iddle Miocene vertebrates from the Amazonian Madre de Dios Subandean Zone, Perú

HAL Id: hal-00956894

https://hal.archives-ouvertes.fr/hal-00956894

Submitted on 7 Mar 2014

HAL

is a multi-disciplinary open access

archive for the deposit and dissemination of sci-

entific research documents, whether they are pub-

lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire

HAL

, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

Middle Miocene vertebrates from the Amazonian Madre

de Dios Subandean Zone, Perú

Pierre-Olivier Antoine, Martin Roddaz, Stéphanie Brichau, Julia Tejada-Lara,

Rodolfo Salas-Gismondi, Ali Altamirano, Mélanie Louterbach, Luc Lambs,

Thierry Otto, Stéphane Brusset

To cite this version:

Pierre-Olivier Antoine, Martin Roddaz, Stéphanie Brichau, Julia Tejada-Lara, Rodolfo Salas-

Gismondi, et al.. Middle Miocene vertebrates from the Amazonian Madre de Dios Subandean

Zone, Perú. Journal of South American Earth Sciences, Elsevier, 2013, vol. 42, pp. 91-102.

10.1016/j.jsames.2012.07.008. hal-00956894

pen

rchive

OULOUSE

rchive

uverte (

OATAO

)

OATAO is an open access repository that collects th

e work of Toulouse researchers and

makes it freely available over the web where possib

le.

This is an author-deposited version published in :

http://oatao.univ-

toulouse.fr/

Eprints ID : 11085

To link to this article

: doi:10.1016/j.jsames.2012.07.008

URL :

http://dx.doi.org/10.1016/j.jsames.2012.07.008

To cite this version

: Antoine, Pierre-Olivier and Roddaz, Martin and

Brichau, Stéphanie and Tejada-Lara, Julia and Salas

-Gismondi,

Rodolfo and Altamirano, Ali and Louterbach, Mélanie

and Lambs, Luc

and Otto, Thierry and Brusset, Stéphane Middle Mioc

ene vertebrates

from the Amazonian Madre de Dios Subandean Zone, Pe

rú. (2013)

Journal of South American Earth Sciences, vol. 42 .

pp. 91-102. ISSN

08959811

Any correspondance concerning this service should b

e sent to the repository

administrator:

staff-oatao@listes-diff.inp-toulouse.fr

Middle Miocene vertebrates from the Amazonian Madre de Dios Subandean Zone,

Perú

Pierre-Olivier Antoine

*

, Martin Roddaz

, Stéphanie Brichau

, Julia Tejada-Lara

Rodolfo Salas-Gismondi

, Ali Altamirano

, Mélanie Louterbach

, Luc Lambs

, Thierry Otto

Stéphane Brusset

Institut des Sciences de l

’

Évolution, UMR-CNRS 5554, CC064, Université Montpellier 2, Place Eugène Bataillon, F-34095 Montpellier, France

Géosciences-Environnement Toulouse, Université de Toulouse, UPS (SVT-OMP), LMTG, CNRS, IRD, 14 Avenue Édouard Belin, F-31400 Toulouse, France

Museo de Historia Natural-Universidad Nacional Mayor San Marcos, Departamento de Paleontología de Vertebrados, Avenida Arenales 1256, Lima 11, Perú

Institut Français d

’

Etudes Andines, Avenida Arequipa 4500, Lima 18, Perú

Institut Polytechnique Lasalle Beauvais, Département Géosciences, 19 rue Pierre Waguet, F-60026 Beauvais Cedex, France

EcoLab, UMR 5245 CNRS-UPS-INPT, Université de Toulouse, 118 Route de Narbonne, F-31062 Toulouse Cedex 9, France

Keywords:

Colloncuran-early Laventan

Marsupialia

Rodentia

Biochronology

Fission track age

Biogeography

abstract

A new middle Miocene vertebrate fauna from Peruvian Amazonia is described. It yields the marsupials

Sipalocyon

sp. (Hathliacynidae) and

Marmosa

(

Micoureus

) cf.

laventica

(Didelphidae), as well as an

unidenti

fi

ed glyptodontine xenarthran and the rodents

Guiomys

sp. (Caviidae),

Scleromys

sp., cf.

quadrangulatus-schurmanni-colombianus

(Dinomyidae), an unidenti

fi

ed acaremyid, and cf.

Micro-

steiromys

sp. (Erethizontidae). Apatite Fission Track provides a detrital age (17.1

2.4 Ma) for the locality,

slightly older than its inferred biochronological age (Colloncuran-early Laventan South American Land

Mammal Ages:

w

15.6

e

13.0 Ma). Put together, both the mammalian assemblage and lithology of the

fossil-bearing level point to a mixture of tropical rainforest environment and more open habitats under

a monsoonal-like tropical climate. The fully

fl

uvial origin of the concerned sedimentary sequence

suggests that the Amazonian Madre de Dios Subandean Zone was not part of the Pebas mega-wetland

System by middle Miocene times. This new assemblage seems to reveal a previously undocumented

spatiotemporal transition

between the late early Miocene assemblages from high latitudes (Patagonia

and Southern Chile) and the late middle Miocene faunas of low latitudes (Colombia, Perú, Venezuela, and

?Brazil).

1. Introduction

The middle Miocene period is characterised by the last climatic

optimum (MMCO, for Middle Miocene Climatic Optimum) before

a sustainable deterioration culminating with Pliocene

e

Pleistocene

ice ages (

Zachos et al., 2001

2008

). In Northern South America, this

MMCO coincides with the emergence of the

Pebas system

, a large

wetland with marine in

fl

uence partly engul

fi

ng what is lowland

Amazonia today (for review, see

Hoorn et al., 2010a

This environment was particularly favourable to biodiversity

and it allowed a wide array of organisms, such as molluscs,

arthropods, and plants to be fossilised (e.g.,

Hoorn, 1993

;

Antoine

et al., 2006

;

Pons and De Franceschi, 2007

;

Jaramillo et al., 2010

;

Wesselingh and Ramos, 2010

). By contrast, and with the notable

exceptions of La Venta (late middle Miocene, Colombia;

Kay et al.,

1997

) and Fitzcarrald local fauna (eastern Perú;

Antoine et al.,

2007

;

Goillot et al., 2011

;

Pujos et al., in press

), middle Miocene

mammals are virtually unknown in tropical-equatorial South

America (e.g.,

Negri et al., 2010

The present work aims to report a new vertebrate locality from

the Subandean Zone of Southwestern Peru, designated MD-67

(S12

38.683

; W71

19.284

;

w

428 m Above Sea Level) and doc-

umenting the concerned interval. MD-67 was discovered by one of

us (MR) nearby Pilcopata (Cusco) in 2007 (

Fig. 1

). The corre-

sponding results are exposed hereunder in systematic, bio-

stratigraphical, and biogeographical perspectives.

1.1. Geological context

The western Amazon drainage basin extends today from

southern Colombia to northern Bolivia (

Hoorn et al., 2010a

). Since

Pliocene times (

Espurt et al., 2007

2010

), the Amazonian foreland

*

Corresponding author. Tel.:

33 467 143 251; fax:

33 467 143 640.

E-mail address:

pierre-olivier.antoine@univ-montp2.fr

(P.-O. Antoine).

http://dx.doi.org/10.1016/j.jsames.2012.07.008

basin has been divided into two foreland basin systems (

sensu

DeCelles and Giles, 1996

): the North Amazonian foreland basin

system and the South Amazonian foreland basin, separated by the

Fitzcarrald Arch (

Roddaz et al., 2005

). The South Amazonian fore-

land basin system comprises the Southern Peruvian and Northern

Bolivian foreland basins. The Subandean Zone is part of the Madre

de Dios foreland basin (

Fig. 1

). The southernmost part of the Sub-

andean Zone is structured by a syncline (Salvación Syncline) fol-

lowed by a thrust-related anticline (Pantiacolla Anticline) and the

Madre de Dios blind thrust front (

Fig. 1

The fossiliferous outcrop dips 35

SW with a N130 strike (

Fig. 2

It crops out in the southern

fl

ank of the Pantiacolla anticline and

was originally mapped as part of the Paleocene

e

Eocene Yahuar-

ango Formation, based on sedimentary facies, but without any

biostratigraphic constraint (

Vargas and Hipólito, 1998

). The

Yahuarango Formation (northern Perú) is poorly dated and it

consists mainly of red siltstones and mudstones forming distal

fl

uvial deposits (see

Roddaz et al., 2010

for a review).

The fossiliferous level corresponds to a 1 m-thick sand-and-

gravel channel of

fl

uvial origin, with iron-rich pisolites and top-

ped by a ferruginous duricrust (

Fig. 2

A). These features coincide

with a lateritic pro

fi

le (e.g.,

Tardy, 1992

). The fossiliferous channel

developed over paleosol mudstones (

Fig. 2

A and B).

2. Material and methods

2.1. Paleontology

All the vertebrate fossil remains described here were recovered

either by hand-picking on the ferruginous crust topping the MD-67

locality during the exploratory 2007

fi

eld trip, or by excavating and

by screen-washing of the corresponding sediment during the 2009

fi

eld season. In 2009, ca. 200 kg of rough sediment were screen-

washed, using a 1 mm mesh. A new survey in 2011 unfortunately

showed the fossil-yielding outcrop had been washed away by the

Madre de Dios River in the meantime (

Fig. 2

C).

ragments of turtle plates and crocodile teeth were also recov-

ered in MD-67. As they do not display diagnostic features, they will

not be described here. By contrast, neither

fi

sh nor plant remains

were unearthed in MD-67.

Among mammalian remains, a mesio-labial fragment of a hyp-

selodont upper tooth referable to a toxodontid notoungulate was

recognised in MD-67. This specimen, of poor biochronological and

environmental use, is the only remain unambiguously assignable to

native ungulates in the concerned locality.

The nomenclature used for marsupial dentition is adapted from

that of

Goin and Candela (2004)

. Morphological features of the

cavioid rodent are described and diagnosed following the termi-

nology and phylogenetic characters proposed by

Pérez (2010)

. For

dental features of the dinomyid, octodontoid, and erethizontid

rodents, we follow the nomenclature of

Wood and Wilson (1936)

as modi

fi

ed by

Antoine et al. (2012)

Except when mentioned, dimensions are given in mm.

2.2. Apatite

fi

ssion track analysis (AFTA)

Apatite grains were mounted and polished for etching to several

the natural spontaneous

fi

ssion tracks. Apatites were etched using

5 N HNO

at 20

C for 20 s. Etched grain mounts were packed with

mica external detectors and corning glass (CN5) dosimeters and

irradiated in the FRM 11 thermal neutron facility at the University

of Munich in Germany. Following irradiation the external detectors

were etched using 48% HF at 20

C for 25 min. Analyses were

carried out on a Zeiss Axioplan microscope at a magni

fi

cation

1250, using a dry (

100) objective. Con

fi

ned track length

measurements were made using a drawing tube and digitising

tablet, calibrated against a stage micrometre. Single-grain AFT ages

were calculated using the external detector method and the zeta

calibration approach, as recommended by the I.U.G.S. Subcom-

mission on Geochronology (

Hurford, 1990

). Track length

measur

ements were restricted to con

fi

ned tracks parallel to the c-

crystallographic axis.

2.3. Institutional abbreviations

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

IGM, Ingeominas (Instituto Nacional de Investigaciones en Geo-

ciencias, Minería y Química, Museo Geológico, Bogotá, Colombia;

ISE-M, Institut des Sciences de l

’

Évolution, Montpellier, France;

MACN, Museo Argentino de Ciencias Naturales Bernardino Riva-

davia, Buenos Aires, Argentina; MLP, Museo de Ciencias Naturales

de La Plata, La Plata, Argentina; MNHN, Muséum National d

’

Histoire

Naturelle, Paris; MUSM, Museo de Historia Natural de la Uni-

versidad Nacional Mayor San Marcos, Lima, Perú; UCMP, University

of California Museum of Paleontology, Berkeley, USA; UFAC, Labo-

ratório de Pesquisas Paleontológicas, Universidade Federal do Acre,

Rio Branco, Brazil.

Fig. 1.

Location and geological map of the studied area, in the Amazonian Madre de

Dios Subandean Zone, Perú; modi

fi

ed after

Vargas and Hipólito (1998)

2.4. Other abbreviations

dist., distal; est., estimated; FAD, First Appearance Datum;

height; HI, Hypsodonty Index (

/MDL);

, length; max, maximal;

LLO, Last Local Occurrence; MDL, mesio-distal length; mes, mesial;

SALMA, South American Land Mammal Age;

, width.

3. Systematic Paleontology

Class MAMMALIA Linnaeus, 1758

Cohort MARSUPIALIA Illiger, 1811

Order DIDELPHIMORPHIA Gill, 1872

Suborder SPARASSODONTA Ameghino, 1894

Superfamily BORHYAENOIDEA Ameghino, 1894

Subfamily HATHLIACYNIDAE Ameghino, 1894

Sipalocyon

Ameghino, 1887

Sipalocyon

sp.

MUSM 1967 is a much worn and eroded tooth, with a tribos-

phenic bunodont pattern and three broken roots (one lingual and

two labial ones). Following the work of

Forasiepi et al. (2006)

, this

tooth is interpreted as a right M3 of a carnivorous marsupial. It is

roughly triangular in occlusal view, with a straight mesial side (mes

5.10), a long and bilobed distolingual border (max

6.69),

and a concave labial side (MDL

4.67). Crown height equals

3.24 mm. The paracone and the metacone have their bases twin-

ned, thus determining a very short premetacrista. Even if its tip is

broken away and eroded, the paracone was low, with a faint pre-

paracrista; the most prominent cusp is the metacone, which

occupies a central position in occlusal view (

Fig. 3

A and B). The

paraconule and the metaconule are equally developed, and closely

appressed to the paracone and metacone, respectively (

Fig. 3

A and

B). The protocone is bulbous, much lower than the latter cusps, and

widely remote from them, which further determines a wide

triangular basin which is bordered by a faint preprotocrista and

a low postprotocrista. The postmetacrista is strongly developed and

distally convex, with a

fl

at and high vertical wear facet but without

carnassial notch (

Fig. 3

B). The stylar shelf is smooth and reduced,

with faint or absent cusps. Yet, a rounded crest surrounds the

mesiolabial corner of the tooth, which was either interpreted as an

ectocingulum or a

Cusp B

(

Marshall, 1978

;

Forasiepi et al., 2006

The main morphological features of this tooth (tribospheny,

reduced protocone, paracone and metacone closely appressed, and

reduced stylar shelf and cusps) point to a carnivorous marsupial, as

fi

ned by

Muizon (1999: 502)

. The presence of a thick and low

protocone, and of equally developed paraconule and metaconule

allow referring it as to a medium-sized hathliacynid borhyaenoid

(

Forasiepi et al., 2006

). Accordingly, comparison was made with

Cladosictis

Ameghino, 1887 (early Miocene of Argentina and Chile;

Flynn et al., 2002

)

Sipalocyon

Ameghino, 1887 (Colhuehuapian-San-

tacrucian of Patagonia;

Marshall, 1981

Acyon

Ameghino, 1887

(middle Miocene of Southern Bolivia;

Forasiepi et al., 2006

), and

Notogale

Loomis, 1914 (Deseadan of Bolivia and Argentina;

Muizon,

1999

). A similar large-sized lingual basin is only observed in

Noto-

gale mitis

(Ameghino,1897), as illustrated by

Marshall (1981: Figs. 54

and 55)

. Based oncompatible proportions/size and onthe presence of

an ectocingulum, MUSM 1967 shows the closest af

fi

nities with the

M3 of

Sipalocyon gracilis

Ameghino, 1887 (AMNH 107401-001; FM

9254-001; Santacrucian) and

Sipalocyon externa

(Ameghino, 1902)

(Colhuehuapian;

Marshall, 1981

). It differs from

S. externa

in having

a lower and smoother protocone, a very wide lingual basin, labially-

displaced paracone and metacone, and in showing a more developed

paraconule and metaconule (

Marshall, 1981

), but closely resembles

the M3 of

S. gracilis

from

Río Coyle

(maxilla MNHN SCZ 122,

Fig. 2.

Middle Miocene vertebrate locality MD-67, in the Amazonian Madre de Dios Subandean Zone, Perú. A. Detail of the fossil-yielding red sand-and-gravel channel, developing

over red clays (paleosol), as discovered in October 2007. Note the dipping of the topping surface (ferruginous duricrust). Picture by Patrice Baby. B. Detail of the transition between

the underlying red clays (top) and the gravel channel, with iron-rich pisolites (black balls, at the bottom). Picture by Patrice Baby (same day as in

Fig. 2

A). C. The same area in August

2011. The fossil-yielding channel (denoted by the white dotted line) had been washed away by the Madre de Dios River in the meantime. Picture by Laurent Marivaux. (For

interpretation of the references to colour in this

fi

gure legend, the reader is referred to the web version of this article.)

Tournouër Collection; Santa Cruz Formation, Patagonia) in all these

aspects.

Sipalocyon

has a Colhuehuapian

e

Friasian range, so far

restricted to middle and high latitudes (37.5

e

Marshall, 1981

;

Flynn et al., 2008

). Recognition of a hathliacynid closely allied to

Sipalocyon

, here referred as to

Sipalocyon

sp., widely extends north-

ward the geographical range of the genus.

Superfamily DIDELPHOIDEA Gray, 1821

Family DIDELPHIDAE Gray, 1821

Subfamily DIDELPHINAE Gray, 1821

Tribe MARMOSINI Reig, Kirsch and Marshall in Reig, 1981

Genus

Marmosa

Gray, 1821

Fig. 3.

Fossil mammal remains from the MD-67 locality, middle Miocene of the Amazonian Madre de Dios Subandean Zone, Perú. A

e

Sipalocyon

sp., right M3 (MUSM 1967) in

labial (A) and occlusal views (B). C.

Marmosa

(

Micoureus

) cf.

laventica

, left M3 lacking the protocone (MUSM 1968). Reconstructed areas appear in grey (reconstruction based on

Marshall, 1976

: text-Fig. 2). D

e

E. Glyptodontinae indet., isolated osteoderms in external view. D, MUSM 1720; E, MUSM 1585. F

e

Guiomys

sp., left M1/2 (MUSM 1970) in occlusal

(F), labial (G), and mesial views (H). I

e

Scleromys

sp., gr.

quadrangulatus-schurmanni-colombianus

. Left d4 (MUSM 1971) in occlusal view (I). Left m1 (MUSM 1972) in occlusal (J)

and labial views (K). L, Acaremyidae indet., left m1/2 (MUSM 1973) in occlusal view. M

e

N, cf.

Microsteiromys

sp. Left P4 (MUSM 1974) in occlusal view (M). Right ?M3 (MUSM 1975)

in occlusal view (N). Scale bar

2 mm (A

e

B, F

e

K), 5 mm (

e

E), or 1 mm (C, L

e

N).

Subgenus

Micoureus

Lesson, 1842

sensu

Voss, and Jansa, 2009

101

Marmosa

(

Micoureus

) cf.

laventica

Marshall, 1976

MUSM 1968 is a left upper molariform tooth, the protocone of

which was broken away. This tooth has a dilambdodont tribos-

phenic pattern, a large stylar shelf, and four stylar cusps (termed A

e

D). It bears one lingual and two labial roots. The tooth is almost

unworn and its dimensions are very small (MDL

1.75; est

1.8). The metacone is much larger and higher than the para-

cone. There was neither paraconule nor metaconule. All these

features point to a marmosine didelphid rather than to a micro-

biothere (

Marshall, 1976

). MUSM 1968 was compared to the M1

e

M3 series UCMP 108563 of

laventica

Marshall, 1976

from the

middle Miocene of La Venta, Colombia, described and illustrated by

Marshall (1976: text-Fig. 2)

, as well as to the isolated M1 IGM

251011 and M3 IGM 250278 of the same species and area, as

described and

fi

gured under the name

laventicus

Goin

(1997: Fig. 11.2)

. Its isometric proportions, moderate dissymmetry

(with a lingual part slightly mesially displaced), and concave labial

border in occlusal view allow identifying it as an M3. The missing

part of the tooth was virtually reconstructed, based on available

illustrations of

(

laventica

(

Fig. 3

C). There was neither junc-

tion between the preprotocrista and the paracone nor between the

postprotocrista and the metacone. The stylar cusp A is disconnected

from the paracone and from the preparacrista. The stylar cusp B

appears as the strongest cusp, as it does in fresh teeth of most

marmosines (

Rossi et al., 2010

: 22). There is a deep mesiolabial

notch for the metastyle of M2, but neither labial, nor distal

cingulum, as in

(

laventica

(

Marshall, 1976

) and the living

species

(

demerarae

Thomas, 1905 (ISE-M V-1590). A thick

cingulum, running along the labial half of the mesial side, was

connecting the stylar cusp A and the protocone through the pre-

protocrista, as in the isolated M3 IGM 250278 of

(

laventica

(

Goin, 1997

). This cingulum is much stronger and the tooth ca. 30%

smaller than in

(

demerarae

. Even if their overall size is

comparable, the stylar cusps are much stronger and the notch for

the metastyle of M2 is much deeper in MUSM 1968 than in the M3s

of the living

Marmosa

(

Marmosa

)

murina

Linnaeus,1758 (Linnaeus

’

Mouse Opossum; ISE-M V-1109).

Basically, the size, proportions, and morphological features of

MUSM 1968 closely match those of the M3 of

(

laventica

from

La Venta area, Colombia, i.e., the only fossil species described and

named for this genus so far (

Goin, 1997

). The stratigraphical range

(

laventica

spans the La Victoria Fm. and the base of the

Baraya Mb. of the overlying Villavieja Fm. in La Venta region,

Colombia (

w

13.5

e

13.0 Ma;

Goin, 1997

;

Madden et al., 1997

: 511;

Gradstein et al., 2005

). If the close af

fi

nities of MUSM 1968 with

Marmosa

(

Micoureus

) are con

fi

rmed, this would be the second fossil

occurrence of this living (sub-) genus (

Goin, 1997

Infraclass EUTHERIA Gill, 1872

Order XENARTHRA Cope, 1889

Suborder CINGULATA Illiger, 1911

Superfamily GLYPTODONTOIDEA Gray, 1869

Family GLYPTODONTIDAE Gray, 1869

Subfamily GLYPTODONTINAE Gray, 1869

Genus and species indet.

Two isolated carapace osteoderms were found. They are small

(MUSM 1720:

16, preserved

22; MUSM 1585:

13,

17),

quadrangular to pentagonal in shape, with large round principal

fi

gures near the posterior edge. In both osteoderms, the principal

fi

gure is completely surrounded by peripheral ones

e

eight in MUSM

1720 (

Fig. 3

D) and six in MUSM 1585 (

Fig. 3

e

although medial,

lateral, and posterior

fi

gures are reduced in size. The principal

fi

gure

is slightly convex on MUSM 1720, while it is

fl

at on MUSM 1585.

Osteoderms are thin (MUSM 1720: 3.5 mm; MUSM 1585: 4.8 mm),

with a punctuated surface, and serrated edges. The principal and

radial sulci are wide and shallow. Piliferous pits are large and located

in some intersections of the principal sulcus and the radial sulci

(

Fig. 3

e

E).

The concerned osteoderms cannot be con

fi

dently identi

fi

ed at

genus or species level because their characteristics are widely

distributed in several clades. However, they are likely to document

a single small and unknown taxon within the Glyptodontinae

(smaller than the oldest glyptodontine genus known so far,

Bor-

eostemma

from Laventan SALMA), because of the combination of

the following characters: (i) osteoderms

fl

at or smoothly convex,

(ii) surface punctuated, (iii) principal

fi

gure completely surrounded

by peripheral

fi

gures, and (iv) quadrangular

-shaped cross

section of the sulci. The latter feature has been suggested as

a noncranial character that differentiates Glyptodontinae from

Propalaehoplophorinae within Glyptodontidae (

Carlini et al., 2008

a consequence, a reduced size

e

almost twice smaller than the

corresponding osteoderms in

Boreostemma

e

and the presence of

thin osteoderms might represent a plesiomorphic condition for the

Glyptodontinae. On the other hand, the posteriorly displaced

position of the principal

fi

gure, as well as the reduction in size of

posterior, medial, and lateral peripheral

fi

gures, appear as tied to

the location of the osteoderms in the carapace (i.e., submarginal),

instead of representing a plesiomorphic trait (i.e., principal

fi

gure

close to the posterior margin), as seen on glyptatelines and

Para-

propalaehoplophorus

(see

Croft et al., 2007

Order RODENTIA Bowdich, 1821

Infraorder HYSTRICOGNATHI Tullberg, 1899

Parvorder CAVIOMORPHA Wood, 1955

Superfamily CAVIOIDEA Fischer de Waldheim, 1817

Family CAVIIDAE Fischer de Waldheim, 1817

Guiomys

sp.

MUSM 1970 is a very small tooth (

2.93;

2.61), hyp-

selodont (i.e.,

euhypsodont

following

Pérez, 2010

), preserving the

occlusal surface, but the basal most part of which crown is broken

away (

Fig. 3

e

H). This tooth is interpreted as a left M1 or M2. It is

bilobed, with lozenge-shaped lobes connected by a thick labial

bridge; the mesial lobe is slightly smaller than the distal one

(

Fig. 3

F). The occlusal surface of both lobes lacks any fossette and

the apex of each lobe is not constricted at the current wear stage.

The hypo

fl

exus is transversely developed (reaching c. 80% of the

transverse width of the crown) and funnel-shaped; cement is

fi

lling

the bottom of the hypo

fl

exus, which is pointing distally and labially.

A shallow furrow faces the tip of the hypo

fl

exus. There is no

transverse dentine crest crossing each lobe (

Fig. 3

F). Enamel is

present all around the crown but interrupted on the labial most

part of each lobe.

The bilobed pattern of this molar is characteristic of cavioid

caviomorph rodents (

Pérez, 2010

;

Croft et al., 2011

). Among Cav-

ioidea, the hypselodont pattern of this tooth impedes its referral to

Asteromys

Luantus

Chubutomys

, and

Phanomys

(

Pérez, 2010

). The

funnel-shaped hypo

fl

exus points to the clade including

Eocardia

robusta

Guiomys unica

, and crown cavioids (

Pérez, 2010

). The

absence of transverse dentine crest on each lobe discards referring

this tooth as to

Prodolichotis pridiana

Orthomyctera chapadma-

lense

, while the funnel-shaped hypo

fl

exus further distinguishes

MUSM 1970 from

P. pridiana

(

Walton, 1997

;

Pérez, 2010

). The

presence of a single labial furrow makes it distinct from upper

molars of

robusta

. To sum up, the morphology of this specimen

points to a stem cavioid and, together with its small size, it is only

consistent with

Guiomys

Pérez, 2010

, the smallest and earliest

caviid known so far, from the middle Miocene of Patagonia (?Col-

loncuran and Laventan SALMAs;

Pérez, 2010

;

Pérez and Vucetich,

) and the Laventan of Quebrada Honda, Bolivia (

Croft et al.,

2011

). This left molar is provisionally referred to as

Guiomys

sp.,

as it is 15

e

20% smaller than the teeth of the type and only species of

the latter genus (

Pérez, 2010

;

Croft et al., 2011

Superfamily CHINCHILLOIDEA Kraglievich, 1940

Family DINOMYIDAE Peters, 1873

Subfamily POTAMARCHINAE Kraglievich, 1926

Scleromys

sp., gr. quadrangulatus-schurmanni-colombianus

MUSM 1971 is a left d4. It is low-crowned and elongate mesio-

distally, with a trapezoid occlusal outline, widening distally

(MDL

5.04;

2.81;

3.10; HI

0.62). It has an

F-

G

-I-I

lophid

pattern, with distinctive features such as taeniodonty (see

Antoine

et al., 2012

), elongate cristids connected distolingually to the ante-

rolophid and the metalophid, a circular supernumerary cuspid

mediodistal to the anterolophid, and a short transverse supernu-

merary lingual lophid distal to the metalophid (

Fig. 3

I). The hypo-

lophid and the posterolophid are straight, oblique, distinct, and

parallel one to another. The hypolophid displays a small mesiolabial

spur. The enamel band is crenulated on the mesial side of the

posterolophid. Such crenulations occur frequently in potamarchine

teeth (especially in

Potamarchus murinus

;

Sant

’

Anna Filho, 1994

This tooth is twice smaller than d4s of

Olenopsis

/Drytomomys

from

the Middle Miocene of La Venta (Colombia;

Fields,1957

: 330, Fig.18;

Candela and Nasif, 2006

) and of Fitzcarrald (Perú;

Antoine et al.,

2007

). In addition, MUSM 1971 is quite distinct from the d4 of

murinus

from the middle to late Miocene of the Upper Juruá, Brazil

(AMNH58535;

Sant

’

Anna Filho,1994

), in having a trapezoid occlusal

outline and in being tetralophodont, instead of being quadrangular

and pentalophodont, respectively. MUSM 1971 is tetralophodont,

half the size of, and much less hypsodont (HI

0.62) than the

pentalophodont p4s referred to

Simplimus indivisus

(MLP 15-244a:

w

2.92; Laguna Blanca/Río Fénix, Friasian SALMA, Patagonia;

Vucetich, 1984

) and to

Simplimus

sp. (UFAC DGM 533M: HI

w

2.88;

Upper Juruá, middle-late Miocene, Brazil;

Sant

’

Anna Filho,1994

). On

the other hand, its size and proportions recall those of d4s of

schurmanni

from La Venta, Colombia, as illustrated by

Fields (1957:

285, Fig. 5)

. The pattern is quite similar, especially for the distal

lophids. Yet, the occlusal pattern of mesial lophids and cuspids

observed in MUSM 1971 prevents from referring it to

schur-

manni

. Teeth of

colombianus

(La Venta;

Fields, 1957

;

Walton,

1997

) are ca. 20% larger than both

schurmanni

and MUSM 1971.

MUSM 1972 is a left lower molariform and hypsodont tooth,

with a quadrangulate occlusal contour and a straight anterior

border (

Fig. 3

J). It is high, elongate mesiodistally and at an early

stage of wear (MDL

4.14;

3.72;

8.4; HI

2.03). The

mesiodistal and labiolingual lengths of the tooth diminish and

increase strongly with wear, respectively, which allows interpret-

ing it as an m1 at

stage of wear n

, by comparison with serial

sections of molariform teeth of

colombianus

as proposed by

Fields (1957: 318

e

319, Fig. 14)

. Roots are still developed and the

neck is well-marked (

Fig. 3

K). This tooth is tetralophodont and

taeniodont, with oblique and large lophids. The short metalophulid

I and the longer metalophulid II are connected at both lingual and

labial ends. The hypolophid and the posterolophid (damaged in its

distolingual part) are not connected lingually at the current stage of

wear, but they would join in later stages. Contrary to other lophids,

the posterolophid is curved and concave mesiolingually in occlusal

view, joining a thick and mesially-displaced hypoconid.

These two lower teeth have a typical dinomyid occlusal pattern

(i.e., high-crowned, tetralophodont, taeniodont, with oblique

lophids, and a posterolophid isolated before wear). By their

dimensionsand pattern,theyare provisionallythought to document

a single taxon.

Scleromys

Ameghino, 1887 is the earliest represen-

tative of Dinomyidae (

Horovitz et al., 2006

). The high crown,

mesiodistal elongation, quadrangular contour, large and persistent

fossettids, and straight aspect of the mesial border of MUSM 1972

discard any referral to the typical Santacrucian species of

Scleromys

i.e., the type species

Scleromys angustus

Ameghino, 1887 and

Scle-

romys osbornianus

Ameghino 1894 (Argentinian Patagonia). The

present specimen seems to be somewhat closer to

S. quadrangulatus

(pre-Santacrucian

Pinturan

age, Patagonia, Argentina;

Kramarz,

2006

) and to

Scleromys

sp. (Mariño Formation, ?early Santa-

crucian, Mendoza, Argentina;

Cerdeño and Vucetich, 2007

) in

sharing both a small size and a quadrangular contour. MUSM 1972

further resembles

S. quadrangulatus

in having a straight mesial

border, and a mesiodistal elongation. However, the lower molars

referred to

S. quadrangulatus

are lower-crowned than MUSM 1972

and they lack persistent fossettids. The latter feature is so far only

observedin

.sp.fromtheMariñoFormation(

CerdeñoandVucetich,

2007

schurmanni

Stehlin, 1940 and

colombianus

Fields,

1957

, from the Laventan of Colombia and Peru (

Fields, 1957

;

Walton, 1997

;

Antoine et al., 2007

). Similarly, in the Patagonian

species of

Scleromys

, the metalophulid II is typically reduced, being

shorter than the metalophulid I. In that aspect, MUSM 1972 seems to

be closer to the Laventan species from Colombia in having a larger

metalophulid II. Nevertheless, in the latter taxa (the generic

assignment of which is challenged by most authors; e.g.,

Patterson

and Wood, 1982

;

Walton, 1997

), teeth are noticeably larger,

higher-crowned and wider labiolingually than in MD-67. Regardless

of thelingualhypolophid

e

posterolophidjunction (notyetoccurring

in MUSM 1972), the MD-67 specimen matches an m3 referred to

Scleromys

cf.

S. schurmanni

from the Laventan Fitzcarrald local

fauna of Peruvian Amazonia (MUSM 1566;

Antoine et al., 2007

;

Negri et al., 2010

), being only narrower labiolingually and lower-

crowned. In addition, this lower molar has similar proportions and

lophid pattern as an m2 from the middle to late Miocene of the

Upper Juruá, Brazil (UFAC DGM 582M;

Sant

’

Anna Filho, 1994

)

referred to

colombianus

, by reference to specimens from La

Venta, Colombia. However, the tooth from MD-67 is distinctly

smaller, lower-crowned, and less prismatic.

To sum up, the dental pattern and the occlusal contour of MUSM

1972 match those of several species referred to

Scleromys sensu lato

either of pre-Santacrucian (

S. quadrangulatus

) or of Laventan ages

(

colombianus

and

schurmanni

;

Fields, 1957

). These taxa are

likely to form a

lineage

distinctive from the typical Santacrucian

Scleromys

cluster, as hypothesised by

Kramarz (2006: 59)

. Following

that scheme, the crown height, proportions, and fossettid develop-

ment of MUSM 1972 shall coincide with a

transitional evolutionary

stage

between

quadrangulatus

from the early Miocene of

Argentina and the representatives of

Scleromys

from the late

middle Miocene of Northern South America (

Fields,1957

;

Sant

’

Anna

Filho, 1994

;

Walton, 1997

;

Kramarz, 2006

;

Antoine et al., 2007

;

Cerdeño and Vucetich, 2007

). As such, these specimens, provision-

ally referred to as

Scleromys

sp., may document a

Friasian

Colloncuran morphological grade for potamarchine dinomyids.

Superfamily OCTODONTOIDEA Waterhouse, 1839

Family ACAREMYIDAE Ameghino, 1902

Genus and species indet.

MUSM 1973 is a diminutive lower molar (MDL

1.72;

2.10),

roughly square in occlusal view, and with a rather high crown

(

mesodont

sensu

Vucetich and Kramarz, 2003

). It is interpreted as

a left m1 or m2. This tooth has a tetralophodont pattern, with bulky

lophids and no cuspid well individualised (

Fig. 3

L). Enamel is thick,

especially on the labial half of the molar. In occlusal view, the mesial

margin is straight and the distal one convex. The lophids are thick,

transversely oriented, and separate by deep and wide

fl

exids. The

hypo

fl

exid is deep (it almost reaches the lingual half of the tooth),

nearly transverse, and

U-shaped

in occlusal view. The talonid is

narrower than the

trigonid

, due to the weak lingual development

of the posterolophid, thus providing a trapezoid shape to the

occlusal contour of the tooth. The metalophulid I is complete and

straight, with a thick mesiolabial projection, pointing linguodistally

and joining the metalophulid II. The metalophulid II is long, and it

also joins the metalophulid I at the lingual margin of the tooth,

determining a mesial fossettid, oval and elongate transversely

(closed antero

fl

exid). The junction between both metalophulids is

somewhat constricted lingually, mesial to the mesostylid. The

ectolophid is distinct and oblique. The hypolophid is transverse,

and it connects lingually a smoothly constricted entoconid. The

meso

fl

exid is open lingually until a late stage of wear. The meta-

fl

exid is also open lingually, and it would not get closed even with

heavy wear. The posterolophid is curved, with a strong hypoconid

and a short lingual arm (getting longer with wear).

Both the tetralophodont pattern and small size of MUSM 1973

allow attributing it to an early diverging octodontoid. The absence of

an eight-shaped-pattern due to the lingual opening of the meta

fl

exid

discards referring this tooth to the Octodontidae (

Vucetich and

amarz, 2003

). Among Echimyidae, the transverse orientation of

the lophids impedes any referral to Adelphomyinae, while tetralo-

phodonty distinguishes it from Echimyinae (

Vucetich et al., 1993

;

Kramarz, 2001

). MUSM 1973 resembles

Protacaremys

(Colhuehua-

pian-Colloncuran SALMAs of Patagonia and Chile;

Flynn et al., 2008

)

in having deep, transversely elongate and persistent fossettids/

fl

ex-

ids (notably the antero-fossettid/-

fl

exid), but differs from it in the

absence of pinched and cuspidate lophids, in the U-shaped, trans-

versely oriented and smooth hypo

fl

exid, in the complete and early

lingual junction between the metalophulids I and II (anterolophid

and mesolophid

sensu

Vucetich and Kramarz, 2003

), and in lacking

fi

gure-eight occlusal pattern (

Kramarz, 2001

On the other hand, the dental pattern of MUSM 1973 is strongly

reminiscent of Acaremyidae (

Acaremys

Sciamys

, and

Galileomys

;

Vucetich and Kramarz, 2003

). It differs from the m1s and m2s of

Galileomys

in a marked mesodonty, the presence of transverse

lophids,

fl

exids, and fossettids, a stronger transverse development

of the U-shaped hypo

fl

exid, and in the lingual connection between

the two mesial lophids (

Vucetich and Kramarz, 2003

;

Kramarz,

2004

). It resembles more

Sciamys

and

Acaremys

. Nevertheless, it

can be distinguished from

Sciamys

, as illustrated by

Arnal and

Vucetich (2011)

, by the quadrangular occlusal contour (more elon-

gate mesiodistally in

Sciamys

), by a shorter labial arm of the post-

erolophid, and by smaller dimensions. It differs from known species

Acaremys

in having a mesolophid reaching the lingual wall and

determining a transversely elongate anterofossettid persistent until

late stages of wear, two features which are only observed in

Sciamys

among acaremyids (

Vucetich and Kramarz, 2003

). It further differs

from

Acaremys

in having a much smaller metaconid, a narrower and

transversally longer meso

fl

exid, a less penetrating hypo

fl

exid, and

a posterolophid narrower mesiodistally. As a result, MUSM 1973 is

likely to document a late representative of Acaremyidae with closer

fi

nities to

Sciamys

and

Acaremys

than to

Galileomys

, provisionally

identi

fi

ed as

Acaremyidae indet

Superfamily ERETHIZONTOIDEA Thomas, 1897

Family ERETHIZONTIDAE Thomas, 1897

cf.

Microsteiromys

sp.

Two upper molariform teeth are assigned to a small

erethizontid.

MUSM 1974 is a left P4 with a circular occlusal outline

(

2.14;

2.20). It is preserved at a very early stage of wear

(just erupted). Accordingly, there is no wear facet due to the

contact with a mesial or a distal tooth (

Fig. 3

M). The crown is

much higher lingually than labially. The tooth was biradiculate,

with lingual and labial roots mesiodistally elongate. The occlusal

pattern is restricted to the apical-most part of the crown, with

distinct cusps. The tooth is tetralophodont, with a thin but

complete anteroloph, curved distolingually, with a prominent

protocone. The para

fl

exus is triangular in occlusal view, and open

both lingually and labially at the current stage of wear. The pro-

toloph is oblique, short lingually, and independent from adjacent

lophs at both ends; the lingual protoloph is a low and constricted

crest. The paracone is distinct and bulky. The hypo

fl

exus is deep

and wide (i.e., there is no endoloph). The lingual cusps, still

distinct at the current stage of wear, would be coalescent after

a smooth wear. The meso

fl

exus is open lingually and medially;

the mure is low and constricted, interrupted at the current stage

of wear. The third loph is very short, restricted to the centro-

distal part of the crown, and interpreted as a mesolophule; it is

connected lingually to the anterior arm of the hypocone and

labially, to the mesoloph. The posteroloph is thin, with a hypo-

cone and a metacone distinct and bulky. It extends mesiolabially

to the metacone until a faint cusp, interpreted as a mesostyle and

facing the paracone. The postero

fl

exus, still open mesiolabially,

would be quadrate at later stages of wear.

A right pentalophodont molar (MUSM 1975;

2.44 mm;

2.11 mm) has three broken roots, with two labial roots

(circular cross section) and a mesiodistally elongate lingual one,

which point to an upper molar (

Fig. 3

N). A contact facet is visible

on the mesial side of the crown, but not on the distal side. Given

the stage of occlusal wear, it allows identifying this tooth as

a probable M3, even if the tooth is only slightly tapering distally (as

in M2s) and if the hypocone is also located as in an M2, i.e.,

lingually displaced with respect to what is generally observed in

M3s. The occlusal pattern is longer than wide, with a

EUJ

-shaped

occlusal design. All the lophs are moderately worn, with slight

constrictions, and a low obliquity. The anteroloph is the most

developed loph, and it connects a labiolingually compressed and

oblique paracone. The para

fl

exus is oval, elongate transversely,

open labially but closed lingually by a low and constricted lingual

protoloph. The protoloph is straight, slightly oblique, and it

connects a strong paracone, labially. It joins the mesolophule

lingually, through a strong but very short mure, oriented sagittally.

The mesostyle and the paracone are remote and the meso

fl

exus is

wide and U-shaped in occlusal view (open labially). The meso-

lophule is transversely oriented and S-shaped, with a labiodistally-

oriented mesostyle. The meta

fl

exus is wide, comma-shaped, still

labially open at this stage of wear, and closed lingually by a labio-

lingually compressed hypocone and its strong anterior arm. The

posteroloph is short, but it extends mesiolabially until the meta-

cone, thus forming a distolabial wall. The posteroloph is somewhat

constricted labially to the hypocone and lingually to the metaloph.

The latter is short, oriented mesiolabially, connected to the mid-

posteroloph and disconnected to both the distolabial wall and

the metacone, thus forming a small posterofossette joined labially

to the meta

fl

exus.

The pentalophodont pattern of the upper molar, the low loph-

obliquity, and the circular occlusal outline of the P4, as well as the

low crown and thick enamel of both teeth, point to erethizontids.

Given their compatible size and pattern, both teeth are assumed to

document a single diminutive taxon. The fossil record of Erethi-

zontidae ranges from the Late Oligocene up to Recent times

(

Vucetich et al., 1999

;

Candela and Morrone, 2003

). Accordingly

and whenever possible, these teeth were compared to those of

Protosteiromys

(Deseadan, Patagonia;

Wood and Patterson, 1959

Eosteiromys

Ameghino, 1902,

Hypsosteiromys

Patterson, 1958

Parasteiromys

Ameghino, 1903, and

Branisamyopsis

Candela, 2003

(Colhuehuapian, Argentina;

Patterson, 1958

;

Candela, 1999

2003

;

Dozo et al., 2004

;

Kramarz, 2004

;

Kramarz and Bellosi,

2005

;

Vucetich et al., 2010

), of

Steiromys

Ameghino, 1887 (Santa-

crucian-?Laventan; Patagonia, ?Colombia;

Fields, 1957

;

Walton,

1997

;

Candela, 1999

), of

Neosteiromys

Rovereto, 1914 (?Collon-

curan-Huayquerian; Northwestern Argentina;

Candela, 1999

2004

), of

Microsteiromys

Walton, 1997

(Laventan; Colombia;

Walton, 1997

), and of

Erethizon

Cuvier, 1823 (Recent;

Candela,

1999

The teeth from MD-67 are half the size of than those of all

known Recent and fossil Erethizontidae but

Microsteiromys jacobsi

Walton, 1997

, of compatible size (

Walton, 1997

). No P4 is described

from any Deseadan erethizontid. The P4 from MD-67 differs from

all known Recent and fossil Erethizontidae but

Eosteiromys

homogenidens

Ameghino, 1902, in having a labially-displaced

hypocone. A circular occlusal outline is only observed in

E. homogenidens

(holotype MACN A-52-165) and ?

Steiromys

from La

Venta (

Walton, 1997

). Tetralophodonty and low loph obliquity also

characterises the P4 of

Eosteiromys

, ?

Steiromys

, and

Steiromys

detentus

Ameghino, 1887 (MLP 15

e

17). MUSM 1974 further differs

from the dP4 of

Parasteiromys uniformis

(Ameghino,1903) in having

a marked mure, and in bearing

fl

exi labiolingually compressed

(

Candela, 1999

). Its para

fl

exus, which is closed labially, makes it

distinct from the P4 of

Branisamyopsis praesigmoides

Kramarz, 2004

and

detentus

(

Kramarz, 2004

: 14, Fig. 7). It also differs from the P4

Neosteiromys bombifrons

Rovereto,1914 in having no lingual wall

(

Candela, 2004

: 61, Fig. 3).

The upper molar from MD-67 differs from all known erethi-

zontids in its labially-interrupted metaloph and in its mesiodistal

elongation (e.g.,

Candela, 1999

2004

;

Vucetich et al., 2010

). Its size

is also very distinctive, as it is twice smaller than in all known

taxa, except

jacobsi

, of similar dimensions (

Walton, 1997

). The

pentalophodont pattern makes it distinct from

Protosteiromys

medianus

Hypsosteiromys axiculus

, and

Steiromys

(

Patterson, 1958

;

Wood and Patterson, 1959

;

Candela, 1999

;

Dozo et al., 2004

). The

presence of a mure and the late labial closure of the para

fl

exus

discard any referral to

Parasteiromys

Neosteiromys

(

Candela,

1999

The presence of a distal constriction labially to the hypocone and

lingually to the metaloph is only observed in

Parasteiromys friantae

Candela, 1999

(but not in

P. uniformis

;

Candela, 1999

: Fig. 2),

homogenidens

(type, MACN A-52

e

165) and

. cf.

homogenidens

(

Kramarz, 2004

). However, all these taxa are characterised by

lower-crowned teeth, a much larger size, and a complete metaloph,

i.e., reaching the labial side of the tooth.

To sum up, the pattern and morphological features of the P4 and

of the upper molar from MD-67 show certain af

fi

nities with

homogenidens

(Colhuehuapian), and

Steiromys

(Santacrucian-?

Laventan) to a lesser extent, while size is only compatible with

Microsteiromys

(

Walton, 1997

Microsteiromys

Walton, 1997

is the

smallest known representative of New World porcupines, with

a geographic and stratigraphic range thus far restricted to the

Baraya Member of the Villavieja Fm. of Colombia (c. 13

e

12.5 Ma,

middle Laventan;

Madden et al., 1997

;

Candela and Morrone,

2003

). This diminutive monotypic genus is only known by two

mandibles, which makes it impossible to compare with the avail-

able teeth from MD-67 (

Walton, 1997

). However, we assume that

these teeth are likely to document a close ally of

Microsteiromys

Walton, 1997

, here referred to as cf.

Microsteiromys

sp. Given their

morphological af

fi

nities with taxa spanning the early and middle

Miocene interval, these specimens cannot be used with much

con

fi

dence in a biochronological perspective.

4. Discussion

4.1. Biochronological age

Glyptodontid xenarthrans are of poor biochronological interest,

especially when identi

fi

ed above genus level. Caviomorph rodents

are the most diverse group in MD-67, with four taxa encompassing

four superfamilies (Octodontoidea, Erethizontoidea, Cavioidea, and

Chinchilloidea;

Vucetich et al.,1999

). Caviidae have their FAD in the

lat

est (or post-) Colloncuran of Patagonia, with

Guiomys

(

Pérez,

2010

;

Pérez and Vucetich, 2011

), a genus which is recognised at

MD-67 (

Fig. 3

e

Table 1

). Acaremyidae had a Colhuehuapian-

Colloncuran stratigraphical range (

Vucetich and Kramarz, 2003

)

while the hathliacynid marsupial

Sipalocyon

had only an early

Miocene record so far (Colhuehuapian-Santacrucian;

Marshall,

1981

). As such, MD-67 might extend upward the known range of

both taxa, and thus represent their Last Local Occurrence (

Table 1

On the other hand, the recognition of

Marmosa

(

Micoureus

) and of

Microsteiromys

-like dwarf porcupine in MD-67 predates the

previous FAD of both taxa, formerly restricted to the Laventan

Table 1

Stratigraphic range of the middle Miocene vertebrate locality MD-67 as inferred by

mammalian biochronology and

fi

ssion track datings (see

Supplementary Data

). Data

from

Walton (1997)

Vucetich et al. (1993

2010

Vucetich and Kramarz (2003)

Kramarz (2004

2006)

Gradstein et al. (2005)

Kramarz and Bellosi (2005)

Cerdeño

and Vucetich (2007)

Croft et al. (2011)

, and

Pérez and Vucetich (2011)

SALMA in the La Venta area of Colombia (

Table 1

;

Marshall, 1976

;

Goin, 1997

;

Madden et al., 1997

;

Walton, 1997

fi

rst sight, the biochronological age for MD-67 is middle

Miocene, i.e., Colloncuran-Laventan (

w

15.6

e

11.6 Ma;

Kay et al.,

1997

;

Madden et al., 1997

), but the occurrence of a typical early

Miocene genus such as

Sipalocyon

and the potentially

plesiomor-

phic evolutionary stage

Scleromys

with respect to its Laventan

counterparts (

Walton, 1997

) would tend to favour a Colloncuran-

early Laventan age for this new locality (

w

15.6

e

13.0 Ma;

Table 1

4.2. Apatite

fi

ssion track age of MD-67: 17.1

2.4 Ma

fi

ssion track age was calculated on 11 apatite grains. The

corresponding results are detailed in the

Supplementary Data

. The

c

test (

Galbraith, 1981

;

Green, 1981

) is currently used to discrim-

inate between concordant (

(

c

)

>

5%) and discordant (

(

c

)

<

5%)

grain-age distributions. The MD 67 sample passes the

c

test

(

c

)

>

5%) indicating that the apatite grains display a concordant

population and that they are derived from homogeneous sources

(

Galbraith, 1981

;

Green, 1981

The central age (17.1

2.4 Ma), pointing to a Burdigalian

e

Langhian age for the concerned grains (

Table 1

), is only partly in

agreement with the faunal content of MD-67, interpreted as doc-

umenting a Colloncuran-early Laventan age (

w

15.6

e

13.0 Ma;

Table 1

In other words, Apatite Fission Track provides a detrital age

(17.1

2.4 Ma) for the vertebrate-yielding locality, slightly older

than its inferred biochronological age (Colloncuran-early Laventan

South American Land Mammal Ages:

w

15.6

e

13.0 Ma). Be as it may,

the middle Miocene age of the concerned outcrop is fully contra-

dictory to its original assignment to the Paleocene-Eocene

Yahuarango Formation (

Vargas and Hipólito, 1998

4.3. Paleoenvironment

At regional scale, the concerned area mostly yielded terrestrial

habitats of low elevation, with a moist forest assumed as resem-

bling the modern Amazonian rainforest, in terms of composition

and biodiversity (

Hoorn, 1993

;

Antoine et al., 2006

;

Pons and De

Franceschi, 2007

;

Hoorn et al., 2010a

Unfortunately, MD-67 yields no pollen, spores, or plant fossils.

Available proxies (dental morphology and habitat preferences of

living and fossil mammal analogues; depositional environment)

thus provide only indirect information on the environment of the

concerned area by the time fossil mammals accumulated.

Most rodents from the current assemblage, including a meso-

dont acaremyid, a hypsodont dinomyid (

Scleromys

sp.), and

a hypselodont caviid (

Guiomys

sp.), can be interpreted as an indi-

cator of open and dusty environments under a quite dry and windy

climate (

Candela and Vucetich, 2002

). Yet, the dwarf

Micro-

steiromys

-like erethizontid of MD-67 was most probably fully

arboreal, like all living New World porcupines and their fossil kin,

as proposed by

Candela and Picasso (2008)

. Accordingly, all living

marmosine marsupials, such as the representatives of the subgenus

Marmosa

(

Micoureus

) are forest dwellers, being either arboreal or

ground foragers, in moist habitats related to tropical evergreen or

mossy forests (

Emmons and Feer, 1997

The appendicular skeleton of

Sipalocyon

is poorly known, but

functional anatomy suggests arboreal and potential grasping

capabilities for this carnivorous marsupial (

Argot, 2003

2004

). The

Santa Cruz Beds of Patagonia, which yield most remains of

Sipalo-

cyon

, are interpreted as originating from a temperate coastal plain

(

Bown and Fleagle, 1993

), with moderate rainfall and a mixture of

forested habitats/open areas/bushland as suggested by mammals,

pollen, and terrestrial arthropods (

Vizcaíno et al., 2010

Lithology of the fossil-bearing beds (i.e., channel-iron deposits,

with Fe-rich pisolite gravels) provides valuable information on the

depositional environment and associated diagenetic processes:

pisolites might have formed in the ground by alteration and

concretion of highly ferruginous groundwater solutions under warm,

humid, and seasonally-contrasted conditions (

Tardy, 1992

). Accord-

gly, isotopic analyses performed on coeval mollusc shells from the

Iquitos area (

w

1000 km more to the North) show the region was

experiencing a seasonal water in

fl

ux under a monsoonal-like tropical

climate by that time (

Kaandorp et al., 2006

;

Wesselingh et al., 2006

Both proxies are therefore in good agreement.

One of the striking features of the Miocene of Amazonia is the

presence of a large and long-lasting

mega-wetland

. The Amazo-

nian mega-wetland reached its maximum extent during the Middle

Miocene (also called

Pebas phase

sensu

Hoorn et al., 2010a

) and

it consisted of a complex mosaic of lakes, embayments, swamps,

rivers, and

fl

uvio-tidal environments (see review in

Hoorn et al.,

2010a

). Our data suggest the absence of this megawetland in

the Amazonian Madre de Dios Subandean Zone of Perú (

Fig. 4

while other coeval localities such as IQ-26 and NA069 (nearby

Iquitos;

Antoine et al., 2006

;

Pujos et al., 2009

) or the Fitzcarrald

Local Fauna (

Antoine et al., 2007

;

Goillot et al., 2011

) were under its

fl

uence during the same period (

Fig. 4

). This environmental

contrast might in turn have played some role in the faunal

discrepancies as observed in middle Miocene times between

Northern and Southern South America (

Madden et al., 1997

4.4. Biogeography

The earliest undisputable representatives of Glyptodontinae,

referred to

Boreostemma

, originate from middle Miocene Laventan

localities of La Venta and Fitzcarrald (Colombia and Perú, respec-

tively; e.g.,

Antoine et al., 2007

). In southern South America, the

earliest record occurs much later, with

Glyptodontidium tuberifer

from the late Miocene-Pliocene of NW Argentina (e.g.,

Oliva et al.,

2010

), which is most probably tied to environmental reasons

(

Carlini et al., 2008

). If con

fi

rmed, the referral of the osteoderms of

MD-67 to Glyptodontinae would still extend the gap between the

fi

rst local occurrences of the group in low and high latitudes.

The cavioid rodent

Guiomys

Pérez, 2010

was so far restricted to

the middle Miocene of Patagonia and Southern Bolivia, with a ?

Colloncuran (

w

14 Ma; Patagonia)-Laventan (Patagonia

Southern

Bolivia) stratigraphic range (

Fig. 4

;

Pérez, 2010

;

Croft et al., 2011

;

Pérez and Vucetich, 2011

). Accordingly,

Sipalocyon

was so far

restricted to high Southern latitudes, from Patagonia to Southern

Chile (51.5

e

37.5

Marshall, 1981

;

Forasiepi et al., 2006

;

Flynn

et al., 2008

). The co-occurrence of this stem caviid and of

Sipalo-

cyon

(37.5

e

S) in Peruvian Amazonia extends the geographical

range of

w

and

w

more to the North, respectively (

Flynn

et al., 2008

;

Croft et al., 2011

). On the other hand, the small mar-

mosine

. (

laventica

and the diminutive erethizontid

Micro-

steiromys

were only known in the Huila Department of Colombia

during a very short Laventan interval (

Fig. 4

;

Goin, 1997

;

Walton,

1997

), i.e., later and

w

more to the North than in MD-67.

In other words, the MD-67 assemblage both (i) postdates the

formerly known range of taxa of mid- and high latitude af

fi

nities

(

Sipalocyon

; Acaremyidae) and (ii) predates that of taxa of low latitude

fi

nities (

Marmosa

(

Micoureus

);

Microsteiromys

; Glyptodontinae).

This phenomenon is summarised by the spatiotemporal range of the

small dinomyid recognised in MD-67 and its kin, if close phylogene

tic

fi

nities are con

fi

rmed between the representatives of

Scleromys

sensu stricto (early Miocene of Patagonia and Southern Chile;

Flynn

et al., 2008

) and those of

Scleromys

sensu

Walton, 1997

(late

middle Miocene of Colombia and Peru;

Walton, 1997

;

Antoine et al.,

2007

). Such assertion may be an artefact tied to the scarce fossil