Metabolic rate and body temperature (Tb) reduction during torpor can provide significant energy savings for bats during inclement weather and food scarcity. However, torpor use may slow down biochemical processes including fetal and juvenile development and sperm production. Sex-differences in the timing of reproductive activity of bats in the temperate climate zone should result in differences of the thermoregulation behaviour by males and females during summer. To test this hypothesis, we studied thermoregulation of free-ranging, tree-dwelling gleaning bats (Myotis bechsteinii) and trawling bats (M. daubentonii) during different reproductive periods. Gleaners and trawlers are able to forage on prey which is sitting on vegetation and the ground (gleaning) or which is slowly moving over water bodies (trawling). This prey is characterized by lower ambient temperature (Ta) dependent abundance than flying prey. We used temperature-sensitive radio transmitters to measure skin temperature (Tsk). Temperature telemetry over 144 census days revealed a significant effect of reproductive period and sex on Tsk. Pre-spermatogenic males exhibited a significantly greater Tsk reduction than females in early pregnancy. Males at the beginning of sperm production and in main spermatogenesis exhibited much more frequent and deeper temperature reductions than females in late pregnancy and in lactation. Lactating females maintained the highest Tsk of all bats. Post-lactating females reduced Tsk to the same extent or even more than males in advanced spermatogenesis.
Our findings indicate that the thermoregulation of gleaning and trawling temperate bats is likely to be much less influenced by environmental conditions than that of aerial hawking bat species. We suggest that both sexes of Bechstein's bats and Daubenton's bats primarily adapt their thermoregulation in response to current reproductive activity.
Introduction
Small-sized endotherms with large relative surface areas generally lose considerable heat when maintaining elevated Tb. Remaining homeothermic entails substantial energetic costs which are exacerbated during adverse weather conditions (Speakman & Rowland 1999, White & Seymour 2003, Clarke & Rothery 2008). Many small mammals and birds can reduce energy expenditure through reduction of Tb and metabolic activity during torpor (Geiser 2004, Heldmaier et al. 2004, Speakman 2007). For example, insectivorous bats from temperate regions hibernate during the cold winter season when food is typically not available and commonly use torpor during the active season from spring to autumn (Turbill et al. 2003a, Willis et al. 2005). Although the extent of heterothermy and hypometabolism is usually less pronounced during short bouts of torpor than in hibernation (Geiser & Ruf 1995, Wojciechowski et al. 2007), the short-term decrease in metabolism in 5 summer also leads to profound energy savings (Geiser 2004, Heldmaier et al. 2004). Entering torpor may lead to considerable physiological and ecological costs (Wojciechowski et al. 2007). Possibly, the greatest potential cost of employing torpor in summer is suppression of reproductive processes, in particular fetal development, milk production and body heat transfer from mother to offspring (Racey & Swift 1981, Wilde et al. 1999). In temperate latitudes, late birth and late weaning reduce the time necessary for females and young to acquire fat to fuel hibernation. The lack of sufficient fat deposition may result in higher mortality of young bats, particularly over winter (Thomas et al. 1990). Furthermore, prolonged use of torpor by pregnant females might lead to resorption or abortion of offspring (Roer 1973, Grindal et al. 1992). Male bats using torpor have to cope with costs including reduced testicular growth and maintenance, suppressed accessory gland activity and diminished spermatogenesis and copulation activity (Jolly & Blackshaw 1988, Hamilton & Barclay 1994, Grinevitch et al. 1995, Entwistle et al. 1998, Dietz & Kalko 2006). In temperate zone bats, reproductive stages are asynchronous between sexes (Racey 1974, Entwistle et al. 1998, Dietz & Kalko 2006). While females are pregnant in spring, lactating in mid-summer and post-lactating in late summer, males are sexually inactive in early spring. They exhibit increasing spermatogenetic activity from early to late summer when mating season begins (Racey 1974, Entwistle et al. 1998, Encarnação et al. 2004).
Sex-differences in reproductive activity should result in differences in thermoregulatory strategies by males and females during summer (Dietz & Kalko 2006). Most studies about thermoregulation of bats focus on only one sex (e.g. Audet & Fenton 1988, Entwistle et al. 1998, Lausen & Barclay 2003, Turbill et al. 2003a, b, Willis et al. 2005, Solick & Barclay 2006), encompass only one reproductive period (e.g. Cryan & Wolf 2003, Turbill 2006a, b) or were conducted in the laboratory (Kurta 1986, Kurta & Kunz 1987, Geiser et al. 2000, e.g., Geiser 2004). Results of the field studies about sex-differences in thermoregulation of Nearctic bats show that males use torpor more frequently than reproductive females (Hamilton & Barclay 1994, Grinevitch et al. 1995). However, given that the data are mostly limited to studies of bats living in anthropogenic structures, information on bats roosting in trees are lacking even though trees are the most common natural roost sites (Kunz & Lumsden 2003).
The goal of our study was to investigate thermoregulation during the entire breeding season by Palae arctic, tree roosting Bechstein's bat (Myotis bechsteinii) and Daubenton's bat (M. daubentonii). For both European bat species, old growth deciduous forests are important habitats because they provide many tree species in multiple stages of growth and decay for roosting (Schlapp 1990). M. bechsteinii differs from the trawling M. daubentonii mainly in the manner it forages, specifically in that they mostly glean arthropods e.g., Lepidoptera and Coleoptera from vegetation and the ground (Wolz 2002, Siemers & Kerth 2006, Siemers & Swift 2006). M. daubentonii is mainly adapted to hunt for insects above water surfaces where it feeds mostly on swarming or emerging Chironomidae (Swift & Racey 1983, Kalko & Schnitzler 1989). However, in contrast to aerial hawking bats that forage on flying prey in strong dependence of Ta conditions, gleaning and hawking bats like M. bechsteinii and M. daubentonii also forage successfully during periods with low Ta (Barclay 1991, Chruszscz & Barclay 2003, Dietz & Kalko 2007). For both bat species, we expected to find comparable sex-differences in thermoregulation. We expected Ta would affect thermoregulation of reproductive active bats less than actual reproductive condition. We hypothesised that in order to maximise fetal development and milk production, females should maintain high Tb during pregnancy and lactation period while torpor would be used predominantly in the post-lactation period. In contrast, adult males are expected to reduce Tb more often especially at low Ta during the energetically costly period for females (pregnancy and parturition), because they are not involved in the development and rearing of the young. However, as sperm production commences, thermoregulatory behaviour of the males should change accordingly.
Methods
Study sites
Our study of thermoregulation of Daubenton's bats took place in Philosophenwald, a deciduous forest near the city of Giessen (highest elevation: 274 m a.s.l.) in central Germany (50.35 N; 8.40 E). Mean annual precipitation and temperature of Giessen are 590 mm and 9.1°C respectively. Several ponds and the River Lahn are the most important feeding sites of Daubenton's bats that catch emerging and swarming prey over water surfaces. Philosophenwald is dominated by Fagus sylvatica which contain the woodpecker cavities, brunch breaks and crevices which M. daubentonii use as roosts. Thermoregulation of Bechstein's bats was investigated in a deciduous forest (“Friemholz”) in Luxembourg near the city of Echternach (49.80 N; 6.43 E). The forest is located at 350 m a.s.l. on a montane plateau in the “Little Switzerland” region of Luxembourg. Average annual precipitation and temperature of Echternach are 750 mm and 8.5°C. Friemholz is dominated by Quercus robur, Carpinus betulus and Fagus sylvatica. These tree species offer M. bechsteinii day and night roosts in form of woodpecker cavities and crevices.
Table 1.
Summary of radio-tracking data of Myotis bechsteinii and M. daubentonii sampled in Luxembourg and in Germany, respectively. Reproductive periods (RP) of males are pre-spermatogenesis (PS), early spermatogenesis (ES), main spermatogenesis (MS) and advanced spermatogenesis (AS). RPs of females are early pregnancy (EP), late pregnancy (LP), lactation (Lac) and post-lactation (Post-Lac).
Radio telemetry
With reference to studies on M. bechsteinii (Kerth et al. 2001, Dietz & Pir 2009), M. daubentonii (Encarnação et al. 2004) and other European Myotis species (Racey 1974) we divided the reproductive cycle of the bats into four periods. For females, we differentiated between early pregnancy (Tracking period I), late pregnancy (Tracking period II), lactation (Tracking period III) and post-lactation (Tracking period IV). Reproductive periods of males were delineated as pre-spermatogenesis (Tracking period I), early spermatogenesis (Tracking period II), main spermatogenesis (Tracking period III) and advanced spermatogenesis (Tracking period IV). For determination of reproductive status we caught bats in mist nets set near and around known roost trees and used the methods described by Racey (1974). Pregnant bats were determined by gently palpating the abdomen. Lactating bats were identified by a bare patch around their swollen nipples. Size and form of the nipples and growth of hair served as indicator for post-lactation. The reproductive status of males was determined by assessing the size of the testes, distension of the epididymis and coloration of the tunica vaginalis (Racey 1974, Encarnação et al. 2004). The bats were weighed to the nearest 0.1 g using a digital scale (Kern digital scale CM 60-2N, Kern and Sohn GmbH, Balingen-Frommem, Germany). Forearm length was measured with dial callipers (accuracy: 0.1 mm). The age of the bats was determined by evaluating the closure of the epiphysis (Anthony 1988), length of wrists, abrasion of teeth and evaluation of a chin spot (Richardson 1994).
We tagged Bechstein's bats and Daubenton's bats (Table 1) with LB-2T temperature-sensitive radio-transmitters (Holohil Systems Ltd., Carp., Ontario, Canada) to measure Tsk and to locate day roosts. The usefulness of temperature-sensitive transmitters for investigations of thermoregulatory behaviour of bats in the field was first demonstrated by Audet & Thomas (1996) and Barclay et al. (1996). They found that externally measured Tsk accurately reflects Tb and Tsk values are only slightly affected by Ta. Their study revealed Tsk to be within 2.0°C of rectal temperature with a maximum deviation of only 3.3°C. Willis & Brigham (2003) also showed that Tb is only slightly higher than Tsk. Transmitter mass represented 3.4–6.8% of the bat's body mass, well below and in a few cases slightly above the 5% suggested for radio-tracking studies (see Aldridge & Brigham 1988, ASM guidelines 1998). We glued the transmitters onto the skin between the shoulder blades of the bats using Skinbond surgical adhesive (Smith & Nephew United, Inc., Largo, Florida, USA). Roosting bats were tracked using Yaesu receivers (VR-500, modified by Wagener, Cologne, Germany) with 2-element Yagi antennae (HB9CV) and Tsk measured manually. To record Tsk, we timed the pulse rate of the transmitter through pulse counting in three successive 1-minute-sessions using a stop watch. The measurements were taken 2-3 times per hour on all days a transmitter remained on a bat (3-20 days). Pulse rates were translated into Tsk using transmitter-specific calibration curves provided by the manufacturer. Hourly measurements of Tsk were averaged leading to a total sample of 2062 hourly skin temperature recordings from 144 bat-days (one bat measured on one day). In addition, we recorded hourly Ta during 24 hours at all bat days using Thermochron iButtons®) (Dallas Semiconductor Corp., Dallas, Texas, USA, resolution ± 0.5°C) placed 2 m above the ground on a shaded tree trunk in the centre of the study forest.
Data analysis
For comparisons of thermoregulatory patterns between sexes and periods we defined the time of our measurements of Tsk as bat day. A bat day depended upon emergence and return time to the day roost and excluded all hours during which bats were foraging out of the roost. Bat days lasted 14 hours (6:00 am– 8:00 pm) in May and in August and 15 hours (6:00 am–9:00 pm) in June and in July, respectively. For all temperatures (mean, min and max Tsk, differences between Tsk and Ta) and environmental conditions (Ta, humidity) we calculated individual mean daily values and standard deviations (± SD) during the time an individual roosted. Subsequently, we averaged the mean values for each reproductive stage (early pregnancy, late pregnancy, lactation and post-lactation versus pre-spermatogenesis, early spermatogenesis, main spermatogenesis and advanced spermatogenesis). We tested the data for normality using Kolmogorov-Smirnov tests. The influence of gender, reproductive period and environmental conditions on individual Tsk was tested using General Linear Model analysis (GLM). All data were non-normal, thus we applied Kruskal-Wallis nonparametric one-way ANOVAs to test for significant differences between reproductive periods of females and males and a Mann-Whitney U-test to test for significant differences between the thermoregulation of two groups (either between different reproductive periods of one sex or between both sexes during one tracking period). We calculated Spearman rank correlations to assess the relationship between Tsk and Ta. Probability level determining significance was set to p < 0.05. All statistical analyses were performed using Statistica 6.0 (SigmaStat).
Results
Influence of reproduction period on thermoregulation
General linear model analyses (GLMs) revealed that min Tsk, mean Tsk and max Tsk as well as Tsk - T each varied significantly within the whole data of 2062 hourly Tsk recordings from 144 census days. Reproductive period and especially the interaction between reproductive period and sex and the interaction between reproductive period and Ta explained most of the variation in our data set (Table 2). In contrast, environmental conditions like air humidity influenced thermoregulation either very weakly or not at all. Kruskal-Wallis ANOVAs further confirmed reproductive condition as major factor influencing thermoregulation of both sexes (Table 3). Species-specific comparisons showed that females of both species reached very similar daily min, mean and max Tsk in every investigated reproductive period. In females of both bat species, daily mean, min and max Tsk were highest and the values remained nearly equally stable during late pregnancy and lactation (Fig. 1 and 2). During early pregnancy and post-lactation period daily mean, min and max Tsk dropped to significantly lower levels (Mann-Whitney-U-tests, p < 0.05). For instance, for Bechstein's bats and Daubenton's bats, in post-lactation, daily min Tsk was 23.4 ± 0.9°C and 22.5 ± 0.9°C respectively. In contrast, late pregnant and lactating females of both bat species reached daily min Tsk between 31.8 ± 2.7°C and 33.0 ± 2.5°C. Compared to females, mean, min and max Tsk of male Bechstein's bats and Daubenton's bats in the same reproductive period were not that synchronous. However, in case of male Daubenton's bats, daily min Tsk increased continuously from early spermatogenesis (17.7 ± 0.8°C) over main spermatogenesis (21.1 ± 6.6°C) to advanced spermatogenesis (30.3 ± 4.9°C). Min Tsk of male Bechstein's bats similarly increased continuously from spring (13.0 ± 2.1°C) over summer (18.9 ± 2.0°C) to autumn (20.3 ± 1.5°C).
Table 2.
Results of GLM analysis to test the dependence of daily minimum, mean and maximum Tsk and daily Tsk - Ta differences of tree-dwelling Myotis bat species (Myotis bechsteinii and M. daubentonii) regarding species, gender, reproductive period and environmental conditions (daily mean Ta, humidity, wind speed). Tracking days are listed in Table 1. n.s. = not significant. Note: F = Fit, * = p < 0.05; ** = p < 0.01; *** = p < 0.001.
Table 3.
Kruskal-Wallis-ANOVA-based comparisons of daily minimum (min), mean and maximum (max) skin temperature (Tsk) and Tsk-Ta differences between all investigated reproductive periods of male and of female Bechstein's bats (Myotis bechsteinii) and Daubenton's bats (M. daubentonii). For each tested thermoregulation describing variable, H-values with significance levels of the inter-periodical differences are given, df = degrees of freedom. * = p < 0.05; ** = p < 0.01; *** = p < 0.001.
Influence of gender on thermoregulation
Sex-specific comparisons showed that mean, min and max Tsk differed significantly between female and male bats of both Myotis species especially during tracking periods I, II and III (Fig. 1 and 2, Mann-Whitney-U-tests comparing sexes: p < 0.05). In tracking period I, mean Tsk of female Bechstein's bat in early pregnancy exceeded those of the male in pre - spermatogenesis on average about 9°C. In tracking period II, female Bechstein's bats in late pregnancy on average had an 8°C higher daily mean Tsk of the males in early spermatogenesis. Similarly, daily mean Tsk of female Daubenton's bats in late pregnancy and in lactation exceeded those of male Daubenton's bats in early and in main spermatogenesis on average about 12.5°C and 7.7°C, respectively. For both species, sex-based differences in daily mean, min and max Tsk reversed or at least declined to a non-significant level during tracking period IV. For example, daily mean Tsk of post-lactating females of both Myotis species (Bechstein's bats: 26.5°C ± 1.9, Daubenton's bats: 26.7°C ± 0.5) exceeded those of males in advanced spermatogenesis only about 3°C (male Bechstein's bats: 23.5°C ± 0.7) or fell 6°C below daily mean Tsk of males (male Daubenton's bats: 32.8°C ± 2.5).
Influence of ambient temperature on thermoregulation
GLM analyses revealed Ta as the environmental condition with the most influence on thermoregulation of the investigated bats. Differences between daily mean Tsk and daily mean Ta (Tsk - Ta) varied significantly between reproductive periods of females as well as between reproductive periods of males of both Myotis species (Kruskal-Wallis ANOVAs, Table 3). Tsk - Ta of females in late pregnancy and in lactation (Bechstein's bats 16.6°C ± 0.7 and 14.9 ± 1.5°C, Daubenton's bats 18.5 ± 1.4°C and 15.2 ± 3.0°C) were similar and significantly higher (Mann-Whitney-U-tests, p < 0.05) than those of early pregnant (Bechstein's bats 9.2°C ± 5.2) and post-lactating bats (Bechstein's bats 11.0 ± 1.8°C, Daubenton's bats 11.0 ± 1.1°C).
Species-specific comparisons confirmed that within the same reproductive periods, Tsk - Ta of females and males were similar between the two investigated Myoris-species with the exception of advanced spermatogenesis during which Tsk - Ta of male Daubenton's bats were significantly higher (about 4–5°C) than Tsk - Ta of male Bechstein's bats (Mann-Whitney-U-tests, Table 4). In males, Tsk - Ta of Bechstein's bats increased significantly from pre-spermatogenesis (2.0 ± 4.4°C) over early spermatogenesis (8.2 ± 0.8°C) to advanced spermatogenesis (9.5 ± 1.1°C) and Tsk - Ta of Daubenton'sbats increased from early spermatogenesis (5.5 ± 0.9°C) over main spermatogenesis (7.2 ± 3.6°C) to advanced spermatogenesis (14.1 ± 1.0°C) (Mann-Whitney-U-tests, p < 0.05).
Table 4.
Mann-Whitney U-test-based species-specific comparisons of daily minimum (min), mean and maximum (max) skin temperature (Tsk) and differences of mean daily Tsk and Ta (Tsk - Ta) of male and female Bechstein's bats (Myotis bechsteinii) and Daubenton's bats (M. daubentonii) during different tracking periods. For each variable U-values with significance levels of species-specific differences are given. * = p < 0.05; ** = p < 0.01; *** = p < 0.001.
Fig. 1.
Hourly Tsk of simultaneously radio-tracked male and female Myotis bechsteinii in different tracking periods of their life cycle. Dark bars indicate scotophase.
Fig. 2.
Hourly Tsk of simultaneously radio-tracked male and female Myotis daubentonii in different tracking periods of their life cycle. Dark bars indicate scotophase.
Tsk of males and females of both bat species usually exceeded Ta to the highest extent during evening and morning hours when Ta was lowest and bats were active after returning from foraging to their roosts or prior of leaving roosts for foraging (Fig. 3). During midday hours, Tsk - Ta differences were lowest because of highest Ta and lowest Tsk at this time of the day (Fig. 3). In case of Bechstein's bats, Spearman rank correlation analysis between hourly Tsk and Ta showed that hourly Tsk - Ta correlations of females in early pregnancy (rs, = 0.330, p = 0.005, n = 70), in late pregnancy (rs= 0.199, p = 0.003, n = 225), in lactation (rs= -0.043, p = 0.428, n = 330) and in post-lactation (rs= 0.330, p < 0.001, n = 294) were clearly weaker than hourly T sk - Ta correlations of the simultaneously tracked males although these decreased continuously from pre-spermatogenesis (rs = 0.787, p < 0.001, n = 280) over early spermatogenesis (rs = 0.708, p < 0.001, n = 75) to advanced spermatogenesis (rs = 0.555, p < 0.001, n = 139). Hourly Tsk - Ta correlations of male Daubenton's bats increased slightly from early spermatogenesis (rs, = -0.109, p = 0.253, n = 162) over main spermatogenesis (rs= 0.268, p = 0.021, n = 75) to advanced spermatogenesis (rs, = 0.363, p < 0.001, n = 98) and were similarly weak as the hourly Tsk - Ta correlations of female Daubenton's bats in late pregnancy (r = 0.107, p = 0.117, n = 197), in lactation (rs= -0.256, p = 0.048, n = 60) and in post-lactation (rs = -0.262, p = 0.009, n = 98).
Discussion
The study revealed that both tree-roosting Myotis species showed a very similar thermoregulatory behaviour that varied between sexes and reproductive periods. As postulated, females maintained constantly high Tb during late pregnancy and lactation period whereas early pregnant and post-lactating females reduced Tb to conserve energy significantly more often. Males allowed Tb to decrease less frequently, to a smaller extent and for shorter duration per day during advanced spermatogenesis compared to earlier in the year, when there was less sexual activity. It is obvious that both sexes adapted thermoregulation primarily to reproductive activity as it was shown for Nearctic bats roosting in buildings (Hamilton & Barclay 1994, e.g. Grinevitch et al. 1995, Cryan & Wolf 2003). Environmental conditions barely explained the observed Tsk fluctuations of the bats in our study especially because Tsk significantly varied between females and males which were simultaneously radio-tracked under similar environmental conditions.
Fig. 3.
Differences of mean daily Tsk and Ta (Tsk -Ta) of male and female Bechstein's bats (Myotis bechsteinii) and Daubenton's bats (M. daubentonii) during different reproduction periods compared to ambient temperature (Ta).
In contrast to our findings, studies on the thermoregulation of other temperate bat species suggest that the individuals' thermoregulatory behaviour depends to a much higher extent on environmental conditions (e.g. Audet & Fenton 1988, Hamilton & Barclay 1994, Grinevitch et al. 1995, Lausen & Barclay 2003, Turbill et al. 2003a, b, Turbill & Geiser 2006, Willis et al. 2006). This may possibly be related to the primarily aerial hawking foraging strategy of those bats (Barclay 1991, Chruszscz & Barclay 2003). Food intake by aerial hawking during night hours with low Ta is likely to be low as flight activity of winged prey is limited under these conditions (Racey & Swift 1985, Hickey & Fenton 1996). Therefore, aerial hawkers forage mainly during warm night hours during dusk and dawn (e.g. Rydell 1993, Swift 1997). Trawling and gleaning bats which we investigated here can forage successfully all night long, even when Ta is low (Chruszscz & Barclay 2003). Foraging behaviour of Bechstein's bats and Daubenton's bats was not significantly correlated with environmental conditions throughout the night (Dietz & Kalko 2007, Dietz & Pir 2009) suggesting a mostly Ta independent foraging activity as it was found for gleaning Myotis evotis (Chruszscz & Barclay 2003). M. daubentonii mainly feeds on swarming chironomids, capturing them close to water surfaces (Kalko & Schnitzler 1989). Insect density over water surfaces is less affected by short-term low-Ta because of the heat saving capacity of water (Dietz & Kalko 2007). M. bechsteinii is able to hawk and glean non-volant prey from vegetation and the ground and may catch sitting arthropods e.g., Lepidoptera-larvae and Coleoptera that are characterized by lower Ta-dependent activity than flying insects (Wolz 2002, Siemers & Kerth 2006, Siemers & Swift 2006).
In addition to energy intake by foraging, roosting habits may also influence thermoregulation (Solick & Barclay 2006, Boyles 2007, Willis & Brigham 2007). For example, roosting solitary and in exposed roosts incurs much higher costs for maintenance of homeothermy compared to roosting in maternity colonies in well isolated roosts (Willis & Brigham 2007). This probably explains why bats roosting in rock crevices or in foliage and why solitary roosting pregnant and lactating temperate bats (Lausen & Barclay 2003, Solick & Barclay 2006, Willis et al. 2006) reduce body temperature much more frequently than reproducing female Bechstein's bats and Daubenton's bats that communally roost in tree cavities (Dietz & Kalko 2006). For instance, rock-roosting Eptesicus fuscus usually become torpid on at least 40% of bat days, independently from reproductive condition (Lausen & Barclay 2003).
During our study, female M. bechsteinii and M. dcnibentonii formed maternity colonies in cavities made by woodpeckers in oaks and beeches. Males roosted solitary in natural crevices of hornbeams and beeches. These sex-differences in roosting habits are also directly related to differences in thermoregulation (Hamilton & Barclay 1994, Grinevitch et al. 1995, Kerth & Morf2004). Roosting solitary in crevices with high temperature fluctuations might be advantageous for the solitary males as lower temperatures improve conditions for cooling down and going into torpor and higher temperature fluctuations during day roosting foster passive rewarming (Hamilton & Barclay 1994, Grinevitch et al. 1995, Turbill et al. 2003a, 2006a). Compared to the beech and hornbeam crevices with large entrances, well isolated woodpecker-made cavities with small openings have a much more stable microclimate (Sedgeley & O'Donnell 1999, Ruczynski & Bogdanowicz 2005). Hence, in addition to social thermoregulation through clustering, roosting in woodpecker-made cavities may markedly reduce the females' costs for maintaining high Tb during reproduction (Solick & Barclay 2006, Boyles 2007, Willis & Brigham 2007).
However, our findings for female bats suggest that the need for avoiding Tb reduction is lower during early pregnancy, when embryo mass is relatively small, and during post-lactation, when juveniles have been weaned, than during late pregnancy and lactation when the embryo and the neonate reach the high body mass that is typical for bats (Kurta & Kunz 1987, Cretekos et al. 2005). Females of other bat species similarly decreased Tb reduction from early pregnancy to late pregnancy (Audet & Fenton 1988, Hoying & Kunz 1998) and lactation (Audet & Fenton 1988, Hamilton & Barclay 1994, Grinevitch et al. 1995, Chruszcz & Barclay 2002, Lausen & Barclay 2003, Solick & Barclay 2006). We assume that to overcome the energetically demanding periods of late pregnancy and lactation, an adequate amount of fat reserves saved through regular Tb reduction in early pregnancy might be very advantageous (Geiser & Masters 1994). Female and juvenile temperate bats profit from regular strong Tb reduction during post-lactation because autumnal energy-saving and conservation of fat stores significantly decrease winter mortality (Audet & Fenton 1988, Barclay 1991). Moreover, the quantity of autumnal deposited fat that remains after hibernation is linked to successful ovulation in spring and hence contributes to successful fertilization and beginning of pregnancy in the following year (Kunz et al. 1998).
In case of male Myotis bats, we assume that high Tb are most important during the spermatogenic period in summer and autumn. Testicular recrudescence and sperm production are inhibited by use of torpor in small mammals (Meistrich et al. 1973, Barnes et al. 1988, Fietz et al. 2004). The same may be true for bats. Supporting this, low Tb markedly diminished the number of spermatozoa that male sharp-nosed tomb bats (Taphozous georgianas) produce per day (Jolly & Blackshaw 1988). Moreover, sperm supply may limit reproductive success of male bats (e.g. Shapiro & Giraldeau 1996), especially under conditions of sperm competition. These are likely to occur in Bechstein's bats and Daubenton's bats where one female potentially mates with several males (Kerth & Morf 2004).
Males compensated the energetic costs of maintaining homeothermy through increased foraging activity, hence increased intake of food (Dietz & Kalko 2007) and through Tb reduction in the morning hours, when temperatures in natural tree roost were lowest (own unpublished data). During periods of elevated reproductive activity, bats were observed to remarkably lose body mass if not increasing foraging or torpor use (Entwistle et al. 1998). As our data indicate, to prepare for the energy demanding spermatogenic period, male Myotis bats exhibit energy saving and fat conservation through regular strong Tb reductions during pre-spermatogenesis in spring. Consistent with this, short bouts of deep torpor not only seem to be ubiquitous amongst aspermatogenic male bats in temperate climate zones but also in regions where milder, subtropical climates guarantee favourable Ta for foraging in spring (Turbill et al. 2003a, 2006a, b). Tb was regularly reduced by male Bechstein's bats in August, too, indicating that the use of torpor might also outweigh the benefit of whole-day normothermy during advanced spermatogenesis when most sperms and gonadal tissue have already been produced (Racey & Tam 1974). We found males with filled epididymes in tree roost at the end of July. Becoming torpid during this time would reduce copulation frequencies, especially in case of Daubenton's bats because females of this species that have gained enough weight for hibernation already leave their breeding areas at the beginning of September to settle in the hibernation site (Harrje 1994). In addition, the minimized extent of Tb reduction by males during advanced spermatogenesis may reflect a reduced ability to lower Tb when testosterone levels are high (Entwistle et al. 1998).
However, even small temperature reductions as in shallow torpor can already result in substantial energy savings for bats (Webb et al. 1993) and during late summer might especially contribute to saving fat in preparation for hibernation (Grinevitch et al. 1995). Overall we conclude that female and male Bechstein's bat and Daubenton's bats as well as other bats in the temperate climate zone actively adapt their thermoregulatory behaviour to optimize theirreproductive success. Furthermore, the feeding strategy of “gleaning” and “trawling” contributes to the observed plasticity of thermoregulation because gleaning and trawling bats are likely to be much less influenced by environmental conditions than aerial hawking bats.
Acknowledgements
We are very grateful to Prof. Dr. Elisabeth K.V. Kalko for suggestions and discussions about the ecological function of thermoregulation in temperate bats. We are also grateful to Dr. Justin Boyles and a second reviewer for very helpful comments to the manuscript.
