**3. Caves as hibernacula**

Hibernation, an optimal adaptation to a prolonged fall in temperature and reduction in prey availability, is a characteristic of the annual cycle of insectivorous temperate zone bats [16]. Selection of a suitable hibernation site is crucial for overwinter survival and, in temperate zone, caves and mines tend to be the most common hibernacula. Caves can be divided into three basic types based on microclimate and use by bats: (1) warm caves used during the summer, including maternity colonies, (2) relatively cold hibernacula with a stable micro‐ climate and (3) caves used during the autumn swarming [6]. Of course, both warm caves and hibernacula can also be used during the spring and autumn migrations too. At higher latitudes, cave temperatures are too low and they tend to be used mainly during torpor and hibernation. Note, however, that while thousands of bats can hibernate at such sites, those sites with lower numbers may be very important locally and their overall contribution to bat population great [7].

(4) *Summer period (mid‐June–end of July)*. During this period, the cave is used only sporadically (**Figure 2**), though the bats visiting the roost stay the whole night, i.e. they enter before mid‐ night and leave after midnight. This type of activity suggests that, during this period, the cave may be being used as a night roost between peaks in foraging activity or as a transitional day roost [11, 18]. At this time, the cave entrance is visited almost exclusively by males [34, 35] as adult females occupy maternity roosts during lactation and return to these between foraging

Flight activity at the night roost entrance is influenced by fluctuation in ambient temperature, rather than any absolute temperature threshold, the higher the difference between maximum and minimum daily temperature, the higher the activity level. This corresponds with a model proposing that activity changes in temperate insectivorous bats reflect changes in insect activ‐ ity [8], i.e. if day‐insect abundance is high due to warmer nights, bat foraging activity may continue overnight with no visits registered at the cave entrance (low activity). On the other hand, when nights are cooler and the daily temperatures range is higher, bats will tend to spend more time in the night roost. Foraging activity is highest at dusk and just before dawn, after which the bats return to the day roost [36]. This model is also supported by the influence of rainfall, with flight activity at the cave entrance increasing as rainfall increases whether the

(5) *Autumn migration or swarming period (late July–mid‐November)*. This period is typified by very high general activity and an increasing number of bats entering the cave. With the break‐ up of the summer breeding colonies, activity at the cave entrance gradually increases as adult females and juveniles arrive [9, 38], often in small groups of 2–12. The majority of bats does not roost in the cave and probably arrive after the first foraging period; hence, peak activ‐ ity tends to occur around midnight. Activity around the cave entrances in autumn probably enables juveniles to recognize potential hibernacula and to meet individuals of the opposite sex, which live separately during summer (e.g. [15]). Activity level is positively related to average daily temperature, atmospheric pressure and rainfall. Thus, when nights are warm and insect activity is high (high atmospheric pressure), the bats will quickly catch enough prey and will search for the cave entrances (swarming sites) in order to mate or obtain shelter

Hibernation, an optimal adaptation to a prolonged fall in temperature and reduction in prey availability, is a characteristic of the annual cycle of insectivorous temperate zone bats [16]. Selection of a suitable hibernation site is crucial for overwinter survival and, in temperate zone, caves and mines tend to be the most common hibernacula. Caves can be divided into three basic types based on microclimate and use by bats: (1) warm caves used during the summer, including maternity colonies, (2) relatively cold hibernacula with a stable micro‐ climate and (3) caves used during the autumn swarming [6]. Of course, both warm caves and hibernacula can also be used during the spring and autumn migrations too. At higher

bouts, night roosts being used sporadically and for brief periods [36, 37].

nights are warm or cold.

56 Cave Investigation

it be raining [14, 22].

**3. Caves as hibernacula**

More than 1200 caves are located in the Moravian and Javoříčský Karst regions of the Czech Republic, many of which host significant and regularly monitored bat hibernacula (**Figure 4**). Three of these cave systems (Javoříčské, Sloupsko‐šošůvské and Býčí skála) represent the largest bat hibernacula in the Czech Republic [39], with 17 bat species registered during hiber‐ nation, including rare species such as *Rhinolophus ferrumequinum*, the northern bat *Eptesicus nilssonii*, and the pond bat *Myotis dasycneme*. A similarly rich bat fauna has only been found in caves in the Slovak Karst and the Muránská Planina [40], both of which are also located along the northern distribution border of some bat species (e.g. Geoffroy's bat *Myotis emarginatus* or the lesser mouse‐eared bat *Myotis blythii*).

Both of these karst systems have a long history of bat research, beginning with speleologi‐ cal research of caves made by Dr. Friedrich Anton Kolenati in the second half of nineteenth century [41]. Modern bat research in the region was initiated by Prof. RNDr. Jiří Gaisler in the 1950s and it continues, including our long‐term research of bat hibernation, to the pres‐ ent day. As a result, some of these hibernacula have been monitored for almost 50 years [42].

**Figure 4.** Main entrance of Sloupsko‐šosůvské cave representing one of the largest bat hibernacula in the Czech Republic. Photo by Leos Stefka.

As one of the main requirements of our own research was to avoid any disturbance to hibernating bats, we used visual censuses only (including night censuses using Pathfinder 2000s night‐vision scope) with no handling or marking [10, 29]. Thermal profiles were also undertaken to evaluate physiological condition. Fur surface body temperature, which is correlated with core body temperature, was measured using a Raynger MX2 non‐contact IR thermometer (Raytek Corporation, USA). Two major model species were regularly moni‐ tored in the caves, the greater mouse‐eared bat (*Myotis myotis*) and the lesser horseshoe bat (*R. hipposideros*) (**Figure 5**), these being typical members of the bat community hibernating in the Moravian Karst [9, 10, 43].

### **3.1. Model of bat hibernation in natural caves**

In late summer and early autumn, bats undergo a preparation phase for hibernation dur‐ ing which they rapidly accumulate body fat deposits [44] needed for surviving the torpor period. The fat is accumulated by energy savings achieved through increasingly longer daily torpor bouts during the diurnal resting period. Hibernation is usually interrupted by periodic arousals [45, 46], usually related to drinking, feeding (in mild periods) or even mating [23, 35]. As part of the fat deposits must be metabolized for torpid individu‐ als to become physiologically active during winter, such arousals are energetically costly [47, 48].

These arousals, and any subsequent activity, will be mirrored in ecological parameters such as community structure, bat population abundance, shelter selection or total movement activ‐ ity. Monitoring of hibernating bats in the Moravian Karst has confirmed that the ratio of 'visible' bats changes through the winter, i.e. bats may move from inaccessible shelters to places where they can be monitored by investigators [9, 49]. The total number of hibernating bats grows continuously from October, with highest abundances occurring in February or

**Figure 5.** Hibernating clusters of two bat species regularly monitored in the Moravian Karst. (A) The greater mouse‐ eared bat (*Myotis myotis*) (body length of 6.5–8 cm) and (B) the lesser horseshoe bat (*Rhinolophus hipposideros*) (body length of 3.5–4.5 cm).

March, depending on community structure. Any increase in abundance will be influenced by immigration of newcomers during the pre‐hibernation period only (mid‐November–mid‐ December). Switching of hibernation sites during the deep hibernation period (i.e. leaving the hibernaculum) has only been registered exceptionally [21]. In April, there is a gradual but relatively rapid emergence from the hibernation sites (approximately 3 weeks), with bat abundance in cave decreasing to a minimum.

As one of the main requirements of our own research was to avoid any disturbance to hibernating bats, we used visual censuses only (including night censuses using Pathfinder 2000s night‐vision scope) with no handling or marking [10, 29]. Thermal profiles were also undertaken to evaluate physiological condition. Fur surface body temperature, which is correlated with core body temperature, was measured using a Raynger MX2 non‐contact IR thermometer (Raytek Corporation, USA). Two major model species were regularly moni‐ tored in the caves, the greater mouse‐eared bat (*Myotis myotis*) and the lesser horseshoe bat (*R. hipposideros*) (**Figure 5**), these being typical members of the bat community hibernating

In late summer and early autumn, bats undergo a preparation phase for hibernation dur‐ ing which they rapidly accumulate body fat deposits [44] needed for surviving the torpor period. The fat is accumulated by energy savings achieved through increasingly longer daily torpor bouts during the diurnal resting period. Hibernation is usually interrupted by periodic arousals [45, 46], usually related to drinking, feeding (in mild periods) or even mating [23, 35]. As part of the fat deposits must be metabolized for torpid individu‐ als to become physiologically active during winter, such arousals are energetically costly

These arousals, and any subsequent activity, will be mirrored in ecological parameters such as community structure, bat population abundance, shelter selection or total movement activ‐ ity. Monitoring of hibernating bats in the Moravian Karst has confirmed that the ratio of 'visible' bats changes through the winter, i.e. bats may move from inaccessible shelters to places where they can be monitored by investigators [9, 49]. The total number of hibernating bats grows continuously from October, with highest abundances occurring in February or

**Figure 5.** Hibernating clusters of two bat species regularly monitored in the Moravian Karst. (A) The greater mouse‐ eared bat (*Myotis myotis*) (body length of 6.5–8 cm) and (B) the lesser horseshoe bat (*Rhinolophus hipposideros*) (body

in the Moravian Karst [9, 10, 43].

[47, 48].

58 Cave Investigation

length of 3.5–4.5 cm).

**3.1. Model of bat hibernation in natural caves**

Movement activity of bats inside the hibernaculum, expressed as the percentage of new find‐ ings during a visit, is registered throughout the winter, with levels fluctuating in our spe‐ cies‐specific models. Hibernation activity of *R. hipposideros,* for example, could be divided into three distinct periods reflecting early, deep, and late hibernation; while *M. myotis* movement activity remained relatively high throughout the season [30]. A continuous arrival of bats at the hibernaculum means that *R. hipposideros* abundance increased gradually over the 6–8 weeks leading to mid‐December, and decreased again from mid‐March as they gradually left (**Figure 6**). The deep hibernation period was characterized by low movement activity in the cave and minimal changes in abundance, as also confirmed by detection of ultrasound sig‐ nals [17, 27]. Even in the middle of winter, when the conditions outside were suitable, some awakened *R. hipposideros* became aroused and left the cave, shortly to return again [17, 50].

**Figure 6.** Changes in abundance of two model species during hibernation period 2011/2012. Data presented in 2‐week periods.

Abundance of *M. myotis*, on the other hand, increased continuously throughout the winter, eventually dominating the bat community by the end of hibernation period. In comparison, this species leave the hibernacula rapidly, all bats having disappeared over a period of just 3 weeks [49, 51].

Our two model species accounted for more than 80% of all bat observations in the caves. Bat netting at the cave entrances during spring and autumn migrations, however, confirmed a much higher diversity than during hibernation, with other bat species showing a higher dom‐ inance. Small species of genus *Myotis,* such as *M. emarginatus*, *M. daubentonii*, Natterer's bat *Myotis nattereri*, and Bechstein's bat *Myotis bechsteinii*, are often underestimated during winter monitoring [9] as they tend to use more or less inaccessible roost sites (e.g. deep crevices) [34], depending on the local microclimate, species‐specific requirements, season or weather. We found that around 20% of all bats hibernating in natural caves need to be monitored during winter as the cumulative number of bats entering the cave (calculated using a double IR‐light logging system) was much higher (**Figure 7**).

#### **3.2. Shelter selection during hibernation**

As roost site characteristics can play an important role in bat thermoregulation, choice of site will undoubtedly influence bat fitness and survival. Ransome [52] classified caves used

**Figure 7.** Cumulative number of bats entering the cave (winter season 2000/2001) recorded by the automatic IR‐light logging system (area) and the numbers of bats hibernating inside the cave monitored during winter counts (bars).

as hibernation sites into three basic types depending on temperature fluctuation: (1) caves displaying a constant temperature regime, (2) caves with dynamic temperatures and (3) caves with fluctuating low temperatures. Note, however, that numerous factors affect the climate of individual caves; and that each cave will be unique in its geomorphologic and microcli‐ matic parameters [6]. Caves with more or less constant temperatures over the year (averaging between 6 and 10°C) usually have just one entrance and temperature fluctuation tends to occur in the outer entrance parts only due to high air flow. Thermally dynamic caves are char‐ acterized by large passages with different temperatures. Such caves tend to have two or more entrances, their mutual positions influencing internal temperature conditions. As any two caves will differ significantly, therefore, it will be difficult to specify an average annual tem‐ perature. In general, average annual temperature will be in the range of 3–14°C. Hibernating bat communities sheltering in such caves tend to show the most stable abundances. The third cave type always tends to display fluctuating temperatures, despite usually having just one entrance. During winter, air temperature will decrease significantly due to cold air flowing in from the cave entrance [53, 54].

Abundance of *M. myotis*, on the other hand, increased continuously throughout the winter, eventually dominating the bat community by the end of hibernation period. In comparison, this species leave the hibernacula rapidly, all bats having disappeared over a period of just 3

Our two model species accounted for more than 80% of all bat observations in the caves. Bat netting at the cave entrances during spring and autumn migrations, however, confirmed a much higher diversity than during hibernation, with other bat species showing a higher dom‐ inance. Small species of genus *Myotis,* such as *M. emarginatus*, *M. daubentonii*, Natterer's bat *Myotis nattereri*, and Bechstein's bat *Myotis bechsteinii*, are often underestimated during winter monitoring [9] as they tend to use more or less inaccessible roost sites (e.g. deep crevices) [34], depending on the local microclimate, species‐specific requirements, season or weather. We found that around 20% of all bats hibernating in natural caves need to be monitored during winter as the cumulative number of bats entering the cave (calculated using a double IR‐light

As roost site characteristics can play an important role in bat thermoregulation, choice of site will undoubtedly influence bat fitness and survival. Ransome [52] classified caves used

**Figure 7.** Cumulative number of bats entering the cave (winter season 2000/2001) recorded by the automatic IR‐light logging system (area) and the numbers of bats hibernating inside the cave monitored during winter counts (bars).

weeks [49, 51].

60 Cave Investigation

logging system) was much higher (**Figure 7**).

**3.2. Shelter selection during hibernation**

Survival of hibernating bats will be influenced not only by the selection of a suitable hibernac‐ ulum but also by the specific microhabitat conditions within. The correct choice will be crucial for the efficient use of stored energy and for the appropriate timing of flight activity. Indeed, studies have shown that bats are able to regulate length and depth of torpor by selecting favourable sites [46]. During our own monitoring of hibernating bats, we monitored a range of parameters including site type (exposed, semi‐exposed and hidden), relative height above the floor and position in the cave [30]. During hibernation, *R. hipposideros*, a thermophilic species, were registered in practically all parts of the caves under study (Kateřinská and Sloupsko‐ šošůvské), with the exception of the entrances, which have more dynamic microclimates [30, 49]. The bats tended to prefer low shelters (under 3 m from the floor) and always hibernates hanging free in open unprotected sites, regardless of season or hibernaculum type. The sites selected by *R. hipposideros* had stable temperature and humidity conditions with minimal air flow and, as a result, the species showed very low movement activity levels during deep hibernation. The pre‐ and post‐hibernation periods, on the other hand, are typified by high movement activity. Changes in shelter types during winter corresponded to phases in the annual cycle and the physiological status and behaviour of bats only and not to any changes in the environment, there being no temperature fluctuation in the deep cave sections used by *R. hipposideros* [16, 55]. Despite hibernating next to a footpath frequently used by tourists or spe‐ leotherapy patients, we failed to register any vulnerability of *R. hipposideros* to human activity.

Euryvalent *M. myotis*, on the other hand, were registered throughout the cave systems during hibernation, using all shelter types indiscriminately (exposed and hidden, ceiling and walls) and showing high seasonal dynamics. During deep hibernation (mid‐December–early April), these bats are continuously moving into the outer parts of the cave where they select specific sites for the formation of clusters (**Figure 8**). Over 80% of all *M. myotis* hibernating in the Sloupsko‐šošůvské caves, for example, were found in one specific area during late hiberna‐ tion [49]. A shift towards the cave entrance has also been reported in other European hiber‐ nacula [56, 57]. Movement activity of *M. myotis* was relatively high in hibernacula throughout the hibernation period and could not be divided into specific periods [30]. In the absence of

**Figure 8.** Percentage of the greater mouse‐eared bats *Myotis myotis* hibernating in the Kateřinská cave entrance during winter 1992/1993. Data presented in 2‐week periods.

food, *M. myotis* select sites with a constant temperature for deep hibernation in order to maxi‐ mize energy savings. On the other hand, the bats most likely shift to sites with more dynamic temperature regimes during the late hibernation period as the changes in ambient tempera‐ ture help bats synchronize arousals with actual weather conditions. In doing so, aroused bats are able to couple emergence activity with favourable climatic conditions for foraging.

It is apparent that neither all individuals nor all populations have the same model of hiberna‐ tion [58]. Our studies suggest that *M. myotis*, at least at population level, may not follow the same hibernation model and that a range of hibernation strategies (i.e. level of movement activity, preference for different shelter types) may be used depending on the prevalent dif‐ ferent microclimate profile (i.e. dynamic or stable). Populations in different hibernacula will exhibit responses tuned to that environment, while individuals of the same species may vary in the strategy used to survive hibernation. In this way, the bats optimize their depletion of energy reserves and improve their chances of surviving in the winter [59]. High fidelity to particular underground shelters also suggests that the adopted hibernation strategy may limit bats to repeated use of particular hibernacula [60].

#### **3.3. Cave temperature and bat hibernation**

The length of time that temperate bats can survive without feeding will be dictated by the temperature, at which they hibernate. In general, by hibernating in caves where temperatures are low but above freezing (i.e. between 2 and 5°C), the bat's metabolism rate is maintained at an efficient level. While the actual temperatures at which different bats hibernate is species specific [61, 62], the interspecific differences are very small due to the low metabolism and small body mass of temperate bats. Such species‐specific differences vary seasonally, being somewhat smaller during deep hibernation and greater during the pre‐ and late hibernation periods. Bats also display intraspecific variations in preferred temperature, as individuals will select locations based on their energy reserves [63].

Bat arousal may occur as a result of temperature changes in hibernacula, following which the bats may move to a more suitable location [64]. In general, bats prefer to start hibernation at sites with higher temperatures as those with low temperatures may reach freezing point over the coldest months. An optimizing strategy of such type has been observed in *M. myotis* in natural karstic caves (**Figure 9**). As bats often return to the same sites year‐after‐year, this could suggest the use of prior experience, learning from others, and/or olfactory clues in microhabitat selection. Arousals are also temperature dependent, with the length and fre‐ quency of bat arousals increasing with temperature increases over 10°C [46].

During hibernation, bat body temperature falls to within 1–2°C of ambient temperature and meta‐ bolic processes slowdown, thereby reducing energy requirements. As a result, hibernation incurs physiological costs, including the build‐up of metabolic wastes, dehydration, reduced motor func‐ tion, altered immune response, and sleep deprivation [65]. Hibernation may also impose ecologi‐ cal costs such as decreased detection and response to predators [66] and an increased likelihood of freezing [67]. At the cellular level, cold stress changes cellular membrane lipid composition and suppresses the rate at which protein synthesis and cell proliferation takes place [68]. We examined

food, *M. myotis* select sites with a constant temperature for deep hibernation in order to maxi‐ mize energy savings. On the other hand, the bats most likely shift to sites with more dynamic temperature regimes during the late hibernation period as the changes in ambient tempera‐ ture help bats synchronize arousals with actual weather conditions. In doing so, aroused bats are able to couple emergence activity with favourable climatic conditions for foraging.

**Figure 8.** Percentage of the greater mouse‐eared bats *Myotis myotis* hibernating in the Kateřinská cave entrance during

It is apparent that neither all individuals nor all populations have the same model of hiberna‐ tion [58]. Our studies suggest that *M. myotis*, at least at population level, may not follow the same hibernation model and that a range of hibernation strategies (i.e. level of movement activity, preference for different shelter types) may be used depending on the prevalent dif‐ ferent microclimate profile (i.e. dynamic or stable). Populations in different hibernacula will exhibit responses tuned to that environment, while individuals of the same species may vary in the strategy used to survive hibernation. In this way, the bats optimize their depletion of energy reserves and improve their chances of surviving in the winter [59]. High fidelity to particular underground shelters also suggests that the adopted hibernation strategy may

The length of time that temperate bats can survive without feeding will be dictated by the temperature, at which they hibernate. In general, by hibernating in caves where temperatures are low but above freezing (i.e. between 2 and 5°C), the bat's metabolism rate is maintained at an efficient level. While the actual temperatures at which different bats hibernate is species

limit bats to repeated use of particular hibernacula [60].

**3.3. Cave temperature and bat hibernation**

winter 1992/1993. Data presented in 2‐week periods.

62 Cave Investigation

**Figure 9.** Changes in greater mouse‐eared bat *Myotis myotis* body temperature and shelter temperature during the winter of 2002/2003. Explanations: box = interquartile range; middle point = median; whiskers = non‐outlier range; circles = outliers; stars = extremes; continuous line = average shelter temperature.

the ability of primary skin fibroblast cells from the flying membrane of a hibernating *M. myotis* to proliferate under torpor and euthermia. After loosening the tissue mechanically (without pro‐ teases), the cells were identified as fibroblasts based on their spindle shape, positive staining for the vimentin mesenchymal marker, and the presence of typical stress‐fibre organization in the actin cytoskeleton. Cell numbers for the assay started with 20,000 cells per well and these were incubated at 9 or 37° C for 6 days in a 5% CO<sup>2</sup> humidified environment for the experiment. Cells were detached from the cultivation wells and recalculated daily with 30 times repetition. While bat fibroblasts cultured at 37°C were elongated reached high numbers in 6 days, and attached success‐ fully to the well substrate; those cultured at 9°C were spherical, reduced in number and took time to attach (**Figure 10**). Extrapolation from this cellular *in vitro* model suggests that bat fibroblasts have some proliferative capacity at the temperature conditions prevalent during torpor, though wound healing capacity would be much slower than in euthermic animals. Such a physiological response of bat cells may help explain the movements registered in *M. myotis* at low fur tempera‐ tures (*T*flow < 5°C) [69], which would allow bats to save energy long‐term and prolong torpor bouts. All *T*flow events were recorded during late hibernation, when bats are faced with an acute shortage of energetic reserves and enormous metabolic requirements. In most cases, *T*flow events were rep‐ resented by slow displacements between clusters of bats, though departure or arrival to and from clusters was also recorded with no elevation in body temperature (**Figure 11**). Repeat appearances suggest that *T*flow movements may represent a regular part of bat hibernation tactics.

**Figure 10.** Numbers of primary cell (fibroblast) cultured at two different temperatures showing some proliferative capacity at the temperature conditions prevalent during torpor.

the ability of primary skin fibroblast cells from the flying membrane of a hibernating *M. myotis* to proliferate under torpor and euthermia. After loosening the tissue mechanically (without pro‐ teases), the cells were identified as fibroblasts based on their spindle shape, positive staining for the vimentin mesenchymal marker, and the presence of typical stress‐fibre organization in the actin cytoskeleton. Cell numbers for the assay started with 20,000 cells per well and these were

were detached from the cultivation wells and recalculated daily with 30 times repetition. While bat fibroblasts cultured at 37°C were elongated reached high numbers in 6 days, and attached success‐ fully to the well substrate; those cultured at 9°C were spherical, reduced in number and took time to attach (**Figure 10**). Extrapolation from this cellular *in vitro* model suggests that bat fibroblasts have some proliferative capacity at the temperature conditions prevalent during torpor, though wound healing capacity would be much slower than in euthermic animals. Such a physiological response of bat cells may help explain the movements registered in *M. myotis* at low fur tempera‐ tures (*T*flow < 5°C) [69], which would allow bats to save energy long‐term and prolong torpor bouts. All *T*flow events were recorded during late hibernation, when bats are faced with an acute shortage of energetic reserves and enormous metabolic requirements. In most cases, *T*flow events were rep‐ resented by slow displacements between clusters of bats, though departure or arrival to and from clusters was also recorded with no elevation in body temperature (**Figure 11**). Repeat appearances

suggest that *T*flow movements may represent a regular part of bat hibernation tactics.

**Figure 10.** Numbers of primary cell (fibroblast) cultured at two different temperatures showing some proliferative

capacity at the temperature conditions prevalent during torpor.

humidified environment for the experiment. Cells

incubated at 9 or 37° C for 6 days in a 5% CO<sup>2</sup>

64 Cave Investigation

**Figure 11.** Examples of low body temperature movements, showing the bat moving between clusters from left to right. The upper thermal images correspond with lower images from photo‐traps recorded simultaneously (a–c). The rectangles in the lower photo‐trap images indicate the position of the thermal image, while the dotted circles indicate the moving bat. *Source*: [69].
