Taxonomic revision of the genus Galictis
(Carnivora: Mustelidae): species delimitation,
morphological diagnosis, and refined mapping
of geographical distribution
RENATA BORNHOLDT1, KRISTOFER HELGEN2, KLAUS-PETER KOEPFLI3,
LARISSA OLIVEIRA4, MAURO LUCHERINI5 and EDUARDO EIZIRIK1,6*
1Faculdade de Bioci?ncias ? PUCRS, Av. Ipiranga, 6681 pr?dio 12, Porto Alegre RS, 90619-900,
Brazil
2Smithsonian Institution, Division of Mammals, National Museum of Natural History, NHB 390,
MRC 108, P.O. Box 37012, Washington DC, USA
3NIH, Laboratory of Genomic Diversity, 1050 Boyles St Bldg 560 Rm 11-33, Frederick, MD 21702,
USA
4UNISINOS, Laboratorio de Ecologia de Mamiferos, S?o Leopoldo, Brazil
5Grupo de Ecolog?a Comportamental de Mam?feros, C?tedra de Fisiolog?a Animal, Departamento de
Biolog?a, Bioqu?mica y Farmacia, Universidad Nacional del Sur, San Juan 670, 8000 Bah?a Blanca,
Argentina
6Instituto Pr?-Carn?voros, Atibaia, SP 12945-010, Brazil
Received 14 February 2012; revised 10 July 2012; accepted for publication 16 July 2012
Although critical for enabling in-depth evolutionary, ecological, or conservation-orientated studies, taxonomic
knowledge is still scarce for many groups of organisms, including mammals of the order Carnivora. For some
of these taxa, even basic aspects such as species limits and geographical distribution are still uncertain.
This is the case for the Neotropical mustelid genus Galictis, considered one of the least studied carnivoran
genera in the Americas. To address this issue, we performed a comprehensive assessment of morphological
and molecular characters to test the number of species within Galictis, and to characterize their distinc-
tiveness and evolutionary history. In addition, we reviewed and consolidated the available information on
the taxonomy of this genus, so as to provide a historical framework upon which we could interpret our
data. Our analyses demonstrated that two Galictis species can be clearly delimited and diagnosed using
metric and nonmetric morphological characters as well as DNA sequences from mitochondrial and nuclear
gene segments. On the basis of this clarified species-level delimitation, we reassessed the geographical range
of each Galictis taxon, identifying possible areas of sympatry between them. These results provide a solid
taxonomic framework for Galictis, enabling the development of additional studies focusing on this poorly
known taxon.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472.
doi: 10.1111/j.1096-3642.2012.00859.x
ADDITIONAL KEYWORDS: America ? greater grison ? lesser grison ? morphometrics ? Neotropical region;
phylogeny ? skin variation ? skull variation ? taxonomy.
*Corresponding author. E-mail: eduardo.eizirik@pucrs.br
Zoological Journal of the Linnean Society, 2013, 167, 449?472. With 8 figures
bs_bs_banner
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472 449
INTRODUCTION
Taxonomy forms the basis for all biodiversity sci-
ences, as it provides the overall framework upon
which one can describe and characterize spatial or
temporal patterns of population or community
changes. The science of taxonomy is thus a prerequi-
site for downstream studies aimed toward better
understanding and conserving biodiversity (Wilson,
1992, 2005; Reeder, Helgen & Wilson, 2007). Even
amongst some of the best-known organisms, such as
mammals, there is still much to be done in terms of
basic taxonomy (Patterson, 2000, 2001), which is
illustrated by the rate of discovery of new species
observed in the last few decades (Collen, Purvis &
Gittleman, 2004; Reeder et al., 2007). Interestingly,
the spatial pattern of mammalian species description
in the last two decades indicates that such findings
derive in many cases from a combination of poor
historical knowledge of regional faunas (relative to
their diversity) and recently increased taxonomic
efforts (Schipper et al., 2008). This implies that an
improved understanding of mammalian biology is still
positively related with taxonomic engagement, espe-
cially in megadiverse regions whose biotas remain
poorly characterized.
Amongst mammals, carnivores are supposed to be
well known, but is that really the case? Some genera
and species have received minimal study, which
means that basic knowledge such as species limits
and geographical distribution are still uncertain
(Collen et al., 2004). The paucity of such information
implies that the biology and conservation status of
these species are also unclear, which poses a serious
impediment to the design of management strategies
for a group that tends to be particularly susceptible to
anthropogenic threats (Miller et al., 1999; Valenzuela-
Galv?n, Arita & Macdonald, 2007). Curiously, much of
the recent improvement in species-level taxonomic
knowledge on carnivores derives from revisions based
on previously available museum specimens, instead of
resulting from actual field-based discoveries of new
taxa (Patterson, 2000). This is illustrated by the
recent recognition of new carnivoran species using
morphological and/or molecular data sets (Helgen,
Lim & Helgen, 2008; Helgen et al., 2009; del Cerro
et al., 2010; Goodman & Helgen, 2010) derived at
least partially from museum material. In addition to
the description of new species, other aspects of basic
carnivoran taxonomy may be clarified with thorough
revisions of museum specimens, especially in the case
of poorly known genera from understudied regions of
the world.
The Neotropical region spans a large portion of the
Americas, extending from central Mexico to southern
Argentina and Chile, and harbouring one of the
richest hotspots of mammalian diversity worldwide
(Olson et al., 2001; Schipper et al., 2008). Amongst
mammalian species, some carnivoran genera are
endemic to the Neotropics, including some groups
that have so far received very little taxonomic and
biogeographical attention. This is the case for the
grisons (genus Galictis, Carnivora, Mustelidae) for
which very little information has so far been pub-
lished, rendering it one of the least studied carnivo-
ran genera in the Americas. Not surprisingly, the
acquisition of basic information about Galictis has
already been cited as a priority for carnivore studies
in the Neotropics (Oliveira, 2009).
Currently, there are two species recognized in this
genus: Galictis vittata (greater grison) (Schreber,
1776) and Galictis cuja (lesser grison) (Molina, 1782)
(Wozencraft, 2005), but the taxonomic history of these
forms is relatively complex. The name Galictis first
appeared in 1826 in a short publication by the
English naturalist Thomas Bell (Bell, 1826). This
publication contained a brief description of the habits
of a living animal caught in Guyana that the author
himself observed for several months. Bell, however,
used the name Galictis without providing any
detailed description of the precise locality of collection
or its distinctive characteristics. Finally, in the last
lines of the publication, Bell mentioned his intention
to characterize better the animal and designate it as
the type for a new genus. Eleven years after this
report, Bell published a more detailed description of
this carnivoran taxon, in which he examined a speci-
men from the collection of the Zoological Society and
recognized essential similarities between this indi-
vidual and the one that he had observed previously,
thus officially validating the generic name Galictis
(Bell, 1837). However, prior to the establishment of
Bell?s genus, these Neotropical carnivores were
already known to science, as previous authors had
described several taxa under different generic names.
The first reference to grisons in the natural history
literature was that of the Swiss naturalist Jan Nico-
laas Allamand, who presented in 1771 a drawing of a
new mammal from Suriname, for which he coined the
name ?grison? (Buffon, 1776). The official description of
that species came some years later with the work of
Johann Christian von Schreber, who described it
based on Allamand?s drawing, and named it Viverra
vittata (Schreber, 1776). In parallel, Chilean natural-
ist Juan Ignacio Molina described two species of car-
nivores from Chile, Mustela cuja and Mustela quiqui
(Molina, 1782). The author, however, presented these
species without any specification of their type locality.
Nevertheless, it was not until the beginning of the
19th century that the species described by Schreber
(1776) and by Molina (1782) were recognized as
belonging in the same genus. Shaw (1800) combined
450 R. BORNHOLDT ET AL.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
the species described by both authors in the genus
Viverra using the names Viverra vittata, Viverra cuja,
and Viverra quiqui. This author considered Molina?s
M. cuja and M. quiqui to be synonyms of V. cuja and
V. quiqui, respectively. In addition, he considered the
name ?grison? to remain attached only to V. vittata.
Some years later, Oken (1816) described a new
genus that he named Grison, with V. vittata as the
type species. This generic name was followed by some
authors over many years (Thomas, 1907; Ihering,
1911; Goldman, 1920; Osgood, 1943; Goodwin, 1946),
but Oken?s (1816) names have subsequently been
ruled invalid by the International Commission on
Zoological Nomenclature (ICZN, 1956). Various other
generic appellations were also applied to the grisons
over the years, such as Gulo (Desmarest, 1820), Ursus
(Thunberg, 1820), and Lutra (Traill, 1821), names
formerly employed in much broader taxonomic con-
texts, as well as the names Grisonia (Gray, 1865) and
Grisonella (Thomas, 1912), which currently represent
synonyms of Galictis (see Yensen & Tarifa, 2003a, for
a recent review).
Although the validity of Galictis as a taxonomic
entity has been solidly established, much more
uncertainty has surrounded its species-level compo-
sition. When Bell (1837) officially described Galictis,
he recognized Schreber?s (1776) V. vittata as a
member of the new genus, and also described a
second species, Galictis allamandi, based on the
museum specimen that he had observed. Bell did not
include in the new genus the two Chilean species
described by Molina (1782). In a comprehensive zoo-
logical list from Chile, Claudio Gay (1847) considered
Molina?s M. quiqui to be a synonym of G. vittata, but
did not mention M. cuja. As far as we know, this is
the first synonymy list that included any of Molina?s
Chilean species in the genus Galictis. Subsequently,
some authors recognized M. cuja and M. quiqui to be
synonyms of G. vittata (Gray, 1865, 1869). A few
decades later, Thomas (1907) separated G. vittata
from the Chilean species and considered the latter
two (G. cuja and G. quiqui) to be synonyms of each
other. However, in 1912 Thomas again raised the
issue of G. cuja and G. quiqui potentially being sepa-
rate species, but regarded the issue as unresolved
(Thomas, 1912). In addition to this on-going discus-
sion with respect to the status of the two Chilean
species of Galictis, several other species were
described for this genus during the 20th century,
leading to considerable variation in the taxonomic
literature addressing this group (see Yensen &
Tarifa, 2003a, b for exhaustive synonym lists).
Overall, although some authors have continued to
recognize G. allamandi as a potentially valid species
(e.g. Eisenberg, 1989), most current sources only rec-
ognize G. vittata and G. cuja (e.g. Yensen & Tarifa,
2003a, b), although their exact limits remain poorly
defined (see below).
Modern systematic assessments of grison morphol-
ogy and species delimitation may be seen as begin-
ning only with Thomas?s contributions in the early
20th century (e.g. Thomas, 1907, 1912), as this author
was the first to recognize diagnostic characters that
are still perceived as valid today (e.g. Yensen &
Tarifa, 2003a, b). For example, he noted that speci-
mens from south-eastern Brazil (Minas Gerais state)
lacked a ?supplementary internal cusp? in their first
lower molar that seemed to be present in grisons from
northern South America (Thomas, 1907: 162). Five
years later, he proposed that there were two forms of
grisons: a larger one presenting this internal cusp in
the first lower molar (and occurring mostly in the
northern Neotropics), and a smaller one lacking this
internal cusp and inhabiting more southerly areas
of the region (Thomas, 1912). Although these ideas
have formed the basis for present-day recognition
of G. vittata vs. G. cuja (e.g. Yensen & Tarifa, 2003a),
to our knowledge they have never been thoroughly
reassessed since Thomas?s original studies, so that
their effectiveness and general applicability remain
unknown.
The uncertainty surrounding the exact morphologi-
cal distinction between the two Galictis species is a
challenge to the identification of specimens collected
or observed throughout the Neotropics. As a conse-
quence, there is considerable controversy in the lit-
erature regarding the geographical distribution of
these taxa. For instance, some authors have consid-
ered the southern limit of G. vittata to be in north-
eastern Brazil (e.g. Mares et al., 1981; Emmons &
Feer, 1997; Larivi?re & Jennings, 2009), whereas
others have indicated that it is in south-eastern
Brazil (e.g. Nowak, 1991; Eisenberg & Redford, 1999)
or even southern Brazil (Yensen & Tarifa, 2003a). In
stark contrast, Brazil is not even listed as a country
of occurrence for this species in a recent taxonomic
and geographical reference source for mammals
(Wozencraft, 2005). Similar confusion surrounds the
distribution of G. cuja, with some sources listing its
occurrence as far north as north-eastern Brazil
(Yensen & Tarifa, 2003b), whereas others indicate its
northernmost limit on the Atlantic coast to be in
south-eastern Brazil (e.g. Eisenberg & Redford, 1999).
Such discrepancies are apparent when one compares
the distribution maps for these species published in
recent years (e.g. Emmons & Feer, 1997; Eisenberg &
Redford, 1999; Yensen & Tarifa, 2003a, b; Canevari
& Vaccaro, 2007; Cuar?n, Reid & Helgen, 2008; Reid
& Helgen, 2008; Larivi?re & Jennings, 2009), high-
lighting the current lack of knowledge regarding
their geographical range and actual areas of known or
potential sympatry.
TAXONOMIC REVISION OF GENUS GALICTIS 451
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
Given the long and convoluted taxonomic history
of Galictis, as well as the controversial views about
the geographical distributions of G. vittata and
G. cuja, it is critical to address and characterize in
detail the species limits within this genus, which
should also clarify the actual range of the emerging
taxa. Our goals here were thus to test whether two
or more species of Galictis can be recognized based
on morphological and molecular characters, as well
as to provide reliable diagnostic features for them
and a thorough reassessment of their geographical
distribution.
MATERIAL AND METHODS
In order to achieve a comprehensive assessment of
genus Galictis, we analysed morphological aspects of
museum specimens, as well as molecular data col-
lected from fresh tissue samples. On the basis of the
ascertained geographical records derived from the
morphological and molecular data sets, we produced
an updated map of the distribution of each grison
species.
MORPHOLOGICAL ANALYSES
We examined skins and skulls of Galictis specimens
deposited in 22 mammalian collections: Museu de
Ci?ncias e Tecnologia da Pontif?cia Universidade
Cat?lica do Rio Grande do Sul, Porto Alegre, Brazil
(MCT-PUCRS); Museu de Ci?ncias Naturais da
Funda??o Zoobot?nica do Rio Grande do Sul, Porto
Alegre, Brazil (FZB/RS); Museu de Ci?ncias Naturais
da Universidade Luterana do Brasil, Canoas, Brazil
(ULBRA); Laborat?rio de Mam?feros Aqu?ticos da
Universidade Federal de Santa Catarina, Florian?po-
lis, Brazil (LAMAq-UFSC); Museu de Zoologia
da Universidade de S?o Paulo, S?o Paulo, Brazil
(MZUSP); Museu Nacional de Hist?ria Natural, Rio de
Janeiro, Brazil (MNHN); Cole??o de Mam?feros da
Universidade Federal de Pernambuco, Recife, Brazil
(UFPE); Museu Paraense Emilio Goeldi, Bel?m, Brazil
(MPEG); Museo Nacional de Historia Natural y
Antropologia, Montevideo, Uruguay (MNHNA); Museo
Argentino de Ciencias Naturales Bernardino Rivada-
via, Buenos Aires, Argentina (MACN); Museo de La
Plata, La Plata,Argentina (MLP); National Museum of
Natural History, Smithsonian Institution, Washing-
ton, D.C., USA (USNM); American Museum of Natural
History, New York, USA (AMNH); The Academy of
Natural Sciences of Philadelphia, Philadelphia, USA
(ANSP); Museum of Comparative Zoology, Harvard
University, Cambridge, USA (MCZ); The Field
Museum of Natural History, Chicago, USA (FMNH);
Yale Peabody Museum, Yale University, New Haven,
USA (YPM); Natural History Museum from the Uni-
versity of Kansas, Lawrence, USA(NHMKU);Museum
of Vertebrate Zoology, Berkeley, USA (MVZ); The Loui-
siana State University Museum of Natural Science,
Baton Rouge, USA (LSUMNS); Natural History
Museum, London, England (BMNH); and Staatliches
Museum f?r Tierkunde, Dresden, Germany (SMT).
The complete list of specimens is provided in
Appendix S1.
For each of these institutions, all Galictis specimens
were examined, especially those represented by skulls
and/or skins. Skulls were categorized as adult vs.
non-adult, and also assessed in terms of their integrity
(i.e. whether they were intact or broken, and if mea-
surements could be accurately taken). Only adult
skulls with known sex and sufficient integrity to
undertake measurements were included in statistical
analyses. These skulls, along with all others that could
be reliably identified [i.e. including those that pre-
sented diagnostic features (see Results) but belonged
to non-adults and/or were broken, as well as those with
unknown sex] were used for geographical analyses
aimed at reassessing the distribution of the Galictis
species. Skin variation was also surveyed and charac-
terized, especially in cases where the associated skull
could provide reliable species identification (see
Results). Specimens represented only by a skin were
not included in the geographical analysis (see below).
For the morphological analyses, we defined 15 cran-
iodental measurements: skull ? greatest length of
skull (GLS), nasal length (NL), zygomatic breadth
(ZB), mastoid breadth (MB), braincase breadth (BB),
interorbital constriction (IC), postorbital constriction
(PC), palatal width (PW), braincase height (BH),
mandible length (MAL), and mandible height (MH);
teeth ? length of maxillary toothrow (C-M2), external
alveolar distance between upper canines (C-C), exter-
nal alveolar distance between upper molars (M2-M2),
and length of mandible toothrow (c-m2). Adult speci-
mens were recognized as those presenting a fully
erupted permanent dentition along with a total fusion
of the skull sutures. All measurements were taken by
the first author, except for the specimens from BMNH
and SMT, which were measured by the second author.
One categorical variable was recorded for all speci-
mens: the presence or absence of a metaconid in the
first lower molar.
In addition to the skull measurements, we also
recorded the total body length (TL) from several indi-
viduals that contained this information in their skin
tags. To increase the sample size for this external
variable, we combined our data with TL measure-
ments reported in previous studies focusing on Gal-
ictis (Thomas, 1903, 1907, 1912, 1921; Goodwin, 1946;
Greer, 1966; Husson, 1978).
Considering that sexual size dimorphism is promi-
nent in many mustelids, with males being larger than
452 R. BORNHOLDT ET AL.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
females (Dayan et al., 1989; Dayan & Simberloff,
1994; Thom, Harrington & Macdonald, 2004; Rozhnov
& Abramov, 2006; Monakhov, 2009), we performed
t-test comparisons (with a Bonferroni correction) to
assess whether any measurement differed signifi-
cantly between the sexes within each putative Galic-
tis species (defined a priori based on a meristic
character ? see Results). As we observed a strong
pattern of sexual size dimorphism in this genus (see
Appendix S2), we performed all subsequent statistical
analyses separately for males and females, as
described below.
To investigate whether skull measurements con-
tained information that supported segregation
between different Galictis clusters, we conducted mul-
tivariate analyses using all craniodental measure-
ments. Initially, all measurements were log-
transformed so as to reduce their variance and thus
perform a more conservative assessment of group
differences. We then performed a principal component
analysis (PCA) of the correlation matrix of log-
transformed variables. We used a plot of principal
component 1 (PC1) against PC2 to visually assess the
number of distinctive clusters. Then we conducted
discriminant function analyses (DFA) using two differ-
ent approaches: (1) two-group discriminant function,
where we analysed male and female data sets sepa-
rately, in each case considering the two species-level
clusters identified by the PCA (see Results); (2)
multiple-group discriminant function, in which we
performed a joint analysis of the full data set consid-
ering four grouping variables, corresponding to males
and females of each species. In both cases, the assump-
tions of the DFA (normality of the variables, homoge-
neity amongst the covariance matrices, low impact of
multicolinearity, and independence of the samples)
were assessed and found to be adequately met by our
data sets.
Groups identified by the PCA were then used for
subsequent comparisons. We initially calculated stan-
dard descriptive statistics (mean, standard deviation,
and range) for each measurement in each of the
identified groups, and assessed whether they over-
lapped between them. We then compared the mean of
each skull measurement and the TL between the
groups using t-tests with a Bonferroni correction for
multiple comparisons (with an adjusted significance
level of P = 0.05/16). These analyses were performed
separately for males and females. All statistical pro-
cedures based on morphological data were performed
with SPSS v. 17 (SPSS Statistics for Windows, Rel.
17.0.0, Chicago, USA).
MOLECULAR ANALYSES
Tissue samples were obtained from ten individuals of
G. cuja [five obtained from different Brazilian states
(Rio Grande do Sul, Paran?, S?o Paulo, Minas Gerais,
and Bahia) and five from Argentina] and three indi-
viduals of G. vittata from Peru (see Appendix S1 for
sample details). Genomic DNA was extracted from all
samples using a standard phenol-chloroform protocol
(Sambrook, Fritsch & Maniatis, 1989) or the DNeasy
Blood and Tissue Kit (Qiagen), followed by quality
checking on an agarose gel. Available sequences from
two related species were used as outgroups: Ictonyx
striatus and Poecilogale albinucha (see Appendix S3
for GenBank accession numbers). These four taxa are
part of the Ictonychinae, one of the mustelid subfami-
lies defined on the basis of recent analyses of DNA
sequences (Koepfli et al., 2008; Wolsan & Sato, 2010;
Sato et al., 2012).
Mitochondrial gene segments
We amplified a segment of the mitochondrial gene
nicotinamide adenine dinucleotide dehydrogenase
subunit 5 (ND5) via PCR using the primers ND5-DF1
and ND5-DR1 (Trigo et al., 2008). PCR reactions were
performed in a 20 mL final volume containing
10?100 ng of genomic DNA, 1 ? PCR Buffer (Invitro-
gen), 2 mM MgCl2, 0.2 mM deoxynucleotide triphos-
phates (dNTPs), 1 U of Taq Platinum DNA
polymerase (Invitrogen), and 0.2 mM of each primer.
The PCR conditions were the following: ten touch-
down cycles of 94 ?C for 45 s, 60?51 ?C for 45 s (with
a decrease in annealing temperature of 1 ?C per
cycle), and 72 ?C for 1.5 min; followed by 30 cycles of
94 ?C for 45 s, 50 ?C for 45 s, and 72 ?C for 1.5 min,
and a final extension of 72 ?C for 3 min. PCR products
were visualized on a 1% agarose gel stained with
GelRed (Biotium) and purified by precipitation using
ammonium acetate. Purified PCR products were
sequenced in both directions using the DYEnamic ET
Dye Terminator Sequencing Kit (GE Healthcare) and
subsequently analysed in a MegaBACE 1000 auto-
mated sequencer (GE Healthcare).
The forward and reverse chromatograms were
assembled, visualized, and checked using the Phred/
Phrap/Consed package (Ewing et al., 1998; Gordon,
Abajian & Green, 1998). Consensus sequences were
deposited in GenBank (accession numbers JX570686?
JX570694) and aligned using the ClustalW algorithm
implemented in MEGA 4.0 (Tamura et al., 2007). The
alignment was checked and edited by hand using
MEGA, which was also used to assess genetic dis-
tances amongst sequences. Phylogenetic analyses
were performed using three different criteria:
maximum parsimony (MP); maximum likelihood
(ML); and Bayesian inference (BI). The MP and ML
analyses were conducted using PAUP* 4.0b10 (Swof-
ford, 2002), whereas MrBayes 3.1 (Huelsenbeck &
Ronquist, 2001) was used to reconstruct the BI tree.
The best-fit model of nucleotide substitution for the
TAXONOMIC REVISION OF GENUS GALICTIS 453
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
data was estimated using the Akaike information
criterion (AIC) implemented in MODELTEST 3.7
(Posada & Crandall, 1998). The selected model
(general-time-reversible with a proportion of invari-
able sites) was implemented in the ML and BI
analyses. The ML analysis employed a heuristic
search using ten replicates of random taxon addition
(keeping one tree per replicate), followed by tree
bisection-reconnection (TBR) branch-swapping. The
MP phylogeny was based on a heuristic search using
50 replicates of random taxon addition (also keeping
one tree per replicate) and TBR branch-swapping.
Nodal support for the ML and MP methods was
assessed with 500 pseudoreplicates of nonparametric
bootstrapping. The Bayesian analysis used two inde-
pendent replicates of the Metropolis-coupled Markov
chain Monte Carlo (MCMCMC) procedure, each con-
taining four chains (one cold, three heated) run for
106 generations, with trees and parameters sampled
every 100 steps, and the initial 25% discarded as
burn-in.
Nuclear gene segments
We amplified 12 nuclear gene segments (Table 1) with
either one of two (A and B) PCR touchdown protocols:
(A) one cycle of 95 ?C for 3 min; followed by six cycles
of 94 ?C for 15 s, 60 to 50 ?C for 30 s, with a decrease
in annealing temperature of 2 ?C per cycle, and 72 ?C
for 45 s; followed by 30 cycles of 94 ?C for 15 s, 50 ?C
for 30 s, and 72 ?C for 45 s; and one cycle of 72 ?C for
30 min. Each 15 mL reaction contained 6.98 mL of
sterile double-distilled water, 1.5 mL of 10 ? PCR Gold
Buffer, 1.2 mL 25 mM MgCl2, 1.2 mL of 10 mM dNTP
mix, 1.5 mL of both 2 mM forward and reverse primers,
1 U of AmpliTaq Gold Taq polymerase (Applied Bio-
systems, Foster City, CA, USA), and 100 ng genomic
DNA; (B) one cycle of 95 ?C for 10 min; followed by 16
cycles of 94 ?C for 1 min, 63 to 50.2 ?C for 1 min, with
a decrease in annealing temperature of 0.8 ?C per
cycle, and 72 ?C for 1.5 min; followed by 30 cycles of
94 ?C for 1 min, 50 ?C for 1 min, and 72 ?C for
1.5 min; and one cycle of 72 ?C for 5 min. Each 25 mL
reaction contained 14.9 mL of sterile double-distilled
water, 2.5 mL of 10 ? PCR Gold Buffer, 1.5 mL 25 mM
MgCl2, 1.0 mL dimethyl sulphoxide, 2.0 mL of 10 mM
dNTP mix, 1.0 mL of both 2 mM forward and reverse
primers, 1 U AmpliTaq Gold Taq polymerase, and
100 ng genomic DNA. A negative control (no DNA)
was included with all PCRs. Amplification products
were electrophoresed in 1% agarose/Tris-acetic acid-
ethylenediaminetetraacetic acid gels and stained with
ethidium bromide. A 100 bp molecular weight marker
(Promega, Madison, WI, USA) was run with all PCR
products to check that the correct product size was
amplified. PCR products were purified with Exonu-
clease I and shrimp alkaline phosphatase (Affymetrix,
Cleveland, OH, USA). Purified products were then
cycle sequenced in both directions using the BigDye
Terminator v1.1 Cycle Sequencing Kit (Applied
Biosystems) and the original amplification primers.
Cycle sequencing reactions were purified using Agen-
court Cleanseq (Beckman Coulter Inc., Brea, CA,
USA) and then run on an Applied Biosystems 3730xl
DNA Analyzer.
Forward and reverse sequence chromatograms
were assembled, checked, and edited using the
GENEIOUS v. 5.3 software package (Drummond
et al., 2010). Owing to the presence of two mononucle-
otide repeats within the Wilms tumour 1 (WT1) gene
segment, forward and reverse sequences could not be
assembled and were therefore checked individually on
either side of the repeats. Sequences for some locus-
species combinations were taken from Koepfli et al.
(2008) and added to their respective alignments (a
summary of the sequence sources for nDNA analyses
is provided in Table 1). Newly generated sequences
were deposited in GenBank (accession numbers
KC523394?KC523454). Haplotype networks were
built by hand for each nuclear gene segment, and
phylogenies were reconstructed from a concatenated
supermatrix combining all fragments. These analyses
employed the same criteria used for the mitochondrial
data: MP and ML implemented in PAUP*, along with
BI performed with MrBayes. The best-fit model of
nucleotide substitution (selected by the AIC as imple-
mented in MODELTEST) was that of Hasegawa,
Kishino & Yano (1985) with a proportion of invariant
sites. The ML analysis employed a heuristic search
starting with a neighbour-joining tree and followed by
TBR branch-swapping. Nodal support was estimated
by performing 100 bootstrap pseudoreplicates based
on the same search strategy. The MP trees were
retrieved with a heuristic search using 50 replicates
of random taxon addition (keeping one tree per rep-
licate) and TBR branch-swapping. Branch support for
the MP tree was assessed with 500 replicates of
bootstrapping incorporating the same search profile.
The Bayesian analysis used two independent repli-
cates of the MCMCMC procedure, each containing
four chains (one cold, three heated) run for 106 gen-
erations, with trees and parameters sampled every
100 steps, and the initial 25% discarded as burn-in.
GEOGRAPHICAL ANALYSES
The recorded localities for all reliably identified Gal-
ictis specimens (see above for exclusion criteria) were
used to build a geographical database of species
occurrence. For this purpose, we used the exact local-
ity information available in all specimen tags that
included geographical coordinates. However, in most
cases we had to assign coordinates based on other
454 R. BORNHOLDT ET AL.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
T
a
b
le
1.
G
en
e
ab
br
ev
ia
ti
on
an
d
fu
ll
n
am
e,
pr
im
er
se
qu
en
ce
s
(f
or
w
ar
d
an
d
re
ve
rs
e)
,d
es
cr
ip
ti
on
of
ge
n
e
se
gm
en
ts
,a
n
d
li
te
ra
tu
re
so
u
rc
e
fo
r
ea
ch
n
u
cl
ea
r
se
gm
en
t
u
se
d
in
th
e
st
u
dy
A
bb
re
vi
at
ed
ge
n
e
n
am
e
G
en
e
n
am
e
F
or
w
ar
d
pr
im
er
R
ev
er
se
pr
im
er
R
eg
io
n
R
ef
er
en
ce
A
A
M
P
2
A
n
gi
o-
as
so
ci
at
ed
,
m
ig
ra
to
ry
ce
ll
pr
ot
ei
n
A
G
C
T
G
C
T
C
T
T
T
G
A
G
T
G
T
G
C
C
A
G
C
A
C
A
A
G
T
A
A
C
A
G
A
G
T
C
E
xo
n
-i
n
tr
on
V
?l
i
et
al
.,
20
08
A
D
O
R
A
3
A
3
ad
en
os
in
e
re
ce
pt
or
A
C
C
C
C
C
A
T
G
T
T
T
G
G
C
T
G
G
A
A
G
A
T
A
G
G
G
T
T
C
A
T
C
A
T
G
G
A
G
T
T
E
xo
n
M
u
rp
h
y
et
al
.,
20
01
A
P
O
B
A
po
li
po
pr
ot
ei
n
B
G
T
G
C
C
A
G
G
T
T
C
A
A
T
C
A
G
T
A
T
A
A
G
T
C
C
A
G
C
A
A
A
A
T
T
T
T
C
T
T
T
T
A
C
T
T
C
A
A
E
xo
n
A
m
ri
n
e-
M
ad
se
n
et
al
.,
20
03
G
N
A
T
1
R
od
tr
an
sd
u
ci
n
al
ph
a-
su
bu
n
it
A
G
C
A
C
C
A
T
C
G
T
C
A
A
G
C
A
G
A
C
T
G
G
A
T
A
C
C
C
G
A
G
T
C
C
T
T
C
E
xo
n
-i
n
tr
on
B
ro
u
il
le
tt
e,
A
n
dr
ew
&
V
en
ta
,
20
00
J
A
K
1
Ja
n
u
s
ki
n
as
e
1
G
A
T
C
T
C
T
T
C
A
T
G
C
A
C
C
G
G
A
A
C
A
T
T
T
C
C
A
T
G
G
A
C
C
A
G
G
T
C
T
T
E
xo
n
-i
n
tr
on
H
ou
sl
ey
et
al
.,
20
06
M
A
C
F
1
M
ic
ro
tu
bu
le
-a
ct
in
cr
os
sl
in
ki
n
g
fa
ct
or
1
C
C
A
T
C
T
G
C
T
G
A
G
T
A
T
A
A
A
G
T
G
G
T
G
A
A
G
C
C
T
C
C
T
T
C
T
G
C
T
T
G
A
A
G
C
A
E
xo
n
-i
n
tr
on
H
ou
sl
ey
et
al
.,
20
06
P
F
K
F
B
1
6-
ph
os
ph
of
ru
ct
o-
2-
ki
n
as
e/
fr
u
ct
os
e-
2,
6-
bi
ph
os
ph
at
as
e
1
C
A
G
A
G
A
A
C
G
A
C
G
G
T
C
A
C
T
G
A
T
G
G
T
C
A
T
T
A
C
A
A
A
T
G
G
A
C
T
C
A
A
T
G
A
E
xo
n
-i
n
tr
on
H
ou
sl
ey
et
al
.,
20
06
P
T
P
N
4
P
ro
te
in
ty
ro
si
n
e
ph
os
ph
at
as
e,
n
on
-r
ec
ep
to
r
ty
pe
4
C
C
A
G
T
A
T
T
T
T
T
T
G
C
A
A
A
T
T
A
A
A
C
A
A
G
A
A
G
G
A
A
T
G
A
A
A
G
A
A
T
A
A
T
C
T
G
A
G
A
G
G
T
E
xo
n
-i
n
tr
on
H
ou
sl
ey
et
al
.,
20
06
R
A
G
1
R
ec
om
bi
n
at
io
n
ac
ti
va
ti
n
g
pr
ot
ei
n
1
G
C
T
T
T
G
A
T
G
G
A
C
A
T
G
G
A
A
G
A
A
G
A
C
A
T
G
A
G
C
C
A
T
C
C
C
T
C
T
C
A
A
T
A
A
T
T
T
C
A
G
G
E
xo
n
Te
el
in
g
et
al
.,
20
00
R
H
O
1
R
h
od
op
si
n
T
A
C
A
T
G
T
T
C
G
T
G
G
T
C
C
A
C
T
T
T
G
G
T
G
G
G
T
G
A
A
G
A
T
G
T
A
G
A
A
E
xo
n
-i
n
tr
on
V
en
ta
et
al
.,
19
96
T
R
H
D
E
T
h
yr
ot
ro
pi
n
-r
el
ea
si
n
g
h
or
m
on
e
de
gr
ad
in
g
en
zy
m
e
C
T
G
G
A
T
G
A
G
G
A
T
G
T
C
T
G
G
G
A
T
G
A
A
A
A
A
C
T
T
C
C
A
G
G
C
A
A
G
G
T
C
E
xo
n
-i
n
tr
on
H
ou
sl
ey
et
al
.,
20
06
W
T
1
W
il
m
s
tu
m
ou
r
1
G
A
G
A
A
A
C
C
A
T
A
C
C
A
G
T
G
T
G
A
G
T
T
T
T
A
C
C
T
G
T
A
T
G
A
G
T
C
C
T
E
xo
n
-i
n
tr
on
V
en
ta
et
al
.,
19
96
TAXONOMIC REVISION OF GENUS GALICTIS 455
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
types of locality data available on the tag (which often
consisted of information restricted to municipality,
state, or even country, as well as the name of a lake,
river, or mountain for some specimens). This assign-
ment was performed using GLOBAL GAZETTEER v.
2.2 (available at http://www.fallingrain.com) for all
data points containing municipality data. In a few
cases where this was not available, GOOGLE EARTH
6.1.0 (available at http://www.googleearth.com) was
used to identify other landscape features within a
delimited country or region. In cases where the local-
ity of origin was not precise, but could still provide a
reasonable region of placement within an approxi-
mate radius of 500 km (e.g. unspecified state-level
records in Brazil, or tags referring to countries such
as Uruguay or Panama) we used the central coordi-
nates of the identified region. Still, some specimens
contained such vague geographical information (e.g. a
large country or a very long river) that they could not
be reliably used for geographical referencing and were
thus excluded from this analysis. The resulting data-
base was then analysed with a geographical informa-
tion system (GIS) using the biome map reported by
Olson et al. (2001) and the software ArcGIS 9.3
(ESRI, 2009). We then used the distributional maps to
characterize in detail the geographical range of each
species, to assess possible areas of sympatry between
them, and to investigate the presence of these taxa
across Neotropical biomes.
RESULTS
MORPHOLOGICAL DATA
Our first assessment of cranial material consisted of a
survey for the presence of the metaconid in the first
lower molar that has been proposed to be a diagnostic
character distinguishing these species (see Introduc-
tion). These initial surveys revealed that indeed there
was a pool of specimens bearing such a feature (almost
all of which had been sampled in the northern portion
of the genus?s range), and another lacking the meta-
conid (most of which had been collected in the southern
portion of the Galictis range) (Fig. 1). Based on this
character, along with prior identification of museum
specimens, we provisionally assigned each individual
of the former group to G. vittata, and the latter to
G. cuja, so as to provide a taxonomic hypothesis that
could be tested with statistical morphometrics.
We then performed a PCA based on the 15 cranio-
dental measurements, aiming to test whether two or
more distinct clusters could be distinguished, and if
such groups would be congruent with the a priori
taxonomic partition outlined above. The PCA showed
the presence of two distinct clusters in both male
(Fig. 2A) and female (Fig. 2B) data sets, which corre-
sponded very well to the G. cuja and G. vittata parti-
tions previously hypothesized on the basis of the
metaconid character. For both sexes, the distinction
between the two groups was essentially along the first
principal component, which explained 85.15% of the
total variance for males, and 88.64% for females
(Table 2). Taken together, the first two components
explained 89.19 (males) and 92.38% (females) of the
total variation. The variables with the greatest con-
tribution to the first component were GLS, ZB, MB,
MAL, and c-m2 (males) and GLS, ZB, MAL, and c-m2
(females), with this separation thus reflecting differ-
ences in overall skull size.
The DFA corroborated and extended the results of
the PCA. The means of all variables (craniodental
measurements) were statistically different between
the groups (species), indicating that all of them were
highly significant with respect to their ability to seg-
regate the species (Table 3). The classification analy-
sis (using the conservative cross-validation procedure)
with the two-group approach (which considered each
sex separately) was 100% accurate in correctly clas-
sifying the males In the case of females, there was a
single G. cuja misclassified as G. vittata (see Appen-
dix S4 for details of the classification statistics).
The multiple-group DFA yielded three discriminant
functions, but the chi-square test indicated that only
the first two canonical variates (or roots) were signifi-
Figure 1. First and second lower molars from (A) Galictis
vittata and (B) Galictis cuja in lingual view, emphasizing
the presence of the metaconid (indicated by an arrow) in
the former species. The metaconid constitutes a diagnostic
character to distinguish the two Galictis species (see main
text for details).
456 R. BORNHOLDT ET AL.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
cant (Table 2). The first canonical variate (CV1)
accounted for 90.8% of the discriminatory power
between clusters, whereas the second (CV2)
accounted for 7.9%. Based on the standardized coef-
ficients for the 15 variables, the most relevant vari-
ables for group segregation along CV1 were: c-m2,
BB, and MB. The means of population groups repre-
sented by the group centroid indicated two clear clus-
ters (varying along CV1), again corresponding to
G. cuja and G. vittata (Fig. 3). The segregation
between males and females within each species was
mainly seen along the CV2 axis, although this segre-
gation was not perfect and there was considerable
overlap between males and females within each
species.
The classification analysis performed with the
multiple-group DFA yielded an overall correct clas-
sification of 76.3% (using the cross-validation proce-
dure). Almost all cases of misclassification occurred
between sexes of the same species, with a single
event of exchange between G. cuja and G. vittata. In
this case, one G. cuja female was classified as a
G. vittata female. The complete function extracted
from the classification coefficients is shown in
Appendix S4.
Given the identification of distinct clusters by the
PCA, we surveyed the range of all linear measure-
ments observed in each group, keeping male and
female data sets separate. We observed that some
variables do not overlap between the two groups
(Table 4), highlighting the size distinction between
these clusters. Although the magnitude of the distinc-
tion was in some cases modest, the following vari-
ables enabled the direct segregation of Galictis
specimens into two species-level clusters (G. cuja and
G. vittata): BB, c-m2, and TL for males; and GLS, ZB,
BB, IC, MAL, and c-m2 for females (see Table 4). In
addition to the observation of non-overlapping ranges
in some measurements, we noted for both males and
females that the mean of all variables was clearly
different between G. cuja and G. vittata, with the
latter being considerably larger (skull size compari-
son is represented in Fig. 4). The t-test results showed
significant differences in all linear measurements
(craniodental and TL), even after applying the con-
servative Bonferroni correction (Table 5).
The analysis of pelage features revealed a trend
that seems to fit the species-level distinction demon-
strated by the linear measurements, although we did
not perform statistical tests here because of the high
variability in skin age and preparation that could bias
this assessment. We observed that all specimens iden-
tified as G. cuja presented a denser and/or longer
pelage, leading to a ?furrier? appearance. This feature
was never observed in any individual identified as
G. vittata, whose fur was always shorter (Fig. 5). In
addition, G. vittata individuals were always greyish
(varying from pale grey to medium-dark grey), never
Figure 2. Principal component analysis projection based on scores from the first (PC1) and second (PC2) principal
components for 15 craniodental measurements (see Table 2 for details) from specimens of genus Galictis: Galictis cuja
(black circles) and Galictis vittata (white diamonds). A, male data set; B, female data set.
TAXONOMIC REVISION OF GENUS GALICTIS 457
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
exhibiting yellowish hues (see Fig. 5). In contrast,
G. cuja specimens tended to be yellowish, although
some individuals could be as greyish as G. vittata,
thus overlapping in coloration with the latter species.
Therefore, a combination of fur length/density with
coloration may be a reliable external character to
diagnose these species, although caution should be
taken given the observed overlap in the latter.
MOLECULAR DATA
Mitochondrial and nuclear gene segments provided
clear evidence of differentiation between two distinct
Galictis clades (Fig. 6). All phylogenetic analyses per-
formed with the mtDNA data set resulted in high
support for these two clades, whose geographical
occurrence indicated that they corresponded to
G. cuja and G. vittata (Fig. 6A). The genetic distance
between these groups was substantial, e.g. 12.7%
mean uncorrected divergence (p-distance) with the
ND5 data set. All 12 nuclear segments provided con-
cordant evidence for this separation, even when
analysed individually (Fig. 7, Appendix S5). There
was no nuclear haplotype sharing between the two
geographical groups corresponding to G. cuja and
G. vittata, which were found to be reciprocally mono-
phyletic in every case. Depending on the locus analy-
sed, these two groups were differentiated by one
(JAK1 and MACF1), two (TRHDE, AAMP2, ADORA3,
PFKFB1, and PTPN4), three (GNAT1, APOB, and
Table 2. Summary of results from the principal component analysis (PCA) and multiple-group discriminant function
analysis (DFA): male and female factor loading for each craniodental measurement, eigenvalue and cumulative variance
of the first two principal components in the PCA (represented in Fig. 2), and discriminant loadings for each craniodental
measurements, Chi-square statistics, canonical correlation, eigenvalue and cumulative variance of the first two canonical
variates in the DFA (represented in Fig. 3)
Variable
PCA
Multiple-group DFAMales Females
PC1 PC2 PC1 PC2 CV1 CV2
GLS 0.977 -0.059 0.982 -0.064 -0.328 -0.111
NL 0.808 0.212 0.817 -0.357 0.242 0.127
ZB 0.969 -0.117 0.983 -0.035 0.024 0.489
MB 0.969 -0.132 0.979 -0.065 0.673 -0.516
BB 0.918 0.177 0.962 0.139 0.807 -0.936
IC 0.928 -0.136 0.967 0.014 0.453 -0.243
PC 0.746 0.622 0.771 0.580 -0.361 0.537
PW 0.880 0.094 0.844 0.159 -0.415 -0.470
BH 0.932 0.095 0.949 0.106 -0.076 0.045
MAL 0.972 -0.085 0.983 -0.091 -0.437 0.515
MH 0.882 -0.096 0.955 -0.037 -0.609 0.985
C-M2 0.939 -0.016 0.960 -0.067 0.019 -0.220
C-C 0.959 -0.190 0.973 -0.096 -0.512 0.319
M2-M2 0.961 -0.076 0.978 -0.019 0.269 0.154
c-m2 0.969 -0.111 0.984 -0.077 1.045 -0.083
Wilks? lambda - - - - 0.052 0.504
Chi-square statistic - - - - 18.684 73.575
d.f. - - - - 45 28
P-value - - - - 0.000* 0.000*
Canonical correlation - - - - 0.947 0.657
Eigenvalue 12.77 0.60 13.29 0.56 8.778 0.760
Cumulative variance (%) 85.15 89.19 88.64 92.38 90.8 98.7
Values highlighted in bold represent those with the greatest contribution to the first principal component and first
canonical variate. Asterisks indicate results that are statistically significant with the Chi-square statistic of the DFA.
GLS, greatest length of skull; NL, nasal length; ZB, zygomatic breadth; MB, mastoid breadth; BB, braincase breadth; IC,
interorbital constriction; PC, postorbital constriction; PW, palatal width; BH, braincase height; MAL, mandible length;
MH, mandible height; C-M2, length of maxillary toothrow; C-C, external alveolar distance between upper canines;
M2-M2, external alveolar distance between upper molars; c-m2, length of mandible toothrow.
458 R. BORNHOLDT ET AL.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
RHO1), or even six mutational steps (RAG1 and WT1)
(see Table 1 for full names of genes). In addition to
the individual-gene assessments, we also concat-
enated the 12 segments for each individual (ignoring
heterozygous sites), and observed very high support
for the two Galictis clades with all phylogenetic
methods employed (Fig. 6B). Overall, considering
both the mtDNA and nuclear data sets, nodal support
for these clades was > 90% with all analytical
methods.
GEOGRAPHICAL DISTRIBUTION
As the morphological and molecular approaches cor-
roborated the recognition of two clusters in Galictis
corresponding to G. vittata and G. cuja, we employed
our data sets to reassess and refine the geographical
distribution of both of these species. The resulting
maps indicated that G. vittata is distributed in the
northern range of the genus, from the extreme north
of the Neotropics, in Mexico, to the central region of
South America. In contrast, G. cuja is distributed in
the southern range of the genus, from the extreme
south of Peru, southern Bolivia, and north-eastern
Brazil to southern Chile and Argentina (Fig. 8,
Appendix S6).
The refined distributional maps allowed an assess-
ment of potential areas of sympatry between the two
species. The range of G. vittata extends southwards
to Peru (throughout Loreto, Amazonas, Ucayali,
Pasco, Jun?n, and Cuzco departments), Bolivia
(Santa Cruz department), Paraguay (Guaira depart-
ment), and north-western Brazil (Amazonas, Amap?,
Par?, and Rond?nia states). The occurrence in these
countries suggests some overlap with the north-
western limit of the G. cuja range, documented
in Peru (Puno department), Bolivia (Cochabamba,
Santa Cruz, and Tarija departments), and through-
out central-western to north-eastern Brazil (includ-
ing Alagoas, Bahia, Cear?, Goi?s, Para?ba, and
Pernambuco states and the Distrito Federal) (see
Fig. 8 and Appendix S6).
A finer-scale analysis reveals less evidence of
geographical overlap, although some instances may
Table 3. Equality test of group mean results for two-group (males and females analysed separately) and multiple-group
discriminant function analysis (DFA): independent contribution (differentiation power) of each variable to discriminate
the groups
Variable
Two-group DFA
Multiple-group DFAMales Females
Wilks?
lambda F P-value
Wilks?
lambda F P-value
Wilks?
lambda F P-value
GLS 0.338 113.413 0.000 0.171 270.764 0.000 0.226 130.359 0.000
NL 0.616 36.111 0.000 0.416 78.666 0.000 0.464 43.929 0.000
ZB 0.356 104.882 0.000 0.161 292.247 0.000 0.233 125.126 0.000
MB 0.289 142.597 0.000 0.166 281.393 0.000 0.209 143.734 0.000
BB 0.300 135.408 0.000 0.173 267.036 0.000 0.229 127.746 0.000
IC 0.256 168.879 0.000 0.187 243.061 0.000 0.216 137.752 0.000
PC 0.665 29.197 0.000 0.575 41.318 0.000 0.604 24.902 0.000
PW 0.589 40.441 0.000 0.461 65.578 0.000 0.518 35.323 0.000
BH 0.339 113.105 0.000 0.287 139.093 0.000 0.296 90.212 0.000
MAL 0.385 92.564 0.000 0.167 280.263 0.000 0.238 121.628 0.000
MH 0.649 31.319 0.000 0.257 161.867 0.000 0.363 66.654 0.000
C-M2 0.331 117.029 0.000 0.229 189.052 0.000 0.261 107.477 0.000
C-C 0.308 130.511 0.000 0.226 191.826 0.000 0.243 118.482 0.000
M2-M2 0.319 123.720 0.000 0.154 306.761 0.000 0.218 136.472 0.000
c-m2 0.219 207.258 0.000 0.149 319.806 0.000 0.172 183.571 0.000
F = F-value; P-value = significance.
Asterisks indicate morphometric variables that are statistically significant between the groups.
GLS, greatest length of skull; NL, nasal length; ZB, zygomatic breadth; MB, mastoid breadth; BB, braincase breadth; IC,
interorbital constriction; PC, postorbital constriction; PW, palatal width; BH, braincase height; MAL, mandible length;
MH, mandible height; C-M2, length of maxillary toothrow; C-C, external alveolar distance between upper canines;
M2-M2, external alveolar distance between upper molars; c-m2, length of mandible toothrow.
TAXONOMIC REVISION OF GENUS GALICTIS 459
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
remain. When the presence of each species was
assessed with respect to different biomes or ecore-
gions (Olson et al., 2001), we observed that in Peru
G. vittata seems to occur in the broad region of the
Peruvian Amazon, but not in the mountainous region
of the Andes or the drier region of the coast. In
contrast, the G. cuja records from that country are
from the south, in the Montane Grassland and Scru-
blands (Puna grasslands) region, up in the Andes. In
Bolivia, the only record of G. vittata originates from
the Tropical and Subtropical Dry Broadleaf Forest
(Chiquitano), whereas those for G. cuja include this
same biome as well as the Amazonian region (Tropical
and Subtropical Moist Broadleaf Forest). In Para-
guay, both species seem to co-exist in the Tropical and
Subtropical Moist Broadleaf Forests (Atlantic Forest)
in the southern region of the country, although
G. cuja also occurs in the Savanna (Chaco). Within
Brazil, there seems to exist a clear biogeographical
division between the species, with G. vittata occurring
exclusively in the Amazonian region and G. cuja
occurring in the other biomes, including the drier
Cerrado (Savanna) and Caatinga (Deserts and Xeric
Shrublands) of the north-east, as well as the Atlantic
Forest throughout the eastern seaboard, and the
pampas grasslands towards the south (see Fig. 8).
DISCUSSION
Our analyses demonstrated that two Galictis species
can be clearly delimited and diagnosed using metric
and nonmetric morphological characters (Table 6), as
well as DNA sequences from mitochondrial and
nuclear gene segments. On the basis of this clarified
species-level delimitation, we reassessed the geo-
graphical range of each Galictis taxon in the Neotro-
pics, identifying possible areas of sympatry between
them. Each of these topics will be discussed in detail
below.
NUMBER OF EXTANT SPECIES IN GENUS GALICTIS
Although the presence of two grison species has been
commonly mentioned in literature sources reviewing
mammalian taxonomy (e.g. Yensen & Tarifa, 2003a,
b; Wozencraft, 2005; Larivi?re & Jennings, 2009)
this hypothesis has never been formally tested. In
addition, the possible existence of a third species
(G. allamandi) in northern South America and/or
Central America is mentioned by some sources,
including an influential set of reference books on
Neotropical mammals (Eisenberg, 1989; Redford &
Eisenberg, 1992; Eisenberg & Redford, 1999). In
these publications, the authors mentioned that G. al-
lamandi might be a synonym of G. vittata, but ten-
tatively recognized it as a distinct taxon. Moreover,
in Eisenberg (1989), a range map for G. allamandi is
provided, implying that such a taxon might indeed
warrant recognition. As our morphometric analyses
support the recognition of only two Galictis species,
with no evidence for a third cluster in Central
America or northern South America, we review
below the taxonomic history of G. allamandi, so as to
clarify its status.
When Schreber (1776) described V. vittata based on
the drawing by Allamand (Buffon, 1776) of an animal
originating from Suriname, he seems not to have
analysed any actual specimen (museum material or
living animal), implying that the grison features dis-
cussed were originally tied to the drawing alone.
When Bell (1837) described the new species G. alla-
mandi as distinct from G. vittata (albeit in a new
genus), he also tied the description of the former to
the drawing by Allamand, along with a single
museum specimen that he had examined personally,
saying that both ?may perhaps be identical? (Bell,
1837: 45). For the latter species, however, he took
the name ?vittata? (originally tied to the Allamand
drawing) and associated it with the living animal
from Guyana that he had observed (see Introduction),
suggesting that it exhibited a ?distinct specific differ-
ence? (Bell, 1837: 45) from G. allamandi. Therefore,
we can conclude that Bell (1837) untied the Allamand
Figure 3. Multiple-group discriminant function analysis
plot based on scores from the first and second canonical
variates (CV1 and CV2) for 15 craniodental measurements
(see Table 2 for details) from specimens of genus Galictis:
Galictis cuja males (black circles); G. cuja females (white
circles); Galictis vittata males (black diamonds), and
G. vittata females (white diamonds).
460 R. BORNHOLDT ET AL.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
Table 4. Descriptive statistics (mean, standard deviation, range, and sample size) for 15 craniodental measurements and
one external variable comparing Galictis cuja and Galictis vittata, with males and females treated separately (measure-
ments in mm)
Variable G. cuja  G. vittata  G. cuja  G. vittata 
GLS 76.28 ? 4.03 88.80 ? 3.93 69.50 ? 3.49 85.44 ? 3.12
64.93?83.52 81.85?94.76 63.40?77.14 79.10?92.90
N = 60 N = 19 N = 44 N = 26
NL 21.45 ? 1.76 24.74 ? 1.22 19.83 ? 1.36 23.96 ? 1.48
17.20?26.47 22.81?27.41 17.40?23.63 21.13?27.76
N = 67 N = 22 N = 43 N = 31
ZB 43.02 ? 3.12 52.87 ? 2.86 39.21 ? 2.18 50.13 ? 2.40
35.0?50.09 47.18?56.81 34.16?44.04 44.72?54.50
N = 60 N = 20 N = 44 N = 31
MB 39.76 ? 2.75 50.04 ? 2.47 35.62 ? 2.33 46.89 ? 2.61
32.52?46.83 44.80?54.36 31.28?41.74 41.61?52.29
N = 63 N = 20 N = 43 N = 31
BB 34.68 ? 1.83 40.68 ? 1.20 33.10 ? 1.54 39.70 ? 1.39
30.34?37.98 38.34?43.05 29.70?36.51 37.34?42.37
N = 63 N = 21 N = 45 N = 31
IC 16.55 ? 1.30 21.88 ? 1.62 15.26 ? 1.21 20.29 ? 0.97
13.20?19.50 18.50?26.98 13.0?18.17 18.41?21.90
N = 66 N = 22 N = 44 N = 31
PC 17.72 ? 1.33 19.92 ? 0.76 16.83 ? 1.40 19.28 ? 0.98
14.39?21.66 18.81?21.36 12.96?19.76 17.81?20.88
N = 64 N = 21 N = 45 N = 31
PW 11.68 ? 0.96 13.58 ? 0.90 11.07 ? 0.87 13.05 ? 0.77
9.51?13.50 12.05?15.60 9.62?12.73 10.77?14.71
N = 65 N = 22 N = 46 N = 31
BH 25.43 ? 1.69 30.73 ? 1.15 23.17 ? 2.08 29.06 ? 1.30
21.91?28.94 28.81?33.14 18.67?29.35 26.90?32.23
N = 63 N = 20 N = 44 N = 30
MAL 44.79 ? 2.98 54.11 ? 2.50 40.32 ? 2.22 50.83 ? 2.63
36.73?51.70 48.74?57.80 36.02?45.35 46.76?58.38
N = 59 N = 20 N = 39 N = 28
MH 21.80 ? 1.84 25.10 ? 1.95 19.05 ? 1.58 23.62 ? 1.20
17.16?25.99 20.47?27.94 16.02?22.77 21.27?26.65
N = 72 N = 21 N = 47 N = 31
C-M2 20.98 ? 1.50 25.94 ? 1.14 19.08 ? 1.66 24.44 ? 1.11
17.22?27.96 23.82?27.56 16.20?25.90 22.01?26.86
N = 69 N = 22 N = 47 N = 31
C-C 16.75 ? 1.41 21.23 ? 1.42 14.91 ? 1.29 19.41 ? 1.24
13.30?21.49 17.40?23.47 12.30?17.91 17.03?23.03
N = 67 N = 22 N = 47 N = 30
M2-M2 23.97 ? 1.64 29.76 ? 1.29 21.90 ? 1.55 28.11 ? 1.16
20.02?27.91 27.81?31.51 19.64?27.55 26.20?30.51
N = 67 N = 22 N = 45 N = 31
c-m2 26.14 ? 1.52 33.24 ? 1.49 23.74 ? 1.56 30.86 ? 1.50
21.71?29.18 30.30?35.72 20.0?27.40 28.51?35.34
N = 69 N = 21 N = 46 N = 30
TL 601.66 ? 40.62 722.16 ? 19.18 531.36 ? 55.67 658.12 ? 38.62
525.0?657.0 700.0?755.0 443.0?645.0 600.0?706.0
N = 12 N = 6 N = 11 N = 8
Minimum and maximum values in bold indicate variables whose range does not overlap between the species.
GLS, greatest length of skull; NL, nasal length; ZB, zygomatic breadth; MB, mastoid breadth; BB, braincase breadth; IC,
interorbital constriction; PC, postorbital constriction; PW, palatal width; BH, braincase height; MAL, mandible length;
MH, mandible height; C-M2, length of maxillary toothrow; C-C, external alveolar distance between upper canines;
M2-M2, external alveolar distance between upper molars; c-m2, length of mandible toothrow; TL, total body length.
TAXONOMIC REVISION OF GENUS GALICTIS 461
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
drawing from the Schreber description in order to
recognize G. allamandi and G. vittata, whose distinc-
tion was ultimately based on a single museum speci-
men compared to a single living animal. Given this
context, it is apparent that there was no strong evi-
dence that distinguished these species upon their
proposition by Bell (1837). Indeed, our results, which
are based on a large sample of individuals collected
Figure 4. Representative skull views (dorsal, ventral, and lateral skull views and dorsal and lateral mandible views)
from Galictis specimens, emphasizing size differences amongst the groups: A, Galictis vittata male, LSUMNS 2443; B,
Galictis cuja male, MVZ 114774; C, G. vittata female, USNM 180224; and D, G. cuja female, LSUMNS 16948. Scale
bar = 10 mm.
462 R. BORNHOLDT ET AL.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
throughout the distribution of Galictis, did not reveal
any evidence for a third cluster in that region. Thus,
the uncertainty involving G. allamandi seems to have
been a result of taxonomic confusion based on
intraspecific variation, supporting the view that this
species should be in fact considered a junior synonym
of G. vittata.
SPECIES DELIMITATION ? G. CUJA VS. G. VITTATA
The magnitude of morphological and genetic differen-
tiation between the two Galictis species was similar
to results reported by recent studies targeting the
taxonomic status of other mustelid genera (e.g.
Helgen et al., 2008; Harding & Smith, 2009; Jacques
et al., 2009; del Cerro et al., 2010). Our observed dis-
crimination between the two clusters in both types of
multivariate analysis (PCA and DFA) was almost
perfect, with the exception of two male records of
G. cuja [one from Rio Grande do Sul state, southern
Brazil (MZUSP 1044), and another from the coast of
Uruguay (MNHNA 2696)] that were located within
the G. vittata cluster in the PCA (see Fig. 2A).
Observing these two individuals in detail, we noticed
that both were exceptionally large, exhibiting all
linear measurements longer (in some cases much
longer) than the mean for all G. cuja males. Given
their size, it is not surprising that they would deviate
towards the G. vittata cluster. This observation is
rather interesting, and illustrates that the smallest
G. vittata males could overlap in size with the largest
G. cuja males, possibly leading to misidentification if
only some linear measurements are employed. It is
noteworthy that several other male specimens from
Uruguay and southern Brazil were also analysed, all
of which lay within the G. cuja cluster, thus indicat-
ing that the large size of those two specimens may not
be because of any particular geographical trend, but
rather be attributed to within-population inter-
individual variation. In spite of these two exceptions,
the multivariate classification analyses exhibited
extremely high accuracy when distinguishing the two
species, and remained effective even with individuals
whose gender was unknown (see Results and Appen-
dix S4). Another line of evidence that supported clear
species-level delimitation derived from the univariate
analyses. For all 15 craniodental measurements, the
differences between species were highly significant,
corroborating the interpretation that their distinc-
tiveness is strongly related to divergence in size. Such
a pattern was also observed in the external measure-
ment compared here (TL), indicating that this
Table 5. Male and female results of the t-test comparison between Galictis cuja and Galictis vittata based on the mean
of 15 craniodental measurements and one external variable
Variable
Males Females
t d.f. P-value t d.f. P-value
GLS -11.862 77 0.000 -19.153 68 0.000
NL -8.099 87 0.000 -12.375 72 0.000
ZB -12.464 78 0.000 -20.458 73 0.000
MB -14.868 81 0.000 -19.488 72 0.000
BB -13.991 82 0.000 -19.024 74 0.000
IC -15.595 86 0.000 -19.097 73 0.000
PC -7.141 83 0.000 -8.373 74 0.000
PW -8.119 85 0.000 -10.137 75 0.000
BH -13.035 81 0.000 -13.704 72 0.000
MAL -12.549 77 0.000 -17.638 65 0.000
MH -7.146 91 0.000 -13.637 76 0.000
C-M2 -14.144 89 0.000 -15.711 76 0.000
C-C -12.882 87 0.000 -15.078 75 0.000
M2-M2 -15.062 87 0.000 -18.876 74 0.000
c-m2 -18.751 88 0.000 -19.746 74 0.000
TL -6.817 16 0.000 -5.526 17 0.000
t = t-value; d.f. = degrees of freedom; P-value = significance.
GLS, greatest length of skull; NL, nasal length; ZB, zygomatic breadth; MB, mastoid breadth; BB, braincase breadth; IC,
interorbital constriction; PC, postorbital constriction; PW, palatal width; BH, braincase height; MAL, mandible length;
MH, mandible height; C-M2, length of maxillary toothrow; C-C, external alveolar distance between upper canines;
M2-M2, external alveolar distance between upper molars; c-m2, length of mandible toothrow; TL, total body length.
TAXONOMIC REVISION OF GENUS GALICTIS 463
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
size-based discrimination may also be feasible for
identifying live animals.
In addition to the morphological data that sup-
ported clear-cut species-level separation, the molecu-
lar data sets also indicated that the two Galictis taxa
are substantially differentiated. All trees portrayed
two reciprocally monophyletic lineages, one of which
included only samples from Peru (almost certainly
corresponding to G. vittata) and the other comprising
lineages sampled in Brazil and Argentina (including
areas where only G. cuja was found to occur). The
magnitude of evolutionary differentiation between
these two groups (e.g. 12.7% p-distance with ND5)
was sufficient to induce reciprocal monophyly at all 12
surveyed nuclear loci, a pattern that is often not
observed in closely related taxa (e.g. Syring et al.,
2007; Degnan & Rosenberg, 2009). Such nuclear dif-
ferentiation is compatible with the estimate of diver-
gence time between G. cuja and G. vittata, which was
inferred to be c. 2 to 3 million years ago (Mya)
(Koepfli et al., 2008; Sato et al., 2012). The fossil
record indicates that grisons probably originated in
North America, where they may have shared a
common ancestor with the Pliocene genus Trigonictis
(Reig, 1957; Yensen & Tarifa, 2003a). Representatives
of genus Galictis then colonized South America
during the Great American Biotic Interchange
(GABI), with their first record in this sub-continent
appearing in Argentina in the Vorohuean subage of
the late Pliocene (Reig, 1957; Hunt, 1996; Woodburne,
2010), dated at 3.0?2.5 Mya (Cione & Tonni, 1995;
Prevosti & Soibelzon, 2012). Integrating our results
with the available fossil and molecular information,
we can hypothesize that a single Galictis ancestor
invaded South America via the Panamanian isthmus
early during the GABI (Hunt, 1996; Woodburne,
2010; Eizirik, 2012), soon afterwards giving rise to
the two extant species, as well as additional extinct
members of the genus (Reig, 1957). Further analyses
addressing the tempo and mode of this speciation
process should yield interesting insights into the
history of this lineage and associated components of
the Neotropical biota.
MORPHOLOGICAL DIAGNOSIS OF G. CUJA AND
G. VITTATA
Skull
Historically, various literature sources (e.g. Thomas,
1912; Yensen & Tarifa, 2003a, b) have mentioned that
the two extant Galictis species can be distinguished
on the basis of size differences along with the
presence/absence of the metaconid in m1. However,
this distinction had so far not been formally tested
with a broad sample from across the genus?s range.
The results from this study corroborated the useful-
ness of the metaconid as a diagnostic feature between
these species (Fig. 1), as it was present in 100% of the
G. vittata skulls, and absent in virtually all G. cuja
skulls. The single potential exception was a very
small G. cuja male (AMNH 33281) from Temuco,
Chile, that bore a subtle cusp on the left m1, which
was similar but smaller than the metaconid seen in
G. vittata. In addition to the presence/absence of the
Figure 5. Dorsal view from representative skins for the
two Galictis species, emphasizing size and general appear-
ance differences between the groups: A, Galictis vittata
and B, Galictis cuja. A, AMNH 76630, adult female from
Peru. B, AMNH 38983, adult female from Bolivia. Scale
bar = 10 cm.
464 R. BORNHOLDT ET AL.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
metaconid, we observed that the two extant grisons
were also very different with respect to size. The PCA
revealed that the split between G. cuja and G. vittata
occurred along PC1, highlighting size as the main
factor of overall cranial differences. All subsequent
analyses supported the importance of this factor in
the segregation between species, and in all cases
G. vittata was larger than G. cuja. When all analyses
were assessed together, we noticed some linear mea-
surements that contributed the most to such size
segregation: GLS, ZB, MB, BB, MAL, and c-m2. Of
these variables, GLS and MAL describe overall skull
size, yielding very clear differences between species.
The measurements ZB, MB, and BB are the most
important to describe general skull width, indicating
that G. vittata not only has a longer but also a
broader skull, leading to a more robust cranial design
than its congener (see Fig. 4).
Skin
Pelage variation may also be used to aid in species
diagnosis within the genus Galictis, although our
results indicate that it is not as clear-cut as the skull
characters. The marked difference in fur length/
density between G. cuja and G. vittata seems to be a
potentially reliable character to distinguish these
species, as all G. cuja had long and dense fur and all
G. vittata had a short coat. The same precision was
not observed with respect to fur colour, given the
intraspecific variation observed in G. cuja. Still, the
tendency of the latter species to bear yellowish fur
and of G. vittata to have greyish fur can contribute to
species diagnosis, but only when used in combination
with fur density/length. This may be especially rel-
evant in the context of visual (or photographic) iden-
tification of specimens in the field. Our results
indicate that it may be possible to perform reliable
Figure 6. Maximum likelihood (ML) phylogenies depicting the evolutionary relationships amongst Galictis cuja (bGcu or
Gcu sample IDs) and Galictis vittata (Gvi sample IDs) individuals (see Appendix S1 for more details on sample IDs). A,
mitochondrial DNA (nicotinamide adenine dinucleotide dehydrogenase subunit 5 gene segment) data set; B, concatenated
nuclear data set containing 12 different segments (see Table 1 for more information on nuclear segments). Values at
internodes indicate support based on ML/maximum parsimony/Bayesian inference methods, respectively (Bayesian
posterior probabilities are indicated as percentages). Support is represented only for major nodes. See text for additional
details.
TAXONOMIC REVISION OF GENUS GALICTIS 465
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
identification of Galictis species in cases where
overall body size (e.g. TL) can be assessed in combi-
nation with coat density/length, leaving pelage colora-
tion as a third and less clear-cut criterion. Another
potential feature that may distinguish these species is
tail length, as suggested by some previous authors
(e.g. Emmons & Feer, 1997; Yensen & Tarifa, 2003a,
b). We did observe that G. cuja tended to have a
proportionally longer tail (e.g. see Fig. 5), as previ-
ously suggested, but did not test this feature statis-
tically given the small sample size of complete
skeletons and potential biases induced by skin prepa-
rations. Further studies focusing on live individuals
whose reliable identification is available (e.g. based
on the molecular characters that we report here)
should help ascertain the error rate associated with
such field-orientated diagnosis strategies.
GEOGRAPHICAL DISTRIBUTION OF G. CUJA AND
G. VITTATA
The records obtained in this study cover the complete
range of the genus Galictis, from the northern limit of
the Neotropical region, in Mexico, to Patagonia in
Chile and Argentina [except for the southernmost
recorded points (for a review focusing on Patagonia,
see Prevosti & Travaini, 2005)]. The northernmost
limit of G. vittata corroborated the view provided in
the literature, i.e. the Mexican provinces of San Luis
Potosi and Veracruz (Wozencraft, 2005). This region is
the boundary between humid and semihumid forests
in southern Mexico to drier and more open regions in
northern Mexico (where no G. vittata were recorded).
This is an interesting distributional pattern because
it is very similar to that observed in the southernmost
limits of the species. The range of G. vittata extends
from mid-southern Mexico through Central America
into northern South America, where it occupies
tropical and subtropical forests, including the entire
Amazon basin. This pattern excludes adjacent
biomes, such as savannahs, deserts, and montane
grasslands, as well as most of the Atlantic Forest in
eastern South America (see Fig. 8). In our data set,
the latter biome is marginally represented by a single
record in southern Paraguay (AMNH 77695, skull
only, with specific locality given as ?east of Villarica?,
Guaira). Interestingly, Yensen & Tarifa (2003a) did
not include Paraguay in the range of G. vittata, prob-
ably because this single record was not analysed
in the literature sources reviewed by those authors.
This specimen did not contain gender information;
thus it was excluded from the morphometric analy-
ses. However, its large size (e.g. BB = 40.4 mm and
c-m2 = 33.96 mm) and the presence of the metaconid
in the lower carnassial identified it as G. vittata. In
addition to this individual, there has been a report of
a G. vittata specimen from Misiones province, north-
eastern Argentina (Chebez & Massoia, 1996; D?az &
Lucherini, 2006; F. Prevosti, pers. comm.). Although
this individual was not analysed in the present study,
its presence in the south-western edge of the Atlantic
Forest (see Fig. 8) lends support to the hypothesis
that G. vittata may indeed extend its range into this
area, implying potentially considerable geographical
overlap with G. cuja in the region. Further scrutiny of
the geographical limits of these species in this area is
warranted, so as to clarify their actual range overlap
and degree of habitat segregation when in sympatry.
Figure 7. Representative haplotype network for one nuclear gene segment (recombination activating protein 1, RAG1).
Each rectangle represents a distinct haplotype. Haplotypes sampled only in Galictis cuja are depicted in grey, whereas
that sampled only in Galictis vittata is shown in white. The number of copies sampled for each haplotype is indicated
inside the respective rectangle. Hatches across branches indicate mutational steps. The arrow indicates the position of
the root, based on comparison with two outgroup species (Poecilogale albinucha and Ictonyx striatus). Haplotype networks
for the remaining 11 nuclear segments, as well as the underlying sequence information, are given in Appendix S5.
466 R. BORNHOLDT ET AL.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
Figure 8. Map showing the distribution of Galictis vittata (squares) and Galictis cuja (circles) based on the geographical
origin of individuals with ascertained species-level identification. Galictis vittata records represented by red squares with
an inner black dot are based on skulls and skins analysed in this study. The record represented by a solid red square is
an individual collected in Misiones, Argentina (Chebez & Massoia, 1996), that was not analysed directly (see text for
details). Galictis cuja records represented by black circles with a white border are based on skulls and skins analysed
here, whereas those based on DNA samples are depicted as solid black circles. In addition, we also show open circles
representing G. cuja records reported by Prevosti & Travaini (2005) from the southernmost limit of the species? range. The
distributional points are overlaid on a map of Neotropical biomes (defined according to Olson et al., 2001) to allow a visual
assessment of the species? ranges. Records indicate distinct locations where specimens were found (i.e. repeated
coordinates were collapsed into a single point).
TAXONOMIC REVISION OF GENUS GALICTIS 467
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
In addition to Paraguay and north-easternArgentina,
other potential areas of distributional overlap may
include southern Peru, central Brazil, and portions of
Bolivia. The observed range of these taxa in Peru was
consistent with literature sources (e.g. Pacheco et al.,
1995; Eisenberg & Redford, 1999; Yensen & Tarifa,
2003b; Larivi?re & Jennings, 2009), i.e. G. cuja occur-
ring exclusively in the extreme south, associated with
the Andes, and G. vittata occurring throughout the
tropical forests northward. In Brazil, very few data
points were available from the Cerrado biome, ham-
pering any in-depth assessment of the exact boundary
between these species in this region. However, some
records from eastern Amazonia (Maranh?o state) indi-
cate that G. cuja and G. vittata do coexist in this region
(Oliveira, 2009), although the geographical extent of
this overlap is presently unknown. In Bolivia, Cu?llar
& Noss (2003) reported both species in the south, but
our results extend northward the range of G. cuja,
leading to a distributional pattern similar to that
presented by Yensen & Tarifa (2003b) and Eisenberg &
Redford (1999) (see Fig. 8).
In spite of these remaining uncertainties, our results
clarified the southern and eastern limits of the G. vit-
tata distribution and allowed us to provide a more
precise geographical perspective of its range. Our data
indicate that G. vittata is adapted to tropical forests in
Central and South America rather than dry and open
landscapes or high-elevation vegetation associated with
colder temperatures. We examined a total of 67 speci-
mens of G. vittata, whose identification was confirmed
based on morphological characters. We found no evi-
dence corroborating the occurrence of G. vittata in
north-eastern, south-eastern, or southern Brazil, in
contrast to species lists and range maps reported in
previous studies (e.g. da Fonseca et al., 1996; Emmons
& Feer, 1997; Eisenberg & Redford, 1999; Guedes
et al., 2000; Briani et al., 2001; Yensen & Tarifa, 2003a;
Cherem et al., 2004; Larivi?re & Jennings, 2009). Such
a clarification is important, as it indicates that G. cuja
is the only grison species occurring throughout eastern
Brazil, where it occupies a wide variety of biomes,
including the Atlantic Forest, the Cerrado, and the
Caatinga.
Overall, our analysis of the G. cuja distribution
produced a map that was very similar to that
reported by Yensen & Tarifa (2003b), but different
from those presented by other authors (e.g. Eisenberg
& Redford, 1999), mainly with respect to the occur-
rence of this species in north-eastern Brazil. The
presence of G. cuja in this region has been controver-
sial and/or poorly documented, with some reference
maps ignoring its occurrence in the area (e.g. Lariv-
i?re & Jennings, 2009), whereas other sources (e.g. de
Freitas & Silva, 2005; Oliveira, 2009) reported that it
does exist in at least some of the included biomes,
such as the Caatinga. Our data corroborate this view,
and provide conclusive evidence that G. cuja indeed
occurs throughout north-eastern Brazil.
EVOLUTIONARY CONSIDERATIONS
The two species of Galictis are segregated mainly by
size, with G. vittata being consistently larger and
G. cuja smaller. It is thus interesting to hypothesize
about the evolutionary pressures that have shaped
such a size-based distinction between them. Consid-
ering that the two grison species do not show exten-
sive range overlap (see Fig. 8 and Appendix S6),
which could induce and maintain pervasive character
displacement between them, we postulate that other
evolutionary processes underlie this observed pattern
of size segregation. One possibility is that their geo-
graphical ranges overlapped much more in the past
than they do today, implying that their size distinc-
tion was indeed generated by character displacement
between them in sympatry, followed by more recent
range shifts in one or both species (Davies et al.,
2007). Another hypothesis is that their size difference
is mostly influenced by trophic competition with other
living or extinct mustelids, and not with each other.
Amongst the extinct taxa that might have driven such
size evolution are other species of the genus Galictis,
whose presence is documented in the South American
fossil record (e.g. Reig, 1957; Prevosti & Soibelzon,
2012). Given the age of the separation between
G. cuja and G. vittata (see above), and the concomi-
Table 6. Summary of nonmetric (skull and external
aspects) and metric (craniodental and external measure-
ments) morphological features that distinguish Galictis
cuja from Galictis vittata. Six linear measurements were
of major importance in the diagnosis of these species (see
main text for additional details), but here we only include
those that show no overlap between the two species
Character G. cuja G. vittata
Metaconid Absent Present
Dominant pelage colour Usually
yellowish
Greyish
General pelage
appearance
Dense/long fur Short fur
Greater length of skull
(mm)
 63.4?77.1  79.1?92.9
Zygomatic breath (mm)  34.1?44.0  44.7?54.5
Braincase breath (mm)  30.3?37.9  38.3?43.0
 29.7?36.5  37.3?42.3
Mandible length (mm)  36.0?45.3  46.7?58.3
Length of mandible
toothrow (mm)
 21.7?29.1  30.3?35.7
 20?27.4  28.5?35.3
Total body length (mm)  525?657  700?755
468 R. BORNHOLDT ET AL.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
tant existence of these congeneric species throughout
the late Pliocene and Pleistocene, it is plausible to
hypothesize that they may have played a competitive
role in shaping the size of the two living lineages. A
third hypothesis would postulate that the differences
in size and geographical range between G. cuja and
G. vittata are induced by ecological sorting, with each
species being adapted to a distinct set of environ-
ments. Galictis vittata seems to be rather restricted to
humid rain forests, whereas G. cuja occurs in a much
broader array of habitats (see below). Adaptation to
these different environments and their constituent
prey (along with competition with other living or
extinct carnivore species within these different habi-
tats) could have acted in concert to shape the
observed body-size patterns. In addition, both species
might compete for resources at the boundary of their
ranges, which would inhibit pervasive geographical
overlap (Davies et al., 2007). Testing these hypotheses
with diverse approaches should provide an interest-
ing avenue of research in the future.
Another aspect that could be explored in future
evolutionary studies targeting Galictis is coat colour
variation. Galictis cuja, which usually exhibits yellow-
ish and dense fur, occurs mainly in open and drier
landscapes, such as the barren Caatinga and the
Cerrado in north-eastern Brazil, the savannahs and the
grasslands throughout Argentina, and the dry coast of
Uruguay. It could be hypothesized that the yellowish fur
provides an amber appearance that could be favoured as
camouflage in such landscapes. However, G. cuja also
occupies forests, and adaptation to different habitats
might have led to its observed variation in coat colour.
Additionally, this species often reaches high altitudes
(Osgood, 1943; Greer, 1966; M. Lucherini, C. Tellaeche,
J. Reppucci & E. Luengos Vidal, unpubl. data) and
latitudes (Quintana, Ya?ez & Valdebenito, 2000;
Parera, 2002; Prevosti & Travaini, 2005), which may
have led historically to selective pressures for denser
pelage. In contrast, G. vittata exhibits more constant
skin colour and density. The mixture between black
and white fur, producing a pale grey pelage, might be
favoured in dense vegetation and darker landscapes,
such as tropical forests, the main type of biome occu-
pied by G. vittata. Additionally, the shorter and
sparser fur in this species is likely to be an adaptation
to the warm temperatures that are prevalent through-
out its geographical range (see Fig. 8). As in the case of
the size-based differences, in-depth ecological work is
required to test these hypotheses, so as to shed light
onto the evolutionary processes that have shaped these
phenotypes. As the first step towards this goal would
be to robustly delimit and diagnose these species, as
well as to define better their geographical range, the
present study should contribute to establishing such
baseline aspects, and to identifying patterns that have
the potential to spur additional research into these
little-known carnivorans.
ACKNOWLEDGEMENTS
The authors wish to thank the people in charge of the
mammalian collections analysed in this study, who
made this research project possible: Carla S. Fontana
(MCT-PUCRS), M?rcia A. Jardim (FZB/RS), Felipe
Peters (ULBRA), Maur?cio Graipel (LAMAq-UFSC),
Mario de Vivo (MZUSP), Jo?o A. Oliveira (MNHN),
Diego A. de Moraes (UFPE), Jo?o Alberto de Queiroz
(MPEG), Enrique Gonz?lez (MNHNA), David Flores
(MACN), Hugo L?pez (MLP), Linda Gordon (USNM),
Eileen Westwig (AMNH), Ned Gilmore (ANSP),
Catherine Weisel (MCZ), Bill Stanley and Tracy
Damitz (FMNH), Kristof Zyskowski (YPM), Robert
Timm (NHMKU), Chris Conroy (MVZ), Mark Hafner
(LSUMNS), Louise Tomsett and Paula Jenkins
(BMNH), and Clara Stefen (SMT). We also gratefully
acknowledge Cladinara Sarturi and Priscilla Mena
Zamberlan (PUCRS, Brazil), as well as Jessica Tomlin
and Hannah Zinnert (both supported through the
Werner H. Kirsten Student Intern Program at the
NCI, Frederick, Maryland, USA) for excellent labora-
tory assistance. We also thank all members from the
Grupo de Ecolog?a Comportamental de Mam?feros
(GECM) who participated in sample collection in
Argentina, especially Estela Luengos Vidal, Diego
Castillo, Claudia Manfredi, and Emma B. Casanave.
Finally, we are especially grateful to Cristine S.
Trinca, Andr?a T. Thomaz, Jo?o Feliz, Henrique V.
Figueir?, Lucas G. da Silva, and Tiago F. da Silva for
support at various stages of this research. This
project was supported by the Pontif?cia Universidade
Cat?lica do Rio Grande do Sul (PUCRS, Brazil),
Coordena??o de Aperfei?oamento de Pessoal de N?vel
Superior (CAPES, Brazil), Funda??o de Amparo ?
Pesquisa do Estado do Rio Grande do Sul (FAPERGS,
Brazil), and the Smithsonian Institution (USA).
REFERENCES
Amrine-Madsen H, Koepfli K-P, Wayne RK, Springer
MS. 2003. A new phylogenetic marker, apolipoprotein B,
provides compelling evidence for eutherian relationships.
Molecular Phylogenetics and Evolution 28: 225?240.
Bell T. 1826. Zoological Club, 10 January 1826. The Zoologi-
cal Journal 2: 548?554.
Bell T. 1837. Observations on the genus Galictis, and the
description of a new species (Galictis allamandi). Proceed-
ings of the Zoological Society of London 2: 201?206.
Briani DC, Santori RT, Vieira MV, Gobbi N. 2001. Mam?f-
eros n?o-voadores de um fragmento de mata mes?fila
semidec?dua, do interior do Estado de S?o Paulo, Brasil.
Holos Environmental 1: 141?149.
Brouillette JA, Andrew JR, Venta PJ. 2000. Estimate of
TAXONOMIC REVISION OF GENUS GALICTIS 469
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
nucleotide diversity in dogs with a pool-and-sequence
method. Mammalian Genome 11: 1079?1086.
Buffon MC. 1776. Histoire Naturelle g?n?rale et particuli?re:
servant de suite ? l?histoire des Animaux quadrup?des.
Paris: De L?imprimerie Royale.
Canevari M, Vaccaro O. 2007. Gu?a de mam?feros del sur de
Am?rica del Sur. Buenos Aires: Literature of Latin America.
del Cerro I, Marmi J, Ferrando A, Chashchin P, Taber-
let P, Bosh M. 2010. Nuclear and mitochondrial phylog-
enies provide evidence for four species of Eurasian badger
(Carnivora). Zoologica Scripta 39: 415?425.
Chebez JC, Massoia E. 1996. Mamiferos de la Provincia de
Misiones. In: Chebez JC, ed. Fauna Misionera, Catalogo
Sistematico y Zoogeografico de los Vertebrados de la Pro-
vincia de Misiones (Argentina). Buenos Aires: Literature of
Latin America. 180?206.
Cherem JJ, Sim?es-Lopes PC, Althoff S, Graipel ME.
2004. Lista dos mam?feros do estado de Santa Catarina, sul
do Brasil. Mastozoolog?a Neotropical 11: 151?184.
Cione AL, Tonni EP. 1995. Chronostratigraphy and ?Land-
Mammal Ages? in the cenozoic of southern South America:
principles, practices, and the ?Uquian? problem. Journal of
Paleontology 69: 135?159.
Collen B, Purvis A, Gittleman JL. 2004. Biological corre-
lates of description date in carnivores and primates. Global
Ecology and Biogeography 13: 459?467.
Cuar?n AD, Reid F, Helgen KM. 2008. Galictis vittata. In:
IUCN, 2011. IUCN red list of threatened species. Version
2011.2. Available at: http://www.iucnredlist.org
Cu?llar E, Noss A. 2003. Mam?feros del Chaco y de la
Chiquitania de Santa Cruz, Bolivia. Santa Cruz de la
Sierra: Editorial FAN.
Davies TJ, Meiri S, Barraclough TG, Gittleman JL. 2007.
Species co-existence and character divergence across carni-
vores. Ecology Letters 10: 146?152.
Dayan T, Simberloff D. 1994. Character displacement,
sexual dimorphism, and morphological variation among
British and Irish mustelids. Ecology 75: 1063?1073.
Dayan T, Simberloff D, Tchernov E, Yom-Tov Y. 1989.
Inter- and intraspecific character displacement in mustel-
ids. Ecology 70: 1526?1539.
Degnan JH, Rosenberg NA. 2009. Gene tree discordance,
phylogenetic inference and the multispecies coalescent.
Trends in Ecology and Evolution 24: 332?340.
DesmarestMAG.1820.Mammalogie ou description des esp?ces
de mammif?res. Paris: Veuve Agasse, Imprimeur Libraire.
D?az MM, Lucherini M. 2006. Familias Mephitidae, Mus-
telidae, Procyonidae. In: Barquez RM, D?as MM, Ojeda RA,
eds. Mam?feros de Argentina, sistem?tica y distribuci?n.
Tucum?n: SAREM, 100?107.
Drummond AJ, Ashton B, Buxton S, Cheung M,
Cooper A, Heled J, Kearse M, Moir R, Stones-Havas S,
Sturrock S, Thierer T, Wilson A. 2010. Geneious v5.3.
Available at: http://www.geneious.com
Eisenberg JF. 1989. Mammals of the Neotropics, the north-
ern Neotropics, volume 1: Panam?, Colombia, Venezuela,
Guyana, Suriname, French Guiana. Chicago and London:
The University of Chicago Press.
Eisenberg JF, Redford KH. 1999. Mammals of the Neotro-
pics, the central Neotropics, volume 3: Ecuador, Peru, Bolivia,
Brazil.Chicagoand London:The University of Chicago Press.
Eizirik E. 2012. A molecular view on the evolutionary history
and biogeography of Neotropical carnivores (Mammalia, Car-
nivora). In: Patterson BD, Costa LP, eds. Bones, clones, and
biomes: an extended history of recent Neotropical mammals.
Chicago: The University of Chicago Press, 123?142.
Emmons LH, Feer F. 1997. Neotropical rainforest mammals:
a field guide. Chicago: The University of Chicago Press.
ESRI. 2009. ArcGIS desktop: release 9.3. Redlands, CA: Envi-
ronmental Systems Research Institute.
Ewing B, Hillier L, Wendl MC, Green P. 1998. Base-calling
of automated sequencer traces using phred. I. Accuracy
assessment. Genome Research 8: 175?185.
da Fonseca GAB, Herrmann G, Leite YLR, Mittermeier
RA, Rylands AB, Patton JL. 1996. Lista Anotada dos
Mam?feros do Brasil. Occasional Papers in Conservation
Biology 4: 1?38.
de Freitas MA, Silva TFS. 2005. Guia Ilustrado: Mam?feros
na Bahia, esp?cies continentais. Pelotas: Editora USEB.
Gay C. 1847. Historia fisica y politica de Chile, Zoologia.
Paris: Imprenta de Maulde y Renou.
Goldman EA. 1920. Mammals of Panama. Smithsonian Mis-
cellaneous Collections 69.
Goodman SM, Helgen KM. 2010. Species limits and distri-
bution of the Malagasy carnivoran genus Eupleres (Family
Eupleridae). Mammalia 74: 177?185.
Goodwin GG. 1946. Mammals of Costa Rica. Bulletin of the
American Museum of Natural History 87: 271?474.
Gordon D, Abajian C, Green P. 1998. Consed: a graphical
tool for sequence finishing. Genome Research 8: 195?202.
Gray JE. 1865. Revision of the genera and species of Mus-
telidae contained in the British Museum. Proceedings of the
Zoological Society of London 1865: 100?154.
Gray JE. 1869. Catalogue of carnivorous, pachydermatous
and edentate Mammalia in the British Museum. London:
Taylor and Francis.
Greer JK. 1966. Mammals of Malleco Province Chile. Publi-
cations of the Museum, Michigan State University Biological
Series 3: 49?152.
Guedes PG, da Silva SSP, Camardella AR, de Abreu
MFG, Borges-Nojosa DM, da Silva JAG, Silva AA. 2000.
Diversidade de mam?feros do Parque Nacional de Ubajara
(Cear?, Brasil). Mastozoolog?a Neotropical 7: 95?100.
Harding LE, Smith FA. 2009. Mustela or Vison? Evidence
for the taxonomic status of the American mink and a dis-
tinct biogeographic radiation of American weasels. Molecu-
lar Phylogenetics and Evolution 52: 632?642.
Hasegawa M, Kishino H, Yano T. 1985. Dating of the
human-ape splitting by a molecular clock of mitochondrial
DNA. Journal of Molecular Evolution 22: 160?174.
Helgen KM, Kays R, Helgen LE, Tsuchiya-Jerep MTN,
Pinto CM, Koepfli K-P, Eizirik E, Maldonado JE.
2009. Taxonomic boundaries and geographic distributions
revealed by an integrative systematic overview of the moun-
tain coatis, Nasuella (Carnivora: Procyonidae). Small Car-
nivore Conservation 41: 65?74.
470 R. BORNHOLDT ET AL.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
Helgen KM, Lim NT-L, Helgen LE. 2008. The hog-badger is
not an edentate: systematics and evolution of the genus
Arctonyx (Mammalia: Mustelidae). Zoological Journal of the
Linnean Society 154: 353?385.
Housley DJE, Zalewski ZA, Beckett SE, Venta PJ. 2006.
Design factors that influence PCR amplification success of
cross-species primers among 1147 mammalian primer pairs.
BMC Genomics 7: 253?263.
Huelsenbeck JP, Ronquist F. 2001. Mrbayes: Bayesian
inference of phylogenetic trees. Bioinformatics 17: 754?
755.
Hunt RM Jr. 1996. Biogeography of the order Carnivora. In:
Gittleman JL, ed. Carnivore behavior, ecology, and evolu-
tion. Ithaca and London: Cornell University Press, 485?541.
Husson AM. 1978. The mammals of Suriname. Leiden: E.J.
Brill.
Ihering H. 1911. Os mam?feros do Brasil meridional, Car-
nivora: Felidae, Canidae, Procyonidae, Mustelidae. Revista
do Museu Paulista 8: 147?272.
International Commission on Zoological Nomenclature
(ICZN). 1956. Rejection for nomenclatorial purposes of
volume 3 (Zoologie) of the work by Lorenz Oken entitled
Okens Lehrbuch de Naturgeschichte published in 1815?
1816. Opinions and Declarations Rendered by the Interna-
tional Commission on Zoological Nomenclature 14: 1?42.
Jacques H, Veron G, Alary F, Aulagnier S. 2009. The
Congo clawless otter (Aonyx congicus) (Mustelidae: Lutri-
nae): a review of its systematics, distribution and conserva-
tion status. African Zoology 44: 159?170.
Koepfli K-P, Deere K, Slater GJ, Begg C, Begg K, Grass-
man L, Lucherini M, Veron G, Wayne RK. 2008. Multi-
gene phylogeny of the Mustelidae: resolving relationships,
tempo and biogeographic history of a mammalian adaptive
radiation. BMC Biology 6: 10.
Larivi?re S, Jennings AP. 2009. Family Mustelidae
(weasels and relatives). In: Wilson DE, Mittermeier RA, eds.
Handbook of the mammals of the world. 1. Carnivores.
Barcelona: Lynx Edicions, 564?656.
Mares MA, Willig MR, Streilein KE, Lacher TE Jr. 1981.
The mammals of Northeastern Brazil: a preliminary assess-
ment. Annals of the Carnegie Museum of Natural History
50: 81?137.
Miller B, Ralls K, Reading RP, Scott JM, Estes J. 1999.
Biological and technical considerations of carnivore trans-
location: a review. Animal Conservation 2: 59?68.
Molina GI. 1782. Saggio sulla storia naturale del Chili.
Bologna: Stamperia di San Tommaso d.
Monakhov VG. 2009. Is sexual dimorphism variable? Data
on species of the genus Martes in the Urals. Biology Bulletin
36: 45?52.
Murphy WJ, Eizirik E, Johnson WE, Zhang Y-P, Ryder
OA, O?Brien SJ. 2001. Molecular phylogenetics and the
origin of placental mammals. Nature 409: 614?618.
Nowak RM. 1991. Walker?s mammals of the world, 5th edn.
Baltimore: The Johns Hopkins University Press.
Oken L. 1816. Lehrbuch der Naturgeschichte (Zoologie). Jena:
August Schmidt und Comp.
Oliveira TG. 2009. Notes on the distribution, status, and
research priorities of little-known small carnivores in
Brazil. Small Carnivore Conservation 41: 22?24.
Olson DM, Dinerstein E, Wikramanayake ED,
Burgess ND, Powell GVN, Underwood EC, Damico
JA, Itoua I, Strand HE, Morrison JC, Loucks CJ,
Allnutt TF, Ricketts TH, Kura Y, Lamoreux JF, Wet-
tengel WW, Hedao P, Kassem KR. 2001. Terrestrial
ecoregions of the world: a new map off life on Earth. Bio-
science 51: 933?938.
Osgood WH. 1943. The mammals of Chile. Field Museum of
Natural History, Zoological Series 30: 1?268.
Pacheco V, Vivar HME, Ascorra C, Arana-Card? R,
Solari S. 1995. Lista Anotada de los mam?feros peruanos.
Occasional Papers in Conservation Biology 2: 1?35.
Parera A. 2002. Los mamiferos de la Argentina y la regi?n
austral de Sudam?rica. Buenos Aires: El Ateneo.
Patterson BD. 2000. Patterns and trends in the discovery of
new Neotropical mammals. Diversity and Distributions 6:
145?151.
Patterson BD. 2001. Fathoming tropical biodiversity: the
continuing discovery of Neotropical mammals. Diversity and
Distributions 7: 191?196.
Posada D, Crandall KA. 1998. Modeltest: testing the model
of DNA substitution. Bioinformatics 14: 817?818.
Prevosti FJ, Soibelzon LH. 2012. Evolution of the South
American carnivores (Mammalia, Carnivora): a paleonto-
logical perspective. In: Patterson BD, Costa LP, eds. Bones,
clones, and biomes: an extended history of recent Neotropical
mammals. Chicago: The University of Chicago Press, 102?
122.
Prevosti FJ, Travaini A. 2005. New records of Galictis cuja
(Molina, 1782) (Carnivora, Mustelidae) in Southern Patago-
nia. Mammalian Biology 70: 317?320.
Quintana V, Ya?ez J, Valdebenito M. 2000. Orden Car-
nivora. In: Pedreros AM, Valenzuela JY, eds. Mamiferos de
Chile. Valdivia: Ediciones CEA, 155?187.
Redford KH, Eisenberg JF. 1992. Mammals of the Neotro-
pics, the southern cone, volume 2: Chile, Argentina, Uruguay
and Paraguay. Chicago and London: The University of
Chicago Press.
Reeder DM, Helgen KM, Wilson DE. 2007. Global trends
and biases in new mammal species discoveries. Occasional
Papers 269: 1?36.
Reid F, Helgen KM. 2008. Galictis cuja. In: IUCN, 2011.
IUCN red list of threatened species. Version 2011.2. Avail-
able at: http://www.iucnredlist.org
Reig OA. 1957. Un mustelideo del genero Galictis del eoc-
uartario de la provincial de Buenos Aires. Ameghiniana 1:
33?47.
Rozhnov VV, Abramov AV. 2006. Sexual dimorphism of
marbled polecat Vormela peregusna (Carnivora: Mustel-
idae). Biology Bulletin 33: 144?148.
Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular
cloning: a laboratory manual, 2nd edn. New York: Cold
Spring Harbor Laboratory Press.
Sato JJ, Wolsan M, Prevosti FJ, D?El?a G, Begg C, Begg
K, Hosoda T, Campbell KL, Suzuki H. 2012. Evolution-
ary and biogeographic history of weasel-like carnivorans
TAXONOMIC REVISION OF GENUS GALICTIS 471
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472
(Musteloidea). Molecular Phylogenetics and Evolution 63:
745?757.
Schipper J et al. 2008. The status of the world?s land and
marine mammals: diversity, threat, and knowledge. Science
322: 225?230.
Schreber JCD. 1776. Die S?ugthiere in Abbildungen nach der
Natur mit Beschreibungen. Erlangen: Wolfgang Walther.
Shaw G. 1800. General zoology or Systematics natural history
(Mammalia). London: Printed for G. Kearsley.
Swofford DL. 2002. PAUP*: phylogenetic analysis using par-
simony (*and others methods), version 4. Sunderland, MA:
Sinauer Associates.
Syring J, Farrell K, Businsk? R, Cronn R, Liston A.
2007. Widespread genealogical nonmonophyly in species
of Pinus Subgenus Strobus. Systematic Biology 56: 163?181.
Tamura K, Dudley J, Nei M, Kumar S. 2007. MEGA4:
Molecular Evolutionary Genetics Analysis (MEGA) software
version 4.0. Molecular Biology and Evolution 24: 1596?1599.
Teeling EC, Scally M, Kao DJ, Romagnoli ML, Springer
MS. 2000. Molecular evidence regarding the origin of
echolocation and flight in bats. Nature 403: 188?192.
Thom MD, Harrington LA, Macdonald DW. 2004. Why
are American mink sexually dimorphic? A role for niche
separation. Oikos 105: 525?535.
Thomas O. 1903. Notes on South-American monkeys, bats,
carnivores, and rodents, with descriptions of new species.
Annals and Magazine of Natural History 12: 455?464.
Thomas O. 1907. On Neotropical mammals of the genera
Callicebus, Reithrodontomys, Ctenomys, Dasypus, and Mar-
mosa. Annals and Magazine of Natural History 20: 161?168.
Thomas O. 1912. Small mammals from South America.
Annals and Magazine of Natural History 10: 44?48.
Thomas O. 1921. The ?huron? of the Argentine. Annals and
Magazine of Natural History 8: 212?213.
Thunberg CP. 1820. Ursus brasiliensis, nova quaedam
species, descripta et delineata. M?moires de l?Acad?mie
Imp?riale des Sciences de St. P?tersbourg 1820: 400?402.
Traill TS. 1821. Account of the Lutra vittata, and of the
Viverra poliocephalus. Memoirs of the Wernerian Natural
History Society 3: 437?441.
Trigo TC, Freitas TRO, Kunzler G, Cardoso L, Silva
JCR, Johnson WE, O?Brien SJ, Bonatto SL, Eizirik E.
2008. Inter-species hybridization among Neotropical cats of
the genus Leopardus, and evidence for an introgressive
hydrid zone between L. geoffroyi and L. tigrinus in southern
Brazil. Molecular Ecology 17: 4317?4333.
Valenzuela-Galv?n D, Arita HT, Macdonald DW. 2007.
Conservation priorities for carnivores considering protected
natural areas and human population density. Biodiversity
and Conservation 17: 539?558.
V?li U, Einarsson A, Waits L, Ellegren H. 2008. To what
extent do microsatellite markers reflect genome-wide
genetic diversity in natural populations? Molecular Ecology
17: 3808?3817.
Venta PJ, Brouillette JA, Yuzbasiyan-Gurkan V, Brewer
GJ. 1996. Gene specific universal mammalian sequence-
tagged sites: application to the canine genome. Biochemical
Genetics 34: 321?341.
Wilson EO. 1992. The diversity of life. Cambridge: Belknap
Press of Harvard University.
Wilson EO. 2005. Systematics and the future of biology.
Proceedings of the National Academy of Sciences, USA 102:
6520?6521.
Wolsan M, Sato JJ. 2010. Effects of data incompleteness
on the relative performance of parsimony and Bayesian
approaches in a supermatrix phylogenetic reconstruction of
Mustelidae and Procyonidae (Carnivora). Cladistics 26: 168?
194.
Woodburne MO. 2010. The Great American Biotic Inter-
change: dispersals, tectonics, climate, sea level and holding
pens. Journal of Mammalian Evolution 17: 245?264.
Wozencraft WC. 2005. Order Carnivora. In: Wilson DE,
Reeder DR, eds. Mammal species of the world: a taxonomic
and geographic reference, 3th edn. Baltimore: Johns
Hopkins University Press, 532?628.
Yensen E, Tarifa T. 2003a. Galictis vittata. Mammalian
Species 727: 1?8.
Yensen E, Tarifa T. 2003b. Galictis cuja. Mammalian
Species 728: 1?8.
SUPPORTING INFORMATION
Additional Supporting Information may be found in the online version of this article at the publisher?s web-site:
Appendix S1. Galictis specimens used in the morphological, genetic and geographic analyses listed per species
per institution.
Appendix S2. Results of the t-test comparison for sexual size dimorphismwithin Galictis cuja andwithin G. vittata.
Appendix S3. GenBank accession numbers for previously published mitochondrial and nuclear sequences of
two mustelids used as outgroups in the current study (Ictonyx striatus and Poecilogale albinucha).
Appendix S4. Classification coefficients from the classification analyses for each Galictis species (G. cuja and
G. vittata) extracted from the two-group and multiple-group discriminant function analysis (DFA).
Appendix S5. A. Table depicting the variable sites found within and between Galictis cuja and G. vittata in 12
nuclear gene segments.
B. Haplotype networks derived from the nucleotide information described in A for 11 of the 12 nuclear segments.
Appendix S6. Geographic distribution map of Galictis vittata and G. cuja throughout the Neotropical region
based on skulls, skins and DNA samples.
Appendix S7. Results from the multiple-group Discriminant Function Analysis (DFA) applying the stepwise
method.
472 R. BORNHOLDT ET AL.
? 2013 The Linnean Society of London, Zoological Journal of the Linnean Society, 2013, 167, 449?472