pallidus, Arnold & Wilkinson, 2011), and common noctules (Nyctalus noctula, Furmankiewicz et al., 2011)) and/or they use high-amplitude echolocation calls from conspecifics to locate roosts more efficiently (e.g., common noctules and Daubenton's bat (M. daubentonii, Ruczyński & Kalko, 2007, Ruczyński et al., 2009)).

Far less is known about whether bats use scent cues to locate potential roosts, despite olfaction being an important sensory modality for almost all terrestrial mammals (Hayden et al., 2010). In this review, we discuss some reasons that bats would, or would not, use scent cues in habitat selection. We focus on studies testing the use of inadvertent cues from guano and urine because these scents are present at most occupied roosts, and anecdotal reports suggest these cues might be used by bats in roost selection. For example, some people have applied guano to artificial roost boxes to increase the chance of occupancy (Murphy, 1993;Ober, 2008). If guano staining was an effective olfactory lure for artificial bat roosts, this would be a helpful tool for bat management and conservation. Although one can buy chemical lures for bat houses based on distinct odors present in some bat roosts (Nielsen et al., 2006), there is no clear evidence of the effectiveness of any olfactory lure. More generally, a better understanding of the use of scent cues for roost-finding in bats can provide insights linking roosting ecology, social behavior, and sensory ecology in bats.

To assess the capacity of bats to use scent in roost selection, it is useful to review the basic structures for perceiving odors. In mammals, the key olfactory structures are the main olfactory system and the accessory olfactory system (AOS), which can include the vomeronasal organ (VNO) and the accessory olfactory bulb (AOB). Greater complexity of the epithelial structures in these systems can increase the amount of surface area, allowing for greater absorption of odorant molecules (Hecker et al., 2019). The vomeronasal system plays a large part in assessment of chemical cues involved in social behaviors such as mate attraction (Kelliher, 2007).

Nocturnal mammals tend to have larger olfactory structures (Barton et al., 1995) compared to other terrestrial mammals, but most bats have reduced olfactory systems, such as a reduced VNO, a reduced AOB, or the complete loss of one or both of these structures (Hecker et al., 2019). The exception is that an intact VNO and AOB exists in most, or all, species within two of 20 sampled families of bats (Phyllostomidae and Miniopteridae) and in the genus Pteronotus in the family Mormoopidae (Yohe & Dávalos, 2018;Zhao et al., 2011). However, many bat species (e.g., most bats in the family Pteropodidae and Hipposideridae) still have not had their accessory olfactory structures examined (Zhao et al., 2011). Along with reductions in olfactory structures, most bat lineages have lost protein-coding genes-such as Trpc2 and S100z-linked to pheromone perception (Hecker et al., 2019;Yohe et al., 2017Yohe et al., , 2019)). Since many species of bats exhibit behaviors that involve pheromones, it is possible that the main olfactory system in bats incorporates the functions of these accessory systems (Yohe et al., 2017).

Diet plays an important role in the evolution of mammalian olfactory systems (Barton et al., 1995;Brokaw & Smotherman, 2020;Eiting et al., 2014). Phyllostomid bats have the most diverse diet among the bat families (with both generalist omnivores and specialists on insects, fruit, nectar, blood, or small vertebrates), and of those studied to date, all but two phyllostomid species (Brachyphylla cavernarum and Choeroniscus godmani) have maintained all parts of their ancestral olfactory system. In contrast, in 18 of the other bat families, the sampled species have lost parts of these structures (Cooper & Bhatnagar, 1976;Frahm & Bhatnagar, 1980;Schmidt, 1985;Yohe et al., 2017;Yohe & Dávalos, 2018;Zhao et al., 2011). For example, the vespertilionid bats are primarily composed of insectivores that probably rely less on scent, and 11 of 13 sampled vespertilionid species have lost both the VNO and AOB (Bhatnagar & Kallen, 1974;Cooper & Bhatnagar, 1976;Wible & Bhatnagar, 1996;Yohe & Dávalos, 2018). The loss of the AOS in so many bat lineages may be due to ecological differences, but the specific factors are not well understood (Yohe et al., 2017).

Due to the strong link in other mammals between the vomeronasal system and the processing of social cues, bats with more developed VNOs and AOBs are expected to more frequently use olfaction during social interactions (Hecker et al., 2019). However, the relative roles of social and ecological factors as selective pressures are difficult to disentangle because (1) the main olfactory system in bats might coopt functions of the vomeronasal system, (2) there is a paucity of data on olfaction-based social communication in bats, and (3) social and ecological predictors can be statistically confounded. To take one example, the common vampire bat has a well-developed vomeronasal organ, one of the largest relative AOBs recorded in bats, and the most V1rs (Vomeronasal type-1 receptors) recorded to date (Yohe et al., 2019). As one might predict, this species is exceptional in its social complexity (Carter & Leffer, 2015;Wilkinson et al., 2016), but it also appears to use olfaction to find its hosts, which include a large diversity of species (Bahlman & Kelt, 2007;Carter et al., 2020), so it is unclear whether vampire bats evolved enhanced olfaction for social recognition, host recognition, or both.

Roost choice by bats involves multiple sensory cues (Hernández-Montero et al., 2020;Kerth, 2008;Ruczyński & Kalko, 2007;Ruczyński et al., 2009), and there are several reasons why we might expect roost-finding bats to use scent-based social cues alongside other social cues. First, bats use scent in other contexts, such as social recognition, foraging, and mate selection (Bloss et al., 2002;Bouchard, 2001;Chaverri et al., 2018;De Fanis & Jones, 1995;Theis et al., 2016). Olfaction can be used in combination with spatial memory and other senses depending on proximity to the target (Gomes et al., 2016;McCracken, 1993). For example, when searching for their pup in a creche with potentially hundreds or thousands of other pups, Brazilian free-tailed bats (Tadarida brasiliensis) seem to use spatial memory and isolation calls to first locate the general area of their pup, and then use scent to identify their specific pup (Gustin & McCracken, 1987). This multi-modal integration has also been shown in the lesser long-nosed bat (Leptonycteris yerbabuenae) when it uses both echolocation and olfaction to locate cactus flowers (Gonzalez-Terrazas et al., 2016). Similarly, fruit-eating bats are thought to rely on olfaction to locate a fruiting tree and then a combination of echolocation and olfaction to pick ripe fruit (Kalko et al., 1996;Korine et al., 2000). Within interior forests or other high-clutter areas where high-frequency calls do not transmit well (Martens, 1980), bats might use olfactory cues of an unfamiliar, large roost to travel to the general area and then use social calls as a beacon to find the roost entrance once they are closer.

A second reason that scent cues are potentially useful is that they persist over time, whereas sounds convey the presence of bats only at that moment. Scent cues might be important for finding roosts in the absence of conspecifics, as when temperate bats make seasonal movements to summer sites that are vacant at the end of winter, or when a bat comes across a roost that is unoccupied during the night because the occupants are foraging.

Third, scent cues at roosts can be used for social recognition. Bats often preferentially roost with certain individuals, even in colonies where individuals frequently switch roosts and group composition changes day to day (Kerth et al., 2011;Popa-Lisseanu et al., 2008;Rhodes, 2007;Wilkinson et al., 2016;Willis & Brigham, 2004). Maintaining preferred social associations is particularly important for cooperative behavior in some phyllostomid bats (Carter et al., 2020;Wilkinson, 1984;Wilkinson & Boughman, 1998). Chemicals collected from various scent glands contain information about individual identity, and scent profiles might also be distinctive at the colony level in some species (e.g., the Bechstein's bat (Safi & Kerth, 2003) and the greater bulldog bat (Noctilio leporinus, Brooke & Decker, 1996)). In the Bechstein's bat, chemical profiles were similar within haplotypes and groups, but these similarities were not driven by pairwise kinship, because mean within-colony relatedness was only 0.02, and the scent profiles of mothers and adult daughters were not more similar than among other groupmates (Safi & Kerth, 2003). Given that different colonies had distinct chemical profiles, and bats from the more chemically different colonies were more aggressive to each other, colony-level differences in chemical profiles might also predict the outcomes of between-colony social encounters (Safi & Kerth, 2003).

When given a choice to associate with scent samples, bats preferred the scent of familiar versus unfamiliar individuals in the Angolan free-tailed bat (Mops condylurus, Bouchard, 2001), little free-tailed bat (Chaerephon pumilus, Bouchard, 2001), common pipistrelle (Pipistrellus pipistrellus, Bartonicka et al., 2010;De Fanis & Jones 1995), soprano pipistrelle (P. pygmaeus, Bartonicka et al., 2010), greater sacwinged bat (Saccopteryx bilineata, Voigt & von Helverson, 1999), big brown bat (Eptesicus fuscus, Bloss et al., 2002), and Brazilian free-tailed bat (Englert & Greene, 2009;Gustin & McCracken, 1987;Loughry & McCracken, 1991). Males also responded differently to female versus male scents in the common pipistrelle (Bartonicka et al., 2010), soprano pipistrelle (Bartonicka et al., 2010), little free-tailed bat (Bouchard, 2001) and Angolan free-tailed bat (Bouchard, 2001). In a study with Egyptian fruit bats (Rousettus aegyptiacus), the subject chose between an unfamiliar and a familiar bat placed into two chambers, and it was suggested that the choice was based on olfaction due to the absence of visual and acoustic cues (Mann et al., 2011). All of these studies suggest a level of individual recognition through scent cues.

Fourth, roost-finding bats might use scent cues because in some species male bats use glands to scent mark roost entrances (e.g., Pacific flying fox (Pteropus tonganus, Grant & Bannack, 1999), the black flying foxes (P. gouldii, Moulton, 1967;Nelson, 1965), gray-headed flying fox (P. poliocephalus, Moulton, 1967;Nelson, 1965), greater sac-winged bat (Caspers et al., 2009), pale spear-nosed bat (Phyllostomus discolor, Höller & Schmidt, 1996), and lesser sac-winged bat (S. leptura, Caspers et al., 2009). Females can recognize quite a few individual characteristics from this scent marking, such as age, sex, social status, and identity, that can help them select a mate (Muñoz-Romo et al., 2021). While the function of scentmarking in males is to increase mating opportunities by signaling to females or rival males, scent marks at the entrance of a roost could also act as a cue to whether that roost was occupied by familiar or unfamiliar conspecifics.

Fifth, even without scent marking, many entrances to bat roosts accumulate guano and urine that provide inadvertent olfactory cues. Large, occupied bat roosts can accumulate urine and guano over months or years. Predators and trained dogs can locate bat roosts using scent (Chambers et al., 2015;Threlfall et al., 2013). Some owners of occupied bat houses have recommended applying guano to artificial roost boxes to help attract bats to them (e.g., Murphy, 1993;Ober, 2008). Beyond serving as a cue for roost location, feces and urine can communicate important social information in many other mammals (Beynon & Hurst, 2004;Delahay et al., 2000;Eppley et al., 2016;Ferkin & Johnston, 1995;Heth et al., 1998;Hurst et al., 2001;Ramsay & Giller, 1996). The use of these scent cues does not imply any adaptive trait for enhanced olfaction, because bats with a weak sense of olfaction could learn to associate the scent of guano and urine with the reward of finding a new roost site.

Finally, although bats that do not roost in groups may be less likely to use conspecific scent cues to choose roosts, they could also use scent cues to find previously occupied sites that are now empty. Cues of previous occupancy could indicate a suitable roost microclimate but also an increased risk of exposure to predators or parasites.

There are also several reasons why we might not expect bats to use scent to find roosts. As mentioned above, the reduction in the accessory mammalian olfactory systems may indicate less reliance on olfaction relative to other terrestrial mammals. Second, if the primary benefit of a roost is the presence of other bats, then olfaction may not be a useful cue because, unlike bat calls, odors do not convey the immediate presence of other bats, and an empty but scented roost could have degraded since it was last occupied. Third, not all bat roosts accumulate guano and urine. For example, for foliage-roosting bats (as opposed to cavity-roosting bats) there is often not a bottlenecked roost "entrance" where guano would accumulate. Fourth, scent cues would be most useful for bats that need to find new roosts (i.e., no use of spatial memory), but the species that most frequently switch to new roosts (e.g., leaf-roosting bats) may not occupy them long enough to leave large concentrations of guano and urine. Finally, the range at which a bat can perceive cues will clearly play a role in a bat's ability to locate new roosts (Ruczyński & Barton, 2012), and in some contexts like open habitats, even strong scent cues might have a more limited detection range compared with echolocation or social calls. Although bats possess olfaction and could plausibly use scent cues in roost choice, other sensory cues might be more easily detected or more salient. Note. 1. Barclay et al., 1988;2. Boonman, 2006;3. Brown et al., 2020;4. Chinnasamy et al., 2011;5. Cooper & Bhatnagar, 1976;6. Davis et al., 1962;7. Dechmann et al., 2010;8. Englert & Greene, 2009;9. Entwistle et al., 1997;10. Finn, 1997;11. Frahm & Bhatnagar, 1980;12. Fraze & Wilkins, 1990;13. Humphrey, 1966;14. Keeley & Keeley, 2004;15. Kilgour et al., 2013;16. Menzel et al., 2001;17. Mueller, 1966;18. Ruczyński & Bogdanowicz, 2005;19. Ruczyński & Kalko, 2007;20. Ruczyński et al., 2009;21. Selvanayagam & Marimuthu, 1984;22. Silvis et al., 2014;23. Wible & Bhatnagar, 1976;24. Wilkinson, 1992;25. Wilkinson, 1984;26. Yohe et al., 2017;27. Zhao et al., 2011 To assess evidence for and against these arguments, we conducted a literature review, using combinations of the search terms: bats, olfaction, scent, sensory ecology, roost, roosting ecology, olfactory bulb, recognition, homing, latrines, habitat selection, mammals, and habitat. We found 119 papers on olfaction, sensory ecology, and roosting ecology in bats. Below, in chronological order, we focus on the eight studies (with 12 species from four different families) with data that could address whether bats used guano and urine cues in roost selection (Barclay et al., 1988;Brown et al., 2020;Englert & Greene, 2009;Finn, 1997;Mueller, 1966;Ruczyński & Kalko, 2007;Ruczyński et al., 2009;Selvanayagam & Marimuthu, 1984). For each species in these studies, we summarized information about diet, social organization, roost type, and system (Table 1).

To test the importance of olfaction in the homing ability of bats, Mueller (1966) covered the nares of 35 little brown bats (M. lucifugus) and released 10 of these treated bats at a distance of 10 km from their hibernaculum, and another 25 treated bats at a distance of 32 km. As a control group, he also released 25 untreated bats at each distance. Bats were marked with paint to distinguish between treatment groups. He repeated this study with Indiana bats (M. sodalis) by releasing 100 bats with their nares covered and 100 control bats at a site 32.2 km from their hibernacula. He did not analyze these data but argued that bats were at least capable of returning to a hibernaculum without the use of scent. A contingency test of the data (Chisquare test with Yates Correction), however, finds evidence that bats with covered nares were less likely to return to their roost at 32 km. When little brown bats were released at a distance of 10 km, there was no clear treatment effect (13 of 25 control bats returned, 3 of 10 treated bats returned with nares covered, and 5 more treated bats returned after removing the treatment from their nares). When released at a distance of 32 km, none of the 25 treated bats returned with nares covered (six returned with uncovered nares), whereas 16 of the 25 control bats returned (X 2 (N=44, df=1) = 11.6, p < .001). A similar pattern was found with Indiana bats released at 32 km: only four of the 100 treated bats returned with nares covered (and 18 returned with uncovered nares), whereas 18 of the 100 control bats returned ((X 2 (N=182, df=1) = 6.1, p = .01). These results are consistent with an effect of olfaction on homing, but other explanations are also possible, such as that covering the nares affected survival, echolocation, or flight performance.

It is implausible that homing bats rely entirely on olfaction when locating hibernacula because they travel long distances, up to 500 km (Klüg-Baerwald et al., 2017;Norquay et al., 2013), but olfaction could aid in navigation. According to the empirically-supported olfactory navigation hypothesis (Wallraff, 2004), birds, such as homing pigeons, exploit natural odors across the landscape to aid their navigation (Gagliardo et al., 2011;Papi et al., 1973;Wallraff, 2004). While little brown bats do not have a VNO or AOB, they may still process and recognize certain scent cues close to their hibernaculum through their main olfactory system (Table 1). The results from Mueller (1966) are consistent with this hypothesis, but compared to birds, virtually nothing is known about the use of scent cues in homing by bats. Selvanayagam and Marimuthu (1984) tested the impact of removing scent cues on the roosting and territorial behavior of the group-roosting Schneider's leaf-nosed bats (Hipposideros speoris). After marking 40 bats and observing them returning to their roosting locations in a cave, the authors noticed that males often sniffed the substrate near their previous roosting position. To determine if bats were indeed using scent to choose their roosting positions, the researchers washed the cave walls at these spots. Normally, it took bats an average of seven minutes to settle at roosting sites (12 observations from eight bats), but when the area was washed, it took them two to three hours (five observations from five bats). Further experiments in captivity showed that males and females were able to identify previous roosting locations, at least in part through sniffing behaviors. The bats also fought for the darkest roosting locations, and when the researchers moved the light around, the bats shifted their positions so the most dominant individuals were again in the darkest corners. Once the bats had chosen new positions, the males marked their area with urine. The bats appeared to discriminate their own urine from the urine of other individuals, because when urine from other males was placed in the roost site of a focal male, it would respond antagonistically upon its return and cover the mark with its own urine. Females did not mark roosting sites with their urine, but they sniffed at the roof of the cave in the wild and the cage in captivity, possibly using scent from neighbors choose their roosting position. All these observations suggest males use urine marks to advertise their territory and re-locate their roosting spots within a single cave. This indicates that these bats are likely able to process pheromones in the urine, possibly through their VNO, but it does not indicate how they initially find and select their roosting sites (Table 1). Barclay et al. (1988) tested individual silver-haired bats (Lasionycteris noctivagans) in a flight cage with two roosts that were identical except that one had a conspecific bat placed inside it for at least 24 hr before the trial. Eleven bats selected the scented roost in 12 of 21 trials. Although silver-haired bats typically roost in tree crevices that could become scented, an attraction to roosts with unfamiliar scent cues is not strongly expected because they do not normally roost with conspecifics (Barclay et al., 1988; Table 1).

Finn (1997) studied whether guano and urine could be used to attract bats, but the results were unclear. In 13 field trials, bats were trapped in a roost box where their scent was able to accumulate. Groups of Brazilian free-tailed bats (mean group size = 23 bats, range = 6 to 110 bats), and evening bats (mean group size = 4 bats, range = 1 to 9 bats), were each placed inside a roost box near an established roost at dawn, and most bats then stayed inside until sunset. After the bats left the roost box in the evening, no bats were observed there for the next two days. In a captive experiment, six groups of 14 to 16 bats (six trials, 79 Brazilian free-tailed bats and 10 evening bats) were each given the option of different roosts in a flight cage, including four plywood bat houses, two pine bark structures, three bricks with roost holes, and two scented roost sites. The scented sites contained a cotton rag with guano and urine from a small colony of evening bats and a nylon stocking tube (folded in such a way that a bat could crawl into the middle) filled with guano from a colony of Brazilian free-tailed bats. Roosting choices were recorded the next morning. In two different trials, one Brazilian free-tailed bat roosted on the nylon stocking. Without proper experimental control conditions for comparison, these observations are difficult to interpret.

Two experiments tested the sensory basis for selecting new roosts in three group-roosting vespertilionid species, the common noctule bat, brown long-eared bat, and Daubenton's bat, (Ruczyński & Kalko, 2007;Ruczyński et al., 2009; see Table 1). These studies aimed to isolate and test how the success of locating new roosts was affected by sound (echolocation call playbacks), temperature (a heated roost), vision (presence of light), and scent (a cloth treated with some of the bat's feces placed into each bat's cage for 24-48 hr). In training trials, each bat was trained to find one of eight cavities drilled into a log by rewarding them with mealworms when they found a cavity. A bat was included in the study when it consistently found a cavity within six minutes. In each test trial, only one of the eight cavities contained the stimulus. Each subject (eight noctule bats, six long-eared bats, and nine Daubenton's bats) was tested eight times for each of the four stimuli (32 trials per bat). Researchers measured latency for a bat to locate cavities with or without the stimulus. For the scent treatment, bats did not locate the scented roosts more often or faster than unscented roosts, regardless of whether they circled the log in flight or landed on the log and crawled around. While the scent cues did not decrease search time for cavities, conspecific echolocation calls clearly did improve search times. In this study, the scent cue was from the subject itself, rather than the scent of a conspecific which would better represent a bat locating a new roost. Englert and Greene (2009) collected scents by placing felt squares in roost pouches within two captive colonies of Brazilian free-tailed bats for five days. Individual bats (N = 20) chose between three roosting pouches: one with scent from the familiar colony, one with scent from the non-familiar colony, and a clean control pouch. Bats preferred roost pouches with the familiar scents over the other pouches. Although it is possible that the felt squares were marked with guano or urine, the source of olfactory cues used for roost choice is not clear. Brown et al. (2020) investigated the use of conspecific guano and urine cues in the location of new roosts in three group-living bat species (vampire bats, Desmodus rotundus; velvety free-tailed bats, Molossus molossus; and big brown bats, Eptesicus fuscus; see Table 1). Eight experiments tested a range of situations, all of which offered individual bats the choice between a scented or unscented roosting location either in the field using roost boxes or in captivity using a dark chamber, a lit room with dark refuges, or a y-maze. First, five captive common vampire bats repeatedly chose scented tubes more often than unscented tubes, when escaping from light. However, in three other experiments, vampire bats (N = 20, 22, and 33) chose between scented and unscented roosting locations in a small chamber that was entirely dark, and there was no clear evidence of a bias under these conditions. There was also no evidence for attraction to guano and urine scent when 45 vampire bats were tested in a y-maze, although the same individuals were clearly attracted to playbacks of echolocation and social calls. Velvety free-tailed bats (N = 18) also showed no clear preference for hanging on surfaces marked with guano and urine.

To test the effectiveness of a potential olfactory lure for attracting wild bats, Brown et al. ( 2020) also placed paired artificial roosts, one treated with guano and urine and one as a control, for one week at each of 16 sites in Panama and seven sites in Ohio, U.S.A., near existing and occupied artificial roosts. These roosts were monitored with acoustic detectors to determine when bats were visiting and what species were visiting. These field experiments failed to provide clear results because only one visit was recorded, at a scented box in Ohio. The artificial roosts were placed temporarily near existing roosts, so it is possible that sensory cues from the occupied natural roost overshadowed the cues from the artificial roosts, and since in most cases, the scent cues were from other sites, it is also possible that bats were repulsed by scents of unfamiliar conspecifics.

Taken together, what do these studies tell us? First, although bats clearly use scent in many contexts, the sparse evidence that exists suggests that the scent of guano and urine does not act as a dramatic olfactory beacon that lures bats towards roost locations to the same degree as bat calls can provide an acoustic lure (e.g., feeding buzzes, social calls, or even distress calls; Braun De Torrez et al., 2017;Carter et al., 2015;Gillam, 2007;Samoray et al., 2019). In fact, three of the studies that looked at the effect of guano and urine on roost location also looked at the effect of acoustic cues and found that while guano did not readily attract bats, acoustic cues had strong and clear effects on attracting bats to roosts (Brown et al., 2020;Ruczyński & Kalko, 2007;Ruczyński et al., 2009). This finding is consistent with the hypothesis that scent cues are not as reliable of an indicator of the immediate presence of conspecifics or the current viability of a roost, when compared to the acoustic cues of bat calls.

There is still much uncertainty regarding if and how different bats use guano and urine or other scents in roost selection. Most of the studies we included were statistically underpowered with limitations in sampling or deficiencies in the experimental design. Bats might rely on scent cues from guano and urine only within specific contexts or motivational states, such as males searching for females, or young bats searching for conspecifics after dispersing into new areas. In the study by Selvanayagam and Marimuthu (1984), for instance, the male bats may have been using urine to locate their own roosting locations, to advertise their roosting locations to other bats, or both. Similarly, Brown et al. (2020) appeared to have detected a strong effect of guano and urine on roost selection when bats were seeking a dark refuge, but the effect was not detected in other experiments when the entire experimental arena was dark. Some scent cues clearly do influence roost selection (Englert & Greene, 2009), but even if guano and urine is an indicator of the presence of bats at a roost, it is possible that these cues do not contain sufficient social information to be used in roost selection. While guano and urine are present at most bat roosts, scent signals, such as gland secretions, might play a larger role in roost location.

The best evidence for an effect of urine on roost selection comes from a study that used a design that controls for the effects of spatial memory by testing the bats' reaction to the removal of the scent at a familiar location (Selvanayagam & Marimuthu, 1984). Therefore, an experimental design that manipulates scent at a familiar location, rather than at new locations, seems like a promising approach for future experiments. For example, one approach would be to present a scented and unscented entrance to a roost in equal proximity to the familiar entrance which is now blocked and test which new entrance is used (Brown et al., 2020). Another approach would be to swap guano and urine scents that are familiar versus unfamiliar for a random half of bats trained to return to the same familiar roost and then to measure each bat's latency to enter the roost.

In some rodents (Böcskei et al., 1992;Hurst et al., 2001) and carnivores (Hradeck, 1985), social information, such as individual identity or genetic relatedness, is conveyed through major urinary proteins (Beynon & Hurst, 2004;Zhou & Rui, 2010). Therefore, it would be pertinent to identify the social information, if any, that is present in the guano or urine of bats. To test the perception of social information in guano or urine, bats could be tested in a y-maze where either the subject or the source of the sample varies by identity, sex, or reproductive status. To determine the persistence of chemical information over time, similar tests could be repeated with fresh and old (decayed) samples. Testing the production of social information, requires chemically analyzing samples (e.g., using gas-chromatography mass spectrometry) to identify the combinations of compounds encoding social information. Finally, the eight studies included in this review tested only four different bat families and only one known to have a well-developed main and accessory olfactory system. Behavioral data from more bat species is clearly needed to resolve the role of olfaction in the roosting and sensory ecology of bats.