As fleshy fruits ripen, their sugars are converted to ethanol by in situ fermentation, mediated by microorganisms ^[1]?^[3]. Ripe fruits generally have higher concentrations of ethanol than unripe ones ^[3]?^[4]. Although studies on ethanol concentrations of fruits in natural ecosystems are limited ^[4]?^[5], available information suggests that in Central America, ethanol concentrations in fermenting fruits can range from 0.6% up to 4.5% ^[3]. Thus frugivores and nectarivores in the tropics may be exposed to fruits with a wide range of concentrations of ethanol and, while foraging, may naturally encounter it in high concentrations ^[6]. Frugivorous animals display diverse responses to consuming ethanol (reviewed in ^[3]?^[5]), and where an individual falls on the spectrum of tolerance could influence its ability to exploit certain foods.

The effects of ethanol consumption can be negligible or beneficial for some frugivores (reviewed in ^[3]?^[4]). Pentail treeshrews (Ptilocercus lowii) feeding on nectar of bertam palms (Eugeissona tristis) often consume ethanol in doses exceeding 1.4 g ethanol /kg body mass (enough to legally intoxicate humans) without demonstrating signs of intoxication or deleterious effects ^[6]. Conversely, several fruit-eating animals (e.g., cedar waxwings, Bombycilla cerdorum) are adversely affected when they eat large quantities of fruit rich in ethanol ^[7]?^[11]. In humans and some other primates, elevated blood ethanol concentrations (BACs) can affect performance, impairing reaction times, manoeuvrability, spatial orientation, and reasoning ^[12]?^[14], which may lead to reduced survival. Frugivorous bats use odour ^[15], and sometimes echolocation cues ^[16]?^[17], to locate fruits or assess ripeness ^[18]?^[22]. Many frugivorous and nectarivorous animals find overripe fruit unpalatable and may avoid inebriation by consuming fruit with low ethanol content, or consuming smaller quantities of alcohol-rich food ^[7], ^[18], ^[23].

The effects of ethanol are not entirely deleterious. Ethanol can help stimulate appetite and increase net energy intake in humans ^[24]?^[26]. Ethanol plumes emitted by ripe fruits provide olfactory cues that may potentially increase animals' foraging success ^[2], ^[27]. Given that ethanol has a high energy content, exposure to it should select for processes that aid in its metabolism while minimizing costs such as motor impairment ^[3]. After consuming ethanol, honeybees (Apis mellifera) increase flower visitation rates but show no signs of motor impairment ^[28]. While trade-offs between the costs of consuming toxic foods and the energy rewards has received some attention ^[29]?^[30], previous studies on ethanol ingestion by frugivorous bats have used captive animals repeatedly exposed to ethanol ^[18], ^[29], ^[31]?^[32], and no published studies have examined the flight-motor impairment of bats consuming large amounts (?1%) of ethanol. By examining behavioural responses of wild bats that may have built up natural tolerances for ethanol, we hope to expand the ecological scope of intoxication studies.

We predicted that consumption of ethanol by bats would lead to levels of intoxication (?0.11 g.100 ml^?1 blood alcohol content) that would impair both flight performance (motor control) and echolocation (orientation). Specifically, we predicted that bats fed ethanol would require more time to manoeuvre through an obstacle course, collide with more obstacles, land to rest more often, and not complete the course as often as our control group ^[33]. If ethanol impairs echolocation, then its consumption may alter features of bats' echolocation calls such as pulse durations, interpulse intervals and frequency characteristics of calls compared to those of control individuals ^[34].

We found no significant differences in flight performance or echolocation behaviour between bats fed sugar water and bats fed ethanol, suggesting that wild frugivorous and nectarivorous phyllostomids tolerate ethanol at the concentration we studied. Naturally fermenting fruits can range from 0.6% up to 4.5% ethanol in Central America ^[3]. The 1.5% ethanol concentration we fed bats does not appear to have behavioural repercussions, and may be encountered regularly by foraging bats.

In humans and rodents, genetic factors may explain tolerance to ethanol ^[48]?^[49], Variation in expression of alcohol dehydrogenase and aldehyde dehydrogenase, the primary enzymes involved in the breakdown of alcohol, is a good predictor of individual tolerance of ethanol. Environmental factors, such as long-term dietary exposure to ethanol, may also be involved ^[50]. By examining wild caught bats' behavioural responses to ethanol, we apply ecologically relevant interpretations, and examine the possibility of susceptibility to environmental influence. This avoids confounding artefacts manifest in captive studies, such as training periods and development of unnatural ethanol tolerance from repeated exposures.

Orbach (unpublished observations from July 2008) observed that flying captive Rousettus aegyptiacus (Pteropodidae) were less able to successfully negotiate an obstacle course (obstacles 1 wingspan apart) after consuming fruit juice spiked with ethanol than bats just fed fruit juice. Fleshy fruits in the Neotropics may ferment faster and have higher ethanol concentrations than those in Mediterranean habitats, possibly reflecting differences in humidity, fruit abundance, and prevailing temperatures ^[51]?^[52]. Thus foraging Neotropical fruit- and nectar-feeding bats may naturally be exposed more often to comparatively high levels of ethanol (up to 4.5% ^[3]) compared to those in Mediterranean areas (up to 0.7% ^[31]), and therefore may have higher tolerances for ethanol. However, more research determining naturally occurring concentrations of ethanol in fruits vulnerable to fermentation are necessary to validate this hypothesis.

Feeding states may also be important in determining a bat's likelihood to consume foods rich in ethanol. R. aegyptiacus frequently eat fruits containing 0.1?0.7% ethanol concentrations ^[31], and during food-choice experiments, avoid foods with >1% ethanol ^[18], ^[29], ^[31]. However, food-deprived R. aegyptiacus do not discriminate against foods based on ethanol content and eat foods with high (?1%) concentrations of ethanol ^[29], presumably to compensate for dietary energy shortages, as has been proposed for food-deprived hamsters and mice ^[53]?^[54]. When availability of palatable fruit is limited by seasonality or weather, some frugivorous bats supplement their diet with insects, pollen, nectar, flowers, or leaves ^[55]?^[59]. As had been proposed for R. aegyptiacus^[29], we expect that frugivorous phyllostomids may consume more fruits rich in ethanol more often when other food is unavailable.

Food-deprived rats have slowed ethanol metabolism rates compared to controls ^[60] and the same may be true of food-deprived bats. This would mean longer effects of ethanol compared to well-nourished bats or rats. In places like Israel, where ripe fruit availability is seasonal, the possibility of food-restriction is real ^[57], and bats may be susceptible to predation risks and intoxication. The wild phyllostomids we studied had been captured while foraging and may have had residual food in their stomachs to help absorb ethanol and mitigate its impact ^[61]. Future studies using food-deprived phyllostomids or increased concentrations of ethanol may reveal greater behavioural impact of ethanol that could be important to understanding survival probabilities.

Interspecific variation we observed in flying time, course completion, and return to the release point, irrespective of feeding treatments, may be part of a suite of interspecific dietary ^[62]?^[63] and habitat differences ^[35] among phyllostomids. Body mass does not appear to affect flight performance, although it may account for changes in echolocation behaviour. Differences among species in echolocation call characteristics could reflect morphological variation, as the call frequencies of bats increase as their body mass decreases [(Table 1); 64].

Our bats appeared to be behaviourally unaffected by ethanol at the concentration we used, although we observed interspecific variation in blood alcohol concentrations. By sampling widely across phyllostomid species of different masses, we demonstrate that the patterns of exposure to, and tolerance of ethanol are broad, and suggest that sensitivity to ethanol may be important in the adaptive radiation of frugivorous and nectarivorous bats. Across the tropics there is variation in the diversity of frugivorous bats, ranging from 25 genera in the Neotropics to 11 in the Ethiopian tropics to 21 in the East Indies ^[65]. The properties of fruits, such as their temporal and spatial distribution, size, and resistance to mechanical deformation, may contribute to this diversity within regions [reviewed in 66]. For example, variation in biting styles and dentition may allow some bats to exploit harder fruits inaccessible to sympatric species ^[59]. Future long-term studies contrasting the tolerance of fruit- and insect-feeding phyllostomids to ethanol may provide another factor explaining the diversity of New World fruit-feeding bats, and confirm the role of alcohol tolerance in the evolution of frugivory. We advise using alternative methods to assess the BACs of bats other than alcohol-sensitive reagent pads due to the highly subjective nature of interpreting colour blots. If tropical and subtropical fruits contain higher levels of alcohol more often than those in temperate environments (e.g., Mediterranean climates), ethanol tolerance of other frugivores and nectarivores elsewhere in the tropics and subtropics may indicate parallels among Old World fruit-feeding bats. Furthermore, species like R. aegyptiacus, which occur widely from South Africa to Turkey, may show geographic variation in ethanol tolerance reflecting a range that brackets tropical and subtropical climates.

Competing Interests: The authors have declared that no competing interests exist.

Funding: This project was supported by a grant from the University of Western Ontario (Graduate Thesis Research Award) to DN Orbach, and grants from the Kenneth M Molson Foundation and the Natural Sciences and Engineering Research Council of Canada (Discovery Grant R3516A03) to MB Fenton. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

We thank Amanda Adams, Reese Arh McMillan, Laura Barclay, Beth Clare, Dalal Dahrouj, Christina Davy, Leif Einarson, Erin Fraser, Lauren Hooton, Craig Harding, and Lasse Jakobsen for their assistance collecting bats and building the flight room. We thank the staff of Lamanai Outpost Lodge for their hospitality. We are grateful to Hugh Aldridge, Mark Brigham, Daniel Riskin, and John Ratcliffe for reviewing earlier drafts of this manuscript.