Morphological symmetry of Rhipidomys mastacalis (Mammalia, Rodentia, Cricetidae) in fragmented habitats of the Atlantic Forest in Northeastern Brazil: a study on the influence of the environment on an endemic species

1 Introduction

The Brazilian Atlantic Forest is known for its exceptional flora and fauna; however, it is also one of the world’s most threatened ecosystems due to the high deforestation rates. As a result, the landscape has become highly fragmented, where silviculture, monoculture, agriculture, and urban systems coexist (Ribeiro et al. 2009). Studies conducted between 2019 and 2020 indicate that deforestation has intensified in 10 out of the 17 federal units within the biome’s geographical distribution, affecting an area of 130 km² (Fundação SOS Mata Atlântica 2021).

The literature reports that the Northeastern region of Brazil is being transformed into a matrix composed of native Atlantic Forest, cocoa plantations, monocultures (e.g., plantations of Hevea brasiliensis, Erythrina spp., and Saccharum spp.), pastures, and urban areas, with each type of vegetation covering different percentages of the region (Cassano et al. 2021; Pardini 2004). Consequently, these contrasting scenarios can affect many species of small non-volant mammals, particularly specialist habitat species, that are highly sensitive to environmental disturbance (Ochoa 2000; Pardini 2004).

Evaluating whether the fragmentation, modification, or loss of habitat in the Atlantic Forest can alter bilateral symmetry in small non-volant mammal species is useful for monitoring environmental integrity. Morphologically, this group has a low ability to disperse, restricting most of their mobility to small areas (De-La-Cruz et al. 2019; Diffendorfer et al. 1995; Machado 2019; Pires et al. 2002). Moreover, they are habitat specialists, making them more vulnerable to environmental stress, which refers to any condition that imposes disturbance or pressure on organisms and may affect their developmental stability (Benítez and Parra 2011).

Several studies in forested areas that evaluated the impacts of fragmentation and habitat loss on non-volant mammal populations reported an effect of environmental stress associated with changes or fragmentation of forest cover (Badyaev et al. 2000; Bonvicino et al. 2002; Caccavo et al. 2021; Cassano et al. 2009). Thus, it has been postulated that this type of alteration is correlated with negative effects on morphology and loss of genetic variability, affecting developmental stability, which hypothetically would lead to a decrease in growth and reproduction rates of a population, inbreeding, local extinction, or the total disappearance of a population or species (Badyaev et al. 2000; Caccavo et al. 2021; Coda et al. 2016, 2017; Wauters et al. 1996).

Developmental stability refers to an organism’s ability to develop an ideal phenotype despite environmental and genetic pressures during ontogeny (Benítez and Parra 2011). However, when natural systems experience rapid changes due to human disturbances, the development systems may produce morphological changes known as developmental instability or developmental noise (Benítez and Parra 2011).

Symmetrical indicators on the body plane of organisms can be used to detect whether development systems are affected (Klingenberg 2015). Three types of symmetrical indicators are classified in populations depending on the values of the right and left sides of a structure: fluctuating asymmetry, directional asymmetry, and antisymmetry (Benítez et al. 2020; Klingenberg 2015; Leary and Allendorf 1985). These indicators differ in their mathematical and statistical properties, as well as in their origins and biological implications (Benítez et al. 2020).

Fluctuating asymmetry is particularly useful for detecting morphological changes in bilateral organisms due to small differences between the right and left sides resulting from random processes of development alterations associated with environmental and genetic stress (Caccavo et al. 2021; Graham et al. 2010; Klingenberg 2015; Leung et al. 2000). Environmental stressors include temperature, forest cover, parasitism, competition, predation, and resource availability, while genetic stressors include loss of genetic variability, inbreeding, low heterozygosity, and hybridization (Benítez and Parra 2011; Coda et al. 2016; Leary and Allendorf 1985). Anthropogenic stressors, such as habitat fragmentation or loss, toxic products used in insect control on agricultural crops, and contaminated areas resulting from environmental disasters, also generate stress (Allen and Leamy 2001; Anciães and Marini 2000; Caccavo et al. 2021; Cuervo and Restrepo 2007; Oleksyk et al. 2004).

There are two types of analyses to detect fluctuating asymmetry in the morphological structures of an organism: object asymmetry and correspondence asymmetry (Klingenberg 2015). Object asymmetry involves dividing the structure into an imaginary sagittal plane and comparing the two sides (left and right), while correspondence asymmetry involves duplicating the structure in the organism, such as the mandibles or the locomotion limbs of vertebrates, and comparing the two structures that represent the right and left sides (Graham et al. 2010; Klingenberg 2015).

Recent studies that used fluctuating asymmetry with linear techniques (size asymmetry) or with geometric morphometrics (shape asymmetry) found that animal populations under natural conditions tend to have low levels of fluctuating asymmetry, while populations exposed to environmental, genetic, and anthropogenic stress have high levels of this asymmetry, indicating a negative relationship between morphological changes and stressor variables (Caccavo et al. 2021; Leung et al. 2000; Maestri et al. 2015; Oleksyk et al. 2004; Shadrina and Vol’pert 2016).

The present study aimed to compare the presence and levels of fluctuating asymmetry in anatomical structures (skulls, mandibles, scapulae, and pelvis) of Rhipidomys mastacalis (Lund 1840) in different vegetation classes of the Atlantic Forest biome in Northeastern Brazil.

2 Materials and methods

R. mastacalis (Rodentia, Sigmodontinae, Thomasomyini) is a small neotropical mammal (head-body length: 125–155 mm; Tribe 2015) with a wide geographic distribution in Brazil, ranging from the east of the state of Paraíba to the north of Paraná, along the entire coast, and also present in Central Bahia, Central-Eastern Minas Gerais, and the entire state of São Paulo (Tribe 1996). This species is endemic and arboreal, with nocturnal activity, and is commonly found in large fragments of the Atlantic Forest (Tribe 2015). Previous studies have shown that R. mastacalis is a generalist habitat rodent (Estavillo et al. 2013) and can be found in forests as well as cocoa plantations, where it is considered a very common species (Pardini 2004; Silva et al. 2020). However, its abundance is low in better-preserved areas of its geographic distribution, and it is absent in secondary forests (Pardini and Umetsu 2006).

The taxonomic identities of the samples used in this study were verified through direct comparison with the specimens deposited in the Alexandre Rodrigues Ferreira Mammal Collection (CMARF), at the Universidade Estadual de Santa Cruz, Bahia. For the remaining specimens, it was assumed that the identifications were correct based on the available identification labels in each collection.

We reviewed a total of 165 individuals of R. mastacalis with available skulls, mandibles, and postcranial skeletons (scapulae and pelvis), although not in equal quantities. For instance, much of the material consisted only of skulls and mandibles. The biological material is currently deposited in the following Brazilian collections: Coleção de Mamíferos Alexandre Rodrigues Ferreira (CMARF), Universidade Estadual de Santa Cruz (UESC), Bahia State; Universidade Federal de Paraíba (UFPB), Paraíba State; and Museu Nacional (MN), Universidade Federal do Rio de Janeiro (UFRJ), Rio de Janeiro State (Appendix 1).

Firstly, all individuals were classified by age based on the wear of their molar teeth (Costa et al. 2011). After classification, only adult individuals were selected to avoid possible ontogeny effects (Voss et al. 1990). The selected sample consisted of 136 individuals (classes 2 and 3). Individuals in these two age classes were characterized by the presence of the third molar in the upper row and the evident exposure of dentin in the lobes of all molars (Costa et al. 2011). Studies on the genus Rhipidomys that evaluated sexual dimorphism did not find differences between sexes, and their analyses were based on males and females grouped as a single set (e.g., Costa et al. 2011; López-Fuster et al. 2001; Tribe 1996). Therefore, both sexes were grouped for all analyses in this study.

Subsequently, the specimens were grouped by vegetation classes based on land use and vegetation cover in five Brazilian federative units (Figure 1). For this purpose, each record was superimposed on a map of Brazilian land use/cover updated in 2021 (IBGE 2021) using the Quantum GIS program (QGIS, version 3.20), and the following classifications were obtained: (1). Forested vegetation: areas occupied by forests, including dense ombrophilous forest, open ombrophilous forest, seasonal forest, and the mix of ombrophilous forests with other areas undergoing regeneration (IBGE 2021). This classification was assigned to records from the Serra Bonita Natural Private Reserve, located in the Camacan County, Bahia State, and the Parque Natural Municipal Professor João Vasconcelos Sobrinho-Serra Dos Cavalos, Pernambuco State. (2). Mosaic occupancy in forested areas: areas characterized by agriculture, sugarcane monoculture, pasture or forestry associated or not with forest remnants. It also includes areas with natural and anthropic disturbances (IBGE 2021). This classification was assigned to records from the counties of Itapetinga, Maracás, and Una in Bahia State, the Área de Proteção Ambiental Municipal do Inhamum in Maranhão State, Murici in Alagoas State, Parque Estadual Mata de Pau-Ferro and Mata do Engenho, Bujari in Paraíba State, and Mata de Biturry, Vertentes, Caruaru, and Bonito in Pernambuco State. (3). Cocoa plantations: known regionally as “Cabrucas” in the Southern parts of Bahia State, they are small areas of lowland ombrophilous vegetation intended for sustainable cocoa production (Theobroma cacao, Malvaceae). In these areas, the original understory vegetation was eliminated, leaving only the highest layer of the forest (the canopy and emerging trees), to provide shade for this type of crop (Cassano et al. 2009; Silva et al. 2020). All records were from the counties of Ilhéus, Uruçuca, and Una in Bahia State.

Figure 1:

Map of Northeastern Brazil, showing the geographical location of the Rhipidomys mastacalis samples selected for this study.

To detect fluctuating asymmetry values in the morpho-geometric analysis, two types of protocols were utilized (Benítez et al. 2020): object asymmetry was calculated on the occlusal view of the skulls, while correspondence asymmetry was calculated on the lateral views of the mandibles and pelvis, and occlusal views of the scapulae (Figure 2). To access asymmetry, both sides (left and right) were analyzed. All images were captured by the same person using a digital microscope (KKmoon 1600×, equipped with eight LED and an endoscopic camera), aided by a tripod and a millimeter-graded ruler for reference scaling.

Figure 2:

Anatomical structures (left and right), showing the location of morphological landmarks in a specimen of Rhipidomys mastacalis (CMARF–1701) from Brazil. (A) Occlusal view of skulls, (B) mandibles, (C) scapulae, and (D) pelvis.

Two-dimensional homologous morphological landmarks (types I and II; Bookstein 1991) were digitized using software tpsUtil and tpsDig (Rohlf 2010, 2015). The position and nomenclature of each landmark followed criteria outlined by Caccavo et al. (2021) for mandibles, and a combination of criteria used in similar studies (Astúa 2008; García and del Valle Alvarez 2021) for the other anatomical structures. In total, 20 landmarks were placed on the occlusal view of the skulls, 12 on the mandibles, 7 on the scapulae, and 12 on the pelvis, for a total of 1292 duplicate image sets processed: 318 (skulls), 492 (mandibles), 238 (scapulae), and 244 (pelvis).

Subsequently, the digitization and location of the morphological landmarks in each photograph were verified by reviewing the possible presence of atypical landmarks (outliers) and calculating the Type 1 error, using both the set of duplicate images and the total morphological landmarks. The variances in the resulting matrices were analyzed with a two-way Procrustes Analysis (ANOVA-Procrustes). Well-digitized data is indicated by lower mean values of Type 1 error in ANOVA-Procrustes compared to the mean values of mean squares values of all individuals (Souto et al. 2019).

The entire procedure for extracting fluctuating asymmetry values from shape structures was based on the flowchart proposed by Benítez et al. (2020). Fluctuating asymmetry was calculated as the deviation of the configuration of original morphological landmarks from symmetrical consensus (Benítez et al. 2020). To obtain this, ANOVA-Procrustes was performed on the combined duplicate image dataset (Benítez et al. 2020). The results of the “Individual × Side” and “Mean squares” interactions, corrected for variance error, determined the presence or absence of fluctuating asymmetry (Benítez et al. 2020; Klingenberg 2015).

To compare whether there were differences in fluctuating asymmetry among the anatomical structures (skulls, mandibles, scapulae, and pelvis) by the vegetation classes (forested vegetation, occupancy mosaics in forested areas, and cocoa plantations), the fluctuating asymmetry values of shape (“Shape FA Scores”) were obtained from the Procrustes distance results, and a one-way analysis of variance (one-way ANOVA) was applied with post-hoc verification (Tukey’s pairwise).

Finally, to assess the relationship between vegetation cover and fluctuating asymmetry in each structure, a linear Gaussian distribution model was employed. Fluctuating asymmetry values were used as the response variable (dependent variable), while vegetation cover and bone structures served as predictors (independent variables). To accomplish this, a circular buffer with a radius of 2.5 km was calculated around the collection sites, taking into account the maximum habitat range of R. mastacalis at 0.0022 km² (Machado 2019). Within each buffer, the percentage of vegetation cover was calculated using the land use map and the cover raster from MapBioma 6.0 (Souza et al. 2020). This analysis was conducted using the Quantum GIS software (QGIS, version 3.20).

All statistical analyses were tested with an alpha level less than or equal to 0.05 (P ≤ 0.05), and performed with MorphoJ, PAST and R programs (Hammer 2020; Klingenberg 2011; R Core Team 2020).

3 Results

The four anatomical structures (skulls, mandibles, scapulae and pelvis) analyzed in the present study had significant levels of fluctuating asymmetry, evidenced by the results of Fisher’s statistical test and its significance value from the ANOVA-Procrustes for the “Individual × Side” interaction as well as by the values of Type 1 error, lower than the mean squares values in all analyzed structures (Tables 1 and 2).

When comparing the values of fluctuating asymmetry for each anatomical structure among the vegetation classes, we found that the anatomical structures exhibited similar variances in their asymmetry levels, as indicated by the one-way ANOVA test (Figure 3). For skulls, the inter-group differences were as follows: SS: 0.0000141123; d.f: 2; MS: 0.00000705613; F: 0.506; P-value: 0.6048. For mandibles, the differences were: SS: 0.000162143; d.f: 2; MS: 0.0000810716; F: 0.2687; P-value: 0.7652. For scapulae, the differences were: SS: 0.000663512; d.f: 2; MS: 0.000331756; F: 1.125; P-value: 0.3318. Finally, for pelvis, the differences between groups were: SS: 0.000380874; d.f: 2; MS: 0.000190437; F: 1.523; P-value: 0.2267.

Figure 3:

Box plot comparing the fluctuating asymmetry (FA) in Rhipidomys mastacalis from three vegetation classes in Northeastern Brazil. (A) Skulls, (B) mandibles, (C) scapulae, and (D) pelvis. The horizontal lines outside the boxes represent the smallest and largest variance for each population, and the horizontal line inside each box represents the mean value.

The results of the linear model between vegetation cover and the anatomical structures showed that there was no direct relationship between the fluctuating asymmetry values of each structure and the variable tested, vegetation cover (Figure 4): R² adjusted = 0.47; Vegetation cover-Skulls (regression coefficient (β): <0.0001; P-value: 0.926); Vegetation cover-Scapulae (regression coefficient (β): <0.0001; P-value: 0.402); Vegetation cover-Mandibles (regression coefficient (β): <0.0001; P-value: 0.949); Vegetation cover-Pelvis (regression coefficient (β): <0.0001; P-value: 0.731).

Figure 4:

Scatter diagrams based on a simple linear model (fluctuating asymmetry vs. vegetation cover), for four anatomical structures of Rhipidomys mastacalis three vegetation classes in Northeastern Brazil. The trend line is shown in black; the gray area represents the 95 % of confidence intervals.

4 Discussion

The analyses showed statistically significant differences in fluctuating asymmetry among different anatomical structures in all types of vegetation analyzed. Forested vegetation had significantly lower values of fluctuating asymmetry in skulls compared to mandibles, scapulae, and pelvis. Similar trends were observed in the occupancy mosaics in forested areas and in cocoa plantations, with skulls also obtaining the lowest values. For the remaining anatomical structures, the correlations between shape and levels of fluctuating asymmetry were low or medium and not significant.

While the data did not show a correlation between fluctuating asymmetry values and vegetation cover, previous studies on other rodent species have indicated associations between fluctuating asymmetry, cytogenetics, and environmental factors. In a study that evaluated the effect of fluctuating asymmetry on the mandible shape of a rodent, Nectomys squamipes (Cricetidae, Sigmodontinae) and correlated this information with cytogenetics, no association was found between levels of fluctuating asymmetry, endogamy, and low heterozygosity. However, a negative association was observed with the surface area of forested areas (Caccavo et al. 2021).

On the other hand, in a study conducted in Belgium that examined the association between DNA microsatellite fragments and fluctuating asymmetry values in the hind limbs of a habitat-specialist tree rodent (Sciurus vulgaris), no correlation was found between genetics and fluctuating asymmetry (Wauters et al. 1996). This suggests that the asymmetric effect in the evaluated anatomical structures of R. mastacalis may be influenced by other environmental or chemical stressors. An example of this latter hypothesis is the study of Costa et al. (2023), who determined the concentrations of heavy metals in small non-volant mammals (including R. mastacalis) in two different types of habitats: traditional cocoa agroforests and Atlantic Forest fragments, located in the southern region of Bahia State, specifically in the municipalities of Ilhéus and Una, Brazil. The study documented the presence of lead, nickel, manganese, and copper in both habitats. Although the results of Costa et al. (2023) did not show a negative association between body mass, assessed habitats, and heavy metal values, the high levels of heavy metals observed in these areas (hat include specimens analyzed in the present study), indicate that the anatomical structures of R. mastacalis could be more affected by the bioaccumulation of agrochemicals than by the loss of vegetation cover.

Most of the published information regarding the ecological responses of R. mastacalis to disturbances is limited to data related to the presence or absence (as well as relative abundances) of populations in natural, fragmented, and transformed areas, using only capture-mark-recapture techniques (Pardini 2004; Pardini and Umetsu 2006). However, these studies do not take into account the results of other approaches, such as genetic markers, niche modeling, and geometric morphological analyses, which have been implemented in small mammals to document the effects of environmental stress in landscapes with varying levels of disturbance (Caccavo et al. 2021; Kubiak et al. 2017; Maestri et al. 2015).

Studies of non-volant small mammals that have evaluated different morphological structures such as the locomotion limbs and mandibles, indicate the need for further investigation of this topic by adding other morphological and environmental variables and evaluating the presence of agrochemicals (Maestri et al. 2015). Time scale is also a variable that must be considered (Pergams and Lawler 2009). For instance, possible changes in a population during different seasons or years should be evaluated, and from that, correlations drawn to determine whether there is a proclivity to increase the intensity of fluctuating asymmetry in several generations. This could be due to the transformations occurring in the same habitats over the years.

Research on the population dynamics of R. mastacalis in Northeastern Brazil indicates that it is a common and abundant small mammal species in the Atlantic Forest community (Gonçalves et al. 2018; Pardini 2004; Pardini and Umetsu 2006; Silva et al. 2020). However, in certain areas, it has been identified as a pest for locally important crops (Cassano et al. 2021), and it has also been found to be a vector of yellow fever, having been collected in houses (Tribe 2015). Previous studies have suggested that R. mastacalis is an arboreal specialist that relies on complex habitats with dense understory, but is also tolerant of environments with different types of regeneration (Calazans and Bocchiglieri 2020; Grelle 2003). As such, the species appears to be a generalist habitat species with a high degree of tolerance to fragmented environments and wide plasticity in the use of vertical space, including understory, canopy, and soil (Calazans and Bocchiglieri 2020; Grelle 2003; Pardini 2004). The present study provides new insights by documenting, for the first time, the presence of asymmetric structures, as indicated by the indicator of fluctuating asymmetry in morpho-functional structures essential for the species’ survival.

5 Conclusions

Overall, this study contributes new insights by documenting the presence of asymmetric structures, indicated by fluctuating asymmetry, in essential morpho-functional structures for the survival of R. mastacalis. Further investigations are needed, incorporating additional variables and considering the timescale, to better understand the effects of morphological, environmental, and genetic factors on fluctuating asymmetry in small mammal populations.