Morphological variation in bony structures of Hylaeamys seuanezi (Rodentia, Cricetidae, Sigmodontinae) in Southern Bahia, Brazil

1 Introduction

Hylaeamys seuanezi (Weksler et al. 1999), is an omnivorous rodent belonging to the Oryzomyini tribe (Weksler et al. 2006). Morphological characters that diagnose the species include a dorsal fur varying from yellowish brown to more intense brown and a greyish white ventral fur; the tail is slightly smaller or equal to head and body length (126–174 mm), and the hind feet are long and robust, with fleshy hypothenar pads and sparse ungual tufts shorter than claws (Brennand et al. 2013; Percequillo 2015; Weksler et al. 1999). H. seuanezi, has a small skull, with the supraorbital region beaded, the palatal processes developed and frequently not fused with the maxillary bones. The incisive foramina are large with a maxillary septum long and wide. The Postglenoid and stapedial foramina are present and the sphenofrontal foramen and squamosal channel are absent (Weksler et al. 1999). The body mass averages between 51 and 97 g and the karyotype reported has a chromosomal number of 2n = 48 and a number of autosomal arms of FN = 60 (Brennand et al. 2013; Percequillo 2015; Weksler et al. 1999).

The species was originally described with samples from eight localities in Brazil, encompassing the Eastern portion of the Atlantic forest, in the States of Bahia, Minas Gerais, Espirito Santo, and Rio de Janeiro (Weksler et al. 1999). Currently, H. seuanezi is considered an endemic taxon from the Atlantic forest, inhabiting from sea level to 600 m and with a wide geographic distribution in lowland vegetation and on the slopes of highland forests; both associated with coastal Brazil (Brennand et al. 2013; Prado and Percequillo 2013; Weksler et al. 1999).

The habitats currently known for this rodent include agroforestry systems, traditionally established for the production of cocoa (Theobroma cacao Linnaeus) in the Southern State of Bahia, locally called as “cabrucas” (Cassano et al. 2011; Silva et al. 2020); lowland evergreen forests in Espirito Santo and Minas Gerais, as well as restinga vegetation and seasonally flooded evergreen forests in Rio de Janeiro (Brennand et al. 2013).

The taxonomic history of H. seuanezi has been confusing and with various nomenclatural arrangements in the past, as has occurred with most representatives of the Oryzomyini tribe (Hershkovitz 1960). Before its original description, the individuals of that species were considered within the paraphyletic group, Oryzomys capito (Musser et al. 1998), with the name of Oryzomys laticeps (Lund). Then, Weksler et al. (1999), carried out a morphological and metric review, accompanied with cytogenetic information for the populations from the Amazonian lowland, Cerrado and Atlantic Forest, defined by them as belonging to the “O. capito-O. laticeps” group and describe a new species named Oryzomys seuanezi, which is closely related to Oryzomys oniscus (Thomas) and O. capito (Hershkovitz). However, the specific epithet of O. laticeps continued being used in taxonomic lists, as the valid name for the populations from those areas of geographic distribution referred for O. seuanezi (e.g., Pardini 2004; Prado and Percequillo 2013).

Subsequently, with the support of morphological, cytogenetic, and molecular evidence, the genus Oryzomys was divided into 10 new genera (Weksler et al. 2006), and O. seuanezi was included as a synonym of O. laticeps, forming part of Hylaeamys, that currently contains the species of Oryzomys megacephalus group (Weksler et al. 2006): H. acritus (Emmons and Patton); Hylaeamys laticeps (Lund); H. megacephalus (Fischer); H. oniscus (Thomas); H. perenensis (J.A. Allen); H. tatei (Musser, Carleton and Gardner), and H. yunganus (Thomas).

On the basis of morphometric, morphologic, and cytogenetic evidence the name of H. seuanezi was recovered from the synonymy with H. laticeps and considered as a valid species occurring on the Southern bank of the São Francisco river, from the Southern portion of Bahia to the Northern portion of Rio de Janeiro (Brennand et al. 2013). However, in the last taxonomic and geographic list of rodents from South America (Patton et al. 2015), it has been treated—without explanation—as a junior synonym of H. laticeps (Percequillo 2015); name that is not recognized in the updated list of Brazilian mammals (Quintela et al. 2020). Quintela et al. (2020), maintained the name of H. seuanezi proposed by Weksler et al. (1999) and supported by Brennand et al. (2013), but they suggested a deeper review of morphological characters and implementation of additional tools, such as molecular techniques to try to stabilize the nomenclature of the species. It is currently considered a monotypic species with wide geographic distribution in Northeastern and Southeastern Brazil and a high abundance of its individuals in some localities where it has been recorded (Figueiredo et al. 2017; Pardini 2004; Percequillo 2015; Silva et al. 2020).

Due to different geometric morphometric studies on mammals have yielded interesting data on the variation in the shape of bony structures and their relationship with different habitats or food resources (Kubiak et al. 2017; Marchán-Rivadeneira et al. 2010, 2012; Samuels 2009), it would be interesting evaluate if the loss, fragmentation or modification of habitats that has occurred in the Atlantic forests of the South of Bahia in Brazil, since ancestral times, has modified the morpho-anatomical conformation for different species of small mammals that coexist in those environments. Recent biological material collected within the geographic range of H. seuanezi (Bahia State, Brazil), and deposited in natural history collections as “voucher specimens” offers an opportunity to answer these types of questions, in addition to continue documenting taxonomic and geographic information and to try to stabilize the taxonomic status of H. seuanezi as a full species. Consequently, the present study aimed to provide information on intraspecific morphological variation in bony structures (skulls, mandibles and postcranial skeleton), of different populations of this rodent, using for the first time, geometric morphometric methods (Bookstein 1991).

2 Materials and methods

We review a total of 145 individuals of H. seuanezi with skulls, mandibles, and postcranial skeletons (scapula and pelvis) available, but not in equal quantities per individual. All biological material is housed in the Coleção de Mamíferos “Alexandre Rodrigues Ferreira” (CMARF), from the Universidade Estadual de Santa Cruz (UESC), State of Bahia, Brazil (Appendix A). According to age criteria based on toothwear (Voss et al. 1990), and evidence of reproductive activity (scrotted testicles in males and pregnant, lactating, or post-lactating females), individuals were classified into four age classes: Class 1 (juveniles), class 2 (young adults), class 3 (adults), class 4 (old adults).

Previous to morphometric geometric analyzes, we evaluate the possible effect of joining all adult age classes (2–4) and analyzing them as a set (Voss et al. 1990). We also, evaluate the sexual dimorphism and the results of both comparisons allowed us to discriminate or not the samples of 145 individuals for the subsequent analyses. For both comparisons, we use 13 cranial metric variables reported in Brennand et al. (2013).

The three age classes considered as adults were compared with a 1 – way analysis of variance (ANOVA), and the differentiation between sexes (sexual dimorphism), was evaluated with a t-test. All statistical analyzes in this phase were performed with the help of the program PAST (Hammer 2020) and with a probability value of alpha less than or equal to 0.05 (p ≤ 0.05). Subsequently, the specimens were then sorted into groups or morphotypes, according to the geographic proximity criteria and similarity between external-cranial morphological characters (Brennand et al. 2013; Musser 1968). The grouping resulted in samples more robust from various environments located in four municipalities from Southern Bahia, Brazil (Figure 1); all belonging to the Atlantic forest, characterized by flora and fauna elements unique of that ecosystem and with landscapes highly fragmented by anthropic activities (Pardini 2004).

Figure 1:

Geographic location of four populations of Hylaeamys seuanezi, from the Southern Bahia, Brazil, used in this study.

The selected localities were: Cabruca agroforestry systems (Ilhéus municipality), located on the University campus of the Universidade Estadual de Santa Cruz (UESC); it is an area of cocoa production (T. cacao), with sustainable management to protect the remnants or patches of Atlantic forest (Cassano et al. 2011); Fazendas Carirí, Nova Angelica, Juerana and Dona Mihuke (Una municipality), located in an area that maintains approximately 50% of the native Atlantic forest (Pardini 2004) and where there are currently two Conservation Units that restrict the use of its natural resources (Wildlife Refuge and Biological Reserve of Una); Fazendas Oro Verde and Boca de Corrego (Belmonte municipality), are a heterogeneous matrix of native Atlantic forest patches, with Cabruca agroforestry systems and open areas (pastures, secondary forests); Fazendas Reunidas do Vale do Juliana (Igrapiúna municipality), also include an area of ​​remnants of Atlantic forest with areas impacted by various anthropic activities (silvicultural activities and pastures).

2.1 Morphometric geometric analysis

Dorsal views of each skull, as well as the labial views of the mandibles, occlusal right scapulae, and lateral right pelvis, were selected for digitization. All photographs were captured by the same person using a digital microscope (KKmoon 1600×, with eight Led and endoscope camera), and a tripod. A ruler graduated in millimeters was placed next to each bony structure as reference. Type I and II bidimensional homologous morphological landmarks were digitized (Bookstein 1991), using the softwares tpsUtil and tpsDig (Rohlf 2010, 2015). Landmark names and locations were to those documented by Astúa (2008), in scapulae, and Astúa et al. (2015), to skulls and mandibles. Landmark locations on the pelvis were selected to encompass zones that have taxonomic diagnostic value for H. seuanezi. The total number of landmarks digitized on each view were as follows: Dorsal skulls (18), mandibles (11), right scapulae (10), and right pelvis (27) (Figure 2).

Figure 2:

Location of morphological landmarks (type I and II) in a specimen of Hylaeamys seuanezi (CAMRF: 3193). (A) Dorsal view of the skull; (B) labial view of the mandible; (C) occlusal view of the scapula; and (D) lateral view of the pelvis. Scale = 5 mm.

The morphometric geometric and statistical analyses were conducted using the program MorphoJ (Klingenberg 2011). First, we verify the optimal digitization and location of the morphological landmarks in each photograph, reviewing the possible presence of outliers or atypical landmarks and then we calculate the margin of error; to do that, each view was digitized in duplicate and the total morphological landmarks and the two resulting data matrices were verified with a Procrustes analysis of variance (ANOVA-Procrustes). It is assumed that when the mean square value of the error in the ANOVA-Procrustes is less than the value of the mean squares of the individuals, the data are well digitized (Souto et al. 2019); in our case, the values fit that assumption and therefore either set of photos was used.

Morphological landmarks were associated with Cartesian coordinates (x, y) that represented the geometric configuration of each bony structure. The coordinates were subject to a Procrustes adjustment, which removes the variations in size due to the position, orientation, and scale of each image (Klingenberg 2011), and generated three matrices known as Procrustes residuals (Canonical coefficients, Centroid size, and Procrustes coordinates), which were used in subsequent exploratory analyzes.

With the matrix of Canonical Coefficients, we calculate the maximum variation of the different individuals by structures between groups, using a Multivariate Analyses of Variance (MANOVA), permuted at 10.000 iterations pairwise. With the matrices of Centroid Size and Procrustes Coordinates, we calculate equally, the average shape in structures with a Principal Component Analysis (PCA). The statistic tests of distances of Mahalanobis and Procrustes generated in the exploratory analyses were verified with a probability value of alpha less or equal to 0.05 (p ≤ 0.05), on a previous classification of H. seuanezi by municipalities of Bahia State, Brazil: H. seuanezi (Igrapiúna), H. seuanezi (Ilhéus), H. seuanezi (Una) and H. seuanezi (Belmonte).

To test the possible differentiation between sexes for each evaluated structure, discriminant functions were applied, also using the statistic tests of distances of Mahalanobis and Procrustes (p ≤ 0.05), with cross-validation permuted at 10.000 iterations pairwise. Finally, to observe the changes of the average shape in the different structures, graphs of the type Wireframe and deformation grids were used.

3 Results

The age analyses yielded the following classification in four age classes: Class 1 (33 individuals), Class 2 (84 individuals), Class 3 (17 individuals), and Class 4 (11 individuals). In the juveniles (Class 1), the majority of the individuals had usually the third molar nonerupted or newly erupted, with their lingual and labial cusps very close to each other and without dentin exposure. In young adults (Class 2), the 1st and 2nd molars already showed dentin exposure and separation of their cusps, with the third molar showing greater wear than the rest; the anteroloph and mesoloph were connected to paracone and the posteroloph was fused to the metacone. In full adults (Class 3), the molar toothrow was marked worn, with a nearly flat surface, with greater exposure of dentin on its surfaces and with the anteroloph, and mesoloph fused marginally to paracone; the posteroloph was completely fused to the metacone. In older adults (Class 4), the cusps were indistinct on all molars, with the dentin being the only identifiable feature on the entire tooth surface; the anteroloph, mesoloph, and posteroloph were indistinct and fused to major cusps. In Figure 3, the comparisons between the rows of upper molars of the different age classes found in the examined sample are shown together with the average size of their skulls.

Figure 3:

Visual representation of age classes in populations of Hylaeamys seuanezi, from samples analyzed from the Southern Bahia, Brazil. (A) Age classes based on toothwear pattern of the upper molar row (1600× magnified); (B) occlusal view showing differentiation in skull size associated with four age classes. Scale = 10 mm.

The results of ANOVA’s test showed significant differences between adults in age classes 2–4 for eight of 13 variables (Condylo-incisive length: F, 7.956, p = 0.0005901; Length of diastema: F, 32.91, p = 5.993⁻¹²; Length of incisive foramina: F, 6.543, p = 0.00206; Length of nasals: F, 18.13, p = 1.534⁻⁰⁷; Length of bony palate: F, 10.68, p = 5.713⁻⁰⁵; Greatest zygomatic breadth: F, 3.603, p = 0.03048; Breadth of zygomatic plate: F, 3.587, p = 0.03094; Orbital fossa length: F, 19.76, p = 4.571⁻⁰⁸). Therefore, the subsequent analyzes were carried out with the age class best represented in each locality; in this case, the age class 2 (n = 84).

The results of Student’s t-test showed significant differences between males and females in age class 2 only in the variable Breadth of 1st upper molar (t, 2.3176, p = 0.022962). For the rest of the 12 variables, there were no statistically significant differences in sexual dimorphism: Condylo-incisive length: t, 0.27937, p = 0.78066; Length of diastema: t, 0.88294, p = 0.37985; Crown length of maxillary toothrow: t, 0.7353, p = 0.46425; Length of incisive foramina: t, 0.14818, p = 0.88257; Breadth of incisive foramina: t, 0.82308, p = 0.41285; Breadth of rostrum: t, 1.3798, p = 0.17141; Length of nasals: t, 0.27161, p = 0.7866; Length of bony palate: t, 0.76817, p = 0.44459; Interorbital breadth: t, 1.066, p = 0.28956; Greatest zygomatic breadth: t, 1.737, p = 0.86135; Breadth of zygomatic plate: t, 0.26349, p = 0.79283; Orbital fossa length: t, 1.3086, p = 0.19432.

Regarding the morphometric geometric analyzes, the discriminant functions in the dorsal views of skulls, labial of mandibles, and occlusal of scapulae showed that the female and male individuals are not different (Figure 4). However, the lateral view of the pelvis was statistically different (Procrustes distances: 0.03014843, p = 0.0040; Mahalanobis distances: 8.9283; p = <0.0001). Therefore, supported by cranial metric data and geometric morphometric comparisons both sexes were grouped for subsequent analyzes related to the skull, mandible, and scapula, and in the case of the pelvis, the data were analyzed separately between males and females.

Figure 4:

Histograms of frequencies of discriminant functions resulting from comparisons between males and females of Hylaeamys seuanezi for four bony structures from populations from the Southern Bahia, Brazil.

The variation between the groups derived from the MANOVA, for the first two axes, produced the following values of eigenvectors and accumulated variation percentages: Dorsal view of the skull (λ1: 2.62461704, λ2: 1.18718139; S ²: 84.576), labial view of the mandible (λ1: 2.26548694, λ2: 0.98421634; S ²: 92.107), occlusal view of the scapula (λ1: 1.38815982, λ2: 0.89410211; S ²: 81.242), and lateral view of the pelvis (λ1: 82.14621713, λ2: 14.41551863; S ²: 78.548). In Table 1, are documented the comparisons with the cross-validation permuted values ​​between each of the populations, associated with each view. In Figure 5, is observed the dispersion graph derived from the MANOVA, for the four populations of H. seuanezi studied.

Figure 5:

Factorial diagram of the Analysis of Canonical Variables, showing the distribution in the morphospace and correlation of four bone structures in populations of Hylaeamys seuanezi from the Southern Bahia, Brazil. The colors indicate the study municipalities: red = Igrapiúna, green = Ilhéus, black = Una, yellow = Belmonte. In the lateral view of the pelvis, the red and yellow colors are assigned for females and males from Igrapiúna respectively; dark green (female) and light green (males) from Ilhéus; black (females) and gray (male) from Una, dark blue (female) and light blue (males) from Belmonte.

The variation between the average shape of the groups derived from the PCA, for the first two components, produced the following values of eigenvectors and accumulated variation percentages: Dorsal view of the skull (λ1: 0.00025875, λ2: 0.00017869; S ²: 36.917), labial view of the mandible (λ1: 0.00045074, λ2: 0.00004903; S ²: 95.017), occlusal view of the scapula (λ1: 0.00106198, λ2: 0.00029495; S ²: 91.190), and lateral view of the pelvis (λ1: 0.00064953, λ2: 0.00029881; S ²: 91.181). In Figure 6 is observed the dispersion of the different populations related to the average shape of each evaluated structure.

Figure 6:

Morphospace of the principal component analysis, showing the separation derived from the comparison of four bony structures in populations of Hylaeamys seuanezi from the Southern Bahia, Brazil. The figures in dispersion graphs represent the average shape for each structure. In the lateral view of the pelvis, the color black is referred to females and gray for males.

Regarding bony morphology by view, the dorsal skull average shape of H. seuanezi from Igrapiúna differs from populations of adjacent areas by having a different morphology in the posterior region of the parietals (morphological landmarks: 7–8 and 12–13), anterior and posterior region of orbital (landmarks 4–5 and 15–16), and the most posterior part of the nasal bones (landmark: 1). In the three remaining populations (Ilhéus, Una, and Belmonte), such morphological differences are not marked in parietals and interorbital region; however, in the case of the Belmonte population, the nasal bones are shorter; this derived from the vector orientation of landmark 1 (Figure 7).

Figure 7:

Wireframe graphs (left) and deformation grids (right), showing the average shape of the skull in four populations of Hylaeamys seuanezi, from the Southern Bahia, Brazil.

Likewise, the morphology of the mandible in H. seuanezi from Igrapiúna was different in the anterior (anterior-posterior alveolar region, landmarks: 1–2–10–11) and posterior part (condyles, landmarks: 5–6–7; angular process, landmarks: 8–9–10), concerning the three remaining populations (Ilhéus, Una, and Belmonte). The anterior-posterior alveolar region (landmarks: 1–2–10–11) and condylar region (landmarks: 5–6–7) from the Ilhéus population differs subtly from the Una and Belmonte populations, which are morphologically the most similar (Figure 8).

Figure 8:

Wireframe graphs (left) and deformation grids (right), showing the average shape of the mandible in four populations of Hylaeamys seuanezi, from the Southern Bahia, Brazil.

In scapular morphology was observed a trend of three shape patterns with the populations of H. seuanezi from Igrapiúna and Ilhéus, both with marked difference in the anterior region of the scapula (caracoid process, acromion, and meta-acromia, landmarks: 1–2–3–10), upper and lower margins of neck of scapula (landmarks: 4–9) and at the axillary edge of infraspinatus fossa (landmark: 5) and the populations from Una with Belmonte being more similar in those same areas (Figure 9).

Figure 9:

Wireframe graphs (left) and deformation grids (right), showing the average shape of the right scapula in four populations of Hylaeamys seuanezi, from the Southern Bahia, Brazil.

The morphology of the pelvis varied between females and males from the same locality, as well as between the same sexes from different places (Figure 10). In Igrapiúna, the average shape of the pelvis in females was different concerning males in the anterior region (between the gluteal fossa, the iliac fossa, and the greater sciatic notch, landmarks: 1–2–3–19–20), and in the obturator foramen (landmarks: 21–22–23–24–25–26–27). In Ilhéus, females and males were subtly differentiated in the posterior region of the lesser sciatic notch and at the lower part of that notch (landmarks: 12–14). In Una, both sexes were equally differentiated in the lower part of the lesser sciatic notch (landmark: 12) and the anterior region of the iliac fossa (landmark: 20). In Belmonte, both sexes were different in the lower part of the lesser sciatic notch (landmark: 12), and the obturator foramen (landmark: 25).

Figure 10:

Wireframe graphs (left) and deformation grids (right), showing the average shape of the pelvis between sexes in four populations of Hylaeamys seuanezi, from the Southern Bahia, Brazil.

The Igrapiúna females differed from other localities in the obturator foramen (landmarks: 21–22–23–24–25–26–27). The females from Ilhéus differentiated from the females from Una in the region of the iliac fossa and greater sciatic notch (landmark: 19), also in the lower part of the lesser sciatic notch (landmark: 12), and with the ones from Belmonte in the region of the iliac fossa (landmark: 19), the lower part of the lesser sciatic notch (landmark: 12) and in the obturator foramen (landmark: 25). Likewise, the males from Igrapiúna differed from the rest of the males of the other localities in the obturator foramen (landmarks: 21–22–23–24–25–26–27). The males from Ilhéus differed from the ones from Una in the lower part of the lesser sciatic notch (landmark: 12), and in the region of the iliac fossa (landmark: 19), and with the males from Belmonte in the greater sciatic notch (landmark: 19). Males from Una and Belmonte have similar pelvic morphology but some differences were observed in the region of the iliac fossa and the greater sciatic notch (landmark: 19).

4 Discussion and conclusion

Our results showed four morphological patterns, derived from the average shape in bony structures of H. seuanezi populations from the Southern Bahia, Brazil considered as young adults. Unfortunately, the low samples available from the other adult age classes and the significant bias that exists when mixing different age classes in small mammals (Voss et al. 1990), did not allow a more detailed analysis with the entire adult set. However, our analysis of class 2 suggests that the populations inhabiting Igrapiúna are the most differentiated. The Ilhéus population was more similar to Igrapiúna in almost all the structures studied; while the populations of Una and Belmonte tended to be more similar to each other and different from Igrapiúna and Ilhéus respectively.

Morphologically, H. seuanezi from Igrapiúna, has an elongated and markedly narrow braincase in the parietal region, with a narrow interorbital region and long nasals (evidenced by a lengthening in the posterior region of nasals towards the frontal bones of the skull). The remaining populations have skulls with globose shapes and with differences in shape between them; not as marked as in the Igrapiúna population, but evident. The mandible of the Igrapiúna population is characterized by having angular, coronoid, and masseteric processes wide, in addition to an anterior-posterior alveolar region equally wide; while, in the remaining three populations, the mandibular shape is more similar.

The scapulae of the Igrapiúna and the Ilhéus populations are markedly different, and Una and Belmonte share a similar morphology in this structure. In Igrapiúna, the upper and lower margins of the scapular neck and the caracoid, acromian and metachromic processes are narrow with a wide vertebral border, giving the appearance of a large and elongated scapula. In Ilhéus, the scapular morphology is equally large but also robust. For Una and Belmonte, the narrowness of the axillary border is similar, but they are different from each other, in the caracoid, acromian, and metachromic processes.

In the pelvic shape, the obturator foramen is the best differentiated area and the Igrapiúna females stand out for having a distinctive foramen concerning the other females, as well as from all the males. This foramen is large in the Igrapiúna population, and this makes the pelvis appear more robust than the rest. There are also differences associated with the pubic region where the posteroventral extension is elongated for Igrapiúna and Ilhéus females and shorter for Una and Belmonte. Similarly, the anterior part of the Ilion is wider for Igrapiúna and Ilhéus and the opposite for Una and Belmonte. Concerning males, the shape is similar to that found for females in the pubic region, but Ilhéus males have a wider Ilion than the rest.

Comparing the high morphological variation evidenced in our study to H. seuanezi, with similar studies on morphological variation in mammals documented in the past, most of those studies generated information that supported the recognition of the species, with previous controversies in their taxonomic identities (Louzada and Pessôa 2013; Prevosti et al. 2013), in addition to documenting intraspecific geographic variation for a better understanding of geographic distribution patterns and possible taxonomic arrangements at lower hierarchical levels, such as subspecies (Damasceno and Astúa 2016). Another important aspect related to document the presence of morphological variation, are adaptations of the species to various environmental pressures that may be occurring throughout the geographic distribution of an organism (Cardini et al. 2007; Zelditch et al. 2004).

Concerning sexual dimorphism, the morphometric geometric results for the dorsal view of skulls and mandibles coincide with the findings of Brennand et al. (2013), in the cranial non-differentiation between males and females using cranial linear measurements as discriminant variables for H. seuanezi. Similar results of non-differentiation between sexes in skulls and mandibles were found for the aquatic marsupial, Chironectes minimus throughout its gradient of geographic distribution in Central and South America, combining linear and geometric morphometric techniques (Damasceno and Astúa 2016). On the other hand, there are opposite reports for other groups of mammals. In species that were previously considered non-dimorphic with linear variables, cranial differences between the sexes were found using geometric morphometry (Hood 2000). Realize that when we applied discriminant functions on geometric variables from other bony structures, as the scapula and pelvis, the pelvis evidenced the differences between sexes.

A similar case was reported by Schutz et al. (2009), using geometric morphometry in skulls, mandibles, and pelvis to compare sexual dimorphism in two species of canids (Urocyon littoralis and Urocyon cinereoargenteus). These authors didn’t find evidence of sexual dimorphism in skulls and mandibles, but just in the pelvis of U. littoralis. Therefore, it is advisable to add, whenever available, other bony structures in addition to the skull. The post-cranial skeleton provides more accurate information about intraspecific variation (Chemisquy et al. 2015; Morgan 2008) and corroborates that using different anatomical structures morphological characterization of organisms is improved (Cordeiro-Estrela et al. 2008; Grilliot et al. 2009).

It is possible that the different environmental and ecological conditions in each of the investigated areas may be playing a selective pressure in the phenotypic plasticity evidenced in this study for the populations of H. seuanezi. The literature described the Southern Bahia as a heterogeneous landscape composed of remnants of native Atlantic forests, cacao agroforests, monocultures of rubber plantations (Hevea brasiliensis) and Erythrina spp., and open areas (pastures, annual plantations, and urban areas), where the extensions of each type of vegetation vary in percentages of land covered (Faria et al. 2007). Therefore, these contrasting scenarios may affect the natural history of many mammals classified as indicators of potential sensitivities (Ochoa 2000), but notedly for H. seuanezi as forest specialist species (Estavillo et al. 2013; Pardini et al. 2010) that may be responding through morphological architecture and plasticity to environmental constraints.

Having that in mind, the great differentiation the Igrapiúna population showed in the evaluated structures could be the product of adaptations due to loss of forest cover, or consequences of global climate change or perhaps ecological interactions (such as intra and interspecific competition). It has already been reported using geometric morphometry in some rodents species that there is a correlation of anthropogenic, climatic, ecological, and competitive pressures on phenotypic and genotypic characteristics, deriving in contrasting shapes for individuals of the same species or from different genera (Bubadué et al. 2016; Chemisquy et al. 2015; García et al. 2014; Gay and Best 1996; Hendges et al. 2016; Kubiak et al. 2017; Ledevin and Millien 2013; Maestri et al. 2015; Monteiro et al. 2003; Pergams and Lawler 2009; Wilder et al. al. 2005). Additionally, different ecological studies that evaluated the responses of small mammal communities (including H. seuanezi) to disturbances of their habitats documented negative responses associated with low abundances and richness or asymmetry in different orders of this group, product of environmental modifications and forest fragmentations (Cosson et al. 1999; Filho et al. 2018; Goodman and Rakotondravony 2000; Laurance 1994; Pardini 2004).

In our case, the ideal scenario would be to measure in the possible future effects of some ecological and environmental factors on the populations of H. seuanezi on a larger time scale. Unfortunately, all the samples analyzed in this study were collected in recent years (2016–2019), which did not allow us to have a broader panorama on a long-timescale of the same areas with different collection years; This type of study would be interesting to compare the shapes with the changes of habitats; for example, between decades, as has been done in some surveys with small rodents (e.g., Pergams and Lawler 2009). Also, it is important to mention that the Atlantic Forest biome is one of the most threatened on the planet and it is likely that the increase in loss of extensions of forest cover will continue and therefore, ecosystems will continue to change with stronger threats to the biota of that place (Pardini 2004; Ribeiro et al. 2009).

In conclusion, the morphological variation documented with geometric morphometry for different populations of H. seuanezi is a line of evidence that contributes to the taxonomic knowledge of the species and therefore is necessary to expand this type of study along with a detailed review of external characters, such as integuments, scales, and color patterns for the populations studied, and also include and compare geometrically samples from the other localities reported for this taxon in Northeastern and Southeastern Brazil, for which information on external and metric characters has already been generated (Brennnand et al. 2013; Weksler et al. 1999).