Ontogenetic variation of an omnivorous generalist rodent: The case of the montane akodont (Akodon montensis)

Abstract

Cranial ontogenetic analyses improve our understanding of function from developmental, ecological, and
evolutionary perspectives. Akodon montensis is an abundant, omnivorous forest species that occupies many
habitat types. We used traditional and geometric morphometric approaches to describe ontogenetic variation
in skulls of A. montensis. We tested for sexual dimorphism and described patterns of variation associated with
both age and size based on 6 postweaning age classes. We found no evidence for sexual dimorphism. Growth
patterns showed an initial narrowing of the braincase, and associated changes in the rostrum as specimens reached
adulthood. Older animals had an elongated rostrum and palate. Geometric morphometric analysis revealed
allometric variation associated with the basicranium for the entire age series, while traditional morphometric
analyses showed allometric variation in the facial component. The patterns found for A. montensis are similar to
those of other species of Akodon. We characterize ontogenetic patterns for Akodontines, the second most diverse
Sigmodontine tribe, and a model group for studies of shape change in generalist rodents.

Full text

1
Ontogenetic variation of an omnivorous generalist rodent: the case
of the montane akodont (Akodon montensis)
Gabriel Hernandez, Soraida Garcia, Júlio F. Vilela, and noé U. de la SancHa*
Department of Biological Sciences, Chicago State University, 9501 S. King Drive, Chicago, IL 60628, USA (GH, NUDLS)
Integrative Research Center, The Field Museum of Natural History, 1400 S. Lake Shore Drive, Chicago, IL 60605, USA (GH,
JFV, NUDLS)
Department of Ecological Sciences and Engineering, Purdue University, 610 Purdue Mall, West Lafayette, IN 47907, USA (SG)
* Correspondent: delasancha@msn.com
Cranial ontogenetic analyses improve our understanding of function from developmental, ecological, and
evolutionary perspectives. Akodon montensis is an abundant, omnivorous forest species that occupies many
habitat types. We used traditional and geometric morphometric approaches to describe ontogenetic variation
in skulls of A. montensis. We tested for sexual dimorphism and described patterns of variation associated with
both age and size based on 6 postweaning age classes. We found no evidence for sexual dimorphism. Growth
patterns showed an initial narrowing of the braincase, and associated changes in the rostrum as specimens reached
adulthood. Older animals had an elongated rostrum and palate. Geometric morphometric analysis revealed
allometric variation associated with the basicranium for the entire age series, while traditional morphometric
analyses showed allometric variation in the facial component. The patterns found for A. montensis are similar to
those of other species of Akodon. We characterize ontogenetic patterns for Akodontines, the second most diverse
Sigmodontine tribe, and a model group for studies of shape change in generalist rodents.
Los análisis craneales ontogenéticos mejoran nuestra comprensión sobre la función craneal desde una perspectiva
de desarrollo, ecológica y evolutiva. Akodon montensis es una especie omnívora y abundante de bosque que
también se encuentra en muchos otros tipos de hábitats. Se han utilizado enfoques morfométricos tradicionales
y geométricos para describir la variación ontogenética de cráneos de A. montensis. Se evaluó el dimorfismo
sexual y se describieron los patrones de variación asociados con la edad y el tamaño basados en 6 clases de edad
diferentes posteriores al destete. No se encontró evidencia de dimorfismo sexual. Los patrones de crecimiento
mostraron un estrechamiento inicial del cráneo, así como cambios asociados con el rostro a medida que los
especímenes alcanzan la edad adulta. Los animales de mayor edad tienen rostros y paladares alargados. El
análisis morfométrico geométrico reveló una variación alométrica asociada con el básicranio para toda la serie
de edades estudiadas, mientras que los análisis morfométricos tradicionales mostraron una variación alométrica
en el componente facial del cráneo. Los patrones encontrados para A. montensis son similares a los de otras
especies de Akodon. Se caracterizó patrones ontogenéticos para Akodontinos, la segunda tribu más diversa de
Sigmodontinos, y el grupo modelo de roedores generalistas para los estudios de variación de forma.
Key words: Akodontini, allometry, Atlantic Forest, Cricetidae, geometric morphometrics, ontogeny, Paraguay, Rodentia,
Sigmodontinae
Ontogeny can be described as an amalgamation of 3 processes:
development, growth, and allometry (Klingenberg 1998). In
this context, development is the change in shape as a func-
tion of age; growth is the change in size as a function of age;
and allometry is the change in shape as a function of size
(Klingenberg 1998). Among the most relevant insights from
studies of ontogenetic variation is that all evolutionary change
stems from ontogeny (Zelditch et al. 2012). The various com-
ponents of ontogeny can be studied by comparing conspecifics
of different age classes or sizes (Naroll and Von Bertalanffy
Journal of Mammalogy, xx(x):1–12, 2017
DOI:10.1093/jmammal/gyx135
© 2017 American Society of Mammalogists, www.mammalogy.org

2 JOURNAL OF MAMMALOGY
1956; Klingenberg 1998). Descriptions of ontogenetic patterns
are important in minimizing the confounded effects of geo-
graphic variation or interspecific variation (Myers 1989). This
is especially true for species that show seasonal fluctuations in
morphology (Myers 1989). Failure to account for ontogenetic
variation may lead to comparisons of age variants rather than
taxonomic units and may consequently produce incorrect taxo-
nomic or ecological patterns. These complications are likely
exacerbated in areas with closely related and morphologically
similar sympatric species.
Currently, studies of ontogeny in the Sigmodontinae have
involved only a small fraction of the 377 species in this
group, but have included many tribes and evolutionary lin-
eages: tribe Sigmodontini, Sigmodon fulviventer (Zelditch et
al. 1993), S. alstoni, and S. hispidus (Voss and Marcus 1992);
tribe Phyllotini, Calomys expulsus (Hingst-Zaher et al. 2000);
tribe Thomasomini, Rhipidomys mastacalis, R. venezuelae
(Voss and Marcus 1992), and R. latimanus (López-Fuster et
al. 2001); tribe Oryzomyini, Aegialomys xanthaeolus (Prado
and Percequillo 2011), the genus Zygodontomys (Voss et al.
1990; Voss and Marcus 1992), Oryzomys capito (presumably
Hylaeamys megacephalus) and Oryzomys talamancae (pre-
sumably Transandinomys bolivaris—Voss and Marcus 1992);
tribe Ichthyomyini, Rheomys thomasi (Voss 1988; Voss and
Marcus 1992), and R. mexicanus (Voss and Marcus 1992); and
tribe Akodontini, Oxymycterus rutilans and Oxymycterus sp.
(Voss and Marcus 1992), the Akodon “boliviensis group,” A.
toba, A. s. simulator (Myers 1989), and A. urichi (Voss and
Marcus 1992). The revision of the A. “boliviensis group” evalu-
ated ontogenetic patterns and sexual dimorphism for 5 Akodon
species (Myers et al. 1990) and the revision of the A. varius
group included A. toba and A. s. simulator (Myers 1989).
The genus Akodon is the most diverse taxon within the tribe
Akodontini, which is the second most diverse Neotropical
Sigmodontine rodent tribe, behind Oryzomini (D’Elía and
Pardiñas 2015). It is estimated that only 70% of the living spe-
cies of Akodon have been described formally (Gonçalves et al.
2007). This genus tends to include vole-like cursorial forms
(D’Elía and Pardiñas 2015) that are widely distributed through-
out South America, from Colombia to southern Argentina
and Chile (Musser and Carleton 2005). Akodon montensis is
a terrestrial habitat generalist (Püttker et al. 2008; Galiano
et al. 2013) and an omnivore (Paglia et al. 2012), with a diet
that varies throughout the year (Vieira et al. 2006; Talamoni
et al. 2008). The ecology of A. montensis is poorly known
relative to its close relatives A. cursor (Pardiñas et al. 2015)
and A. paranaensis (de la Sancha 2014), and these species
are difficult to differentiate in the field. Akodon montensis is
restricted to southeastern Brazil, Misiones Argentina, and east-
ern Paraguay, inhabiting primarily the Interior Atlantic Forest
(IAF) and Araucária Forest (Dalmagro and Vieira 2005; Lima
et al. 2010; Galiano et al. 2013; Valdez and D’Elía 2013); it
occurs in campo, cerradão (Alho 2005), and ombrophilous
forests (Antunes et al. 2010), but it is also known from open,
savanna-like biomes (Talamoni and Dias 1999; Couto and
Talamoni 2005). Additionally, it has been reported in a variety
of agricultural systems and other disturbed habitats, such as
eucalyptus plantations and buildings in rural areas (Umetsu and
Pardini 2007).
Morphometric data are an important source of information
when trying to understand relationships among evolutionary
units. Morphometric techniques make it possible to quantify and
compare shape variation within and between groups (Adams
et al. 2004). Traditional morphometric methods (TMMs) have
been used to quantify structural differences based on linear
measurements via an assortment of multivariate statistical
techniques (Marcus 1990; Adams et al. 2004). However, this
approach provides little information about shape per se, and it
is difficult to disentangle size and shape and to visualize shape
differences (Adams et al. 2004). These methods result in numer-
ical descriptions of shape that may not be inherently intuitive
and are difficult to interpret (Zelditch et al. 2012). Studies of
allometry have 2 basic types of data sets: longitudinal or cross-
sectional. Longitudinal data sets include measurements on each
individual in the study, at multiple times during their growth
(Klingenberg 1996; Maunz and German 1997). Cross-sectional
data sets are the most common and include many specimens
measured at a single stage and thus allometric trajectories are
measured from averages of many individuals across the various
allometric stages evaluated (Klingenberg 1998).
Geometric morphometric methods (GMMs) have changed
dramatically how morphological information is quantified,
using robust tools to detect morphological differences (Muñoz-
Muñoz and Perpiñán 2010). GMM methods are based on
robust statistical approaches to shape analysis using superim-
position of homologous landmarks (Adams et al. 2013). These
approaches simplify the effective visualization and interpreta-
tion of shape information, and enable discrimination within and
between groups.
Small rodents like A. montensis are ideal for studies of
ontogeny because of their abundance and rapid population
turnover (Gentile 2000). Because form and function are related,
it is reasonable to assume that to some extent shape reflects
niche use. Patterns of ontogenetic shape change should provide
information about niche use or the ability to exploit resources.
We hypothesize that ontogenetic change in A. montensis will
reflect similar shape patterns and ontogeny as other general-
ist omnivorous rodents. Furthermore, understanding morpho-
logical variation in A. montensis can improve our ability to
differentiate A. montensis from other cryptic Akodon species
like A. cursor and A. paranaensis. Given that quantification of
biodiversity involves proper identification of species and abun-
dances of individuals per species (de la Sancha 2014), and that
Akodon species tend to be among the most abundant species
in forests where they are found (Pardini 2004; de la Sancha
2014), improved ability to correctly identify Akodon species
is valuable for biodiversity, biogeography, landscape ecology,
community ecology, and conservation-related questions.
In this study, we address 4 major questions: 1) What are
the morphological changes in the cranium of A. montensis
associated with postnatal ontogeny? 2) Is there sexual dimor-
phism in A. montensis? 3) What are the allometric patterns in

HERNANDEZ ET AL.—ONTOGENETIC VARIATION IN AKODON MONTENSIS 3
A. montensis? 4) How do patterns of allometric and develop-
mental variation compare to other Sigmodontines in the genus
Akodon using similar niches?
Materials and Methods
In this study, we implemented the cross-sectional approach. We
analyzed 133 A. montensis crania (Supplementary Data SD1)
representing a robust postnatal ontogenetic series (61 males,
72 females). These specimens were collected in 4 of the larg-
est forest remnants of eastern Paraguay, including Mbaracayú
Natural Forest Reserve (MB), San Rafael Managed Resource
Reserve (SR), Morombí Natural Private Reserve (MO), and
Limoy Biological Reserve (LY; see de la Sancha 2014 for
details). Specimens also were collected from 6 forest rem-
nants of various sizes at Reserva Tapyta (TA). All specimens
are housed at the mammal collection of the Field Museum
of Natural History (FMNH), in Chicago, Illinois or will be
deposited at the Colección de Zoología, Facultad de Ciencias
Naturales y Exactas (CZ), Universidad Nacional de Asunción,
San Lorenzo, Paraguay. Handling and euthanization of all spec-
imens followed Sikes et al. (2016) and methods were approved
by the institutional animal care and use committee (IACUC) at
Texas Tech University.
Specimens were assigned to 1 of 6 age classes based on molar
eruption and tooth wear (Fig. 1A) beginning with age class
2. Age class 1 was a place holder for future analyses includ-
ing prenatal specimens. Age class 2 (pooled males and females,
n = 3) included specimens with M1 and M2 fully erupted and
M3 not present. Age class 3 (n = 8) included juveniles with
class 2 characteristics and M3 present, however not completely
erupted (Fig. 1A). Age classes 4 to 7 included specimens with
fully erupted M1–M3 with increasing molar wear. Age class
4 (n = 19) included specimens that showed no obvious cusp
wear. Age class 5 (n = 20) included individuals with wear on
only 1 molar and perhaps slight wear on 2 molars. Age class
6 (n = 13) included specimens with clear wear on 2 or more
molars. Age class 7 (n = 9) included specimens with extensive
molar wear where the molar cusps were completely indistin-
guishable (Fig. 1A).
Morphometric analyses.—We used both traditional (TMM)
and geometric (GMM) morphological methods to describe
shape variation in an ontogenetic series of crania. Fourteen cra-
nial measurements were taken to the nearest 0.01 mm using
digital calipers (Table 1). Measurements were taken with a ste-
reoscope to ensure consistent use of homologous points. The
descriptive utility of these measurements has been established
in other studies of ontogeny (Voss 1988; Flores et al. 2015).
In an effort to maintain precision, missing data were estimated
using a maximum-likelihood expectation-maximization algo-
rithm (Dempster et al. 1977) using a Matlab script created by
R. E. Strauss (http://www.faculty.biol.ttu.edu/Strauss/Matlab/
Matlab.htm). This method provides reasonably accurate esti-
mates of missing values (Strauss et al. 2003), particularly
when, as in our study, missing data are few (0.01% missing in
our data set).
We used principal component analysis (PCA) of natural log-
transformed morphological data using the covariance matrix
(PCAlin) in an effort to provide an ordination of the data and a
set of orthogonal variables for use in a subsequent multivariate
allometry analysis (see below). PCA produces monotonically
decreasing eigenvalues, and the first few eigenvectors generally
capture most of the variance in the data (Strauss 2010). PC1
has been proposed as a measure of simple allometry (Jolicoeur
1963). Importance of PC loadings was determined following
Hair et al. (1987). For our data, loadings > 0.30 and < −0.30
were classified as important and loadings with > 0.50 and <
−0.50 as highly important.
We used discriminant function analysis (DFA) to discrimi-
nate shapes of age groups and sexes using a log-transformed
matrix (DFAlin for TMM) in an effort to evaluate development,
by comparing shapes of different age classes. We used non-
parametric multivariate analysis of variance (MANOVA) with
10,000 permutations (Ordoñez-Garza et al. 2010) in Matlab
function “Dfa” (Strauss 2010). This approach differs from a
PerMANOVA sensu Anderson (2001). This analysis was fol-
lowed up by a size-free DFA (Dos Reis et al. 1990) using
Matlab function “DfaResids.” Thus, we were able to disen-
tangle differences between sexes from those due to size alone.
We used the method of Giannini et al. (2004, 2010) within
program R to evaluate allometry. A PCA of log-transformed
data with a jackknife resampling (Giannini et al. 2004) that
generated confidence intervals (CIs) for the original coef-
ficients was performed. Here, isometry is present when there
is deviation of the corresponding eigenvector element from a
hypothetical isometric value, calculated as 1/p0.5 (p = the num-
ber of variables). Deviations from this value indicate allometry.
Inclusion of the raw loadings within the calculated CIs indicted
isometry for that variable, while values more extreme than the
CIs indicate positive allometry (above the confidence limit) or
negative allometry (below the confidence limit—Giannini et al.
2004, 2010).
Geometric morphometrics.—Crania were digitized using a
Konica Minolta DiMAGE Z6 camera using the super-macro
function. The camera was mounted on a cop stand, at a fixed
distance with the plane of the sensor parallel to the base of the
copy stand. Standardization of the distance of the camera from
the specimens ensures that warping of the pictures from curva-
ture of the camera lens is standardized. The fixed distance of
the lens from specimens established a consistent scale for anal-
yses of size variation. Specimen orientation was standardized
by resting the zygomatic arches of each specimens on rubber
bands so that each cranium was supported by the same struc-
tures, thus minimizing deviations in orientation angles.
Ventral landmarks were digitized using TPSDig2 (Rohlf
2015). We used 16 landmarks (Table 2) that mirrored the TMM
measurements (Fig. 1C). We placed landmarks on half of the
cranium to avoid issues related to symmetry (Cardini 2016)
and inflation of error terms. Additionally, because age classes
are typically large and samples sizes for some age classes are
small, ontogenetic analyses are usually performed only on 1
side (Cardini 2016).

4 JOURNAL OF MAMMALOGY
Specimens were aligned using a Generalized Procrustes Fit
(Adams et al. 2013) within program MorphoJ (Klingenberg
2011). Subsequently, we performed both a PCA and DFA
using the Procrustes coordinates in order to explore patterns
of shape variation, and to determine shape differences between
age classes. We then compared the GMM PCA (PCAgeo)
with that of the TMM (PCAlin) using Spearman correla-
tions. This nonparametric statistic is insensitive to deviations
from linearity, normality, and homoscedasticity compared
to parametric approaches (McDonald 2008). Shape differ-
ences between age groups and sexes also were assessed using
DFA (DFAgeo) on landmarks within MorphoJ. The DFAgeo
was followed up with permutation tests (n = 10,000 itera-
tions) to evaluate the null hypothesis of no difference between
Procrustes coordinates for age groups and sexes. The DFAgeo
also was followed up by a multivariate analysis of covariance
Fig. 1.—A) Illustrations of tooth wear patterns for each age group, 2–7 respectively, from left to right. Age classes 2 and 3 do not show full erup-
tion of M3. Age classes 4–7 show complete and full eruption of M3. Age class 4 shows high cusps with no wear. Age class 5 shows some wear on
only 1 molar, most of the other molars show no visible wear. Age class 6 shows clear wear on more than 1 molar, however, cusps are still visible.
Finally, age class 7 is reserved for specimens with considerable and excessive molar wear, cusps are not visible. B) Shows representatives of each
age group crania in our allometric series of Akodon montensis. C) Landmark placement associated with Table 2 used for the geometric morpho-
metric analysis. Subsequent pictures show geometric discriminant function analysis (DFA) comparisons of shape changes from age class 2 with
the other age classes showing the change in cranium morphology shape as a function of size.

HERNANDEZ ET AL.—ONTOGENETIC VARIATION IN AKODON MONTENSIS 5
(MANCOVA) (Viscosi and Cardini 2011) to test for differ-
ences in allometric trajectory between sexes, in light of the fact
that we eliminated sexual dimorphism effects of size on shape
(Viscosi and Cardini 2011). This approach is equivalent to the
size-free DFA used on the linear morphometrics. Finally, we
performed a Procrustes allometric regression for shape co-vari-
ation data comparing the Procrustes coordinate scores against
the log-transformed centroid size of specimens using MorphoJ.
results
Linear morphometrics.—The PC1lin and PC2lin accounted
for 70.9% and 6.1% of the variation found among all age groups
(Fig. 2). PC1lin was associated strongly with LN, BIF, LIF, and
very strongly with BIT, whereas PC2lin was best described by
LIF, and very strongly by BIF and BIT (Table 3). The linear DFA
did not discriminate completely between all age classes, but did
show clear discrimination between every other age class (Fig. 2).
DF1 accounted for 88.6% of variation between groups and DF2
accounted for 6.1%. The MANOVA revealed no significant dif-
ference between males and females (Wilks’ lambda = 0.845,
F14,118 = 1.54, P = 0.105), but showed significant differences
between age classes (Wilks’ lambda = 0.076, F70,546.8 = 5.62,
P < 0.0001). Pairwise comparison between most age classes
were significant (Table 4), but not between age classes 2 and 3
(Wilks’ lambda = 0.388, F14,3 = 1.57, P = 0.3883), classes 5 and
6 (Wilks’ lambda = 0.667, F14,50 = 1.79, P = 0.0674), and classes
6 and 7 (Wilks’ lambda = 0.599, F14,38 = 1.82, P = 0.0689).
Size-free DFA showed no evidence for sexual dimorphism
nor did the subsequent MANOVA (Wilks’ lambda = 0.863,
F13,119 = 1.45, P = 0.135).
Allometry.—Because we found no sexual dimorphism, we
pooled the sexes in our analysis of statistic allometric trajec-
tories. Overall, 11 variables showed significant departure from
isometry (Table 5). Positive allometry was detected in the
length of nasals, breadth of incisive foramina, length of inci-
sive foramina, and breadth of incisor tips. Negative allometry
was found for breadth of braincase, zygomatic breadth, least
interorbital breadth, orbit length, breadth of occipital condyles,
breadth of M1, and length of maxillary molars M1 and M2.
Condyle-incisive length, breadth of nasals, and breadth of pala-
tal bridge were isometric (Table 5).
Geometric analysis.—The greatest amount of shape variation
was associated with the braincase (Fig. 1C). PC1geo and PC2geo
explained 49.6% and 11.6% of the total variance, respectively (Fig.
3). PC1geo was very strongly associated with landmark 6, while
landmark 9 was important, sensu Hair et al. (1987). However,
landmarks 7, 12, 13, and 15 also were important variables (Table
2; Fig. 3). PC2geo was loaded most strongly by landmark 6, while
landmarks 4, 7, 8, and 16 were important (Supplementary Data
SD2). PC1geo demonstrated a shape gradient directly associated
with age, with juveniles associated with positive PC1geo val-
ues and old adults associated with negative values (Fig. 3). Age
DFAgeo (Fig. 3) showed complete discrimination between and
among juvenile classes 2 and 3, and between juvenile classes and
adult classes 4–7. There was no clear discrimination among adults,
Table 1.—Description of the 14 linear measurements taken for the traditional morphometric method (TMM) analyses following Voss (1988) and
Flores et al. (2015).
Measurement Description
Breadth of braincase (BB) Breadth across the smooth lateral surface of the braincase immediately posterodorsal to the squamosal zygo-
matic processes.
Breadth of nasal (BN) The greatest breadth across both nasal bones.
Breadth of the incisive foramina (BIF) The greatest breadth across both incisive foramina.
Breadth of the incisor tips (BIT) Measured across the enameled tips of both upper incisors.
Breadth of the occipital condyles (BOC) The greatest breadth across the dorsal lobes of both occipital condyles.
Breadth of the palatal bridge (BPB) Measured between the protocones of first maxillary molars.
Breadth of first upper molar (BM1) Measured across the protocone-paracone cusp pair on the first maxillary molar.
Condylo-incisive length (CIL) Measured from the greater curvature of an upper incisor to the articular surface of the occipital condyle.
Least interorbital breadth (LIB) The least distance across the frontal bones at the orbital fossae.
Length of nasal (LN) Greatest length of nasal bones.
Length of maxillary molars 1 and 2 (M1–M2) The occlusal length of the maxillary molar series.
Length of incisive foramina (LIF) The length of one incisive foramen.
Orbit length (ORB) Length from anterolateral corner of frontal bone to intersection of zygomatic process of maxilla with jugal.
Zygomatic breadth (ZB) The greatest breadth across the zygomatic processes.
Table 2.—Description of cranial landmarks used in this study.
Additionally, see Fig. 1C.
LM Description of landmark
1 Lateral margin of upper incisive alveolus
2 Posterior-most point of junction between upper incisors
3 Anterior-most point of the incisive foramen
4 Premaxilla-maxilla suture lateral to incisive foramen
5 Posterior margin of zygomatic plate
6 Posterior-most point of suture between jugal and squamosal
7 Posterior-most point of M3
8 Labial-most point of M1
9 Anterior-most point of M1
10 Posterior-most point of the incisive foramen
11 Posterior-most point of the palate
12 Medial point of the junction between tympanic bullae and
Eustachian tube
13 Anterior-most point of the foramen magnum (ventral view)
14 Posterior-most point of the occipital condyle
15 Anterior-most point of junction between upper incisors
16 Anterior-most point of M3

6 JOURNAL OF MAMMALOGY
but the youngest adults (class 4) and the oldest adults (classes 6
and 7) were different. Procrustes distances for the ontogenetic
series differed significantly for pairwise comparisons of all age
groups (P ≤ 0.001) except 2 and 3 (P = 0.093). The MANCOVA
did not show evidence of sexual dimorphism (Wilks’ lambda =
0.591, F65,67 = 0.713, P = 0.4613). The first PCA axes for PCAlin
and PCAgeo were highly correlated and significant (−0.90 cor-
relation coefficients and P < 0.001).
Geometric morphometric allometric trends.—GMM trends
paralleled those of the TMM. Based on geoDFA comparisons, we
found that postweaning models for shape change as a function of
age showed that most of the shape variation was associated with
the braincase. Here, there was general elongation of the braincase
from juvenile forms to adults (landmarks 6, 13, and 14), coupled
with general elongation, particularly on the posterior end (land-
marks 9, 10, and 11) and on the anterior portion (landmarks 1,
2, and 15) of the palate, with a contraction of the breadth of the
rostrum (landmark 5). In general, landmarks associated with the
molars showed a general constriction (landmarks 7, 8, 9, and 16).
However, these are more indicative of the elongation of the brain-
case and palate than changes in the molars (Fig. 3).
Allometric trends based on all specimens showed a positive
trend in change in shape as a function of size (Fig. 4) that was
highly significant (P < 0.0001). Size predicted 45.4% of shape,
and total sums of squares (SS) = 0.137, residual SS = 0.075,
and predicted SS = 0.062.
Fig. 2.—Top) Results of principal component analysis (PCA) of 14 log-transformed linear cranial measurements for 131 specimens of Akodon
montensis. Specimens are identified by sex (males circles; females triangles) and size according to age group. The smallest-sized shapes represent
age class 2 and the largest represent age class 7 with subsequent intermediate sizes. Additionally, open shapes = age class 2, lightest gray = age
class 3, gray = age class 4, dark gray = age class 5, heavy dark gray = age class 6, and black = age class 7. Bottom). Measurements (right panel)
are described in Table 1. Results of discriminant function analysis (DFA) for 14 log-transformed linear measurements between age classes 2–7 of
A. montensis including both sexes. Age classes are identified with the same characters and colors as females on PCA analysis.

HERNANDEZ ET AL.—ONTOGENETIC VARIATION IN AKODON MONTENSIS 7
discussion
Sexual dimorphism.—Previous work on Akodon has returned
mixed results for the presence of sexual dimorphism. Within
A. boliviensis, there is evidence for significant sexual dimor-
phism (Myers and Patton 1989). This is true for A. subfuscus
populations from Arequipa Department in Peru as well, but
not for subfuscus from Cusco and Puno Departments, nor for
A. puer lutescens (Myers and Patton 1989). Analysis of A. var-
ius found no significant cranial sexual dimorphism in any of
the populations evaluated, although males tended to be between
0–3% larger than females in most dimensions (Myers 1989).
Both TMM and GMM approaches have revealed sexual dimor-
phism in A. cursor and A. montensis in Brazil (Geise et al.
2005; Astúa et al. 2015). However, using GMM, Astúa et al.
(2015) found that dorsal and lateral views of the skull revealed
dimorphism, whereas a ventral view did not.
Neither our TMM nor GMM analyses revealed sexual dimor-
phism in A. montensis. We propose 2 explanations for why our
findings differed from some previous studies. First, as suggested
by findings for A. subfuscus, lack of dimorphism in the sample
analyzed may be a consequence of geographic variation (Myers
et al. 1990). Populations in Paraguay represent some of the
southernmost and interior-most populations of A. montensis. IAF
assemblages differ from coastal assemblages (de la Sancha 2014;
de la Sancha et al. 2014), and these differences likely reflect hab-
itat differences that ultimately are expressed morphologically at
the population level. However, there does not seem to be evi-
dence for interspecific competition between the sexes in forest
remnants of Paraguay. It is noteworthy to point out that many of
the forest remnants in Paraguay included in our study are among
the largest forest remnants in Paraguay. Perhaps in smaller forest
remnants, as was observed in Brazil, there is increased competi-
tion for resources that results in sexual dimorphism.
A second explanation for why our analysis did not recover
evidence for sexual dimorphism in contrast to other groups is
that small sample sizes may increase variance, and thus reduce
the statistical power of the DFA (Strauss 2010). A general rule
of thumb for DFA is that the number of specimens should be
several times the number of variables (Struass 2010). Thus,
either small sample sizes or increased number of variables can
independently result in discrimination of groups when these
are not different (Strauss and Atanassov 2006). Each landmark
represents 2 variables. Thus, a minimum sample size of 40
specimens would be preferred. Significant sexual dimorphism
reported by other studies may be a consequence of small sample
sizes coupled with too many variables (landmarks). Our sample
sizes were sufficient for both the TMM and GMM analyses, and
we have confidence in our finding of no sexual dimorphism.
Patterns of ontogeny.—Rodents account for at least 45%
of all mammals in South America (Solari et al. 2012) and
the Sigmodontine rodents (defined as complex-penis Cricetid
rodents including Sigmodon, Akodon, Phylotis, and Orzyomys)
contain about 85 genera and at least 400 species (D’Elía
and Pardiñas 2015). Akodontini is the second most speciose
and widespread tribe within the Sigmodontinae (D’Elía and
Pardiñas 2015). In spite of the diversity, ontogenetic patterns
have been described for only a fraction of the Sigmodontines,
and many of these analyses are based on TMM methods. These
studies have used a variety of statistical approaches to quantify
ontogenetic or allometric patterns. Often, PCA is used with all
age classes of a single taxon. In this context, if most of the
variation within a species is due to growth, then PC1 provides
a proxy for quantifying allometry (Voss and Marcus 1992). In
this sense, PC1 is a strong size factor, and individual scores
indicate a measure of overall body size; character PC1 loadings
are proportional to the allometric coefficients with respect to
size (Bookstein et al. 1985). This concept is controversial, and
modern approaches have been developed to improve the quan-
tification of allometry via TMM (e.g., Giannini 2004, 2010)
and GMM (Zeldith et al. 2012), and these are used in this study.
Table 3.—Principal component (PC) loadings and percent variance for
each linear morphometric PC axis reflected in the cranial morphology
of 14 morphological measurements of Akodon montensis. Importance
of values for loadings are based on the criteria of Hair et al. (1987),
which described PC loadings > 0.30 and < −0.30 as important*, and
loadings with > 0.50 and < −0.50 as highly important**.
Variables Loadings
PC1 PC2 PC3 PC4
BN −0.246 0.144 −0.110 0.184
LN −0.383* −0.100 0.001 0.259
BB −0.059 0.117 −0.080 0.316*
ZB −0.232 −0.019 −0.019 0.128
LIB −0.061 0.020 −0.058 0.016
ORB −0.247 −0.059 −0.012 0.158
CIL −0.271 0.271 −0.115 0.247
BIF −0.390* −0.542** −0.234 −0.534**
BPB −0.291 −0.209 0.182 0.181
BOC −0.085 0.019 −0.118 0.329*
BMI −0.134 0.257 −0.831** −0.125
M1M2 −0.002 −0.136 −0.008 0.269
LIF −0.325* −0.327* 0.144 0.124
BIT −0.479** 0.590** 0.389* −0.409*
Variance (%) 0.709 0.061 0.045 0.038
Table 4.—MANOVA (permutated) results for pairwise comparisons
between age classes in Akodon montensis.
Comparisons Wilks’ λF P
All 0.076 5.6270,547 < 0.0001
Class2_3 0.120 1.5714,3 0.39
Class2_4 0.239 3.8714,17 0.006
Class2_5 0.203 7.5714,27 < 0.0001
Class2_6 0.131 9.4814,20 < 0.0001
Class2_7 0.022 47.6614,15 < 0.0001
Class3_4 0.273 4.3814,23 < 0.0001
Class3_5 0.161 12.2614,33 < 0.0001
Class3_6 0.120 13.6714,26 < 0.0001
Class3_7 0.019 76.3614,21 < 0.0001
Class4_5 0.454 4.0314,47 < 0.0001
Class4_6 0.243 8.8814,40 < 0.0001
Class4_7 0.095 23.7014,35 < 0.0001
Class5_6 0.667 1.7914,50 0.07
Class5_7 0.330 6.5314,45 < 0.0001
Class6_7 0.599 1.8214,38 0.07

8 JOURNAL OF MAMMALOGY
The revision of the Andean A. boliviensis group (Myers et al.
1990) showed that most cranial lengths and external charac-
ters seem to grow rapidly and continuously even after sexual
maturity. However, variables associated with breadth of the
braincase, the orbits, tooth rows, and the hind foot appeared to
increase more slowly, while the cranial depth showed no dif-
ferences (Myers et al. 1990). Using PCA, Myers et al. (1990)
suggested that qualitatively, the cranium and rostrum become
elongated and narrower while the zygomatic plates deepen
and distance between zygomatic arches increases. In compari-
son, Myers’ (1989) revision of the mid-elevation and lowland
A. varius group found a pattern in which measurements from
the anterior portion of the cranium increase faster than those
from the posterior portions of the cranium (i.e., the braincase
and palate).
Our TMM analyses of ontogenetic variation in A. montensis
allowed us to assess individual portions of the crania. Using
the criteria of Gianini et al. (2004) to detect positive and nega-
tive allometry, we found some variation in individual measure-
ments but an overall pattern of a fast-changing rostrum. Our
multivariate ontogenetic analysis (Gannini et al. 2004) suggests
that overall cranium size (CIL) and breadth were isometric (BN
and BPB) for the entire data set (Table 5). This is consistent
with findings for other Akodon species (Myers et al. 1990),
where anterior portions of the cranium show positive allometry.
However, negative allometry was observed on the posterior
portion of the cranium (BB, BOC, BM1), M1M2, and areas
along the orbits and zygomatic arches (ORB, LIB, and ZB).
Again, this finding is consistent with the ontogenetic patterns
described for other Akodon species (Myers 1989; Myers et al.
1990).
Postnatal development.—The DFAlin revealed that the great-
est differences occur in age groups 2 and 3. Beginning at class 4
(adults with fully erupted molars) these differences are dimin-
ished, and by age class 5 and older there is little variation in
overall shape. There are considerable differences in morphol-
ogy between the initial “adult” molar age class (class 4) and
the oldest specimens (class 6 or 7) that resulted in significant
pairwise differences (Fig. 1). This pattern is apparent in both
the TMM and GMM analyses.
Our GMM results were consistent with the patterns revealed
by the TMM analysis. An important advantage of GMM is
the ease of visualization of overall shape and size differences
between groups. Parsons et al. (2015) suggested that mamma-
lian crania are comprised of 3 modules (basicranium, calvar-
ium, and face) that are established embryonically and exhibit
varying modes of growth and function throughout ontogeny.
Our analyses reflect this modularity for the basicranium and
face, but we have no data to allow analysis of the calvarium.
It is the basicranium that shows the most change in younger
age classes, while the face exhibits the greatest change in adult
stages. As in other analyses of Akodon, we found the greatest
shape differences associated with the cranium were an elonga-
tion of the braincase and the anterior portion of the cranium.
In the TMM analysis, the greatest overall differences were
associated with the braincase or the bascranial module. Shape
variation in the cranium begins midface, near M1 and posterior
portions of the incisive foramina, with elongation and narrow-
ing of the midportion of the rostrum.
Our GMM analyses reveal negative allometry of the molar
regions that might reflect elongation of the palate and the incisive
foramina. In specimens of older individuals, this may include
elongation of the braincase. Negative allometry of the molars
reflects the conservative size of molars. Once fully erupted,
molars no longer change in size. The anterior portion of the
incisive foramen seemed to elongate anteriorly, while the pos-
terior portion of the palate elongated posteriorly slightly, poten-
tially indicative of the narrowing and elongation of the rostrum.
By age class 4, we see a continued elongation of the snout, the
region adjacent to M1 and the palatal foramina begin to expand
posteriorly, while the region around the incisors expands anteri-
orly. We begin to see the elongation and narrowing of the basi-
cranium region. At age class 5, we see a continued narrowing
and elongation of the palatal region both anteriorly and more
significantly posteriorly. At this point, we also see consider-
able narrowing and elongation of the basicranial region both
Table 5.—Results of multivariate analysis of cranial allometry in Akodon montensis for the all specimens (male and female pooled) implementing
the technique developed by Giannini et al. (2004). Allometric trend describes the general trend per individual cranial measurement including ISO
(isomentry), NEG (negative allometry), and POS (positive allometry).
Observed Departure Unbiased estimate Bias CI low CI high Allometric trend
Breadth of nasal BN 0.246 −0.021 0.246 0.000 0.217 0.276 ISO
Length of nasal LN 0.383 0.109 0.377 0.003 0.353 0.400 POS
Breadth of braincase BB 0.059 −0.214 0.053 0.003 0.029 0.077 NEG
Zygomatic breadth ZB 0.232 −0.036 0.232 0.000 0.220 0.243 NEG
Least interorbital breadth LIB 0.061 −0.209 0.058 0.001 0.036 0.080 NEG
Orbit length ORB 0.247 −0.021 0.246 0.000 0.232 0.261 NEG
Condylo-incisive length CIL 0.271 0.000 0.267 0.002 0.237 0.298 ISO
Breadth of the incisive foramina BIF 0.390 0.126 0.393 −0.001 0.353 0.433 POS
Breadth of the palatal bridge BPB 0.291 0.029 0.296 −0.002 0.259 0.333 ISO
Breadth of the occipital condyles BOC 0.085 −0.181 0.086 −0.001 0.064 0.108 NEG
Breadth of M1 BMI 0.134 −0.130 0.137 −0.002 0.102 0.172 NEG
Length of M1 and M2 M1M2 0.002 −0.321 −0.053 0.028 −0.076 −0.031 NEG
Length of incisive foramina LIF 0.325 0.055 0.322 0.001 0.291 0.354 POS
Breadth of the incisor tips BIT 0.479 0.216 0.483 −0.002 0.443 0.523 POS

HERNANDEZ ET AL.—ONTOGENETIC VARIATION IN AKODON MONTENSIS 9
anteriorly and posteriorly. By age class 6, we see considerable
narrowing and elongation of the basicranial region; the change
of the posterior portion of the palate and molar region is mini-
mal, but there is considerable elongation of the anterior portion
of the palate. Changes at age class 7 follow the same trajectory
as age class 6 but become more extreme. The patterns observed
here are consistent with qualitative descriptions of ontogenetic
changes in Akodon noted by others (Myers 1989; Myers et al.
1990).
Omnivorous generalists.—Work by Vieira et al. (2006),
Talamoni et al. (2008), and Galetti et al. (2015) demonstrates
that Akodon are omnivorous generalists. Akodon montensis
shows some morphological characteristics that are typical of
omnivores. Akodon montensis has an elongated rostrum, espe-
cially in older specimens, and when compared to the shorter
and more robust rostrum typical of other omnivores like Rattus
spp. However, the rostral elongation is not as extreme as in
typical insectivores such as Oxymycterus spp. (Samuels 2009).
Like other species of Akodon, A. montensis has relatively deli-
cate zygomatic arches compared to other omnivores. However,
they do not show the extreme slenderness of zygomatic arches
found in true insectivorous rodent species. Comparatively, A.
montensis has narrow incisor blades, typical of insectivores that
feed on hard-shelled invertebrates (Strait 1993), with moder-
ate crown heights on the cheek teeth (typical of omnivores), in
contrast to true insectivores that tend to possess considerably
Fig. 3.—Top) Results of principal component analysis (PCA) of 16 Procrustes-transformed landmarks for 131 specimens of Akodon montensis.
Specimens are identified by sex (males circles; females triangles) and size according to age group. The smallest-sized shapes represent age class 2
and the largest represent age class 7 with subsequent intermediate sizes. Additionally, open shapes = age class 2, lightest gray = age class 3, gray
= age class 4, dark gray = age class 5, heavy dark gray = age class 6, and black = age class 7. Bottom) Results of discriminant function analysis
(DFA) for 14 log-transformed linear measurements between age classes 2–7 of A. montensis including both sexes. Age classes are identified with
the same characters and colors as females on PCA analysis.

10 JOURNAL OF MAMMALOGY
reduced molars (Samuels 2009). While A. montensis has been
classified as an insectivore, this species exhibits characters
associated with omnivory. We believe our description of onto-
genetic shape change by a generalist omnivorous rodent will
provide a standard for comparison with other Akodontines and
Sigmodontine rodents, and will form the basis for future work
of form and function in evolutionary ecology.
acknowledgMents
We dedicate this paper to the memory of William “Bill” Stanley.
We thank S. Boyle, who helped with specimen collecting and
partial financial support of fieldwork. N. Giannini generously
and kindly provided his R code for linear allometry analy-
ses. We thank all the field crews in PY, with special thanks to
P. Pérez, M. L. Ortiz, L. Valdez, J. Torres, and M. Maldonado.
We thank staff at La Fundación Moisés Bertoni, Itaipú
Binacional, Guyra Paraguay, and Reserva Morombí for logisti-
cal support and access to the Paraguayan reserves. Financial
support was partially provided by a Fulbright Fellowship (US
Department of State), the Marshall Field Collection Fund of the
Field Museum of Natural History, the Mary Rice Foundation,
Lewis and Clark Exploration Fund (American Philosophical
Society), a Latin American Award (American Society of
Mammalogists), research assistantships from the Department
of Biological Sciences, (TTU), a Texas Tech Association
of Biologists Minigrant (TTUAB), the J. Knox Jones, Jr.
Memorial Endowed Scholarship (TTU), the Michelle C. Knapp
Memorial Scholarship (TTU), an AT&T McNair Fellowship
(TTU), a Hispanic Scholarship Fund Award, Sophie Danforth
Conservation Biology Fund (SDCBF) grant, and a seed grant
from the Center for STEM education and research at Chicago
State University (NDLS). Specimens were obtained and trans-
ported under collecting permit AJ No. 77/07 and exportation
permits No. 04/07; 01/09; 02/09; 07/09, authorized by the
Secretaria del Ambiente, Asunción Paraguay. Curatorial assis-
tance with specimens by B. Patterson, B. Stanley, and J. Phelps
(FMNH) is greatly appreciated. This research, in part, is based
upon work conducted using the Rhode Island Genomics and
Sequencing Center, supported, in part, by the National Science
Foundation (MRI Grant No. DBI-0215393 and EPSCoR Grant
Nos. 0554548 and EPS-1004057), the US Department of
Agriculture (Grant Nos. 2002-34438-12688 and 2003-34438-
13111), and the University of Rhode Island. Supported to JFV
included the CNPq – Brazil Science without Borders Program
(Processo: 206882/2014-9) and the Barbara E. Brown Fund for
Mammal Research. Finally, we thank J. Scheibe, A. Cardini,
and 2 anonymous reviewers whose edits and commentary con-
siderably improved the final version of this manuscript.
suppleMentary data
Supplementary data are available at Journal of Mammalogy
online.
Supplementary Data SD1.—List of specimens of Akodon
montensis used for ontogenetic analyses. All specimens were
collected in various reserves in eastern Paraguay. Locations are
listed by reserve, followed by the latitude and longitude (deci-
mal degrees) of each grid where the specimen was collected,
followed by the corresponding field ID number.
Supplementary Data SD2.—Outlines of the percent variance
and loadings of the top 8 geometric morphometric principal
component (PC) of 16 Procrustes transformed landmarks of
Akodon montensis. 1Importance of values for loadings are based
on the criteria by Hair et al. (1987), which described PC load-
ings > 0.30 and < −0.30 as important*, and loading with > 0.50
and < −0.50 as highly important**.
literature cited
adamS, D. C., F. J. roHlF, and E. S. denniS. 2004. Geometric mor-
phometrics: ten years of progress following the ‘revolution’.
Hystrix, Italian Journal of Mammalogy 71:5–16.
adamS, D. C., F. J. roHlF, and E. S. denniS. 2013. A field comes of
age: geometric morphometrics in the 21st century. Hystrix, Italian
Journal of Mammalogy 24:7–14.
alHo, C. J. R. 2005. Integration of habitats of non-volant small
mammals in the patchy Cerrado landscape. Arquivos do Museu
Nacional 63:41–48.
anderSon, M. J. 2001. A new method for non-parametric multivariate
analysis of variance. Austral Ecology 26:32–46.
antUneS, P. C., M. A. A. campoS, L. G. R. oliVeira-SantoS, and
M. E. Graipel. 2010. Population dynamics of Akodon montensis
(Rodentia, Cricetidae) in the Atlantic forest of southern Brazil.
Mammalian Biology 75:186–190.
aStúa, D., I. bandeira, and L. GeiSe. 2015. Cranial morphomet-
ric analyses of the cryptic rodents species Akodon cursor and
Akodon montensis (Rodentia, Sigmodontinae). Oecologia Australis
19:143–157.
bookStein, F. L., B. cHernoFF, R. L. elder, J. M. HUmpHrieS, Jr., G.
R. SmitH, and R. E. StraUSS. 1985. Morphometrics in evolution-
ary biology. Academy of Natural Sciences of Philadelphia, Special
Publication 15.
cardini, A. 2016. Lost in the other half: improving accuracy in geo-
metric morphometric analyses of one side of bilaterally symmetric
structures. Systematic Biology 65:1096–1106.
coUto, D., and S. A. talamoni. 2005. Reproductive condition of
Akodon montensis Thomas and Bolomys lasiurus (Lund) (Rodentia,
Fig. 4.—Allometric trajectory for geometric morphometric analysis
using the log centroid size versus regression scores. Age classes are
represented with lighter gray for age class 2 progressively to age class
7 in black. This regression shows the general change in shape as a
function of size for Akodon montensis.

HERNANDEZ ET AL.—ONTOGENETIC VARIATION IN AKODON MONTENSIS 11
Muridae) based on histological and histometric analyses of testes
and external characteristics of gonads. Acta Zoologica 86:111–118.
d’elía, G., and U. F. J. pardiñaS. 2015. Tribe Akodontini Vorontsov
1959. Pp. 463–475 in Mammals of South America. 2. Rodents (J.
L. Patton, U. F. J. Pardiñas, and G. D’Elía, eds.). University of
Chicago Press, Chicago, Illinois.
dalmaGro, A., and E. M. Vieira. 2005. Patterns of habitat utilization
of small rodents in an area of Araucaria forest in Southern Brazil.
Austral Ecology 30:353–362.
de la SancHa, N. U. 2014. Patterns of small mammal diversity in frag-
ments of subtropical Interior Atlantic forest in eastern Paraguay.
Mammalia 78:437–449.
de la SancHa, N. U., C. L. HiGGinS, S. J. preSley, and R. E. StraUSS.
2014. Metacommunity structure in a highly fragmented forest: has
deforestation in the Atlantic Forest altered historic biogeographic
patterns? Diversity and Distributions 20:1058–1070.
dempSter, A. P., N. M. laird, and D. B. rUbin. 1977. Maximum like-
lihood from incomplete data via the EM algorithm. Journal of the
Royal Statistical Society, Series B (Methodological) 39:1–38.
doS reiS, S. F., L. M. peSSoa, and R. E. StraUSS. 1990. Application of
size-free canonical discriminant analysis to studies of geographic
differentiation. Revista Brasileira de Genética 13:509–520.
FloreS, D. A., F. abdala, G. M. martin, N. P. Gianinni, and J.
M. martinez. 2015. Post-weaning cranial growth in shrew opos-
sums (Caenolestidae): a comparison with bandicoots (Peramelidae)
and carnivorous marsupials. Journal of Mammalian Evolution
22:285–303.
Galetti, M., R. GUeVara, L. A. Galbiati, C. L. neVeS, R. R. rodarte,
and C. P. mendeS. 2015. Seed predation by rodents and implica-
tions for plant recruitment in defaunated Atlantic forests. Biotropica
47:521–525.
Galiano, D., B. B. kUbiak, R. J. marinHo, and T. R. ocHotorena
de FreitaS. 2013. Population dynamics of Akodon montensis and
Oligoryzomys nigripes in an Araucaria forest of Southern Brazil.
Mammalia 77:173–179.
GeiSe, L., D. A. de moraeS, and H. S. da SilVa. 2005. Morphometric
differentiation and distributional notes of three species of Akodon
(Muridae, Sigmodontinae, Akodontini) in the Atlantic coastal area
of Brazil. Arquivos do Museu Nacional 63:63–74.
Gentile, R., P. S. d’andrea, R. cerqUeira, and L. S. maroJa. 2000.
Population dynamics and reproduction of marsupials and rodents
in a Brazilian rural area: a five-year study. Studies on Neotropical
Fauna and Environment 35:1–9.
Giannini, N. P., F. abdala, and D. A. FloreS. 2004. Comparative
postnatal ontogeny of the skull Dromiciops gliroides (Marsupialia:
Microbioheriidae). American Museum Novitates 3460:1–17.
Giannini, N. P., V. SeGUra, M. I. Gianinni, and D. FloreS. 2010.
A quantitative approach to the cranial ontogeny of the puma.
Mammalian Biology 75:547–554.
GonçalVeS, P. R., P. myerS, J. F. Vilela, and J. A. oliVeira.
2007. Systematics of species of the genus Akodon (Rodentia:
Sigmodontinae) in southeastern Brazil and implication for the bio-
geography of the Campos De Altitude. Miscellaneous Publications.
Museum of Zoology, University of Michigan 197:1–24.
Hair, J. F., R. E. anderSon, and R. L. tatHam. 1987. Multivariate
data analysis. 2nd ed. MacMillan Publishing Company, New York.
HinGSt-zaHer, E., L. F. marcUS, and R. cerqUeira. 2000. Application
of geometric morphometrics to study of postnatal size and shape
changes in the skull of Calomys expulsus. Hystrix 11:99–113.
JolicoeUr, P. 1963. The multivariate generalization of the allometry
equation. Biometrics 19:497–499.
klinGenberG, C. P. 1996. Individual variation of ontogenies: a longi-
tudinal study of growth and timing. Evolution 50:2412–2428.
klinGe nberG, C. P. 1998. Heterochrony and allometry: the analy-
sis of evolutionary change in ontogeny. Biological Reviews
73:79–123.
klinGenberG, C. P. 2011. Morphoj: an integrated software pack-
age for geometric morphometrics. Molecular Ecology Resources
11:353–357.
lima, D., B. azambUJa, V. camilotti, and N. cácereS. 2010. Small
mammal community structure and microhabitat use in the austral
boundary of the Atlantic forest, Brazil. Zoologia 27:99–105.
lópez-FUSter, M. J., R. pérez-Hernandez, and J. VentUra. 2001.
Variación cranéometrica de Rhipidomys latimanus venezuelae
(Muridae, Sigmodontinae). Orsis 16:111–120.
marcUS, L. F. 1990. Traditional morphometrics. Proceedings of the
Michigan Morphometrics Workshop 2: 77–122.
maUnz, M., and R. Z. German. 1997. Ontogeny and limb bone scal-
ing in two new world marsupials, Monodelphis domestica and
Didelphis virginiana. Journal of Morphology 231:117–130.
mcdonald, J. H. 2008. Handbook of biological staistics. 2nd ed.
Sparky House Publishing, Baltimore, Maryland.
mUñoz-mUñoz, F., and D. perpiñán. 2010. Measurement error in
morphometric studies: comparison between manual and computer-
ized methods. Annales Zoology Fennici 47:46–56.
mUSSer, G. G., and M. D. carleton. 2005. Order Rodentia. Pp. 745–
1600 in Mammal species of the world: a taxonomic and geographic
reference (D. E. Wilson and D. M. Reeder, eds.). Johns Hopkins
University Press, Baltimore, Maryland.
myerS, P. 1989. A preliminary revision of various groups of Akodon
(A. dayi, dolores, molinae, neocenus, simulator, toba and varius).
Advances in Neotropical Mammalogy 1989:5–54.
myerS, P., and J. L. patton. 1989. Akodon of Peru and Bolivia-
revision of the fumeus group (Rodentia: Sigmodontinae).
Occasional Papers of the Museum of Zoology, The University of
Michigan 721:1–35.
myerS, P., J. L. patton, and SmitH, M. F. 1990. A review of the
boliviensis group of Akodon (Muridae: Sigmodontinae) on Peru
and Bolivia. Miscellaneous Publications, Museum of Zoology,
University of Michigan 117:1–104.
naroll, R. S., and L. Von bertalanFFy. 1973. The principle of allom-
etry in biology and social sciences. Ekistics 36:244–252.
ordoñez-Garza, N., J. O. matSon, R. E. StraUSS, R. D. bradley,
and J. Salazar-braVo. 2010. Patterns of phenotypic and genetic
variation in three species of endemic Mesoamerican Peromyscus
(Rodentia: Cricetidae). Journal of Mammalogy 91:848–859.
paGlia, A. P., et al. 2012. Annotated checklist of Brazilian mammals.
Occasional Papers in Conservation Biology 6:1–76.
pardiñaS, U. F. J., et al. 2015. Genus Akodon Meyen 1833. Pp 446–
498 in Mammals of South America. 2. Rodents (J. L. Patton, U.
F. J. Pardiñas, and G. D’Elía, eds.). University of Chicago Press,
Chicago, Illinois.
pardini, R. 2004. Effects of forest fragmentation on small mammals
in an Atlantic Forest landscape. Biodiversity and Conservation
13:2567–2586.
parSonS, T. E., C. M. downey, F. R. Jirik, B. HallGrimSSon, and
H. A. Jamniczky. 2015. Mind the gap: genetic manipulation of
basicranial growth within synchondroses modulates calvarial and
facial shape in mice through epigenetic interactions. PLoS One
10:e0118355.
prado, J. R., and perceqUillo, A. R. 2011. Ontogenetic and sexual
variation in cranial character of Aeglialomys xanthaeolus (Thomas,

12 JOURNAL OF MAMMALOGY
1984) (Cricetidae: Sigmodontinae) from Ecuadoran and Peru.
Papéis Avulsosde Zoologia 51:155–177.
püttker, T., R. pardini, Y. meyer-lUcHt, and S. Sommer. 2008.
Responses of five small mammal species to micro-scale varia-
tions in vegetation structure in secondary Atlantic forest remnants,
Brazil. BMC Ecology 8:9.
roHlF, F. J. 2015. The tps series of software. Hystrix, the Italian
Journal of Mammalogy 26:9–12.
SamUelS, J. X. 2009. Cranial morphology and dietary habits of
rodents. Zoological Journal of Linnean Society 156:864–888.
SikeS, R. S., & tHe animal care and USe committee oF tHe american
Society oF mammaloGiStS. 2016. Guidelines of the American
Society of Mammalogists for the use of wild mammals in research
and education. Journal of Mammalogy 97: 663–688.
Solari, S., P. M. Velazco, and B. D. patterSon. 2012. Hierarchical
organization of Neotropical mammal diversity and its historical
basis. Pp. 145–156 in Bones, clones, and biomes: the history and
geography of recent Neotropical mammals (B. D. Patterson and L.
P. Costa, eds.). University of Chicago Press, Chicago, Illinois.
Strait, S. G. 1993. Molar morphology and food texture among
small-bodied insectivorous mammals. Journal of Mammalogy
74:391–402.
StraUSS, R. E. 2010. Discriminating groups of organisms. Pp. 73–91
in Morphometrics for morphometricians (A. M. T. Elewa, ed.).
Springer Press, Berlin, Germany.
StraUSS, R. E., and M. N. antanaSSoV. 2006. Determining best com-
plete subsets of specimens and characters for multivariate morpho-
metric studies in the presence of large amounts of missing data.
Biological Journal of the Linnean Society 88:309–328.
StraUSS, R. E., M. N. antanaSSoV, and J. A. de oliVeira. 2003.
Evaluation of principal-component and expectation-maximization
methods for estimating missing data in morphometric studies.
Journal of Vertebrate Paleontology 23:284–296.
talamoni, S. A., D. coUto, D. A. cordeiro Júnior, and F. M. diniz.
2008. Diet of some species of Neotropical small mammals.
Mammalian Biology 73:337–341.
talamoni, S. A., and M. M. diaz. 1999. Population and community
ecology of small mammals in southeastern Brazil. Mammalia
63:167–182.
tHomaS, O. 1913. New forms of Akodon and Phyllotis, and a new
genus for “Akodon” teguina. American Magazine of Natural
History 8:404–409.
UmetSU, F., and R. pardini. 2007. Small mammals in a mosaic of for-
est remnants and anthropogenic habitats - evaluating matrix quality
in an Atlantic forest landscape. Landscape Ecology 22:517–530.
Valdez, L., and G. d’elía. 2013. Differentiation in the Atlantic forest:
phylogeography of Akodon montensis (Rodentia, Sigmodontinae)
and the Carnaval-Moritz model of Pleistocene refugia. Journal of
Mammalogy 94:911–922.
Vieira, E. M., G. paiSe, and P. H. D. macHado. 2006. Feeding of
small rodents on seeds and fruits: a comparative analysis of three
species of rodents of the Araucaria forest, southern Brazil. Acta
Theriologica 51:311–318.
ViScoSi, V., and cardini A. 2011. Leaf morphology, taxonomy and
geometric morphometrics: a simplified protocol for beginners.
PLoS One 6:e25630.
VoSS, R. S. 1988. Systematic and ecology of Ichthyomyine rodents
(Muroidea): patterns of morphological evolution on small adaptive
radiation. Bulletin of the American Museum of Natural History
188:259–493.
VoSS, R. S., and L. F. marcUS. 1992. Morphological evolution
in Muroid rodents II. Craniometric factor divergence in seven
Neotropical genera, with experimental results from Zygodontomys.
Evolution 46:1918–1934.
VoSS, R. S., L. F. marcUS, and P. eScalante. 1990. Morphological
evolution in Muroid rodents I. Craniometric factor diveren-
cein seven neotropical genera, with experimental results from
Zygodontomys. Evolution 44:1568–1587.
zelditcH, M. L., F. L. bookStein, and B. L. lUndriGan. 1993. The
ontogenetic complexity of developmental constrains. Journal of
Evolutionary Biology 6:621–641.
zelditcH, M. L., D. L. SwiderSki, and H. D. SHeet. 2012. Geometric
morphometric for biologist: a primer. 2nd ed. Academic Press, San
Diego, California.
Submitted 14 January 2017. Accepted 22 September 2017.
Associate Editor was John Scheibe.