Heterothermy in small, migrating passerine birds during stopover: use of hypothermia at rest accelerates fuel accumulation

Each year, billions of birds fly in autumn from their breeding grounds in Eurasia to wintering grounds in Africa and return in spring, often covering many thousands of kilometers (Moreau,1972; Biebach et al.,2000). In small passerine birds, fuel for migratory flight consists mainly of fat (between 85 and 95%) and to a lesser extent of protein(5–15%) (Klaassen and Biebach,1994; Klaassen et al.,2000). Fuel used during flight is amassed before migration and restored while birds sojourn at stopover sites. During migration, most of these normally diurnal birds change their activity rhythm and fly at night(Berthold, 1996). If a bird's fuel reserves are adequate, it takes off at dusk and flies non-stop, often for several hundred kilometers at a stretch(Biebach et al., 1986; Biebach et al., 2000). However,this daily rhythm changes when fuel stores run low, and the bird lands to refuel. At stopovers, birds return to diurnality(Berthold, 1996), and may maintain this activity rhythm for several days, until restored energy reserves are sufficient to resume flight (Biebach et al., 1986). Such changes in activity rhythms may occur several times during the migratory journey and are quantitatively related to a bird's fuel reserves (Biebach,1996).

During migration, songbirds apparently expend more than twice as much energy at stopovers as in flight because the total time spent at stopovers exceeds the time spent in flight by as much as sevenfold(Hedenström and Alerstam,1997; Wikelski et al.,2003; Bowlin et al.,2005). It is unlikely that birds can directly reduce the energy spent in flapping flight, although they can save energy by choosing when and where in the air-column to fly (Carmi et al., 1992). Theoretically, however, there are ways in which a bird might directly reduce energy expenditure during stopover, either while resting or while feeding to refuel. Studies on the rates of fat accumulation in birds,both before and during migration, have addressed potentially important means of saving energy, including changes in feeding behavior (e.g. Gwinner et al., 1985), food choice and digestive physiology (for reviews, see Bairlein, 2002; McWilliams et al., 2004),endocrine regulation of fattening(Wingfield et al., 1990), and energy balance during the fattening process(Klaassen and Biebach, 1994). However, whether migrating birds are able to reduce energy costs while refueling at stopovers, by becoming hypothermic or even entering torpor at rest, has not been explored other than in hummingbirds(Carpenter and Hixon, 1988; Hiebert, 1993). Facultative reductions in body temperature (T_b) do, however, seem to be common in birds (McKechnie and Lovegrove, 2002).

Passerine birds typically maintain T_b between 39°C and 44°C, which, in terms of metabolic energy, is expensive(Prinzinger et al., 1991). However, T_b normally fluctuates through the day and usually decreases by 1–3°C during normothermic rest, with a concomitant saving of energy through decreased metabolic rate (MR)(Prinzinger et al., 1991; Dawson and Whittow, 2000). However, when faced with unfavorable environmental conditions, many birds may decrease their rest-phase T_b significantly below normothermic values. This state, depending on its depth, is called rest-phase hypothermia (T_b lowered by 3–10°C) or torpor(T_b lowered by >10°C)(Reinertsen, 1996; McKechnie and Lovegrove, 2002)[but see also Schleucher (Schleucher,2004) and Schleucher and Prinzinger(Schleucher and Prinzinger,2006)]. In non-migrating passerine species, T_bwas found to decrease in birds exposed to very low air temperatures(T_a), or after several weeks of acclimatization to winter conditions (for reviews, see Reinertsen,1996; Welton et al.,2002). In passerines, torpor has been reported only in nectarivores and insectivores and was described as an adaptation related to variable food availability (McKechnie and Lovegrove, 2003). Here, we adhere to the nomenclature in the`Glossary of Terms for Thermal Physiology'(IUPS Thermal Commission,2001), wherein heterothermy is defined as `a pattern of T_b regulation that exceeds in range that characteristic for homeothermy', while hypothermia is `the condition of a temperature regulator when core temperature is below its range specified for the normal active state of the species', being either regulated or forced(pathological).

Bearing in mind that several species of small, non-migrating passerine birds are capable of significantly lowering T_b at rest,and that hummingbirds use torpor to facilitate fuel accumulation during migration (Carpenter and Hixon,1988; Hiebert,1993), we investigated whether passerine migrants use hypothermia while preparing for their next flight-leg at stopovers. We did this by measuring daily changes in T_b in blackcaps (Sylvia atricapilla), a very common Palearctic passerine migrant, under semi-natural conditions using radiotelemetry at a stopover site at Midreshet Ben-Gurion, Israel, while they fed in a large outdoor aviary. We hypothesized that, while at rest during a stopover, blackcaps lower their T_b below their normotheric rest-phase T_b and become hypothermic, thereby facilitating fuel deposition by reducing nocturnal energy expenditure. We tested the prediction that birds that use hypothermia at rest would increase body mass(m_b) faster during stopover than those that do not.

Blackcaps (Sylvia atricapilla L. 1758) breed throughout Europe,wintering mainly in northern Africa, south to the Sahel(Cramp, 1992), and in spring are one of the most numerous passerine species that stopover in Israel. Body mass of blackcaps during the breeding season is 16–20 g(Cramp, 1992), but during migration m_b may range from <13 g observed in very lean individuals to >27 g in the heaviest birds captured after a 10-day stay at a springtime stopover in Eilat, Southern Israel (personal observations). In April 2007, during the spring migration season, we mist-netted blackcaps in a plantation of Pistacia atlantica on the Sede Boqer Campus of Ben-Gurion University at Midreshet Ben-Gurion (30°52′N,34°46′E) in the Negev Desert highlands. Upon capture, the birds weighed between 14.9 and 17.4 g and their wrist-to-wing tip length ranged between 71 and 80 mm; we found no correlation between these two measures(Pearson product moment correlation: P=0.9). After capture, birds were ringed with standard aluminum rings and transferred to outdoor flight cages (five to a cage 6×2×2.5 m, and three to a cage 3×2×2.5 m). Birds were kept in the flight cages for at least 2 days to allow them to habituate to the semi-natural conditions, which we verified by observing that they ate freely and maintained body mass. The flight cages had a groundcover of grasses and annuals to attract insects. Since blackcaps eat fruit and insects during migration (personal observations)(Karasov and Pinshow, 2000),we supplied them daily with fruit ad libitum and 5–10 g of mealworms, offered in small feeders. The flight cages were covered with 70%shade netting and contained branches to perch on; the ground cover provided shelter and additional shade for birds.

Body temperature was measured with implantable radio transmitters (model BD-2N, Holohil Systems, Carp, Ontario, Canada). Transmitters were previously calibrated between 10°C and 50°C in a controlled-temperature water bath, against a mercury-in-glass thermometer with an accuracy of±0.1°C traceable to the US NIST. Thereafter, we entered the calibration data (temperature and corresponding pulse interval) into the memory of the receiver that calculated T_b with built-in software. Before implantation, transmitters were coated with sterilized, pure paraffin wax, and their final mass ranged between 0.8 g and 0.9 g. We implanted transmitters intraperitoneally in all eight birds (five males and three females) under inhalation anesthesia (Isoflurane®; Minrad, Inc.,Bethlehem, PA, USA), and allowed the birds to fully recover from anesthesia before returning them to the aviaries and beginning temperature recording. Radio signals were recorded with a logging receiver (Lotek model SRX-400A W21AST with Event_Log© software, Newmarket, Ontario, Canada).

We observed daily changes in T_b while birds were rebuilding their body stores as they fed in the aviary, simulating a migratory stopover. We recorded T_b in blackcaps over periods of 4–12 days per bird, resulting in 59 bird-days and 34 bird-nights of observations. During observations in the aviaries, birds were exposed to the natural photoperiod, namely sunrise at 06:00 h (±15 min), sunset at 19:00 h (±10 min), and T_a. T_awas recorded continuously with two calibrated iButton® data loggers (model DS1921, Maxim Integrated Products, Sunnyvale, CA, USA), each suspended from the ceiling of a cage in a well-ventilated, white, open-ended cardboard tube(20 cm long, and 8.5 cm diameter) to minimize effects of incident radiation.

Every day, at approximately 09:00 h, each bird was caught with a hand net and weighed to ±0.1 g with an electronic balance (Scout; Ohaus Corporation, Florham Park, NJ, USA), and the previous day's T_b data were downloaded.

Metabolic rates were measured by indirect calorimetry with a multiple-channel, open-flow respirometry system (Qubit Systems, Kingston, ON,Canada), comprising a differential oxygen analyzer (DOX, S104 Differential Oxygen Analyzer, Qubit Systems, Kingston, ON, Canada) and a CO₂analyzer (LI-6252, LI-COR Inc., Lincoln, NE, USA). In brief, dry,CO₂-free air was supplied to up to four 2.1 l air-tight metabolic chambers at a time, with flow rates controlled upstream and maintained at 1.2–1.5 l min^–1 through each chamber. Three or four chambers, each containing a single bird, were placed in a controlled temperature cabinet (Precision Incubator 850, Thermo Scientific, Waltham, MA,USA), for overnight measurement. Gases leaving the metabolic chambers were selected sequentially for analysis for 20 min, with reference air being sampled for 8.3 min between birds. Thus, the air from the chamber of each bird was sampled every 64–104 min per night. Full details of the system are provided by Marom et al. (Marom et al.,2006). During gas exchange measurements, birds were held at T_a=15°C, approximating the average natural night-time T_a in the outdoor flight cages.

We measured MR and T_b during the rest phase in six out of the eight blackcaps. Food deprivation, which leads to depletion of body energy reserves, serves as a cue to reduce T_b in heterothermic animals [for avian examples see Graf et al.(Graf et al., 1989); Hohtola et al. (Hohtola et al., 1991);McKechnie and Lovegrove (McKechnie and Lovegrove, 2003)] and mimics the physiological changes that occur during the migratory flight (Hume and Biebach, 1996; Karasov and Pinshow, 1998; Karasov et al.,2004). Thus, before measurement, depending on their initial m_b, blackcaps were deprived of food for 6 to 24 h to stimulate them to lower T_b in response to reduced energy reserves, but ensuring that no blackcap weighed less than ∼15 g when it was placed in a metabolic chamber. We began measurements of MR approximately 0.5 h before sunset; T_b was measured by telemetry, as described above. Birds were weighed before and after gas exchange measurements, and to calculate mass-specific MR we assumed that m_b decreased linearly between weighings.

We distinguished between nocturnal normothermy and hypothermia following McKechnie et al. (McKechnie et al.,2007), by assuming that normothermic nocturnal T_b is normally distributed and centered on the modal nocturnal T_b value. Using individual T_b data averaged every minute for each of seven out of eight birds, we fitted the normal distribution curve shaped by the data equal to or higher than the mode. To account for variability in the data sets that could be caused by T_b values of active birds that were in migratory restlessness (Zugunruhe), we calculated the average modal value for all birds and the average standard deviation (s.d.) for all curves. Using these averages we set the lower limit for the normothermic T_b values as the average modal value minus 2 averaged standard deviations.

We assessed the relationship between patterns of thermoregulation and refueling rates by calculating the m_b change between consecutive weighings of the birds caged outdoors, and related this difference to the difference between minimum T_b at night and the concurrent T_a. Since birds were weighed only once a day at approximately 09:00 h, we conservatively used morning m_bdata to examine the relationship between the m_b of individuals and their nocturnal minimum T_b recorded on the night preceding the weighing.

To evaluate whether the observed relationship between T_b and m_b did not result simply from the size differences among individual birds, we additionally analyzed the relationship between minimum T_b at night and a calculated body condition index. The body condition index used was the ratio of m_b to a relatively unchanging linear measure of the animal's size, i.e. the length of a folded wing, from a carpal joint to the tip of the longest primary feather, which is a good body size predictor in blackcaps (Gosler et al.,1998). Since changes in m_b during a stopover result from simultaneous and predictable changes in lean and fat masses(Gannes, 2002; Wojciechowski et al., 2005),we chose to analyze and report changes of total m_b.

Our objective was to determine whether blackcaps use hypothermia at stopovers and, if so, what are the potential benefits of this behavior. Thus,we treated each bird-night as an independent data point and indicated the number of observations (N) in our analyses of the differences between normothermic and hypothermic T_b, the effects of T_b and T_b–T_adifferences on m_b change, and the correlation between MR and minimum T_b. For the analysis of minimum metabolic rates, for each bird we used the lowest MR recorded for a 20 min interval at night.

Since it appeared that the influence of m_b on minimum nocturnal T_b may be different in lighter and in heavier blackcaps, we applied the method of Pinshow et al.(Pinshow et al., 1976) to objectively divide data into two subsets with which we could calculate separate correlations between m_b and T_b. In brief, a range of body masses was chosen that was broad enough to clearly include the point at which the relationship between m_b and T_b changed. The data points were then successively divided into two groups, a higher m_b group and a lower m_b group, and the corresponding pair of least-squares linear regression equations, and their pooled mean squares (PMS) were calculated. This process was repeated, and the pair of regression lines with the lowest PMS was considered to best determine the two subsets of data, one above and the other equal to or below the highest m_b of lighter birds. Then, we tested the correlation between m_b and nocturnal minimum T_bwithin the given subsets using methods described below. If the data were normally distributed and homoscedastic, we used Student's t-test to compare means, or the Pearson product moment correlation to describe the correlation between two variables. For data that were not normally distributed or homoscedastic, we used the Mann–Whitney U- or Spearman rank order correlation tests. A probability of P<0.05 was chosen as the lowest acceptable level of significance, and 0.05<P≤0.1 was taken to indicate a trend. Data are presented as means ± s.d.

Air temperature during data collection ranged between 6.0°C and 35.0°C. The highest daytime (35.0°C) and nighttime (21.0°C) T_a values were associated with the occurrence of Sha'arav conditions, i.e. hot, dry desert winds. The lowest night-time T_a values (∼6.0°C) occurred on the same days as the lowest daytime maxima (18.5°C and 20.5°C). During this period,there were two afternoon hail storms which resulted in a substantial drop in T_a by ∼8°C in 30 min. As a result, during the period of experiments, the circadian amplitude of T_avaried between 7.5°C and 21°C.

In distinguishing between nocturnal normothermy and hypothermia, the average modal T_b value was 39.7±0.7°C for all birds, and the average standard deviation for all curves was 1.2±0.4°C. We calculated the lower limit for normothermic T_b to be 37.4°C(Fig. 1). This is >5°C lower than the average normothermic T_b observed during the day in all birds (see next section). On several occasions, while making pilot measurements of T_b in a climate chamber at T_a=15.0±2.0°C, we observed the behavior of birds as they became hypothermic, eventually lowering their T_b to approximately 34°C. During that time, the responses of hypothermic blackcaps to auditory or tactile stimuli were sluggish compared with those of normothermic individuals.

By day, the birds' T_b values were relatively constant and averaged 42.5±0.4°C (N=59). In all blackcaps, T_b began to decrease almost immediately after dusk and was labile at night. Often, irrespective of the lowest T_battained, the decrease in T_b was continuous until birds commenced rewarming, a process that lasted between 1 h and 8.5 h. Examples of T_b and T_a recordings for two blackcaps, one recorded for 5 days and one for 3 days, are shown in Fig. 2. We recorded patterns of T_b on eight nights in five individuals, which indicated that these birds aroused for part of the night and remained active with T_b relatively high, yet significantly lower than their own day-time active values [41.4±0.7°C (N=8) and 42.3±0.4°C (N=38), respectively, Student's t-test: t=-4.81, d.f.=44, P<0.001)]. Although few, our observations of the birds' behavior suggest that these high nocturnal T_b values reflect activity associated with Zugunruhe (Fig. 3)because we observed them to fly skyward in their flight cages. These individuals began their rest-phase drop in T_b before dusk,but aroused after an hour or two. Nevertheless, nocturnal restlessness did not preclude them from lowering their T_b in the later part of the night to levels characteristic of nocturnal hypothermia(Fig. 3).

Minimum rest-phase normothermic T_b values (those above 37.4°C) in the flight cages averaged 38.8±0.8°C (N=22)whereas those of hypothermic birds were reduced by some 3°C, and averaged 35.5±1.2°C (N=12; Student's t-test: t=9.82, d.f.=32, P<0.001). The minimum T_b observed during the study was 33.0°C in one individual at a mean night-time T_a of 11.4°C. Birds generally began to increase their T_b approximately 1.5 h before civil twilight, and were normothermic at daybreak. Blackcaps rewarmed to normothermy at an average rate of 0.10±0.08°C min^–1 (N=32). We found that minimum T_b at night did not correlate with T_a.

In blackcaps of m_b <16.3 g there was a significant correlation between minimum T_b at night and m_b measured the following morning (Pearson product moment correlation: r=0.67, P<0.01, N=17; Fig. 4A), whereas in heavier birds, we found no such correlation. The lowest T_b values(33.0°C and 34.5°C) were recorded in the lightest blackcaps (13.5 g and 13.7 g, respectively). Since we did not weigh birds in the evening, we do not know the relationship between evening m_b and minimum T_b at night. The positive correlation between minimum T_b at night and a ratio of changing m_bto the constant measure of body size (length of a folded wing) in individual blackcaps supports the above observation. Birds with a better body condition index maintained higher minimum T_b at night (Pearson product moment correlation: r=0.44, P<0.05, N=32; Fig. 4B).

To examine whether the use of hypothermia could potentially affect the rate of m_b increase in refueling blackcaps, we plotted the change in m_b over 24 h against the difference between the minimum nocturnal T_b and the corresponding T_a on the night between weighings. When all available data for 24 h changes of m_b and T_b–T_a differences were analyzed(Fig. 5A), we found no correlation between the variables. However, since m_b may either increase or decrease during a stopover, in analyzing only data where m_b increased between weighings on consecutive days, we found a negative trend, indicating that the reduction of night-time T_b–T_a may affect m_b gain during refueling (Pearson product moment correlation: r=–0.40, P=0.1, N=18; Fig. 5B). Next, we selected only data for birds that lowered their T_b below normothermic rest-phase values and found that during those bird-nights the change in m_b was negatively correlated with T_b–T_a (Pearson product moment correlation: r=–0.77; P<0.05, N=7; Fig. 5C).

During 15 measurements of MR at T_a=15°C, the blackcaps' minimum T_b ranged between 33.4°C and 39.1°C, not differing from the minimum T_b values recorded overnight in the caged birds (Mann-Whitney U-test: P>0.05). By our criteria, five individuals became hypothermic and their minimum MR was correlated with their minimum nocturnal T_b (Spearman rank order correlation: r=0.54, P<0.05, N=15; Fig. 6). By becoming hypothermic, the birds significantly reduced their energy expenditure from 30.22±1.31 mW g^–1, observed in normothermic birds, to 25.94±3.00 mW g^–1 (Mann-Whitney U-test: Z=–2.45, P>0.05, N=15). However,the difference between the highest (32.74 mW g^–1) and the lowest (21.82 mW g^–1) minimum MR values at night was greater,and translates to a >30% reduction in minimum MR in birds that became hypothermic compared with ones that did not. There was also a positive correlation between minimum T_b at night and m_b before MR measurements (Spearman rank order correlation: r=0.70, P<0.01, N=15).

The present data support our hypothesis that small passerine birds are heterothermic and use hypothermia while resting at a migratory stopover,decreasing their T_b by up to 10°C below normothermic rest-phase T_b. While active, blackcaps defended a stable,high T_b in the range typical of passerine birds(Prinzinger et al., 1991). A decrease in rest-phase T_b of approximately 5°C at T_a=15°C was associated with an over 30% reduction in energy expenditure, quantitatively similar to the energy savings observed by McKechnie and Lovegrove (McKechnie and Lovegrove, 2001) in food-restricted, heterothermic speckled mousebirds under similar conditions.

The fact that in the present study the lowest T_b values were observed in the birds of least m_b could be construed to indicate that our observations were of a pathological hypothermic response in emaciated or protein starved birds that do not have large enough energy resources to maintain nocturnal normothermy. However, if that were true, we would expect them not be able to return to normothermy of their own accord,but they did.

Birds, especially those that have crossed wide ecological barriers, land at stopovers with diminished body energy reserves. Blackcaps with m_b <13 g are not unusual in Eilat during spring migration (Reuven Yosef, International Birding and Research Center, Eilat, and our own unpublished data). Additionally, because they exploit splanchnic protein during flight, birds commence their stopover with atrophied gastrointestinal organs (Karasov et al.,2004). Consequently, their digestive efficiency is significantly lower than in the later part of the stopover when net fuel accumulation occurs(Hume and Biebach, 1996; Karasov and Pinshow, 2000; Gannes, 2002). In this initial stage, hypothermia may play a crucial role; birds that forage by day would benefit from lowered energy expenditure at night while waiting for sunup to recommence rebuilding their tissues and energy stores.

Our experiments on caged birds did not reveal whether either the low m_b, usually associated with low body energy stores(Krementz and Pendleton, 1990; Wojciechowski et al., 2005),or the need for reduced energy expenditure during refueling determine the decrease in T_b at night. However, we did find that lighter birds with lower body condition indices regulate their T_bat a lower level (Fig. 4A,B),and that the rate of increase in m_b may be related to the reduction in the T_b–T_a gradient,the latter leading to decreased MR. Moreover, focusing only on blackcaps that lowered T_b below normothermic levels, we found that reducing the difference in T_b–T_aat night led to a greater increase of m_b over those specific 24 h (Fig. 5C). Taking into account that small migrating passerine birds experience large variations in m_b during their flight–refueling cycle, we presume that under natural conditions these birds benefit from hypothermia during stopover while rebuilding their energy stores. By their nature, long flights over broad ecological barriers lead to depletion of a bird's fat reserves within a short period (Schaub and Jenni, 2000; Battley et al.,2001; Pennycuick and Battley,2003). In addition to fat, protein in organs that are inactive during flight also serve as a source of fuel for flight(Hume and Biebach, 1996; Karasov et al., 2004; Bauchinger et al., 2005). Organs of the gastrointestinal tract undergo acute atrophy(Karasov et al., 2004; Bauchinger et al., 2005) that may lead to their dysfunction during the beginning of a stopover(Hume and Biebach, 1996; Gannes, 2002; Karasov and Pinshow, 2000; Tracy et al., 2005). Thus, low body mass upon arrival at a stopover, associated with impaired capacity for assimilation of nutrients, mean that birds may compensate for the high costs of T_b regulation and for low body energy stores by entering controlled hypothermia while resting at the stopover.

Although small migrating birds could benefit in terms of energy by lowering their T_b below normothermic levels while resting at stopovers, this connection has been made only for migrating hummingbirds(Carpenter and Hixon, 1988; Hiebert, 1993). During their autumn migration, the incidence of torpor in rufous hummingbirds(Selasphorus rufus) was directly related to mean body mass,suggesting that they conserve energy by using torpor during premigratory fattening (Hiebert, 1993). There is also evidence that use of daily torpor in heterothermic mammals, such as microchiropteran bats, facilitates pre-hibernatory fattening(Speakman and Rowland, 1999). More recently, Butler and Woakes (Butler and Woakes, 2001) suggested that lowered T_bmight be an important mechanism of energy conservation in migrating barnacle geese (Branta leucopsis).

Heterothermy is not the only mechanism whereby small passerines can reduce their energy requirements at stopover sites. Birds may also save energy by behavioral adjustments, among which huddling is important(Le Maho, 1977; Boix-Hinzen and Lovegrove,1998; Heinrich,2003; McKechnie et al.,2006). Huddling was previously observed in many passerines, but was mainly perceived as a strategy to survive under harsh, wintertime conditions (Steen; 1958; Smith, 1972; Heinrich, 2003). Recently, we described this behavior in blackcaps huddling at rest at migratory stopovers in Israel and Poland (Wojciechowski et al., 2008). Based on our limited observations of huddling, in the present study we cannot quantitatively determine whether huddling significantly contributed to the observed changes in T_band estimated energy savings, but since birds could huddle while hypothermic,doing both together would probably lead to additive energy savings.

Based on our initial respirometry measurements (M.S.W. and B.P.,unpublished), we calculated that huddling blackcaps save up to 30% energy overnight compared with the energy expenditure of normothermic solitary blackcaps. This is consistent with data for other species, where, depending on group size and T_a, birds may reduce their metabolic demands by 50% while resting in a huddle compared with birds resting alone(Chaplin, 1982; Du Plessis and Williams, 1994; Boix-Hinzen and Lovegrove,1998; McKechnie and Lovegrove,2001).

McKechnie and Lovegrove (McKechnie and Lovegrove, 2001) pointed out that patterns of thermoregulation,where T_b is labile and lacking a clearly defended set point as in speckled mousebirds, may be associated with huddling behavior. Comparing our data with that of McKechnie and Lovegrove(McKechnie and Lovegrove,2001) and of McKechnie et al. (McKechnie et al., 2004) leaves an impression of close similarity between the patterns of thermoregulation in blackcaps at migratory stopover and that of speckled and white-backed mousebirds in terms of interactions between physiological and behavioral mechanisms of energy conservation. Both blackcaps and mousebirds lack an obvious set-point for nocturnal T_b regulation, both let their T_b drop well below the normothermic levels observed in other avian taxa, but not as far as in hummingbirds that undergo deep,mammal-like torpor with T_b decreasing by as much as 30°C (for a review, see McKechnie and Lovegrove, 2002). Both, blackcaps and mousebirds, also huddle while resting, which may be additionally beneficial for maintaining T_b.

The rate of rewarming from torpor to normothermic T_bis, on average, lower in birds than in mammals and this difference increases with increasing m_b(McKechnie and Wolf, 2004). It was found in mammals that active thermogenesis at low T_bvalues is a prerequisite for the arousal from torpor(Geiser et al., 2002; Jefimow et al., 2004). Although there is some evidence for non-shivering thermogenesis (NST) in birds, so far no specialized organ has been identified and it appears that, in birds, muscles are its main source (Bicudo et al., 2001). The average rate of rewarming observed in blackcaps was much slower than allometrically predicted for a 15 g bird by McKechnie and Wolf (McKechnie and Wolf,2004), namely, 0.10±0.08°C min^–1versus 0.49°C min^–1. Keeping in mind that migratory flight brings about significant atrophy of flight muscles and other organs that could potentially be involved in both shivering and non-shivering thermogenesis (Karasov et al.,2004; Bauchinger et al.,2005), we suggest that the observed deviations from normothermy,as well as huddling behavior, might result from decreased capacity for thermoregulatory heat production in heterothermic birds, blackcaps among them.

The frequent use of heterothermy by blackcaps implies that its benefits exceed its potential costs, as suggested for small mammals by Humphries et al.(Humphries et al., 2003). We propose that both heterothermy, known to accelerate fattening prior to hibernation in mammals and in refueling migratory hummingbirds, as well as huddling, serve to save energy in refueling passerine migrants that, in turn,might shorten their stopover sojourn.

LIST OF ABBREVIATIONS

 
• m_b

  body mass

 
• MR

  metabolic rate

 
• PMS

  pooled mean squares

 
• T_b

  body temperature

 
• T_a

  air temperature