DISCUSSION
In our study, we found that light intensity at emergence, return, sunset, andd sunrise influenced the emergence and return times ofV. sinensis in different ways, supporting our first hypothesis. Moreover, V. sinensis departed from the roost earlier regardless of the first emergence and mid-emergence times during the lactation period, suggesting support for our second hypothesis. Finally, the mid-emergence time of V. sinensis was earlier when predators were hunting, but the final return time was later when predators were present. This result was inconsistent with our third hypothesis.
Factors affecting an animal’s energy status and the value of additional energy intake are major determinants of risk-taking behaviour (Brown, 1988; Caro, 2005; Lima, 1998). In general, the first emergence of bats occurs after sunset (Acharya, Racey, McNeil, Sotthibandhu, & Bumrungsri, 2015; Arndt. et al., 2018; Lee & McCracken, 2001; Welbergen, 2006) and return events occur before sunrise (Acharya et al., 2015; Lee & McCracken, 2001). During lactation, reproductive female bats would emerge before sunset because of the higher energy demand (Arndt. et al., 2018; Lee & McCracken, 2001). This behaviour has been noted in hoary bats, Lasiurus cinereus(Barclay, 1989), Mexican free-tailed bats,Tadarida brasiliensis(Lee & McCracken, 2001), the grey-headed flying fox, Pteropus poliocephalus(Welbergen, 2006), Daubenton’s bats,Myotis daubentonii (Lučan, 2009), dawn nectar bats, Eonycteris spelaea(Acharya et al., 2015), and Indiana bats, Myotis sodalist(Arndt. et al., 2018). Additionally, during lactation, reproductive female bats also return later to the day roost than during post-lactation, such as in T. brasiliensis(Lee & McCracken, 2001) and Dawn nectar bats, E . spelaea (Acharya et al., 2015). Consistent with these previous studies, our results showed that the lactation period significantly influenced the first emergence and the mid-emergence of V. sinensis during the lactation period. Specifically, 64.29% of emergence events started before sunset during the lactation period, while only 5.88% of emergence events started before sunset during the post-lactation period. The earlier onset of activity and the delay of ending activity increases the foraging time at dusk and dawn (Rydell et al., 1996). To summarize, the higher energy demands may prompt bats to depart from the roost earlier and return to the roost later during the lactation period.
There are many benefits to an early start to night activity, but for bats, doing so in bright light may be a risky option. Large flocks of bats emerging from the day roost and returning to the day roost would attract the attention of diurnal avian predators (Fenton et al., 1994). Predation on bats by birds is seemingly rare and opportunistic at dusk and dawn, especially in temperate zones, but predation would affect many aspects of bats’ behaviour, such as roost selection and foraging activity (Fenton et al., 1994; Lima & O’Keefe, 2013). Bats emerging around dusk before sunset under high light intensity would be exposed to higher predation risk than the bats emerging in true night under low light intensity. As for earlier emerging bats, later returning bats also would be exposed to high predation risk from diurnal predators (Fenton et al., 1994). However, previous studies have mainly assessed predation risk without observing real predators. In this case, the effects of predation on activity rhythms of bats may not be universal. In this study, we found that the mid-emergence time of V. sinensis was earlier when predators were hunting, but the final return time was later when predators were present. The results may challenge the previous view that high predation risk leads to later emergence and the earlier return. This can be explained by the following reasons: First, the previous studies investigated the effects of predation risk on behaviour of bats based on empirical estimation or predator models, but here F. amurensis hunt V. sinensis every day at this roost. Thus, the different results may not be inconsistent. Second,V. sinensis usually gives birth to twins in each litter (Jin. et al., 2012), and the energy demands are greater during the lactation period than in other periods or in bat species with one offspring per litter. Thus, even though V. sinensis may suffer higher risk of predation, they may need to depart from the roost earlier and return to the roost later during the lactation period due to the higher energy demands. The hierarchical partitioning analysis also confirmed that lactation periods had a greater contribution to the mid-emergence time and the final return time in V. sinensis than predation risk. Interestingly, predation risk only influenced the mid-emergence time rather than the first emergence time in V. sinensis. This may be because V. sinensis had to depart in quantity around the mid-emergence time due toF. amurensis successfully preying on bats in dim light (light intensity below 10 lux); this also may reduce the risk of predation on individuals via the dilution effect of a large flock of bats.
Bright light conditions may present a risky option for nocturnal animals, especially for bats. In this study, light intensity was the most important factor affecting the activity rhythm of V. sinensis t (Fig. 4). Specifically, differences between first emergence time and sunset time, and between mid-emergence time and sunset time, were significantly and positively associated with light intensity at sunset (Fig. 1a; Fig. 2a). A similar trend was detected in the relationship between final return time and sunrise (Fig. 3a). These results showed that V. sinensis may adjust the emergence and return times based on sunset and sunrise times, respectively. This may be helpful for avoiding F. amurensis for the following reasons. With the decrease of light intensity, the visual sensitivity of falcons decreases significantly (Fox., Lehmkuhle., & Westendorf., 1976). Therefore, it was clearly a safer behavioural strategy for bats to emerge from the roost under dim light conditions. Additionally, bright light exerts inhibitory effects on the activity of bats (A.Gutierrez, F.Pessoa, Aguiar, & M.A.Pesso, 2014). However, here the opposite trends of the effect of light intensity of first emergence on the emergence time (Fig. 1b; Fig. 2b) and of light intensity at final return on the final return time (Fig. 3b) were observed. With increases in light intensity of first emergence, both the first emergence and mid-emergence events became increasingly earlier relative to sunset, then changed slowly (Fig. 5a, b). Moreover, with increases in light intensity at final return, the final return events became increasingly later relative to sunrise, then changed slowly (Fig. 5c). These findings may be due to the high energy demands in V. sinensis during the lactation period prompting the bats to depart from the roost earlier and return to the roost later despite the high light intensity. We observed that 65.67% of first emergence events (44/67) started before sunset during the lactation period, supporting this view. Additionally, there is a threshold of light intensity in bat activity rhythms (Erkert, 2004), which may reflect the fact that leaving the day roost early at dusk or returning to the roost late at dawn was not the best option for the bats.
Weather conditions may also affect the emergence behaviour of bats, but the effects in the previous studies were not severe (Erickson & West, 2002; Frick et al., 2012; Welbergen, 2008). For example, the relationship between temperature and the start of emergence of bats depended on summer climatic conditions; therefore, the influence of daily temperature in drought conditions may be different than in normal or unusually moist years (Frick et al., 2012). Below a critical minimum temperature (8-10°C), the foraging behaviour of bats would be less beneficial than remaining torpid (Avery, 1985). Drought conditions were associated with low insect abundance (Hawkins et al, 1998), and thus were associated with earlier emergence and displaying risker behaviour (Frick et al., 2012). Consistent with these previous studies, our results indicated that temperature and humidity did not influence emergence or return times of V. sinensis. This may be because environmental temperatures at the roost of V. sinensis were higher than the critical minimum temperature during the monitoring period, and there was sufficient rainfall in summer. The relatively stable environment led to the relatively high insect abundance. In this case, the effects of temperature and humidity on activity rhythms of bats may be relatively weak.