Content uploaded by Fernando Abdala
Author content
All content in this area was uploaded by Fernando Abdala
Content may be subject to copyright.
Cranial ontogeny of Caluromys philander (Didelphidae:
Caluromyinae): a qualitative and quantitative approach
DAVID A. FLORES,* FERNANDO ABDALA,AND NORBERTO GIANNINI
Museo Argentino de Ciencias Naturales ‘‘Bernardino Rivadavia,’’ Av. A
´ngel Gallardo 470, CP 1405, Ciudad de Buenos
Aires, Argentina (DAF)
CONICET. Consejo Nacional de Investigaciones Cientı
´ficas y Te´cnicas, Argentina (DAF, NG)
Bernard Price Institute for Palaeontological Research, Private Bag 3 WITS 2050, University of the Witwatersrand,
Johannesburg, South Africa (FA)
Division of Vertebrate Zoology (Mammalogy), American Museum of Natural History; Programa de Investigaciones de
Biodiversidad Argentina, Universidad Nacional de Tucuma´n, Argentina (NG)
* Correspondent: dflores@macn.gov.ar
The ontogeny of skull allometry has been the subject of research in didelphid, microbiotheriid, and dasyurid
marsupials. We described and compared postweaning stages of cranial development in the woolly opossum
Caluromys philander, a member of a distinct lineage of didelphids. We identified 31 qualitative morphological
changes between juveniles and adults, many related to the trophic apparatus. Early development of the
paracanine fossa in the snout, correlated with the presence of well-developed canines in juveniles, represents a
remarkable difference between the ontogeny of C. philander and other marsupials. We performed bivariate and
multivariate analyses of allometry on a data set composed of 14 cranial measurements and 45 specimens.
Allometric scaling showed that braincase is relatively small in adulthood, whereas other neurocranial
components, particularly the orbit, grew isometrically along its length. The palate becomes elongated by the
combination of the isometry of its length and the negative allometry of its width. The rostrum lengthened, but its
height remained proportionally constant. In relative terms the temporal fossa expanded inward following the
marsupial model that combines the negative allometry of the braincase and the isometry of the zygomatic
breadth. The mandible increased in robustness. The skull ontogeny of C. philander shares 4 allometric trends
with 2 didelphids (Didelphis albiventris and Lutreolina crassicaudata) and two australidelphians (Dromiciops
gliroides and Dasyurus albopunctatus). Three of these trends, breadth of palate, breadth of braincase, and height
of occipital plate, were allometrically negative, whereas height of the dentary was positive. These common
allometric trends may be interpreted as symplesiomorphic for marsupials, suggesting a high degree of
conservatism in the skull development of Caluromys and other marsupials. DOI: 10.1644/09-MAMM-A-291.1.
Key words: anatomy, Caluromyinae, ontogeny, skull allometry
E2010 American Society of Mammalogists
Recent studies on cranial ontogeny in several groups of
marsupials were focused on functional interpretations from
both a qualitative and quantitatively approach (Abdala et al.
2001; Flores et al. 2003, 2006; Giannini et al. 2004). Such
studies detected complex modifications occurring in the skull
during postweaning ontogeny, accompanying the critical
changes of diet (from milk-fed to the adult trophic function)
during this period. Ontogenetic pattern of 2 related amer-
idelphians (the didelphids Didelphis albiventris and Lutreolina
crassicaudata) and 2 australidelphians (the microbiotheriid
Dromiciops gliroides and the dasyurid Dasyurus albopuncta-
tus—Abdala et al. 2001; Flores et al. 2003, 2006; Giannini et
al. 2004) showed some cranial trends common to all taxa and
presumably to the groups to which they belong, whereas other
trends appeared clearly distinct in each of the studied taxa.
The latter probably represent ontogenetic characteristics of
different marsupial lineages. Interactions between growth of
neurocranial and splanchnocranial components modify func-
tions of the skull linked to the transition to the specific diet of
adults, through developmental stages. In a comparative sense
the common patterns of skull growth in postweaning stages in
taxa studied to date may represent plesiomorphies in the
www.mammalogy.org
Journal of Mammalogy, 91(3):539–550, 2010
539
marsupial clade (e.g., isometry of the muzzle height, negative
allometry of the breadth of braincase and the height of
occipital plate, and positive allometry of the height of the
mandible); i.e., a common developmental pattern inherited
from ancestors at the marsupial root.
On the basis of comparisons of the 2 didelphid species
analyzed to date, we advanced the existence of general cranial
growth trends in the family (Flores et al. 2003). For instance,
the basic ontogenetic plan of skull growth in Didelphinae
would consist of an interaction between neurocranial and
splachnocranial growth trends that provide a wide space for
the large temporal muscles (Flores et al. 2003). Thus the
detection of patterns of cranial growth in groups closely
related to didelphines, such as caluromyines, is essential to
assess a common plan in the growth pattern of didelphids in a
more inclusive phylogenetic context.
Caluromys, along with Caluromysiops and Glironia, is one
of the most basal groups in the Didelphidae. An apparent
consensus exists on its phylogenetic position, on the basis
either of morphological data (Flores 2003, 2009), DNA–DNA
hybridization data (Kirsch et al. 1995), DNA sequence data
(Jansa and Voss 2000; Palma and Spotorno 1999), or a
combination thereof (Creighton 1984; Flores 2009; Reig et al.
1987; Voss and Jansa 2003, 2009). All studies agree in the
significant divergence between caluromyines and the remain-
der of Didelphidae. Thus the comparative study of develop-
ment in this taxon is of special importance for understanding
the evolution of the family (Astu´a de Moraes and Leiner
2008). The postweaning cranial development observed in
Caluromys philander, and those of didelphids, dasyurids, and
microbiotheriids, allowed us to test and expand previous
hypotheses of cranial ontogeny in basal marsupials and
improve our prior generalizations made for Didelphidae
(Flores et al. 2003).
MATERIALS AND METHODS
Caluromys philander is a common marsupial of neotropical
rainforests, with well-developed arboreal habits (Atramento-
wicz 1982, 1986, 1995; Charles-Dominique et al. 1981;
Julien-Laferriere 1991; Julien-Laferriere and Atramentowicz
1990). As are most didelphids, the species is omnivorous, with
some marked preference for fruits (Astu´a de Moraes et al.
2003; Charles-Dominique et al. 1981; Julien-Laferriere 1999;
Julien-Laferriere and Atramentowicz 1990; Leite et al. 1996).
Caluromys philander is one of the three species of woolly
opossums living in Central and South America, with body
mass of adult males and females ranging from 300 to 400 g
(Atramentowicz 1995). Breeding patterns and correlation
between absolute age and tooth eruption sequence are well
known (Astu´a de Moraes and Leiner 2008; Atramentowicz
1986). At birth each young is approximately 0.2 g and remains
attached to the teat for about 80 days. Young start eating solid
food at this stage but still depend on milk until approximately
day 120 postpartum, when young become independent
(Atramentowicz 1982, 1995).
Study specimens.—We studied an ontogenetic series of 45
specimens of C. philander of widely different ages housed in
the mammal collection of the American Museum of Natural
History (AMNH). The specimens examined were AMNH
5943-50, 5953, 16955, 16967, 76969, 93529, 94895, 94898,
95525, 95527-8, 95530-5, 95537, 95972, 96623, 96631,
96640, 96650, 96651-2, 139815-6, 169668, 169762, 171380,
174001, 174004, 174006, 203351-2, 208994-5, and 234989.
We focused the study of qualitative changes on the contrast
between juveniles and adults, but we also observed subadults
for anatomical transformations that occur throughout the range
of ages available.
In our sample 14 specimens were collected before achieving
the stage of full adult dentition. Other individuals were adults
of different sizes. The smallest specimen (AMNH 16955, sex
undetermined, condylobasal length of 31.6 mm) has the third
lower (m3) and second upper (M2) molars erupting, and the
third lower (dp3) and upper (dP3) deciduous premolars
present. The largest specimen (AMNH 93529, male) measured
60.1 mm for condylobasal length and exhibited severely worn
cusps. Thus in our sample the smallest individual was only
about one-half the maximum adult size. On the basis of known
sequences of tooth emergence and wear (Astu´a de Moraes and
Leiner 2008; Atramentowicz 1986), our youngest specimen is
approximately 5 months old, whereas the largest specimen
probably was 2 years old (or older).
Study of growth.—In this study we took two descriptive
approaches. First, we contrasted developing features in the
skull of the youngest specimens with those of the oldest on a
qualitative basis (Flores et al. 2006; Table 1). Anatomical
terminology follows primarily Wible (2003). Second, we
followed previous studies on marsupials using a series of 14
linear measurements to estimate allometric growth of skull
components (Abdala et al. 2001; Flores et al. 2003, 2006;
Giannini et al. 2004). These variables represent length, width,
or depth of major skull structures (Fig. 1).
For the quantitative analysis we took 2 methodological
approaches: bivariate and multivariate allometry. In bivariate
studies the scaling of any individual measurement can be
affected strongly by the choice of the independent variable
(Smith 1981; Wayne 1986). In most previous works (Abdala
et al. 2001; Emerson and Bramble 1993; Flores et al. 2003;
Giannini et al. 2004; Wayne 1986) researchers have taken the
total length of the skull, or the condylobasal length, as a proxy
of size for use as an independent variable in allometric studies.
Instead, we used as independent variable one with the least
departure from isometry in our multivariate analyses. This
resulted in an unconventional choice, the length of palate,
which nevertheless guarantees the best approximation to
isometry in the data set. For comparative purposes we also
report results using the condylobasal length as proxy of size.
Bivariate coefficients of allometry can be extremely useful
because they are less affected by sample completeness
(Giannini et al. 2004) and are more directly interpretable in
terms of size-dependent functional relationships (Jungers and
German 1981). In addition, bivariate coefficients can be
540 JOURNAL OF MAMMALOGY Vol. 91, No. 3
derived from simple growth models of each measurement
(Laird 1965; Wayne 1986).
To estimate the change in size of each cranial variable we
used the linear (log
10
) transformation of the power equation of
growth y~xb1b0e, where yis the variable of interest, b
0
the y-
intercept, xthe size proxy, b
1
the slope parameter or bivariate
coefficient of allometry, and ethe error term. We tested
deviations from isometry by means of 2-tailed t-tests, as the
independent variable was normally distributed (Shapiro Wilk
test, w50.98, P50.79). Allometric coefficients are taken to
be isometric whenever they cannot be distinguished statisti-
cally from unity, with afixed at 0.01. Statistically significant
deviations from isometry represented cases of negative
allometry if b
1
,1.0 and positive allometry if b
1
.1.0. We
report allometry coefficients calculated by ordinary least-
squares regression (hereafter LS) and reduced major axis
regression (RMA—Abdala et al. 2001; Flores et al. 2003). The
response of each cranial variable to our chosen proxy of size,
the length of palate, was evaluated by the r
2
coefficient.
In multivariate analysis of allometry size is a latent variable
affecting all measured variables simultaneously. As in our
previous studies, multivariate coefficients of allometry are
obtained from a principal components analysis (PCA) based
on a variance–covariance matrix of all variables (Jolicoeur
1963). This was followed by a resampling strategy using
jackknife pseudovalues of coefficients to generate confidence
intervals (CIs) for the coefficients so obtained (Flores et al.
2006; Giannini et al. 2004). The CI may be inclusive of a
matrix-specific isometric value that depends only on the
number of variables, p, and therefore be statistically
indistinguishable from isometry, or may exclude such value
and therefore be considered significantly allometric. The
isometric value for 14 variables is 0.267 (calculated as 1/p
0.5
),
so instances of allometry included all coefficients whose
TABLE 1.—List of morphological changes detected in a comparison of smallest versus largest specimens of Caluromys philander.
Characters Figure Juvenile Adult
Rostrum
1. Anterior portion of the nasals 3 Wide Narrow
2. Level of upper alveolar line 2 At zygomatic root Ventral to zygomatic root
3. Paracanine fossa 2 Poorly developed Large
4. Incisivomaxillary suture in lateral view 2 Near the anterior margin of the paracanine
fossa, adjacent to last incisor
In the middle of the paracanine fossa,
well separated from last incisor
Palatal region
5. Upper postcanine rows 4 Posteriorly divergent Parallel
6. Major palatine foramen 4 Inconspicuous Small
7. Minor palatine foramen 4 Notch opening laterally Enclosed foramen
8. Postpalatine torus 4 Slightly curved Strongly curved
Orbitotemporal region
9. Postorbital border of frontal 3 Absent Well developed
10. Postorbital constriction 3 Absent Well developed
11. Base of zygomatic process of squamosal 3 Oriented anteriorly Oriented laterally
12. Temporal fossa in dorsal view 3, 4 Narrow Wide
13. Dorsal profile of the skull 2 Higher than level of orbit Level with orbit
14. Temporal line 3 Absent Converging to the caudal tip
15. Anterodorsal portion of the jugal, at the orbit base 4 Not laterally extended Laterally extended
16. Masseteric line of the jugal 2 Absent Present
17. Anterior edge of the orbit 2 Between P2 and dP3 Between P3 and M1
Occipital region
18. Nuchal crest 3 Absent Well developed
19. Occipital condyles 2 Not caudally extended Caudally extended
20. Paracondylar process 2 Small Elongated to the level of the ventral
edge of foramen magnum
Basicranial region
21. Postglenoid process 2 Small Well developed
22. Transverse sinus canal 4 In line with anterior end of tympanic bulla Anterior to the tympanic bulla
23. Glenoid fossa 4 Narrow Laterally expanded
Mandible
24. Angular process 6 Slightly oriented medially Strongly oriented medially
25. Mandibular foramen 5 Large Small
26. Coronoid crest 5 Robust Weak
27. Dorsal margin of the coronoid process 5 Short anteroposteriorly Long anteroposteriorly
28. Posterior shelf of the masseteric fossa 5 Incipient Well developed
29. Ventral margin of mandibular body 5 Straight Convex
30. Diastema between p1-p2 5 Absent Present, long
31. Articular condyle 6 Narrow Laterally expanded
June 2010 FLORES ET AL.—CRANIAL ONTOGENY OF CALUROMYS PHILANDER 541
intervals excluded this value (positive if .0.267, negative if
,0.267).
Giannini et al. (2004) and Flores et al. (2006) followed
Manly’s (1997) suggestion of using trimmed pseudovalues as
an alternative for the calculation of the CI. Trimming the m
largest and msmallest jackknife pseudovalues for each
variable can significantly decrease width of the CI and yield
more realistic allometric estimations (Giannini et al. 2004). If
untrimmed and trimmed CIs differ greatly in width, this can be
interpreted as an indication of extreme pseudovalues affecting
the standard errors. Here we report untrimmed and trimmed
calculations. For the bivariate analyses the computer program
PAST (Hammer et al. 2001) was used. For the multivariate
statistical analysis (PCA +jackknife resampling) the R-script
(R Development Core Team 2008) of Giannini et al. (2009)
was used and is available from the authors.
RESULTS
Qualitative trends.—We identified 31 structures exhibiting
ontogenetic changes between youngest and oldest specimens
of C. philander (Table 1). These included changes in the
rostrum (4 characters), palatal region (4), orbitotemporal
region (9), occiput (3), basicranial region (3), and mandible
(8). Seven of those changes represented the appearance of
structures absent in juveniles (characters 6, 9, 10, 14, 16, 18,
and 30 in Table 1), whereas 11 of them imply completion or
enlargement of structures already present in juveniles
(characters 3, 7, 12, 19, 20, 21, 23, 24, 27, 28, and 31). Three
changes indicated the relative decrease in size of some
structures with age (characters 1, 25, and 26). Finally, 10 more
or less complex reorganizations occurred (characters 2, 4, 5, 8,
11, 13, 15, 17, 22, and 29).
Similar to previous results in morphological analyses of
marsupial ontogeny, several modifications observed in the
skull of C. philander were related to trophic function,
including changes related to the development of masticatory
muscles and safety of the temporomandibular joint at
occlusion. For instance, the paracanine fossa is present but
not well developed in juveniles, and it enlarges considerably
toward adulthood (character 3 in Table 1; Fig. 2). Linked with
this change is the placement of the suture between the
premaxilla and maxilla. In juveniles this suture is located at
the anterior margin of the paracanine fossa, close to the last
upper incisor. In adults the suture has migrated to a position in
the middle of the paracanine fossa (character 4 in Table 1;
Fig. 2). The upper dental series showed a diastema between
the first and second premolars that is present in all the
ontogenetic stages (Fig. 2). By contrast, the first and second
lower premolars are in contact in juveniles and separated by a
large diastema in adults (character 30).
Important morphological changes correlated with the
development of the temporalis muscle include the widening
of the temporal fossae (character 12; Figs. 3 and 4), change in
orientation of the zygomatic process of the squamosal from
anterior to lateral (character 11; Fig. 3, arrow), and the strong
development of the postorbital process (character 9; Fig. 3)
and the sagittal and nuchal crests (character 18; Figs. 2 and 3).
Of these structures, the nuchal crest experienced the most
extreme transformation, perhaps because of its dual condition
of origin and insertion of muscles of the mandible and the
neck, respectively. Associated with this change is the
widening of the dorsal margin of the coronoid process of the
mandible (character 27; Fig. 5).
The development of the masseter muscle also was evident
in several characters, specifically the masseteric line of the
jugal (character 16; Fig. 2), site of origin of the superficial
masseter and present only in adults, and the masseteric fossa
of the mandible, represented as an incipient shelf in juveniles
and a strong and well-developed structure in adults (character
28; Fig. 5). In correlation with the development of the cranial
crests, the dorsal margin of the coronoid process is extended
more craniocaudally in adults (character 27; Fig. 5). Finally,
the angular process of the mandible becomes remarkably more
FIG.1.—Cranial measurements of Caluromys philander used in this
study. Abbreviations: BB, breadth of braincase; BP, breadth of palate;
BZ, zygomatic breadth; HD, height of mandibular body; HM, height
of muzzle; HO, height of occipital plate; LC, length of coronoid
process; LD, length of mandible; LN, length of nasals; LO, length of
orbit; LP, length of lower postcanine row; LT, total length of the
skull; PAL, length of palate; UP, length of upper postcanine row.
542 JOURNAL OF MAMMALOGY Vol. 91, No. 3
robust and mesially extended in adults, increasing the surface
for the insertion of the superficial masseter and the internal
insertion of the medial pterygoid muscle (character 24;
Fig. 6).
Modifications related to adjustments of occlusion included
the displacement of the upper alveolar line, which is leveled
with the zygomatic root in juveniles and ventral to the root in
adults (character 2; Fig. 2). Changes also occurred in the
orientation of the upper postcanine row. In ventral view the
rows are posteriorly divergent in juveniles and subparallel,
with a slight labial convexity, in adults (character 5; Fig. 4).
The adult condition in C. philander in this trait represents a
remarkable difference from other marsupials sampled (Abdala
et al. 2001; Flores et al. 2003, 2006; Giannini et al. 2004).
Development of the bony palate involves the closure of the
minor palatine foramen (character 7; Fig. 4), with its function
probably replaced by the opening of a small major palatine
foramen (character 6; Fig. 4) and the increasing curvature of
the postpalatine torus (character 8; Fig. 4).
Also related to the trophic function is the development of
the paracondylar process of the exoccipital, from which the
posterior belly of the digastric muscle originates and whose
action is depressing the jaw, thus opening the mouth. The
knob present in juveniles was transformed in adults into a
strong process extended ventral to the level of the basioccipital
(character 20; Fig. 2).
FIG.2.—Lateral view of the skull of A) a juvenile, AMNH 96652;
B) subadult, AMNH 203352; and C) adult, AMNH 95529 of
Caluromys philander. Abbreviations: astp, alisphenoid tympanic
process; fr, frontal; iof, infraorbital foramen; M2, second upper
molar; M3, third upper molar; M4, fourth upper molar; mlj,
masseteric line of jugal; mx, maxilla; nc, nuchal crest; oc, occipital
condyle; pa, parietal; pcp, paracondylar process; pgp, postglenoid
process; Pm3, third upper premolar; pmx, premaxilla; pcf, paracanine
fossa; so, supraoccipital; sq, squamosal. Arrows in juvenile and adult
indicate change on the dorsal profile of the braincase. Scale bars 5
10 mm.
FIG.3.—Dorsal view of the skull of A) a juvenile, AMNH 96652;
B) subadult, AMNH 203352; and C) adult, AMNH 95529 of
Caluromys philander. Abbreviations: fr, frontal; la, lacrimal; mx,
maxilla; na, nasal; nc, nuchal crest; pa, parietal; pmx, premaxilla;
poc, postorbital constriction; pop, postorbital process; sco, sutura
coronalis; sfn, sutura frontonasalis; sq, squamosal; so, supraoccipital;
ss, sutura sagittalis; tl, temporal line. Arrows in juvenile and subadult
indicate change of flexure of the zygomatic process of the squamosal.
Scale bars 510 mm.
June 2010 FLORES ET AL.—CRANIAL ONTOGENY OF CALUROMYS PHILANDER 543
Some ontogenetic changes reshape the temporomandibular
joint and the mandibular body resistance. These include the
lateral expansion of the articular condyle of the dentary
(character 31; Fig. 6) and the concomitant change of the
glenoid fossa of the squamosal, with an increased contribution
of the jugal (character 23; Fig. 4) and the enlargement of the
FIG.4.—Ventral view of the skull of A) a juvenile, AMNH 96652;
B) subadult, AMNH 203352; and C) adult, AMNH 95529 of
Caluromys philander. Abbreviations: as, alisphenoid; astp, alisphe-
noid tympanic process; bo, basioccipital; bs, basisphenoid; fc,
fenestra cochleae; fo, foramen ovale; gf, glenoid fossa; hf,
hypoglossal foramen; inf, incisive foramen; jf, jugular foramen; ju,
jugal; lpc, lateral palatine crest; M2, second upper molar; M3, third
upper molar; M4, fourth upper molar; mlj, masseteric line of jugal;
mpf, minor palatine foramen; mpn, minor palatine notch; mtmx,
masseteric tuberosity of maxilla; mx, maxilla; pal, palatine; pcf,
paracanine fossa; pcp, paracondylar process; pe, periotic; pgp,
postglenoid process; pmx, premaxilla; ppt, postpalatine torus; pr,
promontorium; sq, squamosal; tc, transverse canal. Scale bars 5
10 mm.
FIG.5.—Lateral view of the lower jaw of A) a juvenile, AMNH
96652; B) subadult, AMNH 203352; and C) adult, AMNH 95529 of
Caluromys philander. Abbreviations: an, angular process; cml, crest
of masseteric line; con, condylar process; cor, coronoid process; m2,
second lower molar; m4, fourth lower molar; maf, masseteric fossa;
mf, mental foramen; ml, posterior shelf of masseteric line. Scale bars
510 mm.
FIG.6.—Occlusal view of the lower jaw of A) a juvenile, AMNH
96652, and B) adult, AMNH 95529 of Caluromys philander.
Abbreviations: an, angular process; con, condylar process; m2,
second lower molar; m4, fourth lower molar. Scale bars 510 mm.
544 JOURNAL OF MAMMALOGY Vol. 91, No. 3
postglenoid process (character 21; Fig. 2). Also important is
the strengthening of the mandibular body (Fig. 5).
The cranium exhibited two remarkable differences between
juveniles and adults: the development of a postorbital
constriction absent in juveniles (character 10; Fig. 3) and the
flattening of the skull roof by which the domed calvaria of
juveniles is leveled with the interorbital surface in adults
(character 13; Fig. 2, arrows). The latter trend could be linked
to changes in the deflection of the rostrum with respect to the
basicranium during ontogeny (Lieberman et al. 2008).
Bivariate allometry.—The fit of the cranial variables
examined to the proxy of size, the length of palate, as
evaluated by the r
2
coefficient in bivariate regression
(Table 2), varied between 91%(height of mandibular body
and height of muzzle) and 99%(total length of the skull). LS
and RMA showed the same allometric trends in 10 of 13
variables. The 2 methods found isometric trends in 7 variables,
positive allometry in length of nasals, and negative allometry
in breadth of braincase and height of occipital plate (Table 2).
RMA differed from LS in 3 variables (breadth of palate, height
of coronoid process, and height of dentary; Table 2). These
discrepancies are more apparent than real: given that RMA
slopes are always equal to or higher than LS slopes by
definition (RMA 5LS/r, with r1), a negative trend in LS
may become isometric under RMA (e.g., breadth of palate),
and isometric slopes may become positive (e.g., height of
coronoid process). Results using total length of the skull as
proxy of size differed in 3 variables when using the least-
squares method, and in only 1 (breadth of palate) under RMA
(Table 3).
Multivariate allometry.—The first principal component
includes 87.8%of the total variance. Observed multivariate
coefficients of allometry varied widely across variables
(Table 4). Two variables, length of palate and length of
mandible, showed the least observed departure from isometry
(20.001). Therefore, we used the cranial variable, length of
palate, as the proxy of size in bivariate allometry (see above).
Breadth of braincase exhibited the largest departure from
isometry (20.131). Average estimated bias (using absolute
jackknife values) across coefficients calculated from trimmed
and untrimmed values were both small (0.002 and 0.008,
respectively). Extreme pseudovalues did affect jackknife
estimates given that isometry could be rejected in only 4
skull variables using the untrimmed set, as compared with 6
variables if the trimmed set of pseudovalues was used. This
strongly suggests that the breadth of CIs could have been
affected seriously in some variables by those extreme values.
Taking the results from the (m51) trimmed analysis, 3
TABLE 2.—Summary of bivariate regressions using the length of palate as proxy of size (independent variable, see text) in Caluromys
philander. Abbreviations: r
2
, adjusted coefficient of correlation; b, coefficient of allometry under least squares (LS) or reduced major axis
(RMA); t
iso
, 2-tailed Student’s t-value under null hypothesis of isometry (expected coefficient of allometry under isometry 51; d.f. 543); P
iso
,
P-value for t
iso
. Symbols indicate isometry (5), positive allometry (+), and negative allometry (2). When in parentheses, the symbol indicates a
marginally nonsignificant allometric trend (0.01 ,P,0.04). Other abbreviations (variable acronyms) as in Fig. 1.
Variable Range (mm) r
2
LS RMA Allometric trends
b
LS
t
iso
P
iso
b
RMA
t
iso
P
iso
LS RMA
UP 10.9–18 0.97 0.93 21.42 0.16 0.98 20.26 0.79 55
BP 9.1–17.6 0.94 0.78 23.37 0.001 0.89 20.26 0.10 25
BB 14.5–20.3 0.96 0.50 216.00 ,0.001 0.54 214.69 ,0.001 22
BZ 18.9–35.9 0.98 1.01 0.32 0.75 1.05 1.2 0.24 55
LN 13.3–25.3 0.98 1.14 2.72 0.009 1.18 3.66 ,0.001 (+)+
HO 8.9–13.9 0.95 0.6 29.42 ,0.001 0.66 27.98 ,0.001 22
LO 9.3–15.9 0.95 0.96 20.69 0.49 1.05 0.71 0.48 55
HM 6–11.3 0.91 0.97 20.26 0.80 1.18 1.75 0.086 55
HC 10.1–20.2 0.94 1.11 1.23 0.23 1.25 2.82 0.007 5(+)
LP 11.2–19.8 0.97 0.94 21.23 0.23 0.99 20.23 0.82 55
HD 3.3–7.7 0.91 1.29 2.18 0.03 1.55 4.18 ,0.001 5+
LD 24.3–43.6 0.98 0.96 20.87 0.39 1.00 0.05 0.96 55
LT 31.6–60.1 0.99 1.04 1.93 0.06 1.05 2.36 0.023 55
TABLE 3.—Comparison of allometric trends using least-squares
(LS) versus reduced major axis (RMA) regression methods for the
length of palate (PAL) versus the total length of the skull (LT) as
independent variable (proxy of body size) in Caluromys philander.
Symbols indicate isometry (5), positive allometry (+), and negative
allometry (2). When in parentheses, the symbol indicates a
marginally nonsignificant allometric trend. Abbreviations for
variables as in Fig. 1. N/A 5nonapplicable.
Variable
LS RMA
PAL LT PAL LT
UP 52 55
BP 22 52
BB 22 22
BZ 55 55
LN (+)5++
HO 22 22
LO 55 55
HM 55 55
LC 55 (+)+
LP 52 55
HD 5+++
LD 52 55
PAL N/A 2N/A 2
LT 5N/A 5N/A
June 2010 FLORES ET AL.—CRANIAL ONTOGENY OF CALUROMYS PHILANDER 545
variables (breadth of the palate, breadth of braincase, and
height of the occipital plate) were negatively allometric,
whereas 3 variables (total length of the skull, length of
coronoid process, and height of mandibular body) were
positively allometric. Notably, the set of allometric variables
in the trimmed analysis also includes the total length of the
skull, although this variable includes the expected value under
isometry in the 99%CI of the untrimmed set (Table 4). A
brief comparison of bivariate and multivariate allometry
(Table 5) indicated that the latter shared all but 1 trend
(breadth of palate) with RMA and all but 2 trends (height/
length of coronoid and height of dentary) with LS.
DISCUSSION
Skull modeling in Caluromys philander.—As a basis for
comparison with other marsupials we collated the various
results from allometric analyzes to interpret the modeling of
the skull in C. philander. The braincase is relatively smaller
toward adulthood, as indicated by the strong negative trend in
breadth of the braincase and height of the occipital plate. By
contrast, another neurocranial component—the orbit—increas-
es its length isometrically, thus keeping the relative size of the
eye socket almost constant in the anteroposterior dimension.
The palate becomes slender as a consequence of the isometry
of its length and the negative allometry of its width. The upper
and lower tooth rows are isometric, despite tooth emergence in
the upper postcanine row being delayed with respect to the
lower row—during ontogeny, tooth count in the lower row
exceeds in 1 element the upper row (Astu´a de Moraes and
Leiner 2008). Taken together, the positive allometry of the
nasal length and the isometry of the height of the muzzle show
that the rostrum elongates while keeping its height propor-
tionally constant. In relative terms the temporal space expands
inward given the braincase negative allometry and the
zygomatic breadth isometry. The mandible changes essentially
by increasing its robustness because the thickness (height) of
mandibular body and the length of the coronoid process
showed a positive trend, whereas the length of the mandibular
body and the lower postcanine row were isometric. The strong
positive allometry of measurements of the mandible is
correlated with the change of diet after weaning to reach an
active mastication in adults, producing critical adjustment in
the mandibular structures committed with the occlusion and its
mechanic demands. For instance, the lateromesial develop-
ment of the mandibular condyle and the positive allometry of
the height of the mandibular body confer great strength and
resistance to the mandible during the postweaning period,
permitting an active mastication of solid food (Figs. 4 and 6).
Although the sagittal crest does not develop to a great extent in
adults, the nuchal crest does develop to a remarkable degree
(Figs. 3 and 4), suggesting the importance of the posterior
fibers of the temporalis muscle acting as part of a lever that
transmits the strongest force component to the canines (Flores
et al. 2003, 2006). This crest also might reflect the presence of
powerful neck muscles (Radinsky 1981).
Comparative cranial ontogeny and allometry.—As in large-
sized marsupials previously studied (Didelphis, Lutreolina,
Dasyurus), most of the ontogenetic changes observed in C.
philander are indicative of a general strengthening of the
masticatory apparatus and cervical musculature. However, the
pattern of development of some features in C. philander is
TABLE 4.—Results of the multivariate analysis of cranial allometry in Caluromys philander. The first three data columns show results using all
specimens. The remainder of the columns show jackknife results calculated with untrimmed and (m51) trimmed sets of pseudovalues (see
Materials and Methods). Allometry coefficient is the corresponding element of the first (unit) eigenvector per variable. The expected coefficient
is the value under isometry (0.267, same for all variables). The observed coefficient is the value obtained with all specimens included (n545).
The resampled coefficient is the first-order jackknife value. Bias is the difference between the resampled and observed coefficients. The
jackknifed 99%confidence interval (CI) is provided; allometric variables are those whose CIs exclude the expected value under isometry
(0.267). Symbols indicate isometry (5), positive allometry (+), and negative allometry (2). When in parentheses, the symbol indicates a
marginally nonsignificant allometric trend. Abbreviations (variable acronyms) as in Fig. 1.
Variables
Expected
allometry
coefficient
Observed
allometry
coefficient
Observed
departure
Untrimmed values Trimmed values
Resampled
allometry
coefficient Bias 99%CI
Growth
trend
Resampled
allometry
coefficient Bias 99%CI
Growth
trend
LT 0.267 0.2808 0.013 0.280 20.0005 0.258–0.303 50.288 0.007 0.276–0.300 +
PAL 0.267 0.2668 20.001 0.266 20.0010 0.246–0.285 50.268 0.001 0.254–0.282 5
UP 0.267 0.2497 20.017 0.251 0.0009 0.206–0.295 50.249 20.001 0.222–0.276 5
BP 0.267 0.2217 20.044 0.224 0.0018 0.152–0.295 50.232 0.010 0.196–0.267 2
BB 0.267 0.1366 20.131 0.136 20.0005 0.107–0.165 20.133 20.003 0.117–0.150 2
BZ 0.267 0.2830 0.014 0.281 20.0016 0.251–0.312 50.271 20.012 0.252–0.290 5
LN 0.267 0.3034 0.036 0.303 20.0005 0.250–0.356 50.292 20.011 0.266–0.318 (5)
HO 0.267 0.1596 20.107 0.160 0.0008 0.111–0.210 20.167 0.008 0.138–0.197 2
LO 0.267 0.2409 20.023 0.244 0.0033 0.175–0.313 50.266 0.025 0.224–0.307 5
HM 0.267 0.2838 0.017 0.284 0.0002 0.222–0.346 50.275 20.009 0.237–0.313 5
LC 0.267 0.3252 0.055 0.322 20.0033 0.268–0.375 (+) 0.317 20.008 0.276–0.359 +
LP 0.267 0.2602 20.009 0.258 20.0024 0.221–0.295 50.255 20.005 0.227–0.284 5
HD 0.267 0.3957 0.122 0.390 20.0061 0.293–0.486 +0.377 20.019 0.304–0.450 +
LD 0.267 0.2657 0.001 0.268 0.0023 0.223–0.313 50.263 20.003 0.230–0.297 5
546 JOURNAL OF MAMMALOGY Vol. 91, No. 3
somewhat different from that of the marsupials previously
studied. Here we interpret selected aspects of qualitative
change and allometry in the comparative context of marsupials
so far studied.
The early development of the paracanine fossa in C.
philander (Fig. 2) contrasts with the condition observed in
Lutreolina and Didelphis, in which the fossa is absent in
juveniles of comparabale developmental stage (Abdala et al.
2001; Flores et al. 2003). In the australidelphian Dasyurus,
juveniles larger than those of C. philander (,39 mm—Flores
et al. 2006) and in a similar stage of dental eruption do not
show any evidence of a paracanine fossa, whereas it is clearly
present in adults. The paracanine fossa is absent altogether in
the small-sized Dromiciops gliroides. The early development
of the paracanine fossa in C. philander (Fig. 2) is correlated
with large canines already present in young. This is suggestive
of an active use of canines early in the life of individuals,
which perhaps incorporate adult food items into their diet
earlier in life that other marsupials studied (adults of C.
philander are omnivorous with some preference for fruit—
Astu´a de Moraes et al. 2003; Charles-Dominique et al. 1981;
Julien-Laferriere 1999; Julien-Laferriere and Atramentowicz
1990; Leite et al. 1996). This is also correlated with changes in
the temporomandibular joint that progressively limit the
rostrocaudal displacement and the transverse rotational
movement of the mandible, thereby reducing mandibular
movements to the dorsoventral plane (Abdala et al. 2001).
The development of a P1-P2 diastema (Fig. 2) also has been
reported in Didelphis albiventris (Abdala et al. 2001) and L.
crassicaudata, in which the spacing appears comparatively
larger (Flores et al. 2003). Two small interdental spaces
appear in adults of Dromiciops gliroides in both the upper and
lower tooth rows, i.e., a small diastema between c and p1 and
p1 and p2, and also between C and P1 and P1 and P2 (Giannini
et al. 2004). The pattern of dental spacing is different in
Dasyurus albopunctatus given that the interdental space is
between p2 and m1 (p3 is absent in this species), and this
diastema is present in all the ontogenetic stages observed (see
Flores et al. 2006).
The changes in the orientation of the upper postcanine row,
which becomes nearly parallel in adults (Fig. 4), and the
strongly curved postpalatine torus (Fig. 4) reflect a complex
remodeling during postweaning. Such changes were not
detected in the ontogeny of any of the marsupial species
studied to date, likely representing ontogenetic patterns unique
to Caluromyinae, as inferred from palate morphology in other
members of the subfamily (Caluromysiops irrupta and
Glironia venusta; D. A. Flores, pers. obs.).
Results from allometric analyses of the skull development
in the didelphids Didelphis albiventris (Abdala et al. 2001), L.
crassicaudata (Flores et al. 2003), and Caluromys philander
(this contribution), and the australidelphians Dasyurus albo-
punctatus (Flores et al. 2006) and Dromiciops gliroides
(Giannini et al. 2004) allow for a basic comparison of
allometric trends among representatives of different marsupial
lineages (Table 6). Four variables show the same trends in the
5 marsupial species sampled: breadth of palate, breadth of
braincase, height of occipital plate (negatively allometric), and
the height of dentary (positively allometric). These 4 variables
seem to conform to a basic growth plan shared at the level of
marsupials (Flores et al. 2006). The negative allometry of the
breadth of palate detected in all groups implies that the
posterior palate is wide in juveniles. This condition is most
likely related to its function as a platform for the tongue in the
suckling opossum (German and Crompton 1996) and therefore
can be extended to other species of marsupials. Pump suckling
is an essential function during the early life of marsupials and
determines many other structural traits in the skull of neonates
(Clark and Smith 1993; Maier 1993). In turn, neurocranial
components such as braincase dimensions are negativelly
allometric as in most mammals (Emerson and Bramble 1993).
Sensory capsules are in general expected to scale negatively
(Emerson and Bramble 1993). In almost all vertebrates
accelerated differentiation of the central nervous system and
sensory capsules produces embryos and neonates with large
braincases, eyes, and auditory regions relative to trophic
components of the skull (Emerson and Bramble 1993).
However, in marsupials virtually all neurogenesis occurs after
birth and during lactation (Smith 1997). In our studies
inspecting marsupial ontogeny after weaning, and in all the
species sampled, the rate of growth of neurocranial compo-
nents such as breadth of braincase and height of occipital plate
shows pronounced negative allometry.
Some allometric trends are specific to 1 of the marsupial
representatives sampled, but most trends are common to 2 or
more taxa. Caluromys shares specific growth trends with
Didelphis or Lutreolina, but no allometric trend is shared by
the 3 didelphids to the exclusion of australidelphians
(Table 6). Likewise, C. philander shares no allometric trend
with the 2 australidelphians to the exclusion of didelphids
(Table 6).
TABLE 5.—Comparison of results across methods used to estimate
cranial allometry in Caluromys philander. Symbols indicate isometry
(5), positive allometry (+), and negative allometry (2). When in
parentheses, the symbol indicates a marginally nonsignificant
allometric trend. Abbreviations: LS, least-squares method; RMA,
reduced major axis method. Other abbreviations (variable acronyms)
as in Fig. 1. Multivariate trends are based on trimmed values under
jackknife resampling.
Variable LS RMA Multivariate
UP 55 5
BP 25 2
BB 22 2
BZ 55 5
LN ++ (5)
HO 22 2
LO 55 5
HM 55 5
LC 5(+)+
LP 55 5
HD ++ +
LD 55 5
LT 5++
June 2010 FLORES ET AL.—CRANIAL ONTOGENY OF CALUROMYS PHILANDER 547
Three variables exhibit the most remarkable differences
between C. philander and the other didelphids (Table 6). First,
the development of the orbit is isometric in Caluromys,asisin
Dromiciops and Dasyurus, and negatively allometric in the
other 2 didelphids (Table 6). Second, the lengths of the upper
and lower postcanine rows are isometric in C. philander and
negatively allometric in the other didelphids (Table 6). In
adult D. albiventris and L. crassicaudata, the lengths of both
postcanine series are nearly the same, but during ontogeny the
emergence of the lower postcanines is usually advanced in 1
element with respect to the upper postcanines (Astu´a de
Moraes and Leiner 2008). This sequence of emergence leads
to allometric differences between upper and lower rows, with
the coefficient of the lower postcanine row being smaller than
that of the upper row. However, the upper and lower rows
exhibit similar allometric coefficients in C. philander
(isometric in all of methods applied), but the lengths of the
lower and upper tooth rows are equally involved in the mutual
adjustment of the postcanine rows during growth.
The space for temporal musculature in adults is provided by
the interaction between the pattern of growth of the braincase
and zygomatic arches. In didelphids the breadth of zygoma is
isometric, whereas the breadth of the braincase is negatively
allometric (Table 6). Therefore, the space for the temporalis
muscle increases principally inward in relative terms (Abdala
et al. 2001). In Dasyurus the space for the temporalis also
increases by lateral expansion (i.e., the positively allometric
zygomatic breadth; Table 6), as observed in large felids as
Puma concolor (Giannini et al. 2009). This extra space for
masticatory musculature is likely associated with carnivorous
habits (Flores et al. 2006).
The comparison of quantitative cranial trends and the
pattern of morphological changes in the ontogeny of C.
philander and the other marsupial taxa considered herein
indicates a series of common developmental trends and some
trends restricted to specific groups. Just as in other mid- and
large-sized marsupials, most of the ontogenetic changes in C.
philander indicate strengthening of the masticatory system and
cervical musculature. However, several morphological char-
acters observed that appear in late ontogenetic stages in the
other marsupials were already present in young specimens of
C. philander, indicating precocial development of such traits
independent of the final adult size. In addition, specific
changes occur in the palate of C. philander, such as the shift of
orientation between the upper postcanine rows and the
development of a strongly curved postpalatine torus, which
are not known in any marsupial studied to date. Caluromys
shares ontogenetic trends with both didelphids and australi-
delphians. The cranial growth patterns obtained from multi-
variate analysis in C. philander show some unexpected trends,
such as the isometry of the orbit and of the upper postcanine
row, which are trends partially shared only with some studied
australidelphian marsupials. The exploration of the allometric
trends in a representative of the caluromyine clade to the
previously studied taxa reveals that no less than 4 ontogenetic
trends could be interpreted as plesiomorphic relative to the
marsupial clade, likely representing part of the basic
developmental plan of Metatheria. Future interpretations of
the postweaning ontogenetic changes in additional neotropical
marsupials phylogenetically distinct, such as Metachirus,
Monodelphis, and some mouse opossums, could reveal
patterns of growth in a phylogenetic context. Finally, the
comparison of the specific patterns detected in neotropical
marsupials with those of additional Australasian clades with
different life history (e.g., Diprotodontia) is essential to
increasing our understanding of the evolution of cranial
postweaning ontogeny in the group.
RESUMEN
La ontogenia craneana ha sido estudiada desde una
perspectiva alome´trica cuantitativa en marsupiales dide´lfidos,
microbiote´ridos y dasyu´ ridos. En continuacio´n con trabajos
previos, describimos y comparamos el desarrollo craneal
postdestete en la comadreja lanosa Caluromys philander.
Identificamos 31 cambios morfolo´gicos cualitativos entre
juveniles y adultos, la mayorı
´a relacionados con el aparato
tro´fico. El temprano desarrollo de la fosa paracanina,
relacionado con la presencia de caninos bien desarrollados
TABLE 6.—Comparison of allometric trends in Caluromys
philander (this study), the didelphids Didelphis albiventris (Abdala
et al. 2001) and Lutreolina crassicaudata (Flores et al. 2003), and the
australidelphians Dromiciops gliroides (Giannini et al. 2004) and
Dasyurus albopunctatus (Flores et al. 2006). Symbols indicate
isometry (5), positive allometry (+), and negative allometry (2).
When in parentheses, the symbol indicates a marginally
nonsignificant allometric trend. The trends for Caluromys are a
composite of results from LS, RMA, and multivariate methods. In the
cases of Didelphis albiventris and L. crassicaudata bivariate methods
were used with skull length as the independent variable; therefore,
allometric trend of skull length is unknown (marked ?). The number
of allometric cranial variables in each species is listed in the last row.
nis sample size for each species. Other abbreviations (species
acronyms) as in Fig. 1.
Variable
Caluromys
philander
Lutreolina
crassicaudata
Didelphis
albiventris
Dromiciops
gliroides
Dasyurus
albopunctatus
UP 52 2(5)2
BP (2)2222
BB 22 222
BZ 55 55(+)
LN (+)(2)++ 5
HO 2(2)22 2
LO 52 252
HM 55(2)(5)5
LC (+)5+(5)+
LP 5(2)22 2
HD ++ +(+)+
LD 55 +(5)5
PAL 55 552
LT (+)? ?+2
n45 32–43 52–61 43–51 (37) 31
Allometric
trends 7 7 11 7 11
548 JOURNAL OF MAMMALOGY Vol. 91, No. 3
en juveniles, es una marcada diferencia en la ontogenia de C.
philander en relacio´n a otros marsupiales estudiados. Reali-
zamos un ana´lisis de alometrı
´a bivariado y multivariado sobre
una matriz compuesta por 14 medidas craneales en 45
especimenes. El escalamiento alome´trico mostro´ que la caja
cerebral es relativamente pequen˜a en adultos, mientras que
otros componentes neurocraneales, particularmente la orbita,
crecieron isometricamente. El paladar se torna elongado por la
combinacio´n de la isometrı
´a de su largo y la alometrı
´a
negativa de su ancho. El rostro se alarga mientras su altura
permanece proporcionalmente constante. En te´rminos relati-
vos, la fosa temporal se expande interiormente siguiendo el
modelo marsupial, que combina la alometrı
´a negativa de la
caja cerebral y la isometrı
´a del ancho cigoma´tico. La
mandı
´bula se hace ma´s robusta. La ontogenia craneana de
C. philander comparte cuatro tendencias alometricas con dos
dide´lfidos (Didelphis albiventris yLutreolina crassicaudata)y
dos australidelfos (Dromiciops gliroides yDasyurus albo-
punctatus) previamente estudiados. Tres de estas tendencias,
ancho del paladar, ancho de la caja craneana y altura de la
placa occipital fueron alometricamente negativas, mientras
que la altura del dentario fue positiva. Estas tendencias
alometricas comunes pueden ser interpretadas como simple-
siomo´rficas en marsupiales, sugiriendo un alto grado de
conservacio´n del desarrollo craneal de Caluromys y otros
marsupiales.
ACKNOWLEDGMENTS
We thank Robert S. Voss for permitting access to collection
specimens and Pablo Teta for drawing Fig. 1. D. Flores and N.
Giannini thank CONICET (Consejo Nacional de Investigaciones
Cientı
´ficas y Te´cnicas, Argentina) for financial support. PAST
(Palaeontological Scientific Trust, Johannesburg, South Africa)
supported a research trip to Argentina for F. Abdala. We thank John
Wible and an anonymous reviewer for comments that improved the
quality of this manuscript.
LITERATURE CITED
ABDALA F., D. A. FLORES,AND N. P. GIANNINI. 2001. Postweaning
ontogeny of the skull of Didelphis albiventris.Journalof
Mammalogy 87:190–200.
ASTU
´ADEMORAES, D., AND N. O. LEINER. 2008. Tooth eruption
sequence and replacement pattern in woolly opossums, genus
Caluromys (Didelphimorphia: Didelphidae). Journal of Mammal-
ogy 89:244–251.
ASTU
´ADE MORAES, D., R. T. SANTORI,R.FINOTTI,AND R. CERQUEIRA.
2003. Nutritional and fibre contents of laboratory-established diets
of neotropical opossums (Didelphidae). Pp. 221–237 in Predators
with pouches: the biology of carnivorous marsupials (M. E. Jones,
C. R. Dickman, and M. Archer, eds.). CSIRO Publishing,
Collingwood, Australia.
ATRAMENTOWICZ, M. 1982. Influence du milieu sur l’activite´
locomotrice et la reproduction de Caluromys philander (L.). Revue
d’Ecologie (La Terre et La Vie) 36:373–395.
ATRAMENTOWICZ, M. 1986. Dynamique de populations chez
trı
´os marsupiaux didelphide’s de Guyane. Biotropica 18:136–
149.
ATRAMENTOWICZ, M. 1995. Growth of pouch young in the bare-tailed
woolly opossum, Caluromys philander. Journal of Mammalogy
76:1213–1219.
CHARLES-DOMINIQUE, P., M. ATRAMENTOWICZ,M.CHARLES-DOMINIQUE,
H. GERARD,C.M.HLADIK,AND M. F. PREVOST. 1981. Les
mammiferes frugivores arboricoles nocturnes d’une foret Guyan-
naise: inter-relations plantes–animaux. Revue d’Ecologie (La Terre
et La Vie) 35:341–435.
CLARK, C. T., AND K. K. SMITH. 1993. Cranial osteogenesis in
Monodelphis domestica (Didelphidae) and Macropus eugenii
(Macropodidae). Journal of Morphology 215:119–149.
CREIGHTON, G. K. 1984. Systematic studies on opossums (Didelphi-
dae) and rodents (Cricetidae). Ph.D. dissertation, University of
Michigan, Ann Arbor.
EMERSON,S.B.,AND D. M. BRAMBLE. 1993. Scaling, allometry
and skull design. Pp. 384–416 in The skull (J. Hanken and
B. K. Hall, eds.). The University of Chicago Press, Chicago,
Illinois.
FLORES, D. A. 2003. Estudio taxono´mico y zoogeogra´ fico de los
marsupiales de Argentina. Ph.D. dissertation, Universidad Nacio-
nal de Tucuma´n, Argentina.
FLORES, D. A. 2009. Phylogenetic analyses of postcranial skeletal
morphology in didelphid marsupials. Bulletin of the American
Museum of Natural History 320:1–81.
FLORES, D. A., N. P. GIANNINI,AND F. ABDALA. 2003. Cranial ontogeny
of Lutreolina crassicaudata (Didelphidae): a comparison with
Didelphis albiventris. Acta Theriologica 48:1–9.
FLORES, D. A., N. P. GIANNINI,AND F. ABDALA.2006.Compara-
tive postnatal ontogeny of the skull in the Australidelphian
metatherian Dasyurus albopunctatus (Marsupialia: Dasyuro-
morpha: Dasyuridae). Journal of Morphology 267:426–
440.
GERMAN, R. Z., AND A. W. CROMPTON. 1996. Ontogeny of suckling
mechanisms in opposums (Didelphis virginiana). Brain, Behavior
and Evolution 48:157–164.
GIANNINI, N. P., F. ABDALA,AND D. A. FLORES. 2004. Comparative
postnatal ontogeny of the skull in Dromiciops gliroides (Marsu-
pialia: Microbiotheriidae). American Museum Novitates 3460:1–
17.
GIANNINI, N. P., V. SEGURA,M.I.GIANNINI,AND D. A. FLORES. In press.
A quantitative approach to the cranial ontogeny of the puma.
Mammalian Biology.
HAMMER, Ø., D. A. T. HARPER,AND P. D. RYAN. 2001. PAST:
palaeontological statistics software package for education and data
analysis. Palaeontologia Electronica 4. http://palaeo-electronica.
org/2001_1/past/issue1_01.htm. Accessed 29 May 2009.
JANSA, S. A., AND R. S. VOSS. 2000. Phylogenetic studies on
didelphid marsupials I. Introduction and preliminary results from
nuclear IRBP gene sequences. Journal of Mammalian Evolution
7:43–77.
JOLICOEUR, P. 1963. The multivariate generalization of the allometry
equation. Biometrics 19:497–499.
JULIEN-LAFERRIERE, D. 1991. Organisation du peuplement de marsu-
piaux en Guyane Francaise. Revue d’Ecologie (Terre et Vie)
46:125–144.
JULIEN-LAFERRIERE, D. 1999. Foraging strategies and food partitioning
in the neotropical frugivorous mammals Caluromys philander and
Potos flavus. Journal of Zoology (London) 247:71–80.
JULIEN-LAFERRIERE, D., AND M. ATRAMENTOWICZ. 1990. Feeding and
reproduction of three didelphid marsupials in two neotropical
forests (French Guiana). Biotropica 22:404–415.
June 2010 FLORES ET AL.—CRANIAL ONTOGENY OF CALUROMYS PHILANDER 549
JUNGERS,W.L.,AND R. Z. GERMAN. 1981. Ontogenetic and
interspecific skeletal allometry in nonhuman primates: bivariate
versus multivariate analysis. American Journal of Physical
Anthropology 55:195–202.
KIRSCH, J. A. W., A. W. DICKERMAN,AND O. A. REIG. 1995. DNA/
DNA hybridization studies of carnivorous marsupials IV. Interge-
neric relationships of the opossum (Didelphidae). Marmosiana,
Acta Teriologica Latinoamericana 1:57–78.
LAIRD, A. K. 1965. Dynamics of relative growth. Growth 29:249–
263.
LEITE, Y. L. R., L. P. COSTA,AND J. R. STA LLINGS. 1996. Diet
and vertical space use of three sympatric opossums in a
Brazilian Atlantic forest reserve. Journal of Tropical Ecology
12:435–440.
LIEBERMAN, D. E., B. HALLGRIMSSON,W.LIU,T.E.PARSONS,AND H. A.
JAMNICZKY. 2008. Spatial packing, cranial base angulation, and
craniofacial shape variation in the mammalian skull: testing a new
model using mice. Journal of Anatomy 212:720–735.
MAIER, W. 1993. Cranial morphology of the therian common
ancestor, as suggested by the adaptations of neonate marsupials.
Pp. 166–181 in Mammal phylogeny: mesozoic differentiation,
multituberculates, monotremes, early therians, and marsupials (F.
S. Szalay, M. J. Novacek, and M. C. McKenna, eds.). Springer-
Verlag, Berlin, Germany.
MANLY, B. F. J. 1997. Randomization, bootstrap, and Monte Carlo
methods in biology. 2nd ed. Chapman & Hall, London, United
Kingdom.
PALMA, R. E., AND A. E. SPOTORNO. 1999. Molecular systematics of
marsupials based on the rRNA 12S mitochondrial gene: the
phylogeny of Didelphimorphia and the living fossil microbiotheriid
Dromiciops gliroides Thomas. Molecular Phylogenetics and
Evolution 13:525–535.
RD
EVELOPMENT CORE TEAM. 2008. R: a language and environment for
statistical computing. R Foundation for Statistical Computing.
http://www.Rproject.org. Accessed 27 April 2009.
RADINSKY, L. B. 1981. Evolution of the skull shape in carnivores. 2.
Additional modern carnivores. Biological Journal of the Linnean
Society 16:337–355.
REIG, O. A., J. A. W. KIRSCH,AND L. G. MARSHALL. 1987. Systematic
relationships of the living and neocenozoic American ‘‘opossum-
like’’ marsupials (Suborder Didelphimorphia), with comments on
the classification of these and of the Cretaceous and Paleogene
New World and European metatherians. Pp 1–89 in Possums and
opossums: studies in evolution (M. Archer, ed.). Surrey Beatty &
Sons, Sydney, Australia.
SMITH, K. K. 1997. Comparative patterns of craniofacial development
in eutherian and metatherian mammals. Evolution 51:1663–1678.
SMITH, R. J. 1981. On the definition of variables in studies of primate
dental allometry. American Journal of Physical Anthropology
55:323–329.
VOSS, R. S., AND S. JANSA. 2003. Phylogenetic studies on didelphid
marsupials II. Nonmolecular data and new IRBP sequences:
separate and combined analyses of didelphine relationships with
denser taxon sampling. Bulletin of the American Museum of
Natural History 276:1–82.
VOSS, R. S., AND S. JANSA. 2009. Phylogenetic relationships and
classification of Didelphid marsupials, an extant radiation of New
World metatherian mammals. Bulletin of the American Museum of
Natural History 322:1–177.
WAYNE, R. K. 1986. Cranial morphology of domestic and wild canids:
the influence of development on morphological change. Evolution
40:243–261.
WIBLE, J. R. 2003. On the cranial osteology of the short-tailed
opossum Monodelphis brevicaudata (Didelphidae, Marsupialia).
Annals of the Carnegie Museum 72:137–202.
Submitted 5 September 2009. Accepted 12 December 2009.
Associate Editor was Elizabeth R. Dumont.
550 JOURNAL OF MAMMALOGY Vol. 91, No. 3
