Two forest reserves, located in the north of Selangor State Park, namely Gading Forest Reserve (FR) and Bukit Kutu FR and one forest reserve outside the park, were selected. Gading FR is one of the largest forests in northern Selangor with a total area of 18,828 ha (Fig. 1). Bukit Kutu FR covered a total area of 11,700 ha and was considered a wildlife sanctuary. Both Gading FR and Bukit Kutu FR are located approximately 20 km apart. Bukit Tarek FR Extension (E) is the third study site. Bukit Tarek FR (E) is an additional forest land to the existing Bukit Tarek FR. It is located south to Gading FR and west to Bukit Kutu FR with a total area of 3,560 ha of forested land. Bukit Tarek FR (E) is located about 10 km from Gading FR and 15 km from Bukit Kutu FR. Bukit Tarek FR (E) is located outside Selangor State Park and is significantly fragmented as the forest is surrounded by agricultural activities, including rubber and oil palm plantations, while Gading FR and Bukit Kutu FR are connected through adjacent forest reserves via Semangkok FR and Hulu Selangor FR. Fig. 1 and Table 1 show the location and land cover of each study site.

Figure 1.  

Location of Selangor state Park and present study sites. The black dots represent the location of present study sites.

A standardised sampling protocol for each site was carried out using the same type, number and arrangement of traps. Sampling sessions were also held in the same area for the same number of days. A total of five-line transects of 200 m, at an altitude of less than 300 m above sea level, were established in each study site. Each line transect was located at least 50 m apart. The sampling sessions were undertaken from 2017 to 2019. Each sampling session comprised of five days of trapping and at least one sampling session was held each month, consisting of a total of seven sampling sessions for each study site.

The trapping activity of small mammals was carried out using several types of traps, including a collapsible trap for non-volant mammals, while harp traps and mist nets were used to capture volant mammals. A total of 100 collapsible cage traps with dimensions of 42 cm × 16 cm × 16 cm were deployed in each line transect and 20 traps were placed along the transect with an interval of 10 m between the two traps. All the traps were baited with oil palm fruit. During the sampling period, the traps were left open all day and checked early in the morning. Any mammal caught was carefully removed and temporarily placed inside a cloth bag before being examined. Meanwhile, three four banks harp traps and 10 mist nets (9 m × 4 m) were placed in suitable and potential fly paths to trap insectivorous and fruit bats, respectively. Harp traps were checked every hour from 7.30 p.m. to 10.30 p.m. while the mist nets were checked every hour from 6.30 a.m. to 11.00 a.m. and then from 7.30 p.m. to 10.30 p.m.

All the small mammals caught were carefully removed from the traps and nets and placed temporarily in a cloth bag. All captured individuals were thoroughly examined and measured as the species identification of volant and non-volant small mammals is largely dependent on specific measurements. The morphologies of each individual, such as body length, hindfoot, ear size, forearm and tail length, were measured to the nearest millimetre and the weight was recorded in grammes. In addition, the sex of each individual (male or female) and reproductive conditions (pregnant or lactating) were noted. Non-volant and volant small mammals were identified using references such as Medway 1983, Khan 1992, Payne et al. 1998, Kingston et al. 2006, Francis and Barrett 2008 and Jayaraj et al. 2012. After all the individuals were carefully examined, most of the species were tagged and released back to their habitat. Only three representative individuals of each species found at each study site were collected and stored in 70% ethanol as voucher specimens. Liver tissues were excised and stored in a vial containing absolute ethanol for further DNA analysis. All specimens were deposited in the Zoological Collection of Forest Research Institute Malaysia (FRIM).

The species diversity indices, Shannon-Wiener Index (H’), Evenness (SI) and Dominance (D) were calculated using the Paleontological Statistics (PAST) software (Hammer et al. 2001). The Shannon Diversity Index was used as a relatively reliable method of measuring species diversity (Gotelli and Colwell 2001, Colwell 2009). Species accumulation curves were calculated using iNEXT Online (Hsieh et al. 2016) to infer the completeness of the inventory for small mammal assemblages in the three study sites. The non-parametric t-test was assessed for differences in the abundance of volant and non-volant small mammals using the PAST programme. The Bray-Curtis Similarity Index was calculated to show the similarity in the composition of small mammals between forest structures inside and outside of the park.

We selected classic taxonomic measures developed by Clarke and Warwick (1999), taxonomic diversity and taxonomic distinctness, which measured merged information relative to taxonomy, number of species and evenness of the sample and these remained insensitive to the sampling efforts (Clarke and Warwick 1999). Given the multivariate nature of the information included in this index (Somerfield et al. 1997), it is expected to be more sensitive to anthropogenic disturbances (Warwick and Clarke 1998, Leonard et al. 2006, Marcantonio et al. 2013), such as forest fragmentation.

All the published data on small mammals conducted along Selangor State Park and other national/state parks were gathered to produce a complete checklist. All data were carefully examined and species nomenclature was updated with current naming. The Sorensen-Dice Analysis was calculated to show how similar Selangor State Park is with other national/state parks in Peninsular Malaysia, based on presence data of small mammals.

Through this study, we have managed to document a total of 34 species of volant small mammals mainly consisting of bats and 16 species of non-volant small mammals which comprised of rodents. Of this, 24 species of insectivore bats were recorded in this study and these species were mainly recorded in Gading FR (21 species) and Bukit Kutu FR (20 species). This can mainly be the result of the behaviour of insectivorous bats that forage understorey in lowland forest and some of these bats are known to forage in groups (Payne et al. 1998; Tingga et al. 2012). Our results suggested that the family Vespertilionidae is the most diverse, while the family Muridae is the most abundant in three study areas. Only two species of Vespertilionidae were caught in Bukit Tarek FR (E), namely, Kerivoula papillosa and Murina suilla (Table 2). The main factor that contributes to fewer Vespertilionidae species in Bukit Tarek (E) FR lies within the habitat itself. Bukit Tarek (E) FR is considered a fragmented forest and is surrounded by rubber and oil palm plantations (refer to Table 1). Palaeotropical bat assemblages are dominated by members of the families Rhinolophidae and Hipposideridae and the Vespertilionidae subfamilies, Kerivoulinae and Murininae and many of them are typically highly adapted for foraging in the clutter of the forest interior (‘narrow-space’) (Schnitzler and Kalko 2001; Tingga et al. 2012). Consequently, these species may be more sensitive to forest loss and exhibit greater avoidance of disturbed and open habitats (Kingston et al. 2003). Due to this dependence on forest, these species were probably adversely affected by deforestation and other forest disturbance events (Lane et al. 2006; Struebig et al. 2008). A significant low abundance of volant small mammals in Bukit Tarek (E) FR deceptively exhibits the diversity of these mammals as other contributing factors, such as limited food resources and high predation, might influence the presence of bats generally (Hodgkison et al. 2004, Popa-Lisseanu and Voigt 2009).

On the other hand, the family Muridae has shown that it is relatively abundant in all three study sites, despite the lack of significant differences in abundance amongst the study areas. Small mammals, particularly non-volant small mammals, have a distinct habitat specialisation and can be classified as forest and open land specialists and habitat generalists, each responding differently to changes in landscape complexity (Gentili et al. 2014). However, a relatively similar diversity of non-volant small mammals was achieved across all study areas despite some changes in the microhabitat structure. A similar finding has also been reported by Nakagawa et al. (2006), where the results showed no variations in the abundance of small mammals present in five different habitat types comprising of primary forest and degraded areas in Sarawak, Malaysia.

Species characteristic may be one of the reasons that derive a similar structure of non-volant small mammals from these study areas. The long-tailed giant rat (Leopoldamys sabanus) was the most abundant species captured in all three study sites and is present throughout the Sunda region of Southeast Asia. It is a common, generalistic species in local assemblages of small mammals (Lim 1970). The abundance of L. sabanus is higher in Bukit Tarek FR (E) relative to the least-impaired habitat in Bukit Kutu FR. Analysis of habitat utilisation by L. sabanus, based on spool-and-line-tracks of Wells et al. (2008), revealed L. sabanus has a strong ability to move between various forest matrices, such as logged and unlogged forests. A higher abundance of L. sabanus in Bukit Tarek (E) proves that the species can forage for its resources in different habitats, including oil palm and rubber plantations and may encourage a healthy reproduction.

Fig. 2 shows the rarefaction curves of species diversity in the three study areas. Although Gading FR had more captures (199 individuals) than all the sites surveyed in this study, Bukit Kutu FR showed the highest species diversity as it had the steepest curve between the other sites. The total sampling effort was still inadequate to document the total diversity of small mammals in Gading FR and Bukit Tarek FR (E) as the cumulative curve was still exponential, but as for Bukit Kutu FR, the curve is approaching the asymptote, indicating the species diversity in the forest reserve may have been successfully identified. The highest species richness recorded in Gading FR is apparent, as Gading FR encompasses a significantly larger area than Bukit Kutu FR and Bukit Tarek FR (E). Areas increase diversity as a larger plot is likely to have more habitats and functions in order to support a greater diversity of species. Besides, many species require a wide range for adequate prey or seed forage. Conversely, the diversity indices have shown that Bukit Kutu FR has the relatively highest diversity values similar to the accumulation curve in Fig. 2. Bukit Kutu FR has the highest Shannon and Evenness values as rarer species have been detected and the abundance of species has been evenly distributed compared to the other two study sites. Although Gading FR was considered to be the richest in the total number of species, there was a decreased discrepancy between the species richness values estimated by the rarefaction and the Chao 1 estimators. These results indicate that Bukit Kutu FR is potentially the richest study site, followed by Gading FR. According to Rabosky (2009), different ecological limits for clades seem to be the main determinants of diversification and, therefore, species richness. Although Bukit Tarek FR (E) recorded the lowest values for both abundance and diversity compared to the other two sites, the differences between all three study areas were found not to be significant (not shown). Such a result in Bukit Tarek FR (E) may be affected by the habitat itself, where the relative abundance, referred to in Pardini et al. (2005) and Cabrini et al. (2013), may be more negatively sensitive to forest fragmentation and isolation than to species richness.

The low species taxonomic distinctness (Δ*) for Bukit Tarek FR (E) also highlights possible forest fragmentation effects in the area; the taxonomic distinctness is quite sensitive for the discriminating species of the affected areas (Salas et al. 2006, da Silva and Batalha 2006). Ricklefs (2001) provided evidence that diversity differences between clades reflect the region’s capacity to support species of a diverse balance of clade diversity. Furthermore, Clarke and Warwick (2001) suggested an equability influence on the Δ index. Bukit Tarek FR (E) was estimated to have an increased number of dominant taxa compared to the other two study areas. Based on the species identified in Bukit Tarek FR (E), the species diversity is justified by the predominance of two major families of Muridae and Pteropdidae, the abundance of which is mainly contributed only by several species (see Table 2).

Based on the similarity dendrogram, it is possible to hypothesise that species composition in Selangor State Park is most similar to those in Endau-Rompin State Park which is followed by Royal Belum State Park. Endau-Rompin is known to support a high diversity of mammals as reported by Aihara et al. 2016 and the small mammals caught from Selangor State Park have a high similarity of almost 80% to the composition of small mammals in Endau-Rompin State Park. The similarity in species richness amongst Selangor State Park, Endau-Rompin State Park and Royal Belum State Park can also be explained due to the proximity of these sites to each other in the west coast of Peninsular Malaysia.

Comparisons were made with the current study representing the northern region of Selangor State Park with Ulu Gombak FR (Sing et al. 2013, William-Dee et al. 2019) as the central region and Hulu Langat FR (refer Lim et al. 2009) at the southern part of the Park (Table 2). Unlike the central and southern regions of the Park, this study is the first attempt to document the small mammal diversity, while the latter studies have been well-studied for several decades. By comparison, both the central and southern regions of the Park have more species than the northern regions; however, the present study shows that there is a wide variety of small mammals inhabiting these habitats. Previous studies have provided comprehensive documentation of mammal diversity; the current study lacks the inclusion of other classes of mammal species, such as medium-sized mammals and large mammals.

The present study is far from reflecting the actual diversity found in the areas studied. Conservation priority should be given, however, as these forest reserves are important habitats for fauna as a whole. In addition, areas that are impaired and fragmented, such as Bukit Tarek FR (E), should also be treated as equally important habitat for fauna. Although Bukit Tarek FR (E) is considerably fragmented and surrounded by plantations, the habitat still provides refuges for moderate volant and non-volant small mammal diversity, which indicates that habitat patches are relevant to include as part of the conservation priority area.