Metabolic response to temperature variation in the great tit: an interpopulation comparison


Juli Broggi, Department of Biology, University of Oulu, PO Box 3000, FIN-90014 Oulu, Finland.Tel: +358 85531267; Fax: +358 85531227; E-mail:


  • 1We studied the resting metabolic rate (MR) from two great tit Parus major (Linnaeus) populations living in different winter regimes. Birds from the two different localities were exposed individually to +25 °C, 0 °C and −10 °C for the night in three consecutive sessions in random order.
  • 2Birds from Lund (Sweden) had a lower basal MR, as measured at thermoneutrality (+25 °C), than had birds from Oulu (Finland). Nevertheless, below thermoneutrality, birds from Oulu spent relatively more energy, especially at −10 °C.
  • 3Although the energy needed for thermoregulation decreased with increasing basal MR this relation is at a higher metabolic cost for birds in Oulu than for birds in Lund.
  • 4The higher basal MR in Oulu is probably a consequence of a higher maximal MR needed in the severe cold. Further, the observed MRs below thermoneutrality are lower than expected from published data. This suggests that all birds were probably hypothermic at −10 °C, particularly Lund birds, and that the use of controlled hypothermia in great tits may be more common than thought previously. Great tits seem to rely primarily on metabolic adjustment to cope with the harsh climatic conditions in the northernmost parts of its distribution.


Small birds that are resident in the temperate zone or at Arctic latitudes are faced with strong seasonal changes in cold exposure and thermostatic costs. At the time when energetic requirements increase, available food decreases together with time available to acquire it. Such factors combine to make winter an energetically stressful period for resident small birds.

Several adaptations help these birds to survive the non-breeding season. By a process of winter acclimatization, which is primarily a metabolic improvement in thermogenic capacity and endurance, resident birds have an enhanced cold resistance in winter compared to that in summer (Swanson 2003). Furthermore, as energy demands increase dramatically, birds typically have a capacity for increased body reserves and digestive efficiency (Rogers, Nolan & Ketterson 1993; Geluso & Hayes 1999).

Plumage insulation also changes seasonally by the development of new feathers during moult, and the quality and density of feathers may increase as a result of the acclimatization processes (Middleton 1986; Root, O’Connor & Dawson 1991; Swanson 1991; Novoa, Bozinovic & Rosenmann 1994; Cooper 2002). Furthermore, insulation can be modulated to some extent with changing conditions, either decreasing as a result of feather deterioration (Root et al. 1991) or increasing by means of plumage ptiloerection (Hohtola, Rintamäki & Hissa 1980).

In addition, birds may develop several energy-saving strategies such as nocturnal hypothermia (Reinertsen 1983).

Several studies have shown a proximate role of winter temperature in regulating metabolism (see Swanson 2003 for review). In free-living as well as in laboratory-acclimated small birds, mass-specific basal metabolic rate (MR) is normally higher in winter than in summer (Swanson 1990; Cooper & Swanson 1994; Saarela, Klapper & Heldmaier 1995; Liknes & Swanson 1996; Liknes, Scott & Swanson 2002). Further, widespread bird populations wintering in temperate climates show a negative relation between basal MR and temperature (Dawson et al. 1983; Cuthill et al. 2000). However, the precise nature of the association between variation in MR and variation in cold tolerance in birds remains obscure (see Swanson 2003 for review).

Winter ranges of many species are limited by thermoregulatory requirements (Root 1988). Further, some species with wide distribution ranges may experience extremely different winter conditions ranging from mild winters to extremely cold ones (Hoffmann & Blows 1994). In some cases, local adaptations may arise, while phenotypic flexibility usually accounts for most of the adjustments to prevailing local conditions (Ricklefs & Wikelski 2002).

The great tit Parus major (Linnaeus) is a newcomer in northern Europe, and evidence from the breeding period suggests that it may be maladapted to the boreal regions (Rytkönen & Orell 2001). This situation may be maintained by gene flow from southern populations which would prevent local adaptations to conditions at northern latitudes (Kvist et al. 1999). Whether this argument applies to winter survival strategies awaits further research. If seasonal changes in temperature or daylength are a major factor driving seasonal adjustments of physiology, then species wintering in cold climates should have an increased seasonal physiological adjustment with respect to their counterparts from milder climates.

We studied resting MR during the non-breeding season in great tits from two locations differing in winter conditions. By comparing populations living in different regimes of winter-severity, we aimed to elucidate which metabolic adaptations could explain differences in cold acclimatization. We expect birds from the northernmost location to exhibit overall higher metabolic capacity to deal with harsher conditions, and to show some adjustment in order to make this strategy less expensive in energetic terms.

Materials and methods

We studied wild individual great tits during the nonbreeding season from January until March 2001. Birds were captured in two different locations in Lund (Sweden) and Oulu (Finland) and their MR measured during 3 consecutive nights at three different temperatures.

In Lund study area (55°40′ N, 13°25′ E), winter daily average temperatures ranged from −3 °C to 7 °C during the study period that lasted from the end of January to mid-March (median 23 February). During that period, night-length decreased from 15 : 30 to 12 : 10 h and snow was present for about 2 weeks. In the Oulu study area (65° N, 25°30′ E) winter daily average temperatures ranged from −14 °C to 0 °C during the study period that lasted until the end of March (median 17 March). During that period permanent snow covered the study area and the night-length decreased from 13 : 30 to 10 : 40 h.

Great tits in Lund study area live year-round in mixed deciduous forests, fragmented by agricultural landscapes and do not rely on feeders for survival. In contrast, great tits in northern Finland breed in mixed deciduous–coniferous forest, and winter close to human settlements. During winter, they feed on human-provided food on which they are probably highly dependent for their winter survival (Orell 1989).

trapping and maintenance

Birds in Lund were trapped soon after dusk while roosting in nestboxes, and brought indoors for measuring during the whole night. Each bird was measured on 3 consecutive nights and kept alone in outdoor aviaries between measurement nights. The outdoor aviaries consisted of 12·8 m2 surface and 2·2 m high cages with several nestboxes available for roosting, and food was a mixture of peanuts, sunflower seeds and animal fat provided ad libitum. At the start of the different temperature treatments (see below), mass of individual birds did not differ (repeated measures anova, F2,54 = 0·07; P = 0·9)

Birds in Oulu were captured by means of funnel traps that were installed permanently in the study area and worked as feeders except when trapping (see Carrascal et al. 1998 for the same procedure). Birds were caught shortly before dusk and kept in outdoor aviaries between measurements as in Lund. All birds where released after the experiments.

metabolic measurements

Resting MR was measured as the average minimal oxygen consumption under post-absorptive digestive conditions during the resting phase of the daily cycle on resting, non-growing, non-reproductive animals. Basal MR was considered to be the resting MR at thermoneutrality (25 °C) (McNab 1997). The energetic cost of thermoregulation (ECT) was measured as the difference between resting MR and basal MR and represents the additional MR necessary for thermoregulation.

Resting MR was measured in terms of oxygen consumption during the night in open-circuit respirometers in both locations. Each bird was placed after dusk in an individual sealed metabolic chamber (1·6 L) and placed in the darkness of a climate cabinet at three different temperatures (25 °C, 0 °C and −10 °C) on 3 consecutive nights. Some of the birds escaped from the aviaries during the measuring period, which explains the varying sample sizes in different treatments.

The respirometer in Lund consisted of a four-channel set with a flow of 200 mL min−1, and is described in Lindström, Klassen & Kvist (1999) and Nilsson & Råberg (2001). The Oulu respirometer consisted of a two-channel set and one oxygen analyser Servomex 1440 (UK) that received air samples of 600 mL min−1 through a valve system. Dried outdoor air was pumped to both metabolic chambers through mass-flow controllers (Bronkhorst Hi-Tec F201C, the Netherlands) and then dried again before analysis. The valve system switched in periods of 30 min between channels and outdoor air. Readings were recorded every minute and later on, minimum night averages were extracted over 3-h periods between 23·00 and 04·00 h for every bird. The closest outdoor-air reading was used as reference in order to control for any possible analyser drift.

In Lund CO2 was measured, whereas in Oulu it was removed from inlet and outlet air, as CO2 was not measured. Appropriate equations for each of these conditions were used according to Hill (1972).

The different modes of capturing the birds did not bias our measurements of MR. In a sample of great tits from Lund in 2000, birds captured either at feeders or from nestboxes, did not differ in basal MR (t-test: t27 = 0·15; P = 0·88).

Birds were measured at one constant temperature each night and the order of temperature treatments was randomized, to control for possible treatment or captivity effects.

All variables full filled the requirements of normality (tested with the Kolmogorov–Smirnov one sample test) and thus parametric statistics were used in all analysis.


The mean mass of great tits, as measured at the first evening after capture, did not differ between the two areas (mean ± SD, Lund: 18·8 g ± 0·74; N = 24, Oulu: 18·5 g ± 1·48; N = 17, t-test: t = 0·87; P = 0·39). The only factor remaining in a multiple regression with area, sex, age, date of capture and ambient mean temperature on the day of capture was sex, males being significantly heavier than females (t = 3·98; P < 0·001).

The MR of birds at 25 °C, i.e. basal MR, differed between the two areas (t-test: t37 = 3·34; P = 0·002), Oulu birds having a higher basal MR than birds from Lund (Fig. 1). In a multiple regression with basal MR as the dependent variable, neither sex, age, date of capture nor ambient mean temperature the day before measurement explained any significant proportion of the variation in basal MR. The only significant factors left in the model was area (t = 4·34; N = 39; P < 0·001) and mass (t = 2·90; N = 39; P = 0·006), which together explained 37·7% of the variation in basal MR.

Figure 1.

Mean (+ SE) metabolic rate (ml O2 min−1) of great tits spending the night in 25 °C (thermoneutrality), 0 °C and −10 °C. Black bars represent birds from Lund and open bars birds from Oulu.

Because the MR of an individual was measured at three different temperatures, we analysed the differences between birds from the two areas with a repeated-measures anova. Of the predictor variables (area, age, sex and date), only area explained a significant part of the between subjects’ variation (F1,26 = 19·7; P < 0·001). Within subjects, MR increased significantly as the night temperature decreased (F = 171·0; P < 0·001; Fig. 1). The interaction between area and night temperature was not significant (F = 0·97; P = 0·39). The same analysis but with morning mass as the dependent variable, resulted in sex (F = 16·8; P < 0·001) and to some extent area (F = 3·26; P = 0·084) being able to explain some of the variation between subjects. The morning mass also decreased with decreasing night temperature (F = 4·98; P = 0·011).

During our measurements of energy consumption at temperatures below thermoneutrality, we assumed that the total energy budget consists only of basal MR and the cost of thermoregulation (ECT, see Methods). Thus, to obtain a measure of the metabolic cost of thermoregulation we subtracted basal MR from the total energy expenditure. We found no difference between the two areas in this cost of thermoregulation at 0 °C (t-test: t27 = 0·38; P = 0·7), nor could sex, age or date explain any of the variation in a multiple regression (P > 0·5). However, at −10 °C, birds from Oulu tended to expend more energy on thermoregulation than did birds from Lund (t-test: t30 = 1·82; P = 0·079). Area was the only factor remaining in a multiple regression also including sex, age and date (sex, age and date; P > 0·4).

One factor that potentially could affect the metabolic cost of thermoregulation is the level of basal MR. A multiple regression with date, area, sex, age and basal MR still did not explain any of the variation in the cost of thermoregulation at 0 °C (N = 29; P > 0·3 in all cases). However, at −10 °C, both area (F = 11·5; N = 32; P = 0·002) and basal MR (F = 10·1; N = 32; P = 0·004) explained a significant part (33·2%) of the variation in the metabolic cost of thermoregulation (the interaction between area and basal MR being non-significant; P = 0·55). Thus, the cost of thermoregulation decreased with increasing basal MR in both areas (Fig. 2), but this relation is at a higher overall metabolic cost of thermoregulation in the birds from Oulu compared to birds from Lund (Fig. 2).

Figure 2.

Relation between the metabolic cost of thermoregulation (ml O2 min−1) at −10 °C and the basal metabolic rate (ml O2 min−1) as measured at 25 °C. Filled circles and bold line represent birds from Lund and open circles and broken line birds from Oulu. The relation was tested separately for Lund and Oulu with regression analyses, Lund: t = −3·62; N = 16; P = 0·003; equation of the line: y = 2·18 − 1·44x; R2 = 0·48, Oulu: t = −2·03; N = 16; P = 0·062; equation of the line: y = 1·97 − 0·90x; R2 = 0·29.


Birds in Oulu had a higher basal MR but additionally an overall higher MR at all temperatures. Birds in the northernmost population have to deal with harsher conditions (lower temperatures and shorter day lengths) and higher weather unpredictability, and they presumably do this by increasing their thermogenic capacity and endurance. Such an increase in energy expenditure is probably achieved by increasing energy acquisition and digestive efficiency, i.e. increased size of the digestive tract (Piersma & Lindström 1997), in order to support increases in maximal MR, which would in turn elevate their basal MR (see Swanson 2003 for review). The decrease in the energetic cost of thermoregulation (ECT) with increasing basal MR probably depends in part on the usage of heat, generated by the metabolism, for thermoregulation and in part because basal MR increases faster than energy available for work with an increase in total MR (Nilsson 2002).

The increased costs of thermoregulation in the northernmost population raise an interesting question. We expected birds in Oulu to show a decreased ECT in order to cope with harsher conditions without incurring exceptionally high energetic costs. However, in contrast to our expectations, ECT was higher at lower temperatures in birds from the northernmost population.

Heat loss can be expressed in terms of insulation and body-environment thermal gradient, both governing the heat loss from the body below the lower critical temperature. Thus, the fact that birds from the southernmost population had a lower ECT can ultimately be explained in two ways. At a constant ambient temperature, variation in ECT could either be due to differences in insulation, in body temperature, or both. Thus, birds from Lund had either a higher insulation capacity or a smaller temperature gradient between their body and the ambient air. In the first case, better insulation could be achieved by higher plumage quality per se or by fresher plumage. Birds from southern latitudes could afford delaying post-nuptial moult, and may also enjoy better foraging conditions resulting in a higher plumage quality at the peak of the winter. On the other hand, considerable differences should be invoked in order to explain the variation found. Furthermore, if insulation differences are to be expected they would rather be in the other direction, as shown in other studies with winter-acclimatized birds (Middleton 1986; Root et al. 1991; Swanson 1993; Novoa et al. 1994; Cooper 2002).

As no body temperature measurements were obtained conclusions on the degree of hypothermia can be drawn, always with caution, only after certain assumptions are made. Considering that measurements were made far below the thermoneutral zone, a constant insulation value that would be maximal before starting thermogenesis could be assumed. In such conditions, the expected MR would be much higher than the ones we obtained, as calculated from allometric equations and previous data on the same species (Hissa & Palokangas 1970; Peters 1983). Thus, both populations were probably hypothermic at −10 °C, birds from Lund probably being in deeper controlled hypothermia than their counterparts from Oulu, as this temperature is close to the minimum ambient temperature they may encounter.

Controlled hypothermia appears to be a last resort to endure cold temperatures for many passerines and probably is connected to important costs (Reinertsen 1983; Grubb. & Pravosudov 1994). As the capacity for thermogenesis also has costs, e.g. an increased basal MR, a trade-off between this capacity and the use of hypothermia may be anticipated. The optimal combination of the two strategies may depend on average environmental conditions resulting in using controlled hypothermia at the lower end of the local temperature variation. This lower end of the temperature variation would be approximately −10 °C in Lund (altogether, 2 nights had this or a lower minimum temperature during the winter of 2001) but at much colder temperatures in Oulu (62 nights with minimum temperatures at −10 °C or below). Furthermore, a higher basal MR that would allow a high maximal MR may reduce the options for a decrease in body temperature. Given the small bird's high capacity for regulatory thermogenesis in general, the need for a higher basal MR appears unnecessary (Dawson & O’Connor 1996). It is unclear whether the higher resting MR is a contributing factor to these improvements in cold tolerance, a by-product of them or a separate response.

In general, differences between the studied populations appear to be based on metabolic adjustments, which could be interpreted as a first step in the acclimatization process to a new environment (Dawson et al. 1983; Swanson 1993; Ricklefs & Wikelski 2002). Further, birds from both populations appear reluctant to become hypothermic, even more so in the northernmost population where individuals keep a high MR whenever possible. Contrary to our expectations, birds from the northernmost population used a more energetically expensive strategy than their southern counterparts, which does not seem sustainable as winter food predictability diminishes with increasing latitude. In general, birds from Oulu may experience higher food predictability than birds from Lund, as they rely on human-provided food during winter, suggesting the possibility that the use of hypothermia may be more costly than keeping a high MR with the concurrent high ECT.

These results provide a framework in which to study the possible origins of interpopulation differences, and to understand the patterns of colonization of new areas or distribution changes due to climatic variation.


We thank P. Karkkäinen and the staff from Oulu University Zoo for their help in the aviaries, K. Andersson for capturing and looking after the birds in Lund. Comments by K. Lahti and two anonymous referees improved earlier drafts of this manuscript. The study was supported by grants to Markku Orell by the Research Council for Biosciences and Environment of the Academy of Finland (grants 3548, 51858, 47195) and the Thule Institute of the University of Oulu.