Species Diversity and Distribution of Non-volant Small Mammal between Restoration, Boundary, Disturbed and Undisturbed Area in Cameron Highlands, Malaysia

Deforestation in Cameron Highlands, Malaysia has increased significantly in the past few years to accommodate the growing population of Cameron Highlands. This led to a rapid urbanisation in Cameron Highlands which increased anthropogenic activities, causing degradation of the natural environment. Such environmental changes highlight the necessity of wildlife and resource inventories of available forested areas to improve existing conservation and management plans, especially for threatened taxa such as the non-volant small mammals. However, very few studies are known to focus on the effect of deforestation on non-volant small mammals, especially in the adjacent forest. This survey aimed to document non-volant small mammals from four habitat types (restoration areas, boundary, disturbed and undisturbed areas) of Terla A and Bertam, and undisturbed forest of Bukit Bujang Forest Reserve, Cameron Highlands, Malaysia. Samplings were conducted in two phases between August 2020 to January 2021. A total of 80 live traps were deployed along the transect lines in all three study sites, and 10 camera traps were set randomly in each forested area. Results demonstrated that species diversity (H′) is higher at Terla A Forest Reserve compared to Bertam and Bukit Bujang Forest Reserve. In contrast, species diversity in the boundary area (S = 8, H′ = 2.025) and disturbed forest area (S = 8, H′ = 1.992) had similar number of species (S) compared to others study habitat; restoration area had the lowest species diversity (S = 3, H′ = 0.950). Berylmys bowersi was the most captured species from trappings and Lariscus insignis was the most frequently recorded species from camera trappings for all study sites. The results of the survey provided new information on non-volant small mammals in Cameron Highlands for future research, conservation, and management.


INTRODUCTION
Malaysia is one of the biodiversity hotspots in the tropical region of Southeast Asia, with a high faunal diversity including small mammals. Small mammals are described as any mammals that weigh less than 5 kg (Lim & Pacheco 2016) and are highly diverse with a range of tolerance to habitat disturbance (Rickart et al. 1991). The diversity of mammals within the Malaysian territory is quite significant with at least 440 species of mammals recorded (Department of Wildlife and National Parks [DWNP] 2016), of which 15% (66 species) are endemic to Malaysia (Dee et al. 2019).
According to Chan (2019), Cameron Highlands hosts more than 700 plant species, where 145 of them are endemic including 32 orchid species. A total of 153 56 mammals, 199 birds, 58 reptile and 14 amphibian species have also been recorded here. From this list, the pitcher plant, Serow and Mountain Peacock-Pheasant are listed in the 2021 IUCN Red List of Threatened Species.
For years, the degradation of nature due to anthropological development of highland areas in Malaysia has been alarming. The foggy climate, cool temperatures and beautiful scenery from the top of the hills have inspired tourism related developments in the highlands. The construction of roads and resorts has changed the habitat for species that inhabit the mountainous area. According to Weebers and Idris (2016), Cameron Highlands was developed for sanatorium and amusement purpose, and also as an agricultural site with tea and vegetables farm. Indiscriminate land clearing for farming and agriculture has created untold damage and pollution to its once pristine environment. Unsustainable agriculture has severe impacts on the environment, wildlife and tourism of Cameron Highlands (Barrow et al. 2005;Razali et al. 2018). The different types of plant present at different habitats may help in identifying the species of non-volant small mammals (Haziq et al. 2021).
Cameron Highlands is rich in diversity of faunas including insects (54 and 59 species from the Order Coleoptera and Lepidoptera, respectively), 13 species of bats and five species of non-volant small mammals (Shahfiz et al. 2011;Abdullah et al. 2011;Nur Amira et al. 2017). Unfortunately, according to Abdullah et al. (2011) and Palanivel (as cited in The Star Online 2013), due to high demand for the tourism industry, Cameron Highlands has been threatened with degradation of nature over the years. According to Siti Salwa (2020), illegal land-clearing and farming activities have been blamed for the landslides, mudslides and floods that frequently occur in Cameron Highlands. The environmental changes highlight the necessity of wildlife and resource inventories at available forested areas to help in existing conservation and management plans, especially for threatened taxa such as the non-volant small mammals. Consequently, these mammals may be more sensitive to forest loss resulting in the mammals avoiding disturbed and open habitats (Kingston et al. 2003). Due to the dependence on forest, these species were probably adversely affected by deforestation and other forest disturbances (Lane et al. 2006;Struebig et al. 2008). However, there are very few studies that have focused on the effect of deforestation on non-volant small mammals especially in the adjacent forest.
The main purpose of this study is to determine species diversity and distribution of non-volant small mammals in different habitats representing restoration or rehabilitated, edge or boundary, disturbed forest and undisturbed forest area. Factors that may influence the observed differences can be possibly identified by comparing the diversity and distribution of non-volant small mammals in the respective habitat types. Basic information on species diversity is essential for forest management and local authorities to develop sound management plans. However, a lack of basic knowledge on biodiversity could lead to non-holistic local planning and would have a negative impact on the environment, especially on fauna diversity.

Study Site
The study areas were classified into four habitats, namely restoration, boundary, disturbed forest and undisturbed forest area. According to Butler and Lawrence (2019), the categorisation of forest sites was done based on the frequency of human activities in the surrounding areas, distance to human settlement or activities, types of forest, and the status of the forest (protected or non-protected). The presence of various types of habitats plays a major role in the richness of Malaysia's diversity (Dee et al. 2019).
This study was conducted in Terla A Forest Reserve (04°35'36.6" N, 101°22'54.7" E) with an elevation of 1300 m-1500 m. Genus Casuarina (rhu bukit) dominated the restoration area. The average diameter and plant height in the restoration area were 10 mm and 0.90 m, respectively. Flowering plants such as genus Physalis (buah letup), wild orchids, Lagenaria siceraria (bottle guard) and genus Nepenthes (pitcher plant) were seen growing wildly near the hill boundaries. The forest is overgrown with Dipterocarp trees, grasses and herbaceous plants. Minimal human disturbance was observed with the presence of a piping line that provide water source to the agricultural farm nearby. The distance from the restoration area and forest area were approximately 5 m apart.
Bertam Forest Reserve (04º25ʹ 15.0ʺ N, 101º26ʹ 41.4ʺ E) has an elevation of 1200 m-1300 m in Cameron Highlands, Pahang (Fig. 1). The restoration trees were dominated by genus Casuarina (rhu bukit). Flowering plants such as Morus (mulberry tree), Mentha (mint plant) and Bambusa (bamboo tree) were seen growing wildly near the hill boundaries. The average diameter and plant height in the restoration area were 10 mm and 0.90 m, respectively. The forest area has a small river as a source of water and is dominated by Eugeissona (bertam) and Musa balbisiana (sweet wild banana).
The study was also conducted at Bukit Bujang Forest Reserve (04º24ʹ 07.06ʺ N, 101º35ʹ37.28ʺ E) which has an elevation of 400 m-500 m to compare the state of the diversity of non-volant small mammals in two different areas namely the Disturbed Forest Area and Undisturbed Forest Area (Fig. 1). Vegetation for Bukit Bujang Reserve Forest has been dominated by trees of 5 m-10 m high. Different type of trees such as genus Eugeissona (bertam), genus Calamus (rotan), genus Licuala (palas) and genus Oncosperma (bayas) dominated this forest. This area has sufficient water source from rivers such as the Lemoi River.
Terla A and Bertam Forest Reserved are restoration areas listed under "Restoration, Rehabilitation, and Reclamation (3RSM)" programme in Cameron Highlands. Cameron Highlands is one of the areas affected by floods in 2014. Therefore, Forestry Department of Peninsular Malaysia initiated a restoration programme in the Eleventh Malaysia Plan (2016-2020) with an estimated area of 1640 ha in works. Almost all the undisturbed forests nearby in Cameron Highlands are located at low altitudes (Forestry Department of Peninsular Malaysia 2020). Restoration of an area is defined as any intentional activity that initiates or accelerates the recovery of an ecosystem from a degraded state (http://www. ipbes.net/). Forest boundary is particularly a transition area between forest and non-forest where a dense forest is gradually opening up to tree-free land, as for example at the timber line or at the boundary of deserts (Kleinn 2007). Forest areas disturbed by human activities in the past may naturally regain features that characterize primary forests while others that don't will require restoration (Lindenmayer et al. 2014).
The survey was based on two phases of sampling between August 2020 to January 2021. The sampling periods included the dry season and the wet season for Cameron Highlands. Sampling and monitoring non-volant small mammals were replicated twice for all three study sites to minimise bias in collecting data.

Sampling Method
In this study, two methods were used, i.e., live trapping and camera trapping that target the more typical terrestrial non-volant small mammals. Advancements in camera-trapping technology have led to the widespread use of this survey method in the study of terrestrial mammals (Wearn & Glover-Kapfer 2017;Jessica et al. 2021). The two methods were used for a more comprehensive inventory documentation of species such as using camera traps for the documentation of less trappable species (Tasker & Dickman 2001;De Bondi et al. 2010;Thomas et al. 2020). It also allowed for a direct comparison between live trapping and camera trapping efforts in the same location and was aimed to increase the chances of recording targeted samples of non-volant small mammals in the study area (Francesco et al. 2010).

Live Trap
According to Hoffman et al. (2010), the most suitable and easiest trapping method for small mammals is to place traps at more or less fixed intervals at parallel and equidistant transect lines along an equal intervals line, which can cover all types of habitats.
In this study, the transect line constructed was different for each habitat where the boundary area had a longer transect line of 750 m and 250 m at restoration, disturbed forest and undisturbed forest. Each transect line was placed at least 50 m apart, which could be adjusted based on the terrain, accessibility and vegetation types (Pearson & Ruggiero 2003).
A distribution of 80 units of wire mesh live traps measuring 25 cm × 11.5 cm × 12 cm (small traps) and 81 cm × 38 cm × 25 cm (medium traps) were deployed along the transect line in each study site based on ecological conditions and topography of the surrounding area (Figs. 2A,2B and 2C). Live traps were also placed on dead logs, semi-open areas that resemble animal trails, near thistle plants and branches of tree (Jambaari et al. 1999;Zakaria et al. 2001;Norfahiah et al. 2012) to increase trapping success likelihood. The distances between traps were 10 m and were each setup at 50 m interval points. The surface of the traps was covered with forest litters to provide thermal insulation for the captured sample and also acting as camouflage (Torre et al. 2004). Six types of baits were equally distributed among the habitats to compare the effectiveness in luring small mammals. Live traps were baited alternately with sweet potato, banana, palm oil fruits, chicken meat left-overs, roasted prawn and salted fish set for five consecutive days and four consecutive nights. Different types of baits were used to increase the number of species attracted. According to Kok et al. (2013) baits are commonly used for surveying small mammal communities, not only because they attract large numbers of these mammals, but also because they provide sustenance for trapped individuals.
Any captured non-volant small mammals were transferred into dark plastic bags (Payne et al. 2008). Cotton wools coated with chloroform were prepared to anesthetize the samples (Barnett & Dutton 1997). Anesthetized captured samples were then measured (weight, head and body length, tail length, ear length, hindfoot length), sexed and species identification was done in reference to Phillipps and Phillipps (2016) and Francis (2019). All non-volant small mammals captured were tagged on the nail using nail polish (Tingga et al. 2021) for identification and then released back to their habitat (Shukor 2001).
Correct identification of species is of primary importance to many studies (Chaval et al. 2010). According to Pimsai et al. (2014) it incorporates a detailed summary of descriptive characters of the external, cranial and dental morphology (i.e., colour patterns, shape of body or head, size) and measurements (i.e., head and body length, tail length, ear length, hindfoot length) for each of the non-volant small mammal species.

Camera Trap
Camera trapping is an established method for the monitoring of medium-to largesized mammals populations (Garden et al. 2007;Trolle et al. 2008;Mccleery et al. 2014;De Bondi et al. 2010), but are rarely used for smaller mammals. However, according to Jackson et al. (2006) camera traps have been used to record a wide range of fauna in various habitats. Camera trapping is a non-invasive method that generally causes minimum disturbance to the target species, can be left unattended for several days, and are ideally suited for studying rare, elusive, and nocturnal/crepuscular animals that avoid humans (Francesco et al. 2010). The big advantage of using camera traps is that it provides real time record of the animal presence. According to Francesco et al. (2010), camera trapping provides information on activity patterns (from the date and time recorded in the image), behaviour and pelage characteristics that enable individual identification.
Ten camera trap brands from three different models, Digital Trail Camera (HC-800 A), Hunting Trail Camera and Wildlife Camera were used in this study. These cameras used infrared camera sensors triggered by heat and motion set at 1-second interval between exposures unless the animal was running at high speed. We programmed five camera traps to record three photographs for every trigger and another five camera traps with 15-second video duration. The three photos were subjectively defined as a single photographic "event." These settings were selected to provide photographs of the individual in different positions and ultimately increase identification accuracy (Thomas et al. 2020). The cameras were installed on suitable trees with a height of 30 cm-50 cm from the ground at optimum angles overlooking the animal trails without the camera view being blocked by any objects. We identified non-volant small mammal species based on several identification guides (Francis & Barrett 2008;Medway 1978).
Camera traps were deployed at random locations in each study sites (Fig. 3) depending on the ecological condition of the area to avoid biases in the data collected of non-volant small mammals that passed through the camera (Francesco et al. 2010). Each camera deployment point was chosen based on the presence of visible animal trails, footprints, scents, activity areas (e.g., big wallows left by Eurasian wild pig ) and tree marks left by wildlife (e.g., scratching marks of sun bear on tree trunks ]), near streams or in the vicinity of streams ). Baits were not used in this study to avoid any specific preferences or bias in surveying medium (>1 kg) and largesized non-volant small mammal species (Tee et al. 2018).
All 10 units of camera traps were left for five consecutive days for each site per sampling time. The total sampling days for camera trap method were 100 (10 traps × 5 days × 2 phases) within two phases of sampling time. Number of each species photo-captured by random cameras were recorded to calculate species richness.

Data Analysis
The species diversity indices, Shannon-Wiener Index (H'), Evenness (E) and Dominance (D) were used to calculate the species richness of each selected localities using the Paleontological Statistics (PAST) software (Hammer et al. 2001). Shannon Weiner index (Shafie et al. 2014;Estrada-Villegas et al. 2012;Dee et al. 2019) was used to analyse the diversity index of non-volant small mammals in the four main habitats of restoration, boundary, disturbed forest and undisturbed forest area.
According to , One-way ANOVA is a statistical technique extended from an independent t-test to compare the mean for more than three groups of an independent variable. Therefore, one-way ANOVA was used to analyse the influence of different habitats on the number of individuals. Moreover, all the variables are considered as significantly different at p < 0.05 . The trapping frequency of each species was calculated by dividing the total number of species captures by the total number of all captures (Ruppert et al. 2021).

Recorded Species of Non-volant Small Mammals in Cameron Highlands
A total of 23 species of non-volant small mammals from six orders and nine families were recorded by live trapping and camera trapping at the three study sites of Terla A, Bertam and Bukit Bujang Forest Reserve. All the recorded nonvolant small mammals were captured in four habitat types: restoration, boundary, disturbed forest and undisturbed forest area.
Based on

Species Richness
Surveys at the three localities at Terla A, Bertam and Bukit Bujang Forest Reserve successfully recorded a total of 23 species representing nine families, from which 17 species were captured by live traps and 12 species from camera trap photos or videos (Table 1). The family Muridae recorded the highest number of non-volant small mammals captured. Almost 47.83% of the individuals captured fall in this order of Muridae with 11 species recorded using live traps and three species recorded using camera traps. Musser (2017) attributed high population of rodents to the availability of food and shelter as they co-exist with humans. On the other hand, the family Muridae showed that it is relatively common in all three study sites, with no significant difference in numbers between the study sites. 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). Both Sciuridae and Viverridae recorded only three species from each family (13.04%). The other six families, namely Erinacidae, Felidae, Herpestidae, Tragulidae, Tupaiidae and Manidae were represented by a single specimen, each only representing 4.34% of all captures. These six families of non-volant small mammals could be rare, as only single individuals were captured for the entire study period.
Muridae was the most common family with 11 species caught, Berylmys bowersi being the species with highest number of individuals caught (seven) and the species mostly frequently caught close to forest litter. Sundamys muelleri came in second with six individuals and 15.38% of the captures. Sundamys muelleri has a wide distribution and was normally caught near rivers (Francis & Barrett 2008;Jayaraj et al. 2012). Additionally, Jayaraj et al. (2012) reported this big rodent was caught in the limestone area of Gua Ikan where there is a river flowing into the cave. In Borneo, this species was caught on ground or on low trees (Wilson et al. 2006). Zakaria et al. (2001) also reported that this species can survive in disturbed habitats.
Ten species, i.e., Paguma larvata, Paradoxurus musangus, Hylomys suilus, Niviventer cameroni, N. fulvescen, Rattus tanezumi, R. exulans, R. tiomanicus, R. norvegicus, Tupaia glis and Callosciurus caniceps were found in the trapping exercise with one individual representing each species. In conjunction with the study, two species from the family Viverridae i.e Paradoxurus musangus and Paguma larvata were caught in the traps as they are common in both pristine and disturbed forests. Both species can adapt in primary and secondary forests, albeit lower in the latter than the former (Nakashima et al. 2010). This could possibly explain the capture of single individuals of both species in the study sites.
Paradoxurus musangus has a wide geographical distribution and global presence due to its adaptability to a wide range of habitats (Duckworth et al. 2016). In Bukit Timah Nature Reserve (BTNR) Singapore, the presence of feral populations that were originally pets deliberately released by or escaped from their owners contributed to the increase in illegal poaching of the species (Chan & Davison 2019).
The adaptation of M. whiteheadi in the undisturbed area (Table 2) shows that this habitat type is preferred by this species (Chapman et al. 2019). According to Francis and Barrett (2008), this species can be found in tall and undisturbed secondary forests, occasionally encroaching disturbed areas in the vicinity of these forests.

Diversity indices and Relative Abundance for Live Trap Method
Species diversity, Shannon Wiener (H') was higher at Terla A Forest Reserve (H' = 2.274) compared to Bertam and Bukit Bujang Forest Reserve. In contrast, the species evenness (E) in Bertam (E = 0.857) was lower than Bukit Bujang Forest Reserve (E = 0.931). The value of Simpson's Dominance Index, D, indicated a low species dominance at Bertam at 0.806 while the value of species dominance is slightly higher in Bukit Bujang FR at 0.809 (Table 3). According to Rabosky (2009), different ecological limits seem to be the main determinants of diversification and, therefore, species richness. The Chao-1 estimator indicated that Terla A was the richest area. The results of this study revealed that a recovered land of the secondary forest had an impact on the diversity and distribution of nonvolant small mammals. The age of the restoration area contributes to the growth of vegetation in each habitat (Derhe et al. 2017). The restoration project for Terla A and Bertam Forest Reserves began in November 2017 and no restoration projects were ongoing in Bukit Bujang (Forestry Department of Peninsular Malaysia 2020).
Although secondary and recovering forests may harbor a similar number of species as mature forests (Poorter et al. 2021), communities in secondary forests are usually dominated by generalist species (Gardner et al. 2007). According to Whitehead et al. (2014), restoration sites were progressing towards becoming a rainforest and deviating from pasture sites in their small-medium mammal composition.  Table 4 shows the species diversity, Shannon Wiener (H') in the boundary area (S = 8, H' = 2.025) and disturbed forest area (S = 8, H' = 1.992) shows the same total of number of species (S) compared to other study habitats; restoration area has the lowest species diversity (S = 3, H' = 0.950). In contrast, the species evenness (E) in the disturbed forest area (E = 0.916) is lower than in the undisturbed forest area (E = 0.930). The value of Simpson's Dominance Index, D, shows low species dominance in restoration area at 0.56 while the value of species dominance is slightly higher in the disturbed forest area at 0.86. The Chao-1 species richness estimator indicated that the boundary area was the richest area. The restoration habitat contributed to the least number of non-volant small mammals compared to other habitats, likely because the habitat was quite open without tree canopies compared to other habitats. According to Yaap et al. (2010), the presence of nonvolant small mammals in this particular habitat was influenced by the availability of feed, water, and shelter provided by the nearby settlement areas that surrounds the forest. Zakaria et al. (2001) also stated that rodents can sustain themselves with seeds, fruits and plant matter from the natural vegetation without any detrimental effects on the habitat/ecosystem. Cold weather and limited food sources such as fruit could be possible factors that lead to low diversity in mountain habitats (Shukor 1997;Butler & Lawrence 2019). Non-volant small mammals such as rodents and insectivores are highly mobile animals whose distribution is influenced by the altitude and vegetation types as well as human disturbance (Mulungu et al. 2008;Sukma et al. 2019). Although the Evenness (E) at Bukit Bujang FR recorded the lowest values for both abundance and diversity compared to the other two sites, the differences between all three study areas were found to be significant. Such a result in Bukit Bujang FR may be affected by the habitat itself, where the relative abundance, in reference 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. One-way ANOVA analysis performed between sites, there were significant differences in non-volant small mammals' diversity for Terla A, Bertam and Bukit Bujang Forest Reserve (F (5,22) = 3.086, p = 0.029). Contrasting vegetation and intensities of anthropogenic activities found at each site may explain the observed variations. In previous studies, Ramli and Hashim (2009) have reported that small mammal populations in tropical forests have seasonal variations and variations in population structure, density, biomass and species richness even if they live within the same region but in different habitat types.

Camera Trapping
The total field effort comprised of 10 camera traps, over a cumulative period of five days for each site, recorded a total of 94 photographs. A total of 12 species from eight families were recorded from three study sites where Bertam FR captured the most photographs with a total of 46 photos (Table 5). Forty squirrels, Lariscus insignis and Callosciurus canicep recorded the highest photos captured. A Sciurids, L. insignis which is considered to be elusive and typically trap shy as it has never been live trapped to date, was successfully recorded by the camera trap. Saiful et al. (2001) reported the home range of L. insignis as too small which posed a challenge for live trapping as only one individual was successfully trapped in their study in Ulu Gombak. Ruppert et al. (2021) also encountered only one individual of Lariscus insignis in their study in Penang Island.
In reference to the species inventory of Cameron Highlands, Tragulus kanchil is a newly recorded species in Cameron Highlands. Based on camera trap results, T. kanchil develops nocturnal pattern behaviour as they became elusive due to fragmentation and frequent encroachment. According to Farida et al. (2006), Tragulus sp. is typically shy and rarely seen in the forest and has only been caught in camera trap foraging the forest floor looking for fruits.
In the present study, camera trapping techniques were unable to identify non-volant small mammal species such as rats and squirrels, presumably, due to their small size but can be classified under the order of Rodentia such as Rattus sp. and Sciurus sp. According to Nick et al. (2021), their size is often insufficient to trigger infra-red sensors, and resultant images may be of inadequate quality for species identification with a possibility that some species not being able to be identified.
Camera trapping eliminates the need to handle an individual physically. It offers a method for detecting rare, elusive, or trap-shy individuals that may be missed by traditional, intensive, shorter-duration live trapping methods (Gray et al. 2017;Rendall et al. 2014). According to Pollock et al. (2002), the major limitation associated with the use of camera traps for terrestrial mammals is that some species may not be detected.

Comparison of The Present and Previous Studies on Non-Volant Small Mammals in Cameron Highlands
Recent study recorded 17 species of non-volant small mammals captured by live traps and 12 species from camera trap photos. Previous study conducted by Shahfiz et al. (2011) in two forest reserves in Cameron Highlands recorded a total of 18 individuals from five species of non-volant small mammals captured in Mentigi and Ulu Bertam Forest Reserve. The common treeshrew (Tupaia glis), grey-bellied squirrel (Callosciurus caniceps) and white-bellied rat (Niviventer fulvescens) were caught in the study where five individuals were recorded for each species. Studies conducted by Shahfiz et al. (2011) andNur Laila (2019) are in parallel with current study where they also found Berylmys bowersi, Niviventer cremoriventer and Tupaia glis in Cameron Highland. Nine individuals from four species in Terla A Forest Reserve, Cameron Highlands comprising of Muridae, Sciuridae and Tupaiidae family were captured by Nur Laila (2019). The most abundant species were N. cremoriventer (four individuals) followed by Dremomys rufigenis (three individuals) and current study in Terla A Forest Reserve also recorded three individuals of Dremomys rufigenis. However, only one individual T. glis was captured using live trap and caught at the boundary area in Terla A Forest Reserve. Tupaia glis can be easily observed foraging and moving around in the presence of human (Noor Aisyah et al. 2016) as they have high tolerance to habitat disturbance (Nur Syazana et al. 2013;Muhammad Hafiz et al. 2015).

CONCLUSION
Twenty-three non-volant small mammal species were recorded in this study. Overall, Terla A Forest Reserved recorded the highest diversity of non-volant small mammals with 14 species. Although the restoration habitat differed greatly from the boundary, disturbed and undisturbed forest area in terms of species richness, it still appears to have an important role in providing habitat for highly adaptable species. Therefore, it is crucial for the authorities to manage these non-protected areas properly as they continue to function as an ecosystem. In addition, some species are endemic such as N. cameroni and N. cremoriventer to specific habitats or elevations (Pimsai et al. 2014), making them a high conservation priority.
One limitation is that although overall species diversity of small non-volant mammals is high, trap success may have been low in some parts of the habitat due to weather conditions during data collection. It was raining during the sampling period which may have affected the activity of non-volant small mammals. This could have impacted our data because the mammals would have been less likely to roam around. Therefore, many species may not have been detected in the rapid survey.