behavioural hypothermia during exercise recovery in frogs

Ambient temperature strongly influences animal behaviourand metabolism, and has long been a primary focus of studieson the comparative physiology of ectothermic vertebrates.Long-term thermal acclimation as well as acute temperaturechanges can have powerful consequences on reproduction,energetics, locomotion and survival (Crawshaw, 1979; Prosser,1991). That temperature is important to biological processes isillustrated by the existence of behavioural thermoregulation inmany ectotherms. Preferred body temperatures oftencorrespond with temperature ranges optimal for reproductivebehaviour, growth, locomotory performance and blood oxygencapacity (Pough, 1976; Huey, 1982; Hutchison and Dupré,1992). However, most studies on behavioural thermoregulation

have focused on fish and reptiles, presumably because theirmore constant habitat (either aquatic or terrestrial) allows moreprecise thermoregulation. Amphibians, in contrast, areconsidered poor thermoregulators, simply trackingenvironmental temperature changes (Brattstrom, 1963, 1979).Regardless of the precision of thermoregulation, changes inpreferred or ambient temperature will undoubtedly alter thelocomotory and metabolic capacity of amphibians, especiallythose overwintering under the ice of lakes and ponds subjectto complex temporal and spatial gradients of temperature(Bradford, 1983; Pinder et al., 1992; Friet, 1993).

Until recently, behaviour in the overwintering aquatic frogRana temporariahas been largely ignored (but see Tattersall

609The Journal of Experimental Biology 202, 609–622 (1999)Printed in Great Britain © The Company of Biologists Limited 1999JEB1849

At the low temperatures of the overwinteringenvironment of the frog Rana temporaria, small changes inambient temperature have large effects on metabolism andbehaviour, especially since Q10 values are often greatlyelevated in the cold. How the overwintering aquatic frogcopes with variable thermal environments in terms of itsoverall activity metabolism and recovery from pursuit bypredators is poorly understood, as is the role of behaviouralthermoregulation in furthering recovery from intenseactivity. Exhaustive exercise was chosen as the method ofevaluating activity capacity (defined by time to exhaustion,total distance swum and number of leg contractions beforeexhaustion) and was determined at 1.5 and 7 °C. Othercohorts of frogs were examined at both temperatures todetermine the metabolic (acid–base, lactate, glucose, ATPand creatine phosphate) and respiratory responses toexercise in cold-submerged frogs. Finally, temperaturepreference before and after exercise was determined in athermal gradient to define the importance of behaviouralthermoregulation on the recovery rates of relevantmetabolic and respiratory processes. Activity capacity wassignificantly reduced in frogs exercised at 1.5 versus7 °C,although similar levels of tissue acid–base metabolites andlactate were reached. Blood pH, plasma PCO∑ and lactatelevels recovered more rapidly at 1.5 °C than at 7 °C;

however, intracellular pH and the recovery of tissuemetabolite levels were independent of temperature. Restingaerobic metabolic rates were strongly affected bytemperature (Q10=3.82); however, rates determinedimmediately after exercise showed a reduced temperaturesensitivity (Q10=1.67) and, therefore, a reduced factorialaerobic scope. Excess oxygen consumption recovered toresting values after 5–6.25 h, and 67 % recovery timestended to be slightly faster at the lower temperatures.Exercise in the cold, therefore, provided an immediatelyhigher factorial scope, which could be involved in the fasterrate of recovery of blood lactate levels in the colder frogs.In addition, exercise significantly lowered the preferredtemperature of the frogs from 6.7 to 3.6 °C for nearly 7 h,after which they returned to their normal, unstressedpreferred temperatures. Thus, a transient behaviouralhypothermia in the skin-breathing, overwintering frog maybe an important strategy for minimising post-exercisestress and maintaining aerobic metabolism duringrecovery from intense activity.

Key words: behavioural hypothermia, temperature selection,thermoregulation, overwintering, amphibian, exhaustive exercise,recovery, frog,Rana temporaria.

Summary

Introduction

DOES BEHAVIOURAL HYPOTHERMIA PROMOTE POST-EXERCISE RECOVERY INCOLD-SUBMERGED FROGS?

GLENN J. TATTERSALL1,* AND ROBERT G. BOUTILIERDepartment of Zoology, University of Cambridge, Cambridge CB2 3EJ, UK

1Present address: Department of Physiology, Northeastern Ohio Universities College of Medicine, Rootstown, Ohio, 44272-0095, USA*e-mail: [email protected]

Accepted 8 December 1998; published on WWW 3 February 1999

610

and Boutilier, 1997). In the past, cold-submergedamphibians were believed to enter a state of cold-torporwhereby they become sluggish and unresponsive (Holzapfel,1937; Willis et al., 1956). However, recent field studies haveshown that overwintering frogs do not remain inactive underthe ice (Friet, 1993; Stinner et al., 1994), and laboratorystudies have begun to back up the contention that behaviourcan play an important role in maintaining aerobicmetabolism and minimising the influence of external stresses(e.g. hypoxia; Tattersall and Boutilier, 1997). Whenencountering ambient hypoxia, cold-submerged frogsrespond by selecting a lower environmental temperature(1.5 °C during hypoxia versus 7 °C during normoxia;Tattersall and Boutilier, 1997). This hypoxia-induceddichotomy in preferred temperatures begs the question: dothese small changes in temperature pose large respiratory,metabolic or activity costs? Unstressed animals may preferhigher temperatures since these enhance locomotoryfunction, even though they may seem to be detrimental forlong-term survival by accelerating the oxidation of thelimiting overwintering fuel stores. However, followingstressful situations, such as intense activity, the decidingfactor for temperature preference may be the need for rapidrecovery of aerobic metabolism or to process the end-products of anaerobiosis. In addition, the internal hypoxiaand acidosis often associated with exercise may also act assignals for altering temperature preference during therecovery period from activity. Indeed, only recently has itbeen appreciated that many ectotherms and even someendotherms become hypothermic in the face of many formsof stress, such as hypoxia, hypercapnia or anaemia (Woodand Gonzales, 1996), thus emphasising the potentialsignificance of this widespread response.

The aquatic overwintering frog encounters conditions wherepredation risks from fish or mammals are substantial (Emeryet al., 1972; Pinder et al., 1992). Pursuit by predators at lowtemperatures could conceivably lead to exhaustion and rapidlactic acid accumulation, since more fast-twitch fibres arerecruited to power swimming at lower temperatures (Rome etal., 1992). Furthermore, the high Q10 values for metabolic rateat low temperatures (Tattersall and Boutilier, 1997), combinedwith the diffusion-limited respiratory exchange surface of thesubmerged frog, mean that both activity and respiratoryrecovery should be limited by different capacities dependingon temperature. A priori, low temperatures could lead morerapidly to exhaustion and a slower recovery as a result of toQ10 effects, whereas elevated temperatures might lead to moreintense activity and the accumulation of high levels of lactate.However, by reaching the maximal capacity for cutaneousoxygen uptake at the higher temperatures, a longer recoveryperiod following exhaustive exercise may be required, makingpredictions based on the normal assumption of faster rates athigher temperatures unreliable. In addition, since recoveryfrom exercise represents a refractory period for normal musclecontraction and resting metabolism, post-exercise frogs mayfind themselves even more vulnerable to predation (Gatten et

al., 1992). Unravelling the effects of temperature on post-exercise recovery metabolism may aid in elucidating the roleof behavioural thermoregulation in ensuring overwinteringsurvival.

Following intensive exercise in frogs, metabolic rates areelevated three- to tenfold above resting levels, and bloodlactate concentrations increase five- to tenfold (Gleeson,1991). However, any apparent causal relationship between therates of oxygen consumption (M

.O∑) and lactate removal is no

longer believed to be wholly responsible for the elevated post-exercise metabolic rate (Gaesser and Brooks, 1984; Gleeson,1996). In anuran amphibians especially, lactate recoveryfollows a more extended time course than M

.O∑ recovery

(Bennett and Licht, 1973; Gatten, 1987; Withers and Hillman,1988; Gleeson, 1996) and is also quite sensitive to temperature(Q10=1.5–2.5; Gleeson, 1991). This relationship isaccentuated in the purely skin-breathing amphibianCryptobranchus alleganiensis, which takes up to 22 h torecover all the blood-borne lactate, as opposed to the 4–8 hrequired for an air-breathing anuran (Boutilier et al., 1980).To our knowledge, the simultaneous effects of temperature onlactate removal and M

.O∑ recovery have only been examined

in the lizard Dipsosaurus dorsalis(Wagner and Gleeson,1996). These authors demonstrated that post-exercisetemperature preference is briefly lowered, even though thisdoes not lead to a faster lactate recovery when compared withlizards at a constant, warm temperature (Wagner and Gleeson,1997). The reasons for this temporary decrease in preferredtemperature are not clear, but a lowering of body temperature(e.g. behavioural hypothermia) in cold-submerged frogswould probably benefit the recovery from exercise, given thelarge thermal sensitivities of biochemical events and thepotentially limiting gas exchange capacity of the skin (Pinder,1987). Studying this phenomenon in an amphibian maydelineate evolutionary differences in physiological exerciserecovery and answer ecological questions relevant to theoverwintering frog.

The objectives of this study of exercise in cold-submergedfrogs were threefold: (1) to determine the thermalsensitivities of behavioural and metabolic correlates ofexhaustive exercise, (2) to establish the effects of temperatureon metabolic and respiratory recovery from exhaustiveexercise, and (3) to investigate whether exhaustive exercisechanged preferred temperatures. These were accomplished bymanually pursuing submerged frogs at 1.5 and 7 °C until theybecame exhausted, and recording the time, distance andnumber of leg contractions to exhaustion. Different groups ofsubmerged frogs were then allowed to recover for 24 h at therespective exercise temperatures (1.5 or 7 °C), and blood andtissue acid–base variables and levels of associatedmetabolites were measured (series I experiments). A separategroup of frogs was exercised in the same manner, and post-exercise rates of oxygen consumption were determined at 1.5and 7 °C for 10–12 h (series II). Finally, temperaturepreference was examined for 12 h before and 12 h afterexercise (series III).

G. J. TATTERSALL AND R. G. BOUTILIER

611Behavioural hypothermia during exercise recovery in frogs

Materials and methodsMaintenance of experimental animals

Male Rana temporaria(mean mass 22 g) were purchasedfrom Blades Biological and were caught from wild populationsin Ireland in October 1997. All frogs were acclimated for atleast 3 weeks at 3–4 °C in aerated water in Living Streams™

(Frigid Units, Inc.) before experimentation. For the duration ofacclimation and experimentation, all frogs were left unfed andkept in reduced light to mimic as closely as possible theoverwintering conditions in the wild. Acid–base andmetabolite determinations (series I) were conducted inDecember 1997, metabolic rate determinations (series II) wereconducted in November 1997, and the temperature preferencedeterminations (series III) were conducted in January 1998. Allexperimental series were conducted on different cohorts offrogs.

Exercise protocol

Submerged frogs from the acclimation tanks (3–4 °C) weretransferred to the experimental temperatures (1.5 or 7 °C;earlier determined to be the preferred temperatures underhypoxia and normoxia, respectively; Tattersall and Boutilier,1997) 8 h before being exercised to minimise the effects oftemperature changes or handling. Individual frogs were placedin a trough (0.12 m×0.1 m×1.25 m) filled with water andmaintained at the experimental temperature. Animals werechased from end to end of the trough until they were exhausted;the exhaustion end-point being taken as the loss of the rightingreflex for at least 10 s, following an initial refusal to swimfurther. For the duration of the exercise period, frogs wereprevented from surfacing to breathe air. The swimmingparameters measured were the duration of exercise (time toexhaustion), the number of leg contractions and the totaldistance swum (m). These measurements enabled thecalculation of swimming speed (non-sustainable), legcontraction frequency and leg contraction efficiency averagedacross the entire exercise period. The same exercise protocolwas applied to all three series of experiments, but theswimming data were only collected from frogs in series I(N=50). All animal experiments were performed underlicensed approval by the Home Office (UK).

Series I: acid–base and metabolite recovery experiments

All frogs were exercised as described above and allowed torecover without access to air in water at the same temperatureat which they had been exercised (1.5 or 7 °C). In addition toa resting control group at both temperatures (N=5; 8 hacclimation at 1.5 or 7 °C), there were four recovery samplinggroups (N=5 each): 0, 2, 8 and 24 h post-exercise recovery. Ateach sampling time following exercise, frogs wereimmediately transferred to an anaesthetic solution (1 % MS-222 buffered with 1.2 % NaHCO3) until unresponsive(approximately 10 min). Blood (up to 300µl) was drawn fromthe left aortic arch into heparinised capillary tubes, sealed andplaced on ice until acid–base measurements were made (within5 min). Subsequent to withdrawing the blood, samples of

gastrocnemius muscle were removed, quickly rinsed in anisotonic solution to eliminate excess blood, freeze-clamped at−196 °C and then stored at −80 °C for up to 1 month, afterwhich intracellular acid–base and metabolite measurementswere conducted.

Blood acid–base variables (pHe and total CO2) andhaematocrit were determined as described previously(Tattersall and Boutilier, 1997), and true plasma bicarbonateconcentration and PCO∑ were calculated using the apparentplasma pK′and CO2 solubilities given by Heisler (1989). Theremaining plasma was frozen and analysed for lactate andglucose along with the extracted tissues. Intracellular pH,tissue bicarbonate level and PCO∑ in the gastrocnemius weredetermined as described previously (Pörtner et al., 1990).

To make metabolite measurements, frozen gastrocnemiustissues were pulverised under liquid nitrogen, and a knownmass was extracted in 500µl of 7 % perchloric acid. Thesupernatant remaining after centrifugation at 15 800g wasneutralised to pH 7 with 2 mol l−1 potassium hydroxide and0.4 mol l−1 sodium imidazole. Plasma was also extracted in asimilar fashion. Tissue metabolite levels were measuredaccording to standard enzymatic techniques (Passoneau andLowry, 1993). In the plasma, only lactate and glucoseconcentrations were determined. In the muscle, lactate,glucose, glycogen, creatine phosphate (PCr), creatine and ATPlevels were measured, although glycogen concentrations weretoo variable in this cohort of frogs to draw useful conclusions.

Series II: metabolic rate recovery experiments

Frogs (N=16 at 1.5 °C and N=13 at 7 °C) were transferredfrom Living Streams at 3–4°C to respirometers (total volume220ml) and left to acclimate to the new experimentaltemperature overnight (8h). The following morning, the rate ofoxygen consumption (M

.O∑) was determined for at least 1h

before swimming the frogs. Frogs were exercised at 1.5 and 7°Cto the point of exhaustion according to the exercise protocoldescribed above. Immediately following exercise, the frogs werereturned to the respirometer, and the post-exercise rate of oxygenconsumption was measured until it reached the previous restinglevels or for up to 12h. Oxygen consumption was determinedby continuously monitoring the decline in PO∑ in a closedrespirometer by passing the water over a thermostattedpolarographic oxygen electrode (Cameron Instruments; Texas,USA). The PO∑ was not allowed to fall below 100mmHg(13.3kPa) (determined to be well above the critical PO∑ foroxygen uptake in cold-submerged frogs; G. J. Tattersall and R.G. Boutilier, in preparation), at which time the respirometerswere gently flushed with aerated water and resealed for furthermonitoring of PO∑. The PO∑ signal from the oxygen meter(OM200, Cameron Instruments) was connected to a computervia an A/D converter, and data were sampled at 60Hz andaveraged over 10s intervals to reduce electrical interference.Subsequently, the negative derivative of PO∑ with respect to timewas calculated using purpose-written software (Datlab,Oroboros, Austria) and converted to rate of oxygenconsumption. Resting metabolic rate was taken as the mean rate

612

of oxygen consumption over the 1h that preceded exercise; inthe final analysis, 15min means of post-exercise metabolic rateswere used to simplify statistical comparisons. Other parametersof recovery metabolism were calculated using the raw oxygenconsumption data (Fig. 1). Maximal metabolic rates werecalculated as the extrapolated ‘zero time’ value after fitting adouble-exponential equation to the oxygen consumption data ofeach individual frog. This equation also enabled a reasonableestimate of the total post-exercise oxygen consumption byintegrating over the entire recovery period. By accounting for aconstant resting metabolic rate across the same experimentalperiod, total excess post-exercise oxygen consumption (EPOC;µmolO2g−1) could be calculated. Similarly, excess oxygenconsumption (EOC; µmolO2g−1) during exercise was estimatedby assuming a maximal metabolic rate across the entire periodof exercise and subtracting the resting rate over the same period.Finally, the time for a 67% recovery of metabolic rate (aboveresting levels) was calculated by rearranging the fitted double-exponential equations for each animal.

Series III: temperature preference recovery experiments

The temperature preference of cold-submerged frogs (N=17)was measured in a linear temperature gradient, as reportedpreviously (Tattersall and Boutilier, 1997). Animal locationalong the gradient chamber was determined using a bank ofinfrared detectors and was subsequently converted to preferredtemperature using the equation describing water temperature (T,°C) versusdistance (d, cm) along the gradient: T=0.099d−0.16(r2=0.995). All data were sampled at 1 min intervals, and thefinal preferred temperatures represented means calculated over15 min periods before and after exercise. Over the 12 h beforeexercise, temperature preference was determined to establishcontrol values. At the end of this 12 h period, individual frogswere swum to exhaustion at 3–4 °C (the overwinteringacclimation temperature, as well as mid-way between the twoexperimental temperatures) and then returned to the thermal

gradient. Since inactivity following exhaustive exercise couldconfound any temperature preference behavioural activity, twogroups of frogs were studied: one group was placed at the coldend (N=8) and another group at the warm end (N=9) of thethermal gradient (referred to as cold and warm recovery inFig. 7), and preferred temperatures were observed for a further12 h. This was performed to ensure that the preferredtemperature during recovery would not simply be an artefact ofanimal inactivity. This effect was later determined to beinsignificant (P=0.8), so all temperature selection data weresubsequently pooled during the final analysis.

Statistical analyses

All metabolite and acid–base results were analysed using two-way analysis of variance (ANOVA), with temperature andexercise (resting and exercise recovery) as factors. Post hoccomparisons of significant factors were conducted usingmultiple comparison tests (Student–Newman–Keuls). Fifteenminute means of post-exercise metabolic rate data werecompared with the pre-exercise resting metabolic rate usingmultiple comparison tests (Student–Newman–Keuls). Othervariables derived from the metabolic rate (maximal, metabolicscope, factorial scope, total post-exercise oxygen consumed,EPOC and recovery time) were analysed using t-tests comparingthe 1.5 and 7°C data. Temperature preference data wereanalysed using repeated-measures two-way ANOVA, withplacement (cold versuswarm end) and exercise status (pre-versus post-exercise) as factors, and time as the repeatedmeasure. Since the side (cold or warm) of the tank in which frogswere placed following exercise was found to be insignificant(P=0.8), the data were pooled and treated as one groupthereafter. Post hocanalysis of temperature preference data wasperformed using pairwise comparisons of each time pointagainst the mean preferred temperature prior to exercise(Student–Newman–Keuls). All statistical tests were consideredsignificant at P<0.05, and data are presented as the mean ±S.E.M.


R E 0 4 5 6 8 9 107

EPOC

.

EOC

.

Met

abol

ic r

ate

(µm

ol O

2 g-1

h-1

)

1.8

1.6

1.4

1.2

1.0

0.8

0.6

0.4

0.2

01 2 3

Time (h)

MO2=0.876e−0.0836T+0.857e−0.965T

.MO2rest

T0.67.

MO2maxFig. 1. Representative resting andpost-exercise metabolic rates (M

.O∑)

of a frog at 1.5 °C. Resting rate (R)was determined for up to 1 h beforeexhaustive exercise (E), andrecovery rate was determinedfor up to 10 h. Maximal M

.O∑

(M.O∑max; 1.74µmol O2g−1h−1) was

determined by extrapolating thedouble-exponential plot to 0 h,enabling the calculation ofscope and factorial scope. Excessoxygen consumed during exercise(EOC; hatched area) [EOC = M

.O∑max minus M

.O∑rest) × exercise

duration = 0.302µmol O2g−1] wasestimated assuming maximal M

.O∑

throughout the exercise period. Excess post-exercise oxygen consumption (6.66µmol O2g−1) (EPOC; unshaded area under the curve) wascalculated by integrating the fitted curve (total area under the curve) and subtracting the area due to resting M

.O∑ (M

.O∑rest; shaded area). Scope

= M.O∑max minus M

.O∑rest= 1.39µmol O2g−1h−1; factorial scope = M

.O∑max/M

.O∑rest= 4.97; 67 % recovery time (T0.67) = 1.78 h.


ResultsSwimming parameters of exhaustive exercise

On average, frogs were capable of swimming for15–18 min before reaching exhaustion, and this was notsignificantly affected by temperature (P=0.17; Table 1). Thetotal number of leg contractions, distance swum, swimmingspeed and leg contraction frequency were all significantlygreater at 7 °C than at 1.5 °C, whereas leg contractionefficiency was not significantly affected by temperature(Table 1). A general observation of exercising frogs wasthat their swimming appeared to be more coordinated at 7 °Cthan at 1.5 °C, with there being more simultaneous legstrokes. In addition, the closer to exhaustion the frogs becameat both temperatures, the less coordinated was theirswimming, such that relative swimming speed slowed downto the point where it became non-sustainable, and thus wasthe end-point for exhaustion. Frogs at 7 °C also madenumerous attempts to surface to breathe air, while frogsswimming at 1.5 °C simply remained submerged throughoutthe exercise period.

Series I: acid–base and metabolite recovery experiments

Exhaustive exercise induced a combined respiratory andmetabolic acidosis in the plasma that was more pronounced at7 °C than at 1.5 °C (P<0.05; Fig. 2; Table 2). At the coldertemperature, plasma acid–base status was restored to normalwithin 2 h; at 7 °C, the restoration of pHe, arterial PCO∑ (PaCO∑)and [HCO3−] occurred after 8 h. Although not significantly so,post-exercise plasma bicarbonate concentrations at 1.5 °C wereconsistently lower than control values, in contrast with the fullrecovery of bicarbonate observed at 7 °C (Fig. 2; Table 2).

Haematocrit was significantly influenced by temperature andexercise (Table 2), increasing from 27.0 % at rest to 34.3 %immediately following exercise at 1.5 °C (P<0.05) and from30.1 % to 44.5 % at 7 °C (P<0.05). Resting haematocrit levelswere re-established at both temperatures by 2 h after exercise.

Gastrocnemius acid–base balance following exercise wasalso disturbed (pHi falling to 7.184–7.267 from 7.426–7.465),resulting in a significant acidosis unaffected by temperature(Fig. 3; Table 3). By 2 h post-exercise, pHi had recovered tothe pre-exercise control values at both temperatures. Tissue

Table 1. Swimming variables in submerged frogs after exhaustive exercise at 1.5 and 7 °C

Contraction ContractionTemperature Exhaustion time Number of leg Swimming Swimming speed frequency efficiency(°C) (min) contractions distance (m) (m min−1) (min−1) (m−1)

1.5 15.2±1.1 384±28 44.0±3.6 2.89±0.12 25.3±0.65 8.86±0.187.0 17.9±1.1 639±32 69.2±3.8 3.92±0.11 36.4±0.98 9.34±0.23

Significant P=0.17 P<0.001 P<0.001 P<0.001 P<0.001 P=0.3temperature effect

Q10 1.35 2.52 2.28 1.74 1.94 1.10

Swimming speed = swimming distance/exhaustion time.Contraction frequency = number of leg contractions/exhaustion time.Contraction efficiency = number of leg contractions/swimming distance.Values are means ±S.E.M., N=50.

Table 2. Plasma and blood properties before and after exercise in submerged frogs at 1.5 and 7 °C

Post-exercise recovery timeTemperature Pre-exercise

Variable (°C) control 0 h 2 h 8 h 24 h Significant effect

Haematocrit 1.5 27.0±2.2 34.3±3.2* 31.7±4.4 24.7±4.2 26.9±4.2 Temperature7.0 30.1±4.2 44.5±3.2*,‡ 37.3±1.8 31.9±1.7 37.5±2.9 Exercise

[Lactate] (mol l−1) 1.5 0.98±0.56 10.79±1.16* 6.91±1.82* 3.02±1.34 1.98±0.21 Temperature7.0 1.68±0.34 12.96±1.16* 11.57±0.88*,‡ 4.00±1.02 3.16±1.24 Exercise

[Glucose] (mol l−1) 1.5 1.46±0.19 2.41±0.15 2.30±0.20 1.94±0.41 1.81±0.34 Exercise7.0 1.51±0.25 2.81±0.32* 2.57±0.21 1.97±0.16 2.53±0.38

*N=5 frogs per treatment level; values are means ±S.E.M.Haematocrit = total cell volume/total blood volume.*Significantly different from the control value at the respective temperature (P<0.05).‡Significant temperature effect at a time point (P<0.05).

614 G. J. TATTERSALL AND R. G. BOUTILIER

Table 3. Biochemical properties of gastrocnemius muscle before and after exercise in submerged frogs at 1.5 and 7 °C

Post-exercise recovery timeTemperature Pre-exercise

Variable (°C) control 0 h 2 h 8 h 24 h Significant effect

[Lactate] (mmol l−1) 1.5 1.05±0.25 4.63±0.93* 2.09±0.86 1.88±0.42 2.15±0.06 Exercise7.0 1.45±0.28 6.38±0.73* 3.07±0.43 1.82±0.51 1.38±0.32

[Glucose] (mmol l−1) 1.5 0.200±0.062 0.370±0.034 0.531±0.112* 0.562±0.076* 0.248±0.034 Exercise7.0 0.143±0.044 0.720±0.125* 0.790±0.041* 0.439±0.125 0.197±0.044

[ATP] (mmol l−1) 1.5 1.90±0.11 1.89±0.23 1.62±0.14 1.88±0.08 1.75±0.28 −7.0 1.79±0.12 1.88±0.13 2.11±0.15 1.98±0.14 2.02±0.03

[Creatine phosphate] 1.5 11.68±0.79 9.12±1.10 8.77±1.03 11.62±0.79 10.24±1.76 Exercise(mmol l−1) 7.0 12.16±0.95 6.85±0.78* 10.51±1.00 10.39±0.73 11.33±0.74

[Creatine] (mmol l−1) 1.5 6.73±1.31 8.89±1.36 7.41±0.95 5.42±0.42 8.73±1.02 Exercise7.0 4.81±0.42 10.25±1.14* 7.33±0.22 6.45±0.63 6.45±0.87

N=5 frogs per treatment level; values are means ±S.E.M.*Significantly different from the control value at the respective temperature (P<0.05).

Post-exercise

A

C

B

34681012

a

bc de

a

bc

deD 2

15

20

25

30

8.58.38.17.97.77.5

pHe

Pre-

*‡

*‡*

*‡121086420P

CO

2 (m

mH

g)

*‡*

* *

30

25

20

15[H

CO

3- ](m

mol

l-1)

R

a

0

b

2

c

8

d

24

e

Time (h)

7.5 7.7 7.9 8.1 8.3 8.5

PCO2 (mmHg)

[HC

O3- ]

(mm

ol l-1

)

Blood pH

Fig. 2. Blood acid–base changes before (R) and afterexercise (shaded area) in cold-submerged frogs at 1.5(filled circles) and 7 °C (open circles). (A) Changes inblood pH, (B) changes in plasma PCO∑ and (C) changes inplasma bicarbonate concentration. Values are means ±S.E.M. (N=5). (D) A Davenport diagram incorporatingmean values from A and C, without the error bars. Thesmall italic letters refer to time points before and afterexercise: a, resting; b, 0 h post-exercise; c, 2 h post-exercise; d, 8 h post-exercise; e, 24 h post-exercise.Curvilinear lines in D represent PCO∑ isopleths for 1.5(solid lines) and 7 °C (dotted lines). *Significantlydifferent from the control value at the respectivetemperature (P<0.05); ‡significant effect of temperatureat this time point (P<0.05). 1 mmHg=0.133 kPa.


PCO∑ and bicarbonate concentrations, however, were morevariable and were not significantly affected by exercise orrecovery.

The post-exercise increase in lactate concentrations mirroredthe decline in pH in both the plasma and the muscle, peakingimmediately after exercise (Fig. 4; Tables 2, 3). Temperaturehad an overall significant effect on plasma lactateconcentration, which was due to the prolonged recovery forlactate at 7 °C compared with that at 1.5 °C (8 h at 7 °C and 2 hat 1.5 °C; 67 % clearance times of 6.38 h and 4.75 h at 7 and1.5 °C, respectively; Table 2). However, the proportionalincrease in lactate concentration was higher at 1.5 than at 7 °C(Fig. 4A). Gastrocnemius lactate levels were not significantlyaffected by temperature, either before or after exercise,although recovery of lactate levels was completed by 2 h atboth 1.5 and 7 °C (Fig. 4B).

Of all the other metabolites measured, the levels ofglucose, creatine phosphate (PCr) and creatine weresignificantly affected by exercise (Table 3); ATP levels wereunaffected by exercise or temperature (Table 3), andglycogen levels were so variable that no effect of either

treatment was observed (data not shown). Plasma glucoselevels were highest immediately after exercise, whereasgastrocnemius glucose levels remained higher than restingvalues for more than 8 h following exercise. GastrocnemiusPCr levels were reduced and creatine levels were increasedfollowing exercise, although this effect was only significantat 7 °C (Table 3). By 2 h post-exercise, these values were notsignificantly different from control levels.

Series II: metabolic rate recovery experiments

Resting and maximal metabolic rates following exercisewere significantly affected by temperature (0.31±0.02 and1.65±0.07µmol O2g−1h−1 at 1.5 °C versus0.65±0.03 and2.19±0.06µmol O2g−1h−1 at 7 °C; Table 4; Fig. 5). This wasaccompanied by a significantly larger aerobic scope formetabolic rate at 7 °C (1.55±0.05 versus1.34±0.06µmol O2g−1h−1 at 1.5 °C), although factorial scopewas significantly reduced at 7 °C (3.47±0.16 versus5.50±0.26at 1.5 °C). The estimated amount of oxygen consumed over the period of exercise (EOC) was also significantly greaterat 7 °C (0.473±0.027µmol O2g−1) than at 1.5 °C

6

8

10

12

abc

d

ea

b

c

d

e

16 14

Post-exercise

A

C

B

D

6

8

10

12

7.5

7.4

7.3

7.2

7.1

pHi

Pre-

*

201612840

PC

O2 (

mm

Hg)

15

10

5

[HC

O3- ]

(mm

ol l-1

)

R

a

0

b

2

c

8

d

24

e

Time (h)

7.1 7.2 7.3 7.4 7.5

PCO2 (mmHg)

[HC

O3- ]

(mm

ol l-1

)

Gastrocnemius pH

*

Fig. 3. Gastrocnemius acid–base changes before (R)and after exhaustive exercise (shaded area) in cold-submerged frogs at 1.5 °C (filled circles) and 7 °C(open circles). (A) Changes in intracellular pH, (B)changes in tissue PCO∑ and (C) changes in tissuebicarbonate concentration. Values are means ±S.E.M.(N=5). D is a Davenport diagram incorporating A andB. See Fig. 2 for further explanation. *Significantlydifferent from the control value at the respectivetemperature (P<0.05). 1 mmHg=0.133 kPa.

616

(0.311±0.024µmol O2g−1), as was the total amount of oxygenconsumed post-exercise (EPOC; 3.16±0.23µmol O2g−1 at1.5 °C and 4.23±0.37µmol O2g−1 at 7 °C; Table 4).

In terms of exercise recovery, metabolic rate tended torecover faster at 1.5 than at 7 °C (67 % return to resting taking2.10 h versus2.88 h; P=0.07; Table 4). However, in terms ofrecovery from resting values, post-exercise metabolic rateswere no longer significantly different after 5–6.25 h at 7 and1.5 °C (Fig. 5). The exponential decrease in metabolic ratefollowing exercise was mirrored by an increase in Q10, which

returned to a resting value of 3.82 by the end of the 10 hrecovery period, starting from a value of 1.67. This departureof the low-temperature metabolic rate from the high-temperature one is also seen by examining the changes inmetabolic rate relative to resting rates over the entire course ofrecovery. Frogs at 1.5 °C increased their metabolic rate 5.5-foldover the resting rate (significantly higher than the rate for frogsat 7 °C for the first hour of recovery), and this rate decreasedrapidly at first during recovery and then exponentially later inthe recovery period. The proportional increase at 7 °C (3.47), incontrast, decreased monotonically over time (Fig. 6).

Series III: temperature preference before and after exercise

Temperature preference in unexercised frogs was unchangedover the initial 12 h period; frogs showed a definite preferencefor the warm end of the thermal gradient (mean 6.7 °C; Fig. 7)and avoided the extremely cold temperatures until they wereexercised. Regardless of which end of the gradient the frogswere placed following exhaustive exercise, their preferredtemperature fell by 3.1 °C to 3.6 °C (i.e. they no longer avoidedcold temperatures) and remained lowered for 6.75 h, after whichthere was no significant difference from the running meantemperature chosen during the pre-exercise period. By the timethis period was over, most frogs had begun to prefer the warmerend of the gradient, as they had when they were unstressed.

DiscussionCold temperatures have profound effects on many aspects of

the physiology and behaviour of overwintering frogs. Intenseactivity and recovery from exhaustive exercise caused differentresponses to temperature, as witnessed in submerged frogsswimming at 1.5 and 7 °C. Frogs at 1.5 °C had lower fatigueresistance (e.g. lower activity capacity) than frogs swimming at7 °C (Table 1), even though they exhibited the same degree ofexhaustion in terms of tissue acid–base status and lactate build-up (Table 3). Although the extent of the muscle lactacidosis wassimilar in both groups of animals, this was not reflected in theblood, because frogs exercised at higher temperaturesdeveloped a greater combined respiratory and metabolicacidosis (Table 2; Fig. 2). A reduced factorial scope at 7 °Ccould well have contributed to the greater extracellular acidosis,


R 2 80

2

4

6

8

02

6

04812

A

B0

4

8

12

16

*‡

Pre- Post-exercise

**

*0 2 8 24Time post-exercise (h)

Frac

tion

ofco

ntro

l val

ue

4

0 2 8 24Time post-exercise (h)

Frac

tion

ofco

ntro

l val

ue

*

*

0Time (h)

24

[Lac

tate

] (m

mol

l-1)

Fig. 4. Lactate concentrations in the plasma (A) and gastrocnemiusmuscle (B) in resting (R) and post-exercise recovering frogs at 1.5 °C(filled symbols) and 7 °C (open symbols). The inset bar graphsrepresent the mean post-exercise lactate concentrations as a fractionof the mean resting (control) values, with the value 1 equal to theresting value. Values are means ±S.E.M. (N=5). *Significantlydifferent from the control value at the respective temperature(P<0.05); ‡significant effect of temperature at this time point(P<0.05).

Table 4. Metabolic rate measurements and derived variables of exercised submerged frogs at 1.5 and 7 °C

Restine Maximal Aerobic Factorial 67%Temperature M

.O∑ M

.O∑ scope aerobic EOC EPOC Recovery

(°C) (µmolΟ2g−1h−1) (µmolΟ2g−1h−1) (µmolΟ2g−1h−1) scope (µmolΟ2g−1) (µmolΟ2g−1) time (h)

1.5 0.31±0.02 1.65±0.07 1.34±0.06 5.50±0.26 0.311±0.024 3.16±0.23 2.10±0.227.0 0.65±0.03 2.19±0.06 1.55±0.05 3.47±0.16 0.473±0.027 4.23±0.37 2.88±0.37

Significant P<0.001 P<0.001 P=0.02 P<0.001 P<0.001 P=0.02 P=0.07temperature effect

Q10 3.82 1.67 1.29 0.43 1.91 1.70 1.78

EOC, estimated excess oxygen consumed during exercise; EPOC, excess post-exercise oxygen consumed during recovery.Values are means ±S.E.M., N=16 at 1.5 °C, N=13 at 7 °C.


since rates of CO2 excretion and O2 uptake would have reachedthe maximum capacity of the skin, but not necessarily themaximum requirements for post-exercise metabolism (e.g.acid–base restoration, lactate metabolism, glycogenresynthesis). Preferred temperatures following exercise weremuch lower than resting values for nearly 7 h (Fig. 7),suggesting that submerged frogs take advantage oftemperatures that promote more rapid recovery from intenseactivity. Thus, although unstressed frogs normally preferwarmer temperatures (Fig. 7; Tattersall and Boutilier, 1997), atwhich they are able to swim faster and more efficiently(Table 1), the increased potential for recovery at very lowtemperatures is borne out by the long period for which frogsremain hypothermic following activity.

Temperature effects on correlates of exercise exhaustion

Frogs showed consistently lower performance correlates ofswimming at 1.5 °C than at 7 °C, demonstrating the importancethat narrow temperature ranges can have on the activitycapacity of overwintering frogs. The total number of legcontractions, swimming distance, swimming speed and legcontraction frequency were all 1.4- to 1.7-fold lower at 1.5 thanat 7 °C (Table 1). Although exhaustion occurred over a shorterswimming distance at 1.5 than at 7 °C, the degree ofintracellular lactacidosis was the same (Table 3; Fig. 4). Thus,while exercise in the cold may be more difficult to sustain, theextent to which anaerobic processes are recruited appears to besimilar. Rome (1990) has suggested that a lower temperatureresults in a ‘compression of recruitment order’ of muscle fibressuch that glycolytic, fast-twitch fibres are mobilised at lower

1

2

3

4

5

6 *

*

*

*

*

Fact

oria

l met

abol

ic s

cope

as

a pr

opor

tion

of R

MR

0 1 2 3 4 5 6 7 8 9 10Time (h)

Fig. 6. Plot of factorial metabolic scope as a proportion of the restingmetabolic rate (RMR) during recovery from exhaustive exercise at1.5 °C (filled circles) and 7 °C (open circles). Values are means ±S.E.M. (N=16 at 1.5 °C, N=13 at 7 °C). Starting values immediatelyafter exercise are 5.50 at 1.5 °C and 3.47 at 7 °C. Colder recoveringfrogs appear to retain a higher factorial scope for at least 6 h,although this is only significantly higher for the first hour post-exercise (*P<0.05).

R 0 1 2 3 40

1

2

3

42.2

2.0

1.8

1.6

Pre- Post-exercise

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0

Met

abol

ic r

ate

(µm

ol O

2 g-1

h-1

)

5 6 7 8 9 10

Q10

for

met

abol

ic r

ate

Time (h)

Fig. 5. Metabolic rates in resting and recovering frogsexercised to exhaustion at 1.5 °C (filled circles) and 7 °C(open circles). Solid lines denote the double-exponentialtrend of metabolic rate recovery in both groups.Metabolic rates were not significantly different fromresting (R) values after 5–6 h of recovery from exercise.Dotted vertical lines show the 67 % recovery of theexcess metabolic rate occurring after 2.10 h at 1.5 °C andat 2.88 h at 7 °C. In addition, Q10 (filled squares)calculated from the mean metabolic rates at eachtemperature is shown and illustrates the temporalhysteresis and the gradual return of the resting thermalsensitivity of metabolic rate (Q10=3.82). See Table 4 forother variables calculated from aerobic metabolic rates.Values are means ±S.E.M. (N=16 at 1.5 °C; N=13 at 7 °C).

618

speeds to compensate for the reduced power output of musclesin the cold. When this is coupled with the lower thermalsensitivity of glycolytic ATP provision compared withoxidative ATP production in muscle fibres (i.e. Q10=1.3 forlactate accumulation versus5.7 for aerobic scope in frogs inthe temperature range 5–10 °C; Gatten et al., 1992), frogsswimming in the cold are unable to swim as far or as intenselyas those swimming at higher temperatures. The fact that thecolder frogs accumulate the same amount of muscle lactate andprotons over shorter swimming distances than warmer frogsmeans that, at least in terms of tissue acid–base status andlactate accumulation, frogs at both temperatures appear to havereached the same exhaustion end-point.

The maximal lactate concentrations attained in the plasmaand gastrocnemius muscle of exercised frogs were notinfluenced by temperature (Tables 2, 3). However, given whatis known about exercise physiology in other animals (Gleeson,1991), it was surprising that the lactate concentrations in theplasma were 2–3 times higher than in the muscle, when theopposite might have been expected (Fournier and Guderley,1992). It is possible that this reversed gradient is a result ofactive mechanisms shunting lactate preferentially into thecirculation as an extracellular fuel for core tissues duringperiods of oxygen limitation, as proposed by Donohoe andBoutilier (1998). However, it is also likely that this distributionis the passive result of the complex ion and pH gradients andthe higher permeability of undissociated lactic acid betweenthe intracellular and extracellular environments, as suggestedby Pörtner (1993).

The corresponding post-exercise lactate values from another

ranid species show plasma and muscle levels of 19.0 and38.9 mmol l−1 respectively (Fournier and Guderley, 1992), asopposed to the much lower values in the present study (cf.Tables 2, 3). The reasons for this may be due to the moreintense activity at the higher temperature or that exhaustion atlow temperatures is more closely related to the disturbancesproduced by protons or phosphate (Wilkie, 1986). However,the actual causes of muscular fatigue are still a matter ofconsiderable debate (Fitts, 1994). The glycogen concentrationsobserved in the cold-submerged frogs in the present study,although variable, did not appear to become depleted to theextent that they were limiting to energy supply (see Materialsand methods). Indeed, muscle ATP concentrations wereunaffected by exercise or temperature, and although PCrconcentration fell following exercise at 7 °C (Table 3),exhaustion in submerged frogs was not necessarily associatedwith a detectable uncoupling of the energy supply to musclecontraction. In fact, it is possible that neuromusculartransmission is so reduced (Takeuchi, 1958; Grainger andGoldspink, 1964; Jensen, 1972) or that muscle relaxation is soprolonged (Marsh, 1990) at and below 7 °C that limitations onthe ability to induce muscle contraction occur (Rome et al.,1992), which lead to exhaustion without a huge lactate load.

In terms of acid–base balance in the blood, exercise at 7 °Cled to a greater combined respiratory and metabolic acidosisthan at 1.5 °C (Table 2; Fig. 2). Thus, exercising at warmertemperatures in a skin-breather leads to greater CO2 retentionand presumably to a larger depletion of blood oxygen stores.The larger increase in blood haematocrit at 7 °C than at 1.5 °Cfurther supports this idea, because increases in haematocrit are


0 3 4 8 9 10 11 E 1 2 3 4 5 6 8 107 120

1

2

3

4

5

6

7

86.75 h

C

0 4 6 80

102030405060

Sele

cted

tem

pera

ture

(°C

)

1 2 5 6 7 9 11

Time (h)

2T (°C)

0 4 6 82T (°C)

0 4 6 82T (°C)

B

A

% O

bser

vatio

ns

Fig. 7. Plot of temperaturepreference in cold-submergedfrogs before and after exercise(E). Prior to exhaustiveexercise (A), frogs preferredthe warm end of the thermalgradient (overall mean 6.7 °C).Immediately after exercise (B),the preferred temperature fellto 3.6 °C and remainedsignificantly different fromresting temperature preferencefor 6.75 h. After this point(C), frogs returned to theirpreviously preferred temperature.Inset graphs are histograms ofpercentage observations atgiven temperatures (T) duringthese three discrete periods (A,B and C). Warm (dotted line)and cold recovery (solid line)refer to two groups of frogs:one group was introduced tothe warm end and the other tothe cold end of the tankimmediately following exercise (this effect was found to be insignificant). Values are means ±S.E.M. (N=17).


often associated with both hypercapnia (Nikinmaa, 1990) andhypoxaemia in amphibians (Malvin et al., 1992; Pinder andSmits, 1993). Why should exercising at warmer temperaturesresult in greater extracellular acidosis, yet not a greater lactateaccumulation? It is likely that the limit for gas exchange isreached in the warmer animals, which have a higher maximalM.O∑, but a smaller factorial scope (Table 4). A more intense

combined respiratory and metabolic acidosis at the highertemperature suggests that gas exchange during andimmediately following exercise was limiting or impaired,leading to the higher PaCO∑ values at 7 °C than at 1.5 °C.

Effects of temperature on exercise recovery

Although exhaustive exercise resulted in similar levels oflactate accumulation at both temperatures, the removal of thisextracellular lactate followed a different time course at the twotemperatures studied. Plasma lactate concentrations were stillsignificantly higher than resting values 2 h after exercise, andthis effect was more noticeable at 7 °C than at 1.5 °C. Ananalogous situation is seen by comparing the more prolongedrecovery of plasma PCO∑ and pHe at 7 °C than at 1.5 °C (Fig. 2;Table 2). The protracted recovery of lactate levels and plasmaacid–base balance at 7 °C probably results from the maximumcapacity for cutaneous gas exchange being reached at thehigher temperature. If, in fact, recovery at 7 °C is equated withsome degree of blood or tissue hypoxia, this may help toexplain the higher plasma lactate concentrations after 2 h ofrecovery at 7 °C, during which time a systemic hypoxia mayhave exacerbated the metabolic removal of circulating lactate.For example, the skin-breathing salamander Cryptobranchusalleganiensis took longer to recover plasma-borne lactatelevels after exercise than did an equally sized air-breathingtoad, Bufo marinus(Boutilier et al., 1980), suggesting thatdiffusion-limited cutaneous gas exchange can result in a slowerlactate recovery. In fact, the respiratory regulation of acid–basebalance by lung ventilation is expected to impart a more rapidacid–base recovery following exercise, and this is borne out bythe faster readjustment of arterial PCO∑ (PaCO∑) in air-breathersthan in skin-breathers (Boutilier et al., 1980). The pattern ofacid–base recovery observed in the gastrocnemius muscle,however, was similar at both temperatures, taking only 2 h toreturn to resting levels (Table 3; Fig. 3). It appears thattemperature has little influence on the recovery of correlates ofexercise in muscle, whereas large temperature effects occur inthe blood, emphasising the importance of gas exchange toblood acid–base regulation.

At the higher temperature examined in this study post-exercise, submerged frogs have probably maximised theircutaneous gas exchange (Fig. 6; Table 4), resulting in a muchlower factorial scope at 7 °C than at 1.5 °C. At bothtemperatures, considerable cutaneous capillary recruitmentmust have occurred, but it is likely that, at colder temperatures,resting metabolic rates are low enough to permit large relativeincreases in oxygen uptake across the skin (up to 5.5 timesresting M

.O∑), which appear to be sustained throughout the

entire recovery period (Fig. 6). Feder (1988) showed that air-

breathing salamanders could increase their metabolic rateduring exercise by two- to fourfold; however, when preventedfrom breathing air, this skin-breathing animal could onlymanage small changes above resting metabolic rate. This wasfurther corroborated by work on naturally lunglesssalamanders; these could only sustain metabolic rates of 1.6–3times above resting rates, while similarly sized lungedsalamanders sustained rates between 3.5 and 7 times higherthan resting rates (Full et al., 1988).

The recovery of metabolic rate following exercise displayedan exponential decrease at both temperatures (Fig. 5).However, the 67 % recovery point of the excess post-exerciseoxygen consumption (EPOC) tended to occur slightly faster at1.5 °C than at 7 °C (2.10 h versus2.88 h; Table 4). This periodis still considerably shorter than that estimated for plasmalactate removal (67 % clearance after 4.75 h at 1.5 °C and after6.38 h at 7 °C; see Results) supporting the assertion of Gaesserand Brooks (1984) that the traditional assigning of oxygen debtdirectly to lactate removal is overly simplistic. Thus, attemptsto correlate post-exercise oxygen consumption with lactateremoval will not necessarily yield a causal relationshipbetween lactate load and total oxygen consumed, especiallyfollowing intense exercise where similar exhaustive end-pointswere reached.

Thermoregulation before and after exhaustive exercise

Unstressed, resting frogs showed a distinct bias for the warmside of a thermal gradient, with a mean preferred temperatureover the 12 h period prior to exercise of 6.7 °C (Fig. 7) identicalto that previously measured by Tattersall and Boutilier (1997).Following exhaustive exercise, temperature preferencechanged for 6.75 h, falling to 3.6 °C (Fig. 7), after which theanimals moved back to their pre-exercise preferredtemperatures. This transitory behavioural hypothermia isreminiscent of that seen in hypoxic, submerged frogs(Tattersall and Boutilier, 1997). Similar benefits of loweringbody temperature would exist for both externally appliedhypoxia and the internally derived hypoxia resulting fromexercise. These benefits include: (1) a reduced metabolic ratethrough Q10 effects promoting aerobic metabolism, (2) anincreased haemoglobin affinity through temperature effectspromoting the loading of oxygen at the skin and themaintenance of a high arterial oxygen saturation, and (3)decreased energetic costs of systemic oxygen delivery at lowertemperatures. All three benefits would help to reduce overallenergy expenditure during periods of stress (Wood and Malvin,1991). In terms of exercise, a fourth benefit of behaviouralhypothermia would be that frogs would maintain a greateroxygen store relative to their total M

.O∑, which could be

immediately devoted to recovery metabolism.Presumably, this behaviourally induced hypothermia

(Fig. 7) is part of a coordinated response to reducing metabolicdemand during periods of internal hypoxic stress. Hypercapniaand hypoxia are both known to induce behaviouralhypothermia in amphibians (Riedel and Wood, 1988; Woodand Malvin, 1991). Both of these conditions are probably

620

present in the exercised frog, and the large increases inhaematocrit at 7 °C following exercise are suggestive of agreater degree of hypoxaemia at the warmer temperature(Malvin et al., 1992; Pinder and Smits, 1993). Thus, a loweringof body temperature would be expected to alleviate this stressin much the same way as occurs during externally appliedhypoxia (Wood, 1995).

In terms of blood lactate, a similar amount is removedinitially more rapidly at 1.5 than at 7 °C (Fig. 4). Although1.5 °C does not correspond to the actual preferred temperaturefollowing exercise, similar conclusions can be drawn regardingthe directional effects of temperature on recovery rates. Alowered body temperature would further reduce the P50 forblood oxygen affinity (G. T. Tattersall and R. G. Boutilier, inpreparation) and ensure better loading of oxygen at the skin.Oxygen release to the tissues following exercise wouldprobably not be a problem given the high levels of PCO∑ andlower pH expected near the tissues (i.e. the Bohr effect), so anincreased affinity should act in assisting oxygen delivery.Another potential benefit of lowering body temperature is theincreased factorial scope, which is evident in diffusion-limitedskin-breathers (Ultsch, 1974; Beckenbach, 1975). As a higherproportion of resting metabolic rate is maintained for at least1 h following exercise (Fig. 6), recovering at a lowertemperature means that frogs may initially devote a greaterproportion of their aerobic metabolic rate to the processes ofrecovery, without sacrificing overall oxygen homeostasis.Since lowering the temperature increases the available scopefor cutaneous conductance of respiratory gases (G. T. Tattersalland R. G. Boutilier, in preparation), selecting a coldenvironment may preclude the need for arterio-venousdifferences to increase dramatically following exercise, andtherefore circulating blood would not be brought to extremelylow PaO∑ levels and a severe and prolonged respiratoryacidosis could be avoided. Once the metabolic rate has fallenfar enough, the frogs can return to their previously preferredhigher temperatures (i.e. after 7 h) without sacrificing theirability to exchange respiratory gases and to repay the large‘oxygen debt’.

A similar transient behavioural hypothermia has beenobserved following activity in the lizard Dipsosaurus dorsalis(Wagner and Gleeson, 1997). Lizards exhaustively exercisedat 40 °C and allowed to recover in a thermal gradient loweredtheir body temperature by 3–4 °C over the initial 90 min, butthereafter returned to the previously preferred warmer bodytemperature during an additional 30 min of recovery. However,in lizards, this behavioural hypothermia response did not leadto more rapid removal of lactate from the blood than if thepreferred temperature had remained constant. Although theauthors concluded that the energetic benefits of recovering ata cold temperature were not adequate to account for majorshifts in thermoregulation, a more convincing case could bemade for cold-submerged frogs, as shown by the relativelyfaster plasma lactate recovery in the cold (Fig. 4) and simplyby the very prolonged behavioural hypothermia after exercise(Fig. 7).

Conclusions and perspectivesUnder normal circumstances, submerged frogs prefer to

remain as warm as aerobically possible (e.g. approximately7 °C in laboratory experiments). By avoiding the coldertemperatures, the animals would maintain a higher activitycapacity and presumably an increased ability to avoidpredation in their natural environment. However, whenstressed by exercise, temperature preference falls for aconsiderable period (6.75 h) before returning to normal. Byselecting a lower temperature, the demands for oxygen uptakeby the skin-breathing frog are greatly reduced, and they canimmediately dedicate a larger proportion of their totalmetabolic rate towards recovery from exercise. Indeed, theycould accelerate their acid–base and lactate recovery bymoving to the cold, a result that would not have been predictedfrom simple temperature effects on metabolism (Bennett andLicht, 1972; Gleeson, 1980; Wagner and Gleeson, 1996). Onemay question whether overwintering frogs are ever provokedinto exhaustive exercise in nature. However, given thatsubmerged frogs do not remain completely dormant during thewinter (Friet, 1993; Stinner et al., 1994) and are capable ofresponding to environmental cues, activity may be quiteintense during pursuit by a predator. In fact, it is remarkablethat frogs at 1.5 °C swam as much as they did, considering thatprevious work on amphibians suggests that temperatures below4 °C usually result in an overall reduction in activity (Bellis,1962; Lotshaw, 1977). If this is the case, it means that frogschose to remain inactive in the cold despite havingconsiderable scope for activity. Such potential scope foractivity may be beneficial in eluding predators (e.g. fish ormammals), especially if they are unable to remaininconspicuous. However, given their greater capacity foractivity, it would not take long for a fish or mammalianpredator to capture a swimming frog, since the frog wouldeither become exhausted sooner or simply be out-distanced.

One consequence of the results of this study is thatoverwintering frogs choosing to live in the extreme cold (e.g.0–1 °C immediately below the ice) would rely more heavilythan would frogs at higher temperatures on their limited fuelstores to support similar levels of intense activity because theywould become exhausted over shorter distances (Table 1).Thus, strategies of habitat selection in unstressedoverwintering frogs should ideally be geared towardsmaintaining a reasonable level of activity and alertness byinhabiting the relatively warm environment at the lake bottom.However, the selection of cooler environments is a sensiblestrategy following the stress of intense activity since it reducesresting aerobic metabolic rate, widens the factorial scope andfacilitates metabolic and acid–base recovery. The long periodrequired to return to normal thermoregulatory patterns istestimony to the importance of lowering body temperature inthe alleviation of stress.

Indeed, all animals experience numerous forms ofenvironmental and physiological stress, including hypoxia,hypercapnia, hypoglycaemia, dehydration, anaemia and toxicinsults. Recent work suggests that there are some general



thermoregulatory principles concerning how both ectothermsand endotherms cope with these stresses. Behaviouralhypothermia and the physiological lowering of bodytemperature are recognised as controlled responses to reducingthe effects of these stresses (Wood, 1995), and it is noteworthythat exercise can also elicit similar reductions in bodytemperature. The widespread occurrence of this responseacross many animal taxa suggests an integrated neurologicalor hormonal control of thermoregulation at new homeostaticset-points and implies a commonality of stress-sensingmechanisms. As yet, however, the biochemical triggers thatelicit hypothermia are unknown, although they are currentlythe focus of much investigation.

The authors would like to thank Tim West for helpfuldiscussions during the experiments and writing. This studywas part of a PhD dissertation at the University of Cambridgeand was supported by a 1967 Centennial NSERC scholarshipto G.J.T., and a BBSRC operating grant to R.G.B.

ReferencesBeckenbach, A. T. (1975). Influence of body size and temperature

on the critical oxygen tension of some plethodontid salamanders.Physiol. Zool. 48, 338–347.

Bellis, E. D. (1962). The influence of humidity on wood frog activity.Am. Midl. Nat. 68, 139–148.

Bennett, A. F. and Licht, P. (1972). Anaerobic metabolism duringactivity in lizards. J. Comp. Physiol. 81, 277–288.

Bennett, A. F. and Licht, P. (1973). Relative contributions ofanaerobic and aerobic energy production during activity inamphibia. J. Comp. Physiol. 87, 351–360.

Boutilier, R. G., McDonald, D. G. and Toews, D. P. (1980). Theeffects of enforced activity on ventilation, circulation and bloodacid–base balance in the aquatic gill-less urodele, Cryptobranchusalleganiensis: a comparison with the semi-terrestrial anuran, Bufomarinus. J. Exp. Biol. 84, 289–302.

Bradford, D. F. (1983). Winterkill, oxygen relations and energymetabolism of a submerged dormant amphibian, Rana muscosa.Ecology 64, 1171–1183.

Brattstrom, B. H. (1963). A preliminary review of the thermalrequirements of amphibians. Ecology 44, 238–255.

Brattstrom, B. H. (1979). Amphibian temperature regulation studiesin the field and laboratory. Am. Zool. 19, 345–356.

Crawshaw, L. I. (1979). Responses to rapid temperature change invertebrate ectotherms. Am. Zool. 19, 225–237.

Donohoe, P. H. (1997). Factors effecting metabolic rate reductionduring hibernation in the frog, Rana temporaria. PhD thesis,Department of Zoology, University of Cambridge, Cambridge.

Donohoe, P. H. and Boutilier, R. G. (1998). The protective effectsof metabolic rate depression in hypoxic cold submerged frogs.Resp. Physiol. 111, 325–336.

Emery, A. R., Berst, A. H. and Kodaira, K. (1972). Under-iceobservations of wintering sites of leopard frogs. Copeia 1972,123–126.

Feder, M. E. (1988). Exercising with and without lungs. II.Experimental elimination of pulmonary and buccopharyngeal gasexchange in individual salamanders (Ambystoma tigrinum). J. Exp.Biol. 138, 487–497.

Fitts, R. H. (1994). Cellular mechanisms of muscle fatigue. Physiol.Rev. 74, 49–94.

Fournier, P. A. and Guderley, H. (1992). Metabolic fate of lactateafter vigorous activity in the leopard frog, Rana pipiens. Am. J.Physiol. 262, R245–R254.

Friet, S. C. (1993). Aquatic overwintering of the bullfrog (Ranacatesbeiana) during natural hypoxia in an ice-covered pond. MScthesis, Department of Biology, Dalhousie University, Halifax.

Full, R. J., Anderson, B. D., Finnerty, C. M. and Feder, M. E.(1988). Exercising with and without lungs. I. The effects ofmetabolic cost, maximal oxygen transport and body size onterrestrial locomotion in salamander species. J. Exp. Biol. 138,471–485.

Gaesser, G. A. and Brooks, G. A. (1984). Metabolic bases of excesspost-exercise oxygen consumption: a review. Med. Sci. Sport Exer.16, 29–43.

Gatten, R. E. (1987). Activity metabolism of anuran amphibians:tolerance to dehydration. Physiol. Zool. 60, 576–585.

Gatten, R. E., Miller, K. and Full, R. J. (1992). Energetics at restand during locomotion. In Environmental Physiology of theAmphibians (ed. M. E. Feder and W. W. Burggren), pp. 314–377.Chicago: The University of Chicago Press.

Gleeson, T. T. (1980). Metabolic recovery from exhaustive activityby a large lizard. J. Appl. Physiol. 48, 689–694.

Gleeson, T. T. (1991). Patterns of metabolic recovery from exercisein amphibians and reptiles. J. Exp. Biol. 160, 187–207.

Gleeson, T. T. (1996). Post-exercise lactate metabolism: acomparative review of sites, pathways and regulation. Annu. Rev.Physiol. 58, 565–581.

Grainger, J. N. R. and Goldspink, G. (1964). The effect ofadaptation temperature on the properties of nerve–musclepreparations and on the performance of the frog Rana temporaria.Helgolander Wiss. Meeresunter. 9, 420–427.

Heisler, N. (1989). Parameters and methods in acid–base physiology.In Techniques in Comparative Respiratory Physiology: AnExperimental Approach, vol. 37 (ed. C. R. Bridges and P. J. Butler),pp. 305–332. Cambridge: Cambridge University Press.

Holzapfel, R. A. (1937). The cyclic character of hibernation in frogs.Quart. Rev. Biol. 12, 65–84.

Huey, R. B. (1982). Temperature, physiology and the ecology ofreptiles. In Biology of the Reptilia, vol. 12 (ed. C. Gans and F. H.Pough), pp. 25–91. New York: Academic Press.

Hutchison, V. H. and Dupré, R. K. (1992). Thermoregulation. InEnvironmental Physiology of the Amphibians (ed. M. E. Feder andW. W. Burggren), pp. 206–249. Chicago: The University ofChicago Press.

Jensen, D. W. (1972). The effect of temperature on transmission atthe neuro-muscular junction of the sartorius muscle of Ranapipiens. Comp. Biochem. Physiol. 41A, 685–695.

Lotshaw, D. P. (1977). Temperature adaptation and effects of thermalacclimation in Rana sylvatica and Rana catesbeiana. Comp.Biochem. Physiol. 56A, 287–294.

Malvin, G. M., Macias, S., Sanchez, M. and Parkrapid, A. (1992).Hypoxia-induced hemoconcentration in the toad Bufo woodhousei:role of spleen and lymph heart. FASEB J. 6, 1529A.

Marsh, R. L. (1990). Deactivation rate and shortening velocity asdeterminants of contractile frequency. Am. J. Physiol. 259,R223–R230.

Nikinmaa, M. (1990). Vertebrate Red Blood Cells: Adaptations ofFunction to Respiratory Requirements. Berlin: Springer-Verlag.262pp.

622

Passoneau, J. V. and Lowry, O. H.(1993). Enzymatic Analysis: APractical Guide. Totawa, NJ: Humana Press. 403pp.

Pinder, A. W. (1987). Cutaneous diffusing capacity increases duringhypoxia in cold submerged bullfrogs (Rana catesbeiana). Respir.Physiol.70, 85–95.

Pinder, A. W. and Smits, A. W. (1993). Mechanisms of acutehemoconcentration in bullfrogs in response to hypoxemia. Am. J.Physiol.264, R687–R695.

Pinder, A. W., Storey, K. B. and Ultsch, G. R.(1992). Estivationand hibernation. In Environmental Physiology of the Amphibians(ed. M. E. Feder and W. W. Burggren), pp. 250–274. Chicago: TheUniversity of Chicago Press.

Pörtner, H.-O. (1993). Multicompartmental analyses of acid–base andmetabolic homeostasis during anaerobiosis: invertebrate and lowervertebrate examples. In Surviving Hypoxia: Mechanisms of Controland Adapation(ed. P. W. Hochachka, P. L. Lutz, T. Sick, M.Rosenthal and G. van den Thillart), p. 570. Boca Raton: CRC Press.

Pörtner, H.-O., Boutilier, R. G., Tang, Y. and Toews, D. P. (1990).Determination of intracellular pH and PCO∑ after metabolicinhibition by fluoride and nitrilotriacetic acid. Respir. Physiol.81,255–274.

Pough, F. H. (1976). The effect of temperature on the oxygencapacity of reptile blood. Physiol. Zool.49, 141–151.

Prosser, C. L.(1991). Temperature. In Environmental and MetabolicAnimal Physiology(ed. C. L. Prosser), pp. 109–165. New York:Wiley-Liss.

Riedel, C. and Wood, S. C.(1988). Effects of hypercapnia andhypoxia on temperature selection of the toad, Bufo marinus. Fedn.Proc. 2, 500A.

Rome, L. C.(1990). Influence of temperature on muscle recruitmentand muscle function in vivo. Am. J. Physiol. 259, R210–R222.

Rome, L. C., Stevens, E. D. and John-Alder, H. B.(1992). Theinfluence of temperature and thermal acclimation on physiologicalfunction. In Environmental Physiology of the Amphibians(ed. M.E. Feder and W. W. Burggren), pp. 183–205. Chicago: TheUniversity of Chicago Press.

Stinner, J., Zarlinga, N. and Orcutt, S. (1994). Overwinteringbehaviour of adult bullfrogs, Rana catesbeiana, in northeasternOhio. Ohio J. Sci.94, 8–13.

Takeuchi, N.(1958). The effect of temperature on the neuromuscularjunction of the frog. Jap. J. Physiol.8, 391–404.

Tattersall, G. J. and Boutilier, R. G.(1997). Balancing hypoxia andhypothermia in cold-submerged frogs. J. Exp. Biol. 200,1031–1038.

Ultsch, G. R.(1974). Gas exchange and metabolism in the Sirenidae(Amphibia: Caudata). I. Oxygen consumption of submergedsirenids as a function of body size and respiratory surface area.Comp. Biochem. Physiol.47A, 485–498.

Wagner, E. L. and Gleeson, T. T.(1996). Low temperature andexercise recovery in the desert iguana. Physiol. Zool. 69,168–190.

Wagner, E. L. and Gleeson, T. T. (1997). Postexercisethermoregulatory behaviour and recovery from exercise in desertiguanas. Physiol. Behav.61, 175–180.

Wilkie, D. R. (1986). Muscular fatigue: effects of hydrogen ions andinorganic phosphate. Fedn. Proc.45, 2921–2923.

Willis, Y. L., Moyle, D. L. and Baskett, T. S.(1956). Emergence,breeding, hibernation, movements and transformation of thebullfrog, Rana catesbeiana, in Missouri. Copeia1956, 30–41.

Withers, P. C. and Hillman, S. S.(1988). A steady-state model ofmaximal oxygen and carbon dioxide transport in anuranamphibians. J. Appl. Physiol.64, 860–868.

Wood, S. C. (1995). Interrelationships between hypoxia andthermoregulation in vertebrates. In Advances in Comparative andEnvironmental Physiology, vol. 22 (ed. N. Heisler), pp. 209–231.Berlin: Springer-Verlag.

Wood, S. C. and Gonzales, R.(1996). Hypothermia in hypoxicanimals: Mechanisms, mediators and functional significance.Comp. Biochem. Physiol.113B, 37–43.

Wood, S. C. and Malvin, G. M.(1991). Physiological significanceof behavioural hypothermia in hypoxic toads (Bufo marinus). J.Exp. Biol.159, 203–215.


behavioural hypothermia during exercise recovery in frogs

Documents