Improving the science and practice of using artificial roosts for bats

Worldwide, artificial bat roosts (e.g., bat boxes, bark mimics, bat condos) are routinely deployed for conservation, mitigation, and community engagement. However, scant attention has been paid to developing best practices for the use of artificial roosts as conservation tools. Although bats readily occupy artificial roosts, occupancy and abundance data are misleading indicators of habitat quality. Lacking information on bat behavior, health, and fitness in artificial roosts, their conservation efficacy cannot be adequately validated. We considered the proximal and ultimate factors, such as evolutionarily reliable cues, that may prompt bats to preferentially use and show fidelity to suboptimal artificial roosts even when high‐quality alternatives are available. Possible negative health and fitness consequences for artificial roost inhabitants include exposure to unstable and extreme microclimates in poorly designed roosts, and vulnerability to larger numbers of ectoparasites in longer lasting artificial roosts that house larger bat colonies than in natural roosts. Bats using artificial roosts may have lower survival rates if predators have easy access to roosts placed in conspicuous locations. Bats may be lured into occupying low‐quality habitats if attractive artificial roosts are deployed on polluted urban and agricultural landscapes. To advance the science behind artificial bat roosts, we present testable research hypotheses and suggestions to improve the quality of artificial roosts for bats and decrease risks to occupants. Because continued loss of natural roosts may increase reliance on alternatives, such as artificial roosts, it is imperative that this conservation practice be improved.


INTRODUCTION
Chiropterans (bats) are of high conservation importance but are undergoing global declines due to human-induced threats, such as roost habitat loss (Frick et al., 2020).Tree-roosting bats are particularly vulnerable because trees used as day roosts, maternity sites, and hibernacula do not readily replenish (Manning et al., 2013;Russo et al., 2023;Vesk et al., 2008).To mitigate habitat loss, conservation practitioners increasingly deploy artificial roosts as surrogates for tree roosts (Rueegger, 2016) (hereafter natural roosts).Artificial roosts, intended to mimic tree cavities, are inexpensive and deployed on almost every continent (Rueegger, 2016).Currently, >70 bat species use artificial roosts.Although some species are cave obligates, most are bats that use trees as natural roosts (Mering & Chambers, 2014;Rueegger, 2016).Although the utility of these structures for providing alternative habitat is evident (Brittingham & Williams, 2000;Flaquer et al., 2006), recent work questions their suitability for bats because some designs generate lethal temperatures (Alcalde et al., 2017;Flaquer et al., 2014;Griffiths, 2022;Griffiths, Rowland, et al., 2017;Hoeh et al., 2018;Martin Bideguren et al., 2019;Rueegger, 2019).Past research has overwhelmingly relied on occupancy and abundance data as a measure of conservation success, and there have been few studies of bat fitness in artifical roosts (Cowan et al., 2021).
Recent research has shown that, when applied under specific scenarios, artificial roosts can harm bats and potentially yield ecological traps (Crawford et al., 2022;Flaquer et al., 2014;Griffiths, 2022;Lausen et al., 2023).We posit that many artificial roosts provide suboptimal roosting conditions and are inadequate substitutes for natural roosts.Due to the rising popularity of artificial roosts (Crawford & O'Keefe, 2021a), it is critical that conservation practitioners identify and correct pitfalls associated with their use, which could otherwise harm at-risk species and lead to misplaced conservation efforts globally (Ford et al., 2021).
On natural and human-altered landscapes, an ecological trap scenario occurs when an animal population experiences detec-tion errors when identifying high-quality habitat and instead preferentially selects suboptimal habitat (Battin, 2004;Robertson & Hutto, 2006).Sudden anthropogenic land-use change can alter the intensity of cues used in habitat selection and create novel ecological traps, lowering reproductive success or survival (Hale & Swearer, 2016).To identify ecological traps, researchers must demonstrate that a population prefers habitat types that lower fitness compared with available alternatives (Battin, 2004).Lacking demographic estimates of fitness, an ecological trap can be obscured by higher population densities and sublethal physiological effects (e.g., Coleman & Barclay, 2011).
Although the idea of artificial roosts acting as ecological traps for bats has yet to be rigorously tested, this concept has been investigated extensively in birds, with some studies documenting evidence of ecological traps (e.g., Courtois et al., 2021;Demeyrier et al., 2016;Klein et al., 2007;Mänd et al., 2005;Olson, 2017) and some not (e.g., Schwartz et al., 2020).Important behavioral and phenological distinctions exist between avian use of nest boxes and bat use of artificial roosts.Bats typically do not leave roosts during the day, likely due to a high perceived risk of predation associated with daytime flight (Lima & O'Keefe, 2013).Adult and fledgling birds may leave nest boxes any time of day and do so regularly when foraging during daylight hours (e.g., Stanton et al., 2016).Further, bats generally give birth to 1−2 pups and, in the temperate zone, are typically restricted to 1 reproductive event per year (Barclay & Harder, 2003), whereas many bird species can lay an additional clutch after nest failure.Social structure is another key aspect that differentiates bird and bat use of artificial roosts.Birds typically form breeding pairs, resulting in nest boxes being occupied by the pair and their young (e.g., Demeyrier et al., 2016).In contrast, bats in artificial roosts form large maternity colonies composed of tens to thousands of females and their pups in a single structure (e.g., Brittingham & Williams, 2000;Lausen et al., 2023;Lourenço & Palmeirim, 2004).These factors may subject bats to a unique set of threats in artificial roosts that may be less relevant to birds using nest boxes.
A significant knowledge gap exists regarding using artificial roosts for bats that cannot be filled by referencing avian literature.Artificial roosts are used globally for bat conservation, community engagement, and mitigation, specifically to offset the loss of tree cavities, attract bats to agricultural landscapes for pest suppression, and house bats excluded from buildings (Brittingham & Williams, 2000;Flaquer et al., 2006;Puig-Montserrat et al., 2015;Rueegger et al., 2019;Smith & Agnew, 2002).Thus, there is an urgent need to identify if and under which scenarios artificial roosts might act as ecological traps.Bats and other animals may occasionally select suboptimal natural roosts and landscapes.However, conservation practitioners have direct control over artificial roost type and placement (Crawford & O'Keefe, 2021b).When such decisions are guided by rigorous empirical science, it is less likely that ecological traps for bats will be created.We propose testable hypotheses regarding why bats might select suboptimal artificial roosts, mechanisms through which bats could experience fitness declines, and how artificial roosts differ from natural roosts.Further, we offer practical research strategies and suggestions to limit the likelihood that artificial roosts will act as traps.

WHY BATS MIGHT SELECT SUBOPTIMAL ARTIFICIAL ROOSTS?
Evolutionarily reliable cues could trick bats into using a suboptimal artificial roost.Visual, acoustic, and landscape cues are important for roost selection (Ruczyński et al., 2007(Ruczyński et al., , 2011)).In many cases, these cues were evolutionarily reliable signals for bats to identify quality roost sites (e.g., cavities or crevices in large, solar-exposed trees).Knowing this, artificial roosts are often designed to appear natural (e.g., Gumbert et al., 2013;Parker et al., 2022).In Arizona, bats are attracted to resinmodel artificial roosts that mimic natural exfoliating bark roosts (Mering & Chambers, 2012).Artificial roosts in solar-exposed locations attract maternal females seeking warm roost sites (Brittingham & Williams, 2000;Kerth et al., 2001).Optimal or exaggerated visual and placement cues could lure bats into using suboptimal roosts because roost appearance and landscape position signal microclimate and suitability (Callahan et al., 1997;Hernández-Montero, Reusch, et al., 2020;Hernández-Montero, Schöner, et al., 2020).Bats using evolutionarily reliable cues could mistakenly infer that a novel artificial roost is optimal when, in fact, the roost imposes fitness costs.
Recently, there have been efforts to increase via acoustic lures (e.g., broadcasting conspecific social or foraging calls) and olfactory cues (e.g., painting roosts with guano or urine) the attractiveness and uptake of artificial roosts.Calls of conspecifics broadcast near artificial roosts increase bat activity (i.e., acoustic or recorded flight behavior) compared with control sites; however, artificial roost uptake is not affected by use of an acoustic lure (Brokaw, 2015).Further, guano and urine olfactory cues are generally ineffective for attracting bats to roost sites (Brown & Carter, 2022).We have much to learn about the context of bat social calls and olfactory signals.Cues from conspecifics could be evolutionarily reliable indicators of highquality roost sites.Thus, caution is needed when manipulating these signals and applying them to novel artificial roosts that may be suboptimal habitat.
Comparing roost attractiveness will be difficult because there is much variation in natural and artificial roost structure and design.However, when artificial roosts are deployed where there is ample natural roosting habitat, radiotelemetry could allow assessment of bat preferences for artificial and natural roosts.Contrasting roost availability (e.g., snag density or tree microhabitats, number of artificial roosts) and roost use (e.g., presence or absence, group size) will improve understanding of roost preferences.In pursuit of this information, in Kentucky (USA), Howe ( 2023) assessed natural and artificial roost availability in an 1010-ha wildlife management area, documenting approximately 14 suitable roost trees per hectare.However, 12 radio-tracked Indiana bats (Myotis sodalis) selected artificial roosts on 96% of 75 tracking days (24 unique artificial roosts), scarcely using natural roosts (3 unique natural roosts).Understanding the landscape contexts under which bats use artificial roosts and to what degree bats can be drawn away from natural options will improve understanding of the risks of deploying artificial roosts on landscapes with ample natural roosts.For landscapes lacking natural roosts, carving chainsaw hollows in live and dead trees and girdling trees to create snags could promote roosting habitats that more closely mimic the thermal properties of naturally occurring cavities and exfoliating bark roosts (Best et al., 2022;Griffiths et al., 2018;Rueegger, 2017;Schroder & Ward, 2022).Coupling the creation of cavities and dead trees with the deployment of artificial roosts allows for control of roost availability and landscape placement.
Bats may alter their behavior where artificial roosts are available due to search image development, extended roost longevity, and dependence on artificial roosts in marginal habitat.The roosting ecology and social networks of bats worldwide are highly diverse, with many unique ecological associations and behaviors (Kunz & Lumsden, 2003) that might be altered or disrupted by artificial roosts.Experienced bats guide naïve or young individuals to roosts (Kerth & Reckardt, 2003;Ripperger et al., 2019;Wilkinson, 1992), which could result in bats developing strong search images for artificial roosts.Preferences for artificial roosts might be reinforced by habitat-copying behavior or natal habitat preference induction (Danich et al., 2004;Davis & Stamps, 2004).Bats born to colonies using artificial roosts could be more inclined to choose those structures in the future if they have no referent for natural roosts or if trial-and-error roosting strategies are perceived as riskier.Multigenerational preferences for artificial roosts could be perpetuated by adult bats unaware of any sublethal physiological effects related to use of suboptimal roosts (e.g., reduced body mass, decreased immunocompetence).
Because artificial roost designs are deployed over broad geographical areas (Mering & Chambers, 2014;Rueegger, 2016), dispersing bats could develop strong search images for these structures.For instance, bats in eastern North America may frequently encounter rocket box and bark-mimic style artificial roosts that have been widely deployed for conservation and mitigation (Adams et al., 2015;Crawford et al., 2022;De La Cruz et al., 2018;Fontaine et al., 2021;Hoeh et al., 2018).Shifts in roost selection preference (from natural to artificial) might lead to a dissociation between cues previously used to identify high-quality natural habitat and exaggerated cues elicited by artificial roosts (e.g., standardized placements and conspicuous shapes), facilitating the selection of artificial roosts.
To augment available habitat, artificial roosts are sometimes deployed on landscapes with limited or absent natural roosting options (e.g., Flaquer et al., 2006;Rueegger et al., 2019;Whitaker et al., 2006).Where bats lack a search image for natural roosting habitat, but long-lasting artificial roosts are present, it is possible that a population could become dependent on artificial roosts.Because recruitment of suitable roost trees can take decades to centuries (Manning et al., 2013;Vesk et al., 2008), a population reliant on artificial roosts could be set up for long-term failure (e.g., increased mortality, decreased reproductive success, or landscape abandonment) in the absence of continued roost deployment.
To explore behavioral change and search image development, we recommend comparing colonization times, abundance, fidelity, and roost-switching behavior of bat populations using natural and artificial roosts (e.g., Bergeson et al., 2020).It is essential to determine the role of experience and learning in bat roost selection.Previous work suggests that information transfer and learning play a part in roost selection (Kerth & Reckardt, 2003;Ripperger et al., 2019;Wilkinson, 1992).Still, it is unclear whether roost preferences are innate or whether experience is more elemental.In an experimental setting, one could present bats with novel and familiar roost designs to assess natal habitat preference of populations largely dependent on a specific artificial roost design.Manipulating roost quality (e.g., changing roost temperature in an artificial setting [Webber & Willis, 2018]) might show how bats respond to changing habitat quality and clarify whether bats will show fidelity to natal habitats even if habitat quality degrades.Currently, artificial roosts are deployed under the assumption that bats will switch to a different, safer artificial roost design, if available, when conditions temporarily deteriorate, such as during a heat wave (Goldilocks approach [Lausen et al., 2023]).However, there is no direct evidence to support this claim.Exposing bats to unsuitably high roosting temperatures in a captive setting would allow one to determine whether bats will switch roosts during the daytime and accurately locate a cooler alternative.Because artificial roosts are often used as a stopgap measure until natural habitats regenerate or are restored (Whitaker et al., 2006), population trends, roost use, and target species composition should be monitored over long periods to assess bat responses to habitat restoration and forest succession on landscapes formerly devoid of natural roosts (Griffiths et al., 2019;Griffiths, Bender, et al., 2017).Although managers may plan to phase out artificial roosts as a landscape is restored, this may be impossible if bats become dependent on these structures.However, if bats are observed to readily transition back to natural roosts as they become available, this would add further validity to the practice of using artificial roosts as a bridge to natural habitat restoration.

POSSIBLE SUBOPTIMAL TRAITS OF ARTIFICIAL ROOSTS
Compared with natural roosts, artificial roosts are more likely to expose bats to extreme temperatures when they have lower thermal mass, are painted dark colors, and are deployed in solar-exposed areas.Although intended to mimic natural roosts, artificial roosts are thermally different from trees (Griffiths et al., 2018;Maziarz et al., 2017).Some tree roosts-for example, exfoliating bark-are thermally labile, which may benefit bats by facilitating passive warming (Turbill, 2006).However, bark roosts generally track ambient temperature (Russo et al., 2017), often even buffering against extremes (Lacki et al., 2013) and gaining less heat than nearby artificial roosts (Bergeson et al., 2021).In contrast, artificial roosts have lower thermal mass, have lower surface reflectance (if painted a dark color, as commonly recommended), lack water that is present in tree stems, and have no leaves to provide shading.As a result of these synergistic factors, many artificial roosts do not buffer heat well and may reach dangerous temperatures during heat waves or even on mild temperature days with high solar radiation (Brittingham & Williams, 2000;Flaquer et al., 2014;Griffiths, Rowland, et al., 2017;Martin Bideguren et al. 2019;Rueegger, 2019;Tillman et al., 2021).
A warm roost microclimate is critical to pup development and, thus, bats often select maternity roosts with high solar exposure (Bergeson et al., 2018;Brittingham & Williams, 2000).Exploiting this behavior, conservation practitioners often deploy artificial roosts in sunny locations.However, artificial roosts have lower thermal mass (and often lower surface reflectance) than natural roosts and, hence, overheat more frequently when in full sun (Crawford et al., 2022;Rowland et al., 2017;Strain et al., 2021).
Because suboptimal artificial roost microclimates could be a chronic widespread problem (Crawford & O'Keefe, 2021a), we recommend assessing microclimates of prospective artificial roost designs in different climates and solar exposures before offering them to bats (Crawford et al., 2022;Griffiths, Rowland, et al., 2017;Hoeh et al., 2018;Tillman et al., 2021).Artificial roosts should be developed that support a gradient of temperatures and are resistant to extreme temperature fluxes (Hoeh et al., 2018;Honey et al., 2021;Lourenço & Palmeirim, 2004;Rueegger, 2016).Insulating artificial roosts and increasing heat storage capacity will increase heat transfer resistance (Bakken et al., 2022;Honey et al., 2021;Larson et al., 2018), facilitate metabolic heat retention (Ve ˇlký et al., 2010), and increase robustness to cold temperatures through thermal phase lag (Bakken et al., 2022).It would be productive to test combinations of heat storage capacity, insulation, and other modifications to increase solar heating (Bakken et al., 2022).Achieving a balance between overheating risk and a reproductively beneficial microclimate will be crucial to improving bat fitness in artificial roosts (Crawford & O'Keefe, 2021a).
Deploying several temperature data loggers throughout roosts will facilitate fine-scale microclimate assessments (e.g., Bakken et al., 2022;Hoeh et al., 2018).By simultaneously collecting data on bat body temperature and roost microclimates, one can estimate metabolic rates for bats (Willis et al., 2004) and understand the energy costs of different roost types (Crawford et al., 2022).
Compared with natural roosts, long-lasting artificial roosts may promote larger ectoparasite aggregations if they host larger bat colonies that switch roosts less often.Although ectoparasites are frequently documented on bats (Webber & Willis, 2016), little is known about their presence and abundance in natural and artificial roosts.Bats that roost in bat boxes are subject to a host of ectoparasites, such as bat bugs (Cimex spp.), bat files (Streblidae), wing mites (Spinturnix spp.), and ticks (Argasidae or Ixodidae) (Bartonička & Růžičková, 2013;Brittingham & Williams, 2000;Evans & Lumsden, 2011).Ectoparasite infestations increase metabolic rates, alter behavior, decrease body mass, and compromise immune response (Christe et al., 2000;Giorgi et al., 2001;Lourenço & Palmeirim, 2007).For bats that roost in dead trees, exposure to ectoparasites may be low if roosts are used only for 1−2 years.However, the extended longevity of artificial roosts (see previous section) may elevate ectoparasite exposure and buildup compared with dead tree roosts because roost permanence is correlated with ectoparasite prevalence (Patterson et al., 2007).By examining 67 bat species representing 9 families in Venezuela, Patterson et al. (2007) showed that bats roosting in stable roosts, such as mines, caves, and rocky outcrops, had the highest prevalence and intensity of bat flies.Although tree cavity roosts last longer than snag roosts, ectoparasite prevalence and intensity are still lower in live tree cavity nests used by birds (<10% infested) when compared with nest boxes (>80% infested [Hebda & Wesołowski, 2012;Wesołowski & Stańska, 2001]).Warmer and drier microclimates in artificial roosts might facilitate ectoparasite reproduction, driving higher loads (Hebda & Wesołowski, 2012;Maziarz et al., 2017).
Because bat bugs survive winters in uninhabited artificial roosts and tolerate temperatures ranging from at least 5 to 35 • C (Bartonička, 2010;Bartonička & Gaisler, 2007), bats showing interannual fidelity to artificial roosts could be exposed to ectoparasites immediately upon spring arrival.Bartonička and Růžičková (2013) found bat bugs accumulated faster in bat boxes not cleaned in the spring, reaching ∼30 bugs in late May compared with <5 in cleaned boxes.We recently documented hundreds to over a thousand live bat bugs during the winter, in the absence of bats, in boxes used by tens to hundreds of bats over 3 previous summers (Figure 2).
Data are needed on ectoparasite abundance and community composition within natural and artificial roosts to address these questions.Although sampling artificial roosts is feasible (Bartonička & Gaisler, 2007;Bartonička & Růžičková, 2013;Evans & Lumsden, 2011), sampling cavity and exfoliating bark roosts require safety measures due to their fragility and inaccessibility.Cleaning artificial roosts when not in use can reduce ectoparasite loads (Bartonička & Růžičková, 2013) but could be challenging with large, complex designs.Novel roost designs or mechanisms for removing ectoparasites from roosts are needed (e.g., hinged panels to facilitate cleaning, traps for ectoparasites).Because snag roosts are ephemeral and occasionally fall, sampling recently downed natural roosts could provide some insight into ectoparasite abundance and species composition in ephemeral natural roosts.To assess the effects of ectoparasites, bats could be captured at roosts and examined in a controlled setting to determine how immunocompetence relates to observed on-body or in-roost ectoparasite loads.Correlating bat group size and roost switching frequency with data on ectoparasite prevalence would provide insights into how limited roost availability or artificial roost use affects bat ectoparasite loads.
Artificial roost use could elevate predation risk for bats if such roosts are conspicuous, easily accessed by predators, and used on roost-limited landscapes.Because artificial roosts are often conspicuous (e.g., placed on buildings, poles, or along forest edges [Mering & Chambers, 2014]), these structures could become targets for repeated predation attempts when bats are at roost or when they emerge (Kelm et al., 2023).In Norway, soprano pipistrelles (Pipistrellus pygmaeus) are repeatedly depredated by corvids and gulls while roosting and emerging from building-mounted boxes; corvids even developed novel strategies for dislodging bats from boxes, but no predation attempts were documented at nearby woodland roosts during an 11-year study (Michaelsen et al., 2014).
Artificial roosts with large chamber spacing, limited vertical space, or close proximity to the ground could make roosting bats more accessible to predators.Although there is much variability in artificial roost dimensions, many designs mimicking natural roosts are ≤30-cm tall (Crawford & O'Keefe, 2021a;Mering & Chambers, 2014;Rueegger, 2016) and are mounted 3−6 m aboveground (Goldingay & Stevens, 2009;Mering & Chambers, 2014).Natural roost entrances could be more difficult for terrestrial predators to detect if they are higher (e.g., average 8.8 m for 13 tree-roosting bats in North America [Drake et al., 2020]).Although bats in tree cavities may access roosts from entrances near the ground, such species may select roosts with features that allow them to evade predators (e.g., smooth walls and ample flight space [Clement & Castleberry, 2013]).In artificial roosts, large chambers and entrances, coupled with limited vertical space, may give terrestrial predators easier access to bats.
If artificial roosts are long lasting, frequently occupied, and on roost-deficient landscapes, odors from accumulating guano and a higher probability of presence could alert predators.For instance, in Australia, potential predators (e.g., birds, possums, and rats) visit artificial roosts seeded with bat guano more often than unseeded controls (Threlfall et al., 2013).Traps installed below roosts for passive monitoring and fecal collection (e.g., Adams et al., 2015;De La Cruz et al., 2018;Hoeh et al., 2018;Mering & Chambers, 2012) may compound the scent problem by accumulating guano over extended periods.
More information is needed on predation rates at different roost types, including artificial versus natural, for different fine-scale placements, and as a function of roost abundance.Video monitoring offers a cost-efficient way to passively monitor roosts for terrestrial predators (e.g., Figure 3; Threlfall et al., 2013).Although predator guards are recommended for artificial bat roosts (Tuttle et al., 2013), empirical data on their efficacy are lacking.We suggest studies on the efficacy of guards and roost design modifications, including providing narrower entrances and roosting space extending past the reach of predators (Goldingay & Stevens, 2009).
Attractive artificial roosts could facilitate bat use of suboptimal landscapes.Currently, little is known about the effects of providing artificial roosts on landscapes lacking quality foraging, drinking, and roosting resources.Artificial roosts are often deployed where natural roosts are lacking (Flaquer et al., 2006;Long et al., 2006;Rueegger et al., 2019;Whitaker et al., 2006), but, in some cases, efforts may not extend beyond provisioning roosting habitat (Lindenmayer et al., 2017;Rueegger et al., 2019).On human-modified landscapes, adding artificial roosts but ignoring broader resource requirements (e.g., foraging habitat and water) could yield an ecological trap by attracting bats to marginal habitats and increasing exposure to environmental pollutants (Hale & Swearer, 2017).
Because bats provide valuable ecosystem services by consuming and suppressing pest arthropods (Kunz et al., 2011), artificial roosts are sometimes deployed to attract bats to agricultural and urban landscapes (Burgar et al., 2021;Flaquer et al., 2006;Long et al., 2006;Puig-Montserrat et al., 2015).Foraging habitat quality may be lower on such landscapes because insect diversity decreases as agricultural intensification and urbanization increase (Marini et al., 2009;Merckx & Van Dyck, 2019).Thus, dietary diversity will be affected (Clare et al., 2011).Attracting bats to urban landscapes could also elevate risks for predation, competition, chemical exposure, and conflict with humans (Russo & Ancillotto, 2015).Bats drawn to agricultural fields may be more likely to bioaccumulate pesticides from arthropod prey or pollutants from contaminated water sources (e.g., Gerell & Gerell Lundberg, 1993), which could lead to cancer, micronucleation, decreased fecundity, endocrine disruption, immunosuppression, impaired thermoregulation and echolocation, and higher mortality rates (Bayat et al., 2014;Cable et al., 2022;Sandoval-Herrera et al., 2021).Compounding the many risks previously mentioned (i.e., temperature, ectoparasites, predation), artificial roosts on low-quality landscapes could be exceedingly dangerous to bats.
Mark-recapture studies could be used to assess bat fitness in artificial roosts on landscapes differing in quality (e.g., Bailey et al., 2017;Griffiths et al., 2019).Additionally, capturing bats allows one to assess reproductive status and health (e.g., Bailey et al., 2017;Frick et al., 2010).Evaluating physiological parameters could identify fitness impacts and subtle trends in population health not readily detectable with traditional demographic surveys (Cooke & Suski, 2008).To ensure that the locations where artificial roosts will be deployed are suitable for bats, long-term management plans should evaluate proposed sites with respect to habitat quality (i.e., prey diversity and nutritional value, foraging habitat availability, and environmental contamination).

CONCLUDING REMARKS
For tree-roosting bats, continued deforestation and anthropogenic land-use changes could drive an increased reliance on artificial roosts, but widespread use without consideration for bat fitness could pose conservation threats worldwide.Occupancy data alone do not validate the success of artificial roosts because these data only show relative preference for a structure, not population-level effects.Such data can obscure the presence of ecological traps.Demographic fitness measures are critical to identifying traps (Battin, 2004) and improving the science and practice of using artificial roosts.Bat responses to artificial roosts likely differ by species, sex, roost design, and geographic region, and some assumptions may not be generalizable across these components.Furthermore, most research on artificial bat roosts has been conducted in temperate regions, limiting the ability to make robust inferences about bat responses in tropical environments.
Some may ask whether providing bats something, in the form of an artificial roost, is better than nothing.Indeed, there are cases where this could be true.For instance, supplying tree-roosting bats with artificial roosts when they are being excluded from a building, which should decrease the likelihood of human-bat contact, could be a better alternative than supplying no alternative (e.g., Brittingham & Williams, 2000).Additionally, ecological traps do not explicitly imply a negative population growth rate; therefore, low-severity traps could still be beneficial source habitats with a positive population growth rate.However, it is important to critically evaluate the context of the artificial roost deployment and identify and correct potential pitfalls before providing such roosts to bats.Uninformed deployment of artificial roosts as a panacea for habitat loss could lead to adverse outcomes for imperiled bat species (Crawford & O'Keefe, 2021b).
In light of decreasing forest cover, artificial roosts are being deployed globally for bat conservation and mitigation.By better understanding behavior, health, and fitness of bats when they use structures intended to mimic their natural roosts, the science and practice of artificial roost use can be improved.A holistic approach to bat roost habitat conservation should also consider practices that protect and promote the development of suitable natural roosts.We encourage researchers and conservation practitioners to pursue answers to the questions we posed and conduct work that advances this conservation practice for bats.

FIGURE 1
FIGURE 1An overheating event in a dark green plywood roost resulted in the death of 30 juvenile large forest bats (Vespadelus darlingtoni) on 13 December 2017 in Melbourne, Victoria, Australia: (a) 17 male pups, (b) 13 female pups, and (c) pup with milk visible in its stomach.Reproduced fromGriffiths (2022) with permission from CSIRO Publishing.This figure was not originally published as Open Access.To reuse this material, please seek approval from CSIRO Publishing.

FIGURE 2
FIGURE 2 Bat bug (Cimex spp.) accumulation in a 2-chamber rocket box style artificial roost used by a maternity colony of Indiana bats (Myotis sodalis) for 3 summers: (a) the inside wall of the inner roost chamber and (b) the surface of a chamber spacer block, which was a narrow crevice inside the assembled roost.Photo credits: (a) R. D. Crawford and (b) L. E. Dodd.

FIGURE 3
FIGURE 3 A raccoon (Procyon lotor) standing at the base of a rocket box style artificial roost in Indiana, USA.Photo credit: K. A. Cotton.