Length

ca. 7.5

8.0

ca 7.0

Proximal width

1.1

Distal width

1.6

Metapodial V:

Length

6.0

6.3

Proximal width

2.15

2.5

Distal width

1.2

Proximal phalanx:

3.4

Source: MNHN, Paris

Remerciements aux rapporteurs

acknowledgements to referees

La Redaction tient a remercier les experts exterieurs au Museum national d’Histoire naturelle dont les noms

suivent, d’avoir bien voulu contribuer, avec les rapporteurs de l’Etablissement, a revaluation des manuscrits

(1986-1995):

The Editorial Board acknowledges with thanks the following referees who, with Museum referees, have reviewed

papers submitted to the Memoires du Museum (1986-1995):

ADKISON D.

Macon

U. S. A.

KRANTZ G.W.

Corvallis

U. S. A.

AKESSON Bertil

Goteborg

Sufcde

KUDENOV Jerry D.

Alaska

U. S. A.

AMIARD Jean-Claude

Nantes

France

LAGARDfcREJ.P

La Rochelle

France

ANDRES H

Hambourg

Allemagne

LANA Paulo Da Cunha

Parana

Bresil

BABA K.

Kumamoto

Japon

LAUBIER Lucien

Paris

France

BACHELEF Guy

Arcachon

France

LAVERDE-CASTILLO JJ.A.

Bogota

Colombie

BAUD C.A.

Geneve

Suisse

LETENDRE L

Courbevoie

France

BELLAN Gc-rard

Marseille

France

LEGAY J.M.

Villeurbanne

France

BEN-ELIAHU Nechama

Jerusalem

Israel

LEVIN Lisa A.

La Jolla

U. S. A.

BERGGREN M

Fiskebackskil

Su&de

MACKJE Andrew

Cardiff

Grande-Bretagne

Espagnc

U. S. A.

BERNET-ROLLANDE M.C.

Puteaux

France

MACPHERSON E

Barcelona

BERNOTL

Anthony

France

MANNING R.

Washington

BHAUD Michel

Banyuls-sur-Mer

France

MARSHALL B

Wellington

Nouvelle-Zelande

BLAKE James A.

Woods-Hole

U. S. A.

MAUCHLINE J.

Oban

Grande-Bretagne

U. S. A.

BOURDON R.

Roscoff

France

MAURER Don

Long Beach

BOURLI&RE F.

Paris

France

MCALPINE J.F.

Ottawa

Canada

BOUROULLEC J

Pau

France

MCKENNA M

New York

U. S. A.

BRESSON F.

Paris

France

MCLAUGHLIN P

Washington

U. S. A.

BROSSET A.

Paris

France

METTAM Chris

Cardiff

Grande-Bretagne

BUTMAN Cheryl Ann

Woods-Hole

U. S. A.

MILLAR R.H.

Oban

Grande-Bretagne

CALDE D.

Toronto

Canada

MUIR Alexander Ian

Londres

Grande-Bretagne

CASTELLI Alberto

Modena

Italie

NAGEL P.

Saarbriicken

Allemagne

CHAREST P.

Quebec

Canada

NEWMAN W. A.

San Diego

U. S. A^

CLARK P.

Londres

Grande-Bretagne

NOEL R.

Pau

France

COMBES C

Perpignan

France

OLIVE Peter James William

Tyne

Grande- Bretagne

CORNELIUS P

Londres

Grande-Bretagne

PATERSON Gordon L.J.

Londres

Grande-Bretagne

DAVIE P

Brisbane

Australie

PATTERSON C.

Londres

Grande-Bretagne

DE BROYER C.

Bruxelles

Belgique

PAXTON Hannelore

North Ryde

Australie

DESBRUYERES Daniel

Brest

France

pErez FARFANTE I.

Washington

U. S. A.

DHAINAUT Andrg

Villeneuve d'Ascq

France

PERKINS Thomas H

Saint Petersburg

Nantes

U. S. A.

DORRESTEIJN Adriaan

Mayence

Allemagne

PERTHUISOT J.P

France

DREUX P

Paris

France

PETERSEN Mary E.

Copenhague

Danemark

DUCHENE Jean-Claude

Banyuls-sur-Mer

France

PE ITIBONE Marian H.

Washington

U. S. A.

DUPUIS Y.

Chatenay Malabry

France

PEYROT-CLAUSADE M

Marseille

France

EIBYE-JACOBSEN Danny

Copenhague

Danemark

PLEUEL Fredrik

Stockholm

Su6de

FAIN A.

Bruxelles

Belgique

POCKLINGTON Patricia

Halifax

Canada

FAUCHALD Kristian

Washington

USA

PONTIER J.

Villeurbanne

France

FISCHER Albrecht

Mayence

Allemagne

POOR G.

Victoria

Australie

FITZHUGH Kirk

Los Angeles

U.S.A.

PUIG H.

Paris

France

FLORET J.J.

Paris

France

PURSCHKE GUnter

Osnabruck

Allemagne

FOREY P.L.

Londres

Grande-Bretagne

PUTHZ V.

Schlitz

Allemagne

FOURNIER Judith

Ottawa

Canada

RAMIL F.

Vigo

Espagne

FRANCOIS Y

Paris

France

REISH Donald J.

Long Beach

r c * v

U. S. A.

FRANSEN C.

GAGNE R.

Leiden

Hollande

RICHER DE FORGES B

Noumea

Nouvelle-Caledonie

Washington

U. S. A.

RIEMAN F.

Brcmerhaven

Allemagne

U. S. A.

GAMBI M. Cristina

G£HU J.M

Napoli

Italie

ROUSE Greg

Washington

Bailleul

France

SAN MARTIN Guillermo

Madrid

Espagne

U. S. A.

GENTIL Frank

Roscoff

France

SARDA Rafael

Blanes

GEORGE David

Londres

Grande-Bretagne

SAVAGE D.E.

Berkeley

GIANGRANDE Adriana

Lecce

Italie

SCHMID M.

Paris

France

GIBBS Peter E.

Plymouth

Grande-Bretagne

SCHROEDER Paul

Pullmann

U. S. A.

GILLET Patrick

Angers

France

SCOTT A C

Surrey

Grande-Bretagne

GLASBY Chris

Canberra

Australie

SIBUET Mvriam

Brest

France

GLEMAREC Michel

Brest

France

S1GVALDADOTTIR Elin

Stockholm

Su&de

GOERKE Helmut

Bremerhavcn

Allemagne

SIMON Joseph L.

Tampa

U S. A.

GOODAY A. J

Surrey

Grande-Bretagne

STORK N.E.

londres

Grande-Bretagne

GRASSHOFF M.

Frankfurt

Allemagne

TAYLOR P D.

I-ondres

Grande-Bretagne

GRASSLE Frederick

New Brunswick

Canada

THURSTON M.H.

Surrey

Grande-Bretagne

GRASSLE Judith

New Brunswick

Canada

TOULMOND Andre

Paris

France

GRUET Yves

Nantes

France

TRICART J.

Strasbourg

France

GUILLAUMET J.L.

Caen

France

VACELET J.

Marseille

France

HAIG J.

Los Angeles

U.S.A.

VAN AMERON H.W.J.

Krefeld

Allemagne

Hollande

HARDEGEJorg Detelf

Oldenburg

Allemagne

VAN SOEST R. W. M.

Amsterdam

HAYWARD P.J.

Swansea

Grande-Bretagne

VOKES E.

New Orleans

U. S. A.

HENSLEY D A.

Puerto Rico

U. S. A.

VOVELLEJean

Paris

France

HILBIG Brigitte

Massachusetts

U. S. A.

VUILLEUMIER F.

New York

U. S. A.

HOLTE Boerge

Tromsoe

Norv^ge

WAGELE J. W.

Bielefeld

Allemagne

HOLTHUIS L.B.

Leiden

Hollande

WARREN Lynda

Cardiff

Grande-Bretagne

HOOPER J.N.A.

Brisbane

Australie

WATSON J

Essendon

Australie

HOVE Harry Ten

Amsterdam

Pays-Bas

WESTHEIDE Wilfried

Osnabruck

Allemagne

U. S. A.

HUTCHINGS Patricia

Sydney

Australie

WILLIAMS A.

Washington

JOUIN-TOULMOND Claude

Paris

France

WILSON Robin

Victoria

Australie

KENDALL Michael

Plymouth

Grande-Bretagne

W ITEM ANN K

Vienne

Autriche

KIELAN-JAWOROWSKA Z.

Oslo

Norvege

ZEVTNA G. B.

Moscou

Russie

KNIGHT-JONES Phyllis

KNIGHT-JONES Wyn

Swansea

Grande-Bretagne

ZIBROWIUS Helmut

Marseille

France

AC H E V E DIM PRIMER

EN NOVEMBRE 1995

SUR LES PRESSES

l’imprimerie F. PAILLART

A ABBEVILLE

Date de distribution : 27 novembre 1995.

Depot legal : Novembre 1995.

N° d'impression : 9389.

Source: MNHN. Paris

Source: MNHN , Paris

DERNIERS TITRES PARUS

RECENTLY PUBLISHED MEMOIRS

A parlir de 1993 (Tome 155). les Memoires du Museum sont publies sans indication de serie.

From 1993 (Volume 155), the Memoires du Museum are published without serial titles.

Tome 164 : Jeanne Doubingkr. Pierre Vetter, J. Langiaux, J. Galtier & Jean Broutin, 1995. — La flore

fossile du bassin houiller de Saint-Etienne. 358 pp. (isbn : 2-85653-218-7) 479,92 ff.

Tome 163 : Alain Crosnier (ed.), 1995. — Resultats des Campagnes Musorstom. Volume 13. 518 pp.

(isbn : 2-85653-224-1) 550 ff.

Tome 162 : Jean-Claude Dauvin, Lucien Laubier & Donald J. Reish (eds), 1994. — Actes de la

4 e Conference internationale sur les Polychetes. 642 pp. (isbn : 2-85653-214-4) 390 ff.

Tome 161 : Alain Crosnier (ed.), 1994. — Resultats des Campagnes Musorstom. Volume 12. 569 pp.

(isbn : 2-85653-212-8) 600 ff.

Tome 160 : Nicole Boury-Esnault, Maurizio Pansini, & Maria Jesus Uriz, 1994. — Spongiaires

bathyaux de la mer d'Alboran et du Golfe ibero-marocain. 174 pp. (isbn : 2-85653-213-6) 300 ff.

Tome 159 : Pierre Robbe, 1994. — Les Inuit d'Ammassalik, Chasseurs de l'Arctique. 389 pp. (isbn :

2-85653-270-1) 360 ff.

Tome 158 : Alain Crosnier (ed.), 1993. Resultats des Campagnes Musorstom. Volume 11. 426 pp.

(isbn : 2-85653-208-x) 500 ff.

Tome 157 : LoTc Matile, Judith Najt & Simon Tillier (eds), 1993. — Zoolosia Neocaledonica.

Volume 3. 218 pp. (isbn : 2-85653-205-5) 280 ff.

Tome 156 : Alain Crosnier (ed.), 1993. — Resultats des Campagnes Musorstom. Volume 10. 491 pp

(isbn : 2-85653-206-3) 580 ff.

Informations sur les Publications Scientifiques du Museum national d'Histoire naturelle :

Informations about the Scientific Publications of the Museum national d'Histoire naturelle:

Internet http://www.mnhn.fr/

Prix hors taxe, frais de port en sus. Vente en France : tva 2,10 %.

Prices in French Francs, postage not included.

Larry G. Marshall is Associate Professor at the University of Arizona (Tucson, USA). He has

revised most of the South American marsupial collections. He began research in Bolivia in 1981 and also

did field work in Argentina, Chile and Brazil.

Christian de Muizon is Director of the Institut Franijais d’Etudes Andines (Lima, Peru) and a

research scientist at the Centre National de la Recherche Scientifique (Paris). He has been working on the

South American land-mammal faunas, especially in Bolivia, since 1980.

Denise Sigogneau-Russell is a research scientist at the Centre National de la Recherche

Scientifique and works at the Museum national d'Histoire naturelle (Paris, France). She spent the last 15

years studying primitive mammals and is the leading French specialist on these groups.

The volume presents a detailed study of the oldest skeletons of didelphid marsupials, those of

Pucadelphys andinus. They are exceptionally well preserved and represent one of the major discoveries in

marsupial paleontology in the last decades. The specimens (several partial skeletons) come from the early

Paleocene of the Santa Lucia Formation at Tiupampa (Bolivia). The first part of the work gives a short

presentation of the locality of Tiupampa, its age, taphonomy and a mammalian faunal list. The second part

deals with the skull and dentition, and presents a survey of the cranial features of the early tribosphenic

mammals. The dental structure of Pucadelphys andinus favours its classification within the family

Didelphidae. The third part of the volume analyses the postcranial skeleton. The latter is compared with

that of the small Recent didelphids, thus permitting an interpretation of the way of life and locomotion

of Pucadelphys andinus.

EDITIONS

DU MUSEUM

57, RUE CUVIER

75005 PARIS

ISBN 2-85653-223-3

ISSN 1243-4442

180 FF TTC (France)

176,30 FF HT (Etranger)