BIOTROPICA 48(3): 373–380 2016 10.1111/btp.12306

Movement Behavior of Native and Invasive Small Shows Logging May Facilitate Invasion in a Tropical Rain Forest

Robin Loveridge1,2,6, Oliver R. Wearn1,3, Marcus Vieira4, Henry Bernard5, and Robert M. Ewers1 1 Department of Life Sciences, Imperial College London, Silwood Park Campus, Buckhurst Road, Ascot, Berkshire, SL5 7PY, U.K. 2 BirdLife International Cambodia Programme, #2, Street 476, Toul Tompung 1, Chamkarmon, PO Box 2686, Phnom Penh, Cambodia 3 Institute of Zoology, Zoological Society of London, Regent’s Park, London, NW1 4RY, U.K. 4 Departamento de Ecologia, Instituto de Biologia, Universidade Federal do Rio de Janeiro, Caixa Postal 68020, 21941-902, Rio de Janeiro, Brazil 5 Institute for Tropical Biology and Conservation, Universiti Malaysia Sabah, Sabah, Malaysia

ABSTRACT Invasive species pose one of the greatest threats to biodiversity. This study investigates the extent to which human disturbance to natural ecosystems facilitates the spread of non-native species, focusing on a small community in selectively logged rain forest, Sabah, Borneo. The microhabitat preferences of the invasive rattus and three native species of small mammal were examined in three- dimensional space by combining the spool-and-line technique with a novel method for quantifying fine-scale habitat selection. These methods allowed the detection of significant differences for each species between the microhabitats used compared with alternative, available microhabitats that were avoided. Rattus rattus showed the greatest preference for heavily disturbed habitats, and in contrast to two native small mammals of the Maxomys, R. rattus showed high levels of arboreal behavior, frequently leaving the forest floor and traveling through the understory and midstory forest strata. This behavior may enable R. rattus to effectively utilize the complex three-dimensional space of the lower strata in degraded forests, which is characterized by dense vegetation. The behavioral flexibility of R. rattus to operate in both terrestrial and arboreal space may facilitate its invasion into degraded forests. Human activities that generate heavily disturbed habitats preferred by R. rattus may promote the establishment of this invasive species in tropical forests in Borneo, and possibly elsewhere. We present this as an example of a synergistic effect, whereby forest disturbance directly threatens biodiversity and indirectly increases the threat posed by invasive species, creating habitat conditions that facilitate the establishment of non-native fauna.

Abstract in Malay is available with online material.

Key words: arboreality; habitat preference; invasion; Rattus rattus; selective logging; synergistic effects.

VAST AREAS OF THE REMAINING NATURAL FOREST IN SOUTHEAST spread of disease (Wells et al. 2014a, b), have been poorly ASIA EXIST IN A LOGGED AND DISTURBED STATE. In Borneo and explored in tropical forests thus far, but have the potential to Sumatra, for example, 46 percent and 72 percent, respectively, of increase the extinction risk of native species (Brook et al. 2008). natural forest is degraded (Margono et al. 2012, Gaveau et al. Small mammals play a key role in biological communities, 2014). Within the Malaysian state of Sabah, Borneo, <2 percent acting as both seed predators and dispersers (Asquith et al. 1997, of the remaining lowland forests remain undisturbed, mostly in Wells & Bagchi 2005, Wells et al. 2009). Small mammals are also just a few sites making up a total of 700 km2 (Reynolds et al. important prey items for large avian and mammalian predators 2011). With the vast majority of forests existing in an altered (Wilting et al. 2006, Puan et al. 2011). Changes in the abundance state, it is essential for conservation to understand the processes of this functionally important group in altered forests may cause that drive community change in modified landscapes (Hansen cascading effects on other trophic levels (Grassman et al. 2005). et al. 2001, Meijard & Sheil 2007). Despite this, we still have only Native species may be particularly threatened by the inva- a rudimentary mechanistic and theoretical understanding of why sion of the black , Rattus rattus (Stokes et al. 2009, Gibson et al. these changes in biological communities occur, and the possible 2013). Rattus rattus is now spread over most of the world (Amori role that invasive species may play (Stokes et al. 2009). In particu- & Clout 2003), and when it has invaded native habitats on lar, synergistic effects between threat processes, such as between islands, it has led to well-documented extinctions of native bird logging and biological invasion, or biological invasion and the fauna (Atkinson 1985, Barnett 2001), disruption to nutrient trans- port and soil fertility (Fukami et al. 2006) and carbon sequestra- Received 14 June 2015; revision accepted 8 October 2015. tion (Wardle et al. 2007). The multiple ecological effects that 6Corresponding author; e-mail: [email protected] cascade from the invasion of R. rattus places a premium on ª 2016 The Association for Tropical Biology and Conservation 373 374 Loveridge, Wearn, Vieira, Bernard, and Ewers

understanding how those invasions occur in order to limit the rat ( sabanus). We compared these species to the newly future spread of this species. invading black rat (Rattus rattus). Ear tissue samples were taken In Borneo, R. rattus was traditionally thought to be restricted from all individuals and stored in 95% alcohol. MtDNA sequenc- to urban areas (Payne & Francis 1985). However, R. rattus has ing of the Rattus species confirmed that it was a member of the been recently observed in a small number of primary and sec- Rattus rattus species complex (sensu Aplin et al. 2011, Pages et al. ondary tropical forests of northern Borneo (Wells et al. 2006, 2010) based on both cytochrome b and CO1 gene sequences (M. 2014b, Cusack et al. 2014) and detected in the oil palm plantation Pages, pers. comm.). matrix surrounding the forests examined in this study (Cusack 2011). Although this species has not been systematically surveyed TRAPPING AND TRACKING INDIVIDUALS.—Three trapping grids were across Borneo, the limited data available suggest it is not yet established at locations that encompass a gradient of logging ubiquitous in natural forest areas (Nakagawa et al. 2007, Wells intensity and forest modification. Grids were located more than a et al. 2007, Bernard et al. 2009, Charles & Ang 2010). This pro- kilometer from the forest edge and were all connected by contin- vides a unique opportunity to examine the niche overlap of uous forest. Trapping was conducted for 3 months between May R. rattus with native species at a relatively early stage in the inva- and July of 2012. Traps were checked in the morning and cap- sion process. tured individuals belonging to the four study species were anes- Previous survey work has demonstrated that the occurrence thetized using diethyl ether and a small fur clip was made of R. rattus increases along a gradient from undisturbed to heavily dorsally between the shoulder blades. A spool was then attached disturbed forest (Cusack et al. 2014). Here, we build on this work to the underfur using acrylamide gel (Loctite). Spools were made by examining whether logging facilitates the invasion of R. rattus of nylon quilting cocoons and weighed between 1 g and 2.5 g, into tropical forests by investigating their movement behavior extending over distances of 100–200 m (Danfield Ltd). Spools and microhabitat selection at small spatial scales. Our hypothesis were wrapped in a thin layer of cling-film, followed by a layer of is that the invasive R. rattus has a stronger preference for heavily electrical tape to prevent snagging on vegetation. The weight of disturbed forest microhabitat compared with native small mam- spools was adjusted to <5 percent of the individual’s body weight mals, hence logging disturbance creates favorable microhabitats following the ratio used in radio-tracking studies (Sikes & Gan- for the establishment of R. rattus in tropical forests. non 2001). In addition, all trapped target species were injected We compare the habitat preferences of the invasive R. rattus with a unique passive induced transponder tag (Francis Scientific with three native species of small mammals regarding (1) vertical Instruments, Cambridge) to enable individual identification. Some space use, (2) preference for more (or less) disturbed forest individuals that were recaptured were spool-tracked multiple microhabitats, and (3) preference for moving along and through times. particular substrates associated with forest disturbance. To effec- tively quantify a species’ habitat preference, habitat measurements HABITAT VARIABLES.—Habitat use was examined at two spatial must be made at the spatial scale at which the species makes scales. The largest spatial scale was the scale at which the avail- habitat selection choices (Manly et al. 2002). We were able to ability of microhabitats changed in the forest environment and study fine-scale microhabitat preference using the spool-and-line was standardized to intervals of 10 m along the ’s path. technique (Miles et al. 1981, Boonstra & Craine 1986), as it The smallest spatial scale was the scale at which substrates (indi- allows even very subtle route choices to be quantified (Harris vidual fallen branches, leaf litter patches, open ground) changed et al. 2006, Wells et al. 2006). In addition, we employed a novel along the animal’s path. This spatial scale varied from approxi- matching methodology to establish actual preference for micro- mately 0.5–5m. habitats, comparing used versus control microhabitats. QUANTIFYING MICROHABITAT PREFERENCES.—The first 10 m of METHODS each spool track were discarded as a flight response (Harris et al. 2006), and the first microhabitat measurements were recorded a STUDY SITE AND SPECIES.—The study was conducted within the further 10 m from the starting point of normal behavior. Prelimi- Stability of Altered Forest Ecosystems (SAFE) project site in nary trials were conducted to estimate the separation distance at Sabah, Malaysian Borneo (Ewers et al. 2011). Small mammal trap- which repeated measures on the spool track became independent ping was carried out in a 7200 ha area of repeatedly logged for- from each other. The autocorrelation between consecutive 10-m est, with high spatial variation in levels of forest disturbance. The observations was assessed using the autocorrelation function continuous forest of the SAFE project area is connected to a (ACF) in R (R Development Core Team, 2014). Negligible corre- large (>1 million ha) area of similar forest to the north and is lation was found at 10-m intervals (correlation = 0.046), so this otherwise surrounded by an oil palm plantation matrix. was subsequently used as the spacing between repeated micro- The four study species belong to the family ( habitat measurements. and mice). Native species included the red spiny rat (Maxomys sur- Microhabitat was defined as the area within a 1-m radius of ifer), a widespread generalist species with a characteristic red the spool and nine microhabitat variables were measured at 10-m pelage, Whitehead’s rat (Maxomys whiteheadi), classified as Vulnera- intervals along the animal’s track (a detailed description of micro- ble on the IUCN red list (IUCN 2014), and the long-tailed giant habitat variables is provided in Table S1). All microhabitat vari- Logging Facilitates Invasion of Small Mammals 375

ables were recorded by the same researcher (R.L.). The density of of eight categories following Wells et al. (2006) and the length of vegetation in four layers of forest strata was estimated visually the step (a detailed description of the different substrate cate- for each 10-m interval along the spool track following Puttker gories is provided in Table S1). Species-specific step lengths were et al. (2008) (near the ground 0–0.5 m, understory 0.5–3.0 m, similar; M. surifer had the longest mean step length at 2.0 m and midstory 3.0–20 m, and canopy >20 m). We assigned a score of M. whiteheadi the shortest at 1.8 m. between 1 and 4 to each stratum to give an index of vegetation As with the microhabitat variables, we recorded paired con- density. Four further variables indicating potential resource avail- trol observations for each step by noting the dominant substrate ability and predation risk were also recorded (total canopy clo- along a control step located off the spool track. Control steps sure, forest quality, tree basal area, and habitat concealment were determined by walking the same distance as the observed score). As covariates that may influence the space use of individ- step on a bearing of 45 °, 135 °, 225 °, or 315 ° relative to the uals, we recorded rainfall during the night and lunar phase. Rain- direction of the spool step, with bearings selected sequentially. fall may influence small mammal movement by masking the noise of travel across complex substrates (Vickery & Bider 1981, STATISTICAL ANALYSIS.—Linear mixed effects models were used to Bowman et al., 2000), and variation in moonshine may influence account for the nested repeated measures taken on the same the visibility of the microhabitats, making some locations less spool track from individually identified small mammals. The R desirable when well lit by a full moon (Kotler et al. 1991). packages nlme and lme4 (Bates et al. 2011, Pinheiro et al. 2011) An assessment of microhabitat preference requires a com- were used for these analyses. The categorical rain and lunar phase parison between the microhabitats that are selected by a species variables were included as covariates in all models but are not with other microhabitats that are available, but not used. To com- reported, as they do not impinge on our hypotheses. The signifi- pare observed versus control microhabitat, measurements taken cance of fixed effects, including interaction terms, was deter- every 10 m on the spool track were paired with measurements at mined from likelihood ratio tests by comparing models with and a spatially matched location off the track. The locations of these without the fixed effect or interaction. control measurements were determined by walking 10 m in a To test for vertical spatial segregation between species, we straight line on a bearing relative to the spool track of 45 °, used a binomial general linear mixed effects model to compare 135 °, 225 °, or 315 °, with bearings selected sequentially repeated measures of the height aboveground at 10 m intervals (Fig. 1). along the spool track across species. Height was categorized into ground (<0.5 m) or aboveground (>0.5 m) for analysis, and spe- QUANTIFYING SUBSTRATE PREFERENCES.—We broke spool tracks cies, rain, and lunar phase were fitted as fixed effects. The nested into a series of discrete steps, where a step was defined as a data structure was explicitly modeled by nesting repeated mea- straight-line section of track with no change of direction >20 °. sures taken on the same spool track within individually identified For each step, we characterized the dominant substrate into one small mammals. The microhabitat variables used in this study examined dif- ferent layers of forest strata and together provided a holistic mea- sure of forest disturbance. The values were highly intercorrelated, so we reduced the number of variables to a smaller set of princi- pal component axes using the vegan package in R (Oksanen et al. 2011). Variable standardization was carried out by dividing the value of each variable by the maximum values for that variable. Much of the variation (35%) was captured within the first princi- ple component (PC1), and we therefore used this as the only variable in subsequent modeling of species’ microhabitat prefer- ence. This single composite variable provided a robust and easy to interpret axis from low forest disturbance (negative PC1 val- ues) to high forest disturbance (positive PC1 values). Negative PC1 values were related to high forest quality scores, high canopy and midstory density, and high tree basal area. Positive PC1 scores were associated with high understory and high ground vegetation density. FIGURE 1. An example spool track of a male M. surifer illustrating the spool To compare the strength of microhabitat preference between track the animal moved along (continuous black line) with observation loca- species, a single ‘microhabitat preference’ response variable was fi tions ( lled circles) and the spatially matched control routes (dotted lines) with generated for each 10-m observation along the spool track. This = = fi control locations (open circles). A release point; AB rst 10-m route of response variable was defined as the difference in PC1 scores fl = spool track that was discarded as a ight response; BCobs observed 10-m between observed and paired control observations. This provided = route of spool track; BCcont spatially matched 10-m control route for seg- a measure of the strength of preference for more or less dis- ment BC. turbed microhabitats in relation to the immediate, surrounding 376 Loveridge, Wearn, Vieira, Bernard, and Ewers

environment. A linear mixed effects model was used to compare microhabitat preference among species, retaining rain and lunar phase as additional fixed effects and nesting repeated measures of tracks within individually identified small mammals as random effects. To compare the use of substrates across species, we first calculated the ratio between the number of times a substrate category was recorded in observed versus matched control steps per spool track. This ratio was then used to standardize the observed frequency of substrate use by the availability of the substrate in the immediate vicinity of each spool track. We log-transformed the ratio so that positive and negative values scaled equally. Positive values of this ‘substrate preference’ score indicate preference for that particular substrate and nega- tive values indicate avoidance of that particular substrate. A linear mixed effects model was used to compare substrate FIGURE 2. Microhabitat preferences of three native (white) and one intro- preference among species and substrate type, retaining rain and duced species (gray) of small mammals. Positive values of the habitat prefer- lunar phase as additional fixed effects and nesting repeated ence score indicate a preference for selecting more disturbed habitats over measures of tracks within individually identified small mammals less disturbed habitats, and negative values indicate the reverse. as random effects. LS = Leopoldamys sabanus,MS= Maxomys surifer,MW= Maxomys whiteheadi, RESULTS RR = Rattus rattus.

SPECIES’ SPATIAL BEHAVIOR.—We attached 53 spools to 41 individ- uals, from which a total of 293 movement steps were recorded: turbed forest (preference = 0.31, SE = 0.06), followed by the 15 M. surifer individuals with 139 steps, 13 M. whiteheadi individu- two Maxomys species that had weaker preferences for disturbed als with 81 steps, nine L. sabanus individuals with 41 steps, and forest (M. whiteheadi: preference = 0.13, SE = 0.01; M. surifer: four R. rattus individuals with 32 steps. The small number of preference = 0.10, SE = 0.01). Pairwise comparison between the R. rattus reflects their relatively low abundance in this habitat at two Maxomys species confirmed that M. whiteheadi showed the what we believe is an early stage in the invasion process. The marginally stronger, though non-significantly different, preference total spool track followed was 3.4 km with a mean spool length for more disturbed habitats (difference in preference of 64 m per individual (SE = 44.3 m). All species were captured score = 0.03, SE = 0.04, df = 27, P = 0.45). Of the four species, on each of the three trapping grids, and individuals of different only L. sabanus had a preference for less disturbed habitats (pref- species displayed a large amount of spatial overlap in their terres- erence = À0.089, SE = 0.01). trial foraging behavior. Spool tracks of different species released The four species showed similar patterns of preference on the same night were occasionally found to cross over each across the substrate types (Fig. 3), with no significant differences other, and individuals of different species were recorded using among species (v2 = 4.33, df = 3, P = 0.228) and neither was the same fallen log or dry riverbed. there a significant interaction effect between species and substrate Vertical segregation of space was observed between species preference (v2 = 24.60, df = 17, P = 0.104), despite an apparent (v2 = 26.43, df = 3, P < 0.001). All observations of both Max- preference for R. rattus to prefer moving through suspended leaf omys species were recorded on the forest floor (<0.5 m above the litter that the three native species showed no preference for ground). However, spool tracks of both R. rattus and L. Sabanus (Fig. 3). Across species, however, there were strong preferences were frequently observed to pass along lianas (woody vines) that for some substrates over others (v2 = 88.97, df = 6, P < 0.001). emanated from the forest floor and penetrated up into the under- Overall, small mammals showed the strongest preference for fal- story and midstory forest strata. Twenty-eight percent of R. rattus len wood (preference value = 0.85, SE = 0.17) and dry stream observations (N = 32) and 27 percent of L. Sabanus observations beds (preference value = 0.83, SE = 0.38), followed by rocky (N = 41) were recorded in the understory (0.5–3 m) and a fur- substrates (preference value = 0.53, SE = 0.18). Species tended ther 9 percent of R. rattus observations and 15 percent of L. sa- to actively avoid leaf litter (preference value = À0.47, SE = 0.07), banus observations were recorded in the midstory of the forest but were not strongly influenced by suspended leaf litter (prefer- (3–10 m). ence value = 0.32, SE = 0.06), vines (preference value = 0.25, SE = 0.06), or bare ground (preference value = À0.06, MICROHABITAT AND SUBSTRATE PREFERENCE.—There were signifi- SE = 0.01). Comparing the two Maxomys species; M. surifer had cant differences among species in their microhabitat preferences the slightly stronger preference for traveling along vines, while (Fig. 2) (v2 = 17.91, df = 3, P < 0.001). Across all species, M. whiteheadi had the stronger preference for traveling over rocky R. rattus had the strongest preference for the most heavily dis- terrain (Fig. 3). Logging Facilitates Invasion of Small Mammals 377

FIGURE 3. Substrate preference values of Leopoldamys sabanus (A), Maxomys surifer (B), Maxomys whiteheadi (C), and Rattus rattus (D) for bare ground (BG), fallen wood (FW), leaf litter (LL), rocky substrates (RK), suspended leaf litter (SLL), dry stream beds (ST), and vines (V). Positive preference values indicate substrates that are preferred and negative values indicate substrates that are avoided.

DISCUSSION of particular concern as nest predation by this introduced omni- vore may negatively impact the survival success of native bird Our results emphasize how species make habitat selection species, as has been the case on other tropical islands such as choices at multiple spatial scales (Manly et al. 2002), and that Hawaii (Amarasekare 1993). small-scale processes can be used to help explain large-scale phe- The behavioral flexibility of R. rattus to utilize both terrestrial nomena such as the invasion of R. rattus into modified forests. and arboreal space may significantly increase the species’ invasion We have confirmed earlier survey work that demonstrated R. rat- capability into the complex three-dimensional forest environment. tus is more likely to be present in heavily modified than unmodi- Indeed, invasive species are more frequently habitat generalists fied tropical rain forest (Cusack et al. 2014) and confirmed our than specialists (Olden et al. 2004, Whitney & Gabler 2008). hypothesis that this pattern appears to arise from specific, small- When entering a new forest habitat R. rattus can potentially scale alterations to microhabitat structure that R. rattus actively access both terrestrial and arboreal foraging resources and a prefers to move in. greater extent of niche space than an obligately terrestrial or obli- gately arboreal species. Additionally, in Sabah, arboreal murid VERTICAL SPATIAL SEGREGATION.—This study represents one of the rodent communities are relatively species poor compared with only efforts to quantify the extent of arboreal space use by R. rat- terrestrial small mammals (Wells et al. 2004). Therefore, the arbo- tus (Key & Woods 1996), and how this ability may facilitate inva- real adaptive zone could have more space for invasion. In addi- sion in logged tropical forests. Neither Maxomys species showed tion, it has also been suggested that arboreal behavior may arboreal behavior; however, both L. sabanus and R. rattus showed reduce exposure to terrestrial predators and present physical bar- significantly different vertical space use compared with these ter- riers for aerial predators (Montgomery & Gurnell 1985, Buesch- restrial species, utilizing higher layers of the forest strata. Indeed, ing et al. 2008). the foot morphology of the two species is very similar, both hav- ing broad feet and prominent plantar pads, adaptations known to MICROHABITAT PREFERENCE.—Both Maxomys species and R. rattus indicate arboreality in murid (Aplin et al. 2003). More- showed significant preference for more disturbed microhabitats, over, the long tails of both species, up to 120 percent of body corroborating the results of a previous study using trapping data length for R. rattus and 135 percent of body length for L. sabanus, at the same study site (Cusack et al. 2014). Here, however, we help with balance and may well be adaptions for climbing (Boon- have demonstrated at a smaller spatial scale that these three spe- song et al. 1988). Therefore, R. rattus is well-adapted to fully cies choose to move through more disturbed habitats than less exploit the three-dimensional complexity of forest habitat. This is disturbed, alternative habitats in the immediate surrounding 378 Loveridge, Wearn, Vieira, Bernard, and Ewers

environment. Disturbed habitats were characterized by a low between the two studies likely arises from differences in method- presence of large trees, allowing the growth of dense ground and ologies. Cusack et al. (2014) used microhabitat variables associated understory vegetation layers. This provides excellent cover from with individual trap sites, representing point locations where indi- predators. Heavily disturbed habitats may maximize both preda- vidual small mammals were captured, whereas we used more tor avoidance and resource availability, making these habitats finely detailed information on the movement tracks of individuals desirable for small mammals. The greater abundance of these through the habitat. This apparent contrast in results, therefore, habitats after human disturbance may also explain why it has reinforces the importance of measuring habitat preference at the been observed that small mammal populations tend to increase spatial scale at which the species makes habitat selection choices after logging activity (Adler & Levins 1994, Pardini 2004). in order to accurately discern patterns of behavior (Manly et al. Among the three species, R. rattus showed the strongest prefer- 2002). ence for more disturbed habitat. This suggests that heavily dis- turbed forest habitat is a more strongly preferred habitat for IMPLICATIONS FOR IMPROVED MANAGEMENT OF LOGGED FOREST R. rattus than it is for the native Maxomys species. Heavily dis- SYSTEMS.—The microhabitat data presented demonstrates how turbed forest habitat may, therefore, facilitate the invasion of this one of the major threats to biodiversity in the tropics, namely species into logged forests in Borneo (Stokes et al. 2009, Cusack habitat degradation and modification (Brown & Brown 1992), et al. 2014). may multiply the likely impact of a second major threat, inva- It is interesting to note that this same preference for dis- sive species (Kot et al. 1996). Specifically, we found that habitat turbed habitat was not replicated by L. sabanus, which showed a degradation may generate habitat that is more suitable for preference for less disturbed habitat. One explanation for this R. rattus, so promoting the spread of this invasive species, which reverse trend is that the semi-arboreal L. sabanus benefits from may in turn threaten the persistence of native fauna. Species exploiting resources present in the more intact canopy of less dis- potentially impacted by the invasion of R. rattus into tropical turbed forest. This species also has a relatively larger body size, rain forests include understory and midstory nesting birds, making it more mobile. This, combined with semi-arboreal which are often vulnerable to nest predation by rodents and behavior, may make this species less susceptible to predation other small mammals through direct competition. In addition, (Montgomery & Gurnell 1985) and therefore more able to the threat of R. rattus to native fauna may be further magnified exploit the microhabitats that lack the dense vegetation layer by the species potentially acting as a conduit for the introduc- associated with disturbed forest sites. tion of novel diseases into native small mammal communities (Wells et al. 2014b). SUBSTRATE PREFERENCE.—All study species showed a clear prefer- The findings of this study agree with those of Cusack et al. ence for moving along fallen wood (Fig. 3). Woody debris con- (2014) in demonstrating that R. rattus shows strong preference tains high concentrations of invertebrates, a key resource for for heavily disturbed habitats. We, therefore, recommend that the small mammals (Wells et al. 2006). Larger pieces of woody debris best practical defense against the invasion of R. rattus into tropi- may also be used as cover in order to reduce visual detection by cal forests is to preserve forest habitats that are less suitable for predators (Bowman et al. 2000). This is supported by the obser- the species (Bernard et al. 2009). This may be achieved through vation that individuals in this study sometimes traveled parallel to reduced impact logging techniques that are designed to minimize fallen wood, close against the side. Traveling on top of woody levels of habitat disturbance and degradation (Gerwing & Uhl debris has also been hypothesized to provide a simple substrate 2002, Putz et al. 2008). to allow for faster, more efficient travel and to allow small mam- Our results also point toward specific new methods for min- mals to scan more effectively for predators while moving (Shad- imizing the likelihood of R. rattus invasion based on the fine-scale bolt & Ragai 2010). Therefore, in terms of foraging and habitat preferences of the species for particular forest structural predation risk, fallen wood may be an important microhabitat elements. It is a common logging practice to undertake some feature for maintaining populations of small mammals. crude shaping of felled wood extracted from the forest prior to Small mammals tended to avoid moving through leaf litter, a transportation to reduce the weight and cost of transporting the pattern that may be explained by the complexity of this substrate wood. This contributes to large quantities of wood fragments inducing slow, noisy travel with greater risk of predation (Shad- being discarded within logged forest stands, at the sides of log- bolt & Ragai 2010). By contrast, the preference for sunken, dry ging roads and along the forest edge. Recent research by Pfeifer river beds may be explained by the more enclosed nature of this et al. (2015) has shown that the volume of deadwood within a substrate, reducing visibility to predators. logged tropical forest increases with greater forest disturbance. Using trap-level habitat data, Cusack et al. (2014) found that We have demonstrated that fallen wood is especially preferred by the occurrence of L. sabanus could be predicted by the presence R. rattus. Therefore, reducing the availability of this favored of fallen wood and low levels of leaf litter. However, these were microhabitat type along the forest edge may be one means of not significant predictors of occurrence for either Maxomys spe- reducing the likelihood of R. rattus invading logged forest stands. cies or R. rattus (Cusack et al. 2014). By contrast, we found that This could be achieved by implementing logging practices that all study species showed a significant preference for fallen wood promote extraction efficiency. For example improving tree selec- and an avoidance of leaf litter. The slight contrast in the findings tion to only log the most suitable tall, straight trees would Logging Facilitates Invasion of Small Mammals 379

minimize the amount of wood discarded. Improved disposal of ATKINSON, I. A. E. 1985. The spread of commensal species of Rattus to ocea- waste wood could also be explored. nic islands and their effects on island avifaunas. In P. J. Moors (Ed.), – Furthermore, cutting back midstory vegetation, such as lia- Conservation of island birds, pp. 35 81. International Council for Bird Preservation, Bristol. nas and knocking suspended leaf litter to the ground may reduce BARNETT, S. A. 2001. The story of rats. Their impact on us, and our impact the ability of invasive small mammals to penetrate the non-terrestrial on them. Allen & Unwin, Crows Nest, Sydney. layers of forest strata. Before committing to such an action, how- BATES, D., M. MAECHLER, AND B. BOLKER. 2011. lme4:Linear mixed-effects ever, it will be important to ensure that this approach will not inad- models using S4 classes. R package version 0.999375-42. Available vertently compromise the movement corridors of other native from http://CRAN.R-project.org/package=lme4 BERNARD, H., J. FJELDSA, AND M. MOHAMED. 2009. A case study on the effects small mammal species, such as scansorial mice in the genera of disturbance and conversion of tropical lowland rain forest on the Chiropodomys and Haeromys that may exploit lianas. Our data, non-volant small mammals in north Borneo: management implica- nonetheless, provide initial support for the use of liana cutting as tions. Mammal Study 34: 85–96. an important component of enrichment planting practices aimed at BOONSONG, L., AND M. JEFFREY. 1988. Mammals of Thailand. Darnsutha rehabilitating the structure and composition of degraded forests Press, Bangkok, Thailand. BOONSTRA, R., AND I. CRAINE. 1986. Natal nest location and small mammal (Kettle 2010, Ansell et al. 2011). tracking with a spool and line technique. Can. J. Zool. 64: 1034–1036. BOWMAN, J. C., D. SLEEP,G.J.FORBES, AND M. EDWARDS. 2000. The associa- ACKNOWLEDGMENTS tion of small mammals with coarse woody debris at log and stand scales. For. Ecol. Manage. 129: 119–124. We thank the Royal Society’s South East Asia Rainforest Research BROOK, B. W., N. S. SODHI, AND C. J. BRADSHAW. 2008. Synergies among extinction drivers under global change. Trends Ecol. Evol. 23: 453– Programme for logistical support and Yayasan Sabah, Maliau 460. Basin Management Committee, the State Secretary, Sabah Chief BROWN,K.S.,AND G. G. BROWN. 1992. Habitat alteration and species loss in Minister’s Departments, and the Sabah Biodiversity Council for Brazilian forests. In T. C. Whitmore, and J. A. Sayer (Eds.). Tropical permission to conduct research. This work was funded by the deforestation and species extinction, pp. 119–142. Chapman & Hall, Sime Darby Foundation. RME was supported by European New York City. BUESCHING, C. D., C. NEWMAN,R.TWELL, AND D. W. M ACDONALD. 2008. Rea- Research Council Project number 281986. We are grateful for the sons for arboreality in wood mice Apodemus sylvaticus and Bank voles comments of T. Coulson and A. M. Valenzuela. Myodes glareolus. Mamm. Biol. 73: 318–324. CHARLES,J.K.,AND B. B. ANG. 2010. Non-volant small mammal community SUPPORTING INFORMATION responses to fragmentation of kerangas forests in Brunei Darussalam. Biodivers. Conserv. 19: 543–561. CUSACK, J. J. 2011. Characterising Small Mammal Responses to Tropical Forest Additional Supporting Information may be found with online Loss and Degradation in Northern Borneo Using Capture-Mark- material: Recapture Methods. Unpublished master’s thesis, Imperial College London. TABLE S1. Microhabitat variables and substrate categories recorded CUSACK, J. J., O. R. WEARN,H.BERNARD, AND R. M. EWERS. 2014. Influence along spool tracks. of microhabitat structure and disturbance on detection of native and non-native murids in logged and unlogged forests of northern Borneo. J. Trop. Ecol. 31: 25–35. LITERATURE CITED EWERS, R. M., R. K. DIDHAM,L.FAHRIG,G.FERRAZ,A.HECTOR,R.D.HOLT, AND E. C. TURNER. 2011. A large-scale forest fragmentation experi- ADLER,G.,AND A. N. D. R. LEVINS. 1994. The Island Syndrome in rodent ment: The Stability of Altered Forest Ecosystems Project. Philos. populations. Q. Rev. Biol. 69: 473–490. Trans. R. Soc. B Biol. Sci. 366: 3292–3302. AMARASEKARE, P. 1993. Potential impact of mammalian nest predators on FUKAMI, T., D. A. WARDLE,P.J.BELLINGHAM,C.P.H.MULDER,D.R.TOWNS, endemic forest birds of western Mauna Kea, Hawaii. Conserv. Biol. 7: G. W. YEATES,K.I.BONNER,M.S.DURETT,M.N.GRANT-HOFFMAN, 316–324. AND W. M. WILLIAMSON. 2006. Above- and below-ground impacts of AMORI,G.,AND M. N. CLOUT. 2003. Rodents on islands: A conservation chal- introduced predators in seabird-dominated island ecosystems. Ecol. lenge. In G. R. Singleton, C. A. Hinds, C. J. Krebs, and D. M. Spratt Lett. 9: 1299–1307. (Eds). Rats, mice and people: Rodent biology and management. GAVEAU, D. L. A., S. SLOAN,E.MOLIDENA,H.YAEN,D.SHEIL,N.K.ABRAM, ACIAR Monograph 96, pp. 63–68. Australian Centre for International et al. 2014. Four decades of forest persistence, clearance and logging on Agricultural Research, Canberra. Borneo. PLoS ONE 9: e101654. doi:10.1371/journal.pone.0101654 ANSELL, F., D. EDWARDS, AND K. HAMER. 2011. Rehabilitation of logged rain GERWING,J.J.,AND C. UHL. 2002. Pre-logging liana cutting reduces liana forests: Avifaunal composition, habitat structure, and implications for regeneration in logging gaps in the eastern Brazilian Amazon. Ecol. biodiversity-friendly REDD+. Biotropica 43: 504–511. Appl. 12: 1642–1651. APLIN, K. P., P. R. BROWN,J.JACOB,C.J.KREBS, AND G. R. SINGLETON. 2003. Field GIBSON, L., A. J. LYNAM,C.J.BRADSHAW,F.HE,D.P.BICKFORD,D.S.WOO- methods for rodent studies in Asia and the Indo-Pacific. Monographs. DRUFF,S.BUMRUNGSRI, AND W. F. L AURANCE. 2013. Near-complete APLIN, K. P., H. SUZUKI,A.A.CHINEN,R.T.CHESSER,J.TEN HAVE,S.C. extinction of native small mammal fauna 25 years after forest frag- DONNELLAN, AND A. COOPER. 2011. Multiple geographic origins of mentation. Science 341: 1508–1510. commensalism and complex dispersal history of black rats. PLoS GRASSMAN, L., M. E. TEWES,J.N.SILVY, AND K. KREETIYUTANOUT. 2005. Ecol- ONE 6: e26357. ogy of three sympatric felids in a mixed evergreen forest in north- ASQUITH, N. M., S. J. WRIGHT, AND M. J. CLAUSS. 1997. Does mammal com- central Thailand. J. Mammal. 86: 29–38. munity composition control recruitment in neotropical forests? Evi- HANSEN, A. J., R. P. NEILSON,V.H.DALE,C.H.FLATHER,L.R.IVERSON,D. dence from Panama. Ecology 78: 941–946. J. CURRIE, AND P. J. B ARTLEIN. 2001. Global change in forests: 380 Loveridge, Wearn, Vieira, Bernard, and Ewers

Responses of species, communities, and biomes. Bioscience 51: PUTTKER, T., R. PARDINI,Y.MEYER-LUCHT, AND S. SOMMER. 2008. Responses of 765–779. five small mammal species to micro-scale variations in vegetation struc- HARRIS,D.B.,S.D.GREGORY, AND D. W. MACDONALD. 2006. Space invaders? ture in secondary Atlantic Forest remnants, Brazil. BMC Ecol. 8: 9. A search for patterns underlying the coexistence of alien black rats PUTZ, F. E., P. SIST,T.FREDERICKSEN, AND D. DYKSTRA. 2008. Forest ecology and Galapagos rice rats. Oecologia 149: 276–288. and management reduced-impact logging: Challenges and opportuni- IUCN. 2014. The IUCN Red List of Threatened Species. Version 2014.1 ties. For. Ecol. Manage. 256: 1427–1433. Available from http://www.iucnredlist.org RDEVELOPMENT CORE TEAM. 2014. R: A language and environment for sta- KETTLE, C. J. 2010. Ecological considerations for using dipterocarps for tistical computing. R Foundation for Statistical Computing, Vienna. restoration of lowland rainforest in Southeast Asia. Biodivers. Conserv. Available from: http://www.R.project.org/ 19: 1137–1151. REYNOLDS, G., J. PAYNE,W.SINUN,G.MOSIGIL, AND R. P. D. WALSH. 2011. KEY,G.E.,AND R. D. WOODS. 1996. Spool and line studies on the beha- Changes in forest land use and management in Sabah, Malaysian Bor- vioural ecology of rats (Rattus spp) in the Galapagos Islands. Can. J. neo, 1990–2010, with a focus on the Danum Valley region. Philos. Zool. 74: 733–737. Trans. R. Soc. B Biol. Sci. 366: 3168–3176. KOT, M., M. A. LEWIS, AND P. VAN DEN DRIESSCHE. 1996. Dispersal data and SHADBOLT,A.B.,AND R. RAGAI. 2010. Effects of habitat fragmentation on the the spread of invading organisms. Ecology 77: 2027–2042. movement patterns and dispersal ability of the brown spiny rat (Max- KOTLER, B. P., J. S. BROWN, AND O. HASSON. 1991. Factors affecting gerbil for- omys rajah) in the Planted Forest Zone of Sarawak, Eastern Malaysia. aging behavior and rates of owl predation. Ecology 72: 2249–2260. Biodivers. Conserv. 19: 531–541. MANLY,B.J.F.,J.L.MCDONALD,D.THOMAS,T.L.MCDONALD, AND W. P. L . SIKES,R.S.,AND W. L. GANNON. 2011. Guidelines of the American Society of ERICKSON. 2002. Resource selection by : Statistical design and Mammalogists for the use of wild mammals in research. J. Mammal. analysis for field studies. 2nd edn. Kluwer Academic Publishers, Dor- 92: 235–253. dretch, the Netherlands. STOKES,V.L.,P.B.BANKS,R.P.PECH, AND R. L. WILLIAMS. 2009. Inva- MARGONO, B. A., S. TURUBANOVA,I.ZHURAVLEVA,P.POTAPOV,A.TYUKAVINA, sion by Rattus rattus into native coastal forests of south-eastern A. BACCINI, ... &M.C.HANSEN. 2012. Mapping and monitoring Australia: Are native small mammals at risk? Austral Ecol. 34: deforestation and forest degradation in Sumatra (Indonesia) using 395–408. Landsat time series data sets from 1990 to 2010. Environ. Res. Lett. VICKERY,W.,AND J. BIDER. 1981. The influence of weather on rodent activity. 7: 034010. J. Mammal. 62: 140–145. MEIJARD, E., AND D. S HEIL. 2007. A logged forest in Borneo is better than WARDLE, D. A., P. BELLINGHAM,T.FUKAMI, AND C. P. H. MULDER. 2007. Pro- none at all. Nature 446: 974. motion of ecosystem carbon sequestration by invasive predators. Biol. MILES, M. A., A. D. SOUZA, AND M. M. POVOA. 1981. Mammal tracking and Lett. 3: 479–482. nest location in Brazilian forest with an improved spool-and-line WELLS, K., AND R. BAGCHI. 2005. Eat in or take away - Seed predation and device. J. Zool. 195: 331–347. removal by rats (muridae) during a fruiting event in a dipterocarp rain- MONTGOMERY,W.I.,AND J. GURNELL. 1985. The behaviour of Apodemus. forest. Raffles Bull. Zool. 53: 281–286. Symp. Zool. Soc. Lond. 55: 89–115. WELLS, K., M. PFEIFFER,M.B.LAKIM, AND E. K. V. KALKO. 2006. Movement NAKAGAWA, M., H. MIGUCHI,K.SATO, AND T. NAKASHIZUKA. 2007. A prelimi- trajectories and habitat partitioning of small mammals in logged and nary study of two sympatric maxomys rats in Sarawak, Malaysia: Spac- unlogged rain forests on Borneo. J. Anim. Ecol. 75: 1212–1223. ing patterns and population dynamics. Raffles Bull. Zool. 55: WELLS, K., R. T. CORLETT,M.B.LAKIM,E.K.V.KALKO, AND M. PFEIFFER. 381–387. 2009. Seed consumption by small mammals from Borneo. J. Trop. OKSANEN, J., F. GUILLAUME,B.ROELAND,K.P.LEGENDRE,P.R.MINCHIN, Ecol. 25: 555–558. R. B. O’HARA,G.SIMPSON,P.SOLYMOS,S.STEVENS, AND H. WAGNER. WELLS, K., E. K. V. KALKO,M.B.LAKIM, AND M. PFEIFFER. 2007. Effects 2011. vegan: Community Ecology Package. R. package version 2.0-2. of rain forest logging on species richness and assemblage composi- Available from http://CRAN.R-project.org/package=vegan tion of small mammals in Southeast Asia. J. Biogeogr. 34: OLDEN,J.D.,N.LEROY POFF,M.R.DOUGLAS,M.E.DOUGLAS, AND K. D. 1087–1099. FAUSCH. 2004. Ecological and evolutionary consequences of biotic WELLS, K., D. B. KOCK,M.B.LAKIM, AND M. PFEIFFER. 2006. Is Rattus rattus homogenization. Trends Ecol. Evol. 19: 18–24. invading the primary rainforest on Borneo? Malayan Nat. J. 59: PAGES , M., Y. CHAVAL,V.HERBRETEAU,S.WAENGSOTHORN,J.F.COSSON,J.P. 73–79. HUGOT, AND J. MICHAUX. 2010. Revisiting the of the Rattini WELLS, K., M. B. LAKIM, AND R. B. O’HARA. 2014a. Shifts from native to tribe: A phylogeny-based delimitation of species boundaries. BMC invasive small mammals across gradients from tropical forest to urban Evol. Biol. 10: 184. habitat in Borneo. Biodivers. Conserv. 23: 2289–2303. PARDINI, R. 2004. Effects of forest fragmentation on small mammals in an WELLS, K., R. B. O’HARA,S.MORAND,J.P.LESSARD, AND A. RIBAS. 2014b. Atlantic Forest landscape. Biodivers. Conserv. 13: 2567–2586. The importance of parasite geography and spillover effects for global PAYNE,J.,AND C. M. FRANCIS. 1985. A field guide to the mammals of Borneo. patterns of host–parasite associations in two invasive species. Divers. Sabah Society, Kota Kinabalu, Malaysia. Distrib. 21: 477–486. PFEIFER, M., V. LEFEBVRE,E.TURNER,J.CUSACK,M.KHOO,V.K.CHEY, WELLS, K., M. PFEIFFER,M.B.LAKIM, AND K. E. LINSENMAIR. 2004. Use M. PENI, AND R. M. EWERS. 2015. Deadwood biomass: An underesti- of arboreal and terrestrial space by a small mammal community mated carbon stock in degraded tropical forests? Environ. Res. Lett. in a tropical rain forest in Borneo, Malaysia. J. Biogeogr. 31: 10: 044019. 641–652. PINHEIRO, J., D. BATES,S.DEBROY,D.SARKAR AND THE RDEVELOPMENT CORE WHITNEY,K.D.,AND C. A. GABLER. 2008. Rapid evolution in introduced spe- TEAM. 2011. nlme: Linear and Nonlinear Mixed Effects Models. cies, “invasive traits” and recipient communities: Challenges for pre- R package version 3.1-102. Available from http://cran.r-project.org/ dicting invasive potential. Divers. Distrib. 14: 569–580. web/packages/nlme/index.html WILTING, A., F. FISCHER,S.A.BAKAR, AND K. E. LINSENMAIR. 2006. Clouded PUAN, C., A. W. GOLDIZEN,M.ZAKARIA,M.N.HAFIDZI, AND G. S. BAXTER. leopards, the secretive top-carnivore of South-East Asian rainforests: 2011. Absence of differential predation on rats by Malaysian barn owls their distribution, status and conservation needs in Sabah, Malaysia. in oil palm plantations. J. Raptor Res. 45: 71–78. BMC Ecol. 6: 16.