B322 PDF

PHYLOGENETIC RELATIONSHIPS AND
CLASSIFICATION OF DIDELPHID MARSUPIALS,

AN EXTANT RADIATION OF NEW WORLD
METATHERIAN MAMMALS
ROBERT S. VOSS
Department of Mammalogy
American Museum of Natural History
SHARON A. JANSA
Department of Ecology, Evolution, and Behavior; and
J.F. Bell Museum of Natural History
University of Minnesota
BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY

Number 322, 177 pp., 54 figures, 18 tables
Issued June 30, 2009
Copyright E American Museum of Natural History 2009 ISSN 0003-0090

Representative opossums (top to bottom, not to the same scale): Caluromys derbianus, Monodelphis
brevicaudata, Marmosa robinsoni, Chironectes minimus, Metachirus nudicaudatus.
CONTENTS
Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 5
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 5
Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 7
Comparative Morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Size and External Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Cranium and Mandible . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Dentition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Gene Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Breast Cancer Activating 1 Gene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Dentin Matrix Protein 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
Interphotoreceptor Retinoid Binding Protein . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Recombination Activating 1 Gene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
von Willebrand Factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Summary, Sequence Characteristics, and Model Fitting . . . . . . . . . . . . . . . . . . . . . . 64
Phylogenetic Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Analyses of Nonmolecular Data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Single-gene Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Analyses of Concatenated Genes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Analyses of Combined Datasets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
A Revised Phylogenetic System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Family Didelphidae Gray, 1821 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Subfamily Glironiinae, new . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
Glironia Thomas, 1912 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Subfamily Caluromyinae Reig et al., 1987 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Caluromys J.A. Allen, 1900 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Caluromysiops Sanborn, 1951 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Subfamily Hyladelphinae, new . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Hyladelphys Voss et al., 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Subfamily Didelphinae Gray, 1821 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Tribe Marmosini Hershkovitz, 1992 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Marmosa Gray, 1821 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Monodelphis Burnett, 1830 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
Tlacuatzin Voss and Jansa, 2003 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Tribe Metachirini Hershkovitz, 1992 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Metachirus Burmeister, 1854 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Tribe Didelphini Gray, 1821. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Chironectes Illiger, 1811 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Didelphis Linnaeus, 1758 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Lutreolina Thomas, 1910 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
Philander Brisson, 1762 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Tribe Thylamyini Hershkovitz, 1992 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Chacodelphys Voss et al., 2004 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Cryptonanus Voss et al., 2005 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
Gracilinanus Gardner and Creighton, 1989 . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Lestodelphys Tate, 1934 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Marmosops Matschie, 1916 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
Thylamys Gray, 1843 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
3
4 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 322
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Appendix 1: Morphological Specimens Examined . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Appendix 2: Specimens Sequenced for Nuclear Genes . . . . . . . . . . . . . . . . . . . . . . . . 160
Appendix 3: Nonmolecular Character Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Appendix 4: Nonmolecular Data Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Appendix 5: Nonmolecular Apomorphy List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Appendix 6: Key to the Didelphid Genera . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
ABSTRACT
This report summarizes a decade of morphological and molecular research on the
phylogenetic relationships of didelphid marsupials (opossums), a substantially intact radiation
of New World metatherian mammals. We review the comparative morphology of Recent
opossums, emphasizing those anatomical systems from which taxonomically useful information
is available for the majority of living genera and species, namely the integument, cranium, and
dentition. Morphological similarities and differences among didelphids and other plesiomorphic
marsupials (caenolestids, microbiotheriids, dasyurids, and peramelids) are also described. These
observations, representing evolved differences in diverse functional-morphological systems,
together with karyotypic information gleaned from the literature, provide the basis for coding
129 phylogenetic characters that we scored for 44 ingroup and seven outgroup taxa.
Published information about the size, internal organization, chromosomal location, and
physiological properties of five nuclear genes (BRCA1, DMP1, IRBP, RAG1, vWF) sequenced
for this study suggest that these loci are unlinked, exist as single copies, are active in different
tissues, and encode protein products with widely divergent functions. All of the sequenced
fragments are long (.900 bp), free of ingroup alignment ambiguities, and translate to open
reading frame. Nucleotide data from a total of 7320 aligned sites were obtained from 43 ingroup
and seven outgroup taxa.
Separate parsimony, likelihood, and Bayesian analyses of these six data partitions
(morphology + karyotypes, five genes) resulted in highly congruent estimates of didelphid
phylogeny with few examples of conflict among strongly supported nodes. Analyses of
concatenated sequences and combined (nonmolecular + sequence) datasets effectively summarize
all of the common signal recovered from separate analyses: a completely resolved ingroup
phylogeny with high support statistics at most nodes. Remaining problems (not conclusively
resolved in this study) include the position of the ingroup root and the relationships of three
genera (Chacodelphys, Cryptonanus, Tlacuatzin) within their respective suprageneric clades.
The history of didelphid classification is reviewed, and all previous systems are found to
contain nonmonophyletic groups. A revised phylogenetic classification consistent with our
analytic results includes the following higher taxa: Glironiinae (for Glironia), Caluromyinae
(Caluromys and Caluromysiops), Hyladelphinae (Hyladelphys), Didelphinae (Marmosini,
Metachirini, Didelphini, and Thylamyini), Marmosini (Marmosa, Monodelphis, and Tlacuatzin),
Metachirini (Metachirus), Didelphini (Chironectes, Didelphis, Lutreolina, and Philander), and
Thylamyini (Chacodelphys, Cryptonanus, Gracilinanus, Lestodelphys, Marmosops, and Thyla-
mys). The probable relationships of several Neogene fossil genera are also discussed. To
facilitate identifications, all Recent genera are redescribed, representative crania are illustrated,
and a key is provided.
INTRODUCTION Asia and western North America (Rougier et

al., 1998; Luo et al., 2003; Kielan-Jawor-
They affirme that there are trees of suche owska et al., 2004), and other early metathe-
byggenes, that xvi men ioyninge handes togyther rians are present in the Tertiary record of
and standinge in coompasse, can scarsely em- Europe and Africa (Crochet, 1980; Kurz,
brase sum of them. Emonge these trees is fownde 2007; Hooker et al., 2008). The metatherian
that monstrous beaste with a snowte lyke a foxe, crown group Marsupialia, however, probably
a tayle lyke a marmasette, eares lyke a batte, evolved in South America and subsequently
handes lyke a man, and feete lyke an ape, bearing
dispersed across Antarctica to Australia and
her whelpes abowte with her in an owtwarde
bellye much lyke vnto a greate bagge or purse. New Guinea (Muizon et al., 1997; Springer et
Richard Eden (1555), describing the first al., 1997a; Rougier et al., 1998). Living
European encounter with a marsupial, by Vicen- marsupials comprise seven major groups,
te Yanez Pinzon and his men on the coast of currently ranked as orders in the Linnaean
Brazil in the year 1500 hierarchy, each of which is endemic to either
the New World or the Old World (table 1).
The oldest known metatherian mammals Although the extant New World marsupial
occur in Cretaceous sediments of eastern fauna consists of three orders, two of these
5
TABLE 1 distributed in the Tertiary than they are

Higher Classification and Geographic Distribution of today (Marshall, 1980, 1982b; Bown and
Recent Marsupialsa Fleagle, 1993; Goin, 1997; Goin and Cande-
la, 2004), these orders are appropriately
Genera Species Distributionb
regarded as relictual elements in modern
DASYUROMORPHIA faunas.
Dasyuridae 20 69 OW Instead, the only substantially intact radi-
Myrmecobiidae 1 1 OW ation of New World marsupials is represent-
Thylacinidae 1 1 OW ed by the family Didelphidae, commonly
DIDELPHIMORPHIA known as opossums, of which 91 Recent
Didelphidae 18c 91c NW
species in 18 genera are currently recognized
(Gardner, 2008). Although didelphids were
DIPROTODONTIA
the first metatherians to be encountered by
Acrobatidae 2 2 OW European explorers (Eden, 1555), the first to
Burramyidae 2 5 OW
be described scientifically (Tyson, 1698), and
Hypsiprymnodontidae 1 1 OW
the first to be classified by taxonomists
Macropodidae 11 65 OW
Petauridea 3 11 OW
(Linnaeus, 1758), zoological interest in opos-
Phalangeridae 6 27 OW sums was soon eclipsed by the discovery of
Phascolarctidae 1 1 OW Australasian marsupials. Nevertheless, steady
Potoroidae 4 10 OW advances in didelphid taxonomy (mostly
Pseudocheiridae 6 17 OW involving the description of new species)
Tarsipedidae 1 1 OW were made throughout the 19th and 20th
Vombatidae 2 3 OW centuries. Although this progress was period-
MICROBIOTHERIA ically summarized by authors (e.g., Thomas,
Microbiotheriidae 1 1 NW 1888; Winge, 1893; Cabrera, 1919, 1958),
there is no modern monographic treatment of
NOTORYCTEMORPHIA
the family, and the current suprageneric
Notoryctidae 1 2 OW
classification (McKenna and Bell, 1997;
PERAMELEMORPHIA Gardner, 2005, 2008) does not reflect current
Chaeropodidae 1 1 OW knowledge about opossum evolutionary rela-
Peramelidae 6 18 OW tionships.
Thylacomyidae 1 2 OW The history of didelphid phylogenetic
PAUCITUBERCULATA research was reviewed by Jansa and Voss
Caenolestidae 3 6 NW (2000), who analyzed DNA sequence data
a
from the nuclear Interphotoreceptor Reti-
Numbers of genera and species after Wilson and
noid Binding Protein (IRBP) gene for 21
Reeder (2005) except as noted.
b
NW 5 New World; OW 5 Old World (Sahul).
species. Subsequent reports were focused on
c
After Gardner (2008). analyses of morphological and karyotypic
character data in combination with IRBP
sequences for larger sets of species (Voss and
are represented by a mere handful of living Jansa, 2003; Voss et al., 2005), on phyloge-
forms: Microbiotheria by a single genus and netic problems associated with individual
species (Dromiciops gliroides), and Paucitu- taxa (Voss et al., 2004a; Jansa and Voss,
berculata by six species in three genera 2005), or on the evolutionary dynamics of
(Caenolestes, Lestoros, Rhyncholestes). Both newly sequenced nuclear loci (Jansa et al.,
groups inhabit wet-temperate Patagonian 2006; Gruber et al., 2007). However, no
forests, but paucituberculates also occur in comprehensive phylogenetic synthesis has yet
climatically similar montane habitats of the been attempted, and much new morpholog-
tropical Andes (Bublitz, 1987; Patterson and ical and molecular character data remain
Gallardo, 1987; Albuja and Patterson, 1996; unanalyzed.
Hershkovitz, 1999; Lunde and Pacheco, This monograph summarizes our collabo-
2003). Because microbiotherians and pauci- rative research to date on didelphid phyloge-
tuberculates were more diverse and widely netic relationships and concludes the first
2009 VOSS AND JANSA: DIDELPHID MARSUPIALS 7
phase of our ongoing evolutionary study of taxa have never been considered close to
the family. Although we take this opportu- didelphids, and because their highly derived
nity to review information that has previous- morphologies raise many unresolved issues of
ly been published elsewhere, many new homology that are beyond the scope of this
results are reported herein. Among these, study.
we describe morphological comparisons of COMPARATIVE MORPHOLOGY: Our mor-
didelphids with other plesiomorphic (poly- phological comparisons are based on skins,
protodont) marsupial clades; we describe and skulls, and fluid-preserved specimens (appen-
illustrate integumental and craniodental dix 1), which we examined for discretely
traits not previously discussed in the litera- varying traits that could be scored for most
ture; we report new sequence data from the of the terminal taxa in our analysis. Addi-
Breast Cancer Activating 1 gene (BRCA1) tionally, we summarize information about
and the von Willebrand Factor gene (vWF); other phenotypic characters that, although
we analyze didelphid phylogenetic relation- not suitable for phylogenetic analysis due to
ships based on a new set of morphological continuous variation (e.g., body size), pro-
and karyotypic characters together with vide taxonomically useful descriptors. Infor-
sequence data from five protein-coding nu- mation about morphological character vari-
clear loci; we propose a new suprageneric ation among nondidelphid marsupials is
classification consistent with our analytic largely based on our scoring of outgroup
results; and we provide morphological de- terminal taxa, but we also include remarks on
scriptions and cranial illustrations of all the morphology of other Old World marsu-
currently recognized genera to facilitate pials where these seem relevant or interesting.
taxonomic identifications. To provide a maximally user-friendly
reference, our morphological observations
Materials and Methods are summarized in two different formats.
First, we describe taxonomic comparisons
PHYLOGENETIC ASSUMPTIONS AND TAXON using normal (nontelegraphic) prose in the
SAMPLING: Our phylogenetic analyses are body of the text. These accounts, organized
based on the assumption of didelphid mono- organ-system-by-organ system, are accompa-
phyly, which is strongly supported by a nied as necessary by illustrations of represen-
variety of nuclear-gene sequence datasets tative specimens. Second, we summarize
(e.g., Jansa and Voss, 2000; Amrine-Madsen morphological descriptors taxon-by-taxon
et al., 2003; Meredith et al., 2008). Our using telegraphic prose in the generic ac-
ingroup terminals (table 2) consist of 44 counts of our classification. Although largely
species representing every currently recog- redundant with our data matrix (appendix 4),
nized didelphid genus and subgenus, includ- these generic descriptions are easier to use for
ing all of the species previously analyzed by taxonomic identifications, and they provide
Jansa and Voss (2000, 2005), Voss and Jansa an opportunity to summarize relevant obser-
(2003), Voss et al. (2004a, 2005), Jansa et al. vations that were not encoded as characters.
(2006), and Gruber et al. (2007). Altogether, Formal character descriptions (in appen-
this is by far the most taxon-dense and dix 3) are accompanied by technical details
character-rich didelphid dataset ever ana- about coding procedures and ordering crite-
lyzed. ria. Our criteria for character choice were
The outgroups for this study include described at length by Voss and Jansa (2003),
representatives of every extant ordinal-level so they are not repeated here. The only
clade that has ever been recovered within one noteworthy methodological change in this
or two nodes of Didelphidae in previous study concerns our coding of polymor-
analyses of marsupial relationships, including phisms. As in our previous study, we ignored
caenolestids (Paucituberculata), dasyurids rare variants (on the assumption that all
(Dasyuromorphia), Dromiciops (Microbio- characters are polymorphic given sufficiently
theria), and peramelids (Peramelemorphia). large samples) and coded only those poly-
We did not include Notoryctes (Notorycte- morphisms represented by nearly equal
morphia) or diprotodontians because these frequencies of alternative states (e.g., when
TABLE 2
Marsupial Terminal Taxa Scored for Phylogenetic Analysisa
Percent complete character datab

Nonmolecular Molecular Combined
INGROUP:
Didelphidae
Caluromys lanatus 99.2 100 100
Caluromys philander 99.2 100 100
Caluromysiops irrupta 93.7 16.5 17.8
Chacodelphys formosa 86.0 0 1.5
Chironectes minimus 99.2 99.8 99.8
Cryptonanus chacoensis 98.4 95.2 95.3
Cryptonanus unduaviensis 93.0 99.8 99.7
Didelphis albiventris 100 99.1 99.1
Didelphis marsupialis 100 99.1 99.1
Didelphis virginiana 100 100 100
Glironia venusta 93.0 100 99.9
Gracilinanus aceramarcae 93.0 99.8 99.7
Gracilinanus agilis 95.3 99.7 99.6
Gracilinanus emiliae 95.3 99.8 99.7
Gracilinanus microtarsus 95.3 95.1 95.1
Hyladelphys kalinowskii 95.3 100 99.9
Lestodelphys halli 97.6 30.4 31.6
Lutreolina crassicaudata 99.2 100 100
Marmosa (Marmosa) lepida 96.1 100 99.9
Marmosa (Marmosa) mexicana 99.2 98.8 98.8
Marmosa (Marmosa) murina 99.2 98.6 98.6
Marmosa (Marmosa) robinsoni 99.2 98.8 98.8
Marmosa (Marmosa) rubra 92.2 99.7 99.6
Marmosa (Micoureus) demeraraec 99.2 98.5 98.5
Marmosa (Micoureus) paraguayanac 96.1 95.7 95.7
Marmosa (Micoureus) reginac 99.2 100 100
Marmosops impavidus 98.4 98.0 98.0
Marmosops incanus 96.8 99.0 99.0
Marmosops noctivagus 98.4 100 100
Marmosops parvidens 95.3 98.8 98.7
Marmosops pinheiroi 95.3 99.7 99.6
Metachirus nudicaudatus 98.4 98.9 98.9
Monodelphis brevicaudata 97.7 100 100
Monodelphis emiliae 96.1 100 99.9
Monodelphis peruvianad 92.3 100 99.9
Monodelphis theresa 89.9 100 99.8
Philander frenatus 97.7 100 100
Philander mcilhennyi 100 99.2 99.2
Philander opossum 99.2 99.1 99.1
Thylamys macrurus 96.1 99.8 99.7
Thylamys pallidior 96.1 98.3 98.3
Thylamys pusillus 98.4 95.7 95.7
Thylamys venustus 96.1 99.5 99.4
Tlacuatzin canescens 99.2 97.5 97.5
OUTGROUPS:
Caenolestidae
Caenolestes fuliginosus 86.9 63.3 63.7
Rhyncholestes raphanurus 85.3 62.5 62.9
TABLE 2
(Continued )
Percent complete character datab

Nonmolecular Molecular Combined
Dasyuridae
Murexia longicaudata 95.3 57.9 58.5
Sminthopsis crassicaudata 96.1 42.1 43.0
Microbiotheriidae
Dromiciops gliroides 99.2 63.3 63.9
Peramelidae
Echymipera kalubu 95.3 63.2 63.8
Perameles gunnii 93.0 63.3 63.8
a
Nomenclature follows Wilson and Reeder (2005) and Gardner (2008) except as noted.
b
Number of filled (total minus empty) cells in the corresponding row of each data matrix, divided by the total number
of cells (N5 129, 7320, and 7449 for nonmolecular, molecular, and combined data, respectively) 3 100. For the
nonmolecular data, empty cells include those scored as missing (?) and those scored as inapplicable (-). For the
molecular data, only unsequenced base pairs were counted as missing (?); gaps (-) were counted as filled data cells.
c
Micoureus was formerly treated as a full genus.
d
Formerly identified as Monodelphis adusta (e.g., by Jansa and Voss, 2000; Voss and Jansa, 2003), our material is
referable to M. peruviana, a distinct species recently resurrected from synonymy by Solari (2007).
presence and absence were both commonly We amplified 2.1 kb of BRCA1 exon 11
observed for a given taxon). However, from genomic DNA in two fragments using
whereas we formerly (Voss and Jansa, 2003) the primers listed in table 3. The first
treated such polymorphism as a separate state fragment, comprising the upstream 1.2 kb
intermediate to the two fixed conditions, with of the exon, was amplified using primers F1
transformations to and from the polymorphic paired with R1218 or F47 paired with R1343.
state weighted as half-steps (the scaled The second, downstream fragment (ca. 1 kb)
option discussed by Wiens, 2000), we here was amplified with primers F1163 or F1163a
conform with the prevailing custom of coding paired with R2078 or R2151. These amplifi-
polymorphisms as taxonomic ambiguities cation products were then used in a second
(e.g., as 0/1 for a binary character). This round of PCR to generate smaller pieces of
coding change has minimal impact on our suitable size for sequencing. For the up-
results (only weakly supported relationships stream fragment, either F1 or F47 was paired
are affected, even in separate analyses of with R743, and F593 was paired with either
morphology) but it facilitates likelihood R1218 or R1343. For the downstream
modeling of morphological character evolu- fragment, F1163 or F1163a was paired with
tion (Lewis, 2001) in Bayesian analyses of our R1780, and F1697 was paired with either
combined datasets (see below). R2078 or R2151. A single fragment of ca.
MOLECULAR SEQUENCING AND HOMOLO- 1 kb was amplified from vWF exon 28 using
GY COMPARISONS: The laboratory procedures either F104 or F120 paired with R1141. This
we used for DNA amplification and sequenc- product was then used in a second round of
ing (including the names and locations of PCR in which F1 or F47 was paired with
primers used in PCR reactions) have already either R665 or R742, and F557 was paired
been described for three of the protein-coding with R1141.
nuclear loci analyzed in this report: IRBP Initial amplifications using genomic DNA
(Jansa and Voss, 2000), Dentin Matrix as template were performed as 20 ml reac-
Protein 1 (DMP1; Jansa et al., 2006), and tions using Ampli-Taq Gold polymerase
Recombination Activating 1 Gene (RAG1; (Perkin-Elmer Corp.) and recommended
Gruber et al., 2007). In addition, we concentrations of primers, nucleotides, buff-
sequenced two other protein-coding nuclear er, and MgCl2. These genomic amplifications
genes as described below. were performed using a four-stage touch-
TABLE 3
Primers Used to Amplify and Sequence BRCA1 and vWF
Primer name Sequence
BRCA1- F1 59 TCATTACTGCCTGAGATCACCAG
BRCA1- F47 59 TATTGCCTAACACAGACAGCAT
BRCA1- F593 59 CAACAATATTGAAGACAAAATATTAGGAAA
BRCA1- F1163 59 ATGARACWGAACTACWGATCGATAG
BRCA1- F1163a 59 AATGAGACTGAACTACAGATCGAT
BRCA1- F1697 59 TTWGATGRTTGTTCATCYRAAAACAC
BRCA1- R743 59 TTGATGAAATCCTCAGGCTGYAGGT
BRCA1- R1218 59 GAAGYCTTCTGCTGCGTCTGA
BRCA1- R1343 59 CTAACATTTGATCACTATCAGTAG
BRCA1- R1780 59 TAAATAYTGGGTRTCRAGTTCACT
BRCA1- R2078 59 GAAATTTCCTGGTTGTTTCCAGCAA
BRCA1- R2151 59 TCCTTTTGATYAGGAACTTGTGAAATT
VWF- F104 59 GGTGTGATGGAGCGTTTACACATCTC
VWF- F120 59 GACTTGGCYTTYCTSYTGGATGGCTC
VWF- F557 59 CCTGGGCTACCTCTGTGACCTGGT
VWF- R655 59 CTTCTAGCACAAACACCACATCCAGAACCA
VWF- R743 59 CTCACATCCATYCGTTGCATCA
VWF- R1141 59 ATCTCATCSGTRGCRGGATTGC
down protocol as described in Jansa and analyses, all molecular characters were treat-
Voss (2000). Reamplification reactions were ed as unordered and equally weighted, and
performed using Taq DNA polymerase (Pro- all tree searches were heuristic with at least 10
mega Corp.) in 25 ml reactions for 30 PCR replicates of random stepwise taxon addition
cycles. The resulting PCR products were followed by tree bisection-reconnection
sequenced in both directions using amplifica- (TBR) branch swapping. To choose the best
tion primers and dye-terminator chemistry on models of nucleotide substitution for ML and
an AB 3700 automated sequencer. Bayesian analyses, we examined the fit of
We searched the draft Monodelphis domes- various models separately for each of our five
tica genome using the BLAT (modified gene partition (IRBP, vWF, BRCA1, DMP1,
BLAST; Kent, 2002) algorithm to determine and first and second codon positions of
the copy number and chromosomal location RAG1) based on neighbor-joining trees of
of query sequences from M. brevicaudata. As Jukes-Cantor-corrected distances using both
reported on the Ensemble database (www. hierarchical likelihood-ratio tests (hLRTs)
ensemble.org/index.html), the available refer- and the Akaike Information Criterion (AIC)
ence genome (assembly MonDom5) was as implemented in ModelTest 3.7 (Posada
released on October 2006 and has a base and Crandall, 1998). Where the two ap-
coverage of approximately 7.33. proaches disagreed on model choice, we used
ALIGNMENT AND PHYLOGENETIC ANALY- the model selected by the AIC for reasons
SIS:We aligned DNA sequences with refer- outlined by Posada and Buckley (2004). For
ence to translated amino-acid sequences both ML and Bayesian searches, the best-fit
using MacClade 4.08 (http://macclade.org). model was specified, but model parameter
Aligned sequences were then analyzed phy- values were not fixed. For ML analyses, we
logenetically using maximum parsimony conducted three independent runs of genetic-
(MP) as implemented by PAUP* ver. algorithm searches in GARLI, with random
4.0b10 (Swofford, 1998), maximum likeli- starting topologies and automatic termina-
hood (ML) as implemented by GARLI ver. tion after 10,000 generations with no im-
0.95 (Zwickl, 2006), and Bayesian inference provement in log-likelihood scores. For
as implemented by MrBayes ver. 3.1.1 Bayesian analysis of each gene partition, we
(Ronquist and Huelsenbeck, 2003). For MP conducted two independent runs of Metrop-
olis-coupled Markov-chain Monte Carlo taxon-addition replicates with TBR branch

(MCMCMC), each with one cold and three swapping, but saved only 10 trees per
incrementally heated chains. For each run, we replicate. We then used this pool of 10,000
assumed uniform-interval priors for all pa- trees as the starting point for an unbounded
rameters, except base composition, which heuristic search. Parsimony analysis of the
assumed a Dirichlet prior. Runs were allowed combined (molecular + nonmolecular) data-
to proceed for 5 3 106 generations, and trees set did not exhibit this problem; therefore, we
were sampled every 100 generations. We analyzed the combined dataset using un-
evaluated the burn-in for each run, and bounded heuristic searches with 1000 repli-
pooled the post-burn-in trees to calculate cates of random-taxon addition and TBR
estimated parameter distributions and poste- branch swapping. For Bayesian analysis of
rior probabilities for each node. We assessed the morphological dataset alone, we specified
nodal support from MP and ML analyses the Mkv model (Lewis, 2001) with a C-
using nonparametric bootstrapping (Felsen- distributed rate parameter and ascertainment
stein, 1985). Bootstrap values for the parsi- bias corrected for omission of constant char-
mony analyses (MPBS) were calculated in acters (lset coding 5 variable). For Bayesian
PAUP* from 1000 pseudoreplicated datasets, analysis of the combined (molecular + non-
each of which was analyzed heuristically with molecular) dataset, we specified this same
10 random-addition replicates with TBR model for the morphological data and applied
branch swapping. Bootstrap values for the the best-fit model to each of the five gene
likelihood analysis (MLBS) were calculated in partitions. As above, we allowed parameters to
GARLI using genetic-algorithm searches of be estimated independently across all parti-
1000 pseudoreplicated datasets, allowing tions and used the same MCMCMC settings.
model parameters to be estimated for each To assess the impact of the large amount of
pseudoreplicate. missing data from Chacodelphys, we per-
We analyzed the combined-gene dataset formed all analyses that included nonmolec-
using ML as implemented in RAxML-VI- ular characters with and without this taxon.
HPC (ver. 2.2.3; Stamatakis, 2006). We ONLINE DATA ARCHIVES: All of the new
specified the GTRMIX model, which per- molecular sequences produced for this study
forms initial tree inference using a GTRCAT have been deposited in GenBank with
approximation, with final topology evalua- accession numbers FJ159278FJ159314 and
tion performed under a GTRGAMMA FJ 159316FJ159370 (for a complete list of
model. We allowed parameters to be estimat- GenBank accession numbers of all analyzed
ed independently across the five gene parti- sequences, old and new, see table 9). All of
tions. To evaluate nodal support for this our datasets, selected ML and Bayesian
combined-gene, mixed-model analysis, we analyses, and associated trees have been
performed 1000 bootstrap replicates, again deposited on TreeBase (http://www.treebase.
allowing model parameters to be estimated org) with accession numbers S2164, M4107,
independently across the five genes. We also and M4108. Our nonmolecular data matrix
performed a Bayesian analysis of the com- has also been deposited on MorphoBank
bined-gene dataset, using the same MCMCMC (http://morphobank.geongrid.org) with ac-
settings given above. For this analysis, we cession number X600.
specified the best-fit model for each gene,
decoupled estimation of substitution param- COMPARATIVE MORPHOLOGY
eters across the partitions, and allowed each
gene to assume a separate rate. The literature on didelphid comparative
We analyzed the nonmolecular (morpho- morphology is widely scattered, and no
logical + karyotypic) data alone and in adequate review of this topic has yet been
combination with the molecular data using published. Although the following accounts
MP and Bayesian approaches. Due to the are far from comprehensive, they include
large number of suboptimal trees recovered most of the anatomical features that have
from parsimony analysis of the nonmolecular been surveyed widely among extant genera
dataset, we first performed 1000 random- and that provide relevant taxonomic infor-
mation.1 In effect, such information comes comprehensive compilations of morphomet-

from commonly available materials that can ric data will probably document a continuum
be examined without dissection or other of didelphid size distributions.
special preparations. Therefore, microscopic By comparison, Old World marsupials
features (e.g., those of the spermatozoa; include some species that are smaller and
Temple-Smith, 1987) and visceral characters others that are much larger than any
(e.g., of the digestive tract; Santori et al., didelphid. The smallest living Australian
2004) are not reviewed below, nor are species, for example, is said to be Planigale
osteological traits that require X-ray com- ingrami, with an adult weight of only about
puted tomography, serial sectioning, or 4 g, and the largest is Macropus rufus, males
destructive methods for their study. of which are said to weigh as much as 85 kg
(Dawson et al., 1989). The extinct Pleistocene
Size and External Features species Diprotodon optatum, however, may
have weighed almost 2800 kg (Wroe et al.,
Most opossums are externally unremark- 2004).
able mammals with pointed muzzles, large RHINARIUM AND MOUTH: The rhinarium,
rhinaria, well-developed vibrissae, prominent a prominent pad of naked glandular skin
eyes, membranous ears, nonspinous pelage, surrounding the nostrils, is divided by a
subequal limbs, pentadactyl feet, and naked median crease or sulcus (sulcus medianus;
tails. In many of these respects they resemble Ade, 1999) that extends from between the
other plesiomorphic marsupials (e.g., Dromi- nares to the upper lip in all examined
ciops and dasyurids) as well as certain didelphids. The part of the rhinarium that
unspecialized placentals (e.g., solenodontids, borders the upper lip (pars supralabialis) is
rice tenrecs, gymnures, and tree shrews). broad, and its ventral margin is notched by
Closer inspection, however, reveals numerous one or two distinct grooves on each side of
distinctive and phylogenetically informative the median sulcus (fig. 1). Two ventrolateral
details of didelphid external morphology. grooves are present on each side in most
SIZE: Didelphids are small to medium- examined taxa, but only a single groove is
sized mammals. The smallest Recent species present in Chironectes, Didelphis, Lestodel-
is probably Chacodelphys formosa, the young phys, Lutreolina, Metachirus, Monodelphis,
adult holotype of which had a head-and- Philander, and Thylamys pallidior. Dasyurids
body length of 68 mm and probably weighed and Dromiciops have rhinaria that are
about 10 g (Voss et al., 2004a). By contrast, essentially similar in gross morphology to
the largest living opossum, Didelphis virgini- those of didelphids, with a pars supralabialis
ana, can measure almost 500 mm in head- that makes broad contact with the upper lip;
and-body length and weigh more than 3000 g however, only a single ventrolateral groove is
(Hamilton, 1958). Most didelphids, however, present in these taxa (Pocock, 1926: figs. 26
range in head-and-body length from about 28). By contrast, the supralabial part of the
100 to 300 mm and weigh between about 20 rhinarium is reduced to a narrow philtrum in
and 500 g (table 4). Although some authors caenolestids and peramelids, such that the
have recognized large, medium, and groove-bearing ventral rhinarial margin of
small opossums, taxonomic assignments other marsupials is effectively absent.
to discrete size categories are often arbitrary The mouth is large in all didelphids, with a
and sometimes misleading. For example, posterior angle (angulis oris; Brown, 1971)
Reig et al. (1987: character 25) scored that extends posteriorly to a point below the
Metachirus as large and Caluromys as eye; the upper and lower oral margins are
medium in their phylogenetic analysis, smooth or irregularly wrinkled and anatom-
but linear measurements and weights that ically featureless. Most other marsupials have
we compiled suggest that these taxa are anatomically featureless oral margins like
indistinguishable in size. More taxonomically those of didelphids, but caenolestids are
uniquely provided with reciprocating fleshy
1
An important exception is postcranial skeletal morphology, lappets of unknown function on the upper
the topic of an independent study (Flores, 2009). and lower lips (Osgood, 1921: pl. 2; Bublitz,
TABLE 4
External Measurements (mm) and Weights (g) of Exemplar Didelphid Speciesa
Nb Head and Bodyc Taild Weight

e
Caluromys (Caluromys) philander 7 261 (224279) 390 (373410) 330 (220390)
Caluromys (Mallodelphys) lanatusf 8 278 (270296) 422 (400446) 412 (349500)
Caluromysiops irruptag 1 260 310 496
Chironectes minimush 6 289 (276307) 348 (316362) 605 (520700)
Cryptonanus unduaviensisi 8 105 (97121) 122 (112135) 25 (1540)
Didelphis marsupialise 9 419 (405446) 434 (366497) 1351 (10251700)
Glironia venustaj 2 194 (188201) 208 (201215) 130 (129130)
Gracilinanus agilisk 11 98 (86109) 137 (121162) 25 (1834)
Hyladelphys kalinowskiie 3 77 (7678) 111 (107113) 16 (1318)
Lestodelphys hallil 1 132 88 76
Lutreolina crassicaudatam 11 295 (241342) 283 (242336) 530 (300800)
Marmosa (Marmosa) murinan 13 133 (118152) 173 (156195) 51 (3580)
Marmosa (Micoureus) reginaf 41 180 (142209) 262 (238294) 118 (70164)
Marmosops noctivagusf 16 138 (118155) 183 (154202) 51 (3070)
Marmosops pinheiroie 11 103 (94121) 149 (137156) 27 (2133)
Metachirus nudicaudatuse 9 262 (249287) 345 (326370) 380 (260480)
Monodelphis emiliaef 6 107 (97113) 50 (4553) 30 (2038)
Philander opossume 11 301 (264346) 306 (280333) 549 (380695)
Thylamys karimiio 33 104 (78129) 80 (69106) 28 (1643)
Thylamys macruruso 6 112 (101126) 144 (136153) 39 (3055)
Tlacuatzin canescensp 7 137 (126149) 137 (131145) (3860)
a
Tabulated statistics are the sample mean (rounded to the nearest whole unit) and the observed range (in parentheses)
of measurements and weights recorded from dentally mature specimens; male and female data were combined to increase
sample size despite apparent sexual dimorphism in some species. Most genera are represented by a single exemplar
species, but several genera with recognized subgenera or that include taxa differing conspicuously in size or body:tail
ratios are represented by additional species.
b
Sample size.
c
Obtained by subtracting length of tail from total length following the standard American protocol.
d
Basal flexure to fleshy tip.
e
French Guianan specimens measured by Voss et al. (2001).
f
Measurements and weights from western Brazilian specimens (Patton et al., 2000).
g
Measurements and weight of AMNH 208101; because this was a zoo specimen that may have been obese, the weight
datum is suspect.
h
From Paraguayan specimens (UMMZ 126289, 134022, 134023, 134025, 134559, 134560).
i
From Bolivian specimens measured by Voss et al. (2005).
j
Measurements and weight of MMD 607 (collected near Iquitos, Peru; to be deposited in the Museo de Historia
Natural de la Universidad Nacional Mayor de San Marcos, Lima; M.M. Daz, personal commun.) and INPA 5237
(collected near Mirassol dOeste, Mato Grosso, Brazil; Santos Filho et al., 2007).
k
From Paraguayan specimens (UMMZ 124675, 126104, 133998134006).
l
Measurements and weight from MVZ 173727.
m
From Paraguayan specimens (UMMZ 126109126111, 126113, 134010, 134011, 134017134021).
n
From Surinamese specimens measured by Voss et al. (2001).
o
From Carmignotto and Monfort (2006).
p
External measurements from Oaxacan specimens unaccompanied by weight data (AMNH 3111/2433, 3111/2434,
3114/2437, 148969, 149104, 165651, 165653); range of weights from Zarza et al. (2003).
1987: fig. 4; Patterson and Gallardo, 1987: Lyne (1959), and Brown (1971), whose
fig. 3). terminology is followed here. All of the taxa
FACIAL VIBRISSAE AND MARKINGS: Didel- we examined (including representative species
phid facial vibrissae are grouped into discrete from every genus) exhibit well-developed
tracts that are easily homologized with those mystacial, submental, interramal, supercili-
described and illustrated by Pocock (1914), ary (supraorbital), and genal vibrissae. La-
the rostrum to the frontal region. In other

didelphids, a chevron of dark coronal fur
sometimes extends anteriorly between the
ears and down the rostral midline, but the
condition seen in Caluromys is distinctive and
apparently nonhomologous. No nondidel-
phid marsupial that we examined has a dark
midrostral stripe.
The fur surrounding the eye is not
distinctively colored in Caluromysiops, Lu-
treolina, or Monodelphis, but most other
didelphids have masklike markings. A circu-
mocular mask or ring of dark (usually
blackish) fur that contrasts sharply with the
paler (usually brownish, whitish, or grayish)
color of the crown and cheeks is present in
Glironia, Lestodelphys, and marmosines
(taxa formerly included in Marmosa sensu
Tate, 1933). Species of Caluromys have
essentially similar reddish-brown eye rings
that contrast with grayish cheeks and crowns.
A blackish mask is likewise present in all
examined species of Didelphis, but this
marking is inconspicuous in D. marsupialis
and D. virginiana. A blackish mask is also
present in Metachirus, Chironectes, and
Philander, but the dark circumocular fur in
these taxa is usually continuous with dark fur
on the crown of the head. Among nondidel-
phid marsupials, dark circumocular masks
are present in Dromiciops, some dasyurids
(e.g., Sminthopsis crassicaudata), a few per-
Fig. 1. Ventral view of rhinarium in Thylamys amelemorphians (e.g., Echymipera kalubu),
pallidior (A, UMMZ 156349) and Marmosa and some small arboreal diprotodontians
robinsoni (B, UMMZ 117236). Both species have (e.g., Petaurus breviceps); only one examined
a broad pars supralabialis (psl) that contacts the outgroup taxon, Echymipera kalubu, has a
upper lip. Only a single groove (g) is present along mask that is more or less continuous with
the ventral margin of the pars supralabialis on dark coronal fur.
each side of the median sulcus (ms) in T. pallidior, A distinct whitish supraocular spot is
whereas two ventrolateral grooves are present in
consistently present in species of Metachirus
M. robinsoni. Scale bars 5 2 mm.
and Philander, resulting in the four-eyed
marking by which these animals are com-
beled illustrations of didelphid facial vibris- monly known. Most other didelphids lack
sae are provided by Pocock (1914: fig. 1A) pale supraocular markings. (An indistinct
and Lyne (1959: fig. 2). Most other plesio- pale bar above each eye in Chironectes
morphic marsupials closely resemble didel- appears to be part of the unique transverse
phids in cranial vibrissal traits, but caenoles- banding pattern in that taxon rather than a
tids lack interramal vibrissae (Lyne, 1959). homologue of the condition seen in Meta-
Didelphids exhibit conspicuous taxonomic chirus and Philander.) We have not examined
variation in facial markings (frontispiece; any nondidelphid marsupial with pale su-
Voss and Jansa, 2003: fig. 3). In Caluromys, praocular spots.
a median streak of dark fur, unconnected GULAR GLAND: As described by Tate
with any other dark marking, extends from (1933: 30), many didelphids have a cutaneous
gular (throat) gland, the presence of which is are much harder to observe on fully furred
indicated on dried skins and fluid-preserved adult specimens, the only material commonly
specimens by a bare median patch of skin; available for most species. We found ulnar-
often, but not invariably, the surrounding fur carpal and medial antebrachial vibrissae on
is discolored. According to Barnes (1977: most examined didelphids, whereas anconeal
390), this secretory region contains hyper- and calcaneal vibrissae were often inappar-
trophied apocrine sudoriferous glands and ent.
sebaceous glands, both confined to the All didelphids have soft (nonspinous) fur
thickened dermis. External signs of glandu- consisting of two or more distinct types of
lar activity tend to be maximally developed in hairs whose density and morphology deter-
fully mature males (loc. cit.). mine the appearance and texture of the coat.
No unambiguously glandular throat patch Some taxa (e.g., Caluromys) have somewhat
was observed in any examined specimens of woolly fur that does not lie flat or exhibit the
Caluromys, Caluromysiops, Chironectes, Hy- glossy highlights typically seen in the pelts of
ladelphys, Lutreolina, Philander, or Tlacuat- many other taxa, but textural differences are
zin. By contrast, fully adult male specimens hard to define by objective criteria that can
of Cryptonanus, Gracilinanus, Lestodelphys, be used for character-state definitions or
and Thylamys usually exhibit well-developed taxonomic diagnoses. The only structural
gular glands. A gular gland is also present on (nonpigmental) feature of didelphid body
the young adult male holotype (and only pelage that seems useful in this context is
known skin) of Chacodelphys formosa. Other the presence of uniquely long, coarse guard
didelphid genera (Marmosa, Marmosops, and hairs that project conspicuously from the
Monodelphis) include some species that con- underfur in species of Didelphis.
sistently develop adult male gular glands and The dorsal body pelage of most didelphids
others that just as consistently show no trace is uniformly colored and unpatterned, usual-
of such organs (Voss and Jansa, 2003). ly some shade of brownish or grayish, but
Although adult male Didelphis often have some taxa are distinctively marked (see
discolored gular fur, no glandular skin is illustrations in Eisenberg, 1989; Redford
macroscopically distinguishable. Because we and Eisenberg, 1992; Perez-Hernandez et
were not able to examine any fully adult male al., 1994; Reid, 1997; Eisenberg and Redford,
specimens of Glironia, the occurrence of gular 1999). Blackish transverse bars connected
glands in this taxon is unknown. middorsally on a pale-grayish background,
Gular (or sternal) glands that are macro- for example, characterize Chironectes; dark
scopically and histologically similar to those scapular stripes are unique to Caluromysiops;
of didelphids are present in most dasyurids three longitudinal stripes are present in
(Cooper et al., 2005), but other plesio- several species of Monodelphis (e.g., M.
morphic outgroup taxa (e.g., caenolestids, theresa); a grayish middorsum contrasting
peramelids, Dromiciops) seem to lack all with reddish flanks is exhibited by other
external traces of glandular activity on the species in that genus (e.g., M. brevicaudata);
throat or chest. and a grayish midbody contrasting with
BODY PELAGE: All didelphids have one or reddish head and rump is seen in others
more tracts of postcranial vibrissae (Brown (e.g., M. emiliae). The subtle but consistently
and Yalden, 1973), including ulnar-carpal diagnostic tricolor shading of Thylamys
vibrissae (at the wrist), medial antebrachial and Lestodelphys was described by Tate
vibrissae (at or near the middle of the (1933: 209):
forearm), anconeal vibrissae (at the elbow),
Instead of the usual bicolor system composed of
and/or calcaneal vibrissae (on the ankle).
a dorsal color, paling a little on the sides, which
Lyne (1959) reported the occurrence of is replaced at a generally well-marked transition
postcranial vibrissae in several didelphid line by a distinct ventral color, the elegans group
species based on his examination of pouch [5 Thylamys] displays three distinct shades,
young, whose sensory hair follicles are easily separated from each other along each side by
seen because they are not obscured by coat two lines of transition. The additional lines are
hairs. Unfortunately, postcranial vibrissae subdorsal, running from a point at the center of
the frons [forehead] past the inner edge of each of live Monodelphis emiliae, for example, has
ear (not including it), and straight backward been described as bright glowing violaceous
through scapulae and hips, where they again pink (Emmons, 1997: 34), a lurid hue that is
approach the median line of the body and not retained in any examined museum
merge with the tail. This pair of lines encloses
specimens. What little is known about such
the major part of the dorsal area of head and
body, the color of the area being very dark
fugitive pigments (some of which fluoresce
brownish-gray or grayish fuscous. The fuscous under ultraviolet light) was summarized by
area is pointed at front, projecting forward Pine et al. (1985).
between the ears, and narrows again to a point Most other marsupials have unremarkable
as it merges with the dark color of the upper body pelage that essentially resembles the
surface of the tail. The second [lateral] area, common didelphid condition as described
light gray in color, frequently tinged with buffy above, but postcranial vibrissae are reduced
or yellowish, extends [on each side] between the or absent in a few clades (Lyne, 1959) and
dark dorsal region and the edge of the belly peramelemorphians have stiff, dorsally grooved
color at the normal transition line. Ventral color guard hairs (illustrated in cross section by
either buffy, grayish, or snowy white.
Lyne and McMahon, 1951: figs. 51, 60) that
impart a distinctively harsh, spinous texture
The hair bases of the dorsal fur are heavily to their fur. Most other marsupials also
pigmented, usually dark gray (or grayish), in resemble didelphids in having unpatterned
most didelphids, but species of Didelphis dorsal fur, although there are noteworthy
uniquely exhibit white dorsal underfur. examples of apparently convergent markings
The ventral pelage also exhibits notewor- (e.g., those of Dromiciops somewhat resemble
thy taxonomic variation among didelphids. Chironectes; Marshall, 1978b) and some
In some species the ventral fur is gray- strikingly divergent ones (e.g., the transverse
based, a descriptor that applies when the sacral barring seen in Perameles gunnii; Lyne,
individual hairs are grayish basally and 1951: pl. 1B). All examined outgroup taxa
abruptly paler (usually whitish, yellowish, have dark dorsal underfur.
or brownish) distally; the overall color of the WRIST: The wrists of males and females
ventral fur then depends on the degree to are morphologically similar and externally
which the dark basal pigmentation shows featureless in most didelphids, but striking
through the paler superficial hue. In other sexual dimorphism is present in certain small
species, the ventral fur is partially or entirely arboreal and scansorial forms (Lunde and
self-colored, a descriptor that applies when Schutt, 1999). Grossly enlarged glabrous
the individual hairs have the same pigmenta- tubercles supported internally by carpal
tion (usually whitish) from root to tip. ossifications are exhibited by large adult
Because marked differences in the patterning males of Cryptonanus, Gracilinanus, Marmo-
of gray-based versus self-colored ventral pel- sops, Tlacuatzin, and some species of Mar-
age can occur among closely related (conge- mosa. Two distinct kinds of tubercles can be
neric) species, such variation is often described distinguished, consisting of lateral (ulnar)
and illustrated in the revisionary taxonomic tubercles supported internally by the pisi-
literature (e.g., Patton et al., 2000: fig. 41). form, and medial (radial) tubercles sup-
However, the existence of numerous interme- ported by the prepollex (op. cit.). Although
diate conditions spanning the entire range of some intraspecific variation in the develop-
taxonomic variation in ventral color patterns ment of carpal tubercles has been document-
(e.g., from entirely gray-based to completely ed, most of it can be attributed to ontogeny:
self-white ventral fur) precludes meaningful tubercles are consistently present in the
phylogenetic scoring of this character. largest adult male specimens of species in
Whereas most of the pelage colors (and which such structures occur, whereas they
color patterns) described above are preserved may be lacking in some smaller (presumably
on museum skins that have been protected in younger) conspecific males. Lunde and
dark cabinets from the bleaching effects of Schutt (1999) plausibly suggest that these
light, other colors that can be striking in life structures function as clasping devices during
fade quickly after death. The ventral pelage copulation.
Fig. 2. Sexually dimorphic wrist morphology of Marmosops pinheiroi. Adult males (left, AMNH
267346) possess an externally obvious lateral carpal tubercle that is supported internally by an enlarged
pisiform bone (arrows); females (right, AMNH 267342) do not exhibit this trait.
Lateral carpal tubercles (fig. 2) are more 1958). The pollex (dI) of Chironectes is set
widespread than medial carpal tubercles, off from the other manual digits by a wide
occurring in all taxa that exhibit any exter- gap and appears to be pseudo-opposable
nally obvious sexual dimorphism in the wrist. (sensu Napier, 1961); in fluid-preserved
We found these structures to be consistently specimens it is often folded across the palm
present in large adult male specimens of (Augustiny, 1942: fig. 16).
Cryptonanus, Gracilinanus, Marmosops, Tla- According to Tate (1947), the third and
cuatzin, and most species of Marmosa. In fourth digits of the didelphid manus are
addition, we observed medial carpal tubercles subequal and longer than the other fingers
(Lunde and Schutt, 1999: fig. 3) in Marmosa (fig. 3C), proportions that correspond to the
demerarae, M. mexicana, M. paraguayana, paraxonic morphotype defined by Brown
M. regina, M. robinsoni, and M. rubra. Other and Yalden (1973). Not all didelphids have
didelphids appear to lack sexually dimorphic paraxonic forefeet, however. Instead, many
carpal tubercles,2 but we were not able to have a mesaxonic manus in which dIII is
determine whether or not such structures are distinctly longer than the other fingers; taxa
present in Chacodelphys and Glironia because that exhibit this condition include Chacodel-
no fully adult male specimens of either genus phys, Chironectes, Didelphis, Lestodelphys,
are currently available for study. Lutreolina, Marmosops (fig. 3B), Metachirus,
The water opossum Chironectes has carpal Monodelphis (fig. 3A), Philander, and Thyla-
tubercles that, uniquely, are neither sexually mys. Yet another alternative condition is seen
dimorphic nor ontogenetically variable. In in Caluromys and Caluromysiops, in which
this taxon, juveniles and adults of both sexes dIV is slightly but distinctly longer than dIII
possess a large, fleshy process on the outside (fig. 3D).
of the wrist, resembling a sixth finger, that is Many didelphids have small, weakly re-
supported internally by the pisiform (Augus- curved manual claws that do not extend
tiny, 1942: fig. 16; Mondolfi and Medina, much (if at all) beyond the fleshy apical pad
1957: fig. 14; Oliver, 1976: fig. 1b). of each digit (fig. 3B, C), but some arboreal
No examined nondidelphid marsupial has taxa such as Caluromys (fig. 3D) and others
carpal tubercles of any kind. that are strictly terrestrial such as Mono-
MANUS: The didelphid hand is provided delphis (fig. 3A) have large manual claws that
with five well-developed clawed digits (fig. 3). extend well beyond the apical pads. The large
When the digits are flexed, their tips converge manual claws of Glironia (not illustrated) are
toward the center of the palm, and it seems unusually deep, laterally compressed, and
likely that at least the arboreal and scansorial strongly recurved. Although illustrated taxo-
formswhich tend to have relatively longer nomic differences in claw length are striking,
fingers than terrestrial taxa (Lemelin, 1999; intermediate morphologies observed among
Kirk et al., 2008)are capable of manual other forms (e.g., Hyladelphys) make it
prehension. In most didelphids, the manual difficult to recognize discrete character states
digits are more or less evenly spaced, but in for taxonomic analysis.
some arboreal taxa (e.g., Caluromys, Calur- Like most other plantigrade mammals
omysiops, Marmosa) the gap between dII and (Whipple, 1904; Brown and Yalden, 1973),
dIII is somewhat larger than the gaps didelphids usually have two well-developed
separating other pairs of adjacent fingers, metacarpal pads (thenar, hypothenar) and
and it is possible that these taxa are four interdigital pads on the hairless ventral
incipiently schizodactylous (sensu Haines, (plantar or volar) surface of the manus. These
six pads encircle a central palmar region that
2
Prochel and Sanchez-Villagra (2003) reported that adult is either smooth or sparsely tubercular (as in
males of Monodelphis domestica, a species that lacks external most didelphids; see Hershkovitz, 1997: fig.
evidence of sexual dimorphism in the wrist, nevertheless have 6A) or densely covered with small convex
significantly larger and more robust pisiforms than females. This tubercles (as in Chacodelphys, Lestodelphys,
condition could obviously be interpreted as a state intermediate
to complete absence of sexual dimorphism in the wrist on the
and most species of Thylamys; see Car-
one hand and the very marked dimorphism exhibited by species mignotto and Monfort, 2006: fig. 3C). In
with male carpal tubercles on the other. almost all species, the epidermis that covers
Fig. 3. Dorsal views of right forefeet of Monodelphis brevicaudata (A, AMNH 140466), Marmosops
incanus (B, MVZ 197629), Marmosa robinsoni (C, AMNH 259983), and Caluromys philander (D, AMNH
7433) illustrating generic differences in claw size and digital proportions. Monodelphis and Marmosops
both have mesaxonic forefeet in which the third digit (dIII) is longest, whereas Marmosa has a paraxonic
forefoot in which dIII and dIV are subequal. In Caluromys, dIV is the longest manual digit.
the plantar pads is sharply differentiated from to be larger and to have more pronounced
that of the surrounding plantar surface because dermatoglyphs in arboreal opossums (e.g.,
it is provided with dermatoglyphs (friction or Caluromys) than in terrestrial forms (e.g.,
papillary ridges; Hamrick, 2001) resembling Monodelphis), but a few didelphids exhibit
those on human fingertips. Plantar pads tend qualitatively different morphologies.
The plantar pads of two Brazilian species

of Thylamys (illustrated by Carmignotto and
Monfort, 2006: figs. 3a, 3b) are fused
together and fill the entire palmar region
rather than encircling a central palmar
surface as in other didelphids. Additionally,
only the apices of the plantar pads are
provided with dermatoglyphs in T. velutinus,
whereas the pads of T. karimii lack derma-
toglyphs completely. Instead, the plantar
pads are mostly (T. velutinus) or entirely (T.
karimii) covered with small convex tubercles
like those that occur on the center of the
palm in other species of Thylamys.
The ventral surface of the manus in the
water opossum uniquely lacks any trace of
plantar pads. Instead, the palm of Chiro-
nectes is essentially flat and densely covered
with microscopically dentate tubercles whose
fine structure was described and illustrated
by Brinkmann (1911) and Hamrick (2001).
Numerous smooth, hemisperical papillae
(not mentioned by Hamrick, 2001), however, Fig. 4. Plantar view of right hind foot of
are scattered at regular intervals among the Gracilinanus marica (redrawn from Boas, 1918: pl.
dentate tubercles and presumably have a 1, fig. 2). Pedal characters that distinguish
different function (Brinkmann, 1911). didelphids from most other marsupials include
Among other plesiomorphic marsupials, eleuthrodactyly, a large opposable hallux (digit I),
dasyurids and Dromiciops most closely re- a grooming claw (Putzkralle) on digit II, and
semble didelphids in manual morphology, prominent dermatoglyph-bearing plantar pads.
both taxa having five well-developed clawed
digits. However, whereas dIII and dIV are PES: Most didelphids have an eleuthro-
subequal in Dromiciops, dIII is distinctly dactylous hind foot with five well-developed
longer than dIV in dasyurids. By contrast, digits that are all separate and freely movable
dI and dV are conspicuously reduced and (fig. 4). The only exception is Chironectes,
nail bearing in caenolestids (Osgood, 1921: whose pedal digits are bound together by
pl. 2, fig. 4), and the same digits are webs of skin to form a paddlelike swimming
vestigiallacking claws or nailsin perame- organ (illustrated by Augustiny, 1942: Mon-
lemorphians (Lyne, 1951: fig. 11); of the dolfi and Medina, 1957; Oliver, 1976; Hersh-
remaining manual digits, dIII is distinctly the kovitz, 1997). Although Bensley (1903), Tate
longest in both of these groups. Manual (1933), and Kirsch (1977b) suggested that
plantar pads are indistinct or absent in all some didelphids exhibit incipient fusion of
examined peramelemorphians but they are dII and dIII, we have not observed any
distinct in dasyurids, caenolestids, and Dro- specimen with a hind foot that even remotely
miciops. Whereas the plantar pads of Dromi- resembles the syndactylous condition seen in
ciops and at least some dasyurids (e.g., peramelemorphians and diprotodontians.3
Murexia) have dermatoglyphs, the plantar
pads of other dasyurids (e.g., Sminthopsis 3
Weisbecker and Nilsson (2008) likewise concluded that
crassicaudata) are tubercular, and those of didelphids do not exhibit incipient syndactyly. The illustration
caenolestids are smooth. The central palmar in Hall (1987: fig. 1n) that appears to show an incipiently
surface is essentially smooth (or irregularly syndactylous foot in Didelphis microtarsus (5 Gracilinanus
microtarsus) is a grotesque cartoon that does not accurately
creased) in caenolestids and Dromiciops, but depict the pedal morphology of that species or any other
it is densely tubercular in examined dasyurids opossum (for an anatomically accurate illustration of the hind
and peramelids. foot of Gracilinanus, see fig. 4).
The didelphid hallux (dI) is well developed, the toes. Under high magnification, a fine
set off at a wide angle from the other pedal plantar pelage of very short hairs partially or
digits, and fully opposable. As in other completely covers the heel in some taxa (e.g.,
marsupials, the hallux is claw- and nailless Marmosops), but intermediate conditions in
in opossums. Although this digit tends to be other forms and variation among conspecific
maximally developed in arboreal forms, it is individuals suggest that such microscopic
always long enough to contact the tips of the details are not a reliable basis for either
other digits when they are flexed, even in phylogenetic analysis or taxonomic diagno-
such strictly terrestrial taxa as Lestodelphys, sis. By contrast, a morphologically distinctive
Lutreolina, and Monodelphis. condition occurs in Lestodelphys and Thyla-
The remaining pedal digits bear well- mys (see Carmignotto and Monfort, 2006:
developed claws that are laterally compressed fig. 3), whose heels are entirely covered by
and dorsoventrally recurved like those of coarse (macroscopically visible) fur.
most other plantigrade mammals (Brown Most didelphids have six separate plantar
and Yalden, 1973). The claws on dIV and pads like many other plantigrade mammals
dV are more or less symmetrically crescentic, (Brown and Yalden, 1973), but several
with an unguis that is equally extensive on patterns of taxonomic variation are notewor-
the axial and abaxial surfaces, and a sub- thy: (1) The position normally occupied by
unguis that is only exposed ventrally. Digit the hypothenar (lateral tarsal) and fourth
II, however, bears a strikingly asymmetrical interdigital pads in most didelphids is occu-
grooming claw (Putzkralle; Boas, 1918) with pied by a single elongate pad in Caluromys,
an abaxially exposed subunguis and a round- Caluromysiops, and Glironia; although this
ed, spoon-shaped apex. The subunguis is large structure is plausibly interpreted as the
sometimes also exposed abaxially on the claw result of fusion (hypothenar + interdigital 4),
of dIII, but the asymmetry is consistently less this scenario is complicated by the occasional
than that of the second digital claw, and the presence (e.g., in AMNH 273038) of a small
third claw never has a rounded tip. proximal metatarsal pad that might be a
In the strictly terrestrial opossums Lesto- vestigial hypothenar. (2) The hypothenar is
delphys, Lutreolina, and Monodelphis, the absent or variably present but clearly vesti-
third pedal digit is longer than the adjacent gial in Monodelphis, Lutreolina, Philander,
second and fourth digits, and the hind foot is and Didelphis. (3) The thenar (medial tarsal)
therefore mesaxonic (sensu Brown and Yal- and first interdigital pads are often in contact
den, 1973). By contrast, dII, dIII, and dIV or indistinguishably fused among arboreal
are subequal (none distinctly longer than the taxa (e.g., Caluromys and Marmosa), but a
others) in Didelphis albiventris and D. virgini- continuum of intermediate conditions pre-
ana. Among all other examined didelphid cludes any unambiguous distinction between
taxa (including Didelphis marsupialis), the these morphologies and the separate condi-
second, third, and fourth pedal digits pro- tions seen in terrestrial forms (e.g., Mono-
gressively increase in length, such that dIV is delphis and Thylamys). (4) Interdigital 2 is
the longest toe opposing the hallux. It is much larger than interdigital 3 in most
noteworthy that the relative lengths of pedal arboreal/scansorial forms (e.g., Marmosa,
digits do not strictly covary with those of the Marmosops), but they are subequal in some
manual digits, as they might be expected to terrestrial taxa (e.g., Monodelphis and Lu-
do if transformations of the pes and manus treolina).
were functionally or developmentally deter- The pedal plantar pads of most didelphids
mined by the same factors. For example, are provided with dermatoglyphs like those
whereas dIV is the longest pedal digit in of the manual plantar pads. However, the
Marmosops, dIII is the longest manual digit dermatoglyph-bearing epithelium is much
in that genus. reduced in Thylamys velutinus, and in T.
In most didelphids (e.g., Philander; Hersh- karimii the pedal plantar pads are entirely
kovitz, 1997: fig. 6B) the entire ventral tubercular (lacking dermatoglyphs complete-
surface of the hind foot is macroscopically ly; Carmignotto and Monfort, 2006: fig. 3).
naked from the apex of the heel to the tips of As in so many other aspects of its appendic-
ular morphology, Chironectes is unique Cabrera (1919), and Marshall (1978a), two
among didelphids in lacking any trace of fluid-preserved parous females that we ex-
pedal plantar pads. amined (UMMZ 166634, USNM 536827)
The eleuthrodactylous hind foot of Dro- had pouches exactly resembling the morphol-
miciops resembles the common didelphid ogy illustrated and described by Krieg (1924:
morphotype in having a large opposable fig. 11). Caluromysiops is said to have a
hallux, a grooming claw on dII, dIV . dIII, pouch (Izor and Pine, 1987; Reig et al., 1987),
and dermatoglyph-bearing plantar pads but no explicit description or illustration of
(Hershkovitz, 1999: fig. 21). By contrast, the the female reproductive anatomy of this
hind foot of peramelids is syndactylous (with genus is available, nor were we able to
fused dII and dIII; Hall, 1987; Weisbecker examine suitably preserved parous female
and Nilsson, 2008); the hallux is small and specimens. The presence or absence of a
effectively nonopposable in caenolestids (Os- pouch remains undocumented for many
good, 1921: pl. 2, fig. 3), dasyurids (Thomas, opossums, notably Chacodelphys. Contradic-
1888: pl. 23, fig. 8), and peramelids (Lyne, tory literature accounts of a pouch as present
1951: fig. 12); dIII and dIV are subequal in or absent in Metachirus were discussed by
examined caenolestids and dasyurids, both of Voss and Jansa (2003).
which groups also lack a grooming claw on The marsupium of Caluromys philander
dII; peramelids do not have distinct plantar uniquely consists of deep lateral skin folds
pads; the plantar pads of caenolestids are that enclose the nursing young and open in
smooth; and the plantar pads of some the midline (resembling the morphology that
dasyurids are tubercular. The plantar epithe- Tyndale-Biscoe and Renfree [1987: fig. 2.8]
lium of the heel is macroscopically naked in incorrectly attributed to didelphids in gener-
most of these outgroup taxa, but the under- al). In Caluromys lanatus, Didelphis, and
side of the heel is coarsely furred in some Philander, however, the lateral pockets are
dasyurids (e.g., Sminthopsis crassicaudata). joined posteriorly, forming a more extensive
POUCH AND MAMMAE: Based on our first- enclosure that opens anteriorly (Enders,
hand examination of parous adult female 1937: fig. 19). Yet another condition is
specimens, pouchlike enclosures for nursing exhibited by Chironectes and Lutreolina, in
young are unequivocally present or absent which the lateral pockets are connected
among didelphids. Although our sample sizes anteriorly, forming a marsupium that opens
were always small, we observed no intraspe- posteriorly (Krieg, 1924: fig. 11A; Oliver,
cific variation in this aspect of female 1976: fig. 1B).
reproductive morphology, nor did we ob- In all marsupials that possess a pouch the
serve any intermediate condition between mammae are contained within it, but the
absence and presence of a pouch. Despite mammae of pouchless taxa are variously
the fact that several distinctly different pouch distributed (Tate, 1933: fig. 3). In most
configurations can be recognized among pouchless didelphids (e.g., Glironia, Mar-
opossums, we provisionally recognize all mosa, Metachirus) the mammae are confined
pouches as homologous in the absence of a to a more or less circular inguinal/abdominal
priori evidence to the contrary. array that occupies the same anatomical
We found no trace of a pouch in suitable position as the pouch in taxa that possess a
material (fluid-preserved specimens and care- marsupium. However, other pouchless opos-
fully prepared skins) of parous adult female sums (e.g., Cryptonanus guahybae, Marmo-
Cryptonanus, Glironia, Gracilinanus, Hyladel- sops incanus) have bilaterally paired mammae
phys, Lestodelphys, Marmosa, Marmosops, that extend anteriorly well beyond the pouch
Metachirus, Monodelphis, Thylamys, and region. Although most of these anterior teats
Tlacuatzin. By contrast, well-developed are not actually located on the upper chest,
pouches were consistently found to be they are usually referred to as pectoral or
present in suitably prepared parous adult thoracic mammae (e.g., by Tate, 1933;
females of Caluromys, Chironectes, Didelphis, Reig et al., 1987).
and Philander. Although Lutreolina was Most didelphids have, in addition to
described as pouchless by Thomas (1888), bilaterally paired mammae, an unpaired
median teat that occupies the ventral midline

approximately in the center of the circular
abdominal-inguinal array (fig. 5B). Mamma-
ry counts for didelphids are therefore usually
odd-numbered, an allegedly diagnostic attri-
bute previously noted by Bresslau (1920),
Osgood (1921), and Tate (1933, 1947, 1948a).
The only exceptions that we have encoun-
tered are Caluromys lanatus, Glironia ve-
nusta,4 and Hyladelphys kalinowskii (fig. 5A),
examined females of which have four mam-
mae in two abdominal-inguinal pairs with no
trace of a median teat. By convention,
didelphid mammary complements are sum-
marized by formulae representing the right-
side (R), median (M), left-side (L), and total
(T) teat counts in the format RML 5 T.
Thus, the mammary complement of the
illustrated specimen of Marmosops parvidens
could be written as 414 5 9, whereas that
of Hyladelphys kalinowskii would be 202 5
4. Occasionally, lateral teats are unpaired,
resulting in even-numbered totals for taxa
that possess an unpaired median teat, but
such anomalies are easily distinguished by
formulae (e.g., 314 5 8 if the anteriormost
right teat of AMNH 267344 were really
missing in fig. 5B).
Among other plesiomorphic marsupials
that we examined, a marsupium is absent
only in caenolestids and some dasyurids (e.g.,
Murexia). The pouch of Dromiciops consists
of lateral skin folds opening medially (Hersh-
kovitz, 1999: fig. 14), whereas the pouches of
other dasyurids are variously configured
(Woolley, 1974), and those of peramelids
open posteriorly. Most nondidelphid marsu-
pials lack an unpaired median teat, and
therefore normally have even-numbered
mammary counts, but an unpaired median
teat is said to be present in Rhyncholestes
(e.g., by Osgood, 1924; Patterson and Gal-
lardo, 1987). Fig. 5. Mammary morphology of adult female
CLOACA AND MALE GENITALIA: In most specimens of Hyladelphys kalinowskii (A, AMNH
didelphids the openings of the urogenital and 267339) and Marmosops parvidens (B, AMNH
rectal ducts are inguinal, closely juxtaposed, 267344). Only two pairs of inguinal-abdominal
teats are present in H. kalinowskii, which is
4
A parous adult female specimen with 202 5 4 mammae
unusual among didelphids in lacking an unpaired
was recently collected near Iquitos, Peru, by M. Monica Daz, median teat; its mammary formula is 202 5 4.
whose observations confirm Marshalls (1978c) report of four teats By contrast, the illustrated specimen of M.
based on an old museum skin (FMNH 41440). The Iquitos parvidens has four pairs of inguinal-abdominal
specimen (with field number MMD 607) will be deposited in the teats plus an unpaired median teat (414 5 9).
Museo de Historia Natural de la Universidad Nacional Mayor de
San Marcos (M.M. Daz, personal commun.).
and share a continuous mucosa. In life, both TABLE 5

openings are normally recessed in a common Relative Tail Length of Exemplar Species from 18
sinus (cloaca), but this feature is obscured in Didelphid Genera
male specimens with everted genitalia. Chir-
LT/HBLa
onectes uniquely lacks a cloaca because the
urogenital and rectal openings are widely Caluromys (Caluromys) philander 1.49
separated by furred inguinal skin (Nogueira Caluromys (Mallodelphys) lanatus 1.52
et al., 2004: fig. 1). Conflicting descriptions of Caluromysiops irrupta 1.19
didelphid cloacal morphology by Hershko- Chacodelphys formosa 0.81b
vitz (1992a, 1992b) were discussed and Chironectes minimus 1.20
refuted by Voss and Jansa (2003). All Cryptonanus unduaviensis 1.16
Didelphis marsupialis 1.04
examined nondidelphid marsupials have an
Glironia venusta 1.07
inguinal cloaca resembling the common Gracilinanus agilis 1.40
didelphid condition, with the conspicuous Hyladelphys kalinowskii 1.44
exception of Dromiciops, in which the cloaca Lestodelphys halli 0.67
is basicaudal (Hershkovitz, 1999: fig. 12). Lutreolina crassicaudata 0.96
Descriptions of didelphid male genitalia Marmosa (Marmosa) murina 1.30
(e.g., by Ribeiro and Nogueira, 1990; Marti- Marmosa (Micoureus) regina 1.46
nelli and Nogueira, 1997; Nogueira et al., Marmosops noctivagus 1.33
1999a, 1999b; Nogueira and Castro, 2003) Marmosops pinheiroi 1.45
Metachirus nudicaudatus 1.32
were recently reviewed and summarized by
Monodelphis emiliae 0.47
Nogueira et al. (2004), who tabulated mor-
Philander opossum 1.01
phological comparisons among 18 species in Thylamys karimii 0.77
12 genera. All examined opossums have a Thylamys macrurus 1.29
bifid penis (contra Hershkovitz, 1992b), but Tlacuatzin canescens 1.00
didelphid male genitalia exhibit conspicuous a
taxonomic variation in length, shape, dispo- Length of tail (LT) divided by length of head and
body (HBL), computed from mean values in table 4
sition of the urethral grooves, presence/
except as noted.
absence of diverticula, and other details. b
From external measurements of the holotype (Voss et
Unfortunately, the genitalic characters of al., 2004a).
many species remain unstudied, and we were
unable to extend Nogueira et al.s (2004)
observations for this report. epidermal scales) extends to a variable extent
TAIL: Most didelphids have a tail that is onto didelphid tails. Body fur extends onto
substantially longer than the combined the tail much farther dorsally than ventrally
length of the head and body, but some taxa in Glironia (see da Silva and Langguth, 1989:
are much shorter tailed (table 5). The known fig. 1), Caluromysiops, and some species of
range of relative tail length in the family is Caluromys and Monodelphis (see Voss et al.,
bracketed on the one hand by Gracilinanus 2001: fig. 29A). By contrast, body fur extends
emiliae, an arboreal species with a tail that onto the tail dorsally and ventrally to about
may be almost twice as long as the combined the same extent (from about one-sixth to
length of its head and body (Voss et al., 2005: about one-third the length of that organ) in
table 2), and on the other by some terrestrial Caluromys philander, Chironectes, Didelphis,
forms (e.g., Monodelphis spp.) with a tail that Lutreolina, Philander, and some species of
may be less than half as long as the head and Marmosa (e.g., M. paraguayana). In most
body. However, arboreal taxa are not always other didelphids, however, body fur does not
longer tailed than terrestrial forms. Tree- extend more than a short distance, if at all,
dwelling Glironia venusta, for example, has a onto the tail base.
tail that is subequal in length to its head and The exposed skin of didelphid tails is
body, whereas ground-dwelling Metachirus variously pigmented. Some species have
nudicaudatus is much longer tailed. uniformly pigmented (concolored) tails that
Body pelage (soft fur composed of ordi- are usually some shade of grayish or brown-
nary coat hairs that are not associated with ish, but pale markings are not uncommon.
Many species, for example, have parti-col- most didelphids, but four or more hairs
ored tails that are paler distally than basally. usually emerge from each caudal scale in
In most taxa with pale tail tips, the basal color Lutreolina and Philander. Although individ-
is grayish or brownish and the transition is ual cutaneous scales are hard to distinguish in
either gradual (the basal color fading to Monodelphis, the caudal hairs of species that
whitish distally; e.g., in Marmosops ocellatus) we examined usually emerge from the skin in
or irregularly mottled (e.g., as in Caluromys triplets, so this genus appears to conform to
derbianus; frontispiece). However, in most the widespread didelphid condition. The
specimens of Chironectes (frontispiece), Di- number of hairs normally associated with
delphis, Lutreolina, and Philander, the tail is each caudal scale is impossible to assess in
blackish basally with an abruptly white tip. Glironia (whose dorsal caudal surface is furry
Dorsoventrally bicolored tails (which are and whose ventral caudal surface is modified
grayish or brownish above and pale on the for prehension; see above and below) and
underside) represent another widespread pat- Lestodelphys (on which no discrete clustering
tern of didelphid caudal marking that is within rows of caudal hairs could be distin-
indistinct in some species but sharply defined guished with confidence). The hairs that
in others. Different patterns of caudal mark- emerge from the posterior margin of each
ings often distinguish closely related species caudal scale are not grossly differentiated,
and can be used in combination with other varying in length but subequal in thickness,
external traits to identify congeneric taxa in in most didelphids. However, in most species
the field (Voss et al., 2004b: table 4). of Marmosops the middle hair of each
The epidermal scales that cover the un- caudal-scale triplet is conspicuously thicker
furred nonprehensile surfaces of didelphid than the lateral hairs, and it is often more
tails differ taxonomically in shape and darkly pigmented (Gardner and Creighton,
arrangement (Tate, 1933: fig. 2). Square or 1989).
rectangular caudal scales in unambiguously Although all didelphids are perhaps capa-
annular series occur in Chacodelphys, Cryp- ble of caudal prehension to some extent,
tonanus, Thylamys, Tlacuatzin, and most external morphological features associated
species of Gracilinanus. Although individual with this behavior are variably developed in
cutaneous scales are indistinct in Lestodel- the family. The unfurred caudal surfaces of
phys and Monodelphis, the annular arrange- Chacodelphys, Chironectes, Lestodelphys, Lu-
ment of caudal bristles in both taxa suggest treolina, Metachirus, and Monodelphis are
that they also conform to the Thylamys covered with unmodified scales from base to
pattern. By contrast, rhomboidal (diamond- tip (for illustrations of Chironectes and
shaped) or hexagonal scales in spiral series Lutreolina, see Hershkovitz, 1997: fig. 7). In
occur in Caluromys, Caluromysiops, Chiro- taxa conforming to this morphology, the tail
nectes, Didelphis, Lutreolina, Marmosops, tip is sometimes provided with a smooth
Philander, and most species of Marmosa. A terminal button but never with a ventrally
few taxa, however, have caudal scales that expanded apical pad bearing dermatoglyphs.
appear to be morphologically intermediate The tails of all other didelphids are provided
between the annular and spiral morphotypes. with a distal prehensile surface that may be
In these taxa (e.g., Hyladelphys, Metachirus, smooth or covered by modified scales (con-
Marmosa mexicana) the caudal scales are spicuously unlike those of the caudal dor-
arrayed in annular series over the vertebral sum) but is always transversely creased and
articulations and in more or less spiral series glabrous; in taxa conforming to this mor-
elsewhere. Glironia cannot be assigned to any phology, the tail tip is invariably provided
of these caudal-scale patterns because the with a ventrally expanded pad bearing
dorsal and lateral surfaces of its tail are dermatoglyphs (for illustrations of Didelphis
covered by body fur and the entire ventral and Philander, see Hershkovitz, 1997: fig. 7).
surface is modified for prehension (da Silva Some taxa additionally exhibit caudal mod-
and Langguth, 1989: fig. 1). ifications for basal prehension. Whereas the
Three stiff, bristlelike hairs emerge from unfurred ventral surface of the tail base in
the posterior margin of each caudal scale in most didelphids is covered by smooth flat
scales, each of which is soft and flexible accounts of the didelphid head skeleton as
(yielding easily to the tip of a probe on fluid- exemplified by two species in a single genus
preserved material), the scales on the underside that is increasingly popular among biomed-
of the base of the tail are heavily cornified, ical researchers.
forming hard raised tubercles in Glironia (see Other didelphids, however, differ from
da Silva and Langguth, 1989: fig 1a), Calur- Monodelphis in numerous aspects of cranial
omysiops, and some species of Caluromys. morphology (Voss and Jansa, 2003), includ-
The tail is a slender, muscular organ in ing some characters that remain undescribed
most didelphids, but Thylamys and Lestodel- in the literature. Additionally, the literature
phys have incrassate tails (Morton, 1980) in contains no comprehensive account of the
which fat is seasonally deposited. Incrassate cranial features by which didelphids as a
tails can be recognized superficially by their group differ from other marsupials, although
characteristically swollen outline and soft important comparative observations were
texture in fresh and fluid-preserved material, made by Osgood (1921), Tate (1947,
and by their flattened, grease-stained appear- 1948b), Archer (1976a), Reig et al. (1987),
ance in most skins. When the external aspect Marshall et al. (1995), Wroe (1997), Muizon
of the tail leaves some room for doubt, the (1998), and Sanchez-Villagra and Wible
presence or absence of subcutaneous adipose (2002), among others. Following a brief
tissue is easily determined by dissection. For consideration of overall cranial shape, the
example, a midventral incision that we made following accounts summarize relevant cranial
near the base of the tail of MZUSP 32097 comparisons in a roughly anterior to posterior
provided unambiguous confirmation of Car- sequence along the dorsolateral contours of
mignotto and Monforts (2006) statement the skull, starting with the rostrum and nasal
that the tail of Thylamys macrurus is incras- cavity and proceeding to the orbital region
sate (contra Palma, 1997; Creighton and and braincase. A second sequence describes
Gardner, 2008). ventral cranial features beginning with the
Among other marsupials, representative palate and proceeding to the basicranium and
caenolestids, dasyurids, and peramelemor- occiput. The principal osteological features of
phians that we examined have nonprehensile the didelphid skull and mandible are illustrat-
tails that are uniformly pigmented, annularly ed in figures 6 and 7.
scaled, and lack basal extensions of body SHAPE VARIATION: Didelphids exhibit con-
pelage; most are slender, muscular organs, spicuous variation in overall cranial shape
but those of Rhyncholestes and some dasyur- that is illustrated in the taxonomic accounts
ids (e.g., Sminthopsis crassicaudata) are in- below (figs. 3754). Although some shape
crassate. The tail of Dromiciops is incrassate, variation is obviously allometric (larger
annularly scaled, and distally prehensile, with opossums tending to have relatively larger
a conspicuous extension of body pelage onto rostra and temporal fossae but relatively
the basal 1/4 to 1/3 of the caudal dorsum. smaller orbits and braincases than smaller
opossums), size-independent shape contrasts
Cranium and Mandible are also obvious. Thus, the rostrum is
relatively short and broad in both Hyladel-
The best general description of didelphid phys and Lutreolina, whereas the rostrum is
cranial morphology is Wibles (2003) account relatively long and narrow in Marmosops and
of the skull of Monodelphis brevicaudata, Metachirus.5 A continuous taxonomic series
which includes a review of the literature on of intermediate proportions makes any qual-
didelphid cranial ontogeny, myology, and
other relevant topics. Whereas Wibles de- 5
We quantified rostral proportions in exemplar skulls from
scription is mostly limited to external cranial taxa that were visually judged to represent morphological
features, Rowe et al. (2005) provided a extremes in this respect. We measured rostral length (RL) from
detailed description of internal structures of the anterior margin of one orbit to the tip of the ipsilateral nasal
bone, and we measured skull size as condylobasal length (CBL).
the nasal skeleton of a closely related species, The values we obtained for the index (RL 4 CBL) 3 100
M. domestica. Together, these publications ranged from about 30% (in Lutreolina) to about 45% (in
provide detailed and abundantly illustrated Metachirus).
Fig. 6. Dorsal and ventral cranial views of Marmosa murina showing principal osteological features
mentioned in the text. Abbreviations: als, alisphenoid; atp, alisphenoid tympanic process; bo, basioccipital;
bs, basisphenoid; cc, carotid canal; ect, ectotympanic; fm, foramen magnum; fpj, frontal process of jugal;
fro, frontal; gf, glenoid fossa; gpa, glenoid process of alisphenoid; ip, interparietal; jug, jugal; lac, lacrimal;
max, maxillary; mp, mastoid process (of petrosal); nas, nasal; occ, occipital condyle (of exoccipital); of,
orbital fossa; pal, palatine; par, parietal; pcf, paracanine fossa; pogp, postglenoid process (of squamosal);
pop, postorbital process; pre, premaxillary; prgp, preglenoid process (of jugal); pro, promontorium (of
petrosal); ps, presphenoid; pt, pterygoid; rtp, rostral tympanic process (of petrosal); sq, squamosal; tcf,
transverse canal foramen; tf, temporal fossa; za, zygomatic arch.
itative coding of such variation an arbitrary inary attempts to explain shape differences in
exercise, but shape differences are visually functional terms (e.g., Medelln, 1991).
conspicuous and are sometimes useful for ROSTRUM: The premaxillae of many di-
diagnosing clades discovered by analyzing delphids (Caluromysiops, Chacodelphys, Chir-
other data. The nascent literature on didel- onectes, Cryptonanus, Didelphis, Glironia,
phid cranial morphometrics is not large, but Hyladelphys, Lestodelphys, Lutreolina, Meta-
it includes studies of intraspecific ontogenetic chirus, Monodelphis, Philander, Thylamys,
variation (Abdala et al., 2001; Flores et al., Tlacuatzin) are short, terminating abruptly
2003), multivariate comparisons of conspe- in front of the incisors, and often lack a
cific species (e.g., Ventura et al., 1998, 2002; definitive suture between the left and right
Cerqueira and Lemos, 2000), analyses of elements (fig. 8B). By contrast, in other
evolutionary rates (Lemos et al., 2001), didelphids the premaxillae are produced
geometric modeling of taxonomic variation anteriorly as a more or less acutely pointed
(Astua de Moraes et al., 2000), and prelim- shelflike process that extends the bony
Fig. 7. Left lateral cranial and mandibular views of Marmosa murina showing principal osteological
features mentioned in the text. Abbreviations: als, alisphenoid; ap, angular process; atp, alisphenoid
tympanic process; conp, condylar process; corp, coronoid process; ect, ectotympanic; exo, exoccipital; fpj,
frontal process of jugal; fr, foramen rotundum; fro, frontal; hpp, hamular process of pterygoid; iof,
infraorbital foramen; ip, interparietal; jug, jugal; lac, lacrimal; lc, lambdoid crest; lf, lacrimal foramina;
maf, masseteric fossa; max, maxillary; mef, mental foramina; nas, nasal; pal, palatine; par, parietal; pc, pars
cochlearis (of petrosal); pcf, paracanine fossa; pm, pars mastoideus (of petrosal); pop, postorbital process;
pre, premaxillary; rmf, retromolar fossa; sq, squamosal; ssf, subsquamosal foramen; sup, supraoccipital;
zps, zygomatic process of squamosal.
rostrum beyond I1 and contains a distinct nasal orifice is consequently not visible from
suture between the right and left bones a dorsal perspective (figs. 3744, 4954). In
(fig. 8A). Taxa that exhibit a well-developed Chironectes, Didelphis, Lutreolina, and Phi-
premaxillary rostral process include Calur- lander, however, the tips of the nasals do not
omys, Gracilinanus, Marmosa, and most extend so far anteriorly, and the nasal orifice
examined species of Marmosops. Aspects of is dorsally exposed (figs. 4548).
rostral process development that impact The nasals are wider posteriorly than
phylogenetic character scoring, including anteriorly in most didelphids, but they are
some discrepancies and ambiguities in the not so wide posteriorly as to contact the
published literature on didelphid premaxil- lacrimals (except as rare, often unilateral,
lary morphology, were previously discussed variants). The transition from the narrow
by Voss and Jansa (2003: 2223). anterior part of each nasal to the broader
Didelphid nasal bones always extend posterior part may be gradual or abrupt and
anteriorly beyond the vertically oriented usually occurs at or near the maxillary-
facial processes of the premaxillae to form a frontal suture. By contrast with this wide-
well-defined median apex, and they always spread condition, a few taxa (Chacodelphys,
extend posteriorly between the lacrimals. Thylamys, Marmosops incanus) have uni-
Despite such consistency, taxonomic differ- formly narrow nasals with subparallel lateral
ences in nasal length can be defined with margins (e.g., fig. 54). Due to the usual
respect to other anatomical landmarks. The absence of nasolacrimal contact, the maxillae
tips of the nasal bones are produced anteri- and frontals are normally in contact on both
orly beyond I1 in most opossums, and the sides of the rostrum.
didelphids; in particular, no fenestra is ever

present at or near the conjunction of the
maxillary, nasal, and frontal bones.
Other marsupials exhibit a wide range of
rostral morphologies. In all examined da-
syuromorphians, for example, the nasals are
very short and do not extend anteriorly
beyond the facial processes of the premaxil-
lae; rather than forming an acute apex, they
are notched medially, and the incisive foram-
ina (in the floor of the nasal orifice) are
exposed to dorsal view (e.g., in Dasyurus;
Flores et al., 2006: fig. 4B). In Recent
peramelemorphians, the nasals do not extend
posteriorly between the lacrimals (Freedman,
1967: fig. 2), whereas Recent caenolestids
have a prominent fenestra on each side of the
rostrum at or near the conjunction of the
nasal, frontal, and maxillary bones (Thomas,
1895; Osgood, 1921: pl. 20, fig. 2). In a few
Old World marsupials (e.g., Lasiorhinus,
Tarsipes) the nasals contact the lacrimals
(separating the maxillae from the frontals),
whereas the premaxillae contact the frontals
(separating the nasals from both the lacrimals
and the maxillae) in dactylopsiline petaurids
(Tate, 1948a). By contrast, many other
marsupials (e.g., Dromiciops) essentially re-
semble didelphids in rostral morphology.
NASAL CAVITY: The complex morphology
Fig. 8. Ventral view of rostrum in Marmosa of the bony turbinals inside the didelphid
rubra (A, MVZ 153280) and Monodelphis brevicau- nasal cavity is difficult to visualize without X-
data (B, AMNH 257203) illustrating taxonomic
ray computed tomography, which has only
differences in premaxillary morphology. A broad,
shelflike rostral process (rp) extends the suture been used to describe the endonasal skeleton
between right and left bones well anterior to the of Monodelphis domestica (see Rowe et al.,
incisors in M. rubra, but the premaxillae form only 2005). However, some turbinal structures are
narrow alveolar rims anterior to I1 in M. brevicau- accessible to direct inspection. The maxillo-
data, where the left and right bones are separated turbinals, for example, are thin, dorsally
by a small tissue-filled gap (not a distinct suture). scrolling sheets of bone that occupy the
lower part of the didelphid nasal fossa, where
Most of the lateral surface of the didelphid they arise from the inner surface of the
rostrum is formed by the maxilla, which is maxilla on each side. In most didelphids, the
prominently perforated above P2 or P3 by the greater curvature of each maxilloturbinal
infraorbital foramen. The foramen is notably scroll throws off secondary lamellae, which
larger in Chironectes than in other didelphids, branch to form tertiary lamellae, some of
presumably to accommodate the hypertro- which branch again to form an elaborate
phied infraorbital nerve that conducts trigem- dendritic mass that fills much of the ventro-
inal sensory fibers from the well-developed lateral part of the nasal lumen. By contrast,
mystacial vibrissae of this semiaquatic taxon the maxilloturbinals are simple, slender,
(Sanchez-Villagra and Asher, 2002). Aside unbranched (or sparsely ornamented) scrolls
from small nutrient foramina that may be in all examined species of Monodelphis.
variably present or absent in some taxa, there The morphology of the nasal cavity
are no other lateral rostral openings in remains to be surveyed among other marsu-
pials. However, some peramelemorphians the alisphenoid in Myrmecobius and many

(e.g., Echymipera, Perameles) appear to have diprotodontians.
simple (unbranched) maxilloturbinals, where- ORBITAL MOSAIC: The anteriormost part
as the maxilloturbinals of examined dasyur- of the didelphid orbit is formed by the
ids (e.g., Murexia, Sminthopsis) and Dromi- lacrimal, which is always prominently ex-
ciops closely resemble the elaborately den- posed in lateral view. The lacrimal is
dritic condition seen in most didelphids. A perforated by one or more lacrimal foramina
monographic study of marsupial endonasal that are sometimes concealed within the orbit
features based on high-resolution X-ray (fig. 9A) but usually open laterally on or near
computed tomography (Macrini and Voss, the orbital margin (fig. 9B). Most didelphids
in preparation) will doubtless provide addi- normally have two lacrimal foramina on each
tional points of comparison among didel- side, but Chironectes, Hyladelphys, and some
phids and other marsupial clades. populations of Didelphis virginiana usually
ZYGOMATIC ARCH: The bones comprising have just one lacrimal foramen, and many
the zygomatic arch exhibit few variable other didelphids that normally have two
features among didelphids. The maxillary- lacrimal foramina occasionally have a single
jugal suture is always more or less straight or foramen on one or both sides of the skull.
irregularly crescentic, a frontal process of the The orbital margin formed by the lacrimal is
jugal is invariably present, and the jugal- smoothly rounded in all didelphids, none of
squamosal suture is always deeply inflected which exhibits lacrimal tubercles (e.g., like
(typically ,- or ,- shaped). Likewise, a those seen in macropodids; Wells and Ted-
faceted preglenoid process of the jugal and a ford, 1995: fig. 9) or distinct crests (as in
well-developed postglenoid process of the Myrmecobius and some peramelemorphians).
squamosal are always present, and a distinct Unlike the condition seen in some Old World
glenoid (entoglenoid) process of the alisphe- marsupials with maxillary-frontal contact on
noid consistently forms part of the posterior the medial wall of the orbit (Flannery et al.,
zygomatic root. Although the zygomatic arch 1987: fig. 3), the didelphid lacrimal is always
tends to be more gracile, to have a more in posteroventral contact with the palatine.6
pronounced suborbital deflection, and to The medial wall of the didelphid orbit is
have a more distinct frontal process in small perforated by several openings, including the
opossums (with relatively large eyes and sphenopalatine foramen (always in the pala-
weakly developed masticatory muscles; e.g., tine bone), the ethmoid foramen (in the suture
Hyladelphys; fig. 40) than in large opossums
between the orbitosphenoid and frontal), the
(with relatively smaller eyes but massive
sphenorbital fissure (between the palatine,
masticatory muscles; e.g., Lutreolina; fig.
orbitosphenoid, alisphenoid, presphenoid,
47), most didelphids exhibit intermediate
and sometimes the pterygoid), and the fora-
zygomatic morphologies.
men rotundum (in the alisphenoid). The
Contrasting features of the zygomatic
configuration of these orbital perforations in
region among other marsupial groups in-
all examined taxa is essentially similar to that
clude: (1) the deeply inflected maxillary-jugal
illustrated for Monodelphis by Wible (2003:
suture of peramelemorphians, which divides
fig. 4). In particular, the foramen rotundum is
the jugal into distinct anterodorsal and
always exposed to lateral view behind the
anteroventral processes flanking a well-de-
sphenorbital fissure, from which it is invari-
veloped nasolabial fossa (Filan, 1990); (2) the
ably separated by a bony partition.7 Many
absence of a frontal process of the jugal in
caenolestids and some peramelemorphians 6
Note that elements of the didelphid orbital mosaic are
(Osgood, 1921); (3) the absence of a faceted incorrectly labeled in some published illustrations (e.g.,
preglenoid process of the jugal in several Old Hershkovitz, 1992b [fig. 19], 1997 [fig. 12]), where the orbital
World clades (e.g., Thylacinus and macro- process of the palatine that contacts the lacrimal is misidentified
podoids); (4) the absence of a distinct as the sphenoid.
7
According to Novacek (1986, 1993), the foramen rotun-
postglenoid process of the squamosal in dum is confluent with the sphenorbital fissure in didelphids, but
Hypsiprymnodon, Tarsipes, and vombatids; these foramina are unambiguously separate openings in all of
and (5) the absence of a glenoid process of the material examined by us and by Wible (2003).
Fig. 9. Detail of the anterior orbital region in Marmosops parvidens (A, AMNH 267359) and M.
pinheiroi (B, AMNH 267345). In M. parvidens the lacrimal foramina (lf) are concealed from lateral view
inside the orbit, whereas the lacimal foramina are laterally exposed anterior to the orbit in M. pinheiroi.
Other abbreviations: fro, frontal; jug, jugal; lac, lacrimal; max, maxillary; nas, nasal; pal, palatine.
other marsupials essentially resemble didel- omys, Caluromysiops, Glironia, Tlacuatzin,

phids in these features, but some exhibit and most species of Marmosa (fig. 11A) the
noteworthy differences. In Dromiciops, for postorbital processes are flattened and more
example, the foramen rotundum is concealed or less triangular, but the postorbital pro-
from lateral view within a common vestibule cesses of Chironectes, Didelphis, Lutreolina,
that it shares with the sphenorbital fissure in and Philander are often bluntly pyramidal or
the rear of the orbit (Giannini et al., 2004). hornlike projections. Because frontal out-
The maxillary and alisphenoid bones are growths tend to develop late in postweaning
usually separated by the palatine on the ontogeny (Abdala et al., 2001), postorbital
orbital floor of most didelphids (fig. 10A). By processes are often absent in juvenile and
contrast, the maxillary and the alisphenoid subadult specimens of taxa that normally
are consistently in contact on the orbital floor exhibit them when fully grown. The postor-
of Lutreolina and Monodelphis (fig. 10B). bital processes of Glironia uniquely consist of
The maxillary and alisphenoid are separated both frontal and parietal moieties (fig. 37).
by the palatine in most nondidelphid marsu- Among didelphids that normally lack
pials, but the maxillary and alisphenoid distinct postorbital processes (Chacodelphys,
broadly contact one another in all examined Cryptonanus, Gracilinanus, Lestodelphys,
macropodoids (e.g., Macropus; Wells and Marmosops, Metachirus, Monodelphis, Thy-
Tedford, 1995: fig. 9A). lamys), the supraorbital margins of the
INTERORBITAL REGION: The morphology frontals may be beaded (with a dorsally
of the interorbital region is highly variable upturned edge) or more or less smooth and
among didelphids. The frontals produce featureless. Taxa with supraorbital beads
distinct postorbital processes in several gen- include some species of Gracilinanus (e.g.,
era, most of which are rather large opossums G. emiliae), Hyladelphys, some species of
(e.g., Caluromys, Philander), but some of Marmosops (e.g., M. noctivagus; fig. 11B),
which are small (e.g., Tlacuatzin). In Calur- Marmosa rubra, and Metachirus. By contrast,
Fig. 10. Oblique dorsolateral view of the left orbital floor in Thylamys venustus (A, AMNH 263562)
and Monodelphis peruviana (B, AMNH 272695) illustrating taxonomic differences in sutural patterns. In
Thylamys (and most other didelphids), the maxillary (max) and alisphenoid (als) bones are separated by
the palatine (pal), but the alisphenoid extends anteriorly across the palatine to contact the maxillary in
Monodelphis. Other osteological abbreviations: fro, frontal; hp, hamular process of pterygoid; jug, jugal;
lac, lacrimal; par, parietal; sq, squamosal.
the supraorbital margins are consistently frontal and midparietal sutures, but the right
smooth and essentially featureless in most and left frontals and parietals are co-ossified
specimens of Cryptonanus, Lestodelphys, in all of the remaining (fully adult) specimens
Thylamys, and other species of Marmosops of Caluromysiops that we examined. Most
and Gracilinanus (e.g., G. agilis; fig. 11C). nondidelphid marsupials have ontogenetical-
Distinct postorbital processes are consis- ly persistent median sutures separating the
tently absent in many nondidelphid marsu- braincase roofing bones, but the left and
pial groups (e.g., caenolestids, peramelemor- right frontals and parietals are co-ossified in
phians, macropodoids), but they are present all examined subadult and adult caenolestids
in some dasyuromorphians (e.g., Thylacinus, (Osgood, 1924: figs. 13; Patterson and
Sarcophilus) and in a few diprotodontians Gallardo, 1987: fig. 1). The midparietal
(e.g., Petaurus breviceps, thylacoleonids). suture is also fused in most examined
Other aspects of interorbital morphology are postjuvenile peramelemorphians.
too variable among nondidelphid marsupial The scars that mark the dorsalmost origin
clades to describe succinctly here, but none of the temporalis muscle on each side of the
appear to provide an unambiguous basis for braincase are widely separated, and no
phylogenetic inference or taxonomic diagnosis. sagittal crest is developed in most didelphids.
DORSOLATERAL BRAINCASE: The right and In large adult specimens of Caluromys,
left frontals and parietals are separated by Lestodelphys, and Monodelphis, however, a
ontogenetically persistent median sutures in small sagittal crest is sometimes developed
most didelphids (fig. 6), but the midfrontal over the interparietal or along the midpar-
suture is incomplete or absent in most ietal suture. By contrast, much larger sagittal
juveniles and in all examined subadult and crests that extend anteriorly onto the fron-
adult specimens of Chironectes (fig. 45), talsan unambiguously different condi-
Didelphis (fig. 46), Lutreolina (fig. 47), and tionare consistently developed in adult
Philander (fig. 48). All examined juveniles specimens of Caluromysiops, Chironectes,
and one young adult specimen (FMNH Didelphis, Lutreolina, and Philander. Sagittal
84426) of Caluromysiops have complete mid- crests are altogether absent in caenolestids,
Fig. 11. Dorsal views of the interorbital region in Marmosa murina (A, AMNH 267368), Marmosops
noctivagus (B, AMNH 262402), and Gracilinanus agilis (C, MVZ 197439) illustrating taxonomic
differences in the development of processes and beads. Distinct postorbital processes (pop) are normally
present in fully adult specimens of most Marmosa species, but they are usually absent in Marmosops and
Gracilinanus. Supraorbital beads (b), consisting of upturned dorsally grooved ridges, are present in some
but not all species of Marmosops. Other abbreviations: ioc, interorbital constriction; poc, postorbital
constriction. Scale bars 5 5 mm.
Dromiciops, and in most small-bodied Old which the lateral braincase morphology is
World marsupials, but well-developed sagit- known, also exhibits squamosal-frontal con-
tal crests are present in adult specimens of tact.8
some dasyuromorphians (Dasyurus, Sarco- In most didelphids (Chironectes, Didelphis,
philus, Thylacinus), some peramelemorphians Glironia, Lutreolina, Marmosa, Metachirus,
(Macrotis), and some diprotodontians (e.g., Monodelphis, Philander, and Tlacuatzin) the
phalangerids, {Wakaleo, {Ekaltadeta). petrosal is only exposed on the occiput
The taxonomic distribution of parietal- (behind the lambdoid crest) and on the
alisphenoid versus frontal-squamosal contact ventral surface of the skull (between the
on the lateral braincase of metatherian exoccipital, basioccipital, and alisphenoid;
mammals has often been discussed by au- see Wible, 1990: fig. 1). In taxa conforming
thors (e.g., Archer, 1976a; Marshall et al., to this morphology, there are no gaps among
1995; Muizon, 1998; Wroe et al., 1998), not the bones that normally form the lateral
all of whom have correctly reported the surface of the braincase (squamosal, parietal,
distribution of these alternative conditions and interparietal; fig. 12A). By contrast, the
among didelphids (Voss and Jansa, 2003). In petrosal capsule that encloses the parafloc-
fact, parietal-alisphenoid contact is exhibited culus and the semicircular canals (5 pars
by all didelphids with the unique exception of mastoideus or pars canalicularis) is consis-
Metachirus (fig. 44). Among other marsupi- tently exposed through a fenestra in the
als, parietal-alisphenoid contact is also found suture between the parietal and squamosal
in caenolestids (Osgood, 1921: pl. 20, fig. 2), in Chacodelphys, Cryptonanus, Gracilinanus,
most dasyuromorphians (e.g., Dasycercus; Thylamys, and some species of Marmosops
Jones, 1949: fig. 8), some fossil peramele- (fig. 12B). Both conditionspresence and
morphians (e.g., Yarala; Muirhead, 2000), absence of a fenestra in the squamosal-
and many diprotodontians, whereas squamo-
8
sal-frontal contact is present in some dasyur- Hershkovitz (1999: fig. 25) erroneously depicted the lateral
omorphians (e.g., Sminthopsis; Archer, 1981: braincase of Dromiciops with alisphenoid-parietal contact and
with a large orbitosphenoid ossification wedged between the
fig. 4), all Recent peramelemorphians, and a frontal, alisphenoid, and parietal bones. The correct morphol-
few diprotodontians (Wroe et al., 1998). ogy of lateral braincase elements in this taxon was illustrated by
Dromiciops, the only microbiotherian for Giannini et al. (2004: fig. 2G).
Fig. 12. Lateral view of posterior braincase in Marmosa murina (A, AMNH 272816) and Marmosops
impavidus (B, AMNH 272709), illustrating the presence of a fenestra (fen) that exposes the petrosal (pet)
between the parietal (par) and squamosal (sq) in Marmosops. The petrosal is not laterally exposed by a
fenestra in the parietal-squamosal suture of Marmosa. Other abbreviations: fc, fenestra cochleae; ssf,
subsquamosal foramen.
parietal sutureoccur as balanced polymor- doid crest) as the sutured interparietals of

phisms (neither state clearly predominating) other marsupials. The didelphid interparietal,
in Lestodelphys, Marmosops incanus, and M. however, is fused with the supraoccipital in
noctivagus. No nondidelphid marsupial that all of the skulls we examined (including
we examined has a fenestrated squamosal- postweaning juveniles assignable to Gard-
parietal suture. ners [1973] age class 1), few of which show
INTERPARIETAL: Although some authors any trace of a suture.9 Fortunately, develop-
(e.g., Novacek, 1993) have stated that mar- mental studies of didelphid pouch young are
supials lack an interparietal bone, a large available to prove the existence of a distinct
interparietal is unequivocally present in interparietal center of ossification. In Didel-
Dromiciops (see Giannini et al., 2004: fig. phis the interparietal first appears around day
3), many diprotodontians (e.g., Tarsipes; 8 postpartum (p) and fuses with the supra-
Parker, 1890: fig. 2), and some dasyuromor- orbital by day 28p (Nesslinger, 1956), where-
phians (e.g., Myrmecobius). In these taxa, the as in Monodelphis these events occur on days
sutures that separate the interparietal from 3p and 8p, respectively (Clark and Smith,
adjacent bones (supraoccipital and parietals) 1993). By contrast, the didelphid interparietal
are clearly visible and ontogenetically persis- never appears to fuse with the parietals, the
tent. In all of the taxa we examined with such sutures between them persisting even in the
distinctly sutured interparietals, the interpa- largest adult specimens we examined.
rietal-supraoccipital boundary coincides
closely with the transverse (lambdoid) crest 9
Vestiges of the interparietal-supraoccipital suture in juvenile
that marks the dorsalmost insertion of the specimens of Monodelphis brevicaudata were described by Wible
neck extensor musculature. (2003). What appears to be a well-defined suture between the
An interparietal is also unambiguously interparietal and supraoccipital of Didelphis albiventris in an
present in didelphids, all of which exhibit a illustration published by Abdala et al. (2001: fig. 3) was
intended to represent the hypothetical boundary between bones
large, unpaired, wedge-shaped or oblong that were indistinguishably fused in all of the specimens
element between the left and right parietals examined by the authors of that report (D.A. Flores, personal
in the same position (anterior to the lamb- commun.).
Fig. 13. Posterior and lateral views of the occipital region in Lestodelphys halli (A, B, UWZM 22422)
and Metachirus nudicaudatus (C, D, AMNH 267009) illustrating taxonomic differences in several cranial
characters discussed in the text, including alternative patterns of contact among the interparietal (ip),
squamosal (sq), parietal (par), and mastoid (5 pars mastoideus of petrosal, mas). Additionally, the incisura
occipitalis (io), a distinct notch in the dorsal margin of the foramen magnum (fm), separates the right and
left exoccipitals in Lestodelphys. By contrast, the incisura occipitalis is absent because medial processes of
the right and left exoccipitals are in contact, excluding the supraoccipital from the dorsal margin of the
foramen magnum in Metachirus. Another conspicuous taxonomic difference illustrated in these views
concerns the paroccipital process (pp), which is a small, inconspicuous bony mass adnate to the mastoid in
Lestodelphys. The paroccipital process of Metachirus is much larger and projects almost straight ventrally.
The didelphid condition (a large interpari- median element fused to the supraoccipital
etal fused to the supraoccipital but not to the that is suturally distinct from the parietals
parietals) appears to be unique among and wedged between their posterior borders.
marsupials. In young dasyurids, for example, The interparietal appears to be completely
the interparietal is small and separated by missing in peramelemorphians, none of
open sutures from neighboring bones, or it is which show any trace of a bony element
absent; no examined dasyurid has a large wedged between the parietals anterior to the
phalangerids uniquely exhibit squamosal-

exoccipital contact (Flannery et al., 1987:
fig. 4C, D).
PALATE: The bony palate is variously
perforated by foramina and fenestrae that
exhibit considerable variation in occurrence,
size, and position among marsupials. Unfor-
tunately, inconsistent terminology has long
inhibited clear communication about taxo-
nomic differences. The nomenclature for
palatal perforations adopted herein is illus-
trated in figure 14 and compared with other
terminology in table 6.
The incisive foramina are prominent slots
in the anterior palate that transmit the
nasopalatine ducts (Sanchez-Villagra, 2001a)
together with nerves and blood vessels that
remain to be identified in marsupials (Wible,
2003). In didelphids, these openings extend
from the upper incisor arcade posteriorly to
or between the canines. Although didelphid
Fig. 14. Palatal morphology of Thylamys
incisive foramina are always bordered poste-
venustus (AMNH 261254) illustrating nomencla-
ture for fenestrae, foramina, and other features riorly and posterolaterally by the maxillary
described in the text. Dental loci (I1M4) provide bones, their dividing septum is formed
convenient landmarks for defining the size and primarily by the premaxillae. By contrast,
position of palatal structures. Abbreviations: if, the enormously elongated incisive foramina
incisive foramen; m, maxillary fenestra; mp, max- of caenolestids extend posteriorly between
illopalatine fenestra; p, palatine fenestra; plpf, the premolar rows (opposite P1 or P2), and
posterolateral palatal foramen. the posterior part of the dividing septum is
formed by the maxillae (Osgood, 1921: pl.
lambdoid crest, even as juveniles. The condi- 20). The long incisive foramina of peramele-
tion in caenolestids is uninterpretable because morphians do not extend behind C1, but the
no sutures persist on the posterodorsal posterior part of their dividing septum is
braincase of any examined specimen. likewise formed by the maxillae. Most other
The didelphid interparietal participates in marsupials (e.g., dasyuromorphians and Dro-
two alternative patterns of contact among miciops) have incisive foramina that essen-
bones of the posterior braincase and the tially resemble those of didelphids, with the
occiput. In most opossums the interparietal conspicuous exception of some diprotodon-
does not contact the squamosal because the tians (e.g., Macropus; Wells and Tedford,
parietal is in contact with the mastoid 1995: fig. 9E) in which these openings are
(fig. 13A, B). However, in Chironectes, Di- completely contained by the premaxillae.
delphis, Lutreolina, Philander, and in one Among Recent didelphids, only Caluromys
examined specimen of Metachirus (fig. 13C, and Caluromysiops consistently lack well-
D) the interparietal contacts the squamosal formed maxillopalatine fenestrae (figs. 38,
and prevents the parietal from contacting the 39). Although small perforations in the
mastoid. Two species of Monodelphis (M. maxillary-palatine sutures occur in most
brevicaudata and M. emiliae) exhibit both examined specimens of both genera, these
conditions with approximately equal fre- are obviously vascular openings (perhaps
quency among the specimens we examined. homologous with the major palatine foram-
Most examined nondidelphid marsupials ina of placental mammals) that do not
exhibit parietal-mastoid contact, but acroba- resemble the nonvascular fenestrae of other
tids and Tarsipes exhibit broad interparietal- didelphids. Glironia is sometimes said to lack
squamosal contact (Parker, 1890: fig. 2), and palatal fenestrae (e.g., by Archer, 1982; Reig
TABLE 6
Terminology for Marsupial Palatal Foramina and Fenestrae
This report Other names
incisive foramen anterior palatal foramen (Tate, 1933), anterior palatine vacuity (Osgood, 1921),
palatine fissure (Rowe et al., 2005), premaxillary vacuity (Archer, 1976a, 1981)
maxillary fenestra mesolateral fenestra (Creighton, 1984)
maxillopalatine fenestra major palatine foramen (Wible, 2003), maxillary vacuity (Archer, 1976a, 1981),
maxillopalatine or mesolateral vacuity (Hershkovitz, 1992b, 1997), palatine
fenestra (Clemens et al., 1989), palatine foramen (Novacek, 1993), posterior
palatine vacuity (Osgood, 1921)
palatine fenestra palatine vacuity (Archer, 1976a, 1981), posteromedial fenestra (Creighton, 1984),
posteromedial or palatine vacuity (Hershkovitz, 1992b, 1997)
posterolateral palatal foramen lateral posterior palatal foramen (Tate, 1933), minor palatine foramen (Wible,
2003), palatine canal (Clemens et al., 1989), posterolateral fenestra (Creighton,
1984), postero-lateral foramen (Osgood, 1921), postero-lateral palatine foramen
(Archer, 1976a), postpalatine foramen (Marshall et al., 1995; Muizon, 1998)
et al., 1987), but distinct maxillopalatine species of Marmosops (e.g., M. pinheiroi;

openings are present in most adult specimens fig. 53), and most species of Marmosa. The
that we examined (e.g., INPA 2570; fig. 37). single intact skull of Chacodelphys has
When present, the left and right maxillopa- palatine fenestrae, but these are incompletely
latine fenestrae of didelphids are invariably separated from the maxillopalatine openings;
separated by a broad median septum, and the normal morphology for this taxon
they never extend posteriorly behind the remains to be determined. Palatine vacuities
molar rows. Maxillopalatine fenestrae are are less widely distributed than maxillopala-
very widely distributed among other marsu- tine openings among other marsupials, but
pials, but they are absent in a few taxa (e.g., they occur in some dasyurids (e.g., Sminthop-
Myrmecobius, Dactylopsila, Phascolarctos). sis; Archer, 1981: fig. 5A), some peramele-
Among nondidelphid marsupials that possess morphians (e.g., Isoodon; Lyne and Mort,
these openings, the right and left fenestrae are 1981: fig. 9), and a few diprotodontians (e.g.,
often confluent because no median septum is Phascolarctos, Vombatus).
developed (as in Rhyncholestes and Perameles Most didelphids with fenestrated palates
gunnii), or they may extend posteriorly well have only maxillopalatine or maxillopalatine
behind the molar rows (as in acrobatids and and palatine openings. A few species, how-
some burramyids). ever, have additional fenestrae that are
Many didelphids have, in addition to located in the maxillary bone between the
maxillopalatine fenestrae, separate openings maxillopalatine fenestra and the toothrow (at
in the posterior palate that are entirely the level of M1 or M2) on each side of the
contained within the palatine bones. Palatine palate. Maxillary fenestrae are normally
fenestrae are consistently present in adult present in Chacodelphys, Gracilinanus, many
specimens of Cryptonanus, Didelphis, Graci- species of Thylamys (e.g., T. pusillus), and in
linanus, Lutreolina, Philander, Thylamys, Tlacuatzin canescens. Most other didelphids
some species of Marmosa, and some species lack maxillary vacuities except as rare (usu-
of Marmosops (e.g., M. creightoni; fig. 52). ally unilateral) variants. These palatal open-
Presence is also the modal condition for other ings are not known to occur in nondidelphid
taxa (e.g., Lestodelphys) in which individuals marsupials.
without palatine perforations are uncommon The posterolateral palatal foramen perfo-
variants (Martin, 2005). By contrast, palatine rates the maxillary-palatine suture behind
fenestrae are consistently absent in Calur- M4 or posterolingual to M4 on each side of
omys, Caluromysiops, Chironectes, Glironia, the skull in most metatherians. According to
Hyladelphys, Metachirus, Monodelphis, some Archer (1976a), this foramen transmits the
minor palatine artery from the maxillary

artery to the ventral surface of the palate. In
didelphids, caenolestids, Dromiciops, perame-
lemorphians, and stem metatherians, this
foramen is completely surrounded by bone,
but the foramen is incomplete or absent in
many dasyurids (e.g., Murexia and Sminthop-
sis) because the maxillary and palatine
processes that form the posterior border of
this opening in other taxa fail to ossify
(Wroe, 1997). In most didelphids the pos-
terolateral palatal foramina are small and
located behind the fourth upper molar, but
these openings are conspicuously larger and
extend lingual to the protocone of M4 on
each side of the skull in Thylamys (fig. 54)
and most specimens of Lestodelphys (fig. 51).
Two alternative morphologies of the pos-
terior palate can be recognized among didel-
phids. In Caluromys, Caluromysiops, and
Glironia, the posterior palate slopes ventrally
without any abrupt inflection, the caudal
palatal margin is usually broadly arched, and
prominent lateral corners are not developed;
behind the palate, the choanae are not
strongly constricted (fig.15A). By contrast,
the posterior palate of most other didelphids
is abruptly inflected ventrally, and the caudal
palatal margin is more or less straight with
prominent lateral corners; behind the palate,
the choanae are strongly constricted (fig.
15B). Most nondidelphid marsupials have
posterior palates that do not conform strictly
to either of these morphotypes, but they more
closely resemble those of Caluromys in
lacking well-developed lateral corners.
NASOPHARYNGEAL AND MESOPTERYGOID
REGION: The roof of the nasopharyngeal
passageway (or meatus; Rowe et al., 2005) is
formed by the vomer, the palatines, and the
presphenoid, of which only the palatines and
the presphenoid are usually exposed to
ventral view. Usually concealed inside the
Fig. 15. Ventral midcranial view of Caluromys nasopharyngeal passageway, the vomer is
philander (A, AMNH 267002) and Marmosa deeply divided by the presphenoid into paired
demerarae (B, AMNH 266428) illustrating taxo-
nomic differences in palatal morphology. In r
caluromyines the posterior palate slopes gently
ventrally, and the palatal margin is arched and the palatal margin is more or less straight with
(concave posteriorly) without strongly projecting projecting lateral corners (arrows in lower panel);
lateral corners; the internal nares (or choanae, in) the internal nares are narrow. Other abbreviations:
are very broad. In didelphines, however, the als, alisphenoid; bs, basisphenoid; max, maxillary;
posterior palate is abruptly inflected ventrally pal, palatine; ps, presphenoid; pt, pterygoid.
lateral processes that only occasionally ex- The openings through which the mandib-
tend posteriorly into the mesopterygoid fossa ular division of the trigeminal nerve (V3) exits
to contact the pterygoids (e.g., in Caluromy- the skull have been variously named by
siops; fig. 39). Most other marsupials, how- systematists. Following Gaudin et al.
ever, exhibit different configurations of the (1996), we use the term foramen ovale for
nasopharyngeal roofing bones. In particular, the primary orifice through which V3 exits
the vomer is undivided and extends caudally the endocranial lumen of the adult skull. In
to underlie the presphenoid (concealing most most didelphids, the foramen ovale is bor-
or all of it from ventral view) in caenolestids, dered by the alisphenoid and the petrosal
dasyurids, peramelemorphians, and Dromi- (fig. 16A), but the foramen is sometimes
ciops. contained entirely within the alisphenoid,
The left and right pterygoid bones are and both conditions can be seen on opposite
widely separated by the presphenoid and the sides of the same skull (Gaudin et al., 1996:
basisphenoid in didelphids. In other marsu- fig. 6). Because it is often difficult to
pials, the vomer may also extend between the determine the position of this opening with
left and right pterygoids, but in Dromiciops respect to relevant endocranial sutures, how-
the pterygoids are in midline contact, under- ever, we did not score the position of the
lying the presphenoid and concealing its foramen ovale per se. Instead, secondary
suture with the basisphenoid from ventral enclosures of V3 by outgrowths of the
view; the conjoined pterygoids of Dromiciops alisphenoid tympanic process provide a more
form a strong sagittal keel, which is contin- accessible basis for taxonomic comparisons.
uous with that formed anteriorly by the Two different (apparently nonhomologous)
vomer and posteriorly by the basisphenoid.10 conditions of secondary nerve enclosure
BASICRANIAL FORAMINA: A foramen that occur among didelphids.
transmits the venous transverse canal perfo- In juveniles and adults of Gracilinanus,
rates the alisphenoid anterolateral to the Lestodelphys, Marmosops, Metachirus, and
carotid canal in most didelphids (Sanchez- Thylamys, the extracranial course of V3 is
Villagra, 2001b; Sanchez-Villagra and Wible, enclosed by an alisphenoid process or strut
2002). Caluromys and Caluromysiops, how- that arises from the anteromedial surface of
ever, consistently lack a transverse canal the bulla and extends anteriorly, medially,
foramen (Archer, 1976a; Sanchez-Villagra and dorsally to span the transverse canal
and Wible, 2002; Voss and Jansa, 2003). foramen. In the smaller species with this
According to Kirsch and Archer (1982: condition (e.g., Marmosops pinheiroi; fig.
character 24), the transverse canal is 16B) the extracranial course of V3 remains
absent in Metachirus nudicaudatus and Phi- unenclosed between this process and the
lander opossum, but most of the specimens we primary foramen ovale, but in larger species
examined of both taxa had a transverse canal (e.g., Marmosops noctivagus) a sheet of bone
foramen on each side of the skull. The produced from the posterior edge of the
transverse canal foramen is normally present process extends caudally to form a more or
in most other marsupials (e.g., caenolestids, less complete canal late in postnatal life.
dasyurids, peramelemorphians, and Dromi- An alternative pattern of secondary fora-
ciops), although its reduction or absence in a men and canal formation is seen in Calur-
few taxa (e.g., Planigale, Tarsipes) has been omysiops, Chironectes, Didelphis, Lutreolina,
noted by authors (Archer, 1976a; Aplin, Philander, and some species of Monodelphis
1990; Sanchez-Villagra and Wible, 2002). (e.g., M. theresa). In these taxa, V3 is broadly
enclosed by an alisphenoid lamina (fig. 16C)
10
The obviously autapomorphic midventral basicranial keel that extends along the posteromedial bullar
of Dromiciops is usually described as formed by the presphenoid surface, but there is no anteromedial process
and the basisphenoid (e.g., by Hershkovitz, 1999; Giannini et spanning the transverse canal foramen. Many
al., 2004), but the presphenoid does not participate in forming juvenile specimens of some taxa scored with
this structure in any specimen that we examined. Note that the
presphenoid is incorrectly labeled as the vomer in some
this condition (e.g., Didelphis) do not have a
illustrated ventral views of the didelphid skull (e.g., Novacek, fully enclosed secondary foramen and canal,
1993: fig. 9.4B). but they usually show some laminar devel-
Fig. 16. Ventral view of left ear region in Marmosa murina (A, AMNH 267368), Marmosops pinheiroi
(B, AMNH 267346), and Monodelphis theresa (C, MVZ 182775) illustrating taxonomic differences in
secondary foramen ovale formation. In Marmosa, the mandibular branch of the trigeminal nerve (V3,
reconstructed course shown by heavy arrow) emerges from the endocranial lumen via the foramen ovale
opment; subadults and young adults show floor of the middle ear, but this structure
progressively more complete enclosure; and exhibits significant taxonomic variation in
old adults (large specimens with heavily worn size. The alisphenoid tympanic process is
teeth) usually have completely enclosed large and extends far enough posteriorly to
foramina and canals (Abdala et al., 2001). closely approximate or contact the rostral
All other didelphid taxa (Caluromys, Chaco- tympanic process of the petrosal in Calur-
delphys, Cryptonanuns, Glironia, Hyladelphys, omys and Caluromysiops (see Reig et al.,
Marmosa, Tlacuatzin, and most species of 1987: fig. 44B, D). In other opossums, a
Monodelphis) lack secondary enclosures of distinct gap separates the alisphenoid tym-
the mandibular nerve except as rare (usually panic process from the rostral tympanic
unilateral) variants. process of the petrosal, such that at least
Although many nondidelphid marsupials part of the floor of the middle ear is
(e.g., caenolestids, Dromiciops, most dasyur- membranous (e.g., in Marmosa and Mono-
omorphians) have only a primary foramen delphis; Reig et al., 1987: fig. 41B, D). In
ovale, secondary foramina ovales are also Lestodelphys and Thylamys a rather indis-
widely distributed. Unfortunately, these traits tinct medial process of the ectotympanic
remain to be widely surveyed, and inconsis- makes a small contribution to the floor of
tent terminology (discussed by Gaudin et al., the middle ear, partially filling in the gap
1996) makes it difficult to interpret some between the tympanic processes of the
published descriptions of marsupial basicra- alisphenoid and petrosal (Reig et al., 1987:
nia. Although other developmental mecha- fig. 42B, D). The roof of the hypotympanic
nisms have obviously been responsible for sinus in all Recent didelphids is almost
secondary enclosures of the mandibular exclusively formed by the alisphenoid, with
nerve in some taxa (e.g., Phascolarctos), the petrosal making only a small, often
putative homologues of both didelphid pat- negligible, contribution. The didelphid squa-
terns of secondary foramen formation can be mosal does not participate in forming any
recognized among certain Old World marsu- part of the floor or roof of the middle ear
pials. At least some specimens of Thylacinus, cavity.
for example, have a secondary foramen Caenolestids, dasyurids, peramelemorphi-
formed by an anteromedial bullar strut that ans, and Dromiciops resemble didelphids in
spans the transverse canal foramen, whereas that the alisphenoid forms most of the
the secondary foramen ovale of Recent anterior part of the floor of the middle ear.
peramelemorphians (e.g., Echymipera, Pera- However, whereas caenolestids and some
meles) is formed by a broad medial bullar peramelemorphians (e.g., Echymipera) have
lamina. a small alisphenoid tympanic process that
EAR REGION: All didelphids have an does not contact the rostral tympanic process
osteologically well-defined middle ear cavity of the petrosal, the alisphenoid tympanic
or hypotympanic sinus (sensu van der process is large and broadly contacts the
Klaauw, 1931: 19). A cup-shaped tympanic petrosal in dasyurids, Dromiciops, and other
process (or wing) of the alisphenoid peramelemorphians (e.g., Perameles). The
invariably forms the anterior part of the hypotympanic sinus roof of dasyurids and
(fo), which is bordered by the alisphenoid (als) and the petrosal (pet); the extracranial course of the nerve is
unenclosed in this taxon. In Marmosops, however, the extracranial course of V3 is partially enclosed by a
bony strut (st) that extends from the anteromedial surface of the alisphenoid tympanic process (atp) across
the transverse canal foramen (tcf); the nerve then emerges from a so-called secondary foramen ovale.
Another kind of secondary enclosure is seen in Monodelphis, where the nerve emerges from a secondary
foramen ovale formed by a medial lamina (lam) of the alisphenoid tympanic process. Other abbreviations:
bs, basisphenoid; bo, basioccipital; cc, carotid canal; ect, ectotympanic. Scale bars 5 5 mm.
Dromiciops is almost exclusively formed by concave process that forms a substantial part
the alisphenoid (as in didelphids), but a of the bullar floor in these taxa.
broad petrosal shelf forms part of the sinus In most didelphids, the fenestra cochleae is
roof in caenolestids and some peramelemor- exposed in lateral or ventrolateral view as
phians. Diprotodontians exhibit a very wide illustrated for Didelphis by Wible (1990: fig.
range of bullar morphologies (reviewed by 1). In others (e.g., Caluromys, Caluromysiops,
Aplin, 1990), many of which include sub- Lestodelphys, Marmosa rubra, Monodelphis
stantial squamosal participation. emiliae, and Thylamys), the fenestra cochleae
The epitympanic recess of didelphids is a is concealed within a chamber or sinus
shallow concavity in the ventral surface of formed by a laminar outgrowth of the pars
the petrosal pars canalicularis that encloses canalicularis that approximates or contacts
the mallear-incudal articulation (Wible, 2003: the rostral tympanic process of the pars
fig. 7D).11 Although this morphology is cochlearis. This lamina, which appears to be
widespread among other marsupials, a few homologous with the caudal tympanic pro-
(e.g., Macrotis, Perameles) have an enor- cess of MacPhee (1981), is invariably fused
mously inflated epitympanic recess that posteroventrally with the paroccipital process
vaults high above the ossicles, leaving the of the exoccipital, which also throws forward
mallear-incudal joint effectively unenclosed a small tympanic outgrowth. The resulting
by bone. auditory chambercorresponding to the
The promontorium of the petrosal pars periotic hypotympanic sinus of Archer
cochlearis is not marked by any vascular (1976a) and the post-promontorial tympan-
grooves in didelphids. Instead, the most ic sinus of Wible (1990)is always small
conspicuous feature of this otherwise and uninflated in opossums.
smooth, bulbous structure is the rostral The fenestra cochleae is also exposed in
tympanic process (sensu MacPhee, 1981; caenolestids and some peramelemorphians
Wible, 1990). The rostral tympanic process (e.g., Echymipera), in which tympanic pro-
is small, more or less conical, and perhaps cesses of the petrosal are either small or
serves only to anchor the posterior (ventral) absent. In other peramelemorphians (e.g.,
limb of the ectotympanic annulus in some Perameles) and in dasyurids, however, the
opossums (e.g., Glironia), but it is anteropos- fenestra cochleae is concealed within a
teriorly expanded, dorsally concave, and postpromontorial sinus formed by the seam-
substantially contributes to the ventral en- less fusion of the rostral and caudal tympanic
closure of the middle ear in others (e.g., processes; in these forms (as in Caluromys
and other didelphids with similar petrosal
Marmosa; Reig et al., 1987: figs. 3842). In
features), the sinus is small and uninflated.
Caluromys and Caluromysiops, the rostral
The fenestra cochleae of Dromiciops is
tympanic process extends along the entire
enclosed in an inflated (bubblelike) postpro-
length of the promontorium from the fenes-
montorial sinus that, although strikingly
tra cochleae to the anterior pole (Sanchez-
unlike the smaller sinuses of other plesio-
Villagra and Wible, 2002). The rostral
morphic marsupials, is also assumed to have
tympanic process of caenolestids and most
resulted from fusion of caudal and rostral
peramelemorphians is small and more or less
tympanic processes (Sanchez-Villagra and
conical, but dasyurids, some peramelemor-
Wible, 2002: fig. 11).
phians (e.g., Macrotis), and Dromiciops have
No other bone (besides the alisphenoid,
an anteroposteriorly extensive and dorsally
ectotympanic, and petrosal) participates in
11 auditory sinus formation among Recent
Because the illustrations in Wible (1990) are so useful in
other respects, it is important to note that the epitympanic didelphids, which have only one or two
recess is mislabeled in two views of the tympanic surface of the hypotympanic cavities as described above.
petrosal of Didelphis virginiana. As noted by Muizon (1998: Caenolestids and Dromiciops are similar to
90), the concavity so labeled in Wibles figure 4A and 4C is didelphids in this respect, but Recent dasyur-
really the posterior part of the hypotympanic sinus roof; the true
epitympanic recess is a smaller depression posterolateral to the
omorphians and some peramelemorphians
hypotympanic sinus roof and separated from it by a low ridge, (e.g., Macrotis and Perameles) have a well-
the petrosal crest. developed epitympanic sinus formed by the
Fig. 17. Oblique ventrolateral view of left ear region in Marmosops impavidus (A, MUSM 13284) and
Philander mcilhennyi (B, MUSM 13299) illustrating taxonomic differences in ectotympanic suspension.
Whereas the ectotympanic (ect) is suspended from the skull by attachments both to the petrosal (pet) and
to the malleus (mal) in Marmosops, the ectotympanic of Philander is suspended only from the malleus
(there is no attachment to the petrosal). Other abbreviations: atp, alisphenoid tympanic process; pro,
promontorium (of petrosal); rtp, rostral tympanic process (of petrosal); sq, squamosal.
squamosal. This structure is a cup-shaped where the squamosal, alisphenoid, and pe-
cavity that, in all skulls retaining vestiges of trosal are juxtaposed behind the postglenoid
auditory soft tissues, appears to be covered process. Where the connection can be seen
by the membrana Shrapnelli (pars flaccida of clearly (dried remnants of soft tissues fre-
the tympanic membrane). Archer (1976a: quently obscure this feature), the actual
304) remarked that, [i]n some didelphids attachment seems to be to the petrosal (fig.
(e.g., Metachirus) there is a depression in the 17A). In six genera (Caluromys, Caluromy-
squamosal which is clearly the homologue of siops, Chironectes, Didelphis, Lutreolina, and
this sinus, but nothing resembling the Philander), however, the anterior limb of the
squamosal epitympanic sinus of dasyurids ectotympanic is not directly attached to the
occurs in any living opossum. skull, and the suspension is indirect, via the
ECTOTYMPANIC AND OSSICLES: Although anterior (tympanic) process of the malleus
didelphids were described by van der Klaauw (fig. 17B). In all didelphids with an indirect
(1931: 26) as having a completely free dorsal connection between the ectotympanic
ectotympanic, two distinct patterns of ecto- and the skull, the tympanic annulus is more
tympanic attachment can be recognized in or less ringlike because the posterior (ventral)
the family. In most didelphids the anterior limb is not expanded to form part of the floor
limb (or crus) of the ectotympanic annulus is of the middle ear cavity. By contrast, in some
directly connected to the skull near the point taxa with direct ectotympanic suspension, the
posterior limb tends to be dorsoventrally certain details of didelphid mallear and

flattened and laterally expanded, forming incudal structure were noted by Sanchez-
part of the floor of the external ear canal to Villagra et al. (2002) and Schmelzle et al.
a greater or lesser extent.12 The anterior and (2005), we did not attempt the requisite
posterior limbs of the didelphid ectotympanic dissections to assess the phylogenetic distri-
are always widely separated by a broad bution of the features they described. The
posterodorsal gap, the incisura tympanica. didelphid stapes, however, exhibits striking
Caenolestids, dasyurids, most peramele- morphological variation that can be deter-
morphians, and Dromiciops essentially re- mined without dissection.
semble didelphids in most aspects of ecto- Despite some intraspecific variation noted
tympanic morphology. The anterior limb is by Gaudin et al. (1996), most didelphids
suspended loosely but directly from the normally exhibit one or the other of two
basicranium, and it is separated from the different stapedial morphotypes defined by
posterior limb by a broad posterodorsal Novacek and Wyss (1986). Most didelphids
incisura tympanica. However, whereas the (e.g., Philander; Novacek and Wyss, 1986:
ectotympanic annulus in didelphids, caeno- fig. 5F) have a more or less triangular or
lestids, dasyurids, and peramelemorphians is stirrup-shaped, bicrurate stapes that is perfo-
prominently exposed on the lateral aspect of rated by a large stapedial (obturator or
the bulla, the ectotympanic is uniquely intercrural) foramen. Other didelphids have
concealed from lateral view within the a columelliform (or columnar) stapes that is
auditory bulla of Dromiciops (see Segall, imperforate (or microperforate: with a fora-
1969a: fig. 3, 1969b: fig. 4). The ectotympa- men whose maximum diameter is less than
nics of Recent diprotodontians differ strik- the width of a surrounding crus; Gaudin et
ingly from those of didelphids and other al., 1996). Taxa with columelliform stapes
marsupials, principally by forming tubelike include Caluromysiops, Lestodelphys, and
extensions of the ear canal, by ankylosis or several species of Monodelphis (e.g., M.
fusion with neighboring bones, and by dorsal peruviana, M. theresa).
migration and narrowing or closure of the The ossicular morphology of nondidelphid
incisura tympanica (Aplin, 1990). marsupials exhibits significant taxonomic
variation. Whereas the mallei of caenolestids,
The outermost auditory ossicles of didel-
dasyurids, and peramelemorphians essential-
phids conform to a pattern that is widespread
ly resemble those of didelphids in the features
among other plesiomorphic marsupials. In
described above, the malleus of Dromiciops
particular, the malleus invariably has a long,
has a short, uninflected neck, no lamina, and
sharply inflected neck, a small orbicular
a manubrium that forms an open angle with
apophysis, a well-developed lamina, and a
the anterior process (Segall, 1969a, 1969b).
manubrium that is approximately parallel to
The stapes of Dromiciops is bicrurate (with a
the anterior (tympanic or gracile) process; the large foramen, like didelphids), but the stapes
body of the incus is distinctly differentiated of caenolestids, dasyurids, and peramelemor-
from the long and short processes of that phians are columelliform and imperforate
bone (Segall, 1969b: figs. 7, 9; Henson, 1974: (Novacek and Wyss, 1986; Schmelzle et al.,
fig. 20A). Although taxonomic differences in 2005). Notoryctes and phalangeriform dipro-
todontians have ossicles that are strikingly
unlike those of didelphids and other plesio-
12
We were previously incorrect in describing the expansion morphic marsupials (Segall, 1970; Sanchez-
of the posterior limb of the ectotympanic as medial in taxa Villagra and Nummela, 2001).
with direct suspension (Voss and Jansa, 2003: 30). As noted by
Wible (2003: 165), the expansion of the posterior limb of the OCCIPUT: Fusion of the supraoccipital
ectotympanic in Monodelphis is lateral to the sulcus tympanicus, with the interparietal is common to all
and therefore contributes to the floor of the external ear canal didelphids and occurs very early in postnatal
(acoustic meatus) rather than to the bullar floor. This is also true life as has already been discussed, so the
of most other didelphids with direct ectotympanic suspension.
Only in Chacodelphys, Cryptonanus, Gracilinanus, Lestodelphys,
dorsal part of the lambdoid crest is not
Thylamys, and some species of Marmosa does the ectotympanic associated with any visible sutures, nor does
also make a small bullar contribution. it develop sesamoids for the insertion of
nuchal muscles (as in peramelemorphians, see Among other plesiomorphic marsupials,

below). The crest itself is only weakly dasyurids are essentially similar to didelphids
developed in juveniles and in adult specimens in most occipital features, but mastoid
of the smaller forms, but it is prominent in processes are indistinct in caenolestids, and
fully adult specimens of all of the larger Dromiciops lacks paroccipital processes. Per-
opossums. The mastoid (pars mastoideus of amelemorphians differ from all other marsu-
the petrosal) is always prominently exposed pials in having unique osseous structures that
on the lateral occiput, where it is bordered by are sutured between the parietals and the
the supraoccipital, exoccipital, squamosal, supraoccipital on each side of the skull.
and sometimes the parietal. All didelphids Commonly illustrated but seldom labeled or
have a well-developed mastoid process at the described by authors, these are the lambdoid
ventrolateral corner of the occiput that is sesamoids of Filan (1990), who hypothesized
sometimes provided with a sesamoid. that they function as force multipliers for M.
The dorsal margin of the foramen mag- rectus capitis dorsalis and other head exten-
num is formed by the exoccipitals and the sors. Another occipital feature that might be
supraoccipital in most didelphids (fig. 13A). unique to peramelemorphians is a sesamoid
In taxa that conform to this morphology, the that is probably associated with the origin of
foramen has a distinct middorsal emargina- the posterior belly of the digastric; this small
tion, the incisura occipitalis, that separates bone occupies the apex of the paroccipital
the left and right exoccipitals. By contrast, process in some peramelemorphians (e.g.,
medial processes of the left and right Perameles), but it attaches to the otherwise
exoccipitals are joined at the midline above smooth ventrolateral corner of the exoccipi-
the foramen magnumexcluding the supra- tal in taxa that lack a paroccipital process
occipital from the dorsal margin of the (e.g., Microperoryctes).
foramen and occluding the incisura (fig. MANDIBLE: The didelphid mandible con-
13C)in adult specimens of Didelphis, Lu- sists of an anteroposteriorly elongate hori-
treolina, Metachirus, and Philander. Juvenile zontal ramus that contains the dental alveoli,
(and some subadult) specimens of these an ascending ramus with well-developed
genera, however, resemble other didelphids coronoid and condylar processes, and a
in their occipital morphology, such that both posteroventral angular process. The mandib-
conditions (presence and absence of a supra- ular symphysis is never fused, and the
occipital margin) can be observed in ontoge- retromolar space (between m4 and the base
netic series of conspecific skulls (Abdala et of the coronoid process) is either imperforate
al., 2001: fig. 3). The medial processes of the or pierced by tiny nutrient foramina of
exoccipitals are closely approximated in inconstant number and position. The masse-
Chironectes, but in all of the specimens we teric fossa (a prominent concavity for the
examined the supraoccipital still forms part insertion of the eponymous muscle on the
of the dorsal margin of the foramen magnum posterolateral surface of the jaw) is always
(contra Sanchez-Villagra and Wible, 2002), imperforate, and its posteroventral border is
and the incisura occipitalis persists as a defined by a distinct shelf. The articular
narrow middorsal notch. condyle is transversely elongate and more or
The paroccipital process of the exoccipital is less semicylindrical. Although substantial var-
inconspicuous in most didelphids, consisting iation in mandibular shape is apparent from
of a low, rounded or subtriangular bony mass the drawings that accompany our systematic
for muscle attachment that is broadly adnate accounts (figs. 3754), other features are more
to the petrosal; the indistinct apex of the readily characterized for taxonomic diagnoses
process in taxa that conform to this wide- and phylogenetic inference.
spread condition is directed posteroventrally Most didelphids have two mental foram-
(fig. 13A, B). By contrast, in Chironectes, ina that perforate the lateral surface of the
Didelphis, Lutreolina, Metachirus, and Phi- horizontal ramus, of which the more anterior
lander the paroccipital process is much larger, is usually larger and located below p1 or p2,
more erect, and points almost straight ventrally and the more posterior is usually smaller and
in most examined specimens (fig. 13C, D). located below p3 or m1. Chironectes, howev-
er, has only a single mental foramen, which is canines, double-rooted premolars and lower
very large and occupies the same position as molars, and three-rooted upper molars; appar-
the anterior foramen in those taxa with two ently, dP3 always has three roots and dp3 has
lateral mandibular perforations. two. Most of the following remarks therefore
Alternative states of the form and orien- concern crown morphologies, which afford
tation of the marsupial angular process were greater scope for taxonomic comparisons.
defined by Sanchez-Villagra and Smith UPPER INCISORS: The didelphid first upper
(1997), but we are unable to consistently incisor (I1) is a more or less styliform,
recognize the distinction between the rod- somewhat procumbent, hypsodont tooth that
like and intermediate conditions they is conspicuously unlike I2I5, from which it
scored for didelphid exemplars. Instead, the is always separated by a small diastema.
common didelphid condition seems more Although taxonomic variation among opos-
appropriately characterized as acute and sums in I1 hypsodonty and diastema forma-
strongly inflected (Voss and Jansa, 2003: tion has been suggested by authors (e.g.,
34). Caluromys and Caluromysiops are the Takahashi, 1974; Archer, 1976b; Creighton,
only didelphids with bluntly rounded and 1984; Wroe et al., 2000), diagnostically useful
weakly inflected mandibular angles. differences are more easily recognized among
Most other plesiomorphic marsupials have the posterior teeth of this series.
mandibles that grossly resemble the common The crowns of I2I5 in most didelphids are
didelphid morphotype, but Dromiciops has approximately symmetrical, labiolingually
only one mental foramen and some perame- compressed rhomboids with subequal anteri-
lemorphians (e.g., Perameles gunnii) often or (mesial) and posterior (distal) cutting edges
have three mental foramina. Additionally, that converge to form a sharp central apex;
Caenolestes has a small foramen in the usually, a distinct anterior angle (mesiostyle)
retromolar space that Sanchez-Villagra et and a posterior angle (distostyle) can be seen
al. (2000) called the retrodental canal (after on unworn teeth (fig. 18A). This is the
Hoffstetter and Villaroel, 1974), and it has a morphology that Takahashi (1974: 414)
small masseteric foramen near the ventral termed premolariform, as exemplified by
margin of the masseteric fossa (Osgood, Monodelphis, Marmosa, and Metachirus in
1921), the lumen of which communicates her study. In many taxa with posterior upper
with that of the much larger mandibular incisors of this type, the tooth crowns increase
foramen on the medial surface of the jaw. in length (anteroposterior or mesiodistal
dimension) from front to back, such that I2
Dentition appears visibly smaller than I5 in labial view.
This tendency is very pronounced in some
All didelphids normally have 50 teeth that taxa with premolariform incisors (e.g., Mar-
most authors now recognize as including I1 mosops, Metachirus) but seems to grade
5, C1, P13, and M14 in the upper dentition imperceptibly (via intermediate morpholo-
and i14, c1, p13, and m14 in the lower gies) to the essentially uniform toothrows of
dentition. As in other metatherians, only the other taxa (e.g., Lestodelphys, Tlacuatzin) in
third upper and lower premolars have which I2 and I5 are subequal in crown length.
deciduous precursors (Flower, 1867; Luckett, An alternative morphology, in which the
1993; Luo et al., 2004; van Nievelt and Smith, crowns of I2I5 are conspicuously asymmet-
2005a, 2005b). Despite this prevailing con- rical, with much longer anterior (mesial) than
sensus, marsupial tooth homologies have posterior (distal) cutting edges, occurs in
long been controversial, with the result that Caluromys, Caluromysiops, Didelphis, Gliro-
several alternative systems have been used to nia, Hyladelphys, Lutreolina, and Philander.
identify didelphid dental loci (table 7). On unworn teeth of these taxa, the anterior
As far as we have been able to determine angle (mesiostyle) is more frequently distinct
without dissections or X-rays (i.e., from than is the posterior angle (distostyle), which
external inspection of teeth in situ, and from is often entirely lacking on I5 (fig. 18B). This
the alveoli of specimens with loose teeth), all is the morphology that Takahashi (1974)
didelphids have single-rooted incisors and labeled incisiform, which she recorded for
TABLE 7
Alternative Hypotheses of Marsupial Tooth Homologies
This studya Thomas (1887)b Archer (1984)c Hershkovitz (1992b)d Luckett (1993)e Goin (1997)f
P1 P1 P1 P1 dP1 P1
P2 P3 P2 P2 dP2 P2
dP3 dP4 M1 M1 dP3 M0
P3 P4 P3 P3 P3 P3
M1 M1 M2 M2 M1 M1
M2 M2 M3 M3 M2 M2
M3 M3 M4 M4 M3 M3
M4 M4 M5 M5 M4 M4
i1 i1 i1 i2 i1 i2
i2 i2 i2 i3 i2 i3
i3 i3 i3 i4 i3 i4
i4 i5 i4 i5
p1 p1 p1 p1 dp1 p1
p2 p3 p2 p2 dp2 p2
dp3 dp4 m1 m1 dp3 m0
p3 p4 p3 p3 p3 p3
m1 m1 m2 m2 m1 m1
m2 m2 m3 m3 m2 m2
m3 m3 m4 m4 m3 m3
m4 m4 m5 m5 m4 m4
a
The system adopted here was used by Tate (1933) and is currently followed by most metatherian researchers (e.g.,
Marshall et al., 1995; Rougier et al., 1998; Wroe et al., 2000; Voss and Jansa, 2003). Note that there is no substantive
inconsistency in the literature regarding the identification of I1I5, C1, or c1. In order to focus these comparisons on real
differences in assumptions about dental homologies, semantically equivalent notations used by authors (e.g., I2 for i2,
pm3 for P3) have been modified as necessary to conform with our usage.
b
Also Thomas (1888), who assumed that marsupials primitively had four premolars, of which the second was lost.
c
Based on the assumption that the replaced teeth are M1/m1 (Archer, 1978), this is the nomenclature followed
throughout most of the marsupial literature for the next 10 years (e.g., Reig et al., 1987).
d
Based on the assumption that marsupials primitively had five lower incisors, of which the first was lost (Hershkovitz,
1982). Note that whereas Hershkovitz (1992b) accepted Archers (1984) system of postcanine homologies, Hershkovitz
(1997, 1999) did not.
e
The only difference between Lucketts (1993) notation and ours concerns his designation of the anterior premolars as
first-generation (formerly deciduous) teeth. Although we do not dispute his interpretation of the developmental data at
hand, we think it confusing to use deciduous notation for unreplaced teeth.
f
Goins (1997) system was based on Hershkovitzs (1982) theory that the ancestral first lower incisor of marsupials is
missing in living taxa, and on Archers (1978) conjecture that replaced teeth in the upper and lower dentition are molars;
by designating these as M0/m0, Goin intended to preserved traditional notation for the permanent molariform teeth.
Chironectes in addition to Caluromys and Most peramelemorphians and Dromiciops

other taxa in her study. However, the unworn resemble didelphids in having five upper
juvenile dentitions of Chironectes that we incisors, but I5 is missing in dasyurids,
examined have approximately symmetrical, caenolestids, Echymipera, and Rhyncho-
rhomboidal crowns. There is a tendency meles.13 The first upper incisor is styliform
(more marked in Glironia than in other 13
Following Thomas (1887), we assume that upper teeth are
genera of this category) for I2I5 to decrease lost from (or added to) the posterior (distal) end of the incisor
in crown length from front to back, such that row. Although there appears to be no developmental data to
I2 is sometimes visibly larger than I5 in labial support this notion, we observed that I5 is small and occupies a
view. Aspects of didelphid upper incisor separate alveolus from I2I4 (whose alveoli are usually at least
partially confluent) in some taxa with five upper incisors (e.g.,
morphology reported by authors whose obser- Glironia, Hyladelphys), and that the loss of such a tooth would
vations we have not been able to replicate were simply enlarge an already existing diastema between the incisors
discussed by Voss and Jansa (2003: 35). and C1 rather than opening a fresh gap in the dental arcade.
notched in Rhyncholestes (see Osgood, 1924:

figs. 3b, c; Bublitz, 1987: fig. 13). In
peramelemorphians, I24 are flat crowned
(like I1), but I5 (when present) is caniniform.
All of the upper incisors are single rooted in
dasyurids, caenolestids, Dromiciops, and
many peramelemorphians, but I5 is double
rooted in Perameles gunnii.
UPPER CANINE: The didelphid upper ca-
nine (C1) is invariably separated from the
posteriormost upper incisor by a wide
diastema that may or may not contain a
distinct paracanine fossa for the reception of
the lower canine. In most didelphids, C1
occupies the premaxillary-maxillary suture,
despite the fact that the maxilla always
extends anterolaterally beyond this tooth,
sometimes reaching the alveolus of I5;
therefore, the premaxilla only contacts the
C1 alveolus medially (fig. 14). In Caluromys
and Caluromysiops, however, the upper
canine alveolus is entirely contained within
the maxilla (figs. 38, 39).
The upper canine is a simple unicuspid
Fig. 18. Unworn premaxillary dentitions of tooth in most didelphids, but some species
Marmosops pinheiroi (A, AMNH 267341) and (e.g., Gracilinanus emiliae; Voss et al., 2001:
Lutreolina crassicaudata (B, AMNH 210422) fig. 12A) have a small posterior accessory
illustrating taxonomic differences in the shape of cusp, and others (e.g., Marmosops pinheiroi;
the incisor crowns. In Marmosops and many other fig. 53) have both anterior and posterior
small didelphines, the crowns of I2I4 are sym- accessory cusps. Some care is needed in
metrically rhomboidal, with subequal anterior
determining canine crown morphology, how-
(mesial) and posterior (distal) cutting edges that
converge to form a sharp central apex; a distinct ever, because accessory cusps are sometimes
posterior corner (distostyle) is always present on obliterated by wear (as inferred from their
I5. By contrast, in Lutreolina (and certain other absence in old adults of species that uniform-
taxa), the crowns of I2I4 are conspicuously ly exhibit such structures as juveniles and
asymmetrical, with longer anterior than posterior subadults). Alternatively, a false posterior
cutting edges; a distinct distostyle is usually absent accessory cusp is sometimes formed when the
on I5. Scale bars 5 1 mm. trailing edge of C1 is notched by occlusion
with p1. Unlike other teeth whose size does
not increase ontogenetically after the crown
in most of these taxa, but I1 is a laterally is fully erupted, didelphid canines often
compressed, distinctly chisellike tooth in appear much larger in older adults than in
caenolestids, and it is a labiolingually com- younger animals because the root is contin-
pressed, flat-crowned tooth in peramelemor- uously extruded from the alveolus through-
phians. The unworn posterior incisors have out life.
approximately rhomboidal crowns with sub- Most other marsupials have single-rooted
equal anterior and posterior cutting edges unicuspid upper canines like those of didel-
that converge to a sharp apex in most phids, but accessory cusps are present in
dasyurids and in Dromiciops, but I24 in some peramelids (e.g., Echymipera) and
caenolestids have strikingly asymmetrical dasyurids (e.g., Antechinomys; Archer,
crowns (appropriately described as hatch- 1976b). Additionally, the canine is double
et-shaped by Osgood, 1921: 117), the rooted and provided with accessory cusps in
posterior margins of which are deeply female Caenolestes and Rhyncholestes and in
between P1 and P2 tends to be individually

and ontogenetically variable, however, and a
continuous range of intermediates exist
between taxa in which it is usually absent
(e.g., Lestodelphys; fig. 51) and those in
which it is normally present (e.g., Philander;
fig. 48). By contrast, no diastema is ever
present between P2 and P3, the crowns of
which are in contact or closely approximated
in all examined opossum dentitions.
Although consistently smaller than the
posterior premolars, P1 in most didelphids
is at least half the height or width of P2,
which it essentially resembles in having a
large, laterally compressed, more or less
triangular central cusp that is usually flanked
by smaller anterior and posterior accessory
cusps. By contrast, P1 is a much smaller
(apparently vestigial) tooth that lacks distinct
occlusal features in Caluromys (fig. 38) and
Caluromysiops (fig. 39). In a few specimens
of the latter genus (e.g., AMNH 244364), P1
is missing.
The second and third upper premolars are
always large and well-developed teeth. Each
is surrounded by a broad basal cingulum in
Caluromys and Caluromysiops, but the basal
cingulum of P2 and/or P3 is often incomplete
or indistinct in other didelphids. The second
upper premolar is distinctly taller than P3 in
Fig. 19. Lateral views of left C1P3 in Hyla- Caluromys, Caluromysiops, and Hyladelphys
delphys kalinowskii (A, AMNH 267338), Gracili- (fig. 19A), but P3 is distinctly taller than P2
nanus agilis (B, AMNH 134005), and Thylamys in Chironectes, Cryptonanus, Didelphis, Les-
pusillus (C, AMNH 275445) illustrating taxonomic todelphys, Lutreolina, Monodelphis, Philan-
differences in the relative heights of P2 and P3
der, and Thylamys (fig. 19C). By contrast, P2
(toothrows are not drawn to the same scale).
and P3 are about the same height in
both sexes of Lestoros, whereas male Caeno- Chacodelphys,14 Glironia, Gracilinanus (fig.
lestes and Rhyncholestes have single-rooted 19B), Marmosa, Marmosops, and Metachirus.
unicuspid teeth (Osgood, 1924; Bublitz, When unworn, P3 is provided with sharp
1987). The position of C1 with respect to anterior and posterior cutting edgeseach of
the maxillary-premaxillary suture has not which extends from the cingulum to the apex
been widely surveyed in Marsupialia, but of the toothin Caluromys, Caluromysiops,
the upper canine alveolus is entirely con- Glironia, and Hyladelphys. In all other
tained by the maxilla in Dromiciops, some didelphids, however, the anterior margin of
peramelemorphians (e.g., Echymipera, Pera- 14
Although we (Voss et al., 2004a) previously reported that
meles), and some dasyuromorphians (Mur- P2 is taller than P3 in Chacodelphys based on our examination of
exia, Sminthopsis). In several diprotodontian the holotype of C. formosa (USNM 236330), subsequently
clades, C1 is absent. collected material (Teta et al., 2006) includes five individuals in
UPPER PREMOLARS: The didelphid first which P2 and P3 are subequal in height. We reexamined
USNM 236330 and concluded that, although the crown of P3
upper premolar (P1) is usually in contact was completely exposed above the alveolus, it had not yet
with or very close to C1, but a diastema is attained its definitive adult position in the toothrow of this
often present between P1 and P2. The gap young adult specimen.
the tooth is rounded, and only the posterior Dromiciops has a large, molariform dP3
cutting edge is well developed. Small anterior resembling the common didelphid condition
blades are variably present near the base of (Marshall, 1982b: fig. 17; Hershkovitz, 1999:
P3 in Chironectes and Philander, but the apex fig. 32). By contrast, dP3 is much smaller,
of the tooth is always rounded anteriorly as more or less vestigial, and structurally
in the other taxa with single-bladed P3s. simplified in caenolestids, dasyurids, and
Most other plesiomorphic marsupials have peramelemorphians (Tate, 1948a; Archer,
three upper premolars, but some dasyurids 1976b; Luckett and Hong, 2000).
(e.g., Dasyurus) have only two. In addition to UPPER MOLARS: Didelphid upper molars
the usual diastema between P1 and P2, some conform to the basic tribosphenic bauplan
adult dasyurids (e.g., Murexia) and all adult (Simpson, 1936) in having three principal
peramelemorphians have a diastema between cuspsparacone, protocone, and meta-
P2 and P3. The third upper premolar (P3) is coneconnected by the usual crests in a
taller than P2 in caenolestids, Dromiciops, more or less triangular array (fig. 20). A
and some dasyurids (e.g., Murexia, Sminthop- broad stylar shelf and an anterolabial cingu-
sis), but P2 is taller than P3 in other dasyurids lum are invariably present; the centrocrista
(e.g., Myoictis). An anterior cutting edge is (postparacrista + premetacrista) and the
apparently absent from P3 in all nondidelphid ectoloph (preparacrista + centrocrista +
marsupials. postmetacrista) are uninterupted by gaps;
UPPER MILK PREMOLAR: Although the the para- and metaconules are indistinct or
deciduous upper premolar (dP3) of didel- absent;15 and there is no posterolingual talon.
phids has consistently been described as large In addition, most didelphids have several
and molariform (Flower, 1867; Thomas, (usually five or six) small cusps on the stylar
1888; Bensley, 1903; Tate, 1948b; Archer, shelf, for which most authors employ alpha-
1976b), there is noteworthy taxonomic vari- betical labels (after Bensley, 1906; Simpson,
ation in the morphology of this tooth. In 1929; Clemens, 1966). Of these, stylar cusp B
most opossums dP3 is, indeed, a sizable (labial to the paracone) is more consistently
tooth: its crown area ranges from about 42% recognizable than the others, but a stylar
to 96% of the crown area of M1, the tooth cusp in the D position (labial to the
immediately behind it (Voss et al., 2001: table metacone) is often subequal to it in size.
5). An obviously functional element of the Didelphids differ conspicuously in the
upper toothrow, dP3 occludes with both the relative width (transverse or labial-lingual
deciduous lower premolar (dp3) and with m1. dimension) of successive molars within tooth-
Hyladelphys, however, has a very small upper rows. In some taxa, the anterior molars tend
milk premolar (,10% of the crown area of to be wide in proportion to more posterior
M1; op. cit.) that appears to be functionally teeth, but in others the posterior molars are
vestigial because it does not occlude with any relatively wider (fig. 21). The ratio obtained
lower tooth. by dividing the width of M4 by the width of
Although molariform, the didelphid upper M1 (M4/M1; table 8) conveniently indexes
milk premolar does not always match the this size-independent shape variation and
teeth behind it in all occlusal details. In most ranges from a minimal value of 0.83 (in
specimens, the paracone of dP3 is located on
the labial margin of the crown, and the stylar 15
The literature is inconsistent on this point, with some
shelf is correspondingly incomplete, whereas authors claiming to have observed significant taxonomic
the paracone is lingual to a continuous stylar variation among Recent didelphids in the occurrence of conules
shelf on all didelphid molars (see below). (see Voss and Jansa, 2003: appendix 4). Indeed, careful
examination of unworn teeth usually reveals a tiny enameled
Among those didelphids whose milk premo- chevron on the postprotocrista that is presumably homologous
lars we examined, only Caluromys has a dP3 with the metaconule of stem metatherians; corresponding
in which the paracone is usually lingual to a structures on the preprotocrista, presumably vestigial paraco-
continuous stylar shelf. nules, are much less frequently observed. The fact that conules
have been scored as present and absent in the same taxon (e.g.,
The morphology of milk premolars re- Didelphis) by different authors (e.g., Reig et al. [1987] versus
mains to be widely surveyed in Marsupialia, Wroe et al. [2000]) sufficiently illustrates the ambiguous
but of those taxa that we examined, only interpretation of such indistinct features.
Fig. 20. Occlusal views of left upper and right lower didelphid molars illustrating features of crown
morphology discussed in the text. Abbreviations: acid, anterior cingulid; alci, anterolabial cingulum; ecf,
ectoflexus; encd, entoconid; encrd, entocristid; hycd, hypoconid; hycld, hypoconulid; hycld.n., hypoconulid
notch; mec, metacone; mecd, metaconid; mecrd, metacristid; pac, paracone; pacd, paraconid; pacrd,
paracristid; pomecr, postmetacrista; popacr, postparacrista; poprcr, postprotocrista; prc, protocone; prcd,
protoconid; prmecr, premetacrista; prpacr, preparacrista; prprcr, preprotocrista; stB, stylar cusp B; stD,
stylar cusp D.
Caluromysiops) to a maximal value of 1.76 (in descriptors of a visually compelling but

Lutreolina). Morphometric comparisons of phylogenetically intractable aspect of didel-
dental measurements (Voss and Jansa, 2003: phid dental variation.16
fig. 12) and visual inspections of toothrows The ectoflexus is a V-shaped labial inden-
suggest that this index is correlated with a tation of the stylar shelf that is either present
pattern of molar occlusal transformation that or absent on tribosphenic marsupial molars.
has been described in the literature as Among didelphids, Caluromys and Caluro-
carnassialization (Reig and Simpson, mysiops are the only taxa that lack any trace
1972: 534) or as an emphasis on postval- of an ectoflexus (e.g., figs. 21A, B). In most
lum-prevalid shear (Muizon and Lange- other opossums, a distinct ectoflexus is
Badre, 1997). These tendencies are redun- present on M3, on M2 and M3, or (rarely)
dantly described by numerous dental ratios on M1M3. When ectoflexi are present on
that have been coded as independent charac- multiple teeth, they invariably increase in
ters in previous phylogenetic studies (for an depth from anterior to posterior.
extended discussion, see Voss and Jansa, The shape of the centrocrista (postpara-
2003). crista + premetacrista) and the descriptive
In effect, didelphids with relatively wide nomenclature associated with its alternative
anterior molars have dentitions that can be states have been a source of some confusion
described as not or only weakly carnassia- among marsupial researchers. Traditionally,
lized, whereas those with relatively wide taxa with a -shaped (labially inflected)
posterior molars have strongly carnassialized centrocrista have been described as dilamb-
teeth. Although the endpoints of this mor- dodont because the ectoloph (preparacrista
phocline are strikingly different, the essen- + centrocrista + postmetacrista) is then W-
tially continuous distribution of intermediate shaped (like two inverted lambdas), whereas
morphologies renders any attempt to make taxa with a straight (uninflected) centrocrista
qualitative distinctions based on molar pro- 16
portions an arbitrary exercise. We therefore We were not so circumspect in earlier phylogenetic
analyses, where computed ratios were used to define qualitative
employ not carnassialized, weakly car- states (e.g., Voss and Jansa, 2003: character 57), but measure-
nassialized, moderately carnassialized, ments from additional taxa (including outgroups not analyzed by
and strongly carnassialized as heuristic us in 2003) have blurred distinctions that once seemed clear.
character states in different phylogenetic

datasets (e.g., Didelphis; see Reig et al.,
1987 [character 1]; Wroe et al., 2000 [charac-
ter 10]; Wible et al., 2001 [character 31]). Our
observations agree with Johansons (1996),
that the shape (linearity versus labial inflec-
tion) of this crest is correlated with its
occlusal relief (apical height above the trigon
basin), and we recognize an intermediate
condition for taxa that do not conform with
either traditionally recognized morphotype.
Among didelphids, only Caluromysiops
has a truly linear centrocrista on M1M3.
In this taxon, the apex of the centrocrista is
essentially level with the floor of the trigon
basin, clearly conforming to the predilamb-
dodont condition defined by Johanson
(1996). By contrast, the centrocrista is
strongly inflected labially (buccally)and
therefore distinctly -shapedin Chacodel-
phys, Cryptonanus, Gracilinanus, Lestodel-
phys, Marmosa, Marmosops, Metachirus,
Monodelphis, Thylamys, and Tlacuatzin. The
apex of the crest is elevated well above the
trigon floor in these taxa, which are unam-
biguously dilambdodont sensu Johanson
(1996). The intermediate condition occurs in
the seven remaining genera (Caluromys,
Chironectes, Didelphis, Glironia, Hyladelphys,
Lutreolina, Philander) in which the centro-
crista has a weak labial inflection with a
slightly elevated apex; these taxa can appro-
priately be described as weakly dilambdo-
dont.
In many didelphidsCaluromys, Caluro-
mysiops, Glironia, Gracilinanus, Hyladelphys,
Marmosa, Tlacuatzin, and some species of
Cryptonanus and Marmosopsthe preproto-
crista passes labially around the base of the
paracone to join with the anterolabial cingu-
Fig. 21. Occlusal views of left upper molar
lum. This results in the formation of a
rows illustrating taxonomic differences in relative
widths of teeth (not drawn to the same scale). A,
continuous shelf along the anterior margin
Caluromysiops irrupta (FMNH 84426); B, Calur- of the tooth (fig. 21AD), and taxa possess-
omys philander (AMNH 267334); C, Glironia ing this feature are sometimes said to have a
venusta (AMNH 71395); D, Marmosa regina complete anterior cingulum (Archer,
(MVZ 190333); E, Marmosops impavidus (MUSM 1976b: 3) or to exhibit double rank pre-
13284); F, Lestodelphys halli (MMNH 15708). vallum-postvallid shearing (Cifelli, 1993:
213). In the alternative morphologyexhib-
have usually been called predilambdodont ited by Chacodelphys, Chironectes, Didelphis,
(e.g., by Reig et al., 1987). Unfortunately, Lestodelphys, Lutreolina, Metachirus, Mono-
neither descriptor applies unambiguously to delphis, Philander, Thylamys, and other
some didelphids (Goin, 1997), and certain species of Cryptonanus and Marmosopsthe
taxa have been coded with contradictory preprotocrista extends only to a point at or
TABLE 8
Upper Molar Measurements and Proportionsa
Widths (mm)c
Ratio
b
N M1 M2 M3 M4 (M4/M1)
Caluromys lanatus 5 2.87 3.09 3.01 2.44 0.85

Caluromys philander 5 2.52 2.84 2.90 2.37 0.94
Caluromysiops irrupta 5 2.99 3.21 3.06 2.48 0.83
Chacodelphys formosa 2 1.23 1.64 1.84 1.88 1.53
Chironectes minimus 5 4.04 4.69 5.33 5.34 1.32
Cryptonanus chacoensis 4 1.30 1.71 1.89 1.85 1.42
Cryptonanus unduaviensis 5 1.39 1.79 2.01 2.02 1.45
Didelphis albiventris 1 4.04 5.01 5.60 5.47 1.35
Didelphis marsupialis 3 4.73 5.50 6.04 6.02 1.27
Didelphis virginiana 5 4.59 5.50 6.54 6.99 1.52
Glironia venusta 4 1.92 2.36 2.56 2.47 1.29
Gracilinanus aceramarcae 4 1.20 1.51 1.64 1.69 1.41
Gracilinanus agilis 5 1.28 1.66 1.87 1.89 1.48
Gracilinanus emiliae 3 1.14 1.49 1.67 1.65 1.45
Gracilinanus microtarsus 4 1.30 1.66 1.84 1.86 1.43
Hyladelphys kalinowskii 5 0.96 1.14 1.26 1.21 1.26
Lestodelphys halli 5 1.80 2.40 2.73 2.97 1.65
Lutreolina crassicaudata 4 2.76 3.36 4.15 4.85 1.76
Marmosa (Marmosa) lepida 5 1.33 1.69 1.90 1.96 1.47
Marmosa (Marmosa) mexicana 5 1.44 1.82 2.08 2.26 1.57
Marmosa (Marmosa) murina 5 1.46 1.78 2.08 2.18 1.49
Marmosa (Marmosa) robinsoni 5 1.71 2.10 2.36 2.53 1.48
Marmosa (Marmosa) rubra 5 1.61 2.05 2.33 2.57 1.60
Marmosa (Micoureus) demerarae 5 1.98 2.51 2.82 2.81 1.42
Marmosa (Micoureus) paraguayana 3 2.11 2.65 3.01 3.06 1.45
Marmosa (Micoureus) regina 5 2.18 2.77 3.12 3.17 1.45
Marmosops impavidus 5 1.58 2.00 2.30 2.46 1.56
Marmosops incanus 4 1.74 2.18 2.42 2.69 1.55
Marmosops noctivagus 5 1.83 2.31 2.67 2.82 1.54
Marmosops parvidens 5 1.34 1.67 1.91 1.90 1.42
Marmosops pinheiroi 5 1.33 1.68 1.92 2.04 1.53
Metachirus nudicaudatus 5 2.84 3.29 3.69 3.62 1.27
Monodelphis brevicaudata 5 2.08 2.46 2.64 2.90 1.39
Monodelphis emiliae 5 1.85 2.21 2.34 2.57 1.39
Monodelphis peruviana 5 1.57 1.87 2.05 2.15 1.37
Monodelphis theresa 2 1.36 1.67 1.84 2.11 1.55
Philander frenatus 3 2.77 3.32 3.92 4.17 1.51
Philander mcilhennyi 3 3.53 4.15 4.71 4.99 1.41
Philander opossum 4 3.18 3.74 4.30 4.55 1.43
Thylamys macrurus 5 1.55 2.05 2.32 2.37 1.53
Thylamys pallidior 5 1.31 1.74 2.04 2.12 1.62
Thylamys pusillus 5 1.34 1.72 1.96 2.05 1.53
Thylamys venustus 5 1.37 1.78 2.00 2.10 1.53
Tlacuatzin canescens 5 1.54 1.98 2.28 2.33 1.51
a
Teeth were measured with digital calipers while viewed under low magnification to ensure that the caliper jaws were
consistently placed.
b
Sample size.
c
Molar width was measured from the stylar shelf (at or near the A position) to the lingual apex of the protocone.
Tabulated values are sample means.
near the base of the paracone; the anterola- The upper molars of Recent peramelemor-
bial cingulum does not converge toward the phians differ from those of didelphids in
preprotocrista in these taxa, but passes several respects, most notably by having a
obliquely dorsally such that the two crests discontinuous centrocrista: instead of form-
are discontinuous on the anterior surface of ing a -shaped (labially inflected) crest with
the tooth crown (fig. 21E, F). Voss and Jansa an apex that is lingual to the stylar shelf, the
(2003: 39) discussed several inconsistencies in postparacrista and the premetacrista of
the published literature concerning the dis- modern peramelemorphians terminate on
tribution of these traits among didelphids. the labial margin of the tooth, where they
The postprotocrista also exhibits notewor- are separated by a small gap (Muirhead and
thy variation among didelphids. In most Filan [1995] referred to this trait as the result
genera, this crest decreases in height and of the centrocrista breaching the ectoloph).
width as it passes posterolabially around the Additionally, all modern peramelemophians
base of the metacone before merging with the have a large posterolingual cusp on M1M3;
posterior surface of the tooth crown, leaving in most genera (e.g., Echymipera, Perameles)
a distinct groove or gap in the posterior wall this is a hypertrophied metaconule, but in
of the trigon basin. By contrast, the post- Macrotis it is the metacone (Archer, 1976b).
protocrista connects directly with the base of Another peramelemorphian trait is that the
the metacone, such that the posterior wall of preparacrista of M1 is a tall crest that passes
the trigon basin has no gap or groove, in posterolabially to a large stylar cusp at or
Chironectes, Didelphis, Lutreolina, and Phi- near the C position (Archer, 1976b). The
lander; instead, the unworn postprotocrista anterolabial cingulum is present in some
of these taxa exhibits a distinct carnassial peramelemorphians (e.g., Echymipera) but
notch near the base of the metacone. not in others (e.g., Perameles).
Dasyurids and Dromiciops have tribosphe- On caenolestid molars the principal labial
nic upper molars that differ from those of cusps are hypertrophied stylar elements in the
didelphids in only minor details. Indeed, B and D positions (Osgood, 1921; Marshall,
1987; Goin and Candela, 2004; Goin et al.,
some dasyurids and didelphids are indistin-
2007). In both Caenolestes and Rhyncho-
guishable in published matrices of upper
lestes, the paracone is absent (or indistin-
molar characters (e.g., Archer, 1976b: table
guishably fused with stylar cusp B). The
1), although there is a tendency for stylar
metacone is also absent in Rhyncholestes but,
cusp D to be larger than stylar cusp B when
although fused to the lingual aspect of stylar
both structures are present in the former
cusp D, the metacone is still clearly recog-
group (Wroe, 1997), and a few dasyurids
nizable in Caenolestes. Given this interpreta-
(e.g., Sarcophilus) are conspicuously diver-
tion of caenolestid cusp homologies, it
gent in other upper molar traits. The upper follows that a low crest along the labial base
molars of Dromiciops differ from those of of the tooth crowns in both genera is a
most didelphids by having a linear centro- neomorphic cingulum rather than a vestigial
crista, indistinct stylar cusps, and a stylar stylar shelf. The first two caenolestid upper
shelf that is conspicuously reduced in width molars also have a well-developed postero-
labial to the paracone on M1.17 lingual cusp that has been interpreted either
as a hypertrophied metaconule (Marshall,
17
These distinctive features are not apparent in some 1987; Goin and Candela, 2004; Goin et al.,
published illustrations of the dentition of Dromiciops, notably 2007) or as a neomorphic outgrowth of the
Marshalls (1982b) figure 16b, where an occlusal view of the postprotocingulum (Hunter and Jernvall,
upper teeth of FMNH 22671 shows -shaped centrocristae and
an unreduced stylar shelf on M1. However, in another 1995). Another distinctive trait of the upper
illustration (based on FMNH 22673; op. cit.: fig. 17b), M1 molars of Recent caenolestids is the complete
is correctly shown with a reduced stylar shelf labial to the absence of an anterolabial cingulum.
paracone, and with a linear centrocrista. We borrowed FMNH UPPER DENTAL ERUPTION SEQUENCES: As
22671 to determine whether this specimen is unusual in any
way and found that it is not: the paracone of M1 is almost on
described by Tribe (1990), P3 is the last upper
the labial margin of the tooth, and the centrocristae of M1M3 tooth to erupt in most didelphids. By
are linear; therefore, Marshalls figure 16b is inaccurate. contrast, P3 erupts before M4, which is the
last upper tooth to emerge in Chironectes, distinct lingual cusp or heel, but this structure
Didelphis, Lutreolina, and Philander. In a few is lacking in at least some of the larger
didelphids (e.g., Metachirus nudicaudatus, carnivorous forms (e.g., Dasyurus). A dis-
Monodelphis brevicaudata), M4 and P3 ap- tinctive feature of unworn i3 in dasyurids is
pear to erupt simultaneously. Molar eruption the consistent presence of a small posterior
sequences have not been widely surveyed lobe on the cutting edge of the tooth.18 As
among other marsupials, but P3 seems to be implied by Bensley (1903: 113), this structure
the last upper tooth to erupt in most appears to be homologous with the much
nondidelphid clades, with the exception of larger posterior lobe of i3 seen in peramele-
Dromiciops and caenolestids, in which M4 is morphians. All peramelemorphian lower
usually last (Hershkovitz, 1999; Luckett and incisors lack a lingual cusp or heel, but i1
Hong, 2000). and i2 are otherwise unremarkable in this
LOWER INCISORS: The four lower incisors group. The lower incisor dentition of Dromi-
of didelphids are subequal in size and ciops is similar to that of didelphids except
essentially similar in coronal shape, but the for the fact that i2 is neither staggered nor
alveolus of i2 is always staggered (dis- buttressed in microbiotherians (Hershkovitz,
placed lingually and dorsally relative to the 1999: figs. 5A, 29E).
alveoli of i1, i3, and i4), and the root of this The anteriormost lower incisor of caeno-
tooth is correspondingly supported by a lestids is a long-crowned procumbent (gliri-
prominent labial buttress of alveolar bone form) tooth that most authors (e.g., Ride,
(Hershkovitz, 1982, 1995). The unworn lower 1962; Marshall, 1980; Sanchez-Villagra,
incisor crowns of most didelphids have a 2001b) assume to be either i1 or i2. Which-
distinct lingual cusp or heel (fig. 22, top) that ever it is, at least two of the four single-
is indistinct or absent in Chironectes, Didel- rooted unicuspids that normally occur be-
phis, Lutreolina (fig. 22, bottom), and Phi- tween the gliriform tooth and the anterior-
lander. This character has the same taxo- most undisputed premolar must also be
nomic distribution among didelphids as incisors.19 None of the lower incisors of
Takahashis (1974) distinction between the Caenolestes or Rhyncholestes has a staggered
subrectangular teeth of Chironectes, Didel- alveolus.
phis, Lutreolina, and Philander on the one LOWER CANINE: The lower canine is an
hand, and the suboval lower incisors of erect unicuspid tooth that is conspicuously
Caluromys, Marmosa, Metachirus, and differentiated from the incisors and premo-
Monodelphis on the other, but we are unable lars in all of the larger didelphids (Caluromys,
to appreciate the basis for her shape descrip- Caluromysiops, Chironectes, Didelphis, Lu-
tors. No didelphid has a lobed or indented
treolina, Philander), and in some of the
cutting edge on any of the lower incisor
smaller forms as well (Glironia, Hyladelphys,
crowns.
Among other marsupials, Dromiciops re- 18
This is the posterobuccal lobe of Archer (1976b: 6) who
sembles didelphids in having four lower stated that it is present in only a few dasyurids. However, it is
incisors, whereas dasyurids and peramele- immediately recognizable in juvenile specimens of all the
morphians have only three. The missing dasyurids we examined. Ontogenetically persistent in some taxa,
it is soon obliterated with wear in others and may be
tooth in dasyurids and peramelemorphians indistinguishable in adults.
is presumably i4, because there is only one 19
Like others who have pondered caenolestid antemolar
tooth behind staggered-and-buttressed i2, tooth homologies (e.g., Thomas, 1895; Ride, 1962; Marshall,
and because it seems more likely that teeth 1980), we assume that no more than one lower canine and three
should be lost from the end of the incisor lower premolars are normally present. Because the presence of
two lower premolars (each with two roots; Osgood, 1921) is
series than from intermediate loci. In all of undisputed, only one of the problematic unicuspids could
these polyprotodont groups, the lower inci- belong to this series. Hershkovitz (1982) implied that none of
sors are short-crowned nonprocumbent teeth the lower incisor alveoli of caenolestids are staggered, but he
that differ from those of didelphids in only later (Hershkovitz, 1995) stated that Recent caenolestids have a
staggered i2 (i3). In all of the specimens of Caenolestes and
minor details. Rhyncholestes that we examined, however, the four unicuspids
The unworn lower incisors of most dasy- behind the gliriform tooth have colinear alveoli; there is no
urids (e.g., Murexia, Sminthopsis) have a suggestion of lingual displacement at the second locus.
Fig. 22. Lingual views of lower incisors. Top, Metachirus nudicaudatus (AMNH 266452) with distinct
posterior accessory cusps on i1i4. Bottom, Lutreolina crassicaudata (AMNH 210424) without distinct
posterior accessory cusps on i1i4.
Lestodelphys, Monodelphis, Tlacuatzin). In all small didelphids. Premolarization of c1 in-

of these taxa, the unworn crown of c1 tapers cludes the development of a flattened anteri-
to a sharp apical point (fig. 23A). By or blade (rather than a sharp tip) and a
contrast, the lower canine is procumbent posterior cingulum that sometimes produces
and tends to become premolariform in other a distinct accessory cusp. Whereas these
tendencies are minimally indicated on the

unworn teeth of Chacodelphys and Meta-
chirus (which are flattened along their ante-
rior margin but have no posterior cingulum),
they are increasingly apparent in Cryptona-
nus, Gracilinanus, Marmosa (fig. 23B), and
Thylamys. Premolarization of the lower
canine is maximally developed in the smaller
species of Marmosops (e.g., M. parvidens and
M. pinheiroi), where c1 and p13 form a
series of four strikingly similar teeth (fig.
23C). Generally speaking, taxa with well-
developed accessory cusps on the upper
canine also have a well-developed posterior
accessory cusp on c1, so variation in these
occluding teeth is at least partially correlated.
Large dasyurids (e.g., Dasyurus) have erect,
unicuspid lower canines, but many small
dasyurids (e.g., Sminthopsis) have procum-
bent, premolariform teeth. The lower canines
are small and blunt in peramelemorphians. In
Dromiciops, c1 is procumbent and more or
less premolariform (Marshall, 1982b: fig. 14).
LOWER PREMOLARS: Didelphid lower pre-
molars are unremarkable teeth, each having a
single dominant cusp and a well-developed
distal cingulum that often produces a poste-
rior accessory cusp. Small diastemata are
sometimes present between p1 and c1,
between p1 and p2, and (rarely) between p2
and p3. The lower first premolar (p1) is
vestigial and occlusally featureless in Calur-
omys and Caluromysiops, but it is well
developed in all of the other genera.
The second lower premolar is distinctly
taller than p3 in most didelphids, but p2 and
p3 are subequal in Marmosops incanus and
most examined species of Monodelphis and
Thylamys. In Lestodelphys and in certain
other species of Monodelphis (e.g., M. emi-
liae), however, p3 is distinctly taller than p2.
Most other plesiomorphic marsupials have
three lower premolars that are essentially
similar to those of didelphids, although some
Fig. 23. Labial views of unworn anterior dasyurids (e.g., Dasyurus) have only two.
mandibular dentitions illustrating taxonomic dif- Among the taxa we examined, p3 is taller
ferences in lower canine (c1) morphology. A,
than p2 in caenolestids, Dromiciops (see
Lestodelphys halli (MMNH 17171) with erect,
acutely pointed, and simple c1. B, Marmosa Marshall, 1982b: fig. 14), some dasyurids
mexicana (AMNH 17136) with semiprocumbent (e.g., Murexia), and some peramelemor-
and flat-crowned but simple c1. C, Marmosops phians (e.g., Echymipera), but p2 and p3 are
pinheiroi (AMNH 267342) with fully premolari- subequal in height or p2 is the taller tooth in
form c1 (procumbent and flat crowned with other dasyurids (e.g., Sminthopsis) and per-
posterior accessory cusp). amelemorphians (e.g., Macrotis).
Fig. 24. Occlusal views of right dp3 and m1 illustrating taxonomic differences in trigonid morphology
of the deciduous tooth. Top, Marmosops impavidus (MUSM 13286) with a complete (tricuspid) dp3
trigonid. Bottom, Marmosa murina (MUSM 15297) with an incomplete (bicuspid) dp3 trigonid.
LOWER MILK PREMOLAR: Taxonomically nid of dp3 is usually distinct in our material
significant variation among didelphids in the of Caluromys philander, the trigonid of this
size and shape of dp3 was reported by Voss et taxon is triangular in occlusal outline and
al. (2001: table 5). Many didelphids (Calur- more closely resembles the fully molariform
omys lanatus, Caluromysiops, Chironectes, condition than the blade-like uni- or bicuspid
Cryptonanus chacoensis, Didelphis, Lutreo- alternative. Among other plesiomorphic mar-
lina, Marmosops, Metachirus, Monodelphis supials that we examined, only Dromiciops
emiliae, Philander) have a fully molariform has a large and fully molariform dp3,
dp3 in which the trigonid is represented by its whereas dasyurids and peramelemorphians
normal complement of three cusps (paraco- have small, structurally simplified and obvi-
nid, protoconid, metaconid) in a more or less ously vestigial lower milk premolars.
triangular configuration (fig. 24, top). A fully LOWER MOLARS: Didelphid lower molars
molariform dp3 is also the modal condition conform to the usual tribosphenic pattern
for Monodelphis brevicaudata, in which seven and are readily described using standard
of nine specimens examined with intact milk nomenclature (fig. 20). The trigonid moiety
dentitions had complete trigonids. By con- of each tooth invariably consists of three
trast, dp3 is only partially molariform in distinct cusps (paraconid, protoconid, and
Cryptonanus unduaviensis, Gracilinanus agilis, metaconid) that are connected by two tall
Lestodelphys, Marmosa, and Thylamys, and deeply notched crests (paracristid and
which have incomplete, blade-like trigonids metacristid); three additional cusps (hypo-
bearing only one or two distinct cusps conid, entoconid, and hypoconulid) surround
(fig. 24, bottom). Although just the protoco- the talonid basin. A well-developed anterior
cingulid (precingulid) is always present, but a hypoconid and much exceeds the adjacent
posterior cingulid (postcingulid) is just as hypoconulid in height and breadth on m1m3
consistently absent.20 The anterior cingulid (fig. 25, top). In Chacodelphys and Monodel-
never extends all the way to the base of the phis, however, the entoconid is very small,
paraconid because it abuts on the hypocon- never more than subequal to the hypoconulid,
ulid of the preceding tooth (dp3 in the case of and often smaller than that cusp (fig. 25,
m1), resulting in a prominent concavity in the bottom); it becomes indistinct or is obliterat-
anterolingual occlusal outline (the hypocon- ed by wear in most adult specimens. Discrep-
ulid notch of Archer, 1976b). The entocris- ant observations in the literature about
tid is a short crest that approximately didelphid entoconid size variation were dis-
parallels the lingual margin of the tooth. cussed by Voss and Jansa (2003).
Didelphid lower molars exhibit taxonomic In most didelphids, the hypoconulid is
shape variation that is correlated with much closer to the entoconid than it is to the
carnassialization as previously described for hypoconid, and is said to be twinned with
the upper teeth. Thus, taxa with uncarnassia- the former cusp. Although twinning of the
lized molars have relatively small trigonids hypoconulid and entoconid is often stated (or
and large talonids, whereas opossums with implied) to be universally present among
highly carnassialized molars have relatively tribosphenic marsupials (e.g., by Clemens,
large trigonids and small talonids. Trigonid 1979; Kielan-Jaworowska et al., 2004), it is
enlargement is invariably accompanied by an not. The hypoconulid and entoconid are not
increase in occlusal relief, the protoconid and closely approximated in Caluromysiops, and
metacristid in particular tending to become in some specimens (e.g., FMNH 121522) the
very tall in the most carnassialized forms hypoconulid is at the center of the posterior
(e.g., Lestodelphys and Lutreolina). The width talonid margin, equidistant to the hypoconid
gradient previously noted for the upper and the entoconid.
molars is also seen in the lowers, m2 being Most other plesiomorphic marsupials have
the widest tooth in uncarnassialized denti- lower molars that resemble those of didelphids
tions whereas m3 or m4 is widest in in common tribosphenic features, but several
progressively more carnassialized forms. As differences are noteworthy. Dasyurids, for
for the upper teeth, the highly divergent example, have a well-developed posterior
endpoints of this conspicuous gradient of cingulid on m1m3, the anterior cingulid of
dental shape variation are connected by an peramelemorphian lower molars lacks a hy-
almost continuous taxonomic series of mor- poconulid notch, the anterior cingulid is
vestigial in Dromiciops (which lacks a hypo-
phological intermediates.
conulid notch on m1), and an entocristid is
Only a few lower molar traits that vary
absent in some dasyurids (e.g., Sminthopsis)
among didelphids seem useful for the pur-
and peramelemorphians (e.g., Perameles).
poses of taxon diagnoses and phylogenetic
As they are in so many other character
analysis. Among these, the labial prominence
complexes, caenolestids are the most diver-
of the hypoconid tends to vary along the
gent basal marsupials in lower molar mor-
toothrow, so we standardized comparisons at
phology. In both Caenolestes and Rhyncho-
the third molar locus. In most opossums, the
lestes the paracristid is unnotched and
hypoconid is labially salient (projecting
decreases gradually in height from the apex
beyond the protoconid or level with it) on
of the protoconid to the anterolingual corner
m3, but the hypoconid is lingual to the
of the tooth; the paraconid is therefore
protoconid in Chacodelphys, Lestodelphys, indistinct. Whereas the paracristid runs in
Lutreolina, and Monodelphis. an uninterrupted arc from the protoconid to
Most didelphids have a well-developed the anterolingual corner of the tooth on m1,
entoconid that is about as tall as the the paracristid on m2 and m3 abuts the
20 hypoconulid of the preceding tooth (m1 and
A posterior cingulid (postcingulum) was illustrated by
Goin and Candela (2004: fig. 2B) on a diagram representing the
m2, respectively), at which point it is abruptly
occlusal features of a didelphimorphian opossum, but no deflected posterolingually. The only distinct
living didelphimorphian has such a structure. trigonid cusp on m4 is the metaconid, from
Fig. 25. Lingual views of right m2 and m3 illustrating taxonomic differences in entoconid size. Top,
Thylamys pallidior (AMNH 262405) with a large entoconid (arrow). Bottom, Monodelphis peruviana
(AMNH 272781) with an indistinct entoconid (arrow).
which a diminutive crest (comprising a Summary

vestigial metacristid and part of the para-
cristid) curves anterolabially to abut the The morphological observations described
hypoconulid of m3. The entocristid on m1 above provide evidence of taxonomic differ-
m3 is a long crest that is deflected labially, ences in many phenotypic attributes suitable
away from the lingual margin of the tooth, for phylogenetic analysis, including unam-
and terminates deep inside the talonid basin. biguously distinguishable aspects of pelage
color and structure; sexually dimorphic skin Breast Cancer Activating 1 Gene
glands and carpal tubercles; manual and
pedal digital proportions; the presence, ab- The human BRCA1 gene consists of 22
sence, and morphology of cutaneous shelters coding exons distributed over ca. 100 kb of
for nursing young; caudal modifications for genomic DNA on the long arm of chromo-
prehensility and fat storage; the sizes and some 17 (Miki et al., 1994). The gene is
shapes of cranial bones and bony processes; expressed in numerous tissues, but it is
the presence, absence, and ontogenetic per- especially active in epithelial cells undergoing
sistence of cranial sutures; the number and high levels of proliferation and differentia-
occurrence of cranial foramina and fenestrae; tion (Miki et al., 1994; Casey, 1997). BRCA1
auditory morphology; and the sizes, shapes is a tumor suppressor gene whose unmutated
and eruption sequences of teeth. Because the product is a large (1863 amino acids) nuclear-
relevant literature on mammalian functional cytoplasmic-shuttling phosphoprotein of un-
morphology is too extensive to review known structure but with apparently essen-
effectively in this report, we can only assert tial roles in transcriptional regulation and
that such taxonomic differences are likely to DNA repair, among other intracellular func-
reflect adaptations for such diverse functions tions (Rodrguez and Henderson, 2000;
as concealment, locomotion, reproduction, Dimitrov et al., 2001; Venkitaraman, 2001).
sense perception, and food reduction. Ana- Nucleotide sequence data from BRCA1
tomical trait variation (together with pub- exon 11 have been analyzed in several
lished information about karyotypic differ- mammalian phylogenetic studies, usually in
ences that are not reviewed above) is formally concatenated datasets with other genes (e.g.,
described as qualitative characters in appen- Adkins et al., 2001; Madsen et al., 2001;
dix 3, where we also explain the criteria used Delsuc et al., 2002; Douady et al., 2002;
to score individual specimens and discuss the Raterman et al., 2006). Homology compar-
justification for ordering multistate transfor- isons of the 2163 bp fragment that we
mation series. The resulting data are summa- amplified and sequenced for this study are
rized in appendix 4. consistent with the presence of a single full-
length copy in Monodelphis domestica, where
GENE SEQUENCES the gene appears to be located on chromo-
some 2. BRCA1 sequence data are available
Of the many genes that have been from 48 terminals in our analyses, including
sequenced to date from one or more didel- 41 didelphids (Caluromysiops and Lestodel-
phid marsupials (including the entire genome phys were not sequenced) and all 7 non-
of Monodelphis domestica; Mikkelsen et al., didelphid outgroups. All of the marsupial
2007), only a few have been sequenced from sequences that we obtained and others that
enough taxa to be useful for phylogenetic we downloaded from GenBank translate to
inference. Below we summarize information open reading frame. Twenty independent
about five protein-coding nuclear loci from insertion-deletion events (indels) ranging in
which long (.900 bp) nucleotide sequences length from 3 to 39 bp were required to align
have now been obtained from numerous didelphid BRCA1 sequences; of these, 9 were
species representing almost all of the current- unique to single taxa and 11 were shared by
ly recognized genera (table 9). The following two or more taxa.
accounts describe the size, internal organiza-
tion, and chromosomal location of each gene, Dentin Matrix Protein 1
as well as the tissues in which it is known to
be active and the physiological activity of its The human DMP1 gene (originally known
translated protein product. We also provide a as AG1; George et al., 1993) consists of six
brief synopsis of previous analyses based on exons spanning about 14 kb of genomic
each gene in mammalian phylogenetic re- DNA on the long arm of chromosome 4
search before describing the sequence char- (Hirst et al., 1997). DMP1 is an acidic
acteristics of the fragments we amplified from calcium-binding secreted phosphoprotein of
didelphids and other marsupials. about 513 amino acids, most of which are
TABLE 9
GenBank Accession Numbers for Molecular Sequences Analyzed in This Reporta
BRCA1 DMP1 IRBP RAG1 vWF
Didelphidae
Caluromys lanatus FJ159278 DQ083120 AF257690 DQ865883 FJ159328
Caluromys philander FJ159279 DQ083121 AY233766 DQ865884 FJ159329
Caluromysiops irrupta N/A N/A AF257675 N/A N/A
Chironectes minimus FJ159280 DQ083122 AF257679 DQ865885 FJ159330
Cryptonanus chacoensis FJ159281 DQ083123 AY957494 DQ865886 FJ159331
Cryptonanus unduaviensis FJ159282 DQ083124 AY957492 DQ865887 FJ159332
Didelphis albiventris FJ159283 DQ083125 AF257683 DQ865888 FJ159333
Didelphis marsupialis FJ159284 DQ083126 AF257703 DQ865889 FJ159334
Didelphis virginiana FJ159285 DQ083127 AF257678 DQ865890 FJ159335
Glironia venusta FJ159286 DQ083128 AF257688 DQ865891 FJ159336
Gracilinanus aceramarcae FJ159287 DQ083129 AY957488 DQ865892 FJ159337
Gracilinanus agilis FJ159288 DQ083130 AY957487 DQ865893 FJ159338
Gracilinanus emiliae FJ159289 DQ083131 AY957486 DQ865894 FJ159339
Gracilinanus microtarsus FJ159290 DQ083132 AF257687 DQ865895 FJ159340
Hyladelphys kalinowski FJ159291 DQ112326 DQ112324 DQ865923 FJ159341
Lestodelphys halli N/A DQ083133 AF257710 N/A N/A
Lutreolina crassicaudata FJ159292 DQ083134 AF257685 DQ865896 FJ159342
Marmosa demeraraeb FJ159304 DQ083146 AF257693 DQ865908 FJ159354
Marmosa lepida FJ159293 DQ083135 AY233769 DQ865897 FJ159343
Marmosa mexicana FJ159294 DQ083136 AY233770 DQ865898 FJ159344
Marmosa murina FJ159295 DQ083137 AF257709 DQ865899 FJ159345
Marmosa paraguayanab FJ159305 DQ083147 AY233778 DQ865909 FJ159355
Marmosa reginab FJ159306 DQ083148 AY233780 DQ865910 FJ159356
Marmosa robinsoni FJ159296 DQ083138 AY233774 DQ865900 FJ159346
Marmosa rubra FJ159297 DQ083139 AY233776 DQ865901 FJ159347
Marmosops impavidus FJ159298 DQ083140 AF257695 DQ865902 FJ159348
Marmosops incanus FJ159299 DQ083141 AY233785 DQ865903 FJ159349
Marmosops noctivagus FJ159300 DQ083142 AF257704 DQ865904 FJ159350
Marmosops parvidens FJ159301 DQ083143 AF257680 DQ865905 FJ159351
Marmosops pinheiroi FJ159302 DQ083144 AF257677 DQ865906 FJ159352
Metachirus nudicaudatus FJ159303 DQ083145 AF257707 DQ865907 FJ159353
Monodelphis brevicaudata FJ159307 DQ083150 AY233782 DQ865912 FJ159357
Monodelphis emiliae FJ159308 DQ083151 AF257692 DQ865913 FJ159358
Monodelphis peruvianac FJ159309 DQ083149 AF257694 DQ865911 FJ159359
Monodelphis theresa FJ159310 DQ083152 AY233783 DQ865914 FJ159360
Philander frenatus FJ159312 DQ083153 AY233787 DQ865915 FJ159362
Philander mcilhennyi FJ159313 DQ083154 AF257706 DQ865916 FJ159363
Philander opossum FJ159314 DQ083155 AY233789 DQ865917 FJ159364
Thylamys macrurus FJ159317 DQ083156 AY957491 DQ865918 FJ159366
Thylamys pallidior FJ159318 DQ083157 AF257689 DQ865919 FJ159367
Thylamys pusillus FJ159319 DQ083158 AY957489 DQ865920 FJ159368
Thylamys venustus FJ159320 DQ083159 AY233791 DQ865921 FJ159369
Tlacuatzin canescens FJ159321 DQ083160 AY233768 DQ865922 FJ159370
Caenolestidae
Caenolestes fuliginosus AF355794 N/A FJ159322 AY243384 AY243403
Rhyncholestes raphanurus AY243454 N/A AY243440 AY243399 FJ159365
Dasyuridae
Murexia longicaudata FJ159311 N/A FJ159325 N/A FJ159361
Sminthopsis crassicaudata FJ159316 N/A FJ159327 N/A AY243413
TABLE 9
(Continued )
BRCA1 DMP1 IRBP RAG1 vWF

Microbiotheriidae
Dromiciops gliroides AY243446 N/A FJ159323 AY243389 AY243407
Peramelidae
Echymipera kalubu AF355796 N/A FJ159324 AY243386 AY243405
Perameles gunnii AY243450 N/A FJ159326 AY243394 AY243411
a
See table 2 for taxonomic conventions. N/A 5 not available (see text).
b
As Micoureus in GenBank.
c
Formerly Monodelphis adusta.
encoded by exon 6 (Hirst et al., 1997). The of genomic DNA on the long arm of
protein is assumed to have some role in tissue chromosome 10 (Liou et al., 1989; Fong et
mineralization based on its chemical proper- al., 1990; Pepperberg et al., 1993). The gene
ties and presence in the extracellular matrix product is a large (1230 amino acids) secreted
of dentine and bone (Butler and Ritchie, glycolipoprotein that is primarily found in the
1995; Feng et al., 2003). However, mRNA extracellular matrix between the retinal pig-
transcripts and the protein itself have also ment epithelium (RPE) and the neural retina
been detected in a wide range of soft tissues, of the eye; the protein apparently functions as
where the gene presumably has other func- a two-way carrier of retinoid between the rod
tions (Terasawa et al., 2004). photoreceptors and the RPE, and therefore
Nucleotide sequence data from DMP1 plays a critical role in visual pigment regen-
exon 6 have been analyzed in several previous eration (Pepperberg et al., 1993).
studies of mammalian intraordinal relation- Nucleotide sequences from IRBP exon 1
ships (e.g., Van Den Bussche et al., 2003; were among the first molecular character data
Reeder and Bradley, 2004; Jansa et al., 2006). to provide convincing support for mammali-
Homology comparisons of the 1176 bp frag- an superordinal relationships (Stanhope et
ment that we amplified and sequenced for al., 1992, 1996; Springer et al., 1997b), and
this study are consistent with the presence of they have also been analyzed in several
a single full-length copy in Monodelphis subsequent studies focused on lower-level
domestica, where the gene appears to be relationships (e.g., Jansa and Voss, 2000;
located on chromosome 5. We obtained Mercer and Roth, 2003; Weksler, 2003; Jansa
DMP1 sequences from 42 didelphids (none and Weksler, 2004). The 1158 bp fragment
is available from Caluromysiops). Fourteen that we amplified and sequenced from didel-
indels ranging in length from 3 to 18 bp are phids is homologous with a single full-length
required to align didelphid DMP1 sequences, copy in Monodelphis domestica, where the
including 6 that are unique to single taxa and gene is apparently located on chromosome 1.
8 that are shared by two or more taxa. IRBP sequences are available from 43 didel-
Unfortunately, didelphid DMP1 sequences phid terminal taxa (including Caluromysiops
cannot be aligned unambiguously with se- and Lestodelphys) and 7 nondidelphid out-
quences from nondidelphid marsupials, so no groups. No indels are required to align
outgroups are available to root phylogenetic didelphid IRBP sequences, all of which
analyses of this gene (Jansa and Voss, 2005; translate to open reading frame (Jansa and
Jansa et al., 2006). Voss, 2000, 2005; Voss and Jansa, 2003).
Interphotoreceptor Retinoid Binding Protein Recombination Activating 1 Gene
The human IRBP gene (also known as the The human RAG1 gene contains a single
Retinol Binding Protein 3 gene, RBP3) uninterrupted exon of 3.1 kb located on the
consists of four exons spanning about 9.5 kb short arm of chromosome 11 adjacent to the
RAG2 locus; the gene product is a large subendothelial matrix (by endothelial cells),
DNA-binding protein of 1043 amino acids or they are packaged into the m granules of
(Schatz et al., 1989; Oettinger et al., 1990, platelets (by megakaryocytes). Multimeric
1992). Recombination activating genes are vWF mediates adhesion of platelets to
expressed only in the nucleus of developing B exposed subendothelium following vascular
and T lymphocytes, where the RAG1 and injury, and it serves as the plasma chaperone
RAG2 proteins act synergistically to make of blood factor VIII, an important regulatory
double-stranded breaks at specific recogni- protein in the coagulatory cascade (Ruggieri
tion sequences, initiating recombination of and Zimmerman, 1987; Ginsburg and Bowie,
variable (V), diversity (D), and joining (J) 1992; Sadler, 1998).
gene segments; completion of the V(D)J Nucleotide sequence data from exon 28 of
rearrangement process yields functional im- the von Willebrand factor gene have been
munoglobulin and T-cell receptor proteins, analyzed in many previous mammalian
an essential step in the development of a phylogenetic studies (e.g., Porter et al.,
mature immune system (Gellert, 2002). 1996; Huchon et al., 1999; Neumann et al.,
Nucleotide sequence data from RAG1 2006). The 963 bp fragment that we ampli-
have previously been analyzed in several fied and sequenced for this study is homol-
vertebrate phylogenetic studies (e.g., Groth ogous with a single full-length copy in
and Barrowclough, 1999; Barker et al., 2002; Monodelphis domestica, where the gene is
Baker et al., 2004; Gruber et al., 2007). The apparently located on chromosome 8. We
2790 bp fragment that we amplified and obtained vWF sequences from 41 didelphids
sequenced for this study is homologous with (excluding Caluromysiops and Lestodelphys)
a single full-length copy in Monodelphis and seven outgroup taxa, all of which
domestica, where the gene is apparently translate to open reading frame. Only three
located on chromosome 5. Although RAG1 indels (two of 3 bp and one of 6 bp) are
occupies the same chromosome as DMP1, required to align didelphid vWF sequences;
the considerable distance that separates these all three indels are unique to single taxa.
loci (.2 3 108 bp) suggests that they belong
to different linkage groups. We obtained full- Summary, Sequence Characteristics, and
length RAG1 sequences from 41 didelphids Model Fitting
(excluding Caluromysiops and Lestodelphys),
and we downloaded shorter (543 bp) out- Available information suggests that the
group sequences from Genbank. Ingroup and five nuclear genes sequenced for this study
outgroup sequences all translate to open exist as single full-length copies in the only
reading frame. Only two indels (of 6 bp each) didelphid genome currently available for
were required to align didelphid RAG1 inspection, that they are unlinked (on sepa-
sequences; one indel is unique to Metachirus rate chromosomes or very widely separated),
nudicaudatus, whereas the other is shared by that they are active in different tissues, and
Marmosops parvidens and M. pinheiroi. that their translated protein products have
widely differing functions (table 10). There-
von Willebrand Factor fore, it is reasonable to expect that taxonomic
patterns of nucleotide substitution at these
The human vWF gene consists of 52 exons loci provide substantially independent evi-
spanning 178 kb of genomic DNA on the dence of phylogenetic relationships. Addi-
short arm of chromosome 12 (Mancuso et tionally, the coding sequences we obtained
al., 1989). The gene is normally active only in appear to be fully functional (all translate to
vascular endothelium and in bone-marrow open reading frame), and none exhibits
megakaryocytes (Ruggieri and Zimmerman, alignment ambiguities among our ingroup
1987). The mature (post-translationally pro- taxa. Other sequence characteristics that
cessed) gene product is a large (2050 amino might affect the utility of these genes for
acids) glycoprotein that is assembled into phylogenetic analysis are considered below.
multimers of various sizes; these multimers Base-compositional heterogeneity, as indi-
are either secreted into the plasma or into the cated by significant taxonomic variation in
TABLE 10
Biological Properties of Genetic Loci Sequenced for This Studya
Gene Protein product

Locationb Active in Location Primary function
BRCA1 chr 2 many tissues nucleus & cytoplasm tumor suppression

DMP1 chr 5 dentine & bone extracellular mineralization
IRBP chr 1 retina extracellular photoreception
RAG1 chr 5 developing B & T lymphocytes nucleus immune response
vWF chr 8 endothelium & megakaryocytes extracellular blood clotting
a
As indicated primarily by research on humans and other placental taxa (see references cited in text).
b
Chromosome number in Monodelphis domestica genome.
GC (versus AT) content, is now widely tion (C), and a proportion of invariant sites
recognized as a potential problem for phylo- (pinv) applies to three genes. Fitted base-
genetic inference (see literature reviewed by compositional parameters (pA, pC, pG, pT)
Gruber et al., 2007). Of the five genes for GTR-modeled loci are relatively even for
analyzed herein, BRCA1 and DMP1 exhibit IRBP, slightly skewed towards A for the
the lowest overall GC content (fig. 26). For RAG1 partition, strongly skewed towards A
BRCA1, GC content at all codon positions is for BRCA1, and strongly biased towards A
remarkably consistent across taxa. However, and G for DMP1. All five genes exhibit
GC content at third codon positions (GC3) of higher rates of transitions than transversions;
DMP1 is substantially higher in two taxa in addition, first and second positions of
(Hyladelphys and Tlacuatzin) than it is in RAG1 exhibit a relatively high rate of A-to-C
other didelphids (Jansa and Voss, 2005). As transversions. The degree of among-site rate
reported elsewhere, third codon positions of heterogeneity was lowest for BRCA1 (a
RAG1 are so highly variable in GC content 51.48), and highest for the RAG1 partition
among didelphids that plausible phylogenetic (a 5 0.41).
results cannot be recovered from datasets
that include these positions (Gruber et al., PHYLOGENETIC RELATIONSHIPS
2007); therefore, only first and second codon
positions of RAG1 are analyzed below. The The six data partitions recognized herein
GC content of vWF is most similar to that of (nonmolecular characters plus nucleotide
RAG1, but vWF does not exhibit the sequences from five loci) were analyzed
exceptionally high taxonomic variation in separately and in combination to explore
GC3 seen in the latter gene. Compared to the relevant patterns of support for monophylet-
other loci, IRBP has relatively high GC ic groups. To assess the sensitivity of our
content at all codon positions, with one results to missing data and to different
taxon (Cryptonanus chacoensis) exhibiting analytic methods, we conducted MP and
elevated GC content at first and second Bayesian analyses of the nonmolecular data
codon positions. with and without Chacodelphys (4 analyses);
Additional sequence characteristics can be MP, ML, and Bayesian analyses of each
effectively summarized in the context of single-gene dataset (15 analyses); MP, ML,
model fitting (Posada and Crandall, 1998). and Bayesian analyses of a concatenated-
The best-fitting model of nucleotide substi- gene dataset (3 analyses); and MP, ML, and
tution for each gene is either a general-time- Bayesian analyses of a combined (nonmolec-
reversible (GTR) model or a modification ular + molecular) dataset with and without
thereof (TVMef) that assumes equal base Chacodelphys (6 analyses). Given the com-
frequencies and equal transition rates (ta- plexity of these results (a total of 28 separate
ble 11). In addition, the best-fitting model for analyses), the following accounts are focused
each gene includes among-site rate heteroge- on issues of generic monophyly and on
neity as approximated by a gamma distribu- higher-level (intergeneric) relationships among
Fig. 26. Scatterplot of percent GC content among didelphid sequences at first and second codon
positions versus percent GC content at third codon positions for the five genes included in this study.
didelphid terminal taxa.21 To simplify com- relationship to type I error rates (Erixon et al.,
parisons of support values among the many 2003; Alfaro and Holder, 2006).
(.1000) ingroup nodes recovered by these
analyses, we restrict our attention to boot- Analyses of Nonmolecular Data
strap frequencies (from MP and ML analyses)
and Bayesian posterior probabilities, both of Despite the large amount of missing data
which have respectable statistical pedigrees from Chacodelphys (table 2), including this
(Felsenstein, 1985; Efron et al., 1996; Huel- taxon in a parsimony analysis of our
senbeck et al., 2002) despite their imperfect nonmolecular (morphological + karyotypic)
characters resulted in fewer equally parsimo-
nious trees (with a correspondingly better-
21
Marsupial interordinal relationships and relationships resolved strict consensus) than we obtained
among congeneric didelphid species will be treated in from a parsimony analysis with Chacodelphys
subsequent reports. excluded (table 12). Among other notewor-
TABLE 11
Results of Maximum-likelihood Model Fitting with Single-gene Datasets
BRCA1 vWF IRBP DMP1 RAG1a
Model GTR + C TVMef + I + C GTR + I + C GTR + C GTR + I + C

Parameter estimates:
pA 0.37 0.25 0.25 0.37 0.31
pC 0.19 0.25 0.26 0.18 0.23
pG 0.21 0.25 0.27 0.26 0.24
pT 0.23 0.25 0.22 0.18 0.22
rAC 0.91 2.03 1.96 1.12 3.03
rAG 4.75 7.01 5.88 2.64 4.59
rAT 0.76 1.26 1.46 0.39 0.77
rCG 1.05 1.07 1.13 0.77 1.73
rCT 5.52 7.01 7.43 6.74 7.22
a 1.48 0.83 0.81 0.69 0.41
pinv 0.00 0.19 0.32 0.00 0.52
Ln-likelihoodb 215078.42 26648.62 26817.82 25943.66 26343.58
a
First and second codon positions only.
b
Of best tree.
thy features of this topology (fig. 27), didel- However, several other groups (e.g., the
phid monophyly is only weakly supported, as genus Thylamys, the genus Marmosops, the
are most patterns of intergeneric relation- genus Caluromys, the genus Philander, and
ships. In fact, only six genera and intergeneric Philander + Didelphis) receive moderate
clades are strongly supported (with bootstrap bootstrap support.
values .75%) by parsimony analysis of these Bayesian analysis of these nonmolecular
data: (1) a group containing Lestodelphys and data (implementing the Mkv model of Lewis,
Thylamys; (2) a group that includes Calur- 2001) provides strong support (posterior
omys and Caluromysiops; (3) the genus probabilities .95%) for most of the same
Didelphis; (4) a group that includes Lutreo- clades that are strongly supported by parsi-
lina, Philander, and Didelphis; (5) a group mony, with the notable exception of (4)
that includes the latter three genera plus among those listed above. In addition, there
Chironectes; and (6) the genus Monodelphis. is strong Bayesian support for the monophy-
TABLE 12
Results of Parsimony Analyses of Datasets with Nonmolecular Characters
Nonmolecular dataa Combined datab

w/ Chacodelphys w/o Chacodelphys w/ Chacodelphys w/o Chacodelphys
Ingroup taxa 44 43 44 43
Outgroup taxa 7 7 7 7
Total characters 129 129 7449 7449
Informative characters 114 114 2073 2073
Minimum-length trees 1068 21,803 12 12
Length of best trees 362 356 5960 5954
Consistency Index 0.508 0.517 0.638 0.639
Retention Index 0.788 0.793 0.777 0.777
Resolved nodes (%)c 37 (77.1) 27 (57.4) 44 (91.7) 43 (91.5)
a
Morphological and karyotypic characters.
b
Nonmolecular data plus concatenated sequences from five loci.
c
In strict consensus of minimum-length trees.
Fig. 27. The strict consensus of 1068 minimum-length trees resulting from parsimony analysis of the
nonmolecular (morphological + karyotypic) dataset with Chacodelphys included. Divided circles at each
resolved node indicate support from parsimony bootstrap and Bayesian analyses. For parsimony
bootstrap analysis (MP), white indicates bootstrap frequencies #50%, grey indicates bootstrap frequencies
between 50% and 75%, and black indicates bootstrap frequencies $75%. For the Bayesian analysis (BPP),
white indicates posterior probabilities ,0.95, whereas black indicates posterior probabilities $0.95.
ly of Marmosops (only moderately supported includes the latter four genera plus Marmo-
by parsimony). However, Bayesian support is sops; and a group that includes Marmosa,
weak for several nodes with moderate parsi- Monodelphis, and Tlacuatzin. As reported
mony bootstrap support, including the genus elsewhere (Jansa and Voss, 2005), there are
Thylamys, the genus Caluromys, and the no examples of hard incongruence to be
genus Philander. found in comparing analytic results between
Most of the obvious differences between IRBP and DMP1 because all conflicting
these results and our previously published nodes (e.g., those resolving the position of
analyses of didelphid morphological charac- Hyladelphys) have weak support from one or
ter data are attributable to rooting. Whereas both genes.
we previously (Voss and Jansa, 2003; Voss et Analyses of first and second codon posi-
al., 2004a; Jansa and Voss, 2005) rooted our tions of RAG1 (fig. 30) support the mono-
estimates of didelphine relationships using phyly of most polytypic didelphid genera
Caluromys, Caluromysiops, and Glironia as (except Caluromys, Didelphis, and Philander),
outgroups, these analyses are rooted using but some genera recovered as monophyletic
nondidelphid marsupial outgroups. Consid- groups (e.g., Thylamys, Marmosops, and
ered as an undirected network, the ingroup Monodelphis) do not receive consistently
topology of figure 27 is almost identical to strong support. Many of the same interge-
that obtained by Jansa and Voss (2005: fig. neric relationships supported by IRBP and
1C) for the same set of didelphid terminal DMP1 were also recovered, notably includ-
taxa. ing the branching patterns among Meta-
chirus, Chironectes, Lutreolina, and Philander
Single-gene Analyses + Didelphis. Additionally, Marmosops was
recovered as the sister group to a clade that
Maximum-likelihood topologies with su- includes Cryptonanus, Gracilinanus, and Thy-
perimposed summaries of nodal statistics lamys. Unfortunately, the uncertain position
from three analyses (MP, ML, and Bayesian) of the root (probably an artifact of incom-
of each single-gene dataset (figs. 2832) plete outgroup sequences; see above) is
illustrate similarities and differences in clade reflected in the lack of strong support for
support among the five loci we sequenced. any basal relationships in this topology.
Because our analytic results for IRBP and Analyses of sequence data from BRCA1
DMP1 have already been published (Jansa produced the most compelling single-gene
and Voss, 2005), only summary comments estimate of didelphid relationships recovered
are warranted here. Briefly, the monophyly to date, with strikingly consistent high nodal
of all genera represented by two or more support values and an almost completely
terminal taxa is strongly supported by both resolved consensus topology (fig. 31).
of these genes (figs. 28, 29) with the exception Among other salient features, every polytypic
of Gracilinanus (weakly to moderately sup- opossum genus with the exception of Didel-
ported by IRBP, not recovered as monophy- phis was recovered with strong support, as
letic by DMP1), Thylamys (only weakly were almost all of the higher-level clades
supported by Bayesian analysis of DMP1), recovered with strong support by any anal-
and Didelphis (inconsistently recovered as ysis of other single-gene datasets (the only
monophyletic by analyses of IRBP). On the exceptions are groups that include Lestodel-
assumption that the DMP1 tree (lacking phys and Caluromysiops, taxa from which
outgroup terminals) is appropriately rooted BRCA1 sequences are not available). Three
within a node or two of Glironia, several of these merit particular attention. One is
patterns of intergeneric relationships are also clade H of Jansa and Voss (2000), which
supported by both genes, including the nested unites the Thylamys cluster with Metachirus
clades (Metachirus (Chironectes (Lutreolina and the large opossums with 22 chromo-
(Didelphis + Philander)))); a group that somes (Chironectes, Lutreolina, Philander,
includes Lestodelphys and Thylamys; a group and Didelphis). Another is the subfamily
that includes Cryptonanus, Gracilinanus, and Didelphinae of traditional usage (excluding
Lestodelphys + Thylamys; a group that Hyladelphys, Caluromys, and Glironia), and
Fig. 28. The tree resulting from maximum-likelihood analysis of the IRBP dataset under its best-fitting
model (GTR + I + C, ln-likelihood 5 26817.82; table 11). Support statistics from a parsimony bootstrap
analysis, a maximum-likelihood bootstrap analysis, and a Bayesian analysis are indicated at each resolved
node. For the parsimony and maximum-likelihood analyses (MP and ML, respectively), white indicates
bootstrap frequencies #50%, grey indicates bootstrap frequencies between 50% and 75%, and black
indicates bootstrap frequencies $75%. For the Bayesian analysis (BPP), white indicates posterior
probabilities ,0.95, whereas black indicates posterior probabilities $0.95.
Fig. 29. The tree resulting from maximum-likelihood analysis of the DMP1 dataset under its best-
fitting model (GTR + C, ln-likelihood 5 5943.66; table 11). Four of the five deletion events that optimize
as synapomorphies on this topology are indicated with vertical gray bars; the fifth (not shown) is shared by
Marmosa regina, M. paraguayana, M. demerarae, M. murina, and M. lepida, but this clade is subtended by
a zero-length branch. Because alignable DMP1 sequences are not available from outgroup taxa (see text),
this tree is rooted with Glironia. Conventions for indicating nodal support are described in the caption to
figure 28. The broken arrow indicates the alternative position of Tlacuatzin recovered by ML bootstrap
analysis (the position for this taxon recovered by ML analysis of the original data was not supported
by bootstrapping).
Fig. 30. The tree resulting from maximum-likelihood analysis of first and second codon positions of
RAG1 under its best-fitting model (GTR + I + C, ln-likelihood 5 26343.58; table 11). The single
synapomorphic deletion event in this gene is indicated with a vertical gray bar. Conventions for indicating
nodal support are described in the caption to figure 28. Asterisks indicate nodes resolved by ML analysis
of the original data that lack bootstrap support.
Fig. 31. The tree resulting from maximum-likelihood analysis of the BRCA1 dataset under its best-
fitting model (GTR + C, ln-likelihood 5 215078.42; table 11). Vertical bars indicate unique and
unreversed synapomorphic insertions (black) or deletions (gray) in this gene. Conventions for indicating
nodal support are described in the caption to figure 28. The broken arrow indicates the alternative position
of Glironia recovered by parsimony analysis.
the third is Hyladelphys + Didelphinae. A nodes not convincingly resolved by BRCA1

novel result (uniquely recovered with strong are the one that clusters Cryptonanus with
support by BRCA1 but not strongly contra- Gracilinanus (recovered with only weak to
dicted by any other gene) is the sister-group moderate support by MP, ML, and Bayesian
relationship between Marmosa and Tlacuat- analyses) and the position of the ingroup
zin. In effect, the only relevant intergeneric root. The latter was placed between Glironia
Fig. 32. The tree resulting from maximum-likelihood analysis of the vWF dataset under its best-fitting
model (TVMef + I + C, ln-likelihood 5 26648.62; table 11). Conventions for indicating nodal support are
described in the caption to figure 28. The broken arrow indicates the alternative position of Marmosa
rubra supported by Bayesian analysis.
TABLE 13
Results of Parsimony Analyses of Six Molecular Datasets
BRCA1 vWF IRBP DMP1 RAG1a Concatenated
Ingroup taxa 41 41 43 42 41 43
Outgroup taxa 7 7 7 0 5 7
Aligned sites 2163 963 1158 1176 1860 7320
Variable sitesb 753 324 267 444 315 2103
% variable sitesb 34.8 33.6 23.1 37.8 16.9 28.7
Informative sitesb 443 202 161 246 210 1262
Minimum-length trees 8 60 121 30 717 8
Length of best trees 2297 967 947 735 584 5583
Consistency Index 0.67 0.60 0.64 0.72 0.66 0.65
Retention Index 0.80 0.75 0.77 0.81 0.80 0.78
Resolved nodes (%) 43 (92.5) 40 (87.0) 38 (79.2) 34 (85.0) 15 (34.1) 45 (92.8)
a
First and second codon positions only.
b
Among ingroup taxa.
c
Among ingroup taxa in strict consensus of minimum-length trees.
and Caluromys with moderate support from by analyses of DMP1). Although Caluromys
MP, but it was placed between Caluromys + was not recovered as monophyletic by
Glironia and other didelphids with weak analyses of RAG1 sequences, this result is
support in ML and Bayesian analyses. plausibly attributable to the unstable position
Analyses of sequence data from vWF of the ingroup root.
(fig. 32) strongly support the monophyly of Several patterns of intergeneric relation-
all polytypic didelphid genera except Didel- ships were also consistently recovered, usu-
phis and Philander. In addition, they strongly ally with strong support. These include the
support some of the same intergeneric clades nested clades (Chironectes (Lutreolina (Didel-
common to our other single-gene results, phis + Philander))) and a group comprised of
including (Chironectes (Lutreolina (Philander Thylamys, Gracilinanus, and Cryptonanus. In
+ Didelphis))); a group that includes Thyla- addition, both loci (IRBP, DMP1) from
mys, Gracilinanus, and Cryptonanus; and which sequence data were available for
another that includes Marmosa, Monodel- Lestodelphys support a sister-group relation-
phis, and Tlacuatzin. A novel resultnot ship between that genus and Thylamys.
strongly supported by any other gene and Lastly, all of the recovered single-gene trees
strongly contradicted by our RAG1 results are consistent with our assumption of didel-
is a sister-group relationship between Graci- phid monophyly.
linanus and Thylamys. No other intergeneric With a single exception, no pattern of
clades are consistently strongly supported by intergeneric relationships strongly supported
parsimony, likelihood, or Bayesian analyses by any analysis of these genes was strongly
of vWF sequence data; indeed, most are contradicted by another. Examples of rela-
weakly supported by two or more analyses. tionships strongly supported by some analy-
In summary, most didelphid genera were ses and not strongly contradicted by any
recovered as monophyletic groups with other include the sister-group relationship of
strong nodal support values from most genes Marmosops to Thylamys + Gracilinanus +
(table 14). Support statistics suggest that the Cryptonanus; the sister-group relationship of
most problematic genera are Didelphis (re- Metachirus to the large opossums with 22
covered as monophyletic with strong support chromosomes; a group that includes Mar-
only by analyses of DMP1), Philander mosa, Monodelphis, and Tlacuatzin; the
(inconsistently recovered as monophyletic group Marmosa + Tlacuatzin; the subfamily
by analyses of RAG1 and vWF), and Didelphinae; and the clade Hyladelphys +
Gracilinanus (not recovered as monophyletic Didelphinae. In effect, the only example of
TABLE 14
Support for Generic Monophyly from Single-gene Datasetsa
IRBP DMP1 RAG1 BRCA1 vWF
Caluromys ** ** NR ** **
Cryptonanus ** ** ** ** **
Didelphis NR ** NR NR NR
Gracilinanus R NR ** ** **
Marmosa ** ** ** ** **
Marmosops ** ** * ** **
Monodelphis ** ** * ** **
Philander ** ** NR ** NR
Thylamys ** * * ** **
a
Key: ** 5 strong support from all analyses (MP, ML, Bayesian); * 5 strong support from at least one analysis; R 5
recovered without strong support; NR 5 not recovered.
hard incongruence among patterns of inter- ingroup root (fig. 34A, B, C) suggest that the
generic relationships supported individually parsimony and likelihood solutions are al-
by these loci concerns the relationships of most equiprobable but slightly favor the
Cryptonanus. This obviously problematic latter.
taxon was recovered with strong nodal The second equivocal node concerns the
support either as the sister group of Gracili- relationships of Marmosa, Tlacuatzin, and
nanus emiliae (in all analyses of DMP1), or as Monodelphis. Although Marmosa and Tla-
the sister group of Gracilinanus (in all analyses cuatzin were recovered as sister taxa, support
of RAG1), or as the sister group of Thylamys for this clade was uniformly weak. Bayesian
+ Gracilinanus (in all analyses of vWF). posterior probabilities for all possible resolu-
tions of this node (fig. 34D, E, F), however,
Analyses of Concatenated Genes suggest a somewhat clearer ranking than was
obtained for alternative placements of the
Maximum parsimony, maximum likeli- ingroup root.
hood, and Bayesian analyses of concatenated The third problematic issue concerns
sequence data from all five genes (7320 bp) relationships among Gracilinanus, Cryptona-
resulted in well-resolved and strikingly sim- nus, and Thylamys + Lestodelphys. Here,
ilar trees with high support values at most parsimony provides moderate bootstrap sup-
nodes (fig. 33). Among other analytic simi- port (58%) for grouping Gracilinanus with
larities, all polytypic genera were recovered Cryptonanus as does likelihood (70%). Al-
as monophyletic groups, as were all of the though Bayesian support for Gracilinanus +
higher-level clades common to two or more Cryptonanus is weak, posterior probabilities
of the single-gene analyses discussed above. clearly favor this clade over other phyloge-
However, three intergeneric nodes remain netic alternatives (fig. 34G, H, I).
weakly or inconsistently supported.
The first of these concerns the position of Analyses of Combined Datasets
the ingroup root. Whereas maximum parsi-
mony weakly supports placing the root Maximum parsimony and Bayesian anal-
between Glironia and other didelphids (which yses of the combined (nonmolecular +
were recovered as a monophyletic group with molecular) data without Chacodelphys recov-
a bootstrap frequency of 74%), maximum er the same higher-level topologies that were
likelihood places the root between (Glironia obtained from analyses of the concatenated-
(Caluromys + Caluromysiops)) and Hyladel- gene data, with only a few noteworthy
phys + Didelphinae; however, likelihood changes in nodal support (fig. 35). For
bootstrap support for the former group is example, parsimony bootstrap support for
weak (,50%). Bayesian posterior probabili- the monophyly of Didelphis increases slightly
ties for three alternative placements of the when nonmolecular character data are added
Fig. 33. The tree resulting from a mixed-model maximum-likelihood analysis of the concatenated (five-
gene) dataset. Conventions for indicating nodal support are described in the caption to figure 28. The
broken arrow indicates the alternative position of Glironia recovered by parsimony analysis.
Fig. 34. Alternative resolutions for three patterns of intergeneric relationships that exhibit nodal
posterior probabilities ,0.95 in the Bayesian analysis of the concatenated five-gene dataset (see fig. 33). A,
B, C, alternative relationships among Glironia, Hyladelphys + Didelphinae, and Caluromys +
Caluromysiops; D, E, F, alternative relationships among Tlacuatzin, Marmosa, and Monodelphis; G, H,
I, alternative relationships among Cryptonanus, Thylamys + Lestodelphys, and Gracilinanus. Black and
grey circles at nodes indicate posterior probabilities $0.95 or ,0.95, respectively. Branch lengths are
shown at the same scale across all trees.
(from 80% to 93%), as does the Bayesian nodes are substantially unaffected by taxon
posterior probability for placing the ingroup addition.
root between (Glironia (Caluromys + Calur-
omysiops)) and other didelphids (from 0.51 to Discussion
0.87). Similarly, adding nonmolecular char-
acters marginally improves the Bayesian Analyses of our combined dataset (exclud-
posterior probability for Marmosa + Tlacuat- ing Chacodelphys; fig. 35) effectively summa-
zin (from 0.63 to 0.83). rize all of the shared phylogenetic signal that
By contrast, including Chacodelphys in a we have found to date in morphology,
combined-data parsimony analysis dramati- karyotypes, and five unlinked protein-coding
cally erodes bootstrap support for five nuclear genes. Additional support is provided
intergeneric nodes (fig. 36). In fact, the strict by 16 parsimony-informative indels (figs. 29,
consensus of 12 minimum-length trees (not 30, 31) that were not coded as phylogenetic
shown) places Chacodelphys as the sister characters but which optimize as synapomor-
group of Monodelphis with weak (MPBS 5 phies on this topology. Lastly, this tree is
49%) support. Bayesian analysis, however, congruent with the results of independent
strongly supports the membership of Chaco- phylogenetic analyses based on sequence data
delphys in a clade that also includes Gracili- from intron 1 of the nuclear transthyretin
nanus, Cryptonanus, and Thylamys + Lesto- (TTR) gene and the mitochondrial 12S
delphys, and posterior probabilities for other rDNA gene (Steiner et al., 2005). To the best
Fig. 35. The 50% majority-rule consensus of post-burnin trees resulting from a mixed-model Bayesian
analysis of the combined (nonmolecular + molecular) dataset without Chacodelphys. Conventions for
indicating nodal support are described in the caption to figure 28. The broken arrow indicates the
alternative position of Glironia recovered by parsimony analysis.
Fig. 36. The 50% majority-rule consensus of post-burnin trees resulting from a mixed-model Bayesian
analysis of the combined (nonmolecular + molecular) dataset including Chacodelphys. Conventions for
indicating nodal support are described in the caption to figure 28. The broken arrow indicates the
alternative position of Glironia recovered by parsimony analysis.
of our knowledge, no strongly conflicting tionary interest. Phenotypically, Cryptonanus

valid results have been obtained in any other is certainly more similar to Gracilinanus than
published phylogenetic analysis of molecular it is to Thylamys or Lestodelphys, whereas
or morphological character data.22 Although Tlacuatzin is undeniably more similar to
this tree is fully resolved with high support Marmosa than it is to Monodelphis. The
values at most nodes, several outstanding positions of these genera in our combined-
problems merit comment. data trees are therefore plausible despite the
As we have previously noted elsewhere, absence of compelling parsimony or Bayesian
Chacodelphys exhibits conflicting patterns of support.
derived morphological similarities with Thy- Although it might seem best to wait until
lamys + Lestodelphys on the one hand and these few remaining issues are convincingly
with Monodelphis on the other (Voss et al., resolved before proposing a formal classifi-
2004a). Weak parsimony support for nodes cation, there is no guarantee that additional
along the phylogenetic path between Thyla- data will soon be forthcoming or useful. In
mys + Lestodelphys and Monodelphis in the the meantime, other relationships that are
combined-data analysis that includes Chaco- strongly supported by our results merit
delphys (fig. 36) presumably reflects such nomenclatural recognition, and the contents
character conflict, although it is noteworthy of some taxa currently recognized as valid
that Bayesian support for the same nodes is need to be revised on the basis of our analytic
unaffected by taxon addition. Although we results. As documented below, no previous
are convinced that Chacodelphys is closely classification is consistent with what is now
related to Lestodelphys and Thylamys by a confidently known about didelphid phyloge-
host of phenotypic resemblances too indefi- netic relationships.
nite to code as characters but too numerous
to ignore, a compelling analytic solution to CLASSIFICATION
this vexing problem is unlikely to be forth-
coming until at least some of the missing Most essays on marsupial classification
molecular data for Chacodelphys can be (e.g., Simpson, 1945; Kirsch, 1977a; Mar-
obtained from fresh material. shall, 1981; Aplin and Archer, 1987; Archer
However, even substantial amounts of and Kirsch, 2006) are uninformative about
sequence data are clearly not enough to the historical development of opossum sys-
resolve other phylogenetic uncertainties. In- tematics. To be sure, the chronological and
deed, it is not a little frustrating that, with bibliographic details of didelphid taxonomy are
.7000 bp of protein-coding nuclear sequence of limited interest, so the following paragraphs
in hand, the relationships of Glironia, Cryp- mention only a few milestones on the long road
tonanus, and Tlacuatzin should still be to the currently accepted classification. Because
problematic. Glironia is of special concern, most of the technical information about how
because the two plausible phylogenetic reso- and why names were formerly applied to taxa
lutions for this genus (fig. 34A, B) determine was recently summarized by Gardner (2008),
the root of the didelphid radiation. The this brief narrative is primarily intended to
alternative phylogenetic resolutions of Cryp- serve as an introduction to the revised phylo-
tonanus and Tlacuatzin within their respective genetic system that we propose.
groups do not seem like comparably weighty Linnaeus (1758) described only five species
issues, but each could affect ecobehavioral of marsupials, all of which were didelphids.
character optimizations of significant evolu- Although the four Linnaean species currently
recognized as valid (marsupialis, philander,
22
DNA-DNA hybridization results showing Gracilinanus opossum, and murina) are now referred to
nested within Marmosops (Kirsch and Palma, 1995; Kirsch et al., different genera, the great Swede placed all of
1997) were based on taxonomic misidentifications (Voss and them in the genus Didelphis. New generic
Jansa, 2003: 57). Analyses of 12S gene sequences by Palma and names for opossums proliferated over subse-
Spotorno (1999) recovered Metachirus and Marmosops as sister
taxa, but this grouping was not found in subsequent analyses of
quent decades of the 18th and 19th centuries
12S sequence data (Steiner et al., 2005) for reasons that remain (table 15), but no consistent binomial usage
unexplained. had emerged prior to Thomass (1888) land-
TABLE 15
Chronological List of Available Generic Names for Recent Didelphids
(Boldface indicates taxa recognized as valid genera in this report.)
Type speciesa Current status (if not valid)

b
Didelphis Linnaeus, 1758 Didelphis marsupialis Linnaeus, 1758
Philander Brisson, 1762 Didelphis opossum Linnaeus, 1758c
Chironectes Illiger, 1811 Lutra minima Zimmerman, 1780d
Memina G. Fischer, 1814 Lutra minima Zimmermann, 1780d synonym of Chironectes Illiger, 1811
Opossum Schmid, 1818 Didelphis marsupialis Linnaeus, 1758e synonym of Didelphis Linnaeus, 1758
Sarigua Muirhead, 1819 Didelphis marsupialis Linnaeus, 1758f synonym of Didelphis Linnaeus, 1758
Marmosa Gray, 1821 Didelphis murina Linnaeus, 1758d
Monodelphis Burnett, 1830 Didelphis brachyuros Schreber, 1777g
Asagis Gloger, 1841 Didelphis murina Linnaeus, 1758b synonym of Marmosa Gray, 1821
Micoureus Lesson, 1842 Didelphis cinerea Temminck, 1824b subgenus of Marmosa Gray, 1821
Peramys Lesson, 1842 Didelphis brachyuros Schreber, 1777b synonym of Monodelphis Burnett, 1830
Thylamys Gray, 1843 Didelphis elegans Waterhouse, 1839d
Grymaeomys Burmeister, 1854 Didelphis murina Linnaeus, 1758b synonym of Marmosa Gray, 1821
Metachirus Burmeister, 1854 Didelphis nudicaudata E. Geoffroy, 1803h
Microdelphys Burmeister, 1856 Didelphis tristriata Olfers, 1818b synonym of Monodelphis Burnett, 1830
Cuica Liais, 1872 Didelphis murina Linnaeus, 1758i synonym of Marmosa Gray, 1821
Gamba Liais, 1872 Didelphis palmata Daudin, 1799i synonym of Chironectes Illiger, 1811
Caluromys J.A. Allen, 1900 Didelphis philander Linnaeus, 1758i
Lutreolina Thomas, 1910 Didelphis crassicaudata Desmarest, 1804d
Glironia Thomas, 1912 Glironia venusta Thomas, 1912j
Leucodidelphis Ihering, 1914 Didelphis paraguayensis J.A. Allen, 1902j synonym of Didelphis Linnaeus, 1758
Marmosops Matschie, 1916 Didelphis incana Lund, 1840j
Metachirops Matschie, 1916 Didelphis quica Temminck, 1824j synonym of Philander Brisson, 1762
Monodelphiops Matschie, 1916 Microdelphys sorex Hensel, 1872j synonym of Monodelphis Burnett, 1830
Holothylax Cabrera, 1919 Didelphis opossum Linnaeus, 1758j synonym of Philander Brisson, 1762
Minuania Cabrera, 1919 Didelphis dimidiata Wagner, 1847j synonym of Monodelphis Burnett, 1830
Mallodelphys Thomas, 1920 Didelphis laniger Desmarest, 1820j subgenus of Caluromys J.A. Allen, 1900
Lestodelphys Tate, 1934 Notodelphys halli Thomas, 1921j
Caluromysiops Sanborn, 1951 Caluromysiops irrupta Sanborn, 1851j
Stegomarmosa Pine, 1972 Marmosa (Stegomarmosa) andersoni Pine, synonym of Marmosa Gray, 1821
1972d
Gracilinanus Gardner & Creighton, 1989 Didelphis microtarsus Wagner, 1842j
Grayium Kretzoi & Kretzoi, 2000 Didelphis murina Linnaeus, 1758k synonym of Marmosa Gray, 1821
Hyladelphys Voss et al., 2001 Gracilinanus kalinowskii Hershkovitz, 1992j
Tlacuatzin Voss & Jansa, 2003 Didelphis (Micoureus) canescens J.A. Allen,
1893j
Chacodelphys Voss et al., 2004 Marmosa formosa Shamel, 1930j
Cryptonanus Voss et al., 2005 Marmosa agilis chacoensis Tate, 1931j
a
For simplicity we have emended all nomenclaturally irrelevant mistakes made by authors, such as incorrect spellings
(e.g., Didelphys for Didelphis) and misattributed authorship and dates for nominal species otherwise eligible for type
fixation (ICZN, 1999: Article 67).
b
Subsequently designated by Thomas (1888) who, however, sometimes used senior synonyms to denote originally
included nominal species (e.g., brevicaudata Erxleben for brachyuros Schreber as the type species of Peramys Lesson).
c
By plenary action of the International Commission on Zoological Nomenclature (ICZN, 1998).
d
By monotypy.
e
Subsequently designated by Brown (2004).
f
Subsequently designated by Cerqueira and Tribe (2008).
g
Subsequently designated by Matschie (1916).
h
Subsequently designated by Thomas (1888); see Hershkovitz (1976).
i
Subsequently designated by Hershkovitz (1949), but authorship and dates misattributed by him are emended herein.
j
By original designation.
k
Grayium was proposed as a replacement name for Marmosa Gray, 1821, on the mistaken assumption that the latter
was preoccupied by Marmosa Rafinesque, 1815, a nomen nudum (Creighton and Gardner, 2008b).
TABLE 16
Classifications of Recent Opossums, 18881958a
Thomas (1888) Matschie (1916) Cabrera (1919) Simpson (1945) Cabrera (1958)
Didelphidae Didelphidae Didelphidae Didelphidae Didelphidae

Didelphis Didelphis Dromiciops Philander Caluromys
Didelphisb Didelphisb Glironia Monodelphis Caluromysiops
Metachirusb Metachiropsb Philander Dromiciops Glironia
Philanderb Metachirusb Marmosa Glironia Dromiciops
Micoureusb Peramysb Marmosab Notodelphisc Monodelphis
Peramysb Micoureusb Thylamysb Marmosa Monodelphisb
Chironectes Caluromysb Peramys Metachirops Minuaniab
Marmosab Minuania Metachirus Lestodelphys
Grymaeomysb Lutreolina Lutreolina Marmosa
Marmosopsb Metachirus Didelphis Marmosab
Thylamysb Holothylax Chironectes Thylamysb
Dromiciopsb Didelphis Philander
Glironiab Chironectes Metachirus
Monodelphisb Lutreolina
Monodelphiopsb Didelphis
Microdelphysb Chironectes
[Chironectes]
a
Incorrect spellings (e.g., Didelphys for Didelphis; Thomas, 1888) have been changed to conform with current usage
throughout, but the sequence of names in each classification has been preserved as originally published. Note that some
names (e.g., Philander) were used by early taxonomists for species that are now placed in other genera (see text).
b
Ranked as subgenera.
c
Not an available name. Notodelphys Thomas, 1921, originally proposed for the taxon now known as Lestodelphys, is
preoccupied by Notodelphys Allman, 1847, a copepod (Tate, 1934).
mark catalog of the marsupials in the British Thomas, 1894) was also included. Matschies
Museum of Natural History. taxonomy is noteworthy for his early recog-
By comparison with Old World marsupi- nition that the so-called murine opossums
als, didelphids appeared to Thomas (1888: (pouchless mouse- and rat-sized species with
315) to be an exceedingly homogeneous circumocular masks and long tails) that
[family], its members presenting a very small Thomas (1888) had lumped together in
range of differentiation. Accordingly, he Micoureus actually represent several distinct
recognized only Didelphis and Chironectes as groups, and for treating those groups as
full genera, but several other taxa were ranked coordinate taxa.
as subgenera of Didelphis (table 16). These Cabreras (1919) classification was not the
included Metachirus (containing species now first to reject Linnaeuss inclusive concept of
referred to Metachirus, Philander, and Lu- Didelphis, but it was influential in establish-
treolina), Philander (for Caluromys), Micour- ing modern binomial usage. His taxonomy
eus (containing species now referred to included 11 genera, only one of which
Marmosa, Marmosops, Thylamys, and Graci- contained subgenera. Like Thomas (1888)
linanus), and Peramys (for Monodelphis). and Matschie (1916), Cabrera made no use of
Although knowledge of didelphid diversity subfamilies, tribes, or other suprageneric
increased rapidly in the years following categories to indicate relationships among
Thomass classification, Matschie (1916) per- living opossums. Unlike Matschie, Cabrera
sisted in referring all nonaquatic opossums to grouped all of the murine opossums
the genus Didelphis. However, Matschie together, using Marmosa as the name for
recognized more subgenera of Didelphis than the taxon that Thomas called Micoureus.
Thomas did, resurrecting old names or Simpson (1945) recognized several sub-
describing new ones to suit his needs. families of didelphids, but he placed all of the
Dromiciops (described as a didelphid by living opossums (plus Dromiciops) in the
subfamily Didelphinae, and he did not use The classification proposed by Hershko-
tribal nomenclature to indicate relationships vitz (1992b) was not supported by any
among them. In effect, his treatment of phylogenetic analysis, nor were phylogenetic
Recent didelphids amounted to no more criteria cited to support any of its novel
than an alphabetic list of generic names, features, including the recognition of four
conveying even less information about in- groups of Recent opossums as separate
ferred relationships than the classifications of families (Marmosidae, Caluromyidae, Gliro-
his predecessors. niidae, Didelphidae). Taxonomic rank aside,
Cabreras (1958) checklist of South Amer- these taxa represent significant departures
ican mammals was one of the last attempts to from previous classifications by grouping
classify Recent opossum diversity by tradi- Metachirus with Monodelphis and the mu-
tional (prephylogenetic) criteria, and it re- rine opossums, by the absence of any formal
mained more or less unchallenged until the relationship between this group (Marmosi-
advent of molecular systematics in the mid- dae) and the large opossums with 22 chro-
1970s. Despite the fact that at least some mosomes (Didelphidae), and by the separa-
contemporaneous researchers (e.g., Reig, tion of Glironia on the one hand from
1955) recognized suprageneric groups among Caluromys and Caluromysiops on the other.
living opossumlike marsupials, Cabrera did Also unlike other classifications, four mu-
not. Instead, only his taxonomic sequence rine genera (Gracilinanus, Marmosops, Mar-
and his use of subgenera for Monodelphis and mosa, Micoureus) were associated in a single
Marmosa hint at inferred patterns of rela- taxon (Marmosinae) that excluded Thylamys
tionships. Most of the generic names that and Lestodelphys.
Cabrera recognized as valid are still used in Kirsch and Palma (1995) formalized the
the same sense today, with the conspicuous results of their DNA-DNA hybridization
exception of Marmosa: whereas the nomino- experiments in a classification that, for the
typical subgenus contained (in addition to first time, associated Metachirus (as a mono-
Marmosa sensu stricto) species now referred typic tribe) with the large 2n 5 22 opossums
to Marmosops, the subgenus Thylamys con- (Didelphini), associated Monodelphis with
tained (in addition to Thylamys sensu stricto) Marmosa and Micoureus, and recognized
species now referred to Chacodelphys, Cryp- the other murine opossums (Gracilinanus,
tonanus, and Gracilinanus. Lestodelphys, Marmosops, Thylamys) as
Reig et al. (1985) provided the first classi- forming a separate clade. All of these taxa
fication of opossumlike marsupials based on were referred to the family Didelphidae,
an explicitly phylogenetic analysis (subse- whereas Glironia, Caluromys, and Caluromy-
quently published by Reig et al., 1987), and siops were referred to the Caluromyidae.
their classification was also the first to McKenna and Bells (1997) classification
incorporate results from molecular and cyto- of Recent opossums followed Reig et al.
genetic research (Kirsch, 1977b; Reig et al., (1985), but they used the name Monodelphini
1977). Not surprisingly, this classification for the group that those authors called
(table 17) departed from its predecessors in Marmosini. This was the system subsequent-
several respects. Among them, Dromiciops was ly adopted by Gardner (2005, 2008) who
once and for all removed to a separate family, lumped all of the old murine genera
caluromyines and didelphines were distin- together with several new taxa in the tribe
guished as separate subfamilies, and several Monodelphini. No subordinate taxa (within
tribes of the latter were recognized: all of the Monodelphini) were recognized by these
large opossums with a diploid number (2n) of authors to indicate relationships among the
22 chromosomes were referred to the Didel- many genera assigned to this group.
phini, all of the smaller taxa with 2n 5 1418
chromosomes were referred to the Marmosini A Revised Phylogenetic System
(thus implying a close relationship between
Monodelphis and the murine opossums), As reviewed above, all previous classifica-
and a monotypic tribe was created for tions of Recent opossums contain nonmono-
Metachirus. phyletic groups. Examples include Marmo-
TABLE 17
Classifications of Recent Opossums, 19852008
Reig et al. (1985)a Hershkovitz (1992b)b Kirsch and Palma (1995) Gardner (2008)c
Didelphidae Marmosidae Didelphidae Didelphidae

Didelphinae Marmosinae Didelphinae Caluromyinae
Didelphini Gracilinanus Didelphini Caluromys
Chironectes Marmosops Didelphis Caluromysiops
Didelphis Marmosa Philander Glironia
Lutreolina Micoureus Lutreolina Didelphinae
Philander Thylamyinae Chironectes Didelphini
Marmosini Thylamys Metachirini Chironectes
Lestodelphys Lestodelphyinae Metachirus Didelphis
Marmosa Lestodelphys Thylamyinae Lutreolina
Micoureus Metachirinae Thylamyini Philander
Monodelphis Metachirus Thylamys Metachirini
Thylamys Monodelphinae Lestodelphys Metachirus
Metachirini Monodelphis Marmosopsini Monodelphini
Metachirus Caluromyidae Marmosops Chacodelphys
Caluromyinae Caluromyinae Gracilinanus Cryptonanus
Caluromys Caluromys Marmosinae Gracilinanus
Caluromysiops Caluromysiopsinae Marmosini Hyladelphys
Glironia Caluromysiops Marmosa Lestodelphys
Glironiidae Micoureus Marmosa
Glironia Monodelphini Marmosops
Didelphidae Monodelphis Micoureus
Philander Caluromyidae Monodelphis
Didelphis Caluromyinae Thylamys
Chironectes Caluromys
Lutreolina Caluromysiops
Glironiinae
Glironia
a
This is also the classification of Reig et al. (1987) and Marshall et al. (1990).
b
The classification of Hershkovitz (1992a) differs from this one in minor details. Hershkovitz (1997) proposed that
Lutreolina and Chironectes each be placed in monotypic subfamilies of Didelphidae.
c
Gardners (2008) suprageneric classification and nomenclature followed McKenna and Bell (1997) but included
subsequently described genera.
sini sensu Reig et al. (1985), Marmosidae ture. Points (1) and (2) should be familiar to
sensu Hershkovitz (1992b); Marmosopsini most readers and, we trust, are not contro-
sensu Kirsh and Palma (1995), and Mono- versial. Point (3) might be disputed by those
delphini sensu Gardner (2005, 2008). There- who feel that cladograms should be com-
fore, a revised phylogenetic system is needed. pletely described by classifications, but the
The classification suggested below is based mnemonic value of names is lost if they
on the following principles: (1) taxa should be proliferate beyond reasonable need, the
demonstrably monophyletic; (2) binomial hierarchical ordering implied by the consis-
nomenclature should remain as stable as tent use of familiar suffixes (e.g., -idae, -inae,
possible consistent with the requirement of -ini) is compromised if many clade names
monophyly; (3) it is not necessary to name with unfamiliar endings are interpolated, and
every node in a cladogram; (4) if future naming weakly supported clades that might
discoveries can be reasonably anticipated, disappear when new character data or taxa
taxa should be named so that these can be are added to future analyses is not in the
incorporated with minimal disruption; and interest of nomenclatural stability. Point (4)
(5) usage should be consistent with the refers to long branches, which can be
International Code of Zoological Nomencla- interpreted as artifacts of extinction or
incomplete taxon sampling (Horovitz, 1999); widely accepted rules of biological nomen-
this notion provides some justification for clature. The lack of species-level revisionary
naming monotypic higher taxa to accommo- studies for most genera is also discussed
date future discoveries of new forms. Point under this heading.
(5) acknowledges the necessity for widely
accepted rules to govern the use of technical Family Didelphidae Gray, 1821
names.
CONTENTS: Glironiinae, Caluromyinae,
Because detailed generic synonymies have
Hyladelphinae, and Didelphinae.
recently been published by Gardner (2008),
we do not provide them here. However, a DIAGNOSIS: Didelphids can be distin-
chronological list of available names with guished unambiguously from other marsupi-
information about type species and current als by their soft (nonspinous) body pelage;
status (table 15) contains most of the infor- five subequal claw-bearing manual digits; five
mation needed to understand current bino- separate pedal digits (loosely connected by
mial usage. Under the indented heading webbing in Chironectes), of which the hallux
Contents we list all of the valid lower-level (dI) is large and opposable and the second
taxa included by each higher-level taxon; for (dII) bears an asymmetrical grooming claw;
example, the genera referred to a tribe, or the inguinal cloaca; prehensile tail (secondarily
species referred to a genus. In lists of lacking external evidence of prehensility in
congeneric species, those currently regarded some taxa); long nasal bones (extending
as valid are in boldface, followed by an anteriorly beyond the facial processes of the
alphabetic list of available junior synonyms premaxillae and posteriorly between the
in parentheses; however, names of doubtful lacrimals); unfenestrated rostrum; uninflect-
application (e.g., brasiliensis Liais, 1872, ed maxillary-jugal suture; distinct foramen
which might be a junior synonym of more rotundum (not confluent or sharing a com-
than one currently recognized species of mon vestibule with the sphenorbital fissure);
Didelphis; see Cerqueira and Tribe, 2008) alisphenoid-parietal contact (except in Meta-
are not included. chirus); large interparietal fused to the
Under Diagnosis we list unique combi- supraoccipital; complete posterolateral pala-
nations of traits that distinguish members of tal foramina; deeply divided vomer that does
suprageneric taxa from other marsupials, not conceal the presphenoid or extend
whether or not such traits are unambiguously posteriorly into the mesopterygoid fossa;
assignable by parsimony as relevant synapo- lack of midline contact between the left and
morphies (listed in appendix 5). Under right pterygoids; unkeeled basisphenoid;
Morphological description (for genera) well-developed tympanic process of the ali-
we list all phenotypic descriptors that apply sphenoid; well-developed rostral tympanic
to included species; traits of particular process of the petrosal; lack of a squamosal
importance for identification are indicated epitympanic sinus; laterally exposed ectotym-
by italics. Ranges of metrical traits (e.g., 170 panic; unspecialized malleus (with long
210 mm) are approximations that match or sharply inflected neck, small orbicular
slightly exceed known minimal and maximal apophysis, and well-developed lamina); dis-
adult values that we believe to be reliable tinct mastoid and paroccipital processes;
based on published and unpublished sources. unreduced dental complement of 50 teeth;
Except as noted otherwise, all descriptions of nongliriform incisors; large milk premolars
cranial traits are based on adult specimens, (except in Hyladelphys); tribosphenic upper
and all dental descriptions are based on molars, each with a continuous stylar shelf,
unworn teeth. uninterrupted centrocrista, and reduced or
Under Remarks we discuss evidence for absent para- and metaconules; and tribo-
monophyly and other issues that relate to sphenic lower molars, each with a well-
taxon recognition and rank. Although ranks developed anterior cingulid (notched for the
are biologically arbitrary, they affect spelling hypoconulid of the preceding tooth), notched
and information retrieval in the prevailing paracristids, and without any trace of a
Linnaean system, and they are regulated by posterior cingulid.
Didelphids have often been compared with the diagnostic endocranial and postcranial
dasyurids, but the literature contains no features described by Sanchez-Villagra et al.
adequate statement of the many characters (2007).23
that distinguish these superficially similar yet REMARKS: A wide range of fossil taxa have
highly divergent clades. Among the external at one time or another been regarded as
and craniodental features treated in this didelphids (e.g., by Simpson, 1935, 1945;
report, dasyurids consistently differ from Clemens, 1979; Marshall, 1981; McKenna
didelphids by their small, nonopposable and Bell, 1997), but phylogenetic analyses
hallux; lack of a grooming claw on pedal suggest that most of the extinct forms once
digit II; a nonprehensile tail (provided with a thought to be closely related to Recent
terminal tuft of hairs that is never present in opossums (e.g., {Alphadon, {Andinodelphys,
opossums); anteriorly truncated and medially {Glasbius, {Herpetotherium, {Jaskhadelphys,
notched nasals; a small interparietal that, {Pediomys) are stem metatherians and not
when present in juvenile skulls, is suturally members of the crown group Marsupialia
distinct from the supraoccipital; incomplete (Rougier et al., 1998; Wible et al., 2001; Luo
posterolateral palatal foramina; an undivided et al., 2003; Sanchez-Villagra et al., 2007).
vomer that extends posteriorly to underlie the Herein we explicitly restrict Didelphidae to
presphenoid within the mesopterygoid fossa; living didelphimorphians, their most recent
well-developed squamosal epitympanic sinus- common ancestor, and all of its descendants.
es; four upper and three lower incisors; a Even more restrictive concepts of Didel-
distinct posterior lobe on the unworn cutting phidae have been proposed, but none is
edge of i3; vestigial milk premolars; and a widely accepted. Hershkovitz (1992b), for
distinct posterior cingulid on the lower example, used Didelphidae to include just the
molars. large opossums with 2n 5 22 chromosomes
Originally described as a didelphid by (Chironectes, Didelphis, Lutreolina, Philan-
Thomas (1894) and long maintained in that der), whereas Kirsch and Palma (1995)
family by subsequent authors (e.g., Simpson, excluded Glironia, Caluromys, and Caluro-
1945), Dromiciops differs from opossums by mysiops from the family. Because Didelphi-
its basicaudal cloaca; a foramen rotundum dae in any of these applications (ours,
recessed in a common vestibule with the Hershkovitzs, or Kirsch and Palmas) is
sphenorbital fissure; frontal-squamosal con- monophyletic, the choice among them must
tact; a large interparietal that is suturally be justified by other criteria. In our view, the
distinct from the supraoccipital; an undivided name Didelphidae and its colloquial equiva-
vomer that underlies the presphenoid within lent (didelphids) are so deeply entrenched
the mesopterygoid fossa; left and right in the literature as referring to all Recent
pterygoid bones that contact one another in
opossums that more restrictive applications
the midline; keeled basisphenoid; a concealed
would serve no adequate compensatory
ectotympanic; a specialized malleus (with a
purpose.
short uninflected neck, no orbicular apoph-
Although didelphid monophyly is impres-
ysis, and no lamina); lack of a paroccipital
sively supported by nucleotide sequence data
process; a discontinuous stylar shelf on M1;
an unstaggered i2 alveolus; and a vestigial 23
As noted above, the posterior cingulid is absent in
anterior cingulid on the lower molars (of didelphids, caenolestids, microbiotherians, and peramelids, so
which m1 lacks a hypoconulid notch). it optimizes as an unambiguous marsupial synapomorphy on all
Caenolestids, peramelids, and other mar- metatherian topologies in which these groups are recovered as
supials are sufficiently distinct from opos- basal clades (e.g., Meredith et al. 2008). Because most
Cretaceous and early Tertiary metatherians are known only
sums that explicit comparisons would be from teeth, this trait provides a potentially useful criterion for
pointless here. Some stem metatherians (e.g., distinguishing stem taxa from members of the crown clade. For
herpetotheriids) that are strikingly similar to example, several of the Paleocene Brazilian fossils classified as
didelphids in most respects (including ear didelphids by Marshall (1987) are described by that author as
having posterior cingulids and might plausibly be regarded on
morphology; Gabbert, 1998) differ from that basis as stem metatherians. By contrast, Marsupialia (in the
didelphids by having a distinct posterior crown-group sense) does not seem to be diagnosable by any
cingulid on the lower molars in addition to upper molar trait.
(e.g., Jansa and Voss, 2000; Amrine-Madsen vertebrate paleontology catalog) there is no
et al., 2003; Meredith et al., 2008), unambig- bone wedged between the parietals and the
uous morphological synapomorphies of the supraoccipital, which share a suture just
family remain to be confidently identified. behind the lambdoid crest. Therefore, the
Because our analyses did not include any available evidence suggests that the didelphid
nonmarsupial outgroup, the position of the condition is unique.
marsupial root node was not determined, and Other alleged morphological synapomor-
the phylogenetic interpretation of character- phies of didelphids include features of the
state transformations on the branch separat- spermatozoa (Temple-Smith, 1987), postcra-
ing didelphids from nondidelphid marsupials nial skeleton (Horovitz and Sanchez-Villagra,
is correspondingly equivocal. If, as most 2003), and petrosal (Ladeveze, 2007). How-
recent analyses suggest, didelphids are the ever, the phylogenetic interpretation of such
basalmost branch of Marsupialia (Nilsson et traits is compromised by sparse ingroup
al., 2004; Amrine-Madsen et al., 2003; sampling. Spermatozooa, for example, have
Horovitz and Sanchez-Villagra, 2003; Asher not been studied from many genera (e.g.,
et al., 2004; Sanchez-Villagra et al., 2007; Caluromysiops, Glironia, Hyladelphys, Lu-
Meredith et al., 2008; Beck, 2008), then such treolina, Marmosops, Thylamys), and pub-
transformations might be didelphid synapo- lished phylogenetic analyses of postcranial
morphies, or they could be synapomorphies and petrosal characters have not included
of the unnamed clade that includes caenoles- basal didelphids (analyzed didelphid terminal
tids and Australidelphia. taxa in both of the osteological studies cited
Comparisons with stem metatherians that above belong to the subfamily Didelphinae).
are believed to be close outgroups to Hopefully, future studies will help fill in
Marsupialia, including {Mayulestes (see Mui- many of these taxonomic gaps and contribute
zon, 1998), {Pucadelphys (see Marshall et al., to a better assessment of anatomical charac-
1995), and {Herpetotherium (see Gabbert, ter support for didelphid monophyly.
1998; Sanchez-Villagra et al., 2007) suggest Only Recent didelphids are formally clas-
that most of the craniodental traits by which sified and described below, but several South
didelphids differ from other marsupials are American Neogene genera represented by
plesiomorphic, but one exception merits well-preserved cranial material ({Hyperdidel-
comment. Bone homologies on the postero- phys, {Thylatheridium, {Thylophorops) are
dorsal braincase have received little attention clearly members of the didelphimorph crown
in the literature, but they appear to be clade. Pending a phylogenetic analysis of
phylogenetically informative at many taxo- their relationships with living forms, we
nomic levels within Metatheria. In particular, follow current paleontological judgments of
the presence of a large undivided interparietal taxonomic affinity in suggesting where these
bone that is wedged between the parietals fossils belong. By contrast, most other fossil
anteriorly and fused to the supraoccipital didelphids are only represented by dental
posteriorly may be a didelphid synapomor- fragments from which few characters can be
phy. According to Muizon (1998: 38, fig. 6), scored, and their relationships to extant taxa
paired interparietal ossifications are fused to are correspondingly ambiguous.
the parietals in {Mayulestes, but the basis for
this interpretation (which has no analog Subfamily Glironiinae, new
among living metatherians) is not explained,
and it seems equally possible that the CONTENTS: Glironia.
interparietal(s) is(are) absent in this taxon. DIAGNOSIS: Members of this clade differ
In {Pucadelphys, paired interparietals (post- from other didelphids by the extension of soft
parietals) are suturally distinct from the fur along the dorsal surface of the tail from
parietals and from the supraoccipital (Mar- base to tip (the dorsal surface of the tail is
shall et al., 1995: 50, fig. 12). The occiput of macroscopically naked distally in other opos-
{Herpetotherium has not previously been sums), by strongly recurved and laterally
described, but in a well-preserved specimen compressed manual claws (manual claws are
that we examined (127684 in the AMNH much less strongly recurved and laterally
compressed in other opossums), and by prehension with raised tubercles near base and
postorbital processes that are formed by the apical pad bearing dermatoglyphs.
frontals and parietals (postorbital processes Premaxillary rostral process absent. Nasals
are absent or are formed only by the frontals long, extending anteriorly beyond I1 (con-
in other opossums). cealing nasal orifice from dorsal view), and
REMARKS: The name Glironiidae as used conspicuously widened posteriorly near max-
by Hershkovitz (1992a, 1992b, 1999) was a illary-frontal suture. Maxillary turbinals
nomen nudum because it was not accompa- elaborately branched. Lacrimal foramina
nied by a statement of diagnostic characters concealed within anterior orbital margin or
(ICZN, 1999: Article 13). To our knowledge, exposed laterally, usually two on each side.
no family-group name based on Glironia is Orbits very large; supraorbital crests well
technically available from any other publica- developed; flattened triangular postorbital
tion. processes present, formed by both frontal
and parietal bones. Left and right frontals
and parietals separated by persistent median
Glironia Thomas, 1912 sutures. Parietal contacts alisphenoid on
Figure 37 lateral braincase (no frontal-squamosal con-
tact). Sagittal crest absent. Petrosal not
CONTENTS: venusta Thomas, 1912 (includ- laterally exposed through fenestra in parie-
ing aequatorialis Anthony, 1926; and criniger tal-squamosal suture (fenestra absent). Pari-
Anthony, 1926). etal-mastoid contact present (interparietal
MORPHOLOGICAL DESCRIPTION: Combined does not contact squamosal).
length of adult head and body probably ca. Maxillopalatine fenestrae usually present,
170210 mm; adult weight probably ca. 100 but small; palatine fenestrae absent; maxil-
200 g (measurements from the only two adult lary fenestrae absent; posterolateral palatal
specimens known to have been measured by foramina small, not extending anteriorly
the American method and accompanied by between M4 protocones; posterior palatal
weight data are in table 4). Rhinarium with morphology Caluromys-like (without promi-
two ventrolateral grooves on each side of nent lateral corners, the choanae not constrict-
median sulcus; dark circumocular mask ed behind). Maxillary and alisphenoid not in
present; pale supraocular spot absent; dark contact on floor of orbit (separated by
midrostral stripe absent; throat gland un- palatine). Transverse canal foramen present.
known (no suitably preserved adult male Alisphenoid tympanic process smoothly glob-
specimens have been examined for this ular, without anteromedial process or postero-
feature). Dorsal body pelage unpatterned, medial lamina enclosing extracranial course of
brownish; dorsal underfur gray; dorsal guard mandibular nerve (secondary foramen ovale
hairs short and inconspicuous; ventral fur absent), and not in contact with rostral
grayish or gray-based whitish, with or tympanic process of petrosal. Anterior limb
without self-whitish pectoral markings. Ma- of ectotympanic suspended directly from
nus paraxonic (dIII 5 dIV); manual claws basicranium. Stapes triangular with large
strongly recurved, laterally compressed, and obturator foramen. Fenestra cochleae exposed,
much longer than fleshy apical pads of digits; not concealed by rostral and caudal tympanic
dermatoglyph-bearing manual plantar pads processes of petrosal. Paroccipital process
present; central palmar epithelium smooth; small, adnate to petrosal. Dorsal margin of
carpal tubercles absent. Pes unwebbed; dIV foramen magnum bordered by exoccipitals
longer than other pedal digits; plantar surface and supraoccipital, incisura occipitalis present.
of heel naked. Pouch absent; mammae 202 Two mental foramina present on lateral
5 4, all abdominal/inguinal; cloaca present. surface of each hemimandible; angular pro-
Tail about as long as combined length of cess acute and strongly inflected.
head and body or a little longer, slender and Unworn crowns of I2I5 asymmetrical
muscular (not incrassate), and densely furred (incisiform), with longer anterior than
from base to tip except along ventral midline; posterior cutting edges; I5 separated from
naked ventral caudal surface modified for I4 by small diastema in some specimens (e.g.,
Fig. 37. Glironia venusta (based primarily on INPA 2570, a subadult male from the Rio Urucu,
Amazonas, Brazil). Because p3 is incompletely erupted on INPA 2570, the morphology and adult position
of this tooth was reconstructed from other specimens. The postorbital processes become more pronounced
with age.
INPA 2570) but not others (e.g., FMNH Subfamily Caluromyinae Reig et al., 1987
41440). Upper canine (C1) alveolus in pre-
maxillary-maxillary suture; C1 simple, with- CONTENTS: Caluromys and Caluromysiops.
out accessory cusps. First upper premolar DIAGNOSIS: Caluromyines can be distin-
(P1) present, smaller than more posterior guished from other confamilial taxa by their
premolars but well formed and not vestigial; long fourth manual digit (dIII is the longest
second and third upper premolars (P2 and manual digit, or dIII and dIV are subequal in
P3) subequal in height; P3 with both anterior all other didelphids); a completely ossified
and posterior cutting edges. Molars weakly palate (maxillopalatine and sometimes addi-
carnassialized (postmetacristae longer than tional palatal fenestrae are consistently pres-
postprotocristae); relative widths M1 , M2 ent in most other didelphids); lack of a
, M3 . M4; centrocrista weakly inflected transverse canal foramen (transverse canal
labially on M1M3; ectoflexus absent or foramina are almost invariably present in all
indistinct on M1 and M2, distinct but usually other didelphids); an alisphenoid tympanic
shallow on M3; anterolabial cingulum con- process that contacts or closely approximates
tinuous with preprotocrista (complete ante- the rostral tympanic process of the petrosal
rior cingulum present) on M3; postproto- (the alisphenoid tympanic process and the
crista without carnassial notch. Last upper rostral tympanic process of the petrosal are
tooth to erupt is P3. widely separated in other didelphids); an
Lower incisors (i1i4) with distinct lingual indirectly suspended anterior limb of the
cusps. Lower canine (c1) erect, acutely ectotympanic (the anterior limb of the
pointed, and simple (without a posterior ectotympanic is directly suspended from the
accessory cusp). Second lower premolar (p2) skull in most other didelphids); a fenestra
taller than p3; morphology of lower milk cochleae that is concealed in a sinus formed
premolar (dp3) unknown (no juvenile speci- by the rostral and caudal tympanic processes
mens examined). Hypoconid labially salient of the petrosal (the fenestra cochleae is
on m3; hypoconulids twinned with entoco- exposed in most other didelphids); a blunt
nids on m1m3; entoconids much taller than and weakly inflected angular process (the
hypoconulids on m1m3. mandibular angle is acute and strongly
DISTRIBUTION: Glironia is currently known inflected in all other didelphids); an upper
from fewer than two dozen specimens col- canine alveolus that is completely contained
lected at widely scattered localities between by the maxilla (C1 occupies the premaxillary-
300 and 1000 m above sea level in Brazil maxillary suture in all other didelphids); a
(Amazonas, Mato Grosso, Para, Rodonia), vestigial or absent first upper premolar (P1 is
eastern Ecuador, eastern Peru, and eastern present and nonvestigial in all other didel-
Bolivia (Daz and Willig, 2004; Santos-Filho phids); a tall second upper premolar that much
et al., 2007; Barkley, 2008). Although the exceeds P3 in height (P2 is either subequal to
genus probably also occurs in southeastern or smaller than P3 in most other didelphids);
Colombia, no specimens are known to have and lack of an ectoflexus on all of the upper
been collected there (Daz and Willig, 2004). molars (a distinct ectoflexus is present, at least
Most records accompanied by definite hab- on M3, in all other didelphids).
itat information are from primary or second- REMARKS: Caluromyine monophyly is
ary Amazonian rainforest, but a single strongly supported by morphological char-
specimen was recently reported from the acters (appendix 5) and IRBP gene sequences
upper Paraguay Basin (Santos-Filho et al., (fig. 28), and this clade is recovered with
2007), and several Bolivian records are from strong support in analyses of concatenated
dry forest (Tarifa and Anderson, 1997; genes (fig. 33) and total evidence (nonmolec-
Emmons, 1998). ular + molecular characters; figs. 35, 36)
REMARKS: Only the type species of Glir- despite the overall incompleteness of our
onia is currently recognized as valid, but it data from Caluromysiops (table 2). Two
seems probable that additional taxa are deletions at the DMP1 locus that currently
represented among the material now pre- optimize as synapomorphies of Caluromys
served in museum collections. (fig. 29) might prove to be caluromyine
synapomorphies when this gene is eventually MORPHOLOGICAL DESCRIPTION: Combined

sequenced for Caluromysiops. Although most length of adult head and body ca. 210
recent authors (e.g., Reig et al., 1985, 1987; 300 mm; adult weight ca. 190500 g. Rhinar-
Kirsch and Palma, 1995; McKenna and Bell, ium with two ventrolateral grooves on each
1997; Gardner, 2005, 2008) have included side of median sulcus; dark circumocular
Glironia in the Caluromyinae (or Caluromyi- mask present; pale supraocular spot absent;
dae), that genus appears to represent an dark midrostral stripe present; throat gland
ancient lineage of basal didelphids that may absent. Dorsal pelage unpatterned, grayish or
not be closely related to Caluromys + reddish brown, or indistinctly mottled with
Caluromysiops (see Jansa and Voss, 2000; same colors; dorsal underfur grayish; dorsal
this report, above). The authorship of family- guard hairs short and inconspicuous; ventral
group names based on Caluromys is often fur gray based or self-colored. Manual digit
attributed to Kirsch (1977b), but no such IV longer than other manual digits; manual
name in that work fulfills the technical claws longer than fleshy apical pads of digits;
criteria for nomenclatural availability (ICZN, dermatoglyph-bearing manual plantar pads
1999: Article 13). Apparently, the first present; central palmar epithelium smooth or
explicit statement of characters purported to sparsely tuberculate; carpal tubercles absent.
differentiate this taxon from other opossums Pedal digits unwebbed; pedal digit IV longer
appeared in Reig et al. (1987: 72). than other pedal digits; plantar surface of
heel naked. Pouch present, consisting of
separate lateral skin folds opening medially
Caluromys J.A. Allen, 1900 (subgenus Caluromys) or a deep pocket
Figure 38 opening anteriorly (subgenus Mallodelphys);
known mammary formulae apparently 202
CONTENTS: derbianus Waterhouse, 1841 5 4 (in C. lanatus) and 313 5 7 (in C.
(including antioquiae Matschie, 1917; aztecus philander); cloaca present. Tail much longer
Thomas, 1913; canus Matschie, 1917; centra- than combined length of head and body,
lis Hollister, 1914; fervidus Thomas, 1913; slender and muscular (not incrassate); body
guayanus Thomas, 1899; nauticus Thomas, fur extends onto tail base to the same extent
1913; pallidus Thomas, 1899; pictus Thomas, dorsally as ventrally (subgenus Caluromys) or
1913; pulcher Matschie, 1917; pyrrhus Thom- much further dorsally than ventrally (subge-
as, 1901; and senex Thomas, 1913); lanatus nus Mallodelphys); naked caudal integument
Olfers, 1818 (including bartletti Matschie, dark (grayish or brownish) basally and whitish
1917; cahyensis Matschie, 1917; cicur Bangs, distally with mottled transition; caudal scales
1898; hemiurus Miranda-Ribeiro, 1936; jivaro in spiral series, each scale with three subequal
Thomas, 1913; juninensis Matschie, 1917; bristlelike hairs emerging from distal margin
lanigera Desmarest, 1820; meridensis Matschie, (the central hair of each caudal-scale triplet
1917; modesta Miranda-Ribeiro, 1936; natter- longer than the lateral hairs but not grossly
eri Matschie, 1917; ochropus Wagner, 1842; swollen or petiolate); naked ventral caudal
ornatus Tschudi, 1845; and vitalinus Miranda- surface modified for prehension, with raised
Ribeiro, 1936); and philander Linnaeus, 1758 tubercles near base (Mallodelphys only) and
(including affinis Wagner, 1842; cajopolin apical pad bearing dermatoglyphs.
Muller, 1776; cayopollin Schreber, 1777; Premaxillary rostral process present. Na-
cayopollin Kerr, 1792; dichurus Wagner, sals long, extending anteriorly above I1
1842; flavescens Brongniart, 1792; leucurus (concealing nasal orifice from dorsal view),
Thomas, 1904; trinitatis Thomas, 1894; and and conspicuously widened posteriorly near
venezuelae Thomas, 1903). maxillary-frontal suture. Maxillary turbinals
These taxa are currently allocated to two elaborately branched. Lacrimal foramina
subgenera, Caluromys J.A. Allen, 1900 (con- exposed to lateral view at or near anterior
taining only C. philander), and Mallodelphys orbital margin, usually two on each side.
Thomas, 1920 (containing C. derbianus and Large, flattened, triangular postorbital pro-
C. lanatus), that can be distinguished unam- cesses of frontals present. Left and right
biguously by the traits noted below. frontals and parietals separated by persistent
Fig. 38. Caluromys philander (based on AMNH 266409, an adult male from Paracou, French Guiana).
median sutures. Parietal and alisphenoid in to frontals) in some large adult specimens.
contact on lateral braincase (no frontal- Petrosal not laterally exposed through fenes-
squamosal contact). Sagittal crest weakly tra in parietal-squamosal suture (fenestra
developed on interparietal and/or along absent). Parietal-mastoid contact present
midparietal suture (not extending anteriorly (interparietal does not contact squamosal).
Maxillopalatine fenestrae usually absent slightly longer than postprotocristae in sub-

(small fenestrae, never extending anteriorly genus Caluromys); relative widths usually M1
beyond M1 or posteriorly beyond M2, are , M2 . M3 . M4 (subgenus Mallodelphys)
uni- or bilaterally present in occasional or M1 , M2 , M3 . M4 (subgenus
specimens); palatine and maxillary fenestrae Caluromys); centrocrista weakly inflected
absent; posterolateral palatal foramina small, labially on M1M3; ectoflexus consistently
not extending anteriorly between M4 proto- absent on all upper molars; anterolabial
cones; posterior palate gently sloping ven- cingulum continuous with preprotocrista
trally, usually with arched caudal margin and (complete anterior cingulum present) on
without prominent posterolateral corners M3; postprotocrista without carnassial
(the choanae unconstricted behind). Maxil- notch. Last upper tooth to erupt is P3.
lary and alisphenoid not in contact on floor Lower incisors (i1i4) with distinct lingual
of orbit (separated by palatine). Transverse cusps. Lower canine (c1) erect, acutely
canal foramen absent. Alisphenoid tympanic pointed, and simple (without a posterior
process more or less conical (acutely pointed accessory cusp). Second upper premolar
ventrally), usually without posteromedial (p2) much taller than p3; lower milk premo-
lamina enclosing extracranial course of man- lar (dp3) trigonid with or without paraconid
dibular nerve (secondary foramen ovale (when present, the paraconid is small and
absent),24 and contacting or closely approx- often indistinct). Hypoconid labially salient
imating rostral tympanic process of petrosal. on m3; hypoconulid twinned with entoconid
Anterior limb of ectotympanic indirectly on m1m3; entoconid much taller than
suspended from basicranium (by malleus). hypoconulid on m1m3.
Stapes triangular with large obturator fora- DISTRIBUTION: Caluromys occurs from the
men (in subgenus Caluromys) or subtriangu- Mexican state of Veracruz southward
lar with small but patent foramen (in throughout most of the forested regions of
subgenus Mallodelphys). Fenestra cochleae Central and South America (including Trini-
concealed in sinus formed by rostral and dad) to eastern Bolivia, eastern Paraguay,
caudal tympanic processes of petrosal. Par- and northeastern Argentina (Hall, 1981;
occipital process small, rounded and sub- Gardner, 2008). Most specimens of C. lanatus
triangular, adnate to petrosal. Dorsal margin and C. philander (e.g., those reported by
of foramen magnum bordered by supraoc- Handley, 1976; Anderson, 1997) are from
cipital and exoccipitals, incisura occipitalis lowland or lower montane rainforest; how-
present. ever, C. philander has also been observed in
Two mental foramina usually present on dry forest (Emmons, 1998), and C. derbianus
lateral surface of each hemimandible; angular ranges from sea level to over 2000 m and
process blunt and weakly inflected. occurs in both rainforest and dry forest
Unworn crowns of I2I5 asymmetrical (Handley, 1966; Bucher and Hoffmann,
(incisiform), with much longer anterior 1980; Sanchez et al., 2004).
than posterior cutting edges. Upper canine REMARKS: Generic monophyly (vis-a-vis
(C1) alveolus contained entirely in maxillary Caluromysiops) is supported by IRBP se-
bone; C1 simple, without accessory cusps. quence data (fig. 28), by the presence of a
First upper premolar (P1) minute, vestigial, distinct midrostral stripe, and by the presence
and lacking any consistent occlusal features; of a small but distinct rostral process of the
second upper premolar (P2) much taller than premaxillae (the latter two traits optimize as
P3; P3 with both anterior and posterior unambiguous generic synapomorphies; ap-
cutting edges; upper milk premolar (dP3) pendix 5).
large and molariform. Molars not carnassia- The genus Caluromys has never been
lized (postprotocristae and postmetacristae revised, and there is substantial divergence
are subequal in subgenus Mallodelphys) or in morphology and/or cytochrome b sequenc-
weakly carnassialized (postmetacristae are es among some allegedly conspecific popula-
tions (Voss et al., 2001; Patton and Costa,
24
We examined one old adult male (MVZ 190249) with a 2003). Indeed, no revisionary study has
secondary foramen ovale formed by a posteromedial lamina. documented the conspecific status of the
many subjective synonyms listed above for ing nasal orifice from dorsal view), and
either C. philander or C. lanatus. It is an open conspicuously widened posteriorly near max-
question as to whether the subgenera Calur- illary-frontal suture. Maxillary turbinals
omys and Mallodelphys will prove to be elaborately branched. One or two lacrimal
reciprocally monophyletic or not in future foramina exposed on each side on or near
phylogenetic analyses based on denser taxon anterior orbital margin. Large flattened,
sampling. triangular postorbital processes of frontals
present, continuous with supraorbital crests.
Caluromysiops Sanborn, 1951 Left and right frontals and parietals co-
Figure 39 ossified in most adult specimens (median
sutures incomplete or obliterated). Parietal
CONTENTS: irrupta Sanborn, 1951. and alisphenoid in contact on lateral brain-
MORPHOLOGICAL DESCRIPTION: Combined case (no frontal-squamosal contact). Sagittal
length of adult head and body probably ca. crest present, prominently developed on pari-
250300 mm (Izor and Pine, 1987); adult etals and extending onto frontals in most adult
weight probably ca. 300500 g. Rhinarium specimens. Petrosal not laterally exposed
with two ventrolateral grooves on each side through fenestra in parietal-squamosal suture
of median sulcus; fur of head uniformly pale, (fenestra absent). Parietal-mastoid contact
without any marking (dark circumocular present (interparietal does not contact squa-
mask, supraocular spots, and midrostral mosal).
stripes absent); throat gland absent. Dorsal Maxillopalatine, palatine, and maxillary
pelage pale overall but prominently marked by fenestrae absent (palate completely ossified);
paired blackish stripes extending from fore- posterolateral palatal foramina small, not
limbs to shoulders and thence posteriorly as extending anteriorly between M4 protocones;
narrowing parallel bands to rump; dorsal posterior palatal morphology conforms to
underfur dark (apparently grayish but dis- Caluromys morphotype (without strongly
colored with age in all examined specimens); developed lateral corners, the choanae un-
dorsal guard hairs short; ventral fur vari- constricted behind). Maxillary and alisphe-
ously colored (gray-based buffy in some noid widely separated by palatine on floor of
specimens, self-buffy in others). Manual orbit. Transverse canal foramen absent.
digit IV longer than other manual digits; Alisphenoid tympanic process smoothly
manual claws longer than fleshy apical pads globular, apparently always with posterome-
of digits; dermatoglyph-bearing manual dial lamina enclosing extracranial course of
plantar pads present; central palmar epithe- mandibular nerve (secondary foramen ovale
lium smooth; carpal tubercles absent. Pedal present), and extending posteriorly to contact
digits unwebbed; dIV longer than other or closely approximate rostral tympanic
pedal digits; plantar surface of heel usually process of petrosal. Anterior limb of ecto-
naked. Pouch said to be present (Izor and tympanic indirectly suspended from basicra-
Pine, 1987) but morphological details un- nium (by malleus). Stapes usually columellar
known; mammary formula unknown; cloa- and imperforate (rarely microperforate). Fe-
ca present. Tail longer than combined nestra cochleae usually concealed in sinus
length of head and body, slender and formed by rostral and caudal tympanic
muscular (not incrassate); densely covered processes of petrosal (but not in USNM
with body fur from base almost to tip except 397626, possibly due to pathology or post-
along ventral midline and posterolaterally; mortem damage). Paroccipital process small,
posterolateral caudal scales in spiral series, adnate to petrosal. Dorsal margin of foramen
each scale with three subequal bristlelike hairs magnum bordered by supraoccipital and
emerging from distal margin; naked ventral exoccipitals, incisura occipitalis present.
surface modified for prehension with raised Two mental foramina usually present on
tubercles near base, and apical pad bearing lateral surface of each hemimandible; angular
dermatoglyphs. process obtuse, weakly inflected.
Premaxillary rostral process absent. Nasals Unworn crowns of I2I5 asymmetrical
long, extending anteriorly above I1 (conceal- (incisiform), with longer anterior than
Fig. 39. Caluromysiops irrupta (a composite illustration based on several imperfect adult specimens,
mostly from zoos: AMNH 208101, 244364; USNM 397626; FMNH 84426).
posterior cutting edges. Upper canine (C1) current binomial usage is likewise unobjec-
alveolus contained entirely within maxillary tionable and serves to separate the mono-
bone; C1 simple (without accessory cusps), phyletic cluster of species currently known as
and very large in adult specimens. First upper Caluromys from its long-branched sister
premolar (P1) absent or minute and vestigial; taxon.
second upper premolar (P2) much taller than
P3; P3 with both anterior and posterior Subfamily Hyladelphinae, new
cutting edges; upper milk premolar (dP3)
large and molariform. Molars not carnassia- CONTENTS: Hyladelphys.
lized (postmetacristae subequal to postproto- DIAGNOSIS: Members of this clade unique-
cristae); relative widths M1 , M2 . M3 . ly differ from other didelphids by their
M4; centrocrista straight (uninflected) on vestigial milk dentition (dP3/dp3 are large,
M1M3; ectoflexus absent on all upper more or less molariform teeth in other
molars; anterolabial cingulum continuous opossums; Voss et al., 2001: table 5).
with preprotocrista (complete anterior cingu- Additional contrasts among hyladelphines
lum present) on M3; postprotocrista without and members of other didelphid subfamilies
carnassial notch. Last upper tooth to erupt were tabulated by Jansa and Voss (2005:
is P3. table 2).
Lower incisors (i1i4) with distinct lingual REMARKS: Given the phylogenetic results
cusps. Lower canine (c1) erect, acutely at hand, a monotypic suprageneric taxon for
pointed, and simple (without a posterior Hyladelphys could either be ranked as a
accessory cusp). Second lower premolar (p2) subfamily (as it is here) or as a tribe (if the
much taller than p3; lower milk premolar genus were referred to the Didelphinae).
(dp3) large and nonvestigial, but paraconid Although the issue of rank is not biologically
not well developed (trigonid incomplete). meaningful, the former option serves to
Hypoconid labially salient on m3; hypoconu- emphasize the intermediate position of this
lid not twinned with entoconid on any lower odd little opossum between basal didelphids
molar; entoconid much taller than hypocon- (Glironia and caluromyines) and the speciose
ulid on m1m3. radiation of lineages that are more closely
DISTRIBUTION: Caluromysiops is currently related to Didelphis. The branch leading to
known from just seven localities in the Hyladelphys is very long in most molecular
rainforested Amazonian lowlands of south- reconstructions of didelphid anagenesis (e.g.,
eastern Colombia, eastern Peru and western fig. 33), suggesting an ancient history of
Brazil (Emmons, 2008). independent evolution accompanied by ex-
REMARKS: Various authors (e.g., Cabrera, tinction of transitional forms; the same
1958; Simpson, 1972; Izor and Pine, 1987) conclusion is suggested by its uniquely
have questioned whether it is useful to reduced milk dentition and by characters of
separate Caluromysiops from Caluromys, the postcranial skeleton (Flores, 2009). In
but none explicitly considered the suite of effect, excluding Hyladelphys from Didelphi-
trenchant morphological characters that dis- nae simplifies the diagnosis of the latter clade
tinguish these taxa. Among other features, and provides a new higher taxon to accom-
Caluromysiops differs from Caluromys by its modate fossil relatives of the former, should
lack of a circumocular mask and dark any be discovered.
midrostral stripe, presence of dark scapular
stripes, shorter tail (table 5), absence of a Hyladelphys Voss et al., 2001
premaxillary rostral process, co-ossified fron- Figure 40
tals, presence of a well-developed sagittal
crest, presence of a secondary foramen ovale, CONTENTS: kalinowskii Hershkovitz, 1992.
globular (versus conical) alisphenoid bullae, MORPHOLOGICAL DESCRIPTION: Combined
columelliform stapes, linear centrocrista, and adult length of head and body ca. 7595 mm;
untwinned hypoconulid. Although there is no adult weight ca. 1020 g. Rhinarium with
cladistic problem with combining these taxa two ventrolateral grooves on each side of
under a single generic name, maintaining median sulcus; dark circumocular mask
Fig. 40. Hyladelphys kalinowskii (based on AMNH 267338, an adult male from Paracou, French
Guiana).
present, extending posteriorly from mystacial illary and alisphenoid bones not in contact
region to base of ear on each side of face; pale on floor of orbit (widely separated by
supraocular spot absent; dark midrostral palatine). Transverse canal foramen present.
stripe absent; throat gland absent. Dorsal Alisphenoid tympanic process smoothly
pelage unpatterned reddish-brown with dark- globular, without anteromedial process or
gray hair bases; dorsal guard hairs short and posteromedial lamina (secondary foramen
inconspicuous; ventral fur self-white. Manus ovale absent), and not in contact with rostral
paraxonic (dIII 5 dIV); manual claws tympanic process of petrosal. Anterior limb
relatively large, strongly recurved, and slight- of ectotympanic suspended directly from
ly longer than fleshy apical pads of digits; basicranium. Stapes triangular with large
dermatoglyph-bearing manual plantar pads obturator foramen. Fenestra cochleae ex-
present; central palmar epithelium smooth or posed, not concealed by rostral and caudal
sparsely tubercular; carpal tubercles absent. tympanic processes of petrosal. Paroccipital
Pedal digits unwebbed; dIV longer than other process small and adnate to petrosal. Dorsal
pedal digits; plantar surface of heel naked. margin of foramen magnum bordered by
Pouch absent; mammae 202 5 4, all supraoccipital and exoccipitals, incisura oc-
abdominal-inguinal; cloaca present. Tail cipitalis present.
much longer than combined length of head One or two mental foramina present on
and body, slender and muscular (not incras- lateral surface of each hemimandible; angular
sate), and apparently naked (without a process acute and strongly inflected.
conspicuously furred base); caudal scales in Unworn crowns of I2I5 asymmetrical
both annular and spiral series, each scale with (incisiform), with much longer anterior
three subequal bristle-like hairs emerging than posterior cutting edges. Upper canine
from distal margin; ventral caudal surface (C1) alveolus in premaxillary-maxillary su-
modified for prehension distally, with apical
ture; C1 simple, without accessory cusps.
pad bearing dermatoglyphs.
First upper premolar (P1) smaller than
Premaxillary rostral process absent. Nasals posterior premolars but well formed and
long, extending anteriorly above I1 (conceal-
not vestigial; second upper premolar (P2)
ing most of nasal orifice from dorsal view),
much taller than P3; P3 with both anterior
and conspicuously widened posteriorly near
and posterior cutting edges; milk premolar
maxillary-frontal suture. Maxillary turbinals
(dP3) very small, vestigial, and lacking
elaborately branched. Lacrimal foramina
distinct occlusal features. Upper molars not
exposed laterally on or near anterior orbital
strongly carnassialized (postmetacristae only
margin, one or two on each side. Orbits very
slightly longer than postprotocristae); relative
large; interorbital region strongly convergent
widths M1 , M2 , M3 . M4; centrocrista
anteriorly, with beaded dorsolateral margins;
postorbital processes absent. Left and right weakly inflected labially on M1M3; ecto-
frontals and parietals separated by persistent flexus indistinct on M1, shallow on M2,
median sutures. Parietal and alisphenoid in distinct on M3; anterolabial cingulum con-
contact on lateral braincase (no frontal- tinuous with preprotocrista (complete ante-
squamosal contact). Sagittal crest absent. rior cingulum present) on M3; postproto-
Petrosal not exposed laterally through fenes- crista without carnassial notch. Last upper
tra in parietal-squamosal suture (fenestra tooth to erupt is P3.
absent). Parietal-mastoid contact present Lower incisors (i1i4) with distinct lingual
(interparietal does not contact squamosal). cusps. Lower canine (c1) erect, acutely
Maxillopalatine fenestrae present; palatine pointed, and simple (without a posterior
and maxillary fenestrae absent; posterolateral accessory cusp). Second lower premolar (p2)
palatal foramina small, not extending anteri- taller than p3; lower milk premolar (dp3)
orly between M4 protocones. Posterior pal- small, vestigial, and lacking distinct occlusal
atal morphology more Didelphis-like than features. Hypoconid labially salient on m3;
Caluromys-like (with moderately well-devel- hypoconulid twinned with entoconid on m1
oped posterolateral corners, the internal m3; entoconid taller than hypoconulid on
choanae distinctly constricted behind). Max- m1m3.
DISTRIBUTION: Currently known from few- BRCA1 sequences (fig. 31), concatenated
er than a dozen specimens, Hyladelphys has sequence data from five genes (fig. 33), and
been reported from just nine localities in combined (nonmolecular + molecular) data
the rainforested lowlands of eastern Peru, when Chacodelphys is excluded (fig. 35).
central Amazonian Brazil, southern Guyana, Subfamilial monophyly is also supported by
and French Guiana (Astua, 2007). Extrapo- a uniquely shared insertion (not coded for
lating from these scanty data is obviously phylogenetic analysis) at the BRCA1 locus
problematic, but it would not be surprising (fig. 31). However, only two morphological
to find this elusive taxon anywhere in traits optimize as unambiguous synapomor-
Amazonia. phies of Didelphinae: an unpaired median
REMARKS: High levels of molecular diver- teat and a strongly labially inflected centro-
gence between sequenced specimens of Hyla- crista (subsequently reversed in Didelphini;
delphys from French Guiana and Peru, appendix 5). In addition to genera referred to
together with geographic variation in mor- the tribes named above, the genus {Hyperdi-
phological characters, suggest that additional delphys (here unassigned to tribe) clearly
species remain to be described in this genus belongs to this subfamily based on characters
(Jansa and Voss, 2005). described and illustrated by Goin and Pardi-
nas (1996) and others subsequently scored for
Subfamily Didelphinae Gray, 1821 phylogenetic analysis by D.A. Flores and the
senior author.
CONTENTS: Marmosini, Metachirini, Di-
delphini, and Thylamyini. Tribe Marmosini Hershkovitz, 1992
DIAGNOSIS: Members of this clade unique-
ly differ from other didelphids by lacking an CONTENTS: Marmosa, Monodelphis, {Thy-
anterior cutting edge on P3 (the third upper latheridium, and Tlacuatzin.
premolar is double bladed in all basal DIAGNOSIS: Marmosines can be distin-
didelphids). In addition, Didelphinae differs guished from other members of the subfamily
from Glironiinae by having a caudal dorsum Didelphinae by lacking a pouch (present in
that is macroscopically naked, at least near all Didelphini), by lacking an anteromedial
the tip; an unpaired median teat; well- bullar process spanning the transverse canal
developed posterolateral palatal corners; foramen (present in Metachirini and most
and postorbital processes that (when present) Thylamyini), by lacking a fenestra in the
are formed only by the frontals. Didelphinae squamosal-parietal suture (present in most
additionally differs from Caluromyinae by Thylamyini), and by having the supraoccip-
having a shorter fourth manual digit (dIII is ital form part of the dorsal margin of the
subequal to or longer than dIV), an incom- foramen magnum (absent in adult Metachir-
pletely ossified palate (maxillopalatine and ini and most adult Didelphini).
sometimes other fenestrae are invariably REMARKS: The monophyly of Marmosini
present), transverse canal foramina, a wide as construed herein is strongly supported by
gap between the alisphenoid tympanic pro- all analyses of BRCA1 (fig. 31), vWF
cess and the rostral tympanic process of the (fig. 32), concatenated sequence data from
petrosal, an acute and strongly inflected five genes (fig. 33), and our combined (non-
mandibular angle, an upper canine alveolus molecular + molecular) dataset excluding
that lies within the premaxillary-maxillary Chacodelphys (fig. 35). However, no mor-
suture, a nonvestigial P1, a larger P3 (sub- phological trait optimizes as an unambiguous
equal to or taller than P2), and a distinct synapomorphy of this molecularly robust
ectoflexus on M3. Didelphinae additionally clade. Instead, all of the diagnostic features
differs from Hyladelphinae by having an listed above appear to be plesiomorphic
unpaired median teat, nonvestigial milk within the subfamily Didelphinae.
teeth, and a larger P3 (subequal to or taller Although the name Marmosini was used
than P2). by Reig et al. (1985, 1987) it was not
REMARKS: The monophyly of Didelphinae accompanied by an explicit statement of
is strongly supported by all analyses of taxonomically differentiating characters. To
the best of our knowledge, the first family- grandis Tate, 1931); mexicana Merriam, 1897
group name based on Marmosa is technically (including mayensis Osgood, 1913; savan-
available from Hershkovitz (1992b). An narum Goldman, 1917; and zeledoni Gold-
alternative name for this clade is Monodel- man, 1917); murina Linnaeus, 1758 (including
phini, which McKenna and Bell (1997) bombascarae Anthony, 1922; chloe Thomas,
attributed to Talice et al. (1961). However, 1907; dorsigera Linnaeus, 1758; duidae Tate,
Talice et al. did not mention any characters 1931; guianensis Kerr, 1792; klagesi J.A.
purported to differentiate Monodelphini Allen, 1900; macrotarsus Wagner, 1842;
from other didelphids, so their name is madeirensis Cabrera,. 1913; maranii Thomas,
unavailable (ICZN, 1999: Article 13). To 1924; meridionalis Miranda-Ribeiro, 1936;
the best of our knowledge, the first family- moreirae Miranda-Ribeiro, 1936; muscula
group name based on Monodelphis is also Cabanis, 1848; parata Thomas, 1911; rorai-
available from Hershkovitz (1992b). As first mae Tate, 1931; tobagi Thomas, 1911; and
revisors in the sense of the Code (ICZN, waterhousei Tomes, 1860); quichua Thomas,
1999: Article 24), we select Marmosini 1899 (including musicola Osgood, 1913);
Hershkovitz, 1992, to have precedence over robinsoni Bangs, 1898 (including casta Thom-
Monodelphini Hershkovitz, 1992. as, 1911; chapmani J.A. Allen, 1900; fulvi-
Our reluctance to recognize any subtribal venter Bangs, 1901; grenadae Thomas, 1911;
distinction between Marmosa and Monodel- isthmica Goldman, 1912; luridavolta Good-
phis is based on the uncertain position of win, 1961; mimetra Thomas, 1921; mitis
Tlacuatzin, which is not convincingly re- Bangs, 1898; nesaea Thomas, 1911; pallidi-
solved despite the large amount of data at ventris Osgood, 1912; ruatanica Goldman,
hand (.2000 parsimony-informative charac- 1911; and simonsi Thomas, 1899); rubra Tate,
ters; table 12). In the event that a sister-group 1931; tyleriana Tate, 1931 (including phelpsi
relationship between Marmosa and Tlacuat- Tate, 1939); and xerophila Handley and
zin (weakly indicated by supermatrix analy- Gordon, 1979.
ses; figs. 3336) were strongly supported by The subgenus Micoureus contains alstoni
some future dataset, it would then make J.A. Allen, 1900 (including nicaraguae Thom-
sense to recognize one subtribe (e.g., Mar- as, 1905); constantiae Thomas, 1904 (includ-
mosina) for those taxa and another (Mono- ing budini Thomas, 1920); demerarae Thom-
delphina) for Monodelphis and {Thylather- as, 1905 (including areniticola Tate, 1931;
idium. Although it seems indisputable that domina Thomas, 1920; esmeraldae Tate,
Monodelphis and {Thylatheridium are closely 1931; limae Thomas, 1920; and meridae Tate,
related (Goin and Rey, 1997), we note that 1931); paraguayana Tate, 1931 (including
this hypothesis has yet to be tested ana- cinerea Temminck, 1824 [preoccupied]);
lytically, nor is it certainly known whether phaea Thomas, 1899 (including perplexa
or not these taxa are reciprocally monophy- Anthony, 1922); and regina Thomas, 1898
letic.25 (including germana Thomas, 1904; mapirien-
sis Tate, 1931; parda Tate, 1931; rapposa
Marmosa Gray, 1821 Thomas, 1899; and rutteri Thomas, 1924).
Figure 41 MORPHOLOGICAL DESCRIPTION: Combined
length of adult head and body ca. 100
CONTENTS: We recognize two subgenera, 210 mm; adult weight ca. 20170 g. Rhinar-
Marmosa Gray, 1821, and Micoureus Lesson, ium with two ventrolateral grooves on each
1842 (see Remarks, below). side of median sulcus; dark circumocular
The subgenus Marmosa contains andersoni mask present; pale supraocular spot absent;
Pine, 1972; lepida Thomas, 1888 (including dark midrostral stripe absent; throat gland
absent in some species (e.g., M. murina) but
25
In fact, it has been suggested that they are not: Considero present in adult males of other species (e.g.,
que Thylatheridium es un genero derivado, en el Plioceno M. mexicana). Dorsal pelage unpatterned,
inferior, de una de las especies de Monodelphis que, ante
posibilidades ecologicas favorables, acelero su ritmo evolutivo,
superficially brownish, reddish, or grayish,
apartandose rapidamente de sus congeneres de evolucion lenta but dorsal hair bases always dark gray;
(Reig, 1958: 90). dorsal guard hairs short and inconspicuous;
Fig. 41. Marmosa murina (based on AMNH 267368, an adult of unknown sex from Paracou, French
Guiana). The inset shows the morphology of the unworn lower canine (with flattened apical blade; based
on AMNH 266416, a young adult female from the same locality).
ventral fur superficially whitish, yellowish, or Left and right frontals and parietals separat-
orange, wholly or partly gray based (appar- ed by persistent median sutures. Parietal and
ently never completely self-colored). Manus alisphenoid in contact on lateral braincase
paraxonic (dIII 5 dIV); manual claws about (no frontal-squamosal contact). Sagittal crest
as long as or slightly longer than fleshy apical never developed. Petrosal usually not ex-
pads of digits; dermatoglyph-bearing manual posed laterally through fenestra in parietal-
plantar pads present; central palmar epithe- squamosal suture (fenestra absent).26 Parietal-
lium smooth or sparsely covered with flat- mastoid contact present (interparietal does
tened tubercles (never densely tuberculate); not contact squamosal).
carpal tubercles completely absent in some Maxillopalatine fenestrae present; palatine
species (e.g., M. murina), or only lateral fenestrae absent in most species (but consis-
carpal tubercles present in adult males (e.g., tently present in some; e.g., M. mexicana);
in M. lepida), or both medial and lateral maxillary fenestrae absent; posterolateral
carpal tubercles present in adult males (e.g., palatal foramina small, never extending
in M. robinsoni). Pedal digits unwebbed; dIV anteriorly between M4 protocones; posterior
longer than other pedal digits; plantar surface palatal morphology conforms to Didelphis
of heel macroscopically naked, but all or part morphotype (with well-developed lateral cor-
of heel covered with microscopic hairs in ners, the choanae constricted behind). Max-
some species (e.g., M. rubra). Pouch absent; illary and alisphenoid not in contact on floor
mammae 313 5 7 (e.g., in M. lepida) to 7 of orbit (separated by palatine). Transverse
17 5 15 (in M. mexicana), all abdominal- canal foramen present. Alisphenoid tympanic
inguinal; cloaca present. Tail always substan- process smoothly globular, without antero-
tially longer than combined length of head medial process or posteromedial lamina
and body; slender and muscular (not incras- enclosing the maxillary nerve (secondary
sate); without a conspicuously furred base in foramen ovale usually absent),27 and not in
some species (e.g., M. murina) or tail base contact with rostral tympanic process of
conspicuously furred to about the same petrosal. Anterior limb of ectotympanic
extent dorsally as ventrally (e.g., in M.
suspended directly from basicranium. Stapes
paraguayana); naked caudal integument uni-
usually triangular, with large obturator
colored (all dark) in most species but mottled
foramen (microperforate and imperforate
distally with white spots and/or white tipped
stapes occur as rare variants in several
in others (e.g., M. paraguayana); caudal
species). Fenestra cochleae exposed in most
scales in spiral series (e.g., in M. murina) or
species, but fenestra concealed in sinus
in both spiral and annular series (e.g., in M.
formed by caudal and rostral tympanic
mexicana), each scale usually with three
processes of petrosal in M. rubra. Paroccipi-
subequal bristlelike hairs emerging from
tal process small, rounded, adnate to petro-
distal margin; ventral caudal surface always
modified for prehension distally, with apical sal. Dorsal margin of foramen magnum
pad bearing dermatoglyphs. bordered by supraoccipital and exoccipitals,
Premaxillary rostral process present in most incisura occipitalis present.
species (absent in M. xerophila). Nasals long, Two mental foramina usually present on
extending anteriorly beyond I1 (concealing lateral surface of each hemimandible (one
nasal orifice in dorsal view), and conspicu- foramen or three foramina occur as rare,
ously widened posteriorly near maxillary-
26
frontal suture. Maxillary turbinals elaborate- We observed a fenestra between the squamosal and parietal
ly branched. Lacrimal foramina usually two on just two skulls, both of which are examples of Marmosa
lepida. This trait occurs bilaterally on MNHN 1998-306 (from
on each side, exposed to lateral view on or French Guiana) and unilaterally on MVZ 155245 (from Peru).
near anterior orbital margin. Supraorbital All of the other six specimens of M. lepida that we examined for
margins with distinct beads or prominent this character resemble other Marmosa spp. in lacking these
crests; flattened, triangular postorbital pro- openings.
27
Two out of 12 examined specimens of Marmosa regina
cesses usually present in fully mature adults have secondary foramina ovales formed by posteromedial bullar
but substantially larger in some species than laminae: this trait occurs bilaterally on MVZ 190326 and
in others (absent or indistinct in M. rubra). unilaterally on MVZ 190328.
usually unilateral variants in some species); rately (figs. 2832), in tandem (fig. 33), and
angular process acute and strongly inflected. in combination with nonmolecular characters
Unworn crowns of I2I5 symmetrically (figs. 35, 36); generic monophyly is also
rhomboidal (premolariform), with subeq- supported by a uniquely shared insertion at
ual anterior and posterior cutting edges, and the BRCA1 locus (fig. 31). Only a single
increasing in length (mesiodistal dimension) morphological character, the possession of a
from I2 to I5. Upper canine (C1) alveolus in rostral process of the premaxillae, optimizes
premaxillary-maxillary suture; C1 simple as an unambiguous generic synapomorphy
(without accessory cusps in most species) or (appendix 5), but even this weak phenotypic
with posterior accessory cusp only (in M. evidence is compromised by the absence of a
lepida). First upper premolar (P1) smaller rostral process in M. xerophila, a species that
than posterior premolars but well formed and we did not score for this study.
not vestigial; second and third upper premo- No published phylogenetic analysis sup-
lars (P2 and P3) subequal in height; P3 with ports the reciprocal monophyly of Marmosa
posterior cutting edge only; upper milk and Micoureus, both of which have been
premolar (dP3) large and molariform. Mo- treated as valid genera by recent authors
lars moderately carnassialized (postmetacris- (e.g., Gardner, 2005; Creighton and Gardner,
tae are visibly longer than postprotocristae); 2008b; Gardner and Creighton, 2008b).
relative widths usually M1 , M2 , M3 , Instead, species of Micoureus have consis-
M4; centrocrista strongly inflected labially on tently been recovered as nested within a
M1M3; ectoflexus indistinct or absent on paraphyletic group of Marmosa species
M1, shallow but usually distinct on M2, and (Kirsch and Palma, 1995; Voss and Jansa,
consistently deep on M3; anterolabial cingu- 2003; Jansa and Voss, 2005; Steiner et al.,
lum continuous with preprotocrista (com- 2005; Jansa et al., 2006; Gruber et al., 2007).
plete anterior cingulum present) on M3. Last Obviously, there are several alternative tax-
upper tooth to erupt is P3. onomic solutions to this problem.
Lower incisors (i1i4) with distinct lingual One solution would be to treat Micoureus
cusps. Unworn lower canine (c1) usually as a junior synonym of Marmosa without
semiprocumbent, with flattened bladelike apex, recognizing any subgenera of the latter.
with or without distinct posterior accessory Another would be to recognize Micoureus
cusp. Second lower premolar (p2) taller than as a subgenus of Marmosa and to name new
p3; lower milk premolar (dp3) trigonid subgenera for other monophyletic clusters of
incomplete (bicuspid). Hypoconid labially Marmosa species. A third would be to
salient on m3; hypoconulid twinned with recognize Micoureus as a genus and to
entoconid on m1m3; entoconid much taller describe new genera as needed to make
than hypoconulid on m1m3. Marmosa monophyletic. Unfortunately, the
DISTRIBUTION: Species of Marmosa collec- first option would result in the loss of a useful
tively range from the Mexican state of name (Micoureus, see below), whereas the
Tamaulipas southward throughout most of second and third options are not currently
Central and tropical South America to workable because many species of Marmosa
Bolivia, Paraguay, and northern Argentina have not been included in any phylogenetic
(Hall, 1981; Creighton and Gardner, 2008b; analysis, and their relationships are corre-
Gardner and Creighton, 2008b). Although spondingly obscure. Our interim solution is
most species inhabit lowland rainforests, a to move the currently intractable problem of
few are restricted to dry habitats (e.g., M. paraphyly from the generic to the subgeneric
xerophila; Handley and Gordon, 1979) and level. Although taxonomic rank is biologi-
some species of the subgenus Micoureus cally arbitrary, it affects binomial usage,
occur in montane rainforest at elevations in which should be conformable with phyloge-
excess of 2000 m (e.g., M. mapiriensis sensu netic relationships insofar as these are
Tate, 1933: 76). known. In effect, because the use of generic
REMARKS: The monophyly of Marmosa names is obligatory under the current Lin-
(including Micoureus) is supported by se- naean system, it is crucial that genera be
quence data from five genes analyzed sepa- monophyletic.
The monophyly of Micoureus has been Doutt, 1938; palliolata Osgood, 1914; per-
supported in every sequencing study to date uviana Osgood, 1913; reigi Lew and Perez-
that has included two or more exemplar Hernandez, 2004; ronaldi Solari, 2004; rubida
species (e.g., Patton et al., 1996; Voss and Thomas, 1899; scalops Thomas, 1888; sorex
Jansa, 2003; this study), and it may often be Hensel, 1872 (including henseli Thomas,
appropriate to indicate this fact in contradis- 1888; itatiayae Miranda-Ribeiro, 1936; lundi
tinction to the paraphyly of the subgenus Matschie, 1916; and paulensis Vieira, 1950);
Marmosa. Where appropriate, this can be theresa Thomas, 1921; umbristriatus Miran-
achieved using double quotes for the latter, as da-Ribeiro, 1936; and unistriatus Wagner,
for the species Marmosa (Marmosa) lepida 1842.
as contrasted with Marmosa (Micoureus) MORPHOLOGICAL DESCRIPTION: Combined
demerarae. Predictably, the subgeneric clas- length of adult head and body ca. 70
sification of Marmosa will be refined as 200 mm; adult weight ca. 15150 g. Rhinar-
future studies based on denser taxon sam- ium with one ventrolateral groove on each side
pling yield increasingly resolved estimates of of median sulcus; dark circumocular mask
relationships within this speciose group. absent; pale supraocular spot absent; dark
Most of the species herein referred to midrostral stripe absent; throat gland present
Marmosa have not been revised since Tate in adult males of most species but possibly
(1933), and some currently recognized syn- absent in some (e.g., M. theresa). Dorsal
onymies are the result of uncritical lumping pelage coloration highly variable, but dorsal
by subsequent authors (e.g., Hershkovitz, hair bases always dark gray; dorsal guard
1951; Cabrera, 1958). Recent analyses of hairs short and inconspicuous; ventral fur
both mtDNA sequence data (e.g., by Patton self-colored or gray based, highly variable in
et al., 2000; Patton and Costa, 2003) and surface pigmentation. Manus mesaxonic (dIII
morphological characters (Rossi, 2005) sug- . dIV); manual claws very long, extending
gest that several nominal taxa currently listed well beyond fleshy apical pads of digits;
as synonyms (e.g., of M. demerarae, M. dermatoglyph-bearing manual plantar pads
murina, and M. robinsoni) are probably valid present, but pads small and dermatoglyphs
species. Therefore, significant changes to the sometimes indistinct; central palmar epitheli-
species-level taxonomy of Marmosa should um smooth or sparsely tuberculate; carpal
be expected soon. tubercles absent in both sexes. Pedal digits
unwebbed; pedal digit III longer than digit IV;
Monodelphis Burnett, 1830 plantar surface of heel naked. Pouch absent;
Figure 42 mammae 414 5 9 (all abdominal-inguinal;
e.g., in M. brevicaudata) to 13113 5 27
CONTENTS: adusta Thomas, 1897 (includ- (including pectoral teats; e.g., in M. sorex);
ing melanops Goldman, 1912); americana cloaca present. Tail much shorter than com-
Muller, 1776 (including brasiliensis Erxleben, bined length of head and body; thick but
1777; brasiliensis Daudin, 1802; trilineata muscular, not incrassate; tail conspicuously
Lund, 1840; and tristriata Illiger, 1815); furred at base to about the same extent
brevicaudata Erxleben, 1777 (including bra- dorsally as ventrally (e.g., in M. emiliae), or
chyuros Schreber, 1777; dorsalis J.A. Allen, caudal fur extends farther dorsally than
1904; hunteri Waterhouse, 1841; orinoci ventrally (e.g., M. brevicaudata), or tail base
Thomas, 1899; sebae Gray, 1827; surinamen- unfurred (e.g., M. peruviana); unfurred caudal
sis Zimmermann, 1780; touan Bechstein, surfaces covered with macroscopic bristlelike
1800; touan Shaw, 1800; touan Daudin, hairs, not naked-appearing; caudal scales often
1802; and tricolor E. Geoffroy, 1803); dimi- inapparent but always in annular series;
diata Wagner, 1847 (including fosteri Thom- relationship between caudal scales and hairs
as, 1924); domestica Wagner, 1842 (including usually obscure, but subequal hairs usually
concolor Gervais, 1856); emiliae Thomas, arranged in triplets; ventral caudal surface not
1912; glirina Wagner, 1842; handleyi Solari, modified for prehension.
2007; iheringi Thomas, 1888; kunsi Pine, Premaxillary rostral process absent. Nasals
1975; maraxina Thomas, 1923; osgoodi long, extending anteriorly beyond I1 (con-
Fig. 42. Monodelphis brevicaudata (based on AMNH 257203, an adult female from San Ignacio de
Yuruan, Bolvar, Venezuela).
cealing nasal orifice from dorsal view), and Two mental foramina present on lateral
conspicuously widened posteriorly near max- surface of each hemimandible; angular pro-
illary-frontal suture. Maxillary turbinals cess acute and strongly inflected.
(viewed through the nasal orifice) simple or Unworn crowns of I2I5 symmetrically
sparsely ornamented scrolls, not elaborately rhomboidal (premolariform), with sub-
branched. Lacrimal foramina (usually two on equal anterior and posterior cutting edges,
each side) prominently exposed on orbital and usually increasing in length (mesiodistal
margin or on face anterior to orbit. Orbits dimension) from I2 to I5. Upper canine (C1)
small, interorbital region more or less parallel alveolus in premaxillary-maxillary suture; C1
sided (usually without conspicuous constric- usually simple (without accessory cusps), but
tions); supraorbital margins smoothly round- small posterior accessory cusp sometimes
ed, without beads or distinct postorbital present (e.g., in M. peruviana). First upper
processes (but blunt, indistinct processes premolar (P1) smaller than posterior premo-
occasionally developed in large specimens of lars but well formed and not vestigial; third
some species; e.g., M. emiliae). Parietal and upper premolar (P3) taller than P2; P3 with
alisphenoid in contact on lateral braincase posterior cutting edge only; upper milk
(no frontal-squamosal contact). Sagittal crest premolar (dP3) large and molariform. Mo-
absent (e.g., in M. theresa) or small (usually lars highly carnassialized (postmetacristae
not extending to frontals; e.g., in M. brevi- much longer than postprotocristae); relative
caudata). Petrosal not exposed laterally widths consistently M1 , M2 , M3 , M4;
through fenestra in parietal-squamosal suture centrocrista strongly inflected labially on M1
(fenestra absent). Parietal-mastoid contact M3; ectoflexus shallow on M1, deeper on M2,
usually present (interparietal seldom contacts and consistently deep on M3; anterolabial
squamosal). cingulum and preprotocrista discontinuous
Maxillopalatine fenestrae present; palatine (anterior cingulum incomplete) on M3. Last
fenestrae usually absent; maxillary fenestrae upper tooth to erupt is P3 in some species (e.g.,
absent; posterolateral palatal foramina small, M. peruviana), or P3 and M4 erupt simulta-
not extending anteriorly between M4 proto- neously (e.g., in M. brevicaudata).
cones; posterior palatal morphology con- Lower incisors (i1i4) with distinct lingual
forms to Didelphis morphotype (with mod- cusps. Second lower premolar (p2) subequal
erately well-developed lateral corners, the in height to p3 (e.g., in M. brevicaudata), or p3
choanae somewhat constricted behind). taller than p2 (e.g., in M. emiliae); lower milk
Maxillary and alisphenoid in contact on floor premolar (dp3) trigonid usually complete
of orbit (not separated by palatine). Trans- (tricuspid). Hypoconid lingual to protoconid
verse canal foramen present. Alisphenoid (not labially salient) on m3; hypoconulid
tympanic process smoothly globular; pos- twinned with entoconid on m1m3; entoconid
teromedial lamina forming secondary fora- subequal to or smaller than hypoconulid on
men ovale present in some species (e.g., M. m1m3.
theresa) or lamina and secondary foramen DISTRIBUTION: Species of Monodelphis
ovale absent (e.g., in M. brevicaudata). occur in lowland and montane rain forests
Anterior limb of ectotympanic suspended and dry forests from eastern Panama
directly from basicranium. Stapes triangular throughout most of tropical and subtropical
with large obturator foramen (e.g., in M. South America to about 37uS in eastern
brevicaudata), or columellar and microperfo- Argentina (see range maps in Pine and
rate or imperforate (e.g., in M. peruviana). Handley, 2008); recorded elevations range
Fenestra cochleae exposed in most species, from sea level to at least 2500 m (on the
but fenestra concealed in sinus formed by eastern slopes of the tropical Andes). The
rostral and caudal tympanic processes of apparent absence of Monodelphis from the
petrosal in M. emiliae. Paroccipital process trans-Andean lowlands of western South
small, rounded, and adnate to petrosal. America is noteworthy, but this is possibly
Dorsal margin of foramen magnum bordered an artifact of inadequate collecting; a repre-
by supraoccipital and exoccipitals, incisura sentative of the M. adusta complex is
occipitalis present. rumored to occur along the Pacific littoral
of Colombia and northern Ecuador (Solari, Premaxillary rostral process absent. Nasals
2007), but no specimen-based records have long, extending anteriorly beyond I1 (con-
yet been published from that region. cealing nasal orifice from dorsal view), and
REMARKS: The monophyly of Monodelphis usually widened posteriorly near maxillary-
vis-a-vis other Recent didelphids is convinc- frontal suture. Maxillary turbinals elaborate-
ingly supported by 10 nonmolecular charac- ly branched. Lacrimal foramina visible in
ters (appendix 5); by a uniquely shared and lateral view at or near anterior orbital
unreversed deletion at the BRCA1 locus margin, usually two on each side. Flattened,
(fig. 31); and by sequence data from five triangular postorbital processes well devel-
genes analyzed separately (figs. 2832), to- oped in fully adult specimens. Left and right
gether (fig. 33), and in combination with frontals and parietals separated by persistent
morphology and karyotypes (figs. 35, 36). median sutures. Parietal and alisphenoid in
contact on lateral braincase (no frontal-
Tlacuatzin Voss and Jansa, 2003 squamosal contact. Sagittal crest absent.
Figure 43 Petrosal not exposed laterally through fenes-
tra in parietal-squamosal suture (fenestra
CONTENTS: canescens J.A. Allen, 1893 absent). Parietal-mastoid contact present
(including gaumeri Osgood, 1913; insularis (interparietal does not contact squamosal).
Merriam, 1908; oaxacae Merriam, 1897; and Maxillopalatine fenestrae present; palatine
sinaloae J.A. Allen, 1898). fenestrae absent; maxillary fenestrae present,
MORPHOLOGICAL DESCRIPTION: Combined usually opposite M2, sometimes partially
length of adult head and body ca. 110 confluent with maxillopalatine openings; pos-
160 mm; adult weight ca. 3070 g. Rhinar- terolateral palatal foramina small, not ex-
ium with two ventrolateral grooves on each tending anteriorly between M4 protocones;
side of median sulcus; dark circumocular posterior palatal morphology conforms to
mask present; pale supraocular spot absent; Didelphis morphotype (with well-developed
dark midrostral stripe absent; throat gland lateral corners, the choanae constricted be-
absent. Dorsal pelage unpatterned, usually hind). Maxillary and alisphenoid not in
pale gray or grayish brown; dorsal underfur contact on floor of orbit (separated by
gray; dorsal guard hairs short and inconspic- palatine). Transverse canal foramen present.
uous; ventral fur self-whitish, -cream, or Alisphenoid tympanic process smoothly
-buffy. Manus paraxonic (dIII 5 dIV); globular, without anteromedial process or
manual claws about as long as fleshy apical posteromedial lamina enclosing extracranial
pads of digits; dermatoglyph-bearing manual course of mandibular nerve (secondary fora-
plantar pads present; central palmar epithe- men ovale absent), and not in contact with
lium sparsely tuberculate; adult males with rostral tympanic process of squamosal. An-
well-developed lateral carpal tubercles but terior limb of ectotympanic directly suspend-
not medial carpal tubercles. Pedal digits ed from basicranium. Stapes usually subtrian-
unwebbed; dIV longer than other pedal gular and microperforate or imperforate
digits; plantar surface of heel naked. Pouch (rarely triangular with a large obturator
absent; mammae 414 5 9 or 515 5 11, foramen). Fenestra cochleae exposed, not
all abdominal-inguinal; cloaca present. Tail concealed by rostral and caudal tympanic
about as long as combined length of head and processes of petrosal. Paroccipital process of
body, slender and muscular (not incrassate); exoccipital small, rounded, adnate to petro-
body pelage present on basal 1/10 or less of sal. Dorsal margin of foramen magnum
tail; naked caudal integument uniformly bordered by supraoccipital and exoccipitals,
grayish or faintly bicolored (darker dorsally incisura occipitalis present.
than ventrally), occasionally parti-colored Two mental foramina present on lateral
(with a whitish tip); caudal scales in annular surface of each hemimandible; angular pro-
series, each scale with three subequal bristle- cess acute and strongly inflected.
like hairs emerging from distal margin; ventral Unworn crowns of I2I5 symmetrically
caudal surface modified for prehension distal- rhomboidal (premolariform), with sub-
ly, with apical pad bearing dermatoglyphs. equal anterior and posterior cutting edges,
Fig. 43. Tlacuatzin canescens (based on primarily on USNM 511261, an adult female from Rancho
Sapotito, Nayarit, Mexico; some dental details were reconstructed from USNM 125659, an adult male
from Los Reyes, Michoacan, Mexico).
and subequal in length (mesiodistal dimen- new name would then be needed for the clade
sion) from I2 to I5. Upper canine (C1) now awkwardly referred to in the literature
alveolus in premaxillary-maxillary suture; as the large 2n 5 22 opossums (Chiro-
C1 simple, without accessory cusps. First nectes, Didelphis, Lutreolina, Philander). Be-
upper premolar (P1) present, smaller than cause many morphological characters unam-
posterior premolars but well formed and not biguously diagnose the latter group, whereas
vestigial; second and third upper premolars only a few diagnose the group that includes
(P2 and P3) subequal in height; P3 with the large 2n 5 22 opossums plus Metachirus
posterior cutting edge only; upper milk (see appendix 5), the present arrangement
premolar (dP3) large and molariform. Mo- seems preferable. Additionally, because the
lars moderately carnassialized (postmetacris- branch leading to Metachirus is a long one
tae longer than postprotocristae); relative (e.g., in fig. 33), we expect that fossils will
widths usually M1 , M2 , M3 , M4; eventually be found to occupy it, and that
centrocrista strongly inflected labially on such discoveries will minimally disrupt the
M1M3; ectoflexus usually absent on M1, suprageneric nomenclature if a tribe is
shallow on M2, deep and distinct on M3; already available to accommodate them.
anterolabial cingulum continuous with pre- The name Metachirini was credited by
protocrista (complete anterior cingulum pre- Gardner and Dagosto (2008) to Reig et al.
sent) on M3; postprotocrista without carnas- (1985), but no family-group name based on
sial notch. Last upper tooth to erupt is P3. Metachirus fulfills the technical criteria for
Lower incisors (i1i4) with distinct lingual availability (ICZN, 1999: Article 13) prior to
cusps. Lower canine (c1) erect, acutely Hershkovitz (1992b), who effectively diag-
pointed, and simple (without a posterior nosed Metachirinae in a key. By the Principle
accessory cusp). Second lower premolar (p2) of Coordination (ICZN, 1999: Article 36),
taller than p3; lower milk premolar (dp3) Metachirini is available with the same
with incomplete (usually bicuspid) trigonid. authorship and date.
Hypoconid labially salient on m3; hypocon-
ulid twinned with entoconid on m1m3; Metachirus Burmeister, 1854
entoconid taller than hypoconulid on m1 Figure 44
m2, usually subequal in height to hypocon-
CONTENTS: nudicaudatus E. Geoffroy, 1803
ulid on m3.
(including antioquiae J.A. Allen, 1916; colom-
DISTRIBUTION: Tlacuatzin is apparently
bianus J.A. Allen, 1900; bolivianus J.A. Allen,
endemic to Mexico, where it occurs in
1901; dentaneus Goldman, 1912; imbutus
tropical dry forests from Sonora southward
Thomas, 1923; infuscus Thomas, 1923; mod-
(principally along the Pacific littoral and
estus Thomas, 1923; myosuros Temminck,
adjacent slopes of the coastal cordilleras) to
1824; personatus Miranda-Ribeiro, 1936;
Oaxaca and Chiapas; isolated populations
phaeurus Thomas, 1901; and tschudii J.A.
also occur in the northern part of the
Allen, 1900).
Yucatan Peninsula and on the Tres Maras
MORPHOLOGICAL DESCRIPTION: Combined
Islands (Hall, 1981; Wilson, 1991; Reid, 1997;
length of adult head and body ca. 250
Zarza et al., 2003).
300 mm; adult weight ca. 250500 g. Rhinar-
Tribe Metachirini Hershkovitz, 1992 ium with one ventrolateral groove on each
side of median sulcus; dark circumocular
CONTENTS: Metachirus. mask present, continuous with dark coronal
DIAGNOSIS: Metachirini is uniquely distin- fur; pale supraocular spot present; dark
guished from other members of the subfamily midrostral stripe absent; throat gland present
Didelphinae (indeed, from all other didel- in adult males. Dorsal pelage unpatterned,
phids) by contact between the frontal and usually some shade of grayish or reddish
squamosal on the lateral surface of the brown; dorsal underfur gray; dorsal guard
braincase. hairs short and inconspicuous; ventral fur
REMARKS: Although Metachirus could self-whitish, -cream, or -buffy. Manus mes-
logically be referred to the Didelphini, a axonic (dIII . dIV); manual claws shorter
Fig. 44. Metachirus nudicaudatus (based on AMNH 267009, an adult male from Paracou, French
Guiana).
than fleshy apical pads of digits; dermato- digits unwebbed; dIV longer than other pedal
glyph-bearing manual plantar pads present; digits; plantar surface of heel naked. Pouch
central palmar epithelium smooth or sparsely absent; mammae 414 5 9, all abdominal-
tuberculate; carpal tubercles absent. Pedal inguinal; cloaca present. Tail longer than
combined length of head and body, slender large, erect, projecting ventrally (not adnate
and muscular (not incrassate), unfurred to petrosal). Dorsal margin of foramen
except at base; naked caudal integument magnum bordered by exoccipitals only,
bicolored basally (brownish or grayish dor- incisura occipitalis absent.
sally, paler ventrally), gradually becoming all Two mental foramina present on lateral
pale distally; caudal scales in both annular surface of each hemimandible; angular pro-
and spiral series (neither pattern predominat- cess acute and strongly inflected.
ing), each scale with three subequal bristlelike Unworn crowns of I2I5 symmetrically
hairs emerging from distal margin; ventral rhomboidal (premolariform), with sub-
caudal surface not externally modified for equal anterior and posterior cutting edges,
prehension. and increasing in length (mesiodistal dimen-
Premaxillary rostral process absent. Nasals sion) from I2 to I5. Upper canine (C1)
long, extending anteriorly above or beyond alveolus in premaxillary-maxillary suture;
I1 (concealing nasal orifice from dorsal view), C1 simple, without accessory cusps. First
and conspicuously widened posteriorly near upper premolar (P1) smaller than posterior
maxillary-frontal suture. Maxillary turbinals premolars but well formed and not vestigial;
elaborately branched. Two lacrimal foramina second and third upper premolars (P2 and
laterally exposed on each side on or just P3) subequal in height; P3 with posterior
anterior to orbital margin. Interorbital region cutting edge only; upper milk premolar (dP3)
very broad, with distinctly beaded supraorbital large and molariform. Molars highly carnas-
margins; postorbital processes absent. Left sialized (postmetacristae much longer than
and right frontals and parietals separated by postprotocristae); relative widths M1 , M2
persistent median sutures. Frontal and squa- , M3 . M4 or M1 , M2 , M3 , M4;
mosal in contact on lateral braincase (no centrocrista strongly inflected labially on
parietal-alisphenoid contact). Sagittal crest M1M3; ectoflexus usually shallow on M1
absent or weakly developed just anterior to and M2 but distinct on M3; anterolabial
occiput. Petrosal not exposed laterally cingulum and preprotocrista usually discon-
through fenestra in parietal-squamosal suture
tinuous (anterior cingulum usually incom-
(fenestra absent). Parietal-mastoid contact
plete)28 on M3; postprotocrista without
normally present (interparietal usually does
carnassial notch. Third upper premolar (P3)
not contact squamosal).
and M4 erupt simultaneously.
Maxillopalatine fenestrae present; palatine
Lower incisors (i1i4) with distinct lingual
and maxillary fenestrae absent; posterolateral
cusps. Lower canine (c1) procumbent, with
palatal foramina small, not extending anteri-
flattened bladelike anterior margin, but
orly between M4 protocones; posterior pal-
usually without a distinct posterior accessory
atal morphology conforms to Didelphis
cusp. Second lower premolar (p2) taller than
morphotype (with well-developed lateral cor-
ners, the choanae constricted behind). Max- p3; lower milk premolar (dp3) trigonid
illary and alisphenoid usually not in contact complete (tricuspid). Hypoconid labially sa-
on floor of orbit (separated by palatine in lient on m3; hypoconulid twinned with
most examined specimens). Transverse canal entoconid on m1m3; entoconid much taller
foramen present. Alisphenoid tympanic pro- than hypoconulid on m1m3.
cess laterally compressed (not globular), with DISTRIBUTION: Metachirus is widely dis-
anteromedial process enclosing extracranial tributed in humid lowland and lower mon-
course of mandibular nerve (secondary fora- tane forests (usually below 2000 m) from
men ovale present), and not in contact with Chiapas, Mexico (Medelln et al., 1992) to
rostral tympanic process of petrosal. Anteri- northern Argentina (Flores et al., 2007).
or limb of ectotympanic suspended directly Published distribution maps (e.g., Reid,
from basicranium. Stapes triangular with 28
large obturator foramen. Fenestra cochleae Because this character is normally quite constant within
species, it is noteworthy that one specimen from Paracou,
usually exposed, not or incompletely con- French Guiana (AMNH 266451), has a complete anterior
cealed by rostral and caudal tympanic cingulum whereas others from the same locality (e.g., AMNH
processes of petrosal. Paroccipital process 267362) have an incomplete anterior cingulum.
1997; Gardner and Dagosto, 2008) indicate slaui Pohle, 1927; cayennensis Turton, 1800;
several range disjunctions, but some of these guianensis Kerr, 1792; gujanensis Link, 1795;
may be artifacts of inadequate collecting langsdorffi Boitard, 1845; palmata Daudin,
effort in suitable habitats. 1802; panamensis Goldman, 1914; paraguen-
REMARKS: Although only a single species sis Kerr, 1792; sarcovienna Shaw, 1800;
of Metachirus is currently regarded as valid, variegatus Olfers, 1818; and yapock Desmar-
recent mtDNA sequencing studies have est, 1820).
revealed deep molecular divergence (.10% MORPHOLOGICAL DESCRIPTION: Combined
at the mitochondrial cytochrome b locus; length of adult head and body ca. 250
Patton et al., 2000; Patton and Costa, 2003) 320 mm; adult weight ca. 510790 g. Rhinar-
among supposedly conspecific geographic ium with one ventrolateral groove on each
populations, and it seems likely that several side of median sulcus; dark circumocular
of the names currently listed as junior syno- mask present, continuous with dark coronal
nyms of M. nudicaudatus will be resurrected as fur (the whole top of the head except for an
valid species by future revisionary studies. indistinct pale bar over each eye is blackish);
pale supraocular spot absent; dark midrostral
Tribe Didelphini Gray, 1821 stripe absent; throat gland absent. Dorsal
pelage pale silvery gray with blackish trans-
CONTENTS: Chironectes, Didelphis, Lutreo-
verse bars connected by blackish middorsal
lina, Philander, and {Thylophorops.
stripe (when fresh; old museum skins become
DIAGNOSIS: Members of the tribe Didel-
brownish or reddish with age); dorsal under-
phini differ from all other members of the
fur pale gray; dorsal guard hairs short;
subfamily Didelphinae by their possession of
ventral fur self-white. Manus mesaxonic (dIII
a well-developed pouch, black-and-white
.dIV); manual claws shorter than fleshy
marked tail, nasal tips that do not extend
apical pads of digits; dermatoglyph-bearing
above I1, co-ossified frontals, well-developed
manual plantar pads absent (plantar epitheli-
sagittal crest, ectotympanic suspension via the
um uniformly and very densely covered with
malleus only, delayed eruption of M4, and
minutely denticulate tubercles among which
lower incisors without a distinct lingual cusp.
are scattered at regular intervals many
REMARKS: The monophyly of Didelphini is
smooth, hemispherical papillae); large, fleshy
strongly supported by most phylogenetic
carpal tubercle supported internally by pisi-
analyses published to date (e.g., Kirsch and
form present in both sexes. Pedal digits
Palma, 1995; Kirsch et al., 1997; Jansa and
webbed from base to terminal phalanges; dIV
Voss, 2000, 2005; Voss and Jansa, 2003, this
longer than other pedal digits; plantar surface
report; Steiner et al., 2005; Jansa et al., 2006;
of heel naked. Pouch present, opening
Gruber et al., 2007). The only exceptions
posteriorly; mammae 212 5 5; cloaca
appear to be primitive parsimony (Wagner)
absent (urogenital and rectal openings widely
trees, unaccompanied by support statistics,
separated by furred skin). Tail longer than
that were based largely on dental characters
combined length of head and body, thick and
(e.g., Reig et al., 1987: figs. 6467). All of the
muscular (not incrassate); body fur extends
diagnostic morphological features listed
onto basal one-sixth or less of tail, to about
above, together with several other cranioden-
the same extent dorsally as ventrally; naked
tal and karyotypic traits, optimize as unam-
caudal integument (beyond furry base) black-
biguous tribal synapomorphies (appendix 5).
ish with abruptly whitish tip (rarely all
The monophyly of Didelphini is additionally
blackish); caudal scales in spiral series, each
supported by a uniquely shared deletion at the
dorsal scale usually with three subequal
DMP1 locus (fig. 29), and by two uniquely
bristlelike hairs emerging from distal margin
shared deletions at the BRCA1 locus (fig. 31).
(ventral scales often have 45 hairs each);
Chironectes Illiger, 1811 ventral caudal surface not externally modi-
fied for prehension.
Figure 45
Premaxillary rostral process absent. Nasals
CONTENTS: minimus Zimmermann, 1780 short, not extending anteriorly above I1
(including argyrodytes Dickey, 1928; bres- (exposing nasal orifice in dorsal view), and
Fig. 45. Chironectes minimus (based on AMNH 96759, an adult male from Ilha do Taiuna, Para,
Brazil).
conspicuously widened posteriorly near max- C1 simple, without accessory cusps. First
illary-frontal suture. Maxillary turbinals upper premolar (P1) smaller than posterior
elaborately branched. One lacrimal foramen premolars but well formed and not vestigial;
usually present on each side just outside third upper premolar (P3) taller than P2; P3
anterior orbital margin. Blunt, hornlike with posterior cutting edge only; upper milk
postorbital frontal processes present in premolar (dP3) large and molariform. Upper
adults. Right and left frontals co-ossified, molars highly carnassialized (postmetacristae
midfrontal suture incomplete or absent; left much longer than postprotocristae); relative
and right parietals separated by persistent widths M1 , M2 , M3 . M4 or M1 , M2
midparietal suture. Parietal and alisphenoid , M3 , M4; centrocrista only weakly
in contact on lateral braincase (no frontal- inflected labially on M1M3; ectoflexus
squamosal contact). Sagittal crest present, shallow but distinct on M1 and M2, deepest
well developed on parietals, and extending on M3; anterolabial cingulum and prepro-
anteriorly onto frontals. Petrosal not laterally tocrista discontinuous (anterior cingulum
exposed through fenestra in parietal-squa- incomplete) on M3; postprotocrista with
mosal suture (fenestra absent). Parietal-mas- carnassial notch. Last upper tooth to erupt
toid contact absent (interparietal narrowly is M4.
contacts squamosal). Lower incisors (i1i4) without distinct
Maxillopalatine fenestrae present; palatine lingual cusps. Lower canine (c1) erect,
and maxillary fenestrae absent; posterolateral acutely pointed, and simple (without a
palatal foramina not extending anteriorly posterior accessory cusp). Second lower
between M4 protocones; posterior palatal premolar (p2) taller than p3; lower milk
morphology conforms to Didelphis morpho- premolar (dp3) large and molariform with a
type (with well-developed lateral corners, the complete (tricuspid) trigonid; hypoconid la-
choanae constricted behind). Maxillary and bially salient on m3; hypoconulid twinned
alisphenoid usually not in contact on floor of with entoconid on m1m3; entoconid much
orbit (uni- or bilateral contacts were observed taller than hypoconulid on m1m3.
as rare variants). Transverse canal foramen DISTRIBUTION: Chironectes occurs along
present. Alisphenoid tympanic process small streams in moist lowland and lower montane
and uninflated, usually with medial lamina forests (the highest elevational record ap-
enclosing extracranial course of mandibular pears to be about 1900 m; Handley, 1976)
nerve (secondary foramen ovale usually from the Mexican state of Oaxaca through-
present), and not in contact with rostral out most of Central America and tropical
tympanic process of petrosal. Anterior limb South America to northeastern Argentina
of ectotympanic suspended indirectly from (Misiones) and Uruguay (Gonzalez and
basicranium (by malleus). Stapes triangular, Fregueiro, 1998). As mapped by Marshall
with large obturator foramen. Fenestra (1978d), the geographic range of this genus is
cochleae exposed, not concealed by rostral strikingly disjunct, but collecting localities
and caudal tympanic processes of petrosal. subsequently reported by Anderson (1997),
Paroccipital process large and erect, not Linares (1998), and Brown (2004) have closed
adnate to petrosal. Dorsal margin of foramen many distributional gaps. As mapped by
magnum bordered by supraoccipital and Stein and Patton (2008a), Chironectes ap-
exoccipitals, incisura occipitalis present but pears to be absent from central Amazonia,
narrow. but an unvouchered record from the lower
One mental foramen present on lateral Tapajos (George et al., 1988) suggests that
surface of each hemimandible; angular process this hiatus may be an artifact of inadequate
acute and strongly inflected. collecting. Although water opossums are
Unworn crowns of I2I5 symmetrically known to inhabit gallery-forested streams in
rhomboidal (premolariform), with sub- the Cerrado (Mares et al., 1989), there are no
equal anterior and posterior cutting edges, published records from the Caatinga and
and subequal in length (mesiodistal dimen- Chaco.
sion) from I2 to I5. Upper canine (C1) REMARKS: Given the very broad geograph-
alveolus in premaxillary-maxillary suture; ic distribution of this unrevised genus, it
seems probable that at least some of the glyph-bearing manual plantar pads present;
many putative synonyms of Chironectes central palmar epithelium smooth or sparsely
minimus will be found to represent valid taxa tuberculate; carpal tubercles absent. Pedal
in future analyses of morphological and/or digits unwebbed; dIV slightly longer than
molecular data. other pedal digits in some species (e.g., D.
marsupialis) or dII, dIII, and dIV subequal
Didelphis Linnaeus, 1758 (e.g., in D. albiventris and D. virginiana);
Figure 46 plantar surface of heel naked. Pouch well
developed, opening anteriorly; mammae nor-
CONTENTS: albiventris Lund, 1840 (includ-
mally 313 5 7 to 616 5 13 or more29;
ing bonariensis Marelli, 1930; dennleri Mar-
cloaca present. Tail slightly longer than
elli, 1930; lechei Ihering, 1892; leucotis
combined length of head and body, slender
Wagner, 1847; paraguayensis J.A. Allen,
and muscular (not incrassate); basal 1/6 to
1902; poecilotis Wagner, 1842; and poecilo-
1/4 densely furred (covered with body pelage);
nota Schinz, 1844); aurita Wied-Neuweid,
naked caudal integument blackish proximally
1826 (including koseritzi Ihering, 1892; long-
and abruptly whitish distally; caudal scales in
ipilis Miranda-Ribeiro, 1935; and melanoidis
spiral series, each scale usually with three
Miranda-Ribeiro, 1935); imperfecta Mon-
subequal bristlelike hairs emerging from
dolfi and Perez-Hernandez, 1984; marsupialis
distal margin; ventral caudal surface variably
Linnaeus, 1758 (including battyi Thomas,
modified for prehension distally (prehensile
1902; cancrivora Gmelin, 1788; caucae J.A.
modifications are more strongly developed in
Allen, 1900; colombica J.A. Allen, 1900;
D. auritus and D. marsupialis than in the other
etensis J.A. Allen, 1902; insularis J.A. Allen,
species), but dermatoglyph-bearing apical
1902; karkinophaga Zimmermann, 1780; par-
pad consistently present.
ticeps Goldman, 1917; richmondi J.A. Allen,
Premaxillary rostral process absent. Nasals
1901; and tabascensis J.A. Allen, 1901);
short, not produced anteriorly above I1
pernigra J.A. Allen, 1900 (including andina
(exposing nasal orifice in dorsal view), and
J.A. Allen, 1902; and meridensis J.A. Allen,
conspicuously widened posteriorly near max-
1902); and virginiana Kerr, 1792 (including
illary-frontal suture. Maxillary turbinals
boreoamericana J.A. Allen, 1902; breviceps
elaborately branched. Lacrimal foramina
Bennett, 1833; californica Bennett, 1833;
one or two on each side, exposed laterally
cozumelae Merriam, 1901; illinensium Link,
on orbital margin or on face just anterior to
1795; pigra Bangs, 1898; pilosissima Link,
orbit. Postorbital processes bluntly pyrami-
1795; pruinosa Wagner, 1843; texensis J.A.
dal, maximally developed in old adults. Left
Allen, 1901; woapink Barton, 1806; and
and right frontals co-ossified (midfrontal
yucatanensis J.A. Allen, 1901).
suture incomplete or absent), but left and
right parietals separated by persistent mid-
length of adult head and body 310495 mm;
parietal suture. Parietal and alisphenoid in
adult weight 6005100 g. Rhinarium with
contact on lateral braincase (no frontal-
one ventrolateral groove on each side of
squamosal contact). Sagittal crest large, well
median sulcus; dark circumocular mask
developed, and extending onto frontals.
indistinct in some species (D. aurita, D.
Petrosal not exposed laterally through fenes-
marsupialis, D. virginiana) but distinct in
tra in squamosal-parietal suture (fenestra
others (D. albiventris, D. imperfecta, D.
absent). Parietal-mastoid contact absent (in-
pernigra); pale supraocular spot absent; dark
terparietal narrowly contacts squamosal).
midrostral stripe absent; throat gland absent.
Maxillopalatine and palatine fenestrae
Dorsal pelage unpatterned; dorsal underfur
present; maxillary fenestrae absent; postero-
white; dorsal guard hairs long, coarse and
lateral palatal foramina not extending ante-
conspicuous, giving the pelage a distinctively
riorly lingual to M4 protocones; posterior
shaggy appearance; ventral fur pale (usually
whitish) tipped with black. Manus mesaxonic 29
Occasional records of Didelphis virginiana with up to 17
(dIII . dIV); manual claws usually longer pouch young (cited by Hamilton, 1958) suggest that teat counts
than fleshy apical pads of digits; dermato- in this species may sometimes exceed 616 5 13.
Fig. 46. Didelphis marsupialis (based on AMNH 266459, an adult male from Paracou, French Guiana).
palate with prominent lateral corners, the canal foramen present. Alisphenoid tympanic
choanae abruptly constricted behind. Maxil- process small and uninflated, usually with
lary and alisphenoid not in contact (separat- posteromedial lamina enclosing extracranial
ed by palatine) on floor of orbit. Transverse course of mandibular nerve (secondary fora-
men ovale usually present), and not in America (except the Rockies, the northern
contact with rostral tympanic process of Great Plains, the Great Basin, the arid
petrosal. Anterior limb of ectotympanic southwest, north-central Mexico, and Baja
suspended indirectly from basicranium (by California; Hall, 1981), all of Central Amer-
malleus). Stapes usually triangular with large ica, and most of South America (to about
obturator foramen but sometimes columellar 40uS latitude in central Argentina; Flores et
and imperforate (varies within species). Fe- al., 2007). South American collection records
nestra cochleae exposed, not concealed by (mapped by Cerqueira and Tribe, 2008)
rostral and caudal tympanic processes of represent almost every non-desert tropical
petrosal. Paroccipital process large, erect (not and subtropical biome on the continent.
adnate to petrosal) and projecting ventrally. REMARKS: Despite contradictory or am-
Dorsal margin of foramen magnum bordered biguous phylogenetic results from most
by exoccipitals only, incisura occipitalis sequenced loci analyzed separately (table 14),
absent. the monophyly of Didelphis is strongly
Two mental foramina usually present on supported by parsimony and Bayesian anal-
lateral surface of each hemimandible (one yses of morphology (fig. 27), a concatenated
foramen is present unilaterally in a single five-gene dataset (fig. 33), and combined
examined specimen of D. albiventris); man- (nonmolecular + molecular) supermatrices
dibular angular process acute and strongly (figs. 35, 36). Revisionary studies of the
inflected medially. genus Didelphis in its modern sense were
Unworn crowns of I2I5 asymmetrical initiated by Allen (1901, 1902) and continued
(incisiform), with much longer anterior by Krumbiegel (1941), Gardner (1973), Cer-
than posterior cutting edges. Upper canine queira (1985), and Lemos and Cerqueira
(C1) alveolus in premaxillary-maxillary su- (2002). Useful summaries of morphological
ture; C1 simple, without accessory cusps. characters that distinguish Didelphis species
First upper premolar (P1) smaller than in local faunas are provided by Mondolfi and
posterior premolars, but well-formed and Perez-Hernandez (1984), Catzeflis et al.
not vestigial; third upper premolar (P3) taller (1997), Cerqueira and Lemos (2000), Flores
than second (P2); P3 with posterior cutting and Abdala (2001), and Ventura et al. (2002).
edge only; upper milk premolar (dP3) large Analyses of mitochondrial DNA sequence
and molariform. Molars highly carnassialized variation within and among South American
(postmetacristae conspicuously longer than species were reported by Patton et al. (2000)
postprotocristae); relative widths M1 , M2 and Patton and Costa (2003).
, M3 . M4 or M3 , M4; centrocrista only
weakly inflected labially on M1M3; ecto- Lutreolina Thomas, 1910
flexus shallow, indistinct, or absent on M1 Figure 47
and M2, but consistently deep and distinct on
M3; anterolabial cingulum and prepro- CONTENTS: crassicaudata Desmarest, 1804
tocrista discontinuous (anterior cingulum (including crassicaudis Olfers, 1818; bonaria
incomplete) on M3; postprotocrista with Thomas, 1923; ferruginea Larranaga, 1923;
carnassial notch. Last upper tooth to erupt lutrilla Thomas, 1923; macroura Desmoulins,
is M4. 1824; paranalis Thomas, 1923; travassosi
Lower incisors (i1i4) without distinct Mirando-Ribeiro, 1936; and turneri Gunther,
lingual cusps. Second lower premolar (p2) 1879).
much taller than p3; lower milk premolar MORPHOLOGICAL DESCRIPTION: Combined
(dp3) with complete (tricuspid) trigonid. length of adult head and body ca. 240
Hypoconid labially salient on m3; hypocon- 345 mm; adult weight ca. 300800 g). Rhi-
ulid twinned with entoconid on m1m3; narium with one ventrolateral groove on
entoconid much taller than hypoconulid on each side of median sulcus; head entirely pale
m1m3. and unmarked (dark circumocular mask, pale
DISTRIBUTION: Didelphis is a quintessen- supraocular spots, and dark midrostral stripe
tially eurytopic genus that ranges from absent); throat gland absent. Dorsal pelage
southern Canada throughout most of North unpatterned, usually some shade of yellowish
Fig. 47. Lutreolina crassicaudata (based AMNH 254513, an adult female from Punta Lara, Buenos
Aires, Argentina).
brown with grayish hair bases; dorsal guard verse canal foramen present. Alisphenoid
hairs short and inconspicuous; ventral fur tympanic process small and usually rounded
gray based or self-buffy. Manus mesaxonic (somewhat more inflated than in Chironectes
(dIII . dIV); manual claws much longer than or Didelphis), with broad lamina enclosing
fleshy apical pads of digits; dermatoglyph- extracranial course of mandibular nerve
bearing manual plantar pads present; central (secondary foramen ovale present), and not
palmar epithelium more or less smooth; in contact with rostral tympanic process of
carpal tubercles absent. Pedal digits un- petrosal. Anterior limb of ectotympanic
webbed; dIII longer than other pedal digits; suspended indirectly from braincase (by
plantar surface of heel naked. Pouch present, malleus). Stapes triangular with large obtu-
opening posteriorly; mammae 414 5 9 to rator foramen; fenestra cochleae exposed
515 5 11; cloaca present. Tail usually (not enclosed by bony laminae). Paroccipital
slightly shorter than combined length of head process large, erect, directed posteroven-
and body, thicker than in other large opos- trally. Dorsal margin of foramen magnum
sums but muscular, not incrassate; furred to bordered by exoccipitals only, incisura oc-
about the same extent dorsally and ventrally cipitalis absent.
for basal one-third to one-half; naked caudal Two mental foramina present on lateral
integument blackish proximally but abruptly surface of each hemimandible; angular pro-
whitish near tail tip; caudal scales in spiral cess acute and strongly inflected.
series, each scale usually with four or five Unworn crowns of I2I5 asymmetrical
subequal bristlelike hairs emerging from (incisiform), with much longer anterior
distal margin; ventral caudal surface not than posterior cutting edges. Upper canine
externally modified for prehension. (C1) alveolus in premaxillary-maxillary su-
Premaxillary rostral process absent. Nasals ture; C1 simple, without accessory cusps.
short, not extending anteriorly above I1 First upper premolar (P1) smaller than
(exposing nasal orifice in dorsal view), and posterior premolars but well formed and
widened posteriorly near maxillary-frontal not vestigial; third upper premolar (P3) taller
suture. Maxillary turbinals elaborately than P2; P3 with posterior cutting edge only;
branched. Lacrimal foramina two on each upper milk premolar (dP3) large and molar-
side, exposed laterally on orbital margin or iform. Upper molars highly carnassialized
on face just anterior to orbit. Short, blunt, (postmetacristae conspicuously longer than
hornlike postorbital processes usually present postprotocristae); relative widths M1 , M2
in large adult specimens. Left and right , M3 , M4; centrocrista only weakly
frontals co-ossified (midfrontal suture incom- inflected labially on M1M3; ectoflexus
plete or absent), but left and right parietals usually distinct only on M3; anterolabial
separated by persistent midparietal suture. cingulum and preprotocrista discontinuous
Parietal and alisphenoid in contact on lateral (anterior cingulum incomplete) on M3; post-
braincase (no frontal-squamosal contact). protocrista with carnassial notch. Last upper
Sagittal crest present, well developed on tooth to erupt is M4.
parietals and extending anteriorly onto fron- Lower incisors (i1i4) without distinct
tals. Petrosal not laterally exposed through lingual cusps. Lower canine (c1) erect,
fenestra in squamosal-parietal suture (fenes- acutely pointed, and simple (without a
tra absent). Parietal-mastoid contact absent posterior accessory cusp). Second lower
(interparietal narrowly contacts squamosal). premolar (p2) taller than p3; lower milk
Maxillopalatine and palatine fenestrae premolar (dp3) large and molariform with
present; maxillary fenestrae absent; postero- complete (tricuspid) trigonid. Hypoconid lin-
lateral palatal foramina not extending ante- gual to protoconid (not labially salient) on m3;
riorly between M4 protocones; posterior hypoconulid twinned with entoconid on m1
palatal morphology conforming to Didelphis m3; entoconid taller than hypoconulid on m1
morphotype (with well-developed lateral cor- and m2, but sometimes subequal in height to
ners, the choanae abruptly constricted be- hypoconulid on m3.
hind). Maxillary and alisphenoid in contact DISTRIBUTION: The known distribution of
(overlying palatine) on floor of orbit. Trans- Lutreolina consists of two disjunct regions of
nonforest vegetation in South America (Stein ium with one ventrolateral groove on each
and Patton, 2008b). Specimens traditionally side of median sulcus; dark circumocular
referred to the subspecies L. crassicaudata mask present, usually continuous with dark
turneri are from widely scattered localities in coronal fur (the coronal fur is not dark in
the savanna lowlands (below 900 m) in some examined specimens of P. frenatus);
eastern Colombia and central Venezuela, pale supraocular spot present; dark midrostral
and from the isolated but adjacent savannas stripe absent; throat gland absent. Dorsal
of southern Venezuela and northern Guyana pelage unpatterned grayish or blackish, or
(Voss, 1991: 91); although specimens of this with grayish flanks and black middorsal
form are currently unknown from Surinam stripe; dorsal underfur gray; dorsal guard
and French Guiana, they could be expected hairs usually short (longer middorsally than
to occur in the as yet poorly surveyed coastal along flanks in P. mcilhennyi); ventral fur
savannas of either (or both) countries and in variously pigmented, usually gray-based
the adjacent Brazilian state of Amapa. buffy or cream, or entirely grayish (variable
Southern South American specimens (repre- within and among examined taxa). Manus
senting several nominal taxa in addition to mesaxonic (dIII . dIV); manual claws about
the nominotypical subspecies L. c. crassicau- as long as fleshy apical pads of digits;
data) are much more commonly collected in a dermatoglyph-bearing manual plantar pads
wide range of tropical, subtropical, and present; central palmar epithelium smooth or
temperate habitats in Paraguay, Uruguay, sparsely tuberculate; carpal tubercles absent.
eastern Bolivia, southern Brazil, and north- Pedal digits unwebbed; dIV longer than other
ern Argentina (Flores et al., 2007; Stein and pedal digits; plantar surface of heel naked.
Patton, 2008b). Pouch present, opening anteriorly; mammae
REMARKS: Although several morphologi- usually 212 5 5 or 313 5 7; cloaca
cally diagnosed subspecies of Lutreolina have present. Tail longer than combined length of
been treated as valid by authors (e.g., head and body, slender and muscular (not
Thomas, 1923; Ximenez, 1967; Graipel et incrassate); furred dorsally and ventrally to
al., 1996; Stein and Patton, 2008b), no about the same extent for basal one-fifth;
published analysis of molecular variation naked caudal integument blackish proximally
has tested the implicit assumption that only and abruptly whitish distally; caudal scales
a single species occurs across the vast range in spiral series, each scale with 46 bristle-
of South American landscapes from which like hairs emerging from distal margin;
specimens have been collected. ventral caudal surface modified for prehen-
sion distally, with apical pad bearing derma-
Philander Brisson, 1762 toglyphs.
Figure 48 Premaxillary rostral process absent. Nasals
short, not extending anteriorly above I1
CONTENTS: andersoni Osgood, 1913 (in- (exposing nasal orifice in dorsal view), and
cluding nigratus Thomas, 1923); deltae Lew widened posteriorly near maxillary-frontal
et al., 2006; frenatus Olfers, 1818 (including suture. Maxillary turbinals elaborately
azaricus Thomas, 1923; quica Temminck, branched. Lacrimal foramina usually two
1824; and superciliaris Olfers, 1818); mcilhen- on each side, exposed laterally on orbital
nyi Gardner and Patton, 1972; mondolfii Lew margin or on face just anterior to orbit.
et al., 2006; olrogi Flores et al., 2008; and Interorbital region smoothly rounded, with-
opossum Linnaeus, 1758 (including canus out supraorbital beads or crests; short, blunt,
Osgood, 1913; crucialis Thomas, 1923; fusco- hornlike postorbital processes usually present
griseus J.A. Allen, 1900; grisescens J.A. Allen, in large adult specimens. Left and right
1901; melantho Thomas, 1923; melanurus frontals co-ossified (midfrontal suture incom-
Thomas, 1899; pallidus J.A. Allen, 1901; plete or absent), but left and right parietals
and virginianus Tiedemann, 1808). separated by persistent midparietal suture.
MORPHOLOGICAL DESCRIPTION: Combined Parietal and alisphenoid in contact on lateral
length of adult head and body ca. 250 braincase (no frontal-squamosal contact).
350 mm; adult weight ca. 280700 g. Rhinar- Sagittal crest present, well developed on
Fig. 48. Philander opossum (based primarily on AMNH 266387, an adult female from Paracou, French
Guiana; minor details were reconstructed from AMNH 266386 and 266994).
parietals and extending anteriorly onto fron- carnassial notch. Last upper tooth to erupt
tals. Petrosal not exposed laterally through is M4.
fenestra in squamosal-parietal suture (fenes- Lower incisors (i1i4) without distinct
tra absent). Parietal-mastoid contact absent lingual cusps. Lower canine (c1) erect,
(interparietal narrowly contacts squamo- acutely pointed, and simple (without a
sal). posterior accessory cusp). Second lower
Maxillopalatine and palatine fenestrae premolar (p2) much taller than p3; lower
present; maxillary fenestrae absent; postero- milk premolar (dp3) large and molariform
lateral palatal foramina small, not extending with complete (tricuspid) trigonid. Hypocon-
anteriorly between M4 protocones; posterior id labially salient on m3; hypoconulid
palatal morphology conforming to Didelphis twinned with entoconid on m1m3; entoco-
morphotype (with prominent lateral corners, nid much taller than hypoconulid on m1m3.
the choanae constricted behind). Maxillary DISTRIBUTION: Species of Philander occur
and alisphenoid usually separated by palatine in lowland and lower montane moist forests
on floor of orbit (maxillary-alisphenoid (to 1600 m; Reid, 1997) from the Mexican
contact occurs unilaterally or bilaterally in a state of Tamaulipas southward throughout
few specimens). Transverse canal foramen most of Central America (Hall, 1981) and
usually present. Alisphenoid tympanic pro- tropical South America (Patton and da Silva,
cess small and uninflated, usually with broad 2008) to northern Argentina (Chaco, For-
lamina enclosing extracranial course of man- mosa, and Misiones; Flores et al., 2007).
dibular nerve (secondary foramen ovale Collecting localities mapped by Anderson
present), and not contacting rostral tympanic (1997), Linares (1998), and Brown (2004)
process of petrosal. Anterior limb of ecto- suggest that Philander is largely absent from
tympanic indirectly suspended from basicra- intervening arid, semiarid, or treeless land-
nium (by malleus). Stapes usually triangular scapes (e.g., the Chaco, Cerrado, Caatinga,
with large obturator foramen. Fenestra and Llanos), with a few exceptions that may
cochleae exposed (not concealed by rostral have been taken in gallery forests.
and caudal tympanic processes of petrosal). REMARKS: Although inconsistently sup-
Paroccipital process large, erect, directed ported by our separate nuclear-gene analyses
posteroventrally. Dorsal margin of foramen (table 14) and by Patton et al.s (1996)
magnum bordered by exoccipitals only, analyses of sequence data from the mito-
incisura occipitalis absent. chondrial cytochrome b locus, the monophy-
Two mental foramina present on lateral ly of Philander is moderately well supported
surface of each hemimandible; angular pro- by nonmolecular characters (fig. 27) and it is
cess acute and strongly inflected. strongly supported by our analyses of con-
Unworn crowns of I2I5 asymmetrical catenated genes and combined datasets
(incisiform), with much longer anterior (figs. 33, 35, 36). The current species-level
than posterior cutting edges. Upper canine classification given above largely follows
(C1) alveolus in premaxillary-maxillary su- Patton and da Silva (1997), who analyzed
ture; C1 simple, without accessory cusps. mtDNA haplotype diversity among 42 Cen-
First upper premolar (P1) smaller than tral and South American specimens that they
posterior premolars but well-formed and interpreted as representing six valid taxa
not vestigial; third upper premolar (P3) taller (andersoni, canus, frenatus, fuscogriseus, mcil-
than P2; P3 with posterior cutting edge only; hennyi, opossum); three unsequenced taxa
upper milk premolar (dP3) large and molar- (azarica, melanurus, pallidus) were also rec-
iform. Molars highly carnassialized (postme- ognized as valid on the basis of morpholog-
tacristae conspicuously longer than postpro- ical traits. Whereas andersoni, frenatus, and
tocristae; relative widths M1 , M2 , M3 , mcilhennyi were recovered as monophyletic
M4; centrocrista only weakly inflected labi- haplotype groups that Patton and da Silva
ally on M1M3; ectoflexus usually distinct recognized as monotypic species, their con-
only on M3; anterolabial cingulum and cept of opossum included several subspecies
preprotocrista discontinuous (anterior cingu- (azarica, canus, fuscogriseus, melanurus, pal-
lum incomplete) on M3; postprotocrista with lida) and was not recovered as a clade. Based
on their phylogenetic analysis (op. cit.: anteromedial bullar process that spans the
fig. 2), it would be logical to recognize transverse canal foramen.
fuscogriseus and canus as valid species, but REMARKS: The monophyly of Thylamyini
the authors reasonably cautioned against (without Chacodelphys, from which sequence
facile taxonomic decisions based on the data are unavailable) is strongly supported
results of analyzing a single maternally by parsimony, likelihood, and Bayesian
inherited gene. Indeed, the genus clearly analyses of IRBP (fig. 28) and BRCA1
requires a more comprehensive revisionary (fig. 31); by likelihood and Bayesian analyses
treatment, with particular attention devoted of DMP1 (fig. 29); and by Bayesian analysis
to obtaining analyzable character data from of RAG1 (fig. 30). It is also strongly
hitherto unrepresented taxa (notably azarica, supported by all analyses of a concatenated
crucialis, grisescens, melanurus, and pallidus). five-gene dataset (fig. 33), and by Bayesian
Our placement of azarica in the synonymy of analysis of a combined (nonmolecular +
P. frenatus uncritically follows Gardner molecular) dataset that includes Chacodel-
(2005), who also allocated nominal taxa not phys. The presence of a fenestra in the
treated by Patton and da Silva (1997) to the parietal-squamosal suture (appendix 5) and
synonymy of P. opossum. Although an a uniquely derived insertion at the BRCA1
alternative species-level classification of Phi- locus (fig. 31) both optimize as thylamyine
lander proposed by Hershkovitz (1997) was synapomorphies.
not based on any explicit analysis of speci- Apparently, the earliest family-group
men data, it raised several issues not ade- name based on Thylamys is technically
quately treated by other authors and drew available from Hershkovitz (1992b), who
attention to the possible existence of unde- erroneously attributed authorship of Thyla-
scribed taxa. New species described by Lew et myinae to Reig et al. (1987).
al. (2006) and Flores et al. (2008) contribute
additional taxonomic complexities and un- Chacodelphys Voss et al., 2004
derscore the need for critical revisionary
CONTENTS: formosa Shamel, 1930 (includ-
work on this difficult genus.
ing muscula Shamel, 1930).
The tiresome nomenclatural controversy
regarding the use of Philander Tiedemann,
length of adult head and body 68 mm
1808, versus Metachirops Matschie, 1916, as
(external measurements are only available
the correct generic name for the gray four-
from the young adult holotype); adult weight
eyed opossums (Pine, 1973; Hershkovitz,
unknown but probably ca. 10 g. Rhinarial
1976; Gardner, 1981) was made moot by a
morphology unknown (no fluid-preserved
recent ruling of the International Commis-
specimens have been examined); dark cir-
sion on Zoological Nomenclature (ICZN,
cumocular mask present but narrow and
1998) that definitively established the avail-
inconspicuous; pale supraocular spot absent;
ability of Philander from Brisson (1762).
dark midrostral stripe absent; gular gland
Tribe Thylamyini Hershkovitz, 1992 present. Dorsal fur unpatterned, brownish,
somewhat darker middorsally than along
CONTENTS: Chacodelphys, Cryptonanus, flanks, but pelage not distinctly tricolored;
Gracilinanus, Lestodelphys, Marmosops, and dorsal underfur gray based; dorsal guard
Thylamys. hairs short and inconspicuous; ventral fur
DIAGNOSIS: Members of the tribe Thyla- gray-based buffy. Manus mesaxonic (dIII .
myini differ from all other didelphids by dIV); manual claws shorter than fleshy apical
having a fenestra in the parietal-squamosal pads of digits; manual plantar pads present,
suture through which the petrosal is visible but presence/absence of dermatoglyphs un-
on the lateral surface of the braincase. known; central palmar surface of manus
Additionally, most thylamyines (Chacodel- densely covered with small convex tubercles;
phys and Cryptonanus are exceptions) differ occurrence of carpal tubercles unknown.
from otherwise similar marmosines by having Pedal digits unwebbed; dIV slightly longer
a secondary foramen ovale formed by an than other pedal digits; plantar surface of
heel naked. Mammary formula, morphology by large obturator foramen. Fenestra cochle-
of pouch (if any), and presence/absence of ae exposed, not concealed by rostral and
cloaca unknown. Tail shorter than combined caudal tympanic processes of petrosal. Par-
length of head and body, slender and occipital process small, adnate to petrosal.
muscular (not incrassate); body pelage not Dorsal margin of foramen magnum formed
extending onto tail base; tail densely covered by supraoccipital and exoccipitals, incisura
with short hairs and distinctly bicolored occipitalis present.
(dark above, pale below); caudal scales Two mental foramina present on lateral
arranged in annular series, each scale with surface of each hemimandible; angular pro-
three subequal bristlelike hairs emerging cess acute and strongly inflected.
from distal margin; ventral caudal surface Unworn crowns of I2I5 symmetrically
not externally modified for prehension. rhomboidal (premolariform), with subeq-
Rostral process of premaxillae absent. ual anterior and posterior cutting edges, and
Nasals long, extending anteriorly beyond I1 increasing in length (mesiodistal dimension)
(concealing nasal orifice from dorsal view), from I2 to I5. Upper canine (C1) alveolus in
with subparallel lateral margins (not widened premaxillary-maxillary suture; C1 simple,
posteriorly). Maxillary turbinals large and without accessory cusps. First upper premo-
elaborately branched. Two lacrimal foramina lar (P1) smaller than posterior premolars but
laterally exposed on each side just anterior well formed and not vestigial; second and
to orbital margin. Supraorbital margins third upper premolars (P2 and P3) subequal
smoothly rounded, without beads or process- in height (see footnote 14, above); P3 with
es; strongly marked interorbital and postor- posterior cutting edge only; upper milk
bital constrictions present. Right and left premolar (dP3) morphology unknown (no
frontals and parietals separated by persistent juvenile specimens examined). Upper molars
median sutures. Parietal and alisphenoid highly carnassialized (postmetacristae con-
bones in contact (no squamosal-frontal spicuously longer than postprotocristae);
contact). Sagittal crest absent. Petrosal ex- relative widths M1 , M2 , M3 , M4;
posed laterally though fenestra in parietal- centrocrista strongly inflected labially on
squamosal suture. Parietal-mastoid contact M1M3; ectoflexus absent on M1, very
present (interparietal does not contact squa- shallow on M2, distinct only on M3; ante-
mosal). rolabial cingulum and preprotocrista dis-
Maxillopalatine fenestrae present and very continuous (anterior cingulum incomplete)
large; palatine fenestrae present but incom- on M3; postprotocrista without carnas-
pletely separated from maxillopalatine open- sial notch. Last upper tooth to erupt is prob-
ings; maxillary fenestrae very small but ably P3.
bilaterally present near M1/M2 commissure; Lower incisors (i1i4) with distinct lingual
posterolateral palatal foramina small, not cusps. Lower canine (c1) semiprocumbent,
extending lingual to M4 protocones; posteri- with flattened bladelike apex, but simple
or palate conforms to Didelphis morphotype (without a distinct posterior accessory cusp).
(with prominent lateral corners, the internal Second lower premolar (p2) taller than p3;
choanae abruptly constricted behind). Max- lower milk premolar (dp3) morphology
illary and alisphenoid not in contact on unknown (no juvenile specimens examined).
orbital floor (separated by palatine). Trans- Hypoconid lingual to protoconid (not labially
verse canal foramen present. Alisphenoid salient) on m3; hypoconulid twinned with
tympanic process probably smoothly globu- entoconid on m1m3; entoconid very small,
lar (broken in both examined specimens), subequal in height to hypoconulid on m1m3.
without anteromedial process or posterome- DISTRIBUTION: Chacodelphys is currently
dial lamina enclosing extracranial course of known from just five localities in the
mandibular nerve (secondary foramen ovale Argentinian provinces of Chaco and For-
absent), and probably not contacting rostral mosa (between 2527u S latitude and 5860u
tympanic process of petrosal. Anterior limb W longitude; Teta et al., 2006). However,
of ectotympanic directly suspended from because the humid savanna habitats that
basicranium. Stapes triangular, perforated occur at these sites resemble those found
elsewhere in northern Argentina, eastern of manus sparsely tubercular (neither smooth

Bolivia, western Paraguay, and southwestern nor densely covered with convex tubercles);
Brazil, it would be reasonable to expect that lateral carpal tubercles present in adult males.
the genus is more widely distributed. Pedal digits unwebbed; dIV longer than other
REMARKS: In the absence of undamaged pedal digits; plantar epithelium of heel
cranial material, we are unable to provide an naked. Pouch absent; mammae 414 5 9
illustration for this genus. However, photo- (all abdominal-inguinal) to 717 5 15 (with
graphs of two imperfect skulls were published pectoral teats); cloaca present. Tail longer
by Voss et al. (2004a: fig. 2) and Teta et al. than combined length of head and body,
(2006: fig. 2). slender and muscular (not incrassate); body
The problematic relationships of this genus pelage not extending more than a few mm
were discussed at length by Voss et al. onto tail base; unfurred caudal integument
(2004a) and seem unlikely to be resolved more or less bicolored (dark above, paler
definitively in the absence of nuclear gene below) in most specimens; caudal scales in
sequence data. As discussed above, the tribal distinctly annular series, each scale with three
membership of Chacodelphys is most com- subequal bristlelike hairs emerging from
pellingly supported by Bayesian analysis of a distal margin; ventral caudal surface modi-
dataset that combines nonmolecular charac- fied for prehension distally, with apical pad
ters with nuclear-gene sequence data (fig. 36), bearing dermatoglyphs.
but it is consistent with subjective assess- Rostral process of premaxillae absent.
ments of similarity based on integumental Nasals long, extending anteriorly beyond I1
and craniodental traits (e.g., by Tate, 1933). (concealing nasal orifice from dorsal view),
Although separate analyses of nonmolecular and conspicuously widened posteriorly near
character data suggest that this genus may maxillary-frontal suture. Maxillary turbinals
be more closely related to Thylamys and large and elaborately branched. Two lacrimal
Lestodelphys than to other thylamyines (fig. foramina laterally exposed on each side on or
27), it seems premature to formalize such just anterior to orbital margin. Supraorbital
weakly supported results by subtribal no- margins rounded, without beads or processes
menclature. (a few old individuals have incipient postor-
bital processes); distinct interorbital and
Cryptonanus Voss et al., 2005 postorbital constrictions usually present in
Figure 49 juveniles and young adults. Left and right
frontals and parietals separated by persistent
CONTENTS: agricolai Moojen, 1943; cha- median sutures. Parietal and alisphenoid in
coensis Tate, 1931; guahybae Tate, 1931; contact on lateral braincase (no squamosal-
ignitus Daz et al., 2002; and unduaviensis frontal contact). Sagittal crest absent. Petro-
Tate, 1931. sal laterally exposed through fenestra in
MORPHOLOGICAL DESCRIPTION: Combined parietal-squamosal suture. Parietal-mastoid
length of adult head and body ca. 80 contact present (interparietal does not con-
120 mm; adult weight ca. 1540 g. Ventral tact squamosal).
margin of rhinarium with two shallow Maxillopalatine fenestrae large; palatine
grooves on each side of median sulcus; dark fenestrae present; maxillary fenestrae absent;
circumocular mask present; pale supraocular posterolateral palatal foramina small, not
spot absent; dark midrostral stripe absent; extending lingual to M4 protocones; posteri-
gular gland present in adult males. Dorsal or palate conforms to Didelphis morphotype
body pelage unpatterned, usually grayish or (with prominent lateral corners, the internal
reddish brown; dorsal fur gray based; dorsal choanae abruptly constricted behind). Max-
guard hairs very short and inconspicuous; illary and alisphenoid not in contact on
ventral fur gray based or self-colored (vary- orbital floor (separated by palatine). Trans-
ing among species). Manus paraxonic (dIII 5 verse canal foramen present. Alisphenoid
dIV); manual claws shorter than fleshy apical tympanic process smoothly globular, without
pads of digits; dermatoglyph-bearing manual anteromedial process or posteromedial lam-
plantar pads present; central palmar surface ina enclosing extracranial course of mandib-
Fig. 49. Cryptonanus unduaviensis (based on AMNH 209152, an adult female from the Ro Baures,
Beni, Bolivia).
ular nerve (secondary foramen ovale absent), protoconid) on m3; hypoconulid twinned
and not contacting rostral tympanic process with entoconid on m1m3; entoconid much
of petrosal. Anterior limb of ectotympanic taller than hypoconulid on m1m3.
directly suspended from basicranium. Stapes DISTRIBUTION: Cryptonanus is known from
triangular, perforated by large obturator mostly unforested tropical and subtropical
foramen. Fenestra cochleae exposed, not biomes south of the Amazon River and east
concealed by rostral and caudal tympanic of the Andes, including the Caatinga, Cer-
processes of petrosal. Paroccipital process rado, Chaco, and northern Pampas; collec-
small, adnate to petrosal. Dorsal margin of tion localities mapped by Voss et al. (2005)
foramen magnum formed by supraoccipital and DEla and Martnez (2006) are from
and exoccipitals, incisura occipitalis present. Paraguay, Uruguay, eastern Bolivia, north-
Two mental foramina present on lateral ern Argentina, and eastern Brazil.
surface of each hemimandible; angular pro- REMARKS: Our species-level taxonomy of
cess acute and strongly inflected. Cryptonanus follows Voss et al. (2005) who,
Unworn crowns of I2I5 symmetrically however, cautioned that this arrangement is
rhomboidal (premolariform), with subeq- provisional and unsupported by rigorous
ual anterior and posterior cutting edges, tests of species limits. In particular, the
slightly increasing in length (mesiodistal genetic distinctness of C. agricolai, C. cha-
dimension) from I2 to I5. Upper canine coensis, C. ignitus, and C. unduaviensis is not
(C1) alveolus in premaxillary-maxillary su- overwhelmingly indicated by available mor-
ture; C1 usually with one or two small phological data and merits careful evalua-
accessory cusps (the posterior accessory cusp tion. Molecular sequence comparisons would
is more consistently distinct than the anterior be an especially welcome supplement to the
cusp). First upper premolar (P1) smaller than scant phenotypic evidence at hand.
posterior premolars but well formed and not
vestigial; third upper premolar (P3) taller Gracilinanus Gardner and Creighton, 1989
than P2; P3 with posterior cutting edge only; Figure 50
upper milk premolar (dP3) large and molar-
iform. Upper molars strongly carnassialized CONTENTS: aceramarcae Tate, 1931; agilis
(postmetacristae conspicuously longer than Burmeister, 1854 (including beatrix Thomas,
postprotocristae); relative widths M1 , M2 1910; buenavistae Tate, 1931; and peruana
, M3 . M4 or M1 , M2 , M3 , M4; Tate, 1931); dryas Thomas, 1898; emiliae
centrocrista strongly inflected labially on Thomas, 1909 (including longicaudus Hersh-
M1M3; ectoflexus shallow or absent on kovitz, 1992); marica Thomas, 1898 (includ-
M1, deeper on M2, and consistently deep on ing perijae Hershkovitz, 1992); and micro-
M3; anterolabial cingulum continuous with tarsus Wagner (1842).
preprotocrista (complete anterior cingulum MORPHOLOGICAL DESCRIPTION: Combined
present) on M3 in some species (e.g., C. length of adult head and body ca. 75
unduaviensis) but anterolabial cingulum and 130 mm; adult weight ca. 1050 g. Rhinar-
preprotocrista discontinuous (anterior cingu- ium with two ventrolateral grooves on each
lum incomplete) on M3 in others (e.g., C. side of median sulcus; dark circumocular
chacoensis); postprotocrista without carnas- mask present; pale supraocular spot absent;
sial notch. Last upper tooth to erupt is P3. dark midrostral stripe absent; throat gland
Lower incisors (i1i4) with distinct lingual present in adult males. Dorsal pelage un-
cusps. Lower canine (c1) procumbent, with patterned reddish or grayish brown with
flattened bladelike apex, usually with small dark-gray hair bases; dorsal guard hairs
posterior accessory cusp (but often absent on short and inconspicuous; ventral fur mostly
even moderately worn teeth). Second lower gray-based buff or orange (e.g., in G.
premolar (p2) taller than p3; lower milk microtarsus), but self-white in some species
premolar (dp3) with complete trigonid (tri- (e.g., G. emiliae). Manus paraxonic (dIII 5
cuspid in in C. chacoensis) or with incomplete dIV); manual claws small, not extending
trigonid (bicuspid in C. unduaviensis). Hypo- beyond fleshy apical pads of digits; der-
conid labially salient (level with labial apex of matoglyph-bearing manual plantar pads
Fig. 50. Gracilinanus agilis (based on MVZ 197437, an adult male from Mato do Vasco, Minas Gerais,
Brazil).
present; central palmar epithelium smooth or ners, the choanae constricted behind). Max-
sparsely tuberculate; lateral carpal tubercles illary and alisphenoid not in contact on floor
present in fully adult males, but medial carpal of orbit (separated by palatine). Transverse
tubercles absent in both sexes. Pedal digits canal foramen present. Alisphenoid tympanic
unwebbed; dIV longer than other pedal process flask shaped or globular (not laterally
digits; plantar surface of heel naked. Pouch compressed and without a well-developed
absent; mammary formula (unknown for ventral process), almost always with antero-
most species) 313 5 7 to 414 5 9; cloaca medial strut enclosing extracranial course of
present. Tail much longer than combined mandibular nerve (secondary foramen ovale
length of head and body, slender and usually present), and not contacting rostral
muscular (not incrassate), and macroscopi- tympanic process of petrosal. Anterior limb
cally naked (furred only near base); naked of ectotympanic directly suspended from
caudal integument usually unicolored (dark) basicranium. Stapes triangular, with large
but sometimes indistinctly bicolored (dark obturator foramen. Fenestra cochleae usually
above, paler below) at base; caudal scales in exposed (partially concealed by caudal tym-
annular series (most species) or annular and panic process of petrosal in some specimens
spiral series (G. emiliae), each scale with three of G. aceramarcae and G. microtarsus).
subequal bristlelike hairs emerging from Paroccipital process small, rounded, and
distal margin; ventral caudal surface modi- adnate to petrosal. Dorsal margin of foramen
fied for prehension distally (with naked magnum bordered by exoccipitals and supra-
median groove and apical pad bearing occipital, incisura occipitalis present.
dermatoglyphs). Two mental foramina present on lateral
Premaxillary rostral process present. Na- surface of each hemimandible; angular pro-
sals long, extending anteriorly beyond I1 cess acute and strongly inflected.
(concealing nasal orifice from dorsal view), Unworn crowns of I2I5 symmetrically
and conspicuously widened posteriorly near rhomboidal (premolariform), with sub-
maxillary-frontal suture. Maxillary turbinals equal anterior and posterior cutting edges,
elaborately branched. Two lacrimal foramina increasing in length (mesiodistal dimension)
present on each side but usually inconspicu- from I2 to I5 (e.g., in G. microtarsus), or I2
ous in lateral view (partially concealed within I5 crowns subequal in length (e.g., in G.
anterior orbital margin). Interorbital region emiliae). Upper canine (C1) alveolus in
more or less parallel-sided, without well- premaxillary-maxillary suture; C1 simple,
developed constrictions; supraorbital mar- without accessory cusps except in G. emiliae
gins smoothly rounded in most species (e.g., (which consistently has a posterior accessory
G. dryas) or with well-developed beads (in G. cusp). First upper premolar (P1) smaller than
emiliae); postorbital processes usually absent posterior premolars but well-formed and not
except in large adult specimens of G. micro- vestigial; second and third upper premolars
tarsus. Left and right frontals and parietals (P2 and P3) subequal in height; P3 with
separated by persistent median sutures. posterior cutting edge only; upper milk
Parietal and alisphenoid in contact on lateral premolar (dP3) large and molariform. Mo-
braincase (no squamosal-frontal contact). lars carnassialized (postmetacristae longer
Sagittal crest absent. Petrosal laterally ex- than postprotocristae); relative widths M1
posed through fenestra in parietal-squamosal , M2 , M3 . M4 or M1 , M2 , M3 ,
suture. Parietal-mastoid contact present (in- M4; centrocrista strongly inflected labially on
terparietal does not contact squamosal). M1M3; ectoflexus indistinct or absent on
Maxillopalatine fenestrae present; palatine M1 and M2 but consistently distinct on M3;
fenestrae present; maxillary fenestrae present anterolabial cingulum continuous with pre-
in most species (but very small and occasion- protocrista (complete anterior cingulum
ally absent uni- or bilaterally in G. emiliae); present) on M3; postprotocrista without
posterolateral palatal foramina small, not carnassial notch. Last upper tooth to erupt
extending lingual to M4 protocones; posteri- is P3.
or palatal morphology conforms to Didelphis Lower incisors (i1i4) with distinct lingual
morphotype (with well-developed lateral cor- cusps. Second lower premolar (p2) taller than
p3; lower milk premolar (dp3) trigonid not resemble any currently recognized taxa.
incomplete (unicuspid in the few specimens Therefore, the species-level diversity of Gra-
we were able to score for this trait). cilinanus seems certain to increase with future
Hypoconid labially salient on m3; hypocon- revisionary research.
ulid twinned with entoconid on m1m3; The monophyly of Gracilinanus in its
entoconid taller than hypoconulid on m1m3. currently restricted sense is strongly support-
DISTRIBUTION: Species of Gracilinanus ed by parsimony, likelihood, and Bayesian
occur in tropical and subtropical moist analyses of RAG1 (fig. 30), BRCA1 (fig. 31),
forests in lowland and montane landscapes vWF (fig. 32), concatenated sequence data
(to at least 3350 m; Voss et al., 2004b: table from five genes (fig. 33), and combined
12) from northern Venezuela to Bolivia, datasets that include both nonmolecular
Paraguay, southeastern Brazil, northeastern and molecular characters (figs. 35, 36); it is
Argentina, and northern Uruguay (DEla also moderately supported by parsimony and
and Martnez, 2006; Teta et al., 2007; likelihood analyses of IRBP (fig. 28). We
Creighton and Gardner, 2008a). Apparently, have no plausible explanation as to why
only one Argentinian record and two Ur- analyses of DMP1 strongly support the
uguayan records of Gracilinanus are valid, conflicting hypothesis that G. emiliae is the
the remaining published localities for this sister taxon of Cryptonanus (fig. 29). Al-
genus from both countries having been based though Gracilinanus is recovered as a clade
on specimens of Cryptonanus (see Voss et al., by parsimony analysis of nonmolecular data
2005; DEla and Martnez, 2006; Teta et al., (fig. 27), only two morphological traits opti-
2007). Gracilinanus is said not to occur in mize as unambiguous generic synapomor-
central Amazonia (Patton and Costa, 2003), phies (appendix 5).
but the species to be expected there (G.
emiliae) is difficult to collect and this Lestodelphys Tate, 1934
distributional hiatus is perhaps an artifact Figure 51
of inadequate faunal sampling.30 CONTENTS: halli Thomas, 1921.
REMARKS: The contents of this genus have MORPHOLOGICAL DESCRIPTION: Combined
changed significantly in the years since it was length of adult head and body probably ca.
first described by Gardner and Creighton 120160 mm; adult weight probably ca. 60
(1989), principally by removal of distantly 100 g (very few reliably measured specimens
related forms to new genera (see Voss et al., accompanied by weights are available). Rhi-
2005: table 1). With the exception of Graci- narium with one ventrolateral groove on each
linanus microtarsus and G. agilis, species of side of median sulcus; dark circumocular
this genus are not common in museum mask present; pale supraocular spot absent;
collections, and geographic distributions re- dark midrostral stripe absent; throat gland
main poorly documented. In addition, cur- present in adult males. Dorsal pelage tricol-
rent species concepts will doubtless change as ored (distinctly darker middorsally than on
more information becomes available. For flanks); dorsal underfur dark gray; dorsal
example, molecular data (summarized by guard hairs short and inconspicuous; ventral
Costa et al., 2003) suggest the presence of fur self-white from chin to anus. Manus
an undescribed, morphologically cryptic spe- mesaxonic (dIII . dIV); manual claws much
cies among the material currently identified longer than fleshy apical pads of digits;
as G. microtarsus; the geographic distribution dermatoglyph-bearing manual plantar pads
of material currently identified as G. emiliae present; central palmar epithelium densely
is improbably broad; and some specimens tuberculate; carpal tubercles absent. Pedal
from northern Colombia and eastern Peru do digits unwebbed; dIII longer than other pedal
digits; plantar surface of heel coarsely furred.
30
All of the central Amazonian records of Gracilinanus Pouch absent; mammae 818 5 17 or 919
emiliae mapped by Hershkovitz (1992b: fig. 1) and Brown
(2004: fig. 20) are either based on misidentified material or
5 19, of which the anteriormost teats are
specimens that have yet to be examined for reliably diagnostic pectoral; cloaca present. Tail shorter than
traits (Voss et al., 2001: 2930). combined length of head and body, incrassate,
Fig. 51. Lestodelphys halli (based on UWZM 22422, an adult male from Clemente Onelli, Ro Negro,
Argentina; and MVZ 173727, an adult male from Lihuel-Calel, La Pampa, Argentina).
and unfurred except at base; caudal integu- form and microperforate or imperforate.
ment bicolored (dark above, distinctly paler Fenestra cochleae concealed in sinus formed
below); tail scales difficult to distinguish but by rostral and caudal tympanic processes of
apparently in annular series; caudal hairs all petrosal. Paroccipital process small, rounded,
subequal in length and thickness; ventral and adnate to petrosal. Dorsal margin of
caudal surface not modified for prehension. foramen magnum bordered by supraoccipital
Premaxillary rostral process absent. Nasals and exoccipitals, incisura occipitalis present.
long, extending anterior to I1 (concealing Two mental foramina usually present on
nasal orifice in dorsal view), and conspicu- lateral surface of each hemimandible (three
ously widened posteriorly near premaxillary- foramina are present unilaterally on two
maxillary suture. Maxillary turbinals elabo- specimens examined); angular process acute
rately branched. Two lacrimal foramina and strongly inflected.
present on each side, exposed in lateral view Unworn crowns of I2I5 symmetrically
anterior to orbital margin. Supraorbital rhomboidal (premolariform), with sub-
margins smoothly rounded, without beads equal anterior and posterior cutting edges,
or crests; postorbital frontal processes usual- slightly increasing in length (mesiodistal
ly absent (indistinct processes are occasion- dimension) from I2 to I5. Upper canine
ally developed in old adult males). Left and (C1) alveolus in premaxillary-maxillary su-
right frontals and parietals separated by ture; C1 simple, without accessory cusps.
persistent median sutures. Parietal and ali- First upper premolar (P1) smaller than
sphenoid bones in contact on lateral surface posterior premolars but well formed and
of braincase (no frontal-squamosal contact). not vestigial; third upper premolar (P3) taller
Sagittal crest usually absent but weakly than P2; P3 with posterior cutting edge only;
developed along midparietal suture (not upper milk premolar (dP3) large and molar-
extending anteriorly to frontals) in some iform. Molars strongly carnassialized (post-
large specimens.31 Petrosal exposed through metacristae much longer than postproto-
fenestra in parietal-squamosal suture in some crista); relative widths M1 , M2 , M3 ,
individuals but not in others. Parietal-mas- M4; centrocrista strongly inflected labially on
toid contact present (interparietal does not M1M3; ectoflexus shallow or absent on M1,
contact squamosal). consistently present and distinct on M2,
Maxillopalatine fenestrae present but consistently deep on M3; anterolabial cingu-
sometimes quite small; palatine fenestrae lum and preprotocrista discontinuous (ante-
variable but usually present (Martin, 2005: rior cingulum incomplete) on M3; postpro-
fig. 2); maxillary fenestrae absent; posterolat- tocrista without carnassial notch. Last upper
eral palatal foramina very long, usually tooth to erupt is P3.
extending anteriorly lingual to M4 protocones; Lower incisors (i1i4) with distinct lingual
posterior palatal morphology conforms to cusps. Lower canine (c1) erect, acutely
Didelphis morphotype (with strongly pro- pointed, and simple (without a posterior
duced posterolateral corners, the choanae accessory cusp). Third lower premolar (p3)
constricted behind). Maxillary and alisphe- taller than p2; lower milk premolar (dp3)
noid not in contact (separated by palatine) on large, but trigonid incomplete (uni- or
floor of orbit. Transverse canal foramen bicuspid). Hypoconid lingual to protoconid
present. Alisphenoid tympanic process (not labially salient) on m3; hypoconulid
smoothly globular, with well-developed an- twinned with entoconid on m1m3; entoco-
teromedial process enclosing extracranial nid taller than hypoconulid on m1m3.
course of mandibular nerve (secondary fora- DISTRIBUTION: Most known specimens of
men ovale present), and not contacting Lestodelphys have been collected in semides-
rostral tympanic process of petrosal. Anteri- ert shrubland and steppe habitats in Patago-
or limb of ectotympanic directly suspended nian Argentina between 41u and 47uS lati-
from basicranium. Stapes usually columelli- tude, but there are two outlying records from
the Monte desert (between 32u and 38uS) that
31
Among the specimens we examined, a small sagittal crest is may represent a relictual population (Sau-
best developed in UWZM 22422. thier et al., 2007); all reported elevations
associated with collected specimens are less tar pads present; central palmar epithelium
than 1000 m above sea level (Birney et al., smooth or sparsely tuberculate; lateral carpal
1996; Martin, 2003). As noted by Flores et al. tubercles externally conspicuous on wrists of
(2007), the distribution of Lestodelphys ex- large adult males; medial carpal tubercles
tends further south than that of any other absent in both sexes. Pedal digits unwebbed;
living marsupial. dIV longer than other pedal digits; plantar
surface of heel macroscopically naked (a
Marmosops Matschie, 1916 microscopic pelage of very short hairs, not
Figures 52, 53 coarse fur, is usually present). Pouch absent;
mammae varying among examined species
CONTENTS: bishopi Pine, 1981; cracens
from 313 5 7 to 717 5 15, anteriormost
Handley and Gordon, 1979; creightoni Voss
pairs pectoral when mammae $11; cloaca
et al., 2004; fuscatus Thomas, 1896 (including
present. Tail longer than combined length of
carri J.A. Allen and Chapman, 1897; and
head and body, slender and muscular (not
perfuscus Thomas, 1924); handleyi Pine, 1981;
incrassate), and macroscopically naked
impavidus Tschudi, 1845 (including caucae
(without a conspicuously furred base); naked
Thomas, 1900; celicae Anthony, 1922; ma-
caudal integument uniformly dark (usually
descens Osgood, 1913; oroensis Anthony, grayish), bicolored (dark above, pale below),
1922; sobrinus Thomas, 1913; and ucayalensis or parti-colored (dark distally, paler distally);
Tate, 1931); incanus Lund, 1840 (including caudal scales in spiral series, each scale with
bahiensis Tate, 1931; and scapulatus Bur- three bristlelike hairs emerging from caudal
meister, 1856); invictus Goldman, 1912; margin; median hair of each caudal-scale
juninensis Tate, 1931; neblina Gardner, 1990; triplet usually much thicker and darker than
noctivagus Tschudi, 1844 (including albiven- lateral hairs; ventral caudal surface modified
tris Tate, 1931; collega Thomas, 1920; dor- for prehension distally (with naked median
othea Thomas, 1911; keaysi J.A. Allen, 1900; groove and apical pad bearing dermato-
leucastrus Thomas, 1927; lugendus Thomas, glyphs).
1927; neglectus Osgood, 1915; politus Cab- Premaxillary rostral process long and well
rera, 1913; purui Miller, 1913; stollei Miran- developed in some species (e.g., M. parvidens,
da-Ribeiro, 1936; and yungasensis Tate, M. pinheiroi) but short in others (e.g., M.
1931); ocellatus Tate, 1931; parvidens Tate, noctivagus) and apparently absent in some
1931; paulensis Tate, 1931; and pinheiroi Pine, (M. incanus). Nasals long, extending anteri-
1981 (including woodalli Pine, 1981). orly beyond I1 (concealing nasal orifice in
MORPHOLOGICAL DESCRIPTION: Combined dorsal view), and conspicuously widened
length of adult head and body ca. 90 posteriorly near maxillary-frontal suture ex-
195 mm; adult weight ca. 20140 g. Rhinar- cept in M. incanus (which has narrow, more
ium with two ventrolateral grooves on each or less parallel-sided nasals). Maxillary tur-
side of median sulcus; dark circumocular binals (viewed through nasal orifice) elabo-
mask present; pale supraocular spot absent; rately branched. Lacrimal foramina, usually
dark midrostral stripe absent; throat gland two on each side, concealed from lateral view
consistently present in adult males of some inside orbit (e.g., in M. parvidens) or exposed
species (e.g., M. incanus, M. noctivagus) but on orbital margin (e.g., in M. pinheiroi).
apparently absent in others (e.g., M. impavi- Interorbital region without abrupt constric-
dus, M. parvidens). Dorsal pelage unpat- tions; supraorbital margins rounded in some
terned, usually some shade of dull grayish species (e.g., M. parvidens), beaded in others
brown, but distinctly reddish or grayish in (e.g., M. noctivagus); postorbital processes
some species; dorsal hair bases always dark usually absent except in old males of some
gray; dorsal guard hairs short and inconspic- species (e.g., M. impavidus32). Left and right
uous; ventral fur self-colored (whitish, buffy, frontals and parietals separated by persistent
or brown) or entirely or partially gray based. median sutures. Parietal and alisphenoid in
Manus mesaxonic (dIII . dIV); manual
claws small, shorter than fleshy apical pads 32
Postorbital processes in this species are best developed in
of digits; dermatoglyph-bearing manual plan- AMNH 272760.
Fig. 52. Marmosops creightoni (based on CBF 6552 and UMMZ 155999, adult males from the valley of
the Ro Zongo, La Paz, Bolivia).
Fig. 53. Marmosops pinheiroi (based on AMNH 267345, a young adult male from Paracou, French
Guiana), a small species that differs from M. creightoni (fig. 52) in several trenchant craniodental features
including laterally exposed lacrimal foramina, absence of palatine vacuities, and presence of accessory
cusps on C1.
contact on lateral braincase (no frontal- height; P3 with posterior cutting edge only;
squamosal contact). Sagittal crest absent. upper milk premolar (dP3) large and molar-
Petrosal consistently exposed laterally through iform. Molars strongly carnassialized (post-
fenestra in parietal-squamosal suture in most metacristae much longer than postprotocris-
species (fenestra polymorphically absent in tae); relative widths usually M1 , M2 , M3
M. incanus and M. noctivagus). Parietal- , M4; centrocrista strongly inflected labially
mastoid contact present (interparietal does on M1M3; ectoflexus shallow or indistinct
not contact squamosal). on M1, shallow but usually distinct on M2,
Maxillopalatine fenestrae present; palatine and consistently deep on M3; anterolabial
fenestrae present in some species (e.g., M. cingulum and preprotocrista discontinuous
creightoni; fig. 52), but absent in others (e.g., (anterior cingulum incomplete) on M3 in
M. pinheiroi; fig. 53); maxillary fenestrae some species (e.g., M. noctivagus) but an-
absent; posterolateral palatal foramina small, terolabial cingulum continuous with prepro-
not extending anteriorly lingual to M4 tocrista (anterior cingulum complete) in
protocones; posterior palatal morphology others (e.g., M. parvidens); postprotocrista
conforms to Didelphis morphotype (with without carnassial notch. Last upper tooth to
prominent lateral corners, the choanae con- erupt is P3.
stricted behind). Maxillary and alisphenoid Lower incisors (i1i4) with distinct lingual
not in contact on floor of orbit (separated by cusps. Second lower premolar (p2) subequal
palatine). Transverse canal foramen present. in height to p3 in some species (e.g., M.
Alisphenoid tympanic process small, often incanus) but distinctly taller than p3 in others
bluntly conical or laterally compressed (never (e.g., M. impavidus); lower milk premolar
smoothly globular), always with a well- (dp3) trigonid complete. Hypoconid labially
developed anteromedial process enclosing salient on m3; hypoconulid twinned with
the extracranial course of mandibular nerve entoconid on m1m3; entoconid usually
(secondary foramen ovale present), and not taller than hypoconulid on m1m3.
contacting rostral tympanic process of petro- DISTRIBUTION: Species of Marmosops oc-
sal. Anterior limb of ectotympanic directly cur in tropical lowland and montane moist
suspended from basicranium. Stapes triangu- forests from central Panama southward
lar, with large obturator foramen. Fenestra along the Andes and throughout Amazonia
cochleae exposed, not concealed by rostral to eastern Bolivia and southeastern Brazil
and caudal tympanic processes of petrosal. (see maps in Gardner and Creighton, 2008a).
Paroccipital process small, rounded, and The genus is apparently unknown from
adnate to petrosal. Dorsal margin of foramen Paraguay, Argentina, and Uruguay. See
magnum bordered by supraoccipital and Mustrangi and Patton (1997) for verified
exoccipitals, incisura occipitalis present. specimen records based on modern revision-
Two mental foramina usually present on ary research in southeastern Brazil, Patton et
lateral surface of each hemimandible; angular al. (2000) for the same in western Amazonia,
process acute and strongly inflected. Voss et al. (2001) for northeastern Amazonia,
Unworn crowns of I2I5 symmetrically and Voss et al. (2004b) for Bolivia. Many
rhomboidal (premolariform), with sub- other published geographic data purporting
equal anterior and posterior cutting edges, to represent records of Marmosops species
increasing in length (mesiodistal dimension) (e.g., Anderson, 1997; Brown, 2004) are
from I2 to I5. Upper canine (C1) alveolus in based on misidentified material (Voss et al.,
premaxillary-maxillary suture; C1 without 2004b).
accessory cusps in some species (e.g., M. REMARKS: The monophyly of Marmosops
creightoni; fig. 52), or with posterior accesso- is uniformly strongly supported by parsimo-
ry cusp (e.g., M. juninensis), or with anterior ny, likelihood, and Bayesian analyses of
and posterior accessory cusps (e.g., M. IRBP (fig. 28), DMP1 (fig. 29), BRCA1
pinheiroi; fig 53). First upper premolar (P1) (fig. 31), vWF (fig. 32), and concatenated
smaller than posterior premolars but well sequence data from five genes (fig. 33). The
formed and not vestigial; second and third genus is also recovered with strong support
upper premolars (P2 and P3) subequal in by parsimony and Bayesian analyses of
combined datasets that include nonmolecular apparently also in adult females of some
and molecular characters (figs. 35, 36), and species (Carmignotto and Monfort, 2006).
by Bayesian analyses of nonmolecular char- Dorsal pelage usually grayish and distinctly
acters (fig. 27) and RAG1 sequences (fig. 30). darker middorally than laterally (tricol-
Spirally arranged caudal scales, the petiolate ored), but dorsal fur brownish and without
morphology of the central hair in each conspicuous patterning in T. karimii and T.
caudal-scale triplet, the presence of a rostral velutinus (Carmignotto and Monfort, 2006);
process of the premaxillae, and unique dorsal hair bases dark gray; dorsal guard
deletions at the DMP1 and BRCA1 loci hairs short and inconspicuous; ventral fur
optimize as unambiguous generic synapo- self-white in some species (e.g., T. pusillus),
morphies (appendix 5; figs. 29, 31). As noted wholly or partly gray-based whitish or
elsewhere, Kirsch and Palmas (1995) sugges- yellowish in others (e.g., T. venustus). Manus
tion that Marmosops is paraphyletic was mesaxonic (dIII . dIV); manual claws about
based on misidentified voucher material as long as fleshy apical pads in most species
(Voss and Jansa, 2003: 57). (but claws extend well beyond apical pads in
Discrepancies between our species-level T. karimii and T. velutinus; Carmignotto and
synonymies for Marmosops and those in Monfort, 2006); dermatoglyph-bearing man-
Gardner (2005) are explained by Voss et al. ual plantar pads present in most species;
(2004b). Few of the currently recognized central palmar epithelium densely covered
species have received critical revisionary with small convex tubercles; carpal tubercles
attention, and it seems likely that several absent in both sexes. Pedal digits unwebbed;
widespread taxa (e.g., M. fuscatus, M. nocti- dIV longer than other pedal digits; plantar
vagus, and M. impavidus) will prove to be surface of heel coarsely furred. Pouch absent;
composite. For regional revisions of Marmo- mammae varying among species from 414
sops see Mustrangi and Patton (1997), Patton 5 9 (in T. karimii; Carmignotto and Mon-
et al. (2000), and Voss et al. (2001, 2004b). fort, 2006) to 717 5 15 (in T. elegans; Tate,
1933), including pectoral teats (when the
Thylamys Gray, 1843 total number of mammae $ 11); cloaca
Figure 54 present. Tail longer than combined length
of head and body in most species (except in
CONTENTS: cinderella Thomas, 1902 (in- T. karimii and T. velutinus; Carmignotto and
cluding sponsorius Thomas, 1921); elegans Monfort, 2006); apparently always incrassate
Waterhouse, 1839 (including coquimbensis (but quantity of stored fat may vary season-
Tate, 1931; and soricinus Philippi, 1894); ally); furred only at base (to about the same
karimii Petter, 1968; macrurus Olfers, 1818 extent dorsally as ventrally); unfurred caudal
(including griseus Desmarest, 1827); pallidior integument bicolored (dark above, pale
Thomas, 1902; pusillus Desmarest, 1804 below) in most species and parti-colored in
(including bruchi Thomas, 1921; citellus some (T. macrurus has an all-white tail tip);
Thomas, 1912; nanus Olfers, 1818; and verax tail scales in annular series, each scale with
Thomas, 1921); tatei Handley, 1957; velutinus three subequal bristlelike hairs emerging
Wagner, 1842 (including pimelurus Rein- from distal margin; ventral caudal surface
hardt, 1851); and venustus Thomas, 1902 modified for prehension distally (with naked
(including janetta Thomas, 1926). median groove and apical pad bearing
MORPHOLOGICAL DESCRIPTION: Combined dermatoglyphs) in most species, but ventral
length of adult head and body ca. 75 prehensile surface reduced or absent in T.
140 mm; adult weight ca. 1065 g. Rhinar- karimii and T. velutinus (Carmignotto and
ium with two ventrolateral grooves on each Monfort, 2006).
side of median sulcus in most examined Premaxillary rostral process absent. Nasals
species (but T. pallidior has only a single long, extending anteriorly above or beyond
ventrolateral groove on each side); dark I1 (concealing nasal orifice in dorsal view),
circumocular mask present; pale supraocular and uniformly narrow (with subparallel lateral
spot absent; dark midrostral stripe absent; margins) in most specimens. Maxillary turbi-
throat gland present in fully adult males, and nals elaborately branched. Two lacrimal
Fig. 54. Thylamys pusillus (based on AMNH 275442, an adult female, and 275445, an adult male, both
from Estancia Bolvar, Tarija, Bolivia).
foramina present on each side, exposed to premaxillary-maxillary suture; C1 without

lateral view on anterior orbital margin in accessory cusps in all examined species (but
most species, but usually concealed inside see Carmignotto and Monfort, 2006).33 First
orbit in others (e.g., T. tatei). Interorbital upper premolar (P1) smaller than posterior
and postorbital constrictions often distinct; premolars but well formed and not vestigial;
supraorbital margins rounded in some spe- third upper premolar (P3) taller than P2; P3
cies (e.g., T. pallidior), squared or beaded in with posterior cutting edge only; upper milk
others; postorbital processes usually absent premolar (dP3) large, molariform, and non-
or indistinct but occasionally present in old vestigial. Molars strongly carnassialized
adults of some species (e.g., T. macrurus). (postmetacristae much longer than postpro-
Left and right frontals and parietals separat- tocristae); relative widths usually M1 , M2
ed by persistent median sutures. Parietal and , M3 , M4; centrocrista strongly inflected
alisphenoid in contact on lateral braincase labially on M1M3; ectoflexus absent or
(no frontal-squamosal contact). Sagittal crest indistinct on M1, usually present but shallow
absent. Petrosal almost always exposed on M2, consistently deep and distinct on M3;
laterally through fenestra in parietal-squa- anterolabial cingulum and preprotocrista
mosal suture. Parietal-mastoid contact pres- discontinuous (anterior cingulum incom-
ent (interparietal does not contact squamo- plete) on M3; postprotocrista without car-
sal). nassial notch. Last upper tooth to erupt is P3.
Maxillopalatine fenestrae present; palatine Lower incisors (i1i4) with distinct lingual
fenestrae present; maxillary fenestrae present cusps. Second and third upper premolars (p2
in some species (e.g., T. pusillus) but absent in and p3) subequal in height; lower milk
others (e.g., T. pallidior); posterolateral pala- premolar (dp3) trigonid usually incomplete
tal foramina very large, extending anteriorly (bicuspid). Hypoconid labially salient on m3;
between M4 protocones; posterior palatal hypoconulid twinned with entoconid on m1
morphology conforms to Didelphis morpho- m3; entoconid much taller than hypoconulid
type (with prominent lateral corners, the on m1m3.
choanae constricted behind). Maxillary and DISTRIBUTION: Species of Thylamys collec-
alisphenoid not in contact on floor of orbit tively range from north-central Peru (An-
(separated by palatine). Transverse canal cash) southward along the Andes and Pacific
foramen present. Alisphenoid tympanic pro- coastal lowlands to central Chile; in the
cess smoothly globular, with anteromedial unforested landscapes south of Amazonia,
process enclosing extracranial course of the genus extends eastward across Bolivia
mandibular nerve (secondary foramen ovale (Anderson, 1997), Paraguay, and Argentina
present), and not contacting rostral tympanic (Flores et al., 2007) to Uruguay (Gonzalez et
process of petrosal. Anterior limb of ecto- al., 2000) and eastern Brazil (Bahia, Pernam-
tympanic directly suspended from basicrani- buco, and Piau; Carmignotto and Monfort,
um. Stapes triangular, with large obturator 2006). Recorded elevations range from near
foramen. Fenestra cochleae concealed in sinus sea level to at least 3800 m (Solari, 2003).
formed by rostral and caudal tympanic pro- REMARKS: The monophyly of Thylamys is
cesses of petrosal. Paroccipital process small, strongly supported by parsimony, likelihood,
rounded, adnate to petrosal. Dorsal margin and Bayesian analyses of IRBP (fig. 28),
of foramen magnum bordered by supraoc- BRCA1 (fig. 31), vWF (fig. 32), and concat-
cipital and exoccipitals, incisura occipitalis
enated sequence data from five genes
present.
(fig. 33). Strong support for this clade is also
Two mental foramina present on lateral
surface of each hemimandible; angular pro- 33
Although several Brazilian species of Thylamys were scored
cess acute and strongly inflected. as polymorphic for presence of posterior accessory canine cusps
Unworn crowns of I2I5 symmetrically by Carmignotto and Monfort (2006: table 4), the increasing
rhomboidal (premolariform), with sub- frequencies that they observed in older age classes suggest that
C1 in these taxa is notched by occlusion with p1 in their
equal anterior and posterior cutting edges, material, and that the structure in question is not homologous
increasing in length (mesiodistal dimension) with the posterior accessory cusp scored from specimens with
from I2 to I5. Upper canine (C1) alveolus in unworn dentitions by Voss and Jansa (2003: character 53).
provided by parsimony and likelihood anal- fresh acknowledgment here. Nondidelphid

yses of DMP1 (fig. 29) and by parsimony marsupial sequences newly reported herein
and Bayesian analyses of RAG1 (fig. 30) and were obtained from samples contributed by
combined datasets that include both nonmo- Steve Donnellan (SAM), Milton Gallardo
lecular and molecular characters (figs. 35, (Universidad Austral de Chile), and the late
36). The monophyly of Thylamys is also Terry Yates (MSB). Monica Daz and Luca
recovered with moderate support by parsi- Luna provided important new morphometric
mony analysis of nonmolecular characters data for table 4.
(fig. 27), but only one morphological trait The illustrations that accompany our text
(uniformly narrow nasals; see appendix 5) are the work of two uniquely talented artists.
optimizes as an unambiguous generic syna- Pat Wynne (who drew figures 121, 23, and
pomorphy. 3754) has worked on this project off and on
We tentatively recognize bruchi as a for almost as long as we have, and is now
synonym of Thylamys pusillus following Voss without doubt the most accomplished illus-
et al. (in press), who explain why the former trator of opossum morphology to be found
name does not belong in the synonymy of T. anywhere. (Indeed, she is responsible for the
pallidior (contra Creighton and Gardner, discovery of several osteological characters
2008c), and we refer sponsorius to the that had escaped our notice until she drew
synonymy of T. cinderella following Braun them.) Pat worked patiently with us on
et al., (2005). The names pulchellus Cabrera, multiple drafts of each figure to get the
1934, and fenestrae Marelli, 1932, might be details right, and we are grateful for the skill,
synonyms of T. pusillus and T. pallidior, professionalism, and unfailing good humor
respectively, but we have not seen the that she invariably brought to this job. Fiona
holotypes and published information about Reid, peerless illustrator of Neotropical rain-
these nominal taxa is insufficient to support forest mammals, likewise tolerated many nit-
any definite conclusions about them. Many picking revisions of her preliminary sketches
other species-level issues in this genus remain before rendering the brilliantly finished final
problematic despite much recent taxonomic product. We thank Michael Goulding for
work (e.g., Palma et al., 2002; Solari, 2003; sending us David Schlesers photograph of
Braun et al., 2005; Carmignotto and Monfort, Marmosa murina (reproduced on the cover),
2006). and we thank Natures Images Inc. for
permission to use it.
ACKNOWLEDGMENTS As always, we are grateful to the support
staff of the Department of Mammalogy for
This protracted research project would their expert assistance at every stage of this
have been impossible without the generosity project, especially Pat Brunauer (who helped
and patience of many colleagues who loaned find many obscure references in our library),
us irreplaceable material, tolerated numerous Neil Duncan (osteological preparator ex-
requests for loan extensions, and hosted our traordinaire), and Eileen Westwig (who
visits to their institutional collections (abbre- managed countless loans and kept the
viated as in appendix 1). Among those whose paperwork in order). Museum travel and
assistance in these respects we have acknowl- other related expenses were supported from
edged in previous reports, we remain espe- the departments Taxonomic Mammalogy
cially grateful to Paula Jenkins (BMNH); Fund (endowed by the late Karl F. Koop-
Bruce Patterson and Bill Stanley (FMNH); man). Lastly, RSV acknowledges the prodi-
Jim Patton and Eileen Lacey (MVZ); Mark gious energies of his curatorial forbears at the
Engstrom and Burton Lim (ROM); Phil AMNH (especially George Tate and Syd
Myers (UMMZ); and Mike Carleton, Al Anderson) who amassed the morphological
Gardner, and Linda Gordon (USNM). Im- collections on which this project was largely
portant tissue samples contributed by Robert based.
Baker, Francois Catzeflis, Guillermo DEla, DNA sequencing work for this project
Louise Emmons, Chris Hice, the late John began at the Monell Molecular Laboratory
Kirsch, and Maria N.F. da Silva also merit of the AMNH, where SAJ was supported by
a Kalbfleisch postdoctoral research grant and Amrine-Madsen, H., M. Scally, M. Westerman,

by the Lewis B. and Dorothy Cullman M.J. Stanhope, C. Krajewski, and M.S. Spring-
Program for Molecular Systematic Studies. er. 2003. Nuclear gene sequences provide
evidence for the monophyly of australidelphian
Subsequent lab work was carried out at the
marsupials. Molecular Phylogenetics and Evo-
Laboratory of Molecular Systematics of the lution 28: 186196.
National Museum of Natural History (where Anderson, S. 1997. Mammals of Bolivia, taxono-
SAJ was supported by a Smithsonian Insti- my and distribution. Bulletin of the American
tution Scholarly Studies grant) and at the Museum of Natural History 231: 1652.
University of Minnesota (with partial sup- Aplin, K.P. 1990. Basicranial regions of diproto-
port from NSF grant DEB-0211952). Chris- dontian marsupials: anatomy, ontogeny, and
tina Schmidt and James Wisker (both sup- phylogeny. Unpublished Ph.D. dissertation,
ported by NSF-REU supplement DEB- University of New South Wales: Sydney.
0322253) helped sequence BRCA1 and vWF Aplin, K.P., and M. Archer. 1987. Recent
advances in marsupial systematics with a new
for this project.
syncretic classification. In M. Archer (editor),
Useful comments on earlier drafts of this Possums and opossums: studies in evolution.
manuscript were provided by Robin Beck, Al Vol. 1, xvlxxii. Sydney: Surrey Beatty.
Gardner, Francisco Goin, Marcelo Sanchez- Archer, M. 1976a. The basicranial region of
Villagra, and Jim Patton. Tom Giarla tested marsupicarnivores (Marsupialia), interrelation-
an early draft of our key (appendix 6) and ships of carnivorous marsupials, and affinities
helpfully pointed out several errors and of the insectivorous marsupial peramelids.
ambiguities. We thank these colleagues for Zoological Journal of the Linnaean Society 59:
their efforts, but hasten to add that any 217322 + 21 pls.
remaining errors are our responsibility alone. Archer, M. 1976b. The dasyurid dentition and its
relationships to that of didelphids, thylacinids,
borhyaenids (Marsupicarnivora), and perame-
REFERENCES lids (Peramelina: Marsupialia). Australian Jour-
nal of Zoology (suppl. ser.) 39: 134.
Abdala, F., D.A. Flores, and N.P. Giannini. 2001. Archer, M. 1978. The nature of the molar-
Postweaning ontogeny of the skull of Didelphis premolar boundary in marsupials and a reinter-
albiventris. Journal of Mammalogy 82: 190200. pretation of the homology of marsupial cheek-
Ade, M. 1999. External morphology and evolution teeth. Memoirs of the Queensland Museum 18:
of the rhinarium of Lagomorpha. Mitteilungen 15764.
aus dem Museum fur Naturkunde in Berlin Archer, M. 1981. Results of the Archbold Expe-
zoologische Reihe 75: 191216. ditions. No. 104. Systematic revision of the
Adkins, R.M., E.L. Gelke, D. Rowe, and R.L. marsupial dasyurid genus Sminthopsis Thomas.
Honeycutt. 2001. Molecular phylogeny and Bulletin of the American Museum of Natural
divergence time estimates for major rodent History 168: 61224.
groups: evidence from multiple genes. Molecu- Archer, M. 1982. A review of Miocene thylacinids
lar Biology and Evolution 18: 777791. (Thylacinidae, Marsupialia), the phylogenetic
Albuja, V.L., and B.D. Patterson. 1996. A new position of the Thylacinidae and the problem of
species of northern shrew-opossum (Paucituber- apriorisms in character analysis. In M. Archer
culata: Caenolestidae) from the Cordillera del (editor), Carnivorous marsupials. Vol. 2, 445
Condor, Ecuador. Journal of Mammalogy 77: 476. Mosman, NSW: Royal Society of New
4153. South Wales.
Alfaro, M.E., and M.T. Holder. 2006. The Archer, M. 1984. Origins and early radiations of
posterior and the prior in Bayesian phyloge- marsupials. In M. Archer and G. Clayton
netics. Annual Review of Ecology, Evolution, (editors), Vertebrate zoogeography and evolu-
and Systematics 37: 1942. tion in Australasia, 477515. Carlisle, WA:
Allen, J.A. 1901. A preliminary study of the North Hesperian Press.
American opossums of the genus Didelphis. Archer, M., and J. Kirsch. 2006. The evolution
Bulletin of the American Museum of Natural and classification of marsupials. In P.J. Armati,
History 14: 149195. C.R. Dickman, and I.D. Hume (editors),
Allen, J.A. 1902. A preliminary study of the South Marsupials, 121. Cambridge: Cambridge Uni-
American opossums of the genus Didelphis. versity Press.
Bulletin of the American Museum of Natural Asher, R.J., I. Horovitz, and M.R. Sanchez-
History 16: 249279. Villagra. 2004. First combined cladistic analysis
of marsupial mammal interrelationships. Molec- bernes Selskaber Biologiske Meddelelser 1(8):

ular Phylogenetics and Evolution 33: 240250. 123 + 2 pls.
Astua, D. 2007 (2006). Range extension and Bown, T.M., and J.G. Fleagle. 1993. Systematics,
first Brazilian record of the rare Hyladelphys biostratigraphy, and dental evolution of the
kalinowskii (Hershkovitz, 1992) (Didelphimor- Palaeothentidae, later Oligocene to early-middle
phia, Didelphidae). Mammalia 2006: 174176. Miocene (DeseadanSantacrucian) caenolestoid
Astua de Moraes, D., E. Hingst-Zaher, L.F. marsupials of South America. Paleontological
Marcus, and R. Cerqueira. 2000. A geometric- Society Memoir 29: iv + 176.
morphometric analysis of cranial and mandib- Braun, J.K., R.A. Van Den Bussche, P.K. Morton,
ular shape variation of didelphid marsupials. and M.A. Mares. 2005. Phylogenetic and
Hystrix 10: 115130. biogeographic relationships of mouse opossums
Augustiny, G. 1942. Die Schwimmanpassung von Thylamys (Didelphimorphia, Didelphidae) in
Chironectes. Zeitschrift fur Morphologie und southern South America. Journal of Mammal-
Okologie der Tiere 39: 276319. ogy 86: 147159.
Baker, M.L., J.P. Wares, G.A. Harrison, and R.D. Bresslau, E. 1920. The mammary apparatus of the
Miller. 2004. Relationships among the families Mammalia in the light of ontogenesis and
and orders of marsupials and the major phylogenesis. London: Methuen.
mammalian lineages based on recombination Brinkmann, A. 1911. Om Hudens Bygning paa
activating gene-1. Journal of Mammalian Evo- Haand og Fod hos Chironectes variegatus.
lution 11: 116. Videnskabelige Meddelelser den naturhistoriske
Barker, F.K., G.F. Barrowclough, and J.G. Groth. Forening i Kbenhavn 1910: 117, 1 pl.
2002. A phylogenetic hypothesis for passerine Brisson, M.J. 1762. Regnum Animale in classes IX
birds: taxonomic and biogeographic implica- distributum, sive synopsis methodica (2nd
tions of an analysis of nuclear DNA sequence ed.). Lugduni Batavorum [Leiden]: Theodorum
data. Proceedings of the Royal Society of Haak.
London B 269: 295308. Brown, B.E. 2004. Atlas of New World marsupi-
Barkley, L.J. 2008. Genus Glironia O. Thomas, als. Fieldiana Zoology (New Series) 102: ivii,
1912. In A.L. Gardner (editor), Mammals of 1308.
South America. Vol. 1. Marsupials, xenar- Brown, J.C. 1971. The description of mammals. 1.
thrans, shrews, and bats, 1214. Chicago: The external characters of the head. Mammal
University of Chicago Press. Review 1: 151168.
Barnes, R.D. 1977. The special anatomy of Brown, J.C., and D.W. Yalden. 1973. The
Marmosa robinsoni. In D. Hunsaker (editor), description of mammals. 2. Limbs and locomo-
The biology of marsupials, 387413. New York: tion of terrestrial mammals. Mammal Review 3:
Academic Press. 107134.
Beck, R.M.D. 2008. A dated phylogeny of Bublitz, J. 1987. Untersuchungen zur Systematik
marsupials using a molecular supermatrix and der rezenten Caenolestidae Trouessart, 1898:
multiple fossil constraints. Journal of Mammal- unter Verwendung craniometerischer Metho-
ogy 89: 175189. den. Bonner zoologische Monographien 23:
Bensley, B.A. 1903. On the evolution of the 196.
Australian Marsupialia; with remarks on the Bucher, J.E., and R.S. Hoffmann. 1980. Calur-
relationships of marsupials in general. Transac- omys derbianus. Mammalian Species 140: 14.
tions of the Linnaean Society of London (2nd. Butler, W.T., and H. Ritchie. 1995. The nature and
Series, Zoology) 9: 83217 + pls. 57. functional significance of dentin extracellular
Bensley, B.A. 1906. The homologies of the stylar matrix proteins. International Journal of Devel-
cusps in the upper molars of the Didelphyidae. opmental Biology 39: 169179.
University of Toronto Studies (Biological Se- Cabrera, A. 1919. Genera mammalium: Mono-
ries) 5: 149159. tremata, Marsupialia. Madrid: Museo Nacional
Biggers, J.D., H.I. Fritz, W.C.D. Hare, and R.A. de Ciencias Naturales.
McFeeley. 1965. Chromosomes of American Cabrera, A. 1958 (1957). Catalogo de los
marsupials. Science 148: 16021603. mamferos de America del Sur [part 1]. Revista
Birney, E.C., J.A. Monjeau, C.J. Phillips, R.S. del Museo Argentino de Ciencias Naturales
Sikes, and I. Kim. 1996. Lestodelphys halli: new Bernardino Rivadavia (Ciencias Zoologicas)
information on a poorly known Argentine 4(1): iiv, 1307.
marsupial. Mastozoologa Neotropical 3: 171 Carmignotto, A.P., and T. Monfort. 2006. Tax-
181. onomy and distribution of the Brazilian species
Boas, J.E.V. 1918. Zur Kenntnis des Hinterfusses of Thylamys (Didelphimorphia: Didelphidae).
der Marsupialier. Kongelige Danske Videnska- Mammalia 2006: 126144.
Carvalho, B.A., L.F.B. Oliveira, A.P. Nunes, and species of gracile mouse opossums, genus
M.S. Mattevi. 2002. Karyotypes of nineteen Gracilinanus (Marsupialia, Didelphidae). Pro-
marsupial species from Brazil. Journal of ceedings of the Biological Society of Washing-
Mammalogy 83: 5870. ton 116: 275292.
Casey, G. 1997. The BRCA1 and BRCA2 breast Creighton, G.K. 1984. Systematic studies on
cancer genes. Current Opinion in Oncology 9: opossums (Didelphidae) and rodents (Criceti-
8893. dae). Ph.D. dissertation, University of Michi-
Catzeflis, F., C. Richard-Hansen, C. Fournier- gan. Ann Arbor, MI: University Microfilms.
Chambrillon, A. Lavergne, and J.-C. Vie. 1997. Creighton, G.K., and A.L. Gardner. 2008a
Biometrie, reproduction, et sympatrie chez (2007). Genus Gracilinanus Gardner and
Didelphis marsupialis et D. albiventris en Guyane Creighton, 1989. In A.L. Gardner (editor),
francaise (Didelphidae: Marsupialia). Mamma- Mammals of South America. Vol. 1. Marsupi-
lia 61: 231243. als, xenarthrans, shrews, and bats, 4350.
Cerqueira, R. 1985. The distribution of Didelphis Chicago: Chicago University Press.
in South America (Polyprotodontia, Didelphi- Creighton, G.K., and A.L. Gardner. 2008b
dae). Journal of Biogeography 12: 135145. (2007). Genus Marmosa Gray, 1821. In A.L.
Cerqueira, R., and B. Lemos. 2000. Morphometric Gardner (editor), Mammals of South America.
differentiation between Neotropical black-eared Vol. 1. Marsupials, xenarthrans, shrews, and
opossums, Didelphis marsupialis and D. aurita bats, 5161. Chicago: Chicago University Press.
(Didelphimorphia, Didelphidae). Mammalia 64: Creighton, G.K., and A.L. Gardner. 2008c (2007).
319327. Genus Thylamys Gray, 1843. In A.L. Gardner
Cerqueira, R., and C.J. Tribe. 2008 (2007). (editor), Mammals of South America. Vol. 1.
Genus Didelphis Linnaeus, 1758. In A.L. Gard- Marsupials, xenarthrans, shrews, and bats,
ner (editor), Mammals of South America. 107117. Chicago: Chicago University Press.
Vol. 1. Marsupials, xenarthrans, shrews, and Crochet, J.-Y. 1980. Les marsupiaux du Tertiaire
bats, 1725. Chicago: Chicago University Press. dEurope. Paris: Editions de la Fondation
Cifelli, R.L. 1993. Theria of metatherian-eutherian Singer-Polignac.
grade and the origin of marsupials. In F.S. Szalay, da Silva, M.N.F., and A. Langguth. 1989. A new
M.J. Novacek, and M.C. McKenna (editors), record of Glironia venusta from the lower
Mammal phylogeny: Mesozoic differentiation, Amazon, Brazil. Journal of Mammalogy 70:
multituberculates, monotremes, early therians, 873875.
and marsupials, 205215. New York: Springer. Dawson, T.J., E. Finch, and L. Freedman., et al.
Clark, C.T., and K.K. Smith. 1993. Cranial 1989. Morphology and physiology of the
osteogenesis in Monodelphis domestica (Didel- Metatheria. In D.W. Walton and B.J. Richard-
phidae) and Macropus eugenii (Macropodidae). son (editors), Fauna of Australia. Vol. 1B:
Journal of Morphology 215: 119149. Mammalia, 451504. Canberra: Australian
Clemens, W.A., Jr. 1966. Fossil mammals of the Government Publishing Service.
type Lance Formation, Wyoming. Part 2. DEla, G., and J.A. Martnez. 2006. Registros
Marsupialia. University of California Publica- uruguayos de Gracilinanus Gardner y Creighton,
tions in Geological Sciences 62: ivi, 1122. 1989, y Cryptonanus Voss, Lunde, y Jansa, 2005
Clemens, W.A., Jr. 1979. Marsupialia. In J.A. (Didelphimorphia, Didelphidae). Mastozoolo-
Lillegraven, Z. Kielan-Jaworowska, and W.A. ga Neotropical 13: 245249.
Clemens (editors), Mesozoic mammals, the first Delsuc, F., M. Scally, and O. Madsen., et al. 2002.
two-thirds of mammalian history, 192220. Molecular phylogeny of living xenarthrans and
Berkeley: University of California Press. the impact of character and taxon sampling on
Clemens, W.A., Jr, B.J. Richardson, and P.R. the placental tree rooting. Molecular Biology
Baverstock. 1989. Biogeography and phylogeny and Evolution 19: 16561671.
of the Metatheria. In D.W. Walton and B.J. Daz, M., and M. Willig. 2004. Nuevos registros de
Richardson (editors), Fauna of Australia. Glironia venusta y Didelphis albiventris (Didel-
Vol. 1B. Mammalia, 527548. Canberra: Aus- phimorphia) para Peru. Mastozoologa Neo-
tralian Government Publishing Service. tropical 11: 185192.
Cooper, N.K., P.C. Withers, and T. Stewart. 2005. Dimitrov, S.D., E. Matouskova, and J. Forjet.
Description of sternal glands in Western Aus- 2001. Expression of BRCA1, NBR1, and NBR2
tralian Sminthopsis and Ningaui (Marsupialia: genes in human breast cancer cells. Folia
Dasyuridae). Journal of the Royal Society of Biologica (Praha) 47: 120127.
Western Australia 88: 4549. Dobigny, G., J.-F. Ducroz, T.J. Robinson, and V.
Costa, L.P., Y.L.R. Leite, and J.L. Patton. 2003. Volobouev. 2004. Cytogenetics and cladistics.
Phylogeography and systematic notes on two Systematic Biology 53: 470484.
Douady, C.J., P.I. Chatelier, and O. Madsen., et the western Brazilian Amazon. Journal of
al. 2002. Molecular phylogenetic evidence con- Mammalogy.
firming the Eulipotyphla concept and in support Felsenstein, J. 1985. Confidence limits on phylog-
of hedgehogs as the sister group to shrews. enies: an approach using the bootstrap. Evolu-
Molecular Phylogenetics and Evolution 25: tion 39: 783791.
200209. Feng, J.Q., H. Huang, and Y. Lu., et al. 2003. The
Eden, R. 1555. The Decades of the newe worlde or dentin matrix Protein 1 (DMP1) is specifically
west India, conteynyng the navigations and expressed in minerallized, but not soft, tissues
conquests of the Spanyardes, with the particular during development. Journal of Dental Re-
description of the moste ryche and large landes search 82: 776780.
and Ilandes lately founde in the west Ocean Filan, S.L. 1990. Myology of the head and neck of
perteynyng to the inheritaunce of the kinges of the bandicoot (Marsupialia: Peramelemorphia).
Spayne [etc.]. Londini: In aedibus Guilhelmi Australian Journal of Zoology 38: 617634.
Powell. [Reprinted and edited by Edward Arber Flannery, T., M. Archer, and G. Maynes. 1987.
(1885) in The first three English books on The phylogenetic relationships of living phalan-
America (Edinburgh: Turnbull and Spears).] gerids (Phalangeroidea: Marsupialia) with a
Efron, B., E. Halloran, and S. Holmes. 1996. suggested new taxonomy. In M. Archer (editor),
Bootstrap confidence levels for phylogenetic Possums and opossums: studies in evolution.
trees. Proceedings of the National Academy of Vol. 2, 477506. Sydney: Surrey Beatty.
Sciences of the United States of America 93: Flores, D.A. 2009. Phylogenetic analyses of
70857090. postcranial skeletal morphology in didelphid
Eisenberg, J.F. 1989. Mammals of the Neotropics. marsupials. Bulletin of the American Museum
Vol. 1. Panama, Colombia, Venezuela, Guyana, of Natural History 320: 181.
Suriname, French Guiana. Chicago: University Flores, D.A., and F. Abdala. 2001. Diferencias
of Chicago Press. morfologicas de craneo y denticion en Didelphis
Eisenberg, J.F., and K.H. Redford. 1999. Mammals albiventris y D. marsupialis (Didelphimorphia:
of the Neotropics. Vol. 3. Ecuador, Peru, Bolivia, Didelphidae) de Argentina y Bolivia. Comuni-
Brazil. Chicago: Chicago University Press. cacoes do Museu de Ciencias e Tecnologia da
Emmons, L.H. 1997. Neotropical rainforest mam- PUCRS (Serie Zoologia) 14: 101110.
mals: a field guide. 2nd ed. Chicago: Chicago Flores, D.A., R.M. Barquez, and M.M. Daz.
University Press. 2008. A new species of Philander Brisson, 1762
Emmons, L.H. 1998. Mammal fauna of Parque (Didelphimorphia, Didelphidae). Mammalian
Nacional Noel Kempff Mercado. In T.J. Killeen Biology 73: 1424.
and T.S. Schulenberg (editors), A biological Flores, D.A., M.M. Daz, and R.M. Barquez.
assessment of Parque Nacional Noel Kempff 2007. Systematics and distribution of marsupials
Mercado, Bolivia. (RAP Working Papers 10), in Argentina: a review. University of California
129135. Washington, DC: Conservation Inter- Publications in Zoology 134: 579669.
national. Flores, D.A., N.P. Giannini, and F. Abdala. 2003.
Emmons, L.H. 2008. Genus Caluromysiops San- Cranial ontogeny of Lutreolina crassicaudata
born, 1951. In A.L. Gardner (editor), Mammals (Didelphidae): a comparison with Didelphis
of South America. Vol. 1. Marsupials, xenar- albiventris. Acta Theriologica 48: 19.
thrans, shrews, and bats, 1112. Chicago: Flores, D.A., N.P. Giannini, and F. Abdala. 2006.
Chicago University Press. Comparative postnatal ontogeny of the skull in
Enders, R.K. 1937. Panniculus carnosus and the Australidelphian metatherian Dasyurus al-
formation of the pouch in didelphids. Journal bopunctatus (Marsupialia: Dasyuromorphia:
of Morphology 61: 126. Dasyuridae). Journal of Morphology 267:
Engstrom, M.D., and A.L. Gardner. 1988. Kar- 426440.
yotype of Marmosa canescens (Marsupialia: Flower, W.H. 1867. On the development and
Didelphidae): a mouse opossum with 22 chro- succession of the teeth in the Marsupialia.
mosomes. Southwestern Naturalist 33: 231233. Philosophical Transactions of the Royal Society
Erixon, P., B. Svennblad, T. Britton, and B. of London 157: 631641 + pls. 29, 30.
Oxelman. 2003. Reliability of Bayesian posteri- Fong, S.-L., W.-B. Fong, T.A. Morris, K.M.
or probabilities and bootstrap frequencies in Kedzie, and C.D.B. Bridges. 1990. Character-
phylogenetics. Systematic Biology 52: 665673. ization and comparative structural features of
Fantin, C., and M.N.F. da Silva. In press. the gene for human interstitial retinol-binding
Karyotype of the bushy-tailed opossum Glironia protein. Journal of Biological Chemistry 265:
venusta (Didelphidae, Didelphimorphia) from 36483653.
Freedman, L. 1967. Skull and tooth variation in Levantamento de mamferos do ParnaTapajos.

the genus Perameles. Part 1. Anatomical fea- Brasil Florestal 63: 3341.
tures. Records of the Australian Museum 27: Giannini, N.P., F. Abdala, and D.A. Flores. 2004.
14765, pls. 1623. Comparative postnatal ontogeny of the skull in
Gabbert, S.L. 1998. Basicranial anatomy of Dromiciops gliroides (Marsupialia: Micro-
Herpetotherium (Marsupialia: Didelphimor- biotheria). American Museum Novitates 3460:
phia) from the Eocene of Wyoming. American 117.
Museum Novitates 3235: 113. Ginsburg, D., and E.J.W. Bowie. 1992. Molecular
Gardner, A.L. 1973. The systematics of the genus genetics of von Willebrand disease. Blood 79:
Didelphis (Marsupialia: Didelphidae) in North 25072519.
and Middle America. Special Publications of the Goin, F.J. 1997. New clues for understanding
Museum Texas Tech University 4: 181. Neogene marsupial radiations. In R.F. Kay,
Gardner, A.L. 1981. Mammals of Surinam [re- R.H. Maddeen, R.L. Cifelli, and J.J. Flynn
view]. Journal of Mammalogy 62: 445448. (editors), Vertebrate paleontology in the Neo-
Gardner, A.L. 2005. Order Didelphimorphia. In tropics: the Miocene fauna of La Venta,
D.E. Wilson and D.M. Reeder (editors), Mam- Colombia, 187206. Washington, DC: Smithso-
mal species of the world: a taxonomic and nian Institution Press.
geographic reference. 3rd ed.: 318. Baltimore: Goin, F.J., and A.M. Candela. 2004. New
Johns Hopkins University Press. Paleogene marsupials from the Amazon basin
Gardner, A.L. (editor). 2008 (2007). Mammals of eastern Peru. Natural History Museum of
of South America. Vol. 1. Marsupials, xenar- Los Angeles County Science Series 40: 1560.
thrans, shrews, and bats. Chicago: Chicago Goin, F.J., and U.F.J. Pardinas. 1996. Revision de
University Press. las especies del genero Hyperdidelphys Ame-
Gardner, A.L., and G.K. Creighton. 1989. A new ghino, 1904 (Mammalia, Marsupialia, Didelphi-
generic name for Tates microtarsus group of dae). Su significacion, filogenetica, estratogra-
South American mouse opossums (Marsupialia: fica, y adaptativa en el Neogeno del Cono Sur
Didelphidae). Proceedings of the Biological sudamericano. Estudios Geolologicos 52: 327
Society of Washington 102: 37. 359.
Gardner, A.L., and G.K. Creighton. 2008a Goin, F.J., and P. Rey. 1997. Sobre las afinidades
(2007). Genus Marmosops Matschie, 1916. de Monodelphis Burnett, 1830 (Mammalia:
In A.L. Gardner (editor), Mammals of South Marsupialia: Didelphidae: Marmosinae). Neo-
America. Vol. 1. Marsupials, xenarthrans, shrews, tropica 43: 9398.
and bats, 6174. Chicago: Chicago University Goin, F.J., M.R. Sanchez-Villagra, A. Abello, and
Press. R.F. Kay. 2007. A new generalized paucituber-
Gardner, A.L., and G.K. Creighton. 2008b (2007). culatan marsupial from the Oligocene of Bolivia
Genus Micoureus Lesson, 1842. In A.L. Gardner and the origin of shrew-like opossums. Pale-
(editor), Mammals of South America. Vol. 1. ontology 50: 12671276.
Marsupials, xenarthrans, shrews, and bats, 7482. Gonzalez, E.M., and G. Fregueiro. 1998. Primer
Chicago: Chicago University Press. registro de Chironectes minimus para Uruguay
Gardner, A.L., and M. Dagosto. 2008 (2007). (Mammalia, Didelphidae). Comunicaciones Zo-
Tribe Metachirini Reig, Kirsch, and Marshall, ologicas del Museo de Historia Natural de
1985. In A.L. Gardner (editor), Mammals of Montevideo 12(192): 18.
South America. Vol. 1. Marsupials, xenar- Gonzalez, E.M., A.M. Saralegui, and G. Fre-
thrans, shrews, and bats, 3539. Chicago: gueiro. 2000. The genus Thylamys Gray, 1843,
Chicago University Press. in Uruguay (Didelphimorphia, Didelphidae).
Gaudin, T.J., J.R. Wible, J.A. Hopson, and W.D. Boletn de la Sociedad Zoologica del Uruguay
Turnbull. 1996. Reexamination of the evidence 12: 4445.
for the Cohort Epitheria (Mammalia, Eutheria). Graipel, M., P.R.M. Miller, and A. Ximenez.
Journal of Mammalian Evolution 3: 3179. 1996. Contribucao a identificao e distribuicao
Gellert, M. 2002. V(D)J recombination: RAG das subespecies de Lutreolina crassicaudata
proteins, repair factors, and regulation. Annual (Desmarest) (Marsupialia, Mammalia). Revista
Review of Biochemistry 71: 101132. Brasileira de Biologia 13: 781790.
George, A., B. Sabsay, P.A.L. Simonian, and A. Groth, J.G., and G.F. Barrowclough. 1999. Basal
Veis. 1993. Characterization of a novel dentin divergences in birds and the phylogenetic utility
matrix acidic phosphoprotein. Journal of Bio- of the nuclear RAG-1 gene. Molecular Phylo-
logical Chemistry 268: 1262412630. genetics and Evolution 12: 115123.
George, T.K., S.A. Marques, M. de Vivo, L.C. Gruber, K.F., R.S. Voss, and S.A. Jansa. 2007.
Branch, N. Gomes, and R. Rodrigues. 1988. Base-compositional heterogeneity in the RAG1
locus among didelphid marsupials: implications Hershkovitz, P. 1982. The staggered marsupial
for phylogenetic inference and the evolution of lower third incisor (I3). Geobios (Lyon) Mem-
GC content. Systematic Biology 56: 114. oire Special 6: 191200.
Haines, R.W. 1958. Arboreal or terrestrial ances- Hershkovitz, P. 1992a. Ankle bones: the Chilean
try of placental mammals. Quarterly Review of opossum Dromiciops gliroides Thomas, and
Biology 33: 123. marsupial phylogeny. Bonner Zoologische Beit-
Hall, E.R. 1981. The mammals of North America. rage 43: 181213.
2nd ed. New York: Wiley, 2 vols. Hershkovitz, P. 1992b. The South American
Hall, L.S. 1987. Syndactyly in marsupialsprob- gracile mouse opossums, genus Gracilinanus
lems and prophecies. In M. Archer (editor), Gardner and Creighton, 1989 (Marmosidae,
Possums and opossums: studies in evolution. Marsupialia): a taxonomic review with notes
Vol. 1, 245255. Chipping Norton, NSW: Sur- on general morphology and relationships.
rey Beatty. Fieldiana Zoology (New Series) 39: ivi, 156.
Hamilton, W.J., Jr. 1958. Life history and Hershkovitz, P. 1995. The staggered marsupial
economic relations of the opossum (Didelphis third lower incisor: hallmark of cohort Didel-
marsupialis virginiana) in New York State. phimorphia, and description of a new genus and
Cornell University Agricultural Experiment species with staggered i3 from the Albian
Station Memoir 354: 148. (Lower Cretaceous) of Texas. Bonner zoolo-
Hamrick, M.W. 2001. Morphological diversity in gische Beitrage 45: 153169.
digital skin microstructure of didelphid marsu- Hershkovitz, P. 1997. Composition of the family
pials. Journal of Anatomy 198: 683688. Didelphidae Gray, 1821 (Didelphoidea: Marsu-
Handley, C.O., Jr. 1966. Checklist of the mammals pialia), with a review of the morphology and
of Panama. In: R.L. Wenzel and V.J. Tipton behavior of the included four-eyed pouched
(editors), Ectoparasites of Panama, 753795. opossums of the genus Philander Tiedemann,
Chicago: Field Museum of Natural History. 1808. Fieldiana Zoology (New Series) 86: 1103.
Handley, C.O., Jr. 1976. Mammals of the Smith- Hershkovitz, P. 1999. Dromiciops gliroides Thom-
sonian Venezuelan Project. Brigham Young as, 1894, last of the Microbiotheria (Marsupia-
University Science Bulletin Biological Series lia), with a review of the family Microbiother-
20(5): 189, fig. 1. iidae. Fieldiana Zoology (New Series) 93: 160.
Handley, C.O., Jr., and L.K. Gordon. 1979. New Hirst, K.L., D. Simmons, and J. Feng., et al. 1997.
species of mammals from northern South Elucidation of the sequence and the genomic
America: mouse opossums, genus Marmosa. In organization of the human dentin matrix acidic
J.F. Eisenberg (editor), Vertebrate ecology in phosphoprotein 1 (DMP1) gene: exclusion of
the northern Neotropics, 6572. Washington, the locus from a causative role in the pathogen-
DC: Smithsonian Institution Press. esis of dentinogenesis imperfecta type II. Geno-
Hayman, D.L. 1977. Chromosome number mics 42: 3845.
constancy and variation. In B. Stonehouse and Hoffstetter, R., and C. Villarroel. 1974. Decou-
D. Gilmore (editors), The biology of marsupi- verte dun marsupial microtragulide (5 argyr-
als, 2748. London: Macmillan. olagide) dans le Pliocene de lAltiplano bolivien.
Hayman, D.L. 1990. Marsupial cytogenetics. Comptes Rendus de lAcademie des Sciences
Australian Journal of Zoology 27: 339349. 278: 19471950.
Henson, O.W. 1974. Comparative anatomy of the Hooker, J.J., M.R. Sanchez-Villagra, F.J. Goin,
middle ear. In W.D. Keidel and W.D. Neff E.L. Simons, Y. Attia, and E.R. Seiffert. 2008.
(editors), Auditory system, anatomy, physiology The origins of Afro-Arabian didelphimorph
(ear): handbook of sensory physiology. Vol. 5/1, marsupials. Paleontology 51: 635648.
39110. Berlin: Springer. Horovitz, I. 1999. A phylogenetic study of living
Hershkovitz, P. 1949. Generic names of the four- and fossil platyrrhines. American Museum
eyed pouch opossum and the woolly opossum Novitates 3269: 140.
(Didelphidae). Proceedings of the Biological Horovitz, I., and M.R. Sanchez-Villagra. 2003. A
Society of Washington 62: 1112. morphological analysis of marsupial mammal
Hershkovitz, P. 1951. Mammals from British higher-level phylogenetic relationships. Cladis-
Honduras, Mexico, Jamaica, and Haiti. Fieldi- tics 19: 181212.
ana Zoology 31: 547569. Huchon, D., F.M. Catzeflis, and E.J. Douzery.
Hershkovitz, P. 1976. Comments on generic names 1999. Molecular evolution of the nuclear von
of four-eyed opossums (family Didelphidae). Willebrand factor gene in mammals and the
Proceedings of the Biological Society of Wash- phylogeny of rodents. Molecular Biology and
ington 89: 295304. Evolution 16: 577589.
Huelsenbeck, J.P., B. Larget, R.E. Miller, and F. Kirk, E.C., P. Lemelin, M.W. Hamrick, D.M.
Ronquist. 2002. Potential applications and Boyer, and J.I. Bloch. 2008. Intrinsic hand
pitfalls of Bayesian inference of phylogeny. proportions of euarchontans and other mam-
Systematic Biology 51: 673688. mals: implications for the locomotor behavior
Hunter, J.P., and J. Jernvall. 1995. The hypocone of plesiadapiforms. Journal of Human Evolu-
as a key innovation in mammalian evolution. tion 55: 122.
Proceedings of the National Academy of Kirsch, J.A.W. 1977a. The classification of mar-
Sciences of the United States of America 92: supials. In D. Hunsaker, II (editor), The biology
1071810722. of marsupials, 150. New York: Academic Press.
ICZN. 1998. Opinion 1894. Regnum animals. ed. Kirsch, J.A.W. 1977b. The comparative serology
2 (M.J. Brisson, 1762): rejected for nomencla- of Marsupialia, and a classification of marsupi-
tural purposes, with the conservation of the als. Australian Journal of Zoology (suppl. ser.)
mammalian generic names Philander (Marsupi- 52: 1152.
alia), Pteropus (Chiroptera), Glis, Cuniculus and Kirsch, J.A.W., and M. Archer. 1982. Polythetic
Hydrochoerus (Rodentia), Meles, Lutra and cladistics, or, when parsimonys not enough: the
Hyaena (Carnivora), Tapirus (Perissodactyla), relationships of carnivorous marsupials. In M.
Tragulus and Giraffa (Artiodactyla). Bulletin of Archer (editor), Carnivorous marsupials. Vol. 2,
Zoological Nomenclature 55: 6471. 595620. Mosman, NSW: Royal Society of New
ICZN. 1999. International code of zoological South Wales.
nomenclature. 4th ed. London: International Kirsch, J.A.W., F.-J. Lapointe, and M.S. Springer.
Trust for Zoological Nomenclature. 1997. DNA-hybridization studies of marsupials
Izor, R.J., and R.H. Pine. 1987. Notes on the and their implications for metatherian classifi-
black-shouldered opossum, Caluromysiops ir- cation. Australian Journal of Zoology 45:
rupta. Fieldiana Zoology (New Series) 39: 211280.
117124. Kirsch, J.A.W., and R.E. Palma. 1995. DNA/
Jansa, S.A., J.F. Forsman, and R.S. Voss. 2006. DNA hybridization studies of carnivorous
Different patterns of selection on the nuclear marsupials. V. A further estimate of relation-
genes IRBP and DMP-1 affect the efficiency but ships among opossums (Marsupialia: Didelphi-
not the outcome of phylogeny estimation for dae). Mammalia 59: 403425.
didelphid marsupials. Molecular Phylogenetics Krieg, H. 1924. Beobachtungen an argentinischen
and Evolution 38: 363380. Beutelratten. Zeitschrift fur Morphologie und
Jansa, S.A., and R.S. Voss. 2000. Phylogenetic Okologie der Tiere 1: 637659.
studies on didelphid marsupials I. Introduction Krumbiegel, I. 1941. Die Saugetiere der Sudamer-
and preliminary results from nuclear IRBP gene ika-Expeditionen Prof. Dr. Kriegs. Zoologischer
sequences. Journal of Mammalian Evolution 7: Anzeiger 134: 2953.
4377. Kurz, C. 2007. The opossum-like marsupials
Jansa, S.A., and R.S. Voss. 2005. Phylogenetic (Didelphimorphia and Peradectia, Marsupialia,
relationships of the marsupial genus Hyladel- Mammalia) from the Eocene of Messel and
phys based on nuclear gene sequences and Geiseltalecomorphology, diversity, and pa-
morphology. Journal of Mammalogy 86: 853 laeogeography. Kaupia 15: 365.
865. Ladeveze, S. 2007. Petrosal bones of metatherian
Jansa, S.A., and M. Weksler. 2004. Phylogeny of mammals from the Late Palaeocene of Itabora
muroid rodents: relationships within and among (Brazil), and a cladistic analysis of petrosal
major lineages as determined by IRBP gene features in metatherians. Zoological Journal of
sequences. Molecular Phylogenetics and Evolu- the Linnean Society 150: 85115.
tion 31: 256276. Lemelin, P. 1999. Morphological correlates of
Johanson, Z. 1996. Revision of the Late Creta- substrate use in didelphid marsupials: implica-
ceous North American marsupial genus Alpha- tions for primate origins. Journal of Zoology
don. Palaeontographica 242: 127184 + 12 pls. (London) 247: 165175.
Jones, F.W. 1949. The study of a generalized Lemos, B., and R. Cerqueira. 2002. Morphological
marsupial (Dasycercus cristicauda Krefft). differentiation in the white-eared opossum
Transactions of the Zoological Society of group (Didelphidae: Didelphis). Journal of
London 26: 409501. Mammalogy 83: 354369.
Kent, W.J. 2002. BLATthe BLAST-like align- Lemos, B., G. Marroig, and R. Cerqueira. 2001.
ment tool. Genome Research 4: 656664. Evolutionary rates and stabilizing selection in
Kielan-Jaworowska, Z., R.L. Cifelli, and Z.-X. large-bodied opossum skulls (Didelphimorphia:
Luo. 2004. Mammals from the age of dinosaurs. Didelphidae). Journal of Zoology (London)
New York: Columbia University Press. 255: 181189.
Lew, D., R. Perez-Hernandez, and J. Ventura. Tasmanian monotremes and marsupials. Papers
2006. Two new species of Philander (Didelphi- and Proceedings of the Royal Society of
morphia, Didelphidae) from northern South Tasmania 1950: 7184.
America. Journal of Mammalogy 87: 224237. Lyne, A.G., and P.A. Mort. 1981. A comparison
Lewis, P.O. 2001. A likelihood approach to of skull morphology in the marsupial bandicoot
estimating phylogeny from discrete morpholog- genus Isoodon: its taxonomic implications and
ical character data. Systematic Biology 50: 913 notes on a new species, Isoodon arnhemensis.
925. Australian Mammalogy 4: 107133.
Linares, O.J. 1998. Mamferos de Venezuela. MacPhee, R.D.E. 1981. Auditory region of pri-
Caracas: Sociedad Conservacionista Audubon mates and eutherian insectivores: morphology,
de Venezuela. ontogeny, and character analysis. Contributions
Linnaeus, C. 1758. Systema naturae per regna tria to Primatology 18: 1282.
naturae, secundum classes, ordines, genera, Madsen, O., M. Scally, and C.J. Douady., et al.
species, cum characteribus, differentiis, synony- 2001. Parallel adaptive radiations in two major
mis, locis. 10th ed. Vol. 1. Stockholm: Laurentii clades of placental mammals. Nature 409:
Salvii. 610614.
Liou, G.I., D.-P. Ma, and Y.-W. Yang., et al. Mancuso, D.J., E.A. Tuley, and L.A. Westfield., et
1989. Human interstitial retinoid-binding pro- al. 1989. Structure of the gene for human von
tein. Journal of Biological Chemistry, 82008206. Willebrand factor. Journal of Biological Chem-
Luckett, W.P. 1993. An ontogenetic assessment of istry 264: 1951419527.
dental homologies in therian mammals. In F.S. Mares, M.A., J.K. Braun, and D. Gettinger. 1989.
Szalay, M.J. Novacek, and M.C. McKenna Observations on the distribution and ecology of
(editors), Mammal phylogeny: Mesozoic differ- the mammals of the cerrado grasslands of
entiation, multituberculates, monotremes, early central Brazil. Annals of the Carnegie Museum
therians, and marsupials, 182204. New York: 58: 160.
Springer. Marshall, L.G. 1977. Lestodelphys halli. Mamma-
Luckett, W.P., and N. Hong. 2000. Ontogenetic lian Species 81: 13.
evidence for dental homologies and premolar Marshall, L.G. 1978a. Lutreolina crassicaudata.
replacement in fossil and extant caenolestids Mammalian Species 91: 14.
(Marsupialia). Journal of Mammalian Evolu- Marshall, L.G. 1978b. Dromiciops australis. Mam-
tion 7: 109127. malian Species 99: 15.
Lunde, D.P., and V. Pacheco. 2003. Shrew Marshall, L.G. 1978c. Glironia venusta. Mamma-
opossums (Paucituberculata: Caenolestes) from lian Species 107: 13.
the Huancabamba region of east Andean Peru. Marshall, L.G. 1978d. Chironectes minimus. Mam-
Mammal Study 28: 145148. malian Species 109: 16.
Lunde, D.P., and W.A. Schutt, Jr. 1999. The Marshall, L.G. 1980. Systematics of the South
peculiar carpal tubercles of male Marmosops American marsupial family Caenolestidae.
parvidens and Marmosa robinsoni (Didelphidae: Fieldiana Geology (New Series) 5: iviii, 1145.
Didelphinae). Mammalia 63: 495504. Marshall, L.G. 1981. The families and genera of
Luo, Z.-X., Q. Ji, J.R. Wible, and C.-X. Yuan. Marsupialia. Fieldiana Geology (New Series) 8:
2003. An Early Cretaceous tribosphenic mam- ivi, 165.
mal and metatherian evolution. Science 302: Marshall, L.G. 1982b. Systematics of the South
19341940. American marsupial family Microbiotheriidae.
Luo, Z.-X., Z. Kielan-Jaworowska, and R.L. Fieldiana Geology (New Series) 10: iviii, 175.
Cifelli. 2004. Evolution of dental replacement Marshall, L.G. 1987. Systematics of Itaboraian
in mammals. Bulletin of the Carnegie Museum (Middle Paleocene) age opossum-like marsu-
of Natural History 36: 159175. pials from the limestone quarry at Sao Jose de
Lyne, A.G. 1951. Notes on the external characters Itabora, Brazil. In M. Archer (editor), Possums
of the barred bandicoot (Perameles gunnii and opossums: studies in evolution. Vol. 1,
Gray), with special reference to the pouch 91160. Chipping Norton, NSW: Surrey Beatty.
young. Proceedings of the Zoological Society Marshall, L.G., J.A. Case, and M.O. Woodburne.
of London 121: 587598. 1990. Phylogenetic relationships of the families
Lyne, A.G. 1959. The systematic and adaptive of marsupials. Current Mammalogy 2: 433505.
significance of the vibrissae in the Marsupialia. Marshall, L.G., C. de Muizon, and D. Sigogneau-
Proceedings of the Zoological Society of Lon- Russell. 1995. Pucadelphys andinus (Marsupia-
don 133: 79132. lia, Mammalia) from the early Paleocene of
Lyne, A.G., and T.S. McMahon. 1951. Observa- Bolivia. Memoires du Museum National dHis-
tions on the surface structure of the hairs of toire Naturelle 165: 1164.
Martin, G.M. 2003. Nuevas localidades para Muirhead, J. 2000. Yaraloidea (Marsupialia,
marsupiales patagonicos (Didelphimorphia y Peramelemorphia), a new superfamily of mar-
Microbiotheria) en el noroeste de la provincia supial and a description and analysis of the
del Chubut, Argentina. Mastozoologa Neo- cranium of the Miocene Yarala burchfieldi.
tropical 10: 148153. Journal of Paleontology 74: 512523.
Martin, G.M. 2005. Intraspecific variation in Muirhead, J., and S.L. Filan. 1995. Yarala
Lestodelphys halli (Marsupialia: Didelphimor- burchfieldi, a plesiomorphic bandicoot (Marsu-
phia). Journal of Mammalogy 86: 793802. pialia, Peramelemorphia) from Oligo-Miocene
Martinelli, P.M., and J.C. Nogueira. 1997. Penis deposits of Riversleigh, northwestern Queens-
morphology as a distinctive character of the land. Journal of Paleontology 69: 127134.
murine opossum group (Marsupialia Didelphi- Muizon, C. de. 1998. Mayulestes ferox, a borhyae-
dae): a preliminary report. Mammalia 61: noid (Metatheria, Mammalia) from the early
161166. Palaeocene of Bolivia: phylogenetic and palaeo-
Matschie, P. 1916. Bemerkungen uber die Gattung biological implications. Geodiversitas 20: 19
Didelphis L. Sitzungsberichte der Gesellschaft 142.
Naturforschender Freunde zu Berlin 1916: Muizon, C. de., R.L. Cifelli, and R.C. Paz. 1997.
259272. The origin of the dog-like borhyaenoid marsu-
McKenna, M.C., and S.K. Bell. 1997. Classifica- pials of South America. Nature 389: 486489.
tion of mammals above the species level. New Muizon, C. de., and B. Lange-Badre. 1997.
York: Columbia University Press. Carnivorous dental adaptations in tribosphenic
Medelln, R.A. 1991. Ecomorfologa del craneo de mammals and phylogenetic reconstruction.
cinco didelfidos: tendencias, divergencias, e Lethaia 30: 353366.
implicaciones. Anales del Instituto de Biologa Mustrangi, M.A., and J.L. Patton. 1997. Phylo-
de la Universidad Nacional Autonoma de geography and systematics of the slender mouse
Mexico (Serie Zoologa) 62: 269286. opossum Marmosops (Marsupialia, Didelphi-
Medelln, R.A., G. Cancino Z., A. Clemente M., dae). University of California Publications in
and R.O. Guerrero V. 1992. Noteworthy Zoology 130: 186.
records of three mammals from Mexico. South- Napier, J.R. 1961. Prehensility and opposability in
western Naturalist 37: 427429. the hands of primates. Zoological Society of
Mercer, J.M., and V.L. Roth. 2003. The effects of London Symposium 5: 115132.
Cenozoic global change on squirrel phylogeny. Nesslinger, C.L. 1956. Ossification centers and
Science 299: 15681572. skeletal development in the postnatal Virginia
Meredith, R.W., M. Westerman, J.A. Case, and opossum. Journal of Mammalogy 37: 382394.
M.S. Springer. 2008. A phylogeny and timescale Neumann, J., Michaux, and V. Lebedev., et al.
for marsupial evolution based on sequences for 2006. Molecular phylogeny of the Cricetinae
five nuclear genes. Journal of Mammalian subfamily based on the mitochondrial cyto-
Evolution 15: 136. chrome b and 12S rRna genes and the nuclear
Miki, Y., J. Swensen, and D. Shattuck-Eidens., et vWF gene. Molecular Phylogenetics and Evo-
al. 1994. A strong candidate for the breast and lution 39: 135148.
ovarian cancer susceptibility gene BRCA1. Nilsson, M.A., U. Arnason, P.B.S. Spencer, and
Science 266: 6671. A. Janke. 2004. Marsupial relationships and a
Mikkelsen, T.S., M.J. Wakefield, B. Aken, and timeline for marsupial radiation in South
C.T. Amemiya., et al. 2007. Genome of the Gondwana. Gene 340: 189196.
marsupial Monodelphis domestica reveals inno- Nogueira, J.C., and A.C.S. Castro. 2003. In M.E.
vation in non-coding sequences. Nature 447: Jones, C.R. Dickman, and M. Archer (editors),
167178. Predators with pouches: the biology of carniv-
Mondolfi, E., and G. Medina P. 1957. Contribu- orous marsupials, 183204. Melbourne: CSIRO
cion al conocimiento del perrito de agua Press.
(Chironectes minimus Zimmermann). Memorias Nogueira, J.C., A.C.S. Castro, E.V.C. Camara,
de la Sociedad de Ciencias Naturales La Salle and B.G.O. Camara. 2004. Morphology of the
17: 141155. male genital system of Chironectes minimus and
Mondolfi, E., and R. Perez-Hernandez. 1984. Una comparison to other didelphid marsupials.
nueva subespecie de zarigueya del grupo Didel- Journal of Mammalogy 85: 834841.
phis albiventris (Mammalia-Marsupialia). Acta Nogueira, J.C., M.N.F. da Silva, and B.G.O.
Cientfica Venezolana 35: 407413. Camara. 1999a. Morphology of the male genital
Morton, S.R. 1980. Ecological correlates of caudal system of the bushy-tailed opossum Glironia
fat storage in small mammals. Australian venusta Thomas, 1912 (Didelphimorphia, Di-
Mammalogy 3: 8186. delphidae). Mammalia 63: 231236.
Nogueira, J.C., P.M. Martinelli, S.F. Costa, G.A. 12S mitochondrial gene: the phylogeny of
Carvalho, and B.G.O. Camara. 1999b. The Didelphimorphia and of the living fossil micro-
penis morphology of Didelphis, Lutreolina, biotheriid Dromiciops gliroides. Molecular Phy-
Metachirus, and Caluromys (Marsupialia, Di- logenetics and Evolution 13: 525535.
delphidae). Mammalia 63: 7992. Palma, R.E., and T.L. Yates. 1996. The chromo-
Novacek, M.J. 1986. The skull of leptictid somes of Bolivian didelphid marsupials. Occa-
insectivorans and the higher-level classification sional Papers of the Museum Texas Tech
of eutherian mammals. Bulletin of the American University 162: 120.
Museum of Natural History 183: 1112. Palma, R.E., and T.L. Yates. 1998. Phylogeny of
Novacek, M.J. 1993. Patterns of diversity in the southern South American mouse opossums
mammalian skull. In J. Hanken and B.K. Hall (Thylamys, Didelphidae) based on allozyme
(editors), The skull. Vol. 2, 438545. Chicago: and chromosomal data. Zeitschrift fur Sauge-
University of Chicago Press. tierkunde 63: 115.
Novacek, M.J., and A. Wyss. 1986. Origin and Parker, W.K. 1890. On the skull of Tarsipes
transformation of the mammalian stapes. In rostratus. Studies from the Museum of Zoology
K.M. Flanagan and J.A. Lillegraven (editors), in University College (Dundee) 1: 7983 + 1 pl.
Vertebrates: phylogeny, and philosophy. (Uni- Patterson, B., and R. Pascual. 1972. The fossil
versity of Wyoming Contributions to Geology mammal fauna of South America. In A. Keast,
Special Paper 3), 3553. Laramie: University of F.C. Erk, and B. Glass (editors), Evolution,
Wyoming. mammals, and southern continents, 247309.
Oettinger, M.A., D.G. Schatz, C. Gorka, and D. Albany: State University of New York Press.
Baltimore. 1990. RAG-1 and RAG-2, adjacent Patterson, B.D., and M.H. Gallardo. 1987.
genes that synergistically activate V(D)J recom- Rhyncholestes raphanurus. Mammalian Species
bination. Science 248: 15171523. 286: 15.
Oettinger, M.A., B. Stanger, and D.G. Schatz., et Patton, J.L., and L.P. Costa. 2003. Molecular
al. 1992. The recombination activating genes, phylogeography and species limits in rainforest
RAG1 and RAG2, are on chromosome 11p in didelphid marsupials of South America. In M.E.
humans and chromosome 2p in mice. Immuno- Jones, C.R. Dickman, and M. Archer (editors),
genetics 35: 97101. Predators with pouches: the biology of carniv-
Oliver, L.R. 1976. The management of yapoks orous marsupials, 6381. Melbourne: CSIRO
(Chironectes minimus) at Jersey Zoo, with Press.
observations on their behavior. Jersey Wildlife Patton, J.L., and M.N.F. da Silva. 1997. Defini-
Preservation Trust 13: 3236. tion of species of pouched four-eyed opossums
Osgood, W.H. 1921. A monographic study of the (Didelphidae, Philander). Journal of Mammal-
American marsupial Caenolestes, with a descrip- ogy 78: 90102.
tion of the brain of Caenolestes by C. Judson Patton, J.L., and M.N.F. da Silva. 2008 (2007).
Herrick. Field Museum of Natural History Genus Philander Brisson, 1762. In A.L. Gardner
Zoological Series 14: frontispiece, 1162, pls. (editor), Mammals of South America. Vol. 1.
122. Marsupials, xenarthrans, shrews, and bats,
Osgood, W.H. 1924. Review of living caenolestids 2735. Chicago: Chicago University Press.
with description of a new genus from Chile. Patton, J.L., M.N.F. da Silva, and J.R. Malcolm.
Field Museum of Natural History Publication 2000. Mammals of the Rio Jurua and the
207(Zoological Series 14): 163172, pl. 23. evolutionary and ecological diversification of
Palma, E. 1995. The karyotypes of two South Amazonia. Bulletin of the American Museum of
American mouse opossums of the genus Thyla- Natural History 244: 1306.
mys (Marsupialia: Didelphidae), from the An- Patton, J.L., S.F. dos Reis, and M.N.F. da Silva.
des, and eastern Paraguay. Proceedings of the 1996. Relationships among didelphid marsupi-
Biological Society of Washington 108: 15. als based on sequence variation in the mito-
Palma, E. 1997. Thylamys elegans. Mammalian chondrial cytochrome b gene. Journal of Mam-
Species 572: 14. malian Evolution 3: 329.
Palma, R.E., E. Rivera-Milla, T.L. Yates, P.A. Pepperberg, D.R., T.-I.L. Okajima, and B. Wig-
Marquet, and A.P. Meynard. 2002. Phylogenet- gert, et al. 1993. Interphotoreceptor retinoid-
ic and biogeographic relationships of the mouse binding protein (IRBP): molecular biology and
opossum Thylamys (Didelphimorphia, Didel- physiological role in the visual cycle of rhodop-
phidae) in southern South America. Molecular sin. Molecular Neurobiology 7: 6185.
Phylogenetics and Evolution 25: 245253. Perez-Hernandez, R., P. Soriano, and D. Lew.
Palma, R.E., and A.E. Spotorno. 1999. Molecular 1994. Marsupiales de Venezuela. Caracas: Cua-
systematics of marsupials based on the rRNA dernos Lagoven.
Pine, R.H. 1973. Anatomical and nomenclatural Reig, O.A. 1958. Comunicacion preliminar sobre
notes on opossums. Proceedings of the Biolog- nuevas especies del genero Thylatheridium Reig
ical Society of Washington 86: 391402. (Mammalia, Didelphidae). Neotropica 4: 8995.
Pine, R.H., and C.O. Handley, Jr. 2008 (2007). Reig, O.A., A.L. Gardner, N.O. Bianchi, and J.L.
Genus Monodelphis Burnett, 1830. In A.L. Patton. 1977. The chromosomes of the Didel-
Gardner (editor), Mammals of South America. phidae (Marsupialia) and their evolutionary
Vol. 1. Marsupials, xenarthrans, shrews, and significance. Biological Journal of the Linnaean
bats, 82107. Chicago: Chicago University Society of London 9: 91216.
Press. Reig, O.A., J.A.W. Kirsch, and L.G. Marshall.
Pine, R.H., J.E. Rice, J.E. Bucher, D.H. Tank, Jr., 1985. New conclusions on the relationships of
and A.M. Greenhall. 1985. Labile pigments and the opossum-like marsupials, with an annotated
fluorescent pelage in didelphid marsupials. classification of the Didelphimorphia. Ame-
Mammalia 49: 249256. ghiniana 21: 335343.
Pocock, R.I. 1914. On the facial vibrissae of Reig, O.A., J.A.W. Kirsch, and L.G. Marshall.
Mammalia. Proceedings of the Zoological 1987. Systematic relationships of the living and
Society of London 1914: 889912. Neocenozoic American opossum-like marsu-
Pocock, R.I. 1926. The external characters of pials (suborder Didelphimorphia), with com-
Thylacinus, Sarcophilus, and some related mar- ments on the classification of these and of the
supials. Proceedings of the Zoological Society of Cretaceous and Paleogene New World and
London 1926: 10371084. European metatherians. In M. Archer (editor),
Porter, C.A., M. Goodman, and M.J. Stanhope. Possums and opossums: studies in evolution.
1996. Evidence on mammalian phylogeny from Vol. 1, 189. Sydney: Surrey Beatty.
sequences of exon 28 of the von Willebrand Reig, O.A., and G.G. Simpson. 1972. Sparassocy-
factor gene. Molecular Phylogenetics and Evo- nus (Marsupialia, Didelphidae), a peculiar
lution 5: 89101. mammal from the late Cenozoic of Argentina.
Posada, D., and T. Buckley. 2004. Model selection Journal of Zoology (London) 167: 511539.
and model averaging in phylogenetics: advan- Ribeiro, M.G., and J.C. Nogueira. 1990. The penis
tages of Akaike Information Criterion and morphology of the four-eyed opossum Philan-
Bayesian approaches over likelihood ratio tests. der opossum. Anatomischer Anzeiger 171:
Systematic Biology 53: 793808. 6572.
Posada, D., and K.A. Crandall. 1998. ModelTest: Ride, W.D.L. 1962. On the evolution of Australian
testing the model of DNA substitution. Bioin- marsupials. In: G.E. Leeper (editor), The
formatics 14: 817818. evolution of living organisms, 281306. Mel-
Prochel, J., and M. Sanchez-Villagra. 2003. Carpal bourne: Melbourne University Press.
ontogeny in Monodelphis domestica and Calur- Rodrguez, J.A., and B.R. Henderson. 2000.
omys philander (Marsupialia). Zoology 106: Identification of a functional nuclear export
7384. sequence in BRCA1. Journal of Biological
Raterman, D., R.W. Meridith, L.A. Ruedas, and Chemistry 275: 3858938596.
M.S. Springer. 2006. Phylogenetic relationships Rofe, R., and D. Hayman. 1985. G-banding
of the cuscuses and brushtail possums (Marsu- evidence for a conserved complement in the
pialia: Phalangeridae) using the nuclear gene Marsupialia. Cytogenetics and Cell Genetics 39:
BRCA1. Australian Journal of Zoology 54: 4050.
353361. Ronquist, F., and J.P. Huelsenbeck. 2003.
Redford, K.H., and J.F. Eisenberg. 1992. Mam- MrBayes 3: Bayesian phylogenetic inference
mals of the Neotropics. Vol. 2. Chile, Argen- under mixed models. Bioinformatics 19: 1572
tina, Uruguay, Paraguay. Chicago: Chicago 1574.
University Press. Rossi, R.V. 2005. Revisao taxonomica de Mar-
Reeder, S.A., and R.D. Bradley. 2004. Molecular mosa Gray, 1821 (Didelphimorphia, Didelphi-
systematics of neotomine-peromyscine rodents dae). Unpublished Ph.D. dissertation, Universi-
based on the dentin matrix protein 1 gene. dade de Sao Paulo. 2 vols.
Journal of Mammalogy 85: 11941200. Rougier, G.W., J.R. Wible, and M.J. Novacek.
Reid, F.A. 1997. A field guide to the mammals of 1998. Implications of Deltatheridium specimens
Central America and southeast Mexico. New for early marsupial history. Nature 396:
York: Oxford University Press. 459463. [Data matrix and character descrip-
Reig, O.A. 1955. Noticia preliminar sobre la tions were published electronically as supple-
presencia de microbiotherinos vivientes en la mentary information at www.nature.com. The
fauna sudamericana. Investigaciones Zoologicas first paper publication of this information
Chilenas 2: 121130. appeared as appendix 1 of Wible et al. (2001).]
Rowe, T.B., T.P. Eiting, T.E. Macrini, and R.A. Santori, R.T., D. Astua de Moraes, and R.
Ketcham. 2005. Organization of the olfactory Cerqueira. 2004. Comparative gross morpholo-
and respiratory skeleton in the nose of the gray gy of the digestive tract in ten Didelphidae
short-tailed opossum Monodelphis domestica. marsupial species. Mammalia 68: 2736.
Journal of Mammalian Evolution 12: 303336. Santos-Filho, M., M.N.F. da Silva, B.A. Costa,
Ruggieri, Z.M., and T.S. Zimmerman. 1987. Von C.G. Bantel, C.L.G. Vieira, D.J. Silva, and
Willebrand factor and von Willebrand disease. A.M.R. Franco. 2007. New records of Glironia
Blood 70: 895904. venusta Thomas, 1912 (Mammalia, Didelphi-
Sadler, J.E. 1998. Biochemistry and genetics of von dae) from the Amazon and Paraguay basins,
Willebrand factor. Annual Review of Biochem- Brazil. Mastozoologa Neotropical 14: 103105.
istry 67: 395424. Sauthier, D.E.U., M. Carrera, and U.F.J. Pardi-
Sanchez, F., P. Sanchez-Palomino, and A. Cadena. nas. 2007. Mammalia, Marsupialia, Didelphi-
2004. Inventario de mamferos en un bosque de dae, Lestodelphys halli: new records, distribution
los Andes centrales de Colombia. Caldasia 26: extension and filling gaps. Check List 3:
291309. 137140.
Sanchez-Villagra, M.R. 2001a. Ontogenetic and Schatz, D.G., M.A. Oettinger, and D. Baltimore.
phylogenetic transformations of the vomerona- 1989. The V(D)J recombination activating gene,
sal complex and nasal floor elements in marsu- RAG-1. Cell 59: 10351048.
pial mammals. Zoological Journal of the Schmelzle, T., S. Nummela, and M.R. Sanchez-
Linnean Society 131: 459479. Villagra. 2005. Phylogenetic transformations of
Sanchez-Villagra, M.R. 2001b. The phylogenetic the ear ossicles in marsupial mammals, with
relationships of argyrolagid marsupials. Zoo- special reference to diprotodontians: a character
logical Journal of the Linnean Society 131: analysis. Annals of the Carnegie Museum 74:
481496. 189200.
Sanchez-Villagra, M.R., and R.J. Asher. 2002. Segall, W. 1969a. The middle ear region of
Cranio-sensory adaptations in small faunivor- Dromiciops. Acta Anatomica 72: 489501.
ous semiaquatic mammals, with special refer- Segall, W. 1969b. The auditory ossicles (malleus,
ence to olfaction and the trigeminal system. incus) and their relationships to the tympanic: in
Mammalia 66: 93109. marsupials. Acta Anatomica 73: 176191.
Sanchez-Villagra, M.R., S. Gemballa, S. Num- Segall, W. 1970. Morphological parallelisms of the
mela, K.T. Smith, and W. Maier. 2002. Onto- bulla and auditory ossicles in some insectivores
genetic and phylogenetic transformations of the and marsupials. Fieldiana Zoology 51: 169205.
ear ossicles in marsupial mammals. Journal of Sharman, G.B. 1973. The chromosomes of non-
Morphology 251: 219238. Eutherian mammals. In A.B. Chiarelli and E.
Sanchez-Villagra, M.R., R.F. Kay, and F. Anaya- Capanna (editors), Cytotaxonomy and verte-
Daza. 2000. Cranial anatomy and paleobiology brate evolution, 485530. London: Academic
of the Miocene marsupial Hondalagus altipla- Press.
nensis and a phylogeny of argyrolagids. Pa- Simpson, G.G. 1929. American Mesozoic Mam-
laeontology 43: 287301. malia. Memoirs of the Peabody Museum 3(1):
Sanchez-Villagra, M.R., S. Ladeveze, and I. ixv, 1171, 32 pls.
Horovitz., et al. 2007. Exceptionally preserved Simpson, G.G. 1935. Note on the classification of
North American Paleogene metatherians: adap- Recent and fossil opossums. Journal of Mam-
tations and discovery of a major gap in the malogy 16: 134137.
opossum fossil record. Biology Letters 3: 318 Simpson, G.G. 1936. Studies of the earliest
322. mammalian dentitions. Dental Cosmos (Au-
Sanchez-Villagra, M.R., and S. Nummela. 2001. gustSeptember), 124.
Bullate stapedes in some phalangeriform mar- Simpson, G.G. 1945. The principles of classifica-
supials. Mammalian biology 66: 174177. tion and a classification of mammals. Bulletin of
Sanchez-Villagra, M.R., and K.K. Smith. 1997. the American Museum of Natural History 85:
Diversity and evolution of the marsupial man- ixvi, 1350.
dibular angular process. Journal of Mammalian Simpson, G.G. 1972. Didelphidae from the Cha-
Evolution 4: 119144. padmalal formation in the Museo Municipal de
Sanchez-Villagra, M.R., and J.R. Wible. 2002. Ciencias Naturales of Mar del Plata. Publica-
Patterns of evolutionary transformations in the ciones del Museo Municipal de Ciencias Natur-
petrosal bone and some basicranial features in ales Mar del Plata 2: 140.
marsupial mammals, with special reference to Solari, S. 2003. Diversity and distribution of
didelphids. Journal of Zoological Systematics Thylamys (Didelphidae) in South America, with
and Evolutionary Research 40: 2645. emphasis on species from the western side of the
Andes. In: M.E. Jones, C.R. Dickman, and M. marsupials: chromosome banding and in-situ
Archer (editors), Predators with pouches: the hybridization. Chromosome Research 7: 267
biology of carnivorous marsupials, 82101. 275.
Melbourne: CSIRO Press. Swofford, D.L. 1998. PAUP*: phylogenetic anal-
Solari, S. 2007. New species of Monodelphis ysis using parsimony, beta version 4.0b10.
(Didelphimorphia: Didelphidae) from Peru, Sunderland, MA: Sinauer Associates.
with notes on M. adusta. Journal of Mammal- Swofford, D.L., and W.P. Madison. 1987. Recon-
ogy 88: 319329. structing ancestral character states under
Springer, M.S., J.A.W. Kirsch, and J.A. Case. Wagner parsimony. Mathematical Biosciences
1997a. The chronicle of marsupial evolution. In: 87: 199229.
T.J. Givnish and K.J. Sytsma (editors), Molec- Takahashi, F. 1974. Variacao morfologica de
ular evolution and adaptive radiation, 129161. incisivos em didelfdeos (Marsupialia, Didelphi-
Cambridge: Cambridge University Press. nae). Anais Academia Brasileira de Ciencias 46:
Springer, M.S., A. Burk, J.R. Kavanagh, V.G. 413416.
Waddell, and M.J. Stanhope. 1997b. The Talice, R.V., S.L. de Mosera, and T. Machado.
Interphotoreceptor Retinoid Binding Protein 1961 (1960). Observaciones sobre Monodel-
gene in therian mammals: implications for phis dimidiata. Actas y Trabajos del Primer
higher-level relationships and evidence for loss Congreso Sudamericano de Zoologa. Tomo 4:
of function in the marsupial mole. Proceedings 149156.
of the National Academy of Science of the Tarifa, T., and S. Anderson. 1997. Two additional
United States of America 94: 1375413759. records of Glironia venusta Thomas, 1912
Stamatakis, A. 2006. RAxML-VI-HPC: maximum (Marsupialia, Didelphidae) for Bolivia. Mam-
likelihood-based phylogenetic analyses with malia 61: 111113.
thousands of taxa and mixed models. Bioinfor- Tate, G.H.H. 1933. A systematic revision of the
matics 22: 26882690. marsupial genus Marmosa with a discussion of
Stanhope, M.J., J. Czelusniak, J.-S. Si, J. Nick- the adaptive radiation of the murine opossums
erson, and M. Goodman. 1992. A molecular (Marmosa). Bulletin of the American Museum
perspective on mammalian evolution from the of Natural History 66: 1250 + 26 pls.
gene encoding Interphotoreceptor Retinoid Tate, G.H.H. 1934. New generic names for two
Binding Protein, with convincing evidence for South American marsupials. Journal of Mam-
bat monophyly. Molecular Phylogenetics and malogy 15: 154 pp.
Evolution 1: 148160. Tate, G.H.H. 1947. Results of the Archbold
Stanhope, M.J., M.R. Smith, V.G. Waddell, C.A. Expeditions. No. 56. On the anatomy and
Porter, M.S. Shivji, and M. Goodman. 1996. classification of the Dasyuridae (Marsupialia).
Mammalian evolution and the interphotorecep- Bulletin of the American Museum of Natural
tor retinoid binding protein (IRBP) gene: History 88: 97156.
convincing evidence for several superordinal Tate, G.H.H. 1948a. Results of the Archbold
clades. Journal of Molecular Evolution 43: Expeditions. No. 59. Studies on the anatomy
8392. and phylogeny of the Macropodidae (Marsupi-
Stein, B.R., and J.L. Patton. 2008a (2007). alia). Bulletin of the American Museum of
Genus Chironectes Illiger, 1811. In A.L. Gard- Natural History 91: 233352.
ner (editor), Mammals of South America. Tate, G.H.H. 1948b. Results of the Archibold
Vol. 1. Marsupials, xenarthrans, shrews, and Expeditions. No. 60. Studies in the Peramelidae
bats, 1417. Chicago: Chicago University Press. (Marsupialia). Bulletin of the American Muse-
Stein, B.R., and J.L. Patton. 2008b (2007). um Natural History 92: 313346.
Genus Lutreolina O. Thomas, 1910. In A.L. Temple-Smith, P. 1987. Sperm structure and
Gardner (editor), Mammals of South America. marsupial phylogeny. In M. Archer (editor),
Vol. 1. Marsupials, xenarthrans, shrews, and Possums and opossums: studies in evolution
bats, 2527. Chicago: Chicago University Press. Vol. 1, 171194. Sydney: Surrey Beatty.
Steiner, C., M. Tilak, E.J.P. Douzery, and F.M. Terasawa, M., R. Shimokawa, and T. Terashima.,
Catzeflis. 2005. New DNA data from a trans- et al. 2004. Expression of dentin matrix protein
thyretin nuclear intron suggest an Oligocene to 1 (DMP1) in nonmineralized tissues. Journal of
Miocene diversification of living South Ameri- Bone and Mineral Metabolism 22: 430438.
can opossums (Marsupialia: Didelphidae). Mo- Teta, P., E. Muschetto, S. Maidana, C. Bellomo,
lecular Phylogenetics and Evolution 35: 363 and P. Padula. 2007. Gracilinanus microtarsus
379. (Didelphimorphia, Didelphidae) en la provincia
Svartman, M., and A.M. Vianna-Morgante. 1999. de Misiones, Argentina. Mastozoologa Neo-
Comparative genome analysis in American tropical 14: 113115.
Teta, P., U.F.J. Pardinas, and G. DEla. 2006. van Nievelt, A.F.H., and K.K. Smith. 2005b. To
Rediscovery of Chacodelphys: a South American replace or not to replace: the significance of
marsupial genus previously known from a single reduced functional tooth replacement in marsu-
specimen. Mammalian Biology 71: 309314. pial and placental mammals. Paleobiology 31:
Thomas, O. 1887. On the homologies and succes- 324346.
sion of the teeth in the Dasyuridae, with an Venkitaraman, A.R. 2001. Functions of BRCA1
attempt to trace the history of the evolution of and BRCA2 in the biological response to DNA
mammalian teeth in general. Philosophical damage. Journal of Cell Science 114: 35913598.
Transactions of the Royal Society of London Ventura, J., R. Perez-Hernandez, and M.J. Lopez-
178B: 443462. Fuster. 1998. Morphometric assessment of the
Thomas, O. 1888. Catalogue of the Marsupialia Monodelphis brevicaudata group (Didelphimor-
and Monotremata in the collection of the British phia: Didelphidae) in Venezuela. Journal of
Museum (Natural History). London: Trustees Mammalogy 79: 104117.
of the British Museum (Natural History). Ventura, J., M. Salazar, R. Perez-Hernandez, and
Thomas, O. 1894. On Micoureus griseus Desm., M.J. Lopez-Fuster. 2002. Morphometrics of the
with the description of a new genus and species genus Didelphis (Didelphimorphia: Didelphi-
of Didelphyidae. Annals and Magazine of dae) in Venezuela. Journal of Mammalogy 83:
Natural History (ser. 6) 14: 184188. 10871096.
Thomas, O. 1895. On Caenolestes, a still existing Voss, R.S. 1991. An introduction to the Neotrop-
survivor of the Epanorthidae of Ameghino, and ical muroid rodent genus Zygodontomys. Bulle-
the representative of a new family of Recent tin of the American Museum of Natural History
marsupials. Proceedings of the Zoological Soci- 210: 1113.
ety of London 1895: 870878, pl. 50. Voss, R.S., A.L. Gardner, and S.A. Jansa. 2004a. On
Thomas, O. 1923. The geographical races of the relationships of Marmosa formosa Shamel,
Lutreolina crassicaudata. Annals and Magazine 1930 (Marsupialia: Didelphidae), a phylogenetic
of Natural History (ser. 9) 11: 583585. puzzle from the Chaco of northern Argentina.
Thotakura, S.R., N. Karthikeyan, T. Smith, K. American Museum Novitates 3442: 118.
Liu, and A. George. 1999. Cloning and charac- Voss, R.S., and S.A. Jansa. 2003. Phylogenetic
terization of rat dentin matrix protein 1 (DMP1) studies on didelphid marsupials II. Nonmolecular
gene and its 59-upstream region. Journal of data and new IRBP sequences: separate and
Biological Chemistry 275: 1027210277. combined analyses of didelphine relationships
Tribe, C.J. 1990. Dental age classes in Marmosa with denser taxon sampling. Bulletin of the
incana and other didelphoids. Journal of Mam- American Museum of Natural History 276: 182.
malogy 71: 566569. Voss, R.S., D.P. Lunde, and S.A. Jansa. 2005. On
Tyndale-Biscoe, H., and M. Renfree. 1987. Re- the contents of Gracilinanus Gardner and
productive physiology of marsupials. Cam- Creighton, 1989, with the description of a
bridge: Cambridge University Press. previously unrecognized clade of small didel-
Tyson, E. 1698. Carigueya, seu marsupiale amer- phid marsupials. American Museum Novitates
icanum. Or, the anatomy of an opossum, 3482: 134.
dissected at Gresham College by Edw. Tyson, Voss, R.S., D.P. Lunde, and N.B. Simmons. 2001.
M.D. Fellow of the College of Physicians, and The mammals of Paracou, French Guiana: a
of the Royal Society, and Reader of Anatomy at Neotropical rainforest fauna. Part 1. Nonvolant
the Chyrurgeons-Hall in London. Philosophical species. Bulletin of the American Museum of
Transactions 20: 105164. Natural History 263: 1236.
Van Den Bussche, R.A., S.A. Reeder, E.W. Voss, R.S., P. Myers, F. Catzeflis, A.P. Car-
Hansen, and S.R. Hoofer. 2003. Utility of the mignotto, and J. Barreiro. In press. The six
dentin matrix protein 1 (DMP1) gene for opossums of Felix de Azara: identification,
resolving mammalian intraordinal phylogenetic taxonomic history, neotype designations, and
relationships. Molecular Phylogenetics and Evo- nomenclatural recommendations. Bulletin of the
lution 26: 89101. American Museum of Natural History.
van der Klaauw, C.J. 1931. The auditory bulla in Voss, R.S., E. Yensen, and T. Tarifa. 2004b. An
some fossil mammals, with an introduction to introduction to Marmosops (Marsupialia: Di-
this region of the skull. Bulletin of the American delphidae), with the description of a new species
Museum of Natural History 62: 1352. from Bolivia and notes on the taxonomy and
van Nievelt, A.F.H., and K.K. Smith. 2005a. distribution of other Bolivian congeners. Amer-
Tooth eruption in Monodelphis domestica and ican Museum Novitates 3466: 140.
its significance for phylogeny and natural Wainberg, R.L., T. Gentile de Fronza, and J.G.
history. Journal of Mammalogy 86: 333341. Garca. 1979. Cromosomas marsupiales del
genero Marmosa: M. pusilla bruchi, M. agilis Winge, H. 1893. Jordfundne og nulevende Pung-
chacoensis, y M. microtarsus (Marsupialia: dyr (Marsupialia) fra Lagoa Santa, Minas
Didelphidae). Physis 38: 3338. Geraes, Brasilien. E Museo Lundii 2(2): 1132,
Weisbecker, V., and M. Nilsson. 2008. Integration, pls. 14.
heterochrony, and adaptation in pedal digits of Woolley, P. 1974. The pouch of Planigale sub-
syndactylous marsupials. BMC Evolutionary tilissima and other dasyurid marsupials. Journal
Biology 8: 14 pp. [no pagination; published of the Royal Society of Western Australia 57:
online only], doi:10.1186/1471-2148-8-160. 1115.
Weksler, M. 2003. Phylogeny of Neotropical Wroe, S. 1997. A reexamination of proposed
oryzomyine rodents (Muridae: Sigmodontinae) morphology-based synapomorphies for the
based on the nuclear IRBP exon. Molecular families of Dasyuromorphia (Marsupialia). I.
Phylogenetics and Evolution 29: 331349. Dasyuridae. Journal of Mammalian Evolution
Wells, R.T., and R.H. Tedford. 1995. Sthenurus 4: 1952.
(Macropodidae: Marsupialia) from the Pleisto- Wroe, S., J. Brammall, and B.N. Cooke. 1998. The
cene of Lake Callabonna, South Australia. skull of Ekaltadeta ima (Marsupialia, Hypso-
Bulletin of the American Museum of Natural prymnodontidae?): an analysis of some marsu-
History 225: 1111. pial cranial features and a re-investigation of
Whipple, I.L. 1904. The ventral surface of the propleopine phylogeny, with notes on the
mammalian chiridium, with special reference to inference of carnivory in mammals. Journal of
the conditions found in man. Zeitschrift fur Paleontology 72: 738751.
Morphologie und Anthropologie 7: 261368. Wroe, S., M. Crowther, J. Dortch, and J. Chong.
Wible, J.R. 1990. Petrosals of Late Cretaceous 2004. The size of the largest marsupial and why
marsupials from North America and a cladistic it matters. Proceedings of the Royal Society of
analysis of the petrosal in therian mammals. London B (suppl.) 271: S34S36.
Journal of Vertebrate Paleontology 10: 183205. Wroe, S., M. Ebach, S. Ahyong, C. de Muizon,
Wible, J.R. 2003. On the cranial osteology of the and J. Muirhead. 2000. Cladistic analysis of
short-tailed opossum Monodelphis brevicaudata dasyuromorphian (Marsupialia) phylogeny us-
(Didelphidae, Marsupialia). Annals of the Car- ing cranial and dental characters. Journal of
negie Museum 72: 137202. Mammalogy 81: 10081024.
Wible, J.R., G.W. Rougier, M.J. Novacek, and Wroe, S., and A. Musser. 2001. The skull of
M.C. McKenna. 2001. Earliest eutherian ear Nimbacinus dicksoni (Thylacinidae: Marsupia-
region: a petrosal referred to Prokennalestes lia). Australian Journal of Zoology 49: 487514.
from the Early Cretaceous of Mongolia. Amer- Ximenez, A. 1967. Contribucion al conocimiento
ican Museum Novitates 3322: 144. de Lutreolina crassicaudata (Desmarest, 1804) y
Wiens, J.J. 2000. Coding morphological variation sus formas geograficas. Comunicaciones Zool-
within species and higher taxa for phylogenetic ogicas del Museo de Historia Natural de
analysis. In J.J. Wiens (editor), Phylogenetic Montevideo 9(112): 17.
analysis of morphological data, 115145. Wash- Zarza, H., G. Ceballos, and M.A. Steele. 2003.
ington, DC: Smithsonian Institution Press. Marmosa canescens. Mammalian Species 725:
Wilson, D.E. 1991. Mammals of the Tres Maras 14.
Islands. Bulletin of the American Museum of Zwickl, D.J. 2006. Genetic algorithm approaches
Natural History 206: 214250. for the phylogenetic analysis of large biological
Wilson, D.R. and D.M. Reeder (editor). 2005. sequence datasets under the maximum likeli-
Mammal species of the world. 3rd ed. Balti- hood criterion. Unpublished Ph.D. dissertation,
more: Johns Hopkins University Press, 2 vols. University of Texas at Austin.
APPENDIX 1 127107127109, 127156, 155575, 155581. Subadult

skulls: FMNH 70858, 70859, 92299. Juvenile
Morphological Specimens Examined skulls: none. Fluid specimens: AMNH 248261,
248271.
The specimens from which morpholog- Rhyncholestes raphanurusSkins: FMNH 50071,
ical character data were scored for phyloge- 124002. Adult skulls: FMNH 127468, 127474,
129824, 129827, 129828, 129834, 135035. Subadult
netic analysis are listed below. These include skulls: none. Juvenile skulls: none. Fluid specimens:
the skins, skulls, and fluid-preserved mate- FMNH 129825.
rial previously examined by Voss and Jansa Dasyuridae
(2003), Voss et al. (2004a, 2005), and Jansa Murexia longicaudataSkins: AMNH 105022,
108558. Adult skulls: AMNH 101971101973,
and Voss (2005), together with additional 108554, 108555, 108557, 108558. Subadult skulls:
material newly examined for this report. none. Juvenile skulls: AMNH 101970. Fluid
Except as noted, all examined specimens specimens: AMNH 196649.
(including vouchers for sequenced tissues; Sminthopsis crassicaudataSkins: none. Adult skulls:
SAM 8254, 19819, 12332, 12333. Subadult skulls:
see appendix 2) are preserved in collections SAM 8441. Juvenile skulls: SAM 11536. Fluid
identified by the following institutional specimens: AMNH 160413 (parous female),
acronyms: AMNH, American Museum of 160416.
Natural History (New York); AMS, Aus- Didelphidae
tralian Museum (Sydney); ANWC, Austra- Caluromys lanatusSkins: MUSM 15290, 15291;
MVZ 190248, 190249. Adult skulls: AMNH
lian National Wildlife Collection (Can- 230001, 273038, 273059; MVZ 190249, 190251.
berra); BMNH, Natural History Museum Subadult skulls: MUSM 15291. Juvenile skulls:
(London); CM, Carnegie Museum of Nat- AMNH 75913; MUSM 15290; MVZ 190248. Fluid
ural History (Pittsburgh); CNP, Centro specimens: AMNH 273038, 273059; MVZ 191185
(parous female).
Nacional Patagonico (Puerto Madryn); Caluromys philanderSkins: AMNH 266408, 266409,
FMNH, Field Museum of Natural History 267001. Adult skulls: AMNH 266409, 267002,
(Chicago); INPA, Instituto Nacional de 267335267337. Subadult skulls: AMNH 266408,
Pesquizas da Amazonia (Manaus); KU, 267332. Juvenile skulls: AMNH 96629, 96652,
267330, 267333, 267334. Fluid specimens: AMNH
University of Kansas Museum of Natural 266402 (parous female), 266408.
History (Lawrence); MCZ, Museum of Caluromysiops irruptaSkins: AMNH 208101;
Comparative Zoology at Harvard Univer- FMNH 68336; USNM 396061, 397626. Adult
sity (Cambridge); MMNH, Bell Museum of skulls: AMNH 208101, 244364; FMNH 60698,
84426; USNM 396061, 397626. Subadult skulls:
Natural History (University of Minnesota, none. Juvenile skulls: FMNH 68336, 121522. Fluid
St. Paul); MNHN, Museum National dHis- specimens: FMNH 60154, 60398.
toire Naturelle (Paris); MUSM, Museo de Chacodelphys formosaSkins: USNM 236330. Adult
Historia Natural de la Universidad Nacio- skulls: CNP 639, USNM 236330, plus several
nal Mayor de San Marcos (Lima); MVZ, uncataloged cranial fragments recovered from owl
pellets. Subadult skulls: none. Juvenile skulls: none.
Museum of Vertebrate Zoology (University Fluid specimens: none.
of California, Berkeley); ROM, Royal Chironectes minimusSkins: AMNH 164494, 264571,
Ontario Museum (Toronto); SAM, South 264572, 266477, 266478. Adult skulls: AMNH
Australian Museum (Adelaide); TTU, Mu- 33027, 37483, 47190, 62365, 72020, 96759, 96760,
97319, 129704, 164494, 182939, 212909, 264572.
seum of Texas Tech University (Lubbock); Subadult skulls: AMNH 34197. Juvenile skulls:
UMMZ, University of Michigan Museum AMNH 16072, 126979, 127563, 128994, 264573,
of Zoology (Ann Arbor); UMSNH, Uni- 266478. Fluid specimens: AMNH 24411, 150033,
versidad Michoacana de San Nicolas de 266479; UMMZ 134560 (parous female).
Cryptonanus chacoensisSkins: AMNH 167851,
Hidalgo (Morelia); USNM, National Mu- UMMZ 126105, USNM 236329. Adult skulls:
seum of Natural History (Washington, DC); UMMZ 126105, 134552, 137143; USNM 236329.
UWZM, University of Wisconsin Zoologi- Subadult skulls: AMNH 167851. Juvenile skulls:
cal Museum (Madison); V-, voucher collec- BMNH 11.11.19.23. Fluid specimens: UMMZ
tion of F. Catzeflis, Universite de Montpel- 134552 (parous adult female).
Cryptonanus unduaviensisSkins: AMNH 209150,
lier (Montpellier). 209153, 209154; FMNH 114658, 114665. Adult
skulls: AMNH 209150, 209152209155; FMNH
Caenolestidae 114658. Subadult skulls: AMNH 262401, FMNH
Caenolestes fuliginosusSkins: UMMZ 155572, 114664. Juvenile skulls: FMNH 114667. Fluid
155580. Adult skulls: FMNH 70847; UMMZ specimens: AMNH 247754, MSB 70752.
Didelphis albiventrisSkins: AMNH 66779, 132897, 202727, 235546; UMMZ 166634 (parous female);
132898, 132906, 248304, 248305. Adult skulls: USNM 536827 (parous female).
AMNH 132905, 132942, 248303, 248304. Subadult Marmosa (Marmosa) lepidaSkins: AMNH 67279,
skulls: AMNH 66779, 132929. Juvenile skulls: 182937; MNHN 1998.306; MVZ 154750. Adult
AMNH 132894, 238007. Fluid specimens: AMNH skulls: AMNH 78001, 98656, 273186; KU 135118;
202706; UMMZ 134565134567. MNHN 1998.306; MVZ 154750, 155245; USNM
Didelphis marsupialisSkins: AMNH 209177, 461468. Subadult skulls: MNHN 1982.653; MVZ
247654, 266457, 266462. Adult skulls: AMNH 154752. Juvenile skulls: FMNH 140824. Fluid
266457, 266459, 266468, 266471, 266474. Subadult specimens: AMNH 273186; KU 135118; MVZ
skulls: AMNH 266460, 266472. Juvenile skulls: 155245.
AMNH 266464, 266466, 266470, 266473, 266475, Marmosa (Marmosa) mexicanaSkins: AMNH
267367. Fluid specimens: AMNH 210445, 272836. 12454/10763, 243700; ROM 95795, 96090, 99776.
Didelphis virginianaSkins: AMNH 90322, 139915, Adult skulls: AMNH 189209, 189484, 265851;
146632, 180141, 180142. Adult skulls: AMNH ROM 96090, 99608. Subadult skulls: AMNH
217749, 217762, 217772, 217775, 217786, 219223, 17136, 213754. Juvenile skulls: AMNH 189483,
219224, 219231, 219234, 219235. Subadult skulls: 189485. Fluid specimens: AMNH 265851 (parous
AMNH 217750, 217782, 219868. Juvenile skulls: female).
AMNH 217780, 217790, 219869, 219878, 240517. Marmosa (Marmosa) murinaSkins: AMNH
Fluid specimens: UMMZ 103462 (parous female), 273178; MUSM 15296. Adult skulls: AMNH
110758, 114837 (parous female). 272816, 273062, 273063, 273140, 273178, 273188;
Glironia venustaSkins: AMNH 71394, 71395; MUSM 15293. Subadult skulls: AMNH 272870;
FMNH 41440. Adult skulls: AMNH 71394, MUSM 13283, 15294. Juvenile skulls: MUSM
71395; FMNH 41440. Subadult skulls: INPA 15295, 15297. Fluid specimens: AMNH 268214,
2570. Juvenile skulls: none. Fluid specimens: INPA 272816, 273063, 273140; MUSM 15293.
2570. Marmosa (Marmosa) robinsoniSkins: AMNH
Gracilinanus aceramarcaeSkins: AMNH 72568, 36725, 36728, 37890, 147759. Adult skulls: USNM
LSU 17897, MVZ 171411. Adult skulls: AMNH 456812, 456815, 456821, 456826, 456833. Subadult
72568, LSU 17897, MVZ 171411, UMMZ 156004, skulls: USNM 456825. Juvenile skulls: USNM
156005. Subadult skulls: none. Juvenile skulls: 456813, 456814, 456817, 456820, 456848. Fluid
none. Fluid specimens: UMMZ 156004, 156005. specimens: UMMZ 117236, 117237.
Gracilinanus agilisSkins: MVZ 197437197441. Marmosa (Marmosa) rubraSkins: AMNH 71950,
Adult skulls: MVZ 197437197441. Subadult 71952, 71974; MVZ 153280, 153282, 153283,
skulls: none. Juvenile skulls: AMNH 133224. Fluid 154765. Adult skulls: MVZ 153280, 153282,
specimens: MVZ 197649, 197651, 197652, 197655. 153283, 154765. Subadult skulls: none. Juvenile
Gracilinanus emiliaeSkins: AMNH 203363, BMNH skulls: AMNH 68128, 68129; MVZ 154759. Fluid
9.3.9.10. Adult skulls: AMNH 203363, MUSM specimens: none.
15292, ROM 35466. Subadult skulls: AMNH Marmosa (Micoureus) demeraraeSkins: AMNH
267006. Juvenile skulls: ROM 33807, USNM 266427, 266428, 267370, 267818. Adult skulls:
385066. Fluid specimens: AMNH 267006, MUSM AMNH 266427, 266428, 266431, 266432, 267370,
15292 (parous adult female), ROM 35465 (parous 267371. Subadult skulls: AMNH 266433, 267369.
adult female). Juvenile skulls: AMNH 266429, 266430, 266434.
Gracilinanus microtarsusSkins: FMNH 92958, Fluid specimens: AMNH 266427, 266428, 266432
94294, 94295; MVZ 182055, 197436. Adult skulls: (parous female).
FMNH 94294, 94295; MVZ 197436. Subadult Marmosa (Micoureus) paraguayanaSkins: MVZ
skulls: FMNH 92958; MVZ 182055182057. Juve- 182063, 182064; UMMZ 165664. Adult skulls:
nile skulls: none. Fluid specimens: MVZ 182054. MVZ 182063182065; UMMZ 165664. Subadult
Hyladelphys kalinowskiiSkins: FMNH 89991. Adult skulls: none. Juvenile skulls: none. Fluid specimens:
skulls: AMNH 267003, 267338, 267339; FMNH MVZ 182065; UMMZ 134551.
65754, 89991; ROM 34271. Subadult skulls: none. Marmosa (Micoureus) reginaSkins: AMNH
Juvenile skulls: MUSM 11031, and an uncataloged 273164; MUSM 11063, 15316; MVZ 190323,
specimen (field number V-1791) to be deposited at 190332. Adult skulls: MVZ 190323, 190325,
the MNHN. Fluid specimens: AMNH 267338, 190331, 190332. Subadult skulls: MVZ 190333.
267339 (parous adult female). Juvenile skulls: MVZ 190324, 190328. Fluid
Lestodelphys halliSkins: BMNH 28.12.11.207; specimens: MVZ 191201 (parous female).
MMNH 15708, 17171; MVZ 173727; UWZM Marmosops impavidusSkins: AMNH 272760;
22422. Adult skulls: BMNH 28.12.11.207; MVZ MUSM 15298, 15299. Adult skulls: AMNH
160109, 171069, 171070, 179180, 179182, 173727; 272709, 272760, 273151; MUSM 13284, 15299,
UWZM 22422. Subadult skulls: MMNH 15708. 15300. Subadult skulls: AMNH 273050. Juvenile
Juvenile skulls: MMNH 17171; MVZ 17193. Fluid skulls: MUSM 13285, 13286, 15306, 15307. Fluid
specimens: none. specimens: AMNH 272709 (parous female),
Lutreolina crassicaudataSkins: AMNH 133249, 273151; MUSM 13284, 15300.
133250, 143886. Adult skulls: AMNH 133249, Marmosops incanusSkins: MVZ 182061, 182768,
139825, 210420, 254512, 254513. Subadult skulls: 182769; UMMZ 165662, 165663. Adult skulls:
AMNH 133255, 210419. Juvenile skulls: AMNH MVZ 182061, 182768182770; UMMZ 165662,
210421, 210423, 210424. Fluid specimens: AMNH 165663. Subadult skulls: UMMZ 165661. Juvenile
skulls: none. Fluid specimens: MVZ 182771, 13299. Subadult skulls: MVZ 190336, 190338.
197629. Juvenile skulls: AMNH 273039; MUSM 15319
Marmosops noctivagusSkins: AMNH 272775, 15321, 15323. Fluid specimens: MUSM 15319;
273131; MUSM 15305. Adult skulls: AMNH MVZ 191202 (parous female), 191203 (parous
272775, 272782, 272809, 273051; MUSM 13288. female).
Subadult skulls: MUSM 15301, 15303. Juvenile Philander opossumSkins: AMNH 266995, 266996,
skulls: AMNH 272704, 272715; MUSM 13289 266998, 267014. Adult skulls: AMNH 266379,
13292. Fluid specimens: AMNH 272782 (parous 266381, 266386, 266387, 267014, 267328. Subadult
female), 272809, 273034, 273051 (parous female); skulls: AMNH 266389, 266994. Juvenile skulls:
MUSM 13288. AMNH 266394, 266395, 266398, 266400, 266997.
Marmosops parvidensSkins: AMNH 266426, Fluid specimens: AMNH 266380, 266390.
267817. Adult skulls: AMNH 266421, 267344, Thylamys macrurusSkins: MSB 70700, MZUSP
267347, 267353, 267359, 267361. Subadult skulls: 3209432096, UMMZ 125243. Adult skulls:
AMNH 266422, 266425, 267817. Juvenile skulls: BMNH 3.4.7.21, MSB 70700, MZUSP 32094
AMNH 266426, 267350, 267355. Fluid specimens: 32096, UMMZ 125243. Subadult skulls: BMNH
AMNH 267344 (parous female), 267348, 267353, 99.11.17.1. Juvenile skulls: none. Fluid specimens:
267359, 267361. MZUSP 32097.
Marmosops pinheiroiSkins: AMNH 266423, Thylamys pallidiorSkins: AMNH 262406262408;
267007, 267352. Adult skulls: AMNH 267340, FMNH 54255, 162495. Adult skulls: AMNH
267342, 267345, 267346, 267349, 267352, 267357. 262406262408; FMNH 54255, 162495. Subadult
Subadult skulls: AMNH 267007, 267356. Juvenile skulls: AMNH 262405. Juvenile skulls: FMNH
skulls: AMNH 266424, 267008, 267351, 267354. 5100451006. Fluid specimens: UMMZ 156015.
Fluid specimens: AMNH 267342 (parous female), Thylamys pusillusSkins: AMNH 275440, 275442;
267345, 267346, 267356 (parous female), 267357. MSB 67016. Adult skulls: AMNH 246444, 246448,
Metachirus nudicaudatusSkins: AMNH 266450, 275440, 275442, 275445, 275446; MSB 67016,
267009, 267010. Adult skulls: AMNH 266435 87105. Subadult skulls: AMNH 246449; UMMZ
266437, 266440, 266449, 266450, 267009, 267010. 155837, 155840. Juvenile skulls: AMNH 246452;
Subadult skulls: AMNH 266439, 267362. Juvenile UMMZ 155838, 155839, 155841. Fluid specimens:
skulls: AMNH 266451266453, 267364. Fluid UMMZ 156026, 156029, 156030, 156032 (parous
specimens: AMNH 255815, 261283, 261285; adult female).
USNM 461138, 577756 (parous female).
Thylamys venustusSkins: AMNH 263558, 263562;
Monodelphis brevicaudataSkins: AMNH 36317,
FMNH 162505. Adult skulls: AMNH 263558,
48133, 75830, 75831. Adult skulls: AMNH
263561, 263562; FMNH 162505. Subadult skulls:
257203; USNM 393438, 393439, 393441, 461435,
AMNH 262400, 263555, 263557. Juvenile skulls:
578009. Subadult skulls: AMNH 268061; USNM
AMNH 263559, 263560. Fluid specimens: AMNH
393426, 393431. Juvenile skulls: AMNH 16953,
261250, 261251.
267744; USNM 392046, 392049, 393429, 393432,
393434, 393440, 393442, 543303, 568009. Fluid Tlacuatzin canescensSkins: AMNH 24896, 149104,
specimens: AMNH 140465, 140466, 244469. 172128, 185769. Adult skulls: AMNH 24894,
Monodelphis emiliaeSkins: AMNH 268221; MUSM 148969, 148970; USNM 70767, 73320, 125659,
13298; MVZ 190334; USNM 461883, 461884. 125925. Subadult skulls: AMNH 185770, 213753;
Adult skulls: AMNH 268221; MUSM 13298; USNM 37134, 70241, 96754, 508354. Juvenile
MVZ 190334; USNM 461883, 461884, 579574. skulls: AMNH 145237; USNM 70239, 70242.
Subadult skulls: MVZ 190335. Juvenile skulls: Fluid specimens: AMNH 144638, 165653; USNM
AMNH 95816. Fluid specimens: AMNH 95816; 9514.
USNM 579574. Microbiotheriidae
Monodelphis peruvianaSkins: AMNH 272695; Dromiciops gliroidesSkins: AMNH 92147, 97746;
MUSM 13297; USNM 582782, 588019. Adult FMNH 127450. Adult skulls: FMNH 50075,
skulls: AMNH 272695; MUSM 7157, 11654, 127440, 127453, 127460, 127464, 129804, 129807.
13297; USNM 582782, 588019. Subadult skulls: Subadult skulls: FMNH 127437, 127438, 127447.
AMNH 272781. Juvenile skulls: none. Fluid Juvenile skulls: FMNH 22673, 22675, 127445,
specimens: AMNH 272781; MUSM 15318. 127465. Fluid specimens: AMNH 238022; FMNH
Monodelphis theresaSkins: FMNH 25738; MVZ 127439 (parous female), 127456.
182775. Adult skulls: FMNH 25739; MVZ 182775, Peramelidae
182776. Subadult skulls: none. Juvenile skulls: Echymipera kalubuSkins: AMNH 190970, 190976,
none. Fluid specimens: FMNH 25739; MVZ 221664. Adult skulls: AMNH 190970, 190973,
182776. 190984, 190986, 192168, 236834. Subadult skulls:
Philander frenatusSkins: MVZ 182066, 182067, AMNH 221663. Juvenile skulls: AMNH 190977,
183246, 183247. Adult skulls: MVZ 182066, 190978, 221654, 221662. Fluid specimens: AMNH
182067, 183246. Subadult skulls: none. Juvenile 152838, 192693, 192694.
skulls: none. Fluid specimens: MVZ 182068 (par- Perameles gunniiSkins: none. Adult skulls: MVZ
ous female). 127059, 127070127073, 127099, 127100, 132246,
Philander mcilhennyiSkins: AMNH 272818, 273054, 132248. Subadult skulls: none. Juvenile skulls:
273089; MUSM 13299. Adult skulls: AMNH MVZ 127060. Fluid specimens: MVZ 127057
272818, 273040, 273054, 273055, 273089; MUSM (parous female), 127062.
APPENDIX 2 Gracilinanus microtarsus: Brazil, Sao Paulo, Fazenda

Intervales (MVZ 182055 [MAM 38], MVZ 182056
Specimens Sequenced for Nuclear Genes [MAM 49]).
Hyladelphys kalinowskii: French Guiana, Route de
Nuclear gene sequences were obtained Kaw, 30 km SE Cayenne (MNHN, uncataloged
[T4385]).
from the following specimens, listed alpha- Lestodelphys halli: Argentina, Ro Negro, 10 km by
betically by family, genus, and species. The road E Clementi Onelli (UWZ 22422 [Kirsch lab
institutional catalog number of each voucher extract 775]).
specimen is listed first (in parentheses), Lutreolina crassicaudata: Paraguay, Misiones, 2 km
followed by a tissue identifier (if any, in NE Ayolas (UMMZ 134018 [GKC 848], UMMZ
134019 [GKC 849]).
square brackets). See appendix 1 for institu- Marmosa (Marmosa) lepida: Guyana, Potaro-Sipar-
tional abbreviations. uni, Iwokrama Reserve (ROM 107034 [F 38809]);
Peru, Amazonas, Huampami on Ro Cenepa
Caenolestidae (MVZ 155245 [FC 4928, JLP7844]).
Caenolestes fuliginosus: Ecuador, Bolvar, 4 km E Marmosa (Marmosa) mexicana: Guatemala, El Pe-
Cruz de Lizo (MSB 70587 [NK 27708]). ten, Biotopo Cerro Cahui, El Ramate (ROM 99608
Rhyncholestes raphanurus: Chile, Osorno, La Picada [FN 32277]); Guatemala, El Progresso, Ro Uyus,
(Universidad Austral de Chile, unvouchered [LG 5 km E San Cristobal, Acasaguascatlan (ROM
420]). 99776 [FN 34135]).
Dasyuridae Marmosa (Marmosa) murina: Peru, Loreto, Ro
Murexia longicaudata: Papua New Guinea, Morobe, Galvez (AMNH 272816 [RSV 2303], AMNH
Mt. Missim (ANWC M29669 [female 3]); Papua 272870 [RSV 2413]); Surinam, Para, Zanderij
New Guinea, Southern Highlands Province, Bo- (CM 68346 [TK 17359], 68353 [TK 17387]).
bole (AMS M18453 [ABTC 45110]). Marmosa (Marmosa) robinsoni: Panama, Darien,
Sminthopsis crassicaudata: Australia, New South Cana (TTU 39117 [TK 22552], 39118 [TK 22555]).
Wales, Remington Station (ANWC M16816); Marmosa (Marmosa) rubra: Peru, Amazonas, 0.5 mi
Australia, South Australia, Ngarkat Conservation W Huampami on Ro Cenepa (MVZ 153280 [JLP
Park (SAM M20207 [ABTC 27560]). 6930]).
Didelphidae Marmosa (Micoureus) demerarae: Peru, Loreto, Ro
Caluromys lanatus: Ecuador, Napo, Parque Nacional Galvez (AMNH 272667 [RSV 2029], MUSM
Yasun (ROM 104570). 13294 [RSV 2085]).
Marmosa (Micoureus) paraguayana: Brazil, Sao
Caluromys philander: French Guiana, Les Nouragues
Paulo, Fazenda Intervales, Capao Bonito (MVZ
(V-823 [T-1754], V-960 [T-2020]).
182064 [MAM 46], 182065 [MAM 47]).
Caluromysiops irrupta: New York Zoological Society
Marmosa (Micoureus) regina: Brazil, Amazonas,
(AMNH 244364).
Igarape Nova Empresa, left bank Rio Jurua
Chironectes minimus: Guyana, Barima-Waini, Wai-
(MVZ 190323 [JLP 15435]); Brazil, Acre, Igarape
kerebi (ROM 98855 [FN 31677]).
Porongaba, right bank Rio Jurua (MVZ 190332
Cryptonanus chacoensis: Paraguay, Caazapa, Estancia [MNFS 1232]).
Dos Marias (uncataloged specimen returned to Marmosops impavidus: Peru, Loreto, Ro Galvez
Paraguay [GD 521]). (AMNH 272760 [RSV 2202], MUSM 13284 [RSV
Cryptonanus unduaviensis: Bolivia, Pando, Indepen- 2114]).
dencia (AMNH 262401 [NK 14234]); Bolivia, Marmosops incanus: Brazil, Minas Gerais, Estacao
Santa Cruz, Santiago de Chiquitos (AMNH Biologica de Caratinga, Fazenda Montes Claros,
260032 [NK 12313]). 54 km E Caratinga (MVZ 182768 [MAM 201],
Didelphis albiventris: Paraguay, Canendiyu, 13.3 km 182769 [MAM 202]).
N Curuguaty (UMMZ 134041 [GKC 783]); Para- Marmosops noctivagus: Ecuador, Napo, Parque Na-
guay, Presidente Hayes, 24 km NW Villa Hayes cional Yasun (ROM 105316 [F 37644]); Peru,
(UMMZ 134058 [GKC 816]). Loreto, Ro Galvez (AMNH 272775 [RSV 2225],
Didelphis marsupialis: Peru, Loreto, Ro Galvez AMNH 272782 [RSV 2242], AMNH 272809 [RSV
(AMNH 272836 [RSV 2357], MUSM 13282 [RSV 2294], MUSM 13289 [RSV 2224], MUSM 13292
2273]). [RSV 2131]).
Didelphis virginiana: Mexico, Yucatan, 1.5 km N Marmosops parvidens: Guyana, Upper Takutu-Upper
Labna (ROM 96483 [FN 30300]). Essequibo, Karanambo (ROM 97938 [FN 33439]).
Glironia venusta: Brazil, Amazonas, alto Rio Urucu Marmosops pinheiroi: Guyana, Potaro-Siparuni,
(INPA 2570 [MNFS 75]). Iwokrama Reserve (ROM 108920 [F 43900]).
Gracilinanus aceramarcae: Peru, Cusco, Cordillera Metachirus nudicaudatus: Peru, Loreto, Ro Galvez
Vilcabamba (MUSM 13002 [LHE 1342]). (AMNH 272780 [RSV 2236], MUSM 13293 [RSV
Gracilinanus agilis: Brazil, Minas Gerais, Mata do 2329]).
Vasco, 12 km W Nova Ponte (MVZ 197438 [LPC Monodelphis brevicaudata: Surinam, Saramacca, Ra-
363], MVZ 197439 [LPC 364]). leigh Falls (CM 63511 [TK 10244]); Surinam,
Gracilinanus emiliae: Peru, Loreto, Ro Galvez, Nickerie, Kayserberg Airstrip (CM 68359 [TK
Nuevo San Juan (MUSM 15292 [DWF 413]). 17069]).
Monodelphis emiliae: Peru, Loreto, Ro Galvez (1). See text for a discussion of this character and
(MUSM 13298 [RSV 2083]). its taxonomic distribution.
Monodelphis peruvianus (5 M. adusta of previous Character 4: Interramal vibrissae present (0); or absent
reports; see Solari, 2007): Peru, Loreto, Ro Galvez (1). See text for a discussion of this character and
(AMNH 272695 [RSV 2086]). its taxonomic distribution.
Monodelphis theresa: Brazil, Sao Paulo, Ilha de Character 5: Dark midrostral stripe absent (0); or
Sebastiao, Parque Estadual Ilhabela (MVZ present (1). Scoring criteria and information about
182776 [MAM 160]). the distribution of these traits among didelphid
Philander frenatus: Brazil, Sao Paulo, Fazenda terminal taxa were provided by Voss and Jansa
Intervales, Capao Bonito (MVZ 182066 [MAM (2003: character 2). See text for comments about
41], 182067 [MAM 64]). traits expressed by outgroup taxa.
Philander mcilhennyi: Peru, Loreto, Ro Galvez Character 6: Fur surrounding eye not distinctively
(AMNH 272818 [RSV 2310], MUSM 13299 [RSV colored (0); or eye surrounded by mask of dark fur
2153]). (1). Scoring criteria and information about the
Philander opossum: Surinam, Para, Zanderij (CM distribution of these traits among didelphid termi-
68365 [TK 17015]); Surinam, Suriname, Plantation nal taxa were provided by Voss and Jansa (2003:
Clevia, 8 km NE Paramaribo (CM 76743 [TK character 3). See text for comments about traits
17524]). expressed by outgroup taxa.
Thylamys macrurus: Paraguay, Concepcion, 7 km NE Character 7: Circumocular mask contrasts with color-
Concepcion (MSB 70700 [NK 27536). ation of cheeks and crown (0); or dark fur around
Thylamys pallidior: Bolivia, Tarija, roadside above eye continuous with dark coronal fur (1). This
Cieneguillas (FMNH 162495 [NBH 76-97]). character was modified from Voss and Jansas
Thylamys pusillus: Bolivia, Tarija, Estancia Bolvar (2003) character 4 by eliminating an intermediate
(MSB 67016 [NK 25139]). condition for polymorphism, which was recoded as
Thylamys venustus: Bolivia, Tarija, Chuquiaca 0/1 for this analysis. See text for comments about
(FMNH 162505 [BDP 3345]); Bolivia, Tarija, ca. traits expressed by outgroup taxa.
10 km by road W Narvaez (FMNH 162507 [BDP Character 8: Conspicuous pale spot above each eye
3309]). absent (0); or present (1). Scoring criteria and
Tlacuatzin canescens: Mexico, Jalisco, 6 km SE information about the distribution of these traits
Chamela (TTU 37700 [TK 11826]); Mexico, among didelphid terminal taxa were provided by
Michoacan, 1 km E Playa Azul (UMSNH 2993 Voss and Jansa (2003: character 5). See text for
[TK 45085]). comments about traits expressed by outgroup taxa.
Microbiotheriidae Character 9: Gular gland absent (0); or present (1).
Dromiciops gliroides: Argentina, Neuquen, near Es- Scoring criteria and information about the distri-
tancia Paso Coihue (MVZ 184914 [JLP 16545]). bution of these traits among didelphid terminal
Peramelidae taxa were provided by Voss and Jansa (2003:
Echymipera kalubu: Papua New Guinea, Western character 6). See text for comments about traits
Highlands, Trauna Valley (MVZ 138476 [FC 868]). expressed by outgroup taxa.
Perameles gunni: Australia, Tasmania, Ridgley (SAM, Character 10: Dorsal body pelage more or less
unvouchered [ABTC 66105]). uniformly colored, unpatterned (0); or marked by
dark transverse bars connected middorsally (1); or
marked by paired scapular stripes (2); or marked by
APPENDIX 3 one median and two lateral stripes (3); or grayish
middorsum contrasting with reddish or yellowish
Nonmolecular Character Descriptions flanks (4); or grayish midbody (including flanks)
contrasting with reddish head and rump (5); or
tricolored (6); or with transverse sacral bars (7).
EXTERNAL MORPHOLOGY Scoring criteria and information about the distri-
bution of states 06 among didelphid terminal taxa
Character 1: Pars supralabialis of rhinarium broad (0); were provided by Voss and Jansa (2003: character
or reduced to a narrow philtrum (1). See text for a 7). See text for comments about traits expressed by
discussion of this character and its taxonomic outgroup taxa. In the absence of any compelling
distribution. evidence for intermediacy among these alternative
Character 2: Rhinarium with two ventrolateral grooves conditions, this character was treated as unordered
flanking the median sulcus (0); or with a single in all of our analyses.
ventrolateral groove on each side (1). Information Character 11: Dorsal underfur dark (0); or white (1).
about the distribution of these traits among Scoring criteria and information about the distri-
didelphid terminal taxa was summarized by Voss bution of these traits among didelphid terminal
and Jansa (2003: character 1). See text for taxa were provided by Voss and Jansa (2003:
comments about traits expressed by outgroup taxa. character 8). See text for comments about traits
This character was scored as inapplicable (-) for expressed by outgroup taxa.
taxa with a very narrow pars supralabialis (see Character 12: Dorsal pelage hairs not grossly dissimilar
above). in length or coarseness (0); or dorsal pelage with
Character 3: Reciprocating lateral fleshy outgrowths of guard hairs conspicuously longer and coarser than
upper and lower oral margins absent (0); or present underfur (1); or dorsal pelage with stiff, dorsally
grooved guard hairs (2). Scoring criteria and (2003: character 10). See text for comments about
information about the distribution of states 0 and traits expressed by outgroup taxa.
1 among didelphid terminal taxa were provided by Character 18: Manus with well-developed plantar pads
Voss and Jansa (2003: character 9). See text for (0); or manus without distinct plantar pads (1). See
comments about traits expressed by outgroup taxa. text for a discussion of this character and its
In the absence of any compelling evidence for taxonomic distribution.
intermediacy among these alternative conditions, Character 19: Plantar pads of manus provided with
this character was treated as unordered in all of our dermatoglyphs (0); or manual plantar pads smooth
analyses. (1); or plantar pads tuberculate (2). See text for a
Character 13: Externally conspicuous lateral carpal discussion of this character and its taxonomic
tubercles absent in both sexes (0); or large adult distribution. This character was scored as inappli-
males with a prominent lateral carpal tubercle cable (-) for taxa lacking plantar pads (Chir-
supported internally by the pisiform (1); or both onectes and peramelids, see text). In the absence of
sexes with a finger-like fleshy tubercle supported any compelling evidence for intermediacy among
internally by the pisiform (2). This character is these alternative conditions, this character was
modified from Voss and Jansas (2003: character treated as unordered in all of our analyses.
12) interpretation of pisiform-supported tubercular Character 20: Central palmar surface of manus smooth,
structures by including a third state for the unique or sparsely provided with large flattened tubercles
condition seen in Chironectes. See text for com- (0); or densely covered with small convex tubercles
ments about traits expressed by outgroup taxa. In (1); or entire plantar surface of manus sandpapery,
the absence of any compelling evidence for covered with microscopically dentate tubercles (2).
intermediacy among these alternative conditions, Scoring criteria and information about the distri-
this character was treated as unordered in all of our bution of these unordered traits among didelphid
analyses. terminal taxa were provided by Voss and Jansa
Character 14: Externally conspicuous medial carpal (2003: character 11). See text for comments about
tubercles absent in both sexes (0); or large adult traits expressed by outgroup taxa.
males with a prominent medial tubercle supported Character 21: Hallux large and opposable (0); or very
internally by the prepollex (1). Scoring criteria and small and effectively nonopposable (1). See text for
information about the distribution of these traits a discussion of this character and its taxonomic
among didelphid terminal taxa were provided by distribution.
Voss and Jansa (2003: character 13). See text for Character 22: Pedal digit III distinctly longer than digit
comments about traits expressed by outgroup taxa. IV (0); or pedal digits III and IV subequal (1); or
Character 15: First manual digit (dI) large and claw- pedal IV distinctly longer than III (2). Scoring
bearing (0); or small and nail-bearing (1); or criteria and information about the distribution
vestigial and lacking an unguis entirely (2). See text and ordering (0 1 2) of these traits among
for a discussion of this character and its taxonomic didelphid terminal taxa were provided by Voss and
distribution. Because state 1 can be plausibly Jansa (2003: character 14). See text for comments
interpreted as intermediate to states 0 and 2 in about traits expressed by outgroup taxa.
each case, both characters were treated as ordered Character 23: Pedal digits free (0); or all pedal digits
(0 1 2) in all of our analyses. bound together by extensive fleshy webs (1); or
Character 16: Fifth manual digit (dV) large and claw- pedal digits II and III fused (2). This character
bearing (0); or small and nail-bearing (1); or differs from Voss and Jansas (2003) character 15
vestigial and without an unguis of any kind (2). See by including a separate state for the syndactylous
text for a discussion of this character and its hind feet of peramelids. In the absence of any
taxonomic distribution. Although functional re- compelling evidence for intermediacy among these
duction in manual digits I and V covary among alternative conditions, we treated this character as
marsupials and might be coded together as a single unordered in all of our analyses.
character on the assumption that the transforma- Character 24: Claws on pedal digits IIV all alike (0);
tions in question are developmentally or function- or claw on pedal digit II asymmetrically modified for
ally dependent, dI and dV exhibit independent grooming (1); or claws of pedal digits II and III
patterns of variation among placental mammals, asymmetrically and reciprocally modified for groom-
suggesting that such changes are not necessarily ing (2). See text for a discussion of this character
linked. (All rodents, for example, have a reduced and its taxonomic distribution. Although state 2 is
and clawless dI, whereas dV is typically large and said to be correlated with syndactyly (Hall, 1987),
claw-bearing.) Because state 1 can be plausibly some syndactylous marsupials (e.g., Phascolarctos,
interpreted as intermediate to states 0 and 2 in Vombatus) do not have grooming claws (Putzkral-
each case, both characters were treated as ordered len) on dII and dIII, an observation that implies
(0 1 2) in all of our analyses. some degree of independence between claw func-
Character 17: Manual digit III longer than other tion and digital fusion. In the absence of any
manual digits (0); or dIII and dIV subequal and compelling evidence for intermediacy among these
longer than other manual digits (1); or dIV longer alternative conditions, we treated this character as
than other manual digits (2). Scoring criteria and unordered in all of our analyses.
information about the distribution and ordering Character 25: Plantar epithelium of tarsus macroscop-
(0 1 2) of these traits among didelphid ically naked (0); or densely covered with coarse fur
terminal taxa were provided by Voss and Jansa (1). Scoring criteria and information about the
distribution of these traits among didelphid termi- evidence for intermediacy among these alternative
nal taxa were provided by Voss and Jansa (2003: conditions, this character was treated as unordered
character 16). See text for comments about traits in all of our analyses.
expressed by outgroup taxa. Character 33: Caudal integument uniformly pig-
Character 26: Pouch absent (0); or present (1). mented or indistinctly or irregularly marked (0);
Scoring criteria and information about the distri- or blackish basally and abruptly whitish distally (1).
bution of these traits among didelphid terminal Scoring criteria and information about the distri-
taxa were provided by Voss and Jansa (2003: bution of these traits among didelphid terminal
character 17). See text for comments about traits taxa were provided by Voss and Jansa (2003:
expressed by outgroup taxa. character 22). No examined outgroup taxon
Character 27: Pouch consists of separate lateral exhibits state 1.
skin folds opening medially (0); or lateral folds Character 34: Tail scales in predominantly annular
connected posteriorly, pouch opening anteriorly (1); series (0); or in annular and spiral series (1); or in
or lateral folds connected anteriorly, pouch opening predominantly spiral series (2). Scoring criteria and
posteriorly (2); or folds connected anteriorly and information about the distribution and ordering
posteriorly, pouch opening by a central orifice (3). (0 1 2) of these traits among didelphid
Scoring criteria and information about the terminal taxa were provided by Voss and Jansa
distribution of states 02 among didelphid ter- (2003: character 23). All examined outgroup taxa
minal taxa were provided by Voss and Jansa have annular caudal scale rows.
(2003: character 18). See text for comments Character 35: Unfurred ventral surface of tail base
about traits expressed by outgroup taxa. In covered with smooth scales (0); or provided with
the absence of any compelling evidence for raised tubercles (1). Scoring criteria and informa-
intermediacy among these alternative conditions, tion about the distribution of these traits among
we treated this character as unordered in all of our didelphid terminal taxa were provided by Voss and
analyses. Jansa (2003: character 24). No examined outgroup
Character 28: Mammae all abdominal/inguinal, more taxon has raised tubercles on the ventral surface of
or less confined to pouch region (0); or extending the tail base.
anteriorly beyond pouch region to thoracic region
Character 36: Ventral surface of tail tip covered by
(1). Scoring criteria and information about the
unmodified hair-bearing scales (0); or ventral
distribution of these traits among didelphid termi-
surface of tail tip modified for prehension (1).
nal taxa were provided by Voss and Jansa (2003:
Scoring criteria and information about the distri-
character 19). See text for comments about traits
bution of these traits among didelphid terminal
expressed by outgroup taxa.
taxa were provided by Voss and Jansa (2003:
Character 29: Unpaired median teat absent (0); or
character 25). See text for comments about traits
present (1). See text for a discussion of this
expressed by outgroup taxa.
character and its taxonomic distribution. Our
scoring of Rhyncholestes was based on literature Character 37: Caudal scales bearing three hairs
cited in the text. In the absence of suitably each (0); or more than three hairs each (1). Scoring
preserved parous female specimens or relevant criteria and information about the distribution of
published descriptions of mammary counts, we these traits among didelphid terminal taxa were
scored 11 didelphid terminals as missing (?) for provided by Voss and Jansa (2003: character
this character. 26). Despite the fact that annular scale rows
Character 30: Urogenital and rectal openings are could be distinguished in all examined out-
inguinal (0); or basicaudal (1). See text for a group taxa, individual scales were seldom clearly
discussion of this character and its taxonomic distinguishable within rows, so the number of
distribution. hairs per scale could not be counted. The
Character 31: Urogenital and rectal openings closely exceptions were Caenolestes (which usually has
juxtaposed, and sharing a common mucosa (0); or three hairs per scale; state 0) and Dromiciops
urogenital and rectal openings widely separated (which usually has four or more hairs per scale;
by furred skin (1). Scoring criteria and infor- state 1).
mation about the distribution of these traits Character 38: Tail hairs emerging from each scale not
among didelphid terminal taxa were provided grossly differentiated, varying in length but subequal
by Voss and Jansa (2003: character 20). See text in thickness (0); or central hair much thicker than
for comments about traits expressed by outgroup lateral hairs (1). Scoring criteria and information
taxa. about the distribution of these traits among
Character 32: Body pelage extends onto tail conspic- didelphid terminal taxa were provided by Voss
uously farther dorsally than ventrally (0); or body and Jansa (2003: character 27). Adjacent caudal
pelage extends onto dorsal and ventral surfaces of hairs were similar in gross morphology (state 0) in
tail to about the same extent (1); or body pelage all examined outgroups.
does not extend appreciably onto tail (2). Scoring Character 39: Tail not incrassate (0); or incrassate
criteria and information about the distribution of (1). Scoring criteria and information about the
these traits among didelphid terminal taxa were distribution of these traits among didelphid termi-
provided by Voss and Jansa (2003: character 21). nal taxa were provided by Voss and Jansa (2003:
See text for comments about traits expressed by character 28). See text for comments about traits
outgroup taxa. In the absence of any compelling expressed by outgroup taxa.
CRANIUM AND MANDIBLE Character 50: Foramen rotundum laterally exposed

behind sphenorbital fissure (0); or foramen rotun-
Character 40: Premaxillae not produced anteriorly dum concealed from lateral view, partly or wholly
beyond alveolar margin of I1 (0); or forming a confluent with sphenorbital fissure in rear of orbit
distinct, shelf-like rostral process anterior to I1 (1). (1). See text for a discussion of this character and
Scoring criteria and information about the distri- its taxonomic distribution.
bution of these traits among didelphid terminal Character 51: Maxillary and alisphenoid separate (0);
taxa were provided by Voss and Jansa (2003: or in contact on orbital floor (1). Scoring criteria
character 29). Among our outgroups, only Caeno- and information about the distribution of these
lestes, Dromiciops, and Perameles exhibit state 1. traits among didelphid terminals were provided by
Character 41: Nasals produced anteriorly above or Voss and Jansa (2003: character 43). See text for
beyond I1, obscuring nasal orifice in dorsal view (0); comments about traits expressed by outgroup taxa.
or nasals posterior to I1, exposing nasal orifice Character 52: Postorbital processes absent or indistinct
dorsally (1); or nasals posterior to premaxillary (0); or present (1). This character was modified
facial processes, exposing incisive foramina in dorsal from Voss and Jansas (2003) character 34 by
view (2). Scoring criteria and information about eliminating our former distinction between flat-
the distribution of states 0 and 1 among didelphid tened and not flattened postorbital processes,
terminal taxa were provided by Voss and Jansa which are not consistently distinguishable among
(2003: character 32). See text for comments about nondidelphid marsupials. See text for comments
traits expressed by outgroup taxa. Because state 1 about traits expressed by outgroup taxa.
is obviously intermediate to states 0 and 2, this Character 53: Postorbital processes formed by frontals
character was treated as ordered (0 1 2) in all only (0); or by frontals and parietals (1). In the
of our analyses. context of this study, state 1 is an autapomorphy of
Character 42: Nasals extend posteriorly between Glironia (see text).
lacrimals (0); or truncated anterior to lacrimals Character 54: Left and right frontals separated by
(1). Among the taxa examined for this study, ontogenetically persistent median suture (0); or left
nasals that are truncated anterior to the lacrimals and right frontals partially or completely co-ossified
(state 1) are peculiar to peramelemorphians (see in adults (1). See text for a discussion of this
text). character and its taxonomic distribution.
Character 43: Nasals conspicuously wider posteriorly Character 55: Left and right parietals separated by
than anteriorly (0); or nasals uniformly narrow, ontogenetically persistent median suture (0); or left
their lateral margins subparallel (1). Scoring and right parietals partially or completely co-ossified
criteria and information about the distribution of in adults (1). See text for a discussion of this
these traits among didelphid terminals were pro- character and its taxonomic distribution.
vided by Voss and Jansa (2003: character 33). Character 56: Parietal and alisphenoid in contact on
Among our outgroups, only Echymipera and lateral aspect of braincase (0); or frontal and
Perameles exhibit state 1. squamosal in contact (1). Scoring criteria and
Character 44: Antorbital fenestra between nasal, information about the distribution of these traits
frontal, and maxillary bones on each side of rostrum among didelphid terminals were provided by Voss
absent (0); or present (1). See text for a discussion and Jansa (2003: character 36). See text for
of this character and its taxonomic distribution. comments about traits expressed by outgroup taxa.
Character 45: Maxilloturbinals large and elaborately Character 57: Scars of M. temporalis origin on
branched (0); or small and unbranched (1). Scoring braincase not joined middorsally to form sagittal
criteria and information about the distribution of crest (0); or sagittal crest small, not extending to
these traits among didelphid terminals were pro- frontals (1); or sagittal crest large and extending to
vided by Voss and Jansa (2003: character 31). See frontals (2). Descriptions of character states and
text for comments about traits expressed by out- an account of their taxonomic distribution among
group taxa. didelphid terminals were provided by Voss and
Character 46: Maxillary-jugal suture straight, crescen- Jansa (2003: character 35), but an intermediate
tic, or irregular (0); or suture deeply bifid, dividing state between absence of a sagittal crest and
jugal into dorsal and ventral processes flanking presence of a large crest extending to the frontals
antorbital fossa (1). See text for a discussion of is recognized here. See text for comments about
this character and its taxonomic distribution. traits expressed by outgroup taxa. Because state 1
Character 47: Distinct frontal process of jugal absent is obviously intermediate to states 0 and 2, this
(0); or present (1). See text for a discussion of this character was treated as ordered (0 1 2) in all
character and its taxonomic distribution. of our analyses.
Character 48: Lacrimal margin of orbit smoothly Character 58: Fenestra in squamosal-parietal suture
rounded (0); or forming a distinct crest (1). See absent (0); or present (1). This character is
text for a discussion of this character and its modified from its original formulation (Voss and
taxonomic distribution. Jansa, 2003: character 37) by eliminating an
Character 49: One lacrimal foramen usually present intermediate state for polymorphisms, which were
(0); or two lacrimal foramina usually present (1). recoded as 0/1 for this analysis. See text for
See text for a discussion of this character and its comments about traits expressed by outgroup
taxonomic distribution. taxa.
Character 59: Interparietal absent (0); or present (1). character have been slightly reworded from their
See text for a discussion of this character and its original phrasing in Voss and Jansa (2003: charac-
taxonomic distribution. ter 42), where the distribution of these traits was
Character 60: Interparietal and supraoccipital separat- described for didelphid terminals. See text for
ed by an open suture, at least in juvenile specimens comments about traits expressed by outgroup taxa.
(0); or interparietal-supraoccipital suture fused in Character 70: Presphenoid exposed in roof of nasopha-
juveniles and adults (1). See text for a discussion of ryngeal orifice above posterior palatal margin (0);
this character, which was scored as inapplicable or presphenoid concealed by vomer above posterior
(-) for taxa lacking an interparietal, and as mis- palatal margin (1). See text for a discussion of this
sing (?) for taxa in which it is not known whether character and its taxonomic distribution.
an interparietal is present or absent (see above). Character 71: Left and right pterygoids separate (0);
Character 61: Parietal-mastoid contact present (0); or or in midline contact and forming a sagittal keel (1).
absent (1). See text for a discussion of this See text for a discussion of this character and its
character and its taxonomic distribution. taxonomic distribution.
Character 62: Accessory fenestrae absent (0); or Character 72: Transverse canal foramen absent (0); or
present (1). These nonvascular openings (first present (1). Scoring criteria and information about
apparently named by Lyne and Mort, 1981: 124) the distribution of these traits among didelphid
perforate the maxillary palate at the level of P1 terminals were provided by Voss and Jansa (2003:
and/or P2, occupying a position in between the character 44). See text for comments about traits
incisive foramina and the maxillopalatine fenestrae. expressed by outgroup taxa.
In the context of this study, state 1 is an Character 73: Extracranial course of mandibular nerve
autapomorphy of Perameles. not enclosed by bone (0); or consistently enclosed by
Character 63: Maxillopalatine fenestrae absent (0); or anteromedial strut of alisphenoid bulla (1); or
present (1). This character was modified from its usually enclosed by posteromedial bullar lamina
original formulation (Voss and Jansa, 2003: (2). Scoring criteria and information about the
character 38) by eliminating an intermediate state distribution of these traits among didelphid termi-
for polymorphism, which was recoded as 0/1 for nals were provided by Voss and Jansa (2003:
this analysis. See text for comments about traits character 45). See text for comments about traits
expressed by outgroup taxa. expressed by outgroup taxa. In the absence of any
Character 64: Left and right maxillopalatine fenestrae compelling evidence for intermediacy among these
separated by a median septum (0); or confluent, alternative conditions, this character was treated as
septum absent (1). See text for a discussion of this unordered in all of our analyses.
character and its taxonomic distribution. Character 74: Alisphenoid tympanic process and rostral
Character 65: Palatine fenestrae absent (0); or present tympanic process of petrosal separated by a distinct
(1). Scoring criteria and information about the gap (0); or in contact (1). See text for a discussion
distribution of these traits among didelphid termi- of this character and its taxonomic distribution.
nals were provided by Voss and Jansa (2003: Character 75: Epitympanic recess small, snugly enclos-
character 39). See text for comments about traits ing the mallear-incudal joint (0); or inflated, arching
expressed by outgroup taxa. high above mallear-incudal joint (1). See text for a
Character 66: Maxillary fenestrae absent (0); or discussion of this character and its taxonomic
present (1). Scoring criteria and information about distribution.
the distribution of these traits among didelphid Character 76: Rostral and caudal tympanic processes of
terminals were provided by Voss and Jansa (2003: petrosal separated by a distinct gap (0); or in
character 40). See text for comments about traits contact and separated by a suture (1); or indistin-
expressed by outgroup taxa. guishably fused (2). States 0 and 1 of this character
Character 67: Posterolateral palatal foramina present, represent the same pattern of ingroup variation
with complete bony margins (0); or absent, bony that Voss and Jansa (2003: character 47) described
margins incomplete, represented by open notches in in terms of concealment of the fenestra cochleae.
posterolateral palate (1). See text for a discussion See text for comments about traits expressed by
of this character and its taxonomic distribution. outgroup taxa. Because state 1 is plausibly
Character 68: Posterolateral palatal foramina posterior intermediate to states 0 and 2, this character was
to fourth molars (0); or extending lingual to M4 treated as ordered (0 1 2) in all of our
protocones (1). Scoring criteria and information analyses.
about the distribution of these traits among Character 77: Squamosal epitympanic sinus absent (0);
didelphid terminals were provided by Voss and or present (1). See text for a discussion of this
Jansa (2003: character 41). See text for comments character and its taxonomic distribution.
about traits expressed by outgroup taxa. Although Character 78: Anterior limb of ectotympanic directly
posterolateral palatal foramina are coded as absent attached to skull (0); or indirectly attached via
in dasyurids, their position behind the M4 proto- malleus (1). Scoring criteria and information about
cones in those taxa is indicated by the open notches the distribution of these traits among didelphid
mentioned in the preceding character description. terminals were provided by Voss and Jansa (2003:
Character 69: Posterior palate without distinct lateral character 46). See text for comments about traits
corners, the choanae unconstricted behind (0); or expressed by outgroup taxa.
posterior palate with distinct lateral corners, choanae Character 79: Ectotympanic laterally exposed (0); or
constricted (1). Descriptions of the states of this concealed from lateral view inside bulla (1). See text
for a discussion of this character and its taxonomic Character 88: Masseteric foramen absent (0); or
distribution. present (1). See text for a discussion of this
Character 80: Malleus with long and sharply inflected character and its taxonomic distribution.
neck, well-developed lamina, and manubrium parallel
to anterior process (0); or with short and uninflected
neck, no lamina, and manubrium not parallel to UPPER DENTITION
anterior process (1). See text for a discussion of this
character and its taxonomic distribution. Character 89: First upper incisor styliform or
Character 81: Stapes triangular and bicrurate, perfo- chisellike (0); or I1 mesiodistally expanded and
rated by a large foramen (0); or stapes columelli- flat-crowned (1). See text for a discussion of this
form and microperforate or imperforate (1). See character and its taxonomic distribution.
text for a discussion of this character and its Character 90: Second through fourth upper incisors
taxonomic distribution. Although Schmelzle et al. with approximately symmetrical, rhomboidal crowns
(2005: character 12) scored Didelphis marsupialis as bearing subequal anterior and posterior cutting edges
having a columelliform stapes, most of the (0); or I2I4 conspicuously asymmetrical, with
specimens of D. marsupialis that we examined have much longer anterior than posterior cutting edges
a triangular stapes with a large foramen. (1); or I2I4 hatchet-shaped (2); or I2I4 flat-
Character 82: Lambdoid sesamoids absent (0); or crowned (3). See text for a discussion of this
present (1). See text for a discussion of this character and its taxonomic distribution. In the
character and its taxonomic distribution. absence of any compelling evidence for intermedi-
Character 83: Dorsal margin of foramen magnum acy among these alternative conditions, this char-
formed by exoccipitals and supraoccipital, incisura acter was treated as unordered in all of our
occipitalis present (0); or by exoccipitals only, analyses.
incisura occipitalis absent (1). Scoring criteria and Character 91: Fifth upper incisor present (0); or I5
information about the distribution of these traits absent (1). See text for a discussion of this
among didelphid terminals were provided by Voss character and its taxonomic distribution.
and Jansa (2003: character 49). See text for Character 92: Fifth upper incisor similar to I4 (0); or
explanation of reworded state descriptions and I5 caniniform and double-rooted (1). See text for a
for comments about traits expressed by outgroup discussion of this character and its taxonomic
taxa. distribution. This character was scored as inappli-
Character 84: Paroccipital process absent (0); small cable (-) for taxa lacking I5.
and broadly adnate to the petrosal (1); or large, Character 93: Upper canine alveolus occupies premax-
erect and usually directed ventrally (2). This illary-maxillary suture (0); or C1 alveolus contained
character differs from Voss and Jansas (2003) entirely in maxillary bone (1). This character is
character 48 by including a new state (0) for rephrased from Voss and Jansa (2003: character
absence of the paroccipital process in Dromiciops. 30), who misleadingly described the same taxo-
See text for comments about traits expressed by nomic variation in terms of the presence or absence
other outgroup taxa. Because state 1 is plausibly of a premaxillary palatal process. See text for
intermediate to states 0 and 2, this character was comments about traits expressed by outgroup taxa.
treated as ordered (0 1 2) in all of our Character 94: Upper canine simple, without distinct
analyses. accessory cusps (0); or with distinct posterior
Character 85: Mandible usually with one mental accessory cusp (1); or with distinct anterior and
foramen (0); or with two mental foramina (1); or posterior accessory cusps (2); or with distinct
with three mental foramina (2). This character anterior accessory cusp (3). This character is
differs from Voss and Jansas (2003) character 50 restructured from Voss and Jansas (2003) charac-
by including a new state for the three mental ter 53, which treated states 0, 1, and 2 as stages of
foramina of Perameles gunni. See Jansa and Voss an ordered transformation series. The discovery of
(2005) for a discussion of polymorphism in another condition (state 3) among the outgroup
Hyladelphys, and the text of this report for taxa of this study introduced another hypotheti-
information about outgroup variation. Because cally intermediate condition between the absence of
state 1 is plausibly intermediate to states 0 and 2, any accessory cusps and the presence of both
this character was treated as ordered (0 1 2) anterior and posterior accessory cusps. Although
in all of our analyses. this could be accommodated by a step matrix, we
Character 86: Angular process acute and strongly have also observed dentitions with indistinct
inflected (0); or obtuse and weakly inflected (1). anterior and posterior accessory cusps, suggesting
Scoring criteria and information about the distri- that transitions between states 0 and 2 need not
bution of these traits among didelphid terminals involve states 1 or 3 as intermediate stages.
were provided by Voss and Jansa (2003: character Therefore, we now treat this character as unor-
51). See text for explanation of reworded state dered. We scored sexual dimorphism in canine
descriptions and for comments about traits ex- morphology in caenolestids as polymorphisms (0/3
pressed by outgroup taxa. for Caenolestes, 0/2 for Rhyncholestes; see text).
Character 87: Retromolar fossa imperforate (0); or Character 95: First upper premolar (P1) large, at least
pierced by retrodental canal (1). See text for a one-half the height or width of P2 (0); or vestigial,
discussion of this character and its taxonomic less than one-third the height or width of P2, or
distribution. absent (1). Scoring criteria and information about
the distribution of these traits among didelphid diacy of states 1 and 2 with respect to 0 and 3
terminals were provided by Voss and Jansa (2003: suggest that these traits be scored as elements of a
character 54). See text for comments about traits single ordered transformation series (0 1 2
expressed by outgroup taxa. 3). This character was scored as inapplicable (-)
Character 96: P2 distinctly taller than P3 (0); or P2 for caenolestids, which lack postparacristae and
and P3 subequal in height (1); or P3 distinctly taller premetacristae due to fusion of the paracone and
than P2 (2). Scoring criteria and information metacone with stylar cusps B and D, respectively
about the distribution and ordering (0 1 2) (see text).
of these traits among didelphid terminals were Character 105: Preprotocrista and anterolabial cingu-
provided by Voss and Jansa (2003: character 55). lum joined to form a continuous shelf along anterior
See text for comments about traits expressed by margin of M3 (0); or crista and cingulum separate,
outgroup taxa. not forming a continuous shelf (1). Scoring criteria
Character 97: Third upper premolar (P3) with well- and information about the distribution of these
developed anterior and posterior cutting edges (0); traits among didelphid terminals were provided by
or only posterior cutting edge well developed (1). Voss and Jansa (2003: character 60). Murexia,
Scoring criteria and information about the distri- Dromiciops, and Echymipera exhibit state 0,
bution of these traits among didelphid terminals whereas Sminthopsis exhibits state 1. This character
were provided by Voss and Jansa (2003: character was scored as inapplicable (-) for caenolestids
56). See text for comments about traits expressed and Perameles, which lack an anterolabial cingu-
by outgroup taxa. lum.
Character 98: Deciduous third premolar (dP3) large Character 106: Postprotocrista unnotched (0); or with
and molariform, occluding with dp3 and m1 (0); or
a distinct carnassial notch (1). See text for a
smaller, lacking a distinct protocone, and occluding
discussion of this character and its taxonomic
only with m1 (1); or very small, nonoccluding, and
distribution among didelphid terminals. No exam-
structurally simplified (2). See text for a discus-
ined outgroup taxon has a similarly notched
sion of this character and its taxonomic distribu-
postprotocrista.
tion. Because state 1 is plausibly intermediate to
states 0 and 2, we treated this character as ordered Character 107: Upper molar posterolingual cusp absent
(0 1 2) in all of our analyses. (0); or present (1). See text for a discussion of this
Character 99: Stylar shelf present (0); or absent (1). character and its taxonomic distribution.
See text for a discussion of this character and its Character 108: Third premolar is last upper tooth to
taxonomic distribution. erupt (0); or M4 and P3 erupt simultaneously (1);
Character 100: Neomorphic labial cingulum absent (0); or M4 is last upper tooth to erupt (2). Scoring
or present (1). See text for a discussion of this criteria and information about the distribution and
character and its taxonomic distribution. ordering (0 1 2) of these traits among
Character 101: Upper molars without a distinct didelphid terminal taxa were provided by Voss and
ectoflexus on any tooth (0); or distinct ectoflexus Jansa (2003: character 61). Suitable subadult
present on one or more teeth (1). Scoring criteria specimens were not available to determine eruption
and information about the distribution of these sequences for several species, for which this
traits among didelphid terminals were provided by character was coded as missing (?). Our scoring
Voss and Jansa (2003: character 59). See text for of Rhyncholestes was based on Luckett and Hongs
comments about traits expressed by outgroup taxa. (2000) observation of a subadult specimen that we
This character was coded as inapplicable (-) for did not examine.
taxa lacking a stylar shelf (see above).
Character 102: Anterolabial cingulum present (0); or
absent (1). See text for a discussion of this LOWER DENTITION
character and its taxonomic distribution.
Character 103: Preparacrista of M1 passes anterola- Character 109: Lower incisors with distinct lingual
bially or labially to stylar shelf (0); or M1 cusp (0); or lower incisors without distinct lingual
preparacrista passes posterolabially to stylar shelf cusps (1). Scoring criteria and information about
(1). See text for a discussion of this character and the distribution of these traits among didelphid
its taxonomic distribution. terminals were provided by Voss and Jansa (2003:
Character 104: Centrocrista of M1M3 linear, its apex character 62). See text for comments about traits
almost level with floor of trigon basin (0); or weakly expressed by outgroup taxa.
-shaped, its apex distinctly elevated above floor of Character 110: Anteriormost lower incisor small and
trigon basin (1); or strongly -shaped, its apex high short-crowned (0); or large, long-crowned, and
above floor of trigon basin (2); or discontinuous (3). conspicuously procumbent (1). See text for a
Scoring criteria and information about the distri- discussion of this character and its taxonomic
bution of traits 02 among didelphid terminals distribution.
were provided by Voss and Jansa (2003: character Character 111: Second lower incisor (i2) with stag-
58). Only peramelemorphian outgroups exhibit gered alveolus (0); or I2 alveolus not staggered (1).
state 3 (see text). Although Wroe et al. (2000) See text for a discussion of this character and its
and Wroe and Musser (2001) analyzed the shape of taxonomic distribution. This character was scored
the centrocista and whether or not it is discontin- as missing (?) for caenolestids, for which lower
uous as separate characters, the obvious interme- incisor homologies are unknown.
Character 112: Cutting edge of unworn i3 not bilobed Character 122: Entocristid absent or indistinct (0); or
(0); or distinctly bilobed (1). See text for a present (1). A distinct entocristid is present in all
discussion of this character and its taxonomic examined taxa except Sminthopsis and Perameles.
distribution. This character was scored as missing Character 123: Entocristid a short crest parallel to
(?) for caenolestids, for which lower incisor lingual margin of tooth (0); or entocristid long and
homologies are unknown. labially deflected into talonid basin (1). See text for
Character 113: Fourth lower incisor (i4) present (0); a discussion of this character and its taxonomic
or absent (1). See text for a discussion of this distribution. This character was coded as inappli-
character and its taxonomic distribution. This cable (-) in taxa lacking an entocristid (above).
character was scored as missing (?) for caenoles- Character 124: Hypoconulid at or near posterolingual
tids, for which lower incisor homologies are margin of talonid, twinned with entoconid on m1
unknown. m3 (0); or at midline of tooth, approximately
Character 114: Second lower premolar distinctly taller equidistant to hypoconid and entoconid, not
than p3 (0); or p2 and p3 subequal in height (1); or twinned with the latter cusp (1). Scoring criteria
p3 distinctly taller than p2 (2). Scoring criteria and and information about the distribution of these
information about the distribution and ordering traits among didelphid terminals were provided by
(0 1 2) of these traits among didelphid Voss and Jansa (2003: character 67). All examined
terminals were provided by Voss and Jansa (2003: ougroup taxa exhibit state 0.
character 63). See text for comments about traits Character 125: Lower molars without a posterior
expressed by outgroup taxa. cingulid (0); or posterior cingulid present (1). See
Character 115: Deciduous lower third premolar (dp3) text for a discussion of this character and its
large and occlusally functional, with tricuspid taxonomic distribution.
(complete) trigonid (0); or dp3 smaller but still
occlusally functional, with unicuspid or bicuspid
(incomplete) trigonid (1); or dp3 vestigial, usually KARYOTYPES
lacking distinct occlusal features (2). See text for a
Character 126: Robertsonian equivalents {acr1 + acr5,
discussion of this character and its taxonomic
met1} present as a single metacentric chromosome
distribution. In the absence of juvenile specimens,
(0); or as two acrocentric chromosomes (1). Most
this character was scored as missing (?) for
marsupials, including all of the outgroups in this
several taxa. Because state 1 is plausibly interme-
analysis, have a diploid number (2n) of 14
diate to states 0 and 2, we treated this character as
chromosomes (Sharman, 1973; Hayman, 1977,
ordered (0 1 2) in all of our analyses.
1990; Reig et al., 1977), and this is also the most
Character 116: Hypoconulid notch in anterior cingulum widespread number among didelphids (table 18).
of m2m4 present (0); or absent (1). See text for a By contrast, Glironia and Monodelphis have diploid
discussion of this character and its taxonomic counts of 18 chromosomes, whereas Chironectes,
distribution. Didelphis, Lutreolina, Philander, and Tlacuatzin
Character 117: Paracristid of m1 deeply notched (0); have 2n 5 22. Morphometric comparisons and
or unnotched (1). See text for a discussion of this G-banding studies indicate that each of these
character and its taxonomic distribution. diploid groupings are homogeneous in the sense
Character 118: Paracristid of m2 extends from that different taxa with the same diploid number
protoconid to paraconid without deflection (0); have chromosomes with essentially similar relative
or m2 paracristid abruptly deflected postero- sizes, shapes (arm ratios), and banding patterns
lingually by hypoconulid of m1 (1). See text for a (Reig et al., 1977; Rofe and Hayman, 1985; Svart-
discussion of this character and its taxonomic man and Vianna-Morgante, 1999). By implication, a
distribution. shared history of Robertsonian changes (centric
Character 119: Trigonid of m4 complete (0); or only fissions or centric fusions) could account for the
metaconid distinct on m4 trigonid (1). See text for a taxonomic membership of each diploid category.
discussion of this character and its taxonomic A minimum of four centric fission/fusion events
distribution. is required to account for the difference in diploid
Character 120: Lower third molar hypoconid labially number between karyotypes with 14 chromosomes
salient (0); or m3 hypoconid lingual to salient and those with 22 chromosomes. Based on arm
protoconid (1). Scoring criteria and information homologies suggested by G-banding patterns, the
about the distribution of these traits among following Robertsonian transformations involving
didelphid terminals were provided by Voss and eight acrocentric (acr) autosomes in the 2n 5 22
Jansa (2003: character 65). Caenolestids, Dromi- karyotype and four metacentric (met) autosomes in
ciops, and peramelid outgroups have salient hypo- the 2n 5 14 karyotype are indicated (chromosomes
conids, but the hypoconid is lingual to the are numbered from small to large within each
protoconid in Murexia and Sminthopsis. karyotype): acr1 + acr5 met1, acr2 + acr8
Character 121: Entoconid large and well developed on met2, acr3 + acr10 met3, and acr6 + acr9
m1m3 (0); or very small or indistinct (1). Scoring met4 (Svartman and Vianna-Morgante, 1999). The
criteria and information about the distribution of 2n 5 18 karyotype is intermediate because it has
these traits among didelphid terminals were pro- only two metacentric autosomes, of which the
vided by Voss and Jansa (2003: character 66). All larger can be homologized with acr2 + acr8 of the
examined ougroup taxa exhibit state 0. 2n 5 22 karyotype and with met2 of the 2n 5 14
TABLE 18
Exemplar Didelphid Karyotypes
Taxon 2na References
Caluromys lanatus 14 Reig et al. (1977), Palma and Yates (1996), Patton et al. (2000)
Caluromys philander 14 Reig et al. (1977)
Chironectes minimus 22 Reig et al. (1977), Palma and Yates (1996)
Cryptonanus chacoensisb 14 Wainberg et al. (1979)
Didelphis albiventris 22 Reig et al. (1977), Palma and Yates (1996)
Didelphis marsupialis 22 Gardner (1973), Reig et al. (1977), Palma and Yates (1996), Patton et al. (2000)
Didelphis virginiana 22 Gardner (1973), Reig et al. (1977)
Glironia venusta 18 Fantin and da Silva (in press)
Gracilinanus agilis 14 Carvalho et al. (2002)
Gracilinanus microtarsus 14 Wainberg et al. (1979), Carvalho et al. (2002)
Lestodelphys halli 14 Birney et al. (1996)
Lutreolina crassicaudata 22 Reig et al. (1977), Palma and Yates (1996)
Marmosa (Micoureus) demeraraec 14 Reig et al. (1977), Palma and Yates (1996), Patton et al (2000)
Marmosa (Marmosa) mexicana 14 Biggers et al. (1965)
Marmosa (Marmosa) murina 14 Reig et al. (1977)
Marmosa (Micoureus) paraguayanad 14 Carvalho et al. (2002)
Marmosa (Micoureus) regina 14 Patton et al. (2000)
Marmosa (Marmosa) robinsoni 14 Reig et al. (1977)
Marmosops impavidus 14 Patton et al. (2000)
Marmosops incanus 14 Carvalho et al. (2002)
Marmosops noctivagus 14 Palma and Yates (1996), Patton et al. (2000)
Metachirus nudicaudatus 14 Reig et al. (1977), Palma and Yates (1996), Patton et al. (2000)
Monodelphis brevicaudatae 18 Carvalho et al. (2002)
Monodelphis emiliae 18 Patton et al. (2000)
Philander frenatus 22 Carvalho et al. (2002)
Philander mcilhennyi 22 Reig et al. (1977), Patton et al. (2000)
Philander opossum 22 Reig et al. (1977)
Thylamys macrurus 14 Palma (1995)
Thylamys pallidior 14 Palma and Yates (1996)
Thylamys pusillus 14 Palma and Yates (1998)
Thylamys venustusf 14 Palma and Yates (1996)
Tlacuatzin canescens 22 Engstrom and Gardner (1988)
a
Diploid number.
b
Identified as Marmosa agilis chacoensis by Wainberg et al. (1979).
c
Identified as Marmosa cinerea by Reig et al. (1977), and as Micoureus cinereus by Palma and Yates (1996).
d
The specimens of Micoureus demerarae that Carvalho et al. (2002) karyotyped from Rio Grande do Sul are
referable to this taxon.
e
Some karyotyped specimens previously identified in the literature as Monodelphis brevicaudata (sensu lato) may
represent other species. Based on known geographic ranges (Voss et al., 2001), the Venezuelan karyotypes described by
Reig et al. (1977) were probably obtained from M. palliolata, whereas the Bolivian karyotypes described by Palma and
Yates (1996) were probably obtained from M. glirina.
f
Karyotyped Bolivian material identified as Thylamys elegans by Palma and Yates (1996) was reidentified as T.
venustus by Palma and Yates (1998).
karyotype, whereas the smaller metacentric can be acr8, met2}which can exist in two states: fused
homologized with acr6 + acr9 of the 2n 5 22 (metacentric) or fissioned (acrocentric).
karyotype and with met4 of the 2n 5 14 karyotype Character 127: Robertsonian equivalents {acr2 + acr8,
(op. cit.). As explained by Voss and Jansa (2003: met2} present as a single metacentric chromosome
4244; see also Dobigny et al., 2004), these (0); or as two acrocentric chromosomes (1).
transformational homologies are appropriately Character 128: Robertsonian equivalents {acr3 + acr10,
coded as separate characters, one for each set of met3} present as a single metacentric chromosome
Robertsonian equivalentsfor example, {acr2 + (0); or as two acrocentric chromosomes (1).
Character 129: Robertsonian equivalents {acr6 + acr9, Gracilinanus aceramarcae: 00000 10010 00100 01000
met4} present as a single metacentric chromosome 02010 0-010 02000 10001 00000 01010 00-00 00111
(0); or as two acrocentric chromosomes (1). 00101 10010 01100 00000 00011 00000 00000 11?00
10020 00?00 0000? 00000 01001 ????
Gracilinanus agilis: 00000 10010 00100 01000 02010
APPENDIX 4 ????0 02000 10001 00000 01010 00-00 00111 00101
10010 01100 00000 00011 00000 00000 11000 10020
Nonmolecular Data Matrix 00?00 00001 00000 01001 0000
Gracilinanus emiliae: 00000 10010 00100 01000 02010
The nonmolecular data analyzed in this 0-010 02010 10001 00000 01010 00-00 00111 00101
report is reproduced below. Missing data 10010 01100 00000 00011 00000 00010 11000 10020
are indicated as ? and inapplicable 00000 00001 00000 01001 ????
characters are indicated as -. Parentheses Gracilinanus microtarsus: 00000 10010 00100 01000
enclose observed polymorphisms. An elec- 02010 ????0 02000 10001 00000 01010 01000 00111
00101 10010 01100 00000 00011 00000 00000 11?00
tronic version of these data (in Nexus 10020 00000 0000? 00000 01001 0000
format) can be downloaded from Morpho- Hyladelphys kalinowskii: 00000 10000 00000 01000
Bank (http://morphobank.geongrid.org) 02010 0-000 02010 10000 00000 01000 00-00 00011
with accession number X600. 00100 00010 01000 00000 0001(01) 00001 00000
00200 10010 00000 00002 00000 01001 ????
Caluromys lanatus: 00001 10000 00000 02000 02010 Lestodelphys halli: 01000 10016 00000 00001 00011 0-
11000 00021 10001 00000 01010 01000 01011 000-0 110 02000 0?010 00000 01010 00-00 01(01)11 00101
00000 00010 10100 00011 10(01)01 00101 00000 00110 01100 10000 10011 00000 00000 21000 10021
00010 00000 00000 00000 01001 0000 00000 00021 00001 01001 0000
Caluromys philander: 00001 10000 00000 02000 02010 Lutreolina crassicaudata: 01000 0-000 00000 00000
10010 01020 10001 00000 01010 01000 01011 000-0 00010 12010 01120 01000 10000 01010 11010 02011
00000 00010 10100 00011 10001 00101 00000 00010 10101 00010 01200 00100 00121 00001 00000 21000
00000 00000 00000 01001 0000 10011 10210 00000 00001 01001 1111
Caluromysiops irrupta: 00000 0-002 00000 02000 Marmosa (Marmosa) lepida: 00000 10000 00100
02010 1?0?0 00021 10000 00000 010(01)0 01011 01000 02010 0-010 02020 10001 00000 01010
02011 000-0 00000 00210 10100 10011 10001 00101 01000 00011 00100 00010 01000 00000 00011
00000 00000 00000 00000 00000 01011 ???? 00000 00010 11000 10020 00000 00001 00000
Chacodelphys formosa: ??000 10010 00??0 000?1 02010 01001 ????
????0 ?2000 00000 00100 01010 00-00 00111 00101 Marmosa (Marmosa) mexicana: 00000 10010 00110
10010 01000 00000 00011 00000 00000 01?00 10021 01000 02010 0-010 02010 10001 00000 01010 01000
00?00 0000? 00001 11001 ???? 00011 00101 00010 01000 00000 00011 00000 00000
Chironectes minimus: 01000 11001 00200 001-2 02110 11000 10020 00000 00001 00000 01001 0000
12010 11120 00000 10000 01000 01010 02011 10100 Marmosa (Marmosa) murina: 00000 10000 00000
00010 01200 00100 00020 00000 00000 21000 10011 01000 02010 0-010 02020 10001 00000 01010 01000
10210 00000 00000 01001 1111 00011 00100 00010 01000 00000 00011 00000 00000
Cryptonanus chacoensis: 00000 10010 00100 01000 11000 10020 00000 00001 00000 01001 0000
02010 0-010 02000 10000 00000 01010 00-00 00111 Marmosa (Marmosa) robinsoni: 00000 10010 00110
00101 00010 01000 00000 00011 00000 00010 21000 01000 02010 0-010 02010 10001 00000 01010 01000
10021 00000 00000 00000 01001 0000 00011 00100 00010 01000 00000 00011 00000 00000
Cryptonanus unduaviensis: 00000 10010 00100 01000 11000 10020 00000 00001 00000 01001 0000
02010 ????0 02000 10000 00000 01010 00-00 00111 Marmosa (Marmosa) rubra: ??000 10000 00110 01000
00101 00010 01000 00000 00011 00000 00020 21000 02010 0-0?0 ?2020 10001 00000 01010 00-00 00011
10020 00000 00001 00000 01001 ???? 00100 00010 01000 10000 00011 00000 00000 11000
Didelphis albiventris: 01000 10000 11000 00000 01010 10020 00000 00001 00000 01001 ????
11010 01120 10000 10000 01010 01010 02011 10101 Marmosa (Micoureus) demerarae: 00000 10000 00110
00010 01200 00100 00121 00001 00000 21000 10011 01000 02010 0-010 02020 10001 00000 01010 01000
10210 00000 00000 01001 1111 00011 00100 00010 01000 00000 00011 00000 00000
Didelphis marsupialis: 01000 10000 11000 00000 02010 11000 10020 00000 00001 00000 01001 0000
11010 01120 10000 10000 01010 01010 02011 10101 Marmosa (Micoureus) paraguayana: 00000 10000
00010 01200 00100 00121 00001 00000 21000 10011 00110 01000 02010 0-0?0 01020 10001 00000
10210 00000 00000 01001 1111 01010 01000 00011 00100 00010 01000 00000
Didelphis virginiana: 01000 10000 11000 00000 01010 00011 00000 00000 11?00 10020 00?00 0000?
11010 01120 10000 10000 01000 01010 02011 10101 00000 01001 0000
00010 01200 00100 00121 00001 00000 21000 10011 Marmosa (Micoureus) regina: 00000 10000 00110
10210 00000 00000 01001 1111 01000 02010 0-010 02020 10001 00000 01010 01000
Glironia venusta: 00000 100?0 00??0 01000 02010 0-000 00011 00100 00010 01000 00000 00011 00000 00000
000-1 1??00 00000 01010 01100 00011 00(01)00 11000 10020 00000 00001 00000 01001 0000
00000 01000 00000 00011 00001 00000 10?00 10010 Marmosops impavidus: 00000 10000 00100 00000
00000 0000? 00000 01001 0101 02010 0-010 02020 10101 00000 01010 00-00
00111 00101 00010 01100 00000 00011 00000 00000 10110 01100 10000 00011 00000 00000 21000 10021
11000 10021 00000 00000 00000 01001 0000 00000 00011 00000 01001 0000
Marmosops incanus: 00000 10010 00100 00000 02010 Tlacuatzin canescens: 00000 10000 00100 01000 02010
0-110 02020 1010? 00100 01010 00-00 00(01)11 0-010 02000 10000 00000 01010 01000 00011 00100
00101 00010 01100 00000 00011 00000 00000 11000 10010 01000 00000 00011 00000 00000 11000 10020
10021 00000 0001? 00000 01001 0000 00000 00001 00000 01001 1111
Marmosops noctivagus: 00000 10010 00100 00000 Caenolestes fuliginosus: 1-110 0-000 00001 10010
02010 0-010 02020 10101 00000 01010 00-00 11000 0-000 02000 00001 0001? 00000 00-11 000??
00(01)11 00101 00010 01100 00000 00011 00000 ?0100 00001 01000 00000 10011 01102 1-0(03)0
00000 11000 10021 00000 00000 00000 01001 0000 21?11 -1--- 01211 ???22 01110 01101 0000
Marmosops parvidens: 00000 10000 00100 00000 Rhyncholestes raphanurus: 1-110 0-000 00001 10010
02010 0-010 02020 10101 00000 01010 00-00 11000 0-010 02000 0?010 0001? 00000 00-11 000??
00111 00100 00010 01100 00000 00011 00000 ?0110 00001 01000 00000 10011 00(01)02 1-0(02)0
00020 11000 10020 00000 00000 00000 01001 ???? 21?11 -1--- 01211 ???2? 01110 01101 0000
Marmosops pinheiroi: 00000 10000 00100 00000 02010 Murexia longicaudata: 01000 0-010 00000 00001
0-010 02020 10101 00000 01010 00-00 00111 00100 11000 0-000 02000 0?000 20000 01000 00-0(01)
00010 01100 00000 00011 00000 00020 11000 10021 01010 00100 01001 01010 21000 10011 00000 1-100
00000 00000 00000 01001 ???? 21200 10020 00?00 01122 00001 01000 0000
Metachirus nudicaudatus: 01000 11110 00000 00000 Sminthopsis crassicaudata: 01000 10010 00000 00021
02010 0-010 02010 00000 00000 01010 00-00 10011 11001 13000 02000 0?010 20000 01000 00-00 1000-
00100 00010 01100 00000 00121 00000 00000 11000 00101 01001 01010 21000 10011 00000 1-100 21100
10021 00100 00000 00000 01001 0000 10021 00000 01112 00001 00-00 0000
Monodelphis brevicaudata: 01000 0-014 00000 00000 Dromiciops gliroides: 01000 10001 00000 01000 02010
00010 0-010 00000 00000 00001 01010 10-00 01011 10001 00000 11011 00000 01011 00-00 10010 00100
(01)0100 0(01)010 01000 00000 00011 00000 00000 00011 11010 20011 00000 00000 00100 21000 10000
21000 10021 00100 00010 00001 11001 0101 00200 10020 00000 01001 0000
Monodelphis emiliae: 01000 0-015 00000 00000 00010 Echymipera kalubu: 1-000 11000 02002 201-1 12220
0-0?0 00000 00000 00001 01010 10-00 01011 12000 02000 0?000 11101 10000 00-01 1000- 00100
(01)0100 00010 01000 10000 (01)0011 00000 00001 01200 00000 11011 00013 1-130 21200 10130
00000 21000 10021 00?00 00020 00001 11001 0101 01010 01122 10000 01001 0000
Monodelphis peruviana: 01000 0-010 00000 00000 Perameles gunnii: 1-000 0-007 02002 201-1 12220
00010 0-0?0 02000 00000 00001 01010 10-00 12000 02000 0?001 01101 10100 00-01 1000- 01111
00011 10100 00010 01000 00000 10011 00000 00001 01211 01000 11012 00013 01100 21200 1113-
00000 21?00 10021 00000 0001? 00001 11001 ???? 01?10 01112 10000 00-01 0000
Monodelphis theresa: 01000 0-003 00000 00000 00010
????0 02000 00000 00001 01010 10-00 00011 00100
00010 01200 00000 10011 00000 00000 21?00 10021 APPENDIX 5
00?00 0001? 00001 11001 ????
Philander frenatus: 01000 1(01)100 00000 00000 02010 Nonmolecular Apomorphy List
11010 01120 11000 10000 010(01)0 01010 02011
10101 00010 01200 00100 00121 00001 00000 21?00 Here we list all of the unambiguous
10011 10?10 0000? 00000 01001 1111 changes (those identically located on the
Philander mcilhennyi: 01000 11100 00000 00000 02010 tree under both ACCTRAN and DEL-
11010 01120 11000 10000 01010 01010 02011 10101
00010 01200 00100 00121 00001 00000 21000 10011
TRAN optimizations; Swofford and Mad-
10210 00000 00000 01001 1111 dison, 1987) reconstructed by parsimony
Philander opossum: 01000 11100 00000 00000 02010 on our best-supported ingroup topology
110?0 01120 11000 10000 01010 01010 02011 10101 (without Chacodelphys, fig. 35). In effect,
00010 01200 00100 00121 00001 00000 21000 10011 these morphological and karyotypic trans-
10210 00000 00000 01001 1111
Thylamys macrurus: 00000 10016 00000 00001 02011 formations summarize phenotypic charac-
0-1?0 02000 10010 00100 01010 00-00 00111 00101 ter support for our classification, although
10110 01100 10000 00011 00000 00000 21?00 10021 it should be noted that we do not formally
00000 0001? 00000 01001 0000 recognize all of the nodes in this tree as
Thylamys pallidior: 01000 10016 00000 00001 02011
????0 02000 10010 00100 01010 00-00 00111 00101
taxa.
00110 01100 10000 00011 00000 00000 21000 10021 Internal Branches
00000 00011 00000 01001 0000
Thylamys pusillus: 00000 10016 00000 00001 02011 0- Caluromyinae + Glironia
110 02000 10010 00100 01010 00-00 00111 00101 Character 32. Body pelage on tail: noneRlonger
10110 01100 10000 00011 00000 00000 21000 10021 dorsally than ventrally.
00000 0001(01) 00000 01001 0000 Character 35. Base of tail: not prehensileRprehensile.
Thylamys venustus: 00000 10016 00000 00001 02011 Character 52. Postorbital process 1: absentRpresent.
????0 02000 10010 00100 01010 00-00 00111 00101 Caluromyinae
Character 17. Manual digits III and IV: III 5 IVRIV Character 109. Lower incisors: with lingual cuspR
longest. without lingual cusp.
Character 26. Pouch 1: absentRpresent. Character 126. RES {a1+ a5, m1}: metacentricRa-
Character 57. Sagittal crest: absentRsmall. crocentric.
Character 72. Transverse canal foramen: presentRab- Character 127. RES {a2+a8, m2}: metacentricRacro-
sent. centric.
Character 74. Alisphenoid tympanic process: no Character 128. RES {a3+a10, m3}: metacentricRa-
contact with RTPRcontacts RTP. crocentric.
Character 76. RTP-CTP contact: absent (gap)Rsu- Character 129. RES {a6+a9, m4}: metacentricRacro-
tured. centric.
Character 78. Ectotympanic suspension: from skullR Didelphis + Philander + Lutreolina
from malleus only. Character 65. Palatine fenestrae: absentRpresent.
Character 86. Angular process: acute-strongly inflec- Character 90. Shape I2-I4 crowns: symmetrical
tedRobtuse-weakly inflected. rhomboidsRasymmetrical.
Character 93. C1 alveolus: in premaxillary-maxillary Didelphis + Philander
sutureRin maxillary only. Character 27. Pouch: opens posteriorlyRopens ante-
Character 95. P1 size: largeRvestigial or absent. riorly.
Character 96. Height P2, P3: subequalRP2 taller. Character 36. Tail tip: not prehensileRprehensile.
Character 101. Ectoflexus on M1M3: presentRab- Didelphis
sent. Character 7. Mask: like MetachirusRlike Marmosa.
Caluromys Character 11. Dorsal underfur: darkRwhite.
Character 5. Midrostral stripe: absentRpresent. Character 12. Dorsal guard hairs: shortRlong.
Character 40. Rostral process of premaxillae: ab- Didelphis albiventris + D. marsupialis
sentRpresent. No unambiguous changes.
Hyladelphinae + Didelphinae Philander
Character 69. Posterolateral palatal corners: absent or Character 8. Spot above eye: absentRpresent.
indistinctRpresent. Philander mcilhennyi + Philander opossum
Didelphinae No unambiguous changes.
Character 29. Unpaired median teat: absentRpresent. Thylamyini
Character 104. Centrocrista: weakly -shapedR Character 58. Fenestra in parietal-squamosal suture:
strongly -shaped. absentRpresent.
Didelphini + Metachirini + Thylamyini Cryptonanus + Gracilinanus + Lestodelphys + Thyla-
Character 73. Secondary foramen ovale: absentR mys
formed by anteromedial strut. Character 34. Caudal scales: intermediateRannular.
Didelphini + Metachirini Cryptonanus + Gracilinanus
Character 2. Rhinarium: with 2 groovesRwith 1 groove. Character 17. Manual digits III and IV: III long-
Character 7. Mask: like MarmosaRlike Metachirus. estRIII 5 IV.
Character 36. Tail tip: prehensileRnot prehensile. Cryptonanus
Character 84. Paroccipital process: small, adnate to Character 73. Secondary foramen ovale: formed by
petrosalRlarge and erect. anteromedial strutRabsent.
Character 108. Upper molar eruption: P3 lastRM4 Gracilinanus
and P3 together. Character 40. Rostral process of premaxillae: ab-
Didelphini sentRpresent.
Character 26. Pouch: absentRpresent. Character 66. Maxillary fenestrae: absentRpresent.
Character 32. Body pelage on tail: noneRsubequal. Gracilinanus aceramarcae + G. microtarsus + G. agilis
Character 33. Caudal coloration: uniform or varia- No unambiguous changes.
bleRblack-and-white. Gracilinanus aceramarcae + G. microtarsus
Character 34. Caudal scales: intermediateRspiral. No unambiguous changes.
Character 41. Nasal tips: above or anterior to Lestodelphys + Thylamys
I1Rposterior to I1. Character 10. Dorsal fur markings: unpatternedRtri-
Character 52. Postorbital processes: absentRpresent. colored.
Character 54. Frontals: separateRfused. Character 20. Central palmar surface: smoothRden-
Character 57. Sagittal crest: absentRlarge. sely tubercular.
Character 61. Parietal-mastoid contact: presentRab- Character 25. Ventral tarsal surface: nakedRfurred.
sent. Character 28. Mammary position: abdominal-ingui-
Character 73. Secondary foramen ovale: formed by nalRpectoral.
anteromedial strutRformed by lamina. Character 39. Tail incrassate: noRyes.
Character 78. Ectotympanic suspension: from skullR Character 68. Posterolateral palatal foramina: behind
from malleus only. M4 protoconesRlingual to M4 protocones.
Character 96. Height P2, P3: subequalRP3 taller. Character 76. RTP-CTP contact: absent (gap)Rsu-
Character 104. Centrocrista: strongly -shapedR tured.
weakly -shaped. Character 114. Height p2, p3: p2 tallerRsubequal.
Character 106. Postprotocrista: unnotchedRnotched. Thylamys
Character 108. Upper molar eruption: M4 and P3 Character 43. Nasal width: wider posteriorlyR
togetherRM4 last. uniformly narrow.
Thylamys macrurus + T. pusillus + T. pallidior (branch 22) Caluromys philander:

No unambiguous changes Character 29. Unpaired median teat: absentRpresent.
Thylamys macrurus + T. pusillus Character 32. Body pelage on tail: longer dorsallyR
No unambiguous changes subequal.
Marmosops Character 35. Base of tail: with ventral tuberclesR
Character 34. Caudal scales: intermediateRspiral. smooth ventrally.
Character 38. Caudal hairs: all hairs alikeRmiddle Caluromysiops irrupta:
hair thickest. Character 6. Mask: presentRabsent.
Character 40. Rostral process of premaxillae: ab- Character 10. Dorsal fur markings: unpatternedR
sentRpresent. scapular stripes.
Marmosops impavidus + M. noctivagus + M. incanus Character 54. Frontals: separateRfused.
No unambiguous changes. Character 55. Parietals: separateRfused.
Marmosops impavidus + M. noctivagus Character 57. Sagittal crest: smallRlarge.
No unambiguous changes. Character 73. Secondary foramen ovale: absentR
Marmosops parvidens + M. pinheiroi formed by lamina.
Character 94. Upper canine: simpleRanterior and Character 81. Stapes: bicrurateRcolumelliform.
posterior accessory cusps present. Character 104. Centrocrista: weakly -shapedR
Marmosini linear.
No unambiguous changes. Character 124. Hypoconulid: twinned with entoco-
Marmosa + Tlacuatzin nidRin midline.
Character 13. Lateral carpal tubercles: absentR Glironia venusta:
present in adult males. Character 53. Postorbital processes: frontal onlyR
Character 52. Postorbital process: absentRpresent. frontal and parietal.
Marmosa Character 127. RES {a2+a8, m2}: metacentricRacro-
Character 40. Rostral process of premaxillae: ab- centric.
sentRpresent. Character 129. RES {a6+a9, m4}: metacentricRacro-
Marmosa lepida + M. murina + subgenus Micoureus centric.
No unambiguous changes. Chironectes minimus:
Marmosa murina + subgenus Micoureus Character 10. Dorsal fur markings: unpatternedRlike
No unambiguous changes. Chironectes.
Subgenus Micoureus Character 13. Lateral carpal tubercles: absentR
Character 14. Medial carpal tubercles: absentR present in both sexes.
present in adult males. Character 18. Manual plantar pads: presentRabsent.
Marmosa demerarae + M. paraguayana Character 20. Central palmar surface: smoothRsand-
No unambiguous changes. papery.
Marmosa mexicana + M. robinsoni + M. rubra Character 23. Pedal morphology: eleuthrodacty-
Character 14. Medial carpal tubercles: absentR lousRwebbed.
present in adult males. Character 31. Cloaca: presentRabsent.
Marmosa mexicana + M. robinsoni Character 49. Lacrimal foramina: twoRone.
Character 9. Gular gland: absentRpresent. Character 85. Mental foramina: twoRone.
Monodelphis Didelphis albiventris:
Character 2. Rhinarium: with 2 groovesRwith 1 groove. No unambiguous changes.
Character 6. Mask: presentRabsent. Didelphis marsupialis:
Character 22. Pedal digits III and IV: IV longestRIII No unambiguous changes.
longest. Didelphis virginiana:
Character 36. Tail tip: prehensileRnot prehensile. Character 49. Lacrimal foramina: twoRone.
Character 45. Maxilloturbinals: elaborately bran- Philander frenatus:
chedRsimple scrolls. No unambiguous changes.
Character 51. Maxillary and alisphenoid: separateRin Philander mcilhennyi:
contact. No unambiguous changes.
Character 96. Height P2, P3: subequalRP3 taller. Philander opossum:
Character 114. Height p2, p3: p2 tallerRsubequal. No unambiguous changes.
Character 120. m3 hypoconid position: labially Lutreolina crassicaudata:
salientRlingual to protoconid. Character 6. Mask: presentRabsent.
Character 121. Entoconid size: taller than hypoconu- Character 22. Pedal digits III and IV: subequal or IV
lidRsmall or indistinct. longestRIII longest.
Monodelphis brevicaudata + M. theresa + M. emiliae Character 51. Maxillary and alisphenoid: separateRin
No unambiguous changes. contact.
Monodelphis brevicaudata + M. theresa Character 120. m3 hypoconid position: labially
No unambiguous changes. salientRlingual to protoconid.
Metachirus nudicaudatus:
Terminal Branches Character 8. Spot above eye: absentRpresent.
Character 56. Parietal and alisphenoid: in con-
Caluromys lanatus: tactRseparated.
No unambiguous changes. Cryptonanus chacoensis:
Character 115. dp3 trigonid: uni- or bicuspidRtricus- Character 76. RTP-CTP contact: absent (gap)Rsu-
pid. tured.
Cryptonanus unduaviensis: Tlacuatzin canescens:
No unambiguous changes. Character 66. Maxillary fenestrae: absentRpresent.
Gracilinanus aceramarcae: Character 126. RES {a1+ a5, m1}: metacentricRa-
No unambiguous changes. crocentric.
Gracilinanus microtarsus: Character 128. RES {a3+a10, m3}: metacentricRa-
Character 52. Postorbital process: absentRpresent. crocentric.
Gracilinanus agilis: Monodelphis brevicaudata:
No unambiguous changes. No unambiguous changes.
Gracilinanus emiliae: Monodelphis theresa:
Character 34. Caudal scales: annularRintermediate. Character 9. Gular gland: presentRabsent.
Lestodelphys halli: Character 73. Secondary foramen ovale: absentR
Character 2. Rhinarium: with 2 groovesRwith 1 formed by lamina.
groove. Monodelphis emiliae:
Character 22. Pedal digits III and IV: IV longestRIII Character 76. RTP-CTP contact: absent (gap)Rsu-
longest. tured.
Character 36. Tail tip: prehensileRnot prehensile. Character 114. Height p2, p3: subequalRp3 taller.
Character 57. Sagittal crest: absentRsmall. Monodelphis peruviana:
Character 81. Stapes: bicrurateRcolumelliform. Character 61. Parietal-mastoid contact: presentRab-
Character 114. Height p2, p3: subequalRp3 taller. sent
Character 120. m3 hypoconid position: labially Hyladelphys kalinowskii:
salientRlingual to protoconid. Character 96. Height P2, P3: subequalRP2 taller.
Thylamys macrurus: Character 98. Upper milk premolar: large & molar-
No unambiguous changes. iformRvestigial.
Thylamys pusillus:
No unambiguous changes.
Thylamys pallidior: APPENDIX 6
Character 2. Rhinarium: with 2 groovesRwith 1 groove.
Thylamys venustus: Key to the Didelphid Genera
No unambiguous changes.
Marmosops impavidus:
This key is intended for use on adult
No unambiguous changes. specimens (with completely erupted per-
Marmosops noctivagus: manent dentitions). Juveniles (with dP3/
No unambiguous changes. dp3) and subadults (lacking dP3/dp3, but
Marmosops incanus: with incompletely erupted third premolars
Character 28. Mammary position: abdominal-ingui-
nalRpectoral.
and/or fourth molars) may not key out
Character 43. Nasal width: wider posteriorlyR accurately if they lack traits seen in
uniformly narrow. conspecific adults (e.g., large size, postor-
Character 114. Height p2, p3: p2 tallerRsubequal. bital processes, secondary foramina ovales,
Marmosops parvidens: contact between left and right exoccipitals;
Character 105. Preprotocrista: stops at para-
coneRmeets anterolabial cingulum.
Abdala et al., 2001). Although the follow-
Marmosops pinheiroi: ing couplets are based, insofar as possible,
No unambiguous changes. on characters with unambiguously distin-
Marmosa lepida: guishable alternative states, key results
Character 94. Upper canine: simpleRposterior acces- should always be checked by comparing
sory cusp present.
Marmosa murina: specimens character-by-character against
Character 13. Lateral carpal tubercles: present in generic descriptions and cranial illustra-
adult malesRabsent. tions (figs. 3754). Couplet 10, the only
Marmosa demerarae: step in our key based on a single character,
No unambiguous changes. is obviously problematic if an atypical
Marmosa paraguayana:
Character 32. Body pelage on tail: noneRsubequal. specimen is in hand. Users would be well
Marmosa regina: advised, if an unsatisfactory result is
No unambiguous changes. obtained at any subsequent step, to follow
Marmosa mexicana: the alternative sequence on the hypothesis
Character 65. Palatine fenestrae: absentRpresent. that the species in question is polymorphic.
Marmosa robinsoni:
No unambiguous changes. External and craniodental measurements
Marmosa rubra: used to index size in this key are defined
Character 52. Postorbital process: presentRabsent. and abbreviated as follows: Head-and-
Body Length (HBL) is obtained by sub- cess spanning transverse canal foramen;
tracting length of tail from total length frontal contacts squamosal (no alisphenoid-
(Hall, 1981); Condylo-Basal Length (CBL) parietal contact) . . . . . . . . . . . Metachirus
is measured from the occipital condyles to 4b. Pale supraocular spots present or absent;
the anteriormost point of the premaxillae naked skin of tail blackish basally, becoming
abruptly whitish distally (occasional speci-
(Voss et al., 2004b: fig. 2); and Length of
mens have all-black tails); postorbital pro-
Molars (LM) is measured from the ante- cesses usually present; right and left frontals
riormost labial margin of M1 to the fused (median suture incomplete); sagittal
posteriormost point on M4 (Voss et al., crest well developed, extending to frontals;
2004b: fig. 2). palatine fenestrae present; secondary foramen
ovale usually present, but formed by medial
1a. Plantar surface of manus without distinct lamina that does not span the transverse
plantar pads; finger-like fleshy process on canal foramen; alisphenoid contacts parietal
lateral surface of wrist (resembling a sixth (no frontal-squamosal contact) . . . . . . . . 5
finger) present in both sexes; hind foot 5a. Pale supraocular spots present . . . Philander
conspicuously webbed; dorsal pelage gray 5b. Pale supraocular spots absent . . . . . . . . . 6
boldly marked by transverse bars of black 6a. Dorsal pelage grayish or blackish (never
fur; single lacrimal foramen usually present brownish); dorsal underfur white; dorsal guard
on each side; each hemimandible usually with hairs very long (pelage has a shaggy appear-
a single mental foramen . . . . . . Chironectes ance); maxillary and alisphenoid separated by
1b. Plantar surface of manus with distinct plantar palatine on orbital floor. . . . . . . . . Didelphis
pads; no finger-like fleshy process on wrist; 6b. Dorsal pelage brownish or yellowish-brown;
hind foot not webbed; dorsal pelage variously
dorsal underfur gray; dorsal guard hairs
pigmented, sometimes with contrasting lon-
short; maxillary and alisphenoid in contact
gitudinal stripes, but never boldly marked
on orbital floor . . . . . . . . . . . . . Lutreolina
with transverse blackish bars on a gray
7a. Manual digit IV slightly longer than digit III;
background; two lacrimal foramina usually
postorbital processes well developed; palate
present on each side; each hemimandible
not perforated by any large fenestrae; trans-
usually with two mental foramina. . . . . . 2
verse canal foramen absent; alisphenoid
2a. Dorsal surface of tail covered by long, soft fur
tympanic process contacts or closely approx-
from base to tip; postorbital processes pre-
imates rostral tympanic process of petrosal;
sent, formed by frontals and parietals . . . .
ectotympanic annulus suspended from malle-
. . . . . . . . . . . . . . . . . . . . . . . . . . . . Glironia
us; C1 alveolus contained in maxilla; P1
2b. Dorsal surface of tail macroscopically naked
vestigial or absent; width of M4 less than
for at least one or two centimeters from tip;
width of M1 . . . . . . . . . . . . . . . . . . . . . 8
postorbital processes absent or, if present,
formed only by frontals . . . . . . . . . . . . . 3 7b. Length of manual digit IV subequal to or less
than length of digit III; postorbital processes
3a. Large (HBL .245 mm, weight .260 g, CBL
.55 mm, LM .11 mm); with or without present or absent; palate perforated by at
pale supraocular spots; manual digit III least one set of paired fenestrae; transverse
longer than manual digit IV; dorsal margin canal foramen present (sometimes small and
of foramen magnum formed by right and left hard to see on incompletely cleaned skulls);
exoccipitals only (incisura occipitalis ab- alisphenoid tympanic process and rostral
sent). . . . . . . . . . . . . . . . . . . . . . . . . . . 4 tympanic process of petrosal separated by
3b. Size small or large; never with pale suprao- wide gap; ectotympanic suspended directly
cular spots; manual digit III longer than, from basicranium; P1 smaller than posterior
subequal to, or shorter than digit IV; dorsal premolars, but well-formed and nonvestigial;
margin of foramen magnum formed by C1 alveolus in premaxillary-maxillary suture;
supraoccipital in addition to exoccipitals width of M4 greater than width of M1. . . 9
(incisura occipitalis present) . . . . . . . . . . 7 8a. Circumocular mask and dark midrostral
4a. Pale supraocular spots present; naked skin of stripe present; dorsal body pelage without
tail brownish or grayish basally, fading to dark scapular markings; premaxillary rostral
whitish distally; postorbital processes absent; process present; right and left frontals sepa-
right and left frontals separated by persistent rated by persistent median suture; sagittal
median suture; sagittal crest absent; palatine crest absent (or weakly developed over
fenestrae absent; secondary foramen ovale interparietal); secondary foramen ovale ab-
present, formed by anteromedial bullar pro- sent; centrocrista weakly inflected labially;
hypoconulid twinned with entoconid. . . . . fenestrae absent; fenestra cochleae exposed;

. . . . . . . . . . . . . . . . . . . . . . . . Caluromys P2 and P3 subequal in height. . . Marmosops
8b. Pelage of head uniformly pale, without any 12b. Tail longer or shorter than head-and-body,
dark markings; dark scapular stripes present; with scales (when distinct) in annular series,
premaxillary rostral process absent; right and incrassate (at least seasonally); premaxillary
left frontals fused (median suture incom- rostral process absent; maxillary fenestrae
plete); sagittal crest well developed, extending present or absent; fenestra cochleae concealed
anteriorly to frontals; secondary foramen in a sinus formed by rostral and caudal
ovale present; centrocrista linear, uninflected; tympanic processes of petrosal; P3 taller than
hypoconulid not twinned with entoco- P2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
nid. . . . . . . . . . . . . . . . . . . Caluromysiops 13a. Manual claws long, extending well beyond
9a. Circumocular mask absent; tail shorter than fleshy apical pads of digits; pedal digit III
combined length of head and body, without longer than digit IV; tail shorter than
prehensile modifications; postorbital process- combined length of head and body; ventral
es absent or indistinct; premaxillary rostral surface of tail tip not modified for prehen-
process absent; palatine and maxillary fenes- sion; nasals conspicuously wider posteriorly
trae usually absent; maxillary and alisphenoid than anteriorly; maxillary fenestrae absent;
in contact on orbital floor; secondary fora- stapes columelliform and imperforate or
men ovale absent or formed by medial bullar microperforate; p3 taller than p2; hypoconid
lamina that does not span transverse canal lingual to protocone on m3, not labially
foramen; entoconid subequal to or smaller than salient . . . . . . . . . . . . . . . . . Lestodelphys
hypoconulid on m1m3 (absent or indistinct in 13b. Manual claws short, not extending much, if
worn dentitions) . . . . . . . . . . . . Monodelphis at all, beyond fleshy apical pads of digits;
9b. Dark circumocular mask present (but narrow pedal digit IV longer than digit III; tail longer
and inconspicuous in some taxa); tail longer or shorter than combined length of head-and-
or shorter than combined length of head and body; ventral surface of tail tip modified for
body, with or without prehensile modifica- prehension in most species; nasals uniformly
tions; distinct postorbital processes present or narrow (not much wider posteriorly than
absent; premaxillary rostral process present anteriorly), with subparallel lateral margins;
or absent; palatine and maxillary fenestrae maxillary fenestrae present or absent; stapes
triangular with large obturator foramen; p2
present or absent; maxillary and alisphenoid
and p3 subequal in height; hypoconid labially
separated by palatine on orbital floor; sec-
salient on m3, not lingual to protoco-
ondary foramen ovale absent or formed by
nid. . . . . . . . . . . . . . . . . . . . . . Thylamys
anteromedial bullar process spanning trans-
14a. Manual digit III longer than manual digit IV;
verse canal foramen; entoconid taller than
tail shorter than combined length of head and
hypoconulid on m1m3 (except in Chacodel-
body; nasals narrow, with subparallel lateral
phys) . . . . . . . . . . . . . . . . . . . . . . . . . . 10
margins; maxillary fenestrae present; hypo-
10a. Secondary foramen ovale usually present,
conid not labially salient on m3 (lingual to
formed by anteromedial bullar process span-
protoconid). . . . . . . . . . . . . . Chacodelphys
ning transverse canal foramen . . . . . . . . 11
14b. Manual digits III and IV subequal in length;
10b. Secondary foramen ovale usually absent . .
tail longer than head-and-body; nasals dis-
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
tinctly wider posteriorly than anteriorly;
11a. Manual digits III and IV subequal in length; maxillary fenestrae absent (except in Tlacuat-
tail always longer than combined length of zin); hypoconid labially salient on m3 (not
head and body; premaxillary rostral process lingual to protoconid) . . . . . . . . . . . . . . 15
present; maxillary fenestrae usually present 15a. Tail scales in annular series; rostral process of
(occasionally absent in one species); P2 and premaxillae absent; palatine fenestrae pres-
P3 subequal in height . . . . . . . Gracilinanus ent; petrosal exposed on posterolateral sur-
11b. Manual digit III longer than digit IV; tail face of braincase through fenestra in parietal-
longer or shorter than head-and-body; pre- squamosal suture; unworn C1 usually with
maxillary rostral process present or absent; small accessory cusps; P3 taller than
maxillary fenestrae absent or, if present, then P2 . . . . . . . . . . . . . . . . . . . . Cryptonanus
P3 taller than P2 . . . . . . . . . . . . . . . . . . 12 15b. Tail scales in annual or spiral series; rostral
12a. Tail longer than combined length of head and process of premaxillae present or absent;
body, with scales in spiral series, never palatine fenestrae present or absent; petrosal
incrassate; premaxillary rostral process pres- not exposed through fenestra in parietal-
ent (but short in a few species); maxillary squamosal suture (fenestra absent); C1 with-
out accessory cusps; P2 and P3 subequal in subequal anterior and posterior cutting edges;
height or P2 taller than P3 . . . . . . . . . . . 16 P2 and P3 subequal in height; unworn c1
16a. Very small (HBL ,95 mm; weight ,20 g; usually procumbent and somewhat premo-
CBL ,25 mm; LM ,5 mm); mammae 20 lariform . . . . . . . . . . . . . . . . . . . . . . . . 17
254; rostral process of premaxillae absent; 17a. Caudal scales in spiral series, at least in part
postorbital processes absent; palatine fenes- (the caudal scale arrangement in some species
trae absent; unworn crowns of I2I5 asym- is partly spiral and partly annular); rostral
metrical (incisiform), with much longer ante- process of premaxillae usually present; max-
rior than posterior cutting edges; P2 much illary fenestrae absent; crown of I2 visibly
taller than P3; lower canine (c1) an erect, shorter (mesiodistal dimension) than crown
recurved-conical tooth, not premolarized . . . of I5; unworn c1 procumbent and more or
. . . . . . . . . . . . . . . . . . . . . . . . . Hyladelphys less premolariform. . . . . . . . . . . . Marmosa
16b. Larger; mammae 31357 to 717515; 17b. Caudal scales in unambiguously annular
rostral process of premaxillae present or series; rostral process of premaxillae absent;
absent; postorbital processes usually present maxillary fenestrae present (sometimes con-
(indistinct in some young adults but absent in fluent with maxillopalatine openings); crowns
adults of at least one species); palatine of I2I5 subequal in length; c1 erect and
fenestrae present or absent; unworn crowns recurved-conical, not premolarized . . . . . .
of I2I5 symmetrically rhomboidal, with . . . . . . . . . . . . . . . . . . . . . . . . Tlacuatzin

B322 PDF

Uploaded by

Copyright:

Available Formats

B322 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

B322 PDF

Uploaded by

Copyright:

Available Formats

PHYLOGENETIC RELATIONSHIPS AND

CLASSIFICATION OF DIDELPHID MARSUPIALS,

BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY

Copyright E American Museum of Natural History 2009 ISSN 0003-0090

INTRODUCTION Asia and western North America (Rougier et

TABLE 1 distributed in the Tertiary than they are

Percent complete character datab

Percent complete character datab

Primer name Sequence

olis-coupled Markov-chain Monte Carlo taxon-addition replicates with TBR branch

mation.1 In effect, such information comes comprehensive compilations of morphomet-

Nb Head and Bodyc Taild Weight

the rostrum to the frontal region. In other

The plantar pads of two Brazilian species

median teat that occupies the ventral midline

and share a continuous mucosa. In life, both TABLE 5

didelphids; in particular, no fenestra is ever

pials. However, some peramelemorphians the alisphenoid in Myrmecobius and many

other marsupials essentially resemble didel- omys, Caluromysiops, Glironia, Tlacuatzin,

parietal sutureoccur as balanced polymor- doid crest) as the sutured interparietals of

phalangerids uniquely exhibit squamosal-

This report Other names

et al., 1987), but distinct maxillopalatine species of Marmosops (e.g., M. pinheiroi;

minor palatine artery from the maxillary

posterior limb tends to be dorsoventrally certain details of didelphid mallear and

nuchal muscles (as in peramelemorphians, see Among other plesiomorphic marsupials,

Chironectes in addition to Caluromys and Most peramelemorphians and Dromiciops

notched in Rhyncholestes (see Osgood, 1924:

between P1 and P2 tends to be individually

Caluromysiops) to a maximal value of 1.76 (in descriptors of a visually compelling but

character states in different phylogenetic

Caluromys lanatus 5 2.87 3.09 3.01 2.44 0.85

Lestodelphys, Monodelphis, Tlacuatzin). In all small didelphids. Premolarization of c1 in-

tendencies are minimally indicated on the

which a diminutive crest (comprising a Summary

BRCA1 DMP1 IRBP RAG1 vWF

BRCA1 DMP1 IRBP RAG1 vWF

Interphotoreceptor Retinoid Binding Protein Recombination Activating 1 Gene

Gene Protein product

BRCA1 chr 2 many tissues nucleus & cytoplasm tumor suppression

BRCA1 vWF IRBP DMP1 RAG1a

Model GTR + C TVMef + I + C GTR + I + C GTR + C GTR + I + C

Nonmolecular dataa Combined datab

the third is Hyladelphys + Didelphinae. A nodes not convincingly resolved by BRCA1

BRCA1 vWF IRBP DMP1 RAG1a Concatenated

IRBP DMP1 RAG1 BRCA1 vWF

of our knowledge, no strongly conflicting tionary interest. Phenotypically, Cryptonanus

Type speciesa Current status (if not valid)

Didelphidae Didelphidae Didelphidae Didelphidae Didelphidae

Didelphidae Marmosidae Didelphidae Didelphidae

synapomorphies when this gene is eventually MORPHOLOGICAL DESCRIPTION: Combined

Maxillopalatine fenestrae usually absent slightly longer than postprotocristae in sub-

elsewhere in northern Argentina, eastern of manus sparsely tubercular (neither smooth

foramina present on each side, exposed to premaxillary-maxillary suture; C1 without

provided by parsimony and likelihood anal- fresh acknowledgment here. Nondidelphid

a Kalbfleisch postdoctoral research grant and Amrine-Madsen, H., M. Scally, M. Westerman,

of marsupial mammal interrelationships. Molec- bernes Selskaber Biologiske Meddelelser 1(8):

Freedman, L. 1967. Skull and tooth variation in Levantamento de mamferos do ParnaTapajos.

APPENDIX 1 127107127109, 127156, 155575, 155581. Subadult