UNIVfcRSlTv CP

ILLINOIS I ^ARY

AT URBANA-CHAMPAIGN

L ZOLOGY

CO
CO

tjeotf**

t«R*«

7FIELDIANA
Geology

Published by Field Museum of Natural History

Volume 41, No. 2 December 29, 1978

The Deseadan, Early Oligocene,
Marsupialia of South America

Bryan Patterson

Professor of Vertebrate Paleontology Emeritus
Museum of Comparative Zoology, Harvard University

AND

Research Associate
Field Museum of Natural History

THE LIBRARY OF THE

AND

Larry G. Marshall

JUN07B84

Visiting Curator of Geology UNIVERSITY OF ILLINOIS

Field Museum of Natural History AT URBANA-CHAMPAIGN

ABSTRACT

Deseadan localities are briefly reviewed, and marsupials from Patagonian and
Bolivian local faunas compared. Ten named species of marsupials are known from
deposits of this age, four— Notogale mitis Ameghino, Pharsophorus lacerans Ame-
ghino, Pharsophorus! antiquus Ameghino, Proborhyaena gigantea Ameghino— in
the Borhyaenidae, and six— Pseudhalmarhiphus guaraniticus Ameghino, Palaeo-
thentes lucina Ameghino, P. boliviensis n. sp., P. chubutensis Ameghino, P. praecur-
sor Loomis, Parabderites minusculus Ameghino— in the Caenolestidae. In addition,
there are two unnamed species of Borhyaenidae and one new species of Polydolopi-
dae represented by material too fragmentary for formal description. The polydo-
lopid, from Bolivia, is the latest recorded member of that family. Notogale tenuis
Ameghino is regarded as a nomen vanum, and Pharsophorus tenax Ameghino is
placed in the synonymy of P. lacerans Ameghino.

Notogale mitis is a possible descendant of Patene and an ancestor of Sipalocyon
and, possibly, Cladosictis. A referred partial skull of Pharsophorus lacerans, from
Bolivia, exhibits several resemblances to the saber- tooth thylacosmilids. Phar-
sophorus? antiquus may possibly be involved in the ancestry of Arctodictis. The
gigantic Proborhyaena is the last recorded member of its subfamily. The Deseadan
marks the first appearance of the Caenolestidae. Pseudhalmarhiphus may be
ancestral to Stilotherium, Parabderites minusculus to P. bicrispatus, and several of
the species of Palaeothentes show resemblances to later species of this long-lived
genus. The absence of Didelphidae is a striking feature of the Deseadan fauna.

Library of Congress Catalog Card No.: 78-66785
ISSN 0096-2651

Publication 1292 37

38 FIELDIANA: GEOLOGY, VOLUME 41

INTRODUCTION

During Carlos Ameghino's seventh expedition to Patagonia,
1893-1894, he collected a Deseadan fauna from a locality or area
known as La Flecha in Santa Cruz Province, Argentina (fig. 1). This
site is situated a little distance to the south of the Inlet of the Rio
Deseado. The material collected during the expedition, which in-
cluded some marsupials, formed the basis of Florentino Ameghino's
first paper on the Deseadan (Pyrotherium) fauna, his "Premiere
Contribution ..." (Ameghino, 1895, pp. 603-606; Wood and Patter-
son, 1959, p. 283n). Several years later the French collector, Andre
Tournouer (1903) made a collection of Deseadan fossils from this
same locality (and elsewhere) and sent them to the Museum Na-
tionale d'Histoire Naturelle (MNHN), Paris. These collections were
studied by Gaudry, who in 1906, on the basis of the collection from
La Flecha proposed the name Deseado to replace Ameghino's faunal
name of Pyrotherium beds. La Flecha then, is the type locality of
the Deseadan (Wood and Patterson, 1959, p. 283n).

Between 1894-1896 Carlos Ameghino discovered and collected
from Deseadan beds at the Great Barranca south of Lago Colhue-
Huapi and at Cabeza Blanca. These materials formed the basis for
Florentino 's second contribution to the Deseadan fauna published
in 1897. Additional marsupial taxa from the Deseadan beds of Pata-
gonia were collected on several of Carlos' subsequent trips to these
areas and were described by Florentino in 1899 and 1903. Tournouer
(1903, p. 469) reported having collected a specimen of "Epanorthus"
(= Palaeothentes) and "un Carnivore didelphe" from La Flecha.
These specimens, now in the MNHN, are undescribed. F. B. Loomis
of Amherst College visited Patagonia in 1911 and made a large col-
lection of Deseadan mammals, including several good specimens of
marsupials, from Cabeza Blanca (Loomis, 1914).

Subsequent to that time various workers have discussed Dese-
adan marsupials in taxonomic studies. Included are accounts of
some Borhyaenidae (Cabrera, 1927; Chaffee, 1952), Abderitinae
(Marshall, 1976b), and Caenolestinae (Marshall, 1976a). More
recently Marshall (1978) has revised the borhyaenid subfamily Bor-
hyaeninae, and included description of specimens from Pico Trun-
cado and La Flecha made by the First Marshall Field Expedition to
Patagonia under the leadership of E. S. Riggs in 1922-1924.

In 1968 Hoffstetter mentioned the occurrence of marsupials in the
Salla (Deseadan) beds in the Salla-Luribay Basin- of Bolivia. Bor-

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 39

hyaenidae were reportedly represented by the greater part of a
cranium and associated mandible, tentatively referred to Proborhy-
aena, but to a form smaller than P. gigantea. Several teeth and other
fragments attested to the existence of at least two other genera, one
of medium size and the other smaller. Didelphids were reported as
present but poorly represented.

Hoffstetter et al. (1971) reported marsupials from Deseadan age
beds at Lacayani, Bolivia. These were represented by two mandibu-
lar fragments. More recently Hoffstetter (1976) reported that the
marsupials in the Salla fauna from the Salla-Luribay Basin and
Lacayani, included at least three genera of Borhyaenidae and pos-
sibly Didelphidae, although presence of the latter needed confirma-
tion. All specimens upon which these reports were based are in the
MNHN and none have as yet been described or illustrated.

In this paper we review all Deseadan Marsupialia in Argentinian
(MACN, MLP) and North American (AC, AMNH, FMNH) collec-
tions. We also include description of new materials from the Salla-
Luribay Basin, Provincia Loaza, Departamento La Paz, Bolivia in
Princeton University.

Abbreviations used in the text, figure captions, and tables of
measurements are as follows: C, canine; ca, approximate measure-
ment; D, deciduous; I, incisor; L, length; M, molar; P, premolar; W,
width.

AC Amherst College, Massachusetts

AMNH American Museum of Natural History, New York
FMNH Field Museum of Natural History, Chicago
MACN Museo Argentino de Ciencias Naturales "Bernar-
dino Rivadavia", Buenos Aires, Argentina
MLP Museo de La Plata, La Plata, Argentina
MNHN Museum Nationale d'Histoire Naturelle, Paris,

France
PU Princeton University, Princeton, New Jersey

All measurements are in millimeters (mm.). Higher taxonomic
categories for the Marsupialia follow those adopted by Clemens and
Marshall (1976).

We are indebted to Dr. Donald Baird for permission to study the
Salla collection in Princeton University. Illustration and publication
have been aided by the Gordon Barbour fund of Princeton Univer-
sity.

40 FIELDIANA: GEOLOGY, VOLUME 41

A generous grant (no. 1329) from the National Geographic Soci-
ety, Washington, D.C. made it possible for Marshall to work in
Argentina from September, 1974 to May, 1975 during which time
various aspects of this study were effected. Patterson's studies
there were made possible by John Simon Guggenheim Memorial
Foundation Fellowships.

For access to pertinent collections and other courtesies we wish to
thank Guillermo del Corro and Jose Gallardo, Museo Argentino de
Ciencias Naturales "Bernardino Rivadavia," Buenos Aires, and
Rosendo Pascual, Facultad de Ciencias Naturales y Museo de la
Universidad Nacional de La Plata, La Plata, Argentina. Figures 4
and 5 are by Bill Peterson, the rest are by Elizabeth Liebman.

LOCALITIES

The following is a list of Deseadan fossil localities (fig. 1).

1. La Flecha. 47° 54'S, 66° 15'W. Located about 15-20 km. upriver
and about 5 km. south of mouth of Rio Deseado, Santa Cruz Prov-
ince, Argentina. The fossil beds occur about 15 km. west of the
Estancia now known as Ocho de Julio, and just southwest of the
Puesto La Flecha, on the northern-most side of Lago Dulce. This is
the type locality of the Deseado Formation (Tournouer, 1903, fig. 4;
Bordas, 1945, profile 12; Feruglio, 1949, p. 60; Chaffee, 1952, p. 554,
fig. 9, Loc. 101; Wood and Patterson, 1959, p. 283n).

2. Cabeza Blanca ("Loomis Locality"). 45° 14' 04" S, 67° 29' 80"
W. This locality is a large peanut-shaped hill, about 5 km. southwest
of the Estancia Venter, east of the Rio Chico del Chubut, Chubut
Province, Argentina. From this locality have come most of the
Deseadan specimens in the Ameghino collection (MACN), and all
those of the Loomis collection (AC), and many of those in Field
Museum and the American Museum of Natural History (Simpson,
1948, p. 27, fig. 1, Loc. 6; 1967a, p. 69, fig. 3, Loc. 6; Chaffee, 1952, p.
555, fig. 9, Loc. 6).

3. Pico Truncado. 46° 50' 40"S, 68° 07' 20"W. Located in the large
amphitheater on the southeastern slope of Cerro Pico Truncado,
about 15 km. southwest of the Pueblo Pico Truncado, on the north

■Chaffee (1952, pp. 554-555, fig. 9) shows a locality which he calls "La Flecha" (his
Loc. 11) south and east of Pico Truncado, and north and west of Puerto Deseado.
The true La Flecha locality is the same as his Loc. 10— "Rio Deseado." Chaffee's
Loc. 11 could represent the Cerro Alto locality of Riggs (unpub. field notes), but this
is not certain.

Fig. 1. Map of Bolivia and parts of Argentina showing locations of mammal-bear-
ing fossil localities (circles) discussed in text.

41

42 FIELDIANA: GEOLOGY, VOLUME 41

side of the Rio Deseado, northern Santa Cruz Province, Argentina
(Chaffee, 1952, p. 555, fig. 9, Loc. 7; Simpson, 1948, p. 27, fig. 1,
Loc. 17; 1967a, p. 69, fig. 3, Loc. 17).

4. Great Barranca. 45° 42'S, 68° 42'W. Located about 3 km.
directly south of Lago Colhue-Huapi, Chubut Province, Argentina
(Simpson, 1948, fig. 1, Loc. 1; 1967a, fig. 3, Loc. 1).

5. Neuquen. The first discovery of a Deseadan mammal, Pyrother-
ium romeroi, was made west of the confluence of the rio Limay with
the Rio Neuquen to form the Rio Negro (Ameghino, 1906, pp. 90, 99,
fig. 24). So far as we are aware, no subsequent work has been done at
the locality.

6. Rinconada de los Lopez ("Scarritt Pocket"). 44° 40' S, 68° 25'
W. The Rinconada de los Lopez is on the west side of the Meseta
Canquel, between the Chubut Valley and the Colhue-Huapi Basin,
central Chubut Province, Argentina. The "Scarritt Pocket" is on
the southwestern portion of the Rinconada de los Lopez and on the
west side of the Sierra Canquel (Chaffee, 1952, p. 555, fig. 9, Loc. 1).

7. Paso de los Indios. Laguna de los Machos, Departamento Paso
de los Indios, Chubut Province, Argentina. This locality has yielded
a small Deseadan faunule (Patterson and Pascual, 1968, p. 7).

8. Laguna Payahile. 44° 30' S, 69° 00' W. Fossil beds of Deseadan
age occur around, and especially to the north of Laguna Payahile,
central Chubut Province, Argentina.

9. Curuzu Cuatia. This locality, in Corrientes, Argentina, has
yielded one specimen, the type of "Ameghinotherium curuzucua-
tiense" Podesta ( = Trachytherus spegazzianus Ameghino). As a
note of minor historical interest, the red sandstones in which this
fossil was found formed part of d'Orbigny's "Guaranitic Forma-
tion," a term improperly (by current standards) extended by
Ameghino to include a number of late Cretaceous and early Tertiary
rock units, including the Deseado.

10. Ingeniero Jacobacci. Rio Negro Province, Argentina. A little
known but evidently complex area as regards its Tertiary geology.
Volkheimer (1973) mapped sediments of this age as Colloncuran on
the basis of identifications of mammalian remains collected by him.
Rocks of this age certainly occur in the area, but others do also.
Specimens of Palaeopeltis inornatus ( = Orophodon hapaloides
Kraglievich and Rivas, 1951, nee Ameghino, 1895; Pseudorophodon
kraglievichi Hoffstetter, 1954) and of an archaeohyracid attest to
the presence of Deseadan deposits.

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 43

11. Salla. 17° 05" S, 67° 37" W. The locality is in the Salla-
Luribay Basin, about 90 km. southeast of LaPaz, in the Serrania of
Sicasica, at an elevation of about 4,000 m., Bolivia (see Hoffstetter,
1968; 1976, p. 7, fig. 2).

12. Lacayani. The locality is about 30 km. southeast of La Paz,
and about 50 km. northwest of Luribay, Bolivia (see Hoffstetter et
al, 1971; Hoffstetter, 1976, p. 7, fig. 2).

13. Tremembe. The formation of this name, in the Taubate Basin,
Sao Paulo, Brazil (Paula Couto and Mezzalira, 1971, p. 475, fig. 1),
has long been regarded as late Tertiary or Pleistocene on the basis of
fossil remains, mainly fishes, consistent with but not diagnostic of
such an age. A recent find of specimens of Leontinia (Paula Couto,
in Paula Couto and Mezzalira, 1971) indicates a Deseadan dating
(see also Patterson, in press).

For other possible, but at present dubious, localities for Deseadan
mammals see Chaffee (1952, pp. 554-555).

Marsupialia have been recorded only from the following: La
Flecha, Cabeza Blanca, Pico Truncado, Great Barranca, Rinconada
de los Lopez, Laguna Payahile, Salla, and Lacayani.

AGE OF THE DESEADAN

Conventionally, the Deseadan land mammal age is considered
Early Oligocene (e.g., Wood and Patterson, 1959; Patterson and
Pascual, 1972). Recently Marshall et al. (1977) reported that a basalt
conformably overlying rocks of Deseadan age at Pico Truncado
gave a radiometric date of 33.6 mybp (KA 2917), and another from
Cerro Blanco gave a date of 35.4 mybp (KA 2920). On the basis of
these two radiometric dates these authors tentatively accepted 34.0
mybp as a terminal date for known Deseadan.1 No basal dates for
known Deseadan are yet available and the time period covered by
this age is presently unknown.

Based on additional and younger radiometric dates from basalts
at the Great Barranca south of Lago Colhue-Huapi and at Pico
Truncado, Marshall et al. (1977) tentatively accept a date of 25.0
mybp as a basal age for known Colhuehuapian (Late Oligocene). The
paleontological hiatus between known Deseadan and known Col-
huehuapian is thus in the order of 9.0 my. Unfortunately, no

■For the earlier Tertiary we follow the epoch boundaries of Berggren and Van
Couvering (1974, fig. 1) who date the beginning of the Oligocene at 37.5 mybp.

09
TJ
08

P 2 S

S-H Q,

tf -3 J

§3

•J PL,

s *
o pa

a pq

-8

S

o §

.a 2

Ph h

J

W^O'tfi-ICOrHCNi-ll-li-ll-li-Hi-l

<M ,-, rH ,-H

T3

.s

II

< ft,

e g

o o

i o

ft, Oh

^ «J *~

03 a *r»
lo bo.S--3

e J i a

o 12 e e

a,* 8<

2 t 5 I

>« O «, 8

ft* PQ a, a,

o o
| |

ft. ft,

O ■

3 S

cj 03

CD

a bo

**" as

so C

ej .a

K C

CD CD

O O

c» CD

■I JS

a 03

ft, Ph

T3

«? s*

3 °°

w 5

co 03

3

C bC

s s

8 3
•C |

«d J3

13 O

Si*

ft, Ph

03

tJ

>>

+J

a

CO

03
U

2 &

CO

.2 ^

o

to ®

05

C M

^H

S CO

u

:C0

.— CD

3

CD CO

O

a

CO 0

=1
O

CD W
CO "

H

1 §

C

-fc»

o

O CD

X)

CO

CO

0)

en

2 *
o *

o

pa

* *

H

*

* *

44

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 45

radiometric dates are available for the Mustersan (conventionally
Medial Eocene).

Of the Patagonian marsupial faunas of Deseadan age, the largest
number of specimens (at least 16) are from the Cabeza Blanca, three
from the Rinconada de los Lopez, one from Pico Truncado, three
from La Flecha, one, possibly two, from Laguna Payahile, and one
from the Great Barranca (table 1). An analysis of the relative ages of
these local faunas, one to the other, based solely on comparison of
the marsupials, is not possible at this time. Despite the lack of data
needed for defining the relative ages of these faunas it is unlikely
"that anyone would be prepared to assert that all Deseadan local
faunas were exactly synchronous" (Patterson and Wood, in press).

The marsupial local faunas from Cabeza Blanca, Argentina and
Salla, Bolivia are the most diverse and permit comparisons of their
taxonomic differences and similarities. The most striking feature is
that the two most common borhyaenid species, Notogale mitis and
Pharsophorus lacerans, occur in both local faunas. If the species
Proborhyaena gigantea is also present in the Salla local fauna, as
suggested by two fragmentary maxillae, then three borhyaenid
species are shared.

Palaeothentes boliviensis is slightly smaller than, but is structur-
ally similar to P. chubutensis from Cabeza Blanca. The possibility
exists that P. boliviensis is ancestral to P. chubutensis, making the
fauna from Cabeza Blanca slightly later than that of Salla. It must
be stressed that such an ancestral-descendant relationship cannot
be demonstrated, although it does warrant consideration. It is of in-
terest, in this regard, to note that Hartenberger (1975), based on a
preliminary study of the rodents from these local faunas, also sug-
gested that the Salla local fauna was somewhat earlier than that of
Cabeza Blanca. A more detailed discussion of the relative ages of
Deseadan local faunas is presented by Patterson and Wood (in
press).

There is no doubt that the Bolivian faunas are referable to the
Deseadan Land Mammal Age and aside from the marsupials con-
tain mammalian species and genera inseparable from those in beds
of Deseadan age in Patagonia (see faunal list given by Hoffstetter,
1976). It would appear that during the Deseadan there were few if
any major barriers inhibiting dispersal of terrestrial mammals be-
tween Patagonia and Bolivia, areas separated by some 30° of
latitude.

46 FIELDIANA: GEOLOGY, VOLUME 41

SYSTEMATICS

Order Marsupialia Illiger, 1811

Superfamily Borhyaenoidea (Ameghino, 1894) Simpson, 1930

Family Borhyaenidae Ameghino, 1894

Subfamily Hathlyacyninae Ameghino, 1894

Notogale Loomis, 1914
Notogale Loomis, 1914, p. 216.
Type.—IPharsophorus mitis Ameghino, 1897.
Distribution.— Deseadan, Patagonia, and Bolivia.
Diagnosis.— As for the type and only known species.

Notogale mitis (Ameghino, 1897). Figures 2-7; Table 2

Wharsophorus mitis Ameghino, 1897, p. 504.
Notogale mitis Loomis, 1914, p. 216, figs. 142, 143.

Type.— MACN 52-368, a fragment of a mandibular ramus with
basal portions of M2.3.

Hypodigm.— The type of AC 3060, a fragment of a left mandib-
ular ramus with 1^ and M4 present; AC 3117, a left maxillary frag-
ment originally with part of M2, M3 incomplete, and M4 (M4 now
missing); PU 21867, a fragment of a right mandibular ramus with
talonid of M3, M4 complete; PU 21868, a fragment of a right man-
dibular ramus with M3 (missing tip of protoconid); PU 21869, a frag-
ment of a right mandibular ramus with roots of C, Pj.2, and anterior
root of P3; PU 21871, a fragment of a rostrum with roots (or alveoli)
of left I1'4, and base of right C, alveolus of left C, and roots of right
P1; PU 21872, a fragment of a left mandibular ramus with P2; PU
21874, a fragment of a left mandibular ramus with base of C, and
alveoli of IM; PU 21875, a fragment of a right maxillary with M1;
PU 21876, a fragment of a left maxillary with M1; PU 21877, greater
part of crown of C lacking enamel; PU 21993, a fragment of a right
mandibular ramus with bases of M3.4; and PU 21996, a fragment of a
right mandibular ramus with M2 (broken).

Localities.— The type is probably, and the AC specimens are cer-
tainly, from Cabeza Blanca, Chubut Province, Argentina. PU 21996
is from Branisa Loc. V-5, Salla-Luribay Basin; the other PU speci-
mens are from the general locality of Salla, Bolivia.

Age. —Deseadan.

Diagnosis.— Similar in size and structure to species of Sipalocyon
from beds of Colhuehuapian and Santacruzian age of Argentina.

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 47

Fig. 2. Notogcde mitis (Ameghino, 1897, p. 504). MACN 52-368 (type), a fragment
of a right mandibular ramus with basal portions of M2.3: a, lingual; b, occlusal; c,
labial views. Scale =10 mm.

I 4/3, C 1/1, P 3/3, M 4/4. Upper and lower PI separated from C by
small, but distinct, diastemata. Pl-3 oriented along same axis as
molar series, not set obliquely in jaw. P2 longer and narrower than
P3. P3 with large, distinct posterobasal cuspule, a feature only faint-
ly developed on P2. Protocone of M1 large, paracone about three-
fourths size of metacone; paracone and metacone approximated,

Fig. 3. Notogale mitis (Ameghino, 1897, p. 504). AC 3060, a fragment of a left man-
dibular ramus with P3 and M4 present: a, lingual; b, occlusal; c, labial views. Scale =
10 mm.

48

Fig. 4. Notogale mitis (Ameghino, 1897, p. 504). PU 21867, a fragment of a right
mandibular ramus with talonid of M3, and M4 complete: a, Ungual; b, occlusal; c,
labial views. Scale = 5 mm.

49

50

FIELDIANA: GEOLOGY, VOLUME 41

Fig. 5. Notogale mitis (Ameghino, 1897, p. 504). PU 21872, a fragment of a left
mandibular ramus with P2: a, occlusal; b, labial views. PU 21868, a fragment of a
right mandibular ramus with M3 (missing tip of protoconid): c, occlusal; d, labial
views. Scale = 5 mm.

united basally; parastyle small but distinct; stylar shelf virtually
absent; protocone of M3 large; shallow ectoflex present; metastylar
shear better developed than in M1. Lower molars with small but dis-
tinct anterobasal cingulum; metaconid absent; talonids of M3.4
distinct, narrower than trigonids; talonid of M3 with distinct hypo-
conid, entoconid, and hypoconulid; talonid of M4 essentially uni-
cusped; with enlarged hypoconulid dominant; minute entoconid and
hypoconid visible in unworn teeth (e.g., PU 21867).

Comments.— In the type (MACN 52-368), both teeth have lost all
traces of enamel, their bases are fractured and only a small portion
of the mandibular ramus surrounding them is preserved. Their posi-
tion in the series is not quite certain. Large alveoli can be seen
anterior and posterior to them, and their size and relative propor-
tions indicate that they are certainly molars. We have assumed that
the larger tooth is more posterior in the series. We identify these
teeth as M2.3 although they could possibly, but less probably, repre-

Fig. 6. Notogale mitis (Ameghino, 1897, p. 504). AC 3117, a left maxillary frag-
ment with posterior edge of M and M very broken: a, labial; b, occlusal; c, lingual
views. Scale = 5 mm.

51

Fig. 7. Notogale mitis (Ameghino, 1897, p. 504). PU 21875, a fragment of a right
maxillary with M : a, labial; b, occlusal; c, Ungual views. Scale = 5 mm.

52

I I I
I I I
I I I

CM O

i i i i

i r» oo

i i i i

<N I

oo !

i «i -j i

I 00

! co

S

I

«-■
o

-

lO 05

to <o

i i i i

i i i i

I co i

i i i i

^

s

I I I I

h oo oo oo

CO CM CM CM

ODDD
< a. cu cu

"aj SJ t- oo eo n (O oi

Jin0(0»01t>0)(0
j->7<DCO0OCT>00CJi0O

53

54 FIELDIANA: GEOLOGY, VOLUME 41

sent M^. The proportions and size of the roots of the type match
the alveoli of AC 3060, figured by Loomis (1914, p. 217, fig. 143).
The P2 has been lost since the specimen was figured by him. The two
remaining teeth are P3 and M4; the latter is clearly not M3 as
thought by Loomis, who further complicated matters by labelling it
M2. The M4 was probably isolated from the rest of the jaw when the
specimen was collected and was restored into the M3 position. We
have restored M4 in its correct position (fig. 3).

In the same box with the type of N. mitis (MACN 52-368) is an
isolated right P1 of a borhyaenid. This tooth measures 7.3 mm. in
length and 4.5 mm. in width. It is similar in size and structure to
specimens of Borhyaena tuberata from the Santa Cruz Formation.
Its preservation suggests, however, that it is from Cabeza Blanca
and is thus of Deseadan age. Ameghino makes no mention of this
tooth, and how it came to be associated with the type of N. mitis is
not known. This P1 is clearly not referable to N. mitis, but may
prove referable to a species of Pharsophorus, perhaps P. lacerans.

Notogale mitis is the most abundant species of Borhyaenidae in
the Salla fauna. It is a small borhyaenid with a rather generalized
dentition, suggesting that it may have been a fox-like carnivore. N.
mitis is similar in size to some large species of Pliocene-Pleistocene
Didelphidae (Marshall, 1978) and it is conceivable that the uncon-
firmed reports of didelphids in the Salla fauna by Hoffstetter (1968,
1976) were based on fragmentary remains of this species. N. mitis is
similar in size and structure to species of Sipalocyon from beds of
Colhuehuapian and Santacruzian age and may well be ancestral to
that genus, and/or to the larger Cladosictis.

Notogale tenuis (Ameghino, 1897) nomen vanum

IPharsophorus tenuis Ameghino, 1897, p. 504.
Notogale tenuis Loomis, 1914, p. 217.

Type.— MACN 52-387, an isolated upper? premolar.

Hypodigm.—Type only.

Locality.— No specific locality data, but probably from Cabeza
Blanca, Chubut Province, Argentina.

-Age. —Deseadan.

Comments.— Ameghino (1897, p. 504) believed this tooth to be a
third lower molar. It is, however, a premolar, probably an upper, but
which one is not certain. The crown is low and dentine is exposed on
the lingual? surface. The tooth measures 3.4 mm. long, 2.6 mm.

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 55

wide. Its features suggest possible affinity with other marsupial
groups, notably the Caenolestidae and more specifically the Palaeo-
thentinae. In any case, this tooth is clearly not of a borhyaenid and
the taxon is a nomen vanum.

Subfamily Borhyaeninae (Ameghino, 1894)
Cabrera, 1927

Pharsophorus Ameghino, 1897

Pharsophorus Ameghino, 1897, p. 502; Cabrera, 1927, p. 274; Simpson, 1948, p. 47.

Type.— Pharsophorus lacerans Ameghino, 1897, p. 503.

Distribution.— Deseadan, Patagonia, and Bolivia.

Diagnosis.— \\, C|, Pf, M|; comparable in size to species oiBorhy-
aena. Incisors small, crowded, I2 situated behind lv Lower pre-
molars increasing in size from Pj to P3. Pj implanted obliquely in
jaw. P23 with well-developed posterobasal heel well separated from
protoconid. P3 large, higher than and inclined toward Mx; protoconid
decidedly sloped posteriorly. Lower molars increasing gradually in
size from Mx to M4. Reduced but distinct talonid on MM, decreasing
in size relative to trigonid from Mx to M4. Talonid of Mj with
distinct cusp, in M2 small but basined, in M3.4 cuspate. Metaconid
lacking on M1# M2.4 with small metaconid and anterobasal cingulum;
metaconid increasing in size from M2 to M4. Large mental foramen
below P2. M1"4 with small but distinct protocone and parastyle. M2"3
and probably M1 with reduced but distinct ectocingulum.

Pharsophorus lacerans Ameghino, 1897. Figures 8-11; Tables 3-5.

Pharsophorus lacerans Ameghino, 1897, p. 503, figs. 79, 80; 1906, p. 352, fig. 183;
Loomis, 1914, p. 214, fig. 139; Cabrera, 1927, p. 274; Simpson, 1948, p. 48; Mar-
shall, 1978, p. 32.

Pharsophorus tenax Ameghino, 1897, p. 504; Loomis, 1914, p. 215, figs. 140, 141;
Marshall, 1978, p. 33, figs. 5, 6.

Type.— MACN 52-391, greater part of a left mandibular ramus
with roots of Ix and I3 and alveolus of I2, stump of C, roots of Plt P2
missing tip of crown, P3-Mj complete, and M2.4 present but partially
broken.

Type of P. tenax.— An isolated lower Mlf fide Ameghino. As the
type is apparently lost, Marshall (1978) selected AC 3004 (greater
part of a right mandibular ramus with roots of P2-Mj, and M2.4 pres-
ent but broken) as the neotype.

o -5

56

PATTERSON & MARSHALL: DESEADAN MASUPIALIA

57

Fig. 9. Pharsophorus lacerans Ameghino, 1897, p. 503. MACN 11652, a fragment
of a left mandibular ramus with M1-3: a, lingual; b, occlusal; c, labial views. Scale =
10 mm.

Hypodigm.— The types and neotype as above, and AC 3192, a left
maxillary fragment with complete P3 and M3'4, and broken M1"2;
MACN 52-388, base of a right lower canine; MACN 11652, a frag-
ment of a left mandibular ramus with M^; MACN 11653, a frag-
ment of a right mandibular ramus with posterior half of C alveolus,
alveoli of Plf roots of P2, P3 complete, roots of M1( M2.4 complete; PU
20551, medial portion of skull broken anterior to P2 and posterior to
postorbital process, with left and right P2-M* (all somewhat broken)

ss

If

li

-a "°
c ~

es es

a

-4->

a

CD

C

>

QJ

J*J

-'

o

a

x>

C8
03

S

hi

D

o

A

•o

03

es

°cV

s

'-3

O

O

CO

a

-u

o

3

o

u

"O

s

C
08

1— 1
kQ

a

a>

o

o

<M

s

P

.Q

cu

•4-9

cs

_c

c6
o

£

m

CD

6

ci,

o

CO

=3

00

-S

o*

s

.S«"

X &H

bo

-U

J3
be

<

"C

CO

T3

c

C

a

OS

c

ID

«*-C

o

i

J5

3

J3

o

'% .

« s

o.

03 O

0

8 »

a

2 li

-£

a.

S is co

O O to

58

n i ft t-

<o !

■»r o>
\6 co

» u ra K)
rH rH rH S

f; N »
00 00 t-^

*"< rH

rH O CO t"

■* 1- N N

rH I I rH rH Tf |> |

OS | 00 t-^ t> CO !

J -1

rH !

W NN H I

rH rH rH rH I

rH CN 00

ad oS oi

<N 05 I

t> co !

g

•8

00 H o

H © rH

rH O rH CO CN

M rH rH ^| rH rH

OS t>

to t-;

rH* O

rH CO

co <n !

eo

<N

t^

CO

co

OS

•*<

rH

rH rH I — —

I >A CO

s

CO IO """"J '=?

2° -His!

u

s

-391
652
653

*

^

CN rH rH

»

>o

1

lO rH rH

-r

CO

Z Z Z

c
o

X

o o u

CO

CM

B

< < <

O !~»

*

SSS<

(V

Oh

J

1

i

>.

Q>

— ;

■»■-

n

0^

O

CN
lO

-*•

Z

Id

(J

<

8
2

•

59

60 FIELDIANA: GEOLOGY, VOLUME 41

and roots of M2; and PU 21865, a fragment of a left mandibular
ramus with posterior root of M2 and M3.4 missing their lingual sides.

Localities.— The sites of collection of the types are not known,
although the color and preservation of MACN 52-391 suggest pro-
venance from Cabeza Blanca. MACN 11652 was collected at Cabeza
Blanca by O. Gutierrez and L. G. Marshall in February, 1975.
MACN 11653, is without locality data. The AC specimens were col-
lected at Cabeza Blanca. PU 20551 and 21865 are from Salla,
Bolivia and were collected by L. Branisa.

Age. — Deseadan.

Diagnosis.— Essentially as for genus, smaller than P.I antiquus;
similar in size to the Colhuehuapian Borhyaena macrodonta.

Pharsophorus tenax was considered by Marshall (1978, p. 33) to
be a valid species, although poorly characterized. The neotype is the
smallest in our mandibular series but, as shown in Table 3, the gap
in molar size between it and the type of P. lacerans, the largest indi-
vidual, is bridged by MACN 11652 and 11653. The size range is
comparable to those shown by other similar sized borhyaenids, such
as Borhyaena tuberata, for which good samples are on hand. On the
evidence available we are unable to detect any differences of taxo-
nomic significance between the Patagonian and Bolivian samples.

Description.— In PU 20551 the heel of P3 is narrower than that of
P3 of AC 3192. The palate is unfenestrated and there is a broad naso-
lacrimal contact, as in other borhyaenids. A large, nearly circular in-
fraorbital foramen opens over the middle of P3. The canine roots ex-
tend posteriorly to a point above the anterior edge of P3, resulting in
inflation of the maxillaries in that region. The nasals are restricted
transversely and the dorsal part of the rostrum is narrow trans-
versely. The supraorbital bars are blunt, but distinct, and large
rugose areas, marking sites of muscle attachment, occur on their
dorsal surfaces and extend posteromedially, uniting to form the
sagittal crest.

The lower incisors are small and crowded, their alveoli separated
only by thin films of bone. Ij and I3 are on a line and I2 is directly
behind I:. I: is the smallest and I2 the largest of the series. The
canine is large, oval, and has a narrow but deep vertical groove on
its inner face. The tooth was broken off not far above the gum line
during life. This event occurred some time before death because the
stump was to a considerable extent smoothed by abrasion.
Although the pulp cavity was widely exposed, there is no sign of
decay.

Fig. 11. Pharsophorus lacerans Ameghino, 1897, p. 503. PU 20551, detail of right
P -M : a, labial; b, occlusal; c, lingual views. Scale = 10 mm.

61

62 FIELDIANA: GEOLOGY, VOLUME 41

Table 4. Measurements of mandibular rami of Pharsophorus lacerans.

Depth of ramus below Breadth Depth of ramus below Breadth
labial side of Ml of same labial side of M4 of same

MACN 52-391

40.0

14.5

42.5

17.0

AC 3004

31.2

13.3

34.6

14.9

PU 21865

31.0

15.6

All lower cheek teeth are two-rooted. Pj is implanted obliquely, the
anterior extremity facing anteroexternally. The protoconid of P2 is
largely missing, but enough of its posterior slope is preserved to
reveal that the heel was quite distinct from it. The latter appears to
have been fully as wide as the protoconid; it is highest at the center,
where there is a small median cusp, and, sloping down on either side
inconspicuous cingula run downward and forward from the cusp. P3
has a large protoconid that slopes backward to a marked degree, the
anterior root continuing the slope into the jaw. The base of the
enamel is parallel to the alveolar border of the mandible, showing
that this sloping is natural. Ameghino's figure incorrectly shows
the base of the enamel to be inclined upward anteriorly, giving the
false impression that the sloping was due to post-mortem displace-
ment of the tooth. The heel is in general similar to that of P2, but is
somewhat wider, with a smaller central cusp and less sloping sides.
The premolars steadily increase in size from x to 3.

Mx is slightly shorter and wider than P3, and may have been ap-
proximately equal to it in height when unworn. The protoconid is
large and central in position; the well-developed paraconid is
anterior and only slightly internal to it. Both cusps are truncated by
wear. An anteroexternal cuspule is present at the base of the
paraconid. The base of the enamel on the external face is slightly
smaller, but there is no definite cingulum. No trace of a metaconid is
present. The talonid is wider than the trigonid and extends below it
on the external face. There is a small hypoconid and a minute en-
toconid, but no basin, strictly speaking. The hypoconid is in line
with the protoconid and the external surface of the talonid slopes
sharply downward from it. The remaining molars increase in size
posteriorly and, contrary to Ameghino's figure, are all incomplete,
the protoconids of M2 and 4 and the paraconid of M3 having clearly
been broken off prior to discovery of the specimen. The metaconids
progressively increase in size posteriorly and the paraconids become
more anterointernal in position. The anteroexternal cuspule at the
base of the paraconid is very small on M2 and absent on M4. The

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 63

Table 5. Measurements of PU 20551, Pharsophorus lacerans.

Width across palate between inner surfaces of anterior roots of P 's 22.1

Same between protocones of M 's ca. 24.0

Length from anterior edge of right P to posterior edge of right M ca. 37.5

Dorso ventral diameter of right canine root above posterior root of P 15.0

Transverse diameter of right canine root above posterior root of P 7.4

Dorsoventral diameter of left canine root above posterior root of P 14.5

Transverse diameter of right canine root above posterior root of P2 8.0

Width of naso-lacrimal contact (left side) 15.5

Width of naso-lacrimal contact (right side) 17.0

Diameter from surface of palate between P 's to dorsal edge of nasals 39.0

Diameter of rostrum above P 's 37.0

Diameter across postorbital processes 56.7

talonid becomes progressively smaller; it is wider than the trigonid
on M] and approximately equal to it in width on M2. On M2.3 the
very small entoconid is higher than the hypoconid. A small but
distinct basin occurs on the labial side of the talonid on M2.4. This
feature is lost in early stages of tooth wear, especially on M3.4, as
shown by its absence in MACN 52-391 and 11652, as compared with
the younger MACN 11653.

The mandible is robust, and increases slightly in depth posteriorly
below the cheek teeth. The symphysis is large, extending back to
beneath the center of P3, only slightly rugose, and nearly flat. The
coronoid process and the deep masseteric fossa are, as usual,
posterior to M4. A large mental foramen occurs beneath the anterior
portion of P2 and there are three smaller foramina in a line behind it,
beneath the posterior roots of P2, P3, and M2. Two vascular foramina
are present beneath Ix and I3 on the anterior face of the symphysis,
and a third, somewhat larger one, occurs a short distance below
them.

The paraconid of M3, like the canine, was broken off prior to the
death of the animal, since its stump, too, was clearly smoothed by
subsequent wear. Here, again, there is no evidence of subsequent
decay, although the pulp cavity was also exposed and the split ex-
tended well down into the anterior root. Injuries of this sort appear
to have been rather common among the larger borhyaenids and sug-
gest that at least some of these animals may have been bone
crushers.

The astragalus that was figured in two views by Ameghino (1897,
fig. 80), but not described by him, has not been found in the collec-
tions. This bone is certainly from a borhyaenid, but there is no
positive evidence that it is attributable to Pharsophorus.

64

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 65

Comments.— Cabrera (1927, p. 274) regarded the Mustersan
genus Plesiofelis Roth, 1903 as a junior synonym of Pharsophorus,
relegated Plesiofelis schlosseri Roth to the synonymy of Phar-
sophorus lacerans, and retained Plesiofelis cretaceus Roth as a
distinct species of Pharsophorus.

Marshall (1978) has shown that Pharsophorus and Plesiofelis are
quite distinct, valid genera. Pharsophorus lacerans (fig. 8) differs
from Plesiofelis schlosseri (fig. 12) in the oblique implantation of P^
in P23 being proportionately longer, narrower, and with the
posterobasal heel proportionately larger; in a more gradual increase
in size from Mj to M4; in more reduced talonid cusps; and in absence
of a distinctly basined cingular shelf on lingual side of talonid.
Taken separately these differences are not extreme, yet together
they are sufficient to indicate that the taxa are distinct.

In addition, the provenance of the two described species of
Plesiofelis was given by Roth (1903) as "Formacidn cretacea
superior, Lago Musters (Territorio del Chubut)." Simpson (1936,
1948) has demonstrated that nearly all of Roth's specimens with
this locality data are from the Mustersan horizon at Cerro del
Humo. P. schlosseri is thus almost certainly Mustersan, despite the
argument put forth by Cabrera (1927, p. 275) that it is from the
Deseado horizon at the Great Barranca south of Lago Colhue-
Huapi.

Ameghino (1897) was aware of the close similarity between
species of Pharsophorus and species of Santacruzian Borhyaena. He
noted that P. lacerans was similar to, but was slightly more massive
than, B. tuberata. The presence of a metaconid was also stressed as
a diagnostic feature of P. lacerans. The size differences noted by
Ameghino are of little value as the sample of P. lacerans falls within
the range of the comparative samples of Borhyaena tuberata and B.
macrodonta given by Marshall (1978). Although B. macrodonta and
P. lacerans are structurally quite similar, the latter is distinguished
from B. macrodonta by the presence of a small metaconid on M24
and by a relatively larger talonid. As far as the dentition is concern-
ed, an ancestral-descendant relationship for these taxa is plausible.

A complication arises, however, when the partial skull, PU 20551,
is taken into consideration. This specimen differs decidedly from
Borhyaena and other "dog-like" borhyaenids whose skulls are
known in the marked posterior prolongation of the canine root,
which extends back to a point above the posterior margin of P3; the
inflation of the maxillaries around the canines; the narrowing of the

66 FIELDIANA: GEOLOGY, VOLUME 41

rostrum anterior to P3; the anterior narrowing of the nasals; the
presence of a distinct postorbital process; and of a large, rugose
postorbital ridge extending from it to the sagittal crest. If we are
correct in an assignment of this specimen to P. lacerans, then an
ancestral relationship to Borhyaena would not seem possible. In
fact, there would appear to be no known form that could have fulfill-
ed this role, the Mustersan Plesiofelis being already too specialized
as regards the size gradient of the molars and the greater degree of
talonid reduction (Marshall, 1978).

The most interesting point to be noted is that PU 20551 shows
several resemblances to Thylacosmilus. Differences from
Borhyaena noted above are approaches to thylacosmilid structure
(Riggs, 1934). Derivation of the saber-tooth marsupials from a
borhyaenid ancestry has generally been assumed, although un-
doubted transitional forms are so far unknown. Scott (1937, p. 711)
suggested that the Casamayoran Arminiheringia might have been
involved in the saber-tooth lineage, but Simpson (1948, p. 42) was
sceptical as regards this possibility. PU 20551, whether or not it
was involved in the ancestry, does, as far as the above features are
concerned, demonstrate an intermediate condition between the two
groups.

Pharsophorus? antiquus (Ameghino, 1895). Figures 13, 14.
Worhyaena antiqua Ameghino, 1895, p. 655.

Proborhyaena antiqua Ameghino, 1897, p. 502; Loomis, 1914, p. 219.
Pharsophorus"! antiquus Marshall, 1978, p. 36, fig. 7.

Type.— MACN 52-532, a nearly complete isolated right upper
canine.

Hypodigm — Type and FMNH P13633, crown of a right lower C;
FMNH P 13800, an isolated right M3; and MACN 52-384, an
isolated left M3(?). We associate these specimens on the basis of size
and their general agreement in structure with P. lacerans.

Localities.— The type is probably from La Flecha, Santa Cruz Pro-
vince, Argentina. MACN 52-384 is labeled "Colhuapi Pyroth,"
which means the Deseadan horizon at the Great Barranca. FMNH
P 13633 was collected at La Flecha, and FMNH P 13800 is from
Pico Truncado.

Age. —Deseadan.

Diagnosis.— Very large, similar in size and structure to species of
Arctodictis (Colhuehuapian and Santacruzian); larger and propor-

PATTERSON & MARSHALL: DESEADAN MASUPIALIA

67

Fig. 13. Pharsophorusl antiquus (Ameghino, 1895, p. 655). MACN 52-384, an
isolated left M3(?): a, labial; b, occlusal; c, lingual views. FMNH P 13800, an isolated
right M : d, labial; e, occlusal; f, lingual views. Scale = 10 mm.

tionately more robust than Pharsophorus lacerans.

Description.— On the upper canine a prominent longitudinal
sulcus occurs along the lingual surface of the root. Total tooth
height is 89.5 mm.; crown height, as measured on lingual side, is
35.0 mm.; anteroposterior length of base of crown is 23.7 mm.,
breadth is 16.2 mm. A large wear surface is present on the

68 FIELDIANA: GEOLOGY, VOLUME 41

anteromedial face of the crown, extending from its tip to about half
the distance up the crown surface.

M3 is similar to that in species of Arctodictis. A small but distinct
protocone is present, the paracone is reduced and is connate basally
with the large metacone. A large parastyle is present, and extends,
as a ridge, around the anterior and labial sides of the paracone. A
weak ectocingulum and a shallow ectoflexus occur directly opposite
the metacone. Measurements— maximum length of tooth, 16.2 mm.;
maximum breadth, 13.3 mm.

On the lower canine a deep sulcus extends along the lingual sur-
face and a shallower one along the labial surface of the root. These
sulci extend onto the lowermost edge of the crown. Measurements-
maximum length of tooth, 65.0 mm.; height of enamel portion of
crown on labial side, 43.0 mm.; anteroposterior diameter at base of
crown, 24.5 mm., breadth, 16.7 mm.

On M3 the protoconid and paraconid are large and broad; the
metaconid is small but distinct; the talonid is reduced and is com-
posed of a small posterobasal cuspule with a cingular shelf exten-
ding from it along the labial surface. Measurements— maximum
length of tooth, 16.0 mm.; maximum breadth, 10.8 mm.

Comments.— Ameghino based this species on what he considered
to be a lower canine, although it is certainly an upper right. He was
impressed by the similarity of this specimen to those of species of
Borhyaena from the Santa Cruz Formation. He (1897, p. 502) was,
however, hesitant in referring it to that genus, as the rest of the
fauna with which it was associated differed "profoundly" from that
of the Santacruzian fauna, and later referred this species to Pro-
borhyaena, noting that it could be distinguished from P. gigantea
by its smaller size.

The upper canine, MACN 52-532, is not only smaller than that of
P. gigantea but has a distinct morphology. In the lower canine of P.
gigantea the root is relatively straight and ribbing occurs over the
entire root surface. Although upper canines of P. gigantea are not
known, the upper and lower canines of other borhyaenids are
basically similar.

MACN 52-532 is structurally similar to species of Borhyaena, to
Pharsophorus lacerans and to species of Arctodictis. In fact, the
canines of these three genera resemble each other closely. Canines of
species of Pharsophorus and Borhyaena are virtually indistinguish-
able, while in Arctodictis they are relatively and absolutely larger

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 69

FIG. 14. Pharsophorus? antiquus (Ameghino, 1895, p. 655). FMNH P13633, crown
of a right lower C: a, medial; b, dorsal; c, lateral views. Scale = 2 cm.

and more robust. MACN 52-532 is slightly larger than species of
Borhyaena, suggesting possible affinities to species of Arctodictis.
The several isolated specimens from beds of Deseadan age here
referred to P.(?) antiquus are, like the type, structurally similar to
Pharsophorus lacerans, although larger. In view of the structural
similarity of these specimens to P. lacerans and their occurrence in
beds of similar age, we have tentatively referred them to Pharso-
phorus. Future reference of P.(?) antiquus to Arctodictis cannot be
ruled out, although that genus lacks the metaconid, a cusp present
in MACN 52-384. It is more likely that P.(?) antiquus is referable to
a new genus, which cannot be adequately characterized on the basis
of specimens at hand.

cf. Pharsophorus sp.

cf. Pharsophorus sp. Chaffee, 1952, p. 515; Marshall, 1978, p. 37, figs. 8-10.

70 FIELDIANA: GEOLOGY, VOLUME 41

Material.— AMNH 29591, an incomplete crushed juvenile skull
with left and right C, left and right P1, and unerupted P3; left P2 and
DP3 complete; various tooth fragments, left zygomatic arch, and a
nearly complete left periotic.

Locality.— Rinconada de los Lopez in west side of Sierra Canquel,
Chubut Province, Argentina. The specimen was collected from a
level higher than the lake beds, from a breccia-filled dike near the
top of the intrusive walls.

Age. — Deseadan.

Description.— The canines are erupting and only about 25.0 mm.
of the crowns are exposed. The blades are narrow and weak relative
to the size of the skull and premolars.

P1 is set at an oblique angle in the jaw, a small and narrow pos-
terobasal heel is present. P2 is enormous compared to P1. A large
posterobasal heel is developed, which is elevated and cuspate labial-
ly. P3 is larger than P2. DP3 is almost complete, although the pos-
terolabial corner is broken. Basically, it resembles a diminutive M1,
although the proportions are considerably different. There is a tiny
cingular, crescentic protocone, a small parastyle, a larger paracone
and a still larger metacone. Parastyle, paracone, and metacone are
in a line.

The Ungual half of the left M2 is all that is known of the upper
molars. It has a small but distinct protocone, a reduced paracone, a
small but distinct parastyle, and a large metacone. What remains of
the roots of the other molars suggest that M3 was erupted and M4
erupting. Measurements— maximum combined breadth of nasals
= 58.0 mm., maximum length of nasals = 88.0 mm.; L DP3 = 11.9
mm., W DP3 = 8.0 mm.; L of right P1 = 7.5 mm., W P1 = 3.8 mm.; L
of left P1 = 7.6 mm., W P1 = 4.0 mm.; L P2 = 13.5 mm., W P2 = 7.2
mm.

Comments.— The generic affinities of AMNH 29591 are quite
uncertain. The animal is a juvenile and the canines have not yet at-
tained their adult size. Complete molars, which are so important in
borhyaenid taxonomy, are unknown. Largely on the basis of size,
Chaffee (1952, p. 515) tentatively assigned this specimen to Pharso-
phorus. With reservations we have done likewise. This specimen is
similar in size to Pharsophorus lacerans and appears to resemble
that species more closely than other well known Deseadan species of
Borhyaenidae.

Several features suggest, however, that AMNH 29591 might not

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 71

be referable to Pharsophorus. P2 is considerably larger than P3 in
AMNH 29591, while in the lower dentition of P. lacerans (MACN
52-391) the relative size differences between Pj and P2 are less mark-
ed. Although P3 of AMNH 29591 is not fully developed, it is larger
and more robust than in other known species of Borhyaeninae. Af-
finities of this specimen with some or all of the elements assigned to
P.(?) antiquus cannot be ruled out. In fact, the M2 fragment of
AMNH 29591 is similar in size and structure to the M3 (FMNH
P13800) assigned to P.(?) antiquus. AMNH 29591 may represent a
new genus and species, but this cannot be decided on the basis of
this specimen.

Subfamily Proborhyaeninae Ameghino, 1897, p. 501
Proborhyaena Ameghino 1897

Proborhyaena Ameghino 1897, p. 501.

Type.— Proborhyaena gigantea Ameghino 1897, p. 501.
Distribution.— Deseadan, Patagonia, and possibly Bolivia.
Diagnosis.— As for the type and only known species.

Proborhyaena gigantea Ameghino, 1897. Figures 15-17; Tables 6-9.

Proborhyaena gigantea Ameghino 1897, p. 501, fig. 78; Loomis, 1914, p. 218, fig.
144; Chaffee, 1952, p. 515, pi. 13.

Type.— MACN 52-382, a right mandibular ramus with alveoli of
incisors, C complete, roots of Plt anterior root and posterior alveolus
of P2, P3 complete, M! present but worn, alveoli of M2.3, and roots of
M4.

Hypodigm.— The type and AMNH 29576, a right mandibular
ramus with complete dentition (except for incisors), and associated
root of left C, and complete left M4, and left M23; MLP 71-XI-4-5, a
fragment of a left mandibular ramus with Ml complete but worn,
and anterior root of M2; MLP 73-VII-1-8, a fragment of a right man-
dibular ramus with alveoli of P3, Mx complete, and anterior alveolus
ofM2.

Tentatively referred material.1— MLP 71-XI-4-11 (-13), distal ends

'Alvaro Mones and Martin Ubilla (Comun. Paleontol. Mus. Hist. Nat. de
Montevideo, 1 (7). pp. 151-158) have recently reported and figured a left mandibular
ramus with Pjj-M* of Proborhyaena cf. P. gigantea from the Fray Bentos Fm. at
"paso del Cuello sobre el rio Santa Lucia," Dept. of Canelones, Uruguay.

Fig. 15. Proborhyaena gigantea Ameghino, 1897, p. 501. MACN 52-382 (type), a
right mandibular ramus with alveoli of incisors, C complete, roots of Pj, anterior
root and posterior alveolus of P2, P3 complete, Mx present but worn, alveoli of M2.3,
and roots of M4: a, labial; b, occlusal; c, lingual views. Scale = 50 mm.

72

FIG. 16. Proborhyaena gigantea Ameghino, 1897, p. 501. AMNH 29576, an iso-
lated left M : a, lingual; b, occlusal; c, labial views. Scale =10 mm.

73

o

o

o !

i i

id

i«

i i

<X>

-— 1

i i

i— (

d7

3 '

1 '

lO 1

1 (N

to

10 '

1 *""

t>

i-t

£

S '

1
1

lO

lO

t~

J

^ !

! *4

ire

«

«

CO

e

a
bo
So
a
C

[-^ 00 c-

w

J

03

•<

o

H

05

£

s

i— i

-

o

J

03

* J

* U5 • O -S-

C* i i-H CD «D

00 T m S ?-

cp M _ m ire

oj ^ i> 9* o*

7 -! r* l"T f>J (Nl

0> ~ rH CO

SB Z

6 o6

6 <N

ire w

g> 5

u

opst: .2

sss<<

74

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 75

Table 7. Measurements of mandibular ramus of Proborhyaena gigantea.

MACN 52-382

AMNH 29{

Depth of ramus

below Mx

59.0

ca. 66.9

Breadth of ramus

below Mj

26.5

ca. 29.0

Depth of ramus

below M4

64.0

ca. 74.0

Breadth of ramus

below M4

21.5

ca. 24.0

Symphysis

Breadth

45.5

46.0

Length

92.0

89.0

of three metapodials with attached sesamoids; PU 21866, a frag-
ment of a right maxillary with roots of M1'3, and PU 21992, a frag-
ment of a right maxillary with roots of M3; these specimens agree in
size with the Patagonian ones, but in the absence of any knowledge
of crown structure no positive identification can be made.

Localities.— The type is without precise locality data, although
probably from Cabeza Blanca. AMNH 29576 and MLP 73-VIM-8
are from the Scarritt Pocket, Rinconada de los Lopez. MLP 71-XI-
4-5 and 71-XI-7-11 (-13) were collected from "Zona Sud. Oriental
Laguna Payahile." The PU specimens are from Salla, Bolivia.

Age. — Deseadan.

Diagnosis.— Largest known species of Borhyaenidae. il C\ Pf M|.
Differing from Arminiheringia in larger size, in molars proportion-
ately longer than wide, and in C implanted more perpendicularly.

Description.— The mandibular ramus is robust. The symphysis is
fully fused and extends to a point below the P3-Mx junction. Four
well-developed mental foramina are present in the type, a small one
below the anterior root of P3, a larger one below the anterior root of
P2, and two others of equal size below Mlt placed one above the
other.

Alveoli of the two right incisors and that of the left medial one are
preserved in the type. Both were of equal size and small compared
with other species of Borhyaenidae, so far as can be judged from
their shallow alveoli.

The canine is massive and set almost perpendicularly in the jaw.
Posterodorsally a distinct concave wear facet is present and deep

Fig. 17. Proborhyaena gigantea Ameghino, 1897, p. 501. AMNH 29576, an iso-
lated left M : a, lingual; b, occlusal; c, labial views. Scale = 10 mm.

76

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 77

Table 8. Measurements of upper cheek teeth of Proborhyaena gigantea from
Argentina compared with a specimen from Sal la. Bolivia.

M1 M2 M3

Specimen L W

L

W

L

W

AMNH 29576

25.0

13.9

29.6

17.3

PU 21856 19.2* 15.0

22.8*

15.7*

19.6*

♦Measurements taken on alveoli.

sulci extend down the sides of the tooth, one on the labial and
another on the lingual surface. The root is relatively straight and is
ribbed over its entire surface.

The premolars increase in size from Pj to P3. Pi and P2 are single
cusped and are implanted obliquely, with the anterior roots set
labial to the posterior. P3 has roots aligned in an anteroposterior
direction and the crown carries a small posterobasal cusp or heel.

The molars increase in size from Mt to M4. The talonid is reduced
to a single cusp, which becomes progressively reduced from Mx to
M4, where it is merely a low vestige on the posterobasal edge of the
protoconid. Protoconid and paraconid both increase in size from Mx
to M4, although the paraconid is larger relative to the protoconid in
Mt than in M4. The metaconid is much reduced and is
distinguishable only on M2.3; it lies on the posterolingual edge of the
protoconid and after early stages of wear becomes united in a com-
mon shearing surface with the talonid cusp.

Table 9. Measurements of metapodials of Proborhyaena gigantea.

specimen

MLP71-XI-4-11
MLP 71-XI-4-12
MLP 71-XI-4-13

On the upper molars an ectoflex is well developed on M3, but very
weak on M2. The protocone is reduced to a swelling on the lingual
side and for all practical purposes is absent. The paracone is reduced
and the metacone very large. The metacrista is a prominent feature
on all and a large parastyle is present on M2'3.

The distal ends of the metapodials are very broad; a large keel
bordered by two large sesamoids is present on the ventral surface.

Comments.— Proborhyaena gigantea is the largest and one of the

naximum transverse

maximum dorsoventral

breadth, distal end

depth, distal end

27.0

22.5

30.7

24.5

25.0

78 FIELDIANA: GEOLOGY, VOLUME 41

most specialized of known Borhyaenidae. In its gigantic size,
presence of only two lower incisors in each ramus, oblique implanta-
tion of ^ and P2, large proodont P3, and the parallel ribbing on the
enormous perpendicular lower canines, this species is too specializ-
ed to be ancestral to any known borhyaenid found in beds of
younger age. Ameghino's (1897, p. 502) erroneous reference of his
Worhyaena antiqua to Proborhyaena has been discussed above (p.
68). Possible affinites of Proborhyaena and Arminiheringia have
been discussed elsewhere (Marshall, 1978, p. 23).

Sinclair (1930, p. 38, pi. VIII, figs. 3-3 A) described a fragment of a
right mandibular ramus (FMNH P 13526), which he designated Pro-
borhyaena sp. He stated that it came from the "Deseado Forma-
tion" and the "Astraponotus beds" (= Muster san). Riggs and Pat-
terson (1939, p. 149, n2) pointed out that this was in error and that
the specimen had been collected from Colhue-Huapi beds at the
Great Barranca south of Lago Colhue-Huapi. It is thus younger, not
older than Proborhyaena gigantea; it has three incisors on each side
of the mandibular ramus instead of two; it is smaller than P.
gigantea; and the structure of the canine, mandibular ramus, and
orientation of the premolars are quite different. This specimen is
clearly not referrable to Proborhyaena; it represents a Colhuehua-
pian species of the genus Arctodictis, A. sinclairi, which is possibly
ancestral to A. munizi from the Santacruzian (Marshall, 1978).

A specimen consisting of a cranium and associated mandible,
identified as Proborhyaena, has been recorded from the Salla beds of
the Salla-Luribay Basin, Bolivia (Hoffstetter, 1968, p. 1,095). This
specimen is reportedly smaller than P. gigantea.

Borhyaenidae genus and species indet.

Two specimens of borhyaenids from the Salla fauna cannot be
referred to any of the described species, and their fragmentary
nature prevents precise identification. These are PU 21873, a frag-
ment of a left premaxilla with roots of C and Plf and PU 21997, a
fragment of a right mandibular ramus with roots of two teeth.
These specimens represent an animal intermediate in size between
Notogale mitis and Pharsophorus lacerans. Perhaps the specimen
mentioned by Hoffstetter et al. (1971) from Lacayani, Bolivia, in
which the crown of the M4 had a length of 11.0 mm., should be in-
cluded in this group. These specimens evidently represent a distinct
and so far little known species of Deseadan borhyaenid.

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 79

Superfamily Caenolestoidea

(Trouessart, 1898) Osborn, 1910

Family Caenolestidae Trouessart, 1898

Subfamily Caenolestinae

(Trouessart, 1898) Sinclair, 1906

Pseudhalmarhiphus Ameghino, 1899a
Pseudhalmarhiphus Ameghino, 1899a.
Type.— P. guaraniticus.
Distribution.— Deseadan, Patagonia.
Diagnosis.— As for the type and only known species.

Pseudhalmarhiphus guaraniticus (Ameghino, 1899a)

Halmarhiphus guaraniticus Ameghino, 1899a, p. 7.

Halmariphus guaraniticus Ameghino, 1899b, p. 560, fig. 5.

Halmarhiphus (Pseudhalmarhiphus) guaraniticus Ameghino, 1902a, p. 424, fig. 5.

Pseudhalmarhiphus (Halmarhiphus) guaraniticus Ameghino, 1903, p. 83, fig. 2.

Pseudhalmarhiphus guaraniticus Loomis, 1914, p. 224; Reig, 1955, p. 61; Mar-
shall, 1976a, p. 8.

Type.— Based on a fragment of a right mandibular ramus with
Mj.g. M2 figured by Ameghino, present whereabouts of type
unknown.

Hypodigm.—Type only.

Locality.— Chubut Province, Argentina. No precise locality data
are available.

Age. —Deseadan.

Description and diagnosis.— As figured by Ameghino, the
trigonid of M2 has two well-developed Ungual cusps (paraconid and
metaconid) separated by a deep valley. The trigonid is only slightly
narrower than the talonid, and the cusps are sharp and high.
Trigonid and talonid basins are deep and narrow, and the tooth is
very narrow for its length. An anterobasal cingulum is also well
developed, but not to the degree seen in species of Pichipilus and
Pliolestes (Marshall, 1976a, p. 8). In all of these features
Pseudhalmarhiphus agrees closely with species of Stilotherium
from beds of Santacruzian age in Argentina.

Comments.— Ameghino (1899a, p. 7) gave the length of M^j (his
M4.6) as 3.5 mm., and later (1899b; 1902a; 1903) figured M2 in three
views. The present whereabouts of this specimen, the type, is
unknown. P. guaraniticus is the only caenolestine known from the

80 FIELDIANA: GEOLOGY, VOLUME 41

Deseadan, and is structurally similar, and possibly ancestral, to
Stilotherium from beds of Santacruzian age (Marshall, 1976a, p. 8).
P. guaraniticus is also the oldest known species surely referable to
the Caenolestinae.

Progarzonia notostylopense Ameghino (1904, p. 260), collected
from beds of Casamayoran age at the Great Barranca South of Lago
Colhu6-Huapi, was at one time referred to the Caenolestidae. The
type, MACN 55-14, consists of a left mandibular ramus with a
single rooted P3. Simpson (1967b, p. 9) has noted that this "animal
may be a caenolestid, but in my opinion it is not adequately iden-
tifiable at any taxonomic level below class." We agree.

Subfamily Palaeothentinae Sinclair (1906)

Members of the subfamily Palaeothentinae are known from
deposits of Deseadan through Santacruzian age. Although some 16
generic names for palaeothentines have been proposed (15 based in-
itially on Santacruzian species), their status is extremely dubious.
Apart from Palaeothentes, the first named, they include Acdestis
Ameghino, 1887; Epanorthus Ameghino, 1889; Dipilus Ameghino,
1890; Decastis Ameghino, 1891; Callomenus Ameghino, 1891
Essoprion Ameghino, 1891; Halmadromus Ameghino, 1891
Halmaselus Ameghino, 1891; Palaepanorthus Ameghino, 1902
Metriodromus Ameghino, 1894; Metaepanorthus Ameghino, 1894
Paraepanorthus Ameghino, 1894; Prepanorthus Ameghino, 1894
Cladoclinus Ameghino, 1894; and Pilchenia Ameghino, 1903.

Most of the above genera and included species were distinguished
on the basis of two characters— absolute size and relative size of P3
compared to Mx. In addition, supposed differences in proportions of
M^ were occasionally noted.

Subsequent workers agreed that the Palaeothentinae were overly
split at the generic and specific levels. Simpson (1945, p. 45), for
example, tentatively recognized five genera of Palaeothentinae
(Palaeothentes, Pilchenia, Acdestis, Dipilus, and Halmadromus).
The other genera were either included as synonyms of one of these
five or were regarded as nomina vana. As Simpson (1945, p. 42, 2n)
noted, "Proper generic criteria for this group [Palaeothentinae] have
not yet been worked out, and the published data are inadequate in
several cases." All of these generic names are thus of dubious value
at the present time and must be regarded as such until the entire
subfamily is adequately revised. One of us (L.G.M.) has nearly com-

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 81

pleted a review of the Santacruzian taxa in which five species are
recognized as valid: P. minutus Ameghino, 1887; intermedius
Ameghino, 1887; owenii Ameghino, 1887; lemoinei Ameghino, 1887;
and aratae Ameghino, 1887. These species can be all included in a
single genus PcUaeothentes. Adequate samples for each of these
species are known and variation within each can be evaluated.
Based upon this preliminary study, it was found that such
characters as absolute size, relative size of P3 to Mlf and relative pro-
portions of M14 are useful in diagnosing the different species but are
of very dubious generic value. The characters dictate that either we
recognize five distinct Santacruzian species of Palaeothentes, or we
recognize five monotypic Santacruzian genera. For present pur-
poses we choose the former alternative and have assigned (with
some additional qualifications) the four Deseadan species of
Palaeothentinae to the genus Palaeothentes. We are unable to
detect any differences of generic significance between these and
Santacruzian representatives of the genus.

Palaeothentes Ameghino, 1887

Palaeothentes Ameghino, 1887, p. 5.

Type.— Palaeothentes aratae Ameghino, 187, p. 5.

Distribution.— Deseadan, Colhuehuapian, and Santacruzian of
Patagonia, and Deseadan of Bolivia.

Diagnosis.— ?/2 1/1 3/3 4/4. P13 two-rooted; P1'2 very reduced in
size and crown height; P3 enormous, rivalling M1 in size, with crown
height equal to or greater than that of M1. Posterior end of crown
much broader than anterior. Upper molars decreasing rapidly in size
from M1 to M4. Sharp cutting edge formed along labial sides of M1'2
extending onto P3. Labial sides of M1"2 much higher than lingual. M1
with large posterolingual cusp, conferring rectangular shape to
tooth; remnants of this cusp seen on M2, but not on M3 or M4. Man-
dibular ramus typically long and relatively shallow. One large,
laterally compressed incisor in each jaw followed by four tiny,
single-rooted, vestigial teeth (presumably an incisor, a canine, and
two premolars). P3 either single rooted and styliform with a crown
height much less than that of the Mlt or very large, double rooted,
and equal in height to trigonid crest of Mx

Lower molars decreasing rapidly in size from Mj to M4. Trigonid
region of Mx greatly elongated, with paraconid set far anteriad;
crest connecting paraconid with protoconid blade-like. Trigonid of
Mj typically longer and narrower than talonid. Crista obliqua large,

82

FIELDIANA: GEOLOGY, VOLUME 41

<V

Fig. 18. Palaeothentes lucina (Ameghino, 1903, p. 128). AC 3110, a left mandibular
ramus with P3-M4: a, labial; b, occlusal; c, lingual views. Scale = 5 mm.

well developed. M2.4 bunolophodont; trigonid and talonid regions of
M2.3 distinct, M4 usually ovate.

Palaeothentes lucina (Ameghino, 1903). Figure 18; Table 10.

Pilchenia lucina Ameghino, 1903, p. 128, fig. 49; 1904, p. 259 (said to be new in
1904, but publication in 1903 was prior and valid); Loomis, 1914, p. 222, fig. 146.

Type— MACN 52-371, an isolated left M3.

Hypodigm.— Type and AC 3110, a left mandibular ramus with P3-
M4 complete.

Localities.— The type is from the "Piroteriense" and probably
from Cabeza Blanca; AC 3110 is definitely from that locality.

Age. — Deseadan.

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 83

Diagnosis.— Compared with P. chubutensis and P. boliviensis, P3
similar in relative length compared to Mlf but considerably nar-
rower and lower; larger than in P. praecursor. Decrease in size from
Mj to M4 more gradual than in P. chubutensis.

Comments.— Ameghino erected "Pilchenia" lucina on an isolated
molar which Loomis correctly interpreted as M3. Loomis assigned a
second specimen (AC 3110), consisting of the greater part of a left
mandibular ramus, to this species. We agree with this assignment.

When he proposed Pilchenia, Ameghino did not present criteria
that would distinguish this from any of the 14 named Colhuehua-
pian and Santacruzian genera of Palaeothentinae. We have com-
pared AC 3110 with a large sample of Santacruzian taxa of
Palaeothentes and find no reason to recognize it as distinct at the
generic level and therefore assign lucina to that genus.

Ameghino (1903, p. 128) based "Pilchenia" lobata on a fragment
of a right mandibular ramus with a complete M2 (MACN 52-379).
This specimen was collected from the "Notohippidense" (early San-
tacruzian) horizon at Karaiken, near the eastern end of Lago Argen-
tine Santa Cruz Province, Argentina. As noted by Marshall and
Pascual (1976, p. 113), this specimen is clearly referable to
Palaeothentes and is very similar to, and possibly conspecific with
P. lemoinei from the Santa Cruz beds along the Atlantic coast be-
tween Rio Gallegos in the south and Monte Leon in the north. The
species lobata is placed in Palaeothentes (Marshall and Pascual,
1976).

Palaeothentes boliviensis new species. Figure 19; Table 10.

Holotype.—PU 21977, a fragment of a right mandibular ramus
with P3-Mi (posterolingual corner of Mj missing).

Hypodigm.—Type only.

Locality.— Salla-Luribay Basin (Branisa locality V-2), Bolivia.

Age. — Deseadan.

Etymology.— In reference to Bolivia, boliviensis.

Diagnosis.— Smaller than P. chubutensis; considerably larger and
with more prominent and higher crowned P3 than P. lucina and P.
praecursor.

Description.— The double-rooted P3 is very large relative to Mj and
is equal in height to the trigonid of Mt; the posterior end is much
broader than the anterior, and a small cuspule occurs on the anterior

Fig. 19. Palaeothentes boliviensis new species. PU 21977, a fragment of a right
mandibular ramus with P3-Mj (posterolingual corner of Mj talonid is missing): a,
labial; b, occlusal; c, Ungual views. Scale = 5 mm.

84

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 85

Fig. 20. Palaeothentes chubutensis (Ameghino, 1897, p. 500). MACN 52-378
(type), a right mandibular ramus with posterior root of P2, P3-M2, and M4 relatively
complete, roots of M3 (all teeth are heavily worn): a, labial; b, occlusal; c, lingual
views. Scale = 5 mm.

edge about midway up the crown. The talonid of Mt is only slightly
lower than trigonid.

Comments.— The type of Palaeothentes boliviensis is the only
palaeothentine known from the Salla fauna, and it is the only
palaeothentine yet known outside Patagonia. In its relatively large
size and high, broad P3, P. boliviensis shows closer affinities to P.
chubutensis than to any other known Deseadan palaeothentine. Ex-
cept for the very large Santacruzian species P. aratae, P. boliviensis
is considerably larger than any other known Colhuehuapian or San-
tacruzian species of Palaeothentinae.

Palaeothentes chubutensis (Ameghino, 1897). Figure 20; Table 10.

Epanorthus chubutensis Ameghino, 1897, p. 500, fig. 77.
Palaeopanorthus chubutensis Ameghino, 1901, p. 77.

! -M

00 o
H* <N

! fi

I CO

! -*

d

*

s

hJ

1

1

ic

*

CO

CO

CO

o

1

<N

rH

CN

CO

9

3

J5

CO CO

d

*

224)
n root
■ ramu

CO

G

•
*

.to

»~

•a

to

e
«

3

|

O
O

55
e

■3

s

is (1914, p.
ent based o
mandibular

o
4

a,

O

3

00

t-

•9

CO

a>

to

to

cp

to

CN

ID

CD

t~

a

cs

Loom
lurem
ithof

e

lO

o

e

O

e

C-

c

lO

g

CD

4)

£

CN

O

CD

a>

u

-«

z

o

u

CO

o

CO

o

CN

o

<

>h cct a>

1

3^

Cy

1

<

5
|

CI)

|

-2 ^ Q
< * *

09

a,

a.

^

a.

» * *

86

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 87

Palaeothentes chubutensis Loomis, 1914, p. 221, fig. 145.

Type.— MACN 52-378, a right mandibular ramus with posterior
root of P2, P3-M2, and M4 complete, roots of M3, all teeth being heavi-
ly worn.

Hypodigm.—Type only.

Locality.— Chubut Province, Argentina. The exact locality is not
known, but it is probably from Cabeza Blanca.

Age. — Deseadan.

Diagnosis.— Largest known species of pre-Santacruzian
Palaeothentinae. Differing from similar-sized Santacruzian P.
aratae (e.g., MACN 14) in slightly deeper, more robust mandibular
ramus and larger P3.

Description.— P3 is a very large, double-rooted tooth, equal in
height to the trigonid of Mlf and is much wider anteriorly than
posteriorly.

Comments.— Palaeothentes chubutensis is easily distinguished
from other Deseadan Palaeothentinae by its large size and promi-
nent P3. In P. chubutensis, as in the somewhat smaller P. boliviensis,
P3 is large, prominent, and equal in height to the trigonid of Mv
Palaeothentes lucina and P. praecursor are much smaller than P.
chubutensis and P. boliviensis, and have a proportionately smaller
P3-

Ameghino (1897, fig. 77) figured the type of P. chubutensis in
which M3 was correctly illustrated as missing. Loomis (1914, fig.
145) redrafted Ameghino's figure and in so doing restored the miss-
ing tooth.

Palaeothentes praecursor (Loomis, 1914). Figure 21; Table 10.

Callomenus praecursor Loomis, 1914, p. 223, figs. 147, 148

Acdestis praecursor Pascual and Odreman Rivas, 1971, p. 383; Clemens and Mar-
shall, 1976, p. 69.

Type.— AC 3020, a fragment of a right mandibular ramus with
M2; the crowns of P3 and Mlt originally present, have been lost since
the specimen was figured by Loomis.

Hypodigm. —Type only.

Locality.— Cabeza Blanca, Chubut Province, Argentina.

Age. — Deseadan.

Diagnosis.— P3 two rooted, very small compared with that of other
Deseadan species; about half as high as M1# M12 slightly larger than

88

FIELDIANA: GEOLOGY, VOLUME 41

•; zs/r-

6=

x -•'

2^

Fig. 21. Palaeothentes praecursor (Loomis, 1914, p. 223). AC 3020 (type), a frag-
ment of a right mandibular ramus with crown of M2 and alveoli of P2-Mj: a, labial; b,
occlusal; c, lingual views. Scale = 5 mm.

in P. lucina, considerably smaller than in P. boliviensis and P.
chubutensis.

Comments.— M^ of P. praecursor are heavily worn, preventing
comparison of minor cusp morphology with P. lucina. The small size
of the P3 and large size of Mj.2 serve to separate the species.

Palaeothentes praecursor has the smallest P3 of known Deseadan
Palaeothentinae. In this feature it agrees well with Santacruzian
species assigned by Ameghino to such genera as "Acdestis, "
"Callomenus," and "Decastis." In species of "Acdestis" the P3 is
typically small but two-rooted; while in "Decastis" (e.g., "D. colum-
naris "— MACN 5561) it is very reduced and definitely single rooted.
Because of the structure of P3, Loomis assigned this species to
"Callomenus. "

The structure of P3 is, however, quite variable in the Palaeothen-

PATTERSON & MARSHALL: DESE AD AN MASUPIALIA 89

tinae, and although a useful character in distinguishing species, it
appears to lack generic utility. Mh2 have the typical structure seen
in species of Palaeothentes, and for these reasons we deem it more
parsimonious to regard praecursor, at least for the present, as a
member of this genus.

Subfamily Abderitinae Sinclair, 1906
Parabderites Ameghino, 1902b.
Parabderites Ameghino, 1902b, p. 121.
Type.— Parabderites bicrispatus
Distribution.— Deseadan and Colhuehuapian, Patagonia.

Diagnosis.— Jaw short, deep, but shallower, more gracile than in
Abderites. Lower incisor large, laterally compressed, followed by
four spaced, single-rooted, vestigial teeth, separated from P3 by
distinct diastema. P3 large, double-rooted, blade-like with up to two
serrations on each face and corresponding apical denticles. Trigonid
of M, compressed transversely, blade-like with two distinct serra-
tions on each face and corresponding apical denticles; talonid little
modified, similar to that of M2. M2.3 proportionately longer, nar-
rower than in Abderites; trigonids and talonids distinct. M2 basical-
ly quadrate, slightly broader anteriorly than posteriorly; two labial
cusps, two lower lingual cusps. M4 smaller than M3, oval in occlusal
view.

Parabderites minusculus Ameghino, 1902c. Figure 22.

Parabderites minusculus Ameghino, 1902c, p. 43; Lomis, 1914, p. 224; Marshall,
1976b, p. 79, fig. 10.

Type.— MACN 52-380, a left mandibular ramus with P3-M3 com-
plete, and alveoli of M4.

Hypodigm.— Type only.

Locality.— Probably from Cabeza Blanca.

Age. —Deseadan.

Diagnosis.— Smallest known species of genus; further differing
from Colhuehuapian P. bicrispatus in lacking serrations on P3, with
P3 and trigonid region of Mj inclined forward at greater angle
relative to main horizontal axis of ramus.

Comments.— This species has been discussed at length by Mar-
shall (1976b, p. 79, fig. 10). P. minusculus is the oldest known
species referable to the Abderitinae and is the only known Deseadan
species of that subfamily.

90

FIELDIANA: GEOLOGY, VOLUME 41

Fig. 22. Parabderites minusculus Ameghino, 1902c, p. 43. MACN 52-380 (type), a
left mandibular ramus with P3-M3, and alveoli of M4: a, labial; b, occlusal; c, lingual
views. Scale = 5 mm.

Superfamily Polydolopoidea Clemens and Marshall, 1976
Family Polydolopidae Ameghino, 1897

Genus and species indet.

A single specimen, PU 21998, a fragment of a left mandibular
ramus with the base of a large gliriform tooth, and roots of P2 and P3,
from Salla, Bolivia, is referable to this family (fig. 23). It is the latest
known polydolopid.

Fig. 23. Polydolopid genus and species indet. PU 21998, a fragment of a left man-
dibular ramus with base of large gliriform tooth, and roots of P2 and P3: a, labial; b,
occlusal; c, lingual views. Scale = 10 mm.

91

92 FIELDIANA: GEOLOGY, VOLUME 41

The base of the gliriform tooth is narrow transversely (3.0 mm.)
and deep dorsoventrally (5.5 mm.). It projects laterally from the jaw
at an angle of 35° relative to a line drawn through the centers of P2
and P3. A large, edentulous diastema separates this tooth from P2,
which is very small, situated directly anterior to P3, and double
rooted, the anterior root being smaller than the posterior. The base
of P2 is 2.7 mm. long and 1.8 mm. wide at the posterior end. As
shown by the roots, P3 is a large elongate tooth. Its base is about 9.6
mm. in length, the anterior root being longer (5.8 mm.) than the
posterior (3.8 mm.), with both of about equal breadth (4.0 mm.). The
mandibular symphysis is smooth, suggesting that the rami were
loosely joined and probably capable of movements independent of
each other. A small mental foramen occurs 4.0 mm. below the base
of P2. The depth of the ramus below the middle of P3 on the labial
side is 14.0 mm., the breadth 6.8 mm.

Comparisons.— Included within the family Polydolopidae are five
genera from the early Tertiary of South America. Four of
these— Amphidolops, Eudolops, Polydolops, and Seumadia—are
known only from Patagonia: Eudolops and Amphidolops from the
Casamayoran (conventionally Early Eocene), Seumadia only from
the Riochican (conventionally Late Paleocene), and Polydolops from
Riochican through Muster san (conventionally Mid Eocene) age
(Simpson, 1948; Pascual and Odreman Rivas, 1971). The fifth,
Epidolops, is known only from the Riochican Itaborai fauna near
Rio de Janeiro, Brazil (Paula Couto, 1952).

In possessing three procumbent gliriform teeth in each lower jaw
Epidolops differs markedly from all other known polydolopids
which have only one. Seumadia, is known only from an isolated M3
and therefore cannot be compared directly with the Salla specimen.
Amphidolops, Eudolops, and Polydolops are known from referred
upper and lower dentitions. Polydolops, Amphidolops, Epidolops,
and PU 21998 are characterized by possession of a very large
"plagiaulacoid" P3, while in Eudolops the shearing specialization of
this tooth has been secondarily (?) lost and the tooth greatly reduced
in size. Unfortunately, the lower dentition of Amphidolops anterior
to Mj is unknown and direct comparison of this taxon with the Salla
polydolopid is not possible.

The Salla polydolopid can thus be compared directly only with
species of Polydolops, the most abundant and, whether for this
reason or another, the most varied of known polydolopid genera.
These variations are, however, primarily in size, proportions, and

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 93

minor details of cuspule development. According to Simpson (1948),
they involve no important morphological characters.

In size, PU 21998 is considerably larger than all known species of
Polydolops of Riochican and Casamayoran age. The length of P3 of
Casamayoran and Riochican species has a known range of 4.0 mm.
in P. rothi, to 7.8 mm. in P. clavulus, compared to 9.6 mm. in PU
21998. An extension of the time range of Polydolops to the Muster-
san has been reported by Pascual and Odreman Rivas (1971, p. 381),
but no description of the material has yet appeared, and comparison
with PU 21998 is therefore not possible.

Classification of the Argentinian polydolopid specimens has pro-
ven difficult. Specimens are few, generally fragmentary, and
associated upper and lower dentitions are unknown. More impor-
tantly, the types are not all comparable, and numerous names, ap-
plied to fragmentary materials, can neither be rejected nor
validated (Simpson, 1948). In view of this situation and the
fragmentary nature of PU 21998, affinities of the Bolivian form
with other known genera cannot be firmly established. Of the five
known genera, PU 21998 appears to be structurally most similar to
Polydolops.

Species of Abderites also have a large "plagiaulacoid" tooth, but
in that group it is Mlf while in polydolopids it is P3 (Paula Couto,
1952). In Abderites there are four tiny, spaced, single-rooted,
vestigial teeth between the large gliriform tooth and Mlt and Mj is
always preceded by a styliform P2 (Marshall, 1976b, p. 60). In
polydolopids P3 is preceded only by a two-rooted or incipiently two-
rooted, peg-like P2, as in PU 21998.

In known beds of Casamayoran age, polydolopids are relatively
abundant and are represented by some 10 nominal species (Simp-
son, 1948). Polydolopids are rare and represented only by a single
specimen each in the Mustersan and Deseadan. It is interesting and
probably significant that Abderites, a caenolestid structurally
similar to polydolopids, first appears in beds of Colhuehuapian age
(Marshall, 1976b), following the last appearance of polydolopids in
the Deseadan.

DISCUSSION

The Deseadan was an interesting time in the history of vertebrate
life in South America. At the close of this age various mammalian
groups made their last appearance (e.g., pyrotheriids and isotem-
nids), and others their first (e.g., homalodotheriids, toxodontids,

94 FIELDIANA: GEOLOGY, VOLUME 41

mylodontids). More importantly, those foreign invaders, the
rodents and primates, first occur in beds of this age.

Deseadan time marks for the Marsupialia a number of important
changes, some of which may be linked with the appearance, disap-
pearance or diversification of other groups.

The borhyaenid subfamily Proborhyaeninae reached large size in
the early Tertiary. The last known member of this group, Pro-
borhyaena gigantea, is the largest species of borhyaenid known,
adults having skulls in excess of 2 ft. in length (Marshall, 1978). A
possible explanation for the disappearance of this group exists. The
other occupants of the terrestrial carnivore adaptive zone were
members of a group of large cursorial ground birds, the
Phororhacoidea. Three families, assorted into medium, large, and
gigantic size groups by mammalian standards, are currently
recognized. Phororhacoids are first known from beds of Deseadan
age and ranged in South America through the Monterhermosan
(conventionally Late Pliocene). Two of the families, Psilopteridae
and Phororhacidae, were swift and rather lightly built, while the
third, Brontornithidae (at present unknown subsequent to the San-
tacruzian), included ponderous forms with large massive beaks. The
former two were evidently the dominant cursorial carnivores of
their time (Patterson and Pascual, 1972, p. 262).

As suggested by Marshall (1978), the phororhacoids may have
contributed to the disappearance of such large borhyaenids as Pro-
borhyaena. This is clearly speculative, although it does appear that
these birds either actively or passively replaced, at least in part,
these large Borhyaenidae on the savanna-grasslands of southern
South America.

Representatives of other borhyaenid subfamilies occur in the
Deseadan, and some show affinities with taxa from earlier and later
faunas. Notogale mitis is a small, generalized member of the sub-
family Hathlyacyninae. It could have been derived from the
Casamayoran Patene coluapiensis, and is apparently involved in the
ancestry of Sipalocyon and perhaps that of Cladosictis, both genra
known from the Colhuehuapian and Santacruzian. In the subfamily
Borhyaeninae, Pharsophorus lacerans is of some interest in that cer-
tain cranial characters shown by a referred specimen exhibit some
resemblance to structures encountered in the saber-tooth
thylacosmilids, although dentally the species is close to Borhyaena.
Pharsophorus (?) antiquus may be involved in the ancestry of Arc-
todictis, although this relationship is very uncertain.

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 95

The Caenolestidae, including the subfamilies Caenolestinae,
Palaeothentinae, and Abderitinae, make their first appearance in
the Deseadan. The Caenolestinae are represented by one species,
Pseudhalmarhiphus guaraniticus, the Palaeothentinae by four
species of Palaeothentes (P. lucina, P. chubutensis, P. praecursor,
and P. boliviensis), and the Abderitinae by Parabderites
minusculus. This diversity indicates a considerable, but un-
documented, adaptive radiation prior to the Deseadan. Some of
these Deseadan caenolestids show close structural and presumably
phylogenetic affinity with Colhuehuapian and Santacruzian taxa.
Pseudhalmarhiphus guaraniticus may be involved in the ancestry of
Stilotherium dissimile (Santacruzian); Parabderites minusculus is
probably ancestral to P. bicrispatus (Colhuehuapian); Palaeothentes
chubutensis makes at least an ideal structural ancestor for P.
aratae (Santacruzian); Palaeothentes lucina shows strong affinity to
P. lobata (Colhuehuapian) and P. lemoinei (Santacruzian); and
Palaeothentes praecursor may be ancestral to P. owenii (Santacru-
zian). P. boliviensis does not appear to have left any known descen-
dants in later faunas.

The Polydolopoidea declined in diversity after Casamayoran time
(conventionally Early Eocene), and make their last appearance in
the Deseadan of Bolivia. It is either coincidental or significant that
this group's last appearance is concurrent with the first
documented appearance of Rodentia in South America. It is general-
ly agreed that polydolopids were somewhat rodent-like in structure
and ecology, and may thus have been replaced by rodents in the
Early Oligocene.

One of the most striking features of known Deseadan faunas is
the apparent absence of Didelphidae. As noted above, Hoffstetter
(1968, 1976) tentatively reported didelphids in the Salla fauna of
Bolivia, although these reports are unconfirmed and may possibly
have been based on fragmentary remains of the small borhyaenid,
Notogale mitis. This species is abundant in the PU collection from
Salla, but was not recorded as such by Hoffstetter. The report of a
didelphid from La Flecha by Tournouer (1903) needs confirmation.
The apparent absence of Didelphidae is puzzling, considering that
members of the Didelphinae and Microbiotheriinae are known from
earlier and later faunas. Their absence in Deseadan faunas is not en-
tirely attributable to sampling bias, as very small Caenolestidae
and Rodentia are known from beds of this age and at various
localities.

96 FIELDIANA: GEOLOGY, VOLUME 41

For the most part, the Deseadan Marsupialia show closer
affinities to taxa from later rather than from earlier faunas. In our
opinion, this tends to convey a false impression. For example,
Mustersan Marsupialia are known only from three taxa— Plesiofelis
schlosseri (Borhyaeninae), Procladosictis anomala (Hathlyacyninae),
and an undescribed polydolopoid. This scarcity prevents adequate
comparison between the two faunas.

In contrast, marsupials are relatively abundant in Colhuehuapian
and Santacruzian faunas. The diversity of taxa allow adequate com-
parisons to be made with those in the Deseadan, and in many cases
phylogenetic lineages can be recognized. It should also be stressed
that according to differences between Deseadan and Colhuehuapian
mammals in general, the palaeontological hiatus between these ages
was of less duration than the hiatus between the Mustersan and
Deseadan (Pascual and Odreman Rivas, 1971, 1973).

REFERENCES

Ahlfeld, F. and L. Branisa
1960. Geologia de Bolivia. Inst. Bol. Petroleo., 245 pp.

Ameghino, F.

1887. Enumeraci6n sistematica de las especies de mamiferos f6siles coleccionados
por Carlos Ameghino en los terrenos eocenos de la Patagonia austral y deposita-
dos en el Museo La Plata. Bol. Mus. de La Plata, 1, pp. 1-26.

1889. Contribucibn al conocimiento de los mamiferos f6siles de la Republica Argen-
tina. Actas Acad. Nac. Cienc. C6rdoba, 6, pp. 1-1,027.

1894. Enumeration synoptique des especies de mammiferes fossiles des forma-
tions eocenes de Patagonie. Bol. Acad. Nac. Cienc. C6rdoba, 13, pp. 259-452.

1895. Premiere contribution a la connaissance de la faune mammalogique des
couches a Pyrotherium. Bol. Inst. Geogr. Arg., 15, pp. 603-660 [Volume dated
1894, but this paper probably published in 1895. Ameghino always so cited it.]

1897. Mammiferes Cretaces de l'Argentine. (Deuxieme contribution a la conais-

sance de la faune mammalogique des couches a Pyrotherium). Bol. Inst. Geogr.

Arg., 18, pp. 406-429, 431-521.
1899a. Sinopsis geol6gico-paleontol6gica. Supplemento. Adiciones y correcciones.

La Plata, 13 pp.
1899b. On the primitive type of the plexodont molars of mammals. Proc. Zool. Soc.

London, 1899, pp. 555-571.
1901. Notices preliminaires sur des ongules nouveaux des terrains Cretaces de

Patagonie. Bol. Acad. Nac. Cienc. C6rdoba, 16, pp. 349-426.
1902a. Sur le type primitif des molaires plexodontes des mammiferes. An. Mus.

Nac. Buenos Aires, 3, no. 1, pp. 419-439.
1902b. Premiere contribution a la connaissance de la faune mammalogique des

couches a Colpodon. Bol. Acad. Nac. Cienc. C6rdoba, 17, pp. 71-141 (1902-1904).

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 97

1902c. Notice preliminaires sur les mammiferes nouveaux des terrains Cretaces de
Patagonie. Bol. Acad. Cienc. C6rdoba, 17, pp. 5-70 (1902-1904).

1903. Los diprotodontes del orden de los plagiaulacoideos y el origen de los roe-
dores y de los polimastodontes. An. Mus. Nac. Buenos Aires, 3, no. 1, pp. 81-192.

1904. Nuevas especies de mamiferos Cretaceos y Terciarios de la Republica Argen-
tina. An. Soc. Cien. Arg., 56, pp. 193-208 (1903); 57, pp. 162-175, 327-341 (1904);
58, pp. 35-71, 182-192, 225-291 (1904).

1906. Les formations sedimentaires du Cretace superieur et du Tertiare de Pata-
gonie avec un parallele entre leurs faunes mammalogiques et celles de 1'ancien
continent. An. Mus. Nac. Buenos Aires, (3) VIII, pp. 1-568.

Berggren, W. A. and J. A. Van Couvering

1974. The late Neogene. Paleogeography, Palaeoclimatology, Palaeoecology, 16,
nos. 1-2, pp. 1-216.

BORDAS. A. F.

1945. Geologia estratigrafica de algunas zonas de la Patagonia. An. Mus. Pata-
gonia, 1, pp. 139-184.

Cabrera. A.
1927. Datos para el conocimiento de los dasiuroideos f6siles Argentinos. Revta.
Mus. La Plata, 30, pp. 271-315.

Chaffee. R. G.
1952. The Deseadan vertebrate fauna of the Scarritt Pocket, Patagonia. Bull.
Amer. Mus. Nat. Hist., 98, no. 6, pp. 507-562.

Clemens, W. A. and L. G. Marshall
1976. Fossilium Catalogus. I: Animalia. American and European Marsupialia.
Pars. 123, pp. 1-114.

Feruglio, E.
1949. Descripci6n geolbgica de la Patagonia. Vol. II, pp. 1-349. Minst. Ind. y Com.
de la Naci6n, Div. Gen. Y.P.F.

Gaudry, A.
1906. Fossiles de Patagonie-Les attitudes de quelques animaux. Ann. Paleontol.,
1, no. 1, pp. 1-42.

Hartenberger. J. L.

1975. Nouvelles decouvertes de rongeurs dans le Deseadien (Oligocene inferieur)
de Salla-Luribay (Bolivie). C. R. Acad. Sci. Paris, 280D, pp. 427-430.

HOFFSTETTER. R.

1954. Les gravigrades cuirasses du Deseadan de Patagonie (note preliminaire).
Mammalia, 18, pp. 159-169.

1968. Un gisement de mammiferes Deseadiens (Oligocene inferieur) en Bolivie. C.
R. Acad. Sci. Paris, 267D, pp. 1,095-1,097.

1969. Un primate de l'Oligocene inferieur sud-americain: Branisella boliviano gen.
et sp. nov. C. R. Acad. Sci. Paris, 269D, pp. 434-437.

1976. Rongeurs caviomorphes de l'Oligocene de Bolivie. I. Introduction au Desea-
dien de Bolivie. Palaeovertebrata, Montpellier, 7, no. 3, pp. 1-14.

98 FIELDIANA: GEOLOGY, VOLUME 41

HOFFSTETTER, R., C. MARTINEZ, M. MaTTAUER, and P. TOMASI

1971. Lacayani, un nouveau gisement bolivien de mammiferes Deseadiens
(Oligocene inferieur). C. R. Acad. Sci. Paris, 273D, pp. 2,215-2,218.

Kraglievich, J. L. and S. Rivas
1951. Orophodon Amegh. representante de una nueva superfamilia Orophodontoi-
dea del suborden Xenarthra (nota preliminar). Com. Mus. Argent. Cienc. Nat.,
Cienc. Zool, 2, pp. 7-28.

Lavocat, R.
1976. Rongeurs caviomorphes de l'Oligocene de Bolivie. II. Rongeurs du Bassin
Deseadien de Salla-Luribay. Paleovertebrate, Montpellier, 7, no. 3, pp. 15-90.

Loomis, F. B.

1914. The Deseadan Formation of Patagonia. Concord, New Haven. 232 pp.
Marshall, L. G.

1976a. On the affinities of Pichipilus osborni Ameghino 1890 (Marsupialia, Caeno-
lestinae) from the Santa Cruz beds of southern Patagonia, Argentina. Ameghini-
ana, 13, no. 1, pp. 1-9.

1976b. Revision of the South American fossil marsupial subfamily Abderitinae
(Mammalia, Caenolestidae). Publ. Museo Munic. Cien. Nat. Mar del Plata
"Lorenzo Scaglia," 2, no. 3, pp. 57-90.

1978. Evolution of the Borhyaenidae, extinct South American predaceous mar-
supials. Univ. Calif. Publ. Geol. Sci., 117, pp. 1-89.

Marshall, L. G. and R. Pascual

1976. Nuevos marsupiales caenolestidae del "Piso Notohipidense" (SW de Santa
Cruz, Patagonia) de Ameghino. Sus aportaciones a la cronologia y evolucion de
las comunidades de mamiferos sudamericanos. Publ. Museo Munic. Cien. Nat.
Mar del Plata "Lorenzo Scaglia," 2, no. 4, pp. 91-120.

Marshall, L. G., R. Pascual, G. H. Curtis, and R. E. Drake

1977. South American geochronology— radiometric time scale for Middle to Late
Tertiary mammal-bearing horizons in Patagonia. Science, 195, pp. 1,325-1,328.

Osborn, H. F.

1910. The age of mammals in Europe, Asia, and North America. New York, Mac-
Millan Co., 635 pp.

Pascual, R. and O. E. Odreman Rivas
1971. Evoluci6n de las comunidades de los vertebrados del Terciario Argentino.

Los aspectos paleozoogeograficos y paleoclimaticos relacionados. Ameghiniana,

8, no. 3/4, pp. 372-412.
1973. Las unidades estratigraficas del Terciario portadores de mamiferos. Su

distribucion y sus relaciones con los acontecimientos diastroficos. Actas del

Quinto Congreso Geoldgico Argentino, 3, pp. 293-338.

Patterson, B.
In press. A new pampatheriine armadillo (Mammalia, Edentata) from the early
Oligocene Tremembe formation, Sao Paulo. Papeis Avulsos de Zoologia.

PATTERSON & MARSHALL: DESEADAN MASUPIALIA 99

Patterson, B. and R. Pascual
1968. New echimyid rodents from the Oligocene of Patagonia, and a synopsis of

the family. Breviora, no. 301, pp. 1-14.
1972. The fossil mammal fauna of South America, pp. 247-309. In: Keast, A., F. C.

Erk, and B. Glass. Evolution, mammals and southern continents. State Univ.

of New York Press, Albany.

Patterson. B. and A. E. Wood

In press. Rodents from the Deseadan Oligocene of Bolivia. Bull. Mus. Comp. Zool.

Paula Couto, C.
1952. Fossil mammals from the beginning of the Cenozoic in Brazil. Marsupialia:
Polydolopidae and Borhyaenidae. Amer. Mus. Novit., 1559, pp. 1-27.

Paula Couto. C. de and S. Mezzalira
1971. Nova ccnceituacao geocronol6gica de Tremembe, Estado de Sao Paulo,
Brasil. An. Acad. Bras. Cienc, 43 (suppl.), pp. 473-488. [Geology by Mezzalira,
paleomammalogy by Paula Couto.]

Reig. O. A.
1955. Un nuevo genero y especie de cenolestinos del Plioceno de la Provincia de
Buenos Aires (Republica, Argentina). Rev. Assn. Geol. Argentina, 10, pp. 60-71.

Riggs. E. S.
1934. A new marsupial saber-tooth from the Pliocene of Argentina and its relation-
ships to other South American predaceous marsupials. Trans. Amer. Phil. Soc.,
n.s. 24, pp. 1-31.

Riggs. E. S. and B. Patterson

1939. Stratigraphy of the late Miocene and Pliocene deposits of the province of
Catamarca (Argentina) with notes on the faunae. Physis, 14, pp. 143-162.

Roth, S.
1903. Noticias sobre nuevos mamiferos f6siles del Cretaceo superior y Terciario
inferior de la Patagonia. Rev. Mus. La Plata, 11, pp. 133-158.

Scott. W. B.
1937. A history of land mammals in the Western Hemisphere. Revised edition.
New York, The MacMillan Co., xiv-r 786 pp.

Simpson. G. G.
1936. Notas sobre los mamiferos mas antiguos de la colecci6n Roth. Inst. Mus.
Nac. La Plata, obra del Cincuentenario, 2, pp. 63-94.

1940. Review of the mammal-bearing Tertiary of South America. Proc. Amer.
Phil. Soc., 83, pp. 649-709.

1948. The beginning of the Age of Mammals in South America. Part I. Bull. Amer.

Mus. Nat. Hist., 91, no. 1, pp. 1-232.
1967a. The Ameghinos' localities for Early Cenozoic mammals in Patagonia. Bull.

Mus. Comp. Zool., 136, no. 4, pp. 63-76.
1967b. The beginning of the Age of Mammals in South America. Part II. Bull.

Amer. Mus. Nat. Hist., 137, pp. 1-259.

100 FIELDIANA: GEOLOGY, VOLUME 41

Sinclair, W. J.
1906. Mammalia of the Santa Cruz beds: Marsupialia. Rept. Princeton Univ.

Exped. Patagonia, 4, no. 3, pp. 333-460.
1930. New carnivorous marsupials from the Deseado Formation of Patagonia.

Field Mus. Nat. Hist., Geol. Mem., 1, no, 1, pp. 35-39.

TOURNOUER, A.

1903. Notes sur la geologie et la paleontologie de la Patagonie. Bull. Soc. Geol.
France, ser. 4, 3, pp. 463-473.

Trouessart, E. L.
1898. Catalogus mammalium tarn viventium quam fossilium. New ed. Berlin, pt. 5,
pp. 665-1,264.

VOLKHEIMER, W.

1973. Observaciones geologicas en el area de ingeniero Jacobacci y adyacencias
(Provincia de Rio Negro). Rev. Assn. Geol. Argent., 28, no. 1, pp. 13-36.

Wood, A. E.
1949. A new Oligocene rodent genus from Patagonia. Amer. Mus. Novit., no. 1435,
pp. 1-54.

Wood, A. E. and B. Patterson
1959. The rodents of the Deseadan Oligocene of Patagonia and the beginning of
South American rodent evolution. Bull. Mus. Comp. Zool., Harvard Univ., 120,
no. 3, pp. 281:428.