2.2. Thermogenic mechanisms

2.2.1. Shivering thermogenesis

Shivering thermogenesis is regarded as “an increase in the rate of heat production during cold exposure due to increased contractile activity of skeletal muscles not involving voluntary movements and external work” (IUPS Thermal Commission 2001). Shivering is initiated and maintained by the neuronal system since there is no evidence of enhancement or maintenance of shivering by any blood-borne humoral factors (Mercer & Hammel 1993). Shivering is initiated by the same α-motoneurons that act in voluntary muscle contractions and all shivering thermogenesis, as voluntary contractions, can be blocked by curare poison. The difference exists in the motor control, which in voluntary muscle contractions comes from the motor nuclei of the CNS, while in shivering the motor commands originate in the thermosensitive and integrating part of the CNS. Incipient shivering progresses from thermoregulatory muscle tone to micro-vibrations and eventually to clonic contractions or tremor of both flexor and extensor muscles. True shivering occurs only in mammals and birds (Heath 1968, Kleinebeckel & Klussmann 1990, Ruben 1995). At least in warm-acclimated individuals, it is the only means of regulatory heat production. The term “shivering” has also been used to describe the mechanism some insects use to elevate their thoracic temperature during pre-flight warm-up (e.g. Esch & Goller 1991). As in vertebrate shivering, this mechanism involves the use of muscle contractions for heat production without external work, but it is under the command of a completely differently organized nervous system. In addition, its biochemical basis for heat production may include separate metabolic routes.

The initiation of shivering requires cold stimuli, which may be effective at various sites of the body. For example, cooling of the skin, internal organs, hypothalamus, midbrain brain stem or spinal cord induce shivering in mammals (e.g. Cabanac 1975, Simon et al. 1986). In birds, cooling of the hypothalamus does not stimulate shivering thermogenesis but in contrast, it may lead to the inhibition of shivering (Rautenberg et al. 1972, Snapp et al. 1977, Simon-Oppermann et al. 1978, Mercer & Simon 1984). Extrahypothalamic centers of the brain and spinal cord (Rautenberg et al. 1972, Inomoto and Simon 1981, Martin et al. 1981, Østnes & Bech 1992) and thermosensors in the peripheria, like skin cold-receptors, have a dominant role in eliciting shivering (Necker 1977, Simon et al. 1986, Østnes & Bech 1998).

The energetic efficiency of skeletal muscle work ranges between 20–25% (Wilkie 1960, Prompero et al. 1969, Gibbs & Gibson 1972, Tappy & Guenat 2000). Thus muscles liberate substantial amounts of heat as a by-product of the coupling of chemical energy into mechanical work. During shivering, virtually all the chemical energy of fuels consumed is transformed into heat. Thermogenesis in muscle is initiated by α-motoneuronal stimulation following depolarization of muscle cell membrane and the release of Ca²⁺ from intracellular stores of the sarcoplasmic reticulum. Ca²⁺ results in activation of myosin ATPase and myofibrilar cross-bridge cycling. The increase of ADP in turn accelerates the mitochondrial oxidative phosphorylation increasing the combustion of fuels.

Ion pumping forms another process of heat production during muscle activity. The process involves restoration of the normal polarized state of the sarcolemma and removal of Ca²⁺ from the cytosol. The former is done by Na⁺-K⁺ ATPase and the latter principally via Ca²⁺ ATPase activity of the sarcoplasmic reticulum. Both processes in turn accelerate mitochondrial respiration by altering the phosphorylation state ratio. Some of the free Ca²⁺ in the cytosol is actively taken up by the mitochondria. In the mitochondria, Ca²⁺ stimulates respiration to enhance the rate of ATP production to match demand (McCormack et al. 1990). During prolonged heat production in muscle cells, ~25% of heat released is related to splitting of ATP into ADP and inorganic phosphate residue, and ~75% to the regeneration of ATP in mitochondrial oxidative phosphorylation (Hochachka 1974). Thermogenic reactions in muscle are reviewed in detail by Hochachka (1974), Himms-Hagen (1976), Homsher and Kean (1978), Woledge et al. (1985) and Block (1994).

Shivering closely resembles the normal isometric muscle contraction which occurs in postural tone (Hohtola 1981). Postural tone in mammals and birds is enabled by oxidative twitch-type muscle fibres, while in lower vertebrates it is maintained by slow and graded tonic fibres. The co-existence of sustained aerobic metabolism both in postural tone and shivering thermogenesis suggests that the origin of shivering arises from postural activity. A characteristic of shivering is an asynchronous firing pattern of the motor unit action potentials which prevents gross tremors and thereby possibly decreases convective heat loss (Hohtola and Stevens 1986). Tremors are not a prerequisite for thermogenesis although they emerge when the intensity of shivering increases in parallel with the augmented heat production.

Shivering can be quantified most accurately and easily by measuring the electrical events of a muscle (electromyogram, EMG). Electric currents (action potential, AP) in the muscle fibre are generated when the neurotransmitter acetylcholine is released from motoneuronal synapses. The AP moves along the muscle fibre at a speed of 2–5 m·s^-1 and every part of the muscle fibre several centimeters long will meet the AP within a few milliseconds (Loeb & Gans 1986). The functional unit of the muscle, the motor unit (MU), has a common α-motoneuron which innervates all the muscle fibres of the MU simultaneously. The electrical signal emanating from the activation of the muscle fibres of a single MU is called a motor unit action potential, MUAP (DeLuca 1988). In a single MU, the MUAPs are generated repeatedly, the resulting sequence of MUAPs being termed a motor unit action potential train, MUAPT. The EMG, which is measurable with electrodes, is a result of the summation of MUAPTs in numerous spatially overlapping motor units.

West (1965) showed a linear relationship between the average peak-to-peak EMG and heat production in four species of birds. The use of various other EMG parameters in predicting heat production were systematically studied by Hohtola (1982) with pigeons. The most reliable parameter was obtained by rectifying and averaging the EMG signal, this mean rectified value is equivalent to the mean deviation or mean amplitude around the mean voltage. Another parameter, the root mean square voltage, equivalent to the standard deviation around the mean, was found to be almost as reliable. The correlation between the intensity of the EMG and heat production is not constant in a wide range of ambient temperatures (Hohtola 1982). At low temperatures, where metabolic rates are high, the correlation between the intensity of the EMG and heat production becomes weaker. Since the EMG electrode has a limited field in sensing myoelectric currents, the saturation of this field may be related to the decreasing correlation.

2.2.2. Nonshivering thermogenesis

Nonshivering thermogenesis (NST) comprises metabolic heat production by processes that do not involve skeletal muscle contractions (IUPS Thermal Commission 2001). NST is composed of both obligatory and thermoregulatory (facultative) components. Obligatory NST is independent of short-term changes in ambient temperature and it corresponds to basal metabolic rate. Thermoregulatory NST occurs below the thermo-neutral zone in response to acute cold exposure. The term “nonshivering thermogenesis” is conventionally used to refer thermoregulatory NST and the same practice is adopted in this work.

A simple experiment which is used in mammals to reveal NST is the measurement of oxygen consumption and body temperature in the thermoneutral zone before and after noradrenaline injection. A parallel increase in oxygen consumption and body temperature in response to the injection indicates the existence of NST. In birds, this method cannot be successfully used since noradrenaline generally does not yield a calorigenic response (see Hissa 1988). Another way to reveal NST is to measure oxygen consumption and shivering intensity from muscles simultaneously and to compare the lower critical temperature and shivering threshold temperature (STT). Occurrence of STT lower than LCT is considered as evidence of NST.

In placental mammals, particularly in small-sized species and in neonatal and hibernating animals, the principal effector organ of NST is brown adipose tissue, BAT (Smith & Horwitz 1969, Jansk 1973, Hayward & Lisson 1992). In cold-acclimated or cold-acclimatized individuals, NST in BAT replaces shivering thermogenesis and takes its role of “first line heat productive mechanism”. In mammals, BAT is located primarily in the interscapular region overlying the cervical spinal cord and in smaller quantities in the thoracic, periaortic and perirenal positions (Smith 1964). Unlike white adipose tissue, BAT is characterized by a rich vascularization, which is responsible for its colour, a network of sympathetic fibers around every cell, multilocular adipocytes, and numerous mitochondria with dense cristae (Daniel & Derry 1969, Cannon & Nedergaard 1985). In response to acute cold exposure, the sympathetic nerve-endings of BAT release noradrenaline (NA). This NA binds to membrane β -adrenergic receptors, so stimulating lipolysis and the release of free fatty acids (FFA) (Nicholls & Locke 1984). FFAs released in situ are used as substrates for the respiratory chain and they apparently also activate thermogenesis in BAT. The mitochondrial uncoupling protein (UCP₁, also called thermogenin) makes the mitochondrial inner membrane permeable to protons, so uncoupling the respiratory chain and oxidative phosphorylation. The activated uncoupling protein acts as a proton leak channel and when it short-circuits oxidative phosphorylation, the energy normally stored in the ATP is liberated as heat. Thermogenesis in BAT is also slightly stimulated via α-adrenergic receptors (Mohell et al. 1987). In rodents, BAT thermogenesis contributes to the control of body weight, acting as the regulatory (facultative) part of diet-induced thermogenesis (Rothwell & Stock 1979, Himms-Hagen 1990). Similarly to acute cold exposure, overeating also activates BAT thermogenesis. The role of diet-induced thermogenesis in BAT is the conversion of excess energy into heat in order to prevent weight gain. BAT does not exist in marsupial and monotreme mammals (Hayward & Lisson 1992, Nicol et al. 1997, Rose et al. 1999) or in birds (Johnston 1971, Olson et al. 1988, Saarela et al. 1989, 1991).

Skeletal muscle has been proposed for the site of NST both in mammals and birds. However, the uncoupling of mitochondrial oxidative phosphorylation by UCP₁ has not been found outside of mammalian BAT. In birds, for example, UCP₁ is not even expressed in the skeletal muscles of those birds which have been suggested as candidates for possessing muscular NST (Denjean et al. 1999). Since 1997, new uncoupling protein homologs (UCP₂, UCP₃, UCP₄, UCP₅, avUCP, HmUCP) are found in various animal tissues including muscles and even in plants (Boss et al. 1998, Mao et al. 1999, Yu et al. 2000, Raimbault et al. 2001, Vianna et al. 2001). The two latter studies reported the discovery of avian UCPs that show 55% homology with UCP₁. On the one hand, Raimbault et al. (2001) discovered avian-type UCP (avUCP) using cDNA libraries of chicken skeletal muscle. This avUCP is expressed exclusively in skeletal muscle. A high level of expression of avUCP mRNA was found in cold-acclimated and glucagon treated ducklings and in chickens with a high level of diet-induced thermogenesis. On the other hand, Vianna et al. (2001) reported the discovery of an uncoupling protein (HmUCP) from the swallow-tailed humminbird (Eupetomena macroura). The HmUCP mRNA is primarily expressed in skeletal muscle, but in addition in the heart and liver. Whether UCP₁ is the only true uncoupling protein is unsolved as yet (cf. Nedergaard et al. 1999). The uncoupling activity of other UCP homologs, including avUCP and HmUCP, have not yet been demonstrated. The recent finding that UCP₁-ablated mice do not develop any NST but exhibit shivering when cold acclimated (Golozoubova et al. 2001) indicates that probably only UCP₁ can mediate adaptive NST in the cold.

The existence of NST in birds has been studied intensively during the last few decades. El-Halawani et al. (1970) studied the effects of cold acclimation on oxygen consumption and shivering in the gastrocnemius muscle of domestic chickens. Chickens were not influenced by two-month cold acclimation but after 5 to 9 months shivering was greatly reduced while oxygen consumption was increased. This was regarded as evidence of NST. However, the responses of shivering intensity and oxygen consumption to acute cold-exposure were not studied and thus there was no evidence that regulatory NST really existed. Moreover, the ecological significance of NST achieved by such a long acclimation period can be questioned (Calder and King 1974). The most impressive array of evidence on muscular NST comes from cold-acclimated Muscovy ducklings (Cairina moschata) and king penguin chicks (Aptenodytes patagonicus) (for a review, see Duchamp et al. 1999). In these species, significant differences have been observed between the shivering threshold temperature and lower critical temperature (Barré et al. 1985, Barré et al. 1986a, Duchamp et al. 1989). In these studies, however, shivering was reported merely from gastrocnemius muscle, though unpublished recordings from pectoral muscle were briefly mentioned as showing a similar difference (Barré et al. 1985, Duchamp et al. 1989). Because cold acclimation may result in a shift of shivering to other muscles, the measurement of shivering merely from one muscle group is not extensive enough for revealing the existence of NST. In cold-acclimated Muscovy ducklings, Vittoria and Marsh (1996) also found that shivering in the gastrocnemius muscle was absent during cold exposure, but that in thigh muscles (m. iliofibularis and flexor cruris) muscle activity increased in parallel with the augmentation of the metabolic rate. Another possibility for the reason behind the disappearance of shivering, though only a hypothetical one since the matter has not been studied, is a shift of shivering to deeper regions of the same muscle.

It has been suggested that the endocrine control of avian muscular NST is based on the involvement of several hormones, including the glucagon, catecolamines and thyroid hormones (Duchamp et al. 1999). In vivo and in vitro experiments in some cold-acclimated bird species (muscovy ducklings, king penguin chicks, domestic chickens) indicate that these hormones are potential regulators for NST (e.g. Barré & Rouanet 1983, Barré et al. 1987, Eldershaw et al. 1997, Marmonier et al. 1997, Duchamp et al. 1999). The action of these hormones on the metabolism is generally dose dependent. It is also difficult to distinguish between the direct and indirect effects of these hormones on heat production. At the moment, the existence and mediation of avian NST has not yet been proved and the regulatory pathway has not yet been described. Two separate thermogenic processes have been proposed as accounting for avian muscle NST. The first one is based on the uncoupling of mitochondrial oxidation and phosphorylation (Barré et al. 1986b), and the second mechanism involves increased ATP-dependent sarcoplasmic reticulum Ca²⁺-cycling (Dumonteil et al. 1994). Block (1994) suggested that the futile cycling of Ca²⁺ is a common feature in all kind of muscular thermogenesis. The only net effect of these futile cycles is the loss of ATP and release of heat either by alternating with passive flow of ions through membrane and active re-transport or by catalyzing chemical reactions between two substrates back and forth with different enzymes, e.g. between glucose and glucose-6-phosphate (Surholt & Newsholme 1983).

Both in birds and mammals, the resting muscle oxygen uptake (heat production) has been studied in perfused leg muscles. Muscle metabolism and performance are dependent on the regulation of the blood flow within the muscle. Vasoconstrictors, which increase the perfusion pressure in muscle, yield responses that can be divided into two types depending on their metabolic actions (Clark et al. 2000). In perfused rat hind muscle, increased oxygen and nutrient uptake has been observed in response to the influence of noradrenaline, vasopressine and angiotensin II (Colquhoun et al. 1988, Tong et al. 1997, Newman and Clark 1998) and this is named type A response. Type B response results in decreased muscle oxygen consumption and nutrient efflux (Clark et al. 2000). It has been proposed that muscle has two distinct vascular routes, nutritive and non-nutritive, operating in parallel and regulated by vasoactive substances. The nutritive route is in close contact with muscle cells and the non-nutritive route functions as a vascular shunt leading the blood flow to the connective tissues and associated adipocytes (Clark et al. 2000). The nutritive/non-nutritive flow ratio has a great role in setting the basal metabolic rate. A high nutritive/non-nutritive flow ratio favours the acquisition of nutrients and hormones to the muscle cells and elevates the total metabolism in the muscle. However, when the non-nutritive flow is high, this favours the growth of fat tissue adjacent to the muscle. Although an increased nutritive flow in vitro results in increased muscular thermogenesis, this is unlikely to be thermoregulatory heat production but a part of obligatory heat production.

Specialized thermogenic tissues also exist in some fish species. Specialized heater cells warming the blood going to the brain and eye have been found in the modified eye muscles of the billfish, Xiphiidae and Istiophoridae (Carey 1982, Block 1987, Block 1994). These cells contain numerous mitochondria, hypertrophied T-tubule and sarcoplasmic reticulum membranes but lack organized myofibrillar structures and uncoupling protein. Heat production originates from intense Ca²⁺ cycling in the sarcoplasmic reticulum which is enabled by the rich content of Ca²⁺ ATPase in membranes. Heater cells have not been found in the tissues of birds or mammals.

2.2.3. Postprandial excess heat production

When a fasting animal begins to consume food, the metabolic rate quickly increases above the resting level. This postprandial excess heat production has been variously termed obligatory diet-induced thermogenesis (DIT), the thermic effect of food, specific dynamic action of feeding (SDA), the heat increment of feeding (HIF), or digestion-related thermogenesis (DRT). In a strict sense, many of these terms possess their own narrow specific meanings (see IUPS Thermal Commission 2001). In the literature, these terms are used in general as synonyms.

Postprandial heat production arises from complicated and combined metabolic reactions which are not clearly known. Heat production is dependent both on the amount of food digested and on the time elapsed since the meal. Augmented heat production has been explained as the obligatory utilization of ATP in the metabolic processing of ingested material (Himms-Hagen 1976). Postprandial heat production comprises several factors including: increased muscular activity; fermentation, hydrolysis and absorption in the intestinum; neuronal and hormonal changes; increased active ion transport; increased protein turn-over in cells; and pharmacological effects of nutrients (Blaxter 1989). The magnitude of postprandial heat production is thought to be determined primarily by the rate of protein synthesis and turnover, thus reflecting the metabolic cost of growth (Jobling 1983, Carter & Brafield 1992, Janes & Chappell 1995). The magnitude of postprandial heat production ranges from 30–31% of assimilated energy in protein to 13% in lipid and only 5–6% in carbohydrates (Harper 1971, Ricklefs 1974). Furthermore, postprandial heat production is believed to be mainly the result of the energy released from endogenous energy reserves since release of energy from ingested nutrients in the gut is slower (Schieltz & Murphy 1997). Vagal afferent signals obviously have a significant role in the onset of postprandial heat production (Székely 2000).

There are indications both in favour and against the idea of postprandial heat production substituting for regulatory thermogenesis in cold. The occurrence of substitution can be observed especially by comparing the metabolic rates of fed and starved animals at a wide range of ambient temperatures. The decrease in the ratio of the metabolic rate of fed animals to that of starved ones with decreasing ambient temperature indicates substitution. Bergman and Snapir (1965) compared the ratio in three breeds of domestic fowls and found a decline from 1.2 to 1.0 at ambient temperatures of 32°C and 16°C, respectively. Studies by Misson (1982) and Visser (1991), however, suggest that substitution does not occur in one or two-week-old domestic fowl chicks, respectively. In kestrels (Falco tinnunculus), about 50% of postprandial heat production substitutes for regulatory thermogenesis at temperatures below 10°C (Masman et al. 1989). Biebach (1984) found that substitution occurs in incubating starlings (Sturnus vulgaris) but Klaassen et al. (1989) did not find substitution in Arctic tern chicks (Sterna paradisea). More evidence supporting the role of postprandial heat production as a substitute for regulatory thermogenesis at least partially has been found in free-ranging verdins (Auriparus flaviceps) (Webster & Weathers 1990), in granivorous song birds (Meien-berger & Daubenschmidt 1992), in blue grouses (Dentragapus obscurus) during winter (Pekins et al. 1992), in Adelie penguins (Pygoscelis adeliae) (Janes & Chappell 1995), in house wren chicks (Troglodytes aedon) (Chappell et al. 1997), in Brünnich’s guillemots (Uria lomvia) (Hawkins et al. 1997) and in domestic pigeons during the night (Rashotte et al. 1997, Rashotte et al. 1999). There are also studies which support the idea of substitution in mammals, e.g. in golden hamsters (Simek 1975), in sea otters (Enhydra lutris) (Costa & Kooyman 1984) and in muskrats (Ondatra zibethicus) (MacArthur & Campbell 1994). However, the lack of substitution has been observed too, as in studies with short-tailed shrews (Platt 1974) and with star-nosed moles (Condylura cristata) (Campbell et al. 2000). As a result of all this evidence, it is justified to conclude that postprandial heat production is a true substitute for regulatory thermogenesis at least in some species and in some conditions.

Postprandial heat production is a part of the resting metabolism and is regulated more or less independently of thermoregulation (Schieltz & Murphy 1997). Some studies have also revealed that the active regulation of postprandial heat production for thermo-regulatory purposes may occur during the night (Rijnsdorp et al. 1981, Rashotte et al. 1997). The storage of food in the crop for the night and the consumption of the food when necessary allows for postprandial heat production during the night in a regulated way, thus serving the needs of thermoregulation. The heat production arises from the processing of endogenous energy reserves in response to movements of bulk in the gut rather than from the processing of nutrients contained in the gut itself (Rashotte et al. 1997), a fact proved by the finding that even the feeding of non-digestable cellulose pellets results in increased heat production (Reinertsen & Bech 1994, Geran & Rashotte 1997).

2.2.4. Exercise thermogenesis

During shivering, mammals and birds apparently cannot utilize the full metabolic capacity of their muscles in thermogenesis since locomotion has been observed to have even more heat productive capacity (Marsh & Dawson 1989). Running and flying in birds can maximally yield metabolic rates from 10 to 12 times of the resting values, respectively, though more common values range between 5–10 times (Brackenbury 1984). Corresponding values for cold-induced thermogenesis are approximately 3–8 times the resting values (see Dawson & Whittow 2000).

Similar to postprandial heat production, exercise thermogenesis may be utilized for the substitution of regulatory thermogenesis in cold. As with postprandial thermogenesis, evidence is partly controversial. The extent of substitution varies among the species studied and with the ambient temperatures. Furthermore, all the information comes from adult animals. Signs in favour of substitution have been observed in budgerigars, whose flying oxygen consumption is constant between the ambient temperatures of 37–18°C (Tucker 1968) but whose resting oxygen consumption is doubled (Greenwald et al. 1967). In chaffinches (Fringilla coelebs), the differences between the oxygen consumption of voluntary activity and that of resting decreased between the temperatures of 32–5°C (Pohl 1969), indicating a partial substitution. In redpolls (Carduelis flammea), Pohl and West (1973) showed regulatory thermogenesis being replaced by exercise at very low temperatures (-30 to -45°C) during forced bipedal exercise. In higher temperatures (between -30 and 0°C) the costs of thermoregulation and exercise were additive. Further studies indicating substitution, partial or complete, have involved white-crowned sparrows (Zonotrichia leurophrys gambelii) (Ketterson & King 1977, Paladino & King 1984), Japanese quails (Nomoto et al. 1983a), free-ranging verdins (Auriparus flaviceps) (Webster & Weathers 1990, 2000), Gambel’s quails (Callipepla gambelii) (Zerba & Walsberg 1992), free-ranging juncos (Junco phaeonotus and Junco hyemalis) (Weathers & Sullivan 1993), knots (Calidris canutus) (Bruinzeel & Piersma 1998), domestic pigeons (Hohtola et al. 1998), and Eastern house finches (Carpodacus mexicanus) (Zerba et al. 1999).

In two hummingbird species (Amazilia cyanifrons and A. tzacatl), Schuchmann (1979a) found the difference between the oxygen consumption during resting and hovering to be constant between the ambient temperatures 5–40°C indicating no substitution. Similarly, in another hummingbird, the booted racket-tail (Ocreatus u. underwoodii), substitution does not exist (Schuchmann 1979b). More studies indicating that the metabolic costs of thermoregulation and exercise are additive have involved white-throated sparrows (Zonotrichia albicollis) (Kontogiannis 1968) and dippers (Cinclus cinclus) (Bryant et al. 1985).

Exercise may result in an augmentation of heat loss due to increased forced convection. Goldstein (1983) observed the metabolic rate of Gambel’s quail increasing linearly with increasing wind speed. A positive correlation between the metabolic rate and wind speed also occurs in verdins (Webster & Weathers 1988). When exercising and resting birds are compared in similar net convective conditions in cold, heat produced by exercise is completely used for substituting regulatory thermogenesis (Zerba & Walsberg 1992, Zerba et al. 1999). This may suggest that success in the prevention of excessive heat loss determines whether substitution can occur.

The existence of substitution suggests that exercise has an influence on shivering intensity. Exercise can affect shivering thermogenesis by inhibiting it either directly via neuronal interaction or indirectly via increased body temperature. Nomoto and Nomoto-Kozawa (1985) observed suppressed shivering due to exercise in domestic pigeons. Within one second after the onset of bipedal exercise on a treadmill, shivering in pectoral muscle decreased by 75%. Nomoto and Nomoto-Kozawa (1985) and Nomoto (1989) concluded that a direct neuronal inhibition is present. On the other hand, an exercising pigeon is capable of increasing shivering in the pectorals when the spinal cord is selectively cooled (Nomoto et al. 1983b, Nomoto & Nomoto-Kozawa 1985), which indicates that shivering can coincide with exercise. Hohtola et al. (1998) did not observe direct and abrupt inhibition of shivering in pectorals of pigeons during spontaneous walking or other types of voluntary movements (e.g. preening, feeding and pecking). Hohtola et al. (1998) suggested that startle reactions and direct neural inhibition may be involved in treadmill experiments where forced exercise is used. Hohtola et al. (1998) further concluded that voluntary movements may decrease the need for shivering indirectly by increasing internal heat production and body temperature. In humans, Hong and Nadel (1979) found graded inhibition of shivering during a pedaling exercise using a cycle ergometer, the inhibition increasing at any level of internal temperature when the intensity of exercise increased. The authors concluded that inhibition of a central origin exists and that exercise is accompanied with an integrated arousal response which has precedence over thermoregulatory activities. However, the study of Hong and Nadel (1979) also reports the simultaneous existence of exercise and shivering.