![. Length and V/idth (mrn) of Cheek Teeth of Some Marsupials Ilom Tiuparnpa, Bolivia M2-4/M2-5 Specimen P U CADELP HYS A]\DIÀ'LIS](https://www.researchgate.net/profile/Christian-De-Muizon/publication/279544711/figure/tbl1/AS:644295077941248@1530623325084/Length-and-V-idth-mrn-of-Cheek-Teeth-of-Some-Marsupials-Ilom-Tiuparnpa-Bolivia_Q320.jpg)
Content uploaded by Christian de Muizon
Author content
All content in this area was uploaded by Christian de Muizon on Jun 13, 2016
Content may be subject to copyright.
Larr-v G. Marshall and
Chribtinn de Muizon
The f)avrer of the ASe
of Marnrnals in South
Arnerica
The oldest known mrtrsupial and undisputed plncental mammals in
South America arefrom the Tiupampa bcalfauna of the El Molino For'
matinn, Bolivia, and tLæ Laguna umayo lacalfauna of the Vilqucchbo
Formatinn, Peru, whi.ch are here regarded as Late Cretarcous (Maes-
trbhtian) in age. The bcalfauna at Tiupampa includcs at bast 71 spe-
cics of marsupinls and seven species of placentals, and &t Laguna
IJmayo tncludcs at bast three marsupials and huo placentals. Colbc-
tively these mammals dacument that South America ÿLtas & major the-
ater for the early evolution and cladogenesis of marsupials and
plmerttals, Furthermore, many Late Cretaceous and early Tertiary
mammalirtn distributian patterns in the Westcrn Hemisphere can be
eæplained by envisinning dispersal of stocks îro*, snd not to, South
Amerba. The distributian paüerns of these and ottwr groups of terres-
tria.l and freshwater vertebrates dcmonstrate that opportuniti.es for
disp er s al b etw e en the Ameri.c as w er e av ailnbb on mulüpli o c c asions at
the "Dawn of the Age of Mammals."
C. de.tt*'(
.ÿ"
Larry G. Marshall, seniar research
scientist, Institute of Human Origins,
Berkeley, CA 94709. Christbn de
Muizon, researcher, Centre National
de la Rech.erche Scientifiquc, Institut
de Paléontolaÿ, Muséum natianal
d' Histoire naturellc, 7500 5 Paris'
France.
One of the most intriguing issues in vertebrate history is the origin of the
land mammal fauna of South America. The earliest undisputed record
of mammal life on that continent was, until recentÿ, based on knowl-
edge offossils from rocks of middle Paleocene age (Itaboraian-about
64 million to 61 million years ago) in Argentina and Brazil (Marshall
1985 and references therein). Mammals belon$ng to such gloups as
marsupials, edentates, condylarthsl astrapotheres, notoungulates, and
litopterns were first recorded in these Paleocene faunas. Because of the
specialized nature of these animals, their phylogenetic relationships
among themselves and with grcups of similar age elsewhere in the
world were either unknown or only tenuously established. In addition, it
was not known if some or all of these groups evolved in situ in South
America from long-established ancestral stocks, or if some or all came
from stocks that immigrated to that continent from elsewhere (i.e.,
North America, Africa, or Australia through Antarctica) at a time when
these landmasses \À/'ere eitherjoined directly or in closer proximiÿ (i.e.,
before continental drift produced today's configurations) ' It was long re-
alizedthat these and related Questions about the phylogenetic affinities
and biogeographical histories of earÿ Cenozoic mammals in South
America could confi dentÿ be addressed only by obtaining knowledge of
what mammal life was like during Late Cretaceous time (75 million to
about 66 million years ago) on that continent'
In7967 French geologists and paleontologists announced the discov-
NATIONAL GEOGRAPHIC RESEARCH a(7):23 -55 (7988)
ery of the first Late Cretaceous (Maestrichtian) land mammal fauna in
South America (Grambast et al. 1967; Sigé 7971,7972). This local fauna
is from rocks of the Vilquechico Formation at Laguna Umayo near Lake
Titicaca, Peru. In 1983 Argentine paleontolo$sts announced the discov-
ery of a second Late Cretaceous (Campanian) mammal fauna from
rocks of the Los Alamitos Formation near Arroyo Verde in Rio Negro
proünce, Argentina (Bonaparte €p Soria 1983). Although the mammal
faunas from both the Peruüan and Argentine localities are represented
almost exclusively by isolated teeth and tooth fragrnents, they do offer a
glimpse ofwhat mammal life was like in South America before the Age
of Mammals (66 millionyears ago to Recent).
In November 79BZ a U. S.-French-Boliüan paleontolo$cal expedi-
tion discovered a spectacularÿ rich fauna of fossil mammals in the EI
Molino Formation at Tiupampa, MizQue province, department of Co-
chabamba, in south-central Boliüa (de Muizon, Gayet et al. 79831 7984;
de Muizon 6p Marshall 7985,7987a: b, ci de Muizon, Marshall et al.
1984; Marshall et al. 1983a, 1985). The authors regard this fauna as Late
Cretaceous (Maestrichtian). Fossil mammals and other vertebrates
were collected during this and three subseç[uent visits (August 1 984, Oc-
tober to November 1985, October 19BG) by surface prospectingr ![uarry-
ing, and screen-washing of sediment.
During the four-year progtram, this international paleontolo$cal ex-
peditionhas recovered atleast 18 species oftherian mammals, including
11 marsupials and seven placentals. The marsupials include members
of the families Peradectidae, Microbiotheriidae, Didelphidae, Caroloa-
meghiniidae, and Borhyaenidae; and the placentals include members of
the orders Proteuttreria, Pantodonta, Condylarthra, and Notoungglata.
The proteutherians, pantodonts, and hl,posodontid condylarths are the
onlyknown records ofthese "t1T)ically" Holarctic (i.e., Northern Hemi-
sphere) gfoups in South America.
The 1985 field season yielded spectacular results-the recovery of
four partial skeletons, and four nearÿ complete and three partial skulls
of a new genus and species of didelphid marsupial. These specimens
represent the first skulls and associated skeletons of Cretaceous marsu-
pialsyetknown. Infact, multiple skulls and skeletons ofCretaceous ther-
ian mammals have been collected elsewhere only in Mongolia (Kielan-
Jaworowska et al. 1979).
The Tiupampa local fauna is unç[uestionably the richest - in terms of
both number of species and number and qualiÿ of specimens - Late
Cretaceous mammal localiÿknown in South America and it is proving
to be one of the richest localities of this age in the world. To date, more
than 600 specimens of mammals and several thousand specimens
ofother vertebrate groups have been recovered (Table 1). The fossil-
bearing potential of this localiÿ is virtually unlimited and has yet to be
ade{uately sampled. Tiupampa is proving to be the "Rosetta Stone" for
understanding many debated aspects about the phylogenetic relation-
ships andbiogeographic histories ofCenozoic mammals in SouthAmer-
ica in particular and in the Western Hemisphere in general. Knowledge
of this unusual fauna is revolutionizing current views about mammal
evolution in the Americas at the "Dawn of the Age of Mammals."
Here the authors discuss the followingissues: the age ofthe Tiupampa
local fauna; name, configuration, and preliminary diagnoses of nine
new marsupial genera and species; the phylogeneüc affinities ofthe sev-
en placental taxa; comparisons and contrasts among the mammal taxa
oflocal faunas from Tiupampa, Laguna Umayo, Los Alamitos, and Ita-
borai; aspects of the biogeogfaphic history of the Late Cretaceous and
MARSIIALL €E DE MUIZON
24
Table 1. Vertebrates Collected llom the Mamrnal-bearing Level of the El Molino Formation at Tiuparnpa, Boliüa
Class Osteichthves
Subclass Actinopterugii
Order Ginglynodi
Family Lepisosteidae
LepLsosteus sp.
Order 0steoglossiformes
Famill, Osteoglossidae
cf. Phareodus sp.
Family Hiodontidae
cf. Eohbdon sp.
Order Siluriformes
Family Ariidae
cf. Rhineræter sp.
Famil1,cf. "Ictaluridae" but new
gen. et sp. nov.
Order Characiformes
Family Characidae
cf. RhodsirL sp.
Family Serrasalmidae
cf. Miletes sp.
Familv Enthriniclae
cf. Hoplias sp.
Order Perciformes
Familv Percichthyidae
cf. Percichthys sp.
Subclass Sarcopterygii
Order Dipnoi
Family Ceratodontidae
çn. et sp. indet.
Familv Lepidosirenidae
qen. et sp. indet.
Class Amphibia
Order Anura
Famil,i, Leptodactvlidae
qen. et sp. indet.
Order G'rmnophion;r
Familr,indet.
qen. et sp. indet. (Rage 1986a)
Class Reptilia
Order Chelonia
Family Pelomedusidae.
d. Ror o chc Iy s v ilau ib ns is
cf . P odomemis b r as ilicrc is
Order Squamata
Suborder Lacertilia
Family Iglranidae
gen. et sp. indet.
Family indet.
gen. et sp. indet.
Suborder Ophidia
Famiÿ ?Aniliidae
cî. Coniaphis sp.
Family Boidae
gen. et sp. indet. 1
gen. et sp. indet. 2
Family indet.
gen. et sp. indet.
Order Crocodilia
Suborder Mesosuchia
Family Sebecidae
gen. et sp. nov.
Family Dryosauridae
So k o t osuch us atr. ianwilsoni
Class Mammalia
Ordët Marsupialia
Superfamily Microbiotherioidea
Family Peradecüdae
Peraàcctes austrinum (Sigé 1971)
Family Microbiotheriidae
Subfamily Microbiôtheriinae
Khasb cordilbrensis gen. et sp. nov.
Superfamily Didelphoidea
Family Didelphidae
Subfamily Didelphinae
Puca"dclphys anàiruts gen. et sp. nov.
Inca-delphys antiquus gen. et sp. nov.
Mizqwdelphys pilpinen^sis gen. et sp. nov.
Anàino d.elphy s co chab ahb ens is
gen. et sp. nov.
Subfamily Eobrasiliinae
Tiulardiaflaresi gen. et sp. nov.
Subfamily Derorh;,rrchinae
loskhaàclphys minutus gen. et sp. nov.
Superfamily Caroloame ghinoidea
Fam ily Caroloameghiniidae
Subfamily Protodidelphinae
Kollpanb tiupampina gen. et sp. nov.
Subfamiÿ Caroloameghi ni inae
Roberthoffstetteria
natilnalgeogr aphrca Marshall,
de Muizon €o Si.{é 1983a
Superfamily Borhyaenoidea
Family Borhyaenidae
Subfamily Hathliac1ryrinae
Allqokirus australis gen. et sp. nov.
Order Proteutheria
Family Palaeoryctidae
cî. Cimolestes sp.
Family indet.
gen. et sp. indei.
Order Pantodonta
Family Pantolambdidae
Alc ide dorb igny a inop inato de
Muizon €p Marshall 1987a
Order Condylarthra
Family Hyopsodontidae
Subfamily Mioclaeninae
Tiurla.enw minuhæ de Muizon €p
Marshall 1987c
Molinodus suarezi de Muizon €p
Marshall 1987b
Family Didolodonüdae or
Phenacodonüdae
Andinodus bolivbnsis de Muizon €p
Marshall 1987c
Order Notoungtrlata
Family cf, I{enricosborniidae or
Oldfieldthomasiidae
gen. et sp. indet. (see de Muüon,
Marshall et al. 1984)
early Tertiary mammal fauna of South America in particular and the
Western Hemisphere in general; and the updated taxonomic list of the
vertebrate fauna from the mammal-bearing level at Tiupampa.
Age of Tiupempa" Local Fauna
The El Molino Formation of Bolivia and the Vilquechico Formation of
Pem are regarded as temporal and stratigraphic eQuivalents (Cherroni
Mendieta 1977, Marttnez 7980, Sempéré 1986, Sigé 7972). These rock
units occur on opposite sides of the Andean Basin, a large sedimentary
sû'ucture covering northu/est Argentina, south-central Bolivia, and
southern Peru which includes marine, freshwater, and continental
rocks of Late Cretaceous and early Tertiary age (Martinez 1980). The
absence of associated volcanic rocks \^/ith the mammal-bearing sedi-
ments at Tiupampa (El Molino Formation) and Laguna Umayo (Vil-
Quechico Formation) deters dating and calibration by radioisotopic
methods. However, rock samples for paleomagneüc analysis have been
collected frornTT sites in a 200-m section of the El Molino Formation at
Tiupampa, and from 43 sites in a 400-m section ofthe Vilquechico For-
mation at Laguna Umayo. These samples are being processed by Robert
F. Bufler (Department of Geosciences, IJniversiÿ of Arizona, Tucson) in
an effort to secure age assignments for the mammal-bearing levels with-
LAND MAMMALS IN SOUTH AMERICA 25
in these rock units. Meanwhile, the age of these levels in particular and
of the rock units in general is based on knowledge of faunal content in-
terpreted as Late cretaceous (Maestrichtian). The evidence for this age
assignment follows (de Muizon et al. 19B3, 1984; Marshall et al. 1985).
. correlatinn with dinosaur trarkways within El Molino Formation.
Trackways of at least five differentkinds of dinosaurs (sauropods, thero-
pods) aré known from near the top of the El Molino Formation at Toro-
îoro, about 20 km southwest of Tiupampa. A selacian fauna which
includes six taxa and is regarded as Late Cretaceous (Maestrichtian)
was collected from below the dinosaur trackways at Torotoro. A sela-
cianfaunaidentical to thatfrom Torotoro hasbeen collectedfrom the 1o-
caliÿ of villa viscarra, 2 km south of Tiupampa. Recent detailed
stuatigraphic work by Thierry Sempéré (MISSION ORSTOM, Santa
Cruz, nolivia; Sempêré et aJ. 7987) indicates that the mammal-bearing
fauna at Tiupampâ is stratigraphicalÿ below a dinosaur trackway level
at Torotoro. However, no fossils of dinosaurs are known to have been
found in the selacian fauna below the trackways at Torotoro, in the sela-
cian fauna at Villa Viscarra, or at Tiupampa.
. Correlati.on with dinosaur eg§shellfauna at Laguna Umayo. Frag-
ments of shell from dinosaur eggs have been recovered in association
with mammals in excavations at Laguna Umayo (Kerourio €p Sigé 1984,
Sigé 1968). The Laguna Umayo and the Tiupampa local faunas are re-
grlra"d as temporal equivalents based on the co-occurrence of identical
àctinopterygian fish, the turtle cC. Rorochelys vilavibnsis, and the mar-
supiafneiaàectes antstrinum (Bonaparte 6p Powell1980, de Muizon et
al. 1983, Marshall et al. 1985).
. Correlatianwith dinosaur-bearingfaunas elsewhere in South Ameri-
ca.. The turtles cf. Rorochelys vilnvibnsrs and cf. Podocnemis brasi-
tiensis fiom the mammal-bearing level at Tiupampa are known from
tlre dinosaur-bearing Bauru Formation of BtazTl (Bonapalte 6p Powell
1980, Marshall et al.1gA5); a level 50'm above the selacian fauna at Vila
viscarra (and hence above the mammal-bearing level at Tiupampa) has
yielded cf. Rorochèlysvikwibnsis and a small crocodile rcrerredto cyn-
-odontosuchus cf. rothiwhich is found with dinosaurs in the Neuquén
Group of Argentina (Bonaparte 6p Powell lgB0).
. Stog" of evàtutian of mammals. No mammal faunas of earÿ Paleo-
""rr"àg" ur" known (or recognized) in South America (Marshall 1985).
Howerier, the taxonomically rich middle Patreocene (Itaboraian) fauna
from ItaborairBraztl, includes many small- to medium-sized taxa and
these provide a basis for comparison with the Tiupampa local fauna. As
docufrented below, five ofthe marsupial taxa from Tiupampa represent
potential ancestral forms for taxa at Itaborai, although no gelela or spe-
èi", ,"" shared by these two local faunas. In addition, nopoÿdolopoids
have been discovered in the Tiupampa local fauna; polydolopoids are a
specialized group firstknown in thgmiddle Paleocene faunas of Btaztl
àd a"g"rrtioa. Àmong the placentals, the pantodontis more.plesiomor-
phic thàn its closest miâdle Paleocene relative in North America, and the
mioclaenine hyopsodontids are more generalized than their nearest ear-
ly paleocen" ,ôlâtir"r in North America. The pantodont, hyopsodontid
ôondylarths, and proteutherians are unknown in middle Paleocene or
later faunas in Sôuth America. These data clearÿ suggest that Tiu-
pampa is older than middle Paleocene and is consistentwith a Late cre-
îr""ôrr. stage of evolution. However, since no South American faunas of
early paleoéene age have been recognized, it is impossible to compare
the fauna and dernonstrate thatthe Tiupampa mammals are notearlyor
earliest Paleocene by South American standards.
MARSIIALL €P DE MUIZON
26
. Chicken-and-egg argpment. Some colleagges suggest that the ab-
sence ofassociated dinosaur remains in the Tiupampa local fauna proves
it to be earÿ Paleocene. However, post-dinosaur localities of Maestrich-
tian age do exist in North America (i.e., Montana, see Archibald €p
Clemens 1982; NewJersey, D. Baird personal communication). At the
same time some workers (Sloan et al. 1986) believe they have demon-
strated that dinosauré survived into the earÿ Paleocene of Montana,
NewMexico, India, the $urenees, and Peru (LagunaUmayo). Neverthe-
Iess, the authors contend that biostratigraphic evidence is sufficient to
indicate that the Tiupampa local fauna was contemporaneous with di-
nosaurs elsewhere in South America, and for this reason and the stage of
evolution ofthe mammals in general theyfavor a Late Cretaceous (Maes-
trichtian) age assignment for this local fauna. The non-mammalian taxa
in particular have received only preliminary study, and further work on
these gloups may re(uire reassessment of the biostratigraphic picture
presented above. Even if the evidence swings in favor of an early Faleo-
cene age it will little âlter the phylogenetic and biogeographic conclu-
sions of this paper, or decrease the importance of this local fauna.
Systematics of Mar supials
Unless specificalÿ indicated othenvise, the following mammals were
recovered from between 110 and 140 m above the base ofthe El Molino
Formation at Tiupampa (see location of this level in stratigtaphic sec-
tions in de Muizon, Gayet et al. 1983:fig. 1; Marshall et al. 7985:frg.4,
.ighfl. Nomenclature for marsupial molar stmcture, dental formula
(permanent molars, ffi-M|-MX-ME' deciduous tooth, Yrl, fi*
Archer 7978), and supiageneric ranks follow those used in Marshall
( 198 7) ; and use and chronologr of Souttr American Cenozoic Land Mam-
mal A ges follow Marshall ( 1 985). Abbreviations include : C, canine ; I, in-
cisorl L, length; LU, Laguna Umayo, Faculÿ of Sciences, Universiÿ of
Montpellier, Francel M, molar; MNHN, Muséum national d'Histoire
naturelle, Paris; MN\|, Museu Nacional e Universidade Federal do Rio
deJaneiro, Errazil; P, premolar; W, width; YPFB Pal, paleontologz col-
lection of the Yacimientos Petroleros y Fiscales de Bolivia (YPFB) in the
Cento de Tecnolo$a Petrolera (CTP), Santa Cruz, Bolivia.
ETIMOLOGY;Puca-, forthe Puca Groupwhich includes theEl Molino Forma-
laon (pura or puka k iJ.:.e Quechua Indian word for red); alrd -delphys, in refer-
ence to didelphid marsupials.
TYPE-SPECIES: Pu"caàelphys andinus sp. nov.
DIAGNOSIS: Same as for §.pe and only known species.
ETIMOLOGY; andinus, for the Andes Cordillera of South America.
TYPE-SPECIMEN: YPFB Pal 6105, a nearly complete skull with alveoli of I2-5,
and 11 and C-Ms present on right side; alveoli of 11 and 1s-5, and I2, C, posterior
half of P3 and MLs on left side; right mandibular ramus with I,_r, C-Mu pres-
ent, and alveoli of I, and In; posterior half of left mandibular ramus with M-u;
associated and articulated partial skeleton ofneck and thoracic regions (Fig-
ures 1, top €* 3A).
IIYPODIGM: Tlpe and MNHN Vil 118, a partial isolated left M, (de Muizon,
Marshall et al. 1984:fig. 8); MNHN Vil727, a partial isolated left M5 (de Mui-
zon, Marshall et al. 7984:fig.9); YPFB Pal 6106, nearÿ complete articulated
skeleton without skull (Figure 1, bottom) collected in association with §pe
(\?FB Pal 6105); YPFB Pat 6107, greater part of skull, articulated mandibles,
and most of denütion; YPFB Pal 6108, greater part of partially crushed skull
LAND MAMMALS IN SOUTH AMERICA
Superfamill, Didelphoidea
Farnily Didelphidae
Subfarnily Didelphinae
Pucadelphys $en. nor,"
Pucadelphys qndinus sp. nel§/.
(Figures 1-3)
ary
îiÿre 7. Pucadelphys andinusgen.
et sp. nov.: topt YPFB Pal 6705 (ÿpe),
a nearly compbte skullwith artiru-
lated mandiblcs and partial skebnn
of neck and tLnrrrcir reginns; bot-
tom, YPFB Pal 6706' a nearlY com-
plete arti.culated skebton wîthout
skull. Note presence of "marsupial
bones" on pelvis of Inwer specimm.
Thcse individuals are probablY a
mab-female pair that dicd andwere
fossilized în their burrow whîle in a
snout-rump nesting Posîtion. ToP
specimen fates right, bottom speci-
men faæs lejl. (adual size)
Figpre 2; Pucadelphys andinusgen.
et sp. nov.: toprYWB Pal 6177, par-
tial artbulated skebton of thoradc
regian to below base of tail; bottom,
YPFB Pal 6110, greater part of skull
anà arti.culated mandiblz wîth parts
of thoracic region of skeleton. Proba-
bly a mab-femab nesting pair. Top
specimen fates lcfi
, bottom fares
right. (at:tual size)
with articulated mandibles and most ofdentition; YPFB Pal 6109, rostral part of
skull with right C-P3, left P1-M5, and greater part of left mandible with bases
ofC and P, and Pr-Ms present (Figure 38, C); YPFB PaI 6110, greater part of
skull with articulated mandibles and parts of articulated thoracic regiion of
skeleton (Figure 2, bottom)i YPFB Pal 6111, partial articulated skeleton from
thoracic region to below base o-f tail (Figrre 2, top) collected in association with
YPFB Pal 6110.
MEASUREMENTS: Table 2.
DIAGNOSIS : Dental fonnula 15 / 4 C1 / 7 P 3 / 3 M5 /5 ; small- to medium-sized di-
delphine; skull with distinct frontal-maxillary contact; no distinct eüdence of
palatal vacuities or ossified auditorybullae; C well developed; sharp increase in
size from P1 to P3; P2-" with distinct posterobasal heel; P2-3 with small antero-
basal cuspule and larger posterobasal heel; P3 with weakly developed cingu-
lum along labial and well-developed cingulum along lingual base of crown
28 MARSHALL €P DE MUIZON
Table 2. Length and V/idth (mrn) of Cheek Teeth of Some Marsupials Ilom Tiuparnpa, Bolivia M2-4/M2-5
Specimen
P U CADELP HYS A]\DIÀ'LIS
Upper cheek teeth
l?FB Pal 6105
(right side)
YPFB Pal 6109
(left side)
Lower cheek teeth
\?FB Pal 6105
(right side)
\?FB Pal 6109
(left side)
0.9 0.5
0.9 0.5
0.6 0.3
7.4 0.9
7.5 0.9
'1..4 0.8
1.5 0.7
1.6 1.0
1.7 0.9
1.5 0.8
1.1 0.8
1.3 0.8
t.2 1.0
4.9
5.2
6.8
7.0
INCADELPIIYS ANTIqUUS
Upper cheek teeth
\?FB Pd 6251
(right side)*
(left side)*
Lower cheek teeth
\?FB PaI 6251
(right side)*
M I ZQU E D E L P HY S PILPIÀENSIS
Upper cheek teeth
YIFts T.rI 6 tgG
TIULORDIA FLOR-ESI
Upper cheek teeth
YPFB I']al 6191
(left side)
(right sidc)
Lower cheek teeth
\?FIl Pal 6191 1.0 0.5 1.7 O.7
"M1L - 1.4 nun, l{ - 0.9 rnnt
'!NI'L: 1.1 nun. l\':0.8 ùtrrt
+M L = 1.1 tfLÙl, llt - 0.7 mnt
1.0 0.1
7.2 0.6
7.7 2.0
2.O 1.9
1.7 7.2
1.9 7.7
1.8 2.5
1q ))
1.8 1.4
2.0 .1.3
1.8 2.5 7.3 2.3
1.8 2.3 7.1. 2.1
2.O '1,.3 1.6 1,.2
2.0 1.3 1.9 1.1
1.6 1.3
1.4 1.3 7.4 7.5 1.3 1.6
7.4 1.5 7.3
1.5 0.8 1.6 0.9
1.5 1.8 1.6 1.9 t.4 2.O
connecting anterobasal cuspule with posterobasal heel; Mzr with well-devel-
oped, anteroposteriorly elongate protocone, para- and metaconules; paracone
smaller in size in occlusal vielv and distinctÿ lower in lateral view than meta-
cone; centrocrista dilambdodont (sensu Crochet 1980, V-shaped with apex
pointing labially toward s§;lar cusp C); s§,lar sheHwell developed; metacrista
short, weakly developed; paracrista low, weakly developed, unites Paracone
and sÿlar cusp B; s§ilar cusp B highest of sÿlar cusps; sÿlar cusps A, C, and D
subequal in size and only slightly smaller than s§ilar cusp B; s§ilar cusp A (par-
as§{e) spurlike, ertending anteriorÿ and overhanging posterolabial edge of
preceding tooth; sÿlar cusp A united with paraconule by weakly developed
paracingulum; no clear eüdence ofs§ilar cusp E; ectoflexus absent or very shal-
low on M2, slightÿ deeper on M3, deeper but still shallow on Ma; trigonid cusps
well elevated above talonid; anterobasal cingulum well developed; trigonid
slightly narrower than talonid on M2r subequal in breadth to talonid on M, ta-
lonid slightÿ narrower than trigonid oD Ma: and markedly narrorver on ME;
protoconid highest of trigonid cusps, metaconid next highest, paraconid re-
duced relative to metaconid but still well developed; protoconid and metaconid
united by straight, trenchant protolophid; hl,poconid and entoconid subequal
in height, with hypoconid larger than entoconid in occlusal view; hS.poconulid
subequal in size to, or larger than, entoconid; cingular shelf extends ventrola-
bially from üp of hy.poconulid across posterior surface of hlpoconid; talonid
distinctÿ basined; crista obliqua contacts trigonid at posteromedial surface of
protoconid; on Ms talonid basined but very narrow relative to trigonid, hy.po-
conulid very large (larger than hypoconid), entoconid very small and fused
with anterobasal edge of hypoconulid.
LAND MAMMALS IN SOUTH AMERICA
P2 P3 lsIi! M3
LU/Ltd,L!ÿt\iÿLIÿLIU
29
Fîgure 3. Stereopairs o/ Fucadelphys
andinus gen. et sp. nov.: A, YPFB Pal
6105 (ÿpe), ventral part of skull
sLnwing right sidc of dentitian and
basicranium (posteriar half of Iefi
mandibb is still auarhcd, right man-
dible has been removed) (x 2.5);
B, YPFB Pal 6709, pakttal view of
partial skull showing details of up-
per bfi cheek teelh Pr -u5 (x 5);
C, YPFB Pal 6709, W mandibuktr
ratruts with base of canine and P1n
and P2-M5 present, Inbfutl (lefi) and
occhsal (right) views (x 5).
COMMENTS: P. andinus is the most abundant and completelyknown mam-
mal in the Tiupampa local fauna. Approximatelyg0o/o ofthe skeleton is repre-
sented among the four partial skeletons, trvo of which have nearly complete
articulated skulls (Figgres 7, top ۍ 2' bottom)' The four partial skeletons
30 MARS}IALL & DE MUIZON
(l?FB Pal 6105, 6106, 6110, 6111) were found in pairs, with the individuals of
each pair gfasping one another and facing in opposite directions (Figures 1 €r
2). Tlie authors interpretthis to indicate thatthese animal-pairs (probablymale
and female) died while in their burrows and were fossilized in the snout- rump
nesting position ÿpical of living didelphids. study of aspects of the functional
morphology of these skeletons will enable paleontolo$sts to determine whether
the tàil of P. andiruts had prehensile capabilities as does that of the liüng Vir-
ginia opossurn (Diàelphis virginianus), and whether the hind feet had special-
izaüoni indicaüve ofarboreal or terrestrial life-s§zles. The pelvic region ofYPFB
Pal 6106 (FigUre Trbottom) shows the presence of awell-developed and ossi-
fied "marsupial" or prepubic bone.
P. andinus is also represented by four nearÿ complete and three partial
skulls (Figures 1-3), about a dozen partial upper and lower jaws with teeth,
and several dozen isolated teeth. Features ofthe basicranium and ear re$ons
have proven particularÿ useful in identification of the phylogenetic relation-
ships amongliving and fossil didelphoid marsupials (Reig et al. 7987), and the
neàrÿcomplete skulls ofPucaàclphyswlll enable study ofthese features, which
are rarely preserved in Cretaceous age taxa.
Puca.dclphys is a member of the family Didelphidae and subfamily Didel-
phinae, as these groups are defined by Marshall (7987). The two primary fea-
tures that indicate this affiniÿ are the V-shaped (dilambdodont) centrocrista,
and the small size of the paracone relative to the metacone (Figure 3B)'
Ofthe earÿ Tertiary didelphines that have been recovered from South Ameri-
ca, P. andinui compares closest in size as well as in structute with Stern-
bergiaitaborqiensis Paula couto 1970 from the middle Paleocene (Itaboraian)
fissure fillings in the Itaborai Formation at Sâo José de Itaborai, Brazil. The
"stentbergin-like" taxon recorded by de Muizon, Marshall et al. (1984:figs. B €o
9) was based on isolated teeth here referred to P. andinus (see Hypodigm,
above). S. itaboraiensis was recentÿ restudied by Marshall (7987), based on
reconsideration ofpreüously described material and new specimens of upper
and lower dentitions. P. andiruts differs from s . itaborabnsis in having a larger
hypoconulid on M2-5r in lacking distinct labial cingula on M2--4r and in having
relatively larger para- and metaconules. In these and other features Pttca-
delphys represents a potential structural ancestor fot Sternbergia.
Puca.dctphys also shares some features with species of Alphadon from the
Late Cretaceous ofNorth America which are currently regarded as members of
the famiÿ Peradectidae (sensu Marshall 1987, Reig et al' 1987). For example,
the specimens of A. marshi described and figured by Clemens (1966) are simi-
tar in size to P. andiruts, and both species have well-developed sÿlar cusps,A
tlrrough D. These taxa differ in that Pucadclphys has a V-shaped (dilambdo-
dont) centrocrista and the paracone is notably smaller than the metacone,
while in species of Alphadon the centrocrista is straight (predilambdodont)
and the paracone and metacone are subeç[ual in size and height. These two fea-
tures cuirently serve to define and differentiate members ofthe families Didel-
phidae from those of the Peradectidae (Crochet 1980, Marshall 7987).
The family Peradectidae includes taxa regarded as structurally ancestral to
Didelphidae (Crochet1980, Reigetal. 1987). Clemens (1966) suggestedthatthe
ancestor ofTertiary age didelphids in South America was structurally similar to
such species of Alphadoru as A. marshi and A. Iulli. In all comparable features,
A. marsht does indeed represent a potential structural ancestor for P. andiruæ.
It thus appears probable that didelphids originated from a peradectid or pera-
dectidlike ancestor during the Late Cretaceous in South America.
ET\'1\{OLOGYI Inca-, for the Inca civilizaüon and its descendaàts who inhabit
Tiupampa; and -delphys, in reference to didelphid marsupials.
T\?E-SPECIES; Incadelpltys antiqws sp. nov.
DIAGNOSIS: Same as for §.pe and only known species'
ETI'À4OLOGY: antiquus, Latin for ancient.
TYPE-SPECIMEN: YPFB PaL6257, a Ieft maxillarywithbase of P1, P2-Ma pre-
LAND MAMMALS IN SOUTH AMERICA
Incadelphys.gen nov.
Incadelphys antiquus'sp. nov,
(Figlrre 4)
Figure 4. Stereopairs o/Incadelphys an-
üquus gen. et sp. nov., YPFB Pal 6257
(type): A, Iefi madllary showîng ld-Itf
(Ma is erupting and missing tip of proto-
cone); B, rîght mandibuktr ramus slnw-
ing Pr-M, in occlusal Aefi) and labiol
(rîght) vi.ews. (all x 72)
sent (M4 erupting and missing tip ofprotoconel, unerupted P3; a ri-ght ma-xillar;,
fragrnentwith M1+ 1M, empting) and unerupted P3; a right mandibular ramus
witÏ. Pr-M., talonid of Mn, unerupted P, and Mu; and a left mandlbu^lar rarnus
with taloniâ of Mil Mr-e complete, M, broken, M5 unerupted; all of the same
juvenile indiüdual.
IfYPODIGM: Tyre onlY.
MEASUREMENTS: Table 2'
DIAGNOSIS: Dental formula I?/? C?/? P3l3 M5l5; small-sized didelphine;
premolars enlongate anteroposteriorÿ, qugh longer than wid,e; P1 smaller
ihan p2; p2 w-ith fosterobasa-l cingUlai shelf labially and lingualÿ; M1 molari-
form, protocor" -1*g", trigon '"viah shallow basin, conules small but distinct,
""nt oà.irt, werHyâil.mÉdodont, para- and metacone subeç[ual-in- size and
height, paracingrrllm well developeà, metacrista elongate, sÿlar.sheHvery re-
duéd, àbsent làbia1 to paracone, itylar crsp B absent or fused with paracone,
D largest and base elongate anteroposteriorÿ, cusps A, C, and E also present
but sÀaler than D I on M"H protocone large, trigon well basined, para- and me-
taconules distinctly develofed, paracingulum unites protoconule with sÿIar
cusp A, centrocrista V-shaped, paracone smaller and lower than metacone,
-"à".i.t, *"lI developed ând slightly convex on M3r; paracrista short, high,
unites paracone with iarge stylar cusp B; sÿlar shelf well developed but nar-
MARSIIALL €P DE MUIZON
rows anterioÙ, basin developed between line of sÿlar cusps which are fused
basally and atigrred in the same anteroposterior axis (especially on MF3), and
para- and metacone; sÿlar cusps B and D largest of sÿlar cusps and subequal
in size on M2, B largest s§{ar cusp on M34, sÿlar cusps A and C smaller than B
and D but disünct on ML3, sÿlar cusp E indistinguishable; ectoflexus absent on
M2, very shallow on M3r; posterior ends of Pr-, broader than anterior ends,
crowns inclined anteriorÿ, tip ofcrown set above anterior edge ofanterior root,
Pzwith distinct posterobasal cuspule and about same height as Mri Mr molari-
form, trigonid and talonid sube(ual in breadth, talonid basined, protoconid
highest cuspr metaconid about 50olo the size of protoconid, paraconid very re-
duied andihifted anterolabially, hypoconid and entoconid subequal in size,
hS.poconulid "tw.inned" with entoconid, crista obliqua contacts posterior sur-
faCe of trigonid just labial to protocristid notch, posterolingual edge of meta-
conid swollen into crestlike sbrrcture which is continuous with elevated
entocristid, a feature which increases apparent depth of talonid basin and en-
closes it anteriorÿ; strong cingulids anteriorÿ and posterioÙ; Mz-: with ba-
sined talonid slighfly wider than trigonid, trigonid cusps compressed
anteroposteriorÿ such that trigonid is distinctly broader than lon$; metaconid
intermediate in size between large protoconid and smaller paraconid; as on
M, lingual edge of metaconid swollen posteriorÿ and crestlike edge fused bas-
allywith elevated entocristid; crista obliqua meets posterior wall oftrigonidjust
labial to protocristid notch; hl.poconid larger and higher than entoconid; hypo-
conulid subequal in size to and "twinned" with entoconid; cingulids strong an-
teriorÿ and posteriorÿ.
COMMENTS: Among the other didelphid taxa from Tiupampa, I. antiquus
compares best in size and structure witln Mizquedelphys pilpinensrs (see be-
Iow). These taxa differ in thatlncaàelphys has less robust upper molar crowns;
less distinct conulesl a more reduced but basined sÿIar sheu (especialÿ anteri-
orÿ, labial to paracone - i.e., paracone and sÿIar cuspB are much closer than
in Mizquedelphys); metacrista slightÿ convex on ML31 and s§{ar cusps are
aligned, fused basally, and form an elevated crestlike edge to the sÿlar shelf.
Of the known early Tertiary Didelphidae in South America, L antiquus corn-
paresbestwi+hMarrnosopsis juraàoiPaula Couto 1962 from the middle Paleo-
èene (Itaboraian) of Brazil. upper dentitions of Marmosopsrls are described
and figured by Marshall (7957). These two taxa are remarkably similar in
sûrrcture of the upper molars, differing primarily in that Marmosopsrb has a
relatively shallower basin on the sÿlar sheH (particularÿ anteriorÿ) r a straight
(not concave ventrally) metacrista, and a smaller protocone (i.e., this corelates
n.ith a reduced talonid). In the lower molars these taxa differ in tJtat Marmo-
sopsrls has relatively narrower crowns, paraconid more reduced and set more
anteriorll, such that paracristid is elongate relative to shorter protocristid, pos-
terolingual svvelling of metaconid less distinct, hl.poconid lower relafive to en-
toconid, talonid more reduced and foreshortened anteroposteriorÿ, trigonid
more pelpendicular (not inclined for-ward), and anterior and posterior cingu-
lids less well developed. These differences are minor, and in both size and
strrrcture of upper and lower dentiti ons Incadelphys represents a potential Late
Cretaceous ancestor for Marmosopsis.
ETIIIOLOGY; Mizque-, for the province of Mizque; and -delphys, in reference
to dideiphid marsupials.
TYPE - S PE C I ES : Miz q uedelphy s pilpinens is sp. nov.
DIAGNOSIS: Same as for §.pe and only known species.
ETyMOLOGY: pilpinntsis, for the Pilpina River near Tiupampa.
T1?E-SPECIMEN: YPFB Pal 6196, a left maxilla with P3-M4 (Figure 5).
IIYPODIGM: Tÿpe only.
MEASUREMENTS: Table 2.
DIAGNOSIS: Dental formula unknown; small-sized didelphine; P3 crown nota-
bly elevated aborre molars, with sharp vertical bladelike crest along posterome-
LAND MAMMALS IN SOUTH AMERICA
Mizquedelphys gen. nov.
Miz quedelphy s pilpînens î.s
sp. nov. (Figure 5)
JJ
dial edge, lingual side of posterior root inflated basally; MH constricted
anteroposteriorÿ with well-developed protocone, distinctly basined trigon;
para- and metaconule well developed; paracingulum weakly developed; para-
cone smaller and lower than metacone; centrocrista weakly dilambdodont (V-
shaped); sÿlar shelfmoderately developed; sÿlar cusp B largest ofsÿl€ü cusps,
about same height as paraconei sÿlar cusp D also well dweloped but stightty
smaller than B; A distinct, small; C small, distinctly twinned on M3; E indisünct;
paracrista unites paracone and sÿlar cusp B; metacrista relatively short; ecto-
flexus shallow on MH, deeper on Ma.
COMMENTS : M. pilpinensis averages 20olo smaller th an P. andinus in compa-
rable linear tooth dimensions. Of knourn earÿ Tertiary Didelphidae in South
America, Mizqucdelphys compares best with Marmosopsis juradoi Paula
Couto 1962 from the middle Paleocene (Itaboraian) of BrazrJ Mizquedelphys
and Mafmosopsis are similar in many features of upper molar stucfure, in-
cluding relative size of sÿlar cusps, although there is no clear evidence oftwin-
ning of sÿlar cusp C on .M3 of Marm.osopsis. These taxa diff,er in that
Mizquedclphys has more distinct para- and metaconules, molars proportion-
ately broader than long (i.e., being more constricted anteroposteriorly), and a
broader stylar shelf.
Figpre 5. Stereopair o/ Mizquedelphys
pilpinensisgen. et sp. nov., YPFB Pal
6196 (type); lefi ma:rilktry with Ps-Ma
in occlu.sal vi.ew (x 8)
Andinodelphys gen. nov. ETYMOLOGY; Andina-, for the Andes Cordillera of Boliüa; and -dclphys, in
reference to didelphid marsupials.
TyPE-SPECIES; An^dino d"elphys cochab amb ens is sp. nov.
DIAGNOSIS: Same as for ÿpe and only known species.
Andînodelphys
cochabambensis sp. nov.
(Figure 6)
ET\'1\4OLOGY: cochabambensis, for the department of Cochabamba.
TYPE-SPECIMEN: YPFB PaL6792, an upper right M1? (Figure 6A).
HYPODIGM: Tlpe and YPFB Pal 6194, isolated lower left M*? (Figure 68, C).
MEASUREMENTS: YPFB Pal 6192L:2.9mrn, W:3.2 mm;6794L:3.0 mm,
W: 1.8 mm.
DIAGNOSIS: Dental formula unknown; largest didelphid known in the local
fauna of Tiupampa; protocone, paraconules, and metaconules well developed,
trigon basinedl metacone notably larger and higher than paraconel centro-
crista weakly dilambdodont (V-shaped); paracingulum r.vell developed; s§;lar
shelfbroad and basined; sÿlar cusp B largest of sg,lar cusps and same height as
paracone; paracrista bladelike, unites paracone with anterior edge of sÿlar
cusp B1 s§4ar cusp D very we-Il developed and slightÿ smaller than B; stvlar
cusp A (parastÿle) distinct, but much smaller than B and D; sÿlar cusp C well
developed, similar in size to A, distinctÿ twinned (posterior cuspule slightiy
larger than anterior cuspule); E very smalll metacrista moderately developed;
trigonid moderately elevated above and slightÿ broader than distinctly basined
talonid; anterobasal cingrrlum well developed; metaconid only slightÿ lower
than protoconid; paraconid about halfas large as metaconid; hlpoconid largest
and highest of talonid cusps, eç[ual to entoconid and h;,poconulid combined in
occlusal view; hypoconulid higher than entoconid in lateral üew, about twice
MARSI{ALL 6p DE MUIZON
Figure 6. Stereopairs o/Andinodelphls
cochabambensis gen, et sp. nov.: A,
YPFB PaL 6792 (ÿpe), upper right M4? in
occlusalvian; B, YPFB Pal 6194, lower
l4 MrP in occhrcal vi.eru; C, YPFB Pal
6794 in bbial view. @ll x 8)
A
as large in occlusal view; entoconid very reduced, entocristid absentl moderate-
ly developed cingular sheH extends ventrolabially from tip of hlpoconulid
across posterior surface of hl.poconid.
COMMENTS: A. cochabambensis is the largest dideiphid known in the Tiu-
pampa local fauna, and in linear tooth dimensions averages 30olo lar$er than P.
andinus. Ofknown early Tertiary didelphids in South America, Andinodclphys
compares best with ltaboraidelphys camposi Marshall 6p de Muizon 1984
from the middle Paleocene (Itaboraian) fauna at Itaborai, Brazil. These taxa
differ in that Andinodelphys averages 257o smaller in comparable linear tooth
dimensions and has a smaller entoconid, a less V-shaped centrocristid, and
molars less bulbous in overall appearance. These species share the same rela-
tive size of s§rlar cusps, including a disünctÿ trvinned sÿlar cusp C on Ma as
seen in MNRI 2S78-V (b) of ltaboraiàclphys and the ÿpe of Andinodelplrys'. ln
terms of comparable strucfural features it is thus easy to derive ltaboraidelphys
from Andinodclphys or an Andinodelplrys-hke ancestor.
ln-Andinodelphys, as in Didelphidae (sensu Marshall 7987)' the paracone is
markedly reduced relative to the metacone (in Peradectidae these cusps are
subequal in size and height), while the centrocrista is only weakly V-shaped
and not very different from peradectids in which it is straight (in Didelphidae it
isV-shaped). Marshall (1987) has argued thatthe transition from peradectid or
peradectid-like ancestors to didelphids apparently involved stochastic rates of
change in multiple characters and character complexes, and that the mosaic
distribution of peradectid (i.e., plesiomorphic) and didelphid (i.e., apomor-
phic) character states in some taxa simply documents the transitional nature of
this process. Given this realiÿ, the assignment of taxa to Peradectidae or Didel-
phidae is made on the basis of the first appearance of derived character states
(i.e., paracone reduced relative to metacone) and notby the last appearance of
primitive states (i.e., siraightcentrocrista). Followingthis criterion, the authors
tentaüvely assign Andinodelphys to the didelphid subfamily Didelphinae, al-
though they acknowledge that it could prove referable to a peradectid derivative
distinct from the Didelphidae.
ETYMOLOGY: Tiu-, for the localiÿ Tiupampa; and -lordia, in reference to af-
finities with G aylor dia.
T\?E-SPECIËS: Tiulordiq floresi sp. nov.
DIAGNOSIS: Same as for §rpe and only known species.
LAND MAMMALS IN SOUTH AMERICA
Subfarniÿ Eotrrasiliinae
Tîulordia gien. nov.
35
Figure 7. Stereopair o/Tiulordia floresi
gen. et sp. nov., YPFB Pal 6791; partial
right mandibular ramus tuith C-MB in
Iabial Aefi) and occluial (right) views.
(x 8l
Tiulordia floresi sp. nov.
(Figure 7) ETl'1\4OLOGY:J7oresi, in honor ofWalter Flores, Yacimientos Petroleras y Fis-
cales de Boliüa, for his friendship and assistance in field logistics.
TYPE-SPECIMEN: YPFB Pal 6191, a fragrnent of a right mandibular ramus
\À/ith C - Ms (Figure 7) ; a fragrnent of a right maxillary with posterior edge of M3
and complete M!5; and a fragment of a left maxillary with posterior root of M2,
M3 missing anterolabial corner, and Ma complete; all from the same associated
indiüdual.
HYPODIGM: Tlpe and MNHNViI 724, anisolated left Mr? missing tips ofpro-
toconid, metaconid, and part ofhlpoconuüd (de Muizon, Marshall et al .7984:
fig. 10); and MNHN Vil 128, posterolingual half of an isolated left P, (rle Mui-
zon, Marshall et al. 79$4tftg. 77).
MEASUREMENTS: Table 2.
DIAGNOSIS: Dental formulal?/? C?/7P?/3 M?/?; similar in size and structure
to Gaylordb doellni (Paula Couto 7962); sl.rnphysis extends posteriorÿ to point
below posterior edge of Pr; large mental foramen occurs below posterior root of
Pr; C weakly developed, tip of crown on same level as protoconids of M-r; pre-
molars increase in leng[h and width from P, to P3, tightly packed without dias-
tems separating them from each other or from C or M2; crowrr tips of premolars
Iower than protoconids of M-.; Pr crown markedly inclined anteriorly, much
Iower than P, or P3: with small posterobasal heel; P, low, robust, with broad
posterobasal heel; P, bulbous, specialized for crushing, with broad, shallow-
basined posterobasal heel; M, trigonid and talonid sube{ual intrreadth, M, fi-
gonid slightly broader than talonidi Mz-s othelwise similar, except M,
paraconid more reduced and inflected more anteriad; Mr-, metaconids onJy
itigtrtty to*er than protoconids, paraconids about haras large as metaconidsl
distinct anterobasal cingrrlum; talonids distinctlybasined, markedly foreshofi-
ened anteroposteriorÿ; hypoconid and entoconid very large, subequal in
height, hlpoconid larger in occlusal view; hypoconulid distinct, smaller than
othèr talonid cusps; Ma constricted anteroposteriorÿ; protocone well devel-
oped, trigon basined, para- and metaconules small, disünct; paracone low'er
and smaller than metacone; centrocrista distinctly V-shaped (dilambdodont);
sÿlar shelfbroad, s§llar cusps B and D moderately developed, B slightly larger;
sÿlar cusp A distinct, smaller than B and D, with well-developed, anteriorÿ
projecting parasÿle; s§4ar cusp C very reduced but distinct; no trace ofs§4ar
cusp E; ectoflexus sharp, deep.
COMMENTS ; T. floresi is the "Gaylari-d.a-lke" taxon recorded by de Muizon'
Marshall et al. (1984:figs. 10, 11). Of known early Tertiary age didelphids in
South Ameri ca, T. Jloresi compares most closely in size and structure with Gay-
brdia doeltoi (Paula couto 1962) from the middle Paleocene (Itaboraian) of
Brazil. The hypodigm of G. doelloi includes the tYpe of xenodclphis doelloi P av
la Couto 796i2, tlae lype of Gaylordirt mendesi Paula Couto 1970, and one re-
36 MARSHALL ê9 DE MUIZON
ferred specimen (Marshall 1987). The §.pe species of Gaybrdia is G. macro-
qmodonta Paula Couto 1952a which differs from G. doelbi in that it averages
30olo larger in linear tooth dimensions, and in that P, is §pically larger relative
to M, (Marshalt 1987).
In all comparable features, T. Jlnresi represents a potential structural ances-
tor for Gaylordb in general, and G. doellai in particular. T. floresi differs from
the two species of Gaylordb in having relatively broader and larger talonids
and a less bulbous Pr. Nevertheless, the anteroposterior shortening ofthe molar
talonids and inflation of the P, that characterize Gaylordla are incipiently de-
veloped in Tiubrdb. The only upper molars described for Gaylardia are in the
§.pe of G. macroqmodonta (Marshall 7987tfr9.18), and these are heavilyworn
and lack details of coronal stucture. They do show, however, that MH are, as
in T iulor dia, very broad for their length. Based on knowledge of the upper den-
nnola of Didelphopsis, which is another member of the Eobrasiliinae (see be-
low), Marshatl (19S7) concluded that the upper molars of Gaylordio will
predictably have distinct but not enormous s§ilar cusps B and D, and that B will
be slightly larger. This predicüon is supported by the presence of this arrange-
ment in the proposed Late Cretaceous ancestor of Gaybrdb, T. Jloresi.
Tiubrdia is the earliest known member of the didelphid subfamily Eobrasi-
liinae (sensu Marshall 1987) which was previously represented by three genera
of middle Paleocene (Itaboraian) age from Braztl (i.e,, Didelphopsis Paula
Couto 1952a, Eobrasilia Simpson 7947, and Gaylordb Paula Couto 7952a).
ETYMOLOGY :Jaskhn-, for theJaskha Pata River near Tiupampa; -delphys' in
reference to didelphid marsupials.
TYPE-SPECIES: Jaskhadclphys minutus sp. nov.
DIAGNOSIS: Same for t5.pe and onlyknown species.
Figure 8. Stereopair oflaskhadelphys
minutusgen. et sp, nav., YPFB PaI 6795
(ÿpe);fragment of right ma*illary with
tw1-ltt4 in ôcclusalvia,ry. (x 76)
Sutifarnily Derorh;,,nchinae
Jaskhadelphys gen nov.
Jaskhadelphys mînutus
sp. nov. (FigUre 8)
ETIMOLOGY: minutus, in reference to the tiny size of this species.
T\?E-SPECIMEN: YPFB Pal 6195, a right maxillary fragrnent with M3-4 (Fig-
ure B).
ITYPODIGM: T1.pe only.
MEASUREMENTS: M3L : 1.0 mm,W: 1,Omm; M4L : 1.Omm,W: 1.2mm.
DIAGNOSIS: Dental forrnula unknown; smallest marsupial known in Tiu-
pampa local fauna; protocone disünct, compressed anteroposteriorly, trigon
basined; paraconule larger than metaconulel paracingulum well developed;
paracone and metacone subequal in size and height (paracone broken on M3);
centrocrista V-shaped (dilambdodont); sÿlar shelfbroad, basined; s§;lar cusp
B very large, largest of sÿlar cusps, same height as paracone, unitedwith para-
cone bv sharp paracrista; sÿlar cusp D well developed, smaller than B; sÿlar
cusp A about same size as D; C small but distinct as evidenced by broken basesl
E indistinct; metacrista long, deep, bladelike.
COMMENTS: This tiny species possesses a combination of features not seen in
other didelphids of the Late Cretaceous or early Tertiary in South America: i.e.,
paracone and metacone subequal in size and heigh! centrocrista V-shaped;
protocone constricted anteroposteriorÿ; metacrista lon$r deep, bladelike. The
first feature is regarded as primitive for marsupials (Crochet 1980); the second
is derived and regarded an apomorphy for Didelphidae (Marshall 7987); the
third and fourth are apomorphies which are consistent with those expected in
the upper dentition ofthe tinymiddle Paleocene (Itaboraian) speciesMiruæcu'
Iodelphis minutus Paula Couto 1^962 (i.e,, that taxon is known onlybyits §pe, a
LAND MAMMALS IN SOUTH AMERICA .1,/
Ieft mandibular ramus with M3-5; see Marshall 7987:fig.26).InMinusculodel-
phis tlrre basined talonid is very foreshortened anteroposteriorÿ, the protocris-
tid is verv high and bladelike, and the paraconid projects far anteriorÿ such
that the paracristid is extremely elongated and bladelike. Given the similariÿ in
size ofJaskhadelphys andMinuscubdelphis (the smallest marsupial known in
the Itaborai fauna of Brazil), and the complementariÿ of their dental special-
izations, the authors assigîJa.skhaÀelphys to the didelphid subfamiÿ Deror-
hlmchinae (sensu Marshall 1987). This subfamily was preüously known only
by two Itaboraian taxat Derorhynchus singubtris Paula Couto 1952a and M.
minutus Paula Couto 1962, both ofwhich are represented onlybylower denti-
tions (MarshalI79B7).Jaskhadelphys thus represents the earliest and only Late
Cretaceous member of this subfamilyknown. The authors rccog;nze the possi-
biliÿ that given the subequal size of the paracone and metacone, Jaskha-
delphys and, by inference, Derorhynchtts and Minttsculadclphis' may prove
referable to a peradectid derivative distinct from the Didelphidae.
ETYIV{OLOGYI Nwsia, for the nearby place-name Antn Khasa.
TYPE-SPECIES: Khasin cordilbrensis sp. nov.
DIAGNOSIS: Same as for §.pe and only known species.
Farnily Microbiotheriidae
Subfamily Microbiotheriinae
Khasîq gen. nov.
Khasia cordillerensis sp. nov.
(Figure 9) ET\1{OLOGY: cordilbrensis, for the Andes Cordillera of Bolivia.
TYPE-SPECIMEN: YPFB Pal 6133, a right M3 or Ma missing anterior surface of
protocone (Figure 9A).
HYPODIGM: Type and YPFB Pal 6734, a right M, or M, (Figure 9B).
MEASUREMENTS: YPFB Pal 6133 L : 2. 0 mm, W : 2.6 mm; 6134L : 1.8 mm'
\À/:1.1 mm.
DIAGNOSIS: Dental formula unknown; molar crowns brachl'dont; protocone
large, trigon with shallow basin; para- and metaconules well developed; para-
and metacone high, subequal in size; centrocrista straight; post-protocrista ex-
tends labially to a point below tip of metacone; paracingtrlum distinct and
unites paraconule with s§rlar cusp A; paracrista very short, weak, unites para-
cone with base of s§,;lar cusp A; metacrista weakly developed, unites metacone
with sÿlar cusp E; sÿlar sheHvery reduced particulariy anteriorÿ: sÿlar cusps
very reduced or absent and labial edge of sheHbordered by low, thin, continu-
ous ridge; ectoflexus sharply indented at position of sÿlar cusp C; talonid dis-
tinctly basined, wider than trigonid; anterobasal cingulum weakly developed;
paraconid considerably reduced relative to protoconid, metaconid intermedi-
àte in size between paraconid and protoconid; posterior surface of trigonid set
at oblique angle relative to anteroposterior axis oftooth (i'e., metaconid setpos-
teriad olprotoconid); hlpoconid largest oftalonid cusps in height and in occlu-
sal viewfhypoconulid higher than entoconid, both of equal size when üewed
occlusalÿ; no trace of posterior cingulid'
Figure 9. Stereopairs o/Khasia cordiller-
ensis gen. et sp. nav.: A, YPFB Pal 6733
(type), an isolated upper right M3 or Ma
in occlusal vizr,rt; B, YPFB Pal 6134, iso-
lnted right M3 or M, in occlu.sal Aefr)
and labial (right) views. (both x 72)
A
,ÆilMffiæ
JO MARSIIALL €e DE MUIZON
COMMENTS: Khasia compares best with specimens of Microbiotheriidae of
the subfamily Microbiotheriinae figured in Marshall (7982)' a subfamily that is
endemic to South America, is represented by numerous taxa in rocks of middle
Tertiary (Colhuehuapian and Santacrucian) age in Patagonia, andbythe living
speciesDrombiops australis of central Ch:ILe. Khasia is assigned to the Micro-
biotheriinae (sensuMarshall 1987) based on overall structure ofthe upper mo-
lars (but note similarities betrveen lype of Khasia andM+ of Pedîomys cooki
figured by Clemens 1966:fig. 24) and absence of a posterior cingulid on the low-
er molars. A posterior cingulid is well developed in members of the Pedio-
myinae (sensu Marshall 1987; see Clemens 1966 for figures) but absent in
Microbiotheriinae (see Marshall 1982 for figures). Khasia differs from other
members of the subfamily Microbiotheriinae in having the entoconid reduced
relative to the hlpoconulid. In size and most other sûuctural features of the low-
er denüüon, Khasia compares best with specimens of fumicrobiotherium gau-
dryi descibed and figured by Marshall (7982) from the eaÙ Eocene
(Casamayoran) ofArgentina. Khasia is the oldest named member of the family
Microbiotheriidae and subfamily Microbiotheriinae in South America. A prob-
able microbiotheriid is also reported by Sigé (7972) (: ?Pediomyidae, gen. et
sp. indet. 1) from Laguna Umayo, Peru.
ETYMOLOGYI Kollpa-, for the nearby Quechua place-name Kollpa Ma)Tr.
TYPE-SPECIES: Kollpania tiupampina sp. nov.
DIAGNOSIS: Same as for §.pe and only known species.
Superfamily C aroloameghinoidea
Famiÿ Caroloameghiniidae
Subfarniÿ Protodi delphinae
Kollpania gen nov.
Figure 7O. Stereopair o/Kollpania tiu-
pampinagen. et sp. noÿ., YPFB Pal 6735
(We); *oktedW M4? in occlusalvian.
(x 72)
ET1'1\{OLOGYt tiupampina, for the locality Tiupampa.
T\?E-SPECIMEN: YPFB Pal 6135, an isolated upper left M4? (Figure 10).
III?ODIGM: Tlpe only.
MEASUREMENTS: L:2.6 mm, W:2.4mrn,
DIAGNOSIS : Dental formula unknown; small-sized protodidelphine; brachy-
dont molar crowns; protocone large, with trigon basin position filled by swol-
Ien labial side of protocone; no trace of para- or metaconule, sÿItr cusp C,
paracrista, metacrista, ectoflexus, or centrocrista; para- and metacone very re-
duced, subequal in size, lower than protoconel s§.{ar sheHlarge, cusps B and D
enormous, B slightly larger than D, both compressed transversely; s§;lar cusp A
's,ell developed but much smaller than B and D; E very small; weak cingular
sheHextends from anteromedial surface of protocone to sÿlar cusp A, another
from posterolinguat surface of protocone to sÿlar cusp E.
COMMENTS : K. tiupampina is fhe earliest and only Late Cretaceous member
of the caroloameghiniid subfamÿ Protodidelphinae known. This subfamily
was preüousl), known only from taxa of middle Paleocene (Itaboraian) age in
Brazil (i.e., BobbschafferktPaala Couto 1952a; Guggenhcimia Paula Couto
7952a; Protodidclpltrs Paula Couto 1952a; Zeusdclphys Marshall 1987), and
ReigiaP ascual 1983 from the middle Tertiary ofArgentina (see Marsh all 7987).
Marshall (7952) proposed that Caroloameghiniidae in general and Protodi-
delphinae in parücular gave rise to caenolestoids (sensu Marshall 1980), and
developed in that group quadritubercular upper molars by enlargement and
posterolingual expansion of the metaconule. In all known respects, Kollpania
represents an ideal strucfural ancestor for caenolestoids. The generalized up-
LAND MAMMALS IN SOUTH AMERICA
Ibllpania tiupampina sp. nov.
(Figure 10)
39
æ
per molar structure of caenolestoids as seen in the middle Miocene (santacru-
ôian) genus § tilatherium (see Marshall 1980 :fi g. 5) can easily be derived fiom a
Kollpània-ltkeancestorbyloss ofparacone, posterolingual expansion ofincipi-
enflÿ devetoped cingutar Àheralongthe posterior surface of the protocone, and
antérior displacement ofthe protocone so that it sits directly linguad of the enor-
mous stylaicusp B. Ifthis scenerio is true, then the quadritubercular structure
of the upper môh.s of caenolestoids resulted from expalsion of the posterior
cingular ihey, not the metaconule as proposed by Marshall (19s7). In addiüon,
if ai ancestral-descendant relationship truly exists for Kollpania ald caenoles-
toids, then it is possible that Kollpaninrnay prove referable to the superfamily
Caenolestoidea. The important point for now is rhat Kollpanüa reirrfcrces Mar-
shall,s {7gs7) predicüons that caenolestoids could be derived from protodidel-
phine caroloameghiniids, at least in terms of upper molar strrrcture'
This taxon was the first mammal described and named from the Tiupampa lo-
cal fauna and is the onÿ marsupial from this fauna named prior to this study
(for figures and descriptions of this taxon see de Muizon, Marshall et al. 1984;
Marshall et al. 1983a, 1985)'
Subfamily Caroloarne ghiniinae
Robertholfstetteria
national geo gr ophîc a Marshall,
de Muizon fos Sigé 1983a
Sup erfamily Borhyaenoidea
Famiÿ Borhyaenidae
Subfamily Hathliac;,,ninae
Allqokiru"s gen. nov.
Allqokirus australis sP. nov.
(Figure 11)
ET1.À.{OLOGy: Allqo-, the Quechua wofi(allqô) for dogi and -kiru. the Que-
chua word for tooth.
TYPE-SPECIESI Allqokirus australis sp. nov.
DIAGNOSIS: Same as for §pe and only known species.
ETYMOLOGY; australis, Latin for south or southern.
T\?E-SPECIMEN: YPFB P al 67O4, a fragrnent of a right maxillary with a nearir-
complete M3? (Figure 11,4')
ITYPODIGM: T1.pe and YPFB Pal 6188, a left Mr? missingtips oftrigonid cusps:
}?FB Pal 61S9l; Ieft M, or M, missing trigonid anterior to protocristid; and
YPFB Pal6190, a complete left Mn? (Figlrre 118).
MEASUREMENTS: 6704 L: 3.2 mm, W : 3.4 mm; 6188 L : 3.2 mm, W : 1' 6
mm;6190 L:3.9 mm, W:2.1 mm'
DIAGNOSIS: Dental formula unknown; very small-sized borhyaenid, averages
30olo smaller thanPatenc simpsoni in comparable linear tooth dimensionsl on
M3? protocone well developed, trigon baiined; par-aconule and metaconule
.*uli but distinct; paracone slightly lower ald smaller than metaconel meta-
cone slightly compàssed tranwérsely, hi.gh relatiYe to occlusal surface ofrest of
tooth, u"nità basally by'i'veak11z flsvgl.red straight (predilambdolo1t) centro-
crista; stylar shetfbioaâ; ectofiexus shallorv; metacrista moderately developed;
stylar'cusp B very large, only slightl,v lower than paracone; sÿlar cusp B and
p*..ot "ïttit"d at anterior ends by iow, st-raight p-arlcrista;-sÿlarcusp D very
ir"rny developed and anteropostériorÿ elongated along -labial edge of stÿlar
shelf; stytar cusps C and E abient; sÿlar cusp A very small (i'e', smaller than
stylar cüsp D) andjoined with paraconule !Y narrow palacingulum: protocone
,rL"o* urrt".opoü"torÿ relJtive to labial side of tooth; paracone and meta-
cone set approimately àedially relaüve to labial and lingual edges of crown;
on Mn? pidtoconid'r"ry 1"rg", much highel than paraconid and metaconid;
prr"Â"ia and metaco"nid sirbequal in height, metaconid slightly less robust
Lasally; protocristid bladelike; a:nterobasalcingulid well developed; tigoIriq
notabiÿ.vider than talonid; hl,poconid, the largest ofthe talonid cusps, is united
to postlrointernal surface of piotoconid by sharp crista obliqua-; entoconid very
"eàuced; talonid weakly baJined, and opens lingually üa notch b_etrveen ento-
conid and metaconid; distinct posterocingular shelf extends ventrolabially from
tip of hyroconulid across posterior surface of hypoconid.
CoMMENTS: A. ratstralis is the oldest known member of the endemic south
American marsupial "doglike" family Borhyaenidae. This spe,cies compares
most favorabty t^tt f at"ri simpsonii,auh ôouto 1952b (inclu dirlg lschyrodi--
detphis casteilanosi paula Couto lg52a as a junior s1,'nonlan, see .Marshall
rgar) from the middle Paleocene (Itaboraian) àf eruzi.. P. simpsoniis the old-
40 MARSHALL €P DE MUIZON
Figpre aL. Stereopairs o/Allqokirus
australisgen. et sp. nov.: A, YPFB Pal
670a (Qpe), fragment of right ma.æilbry
with M3Z in occlusalvia»; B, YPFB Pal
6190, complcte W Mn? in occlusal Aefi)
and labial (right) views. hoth x 8)
A
Figure 72. Stereopair olfPeradectes aus-
trinum (Sige 7977), YPFB Pal 6732; lejL
Ms or Ma, rtissin§ anterolabial corner
of crown, in occlusalview. (x 72)
est known species of Borhyaenidae in the Cenozoic and is referred to the sub-
familr, Hathliaoninae fMarshall 1981). HathliaoTrines are dentall]'the most
generalized of knou'n borhvaenids, and earlv forms (i.e.' Patene) make ideal
su'ucrural ancestors for all later members of this group. Allqokirus is now the
most generalized member of the HathliaoT rinae knou'n and ihus has the dis-
rinction of approximatin$ or representing the ancestral borh)'aenid'
Alphadon antstrinum was erected by Sigé (7977) on the basis of three partial
upper molars and a right upper premolar from the Late Cretaceous (Maestrich-
tian) Vilquechico Formaüon at Lagpna Umayo, Peru (it is the first Cretaceous
marsupial described from South America). Detailed descriptions and figures of
the §pe and referred specimens are proüded by Sigé (1972). Crochet (e.9.:
1 980) subsequently place d a ustrinum in the genus P er ade cte s which was pre-
viousÿ known from the early Tertiary of North America and Europe'
An isolated upper molar from Tiupampa (YPFB PaI 6732, a left M3 or Ma
missing anterolabial corner of crown) (Figure 12) is structurally inseparable
from the §pe of P. austrinum and is accordingly referred to that sPecies. The
Tiupampa molar is of generalized peradectid structure (sensu Crochet 1980)
with paracone and metacone subequal in size and height, a straight (predi-
lambdodont) centrocrista, moderately developed metacrista, shallow but dis-
tinct ectoflexus, protocone large with shallow basin on trigon, distinct
paraconules and metaconules, and a distinct paracingtrlum. This tooth is
slightÿlarger (i.e.. L-1.8 mm, W: 2.4rnn;,) thanthe §pe ofP. austrinum as
reported by Sigé {7977:2480) (i.e., a left Ma missing protocone and sÿlar cusp
A; L:7.48 mm). This size difference may be due to the fact that different teeth
are being compared (i.e., the §pe is thought to be an Ma while the Tiupampa
specimen maybe an M3, or vice versa), thatit simply represents individual vari-
ation, or that the §pe and Tiupampa specimen are both incomplete and the one
comparable measurement is approximate.
LAND MAMMALS IN SOUTH AMERICA
Superfamiÿ Microbiotherioidea
Farniÿ Peradectidae
Peradectes austrinum
(Sigé 1971) (Figure 12)
47
Marsupial Evolution in the Late
Cretacizotls of South America
The earliest known records of marsupial (and undisputed placental)
mammals in South America are based on fossils from the El Molino For-
mation at Tiupampa, Bolivia, and the Vilquechico Formation at Laguna
Umayo, Peru (Sigé 797 7, 797 2). The marsupial fauna from Tiupampa is
represented by 77 genera and species: one Peradectidae (Peraàectes
austrinum), one Microbiotheriidae (Khasia cordilbrensrsl, slx Didel-
phidae (Ptrcadelphys andinus, Incaàclphys antiquus, Mizqucdclphys
pilpinensis, Andinodelphys cochabambensis, Tiulardia J'laresi, J ask'
hadelphys minutus), trvo Caroloameghiniidae (Kollpania tiupampina,
Rob erthoffste\teria natbnalgeogr aphic a), and one Borhya erttdae (
All-
qokirus australis) (Table 1). The marsupials from Laguna Umayo are
represented by one Peradectidae (Peradectes attstrinum), one Micro-
biôtheriidae (sensu Marshall 7987;: ?Pediomyidae gen. etsp. indet. 1
of Sigé 7972), and at least one Didelphidae (Sigé 7972; gen. et sp. indet. ).
Comparison of the marsupial faunas of Tiupampa and Laguna
Umayo with that ofthe middle Paleocene (Itaboraian) local fauna at Ita-
boraf, Brazll (Marshall 7987 andreferences therein), reveals the follow-
ing similarities and differences: no genera or species are shared in
common; five taxafrom Tiupampa (i.e., Pura^delphys,Incadelphys, An-
dinodctphys, Tiulnrdia, Allqokirus) are stmctura§ more generalizgd
and represent potential ancestors for Itaborai taxa (i.e., Sternbergio,
Marunôsopsis, Itaboraiàelphys, Gaylnrdia, P atene, respectivel,v) ; the
family Peradectidae (Peradcctes austrinum) is known only from Tiu-
pampa and Laguna lJmayo; the famiÿ Poÿdolopidae (Epi"dalnps
àmeÿinoi Paula Couto 7952b) is known only at Itaborai; the subfami-
lies Eobrasiliinae, Derorhyrrchinae, Caroloameghiniinae, Protodidel-
phinae, and Hathiiacl,ninae are known only from Tiupampa and
itaborai; and the families Microbiotheriidae and Didelphidae are
known from all three, Itaborai, Laggna Umayo, and Tiupampa.
The Late Cretaceous (Maestrichtian) marsupials of South America
compare favorably in "adaptive §pes" with those of similar agein North
America (Clemens 1966,1979): e.g'r Borhrraenidae (i.e., Allqokirus) of
South America have molar specializaüons similar to earÿ Stagodonti-
dae (i.e., Eodetphis) ofNorth America (see Fox 1981 for a reüew ofthis
famiÿ); Caroloameghiniidae (Caroloameghiniinae) have distinctll'
bunodont dental tpeèiaHrations and are known from North (i.e., Glos-
bius) and South (i.e., Roberthffitenerkt) America; Peradectes is
known from the Late Cretaceous of South America and the ear1r, Paleo-
cene of North America (Crochet 1930); Didetphidae (e.$', Pttradelphys)
are similar in size and general structure to species of Peradectidae (e.$.,
Alphadon) of North Americal and the family Microbiotheriidae (sensu
Marshall 1987) is known from both North (i.e., Pediamys, see Clemens
7966,7979) and South (i.e., Khasia) America.
The fossil record of marsupials in South America thus demonstrates
that these animals were abundant and taxonomically diverse by Late
Cretaceous time. The occurrence ofatleastfour superfamilies, five fam-
ilies, seven subfamilies, and 11 genera and species at Tiupampa sug-
gests that marsupials had a substantial but süll undocumented pre-
Maestrichtian history on that continent.
Two Upper Cretaceous mammal-bearing local faunas of pre-Maes-
trichtian àge a"e known in south America, but no marsupials (or un-
doubted plàcentals) have yet been reported from them. Firsl, Mourier et
MARSIIALL €E DE MUIZON
11
aI. (198G) report and figure a trigon of a "therian" upper molar thatwas
collected in associaüon with dinosaurs near Bagua in north-central
Per-u. An upper Santonian to Campanian age has been suggested for this
Iocal fauna. Second, two Eupantotheria (Dryolestoidea, Dryolestidae in-
det.; and Mesungulaltdae, Mesungulatum houssayi), a Triconorlonta
(Triconodontidae indet.), a possible Multituberculata (family indet.),
and a possible Paratheria (Gondwanatheriidae, Gondwanatherium pa-
tagonirum) are reported from the Campanian age Los Alamitos Forma-
tion near Arroyo Verde, Rio Negro province, northeastern Patagonia,
Argentina (Bonaparte 1986; Bonaparte €e Soria 1983, 1985). Both of
these local faunas are süll poorly known and the mammals are repre-
sented only bv isolated teeth or fraglnents of isolated teeth. Thus, the ab-
sence of marsupials (and undoubted placentals) cannot be interpreted
to indicate their nonexistence during the time of accumulation of these
local faunas. In short, absence ofevidence is not evidence ofabsence.
Placental Evolution in the Late
Cretaceous of South America
Undisputed placental mammals are also fïrst known in South America
in the Maestrichtian local faunas of Tiupampa and Laguna Umayo
where members of four orders (Proteutheria, Pantodonta, Condy-
larthra, Notoungulata) have thus far been recorded.
Figpre 73. Stereopairs o/9[ Cimolestes
sp., YPFB Pal 6737, isolated lS tvlr?:
(A) la-bial and (B) occlusalvi.ews. (x 16)
Proteutheria
The only South American record of proteutherians is from Tiupampa
r.r'here the group is represented by a partial isolated left P, and several
isolated lower molars. The P, (MNHN Vil 121; described and figured by
de Muizon, Marshall et al. 1984:figs. 18 €c 19; and Marshall et ai. 1985:
frg. f) has a r,vell-developed metaconid and distinct talonid, suggesting
affinities with species of Gypsonictops and Procerberus. The lower mo-
lars, exemplified by YPFB Pal 6131 (Figure 13), have narrow talonids
and an overall trigonid structure which suggests affiniÿwith members
of the fu-rly Palaeoryctidae (sensu Kielan-Jaworowska et al. 7979).
This particular molar is smaller (L : 1.7 mm, W : 1.1 mm) than in most
species of Cimobstes, butlarger thaninBatodontenuis. The size differ-
ence between the larger partial Pn and the smaller molars, and their
structural affinities with different genera, suggest that two proteuther-
LAND MAMMALS IN SOUTH AMERICA
æ3 B
+J
ians are represented. The taxon represented by the lower molars is
tentatively referred to cC. Cimobstcs sp. and placed in the family Pa-
laeoryctidae, while the taxon represented by the Pn is iisted as Proteuth-
eria, famiÿ, gen. et sp. indet. (Table 1).
Pantodonta
Pantodonts are known in South America only from Tiupampa where
they are represented l:y Abi"dedorbignya inopiruata de Muizon ôp Mar-
shall 1987a, the earliest member of the family Pantolambdidae {sensu
Simons 1960). This animal is represented by its §pe (YPFB Pal6727, a
nearÿcomplete maxillawith most ofP4-M3) (Figlrre 74A), adozen oth-
er partial upper and lower jaws with teeth (e.S.r Figure 148, C), and
abôut four dozenisolated cheek teeth of both adult (e.$., Figure 74D,8)
and juvenile individuals . Abidedorbignya is the most abundant placen-
tal in this local fauna. The partial upper molar IMNHN Vil 103), de-
scribed by Marshall et al. (1983a) and de Muizon, Marshall et al. (1984:
frg.Z) as a right Ma of a large didelphoid marsupial, is actua111r the M3 of
A. inopinata (see de Muizon €p Marshall 7987a).
Pantodonts (sensuSimons 1960, : Taligrada of Chor'v & \\'ang 1979)
are known from the earÿ? Paleocene to earÿ Oligocene of Àsia' and
middle Paleocene to late Eocene of North America. Memtrers of the fam-
iÿ pantolambdidae are known only from the middle Paleocene to late
Ebcene of North America. The middle Paleocene genus Pantolambda
"includes the smallest, oldest and most primitive pantodonts. . . . It ap-
proximates the structural ancestry of the order, . . . [and] Pantoktmbda
bathmodonis muchthe mostprimitive ofthe [known] species" (Simons
796A:77). A. tnopinata differs from P. batLtmodon in being about 600lo
smaller in comparable linear tooth dimensions. and stn:cturallr'more
plesiomorphic in having a straight (predilambdodont) centrocrista. rel-
àtivelybroàder sÿlar shelf, and distinct entoconid. A. inopina_ta is thus
the earliest, smallest, and most generalized member of the order Panto-
donta (sensuSimons 1960), and represents an ideal stn-rcrural ancestor
for aII known Tertiary pantodonts and tillodonts. AbidedorbigVya rhus
favors (contralucas 79BZ) acommon ancestrvfor pantodonts and tillo-
donts as championed by Chow €,, Wang (1979) ' and supports placement
of the Tillodontia (as a suborder) in the order Pantodonta.
Condylarthra
Condylarths are represented at Laguna Umayo by a single undescriLred
tooth (8. Sigé perônal communication) and at Tiupampa b' at least
three speciel, àf *hi"h Tiuclaenus minutus de Muizon €* Nlarshall
1987c ànd, MolinodtLs suarezi de Muizon €p Marshall 1987b are re-
ferred to the family Hyopsodontidae, subfamiÿ Mioclaeninae. and An-
dinodusboliviensis dê uuizon êp Marshall 1987c is classified as famiÿ
Didolodontidae or Phenacodontidae (Table 1)'
T. mirtuttts is the smallest and best-knor'rm condYlarth from Tiu-
pampa; it is represented by a nearÿ complete left mandibular ramus
with most ofthe moderatelyworn dentition (§?e, YPFB PaI 6115, Figlre
15A), and several isolated upper (e.g., YPFB Pal6119, Figure 158) and
lower molars. The lower premolars are elongate anteroposterior\" Pn
has a weakly developed mètaconid, upper molars lack a h1,ocone, Iow-
er molars have a weit-developed pâraconid, trigonid is onl' slightÿ
higher than talonid, talonid is short anteroposteriorly and narrow'er
thàn trigonid, entoconid is smail relative to hypoconirl and subequalirr
size to Ëlrpoconulid, and entoconid and hlpoconulid are fused basalÿ
(Figure iSa). T*o partial lower jaws, one with M2-3 (YPFB Pal6729)
MARSIALL €E DE MUIZON
44
Figure L4. Stereopairs o/Alcidedorbig-
nya inopinata de Muizon Ü Marshall
7987a: A, YPFB Pal 6727 (ÿpe), lejt
masilkrry with P4-M3 în occlusalvia.u;
B,t?FB Pal 6198,fragmmt of right
mandibuktr ramus wîth M3 in occlusal
(right) and ktbial (l4I) views; C' YPFB
Pal 6779, fragmznt of rîght mandibular
ratnus with Pa-M2 in labial Aefi) and
occlusal (right) views; D, YPFB PaI
6724, right Pa in occhtsal via»; Ë, YPFB
Pal 6722, right tw1 in occlusal via,u.
(all x3)
A
C
ffi@r .E
and the otherwith M, and trigonid of M2 (YPFB Pal 6130), have molars
structuralÿ similar to those of T. minutus but average 15olo larger in lin-
ear tooth dimensions. These hvo specimens were found at a different lo-
cality than the §rpe and referred specimens of T. minutus, and may
either prove to represent large (male?) individuals of that species or a
different species of that or a structuralÿ similar genus.
M. s uar ezi, the middle -sized c ondylarth found at Tiupampa, is repre-
sentedbv several partial lowerjawswith two orthree molars each (e.9.,
Figure 16A-D) and several isolated upper molars (e.9., Figure 16E, F).
This species averages 35o/o larger in linear tooth dimensions than T.
minutus, although the overall molar morpholog,r is similar in both. The
only notable difference is that in Mo lino dus the entoconid and hypocon-
ulid are fused basally into a crestlike structure that is enhanced by an en-
tocristid. This feature is present to some degree in most specimens of
North American early and middle Paleocene mioclaenines studied, es-
pecially in species of Promiaclaenus. In fact, in overall strrcture, ?iu-
LAND MAMMALS IN SOUTH AMERICA
ffi
w
45
Figure 75. Stereopairs o/ Tiuclaenus
minutus de Muizon 6? Marshall 7987c:
A,YPFB Pal 6775 (Wr), W mandibular
ramus with an incisor, canine, roots of
Pr, and Pr-Mrpresent in
occlusal (top, bottom)
and labial (middle)
vi.ews (x 3); B, YPFB Pal
6779, fragment of right
ma:rillary with M2 in oc-
clusal vi.ew (x 5).
ckrcrtus compares most closely with species of PromiocLaenus (de
Muizon €p Marshall 7987b).
A. boliviensls, the largest condylarth known from Tiupampa, is repre-
sented only by a pafiial right mandibular ramus with a worn M2 (§pe,
YPFB PaI 6120, Figgre 17).ln comparable linear tooth dimensions this
species averages 30% larger than M. suarezi and 55o/o larger than T.
minutus. The overall molar structure of Andinodus is similar to that of
Tiuclncnrs andMoltnodus, except that the hlpoconulid and entoconid
are more distinct, the entoconid lacks an entocristid, and the talonid ba-
sin is "fïlled" anteriorÿ by a posterobasal srvelling of the metaconid.
This species represents an ideal structural ancestor for North American
members of the family Phenacodontidae (e.g.' Tetraclnenodoru) and
some members of the endemic South American family Didolodontidae
(e.g., Ernestokokenio). As noted by Simpson (7948) and Cifelli (1983),
the similarities betlveen phenacodontids and didolodontids are m€mÿr
and the monophyly of these families has yet to be securely demonstrated.
The only consistent criterion used to classiÿ taxa into one or the other
family is geographic occurrence - North American taxa are placed in
the Phenacodontidae, South American taxa in the Didolodontidae. If
this practice were followed, Andinodus would be classified as a didolo-
dontid, yet if it were found in North America it would unque stionably be
assigned to the Phenacodontidae. The presence of mioclaenine hyopso-
dontids at Tiupampa demonstrates a close "faunal-1ink" with §'pical
NorthAmerican condylarth gXoups. 'lhusAndinodus mayproÿe to be a
"good" phenacodontid despite its occurrence in South America. For this
reason the authors cautiously classiÿ Andinodus as Didolodontidae or
Phenacodontidae (Table 1).
Tiuclncnus andMolinodus are the only securely known members of
the condylarth famiÿ Hyopsodontidae in South America, and are the
oldestmembers ofthe subfamiÿMioclaeninae (seræu CifeIIi 1983) any-
where in the world. Undisputed hyopsodontids are firstknown in North
America in rocks of earÿ Paleocene age and are taxonomicalÿ diverse
in Paleocene and Eocene faunas on that continent (Van Valen 1978). Ar-
46 MARSTIALL €P DE MUIZON
. ;1 .::É é
A
Figpre 76. Stcreopaîrs olf Molinodus
suarezi d.e Muizon Ü Marshall 7987b:
A, B, YPFB Pal 6772 (type), fragncnt of
l"fi mandibular ramus with Mr_uin kt-
bial (A) and occlusal (B) views (x 3);
C, D, YPFB Pal 6173, fragment of right
mandibular ramus ltith M 1-, erupted
end M3 erupting in occlusal (C) and la-
biol (D)'views (x 3); \ YPFB Pal 6777,
W M1 in occlusal vi.ew (x 5); E, YPFB
Pal 6178, right Mz in occlusalview (x 5).
C
E,F
B,
chibald (7982) tentatively assigned Oryprimus erikseni from the Late
Cretaceous (Lancian) of North America to the hÿopsodontid subfamily
Hyopsodontinae and following Van Valen (7978) recognizes the Mio-
claeninae as a distinct family, the Mioclaenidae.
Paula Couto (7952c), based on fossils from the middle Paleocene (Ita-
boraian) of Flraztl, assigned the South American condylarth Asmith-
woodwardia to the North American ft-Ü Hyopsodontidae. He later
(797 B) retracted his view and returned Asm ithwoodwardia to the Dido-
lodontidae (e.g., Simpson 1948 includedAsmithwoodwardia, based on
fossils from the late Paleocene of Argentina, in the Didolodontidae trut
noted similaritiesbehveenit and NorthAmerican members ofthe hyop-
sodontid subfamily Mioclaeninae). McKenna (1956) argued that As-
mithwoodwardia be placed (contra Paula Couto 1952c) in the
Didolodontidae, although he later (McKenna 1987:67) argued in favor
of a possible Asmithwoodwardin-hyopsodontid affinity-"It now
seems to me that (a) Asmithwoodwardkt is closely similar to Didolo-
dus, Ernestokokenkt, and LameToia in marry feafures; @) Hyopsodus
itself is the North American genus most similar to Asmithwoodwardia;
and (c) Hyopsodus andAsmtthwoodwardin could have had a common
ancestor that somehow crossed the Caribbean gap, possibly evenfrom
south to north" (italics, the authors'). More recently, Cifelti (1983) as-
signedAsmithwoodwardiato the South American endemic order Litop-
terna, family Protolipternidae. A detailed systematic revision of aII taxa
currently and preüously referred to the families Hyopsodonüdae, Dido-
lodontidae, or Phenacodontidae, and some other groups (e.9., earÿ Li-
topterna) is clearÿ needed to clariÿ the debated phylogenetic positions
of such taxa as Asmithwoodwardkt andAndinodus.
TWo other Late Cretaceous condylarths have been reported in South
America, although the specimens upon which these identifications were
made have proven referable to other orders. Eirst, Perutherium alti-
planense Thaler in Grambastet d,.7967 was named on the basis ofa left
mandibular fragrnent with talonid of M, and trigonid of M, from Lagu-
LAND MAMMALS IN SOUTH AMERICA
D
lr
47
na Umayo. This taxon was assigned by various workers to the condy-
larth families Didolodontidae (e. 9., Reig 1 981, Tedford 797 4) and Perrp-
ÿchidae, subfamily Perutheriinae (Van Valen 1978). Marshall et al.
(1983b) attempted to demonstrateüatPerutherium is a member of the
order Notoungulata and placed it in its own family, the Perutheriidae
(: Perutheriinae ofVanValenT9TB). Second, Bonaparte €p Soria (1983,
1985) proposed the rarne Mesungulatum houssayi for an isolated tooth
from the Campanian Los Alamitos Formation near Arroyo Verde, Ar-
gentina. Theybelieved it to be an upper M3 and tentatively assigrred it to
the condylarth famiÿ Arctocyonidae. Subsequent work has demonstrat-
ed the tooth to be from a eupantothere (Bonaparte 19BG).
Figure 77. Stereopair o/Andinodus boli-
viensis de Muizon &? Marshall 7987c,
YPFB Pal 6720 (ÿpe);fragnent of riÿt
mandibular ramus with M2in labbl
(lefr) and occlusal (rîght) views. (x 3)
NotoungPlata
Notoungtrlates are represented by two occurrences in the Late Creta-
ceous of South America. First, P. altiplanense from Laguna lJmavo is re-
ferred by Marshall et ai. (1983b) to the mono§pic family Perutheriidae.
And second, a broken upper molar of a notoungulate from Tiupampa is
described and figuredby de Muizon, Marshall et al. (1984: fig. tZ) and
compares with generalized members of the families Henricosborniidae
and Oldfieldthomasiidae. The generic and specific affinities ofthis taxon
are unknown (Tâble 1).
Mones (7987) notes that G ondwanatherium patagpnbum Bonaparte
1986, family Gondwanatheriidae, from the Campanian Los Alamitos
Formation of Argentina,, ar:d Sudamerba ameghinoi Scillato-Yané €o,
Pascual 1984 (see also Scillato-Yané €p Pascual1985), family Sudameri-
cidae, from the middle Paleocene (Itaboraian) of Chubut province, Ar-
gentina, are both characterized by high-crowned (protohypsodont or
euhlpsodont) cheek teeth, thick enamel, and variable occlusal patterns.
He places these two families in a new order, Gondwanatheria, and ten-
tatively assigns it to the cohort Edentata, superorder Xenarthra. If he is
correct, fiien Gondw anatherium r,vould represent üre earliest placental
mammal yet known in South America. There are no known specimens
of Gondwanatheria from either Tiupampa or Laguna Umayo.
Conclusions
The earliest known records of marsupial and undisputed placental
mammals in South America are from the Tiupampa 1oca1 fauna (El Mo-
lino Formation) of Bolivia and the Laguna IJmal.o local fauna (Vilque-
chico Formation) of Peru. Species from Tiupampa include at least 11
marsupials and seven placentals, and those from Laguna Umayo in-
clude at least three marsupials and trvo placentals. These faunas docu-
ment a rich and taxonomicalÿ diverse radiation of therian mammals in
South America's Late Cretaceous (Maestrichtian).
The Tiupampa mammals include the earliest South Americal re-
cords of Didelphidae (Didelphinae, Eobrasiliinae, Derorhlmchinae),
Microbiotheriidae (Microbiotheriinae), Caroloameghiniidae (Caroloa-
me ghiniinae, Protodidelphinae), and Borhyaenidae (Hathliaclnrinae) ;
and the oniy South American records of Proteutheria, Pantodonta, and
rc MARSIIALL €P DE MUIZON
hyopsodontid Condylarthra (Table 1). The lar$est mammals in this fau-
na are the cat-sized pantodont (Abidedorbignya), notoungulate (gen. et
sp. indet.), and didolodontid or phenacodontid condylarth (Andino-
ùn); ttre other taxa are all rat-sized or smaller. The mammals show a
spectrum of dental specializations indicaüng carnivorous (Borhyaeni-
dae), insectivorous (Peradectidae, Microbiotheriidae, Didelphidae,
Proteutheria), omnivorous (Caroloameghiniidae, Condylarthra), and
herbivorous (Pantodonta, Notounguiata) feeding specializations.
The earliest record of Cenozoic mammals in South America is from
localfaunas ofmiddle Paleocene (Itaboraian) age inArgentina and Bra-
zil (Marshall 1gs5 and references therein). The absence ofrecorded ear-
ly Paleocene mammals in South America leaves a hiatus of at least 3
million years between the Maestrichtian faunas of Boliüa and Peru, and
the Itaboraian faunas of Argentina and Btazil, The ranges of subfami-
lies ald families of mammal taxa from the Late Cretaceous faunas of
south America are shown in Figlrre 18, with indication of their known
records in Paleocene faunas on that continent. No species or genus is
known to occur in both Late Cretaceous and middle to late Paleocene
faunasl yet of the eight suprageneric rarrks of marsupials known in the
Late Cretaceous, seven occur in the middle Paleocene of South America,
while ofthe six Late Cretaceous placental ranks only three occur in mid-
d1e Paleocene faunas on that continent. The one marsupial and three
placental groups that apparent\' do not cross the Cretaceous-
Tertiarv boundan'il South America (i. e,, Peradectidae, Palaeoryctidae,
Pantolambdidae, Hvopsodontidae) do however occur in early Tertiary
faunas in Nor-th America. These data suggest that for South America
there were no "real" exlinctions of suprageneric ranks of mammals at
the end ofthe Cretaceous, but that there were shifts or restrictions in geo-
graphic ranges of some groups.
Most of the mammals at Tiupampa and Laguna Umayo are assigned
to families or subfamilies that are also known from similar or slightly
younger rocks in North America [i.e., Peradectidae, Microbiotheriidae,
Caroloame ghiniidae, Palaeon,ctidae, Pantolambdidae, Hyopsodonti-
dae, Phenacodontidae - Didolodontidae). The existence of these distri-
bution patterns can be explained by vicariance (disjunction of
prerdouslywidely diskibuted groups), dispersal, or a combination of the
two. For example, the distribution of the fu-ily Peradectidae has been
attributed to r.icariance (Marshall 7987), while the distribuüon of Pera-
dectes is atu'ibuted to dispersal of a member of this genus from South to
North America in Late Cretaceous or early Paleocene time (Crochet
19S0). The presence of caroloameghiniids (i.e., Ghsbius) in the Late
Cretaceous (Maestrichtianl of North America is atlributed by Marshall
(1987) to dispersal of a stockfrom South to North America in latest Cam-
panian or earliest Maestrichtian time. Didelphids are first known in
North America in rocks of early Eocene age and their presence is athib-
uted to dispersal of a member ofthis family from South to North America
in latest Paleocene time (Crochet 1980).
The paleobiogeographic history of South American proteutherians is
obscure because this group is represented by only a few isolated teeth at
Tiupampa and these have not yet been securely identified to the level of
genus or species. One taxon appears to be a palaeoryctid and is tenta-
tively identi{ied as cf. Cimolèstes sp. (Table 1). The family Palaeorycti-
dae is first known from the Late Cretaceous (middle Campanian) ofAsia
and North America, while the genus Cimolestes is known from rocks of
the Late Cretaceous (Maestrichtian) and earÿ Paleocene (Puercan) in
North America (Kielan-Jaworowska et aL.7979). The other taxon is of
LAND MAMMALS IN SOUTH AMERICA Lq
Figure 78. ktnges of supra§eneric
mammal tara knownfrom the Late
Creta.ceous of &suth Amerira, with
thcir occurrertce in Pabocene faunas
on that contînerut, This illustrates
aspects of mammalktn evolutinn
acro s s the Cretru:eorx - Tertiary
boundary in South Amnrira (erpand-
ed afier Case Ü Woodburne 1986:
fi*. 1). T hE tunt&üve record of Dirlnln-
Late Cretaceous Paleocene
è
è-t
§
§
/è §§
^+-'
s€x §YW
MARSUPIATS
Peradectidae
Microbiotheriidae
Microbiotheriinae
Didelphidae
Didelphinae
Eobrasiliinae
Derorhynchinae
Caroloameghiniidae
Caroloameghinae
Protod idelphinae
Borhyaenidae
Hathliacyninae
PTACENTATS
Palaeoryctidae
Pantolambdidae
Hyopsodontidae
Didolodontidae
Henricosborniidae or
@"
i>
|>
donùâae É based on l\l}(l7rùJs1.ÿnlcn,
as indbated in th.e teït, may be a
Phenarodnnti.d.a.e.
ulolrelotnomasllclae
Condwanatheria
PICKS (Ma) 84.0 74.s 66.4 63.6 61 .o 57.8
less certain affinity, but shares features with both Gypsonbtops and
Procerberuts which are classifiedbyKielan-Jaworow.ska et al. (1979) in
different superfamilies (i. e., Leptictoidea and Palaeoryctoidea, respec-
tively) within the Proteutheria.
The Tiupampa pantodont, Abi-dedorbignya inopinata' is the small-
est, earliest, and structurally most $enerali-ed memtrer of that order
(sensu Simons 1960) and of the fami}r'Pantolambdidae (de Muizon &
Marshall 7987a). Pantodonts are first known in rocks of earÿ? Paleo-
cene age in Asia and in rocks of middle Paleocene a$e in North America
(Chow-€p Wang 7979). The age and genqralized stmcture of Abi"dedor-
big4tyarnakeit an ideal ancestor for the order Pantodonta as defined tra-
ditionally by Simons (1960), and more broadiy by Chow t* Wang (7979)
to include the Tillodontia. The known pantodont record suggests dis-
persal of an Abidedarbignya-ltke stock from South to North America
ând then to Asia in latest Cretaceous or earliest Paleocene time. This
model predicts that the absence of pantodonts in the earlv Paleocene
(Puercàn) of North America is due to sampling bias (i.e., pantodonts
were there at that time but the fossils have not yet been discovered)'
Hyopsodontid condylarths of the subfamily Mioclaeninae are repre-
sented at Tiupamp aby Tiurlnenus minutus and Moltnodus suarezi. In
North America, this subfamiÿ is first recorded in early Paleocene (Puer-
can) rocks (Van Valen 7978). This distribution pattern and the plesio-
morphic structure of Tiuclacnl6 suggest that a mioclaenine stock
dispersed from South to North America in latest Cretaceous or earliest
50 MARSIIALL €P DE MUIZON
rtlt
Paleocene time. The only Late Cretaceous (Maestrichtian) condylarths
known from North America are Protunÿlntum donnae and P. gorgun
(Arctocyonidae, Oxyclaeninae), Ragnarok harbirhtt (Arctocyonidae,
Loxolophinae), Mimatuta morsoth {Penpÿchidae, Anisonchinae),
and Oryprimus erikseni (?Hyopsodonüdae, ?Hyopsodontinae) (Archi-
bald 1982,Johnston & Fox 79B4,Van Valen 7978).
Andinodus boliviensis represents either a member of the famiÿ Dido-
Iodontidae or Phenacodontidae. The monophyly ofthese families hasyet
to be demonstrated, and the structural similariÿ of earÿ members of
both may be due to common ancestqz and not parallelism or conver-
gence in evolution. The authors believe that Andinodus may represent or
approximate an ancestral taxon for some or all didolodontids or phena-
codontids. If this proves correct, then the distribution of these families
can be explained by dispersal of anAndtnodus-like stock from South to
North America in latest Cretaceous or earliest Paleocene time.
The oldest known members of the order Notoungrrlata are repre-
sented by Perutherium altiplanense (famiTy Perutheriidae) at Lagrrna
Umayo, and an indeterminate upper molar (cf. Henricosborniidae or
Oldfieldthomasiidae) at Tiupampa. Notoungulates are fïrst known in
North Americaby Arctostylaps (familyArctos§;Iopidae) in rocks of late
Paleocene (Tiffanian) age, and in Asia (Mongolia) by Palaeosÿbps in
the latest Paleocene (Gingerich 1985). The biogeographic history of no-
toungulates is athibuted by Gingerich (1985) to a late Paleocene dispers-
al of a stock from South to North America and from there to Asia.
Members of the orders Edentata and Dinocerata (sensu Gingerich
1985) are first securelyknown in rocks ofmiddle Paleocene (Itaboraian)
age in South America and in rocks of late Paleocene (late Tiffanian) age
in North America. The distribution ofboth groups is athibuted by Gin-
gerich (1985) to possible dispersal of a South American stock to North
America during earÿ late Paleocene time.
In summary, the fossil land mammal fauna from Tiupampa in south-
central Bolivia provides evidence of a rich and taxonomically diverse
marsupial and placental radiation during Late Cretaceous fMaestrich-
tian) time in South America. These fossils, and those from the temporal-
ly equivalênt Laguna Umayo local fauna of Peru, demonstrate that South
America was a major theater for the earÿ cladogenesis of therian mam-
mals. The record also suggests that many aspects ofthe paleobiogeogra-
phic history of Late Cretaceous and Paleocene marnmals of South
America can be explained by envisioning dispersal of stocks from, and
not to, that continent. Four Late Cretaceous-Paleocene dispersal events
of mammals from South to North America are suggested: late Campa-
nian or early Maestrichtian (Caroloameghiniidae), late Maestrichtian '
or early Paleocene (Peraàcctes, Hyopsodontidae, Pantodonta), eaù
late-Paleocene (Notoungulata, Edentata, Dinocerata), and latest Paleo-
cene (Didelphidae). The disû:ibution ofthese and other groups ofterres-
trial and freshwater vertebrates (Bonaparte 7984a, b; Estes €* Bâez
1985; Rage 7978, 1987,1986b) demonstrate that opportunities for dis-
persal betrveen the Americas were apparentÿ available on multiple oc-
casions at the "Dawn of the Age of Mammals."
Postscript
During J:uLy 79ï7,Jean Dejax (Institut de Paléontolo$e, Muséum nation-
al d'Histoire naturelle, Paris) recovered at least trvo §pes of charophytes
from a level about 25 m above the mammal-bearing unit at Tiupampa.
He is now comparingthemwith the charophles collected in association
with mammals at Laguna Umayo (Sigé 1972) , and with others from oth-
LAND MAMMALS IN SOUTH AMERICA 51
er localities of Late Cretaceous and Paleocene age in South America.
A new mammal localiÿ of Late Cretaceous (Campanian-Maestrich-
tian) age has recently been reported from the Anacleto Member of the
Rio Colorado Formation in Rio Negro province, northern Patagonia, Ar-
gentina (Goin et al. 19BG). A single fragrnent of a right mandibular ra-
mus with alveoli of the last two molars, ofvery small size, is tentatively
idenüfied as a marsupial. However, given the fragrnentarynature ofthis
specimen and lack of diagnostic marsupial characters, the authors re-
gard it as Mammalia incertae sedis.
Aclorowledgtnents
The fieldworkwas made possible by grants from the National Geoglaphic Socieÿr and
an "Action S1Écifique" Ilom the Muséuni national d'Histoire naturelle (MNHNI, Paris,
France. Fossil collectingwas initially carried out under the auspices ofthe Instituto Bo-
liüano de Biolo$a de Altura (IBBA) and the Office de la Recherche Scienüfi{ue et Tech-
nique d'Outre-Mer (Mission ORSTOM) r and recentlv in collaboration with the Centro de
Teènolo$a Petrolera (CTP) of Yacimientos Petroleros y Fiscales de Bolivia (YPFB). The
fossils are the properÿ of CTP (l?FB) and are providéd on loan to the MNHN for cura-
tion and study. Special thanks to Mario Suârez Riglos, Ramiro Suârez Soruco, Carlos
Molina Montes, andJorge Lobo Boneta for their collaboration and logistic support dur-
ing the 1984 and 1985 field seasons. E. Buffetaut, M. Gayet, andJ. -C. Rage kindly pro-
vided updated taxonomic lists of the actinoptery$ans, amphibians, and repüIes.
Completon of tlis research was made possible by a Collaborative Research Grant from
the North Atlantic Trealy Orgarization (NATO), and a "professeur associé" appoint-
ment ofLGM at the MNHN. D. Russell proüded casts for comparative study, and helpful
comments on this paper. The stereophotos were made by Denis Serrette (MNIil\fl.
Bibliography
Archer, M.
1978. The nature of the molar-premolar boundary in marsupials and reinterpre-
tation of the homologr of marsupial cheekteeth. Memoirs oftlæ Quemslanà Mu-
seum 78:757-764.
Archibald,J. D.
aSBZ:. Astudy of Mammalia and geolos/ across the Cretaceous-Tertiarybound-
aryin Garfield Counÿ, Montana. lJruîversiÿ ofCalifomiaPubliratians inGeolog-
. iral ki.ences 122:7-286.
Archibald,J. D.; €c Clemens, W. A. Jr.
a982, Late Cretaceous extinctons. Amerban Scicntist 70:377-385.
Bonaparte,J. F,
lq&qa. El intercambio faunistico de vertebrados continentales entre América del
sur y del Norte a fines del cretâcico. Memair d-el III Congreso Latinoamerbano
P abontoln gîa, Mexico Cily, 43 B - 45O.
aSA4b. Lâte Cretaceous faunal interchange of terrestrial vertebrates tretween the
Americas. Reif, \,V. -E. êc Wesphal, F., editors: Third $tnposium on Terrestrîal
Ëcosysterns. ATTEMPO' Verlag, Tübingen, 79-24.
198à. A new and unusual Late Cretaceous mammal liom Patagonia. Journalof
Vertebrate P aleorûologst 6:264-270.
Bonaparte,J. F.; €o Powell,J' E.
igBO. A continental as'semblage oftetrapods from the upper cretaceous beds of El
Brete, northwestern Argentina (Sauropoda-Coelurosauria-Aves). Sociétc Géo-
logq* de France, Mérnoire Hors Se.ri.e, Paris n-s- 139:19-28.
Bonaparte,J: F,; €P Soria, M. F.
iSaS. nl pri*er mamifero del Cretâcico Argentino. Circukr Infortnaüva, Aso-
ciariôn Pabontol.ôgica Argentina 17t5.
1985. Nota sobre el primer mamifero del Cretâcico Argentino, Campaniano-
Maastrichtiano (Condylarthra). Ameghiniana 27:777 -783.
Case,J. A.; €p Woodburne' M. O.
1986. South American marsupials: a successfiJ crossing of the cretaceous-
Tertiaryboundary. Palnios l:473-476.
Cherroni Mendieta' C.
7977. Ëlsisterna Cretâcicqen la Parte Boliviana de La Cuenca Crètâcica Andina'
Revista TécrûcaYPFB, La Paz 6:5-46.
Chow M.; 6e Wang B.
1979. Rehüànship between the pantodonts and tillodonts and classiflcation ofthe
52 MARSHALL & DE MUIZON
order Pantodonta. Veriebrata PalAsiatica 17:37 -48,
Cifelli, R. L.
1983. The origin and afflrnites ofthe South American Condylarttrra and early Ter-
tiary Litopterna (Mammalia). Arnerban Museum Novitates 2772t'7-49.
Clemens, w. Â.
1966. Fossil mammals of the §pe Lance Formation, Wyoming: pt II. Marsupia-
lia. lJniversîÿ of Californio PublÎcatians in Geolnÿal Sci-ences 62t7-122.
1979. Marsupialia. Lillegpaven, J.A., Kielan-Jaworowska, 2., €o Clemens,
1À/. A. Jr., eütors: Mesozoic Mammals: th.e First T\uo-thirds of Mammalian His-
tory, Universiÿ of California Press, Berkeley, 792-220.
Crochet,J. -Y
tg8b. rcs marsupiaux du fertiaire dEurope. Ëdiüon ..Fondation Singer-
P oligpac, P aris 7:7 - 279.
de Muizon, C.; Gayet, M.; Lavenu, A.; et al.
1983. Late Cretaceous vertebrates including mammals from
central Boliüa. Geobias 76t747-753.
de Muizon, C.; Gayet, M.; Lavenu, A.; et al.-
Tiupampa, souttr-
1984. Obsen ation to the note by Ch. de Muizon, M. Gayet, A, Lavenu, L. G' Mar-
shall. B. Sigé, and C. VilIarroel: "Late Cr.etaceous vertebrates including mammals
from Tiupampa, south-central Bolivia. " Géobîos 77 :257 -252.
de Muizon, C.; €r Marshall, L. G.
1985. Sur 1a piste des premiers mammilères d'Amérique du Sud. La Recherche
76:872-875.
79BZa. Le plus ancien Pantodonte (Mammalia) du Crétacé supérieur de Bolivie'
Comptes rendu-s hebdomadaires des Séances de I'Aca.démie des Sciences, Paris
3O1:2O5-2O8.
79A7b. Le plus ancien Condvlarthre (Mammalia) sud-américain (Crétacé supé-
rieur. Bolir-ie_). Compte rendtts hebdomadaîres des Séances de t Académie des Sci-
ences, PcLris 304:777-774.
L987 c. Deux nouveaux Condt,larthes ( Nlammalia) du Maestrichtien de Tiupampa
(Bolivie). Comptes rendtts hebdornadaires des Séonces de lAcadémie des Sci-
ences, Paris 304:947 -950.
de Muizon, C.; Marshall, L. G.; êp Sigé' B.
1984. The mammal fauna from the EI Molina Formation (Late Cretaceous-
Maestrichtian) at Tiupampa, south-central Bolir'{a. Bulletin du Mttséum natio-
nol d'Histoire naturelb, Paris, sérb 6, section C, 4:327-357.
Estes. R.; êe Bâez. A. M.
1985. Helpetofàunas of North and South America during the Late Cretaceous and
Cenozoic: e'r.idence for interchange? Stelili. F. G. €e \Vebb. S. D.' editors: The
Great Amerban Bintic Interchange. Plenum Press. Nelt, York, 139-196.
Fox. R. C.
1981. Mammals from the Upper Cretaceous Oldman Formation, Alberta. V. Eo-
delphis Matthew, and the evolution of the Stagodontidae (Marsupialia). Cantr-
diant Jottrnal of ùtrth Sciences 18:350-365.
Gingerich. P. D.
1985. South ,tunerican mammals in the Paleocene of North America. Stehli'
F. G. €n \,Vebb. S. D.. editors: T]'Le Great Amerban Biotic Interclrunge. Plenum
Press. New York, 123-137.
Goin, F.J.; Carlini, A. A.; €p Pascual, R.
1986. Un probable marsupial del Cretâcico Tardio del Norte de Patagonia' Argen-
tina. IV Congleso Argentino de Paleontologîa y BinestratigTaJia, Mendoza 2:
*J-+1.
Grarnbast, L.; Martinez. M.l Mattauer, M.; et al.
1967, Perutherium altiplanelzse, nov. gen.) no\r. sp., premier mammilère méso-
zoique d'Amérique du Sud. Comptes rendtts des Séances de I'Académie des Scî-
enc e s, P aris 264:707 -7 70.
Johnston. P. A.; & Fox. R. C.
1984, Paleocene and Late Cretaceous mamtnals from Saskatchewan, Canada.
P alaeonto{r'aphba, A,bteilung A Stuttgart 1,86:763 -222.
Kerourio, P.: 6. Sigé, B.
1984. L'apport des coquilles d'oeufs de dinosaures de Laguna Uma1,o â I'âge de la
Formation \rilquechico (Pérou) et à Ia compréhension de Perutheriurn altiplan-
ense. Nausletter of Sn'atigraplry 13:733-712.
Kielan-Jaworowska, Z.; Bor'r,n. T. NI.; €r Lillegraven,J. A.
1979. Eutheria. Lillegpaven' J. Â., Kielan-Jaworowska, 2., €r Clernens, W A.
Jr., editors: Mesozoic Mamntals: the Fîrst Two-thirds of Mammalian History.
Uni'r'ersity of California Press, Berkelev, 227-258.
LAND MAMMALS IN SOUTH AMERICA 53
Lucas, S. G.
1982. The phylogeny and composition ofthe order Pantodonta (Mammalia, Euth-
eia), Third North Amerban Palcontobÿal Conventian, Proceedings 2:337-342.
Marshall, L. G,
1980. Systematics ofthe South American marsupial family Caenolestidae. FîÊlà-
iana: Geolng't n.s, 5:1-145.
1981. ReviewoftheHathliac;.minae,anextinctzubfamilyofSouthAmerican"dog-
like" marsupials. Fi.el.diana: Geobg, n.s. 7:7-72O,
1982. Systematics of the South American marsupial family Microtriotheriidae.
Ficl.diana: Geolngy' n,s. 10:1-75.
1985. Geochronologzandland-mammalbiochronolog;zofthe transamericanfau-
nal interchange. Stehli, F. G, €p Webb, S, D., editors: The Great Arneriran Bintir
Interchange. Plenum Press, New York, 49-85.
1987. Systematics ofltaboraian (middle PaleoceneJ age oPossum-like marsupials
from ttre limestone quarry at SâoJosé de Itaborai, Braztl'. Archer, M., editor: Pos-
sums and Opo""uoÀ. Royal Zooto$cal Socieÿr New South Wales, Sydney: 91-160.
Marshall, L. G.; €e de Muizon, C.
7984, Anew didelphid marsupial (Itaboraidelphys camposi nov. gen., nov. sP.)
from the middle Paleocene (Itaboraian) ofSâoJosé de Itaborai (Brazil). Comptes
rendus hebdomadaires dcs Séances dE fAcaümîE d.es Scientes, Paris 299:7297-
1300.
Marshall, L. G,; de Muizon, C.; Gayet, M.; et al.
1985. The "Rosetta stone" for mammalian evolution in South America. Natianal
Gimgr aphic Re s e ar ch 7 :27 4 - 2BB.
Marshall, L. G.; de Muizon, C,; €o Sigé, B.
1983a. Late Cretaceous mammals (Marsupialia) from Bolivia. Gbbias 76:739-745.
1983b. PerutLterium alüplantense, un notogplé du Crétacé Supérieur du Pérou'
. Palocovertebrata73:745-755.
Martinez, C.
198O. Strucfure et évolution de la chaîne hergT rienne et de la chaîne andine dans le
nord de la Cordilère des Andes de. Bolivie' Travatt et Doanmzrûs dc
L' O.ks.T.O.M., Pari§ l9:7 -352.
*"*rffi.T.r1;r, ofprimith,e notoungulates and condylarths into the Miocene ofcelom-
bia. AmcricanJournal of kience 254:,736-743.
1981. Early history and biogeograp§ of South America's extinct land mammals'
Ciochon, R L, €p Chiarclli, A. B. , editors: Euo ltttionary Bialag oftLæ New WorA Mon-
keys atul ContinentalDrifi. Plenum Press, New York) 43-77,
Mones,,A..
1987. Gondwanatheria, un nuevo orden de mamiferos sudamericanos (Mammaliai
Edentata: ?Xenarthra). comunicacîaræsPabontolôÿas dcltv'nseo dctlistorinNat-
ural de Montevideo 78:237-2,40.
Mourier, T.;Jaillard, E.; Laubacher, G.; et al.
1986. Découverte de restes dinosauriens et mammalien dâge cÉtacé supérieur à la
base des couches rouges du synclinal de BagUa (Andes nord-1Éruviennes): aspects
stuatigraphiques, sédimentologiques et paléogéogfaphiques concernant lia Égression
fini-crétacée. Bulleün de lo Société Geobgiqtæ du France (B) 2:777-775.
Pascua], R.
1983. Novedosos marsupiales paleôgenos de la Formaciôn Pozuelos (Grupo Pastos
Grandes) de Ia Puna, Salta, Argentina. Amcghinîun2O:2.65-2.8O.
Paula Couto, C, de
7952a, Fossil mammals from the beginning of the cenozoic of Brazil, Marsupialia:
Didelphidae. Amcrban Museum Novîtates 7567'-1-2Â.
aS6[b. Fossil mammals liom the beginning of the Cenozoic of B,raztl. Marsupialia:
Poÿdotopidae and Borhyaenidae. Amerban Musewn Novitates 7559:7-27.
t1ÉZc. Èossil mammals Ilom the treginning of the Cenozoic in Brazil. Condylaritra'
Litoptema, xenungglatal and Astrapotheria. Bulbtîn of the Amcrican Museum of
Natur al History 99 :35 5 - 39 4
1962. Didelfideos frsiles del Paleoceno de BraztL Revista del Museo Argmtirn de
Cierrcias Nahrales "kruwrdirn Riv aàavio," Ciercias 7-nolôgiîas 772:735-7ffi .
197O. News on the fossil marsupials ftom the Riochican of Brazil. AcademinBrasi-
bîra de Ciêncius, Anais, Ria deJaræiro 42:79-34'
1978. Ungulados fosseis do Riochiauense de Itaborai, \f, Brazil, II, condylârt]rra e
Litopterna. Acadsnia Brasibira de Ciênci"as, Anab, Ria dc Janciro 5O:.
209-278.
Rage,J. -C.
a97A. Une connexion continentale ente Amérique du Nord etAmérique du Sud au
54 MARSIIALL €P DE MUIZON
Crétacé supérieur? L'example des vertébrés continentaux. Compte rendu Som-
maire des Séances Société Géologiqlue de France, Paris 6:287-285.
1981. Les continents péri-atlantiques au Crétacé supérieur: mig3ations des faunes
continentales et prcblèmes paléogéographiques. Cretaceous Research 2:65-M.
19BGa. Le plus ancien Amphibien apode (Gl,rnnophiona) fossile. Remarques sur Ia
Épartition et I'histoire paléobiogéographique des Gl,rnnophiones. Comptes ratdtts
læbdomadaires des Séarces d-e tAcaümie des kiences, Paris 302:7033-7036.
19BGb. South American/North American terresfial interchanges in the latest Crtta-
ceous: short comments on Brett-surtran and PauI (1985), with additional data.
Journal of Vertebrate Pabontologit 6:382-383.
Reig, O. A.
1981. Teoria deI origen r- desarrollo de la fauna de mamiferos de América del Sur.
MonogycLphirLe )iaturae, Publbafus por el Mttseo Munbipal de Ciencias Natu-
rabs " Inren=o k aglicL," NIar del Plata 7:7 -762.
Reig, O. À.: Kirsch.J.A.1\''.; f"c ÿarshall' L. G.
1987. Stsremaic rc)aüonslips of the bing and Neocenozojc opossum-like marsr4d-
aüs [suborder Didelphimorphia) with comments on the classification of these and of
dre Crelaceous and Paleogene New World and European Metatherians.
,1lrher. NI., editor: Possums and Opossums. Roval Zoologlca\ Socie§', Nerv South
\l-a-les. Svdneyr 1-89.
killato-Yané, G.J.; €e Pascual, R.
1984. Un peculiar Paratheria, Edentata (Nlamrnalia) de1 Paleoceno de Patagonia.
PrimerasJorttadas Argentinas de Pabontologia de Yertebrados, Restitrcnes, Comis-
iôn de Investigacianes Ci.entificas, La Plata, 75.
1985. Un peculiar Xenanh'a del Paleoceno medio de Patagonia (Argentina). Su im-
portancia en Ia sistemâdca de los Parather-ia. ,$neghinianta 27:773-776.
SemlÉré. T.
1986, Conu-ibuciôn a la estatigrafia de1 Nlesozoico Bolirialo en eI donrinio andino.
htstitut Francais de Redterc]æ kienti.fique, Coru,erio LI,\{$l OÀSTOIÿ1, La Paz, In-
fonræ 7:7-31.
SempÉre. T.; Oiler.J.: Cheroni. C.: et al.
1987. Un ejemplo de cuenca carbonatado en un contexto distensivo de retroarco: pa-
leogeodinârnica de Cetâctico Terminal en la republica de Bolivia (Forrnaci6n El Moli-
na v Equi'r..alentes). II Simpsio, ProgTama Intenm.cional de Corelaciôn Geolôÿa,
no. 212. Tucumân, Argentina, 1-40.
Sige. B.
1968. Dents de MicromammiÊres et fragments de coquilles d'oeufs de Dinosaur-
iens dans la faune de Yertébrés du Crétacé supérieur de Laguna LImavo (Andes
péruriennes). Cornptes rendus hebdorna-daires des Séances de I'Acadénûe des
Sciences, P aris 267 :7495 - 1498.
7971. Les Didelphoidea de Laguna Umal'o (Formation Vilquechico, CÉtacé sulÉr-
ieur, Pérou), etle peuplement marsupial d'Amérique du Sud. Comptes rendus heb-
damadaires des Séarrces de fAcadÉtnie des kiences, Paris 273:2479-2487.
7972. La faunule de mammiêres du Cr.etacé supérieur de Lagrrna Umal'o (Andes
Pérur{erures). Bulbtin du Mttséum natianal d'Histoire naturelb, Paris, no. 99, Sci-
ences de la Terre 19:375-.109.
Simom. E. L.
1960. The Paleocene Partodonta. Transactions ofthe Amerban Phibsophbal So-
cieÿ, Phtladelphza n.s. 50(6) :1 -81.
Simpson, G. G.
a947. A new Eocene marsupial fiom Brazil. Anerican Museum Novitates 7357 :7 -
7.
1tXB. The beginning of the age of mammals in South America, pt, L Bulbtin of the
Amerban Museunt of Natural History 97:7-232.
Sloarr. R r,.; Rigby,J. K Jr.; \r:rn \/alen. L. M.; et al.
1986. Gradual dinosaur s::tinction and simultaneous ungtrlate radiation in the Hell
Creek Formation. Sciente 232:629-633.
Tedford, R- H.
1974. Ivlarstrpialsandthenewpaleogeographl,. Ross,C. A.,eütorl.Paleogeograph-
b Provinces and Provîncioliÿ, special publication 21. Society of Economic Paleonto-
logists and Mineralogists, Tulsa, 109-126.
Van Valen, L.
1978. The beginning of the age of mammals. Evolutionary Theory 4:45-80,
Accepted SJune 1987.
LAND MAMMALS IN SOUTH AMERICA 55
