Search tips
Search criteria 


Logo of icbiolLink to Publisher's site
Integr Comp Biol. 2010 September; 50(3): 358–370.
Published online 2010 May 4. doi:  10.1093/icb/icq033
PMCID: PMC2931312

Thermal Windows on Brazilian Free-tailed Bats Facilitate Thermoregulation during Prolonged Flight


The Brazilian free-tailed bat (Tadarida brasiliensis) experiences challenging thermal conditions while roosting in hot caves, flying during warm daylight conditions, and foraging at cool high altitudes. Using thermal infrared cameras, we identified hot spots along the flanks of free-ranging Brazilian free-tailed bats, ventral to the extended wings. These hot spots are absent in syntopic cave myotis (Myotis velifer), a species that forages over relatively short distances, and does not engage in long-distance migration. We hypothesized that the hot spots, or “radiators,” on Brazilian free-tailed bats may be adaptations for migration, particularly in this long-distance, high-flying species. We examined the vasculature of radiators on Brazilian free-tailed bats with transillumination to characterize the unique arrangements of arteries and veins that are positioned perpendicular to the body in the proximal region of the wing. We hypothesized that these radiators aid in maintaining heat balance by flushing the uninsulated thermal window with warm blood, thereby dissipating heat while bats are flying under warm conditions, but shunting blood away and conserving heat when they are flying in cooler air at high altitudes. We also examined fluid-preserved specimens representing 122 species from 15 of 18 chiropteran families and radiators appeared present only in species in the family Molossidae, including both sedentary and migratory species and subspecies. Thus, the radiator appears to be a unique trait that may facilitate energy balance and water balance during sustained dispersal, foraging, and long-distance migration.


Flight is critical for successful dispersal, foraging, and migration in many bat species. More than 75% of a bat’s energy expended for flight may be converted to heat, which must be dissipated to avoid hyperthermia (Speakman and Thomas 2003). However, at night, flying bats are exposed to high convective and radiative forces that may result in hypothermia under certain environmental conditions (Speakman 1995; J.D. Reichard, unpublished data). Thus, flying and roosting at temporally and spatially variable ambient temperatures (Ta) requires physiological and morphological adaptations for regulating euthermic body temperature (Tb) (Schmidt-Nielsen 1990). Thermal windows and regional heterothermy of appendages are two such adaptations employed by diverse vertebrate taxa that inhabit thermally variable environments (Schmidt-Nielsen 1990; Vogel 2005).

Thermal infrared (TIR) imaging has revealed that surface temperatures (Ts) of bat wings during flight are more similar to Ta than to Tb (Lancaster et al. 1997; J.D. Reichard, unpublished data). Thus, although radiative heat loss to the night sky may be high, the wings of bats fail to maintain a thermal gradient with air that would be expected a priori to facilitate significant loss of heat. This type of regional heterothermy (i.e. relatively cool appendages) is expected to act as a heat-conservation strategy similar to those employed by some birds (Scholander 1955), carnivores (Klir and Heath 1992), aquatic mammals (Scholander and Schevill 1955; Cutright and McKean 2005), and numerous other taxa that inhabit cold or fluid environments. Similar adaptations may also be expected in bats. Although nocturnal lifestyles often require conservation of heat, bats that occur in warmer environments may also encounter high enough Ta to cause hyperthermia during flight, especially if they begin flying before sunset when incident solar radiation may be great (Speakman 1995).

Typically, thermal windows are highly vascularized body regions with reduced insulation that facilitate balance of heat (Schmidt-Nielsen 1990; Klir and Heath 1992). Animals may dissipate heat through vasodilation that directs blood to thermal windows such as the wings of bats, but can retain body heat by reducing blood flow to these areas via vasoconstriction (Cowles 1947; Reeder and Cowles 1951; Kluger and Heath 1970). “Wing fanning” is often used by bats to actively dissipate heat while they are roosting and exposed to solar radiation (Nelson 1965; Ochoa-Acuña and Kunz 1998). Bats are also expected to shunt blood away from their uninsulated wings when they are exposed to cooler environments (Stones and Wiebers 1965), thus reducing heat loss. Thermal windows can help animals depress their Tb during flight, which, in turn, facilitates maintenance of oxygen supply in hypoxic environments such as those encountered at high altitudes (Scott et al. 2008).

The Brazilian free-tailed bat (Tadarida brasiliensis, family Molossidae) migrates up to 1800 km seasonally (Cockrum 1969; Glass 1982) and may commute up to 56 km from a daytime roost during a single foraging bout (Best and Geluso 2003). Evidence from acoustic monitoring and NEXRAD Doppler radar analysis suggest that Brazilian free-tailed bats forage at altitudes up to 3000 m above the ground (Horn and Kunz 2008; McCracken et al. 2008). This species is also known to emerge from caves in south-central Texas well before sunset (Lee and McCracken 2001; Betke et al. 2008; Reichard et al. 2009), when Ta often exceeds 35°C and may approach the thermoneutral zone for this species (~33–38°C; Herreid and Schmidt-Nielsen 1966). This relatively high Ta combined with greater incident solar radiation during daylight hours can reduce a bat’s ability to dissipate heat produced during flight, except at the expense of increased evaporative water loss. Thus, Brazilian free-tailed bats experience relatively extreme temperatures during different periods of their nightly foraging bouts. By contrast, a syntopic bat species, the cave myotis (Myotis velifer, family Vespertilionidae), has a very different ecology and behavior. This species is non-migratory, rarely flies before sunset and remains relatively low to the ground while foraging (Kunz 1974). The overlapping habitats of these two insectivorous species provide an opportunity for direct comparisons of behavioral and physiological responses to temperature. It should be noted that cave temperatures in roosts shared by maternity colonies of Brazilian free-tailed bats and cave myotis may exceed 35°C during summer maternity period and the former species regularly participates in “milling” behavior throughout the day, presumably to avoid high temperatures within clusters (Herreid 1963).

In a study to characterize surface temperatures of different body regions of free-ranging Brazilian free-tailed bats, Reichard (unpublished data) discovered that this species has distinct warm regions, or hot spots, running along the flanks of the their bodies from the antebrachial to the pelvic region, ventral to the extended wings (Fig. 1A). This region of the body is warmer than surrounding furred surfaces on all sides, creating a thermal window, which we have named the “radiator,” that becomes exposed when wings are fully abducted (Fig. 1B).

Fig. 1
(A) Thermal image of a Brazilian free-tailed bat emerging from Eckert James River Cave. The white line on the bat represents the profile of surface temperature that corresponds with the plotted surface temperature. White arrows on the TIR image correspond ...

The objective of the current study was to test the hypothesis that hot spots on Brazilian free-tailed bats observed in TIR images are associated with morphological traits that are characteristic of thermal windows, namely reduced insulation and increased vasculature. We postulated that this structure may be an adaptation for the long-distance, high-altitude flight typical of Brazilian free-tailed bats and thus, may be present in other known long-distance, migrating species [e.g. Straw colored fruit bats (Eidolon helvum, family Pteropodidae; Thomas 1983) and migratory tree bats (Lasiurus cinereus, L. borealis, and Lasionycteris noctivagans, family Vespertilionidae; Cryan et al. 2004)]. We did not expect radiators to be present on non-migratory, low-flying species (e.g. Myotis velifer, family Vespertilionidae; Kunz 1974).


Thermal profiles

We conducted our study between 2006 and 2009 at two caves in south-central Texas; Frio Cave, Uvalde County and Eckert James River Cave, Mason County. Both caves serve as warm-season roosts for large maternity colonies of Brazilian free-tailed bats and cave myotis (Kunz and Robson 1995; McCracken 2003; Betke et al. 2008). We recorded TIR images of Brazilian free-tailed bats and cave myotis with a Merlin Mid TIR camera configured with a 25-mm (22°) lens (Indigo Systems Merlin, Santa Barbara, CA, USA). This camera records in the 1.0–5.4 µm wavelength and is calibrated by the manufacturer for temperatures and distances used in this study. Prior to the onset of nightly emergence of bats, we established the focal plane between 3 and 5 m from the lens outside the cave. Digital TIR images were captured at 60 frames per second using Rtools-Rdac thermal data acquisition software (FLIR Systems Co., Portland, OR, USA), operating on a customized, field-ready personal computer (LC5000 Lunchbox Computers, Duluth, GA, USA). We began recording TIR images after sunset to minimize influences from incident solar radiation and continued recording until emergence activity slowed significantly, usually ~1 h later. We recorded Ta in a shaded location near the mouth of the cave at 5-min intervals throughout the study period with an iButton Hygrochron DS 1923 (Maxim Integrated Products, Dallas Semiconductor, Sunnyvale, CA, USA). TIR data were written directly to an internal hard drive and stored for analysis in the laboratory.

We analyzed TIR digital video using Rtools-Rview thermal image analysis software to identify frames in which bats were in focus and oriented with their ventral surface facing the camera for views of bats with both wings extended. Both species are similar in body size (~12 ± 2 g), but emerge and return through separate openings to the caves or at different times. Although identification of species in TIR images may be difficult, shapes of the wings and locations of emergences meant we were confident about identification. Ts was determined with emissivity = 0.98 for pelage and skin (Monteith and Unsworth 1990). We measured a Ts profile along a transverse segment between the center-points of right and left plagiopatagia (Fig. 1A). For each bat that we measured, Ts was plotted against pixel distance along the profile segment and plots were aligned by shifting the pixel number of the peak Ts at the medial plane of the body. Using a fixed focal plane and limited depth of field meant that pixel dimensions were similar among bats that were used in the analysis.

We analyzed Ts profiles of 50 Brazilian free-tailed bats and 25 cave myotis from the same mean Ta for both species. According to Newton’s law of cooling, heat transfer is determined by the temperature difference between two bodies of matter. Thus, we plotted (TsTa) to illustrate the relative capacity for dry heat loss along the profile: (TsTa) > 0 refers to convective and radiative heat loss from the body, while (TsTa) < 0 depicts heat gain by these fluxes. Because these data were not normally distributed (Shapiro-Wilk test), comparisons between species were made with a non-parametric Wilcoxon test.

Characterizing the radiator

We captured Brazilian free-tailed bats and cave myotis with a 0.5-m diameter, padded hoop net to examine the morphology along the wings and body where the hot spots appeared in TIR images. Bats were photographed in-hand using digital photography (Fig. 1B) and by transilluminating wings with a light box or flashlight (Fig. 1C and D). Patterns of blood vessels were compared to the detailed gross anatomy of vasculature as described for the family Pteropodidae (Kallen 1977).

Taxonomic comparison

To test the hypothesis that radiators are present in other taxa, we examined 122 species representing 15 of the 18 families of bats and available in the fluid-preserved collection at the Museum of Comparative Zoology, Harvard University (Concord, MA, USA; see Supplementary Data online). Each specimen was removed from the fluid, partially dried with a paper towel, and manipulated to view the region beneath the wing, and along the body where radiators were expected. We recorded if the specimen had a hairless gap between hair on the wing and the body and/or vascular patterns similar to those observed on the proximal region of plagiopatium in T. brasiliensis. Because arteries and veins were not always conspicuous in preserved specimens, we assigned them to three categories with similar vasculature. Similar vasculature was identified as “present” when vessels were visible and resembled radiator vessels, and “absent” when vessels were visible and did not resemble radiator vessels. Similar vasculature was “unknown” when vessels were not distinct enough to definitively determine the pattern. Characters were mapped on a supertree phylogeny for the Chiroptera (Jones et al. 2002) limited to the specimens we examined using Mesquite phylogenetic analysis software (version 2.71; Maddison and Maddison 2009). We created mirrored trees and hypothesized ancestral traits by parsimony. Our sampling of museum specimens was weighted toward the family Molossidae. Thus, the proportion of species in each family and genus in our analysis is not representative of the overall relative diversity of these taxa. We assumed vasculature and hair gaps to be species-level traits, thus we examined one preserved individual from each species. Presence of radiator-like morphology at the family level was estimated as the parsimonious ancestral trait predicted from our analysis.


Profiles of surface temperature revealed distinct peaks corresponding to the location of radiators in Brazilian free-tailed bats (Fig. 2A), but not in cave myotis (Fig. 2B). The profiles of Ts for Brazilian free-tailed bats usually had three peaks: one at the median of the body and one on each flank at the margin of the body and the wing membrane (i.e. on the radiators; see Fig. 1A). Several radiators had Ts greater than Ts at the median of the body, but most were less than the maximum Ts along the profile and greater than the adjacent fur. Profiles of Ts for cave myotis usually had a single thermal peak corresponding with the median of the body.

Fig. 2
Overlayed surface temperature profiles for (A) 50 Brazilian free-tailed bats and (B) 25 cave myotis. Each line represents the surface temperature profile along a segment between opposite wing membranes. Variation in peaks and minima relate to different ...

On average, Ts declined more gradually between the body and the wing membrane on flying Brazilian free-tailed bats (N = 50) than on cave myotis (N = 25; Fig. 3). Mean maximum (TsTa) ± standard deviation was 3.91 ± 3.32°C and 2.37 ± 1.96°C for Brazilian free-tailed bats and cave myotis, respectively (Wilcoxon test, P = 0.17; Fig. 3). Minimum Ts of wings, on average, was cooler on Brazilian free-tailed bats than on cave myotis; mean minimum (TsTa) was −4.32 ± 1.57°C and −2.90 ± 2.36°C for Brazilian free-tailed bats and cave myotis, respectively (Wilcoxon test, P = 0.01; Fig. 3).

Fig. 3
Average (TsTa) along profiles on Brazilian free-tailed bats (solid, N = 50) and cave myotis (dashed, N = 25). Values where (TsTa) > 0 represent regions of sensible heat loss from the bat while (TsTa) < 0 will ...

Classification of the blood vessels of the radiator

In Brazilian free-tailed bats, seven prominent veins and arteries, “radiator vessels,” lie perpendicular to the long axis of the body and are connected to larger vessels lying exactly on the ventral margin of the wing and body. To reveal this region, the proximal wing membrane must be pulled taut both laterally and dorsally. The naked patch is then revealed on the proximal wing membrane, with the proximal edge defined by the prominent vessel where the wing membrane joins the furred skin around the thorax. In published records of the vasculature of bats, this descending arteriole is associated with a smaller network of arterioles that do not appear to extend beyond the ribcage (see Kallen 1977), however, in Brazilian free-tailed bats, branching vessels are located all the way to the inferior part of the abdomen. Farther out on the wing, paired arteries and veins of the ulnar and superficial circumflex iliac branches continue distally to form an arc (Slaaf et al. 1987; Fig. 1C). This arc appears to be more robust in Brazilian free-tailed bats than in other species (Fig. 1D). Branches from the ulnar and superficial circumflex iliac systems continue proximally to the furred region of the plagiopatagium and on to the radiator. There is also branching of the radiator arteries as they approach the furred region of the wing.

Phylogenetic comparison

Gaps in the hair on the ventral surfaces of wings surface were visible on 22 of 26 species of the family Molossidae, but this morphology was not observed on species in other families (Fig. 4; see Supplementary Data online). In the molossids, hair gaps were not visible on two of four Chaerephon spp., one of four Molossus spp. and one of three Mops spp. (Fig. 4). Most species in other bat families did not have hair on the ventral surface of the wing. However, migratory lasiurine bats (family Vespertilionidae) that do have hair on portions of their wings did not have hair gaps similar to those observed in the molossids (Fig. 5). Radiator-like morphology was also not apparent on E. helvum, another long-distance migratory pteropodid species.

Fig. 4
Mirrored supertree for the family Molossidae (Jones et al. 2002) showing characters observed on specimens examined at the Museum of Comparative Zoology, Harvard University. Vasculature similar to that seen in the radiator (left) is recorded as present ...
Fig. 5
Mirrored supertree for 15 bat families (Jones et al. 2002) showing characters observed on specimens examined at Harvard Museum of Comparative Zoology. Vasculature similar to that observed in the radiator (left) is recorded as present (black), absent (white), ...

Radiator vessels were difficult to detect on fluid-preserved specimens. Nonetheless, similar vessels were clearly present on about half (12 of 26) of the molossids examined (Fig. 4). All other molossids were categorized as unknown because vascular patterns were not sufficiently discernable. Two of 23 species in the family Phyllostomidae (Leptonycteris nivalis and Glosophaga longirostris), one of 10 in the Emballonuridae (Peropteryx leucoptera), and three of four in the Megadermatidae (Lavia frons, Macroderma gigas, and Megaderma spasma) also appeared to have similar vasculature (Fig. 5; see Supplementary Data online).

Parsimony analysis of ancestral states for each family estimated that only the Molossidae have hair gaps similar to those described for radiators; all other families examined in this study were predicted to have no hair gap as the ancestral state (Fig. 5). The ancestral state for radiator-like vasculature predicted by parsimony was unknown for the families Molossidae, Vespertilionidae, and Mormoopidae; all other families were predicted not to have radiator-like vasculature as their ancestral states (Fig. 5).


We identified radiators on Brazilian free-tailed bats but not on the syntopic cave myotis. These two species overlap in their roosting habits, but are derived from different phylogenetic lineages and exhibit distinctly different flight and foraging behaviors. The former uses high altitude, open airspace (Allison 1937; Herreid and Davis 1966; Vaughan 1966; McCracken et al. 2008), whereas the latter forages amid vegetation and often over water (Kunz 1974, 2004). We postulated that radiators may be an adaptation for migration, but similar morphological characters were absent in other families with known long-distance migrants, including lasiurines and migrating pteropdids (Fleming and Eby 2003). Radiators were also present in more sedentary molossid taxa (including the subspecies T. brasiliensis cynocephala). Several migratory bats from other families have hair on the wing membranes (e.g. lasiurines), but these bats do not have naked strips on the sides of their bodies or a vascular structure closely resembling the radiators observed in the Molossidae. Phylogenetic constraints may explain the lack of radiators in these other taxa. Furthermore, L. cinereus and L. borealis also have a more northern distributional range and even though they may fly in the daylight when northern nights are short, they generally inhabit cooler habitats. Thus, these species probably spend less time flying at high Ta than do tropical and subtropical species, and may face lower risks of hyperthermia.

The general distribution of vessels in the wings of bats is common among chiropteran species (Kallen 1977). However, two distinguishing characteristics of the radiator appear unique to the family Molossidae, and the Brazilian free-tailed bat, in particular. In the illustrated wing vasculature of a pteropodid described by Kallen (1977), a proximal branching arteriole from the ulnar artery (originating from the subclavian artery) runs lateral to the body and appears to have a similar branching pattern to the radiator vessels described herein. This radiator-like branching does not appear to continue beyond the ribcage and, thus, neither the descending arteriole nor the branching vessels appear to be given much significance by Kallen in his description of wing vasculature in pteropodids. However, it should be noted that Kallen’s specimen was significantly larger than a Brazilian free-tailed bat. Because this arteriole has not been identified in previous literature, we refer to it as the “thoracopatagial” arteriole, describing its location and the tissues it supplies (Fig. 1B). Radiator branches in the Brazilian free-tailed bat are relatively evenly spaced from the antebrachial to the pelvic region along the entire flank of the bat’s body. This species also exhibits a strong arc formed by arterioles and venules of the ulnar and superficial iliac circumflex branches that provide dual input to the tensor plagiopatagii muscle (Slaaf et al. 1987). Although we could not fully characterize direction and volumes of blood flow, we can infer from the relatively warm Ts of radiators in TIR images (Fig. 1A) that radiators transport warm blood directly from the torso. Thus, we postulate that the branches originating from the thoracopatagial arteriole shunt warm blood from the thorax and abdomen, through the naked thermal window, thus facilitating the dissipation of heat from flight muscles. Additional histology and examination of tissue in unanesthetized molossid bats are needed to determine the fine vasculature and direction of blood flow in the vessels of the radiator.

While the wings of bats as a whole have characteristics of thermal windows (e.g. naked and highly vascularized), the smaller area of the radiator in molossids appears suitable for finely controlled thermoregulation during flight. To our knowledge, there is no published information on the form and function of this unique trait. We suggest three possible explanations for this deficiency. First, perfusion and the size of various blood vessels are affected by Ta and Tb (Cowles 1947; Reeder and Cowles 1951; Kluger and Heath 1970) and perfusion of the radiator vessels may require specific physiological and thermal conditions. Thus, if observers attempt to characterize wing vessels while the bat is inactive, some vessels may be less apparent. Secondly, radiators appear to be limited to the family Molossidae, thus even the most detailed studies of other bat species would not reveal this morphology (e.g. Kallen 1977). Thirdly, previous studies on circulation in the wings of Brazilian free-tailed bats have used restraining devices that enclose the bat’s body with one wing extended. These restraining devices may exclude the radiators from view while facilitating visualization of other regions (see illustration in Bouskela and Wiederhielm 1979).

Hypotheses on the function of radiators

Radiators in Brazilian free-tailed bats can best be described as a patch of relatively warm, uninsulated skin located beneath the proximal portion of the wing. When the wings are extended and raised during powered flight, this region is exposed and maintains a thermal gradient with the surrounding air, which facilitates dry (sensible) heat loss. In this way, evaporative cooling under heat load may be reduced. The Ts of wings distal to the thermal window cool well below Ta during flight and may contribute to a depression of Tb in free-ranging Brazilian free-tailed bats (J.D. Reichard, unpublished data). Cave myotis had slightly greater maximum Ts on the body, suggesting that the pelage of this species may have poorer insulating properties compared to Brazilian free-tailed bats or that the former maintains a relatively higher Tb during flight. Thus, we would expect cave myotis to experience greater heat loss from the body, but a comprehensive analysis of heat loss during flight in this species invites further study. Additionally, Ts does not drop as far below Ta on wings of cave myotis as it does on wings of the Brazilian free-tailed bat, but both species have Ts consistent with reduced blood flow and cooling from evaporative water loss from the flight membranes and radiative heat loss to the night sky (J.D. Reichard, unpublished data). Differences in altitudes where these two species fly may contribute to these differences in Ts of wings.

Radiators were relatively consistent traits among members of the Molossidae examined in this study. We suggest that the combination of uninsulated skin on proximal wing and a unique vascular structure may be adaptations in this chiropteran lineage, which includes species with skeletal traits and musculature adapted for rapid and prolonged flight (Vaughan 1966). Members of the genus Cheiromeles, family Molossidae (which includes only two, largely hairless species) have evolved an unusual morphological specialization in the proximal wing region. Although we were not able to observe blood vessels in the single museum specimen of C. torquatus to which we had access, this species has evolved pouches in the antebrachial region along the margin of the wings and body (Harrison 1954). The postulated role of pouches in protecting and warming young bats in this species has not been supported (Leong et al. 2009). Instead, Cheiromeles folds its wings into these pouches, which presumably protect delicate membranes and phalanges while it crawls on the ground or on branches (Schutt and Simmons 2001). The pouches are located in the same region of the body as the radiators in T. brasiliensis, but the radiators do not appear to be associated with wing positions while crawling. Notwithstanding, the example from Cheiromeles suggests a propensity for modification or specialization in this region of the body in the family Molossidae. Preliminary results from a molecular phylogenetic analysis of the RAG-2 gene suggest that Cheiromeles is basal to other members of the Molossidae and thus supports the morphological distinction between these taxa (L.K. Ammerman, personal communication).

Interestingly, the morphology of the radiator is similar to that of a brood patch found on many avian species. However, because radiators are present in both females and males, and male Brazilian free-tailed bats do not participate in rearing offspring, we do not expect that transfer of heat between mothers and offspring is the primary function of radiators. Moreover, we would expect a brood patch to be more closely associated with nipples, but the position and size of radiators does not appear optimal for facilitating heat transfer while nursing pups. Moreover, lactating Brazilian free-tailed bats typically roost separately from pups when they are not nursing (Kunz and Robson 1995) and pups cluster together to maintain high Tb (Herreid 1963). It has been suggested that elevated Tb may limit milk energy output in lactating mice (Król and Speakman 2003) and radiators are in a reasonable position for dissipating heat in proximity to the mammary glands, however, this function is also not supported by the presence of radiators in both males and females.

Free-tailed bats are specialized for high-altitude, hawking flight and are abundant in the warmer parts of the world (Vaughan 1966). Thus, species in the family Molossidae may experience high Ta at ground level when they first emerge from their day roosts. During our study, we measured Ta that exceeded 35°C outside the cave, even as bats emerged and dispersed to forage. However, Ta can decrease 6°C for every 1000 m increase in altitude (Moore 1956) and convective forces are potentially greater in prevailing winds. Thus, during a given evenings emergence and foraging flight, Brazilian free-tailed bats often make transitions from conditions requiring heat dissipation to avoid hyperthermia to conditions where conservation of heat may be more beneficial. We suggest that radiators in Brazilian free-tailed bats are highly developed for thermoregulation on the proximal regions of the wings, providing a mechanism for rapid dissipation of heat during prolonged foraging flights and during migratory flights in relatively warm air. This mechanism may also reduce circulation to the distal wing membranes, and thus reduce heat loss in cooler air, especially when wings are exposed to the strong radiative sink to a clear night sky (Léger and Larochelle 2006). Blood may be shunted away from the radiator and toward the body during flight at high altitude and thereby minimize heat loss, but may be passed through the radiator if Tb rises above optimal levels. When Ts exceeds Ta, peripheral blood is cooled rapidly as it passes through the thin, uninsulated membrane of the wing. Thus, bats with radiators would be able to dissipate heat without having to shunt large volumes of warm blood through the distal portions of their wing membranes. The series of lateral radiator vessels could allow for intricate control of blood flow for thermoregulatory purposes, with more radiator vessels being perfused as heat load increases.

High-flying bats, such as Brazilian free-tailed bats, often have limited access to drinking water and may also be highly adapted for conservation of water (Kunz et al. 1995). Descending to drink would likely increase energy expenditure, while at the same time potentially limiting foraging time. Although we have not explicitly demonstrated countercurrent heat exchange, the radiators present in the Brazilian free-tailed bat be a manifestation of this physiological adaptation. When the Ts of wings drop below Ta during flight, blood returning from the wings could overcool the flight muscles and thus reduce flight efficiency. As warm blood flows away from the vital organs and flight muscles through the arterioles in the radiator, it cools rapidly. Additional cooling may be achieved as blood returns from the wings through venules in the radiator. However, we are cautious in postulating the role of these vessels in countercurrent heat (or water) exchange. Although the major arterioles are often paired with venules, this arrangement may only provide limited exchange. Well known countercurrent exchange systems in legs of birds (Midtgård 1981), flippers of marine mammals (Meagher et al. 2008) and other animals (Vogel 2005) exhibit recognizable morphology in which close association of veins and arteries increases the potential area of contact (Schmidt-Nielsen 1990). However, if blood returning through radiator venules is warmed through countercurrent exchange, Brazilian free-tailed bats may be able to rewarm relatively cool blood returning from the wings before it reaches the body. This function may be particularly important in cool air that this species may encounter at high altitudes. Conservation of water could also be achieved through countercurrent diffusion if “drier” venous blood is rehydrated by arteriole blood (Vogel 2005). Maintaining a high temperature gradient between the radiator and the air would mean that more heat can be dissipated by sensible heat loss without high blood flow through the distal wing membranes. Both of these mechanisms could reduce overall transcutaneous loss of water.

Brazilian free-tailed bats move readily from high Ta approaching 40°C in their roosts and at ground level to cool Ta and high radiative and convective heat fluxes at higher altitudes. Thermal windows (i.e. radiators) and intricate control of blood flow to and from the wings may permit rapid dissipation of heat and the avoidance of hyperthermia when activity-related thermogenesis could otherwise exceed heat loss. When heat loss exceeds thermogenesis, however, regional heterothermy and countercurrent exchange may help to sustain euthermia. An analogous rete mirable present in some animals that employ regional heterothermy and thermal windows for thermoregulation may also be present in the wings of molossids, but further investigation is needed to fully characterize the patterns of flow through veins and arteries in these radiators. We suggest that the radiators on Brazilian free-tailed bats represent an adaptation for long-distance, high-altitude flight to facilitate heat, water and energy balance and may have been an important factor in the evolution of migratory behavior in this species. Analogous regional heterothermy is employed for thermoregulation in several mammalian and avian taxa. In arctic environments, for example, regional heterothermy of the limbs in carnivores and of the ears in lagomorphs reduces heat loss by convection or conduction and sustains euthermia (Schmidt-Nielsen 1990; Klir and Heath 1992). Aquatic mammals also employ countercurrent exchange mechanisms and regional heterothermy in their flippers or tails that compensate for strong convective heat loss in colder water (Scholander and Schevill 1955; Cutright and McKean 2005). Similarly, many birds have evolved countercurrent heat-exchange mechanisms in their featherless legs, which have been implicated for their roles in thermoregulation during flight (Martineau and Larochelle 1988; Schmidt-Nielsen 1990). Countercurrent exchange is also employed in maintaining euthermia in warm environments. For example, some ungulates have rete mirabiles in which venous blood cooled near the nasal cavity subsequently cools arterial blood supplying the brain (Baker and Hayward 1968).

The discovery of radiators on Brazilian free-tailed bats presents a promising future for exploring the function of this structure under different physiological and environmental conditions. Continuous TIR imaging of individuals, for example in a wind tunnel (see Ward et al. 1999), may reveal changes in Ts in response to prolonged flight that may also correspond with Ts on the radiators. Moreover, coupling TIR images with respirometry or doubly labeled water experiments could provide important information for distinguishing between energy used for thermoregulation and energy used for locomotion (see Ward et al 2004; Zerba et al. 1999). Such a study would permit researchers to modify Ta and humidity to test effects of these variables on Ts of wings. Studies on captive or semi-captive bats would make it possible to observe individuals of known sex and reproductive condition, which could help identify differences in radiator function among these cohorts. Further characterizations of the circulatory patterns of the radiator are needed to identify origins and destinations of arteries and veins may provide insight into the tissues and organs that are most closely associated temperature regulation during flight. Additionally, histological analysis could be used to search for possible countercurrent exchange mechanisms in this region. Small animal computed tomography (MicroCT) may also improve visualization of the vasculature of intact bats by injecting radio-opaque contrasting agents into the blood stream of a bat (Kindlmann et al. 2005). However, sufficient perfusion of the region of interest can be challenging, especially if the radiator mostly functions when animals are in flight. Birds have been the focus of many previous studies on the physiology of flight at high altitudes (e.g. Scott and Milsom 2006; Scott et al. 2008). Wind tunnel studies have improved our understanding of the kinematics and energetics of flight for both birds (Ward et al. 1999; 2004; Léger and Larochelle 2006) and bats (Carpenter 1986; Thomas et al. 1991) and may be a suitable venue for assessing circulatory changes in the radiator during prolonged flight at different ambient temperatures. Finally, greater understanding of the form and function of radiators in bats may be achieved by comparing morphology, physiology and behavior of different species in the family Molossidae with birds that exhibit similar long-distance foraging or migratory behavior.

Brazilian free-tailed bats move readily from high Ta approaching 40°C in their roosts and at ground level to cool Ta and high radiative and convective heat fluxes at higher altitudes. Thermal windows (i.e. radiators) and intricate control of blood flow to and from the wings may permit rapid dissipation of heat and the avoidance of hyperthermia when activity-related thermogenesis could otherwise exceed heat loss. When heat loss exceeds thermogenesis, however, regional heterothermy and countercurrent exchange may help to sustain euthermia. An analogous rete mirable present in some animals that employ regional heterothermy and thermal windows for thermoregulation may also be present in the wings of molossids, but further investigation is needed to fully characterize the patterns of flow through veins and arteries in these radiators. We suggest that the radiators on Brazilian free-tailed bats represent an adaptation for long-distance, high-altitude flight to facilitate heat, water and energy balance and may have been an important factor in the evolution of migratory behavior in this species.

Supplementary Data

Supplementary data are available at ICB online.


Funding was provided by Boston University’s Center for Ecology and Conservation Biology, the National Park Service (Grant No. H7170040002 to T.H.K., PI), and the National Science Foundation (Grant No. IIS-0326483 to T.H.K., P.I., and M.B., G.F.M., J.K.W., and P.M., Co-PIs) for supporting this research. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Park Service or the National Science Foundation. This study was also made possible, in part, by software available from the National Institute of Health/National Center for Research Resources Center for Integrative Biomedical Computing (Grant No. 5P41RR012553 to C.K.). Additional funding was made available by Sigma Xi Grants-in-Aid to J.D.R.

Supplementary Material

[Supplementary Data]


We thank The Nature Conservancy for access to Eckert James River Bat Cave and I. Marback and W. Cofer for access to Frio Cave. We are grateful to B. French, N. Hristov, L. Allen, C. Casey, L. Gonzalez, and A. Bahadur for assistance and support in the field and the lab. Valuable feedback on this study was provided by K. Warkentin and by participants in the 2009 annual meeting of the North American Society for Bat Research, particularly G. Kwiecinski and W. Lancaster. We thank J. Chupasko and M. Omura for providing access to the mammal collection at the Museum of Comparaive Zoology, Harvard University, and to J. DaCosta who assisted with phylogenetic mapping. Methods used in this study were in accordance with policies of the Texas Parks and Wildlife Department, the American Society of Mammalogists Guidelines for Capture, Handling, and Care of Mammals, and Boston University’s Institutional Animal Care and Use Committee.


  • Allison VC. Evening bat flight from Carlsbad Caverns. J Mammal. 1937;18:80–2.
  • Baker MA, Hayward JN. The influence of the nasal mucosa and the carotid rete upon hypothalamic temperature in sheep. J Physiol. 1968;198:561–79. [PubMed]
  • Best TL, Geluso KN. Summer foraging range of Mexican free-tailed bats (Tadarida brasiliensis mexicana) from Carlsbad Cavern, New Mexico. Southwest Nat. 2003;48:590–6.
  • Betke M, et al. Thermal imaging reveals significantly smaller Brazilian free-tailed bat colonies than previously estimated. J Mammal. 2008;89:18–24.
  • Bouskela E, Wiederhielm CA. Microvascular myogenic reaction in the wing of the intact unanesthetized bat. Am J Physiol. 1979;237:H59–65. [PubMed]
  • Carpenter RE. Flight physiology of intermediate-sized fruit bats (Pteropodidae) J Exp Biol. 1986;120:79–103.
  • Cockrum EL. Migration in the guano bat. Vol. 51. Lawrence, KS: Miscellaneous Publications, Museum of Natural History, University of Kansas; 1969. pp. 303–36.
  • Cowles RB. Vascular changes in the wings of bats. Science. 1947;105:362–3. [PubMed]
  • Cryan PM, Bogan MA, Rye RO, Landis GP, Kester CL. Stable hydrogen isotope analysis of bat hair as evidence for seasonal molt and long distance migration. J Mammal. 2004;85:995–1001.
  • Cutright WJ, McKean T. Countercurrent blood vessel arrangement in beaver (Castor canadensis) J Morph. 2005;161:169–75.
  • Fleming TH, Eby P. Ecology of bat migration. In: Kunz TH, editor. Bat ecology. Chicago: The University of Chicago Press; 2003. pp. 156–208.
  • Glass BP. Seasonal movements of Mexican free-tailed bats, Tadarida brasiliensis mexicana, banded in the Great Plains. Southwest Nat. 1982;27:127–33.
  • Harrison JL. The wing pouches of the naked bulldog bat. Malay Nat J. 1954;9:66–7.
  • Herreid CF II. Temperature regulation and metabolism in Mexican free-tailed bats. Science. 1963;142:1573–4. [PubMed]
  • Herreid CF II, Davis RB. Flight patterns of bats. J Mammal. 1966;47:78–86.
  • Herreid CF II, Schmidt-Nielsen K. Oxygen consumption, temperature, and wter loss in bats from different environments. Am J Physiol. 1966;211:1108–12. [PubMed]
  • Horn JW, Kunz TH. Analyzing NEXRAD Doppler radar images to assess nightly dispersal patterns and population trends in Brazilian free-tailed bats (Tadarida brasiliensis) Integr Comp Biol. 2008;48:24–39. [PubMed]
  • Jones KE, Purvis A, MacLarnon A, Bininda-Emonds ORP, Simmons NB. A phylogenetic supertree of the bats (Mammalia: Chiroptera) Biol Rev. 2002;77:223–59. [PubMed]
  • Kallen FC. The cardiovascular systems of bats: structure and function. In: Wimsatt WA, editor. Biology of bats. New York: Academic Press; 1977. pp. 289–483.
  • Kindlmann GL, Weinstein DM, Jones GM, Johnson CR, Capecchi MR, Keller C. Practical vessel imaging by computed tomography in live transgenic mouse models for human tumors. Molec Imag. 2005;4:417–24. [PubMed]
  • Klir JJ, Heath JE. An infrared thermographic study of surface temperature in relation to external thermal stress in three species of foxes: the red fox (Vulpes vulpes), arctic fox (Alopex lagopus), and kit fox (Vulpes macrotis) Physiol Zool. 1992;65:1011–21.
  • Kluger MJ, Heath JE. Vasomotion in the bat wing: a thermoregulatory response to internal heating. Comp Biochem Physiol. 1970;32:219–26. [PubMed]
  • Król E, Speakman JR. Limits to sustained energy intake VII. Milk energy output in laboratory mice in thermoneuatrality. J Exp Biol. 2003;206:4267–81. [PubMed]
  • Kunz TH. Feeding ecology of a temperate insectivorous bat (Myotis velifer) Ecology. 1974;55:693–711.
  • Kunz TH. Foraging habits of North American insectivorous bats. In: Brigham RM, Kalko EKV, Jones G, Parsons S, Limpens HJGA, editors. Bat echolocation research: tools, techniques, and analysis. Austin, TX: Bat Conservation International; 2004. pp. 13–25.
  • Kunz TH, Robson SK. Postnatal-growth and development in the Mexican free-tailed bat (Tadarida brasiliensis): birth size, growth-rates, and age estimation. J Mammal. 1995;76:769–83.
  • Kunz TH, Whitaker JO, Jr, Wadanoli MD. Dietary energetics of the Mexican free-tailed bat (Tadarida brasiliensis) during pregnancy and lactation. Oecologia. 1995;101:107–15.
  • Lancaster WC, Thomson SC, Speakman JR. Wing temperature in flying bats measured by infrared thermography. J Therm Biol. 1997;22:109–16.
  • Lee YF, McCracken GF. Timing and variation in the emergence and return of Mexican free-tailed bats, Tadarida brasiliensis mexicana. Zool Stud. 2001;40:309–16.
  • Léger J, Larochelle J. On the importance of radiative heat exchange during nocturnal flight in birds. J Exp Biol. 2006;209:103–14. [PubMed]
  • Leong TM, Teo SC, Lim KKP. The naked bulldog bat, Cheiromeles torquatus, in Singapore-past and present records, with highlights on its unique morphology (Microchiroptera: Molossidae) Nature in Singapore. 2009;2:215–30.
  • Maddison WP, Maddison DR. Mesquite: a modular system for evolutionary analysis. 2009 Version 2.71 ( Last accessed 19 April 2006.
  • Martineau L, Larochelle J. The cooling power of pigeon legs. J Exp Biol. 1988;136:193–208.
  • McCracken GF, Gillam EH, Westbrook JK, Lee YF, Jensen ML, Balsley BB. Brazilian free-tailed bats (Tadarida brasiliensis: Molossidae, Chiroptera) at high altitude: links to migratory insect populations. Integr Comp Biol. 2008;48:107–18. [PubMed]
  • McCracken GF. Estimates of population sizes in summer colonies of Brazilian free-tailed bats (Tadarida brasiliensis) In: O’Shea TJ, Bogan MA, editors. Monitoring trends in bat populations of the United States and Territories: problems and prospects. Washington, DC: US Geological Survey, Biological Resources Discipline, Information and Technology Report, USGS/BRD/ITR-2003-003; 2003. pp. 21–30.
  • Meagher EM, McLellan WA, Westgate AJ, Wells RS, Blum JE, Pabst DA. Seasonal patterns of heat loss in wild bottlenose dolphins (Tursiops truncatus) J Comp Physiol B. 2008;178:529–43. [PubMed]
  • Midtgård U. The rete tibiotarsale and arterio-venous association in the hind limb of birds: a comparative morphological study on counter-current heat exchange systems. Acta Zool. 1981;62:67–87.
  • Monteith JL, Unsworth MH. London: Edward Arnold, Hodder Headline PLC; 1990. Principles of environmental physics.
  • Moore JG. The tropospheric temperature lapse rate. Meteorol Atmos Phys. 1956;9:468–70.
  • Nelson JE. Behaviour of Australian Pteropodidae (Megachiroptera) Anim Behav. 1965;13:544–7. [PubMed]
  • Ochoa-Acuña H, Kunz TH. Thermoregulatory behavior in a small island flying fox, Pteropus hypomelanus (Chiroptera: Pteropodidae) J Therm Biol. 1998;24:15–20.
  • Reeder WG, Cowles RB. Aspects of thermoregulation in bats. J Mammal. 1951;32:389–403.
  • Reichard JD, Gonzalez LE, Casey CM, Allen LC, Hristov NI, Kunz TH. Evening emergence behavior and seasonal dynamics in large colonies of Brazilian free-tailed bats. J Mammal. 2009;90:1478–86.
  • Schmidt-Nielsen K. Cambridge: Cambridge University Press; 1990. Animal physiology: adaptation and environment.
  • Scholander PF, Schevill WE. Countercurrent vascular heat exchange in the fins of whales. J Appl Physiol. 1955;8:279–82. [PubMed]
  • Scholander PF. Evolution of climatic adaptation in homeotherms. Evolution. 1955;9:15–26.
  • Schutt WA, Simmons NB. Morphological specializations of Cheiromeles (naked bulldog bats; Molossidae) and their possible role in quadrapedal locomotion. Acta Chiropterol. 2001;3:225–35.
  • Scott GR, Milsom WK. Flying high: a theoretical analysis of the factors limiting exercise performance in birds at altitude. Resp Physiol Neurobi. 2006;154:284–301. [PubMed]
  • Scott GR, Cadena V, Tattersall GJ, Milsom WK. Body temperature depression and peripheral heat loss accompany the metabolic and ventilatory responses to hypoxia in low and high altitude birds. J Exp Biol. 2008;211:1326–35. [PubMed]
  • Slaaf DW, Reneman RS, Wiederhielm CA. Pressure regulation in muscle of unanesthetized bats. Microvasc Res. 1987;33:315–26. [PubMed]
  • Speakman JR, Thomas DW. Physiological ecology and energetics of bats. In: Kunz TH, Fenton MB, editors. Bat ecology. Chicago, IL: The University of Chicago Press; 2003. pp. 430–92.
  • Speakman JR. Chiropteran nocturnality. Symp Zool Soc Lond. 1995;67:187–201.
  • Stones RC, Wiebers JE. A review of temperature regulation in bats (Chiroptera) Am Midl Nat. 1965;74:155–67.
  • Thomas DW. The annual migrations of 3 species of West-African fruit bats (Chiroptera, Pteropodidae) Can J Zool. 1983;61:2266–72.
  • Thomas SP, Follette DB, Farabaugh AT. Influence of air temperature on ventilation rates and thermoregulation of a flying bat. Am J Physiol Regul Integr Comp Physiol. 1991;260:R960–R968. [PubMed]
  • Vaughan TA. Morphology and flight characteristics of molossid bats. J Mammal. 1966;47:249–60.
  • Vogel S. Living in a physical world, V. maintaining temperature. J Biosci. 2005;30:581–90. [PubMed]
  • Ward S, Rayner JMV, Möller U, Jackson DM, Nachtigall W, Speakman JR. Heat transfer from starlings Sturnus vulgaris during flight. J Exp Biol. 1999;202:1589–602. [PubMed]
  • Ward S, Möller U, Rayner JMV, Jackson DM, Nachtigall W, Speakman JR. Metabolic power of European starlings Sturnus vulgaris during flight in a wind tunnel, estimated from heat transfer modeling, doubly labelled water and mask respirometry. J Exp Biol. 2004;207:4291–8. [PubMed]
  • Zerba E, Dana AN, Lucia MA. Temperature and energy expenditure of the eastern house finch (Carpodacus mexicanus) during cold stress. Physiol Biochem Zool. 1999;72:265–76. [PubMed]

Articles from Integrative and Comparative Biology are provided here courtesy of Oxford University Press