Patterns and dynamics of rest-phase hypothermia in wild and captive Blue Tits during winter

Abstract

We evaluated biotic and abiotic predictors of rest-phase hypothermia in wintering blue tits (Cyanistes caeruleus) and also assessed how food availability influences nightly thermoregulation. On any given night, captive blue tits (with unrestricted access to food) remained largely homeothermic, whereas free-ranging birds decreased their body temperature (T
b) by about 5°C. This was not an effect of increased stress in the aviary as we found no difference in circulating corticosterone between groups. Nocturnal T
b in free-ranging birds varied with ambient temperature, date and time. Conversely, T
b in captive birds could not be explained by climatic or temporal factors, but differed slightly between the sexes. We argue that the degree of hypothermia is controlled predominantly by birds’ ability to obtain sufficient energy reserves during the day. However, environmental factors became increasingly important for thermoregulation when resources were limited. Moreover, as birds did not enter hypothermia in captivity when food was abundant, we suggest that this strategy has associated costs and hence is avoided whenever resource levels permit.

Full text

J Comp Physiol B (2009) 179:737–745
DOI 10.1007/s00360-009-0357-1
123
ORIGINAL PAPER
Patterns and dynamics of rest-phase hypothermia
in wild and captive blue tits during winter
Andreas Nord · Johan F. Nilsson · Maria I. Sandell ·
Jan-Åke Nilsson
Received: 21 January 2009 / Revised: 5 March 2009 / Accepted: 19 March 2009 / Published online: 8 April 2009
© Springer-Verlag 2009
Abstract We evaluated biotic and abiotic predictors of
rest-phase hypothermia in wintering blue tits (Cyanistes
caeruleus) and also assessed how food availability inXu-
ences nightly thermoregulation. On any given night, cap-
tive blue tits (with unrestricted access to food) remained
largely homeothermic, whereas free-ranging birds decreased
their body temperature (Tb) by about 5°C. This was not an
eVect of increased stress in the aviary as we found no
diVerence in circulating corticosterone between groups.
Nocturnal Tb in free-ranging birds varied with ambient tem-
perature, date and time. Conversely, Tb in captive birds
could not be explained by climatic or temporal factors, but
diVered slightly between the sexes. We argue that the
degree of hypothermia is controlled predominantly by
birds’ ability to obtain suYcient energy reserves during the
day. However, environmental factors became increasingly
important for thermoregulation when resources were lim-
ited. Moreover, as birds did not enter hypothermia in cap-
tivity when food was abundant, we suggest that this
strategy has associated costs and hence is avoided when-
ever resource levels permit.
Keywords Body temperature regulation · Corticosterone ·
Energy conservation · Food availability · Heterothermy ·
Life history trade-oVs
Introduction
Spending the winter at high latitudes is an energetically
demanding task for most organisms. The situation is espe-
cially problematic for small birds that must cope with a
large surface area to volume ratio and high sustained mass
speciWc metabolic rates but still have to maintain high body
temperatures (Tb) (passerine average: 41.6°C; Prinzinger
et al. 1991). This is especially challenging during times of
low ambient temperatures, short day lengths and restricted
access to food due to snow or ice cover, often experienced
during winters at high latitudes. These diYculties are
ampliWed by the fact that due to their high metabolic
demands, small birds need more energy to survive than is
contained in food carried in their gut. Instead, these birds
utilize subcutaneous fat deposits built up on a daily basis as
overnight metabolic fuel (reviewed by Pravosudov and
Grubb 1997). These fat reserve dynamics results in small
birds (body mass <20 g) having to deposit fat reserves cor-
responding to about 10% of their morning mass as fuel for
the next night (Haftorn 1992). This means that the daily
routine of an overwintering small bird is largely centred on
replenishing energy reserves depleted during the preceding
night by foraging intensively during daylight hours. How-
ever, on cold nights birds can reduce metabolic demands
and consequently daily foraging needs by entering a state of
hypothermia during the resting phase (McKechnie and
Lovegrove 2002).
Rest-phase hypothermia (hereafter referred to simply as
hypothermia), deWned as an actively regulated reduction in
Tb below the normothermic setpoint, is qualitatively similar
to other heterothermic responses (daily torpor, hibernation)
but diVers in the respect that these birds remain responsive
to external stimuli (Reinertsen 1996). Hypothermia can be
an eYcient way of conserving energy as reductions in
Communicated by G. Heldmaier.
A. Nord (&) · J. F. Nilsson · M. I. Sandell · J.-Å. Nilsson
Department of Animal Ecology, Lund University,
Ecology Building, 223 62 Lund, Sweden
e-mail: Andreas.Nord@zooekol.lu.se

738 J Comp Physiol B (2009) 179:737–745
123
metabolic expenditure during hypothermia can be substan-
tial (Mayer et al. 1982; Reinertsen and Haftorn 1984; Mad-
docks and Geiser 1997; Sharbaugh 2001; Cooper and
Gessaman 2005) and even comparatively mild hypothermia
(Tb reduction ·10°C) can decrease metabolic demands by
as much as 50% (Cooper and Gessaman 2005). The degree
to which birds use hypothermia varies predictably with
ambient temperature, season and availability of food
resources (see Reinertsen 1996; McKechnie and Lovegrove
2002; Schleucher 2004 for reviews). Hypothermia thus
appears to be a Wne-tuned physiological mechanism pre-
dictably employed to counter prevailing ambient conditions
on both short and long time scales.
Given the apparent energetic beneWts of hypothermia, it
is puzzling that it is not used routinely by overwintering
birds. In theory at least, this physiological strategy would
be preferred since the decreased energy expenditure at
night would augment survival probability by reducing both
the risk of overnight starvation and predation costs related
to diurnal foraging (Clark and Dukas 2000; Welton et al.
2002; Brodin 2007). However, vigilance and alertness is
reduced during hypothermia (Haftorn 1972; Rashotte et al.
1998; but see Merola-Zwartjes and Ligon 2000 for evi-
dence of no such eVects), which might impair birds’ ability
to detect and escape from nocturnal predators, thereby ren-
dering hypothermic birds more susceptible to overnight
predation (Grubb and Pravosudov 1994; Pravosudov and
Lucas 2000). If the use of hypothermia is constrained by
increased predation costs at night, we predict that the mag-
nitude of Tb decrease will depend on the degree of energetic
stress. If so, a bird settling in for the night must trade-oV the
beneWts of decreased energy expenditure with increased
risks of nocturnal predation, while simultaneously consid-
ering the likelihood of being able to replenish energy
deposits on the subsequent day.
To date, the majority of published data on avian hetero-
thermy is for captive birds (reviewed in McKechnie and
Lovegrove 2002; Schleucher 2004). In contrast, despite
several recent contributions (Brigham et al. 2000; Körtner
et al. 2000, 2001; Dolby et al. 2004; Fletcher et al. 2004;
McKechnie et al. 2004, 2007; Cooper et al. 2008), data on
heterothermia for free-ranging birds are relatively rare.
Consequently, the causes and consequences of the use of
nocturnal hypothermia in natural populations remain
largely unexplored. We used both Weld and laboratory data
to explore biotic and abiotic factors inXuencing the use of
nocturnal hypothermia in blue tits (Cyanistes caeruleus). In
particular, we evaluated the impact of food availability on
hypothermia by comparing free-ranging birds and birds
from the same population being temporarily kept in outdoor
aviaries with unrestricted access to food. Captivity may
impose stress on wild animals, which could be reXected by
changes in the levels of circulating stress hormones (cf. Cyr
et al. 2007; Cyr and Romero 2008; Dickens et al. 2009). In
some cases, exogenous administration of such stress hor-
mones has been found to aVect metabolic rates of birds (e.g.
Astheimer et al. 1992; Spencer and Verhulst 2008), sug-
gesting that there might be a relationship also between
stress hormones and Tb. Because such a relationship could
bias data interpretation, we measured circulating levels of
corticosterone, the major avian stress hormone, in free-
ranging and captive birds, respectively.
The purpose of our study was twofold: (1) to assess the
variation and determinants of nocturnal hypothermia in
free-ranging birds and (2) to investigate the inXuence of
food availability on the use of hypothermia by blue tits.
Materials and methods
General
The blue tit is a small passerine (body mass approximately
12 g) that is resident year round at mid to high latitudes.
Hence, it frequently encounters low ambient temperatures
and constrained foraging opportunities in winter. This, in
combination with the blue tits’ small size, imposes large
challenges for maintaining a positive energy balance at this
time of the year. This makes blue tits well suited for studies
of adaptive energy management and energy conservation
mechanisms during winter.
Field work took place from November 2007 to February
2008 in the Revinge area, about 20 km east of the city of
Lund in southern Sweden (55°42⬘N, 13°28⬘E). The study
area, consisting of small deciduous woodlots and groves
interspersed between pastures and arable Welds, contains
about 500 nest boxes readily used for roosting by blue tits
on winter nights.
Climatic data for the study period were supplied from a
permanent weather station positioned in the centre of the
study area. We measured six related climatic variables;
ambient temperature at the time of sampling, mean temper-
ature during the preceding day (sunrise to sunset) and mean
and minimum temperature during the sampling night and
the night previous to sampling, respectively. The photope-
riod during the study period ranged from approximately 9 h
(early November) to 9.5 h (mid February), reaching an
overall minimum of 7 h in mid December.
Field methods
Starting one hour after sunset, we searched nest boxes for
blue tits. When we encountered a bird, we measured Tb
using a Testo 925 digital thermometer (Testo AG, Len-
zkirch, Germany) equipped with a standard copper thermo-
couple (Ø 0.9 mm; ELFA AB, Järfälla, Sweden) inserted

J Comp Physiol B (2009) 179:737–745 739
123
12 mm into the cloaca. Further insertion did not alter the
temperature readings. Three readings were obtained within
20 s of capture. The intra-sample repeatability coeYcient
was high (r= 0.68, F197,396 = 427.0, P< 0.001; Lessells and
Boag 1987) and the mean of the three measurements was
used in all analyses. Subsequent to body temperature mea-
surements, we marked all unmarked birds with a numbered
aluminium ring and a plastic colour ring, measured mass
(to the closest 0.1 g), tarsus length (to the closest 0.1 mm)
and wing length (to the closest 0.5 mm) and also scored
subcutaneous fat reserves in the tracheal pit on a seven-
grade scale (where 0 = no fat; Pettersson and Hasselquist
1985). Once measurements were completed, the bird was
transferred back into the nest box. The protocol took less
than 5 min to complete. Birds appeared to be undisturbed
by treatment and always remained in the nest box for the
rest of the night. In total, we sampled 142 individuals con-
stituting a random sample with regards to age and sex
(1
2= 0.10; P= 0.75). Individuals were sampled one to
eight times. Birds were weighed and scored for subcutane-
ous fat reserves on all sampling occasions, but measure-
ments of tarsus- and wing lengths were performed during
the Wrst sampling event only.
Between early November and mid December, we mist
netted 27 individuals to measure daytime body temperature
(intra-measurement r= 0.87, F26,54 =191.22, P<0.001,
Lessells and Boag (1987); average used in analyses). The
daytime sample was heavily biased towards yearling
females (nmale, 1 year = 8, nmale, ¸2 years = 0, nfemale,
1 year = 16, nfemale, ¸2 years = 3). Age and sex distribution
was not statistically tested, as data did not fulWl the assump-
tions of formal goodness of Wt tests.
Aviary methods
Two weeks prior to experimental onset, 27 individual blue
tits (nmale, 1 year = 7, nmale, ¸2 years = 0, nfemale, 1 year =
16, nfemale, ¸2 years = 3, nunknown =1; diVerent from the
birds used for measurements of daytime Tb) were mist
netted and transferred to outdoor aviaries at the centre of
the study area. The distance from any of the Weld sites to
the aviary never exceeded 5 km. Birds were housed in pairs
in aviaries measuring 2.0 £3.0 £2.5 m (width £
length £height). We equipped aviaries with multiple
perches, creating a spatially complex structure and also
screened them from the outside to prevent attacks by avian
predators (mainly sparrowhawks Accipiter nisus and
tawny owls Strix aluco). Each bird also received a standard
nest box, identical to those used by the free-ranging
population, for night-time roosting. Birds were kept on a
diet of sunXower seeds and lard balls containing mixed
millet seeds ad libitum. Meal worms were oVered
occasionally.
Beginning 2 weeks after capture, we sampled night-time
Tb (intra-measurement r= 0.82, F115,232 = 562.35, P< 0.001
(Lessells and Boag 1987); mean used in analyses) of roost-
ing captive birds approximately weekly throughout the
study period (n= 7 occasions), commencing measurements
at least 1 h after sunset. Measurements in the aviaries coin-
cided with subsequent sampling in the Weld on all but one
occasion. Birds were weighed and scored for fat reserves on
each but the Wrst two sampling occasions. Birds normally
spent the night in the nest boxes, but occasionally individu-
als roosted unsheltered in the aviary. These birds were not
sampled as they could not be caught and handled within the
same short time frame as birds roosting in nest boxes.
Because the majority of captive birds escaped on the 3rd of
January 2008, we did not include aviary observations sub-
sequent to this date in the analyses. The exclusion of these
observations did not aVect the outcome or interpretation of
statistical analyses.
Corticosterone sampling
We collected blood samples from the jugular vein of 12
aviary birds and 17 birds from the free-living population
not included in the Tb study on two consecutive nights in
December 2007. To avoid the eVect of handling stress on
baseline levels of stress hormones, all blood samples were
drawn within 3 min (mean = 1.05 §0.07 min) of capture
(Schoech et al. 1999; Romero and Romero 2002). Blood
samples were kept on ice for 1–2 h until being centrifuged
at 3,000 rpm for 10 min. After spinning, plasma was sepa-
rated using a Hamilton syringe (Hamilton Bonaduz AG,
Bonaduz, Switzerland) and then immediately frozen at
¡50°C. Samples were analysed for the presence of cortico-
sterone using the methods of WingWeld and Farner (1975).
Tritiated corticosterone was added to each sample to calcu-
late recovery percentages following extraction (approxi-
mately 2,000 cpm/hormone, PerkinElmer Life Sciences,
Upplands Väsby, Sweden). Steroids were extracted with
4 ml diethyl ether and re-suspended in re-distilled ethyl
acetate and isooctane (10:90%). After extraction, cortico-
sterone levels were measured by radioimmunoassay with
tritiated corticosterone and speciWc corticosterone antibod-
ies (Endocrine Science, Calabasas, California). Each sam-
ple was run in duplicate and mean values were compared
with a standard curve. Detection level for corticosterone
was 0.5 ng (ml plasma)¡1. All samples were analysed in a
single assay and intra-assay variation was 10%. Three sam-
ples with corticosterone levels below the detection limit
(naviary =2, nWeld = 1) were conservatively assigned the low-
est detectable plasma corticosterone level [0.05 ng (ml
plasma)¡1] in the ensuing analyses. Moreover, four aber-
rant observations displaying corticosterone levels well
above baseline levels [>12 ng (ml plasma)¡1; naviary =3,

740 J Comp Physiol B (2009) 179:737–745
123
nWeld = 1] were omitted from the data. Inclusion of these
observations in the analyses did not aVect the results.
Statistical analyses
All statistical tests were performed using SAS 9.2 for
Windows. Tb of day- and night-time free-living birds and
captive birds were compared with a linear mixed model
(PROC MIXED) Wtted using restricted maximum likeli-
hood (REML) methods, with Tb as the dependent variable,
origin as a Wxed eVect and bird identity nested within origin
as a random factor. DiVerences between groups were com-
pared using calculated least squares means, with P-values
adjusted for unbalanced multiple comparisons using the
Tukey–Kramer method (Littell et al. 2006). We analysed
factors that could explain the variation in Tb in free-living
and captive birds, respectively, using similar linear mixed
models. Independent factors included all biometric mea-
sures and fat score as well as climatic and linear and
squared temporal (date, minutes from sunset) variables as
covariates and bird identity as a random factor. However,
because of sample bias (see above), we did not include bird
age in the aviary models. Furthermore, to avoid loss of pre-
dictive power of the captive bird model (that was based on
a considerably smaller data set), we did not allow for
squared eVects of date and time of the day in the saturated
model. We included all two-way interactions with age and
sex in the full models of free-living birds and all two-way
interactions with sex in the full model of aviary birds, but
as interaction eVects did not explain any of the variation in
Tb (P> 0.10 in all cases) they are not presented in the Wnal
models. Because measurements were performed by two
persons (A. N., J. F. N.), we included “observer” as a Wxed
factor in all analyses, but it did not explain any variation in
the dependent variable (P> 0.8 in all cases) and was
accordingly excluded from the Wnal models. Full models
were reduced by backward elimination (Seber and Lee
2003) until only signiWcant variables remained in the
model.
Repeatability of within-individual measurements was
calculated following Lessells and Boag (1987), using the
estimates from a one-way general linear model (PROC
GLM) with Tb as the dependent variable and bird identity as
a grouping factor. SigniWcances of random factors were
assessed by comparing the restricted log likelihood ratio of
the reduced and the saturated model to a Chi square distri-
bution with one degree of freedom (Sokal and Rohlf 1995).
For all analyses, denominator degrees of freedom for Wxed
eVects were calculated using the Satterthwaite approxima-
tion (Littell et al. 2006).
Because of multicollinearity between climatic variables,
we Wtted and subsequently reduced full models in both the
aviary and the free-living bird data set with one randomly
chosen climatic variable, using REML methods. When only
signiWcant variables remained, we re-Wtted the model with
maximum likelihood (ML) methods, including each cli-
matic variable in turn. We compared model Wt using the
Akaike Information Criterion, AIC (Seber and Lee 2003)
and used the model with the lowest value of AIC as the
Wnal model.
Plasma corticosterone level between groups was compared
using a general linear model (PROC GLM) with hormone lev-
els as dependent, origin (i.e. free-living or captive) as a Wxed
factor and handling time as a covariate. We included
“observer” (A. N., J.-Å. N.) in the full model, but it did not
explain any variation in plasma corticosterone levels (P>0.7).
All means and intercepts () are reported with their stan-
dard errors (mean/§SE) and all signiWcances, except
those for the log-likelihood ratio tests, are two-tailed.
Results
General
There were large diVerences in Tb between day and night
and also between free-living and aviary birds (PROC
MIXED: F2,120 = 604.10, P< 0.0001; Fig. 1). Post hoc
comparisons revealed that daytime Tb in free-living birds
(= 42.6 §0.2°C) was signiWcantly higher than night-time
Tb in both free-living (=37.8§0.1°C; t237 = 25.1,
P< 0.0001) and captive birds (= 41.9 §0.1°C; t161 =
¡3.0, P= 0.0092). Furthermore, nocturnal Tb was signiW-
cantly lower in free-living compared to captive birds
(t77.9 = 29.2, P< 0.0001). We found no signiWcant diVer-
ence in body mass between captive (mean = 11.76 §
0.11 g) and free-ranging birds (mean = 11.65 §0.06 g;
t test: t154 =0.77, P= 0.19). However, captive birds had
somewhat larger subcutaneous fat reserves (mean = 3.26 §
0.15) than did free-ranging birds (mean = 2.60 §0.09;
Mann–Whitney U test: U= 841.0, P=0.001).
Plasma corticosterone levels did not diVer between
free-living and captive birds (PROC GLM: F1,25 = 0.94,
P=0.65; mean=2.43§0.67 ng (ml plasma)¡1 and
2.67 §0.63 ng (ml plasma)¡1 for captive and free-living
birds, respectively). Handling time did not aVect plasma
corticosterone concentration (PROC GLM: F1,27 = 2.68,
P=0.11).
Field birds
Using backward elimination of non-signiWcant main eVect
terms on a saturated linear mixed model (PROC MIXED)
with Tb as the dependent variable, we excluded, in turn, tar-
sus length (F1,147 =0.14, P= 0.71), minutes from sunset2
(F1,151 =0.33, P= 0.57), mass (F1,122 =0.21, P=0.65), sex

J Comp Physiol B (2009) 179:737–745 741
123
(F1,104 =0.10, P= 0.76), fat score (F5,160 = 0.76, P= 0.58)
and age (F1,123 =0.67, P= 0.42) from the model. However,
Tb was strongly dependent on date (F1,185 =8.52,
P= 0.0039) and date2 (F1,186 =11.63, P= 0.0008), reach-
ing its minimum in mid winter and being relatively higher
in late fall and late winter (Fig. 2). Moreover Tb increased
with mean temperature during the night prior to sampling
(F1,191 = 10.48, P= 0.0014; Fig. 3) and decreased with
minutes from sunset (F1,172 =4.74, P= 0.0309; Fig. 4).
The random eVect term for bird identity was not signiW-
cant (Likelihood ratio test: = 1.4, P= 0.23). Nor was Tb
within an individual signiWcantly repeatable between sam-
pling events (r= 0.10; PROC GLM: F55,142 = 1.15,
P=0.27).
Aviary birds
As for free-living birds, variation in Tb in captive birds was
not explained by tarsus length (PROC MIXED:
F1,20.8 =0.23, P=0.63), mass (F1,37.7 =1.06, P=0.31) and
subcutaneous fat score (F3,37.2 =1.74, P= 0.18). Further-
more, we found no eVect of date (F1,34.3 = 0.15, P=0.70),
mean temperature during the night prior to sampling
(F1,36.9 = 1.24, P= 0.27) and minutes from sunset
(F1,26.3 = 0.67, P= 0.42) on nightly Tb in aviary birds.
However, Tb was signiWcantly dependent on sex of captive
Fig. 1 DiVerences in core body temperature between free-living blue
tits during the day (Weld day), free-living blue tits during the night
(Weld night) and birds from the same population kept in outdoor aviar-
ies and measured during the night (aviary night). In cases of more than
one sample for an individual, the mean of these was used for the illus-
trations. Boxes show medians, 1st and 3rd quartiles. Whiskers extend
to the last observation within 1.5 £the interquartile range, IQR. Open
circles denote observations occurring between 1.5 and 3 IQR. Fille
d
circles denote points greater than 3 IQR
Field (night)Aviary (night)Field (day)
body temperature (°C)
44.0
42.0
40.0
38.0
36.0
34.0
***
***
**
142
27
27
Fig. 2 Nocturnal core body temperature of wintering free-living blue
tits (n= 142) in relation to date (days after 1 October). In cases of mul-
tiple measurements on a single individual, the sample from the coldest
night of the study period is shown in the Wgure. The line represents a
quadratic Wt to the data. The relationship is signiWcant at P= 0.004;
Y= 0.0002x2¡0.0247x+ 38.581; R2=0.081
35
36
37
38
39
40
41
42
43
20 40 60 80 100 120 140
body temperature (°C)
date
Fig. 3 Nocturnal core body temperature of free-living blue tits
(n= 142) as a function of mean ambient temperature one night before
the sampling night. In cases of multiple measurements on a single
individual, the sample from the coldest night of the study period is
shown in the Wgure. The relationship is signiWcant at P= 0.001;
Y= 0.0915x+ 37.717; R2=0.066
35
36
37
38
39
40
41
42
43
-2 0 2 4 6 8 1 0
body temperature (° C)
ambient temper atur e (°C)
Fig. 4 Nocturnal core body temperature of free-living blue tits
(n= 142) in relation to hours from sunset. In cases of multiple mea-
surements on a single individual, the sample from the coldest night o
f
the study period is shown in the Wgure. The relationship is signiWcant
at P= 0.03; Y=¡0.0022x+ 38.577; R2=0.045
body temperature (°C)
35
36
37
38
39
40
41
42
43
2345678
time (h)

742 J Comp Physiol B (2009) 179:737–745
123
birds (F1,17.3 =6.02, P= 0.025; Fig. 5) with males, on aver-
age, maintaining their Tb 0.7°C higher than females
(male =42.5§0.24°C, female = 41.8 §0.14°C).
The random eVect term for bird identity was not signiW-
cant (Likelihood ratio test: =3.0, P= 0.083). However,
Tb was weakly but signiWcantly repeatable within individu-
als between sampling events (r= 0.23; PROC GLM:
F26,65 =1.97, P=0.014).
Discussion
We have shown that free-ranging blue tits decreased Tb
almost 5°C compared to daytime levels, whereas birds from
the same population temporarily kept in outdoor aviaries
remained more or less homeothermic throughout the night
(Fig. 1). Since free-ranging and captive birds were of the
same origin and experienced identical ambient conditions
and photoperiods, the main remaining diVerence between
the groups was the amount of available food. It is widely
recognized that patterns and dynamics of heterothermy
often diVer between free-ranging and captive birds (Geiser
et al. 2000). The discrepancy results from a lower ampli-
tude of diurnal body temperature oscillation in captive indi-
viduals in some species (Buttemer et al. 2003; Cooper et al.
2008) and a much reduced propensity to enter hypothermia
in others (Bech and Nicol 1999; Körtner et al. 2001). Con-
sequently, it has been argued that because of stress, birds in
captivity appear reluctant to use hypothermia, and that the
resulting increased nightly energy expenditure is made pos-
sible by the access to food in captivity (Körtner et al. 2001).
However, we found no diVerences in plasma corticosterone
levels between free-ranging and captive blue tits, implying
that stress levels did not diVer between the groups. We thus
feel conWdent that the observed diVerence in the use of noc-
turnal hypothermia was attributable to eVects of food avail-
ability per se.
There is good reason to assume that food availability is
instrumental in the use of nocturnal hypothermia, because
food deprivation is often necessary to induce hypothermia
in captive birds otherwise fed ad libitum (Reinertsen and
Haftorn 1986; McKechnie and Lovegrove 2003; Laurila
and Hohtola 2005). Incidental evidence for the importance
of food resources in body temperature dynamics is also pro-
vided by the fact that species with a capacity for substantial
decreases in Tb generally are those that rely predominantly
on food sources that are ephemeral, weather-dependent or
diYcult to store endogenously (McKechnie and Lovegrove
2002). Accordingly, we expected that birds feeding mainly
on predictable resources or in non-Xuctuating environments
would be less prone to enter hypothermia, as they would
have a higher relative probability of replenishing energy
reserves the subsequent day (Schleucher 1999, 2001).
The fact that birds seem to avoid hypothermia when food
is abundant suggests that, even though potentially beneW-
cial for energy conservation purposes (e.g. Maddocks and
Geiser 1997; Sharbaugh 2001; Cooper and Gessaman
2005), using heterothermy at night incurs a cost. The risk of
predation has been suggested to be such a cost, inXuencing
optimal nightly Tb in birds as a trade-oV between energetic
constraints and predation risks (Laurila and Hohtola (2005).
We, therefore, argue that captive blue tits in the current
Fig. 5 Nocturnal core body
temperature (Tb) during winter
of male (n= 7) and female
(n= 19) blue tits kept in outdoor
aviaries. Illustrations are based
on one to six observations per
individual and each box repre-
sents one bird. Boxes show
medians, 1st and 3rd quartiles.
Whiskers extend to the last
observation within 1.5 £the
interquartile range, IQR. Open
circles denote observations
occurring between 1.5 and 3
IQR. Dashed lines show Tb
means for females and males,
respectively. DiVerences are
signiWcant at P=0.03
44.0
43.0
42.0
41.0
40.0
39.0 Male
Female
body temperature (°C)

J Comp Physiol B (2009) 179:737–745 743
123
study actively maintained Tb close to daytime levels to min-
imize nightly predation risk and that this was possible as a
result of ad libitum food. Conversely, free-ranging birds
were probably exposed to an omnipresent energetic con-
straint, which made hypothermia more beneWcial regardless
of predation risks. Hence, birds should avoid costs by keep-
ing Tb at high levels whenever possible. However, this is
probably possible only under certain conditions (such as
benign ambient conditions or abundant food resources)
which might never occur during winter in the temperate
region.
Nocturnal Tb varied predictably with date in free-
ranging birds (Fig. 2), reaching an overall minimum in
mid winter and being relatively higher in late fall and
late winter. Changes in Tb regulation with season occur
in free-ranging populations of some species (Carpenter
1974; Chaplin 1976; Reinertsen and Haftorn 1983;
Waite 1991; Brigham et al. 2000; Maddocks 2001) and
may reXect a response to variation in day length
(Reinertsen 1996). During the short days of mid winter
at high latitudes, birds have little choice but to forage
intensively to replenish energy reserves (Haftorn 1992).
As days become longer, more time for foraging is avail-
able and the need for energy conservation at night
decreases (Welton et al. 2002). However, ambient con-
ditions are also subjected to stochasticity and can
change rapidly from one day to the next. In accordance
with previous studies (Merola-Zwartjes 1998; Merola-
Zwartjes and Ligon 2000; Dolby et al. 2004; McKechnie
et al. 2004; Lane et al. 2004), we found Tb to be posi-
tively related to ambient temperature (Fig. 3), which
might reXect a response to such environmental stochas-
ticity. We thus argue that, although Tb regulation in blue
tits may vary seasonally (see above), birds also adjust Tb
in relation to ambient temperature. This allows winter-
ing blue tits to Wne tune energy management and opti-
mize the trade-oV between energy expenditure and
predation risk on a nightly basis.
On a smaller temporal scale, Tb also varied as a function
of progression of the night. More speciWcally, Tb decreased
slowly (0.12°C h¡1) after sunset throughout the sampling
period (Fig. 4). Because we did not sample for more than
about 8 h past sunset, we have no data on arousal. How-
ever, considering the relatively long period of constant
decrease, there is reason to expect the nocturnal hypother-
mic response in wintering free-ranging blue tits does not
show the distinct “entry-maintenance-arousal”-pattern
(McKechnie and Lovegrove 2002) which occurs in other
parid species (Willow tit Poecile montanus: Reinertsen and
Haftorn 1983, 1986; Juniper titmouse Baeolophus ridg-
wayi, Mountain chickadee Poecile gambeli: Cooper and
Gessaman 2005). Rather, our results suggest that in blue tits
there is a slow continuous decrease during the Wrst half of
the night (see e.g. McKechnie and Lovegrove 2001a, b,
2002; Cooper et al. 2008 for examples of such Tb traces).
The low amplitude daily cycles in Tb in aviary birds was
well within the range of the 1–2°C variation ascribed to the
circadian rhythm (Reinertsen and Haftorn 1986). Because
these birds probably did not enter hypothermia, it is no sur-
prise that variation in nocturnal Tb could not be explained
by any of the predictors known to inXuence Tb variation in
free-ranging birds. The absence of a hypothermic response
in captive birds diVers markedly from data in similar stud-
ies, in which nocturnal Tb not only shows pronounced daily
cycles but also varies predictably with both ambient
temperature (Mayer et al. 1982; Reinertsen and Haftorn
1984, 1986; Maddocks and Geiser 1997; Sharbaugh 2001;
McKechnie and Lovegrove 2001b; Downs and Brown
2002; Cooper and Gessaman 2005; Maddocks and Geiser
2007) and season (Carpenter 1974; Chaplin 1976; Reinertsen
and Haftorn 1983; Waite 1991; Maddocks and Geiser
2007). However, the majority of these studies were per-
formed over larger temperature ranges, at higher latitudes
or in colder ambient temperatures than our study. It, there-
fore, seems likely that even though captive conditions
might aVect patterns and use of facultative hypothermia,
this is to some extent dependent on ambient conditions. If
true, ad libitum access to food might suYce to preclude a
hypothermic response during the comparatively mild win-
ters in southern Sweden, but not during more energetically
demanding winters.
Even though our sample was small and heavily skewed,
nocturnal Tb diVered signiWcantly between sexes in the cap-
tive birds (Fig. 5). Testosterone appears to play a signiWcant
role in explaining sex-speciWc variation in thermoregula-
tion in several mammals (Ruby et al. 1993; McKechnie and
Lovegrove 2002 and references therein; Mzilikazi and
Lovegrove 2002). Even though comparable data for birds
are lacking, Merola-Zwartjes and Ligon (2000) found that
torpor in breeding individuals of the Puerto Rican tody
(Todus mexicanus) was restricted to females and proposed
that higher testosterone levels of males might preclude tor-
por. The observed sex diVerences in our captive birds could
potentially be explained if male testosterone production
during winter depends on food levels. This would also
explain the lack of sex diVerences in free-ranging birds.
Conclusions
We have shown that hypothermia in free-ranging blue tits is
a dynamic process that is regulated to ensure an optimal
trade-oV between energy management and costs (e.g. pre-
dation risk) during winter nights. This is supported by the
lack of repeatability of Tb within free-ranging individuals.
Blue tits seem to respond both to predictable factors, such

744 J Comp Physiol B (2009) 179:737–745
123
as day length and to short-term variation in e.g. ambient
temperature. Furthermore, the trade-oV nature of hypother-
mia is consistent with the fact that blue tits remained
largely homeothermic at night in situations with high food
availability. Thus, when hypothermia is not needed to
reduce risks of overnight starvation, blue tits do not pay the
cost of potentially increased predation risk during the night
(Pravosudov and Lucas 2000). Our results, therefore, stress
the overall importance of energetic limitations in regulating
the dynamics of nocturnal hypothermia in overwintering
birds. Surprisingly, in spite of convincing evidence that
food related energetic stress can be a major determinant of
Tb dynamics in birds, experimental evidence from natural
populations is scant (but see Woods and Brigham 2004).
Thus, studies of patterns, dynamics and consequences of Tb
regulation in wild populations are critically needed.
Acknowledgments We are grateful to Charlotta Borell Lövstedt for
supplying climatic data from the study area. Comments from Indrikis
Krams and three anonymous reviewers improved a previous version of
the manuscript. This study was supported by grants from the Swedish
Research Council (to J.-Å. N.). All experimental protocols adhere to
the guidelines of the Swedish Animal Welfare Agency and were
approved by the Malmö/Lund Animal Care Committee, Sweden
(permit nos. M53-06, M237-07).
References
Astheimer LB, Buttemer WA, WingWeld JC (1992) Interactions of cor-
ticosterone with feeding, activity and metabolism in passerine
birds. Ornis Scand 23:355–365
Bech C, Nicol SC (1999) Thermoregulation and ventilation in the
tawny frogmouth, Podargus strigoides: a low-metabolic avian
species. Aust J Zool 47:143–153
Brigham RM, Körtner G, Maddocks TA, Geiser F (2000) Seasonal use
of torpor by free-ranging Australian owlet-nightjars (Aegotheles
cristatus). Physiol Biochem Zool 73:613–620
Brodin A (2007) Theoretical models of adaptive energy management
in small wintering birds. Phil Trans R Soc Lond B Biol Sci
362:1857–1871
Buttemer WA, Nicol SC, Sharman A (2003) Thermoenergetics of
pre-moulting and moulting kookaburras (Dacelo novaeguineae):
they’re laughing. J Comp Physiol B 173:223–230
Carpenter FL (1974) Torpor in an Andean hummingbird: its ecological
signiWcance. Science 183:545–547
Chaplin SB (1976) The physiology of hypothermia in the black-capped
chickadee, Parus atricapillus. J Comp Physiol 112:335–344
Clark CW, Dukas R (2000) Winter survival strategies for small birds:
managing energy expenditure through hypothermia. Evol Ecol
Res 2:473–491
Cooper SJ, Gessaman JA (2005) Nocturnal hypothermia in seasonally
acclimatized mountain chickadees and juniper titmice. Condor
107:151–155
Cooper CE, Körtner G, Brigham M, Geiser F (2008) Body temperature
and activity patterns of free-living laughing kookaburras: the
largest kingWsher is heterothermic. Condor 110:110–115
Cyr NE, Romero M (2008) Fecal glucocorticoid metabolites of
experimentally stressed captive and free-living starlings: impli-
cations for conservation research. Gen Comp Endocrinol
158:20–28
Cyr NE, Earle K, Tam C, Romero LM (2007) The eVect of chronic
physiological stress on corticosterone, plasma metabolites and
immune responsiveness in European starlings. Gen Comp Endo-
crinol 154:59–66
Dickens MJ, Earle KA, Romero LM (2009) Initial transference of wild
birds to captivity alters stress physiology. Gen Comp Endocrinol
160:76–83
Dolby AS, Temple JG, Williams LE, Dilger EK, Stechler KM, Davis
VS (2004) Facultative rest-phase hypothermia in free-ranging
white-throated sparrows. Condor 106:386–390
Downs CT, Brown M (2002) Nocturnal heterothermy and torpor in the
malachite sunbird (Nectarinia famosa). Auk 119:251–260
Fletcher QE, Fisher RJ, Willis CKR, Brigham RM (2004) Free-ranging
common nighthawks use torpor. J Therm Biol 29:9–14
Geiser F, Holloway JC, Körtner G, Maddocks TA, Turbill C, Brigham
RM (2000) Do patterns of torpor diVer between free-ranging and
captive mammals and birds? In: Heldmaier G, Klingenspor M
(eds) Life in the cold: eleventh international hibernation sympo-
sium. Springer, Berlin, pp 95–102
Grubb TC, Pravosudov VV (1994) Toward a general theory of energy
management in wintering birds. J Avian Biol 25:255–260
Haftorn S (1972) Hypothermia of tits in the arctic winter. Ornis Scand
3:153–166
Haftorn S (1992) The diurnal body-weight cycle in titmice Parus spp.
Ornis Scand 23:435–443
Körtner G, Brigham RM, Geiser F (2000) Winter torpor in a large bird.
Nature 407:318
Körtner G, Brigham RM, Geiser F (2001) Torpor in free-ranging
tawny frogmouths (Podargus strigoides). Physiol Biochem Zool
74:789–797
Lane JE, Brigham RM, Swanson DL (2004) Daily torpor in free-rang-
ing whip-poorwills (Caprimulgus vociferus). Physiol Biochem
Zool 77:297–304
Laurila M, Hohtola E (2005) The eVect of ambient temperature and
simulated predation risk on fasting-induced nocturnal hypother-
mia of pigeons in outdoor conditions. J Therm Biol 30:392–399
Lessells CM, Boag PT (1987) Unrepeatable repeatabilities—a com-
mon mistake. Auk 104:116–121
Littell RC, Milliken GA, Stroup WW, WolWnger RD, Schabenberger
O (2006) SAS for mixed models. SAS Institute Inc., Cary
Maddocks TA (2001) The thermal physiology and energetics of
Australian birds. Ph.D. thesis, University of New England
Maddocks TA, Geiser F (1997) Energetics, thermoregulation and
nocturnal hypothermia in Australian silvereyes. Condor 99:104–
112
Maddocks TA, Geiser F (2007) Heterothermy in an Australian passer-
ine, the dusky woodswallow (Artamus cyanopterus). J Ornithol
148:571–577
Mayer L, Lustick S, Battersby B (1982) The importance of cavity
roosting and hypothermia to the energy-balance of the winter
acclimatized Carolina chickadee. Int J Biometeorol 26:231–238
McKechnie AE, Lovegrove BG (2001a) Heterothermic responses in
the speckled mousebird (Colius striatus). J Comp Physiol B
171:507–518
McKechnie AE, Lovegrove BG (2001b) Thermoregulation and the
energetic signiWcance of clustering behavior in the white-backed
mousebird (Colius colius). Physiol Biochem Zool 74:238–249
McKechnie AE, Lovegrove BG (2002) Avian facultative hypothermic
responses: a review. Condor 104:705–724
McKechnie AE, Lovegrove BG (2003) Facultative hypothermic
responses in an Afrotropical arid-zone passerine, the red-headed
Wnch (Amadina erythrocephala). J Comp Physiol B 173:339–
346
McKechnie AE, Körtner G, Lovegrove BG (2004) Rest-phase thermo-
regulation in free-ranging white-backed mousebirds. Condor
106:143–149

J Comp Physiol B (2009) 179:737–745 745
123
McKechnie AE, Ashdown RAM, Christian MB, Brigham RM (2007)
Torpor in an African caprimulgid, the freckled nightjar Caprimul-
gus tristigma. J Avian Biol 38:261–266
Merola-Zwartjes M (1998) Metabolic rate, temperature regulation, and
the energetic implications of roost nests in the bananaquit (Coer-
eba Xaveola). Auk 115:780–786
Merola-Zwartjes M, Ligon JD (2000) Ecological energetics of the
Puerto Rican tody: heterothermy, torpor, and intra-island varia-
tion. Ecology 81:990–1003
Mzilikazi N, Lovegrove BG (2002) Reproductive activity inXuences
thermoregulation and torpor in pouched mice, Saccostomus cam-
pestris. J Comp Physiol B 172:7–16
Pettersson J, Hasselquist D (1985) Fat deposition and migration capac-
ity of robins Erithacus rebecula and goldcrests Regulus regulus at
Ottenby, Sweden. Ring Migr 6:66–76
Pravosudov VV, Grubb TC (1997) Energy management in passerine
birds during the nonbreeding season. A review. Curr Ornithol
14:189–234
Pravosudov VV, Lucas JR (2000) The costs of being cool: a dynamic
model of nocturnal hypothermia by small food-caching birds in
winter. J Avian Biol 31:463–472
Prinzinger R, Pressmar A, Schleucher E (1991) Body temperature in
birds. Comp Biochem Physiol 99:499–506
Rashotte ME, Pastukhov IF, Poliakov EL, Henderson RP (1998)
Vigilance states and body temperature during the circadian cycle
in fed and fasted pigeons (Columba livia). Am J Physiol Regul
Integr Comp Physiol 275:R1690–R1702
Reinertsen RE (1996) Physiological and ecological aspects of
hypothermia. In: Carey C (ed) Avian energetics and nutritional
ecology. Chapman & Hall, New York, pp 125–157
Reinertsen RE, Haftorn S (1983) Nocturnal hypothermia and metabo-
lism in the willow tit Parus montanus at 63 degrees N. J Comp
Physiol 151:109–118
Reinertsen RE, Haftorn S (1984) The eVect of short-time fasting on
metabolism and nocturnal hypothermia in the willow tit Parus
montanus. J Comp Physiol 154:23–28
Reinertsen RE, Haftorn S (1986) DiVerent metabolic strategies of
northern birds for nocturnal survival. J Comp Physiol B 156:655–
663
Romero LM, Romero RC (2002) Corticosterone responses in wild
birds: the importance of rapid initial sampling. Condor 104:129–
135
Ruby NF, Nelson RJ, Licht P, Zucker I (1993) Prolactin and testoster-
one inhibit torpor in Siberian hamsters. Am J Physiol 264:R123–
R128
Schleucher E (1999) Energetics and body temperature regulation in
two convergent dove species from extreme habitats. Ornis Fenn
76:199–210
Schleucher E (2001) Heterothermia in pigeons and doves reduces
energetic costs. J Therm Biol 26:287–293
Schleucher E (2004) Torpor in birds: taxonomy, energetics, and
ecology. Physiol Biochem Zool 77:942–949
Schoech SJ, Ketterson ED, Nolan V (1999) Exogenous testosterone
and the adrenocortical response in dark-eyed juncos. Auk
116:64–72
Seber GAF, Lee AJ (2003) Linear regression analysis. Wiley-Inter-
science, New York
Sharbaugh SM (2001) Seasonal acclimatization to extreme climatic
conditions by black-capped chickadees (Poecile atricapilla) in
interior Alaska (64 degrees N). Physiol Biochem Zool 74:568–
575
Sokal RR, Rohlf FJ (1995) Biometry. W.H. Freeman, New York
Spencer KA, Verhulst S (2008) Post-natal exposure to corticosterone
aVects standard metabolic rate in the zebra Wnch (Taeniopygia
guttata). Gen Comp Endocrinol 159:250–256
Waite TA (1991) Nocturnal hypothermia in gray jays Perisoreus
canadensis wintering in interior Alaska. Ornis Scand 22:107–110
Welton NJ, Houston AI, Ekman J, McNamara JM (2002) A dynamic
model of hypothermia as an adaptive response by small birds to
winter conditions. Acta Biotheor 50:39–56
WingWeld JC, Farner DS (1975) Determination of 5 steroids in avian
plasma by radioimmunoassay and competitive protein binding.
Steroids 26:311–327
Woods CP, Brigham RM (2004) The avian enigma–“hibernation” by
common poorwills (Phalaenoptilus nuttalli). In: Barnes BM,
Carey CM (eds) Life in the cold: evolution, mechanisms, adapta-
tion, and application, Institute of Arctic Biology. University of
Alaska, Fairbanks, pp 231–240