Introduction

Impacts of forest harvest operations on bats (Chiroptera) have been considered detrimental, beneficial or negligible depending on whether sufficient roost structures have been retained (Miller et al. 2003). However, most researchers suggest that harvest operations have a negative impact on habitat quality due to reductions in roost availability (Miller et al. 2003; Russo et al. 2010). The level of roost disturbance and loss is likely to be high in plantation forests compared to other areas because of repeated harvest operations, but this has not been determined empirically (Borkin and Parsons 2010).

When roosts are lost during harvest operations, bat populations may be affected either directly as a result of injury or death (Borkin and Parsons 2010), or indirectly due to loss of available roosting habitat (Jenkins et al. 2007). Several bat species use small roosting areas and few roosts (Carter and Feldhamer 2005; Kapfer et al. 2007; Nixon et al. 2009), which may be located entirely within clear-fell harvested areas. The loss of specific roosts may result in increased energetic costs to bats associated with finding new roosts (Arnold 2007), and the forced use of roosts of poor quality (Sedgeley 2001; Sedgeley and O’Donnell 2004), both of which may reduce survival and reproductive success (Brigham and Fenton 1986; O’Donnell and Sedgeley 2006; Sedgeley 2001; Sedgeley and O’Donnell 2004). Even small declines in bat survival can result in potentially catastrophic population declines (Pryde et al. 2006), especially in threatened endemic island species (O’Donnell 2010; O’Donnell et al. 2010). When rates of either natural or forestry-related roost loss are high, colony sizes may change because of direct losses of bats in harvest operations (Borkin and Parsons 2010) or as the rate of natural replacement of new roosts may not keep up with demand, and bats may have to move into new areas to maintain an array of suitable roosts (Carter and Feldhamer 2005). These movements are likely to be rare because many bat species maintain social groups that have geographically distinct roosting areas (Hillen et al. 2010; Kapfer et al. 2007; Willis and Brigham 2004) and rarely use the roosts of other social groups (O’Donnell 2000a). Thus, colony sizes will likely decline with harvest operations. Therefore, it is important to identify roosting area characteristics, relative rates of roost loss due to forestry operations and natural causes, and replacement rates to understand whether or not net losses are high enough to be of concern, and whether mitigation steps should be taken to improve survival rates. However, studies comparing areas before and after clear-fell harvest are absent from the peer-reviewed literature.

Roost disturbance, logging and habitat loss have been identified as possible causes of decline of New Zealand’s threatened long-tailed bat, Chalinolobus tuberculatus (O’Donnell 2000b; Sedgeley and O’Donnell 2004), which has recently been confirmed as resident within exotic plantation forests throughout New Zealand (Borkin and Parsons 2010). Plantation forests within New Zealand are typically managed on a 26–32 year cycle of clear-fell harvest (Ministry of Agriculture and Fisheries 2010). C. tuberculatus tend to roost in trees close to harvest age in plantation forests (authors’ unpublished data), and are present within trees during harvest operations, potentially resulting in bats being trapped or injured (Borkin and Parsons 2010; Daniel 1981). Trees similar to the those selected as roosts by C. tuberculatus, long dead spars with large sheets of peeling bark, have high rates of natural loss (Carter and Feldhamer 2005), although these rates are likely less than those that occur during clear-fell harvest because of the destruction of multiple roosts simultaneously.

The aim of this research was to investigate the impact of clear-fell harvest on bats by determining patterns of colony size, roosting range size, roost use and roost loss over three summers, using C. tuberculatus as a model species. We quantified roosting ranges and numbers of roosts used by individual bats after harvest operations (post-harvest, P.H.), and in areas where harvest had not occurred (no harvest, N.H.). Over three summers yearly rates of natural and forest-management induced roost loss were also investigated, and colony size was determined using counts of bats emerging from roosts at dusk.

Chalinolobus tuberculatus maintain social groups with traditional roosting areas. They rarely use the roosts of other social groups or change their roosting areas, even when roosts are destroyed (O’Donnell 2000a; O’Donnell and Sedgeley 2006). Thus, we predicted that post-harvest: (1) roosting range sizes would be smaller as suitable habitat for foraging would be lost, (2) fewer roosts would be used due to preferred trees being preferentially harvested, (3) colony size would be smaller due to direct or indirect mortality or as a result of individual bats moving to new areas when roosts were felled.

Materials and methods

Roosts and their use were investigated over three field seasons during the southern hemisphere summer (Season 1 October 2006–March 2007; Season 2 November 2007–March 2008; and Season 3 November 2008–March 2009) in Kinleith Forest, a mainly Pinus radiata plantation forest, South Waikato, New Zealand (37° 47′ S, 175° 53′ E). Bats were captured using mistnets placed across forest roads, or in harp traps or hand nets when emerging from roosts. All captured bats were ringed on the forearm using an individually numbered 2.9 mm band (The Mammal Society, UK). Bats were sexed, and their reproductive status assessed as: never bred; pregnant or lactating female; or adult male. Juvenile bats were identified by unfused epiphyseal phalanges, indicating they were born during the summer of capture and had not yet bred (O’Donnell 2005). Juvenile bats alter their habitat use as they age, meaning differences between males and females are likely to be due to age and not sex at this stage of development (O’Donnell 2001; Racey and Swift 1985). Therefore, juvenile male and female bats were considered to be the same (O’Donnell 2001; Racey and Swift 1985). Pregnant bats were identified by palpation of the abdomen. Enlarged nipples surrounded by a bare patch of skin identified lactating bats. Bats were weighed to measure mass using a pesola scale. Radio-transmitters (Model BD-2, Holohil, Canada) weighed 0.48 g and mean transmitter load was 4.71% (range 3.56–5.65%) of the body mass, conforming to the 5% guideline suggested by Aldridge and Brigham (1988). Radio-transmitters were attached to bats just behind the shoulder blades using ADOS F2 contact adhesive (CRC Industries New Zealand, Auckland, New Zealand) after a small area of hair was clipped to ensure maximum adhesion. Bats were released the same night and in the same location as they were captured. Bats were radio-tracked to their roosts daily using a Yagi antennae and TR4 receiver (Telonics Inc., Arizona, USA), until either their transmitter was no longer detected in the study area (after at least three consecutive days of searching), it was found on the ground, or the transmitter signal persisted for at least 3 days from a roost or other location where it was unable to be retrieved. Exact locations of roosts were determined by either observing bat departure at dusk, or by location of the strongest signal from the radio-transmitter. Additional roosts were found during a search for fecal deposits (n = 1), during observations of emerging bats from suspected roosts (n = 1), and when a P. radiata was felled and two C. tuberculatus fell out (n = 1).

Bats were categorized in relation to harvest history post-radiotracking because the area bats used could not be predicted prior to tracking. Night-time home ranges elucidated in an earlier study were used to categorize bats (authors’ unpublished data). Bats were considered ‘N.H.’ if no clear-fell harvest operations had taken place in the forest stands adjacent to the outer limit (100% minimum convex polygon (MCP)) of their home range for at least 1 year when radiotracking took place. Bats were considered ‘P.H.’ if radiotracking took place within the year after a harvest operation had taken place in the stands adjacent to their home range. Different bats were radiotracked in N.H. and P.H. groups. Roosts were considered N.H. or P.H. depending on the home range of the first bat that was radiotracked to the roost.

Roosting range (ha) was defined as the area containing all identified roosts used by an individual bat during the period its radiotransmitter was attached. The roosting range span (m) was the furthest distance from one edge of the 100% MCP ha to the other. Roosting range and roosting range span were calculated by entering daily roost locations into Ranges 6 v1.217 (Anatrack Ltd, Wareham, UK). Minimum convex polygon (100% MCPs) roosting range calculated to compare with other studies on C. tuberculatus (Harris et al. 1990) did not differ from 95% MCPs (Kendall’s W = 0.0005, χ2(1) = 0.08, P = 1.0). Use of MCPs is the most robust technique when the number of fixes is low, and is the only technique that is strictly comparable between studies (Harris et al. 1990). Visual inspection of graphs produced in incremental analyses were used to determine whether asymptotes were approached for the relationship between individual bat 100% MCP sizes and the number of datapoints (Kenward et al. 2003). Only fully revealed ranges were used in these analyses. Core areas (areas of particularly high use) were determined by visual inspection of graphs of individual bat resource use distributions for sharp discontinuities in area after elimination of outlying locations (Harris et al. 1990; Kenward et al. 2003). The mean percentage of total roosting area corresponding to these discontinuities was calculated and the nearest conventional percentage MCP used to represent bats’ core areas. Consequently, 85% MCPs were used to represent bats’ core ha (mean = 84.74 ± 2.15). However, 85 and 100% MCPs did not vary significantly so only 100% MCPs are reported (Kendall’s W = 0.24, χ2 (1) = 3.77, P = 0.09). The relationship between roosting range size and number of observations, and range span and number of observations were investigated using Spearman’s rank correlation. Roost occupancy (roost fidelity) was determined by dividing the number of days each bat was tracked to a roost by the number of roosts that were found over this period, and is reported as days roost−1.

Percentage roost loss was determined each year by checking the status of all known roosts at the beginning of summer (October). Roosts were considered lost if: (a) there was no longer any trace of the roost (b) the roost had fallen (c) the roost was crushed, or (d) all bark had fallen from the roost if the roost was previously under peeling bark. Causes of roost loss were classified as: (a) harvesting and other forestry-related operations, or (b) natural causes i.e., no obvious forestry operation had taken place and the roost was still present in a position that made it likely that it had fallen naturally. Associations between the specific year and whether roost loss was due to forestry operations or natural causes were investigated using the likelihood ratio χ2 statistic for small sample sizes.

Number of bats present in each roost was determined by counting as they emerged at dusk. Emergence was observed from the ground at a distance that minimized disturbance but still allowed observation of the bats’ departure. When possible, a position was chosen that highlighted emerging bats against the sky. A combination of visual observations and bat detectors were used to identify the number of individuals emerging. Observations began ca. 15 min before sunset and continued until 10 min had passed since the emergence of the previous bat (following Griffiths 2007). Roosts were categorized as colonial if more than one bat was present at emergence. Median colony size for all roosts was calculated over all field seasons.

One-way ANOVA were used to assess differences unless parametric assumptions of normality and equal variances were violated. Consequently, Mann–Whitney tests were used to compare roosting range sizes, spans, numbers of roosts used, roost occupancy times and colony size between N.H. and P.H. Kruskal–Wallis tests were used to determine differences in roosting range sizes and spans between field seasons. Jonckheere-Terpstra’s tests were used to determine trends in the data over the three field seasons. One-tailed tests were adopted when there were a priori predictions in one direction. Untransformed data are reported. Tukey’s hinges were used to calculate percentiles for interquartile ranges (IQR). A t-test was used to investigate differences between the number of days male and female bats were radiotracked. Effect sizes were calculated using Pearson’s product-moment correlation coefficient, r. This is an objective standardised measure of the magnitude of the observed effect, and can be used to suggest importance of effects in practical terms (Field 2005). Cohen’s (1992) guidelines for what constitutes a small or large effect on a population were used, such that a medium effect size (r = 0.3) represents an effect likely to be visible to careful observers (Cohen 1992). A large effect is represented by r ≥ 0.5.

Results

Data integrity

Seventeen bats were radiotracked to a median 2.0 roosts each (IQR = 1.8–4.0, range = 1–6) over the 3 years of the study. These bats represented a range of age and sex classes: three juvenile females, two juvenile males, three pregnant females, three lactating females, and seven adult males (Table 1). A fully revealed roosting range was not achieved for one juvenile female classed N.H. and radiotracked in season 1 so she was not included in roosting range analyses. On average, bats were radiotracked for a median of 5.0 days (IQR = 3.0–11.0, max 19) attaining a total of 131 radiotracking roost-days. No differences were found in the number of days over which bats of different reproductive states were radiotracked to roosts (F (3,14) = 0.67, P = 0.59, r = 0.35). There was no difference in the number of days over which males and females were radiotracked to roosts (t (16) = −0.18, P (2-tailed) = 0.86, r = 0.04). No difference was found in the number of days over which bats were radiotracked between N.H. and P.H. groups (F (1,16) = 0.43, P = 0.52, r = 0.16) so comparisons of N.H. and P.H. roosting range and roost use data were possible. On average, bats in different reproductive states were radiotracked to a similar number of roosts (H(3) = 2.21, P = 0.53), so pooling data across reproductive states was possible for comparisons between N.H. and P.H.. Juvenile bats were radiotracked to the highest number of roosts (median = 4.0, IQR = 1.5–4.0). Pregnant (median = 2.0, IQR = 1.5–2.0), lactating (median = 2.0, IQR = 2.0–4.0) and adult male bats (median = 2.0, IQR = 1.0–3.0) were radiotracked to similar numbers of roosts.

Table 1 Long-tailed bat captures in areas that had not been harvested (N.H.) and post-harvest (P.H.) and within each season
Full size table

Roosting ranges, spans and roost occupancy

As the area of roosting range (100% MCP) increased, span increased (r = 0.88, P (1-tailed) <0.0001, n = 16). There were no positive relationships between the number of observations and the size of the roosting area (100% MCP: r = 0.31, P (1-tailed) = 0.13), or the number of observations and range span (r = 0.20, P (1-tailed) = 0.23). Roosting range size (100% MCP) did not vary between bats of different reproductive states (100% MCP: H (3) = 4.14, P = 0.62; Table 2). Roosting range span did vary between bats of different reproductive states (H (3) = 7.49, P = 0.04, Table 2). Roost occupancy times did not differ between bats of different reproductive states (H (3) = 2.76, P = 0.46, Table 2). Therefore, pooling of range size and roost occupancy data was possible for comparisons between N.H. and P.H. and between field seasons, but this was not possible for span data.

Table 2 Long-tailed bat roosting range sizes and spans for bats in different reproductive states (juvenile, pregnant or lactating female, and adult male)
Full size table

P.H. bats used fewer roosts on average than N.H. bats (U = 16.0, P (1-tailed) = 0.02, N = 17, r = −0.50). N.H. bats used a median of 4.0 roosts each (IQR = 2.0–4.5, max 6, n = 8 bats), whereas P.H. bats used a median 2.0 roosts each (IQR = 2.0–2.0, max 3, n = 9 bats).

Roost occupancy times did not differ between N.H. and P.H. bats (U = 36.0, Z = −0.36, N = 18, P (2-tailed) = 0.75, r = −0.08). On average, bats spent 2.5 days in each roost before moving to a new roost (IQR = 1.5–3.1). N.H. bats moved roosts every 2.3 days (IQR = 1.5–3.2); while P.H. bats moved every 2.5 days (IQR = 1.5–3.0 days).

P.H. roosting ranges were smaller than N.H. roosting ranges (100% MCP: U = 14.5, Z = −1.81, P (1-tailed) = 0.04, N = 16, r = −0.45), being 0.01 ha (IQR = 0.0001–0.3 ha) and 5.7 ha (IQR = 0.07–11.61 ha), respectively.

There were no differences in range sizes amongst field seasons (H (2) = 3.22, P = 0.21). However, over the three field seasons, the size of bats’ roosting ranges decreased (100% MCP: J = 20.5, z = −1.9, P (1-tailed) = 0.03, r = −0.47). In season 1 100% MCP roosting ranges were 6.2 ha (IQR = 0.1–17.1 ha), in season 2 they were 0.1 ha (IQR = 0.0005–0.3 ha), and in season 3 they were 0.0001 ha (IQR = 0.0001–0.08 ha).

Roost loss

On average, each year 20.7% of all known roosts were lost; 14.5% due to forestry operations and 6.2% due to natural tree fall (Table 3). Forestry operations were responsible for 72.2% of the lost roosts; the remaining 27.8% were lost due to natural causes.

Table 3 Contingency table of the association between roost loss and whether loss was due to forestry operations or natural causes
Full size table

There was an association between the year and whether roost loss was due to forestry operations or natural causes (Lχ2 (2) = 9.77, P = 0.02, Cramer’s V = 0.73, P (exact) = 0.02, n = 18, Table 3). Roost loss was highest in the year between field seasons 2 and 3, when ten roosts were lost, nine due to forestry operations (Table 3).

Colony size

We completed 27 emergence counts when one or more bats emerged from the watched roost in Season 1, 43 in Season 2, and 68 in Season 3. Following these counts we classified 16 roosts as colonial in Season 1, 13 in Season 2, and 18 in Season 3. Median colony size was 4.0 bats (IQR = 2.0–8.0, maximum = 20, n = 42 roost emergence counts of colonies). There were differences in colony size between seasons (H (2) = 20.09, P < 0.001), and over the three seasons colony size decreased (J = 120.0, z = −3.9, P (1-tailed) <0.001, n = 42, r = −0.61). Colony size was greatest in season 1 and least in season 3 (Colony sizes: Season 1: 8.0 bats (IQR = 3.5–10.5), n = 16 roost emergence counts; Season 2: 6.0 bats (IQR = 4.0–8.0), n = 13 roost emergence counts; Season 3: 2.0 bats (IQR = 2.0–2.0), n = 13 roost emergence counts).

P.H. colonies were smaller than N.H. colonies (U = 59.0, Z = −3.1, P (2-tailed) = 0.001, N = 35, r = −0.52). N.H. median colony size was 10.0 bats (IQR = 4.0–11.0, maximum = 20, n = 17 roost emergence counts); P.H. median colony size was 4.0 bats (IQR = 2.0–6.0, maximum = 8, n = 18 roost emergence counts).

Discussion

Roost loss due to forestry operations added considerably to relatively high natural rates of roost loss. Therefore, it is not surprising that following harvest operations roosting range size was smaller, fewer roosts were used, and colony sizes were smaller post-harvest. Within plantation forest, it appears as though natural rates of roost loss are higher than in other habitats. For example, roosts in Kinleith Forest were lost at rates 38% higher than those in found in previous research in an agricultural landscape (Sedgeley and O’Donnell 2004). Thus, even without harvest operations bats living in plantation forest are threatened by roost loss.

The data suggest a pattern of declining numbers of roosts used, roosting area and colony sizes coinciding with clear-fell harvest operations. After harvest operations bats used a smaller pool of roosts than those bats in areas where harvest had not occurred recently. Bats in both areas moved among roosts at the same rate. The results of this study are useful when viewed in the context of the absence of similar studies in the peer-reviewed literature (Hayes and Loeb 2007; Russo et al. 2010). A number of factors precluded large sample sizes in this study, and in the case of endangered species smaller sample sizes may be satisfactory because it is very important to identify possible declines and act on them at the earliest time possible (O’Donnell and Langton 2003). In our case, we suggest the study site itself was a major factor influencing capture rates. Kinleith Forest may represent a sink (Pulliam 1988) or pseudo-sink habitat (Watkinson and Sutherland 1995), and population size may be limited by numbers of roosting structures (Sedgeley and O’Donnell 1999a), because over time colony sizes declined, were small, and few roosting structures were used compared to colonies studied in other habitats (Borkin 2010; O’Donnell and Sedgeley 1999; Pulliam 1988; Sedgeley and O’Donnell 2004).

The roosting areas we found were far smaller than those found for C. tuberculatus elsewhere (O’Donnell 2001), presumably reflecting the rarity of roosts within plantations compared to other habitat types (Borkin 2010; Ciechanowski 2005). For example, the median roosting range of adult males was 0.07 ha, far smaller than the 84 ha of adult males from native forest in Fiordland (O’Donnell 2001). Plantation forests are well known as areas with low roost abundance as shown by studies in Poland (Ciechanowski 2005) and Costa Rica (Chaverri et al. 2007). Artibeus watsoni show a similar relationship; in areas of low roost abundance, such as plantation forests, this species has smaller roosting ranges (Chaverri et al. 2007).

Roosting ranges were smaller and bats used fewer roosts post-harvest operations, indicating suitable roosts may be less abundant post-harvest and each individual bat’s pool of roosts is reduced. Certainly, individuals of other bat species display smaller roosting ranges in areas with low roost abundance than those in areas where roosts are plentiful (Chaverri et al. 2007); this suggests our assumption that roosts are less abundant post-harvest is valid. C. tuberculatus are likely to use fewer roosts post-harvest operations, and not seek new roosts in other areas or with other social groups (O’Donnell 2000b) unless forced to move to new areas because of extremely low roost availability and slow rates of roost creation (Carter and Feldhamer 2005). This is because C. tuberculatus maintain social groups that use the same roosting areas over several years, and rarely move into new areas or use the roosting areas of other social groups, even when roost availability declines due to tree loss (O’Donnell 2000b).

The use of fewer roosts by bats post-harvest operations suggests an individual’s pool of available roosts declines with harvest and consequently they have reduced opportunities to choose roosts that optimize energy savings. Bats save energy either by choosing roosts close to profitable foraging grounds (Boland et al. 2009; Encarnação et al. 2005) or that buffer ambient temperature allowing them to either use torpor or maintain the high temperatures required for fetal growth and milk production (Hamilton and Barclay 1994; Hood et al. 2002; Racey and Swift 1981; Russo et al. 2004; Speakman and Thomas 2003; Wilde et al. 1999). An inability to maximize energetic savings can affect fetal and post-natal growth rates (Allen et al. 2010; Sedgeley 2001), and impact on the survival of adults and juveniles, their reproductive fitness and the long-term viability of the population (O’Donnell 2002; O’Donnell and Sedgeley 2006; Sedgeley 2001; Sedgeley and O’Donnell 2004).

The rate of roost loss due to natural causes reported in this study is most likely to be an underestimate of actual roost loss. Roosts may also be abandoned because of declines in suitability due to bark loss (Barclay and Brigham 2001; Carter and Feldhamer 2005) that results in decreased insulation (Nicolai 1986), or their use by predators (Sedgeley and O’Donnell 1999b) or competitors (Barclay and Brigham 2001). Such changes could mean roosts are still standing but no longer functional, but this may remain undetectable (Barclay and Brigham 2001) and occur frequently (Carter and Feldhamer 2005). We expect, therefore, that rates of natural loss are higher again than we report.

Colony sizes reduced over the three summers of study and were smaller after harvest operations had occurred. These changes coincided with a reduction in numbers of known roosts, and the suitability of two other roosts, in the area where the majority of bats were captured and where most of the known communal roosts were located. Eleven known communal roosts were lost between the end of Season 1 and the start of Season 3. A loss of roosts on this scale will impact the local bat population, forcing them to move to new areas (Carter and Feldhamer 2005) if they were not initially injured or trapped when their roost was felled (Borkin and Parsons 2010; Daniel 1981). The loss of multiple roost trees has led to population level declines in other long-tailed bat populations. The number of female bats within one social group dropped dramatically over 2 years of one study, probably due to bat mortality following wind-throw of a patch of trees known to contain roosts (O’Donnell 2000a) or predation (Pryde et al. 2005). A similar decline has occurred in at least one area of Kinleith Forest, and may occur regularly as harvest occurs.

The reduction in colony size with harvest operations, and over time, suggests local bat populations may be ephemeral within Kinleith Forest. It is likely bats continue to use traditional roosting areas and individual roosts, as in native forest (O’Donnell 2000a), until they are lost due to forestry operations or natural causes. C. tuberculatus tend to form socially isolated roosting groups with small exclusive roosting areas, and therefore if a group’s roosting area was largely destroyed it may have limited opportunities to integrate with members of a nearby area (O’Donnell 2000a; Sedgeley and O’Donnell 2004). Such isolation increases their likelihood of local extinction (Theodorou and Couvet 2006).

Plantation forests provide large areas of additional habitat for bats (Borkin and Parsons 2010). However, the impacts of clear-fell harvest and consequent roost loss may mean that plantations are home to bats with low reproductive success (Arnold 2007; Brigham and Fenton 1986; Sedgeley and O’Donnell 2004) that are resident in small colonies. The small colony sizes within plantation forests indicate populations may be more vulnerable to local extinction than those in other habitats where colony sizes are larger (O’Donnell 2002; Theodorou and Couvet 2006). We found fewer roosts and smaller roosting ranges were used after clear-fell harvest operations, indicating roosting habitat is reduced, and smaller colony sizes occur post-harvest. Together, these results suggest that clear-fell harvest operations are negative for C. tuberculatus. However, the presence of C. tuberculatus in this plantation forest for over 30 years (Borkin and Parsons 2009; Daniel 1981), and their occurrence in many other plantations throughout New Zealand (Borkin and Parsons 2010) suggests that bats may have strategies to cope with such impediments. Whether the persistence of C. tuberculatus within Kinleith Forest indicates a self-sustaining population is unknown. Certainly, these results suggest that areas of Kinleith Forest that are undergoing clear-fell harvest may represent sink (Pulliam 1988) or pseudo-sink habitats (Watkinson and Sutherland 1995). However, without data on emigration rates strict classification of these areas would be unwise (Runge et al. 2006). Nevertheless, the survival of such populations may be improved by the creation of reserve areas that are not harvested (Pryde et al. 2006). A primary threat to C. tuberculatus is predation by introduced mammals, such as cats (Felis catus). In native forest intense predator control has been shown to increase survival of New Zealand’s other endemic species of bat, Mystacina tuberculata (O’Donnell et al. 2011), and so is also likely to have positive effects on bat populations in production forest (Pryde et al. 2006). These results should motivate biologists to undertake similar studies to investigate impacts of forest management, so that recommendations to protect bats are justifiable.