The role of past experience in development of feeding behavior in common vampire bats

Introduction

Common vampire bats, Desmodus rotundus (E. Geoffroy, 1810), possess a suite of traits associated with complex foraging skills that are learned gradually over time (Schuppli, Isler & van Schaik, 2012). These traits include flexible foraging, slow development, individualized social relationships, extended parental care, and sharing of food with nonkin (Schuppli, Isler & van Schaik, 2012). Vampire bats are arguably flexible foragers because they rely on a combination of vision, echolocation, passive listening, olfaction, mechanoreception, and thermoperception to extract blood from a diverse range of vertebrate hosts (Greenhall, 1988; Schmidt et al., 1991; Gröger & Wiegrebe, 2006). They have a long lifespan, with records of 29 years in captivity (GGC, personal observation) and at least 17 years in the wild (Delpietro et al., 2017), and are an outlier among bats with regards to their maternal investment and slow reproduction. Gestation requires 5.5 to 7.3 months, and sexual maturity requires up to 9.5 months (Schmidt, 1988a; Delpietro & Russo, 2002). Juveniles are born large (roughly a quarter of the mother’s mass) and depend on their mother for longer than any other bat species; mothers nurse juveniles for 8 to 10 months. In contrast, lactation periods in temperate and other neotropical bats are typically three weeks to three months (Jenness & Studier, 1976; Kunz & Robson, 1995; Hamilton & Barclay, 1998; Crichton & Krutzsch, 2000; Delpietro & Russo, 2002; Kwiecinski, Falzone & Studier, 2003; Chaverri & Kunz, 2006). Female vampire bats form long-term social relationships that involve regurgitations of blood to unfed individuals in need (Wilkinson, 1984; Wilkinson, 1985; Delpietro & Russo, 2002; Carter & Wilkinson, 2013; Delpietro et al., 2017). This cooperative social support is especially important for younger bats because vampire bats will die after three consecutive unfed nights and individuals younger than two years fail to feed on roughly one-third of nights (Wilkinson, 1984). Vampire bats have one of the largest brain and neocortical volumes for their body size among bats (Baron, Stephan & Frahm, 1996; Bhatnagar, 2008). Finally, young vampire bats show increased exploration and decreased neophobia that peaks at 9-10 months (Park, 1990; Vrtilek et al., 2018). This unique combination of social complexity, slow life history, and flexible foraging make vampire bats a good candidate for a large role for learning in foraging. On the other hand, vampire bats are obligate blood-feeders and this extreme dietary specialization can also lead to evolutionary losses in learning ability. For instance, vampire bats have lost the widespread adaptive specialization of taste aversion learning (Ratcliffe, Fenton & Galef, 2003).

How important is learning during development of feeding behavior in vampire bats? To begin to answer this question, we took advantage of an existing captive colony of both captive-born and wild-born bats to test the role of past experience in feeding behavior. In the first test, we asked whether captive-born vampire bats show an innate preference for warm blood. Vampire bats possess unique heat receptors near their nostrils which allow them to detect blood-rich areas on the skin of their hosts at distances of up to 16 cm (Kürten & Schmidt, 1982; Schmidt & Manske, 1982). One previous study found that wild vampire bats will consume blood as cold as 14 °C (Bullard & Shumake, 1973), but the role of past experience in blood-temperature perception and preference remains ambiguous. We compared the blood-temperature preferences of wild-born bats with captive-born bats that had previously only fed on chilled or ambient temperature blood. If learning plays a key role in their use of thermoperception, then we expected that captive-born vampire bats would differ from wild-born bats. Specifically, wild-born bats should show a greater preference for warmer blood (37 °C) over artificially chilled blood (4 °C).

In the second test, we asked whether captive-born vampire bats could successfully feed from a live animal, given their complete lack of experience with this task. If a critical period of early experience or social learning is required for extracting blood from a live animal, then naïve captive-born bats, reared without experiencing feeding on live host animals, might be unable to feed successfully on live hosts. In the wild, younger vampire bats have been observed feeding from the same wound as their mothers (Wilkinson, 1985). Therefore, to test for whether the presence of mothers helps their offspring learn how to feed, we tested naïve captive-born bats alone and again in the presence of their wild-born mothers.

Materials & Methods

Results

Experiment 2. Can captive-born vampire bats feed on live prey?

We found that captive-born vampire bats could successfully bite a live chicken without previous experience or maternal assistance. Their feeding success, however, was ambiguous. Five of the 12 captive-born bats bit the chicken when tested alone, and one of the six captive-born bats bit in the presence of their mother (Table 1, Video S1), yet four of the five captive-born bats that fed did not gain more than 1 g of mass, showing that their feeding performance was poor (Table 2, Data S2). The oldest captive-born vampire bat (16 months old) fed on live prey to an extent that was comparable to the wild-born adults, and it gained a 12% increase in body mass (Table 2), but the other four captive-born bats that were capable of making a wound and feeding from it, did not gain much weight (−3% to +3% change in body mass).

Discussion

Neither the captive-born nor wild-born vampire bats preferred warmer blood. Wild vampire bats will feed from wounds created by other vampire bats, so it is not too surprising that they readily drink from an open spout of blood (Greenhall, Schmidt & Lopez-Forment, 1971; Wilkinson, 1985). It is more surprising, however, that their reliance on thermoperception did appear to not generalize to this novel situation. Vampire bats rely on thermoperception to find blood vessels near the surface of the skin, and will choose to bite the warmer of two rabbit ears (Kürten & Schmidt, 1982; Schmidt & Manske, 1982). Yet a previous study reported that, when eight wild vampire bats were presented with choices of blood at 14 °C, 30 °C, 38.5 °C, and 47 °C, the bats did not show a temperature bias except for an avoidance of the 47 °C blood, possibly due to the denaturing of the blood proteins (Bullard & Shumake, 1973). Our study repeated this test using a more statistically powerful approach: we tested more subjects with a choice between two temperatures that were more different from each other. Our results corroborated the past findings and show that wild-born vampire bats will feed on blood as cold as 5 °C, even in the presence of blood at a more natural warm temperature. Clearly, heat is not the only cue that vampire bats use to select a blood source. The subjects presumably detected and chose spouts using other modalities such as olfaction and echolocation, and thermoperception is thus only one aspect of a vampire bat’s multimodal assessment of a bite site or blood source.

Captive-born vampire bats could feed on a live animal (Table 1), but it is difficult to interpret their feeding performance (Table 2). Rapid urination causes large changes in body mass that can complicate the measurement of blood consumption in vampire bats (Wimsatt & Guerriere, 1962). Nevertheless, the bats in this test appeared to feed less than expected under typical circumstances. Even when only considering the bats that did gain mass, the average gain (∼3 g, Table 2) was still less than half of the mass gains observed in the wild (∼6 g, Turner, 1975) or in captivity (∼9 g, Breidenstein, 1982).

Two of the three wild-born, short-term captive vampire bats fed on the chicken successfully (Table 2). Unexpectedly, the adult long-term captive mothers did not feed on the chicken at all, either when tested alone (three trials), or with their daughter (six trials). We also never observed the adults feeding on the chicken in six pilot trials with alternative experimental setups. There are several possible explanations for why none of the mothers fed nor even seemed motivated to feed. Older captive bats may have been less motivated to feed due to differences in metabolic demands, but this seems unlikely given the susceptibility of vampire bats to fasting (Wilkinson, 1984; Freitas et al., 2013). The adult bats had been fasted in trials for another experiment to induce food sharing (following Carter & Wilkinson, 2013), so they might have habituated or learned that they would eventually be released back into the main cage due to their experiences of being repeatedly isolated overnight without food. However, the captive-born bats had similar experiences.

Perhaps the most likely explanation is that adult vampire bats are more neophobic and less exploratory than younger vampire bats. A previous study showed that the same captive-born bats were an order of magnitude more likely to explore a novel object compared to their adult groupmates (Carter et al., 2018). In the feeding trials of this study, the younger captive-born bats in the experimental arena were more active than the adults with regards to jumping, walking, sniffing, and interacting with the chickens (mean hours of activity for adults = 5.1 h, 95% CI [4.1–6.3] h, n = 10; for young captive-born bats = 7.2 h; 95% CI [6.9–7.5] h, n = 12; Video S2). When tested alone, all 12 captive-born young explored the floor of the experimental area, whereas only one of three mothers did. When tested in mother-daughter pairs, five of the six captive-born daughters and three of the six wild-born mothers moved to the floor. Most of the captive-born bats left the refuge a few minutes after the experimenter left, whereas most of the adult bats did not explore the floor of the test cage at all during the whole night. When tested with her daughter, one of the mothers did not even leave the refuge the entire night. These observations are consistent with past tests showing age-dependent exploration or neophobia (Carter et al., 2018).

An anecdotal observation of age-dependent boldness in a novel feeding context comes from a different captive colony (described in Carter & Wilkinson, 2013). In this colony, one of the authors (GGC) noticed that several of the ten captive-born younger bats would often fly and land on the author when he entered the flight cage, sometimes climbing along his back or up and down his legs. When he held still, one vampire bat attempted to bite his ear. The same behavior was reported by another animal caretaker. In sharp contrast, none of the 22 older adult captive-born bats in the group approached or landed on people entering the flight cage, and they instead typically flew away to a corner or remained vigilant and motionless.

Despite the apparent difference in performance, the actual biting behavior of captive-born bats was generally consistent with the wild-born, short-term captive adults, and with past descriptions of feeding behaviors where individuals took anywhere from a few minutes to 40 min to bite an animal (Greenhall, Schmidt & Lopez-Forment, 1971; Greenhall, 1972; Greenhall, 1988). The chickens were only bitten on the digits, ankle, and areas near the tail that lacked feathers. This is consistent with observations that vampire bats readily target unprotected areas of skin (Greenhall, 1988). To test for more subtle differences in feeding performance, one needs a larger behavioral sample from captive-born and wild-born bats tested under controlled conditions.

In conclusion, naïve young without any relevant past experience can feed on live prey, but it remains unclear how the feeding development and performance of captive-reared vampire bats compares with more experienced wild-reared adults. Our results suggest that social learning from mothers plays at most a supplementary role in the acquisition of the flexible extractive foraging skills of vampire bats.

Conclusions

In this study, we took advantage of a long-term captive colony to test some ideas about the development of feeding behaviors in common vampire bats. We had two unexpected results. First, we observed no temperature preference for blood in either captive-born, long-term captive, or recently wild-capture vampire bats. When vampire bats fed from spouts, they did not generalize their thermal preferences for bite sites on live animals. Thermal cues used during normal feeding are therefore not used in every feeding context. Future studies could incorporate trials where both captive-born and wild-born vampire bats are presented with bite sites on live animals that vary in temperature or other traits (following Schmidt & Manske, 1982).

Second, naïve captive-born common vampire bats can feed on live prey. Further studies could test how lack of experience influences feeding performance. One possible improvement to our study design would be to fully habituate all subjects to the feeding arena, by allowing isolated bats to feed from a dish of blood in the arena over several nights, before presenting a live animal.

The role of learning in other bat species and mammals is not easy to predict. For example, frog-eating bats (Trachops cirrhosus) feed by eavesdropping on the mating calls of several prey species, such as katydids and frogs, some of which are toxic (Page & Jones, 2016). These bats possess traits that suggest adaptation for hunting frogs, such as increased innervation in the part of the cochlea that allows for low-frequency hearing, allowing them to hear in the range of frog advertisement calls (Bruns, Burda & Ryan, 1989). Under such conditions, where predators have adaptations for specific kinds of prey and mistakes are costly, one would expect that preferences for specific frog calls would be fixed. More than a decade of work, however, shows the opposite trend: frog-eating bats are highly flexible, with the ability to learn both asocially and socially, rapidly acquiring and reversing associations between calls and prey quality (Page & Ryan, 2005; Page & Ryan, 2006; Jones et al., 2013; Patriquin et al., 2018). In contrast, other studies show that bats also possess largely innate heuristics under unexpected conditions. For example, naïve captive-born insectivorous bats treat any sufficiently large horizontal smooth surface as water, even a surface that they can fly beneath (Greif & Siemers, 2010). Naïve captive-born flower-visiting Glossophaga bats are strongly and innately attracted to dimethyl disulfide, a component of many neotropical bat-pollinated flowers (Helversen, Winkler & Bestmann, 2000; Carter, Ratcliffe & Galef, 2010). Yet, the performance of these bats when learning simple cue-based associations and generalizing them to new environments is surprisingly bad (Stich & Winter, 2006), and this is because both innate preferences for dimethyl disulfide and learned preferences for novel cues can be overshadowed by spatial memory (Thiele & Winter, 2005; Stich & Winter, 2006; Carter, Ratcliffe & Galef, 2010).

Another purpose of our study was to assess whether captive-born vampire bats would be likely to survive after being released into the wild. Captive-born animals are typically less likely to survive than wild-caught individuals (Stoinski et al., 2003), but there is much variation between species and studies (e.g., Stoinski et al., 2003; Piep et al., 2008; Benson-Amram et al., 2014; Rogers et al., 2016; Abu Baker et al., 2018; Yang et al., 2018). Very little is known about how well captive-reared bats can acquire the necessary skills for foraging in the wild (Courts, 1997; Constantine, 2003; Ruffell & Parsons, 2009; Serangeli et al., 2012). Some insectivorous bat species can apparently go from captive hand-feeding to successful wild foraging, whereas others must first learn how to forage on flying insects in a flight cage (Kelly et al., 2008; Kelly et al., 2012; Serangeli et al., 2012). Captive-reared animals develop a smaller hippocampus (LaDage et al., 2009; Tarr et al., 2009), and are likely to be cognitively and physiologically different in other ways. Six of the captive-born vampire bats from our study were released back into the wild and observed visually and tracked with automated proximity loggers (Ripperger et al., 2016 Ripperger et al., 2019; in prep) until they left the site. We confirmed that all these bats survived for at least three days in the wild and four of the six survived for at least five or six days, before departing the site. We observed no deaths. These observations suggest that the captive-born bats were also capable of feeding on wild prey.

Supplemental Information