Microhabitat use by Rhipidomys mastacalis and Marmosops incanus (Mammalia) in a restinga areas in north-eastern Brazil

Introduction

The occurrence of some species of small mammals in a certain area can be facilitated by habitat characteristics (Prevedello et al. 2008). Because they have relatively smaller range areas, this group demonstrates the ability to discriminate subtle variations in the environment (Fisher et al. 2011), showing different forms of habitat use among species (Prevedello et al. 2008). This differential use can change the species abundance throughout space, influencing the structuring of communities (Gentile & Fernandes 1999; Cerqueira 2000; Arnan et al. 2014). This use can also reduce interspecific encounters, reducing the effect of competition and facilitating the partitioning of resources by co-occurring species (Schoener 1974; Price 1978; Cunha & Vieira 2004; Dalmagro & Vieira 2005). Patterns in habitat use are dependent on the scale of study (Morris 1987; Fisher et al. 2011; Novillo et al. 2017), with several studies highlighting the importance of microscale habitat variations for small mammal species (e.g. Püttker et al. 2008; Naxara et al. 2009; Rocha et al. 2011; Abreu & Oliveira 2014; Arnan et al. 2014; Leimgruber et al. 2014; Delciellos et al. 2015). Morris (1987) was one of the first to define the microhabitat as being composed of environmental variables that affect the behaviour of an individual within its home range. Thus, variations in microhabitat use, including the use of vertical strata, may reflect intrinsic differences in required resources and foraging habits of each species, as well as the influence of other species on the environment (Price 1978; Leimgruber et al. 2014; Passamani & Rosa 2015).

Microhabitat variations, such as the litter characteristics in the Atlantic Forest (Naxara et al. 2009) and the herbaceous stratum in the Cerrado (Rocha et al. 2011) have already been positively related to local species abundance. In restinga environments, coastal ecosystems associated with the Atlantic Forest, whose vegetation is influenced by freshwater or brackish water systems (Cerqueira 2000; Souza et al. 2008), and it was observed that these characteristics might also influence the local abundance of marsupials (Freitas et al. 1997). However, there are few studies that associate habitat characteristics with the abundance of small mammals in these environments (e.g. Freitas et al. 1997; Cerqueira 2000).

Horizontal variations in habitat have already been positively related to the abundance of small mammals such as the amount of litter (Novillo et al. 2017), the density of bromeliads (Abreu & Oliveira 2014) and the herbaceous stratum (Rocha et al. 2011). Gentile and Fernandes (1999) related the increase in the abundance of species of semi-arboreal marsupials to the presence of large trees in the Atlantic Forest. In this same biome, Delciellos et al. (2015) observed that cursorial species may be more abundant in environments with lower vertical stratification, relating the life habit of the species to the exploitation of the habitat and its associated resources.

The quality of the habitat, considered by Hall et al. (1997) as the ability of the environment to provide conditions for individual and population persistence may be an important factor in the composition of small mammal species. The marsupials Marmosops incanus (Lund, 1840) and Marmosa murina Linnaeus (1758) and the rodents Cerradomys vivoi (Percequillo, Hingst-Zaher & Bonvicino, 2008) and Rhipidomys mastacalis (Lund, 1840), for example, have already had their occurrences related to environments with very diverse structuring, from mature forests of Atlantic Forest to early stage of regeneration forests, edge environments and open areas (e.g. Pardini 2004; Püttker et al. 2008; Bezerra & Geise 2015; Delciellos et al. 2015).

The present study analyses the use of the microhabitat by the marsupial M. incanus and the rodent R. mastacalis in areas of arboreal restinga in north-eastern Brazil. This study investigated whether there are interspecific differences in the frequency of use of the vertical strata and if microhabitat characteristics influence the abundance of these species.

Methods

Study area

The study was carried in the Santa Isabel Biological Reserve (REBIO Santa Isabel), in the municipality of Pirambu, state of Sergipe, north-eastern Brazil. REBIO Santa Isabel covers a coastal stretch of approximately 40 km (Brasil 1988) and is among the priority areas for biodiversity conservation in the country (Santos et al. 2017; MMA 2018). During the study period, accumulated annual precipitation in the region was 1254.5 mm and the temperature varied from 28 to 32°C (SINDA 2018). The climate is considered rainy tropical with a dry summer, according to Köppen–Geiger (Alvares et al. 2014).

Three sampling sites (Fig. 1) were selected approximately 600 m from the tide line and separated by approximately 1.5 km (10°42′S, 36°48′W; 10°41′S, 36°47′W; and 10°40′S, 36°47′W). These sites are located in the non-flooded, ‘fruticeto fechado’ phytophysiognomy, which includes tree species to the dunes leeward that form a sparse canopy between 3 and 6 m high (Silva & Britez 2005; Oliveira & Landim 2014).

Results

A total of 109 captures of M. incanus (N = 76) and 29 captures of R. mastacalis (N = 25) were recorded, being only a juvenile of M. incanus captured in pitfall. These species used the vertical strata at different frequencies. Marmosops incanus used the understory (71.3%) more than the ground (χ² = 18.141; P < 0.0001) and all captures of R. mastacalis occurred in the understory.

There was an increase in litter (z = 6.556, P < 0.001) and in the herbaceous stratum (z = −4.558, P < 0.001) in the rainy season, indicating seasonal changes in the microhabitat. The RDA resulted in a biplot (Fig. 2) with two canonical axes: RDA1 explaining 13.4% and RDA2 explaining 1.2% of a total of 14.6% variation (R² = 0.146). The permutation test of the general model indicated that changes in the microhabitat around the traps reflect changes in species abundances (F = 2.943; P = 0.005), with 9.7% of the variance of the species response matrix explained by microhabitat (R² = 0.097). However, the permutation tests for the two generated canonical axes indicated that only RDA1 (F = 16.887; P = 0.011) should be considered.

No collinearity of the data was identified (FVD < 10; Table 1), without the withdrawal of variables from the analysis. The ordistep function indicated significance only for the understory (F = 11.646; P = 0.005), which represents most of the variation of the canonical axis RDA1 (Table 1). Thus, sites with the densest understory have a greater abundance of M. incanus and a smaller abundance of R. mastacalis, indicating that the local variations in the abundances of these species are influenced by this characteristic of the microhabitat.

Discussion

In this study, R. mastacalis and M. incanus used the microhabitat differently in the restinga areas sampled in north-eastern Brazil. Interspecific differences in the frequency of use of vertical strata and microhabitat variations influenced in the abundance of these species.

Regarding the use of vertical strata, R. mastacalis proved to be arboreal for not using the ground in REBIO, a condition also reported by Paglia et al. (2012). However, according to Stallings (1989) and Pardini (2004), this species can exploit the ground in areas of the Atlantic Forest, even though most of the records occur in the upper strata of the vegetation. Other studies in the Atlantic Forest registered the use of the vegetation canopy by R. mastacalis (Grelle 2003; Naxara et al. 2009) and the use of the environment by this species changes in isolated forests (highland marshes) of the Caatinga, where it uses the ground more than the understory (Sousa et al. 2004).

Marmosops incanus, however, seems to keep the scansorial habit as standard, but its frequency of use of the strata can also vary between areas. In REBIO, this species uses more of the understory than the ground, as also reported by Cunha and Vieira (2002), differing from most studies that indicate the opposite in Atlantic Forest forests (Grelle 2003; Pardini 2004; Loretto & Vieira 2008; Prevedello et al. 2008; Passamani & Rosa 2015). Other studies, however, did not observe differences in the use of vertical stratum by this species (Fonseca & Kierulff 1989; Stallings 1989).

In this context, variations in the use of vertical strata by M. incanus and R. mastacalis in different localities suggest a plasticity of these species in relation to vertical space exploration. This use may be influenced by factors such as vegetation type and local interspecific relationships (Price 1978; Jorgensen 2004), given the variation in species of predators and competitors in environments. Differences in space use may also reflect the intrinsic characteristics of the species (Price 1978; Leimgruber et al. 2014; Passamani & Rosa 2015). Variations in the locomotor ability of M. incanus and R. mastacalis, for example, may potentially influence habitat exploration and contributed for resource sharing, as indicated for other species in Vieira (1997), Cunha and Vieira (2002), Delciellos et al. (2006), and Delciellos and Vieira (2006). In addition, food habit is also an intrinsic characteristic that interferes with and directs habitat exploration, since each type of food can have a particular distribution in the environment (Cáceres et al. 2012).

In general, R. mastacalis is considered to be frugivore and predator of seeds and M. incanus is considered an insectivore-omnivore (Paglia et al. 2012). However, although plant material makes up a large part of the diet of R. mastacalis, this rodent also has the ability to exploit arthropods as a food resource (Sousa et al. 2004; Pinotti et al. 2011), while M. incanus may also have fruits as important items in the diet (Bezerra & Geise 2015). In the case of REBIO Santa Isabel, it was observed that these two species consume similar food items in different proportions (Malaquias, E.C. & Bocchiglieri, A., 2018, unpubl. data), and there is competition for food in the locality.

Thus, differences in diet composition may not be sufficient to allow M. incanus and R. mastacalis to coexist in the arboreal stratum of the sampled area. In a review on resource partitioning, Schoener (1974) pointed out that differentiation in habitat use might be more important for coexistence between species than diet. For small mammals, several works also consider habitat partitioning as an important mechanism of coexistence (e.g. Price 1978; Cunha & Vieira 2004; Delciellos et al. 2006; Naxara et al. 2009; Abreu & Oliveira 2014; Leimgruber et al. 2014; Passamani & Rosa 2015; Novillo et al. 2017).

In fact, a greater abundance of M. incanus was observed in places with dense understory at REBIO Santa Isabel. A structured understory can increase vegetation connectivity for better space exploration by species that use this stratum (Camargo et al. 2018). Rhipidomys mastacalis, however, was less abundant with the densification of this understory, either by the structure of the supports for locomotion, as approached by Cunha and Vieira (2002), for example, or by the abundance of the dominant species M. incanus, a potential competitor in these places. Studies have indicated that understory density is also important for M. incanus in more complex areas of the Atlantic Forest, positively influencing its abundance (Püttker et al. 2008; Leiner et al. 2010).

Therefore, the pattern of microhabitat use observed in the study area through the opposite effect of understory thickening on species abundance and differences in the frequency of use of vertical strata may be one way for these species to segregate the microhabitat as a mechanism of coexistence. In addition, there is a possibility that R. mastacalis may use the canopy more frequently in sites with higher density of understory and abundance of M. incanus, since habitat exploration may change with the presence of other species, according to Price (1978) and Delciellos et al. (2006), and M. incanus rarely use this stratum of vegetation (Cunha & Vieira 2002; Grelle 2003; Loretto & Vieira 2008; Prevedello et al. 2008; Passamani & Rosa 2015). This condition would ease competition in the vegetation canopy, and a lower abundance of R. mastacalis in the study area may be a reflection of the absence of sampling in this stratum.

Canopy cover and density of understory and vines are related to the vertical structuring of habitat, important components for the locomotion of species that use the upper strata of vegetation (Cusack et al. 2015). Thus, the understory selection, among other microhabitat variables, points to the importance of vertical vegetation stratification to the most abundant species in the REBIO Santa Isabel community. Both M. incanus and R. mastacalis were more frequent in the understory then ground, revealing that the microhabitat complexity, represented mainly by different degrees of density of this understory, may exert more influence on the abundance of these species than the measured variables related to heterogeneity (number of palms, depth of litter and density of the herbaceous stratum).

In addition to the microhabitat variables considered here, there may be other uninvestigated factors that interfere with the exploration of microscale space by species, whether ecological as the provision of specific food resources or random, as pointed out by Hubbell's (2001). The fact that the understory was the only significant microhabitat variable does not mean that the other variables considered here are not relevant for these species on the meso- and macrohabitat scale.

Detailed ecological data, such as patterns of habitat use, are rare for most species of small mammals, especially in restinga areas. In this sense, the present study makes an important contribution to the understanding of the use of the microhabitat by M. incanus and R. mastacalis, providing new information for this ecosystem still little explored. The differences in the patterns found in other environments suggest a plasticity in relation to the use of vertical strata. In addition, there appears to be an influence of the understory structure on the local abundance of the species within the sampled area. Therefore, the partition of resources between these small co-occurring mammals can be facilitated by the differential use of the microhabitat in restinga environments.