
THE ORIGIN OF AFRO-ARABIAN ‘DIDELPHIMORPH’

MARSUPIALS

by JERRY J. HOOKER* , MARCELO R. SÁNCHEZ-VILLAGRA*� ,

FRANCISCO J. GOIN� , ELWYN L. SIMONS§ , YOUSRY ATTIA– and

ERIK R. SEIFFERT**
*Department of Palaeontology, The Natural History Museum, Cromwell Road, London, SW7 5BD, UK; e-mail: j.hooker@nhm.ac.uk

�Palaeontological Institute and Museum, Karl Schmid-Strasse 4, CH-8006 Zürich, Switzerland; e-mail: m.sanchez@pim.uzh.ch

�División Paleontologı́a Vertebrados, Museo de La Plata, Paseo del Bosque s ⁄ n, 1900 La Plata, Argentina; e-mail: fgoin@fcnym.unlp.edu.ar

§Department of Biological Anthropology and Anatomy, Duke University, and Division of Fossil Primates, Duke Primate Center, 1013 Broad Street, Durham,

North Carolina 27705, USA; e-mail: esimons@duke.edu

–Egyptian Geological Museum, Misr el Kadima, Ethar el Nabi, Cairo, Egypt

**Department of Anatomical Sciences, Stony Brook University, Stony Brook, New York 11794-8081, USA; e-mail: erik.seiffert@stonybrook.edu

Typescript received 16 February 2007; accepted in revised form 5 June 2007

Abstract: New specimens of Peratherium africanum from

Early Oligocene deposits of the Fayum, Egypt, provide key

information on the relationships of the species. These

include the first maxilla to be found and two additional

dentaries. The maxilla can be demonstrated to belong to

the same species as the holotype dentary by study of the

occlusal relationships of upper and lower molars. It can be

shown by several synapomorphies that P. africanum is the

sister species to European Bartonian–Rupelian Peratherium

lavergnense. P. africanum therefore belongs to the ‘didel-

phimorph’ family Herpetotheriidae, not to the peradecti-

morph family Peradectidae. The genus Qatranitherium,

previously erected for this species alone, is here synony-

mized with Peratherium. Comparison with ‘didelphimor-

phian’ taxa from early Paleogene deposits of South

America suggests more remote relationships, indicating an

origin for P. africanum by dispersal from Europe as origi-

nally envisaged. The more precise relationships deduced

here help to constrain the time interval for dispersal to

Afro-Arabia, probably during the earliest Oligocene sea-

level low.

Key words: palaeobiogeography, Peratherium, Eocene, Oli-

gocene, dispersal, mammal.

Extant marsupials are represented by approximately 331

living species distributed in the Americas and the Austral-

asian region (Kirsch et al. 1997; Wilson and Reeder

2005), but the group had a wider distribution in the past,

with fossils now reported from all continents. The marsu-

pial fauna of Africa is the least well known, with signifi-

cantly higher diversity reported from Antarctica (Goin

et al. 1999). Marsupials are only known to have existed in

Africa between the early Eocene and the early Oligocene,

and the short taxonomic history of these forms is already

controversial, as explained below.

The first reported discovery of an extinct Afro-Arabian

marsupial, and still the best known, was from Early

Oligocene deposits of Egypt (Bown and Simons 1984)

and subsequently named Peratherium africanum Simons

and Bown, 1984. Peratherium remains central to discus-

sions of marsupial biogeography, as its presence in Africa

suggests either (1) dispersal to Africa from Europe or (2)

dispersal from South America to Europe via Africa.

Slightly older Late Eocene strata have recently yielded a

possible second Egyptian marsupial of uncertain affinities

(Sánchez-Villagra et al. 2007). Crochet (1984) described

the species Garatherium mahboubii from Early Eocene

deposits of Algeria (see also Mahboubi et al. 1986), but

the marsupial affinities of this species were contested by

Gheerbrant (1995), who aligned G. mahboubii with di-

lambdodont members of Adapisoriculidae, a group of

enigmatic early Paleogene placentals known entirely from

dental remains (Gheerbrant 1991, 1995). McKenna and

Bell (1997) classified Garatherium within the Herpetothe-

riinae, a subfamily of the paraphyletic ‘Didelphidae’

assemblage currently being used by most marsupial tax-

onomists working on isolated teeth from the Paleogene of

Laurasia and Africa (Goin 1993). The genus Qatranitheri-

um was erected by Crochet et al. (1992) to include a

lower deciduous tooth (referred to dP3) from early Oligo-

cene deposits in Oman as well as the P. africanum mate-

rial described by Bown and Simons (1984). The only

other Paleogene African marsupial is Early Eocene Kasse-

rinotherium tunisiense Crochet, 1986, which is known by

two upper molars recovered from Chambi in Tunisia; this

species was referred to the Peradectidae by Crochet. Only

[Palaeontology, Vol. 51, Part 3, 2008, pp. 635–648]

ª The Palaeontological Association doi: 10.1111/j.1475-4983.2008.00779.x 635


two of these four described African marsupial species are

uncontested members of the group: Peratherium africa-

num and Kasserinotherium tunisiense.

In this paper we describe new specimens of Peratheri-

um africanum from an Early Oligocene locality in the

Jebel Qatrani Formation (Fayum Depression, northern

Egypt), including the first upper dentition known. These

specimens help to confirm the generic affinities to Per-

atherium of the species described by Simons and Bown as

P. africanum, and greatly increase our knowledge of the

dental morphology of this species.

Institutional abbreviations. BMNH, Natural History Museum,

London, UK (PD for Palaeontology, ZD for Zoology depart-

ments); CGM, Cairo Geological Museum, Egypt; DPC, Duke

University Primate Center, Durham, USA; MB Ma, Museum für

Naturkunde, Berlin, Germany; NMB, Naturhistorisches Museum

Basel, Switzerland; USTL, Université des Sciences et Techniques

du Languedoc (Montpellier II), France.

SYSTEMATIC PALAEONTOLOGY

‘DIDELPHIMORPHIA’ Gill, 1872

Family HERPETOTHERIIDAE Trouessart, 1879

(sensu Kirsch et al. 1997)

Genus PERATHERIUM Aymard, 1850

Type species. Didelphis elegans Aymard, 1846, Early Oligocene,

Ronzon, France.

Peratherium africanum Simons and Bown, 1984

Plate 1, figure 1; Text-figures 1A–C, 2–4; Tables 1–2

v. 1984 Didelphid marsupial; Bown and Simons,

pp. 447–449, fig. 2.

*v 1984 Peratherium africanus Simons and Bown,

pp. 539–547, figs 2–4.

v. 1992 Qatranitherium africanum (Simons and Bown);

Crochet, Thomas, Sen, Roger, Gheerbrant and Al-

Sulaimani, pp. 539–544.

? 1992 Qatranitherium aff. africanum (Simons and

Bown); Crochet, Thomas, Sen, Roger, Gheerbrant

and Al-Sulaimani, pp. 539–544, pl. 1.

Holotype. Right dentary fragment with P2–M3, roots of C1 and

P1 and alveoli for M4 (CGM 40236) from the upper Jebel Qatra-

ni Formation, Rupelian, Oligocene, Quarry M, Fayum Province,

Egypt.

Paratypes. Left dentary fragment with M2–3, part of M4 and

roots of M1 (CGM 40237) and left edentulous dentary fragment

(DPC 3120), from the same horizon and locality as the holo-

type.

New material. Right maxillary fragment with P2–M3, roots of P1

and parts of alveoli of C1 and M4 (DPC 16946); right dentary

fragment with P3–M4, 2 empty alveoli anterior to P3 and part of

the root of P1? (DPC 3820); and right dentary fragment with

corroded P3–M4 and partial roots for C1–P2 (DPC 8201), all

from the same horizon and locality as the holotype.

Revised diagnosis. Medium-sized Peratherium: length of

M3, 2.14 mm. M1)3 transversely elongate with deep ec-

toflexus, especially on M3 where it emarginates half the

width of the stylar shelf; weakly dilambdodont with meta-

cone much larger than paracone; large, mesially situated

stylar cusp B confluent with cusp A; very small stylar cusp

C on distal flank of cusp B; paracingulum interrupted

adjacent to paracone. Lower molars with: short talonid

with buccally situated cristid obliqua; hypoconulid large,

taller than the entoconid and close to lingual margin;

shallow distolingual area of talonid, after wear becoming

a curved ‘crest’ linking hypoconulid to entoconid (hypo-

conulid-entoconid talonid fold).

Description

Maxilla (Text-fig. 1A–C). Only the lingual wall of the canine

alveolus is preserved, indicating a typically large root. The

root(s) of P1 are completely obscured by matrix. P2)3 are simple

teeth with one main cusp and a cingulum that is interrupted on

the buccal and lingual walls, as is typical for herpetotheriids.

Distally on P2, the cingulum becomes cuspate. The crown is

badly damaged on P3 and yields no useful information. All three

preserved molars have very deep ectoflexi, giving the weak elon-

gate stylar cusp D (StD) an oblique orientation that is nearly

transverse on M3. M1 is damaged, lacking protocone and mesio-

buccal corner, and abrasion generally has obscured stylar cusp

details. M2)3 show more detail. Stylar cusp B (StB) is large and,

although worn, is essentially confluent with stylar cusp A (StA,

parastyle), with no groove separating them on the buccal wall.

Stylar cusp C (StC) is very small and situated on the distal slope

of StB. The mesiobuccal corner of M3 is tilted buccally of a me-

siodistal axis. M1)3 are weakly dilambdodont, the bowing of the

centrocrista impinging little on the stylar shelf, which remains

wide despite the deep ectoflexus. The metacone is much larger

than the paracone and the preparacrista joins StA on all three

teeth (indicated on M1 by its orientation). On M2)3, the proto-

cone lobe tapers lingually and the paraconule and metaconule

are weak. On all three molars, the paracingulum is interrupted

adjacent to the paracone. The infraorbital foramen was relatively

large, its posterior border situated at the level of the anterior

border of M1. A comparative sample of infraorbital foramen size

in small didelphids, a caenolestid and several herpetotheriids

(Table 1), shows that the Fayum specimen is in the uppermost

part of the range. We used for this small survey the M2 as surro-

gate for size. Comparison with the available herpetotheriids

indicates to us that Peratherium is characterized by larger infra-

orbital foramina than either Amphiperatherium Filhol, 1879 or

Herpetotherium Cope, 1873.

636 P A L A E O N T O L O G Y , V O L U M E 5 1


Dentary DPC 3820 (Text-figs 2B, 3B, E). This dentary has teeth

that are less worn than those in the holotype (Text-figs 2A, 3A,

D, Pl. 1, fig. 1) and that are slightly larger (Table 2). The appar-

ent size difference is, however, exaggerated in two ways. Firstly,

the greater wear of the holotype has removed the mesially pro-

jecting part of the paraconids and the distally projecting part of

the hypoconulids, slightly shortening each tooth. Secondly, suc-

cessive teeth show more overlap than in DPC 3820. Thus, in

A 

iof 

B 

C 

D E 

TEXT -F IG . 1 . Maxillae of Peratherium. A, light micrograph. B–E, scanning electron micrographs of gold palladium-coated epoxy

casts. A–C, Peratherium africanum, right maxilla with P2–M3, DPC 16946, upper Jebel Qatrani Formation, Rupelian, Oligocene,

Quarry M, Fayum Province, Egypt. D, Peratherium antiquum, left maxilla with M2)3, BMNH.PD.M2388d, Phosphorites du Quercy,

Late Eocene ⁄ Early Oligocene, Quercy, France. E, Peratherium lavergnense, holotype left maxilla with M1)3, USTL.Lav-1031,

Phosphorites du Quercy, Bartonian, Eocene, Lavergne, France. A–B, lateral ⁄ buccal, and C–E, occlusal views; iof, infraorbital foramen.

Scale bar represents 1 mm.

H O O K E R E T A L . : O R I G I N O F A F R O - A R A B I A N ‘ D I D E L P H I M O R P H ’ M A R S U P I A L S 637


DPC 3820, the hypoconulid of one molar essentially abuts the

paraconid of the succeeding molar, whereas in the holotype the

hypoconulid is overlapped by the mesial tip of the paraconid

of the succeeding molar. This crowding is a phenomenon of

increased wear (Crochet 1980, fig. 5). The relatively unworn

teeth in DPC 3820 show several features not visible on the

holotype. The hypoconulid is larger and taller than in any

other species of Peratherium; it is taller than the entoconid and

prominent in lingual view, not obscured by the paraconid of

the adjacent tooth in occlusal view. The heightening of the

hypoconulid has had the effect of raising the adjacent floor of

the talonid basin, giving it the appearance of a crest when

worn (see below). The posthypocristid of M2 in DPC 3820

appears slightly oblique in contrast to the holotype. However,

although there can be slight intraspecific variation for this fea-

ture, the apparent obliquity of the angle in DPC 3820 is

enhanced by buccal abrasion of the hypoconid. Another differ-

ence from the holotype is the more lingual terminus of the

cristid obliqua on the trigonid. This difference is more appar-

ent than real, and is largely due to the greater wear of the

holotype, exposing a lower level in the buccally dipping cristid

obliqua. M4 is badly corroded, but appears to show a deeper

hypoflexid than on the other molars and a more elongate talo-

nid. P3 is also badly corroded and has lost its talonid, creating

the false appearance of a diastema between P3 and M1. There

are two mental foramina, one below the mesial edge of M1, the

other below its hypoconid.

Dentary DPC 8201 (Text-figs 2C, 3C). This specimen is badly

corroded and little detail can be gleaned from the teeth. Never-

theless, the lower molars show the same large hypoconulid and

vague representation of other features visible in DPC 3820. M4

is broken distally, but the talonid appears broader than in DPC

TABLE 1 . Infraorbital foramen size (in mm) in a sample of small didelphids, a caenolestid (Lestoros) and herpetotheriids, including

Peratherium africanum.

Specimen INF/M2LW Height

InfF

Width

InfF

InfFHxW

(INF)

M2

length

M2

width M2LxM2W

Gracilinanus marica BMNH.ZD. 98.7.124 0.10 0.67 0.34 0.23 1.33 1.71 2.27

Gracilinanus marica BMNH.ZD. 17 0.11 0.77 0.31 0.24 1.3 1.71 2.22

Herpetotherium fugax MB Ma 50672 0.11 1.2 0.39 0.47 1.95 2.19 4.27

Marmosops madescens BMNH.ZD. 27.11.189 0.11 1.15 0.41 0.47 1.71 2.47 4.22

Amphiperatherium

ambiguum

BMNH.PD. M2388a 0.12 1.58 0.46 0.73 2.31 2.69 6.21

Herpetotherium fugax MB Ma 50671 0.12 1.09 0.41 0.45 1.81 2 3.62

Marmosops madescens BMNH.ZD. 23.10.1646 0.12 0.92 0.56 0.52 1.79 2.32 4.15

Marmosops madescens BMNH.ZD. 27.11.194 0.16 1.02 0.59 0.60 1.7 2.18 3.71

Gracilinanus marica BMNH.ZD. 98.7.125 0.16 1.11 0.39 0.43 1.44 1.84 2.65

Peratherium sp. (Branssat) NMB 9537 0.18 1.76 0.6 1.06 2.39 2.5 5.98

Lestoros inca BMNH.ZD. 19.4.381 0.18 0.91 0.55 0.50 1.66 1.7 2.82

Monodelphis scalops BMNH.ZD. 21.8.61 0.19 0.97 0.46 0.45 1.41 1.65 2.33

Monodelphis scalops BMNH.ZD. 14.5.919 0.19 1.16 0.42 0.49 1.66 1.53 2.54

Lestoros inca BMNH.ZD. 22.11.127 0.23 0.86 0.68 0.58 1.61 1.59 2.56

Peratherium sp.

(Egerkingen)

NMB 1913 0.26 1.57 0.55 0.86 1.65 2.03 3.35

Peratherium antiquum NMB 25 0.26 1.79 0.91 1.63 2.14 2.92 6.25

Peratherium africanum DPC 16946 0.27 1.65 0.82 1.35 1.94 2.54 4.93

Peratherium sp. (Quercy) NMB QH 362 0.29 2.04 1.01 2.06 2.32 3.02 7.01

Peratherium sp. (Quercy) NMB QH249 0.30 1.83 0.67 1.23 1.72 2.38 4.09

Peratherium sp. (Branssat) NMB 2955 0.39 2 1.29 2.58 2.35 2.8 6.58

Inf, infraorbital foramen; INF, infraorbital foramen size; L, length; W, width.

TABLE 2 . Measurements of length versus width (in mm) of

teeth of Peratherium africanum; parameters follow Clemens

(1966, p. 4).

No. Tooth Length Width Width

trigonid

Width

talonid

DPC16946 P2 1.60 0.71

P3 2.01 1.08

M1 (1.95) -

M2 1.94 2.54

M3 2.14 2.87

DPC3820 M1 1.70 1.09+ 1.14

M2 2.20 1.21 -

M3 2.20 1.28 1.20

M4 2.30+ 1.15+ -

DPC8201 M1 1.77 1.17+ 1.36

M2 1.90 1.24+ 1.29+

M3 1.95 1.34+ 1.24+

M4 2.10+ 1.23+ 0.98

638 P A L A E O N T O L O G Y , V O L U M E 5 1


3820, with a shallower hypoflexid. DPC 8201 shows more pos-

terior parts of the dentary than the other specimens, including

the opening of the dental canal and beginning of the medially

inflected angle. There are two mental foramina, one below the

junction of P1 and P2 and another below M1.

OCCLUSAL RELATIONSHIPS

The great disparity in size between the small paracone

and the large metacone plus the weakly buccally flexed

centrocrista of M1)3 fit well with the buccally positioned

cristid obliqua and extreme lingual position of the hypo-

conulid of M1–3 (Text-figs 1–2; Pl. 1). The teeth in the

maxilla occlude well with the holotype (Text-fig. 4) and

DPC 3820 lower dentitions, although all three specimens

are clearly from different individuals. Thus, the maxilla

appears slightly too large for the holotype and slightly too

small for DPC 3820. In fact, gerontic crowding of the

molars in the holotype may be responsible for the slight

size discrepancy with the maxilla, but DPC 3820 is genu-

inely slightly larger (Table 2). In particular, for both spec-

imens the obliquity of the upper molar postmetacrista

matches that of the lower molar paracristid, obliquity

lessening distally to a proportionate degree. Likewise, the

near transverse orientation of the upper molar preparacri-

sta matches that of the lower molar protocristid. Also, the

lengths of the upper molar postparacrista and lower

molar cristid obliqua and of the upper molar premetacri-

sta and lower molar posthypocristid respectively corre-

spond closely. Enlargement of the teeth from M1)3 also

corresponds to that of M1–3.

FAMILY AFFINITIES

Simons and Bown (1984) made comparisons between P.

africanum and the three genera of Euramerican Paleogene

herpetotheriids then known (at that time placed in the

family Didelphidae), viz. Herpetotherium from North

America and Peratherium and Amphiperatherium from

Europe. (N.B. Crochet’s interpretation of North American

species attributed to Peratherium as belonging to Herpeto-

therium is followed here). They concluded that the great-

est similarity was with Peratherium; hence their generic

placement of the species africanum. They noted that there

were special similarities to P. cuvieri (Fischer, 1829), P.

antiquum (Blainville, 1840) and P. perrierense Crochet,

1979 on the basis of short lower molar talonids shared

with the first two, shallow hypoflexids and distal molar

size increase shared with P. antiquum and P. perrierense,

and mesially positioned entoconid and its connection to

the metaconid shared with P. cuvieri and P. perrierense.

A 

B 

C 

TEXT -F IG . 2 . Scanning electron micrographs of occlusal views of gold palladium-coated epoxy casts of right lower dentitions of

Peratherium africanum, upper Jebel Qatrani Formation, Rupelian, Oligocene, Quarry M, Fayum Province, Egypt. A, CGM 40236,

holotype with canine, P2–M3, roots of P1. B, DPC 3820, P3–M4, alveoli of P2. C, DPC 8201, P3–M4, roots of canine to P2. Scale bar

represents 1 mm.

H O O K E R E T A L . : O R I G I N O F A F R O - A R A B I A N ‘ D I D E L P H I M O R P H ’ M A R S U P I A L S 639


Crochet et al. (1992) in contrast removed P. africanum

from the Herpetotheriidae (as herpetotheriine Didelphi-

dae, the then current classification) and placed it instead

in the family Peradectidae on the basis of features such as

high molar entocristids, a buccally situated cristid obli-

qua, and a cuspate hypoconulid linked to the entoconid

by a crest. They added a DP3 (with an ‘aff. africanum’

qualification) from Taqah, Sultanate of Oman, which had

the same characters as the Fayum type material, and

placed the species in a new genus, Qatranitherium. There

is certainly some doubt as to whether the DP3 belongs to

the same taxon as the type assemblage, especially as it is

rather large, as Crochet et al. (1992) noted. Nevertheless,

the characters as displayed by the Fayum material do not

support referral of P. africanum to the Peradectidae for

the following reasons.

1. The upper molars in the new maxilla are clearly herpe-

totheriid in their dilambdodonty and enlarged metacone

A 

B 

C 

D 

E 

TEXT -F IG . 3 . Scanning electron micrographs of gold palladium-coated epoxy casts of right lower dentitions of Peratherium

africanum, upper Jebel Qatrani Formation, Rupelian, Oligocene, Quarry M, Fayum Province, Egypt. A, D, CGM 40236, holotype with

canine, P2–M3, roots of P1. B, E, DPC 3820, P3–M4, alveoli of P2. C, DPC 8201, P3–M4, roots of canine to P2. A–C, buccal, and D–E,

lingual views. Scale bar represents 1 mm.

640 P A L A E O N T O L O G Y , V O L U M E 5 1


and they occlude very well with the holotype lower

dentition.

2. The ‘crest’ linking the lower molar hypoconulid to the

entoconid is not the lingual branch of the postcristid

as in Peradectes, but a non-homologous fold, which

describes a buccally convex arc across the distolingual

part of the talonid basin, leaving a distinct postentoconid

sulcus. It is referred to here (see below) as the hypoconu-

lid-entoconid talonid fold (Pl. 1, figs 1–3).

3. The appearance of an entocristid is enhanced by the

advanced wear state and the proximity of the entoconid

to the metaconid, which is a feature of certain species of

Peratherium (see below).

4. A buccally positioned cristid obliqua is typical not

only of Peradectes but also of several species of Perathe-

rium.

5. The lower molar hypoconulid is situated near the

lingual margin, whereas it is positioned more buccally in

peradectids.

COMPARISONS

South American metatherians

It has been suggested that there are special similarities

between some marsupials from the Paleocene Itaborai

fauna in Brazil and Peratherium, and between isolated

postcranials from the same locality and similar remains

from North American and English Eocene deposits (Hoo-

ker 1998, p. 442). If South America was the source for

European species, Africa may have served as the interven-

ing landmass in their dispersal route. For that reason, it

is of interest to compare the Fayum marsupials with

South American forms.

Several Paleogene South American taxa show similari-

ties to Peratherium africanum. For instance, the middle

Paleocene (Tiupampan SALMA) Mizquedelphys pilpinensis

Marshall and de Muizon, 1988 also shows upper molars

that are progressively shorter from M1 to M3, although in

Mizquedelphys the V-shaped centrocrista is less obvious,

the paraconule and metaconule are present, the ectoflexus

is shallower, StC is present, and the preparacrista always

ends buccally at the lingual slope of StB (Muizon 1991).

These same features also apply when P. africanum is com-

pared with Pucadelphys andinus Marshall and de Muizon,

1988 from Tiupampa. Furthermore, StA and StC in Puca-

delphys are quite distinct and the paracingulum, though

short, is wider. On lower molars, talonids are longer than

trigonids (which in M2–4 have mesiodistally compressed

trigonids), the entoconid is proportionally small and even

across M1–3, and the paraconid is reduced.

Andinodelphys cochabambensis Marshall and de Muizon,

1988 has a straight centrocrista, well-developed paracon-

ule and metaconule, and anterobasal cingulum which is

contiguous with the preparaconule crista, well-developed

ectoflexus and quite distinct StA, StB, StC and StD.

Finally, the tiny Jaskhadelphys minutus Marshall and de

Muizon, 1988, also from Tiupampa, clearly departs from

the pattern observable in the African species: in the upper

molars, the protocone body is oddly twisted relative to

the orientation of the rest of the tooth, the paracone and

metacone are more similar in size and set closely to one

another, the ectoflexus is shallow, and all stylar cusps are

distinct.

Among the very rich marsupial fauna of Itaboraian

(Late Paleocene) age, several taxa also show superficial

similarities with Peratherium from the Fayum: Marmosop-

sis Paula Couto, 1962, Sternbergia Paula Couto, 1970,

Derorhynchus Paula Couto, 1952, Monodelphopsis Paula

Couto, 1952, and Itaboraidelphis Marshall and de Muizon,

1984. However, none of them shows the distinct set of

features characterizing the African species. Lower molars

of Marmosopsis, though similar to Peratherium in its trig-

onid-talonid proportions, have notably compressed pre-

molars and molars with much reduced paraconids, small

TEXT -F IG . 4 . The upper (DPC 16946, thick line) and lower (holotype CGM 40236, thin line, drawing reversed) right cheek

dentitions of Peratherium africanum superimposed essentially at full occlusion, as seen in the direction of relative movement (Butler

1972). Broken areas hatched, missing parts of outline reconstructed with dashed lines. Scale bar represents 1 mm.

H O O K E R E T A L . : O R I G I N O F A F R O - A R A B I A N ‘ D I D E L P H I M O R P H ’ M A R S U P I A L S 641


entoconids and talonids that are much lower than trigo-

nids. Upper molars of Marmosopsis have much more buc-

cally flexed centrocristae, small but distinct paraconule

and metaconule, and proportionally smaller StB. Lower

molars of Sternbergia are more slender and laterally com-

pressed, talonids are longer than trigonids, and the meta-

conids and paraconids are subequal in size and much

smaller than the protoconids. Upper molars have much

more buccally flexed centrocristae, a small but distinct

metaconule, wider protocone, shallower ectoflexus and a

distinct StC at least on M2.

Comparisons with Derorhynchus also show significant

differences: lower molars have better developed entoco-

nids and more approximated paraconids and metaconids,

whereas upper molars have distinct StA, StB, StC, StD,

paraconule and metaconule. In Monodelphopsis the lower

first molar has a very reduced metaconid, laterally

compressed entoconid and proportionally smaller hypo-

conulid.

Itaboraidelphys is much larger than the Fayum species,

has better developed pre- and postcingulids, and M4 has

a wider talonid. Despite the similarly buccally flexed

centrocristae, upper molars have proportionally larger

metaconules and clearly recognizable StA, StB, StC, StD,

and shallower ectoflexus.

Goin and Candela (2004) recognized a herpetotheriid

from ?Late Eocene deposits of Santa Rosa, Peru: Rumi-

odon inti Goin and Candela, 2004, which includes upper

and lower teeth. The upper molar differs from the M1 of

African Peratherium in having a wider protocone, shal-

lower ectoflexus, distinct StC and StD, a preparacrista

that is much shorter than the postmetacrista, and in the

lower molars, better developed precingulid, proportionally

longer trigonids, and a much more laterally compressed

entoconid.

Northern Hemisphere herpetotheriids

The African species also differs significantly from North

American herpetotheriids (Korth 1994) in the relative size

and position of the stylar cusps (StB is small and more

mesially placed in Herpetotherium and Copedelphys Korth,

1994, whereas StD is larger and progressively more me-

sially placed from M1 to M3 in these two taxa). Upper

molars have less size difference between paracone and

metacone and lower molars have the cristid obliqua less

buccally situated. On M1 the preparacrista joins StB.

Swaindelphys cifellii Johanson, 1996 from the Torrejo-

nian (middle Paleocene) differs from African Peratherium

in that the lower molars have wider talonids, proportion-

ally smaller hypoconulids, and upper molars have distinct

paraconule, metaconule and StB, C, and D, as well as a

shallower ectoflexus. Another North American herpeto-

theriid is Nortedelphys Case, Goin and Woodburne, 2005

from the late Cretaceous. Nortedelphys shows upper

molars with more buccally flexed centrocristae, distinct

StC on M1–M2, and proportionally shorter postmetacris-

tae, which in M2–M3 are more transverse to the dental

axis. Lower molars have a more approximated paraconid

and metaconid.

Maastrichtidelphys Martin, Case, Jagt, Schulp and Mul-

der, 2005, from the latest Cretaceous of Europe, has a

very distinct StC, a deeply flexed centrocrista, distinct

para- and metaconules and a very small, if present, StB.

There are several records of largely fragmentary herpe-

totheriids from Asia. Asiadidelphis Gabunia, Shevyreva

and Gabunia, 1990 from Early Oligocene deposits in

Kazakhstan has a large StC on M3)4 and twinned cusps C

and D on M2. It is therefore very similar to North Ameri-

can Herpetotherium, as Emry et al. (1995) judged. It there-

fore differs from P. africanum in the same features as

does Herpetotherium. An unnamed marsupial from the

Middle Eocene of Shanghuang, China, is said to be simi-

lar to Asiadidelphis (Qi et al. 1996). An unnamed herpe-

totheriid lower molar from early Eocene deposits in

northern Pakistan (Thewissen et al. 2001) has a relatively

longer talonid, more lingually orientated cristid obliqua

and lower more distally projecting hypoconulid than

P. africanum. Early Eocene Jaegeria from the Gujarat area

of India (Bajpai et al. 2005), based on a lower molar, has

a strong complete ectocingulid and a short, rather pro-

cumbent trigonid more reminiscent of a chiropteran than

of a herpetotheriid marsupial. An unnamed genus from

Early ⁄ Middle Eocene deposits of Turkey (Kappelman

et al. 1996, fig. 2) has lower molars rather like Pera-

therium but upper molars that are strikingly different.

M3)4 are strongly dilambdodont with paracone and meta-

cone of similar height, the centrocrista flexed to join StC

and a large stylar cusp between StB (stylocone) and StC.

Kappelman et al. (1996) identified this large stylar cusp as

StB, but it is preceded by a stylocone (StB), towards

which the preparacrista is directed, and more mesially

still by a parastyle (StA). These upper molars are probably

the specimens reidentified by Gheerbrant and Rage (2006,

p. 238) as the adapisoriculid Garatherium. None of these

Asian taxa shows any special similarities to P. africanum.

The European genus Amphiperatherium differs from P.

africanum and European species of Peratherium as follows

(modified from Crochet 1980). Successive molars show

less enlargement distally along the row. The upper molars

have less size difference between paracone and metacone,

small StB, and there is a tendency for paraconule and

metaconule to be stronger. Lower molars have weaker

postcingulids, and on M4 lingual hypoconulid without

reduction of entoconid. P3 is procumbent. As judged by

Simons and Bown (1984), the closest similarity within

Herpetotheriidae to P. africanum is the genus Peratherium.

642 P A L A E O N T O L O G Y , V O L U M E 5 1


This similarity is based on the following shared charac-

ters: molars enlarging distally along the row (except M4);

enlarged upper molar StB; upper molar metacone much

larger than paracone; deep M3 ectoflexus; lower molars

with relatively buccally situated cristid obliqua; erect

rather than procumbent P3.

European species of Peratherium

Derived species of Peratherium are fairly distinctive, but

more primitive ones from early in the Eocene are less eas-

ily distinguishable from Amphiperatherium. Thus, many

of the characters typically associated with Peratherium were

progressively acquired in the course of evolution during

the Eocene. According to the specific characters delin-

eated by Crochet (1980) and polarized here using the

primitive herpetotheriid genera Swaindelphis, Nortedelphis

and Maastrichtidelphis, Peratherium can be divided into

two main clades: clade A comprising P. monspeliense Cro-

chet, 1979, P. perrierense and P. cayluxi (Filhol, 1873),

and clade B comprising P. sudrei Crochet, 1979, P. bretou-

ense Crochet, 1979, P. lavergnense Crochet, 1979, P. cuvi-

eri, P. elegans and P. antiquum (Text-fig. 5). It is not

clear how the very primitive species P. constans Teilhard,

1927 and P. matronense Crochet, 1979 relate to the oth-

ers. Clade A is characterised by a tendency to enlarge

StD; the more derived species, P. perrierense and P. cay-

luxi, have relatively longer molars. Clade B shows a pro-

gressive acquisition of characters in a sequence of steps

through the middle and late Eocene. All except P. sudrei

are characterized by having upper molars with a weak pa-

racingulum adjacent to the paracone and lower molars

with the entoconid approaching the metaconid and a

buccally positioned cristid obliqua. P. lavergnense, P. cuvi-

eri, P. elegans and P. antiquum are linked by having lower

molars with short talonid and by modifications to the

upper molars, viz. weakening of StC, gaining a crest

between stylar cusps A and B, and even greater enlarge-

ment of the metacone at the expense of the paracone

(N.B. the genus Peratherium is characterised by having

greater size disparity between these cusps than are Amphi-

peratherium or Herpetotherium).

It is clear that the characters defining these four species

provide the best match for P. africanum. The P. cuvieri–

elegans–antiquum clade seems more distant from P. afri-

canum in having a slightly more buccally situated hypo-

conulid, no hypoconulid-entoconid talonid fold (both

primitive) (Pl. 1, fig. 4), a longer upper molar postmeta-

crista and a larger StB (both derived) (Text-fig. 1D). In

contrast, uniquely among European species, P. lavergnense

has a hypoconulid-entoconid talonid fold exactly like that

of P. africanum (Pl. 1, figs 1–3). P. lavergnense also has an

M1–3 hypoconulid that is very close to the lingual tooth

margin and an M3 parastyle that is directed mesiobuccal-

ly. P. lavergnense differs, however, from P. africanum in

having a shallower upper molar ectoflexus, although this

probably represents an autapomorphic reversal from the

moderately deep state that is primitive for the genus. The

TEXT -F IG . 5 . Cladogram of relationships of species of

Peratherium based on definitions given by Crochet (1980) plus

additional features observed herein, not on a parsimony analysis.

Apomorphies (excluding size) at numbered nodes: 1, StB largest

of the stylar cusps, deep M3 ectoflexus emarginating about one-

third of the width of the stylar shelf (reversed at node 11), erect

P3; 2, upper molar metacone nearly twice as tall as paracone; 3,

large StD; 4, molar length ⁄ width proportions increased such

that e.g. M1 is longer than wide, weak StC (paralleled at node

9); 5, M4 hypoconulid median; 6, lower molar paraconid nearly

as tall as metaconid; 7, M3 paracingulum weak adjacent to

paracone, lower molars with entoconid approaching metaconid,

buccally shifted cristid obliqua and metaconid transversely

opposite protoconid; 8, reduction in upper molar centrocristal

bowing; 9, upper molar metacone more than twice as tall as

paracone, crest joining StA to B, weak StC (paralleled at node

4), lower molar talonid shorter than trigonid; 10, M3 parastyle

directed mesiobuccally, lower molars with hypoconulid near

lingual margin and with hypoconulid-entoconid talonid fold; 11,

shallow M3 ectoflexus; 12, weak dilambdodonty, deep M1)2

ectoflexus, very deep M3 ectoflexus emarginating one-half of the

width of the stylar shelf, StA fused to B, StC tiny and close to B,

upper molar paracingulum interrupted, lower molar

hypoconulid taller than entoconid; 13, long upper molar

postmetacrista especially on M3 making the mesiobuccal angle of

the tooth outline obtuse, very large StB exceeding the paracone

in size; 14, StC fused with B in most individuals.

H O O K E R E T A L . : O R I G I N O F A F R O - A R A B I A N ‘ D I D E L P H I M O R P H ’ M A R S U P I A L S 643


deep M1)2 ectoflexus and even deeper M3 ectoflexus, the

tiny StC on the distal flank of StB, the fusion of B to A

and the tall lower molar hypoconulid can be construed as

autapomorphies of P. africanum. Available characters,

therefore, support a sister group relationship between P.

africanum and P. lavergnense. In light of this relationship,

the monotypic genus Qatranitherium Crochet, Thomas,

Sen, Roger, Gheerbrant and Al-Sulaimani, 1992 is judged

unnecessary and is here synonymized with Peratherium.

NATURE OF THE HYPOCONULID-
ENTOCONID TALONID FOLD

Crochet (1980) made no mention of this character in P.

lavergnense. This is probably because it is a relatively

minor structure, consisting really only of a shallowing of

the distolingual area of the talonid basin, and only

becoming obvious with dietary wear. It can nevertheless

be seen in the shading of Crochet’s figure 198b. The

structure may result from the extreme lingual shift of the

hypoconulid and its proximity to the entoconid, causing a

slight local modification of the talonid basin floor (Pl. 1,

fig. 3). In this respect the structure is not a crest, although

it is more distinct in P. africanum than in P. lavergnense

(see above). However, as wear progresses, the upper molar

postprotocrista, during the power stroke, makes a groove

from the mesiobuccal side of the entoconid to the buccal

side of the hypoconulid, isolating the shallow part of the

basin as an apparent crest (Pl. 1, figs 1–2). A similar

structure can be observed in some modern didelphids

such as Marmosops dorothea (Thomas, 1911) and Marmo-

sa murina (Linnaeus, 1758). Other closely related species

of Peratherium, such as P. cuvieri and P. elegans, neverthe-

less lack this structure, there being a distinct valley

between the hypoconulid and entoconid (Pl. 1, fig. 4).

DISCUSSION

Records of marsupials in Africa have the potential to

determine patterns of intercontinental exchange in this

group. From their origins in the Cretaceous of Asia or

North America, it was not until the Tertiary that they

dispersed to all other continents (Cifelli 1993; Rougier

et al. 1998). The early Paleogene faunas of Afro-Arabia

are dominated by endemics, mainly macroscelideans, hy-

racoids, proboscideans and enigmatic ‘insectivorans’. In

the last 15 years, new discoveries in Early–Middle Eocene

strata have added anthropoid (Godinot and Mahboubi

1992; Godinot 1994) and stem strepsirrhine (Hartenber-

ger and Marandat 1992) primates, putative plesiadapi-

forms (Tabuce et al. 2004), zegdoumyid rodents (Vianey-

Liaud et al. 1994), and a ?peradectid marsupial (Crochet

1986; Goin and Candela 2004), indicative of immigration

from Asia, Europe and perhaps South America. The Oli-

gocene record of the herpetotheriid marsupial Peratheri-

um is unequivocal evidence of dispersal from Europe, as

originally proposed by Simons and Bown (1984).

Dating the arrival of Herpetotheriidae in Afro-Arabia

The strata at Quarry M, the site of all the P. africanum

specimens, are within a normally polarised interval identi-

fied as Chron C11n (Seiffert 2006), which ranges in age

from 30.2 to 29.5 Ma (Ogg and Smith in Gradstein et al.

2004) in the late Rupelian, Early Oligocene. This equates

with Paleogene mammal Reference Level MP24 in Europe

(Luterbacher et al. in Gradstein et al. 2004). The sister

taxon of P. africanum, P. lavergnense, ranges in age from

late Bartonian (MP16) to early Rupelian (MP21) (Crochet

1980), thus from c. 37.5 to c. 33 Ma. If P. africanum is

actually derived from P. lavergnense (requiring some

reversals of its autapomorphies) following dispersal from

Europe to Africa, this is unlikely to have occurred later

than the last appearance of P. lavergnense at c. 33 Ma.

This is, therefore, much later than the Thanetian ⁄ Ypresian

date proposed by Gheerbrant and Rage (2006) based on

peradectid affinities. The most likely time for dispersal

would have been during the low sea level at the time of

the Oi-1 glaciation (c. 33.7–33 Ma) (Brinkhuis and Vis-

scher 1995; Luterbacher et al. in Gradstein et al. 2004). If

the Taqah record really represents P. africanum (or at

EXPLANATION OF PLATE 1

Scanning electron micrographs of gold palladium-coated epoxy casts of lower molars of Peratherium in bucco-occlusal views.

Fig. 1. Peratherium africanum, right M2–3 of holotype dentary, CGM 40236, upper Jebel Qatrani Formation, Rupelian, Oligocene,

Quarry M, Fayum Province, Egypt.

Figs 2–3. Peratherium lavergnense, Phosphorites du Quercy, Late Eocene ⁄ Early Oligocene, Quercy, France. 2, left M2–3,

BMNH.PD.M2388e1. 3, right M2–3, BMNH.PD.M2388e3.

Fig. 4. Peratherium cuvieri ⁄ elegans, Phosphorites du Quercy, Late Eocene ⁄ Early Oligocene, Quercy, France, right M1–2, BMNH.PD.M1636.

Abbreviations: entd, entoconid; hetf, hypoconulid-entoconid talonid fold; hyld, hypoconulid; tal, talonid basin; uw, unworn; w, worn.

644 P A L A E O N T O L O G Y , V O L U M E 5 1


PLATE 1

HOOKER et al., Peratherium

entd 

hetf(w) 

hetf(w) 

hetf(uw) 

valley 

tal 

1 

2 

3 

4 

hyld 


least a close relative), this would give the species a range

in time in Afro-Arabia back to c. 31.5 Ma (Seiffert 2006),

which still leaves a gap of 1.5 myr between the two

ranges. An earlier sea-level low at the Middle–Late Eocene

transition is an alternative possibility for the time of dis-

persal. In Africa, it marks the first appearance of anthra-

cotheriid artiodactyls and hystricognathous rodents

(Jaeger et al. 1985; Gheerbrant and Rage 2006). However,

there are no land mammal records from the preceding

Bartonian Stage. Moreover, the morphological diversity of

hystricognathous and anomaluroid rodents recorded at

the earliest Priabonian Birket Qarun Locality 2 (BQ-2)

(ERS, work in prep.) suggests that the dispersal of those

groups into Africa probably occurred long before the sea-

level lowstand at the Bartonian ⁄ Priabonian boundary.

Therefore, there is no firm evidence for a mammalian

dispersal event at the beginning of the Priabonian. Pre-

Oligocene Afro-Arabian faunas with micromammals con-

sistently lack herpetotheriids (e.g. Gheerbrant et al. 1998;

Seiffert et al. 2005; and references therein). However, the

close relationship of P. africanum and P. lavergnense

implies the existence of some form of intermediate in an

unsampled area in either southern Europe or northern

Afro-Arabia. In this context, it is pertinent that Iberia is

known to have been home to the microchoerine omo-

myid primate Pseudoloris as late as the Early Oligocene

(Reference level MP22), 2 myr after the family had

become extinct in the rest of Europe (Köhler and Moyà-

Solà 1999). Moreover, the presence of a rodent genus

in the Majorcan Oligocene, belonging to the otherwise

African family Thryonomyidae (Hugueney and Adrover

1991), indicates the existence of Afro-European dispersal

during this epoch.

Supporting evidence for general dispersal direction

also comes from the scattered records of pre-Eocene

herpetotheriids in North America and Europe: Swaindel-

phis, Nortedelphys and Maastrichtidelphys. Campanian–

Maastrichtian (latest Cretaceous) Nortedelphys also

appears to be the most primitive family member and

suggests a North American origin for the Herpetotherii-

dae (Case et al. 2005). Alternatively, Nortedelphys, Swain-

delphis and Maastrichtidelphys might be considered stem

‘didelphimorphians’. Either way, dispersal appears to

have taken place in three directions: (1) directly to

South America via emerging Central American island

arcs, leading to origin of other ‘didelphimorphian’ fami-

lies there (Case et al. 2005); (2) directly to Asia via the

Bering Straits (Emry et al. 1995); (3) to Afro-Arabia via

Europe. Dispersal to Europe was likely to have been in

two phases, once in the Maastrichtian (Martin et al.

2005), with no evidence of post-Cretaceous survival, and

a second time at the beginning of the Eocene (Hooker

1998). Arrival in Afro-Arabia seems to have been late,

no earlier than Early Oligocene.

Acknowledgements. We thank Scott Moore-Fay (NHM, London)

and P. Chatrath (Duke University) for fossil preparation, A. Zi-

ems and B. Engesser (Naturhistorisches Museum Basel) for

access to specimens, P. Chatrath for managing the Fayum field

camp and providing invaluable technical assistance, and A. Has-

sen Abd el Raouf, Z. El-Alfy, and F. Imbabi for supporting

ongoing work in the Fayum region. Recent fieldwork in the Fa-

yum area has been funded by Leakey Foundation grants to ERS

and US National Science Foundation grants to ELS and to ELS

and ERS. Comments by two anonymous referees have improved

the paper.

REFERENCES

A Y M A R D, A. 1846. Essai monographique sur un genre nou-

veau de mammifère fossile trouvé dans la Haute-Loire et

nommé Entelodon. Annales de la Société d’Agriculture, Sciences,

Arts et Commerce du Puy, 12, 227–267.

—— 1850. Réponse à M. Robert. Annales de la Société d’Agricul-

ture, Sciences, Arts et Commerce du Puy, 14, 80–86.

B A J P A I , S., K A PU R , V. V., T H E W I S S E N , J. G. M.,

T I W A R I , B. N. and D A S , D. P. 2005. First fossil marsupials

from India: Early Eocene Indodelphis n. gen. and Jaegeria n.

gen. from Vastan Lignite Mine, District Surat, Gujarat. Journal

of the Palaeontological Society of India, 50, 147–151.

B L A I N V I L L E , H. M. D. de 1839–64. Ostéographie, ou Descrip-

tion iconographique composée du squelette et du système dentaire

des Mammifères récents et fossiles pour servir de base à la Zoolo-

gie ou à la Géologie. Paris, 8 vols.

B OW N , T. M. and S I M ON S , E. L. 1984. First record of mar-

supials (Metatheria: Polyprotodonta) from the Oligocene in

Africa. Nature, 308, 447–449.

B R I N K H UI S , H. and V I S S CH E R , H. 1995. The upper

boundary of the Eocene Series: a reappraisal based on dinofla-

gellate cyst biostratigraphy and sequence stratigraphy. In

BE R G G R E N , W. A., K E N T, D. V. and H A R DE N B O L , J.

(eds). Geochronology, time scales and global stratigraphic corre-

lation. SEPM Special Publication, 54, 295–304.

B UT L E R , P. M. 1972. Some functional aspects of molar evolu-

tion. Evolution, 26, 474–483.

C A S E , J. A., G OI N , F. J. and W OO D BU R N E , M. O. 2005.

‘‘South American’’ marsupials from the Late Cretaceous of

North America and the origin of Marsupial Cohorts. Journal

of Mammalian Evolution, 12, 461–494.

C I F E L L I , R. L. 1993. Early Cretaceous mammal from North

America and the evolution of marsupial dental characters.

Proceedings of the National Academy of Sciences, USA, 90,

9413–9416.

C L E M E N S , W. A. 1966. Fossil mammals of the type Lance

Formation, Wyoming. Part 2. Marsupialia. University of Cali-

fornia, Publications in Geological Sciences, 62, 1–122.

C OP E , E. D. 1873. Third notice of the extinct vertebrata from

the Tertiary of the Plains. Paleontological Bulletin, 16, 1–8.

C R OC H E T , J.-Y. 1979. Diversité systématique des Didelphidae

(Marsupialia) Européens Tertiaires. Geobios, 12, 365–378.

—— 1980. Les Marsupiaux du Tertiaire d’Europe. Éditions de la

Fondation Singer-Polignac, Paris, 279 pp.

646 P A L A E O N T O L O G Y , V O L U M E 5 1


—— 1984. Garatherium mahboubii nov. gen., nov. sp., marsupial

de l’Eocène inférieur d’El Kohol (Sud Oranais, Algérie).

Annales de Paléontologie (Vert-Invert), 70, 275–294.

—— 1986. Kasserinotherium tunisiense nov. gen., nov. sp., troi-

sième marsupial découvert en Afrique (Eocène inférieur de

Tunisie). Comptes Rendus Hebdomadaires des Séances de

l’Académie des Sciences, Paris, 2, 302, 923–926.

—— T H O M A S , H., S E N , S., R O G E R , J., G HE E RB R A N T ,

E. and A L - S UL A I M A N I , Z. 1992. Découverte d’un Péra-

dectidé (Marsupialia) dans l’Oligocène inférieur du Sultanat

d’Oman: nouvelles données sur la paléobiogéographie des

Marsupiaux de la plaque arabo-africaine. Comptes Rendus

Hebdomadaires des Séances de l’Académie des Sciences, Paris, 2,

314, 539–545.

E M R Y , R. J., L UC A S , S. G., S Z A L A Y , F. S. and T L E U-

B E R DI N A , P. A. 1995. A new Herpetotheriine didelphid

(Marsupialia) from the Oligocene of Central Asia. Journal of

Vertebrate Paleontology, 15, 850–854.

F I L H O L , H. 1873. Sur les vertébrés fossiles trouvés dans les

dépôts de phosphate de chaux du Quercy. Bulletin de la Soci-

été Philomathique de Paris, 6, 10, 87.

—— 1879. Etude des mammifères fossiles de Saint-Gérand-le-

Puy. Annales des Sciences Géologiques, Paris, 10, 1–253, 30

pls.

F I S CH E R , J. B. 1829. Synopsis Mammalium. Stuttgart, 772 pp.

G A B UN I A , L. K., S HE V Y R E V A , N. S. and GA B U N I A , V.

DZH. 1990. On the presence of fossil marsupials (Marsupialia)

in Asia. Soobshcheniya Akademia Nauk, Georgian SSR, 116,

169–171. [In Georgian, English summary].

G H E E R B R A N T , E. 1991. Bustylus (Eutheria, Adapisoriculi-

dae) and the absence of ascertained marsupials in the Palaeo-

cene of Europe. Terra Nova, 3, 586–592.

—— 1995. Les mammifères Paléocènes du Bassin d’Ouarzazate

(Maroc) III. Adapisoriculidae et autres mammifères (Carniv-

ora, ?Creodonta, Condylarthra, ?Ungulata et Incertae sedis).

Palaeontographica, Abteilung A, 237, 39–132.

—— and R A G E , J.-C. 2006. Paleobiogeography of Africa: how

distinct from Gondwana and Laurasia? Palaeogeography,

Palaeoclimatology, Palaeoecology, 241, 224–246.

—— S U DR E , J., S E N , S., A BR I A L , C., M A R A N D A T , B.,

S I G É , B. and V I A N E Y - L I A UD , M. 1998. Nouvelles don-

nées sur les mammifères du Thanétien et de l’Yprésien du

Bassin d’Ouarzazate (Maroc) et leur contexte stratigraphique.

Palaeovertebrata, 27, 155–202.

G I L L , T. 1872. Arrangement of the families of mammals with

analytical tables. Smithsonian Miscellaneous Collection, 11

(230 ⁄ 1), 1–98.

G O DI N OT , M. 1994. Early North African primates and their

significance for the origin of Simiiformes (= Anthropoidea).

235–295. In F L E A G L E , J. G. and K A Y , R. F. (eds). Anthro-

poid origins. Plenum, New York, NY, 708 pp.

—— and M A HB O UB I , M. 1992. Earliest known simian pri-

mate found in Algeria. Nature, 357, 324–326.

G O I N , F. J. 1993. Living South American opossums are not

living fossils. Abstracts, Sixth International Theriological

Congress. Sydney, 112.

—— and C A N DE L A , A. M. 2004. New Palaeogene marsupials

from the Amazon Basin of eastern Perú. 15–60. In C A M P-

B E L L, K. E. (ed.). The Paleogene mammalian fauna of Santa

Rosa, Amazonian Peru. Science Series, Natural History

Museum of Los Angeles County, 40, 163 pp.

—— C A S E , J. A., W OO D BU R N E , M. O., V I Z C A Í N O ,

S. F. and R E G UE R O , M. A. 1999. New discoveries of

‘opossum-like’ marsupials from Antarctica (Seymour Island,

medial Eocene). Journal of Mammalian Evolution, 6, 335–

365.

G R A D S T E I N , F. M., O G G , J. G. and S M I T H , A. G. (eds)

2004. A Geologic Time Scale 2004. Cambridge University Press,

Cambridge, 589 pp.

H A R T E N B E R GE R, J.-L. and M A R A N DA T , B. 1992. A new

genus and species of an early Eocene primate from North

Africa. Human Evolution, 7, 9–16.

H O OK E R , J. J. 1998. Mammalian faunal change across the

Paleocene–Eocene transition in Europe. 428–450. In A U B R Y ,

M.-P., L UC A S , S. and B E R G GR E N , W. A. (eds). Late

Paleocene–Early Eocene climatic and biotic events in the marine

and terrestrial records. Columbia University Press, New York,

NY, 513 pp.

H UG U E N E Y , M. and A D R OV E R, R. 1991. Sacaresia moyae-

ponsi nov. gen. nov. sp., rongeur thryonomyidé (Mammalia)

dans le Paléogène de Majorque (Baléares, Espagne). Geobios,

24, 207–214.

J A E G E R , J.-J., D E N Y S , C. and C O I F F A I T , B. 1985. New

Phiomorpha and Anomaluridae from the late Eocene of

north-west Africa: phylogenetic implications. 567–588. In

L UC KE TT , W. P. and H A R T E N BE R G E R , J.-L. (eds).

Evolutionary relationships among rodents – a multidisciplinary

analysis. Plenum Press, New York, NY, 721 pp.

J OH A N S O N , Z. 1996. New marsupial from the Fort Union

Formation, Swain Quarry, Wyoming. Journal of Paleontolology,

70, 1023–1031.

K A PP E L M A N , J., M A A S , M. C., S E N , S., A L PA G U T , B.,

F OR T E L I U S , M. and L U N KK A , J.-P. 1996. A new early

Tertiary mammalian fauna from Turkey and its paleobiogeo-

graphic significance. Journal of Vertebrate Paleontology, 16,

592–595.

K I R S C H , J. A. W., L A P O I N T E , F. J. and S P R I N G E R ,

M. S. 1997. DNA-hybridization studies of marsupials and their

implications for metatherian classification. Australian Journal

of Zoology, 45, 211–280.

K ÖH L E R , M. and M O Y À - S O L À , S. 1999. A finding of Oli-

gocene primates on the European continent. Proceedings of the

National Academy of Sciences, USA, 96, 14664–14667.

K OR T H , W. W. 1994. Middle Tertiary marsupials (Mammalia)

from North America. Journal of Paleontology, 68, 376–397.

L I N N A E US , C. 1758. Systema naturae per regna tria naturae,

secundum classes, ordines, genera, species, cum characteribus,

differentiis, synonymis, locis. Tenth edition, vol. 1. Laurentii

Salvii, Stockholm, 824 pp.

M A HB O UB I , M., A M E UR , R., CR O C HE T , J.-Y. and J A E -

G E R , J. J. 1986. El Kohol (Saharan Atlas, Algeria): a new

Eocene mammal locality in northwestern Africa. Palaeonto-

graphica, Abteilung A, 192, 15–49.

M A R S H A L L, L. G. and M U I Z ON , C. de 1984. A new di-

delphid marsupial (Itaboraidelphis camposi nov. gen., nov. sp.)

from the Middle Paleocene (Itaboraian) of São José de Itabo-

H O O K E R E T A L . : O R I G I N O F A F R O - A R A B I A N ‘ D I D E L P H I M O R P H ’ M A R S U P I A L S 647


rai (Brazil). Comptes Rendus Hebdomadaires des Séances de

l’Académie des Sciences, Paris, 2, 299, 1297–1300.

—— —— 1988. The dawn of the age of mammals in South

America. National Geographic Research, 4, 23–55.

M A R TI N , J. E., CA S E , J. A., J A G T , J. W. M., S C HU L P, A. S.

and M UL DE R, E. W. A. 2005. A new European marsupial

indicates a Late Cretaceous high-latitude transatlantic dispersal

route. Journal of Mammalian Evolution, 12, 495–511.

M C KE N N A , M. C. and B E L L , S. K. 1997. Classification of

mammals above the species level. Columbia University Press,

New York, NY, 631 pp.

M UI Z O N , C. de 1991. La fauna de mamiferos de Tiupampa

(Palaeoceno Inferior, Formación Santa Lucı́a), Bolivia. 575–

624. In S UA R E Z - S O R UC O, R. (ed.). Fosiles y facies de

Bolivia, Vol. 1: Vertebrados. Revista Técnica de YPFB, Santa

Cruz (Bolivia), 718 pp.

P A UL A C OU TO , C. de 1952. Fossil mammals from the

beginning of the Cenozoic of Brazil. Marsupialia: Didelphidae.

American Museum Novitates, 1567, 1–26.

—— 1962. Didelfı́deos fósiles del Paleoceno de Brazil. Revista del

Museo Argentino de Ciencias Naturales ‘Bernardino Rivadavia’,

Ciencias Zoologicas, 112, 135–166.

—— 1970. News on the fossil marsupials from the Riochican of

Brazil. Anais de Academia Brasileira de Ciências, 42, 19–34.

Q I TAO, B E A R D , K. C., W A N G BANYUE, DA W S ON , M.

R., G UO JIANWEI and L I CHUANKUI 1996. The Shanghu-

ang mammalian fauna, Middle Eocene of Jiangsu: history of

discovery and significance. Vertebrata PalAsiatica, 34, 209–214.

R O UG I E R , G. W., W I B L E , J. R. and N O V A C E K, M. J.

1998. Implications of Deltatheridium specimens for early

marsupial history. Nature, 396, 459–463.

S Á N C H E Z - V I L L A G R A , M. R., S E I F F E R T , E. R., M A R -

T I N , T., S I M O N S , E. L., G UN N E L L , G. F. and A T TI A ,

Y. 2007. Enigmatic new mammals from the late Eocene of

Egypt. Palaeontologische Zeitschrift, 81, 406–415.

S E I F FE R T , E. R. 2006. Revised age estimates for the later

Paleogene mammal faunas of Egypt and Oman. Proceedings of

the National Academy of Sciences, USA, 103, 5000–5005.

—— S I M ON S , E. L., C L Y D E , W. C., R OS S I E , J. B.,

A T TI A , Y., B O W N , T. M., C H A T R A TH , P. and

M A T H I S O N , M. E. 2005. Basal anthropoids from Egypt

and the antiquity of Africa’s higher primate radiation. Science,

310, 300–304.

S I M O N S , E. L. and B O W N , T. M. 1984. A new species of

Peratherium (Didelphidae; Polyprotodonta): the first African

marsupial. Journal of Mammalogy, 65, 539–548.

T A B UC E , R., M A HB O UB I , M., T A F F OR E A U , P. and

S UD R E , J. 2004. Discovery of a highly-specialized plesiadapi-

form primate in the early–middle Eocene of northwestern

Africa. Journal of Human Evolution, 47, 305–321.

T E I L H A R D D E C H A R DI N , P. 1927. Les mammifères de

l’Eocène inférieur de la Belgique. Mémoires du Musée Royal

d’Histoire Naturelle de Belgique, 36, 1–33.

T H E W I S S E N , J. G. M., W I L L I A M S , E. M. and H US S A I N ,

S. T. 2001. Eocene mammal faunas from northern Indo-Paki-

stan. Journal of Vertebrate Paleontology, 21, 347–366.

T H OM A S , O. 1911. New mammals from tropical South

America. Annals and Magazine of Natural History, 8, 7, 513–

517.

T R O UE S S A R T , E. L. 1879. Catalogue des mammifères vivants

et fossiles. Insectivores. Revue et Magasin de Zoologie Pure et

Appliquée, Paris, 3, 7, 219–285.

V I A N E Y - L I A UD , M., J A E G E R , J.-J., HA R T E N B E R G E R ,

J.-L. and M A H BO UB I , M. 1994. Les rongeurs de l’Eocène

d’Afrique nord-occidentale [Glib Zegdou (Algérie) et Chambi

(Tunisie)] et l’origine des Anomaluridae. Palaeovertebrata, 23,

93–118.

W I L S ON , D. E. and R E E D E R , D. M. (eds) 2005. Mammal

species of the world. Third edition. Smithsonian Institution

Press, Washington, DC, 142 pp.

648 P A L A E O N T O L O G Y , V O L U M E 5 1